pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
36134936 | PMC9496194 | pmc | 1,014 | {
"abstract": "In this paper, a superhydrophobic biomimetic composite coating was fabricated on brass by electrochemical etching, brushing PDMS adhesive layer, and depositing carbon soot particles. Due to the microstructure and the optimized ratio of PDMS, the contact angle of the superhydrophobic coating is up to 164° and the sliding angle is only 5°. The results of optical microscopy and morphometric laser confocal microscopy show that the prepared coating surface has a rough hierarchical structure. A high-speed digital camera recorded the droplet bouncing process on the surface of the superhydrophobic coating. The self-cleaning property of the coatings was evaluated by applying chalk dust particles as simulated solid contaminants and different kinds of liquids (including grape juice, beer, cola, and blue ink) as liquid contaminants. The coating remained superhydrophobic after physical and chemical damage tests. This work presents a strategy for fabricating superhydrophobic biomimetic composite coatings with significant self-cleaning properties, durability, and shows great potential for practical engineering applications.",
"conclusion": "4. Conclusions In summary, superhydrophobic coatings with good self-cleaning and durability were prepared by coating the etched brass substrate with optimized proportions of PDMS and then depositing soot particles. The superhydrophobic coating shows a high contact angle of 164° and the low slip angle of 5°. The synergistic effect of microstructure and soot particle deposition makes the superhydrophobic coating self-cleaning, which has a noticeable effect on artificial solid dust pollutants and common liquid pollutants. In addition, the superhydrophobic coating has maintained durability in physical and chemical tests such as sandpaper wear, water flow erosion, and acid and alkaline solutions. Therefore, this can provide a direct way to develop self-cleaning and durability superhydrophobic coatings on engineering metal substrates.",
"introduction": "1. Introduction Brass is an important engineering material and is widely used in the chemical and marine industries due to its excellent thermal stability, good electrical conductivity and, outstanding corrosion resistance [ 1 , 2 ] The preparation of superhydrophobic coatings based on the brass substrate is one of the key research directions in the field of surface engineering [ 3 , 4 , 5 ]. A superhydrophobic surface is a surface with a contact angle greater than 150° and a sliding angle less than 10° [ 6 ]. The artificial superhydrophobic surface was first inspired by the hydrophobic and self-cleaning phenomena on the surface of the stems and leaves of some plants in nature, as well as on the feathers or skin surfaces of some birds and animals. For example, the lotus leaf always keeps its clean leaf surface because of its unique superhydrophobic property. According to a previous study [ 7 ], the upper epidermis of the lotus leaf has micrometer-scale rough papillae structures surrounded by hydrophobic wax-like substances. Benefiting from the special papillary structure and the hydrophobic wax-like substance, the two basic conditions for realizing superhydrophobic function, namely micro-nano structure, and low surface energy, are satisfied. (Literature support) When the water droplets roll off the upper surface of the lotus leaf, it can take away the pollutants on the leaf surface, so as to achieve the effect of self-cleaning, which is the famous “lotus leaf effect” [ 8 ]. At present, the applications of superhydrophobic coatings include digital microfluidics [ 9 ], oil-water separation, [ 10 ] fluid drag reduction [ 11 ], droplet manipulation [ 12 ], anti-fouling [ 13 ], anti-icing [ 14 ], and metal corrosion protection [ 15 ], etc. And it is evident that the research of superhydrophobic coatings with excellent performance has good prospects. To achieve a superhydrophobic surface, two principles need to be satisfied. One is to obtain a rough hierarchical structured surface, and the other is that the material has low surface energy. According to these two principles, researchers have two directions for preparing superhydrophobic surfaces: either to reduce the surface energy of the rough surface or take measures to make the material surface with low surface energy [ 16 ]. To achieve the required coating, researchers have developed various strategies in the preparation process. In terms of techniques, electrostatic spinning [ 17 ], chemical vapor deposition [ 18 ], electrochemical deposition [ 19 ], and laser ablation [ 20 ] are commonly used. In terms of materials, researchers chose to add used nanoparticles such as SiO 2 , ZnO, and TiO 2 to provide roughness or fluorides and non-polar molecules to provide low surface energy [ 21 ]. Although these strategies have made some progress, there are still many problems, such as the expensive equipment and materials involved. Faced with these problems, carbon soot particles, as a cheap hydrophobic material with high porosity and high stratification dimension, are easily accessible and low cost [ 22 ]. However, the direct use of carbon fume particles to prepare superhydrophobic surfaces without any treatment is unsatisfactory. Because there is only a weak physical force between carbon nanoparticles, they are very fragile. They can be easily peeled off from the substrate surface by external mechanical forces, water erosion, etc. The use of binders or crosslinkers can significantly improve the shortcomings of the superhydrophobic character based on carbon soot particles and enhance the adhesion and bonding of carbon soot particles on the surface of the substrate [ 23 ]. Seo et al. [ 24 ] used paraffin wax to fix fragile candle soot. Compared with soot coating without any treatment, the robustness and durability of the coating have been significantly improved. However, the author himself also pointed out that the mechanical stability of paraffin is poor, and it is easy to scratch. Paraffin is also decomposed due to thermal degradation when the temperature exceeds 250 °C. Xiao et al. [ 25 ] connected carbon particles through the hydrolysis of methyl trichlorosilane (MTCS) in the air. While the coating showed excellent resistance to corrosion in acidic and weakly alkaline liquids as well, it was less effective against strong alkaline solutions. Moreover, the chemical vapor deposition method involved in this study has a slow deposition rate and the gaseous by-products are usually very toxic. These drawbacks limit this study to large-scale commercialization. The polymers used in their study by Sutar et al. [ 26 ] require solvents such as toluene, acetone, and o-xylene to dissolve, and the rate of substrate immersion and withdrawal in the polymer solution has to be controlled by the dip coater. These commercial polymer materials and equipment devices undoubtedly indirectly increase the cost of actual industrial production. To overcome the above shortcomings, relatively simple electrochemical etching and direct deposition methods have been adopted by us. Here, superhydrophobic coatings were prepared on brass surfaces by electrochemical etching, brushing the PDMS layer, and depositing carbon soot particles. First of all, the surface of the brass substrate was electrochemically etched, which makes the substrate surface have an uneven microstructure. Subsequently, a 5:1 ratio of PDMS prepolymer and curing agent was used as a binder to coat the etched brass substrate. Then, hydrophobic soot particles were deposited directly in the middle of the candle flame to prepare the superhydrophobic coating. The surface micromorphology and wettability of the prepared superhydrophobic coatings were characterized and investigated, and the self-cleaning properties of the coatings were confirmed by solid and liquid contaminant tests. In addition, corresponding physical and chemical damage tests, including water jets, water droplet impacts, sandpaper abrasion, and acid and alkaline liquid immersion, were designed to further evaluate and verify the durability of the coatings, and the relevant properties of the coatings were enhanced thanks to the embedded and semi-embedded structures formed between the PDMS adhesive layer and the carbon soot particles, and the experiments showed that the prepared coatings can be applied for a long time in real environments.",
"discussion": "3. Results and Discussion 3.1. Surface Morphology and Wettability In this paper, the preparation of super-hydrophobic coating on brass surfaces mainly included electrochemical etching and direct deposition of soot particles. The surface morphology and wettability of the coatings were characterized using optical microscopy, scanning Electron microscope, morphometry microscopy, and contact angle meters. The test results of the bare substrate sample (BS), etched substrate sample (ES), and superhydrophobic coating sample (SC) are shown in Figure 2 . The first column of Figure 2 shows the morphological characteristics and surface contact angles of samples. As can be seen from Figure 2 (a1), the BS surface has only scratches left by sandpaper sanding, and the surface contact angle caused by scratches is 107°, which is shown in the inset of Figure 2 (a1). Similarly, these microstructures on the surface of ES can also be observed in Figure 2 (b1), where the gully-like microstructures created air pockets beneath the water droplets. These air pockets may cause the contact angle of interfacial water droplets to increase to 130°shown in the inset of Figure 2 (b1). BS and ES exhibit a degree of hydrophobicity, probably due to the adsorption of organic compounds from ambient air on the surface of BS and ES, which are mainly non-polar C-C/C-H groups. As pointed out in the literature [ 27 ], the absorption of organic matter from the surrounding atmosphere can cause a conversion of wettability. The synergistic effect of the non-polar groups and the rough surface structure make BS and ES hydrophobic. The contact angle of the SC is 164° shown in Figure 2 (c1) inset, which indicates that the SC surface is superhydrophobic. The second column of Figure 2 shows the surface microscopic images of the three specimens: the BS surface is flat, the ES surface has uneven pits due to electrochemical etching, and the SC surface has a distinct peak and valley shape after the deposition of carbon soot particles. The third column of Figure 3 shows the results of the surface roughness analysis of the specimens, respectively. It can be found that the highest values of arithmetic mean deviation of contour Ra and microscopic unevenness decimal height Rz of ES are 4.836 μm and 22.600 μm, respectively, due to electrochemical etching, indicating that it has the roughest surface. In contrast, the smallest value of the contour cell mean width RSm of SC is 4.792 μm, reflecting the higher density of deposited carbon soot particles on the surface of SC. In order to present further preparation details and results, and to analyze the reasons for the superhydrophobicity, adhesion, and stability of the coating, the pretreatment etching process is presented in more detail. Figure 3 shows the schematic diagram of BS sample electrochemical etching preparation and etching result. Figure 3 (a1,a2) display the process of producing ES sample by electrochemical etching. Figure 3 (a3) shows the results after etching, and it can be seen that there are obvious gully-like microstructures on the surface of ES. The non-smooth microstructure feature can not only trap some air, create the condition for the formation of superhydrophobic character, but also provide a place for the deposition of soot particles. After obtaining ES as shown in Figure 3 above, the PDMS-coated ES was further placed in the middle of the incomplete burning candle flame by direct deposition method to capture hydrophobic soot particles. Since the brushed PDMS mixture is a flat and thin layer, part of the PDMS mixture will penetrate into the etched substrate and will be instantly solidified when burned by the flame, and at the same time, the loose carbon soot particles can be tightly anchored to the surface of the substrate, and the carbon soot particles can completely cover the PDMS surface in the form of embedded or semi-embedded, so there would not be a violent PDMS burning phenomenon. The preparation schematic diagram, morphology, and hydrophobic phenomenon of superhydrophobic coating samples (SC) are shown in Figure 4 below. The mass ratio of the prepolymer and the curing agent in the PDMS mixture is 5:1. Increasing the ratio of the curing agent can improve the hardness and firmness of the PDMS after being solidified [ 28 ]. When depositing carbon soot particles, the PDMS mixture on the substrate surface was heat-solidified while locking the carbon soot particles firmly on the substrate surface. The carbon soot particles on the coating surface can be divided into three main categories according to the relationship between the deposited carbon soot particles and the PDMS adhesive layer. (i) loose carbon soot particles, which are located at the top of the coating and are not in contact with PDMS but are only connected to each other by weak van der Waals forces and can be easily worn off. (ii) Carbon soot particles partially embedded in PDMS, which improve the mechanical stability of the coating through the embedded effect with PDMS. (iii) Completely embedded carbon soot particles, which are integrally embedded in the PDMS adhesive layer as the PDMS is solidified by high temperature. In addition, the microstructure formed on the substrate surface under electrochemical etching can further provide a buffer and shelter for this semi-embedded and embedded structure of PDMS-carbon soot particles to improve the mechanical wear resistance of the coating [ 29 ]. Figure 4 (a1) shows a schematic of ES deposited carbon soot particles. The carbon soot particles are solid by-products of the incomplete combustion of hydrocarbons in the air. Due to the presence of C-H groups, including CH 2 and CH 3 groups, the carbon soot particles in the inner flame show obvious hydrophobicity [ 30 ]. The formation mechanisms of carbon soot particles include uniform nucleation of carbon soot particles, growth and oxidation of carbon soot particles, coalescence of carbon soot particles, and agglomeration of carbon soot particles [ 31 ]. By weak van der Waals forces, soot particles were bound together, forming an irregular network of particles [ 32 ]. In Figure 4 (a2), rough and distributed porous micromorphology were clearly observed on the coating surface. Due to the particular rough, porous structure, air can be trapped in it. So, the contact area between the coating and water droplets is reduced which contributes to the superhydrophobicity of the surface. The structure can form a physical barrier due to the possibility of creating air pockets to reduce the contact area between the substrate and the aqueous medium. Thus an evident silver mirror phenomenon can be observed ( Figure 4 (a3)) [ 33 ]. Superhydrophobic surfaces can be explained by Cassie-Baxter theoretical model: A large number of air capsules on the SC surface reduce the contact area of the liquid-solid interface, and the contact angle of water droplets on the gas-liquid-solid three-phase composite surface can be explained by Cassie-Baxter equation [ 34 ]: (1) cos θ γ = f 1 cos θ 1 + f 2 cos θ 2 \nwhere θ γ is the apparent contact angle; f 1 and f 2 are the fraction of unit apparent area occupied by solid surface and air, respectively ( f 1 + f 2 = 1 ) ; θ 1 and θ 2 are the intrinsic contact angles of water droplets on a solid surface and in the air, respectively. Here, the contact angle of air to water is θ 2 = 180°. Thus, Equation (1) can be further simplified to the following form: (2) cos θ γ = f 1 ( cos θ 1 + 1 ) − 1 According to the experimental measurement results (shown in Figure 2 (b1,c1)), the values of θ 1 and θ γ are 130° and 164°, respectively. Therefore, f 1 is calculated to be 0.1084. This means that 89.16% of the area is water droplets in contact with air. The ability of the SC surface to repel water droplets is powerful, which can be visually observed from the above static silver mirror phenomenon. Considering that most of the practical applications of superhydrophobic surfaces are under dynamic conditions, it is essential to verify the dynamic water repellence of the coating. Here, in order to further illustrate the low surface energy and wettability of the prepared superhydrophobic coating, an experiment of water droplets bouncing against the superhydrophobic coating was conducted to evaluate the water-repelling performance of the coating from a dynamic perspective. The process of water droplet bouncing was recorded with a high-speed camera, and the specific water droplet bouncing phenomenon is shown in Video S1 . As shown in Figure 5 (a1–a8), just like rain droplets bouncing on lotus leaves in nature, the impacting droplets first spread out on the surface to their maximum diameter, then they contract to a certain extent and eventually bounce off the surface. The water droplets underwent the motion process of spreading, shrinking, bouncing back, and rolling down on the prepared superhydrophobic coating. Due to the rough and distributed porous micromorphology on the SC surface, it can trap air near its surface and form air pockets, resulting in a stable Cassie-Baxter composite contact interface [ 35 ]. The water droplets cannot exclude the trapped air when they touch the SC surface. During the water droplet spreading phase ( Figure 5 (a3)), a sub-stable Cassie-Baxter contact state is formed because of the pressure inside the water droplet. In the subsequent shrinkage phase ( Figure 5 (a4)), a recovery from the sub-stable state to the stable Cassie-Baxter contact state is achieved. The air captured the external atmospheric pressure during the droplet rebound balances ( Figure 5 (a5)), thus reducing the viscous resistance of the droplet during its motion on the superhydrophobic surface. This allows the droplet to rebound and roll off from the surface ( Figure 5 (a6–a8)). According to the above analysis, the rebound phenomenon of the impacting droplet on the surface of the SC reflects the excellent dynamic water repellency of the prepared coating. 3.2. Self-Cleaning Property On the surface of the superhydrophobic coating, rolling water droplets can carry away pollutant particles from the surface, to achieve the self-cleaning effect. Here, chalk dust particles (the main components are calcium carbonate and calcium sulfate) were used to simulate solid contaminants to examine the self-cleaning effect of superhydrophobic coatings. The large-size chalk ash particles were first ground by the grinder and subsequently screened with a 200-mesh sieve to ensure that the chalk ash particles all remained around 75 μm in diameter. The results are shown in Figure 6 below. As shown in Figure 6 (a1,b1,c1), BS, ES, and SC were placed on the edge of the glass culture dish at an inclination angle of 10° (the critical value of the sliding angle of the superhydrophobic coating surface) and sprinkled with a layer of chalk dust particles. Then, water drops were dripped on the contaminated surfaces of each sample ( Figure 6 (a2,b2,c2), Video S2 ), respectively. As shown in Figure 6 (a3), the contaminant particles adhered to the BS surface. Similarly, there is still a clear collection of contaminants on the surface of ES after the water flow ( Figure 6 (b3)). In contrast, the droplets on the SC surface appear spherical and the contaminant particles were carried away by the rolling water droplets ( Figure 6 (c3)). The motion states of droplets on the surface of the samples can be divided into three types: sticking, sliding, and rolling [ 36 , 37 ]. Sticking and sliding droplets have a weaker ability to remove solid contaminant particles on the samples. As pointed out in the literature [ 38 ], the pore size of the superhydrophobic surface determines the lower size limit of the contaminant. The self-cleaning property of a superhydrophobic surface can be achieved as long as the particle size exceeds the pore size of the superhydrophobic surface, or the thickness of the contaminant is lower than the protrusion height of the superhydrophobic surface. For SC, since the size of the solid contaminant particles is larger than the size of the porous microstructure on the surface of the SC, the contaminant particles only contact the top of the porous microstructure, resulting in a small actual contact area. The contact form between chalk dust and the superhydrophobic surface is point contact with little surface adhesion, which means that the high capillary force caused by water droplets is higher than the adhesion force between chalk dust particles and the superhydrophobic surface, making the particles easily carried away by water droplets. In contrast, the contact area of the BS or ES surface with the contaminants is much larger, and the water droplets leave the substrate surface in a sliding or sticky manner and do not carry away a large number of contaminants from the surface, most of the contaminant particles are simply redistributed by the action of the water droplets. The above is the self-cleaning of the hydrophobic coating on solid contaminants, and the following is a test on the self-cleaning of liquid contaminants. Here, grape juice, beer, cola, blue ink, and milk were used as simulated liquid contaminants. The result is shown in Figure 7 . The self-cleaning property of the coating was examined by comparing the accumulation of liquid contaminants on the surface of the superhydrophobic coating before and after immersion in common household liquid solutions. As shown in Figure 7 , the superhydrophobic coating was clean before immersion in liquid contaminants. Subsequently, the superhydrophobic coating was removed after being immersed in different liquid contaminants and the surface of the coating was still clean. Video S3 contains more details. As described in the video, the results are similar to those shown in reference [ 39 ], and the coating also performs well against contaminated liquids. 3.3. Durability Superhydrophobic coatings are often exposed to physical and chemical damage when put into practical applications, thus affecting the lifetime of the coatings. Considering the actual application environment, physical damage tests such as simulated rainwater erosion and sandpaper abrasion, and chemical damage tests such as acid and alkali resistance were designed in this paper to evaluate the durability of the prepared coatings. The test process and results are shown in Figure 8 . In order to simulate the impact of natural rainwater on the coating, water jet, and water drop impact tests were set up. As shown in Figure 8 (a1), the syringe was used to draw 10 mL of water, and then the tip of the needle was aimed at the coating and tilted at an angle of 45 to the plane, and the water inside the syringe was pushed out. The water jet test video can be seen in Video S4 . It can be observed from the video that the water jet was repelled by the coating surface and bounced out in the opposite direction due to the water-repellent property of the coating. In general, superhydrophobic coated surfaces will bounce back immediately when impacted by a water jet. This is because the air cushion formed on the superhydrophobic surface blocks the water jet from entering the structure of the surface [ 40 ]. In Figure 8 (a2), due to the high instantaneous water pressure of the water jet, some loose carbon soot particles on the surface are washed away, which is the normal phenomenon, and the coating area position appears cratered and eroded by the water jet. The average contact angle measured in this position was 159° (inset of Figure 8 (a2)), which confirmed that the coating could resist the transient erosion by the water jet. The water droplet impact experiment is shown in Figure 8 (b1). Specific parameters were set as follows: 20 cm distance between the funnel lower end and platform, 45° platform inclination angle, 10 μL water droplet volume, and the speed of 50 drops/min. The sample was fixed on the inclined platform with double-sided tape. The morphology of the sample after the impact test is shown in Figure 8 (b2). After 12 h of continuous water drop impact, most of the carbon soot particles on the coating surface still resided on the substrate surface. Due to the impact of water droplets, some of the carbon soot particles on the coating surface were compacted and aggregated by physical forces which led to a slight decrease in contact angle of about 153° (inset of Figure 8 (b2)). The results showed that the coating has better durability compared to the bare carbon soot coating. In order to evaluate the wear resistance of the prepared coating, a sandpaper abrasion test was designed based on the ISO 8251-2018 standard ( Figure 8 (c1)). Usually, the sandpaper chosen is not less than 1000 mesh, and the larger the mesh of the sandpaper, the smaller the particle size on top of the sandpaper, then the easier it is to destroy the microscopic rough structure of the coating surface, which helps to improve the accuracy of the superhydrophobic coating wear resistance test. Here we choose 1500 grit sandpaper for the test. First, the coating surface was contacted with the upper surface of sandpaper, and then 100 g weight was placed on the base. External thrust was applied to the base to simulate the rubbing and damage effect of external mechanical force on the coating in the natural environment. In the wear test, the single wear stroke is set as 20 cm. Through experiments, it was found that after six wear cycles, some scratches appeared on the surface of the coating, and loose soot particles on the surface were worn away, but most soot particles still adhered to the substrate stably. The morphology of the worn coating is shown in Figure 8 (c2). The contact angle was 154°, which indicated the robust entanglement of the soot particles with the adhesive molecules. The good physical wear resistance of the prepared superhydrophobic coatings is attributed to the surface microstructure produced by electrochemical etching and the overall structure formed between the PDMS mixture and the carbon soot particles. On the one hand, the surface microstructure of the substrate produced by the electrochemical etching provides a buffer and a shelter for the coating material. On the other hand, PDMS acts as an effective binder, which combines with the carbon soot particles to form a strong and systematic monolith. Metal alloys, when exposed to acidic and alkaline environments, are subject to corrosion by acidic and alkaline substances, which can damage the structure and strength of the metal. The superhydrophobic coating can effectively prevent the contact between the water stream containing acid and alkali ions and the metal substrate, so that the substrate can be effectively protected from corrosion. Here, citric acid and sodium carbonate were used to prepare acidic solutions with pH values of 2, 4, and 6, and alkaline solutions with pH values of 8, 10, 12, and 14, respectively. In this paper, samples were submerged in solutions with pH values ranging from 2 to 14 ( Figure 8 (d1)) to test the resistance of the coating to strong acids and bases. As shown in Figure 8 (d2), the coating still maintained a contact angle greater than 150° compared with the coating before the test, and the super-hydrophobicity did not disappear. The above tests show that the prepared superhydrophobic coating can survive in this acidic and alkaline environment and possesses excellent chemical durability."
} | 6,962 |
36224760 | PMC9668097 | pmc | 1,015 | {
"abstract": "Global warming has accelerated in recent decades due to the continuous consumption of petroleum-based fuels. Cyanobacteria-derived biofuels are a promising carbon-neutral alternative to fossil fuels that may help achieve a cleaner environment. Here, we propose an effective strategy based on the large-scale cultivation of a newly isolated cyanobacterial strain to produce phycobiliprotein and biodiesel, thus demonstrating the potential commercial applicability of the isolated microalgal strain. A native cyanobacterium was isolated from Goryeong, Korea, and identified as Pseudanabaena mucicola GO0704 through 16s RNA analysis. The potential exploitation of P. mucicola GO0704 was explored by analyzing several parameters for mixotrophic culture, and optimal growth was achieved through the addition of sodium acetate (1 g/l) to the BG-11 medium. Next, the cultures were scaled up to a stirred-tank bioreactor in mixotrophic conditions to maximize the productivity of biomass and metabolites. The biomass, phycobiliprotein, and fatty acids concentrations in sodium acetate-treated cells were enhanced, and the highest biodiesel productivity (8.1 mg/l/d) was achieved at 96 h. Finally, the properties of the fuel derived from P. mucicola GO0704 were estimated with converted biodiesels according to the composition of fatty acids. Most of the characteristics of the final product, except for the cloud point, were compliant with international biodiesel standards [ASTM 6761 (US) and EN 14214 (Europe)].",
"introduction": "Introduction Industrialization, lifestyle modernization, and significant increases in the number of automobiles have greatly increased the demand for petroleum-based fuels. Currently, approximately 85% of the planet’s primary energy requirement is met by petroleum-based fuels [ 1 ]. However, steady increases in fossil fuel consumption have resulted in global warming because of the build-up of carbon dioxide in the air [ 2 ]. This continuous rise in global temperature may wipe out 39%–43% of the world’s biota [ 3 ]. In the short term, other associated problems may occur, including air quality deterioration, changes in disease patterns, and reduced food supplies [ 4 ]. Therefore, the partial or complete replacement of fossil fuels with renewable clean energy is urgently needed to ensure global stability and human affairs. Biofuels derived from photosynthetic microorganisms are an attractive alternative to petroleum fuels due to (1) the rapid growth rate, (2) space-efficient cultivation, (3) high lipid accumulation ability, and (4) high carbon fixation rate of microalgae [ 1 , 5 , 6 ]. Photosynthetic microorganisms are capable of capturing carbon dioxide and converting solar energy into chemical energy, thus providing an alternative for the production of sustainable and carbon-neutral energy sources [ 5 , 7 ]. Cyanobacteria, a gram-negative prokaryotic autotroph, are microbial organisms ubiquitously found in natural waters that play a pivotal role in biogeochemical cycles [ 8 ]. Furthermore, in addition to photosynthesis, cyanobacteria produce important commercial pigments (astaxanthin, lutein, phycobiliprotein), vitamins, and essential nutrients (notably proteins, carbohydrates, and lipids) [ 9 ]. However, many cyanobacteria also produce a variety of detrimental toxic substances known as cyanotoxins ( e.g. , nodularin and microcystin) that can severely affect human health [ 10 ]. Particularly, recent reports have indicated that β-N-methylamino-l-alanine (BMAA), a neurotoxic chemical found in cyanobacteria can have long-standing and serious health effects [ 11 ]. However, despite producing these toxins, some cyanobacteria might be suitable for biodiesel production since of their high lipid yields, thereby compensating for their toxicity. Biodiesels can thus be derived from cyanobacterial fatty acids followed by transesterification to obtain high-purity, biodegradable, and non-toxic fuels [ 12 ]. The utmost importance in the biodiesel process is the initial selection of cyanobacterial strains. The bioprospecting and isolation of natural endogenous microalgae enables the development of species-specific production of viable chemicals and biodiesels [ 13 ]. Particularly, native isolated strains have significant commercial value for regional large-scale production, owing to their robust growth under conditions to which they are naturally adapted [ 14 ]. The approach also prevents unexpected ecological risks from the potential introduction of invasive species in commercial cultivation fields. In this study, we isolated and identified a native cyanobacterium belonging to the genus Pseudanabaena in South Korea and improved the biomass productivity by optimizing the cultivation conditions. Additionally, a focus on estimating the characteristics of biodiesel obtained from cyanobacterium was conducted, as all biodiesel properties must be considered to meet standard requirements but are often overlooked in many studies for biodiesel production. Here, we report the properties and potential large-scale production of biodiesel derived from a newly isolated natural strain of Pseudanabaena from a quantitative and qualitative perspective.",
"discussion": "Results and Discussion Morphology and Phylogenetic Analysis A cyanobacterial strain was collected from the Gangjeong-Goryeong weir of the Nakdong River in Goryeong, Republic of Korea. Morphological characterization of the strain was conducted under a bright-field microscope ( Fig. 1A ). The native cyanobacterial cells isolated in this study were filamentous and arranged in lines of two to six 3–20 μm cells. The morphology of the cells was largely consistent with that of Pseudananbaena sp. To confirm our morphological analyses, molecular identification was performed and a phylogenetic tree was constructed based on the results of molecular phylogenetic analysis of 16s rRNA sequences ( Fig. 1B ). As expected, a total of 1,484 bp of the 16S rRNA gene sequences were aligned with 16 members of the Pseudanabaena genus. Two Korean populations of P. mucicola (GO0704 isolated in this study, and MN128994) were identical to each other. Moreover, these strains shared high similarity with a Chinese population (KM386852) of P. mucicola , with only a 0.04% dissimilarity. P. mucicola GO0704 was also closely related to other intrageneric species such as P. yagii (0.5-0.8%), P. galeata (0.9%), and P. cinerea (1%). Effect of Sodium Acetate on Native Pseudanabaena mucicola GO0704 Growth Due to the specific shape of Pseudananbaena sp., previous studies on Pseudananbaena sp. mainly used OD value as an indirect indicator of cell growth [ 17 , 18 ]. Here, we also confirmed the strong correlation between OD 650nm and cell number (coefficients R 2 = 0.9976) (Fig. S1). Cyanobacteria are capable of growing in three trophic (photo-/mixo-/hetero-) modes [ 5 ]. The mixotrophic mode uses both organic carbon and CO 2 in the energy cycle [ 19 ]. In this mode, two carbon metabolic ways act in conjunction. Therefore, organic carbon can be used as an energy source in addition to radiant energy [ 20 ]. Next, we identified the most effective carbon source for the growth of P. mucicola GO0704. Equal amounts (1 g/l) of different carbon sources were supplied as nutrients for P. mucicola GO0704: glucose, xylose, galactose, and sodium acetate ( Fig. 2A ). After 72 h of cultivation, sodium acetate was the only carbon source that stimulated the growth of P. mucicola GO0704. Previous studies have demonstrated that sodium acetate can enhance the production of cyanobacterial biomass and metabolites [ 21 ]. In contrast, the addition of galactose and xylose resulted in the death of the cyanobacterial cells, indicating that these carbohydrates adversely affect the cells. Specifically, a previous study reported that xylose induced the death of microalgal cell by blocking carbon cycle of photosynthetic system as a competitive inhibitor [ 22 ]. These results suggest that sodium acetate is an appropriate organic carbon source for mixotrophic cultivation of the cyanobacterium P. mucicola . We also investigated the optimal initial sodium acetate concentration (0, 1, 5, and 10 g/l) for maximizing biomass production. As shown in Fig. 2B , the 1 g/l sodium acetate concentration showed an obvious increase in the growth of P. mucicola GO0704, reaching approximately 45% in 72 h. In contrast, at concentrations above 1 g/l, as with the higher concentration of sodium acetate, P. mucicola growth became inhibited. The OD650 values of the 5 g/l and 10 g/l treatment of sodium acetate were significantly decreased to 0.039 and 0.023, respectively, at 72 h. This inhibition might be caused by damage to the oxygen-evolving complex at the donor side of photosystem II (PSII) by a high concentration of sodium acetate [ 23 ]. PSII is a major protein complex involved in the photosynthesis process and catalyzes the light-induced release of oxygen and the reduction of plastoquinone in water [ 24 ]. Therefore, inhibition of PSII evidently reduces photosynthesis, which is the basic survival strategy of cyanobacteria. Based on the above results, the ideal sodium acetate concentration was 1 g/l. Production of Value-Added Chemicals from Pseudanabaena mucicola GO0704 Next, a scale-up study in batch bioreactor was conducted to examine the effect of sodium acetate addition on metabolite production and its industrial applicability. Fig. 3A illustrates the effect of sodium acetate treatment on biomass production in a 5 L stirred-tank bioreactor. At 144 h, the dry weight of cells supplemented with sodium acetate increased 1.3-fold compared with the control (sodium acetate: 530 mg/l, control: 400 mg/l). Sodium acetate treatment in the photobioreactor promoted the rapid growth of cells, similar to the flask experiment results. The gradual scale-up of bioreactors has recently garnered increasing attention, as this constitutes essential research to assess the feasibility of industrial-scale commercialization. This is because cultivation in flasks and bioreactors can result in different outcomes due to differences in scale and environmental factors such as aeration rate, agitation rate, and shear stress [ 25 , 26 ]. Therefore, we believe that this research is important to evaluate the potential industrial applicability of the mixotrophic mode of P. mucicola . Next, we explored the production of diverse metabolites by P. mucicola GO0704 in a 5 L stirred-tank bioreactor. As illustrated in Fig. 3B , the productivity of value-added metabolites increased with biomass productivity. Pseudanabaena sp. is generally known to produce useful substances such as chlorophyll, phycobiliproteins, and lipids [ 17 , 18 ]. Chlorophyll-a is the most abundant green pigment in cyanobacteria playing a central role in photosynthesis [ 27 ]. It is a commercially important natural dye and is used to color inks, cosmetics, perfumes, liniments, and leather [ 27 ]. Our results showed similar chlorophyll-a productions of 4.26 mg/l and 4.27 mg/l in the control and sodium acetate supplementation groups. Unlike the enhanced biomass productivity in the treatment group, there was no change in the productivity of chlorophyll-a. Sodium acetate treatment appears to reduce chlorophyll a content. chlorophyll a content is known to vary according to the cultivation conditions ( e.g. , type and intensity of light, nutrient composition, and temperature) [ 28 ]. Consumption of organic carbon sources by photosynthetic microorganisms in mixotrophic cultivation can lead to a decrease in chlorophyll content due to changes in photosystem activity [ 29 ]. Previous reports have shown that organic carbon sources cause a decrease in the amount of excitation energy retained in PSII, resulting in a decrease in photosystem II (PSII) activity [ 30 ]. Photosystem II activity represents a photosynthetic efficiency, and thus acts as a factor that can indirectly prove the decrease in the content of chlorophyll a. Phycobiliprotein is a pigment-protein complex responsible for light collection in cyanobacteria and is a value-added substance with diverse potential uses in the cosmetic, food, health, and medical industries [ 31 ]. Phycobiliprotein is composed of phycocyanin (620 nm), allophycocyanin (652 nm), and phycoerythrin (565 nm), which can be distinguished based on the light absorption wavelength [ 31 ]. The phycobiliprotein concentration in the P. mucicola GO0704 cell culture was 52.58 mg/l, which accounted for a 27% increase in mixotrophic conditions. According to ( Table 1 ), most of the phycobiliprotein observed in our study was phycocyanin (PC), and the acetate-treated group produced 52.58 mg/l and the control group produced 31.39 mg/l of PC. There were no significant differences in phycocyanin content between the acetate supplemented group and the control group. Previous research has reported that the accumulation of PC in diverse species was promoted in the mixotrophic condition [ 32 - 34 ]. This discrepancy might be due to the result of a complex PC accumulation regulation mechanism caused by the difference in the types of carbon sources [ 35 ]. it was found that G. sulphuraria grown on glycerol under certain light conditions stimulated the accumulation of PC, but not those grown on glucose and fructose [ 36 , 37 ]. Treatment of glucose, lactose, and galactose also showed an increase in PC content of Anabaena variabilis , whereas treatment of fructose and sucrose showed no significant change in it [ 38 ]. However, when combined with biomass productivity had led to higher PC productivity in the treated group. Among phycobiliprotein, phycocyanin is the most widely studied pigment and has been the focus of many natural bioactive compound screening studies [ 31 ]. Phycocyanin has been reported to possess antioxidant, anti-inflammatory, and immune-stimulating properties [ 31 ]. Moreover, It has been reported to block the cell cycle and could thus be used as an effective anticancer agent [ 39 ]. Our results suggested that the P. mucicola strain GO0704 could enhance the productivity of phycocyanin by acetate treatment and further potentially be used as a natural source of bioactive substances that enhance human health. However, prior to the commercialization of food and biomedical products, candidate compounds must pass a rigorous safety examination. In contrast, biodiesel applications of cyanobacteria such as P. mucicola GO0704 are relatively unaffected by the toxicity of cyanobacteria. Therefore, the utilization of cyanobacteria cells as a biodiesel source is not limited by the safety requirements of products intended to be used in the food and medical industries. Fatty acids are the main components of biodiesel, and the total fatty acid production of P. mucicola GO0704 was 36.17 mg/l. The maximum yield was achieved in the sodium acetate treatment group at 144 h, which constituted a 26% increase compared to the control group ( Fig. 3B ). In the sodium acetate treatment, the fatty acid content varied with the cultivation period. The content of total fatty acids reached 5.8%, 8.1%, and 6.8% at 48 h, 96 h, and 144 h, respectively ( i.e. , the highest fatty acid content was achieved at 96 h) (Fig. S2B). The decrease in fatty acid content observed at 144 h might be related to the complete consumption of sodium acetate, as sodium acetate is often used as a carbon source to enhance total lipid and fatty acid contents [ 40 , 41 ]. Notably, sodium acetate addition at the initial culture stage induced fatty acid accumulation. After all carbon sources were consumed within 144 h, the mixotrophic mode might have gradually shifted to phototrophic growth. The decrease in fatty acids can be attributed to the fact that accumulated metabolites such as fatty acids are often used as energy sources for cell growth. Given that the fatty acids content of P. mucicola GO0704 was reduced after depletion of sodium acetate, additional supplementation of carbon sources could further elevate the maximum contents of fatty acids, which leaves for further research. Biodiesel productivity is a critical variable in selecting the most appropriate species for biodiesel production [ 42 ]. This factor is also a practical indicator for verifying the economic feasibility of biodiesel commercialization [ 43 ]. Biodiesel productivity is calculated as fatty acid content and biomass production divided by the culture time required to achieve a profitable cell density. In this study, the maximum biodiesel productivity rates at 48, 96, and 144 h were 6.2, 8.1, and 6.3 mg/l/d in mixotrophic conditions. Therefore, the P. mucicola GO0704 cultures should ideally be harvested at 96 h to maximize biodiesel yields. Additionally, the biodiesel productivity of most isolated cyanobacteria is lower than that of P. mucicola GO0704 (8.1 mg/l/d), thus highlighting its applicability as a bioenergy source ( Table 2 ). Fatty Acid Profiling in Two Trophic (photo-/mixo-) Culture Modes Next, detailed fatty acid compositions were studied through conventional FAME extraction and conversion procedures, and a total of 8 fatty acids were detected, including 4 saturated fatty acids (SFAs), 2 monounsaturated fatty acids (MUFAs), and 2 polyunsaturated fatty acids (PUFAs). In the case of supplementation with sodium acetate, the percentage of C12:0, C16:0, C16:1, C18:0, and C18:1 increased, whereas the contents of C14:0, C18:2, and C18:3 decreased ( Table 3 ). Along with the decrease in the percentage of total PUFA, the levels of total MUFAs increased. Consistent with these observations, the conversion of PUFAs to MUFAs under mixotrophic conditions has been frequently reported in previous studies [ 44 - 46 ]. The C/N ratio of the medium is known to have a significant influence on the fatty acid composition of microalgae [ 47 ]. Likewise, it can be inferred that PUFAs were more readily converted to MUFAs under mixotrophic conditions in cyanobacteria such as P. mucicola GO0704. Increases in the proportion of MUFAs such as C18:1 improves the ignition performance of biodiesel by effectively negating the poor oxidation stability of PUFAs [ 48 ]. However, the composition of SFAs is also known to have a crucial impact on the low-temperature fluidity of biodiesel [ 49 ]. Therefore, ideal biodiesel production is highly complex, thus requiring a balance of properties such as ignition efficiency, low-temperature fluidity, and oxidation stability. Therefore, in-depth fatty acid composition analyses must be conducted to select optimal biodiesel sources. To accurately evaluate the quality of FAMEs in biodiesel, the expected properties must be carefully cross-checked with existing references. Quality Assessment of Biodiesel Derived from Pseudanabaena mucicola GO0704 The quality of FAMEs derived from P. mucicola GO0704 was predicted using the Biodiesel Analyzer software [ 50 ]. The properties of biodiesels derived from P. mucicola GO0704 and other microalgae ( Euglena ), soybean biodiesel (currently the main source of biodiesel), and international biodiesel standards [ASTM 6761 (USA) and EN 14214 (Europe)] were carefully compared to assess the quality of P. mucicola GO0704-derived biodiesel ( Table 4 ). To this end, important properties often reported in the other studies for the evaluation of biodiesel were considered, including viscosity, cloud point, low-temperature filter plugging point, pour point, cetane number, and iodine value. First, the cetane number of biodiesels derived from P. mucicola was 56.2, which is higher than that of soybean (49.0). In general, a higher cetane number suggests a shorter infusion delay [ 51 ]. P. mucicola biodiesel, which exhibited a higher cetane number than that of soybean, may thus not be affected by delayed infusion. Biodiesel derived from P. mucicola also has a lower viscosity than those of soybean oil and Euglena. High viscosity causes fogging problems and may form engine deposits [ 52 ]. The iodine value (gI 2 100/g) measured in P. mucicola -based biodiesel (65.1) was approximately 0.5 times that of soybean oil (128) and lower than that of Euglena-derived biodiesel (118). The iodine value is an important index of the degree of unsaturation contained in fatty acids and a decrease in unsaturation leads to an increase in oxidative stability [ 53 ]. The cloud point of P. mucicola is 3.9°C, which is higher than the −3°C of the ASTM 6761 (US) standard specification. Diesel fuel crystallizes with apparent solidification and turns turbid at a temperature known as the cloud point. A high cloud point causes clogging of fuel lines and filters due to crystallization [ 52 ]. Euglena biodiesel with a CP point of 15°C poses a greater risk of filter clogging in colder regions. Soybeans with high cold filter plugging point also have the same issue, which could be only possibly reverted with the incorporation of additives into fuels. For example, a nearly 30% decrease in the fuel pour point temperature was achieved when using a 20% ethanol fuel mixture [ 54 ]. Taken together, our findings indicated that the native cyanobacterial strain ( P. mucicola Go0704) isolated and characterized in this study met most of the standard properties of ASTM6761 (USA) and EN 14214 (Europe), and the potential problem of CP could be solved using additives. P. mucicola -derived biodiesel displayed better performance than soybean oil, the first-generation biodiesel, and Euglena-based biodiesel, a representative microalga-derived biodiesel. In summary, the native strain P. mucicola GO0704 is worth considering as a potential biodiesel resource. However, a comprehensive analysis of P. mucicola GO0704-derived biodiesel production from cultivation to conversion is necessary to confirm its feasibility. Feasibility of Mixotrophic Cultivation of Pseudanabaena mucicola GO0704 Using Acetate Obtained from Organic Wastes Acetate can be readily obtained from the acetic acid fermentation process of organic wastes like wastewater and food waste [ 55 ]. Table 5 shows the conversion of organic waste to acetate by anaerobic processes. Food waste (96.0 g COD/L) can be converted into acetate (25.9 g/l), reaching a maximum conversion yield of 27% [ 56 ]. Massive amounts of food waste are generated each year (1.3 billion tons) [ 57 ] and therefore acetate could be obtained from food waste for mixotrophic cultivation. Moreover, given that the cost of organic carbon sources is among the highest expenses in microalgae cultivation [ 58 ], the utilization of acetate obtained from food waste offers a great opportunity to reduce not only the disposal of food waste but also the cost of biodiesel production from cyanobacteria. Our study comprehensively assessed the entire cyanobacteria-derived biodiesel production process, from the isolation and identification of the cyanobacteria strain, the application of the mixotrophic mode in large-scale cultivation, and the analysis of valuable substances with an emphasis on biodiesel performance prediction. The cyanobacteria collected from Goryeong, Republic of Korea, was identified as P. mucicola (strain name, GO0704). Sodium acetate has been shown to induce the mixotrophic mode and enhance the productivity of phycobiliprotein and fatty acids in large-scale cultivation. Sodium acetate treatment improved the quality of biodiesel by enhancing the fatty acid composition. The potential of P. mucicola GO0704 for biodiesel production was also carefully evaluated by analyzing the biodiesel properties and comparing them to international standards. Our study is the first to demonstrate the potential of an indigenous strain of P. mucicola for biodiesel in large-scale mixotrophic cultures and provides a basis for the future development of P. mucicola -derived biodiesel production."
} | 5,974 |
22856464 | null | s2 | 1,017 | {
"abstract": "This review examines the electrochemical techniques used to study extracellular electron transfer in the electrochemically active biofilms that are used in microbial fuel cells and other bioelectrochemical systems. Electrochemically active biofilms are defined as biofilms that exchange electrons with conductive surfaces: electrodes. Following the electrochemical conventions, and recognizing that electrodes can be considered reactants in these bioelectrochemical processes, biofilms that deliver electrons to the biofilm electrode are called anodic, ie electrode-reducing, biofilms, while biofilms that accept electrons from the biofilm electrode are called cathodic, ie electrode-oxidizing, biofilms. How to grow these electrochemically active biofilms in bioelectrochemical systems is discussed and also the critical choices made in the experimental setup that affect the experimental results. The reactor configurations used in bioelectrochemical systems research are also described and the authors demonstrate how to use selected voltammetric techniques to study extracellular electron transfer in bioelectrochemical systems. Finally, some critical concerns with the proposed electron transfer mechanisms in bioelectrochemical systems are addressed together with the prospects of bioelectrochemical systems as energy-converting and energy-harvesting devices."
} | 341 |
29120363 | PMC5712837 | pmc | 1,018 | {
"abstract": "This paper presents a power-generating sensor array in a flexible and stretchable form. The proposed device is composed of resistive strain sensors, capacitive tactile sensors, and a triboelectric energy harvester in a single platform. The device is implemented in a woven textile structure by using proposed functional threads. A single functional thread is composed of a flexible hollow tube coated with silver nanowires on the outer surface and a conductive silver thread inside the tube. The total size of the device is 60 × 60 mm 2 having a 5 × 5 array of sensor cell. The touch force in the vertical direction can be sensed by measuring the capacitance between the warp and weft functional threads. In addition, because silver nanowire layers provide piezoresistivity, the strain applied in the lateral direction can be detected by measuring the resistance of each thread. Last, with regard to the energy harvester, the maximum power and power density were measured as 201 μW and 0.48 W/m 2 , respectively, when the device was pushed in the vertical direction.",
"conclusion": "5. Conclusions In this research, we proposed and demonstrated a power-generating multi-sensor array that can be operated as a triboelectric energy harvester, a capacitive tactile sensor, and a resistive strain sensor. This device is able to detect signals through human movement and generate power by energy harvesting in a single device. The proposed device was implemented by a simple fabrication process such as dip coating, spray coating, or vacuum coating. In addition, the proposed device, which consists of stretchable and flexible materials including hollow tubes, Ag NWs layers, and PDMS and EcoFlex mixtures, is easily bendable, stretchable, foldable, and twistable and has excellent durability. The prototype device was made up of five-row and five-column functional threads, resulting in a 5 × 5 sensor array in an area of 6 × 6 cm 2 . When operated as an energy harvester, the maximum power and power density were measured as 210 μW and 0.48 W/m 2 , respectively, at a 20-MΩ load resistance. In addition, it was demonstrated that the capacitance output increased with a vertical pressurizing force as a tactile sensor and the resistance output increased with a lateral stretching force as a strain sensor. The proposed device is expected to be used in a variety of fields to monitor human motions and activities because the device was fabricated in a woven structure similar to actual clothes and can be extended to a large scale.",
"introduction": "1. Introduction In recent years, wearable devices, which provide a connection between humans and devices anywhere and at any time, have been widely used in various fields such as health care, medical diagnosis, fitness, wellness, entertainment, and military applications to measure the physical conditions of the human body or to check body motion and activity in real time [ 1 , 2 , 3 , 4 , 5 ]. These devices are mainly implemented in the form of wearable accessories such as smart bands and smart glasses, body-attachable devices such as smart lens and smart tattoos, and smart clothes such as sportswear and shoes [ 6 , 7 , 8 , 9 , 10 ]. Among them, smart clothes are a wearable platform that can be equipped with sensors, actuators, data processors, and power sources and has the advantage of being worn like ordinary clothing [ 11 , 12 ]. According to the application and purpose of use, a variety of sensors including pressure sensors, strain sensors, temperature sensors, accelerometers, gyro sensors, gas sensors, and tactile sensors have been utilized in smart clothes [ 13 , 14 , 15 , 16 , 17 ]. Among them, tactile sensors, which detect tactile information through physical touch, have been widely developed to detect various types of information including shape, texture, softness, temperature, vibration, and shear or normal forces. Such a tactile sensor can be implemented by using capacitive, optical, piezoelectric, and piezoresistive transducers to detect the position and magnitude of the input touch force [ 18 , 19 , 20 , 21 ]. Capacitive transducers have been widely adapted for smart clothes applications because they can be relatively easily implemented in a large area in a textile structure. Lee et al. reported on a flexible and sensitive textile-based pressure sensor using conductive fibers coated with rubber materials [ 22 ]. Hasegawa et al. proposed a tactile sensor by weaving hollow tubes and demonstrated a capacitive tactile sensor array in a textile structure [ 23 ]. The strain sensors in smart clothes are used for healthcare systems such as body movement monitoring, medical monitoring, and sports and injury prevention in real time. When the human body deforms intentionally or unintentionally, the strain sensors monitor the magnitude, direction, frequency, and location of the deformation by measuring resistance changes. Wang et al. fabricated wearable strain sensors through a simple carbonization process with woven structures of silk fabric [ 24 ]. In addition, Li et al. and Park et al. reported on highly stretchable and flexible strain sensors composed of carbon-based materials and polymer matrices [ 25 , 26 ]. The devices detected human motion by mounting the sensors on the joint areas of the finger, elbow, and knee. An important issue for these wearable devices and smart clothes is a power source to operate the sensors, electronics, and other components requiring electric power for proper operation. To solve the problem of limited lifetime and the need for periodic recharge of the battery, energy harvesters, which convert mechanical energy into electrical energy and function as semipermanent power sources without the need for an external battery, are being studied [ 27 , 28 , 29 , 30 , 31 , 32 , 33 ]. Recently, several devices having sensors and energy harvesters in a single platform were reported. These devices sense body activity or environmental signals and, at the same time, generate electric power from the human body [ 29 , 30 , 31 , 32 , 33 , 34 ]. Ahn et al. reported on a woven-structured power-generating tactile sensor array in which a piezoelectric energy harvester and tactile sensors were integrated in a single device [ 29 ]. In addition, Saha et al. proposed an electromagnetic generator from human motion and proposed a structure that can be applied to a body-worn sensor [ 33 ]. We have also reported a triboelectric and piezoresistive property of a single hollow tube structure with surface modification [ 34 ]. That report presented that the electrode layers for energy harvester could be utilized as a piezoresistive sensing layers. The triboelectric energy harvester has been widely developed and utilized in converting mechanical energy into electrical energy from periodic contact and separation between materials having different charge affinities. Triboelectric energy harvesters have several advantages including easy fabrication, cost effectiveness, high energy conversion efficiency, and the possibility of using a variety of materials. In addition, since they are easily implemented by using polymer-based or flexible materials, they are suitable for smart-clothes applications requiring self-power generation [ 35 , 36 , 37 , 38 , 39 , 40 ]. Thus, in this work, we propose a device composed of resistive strain sensors, capacitive tactile sensors, and a triboelectric energy harvester in a single platform. The proposed device has a woven textile structure that provides an expandable tactile and strain sensor array. In addition, the output power as a triboelectric energy harvester has been increased by applying surface modifications of the materials using self-assembly monolayer (SAM) coating [ 34 , 41 ].",
"discussion": "4. Results and Discussion 4.1. Energy-Harvesting Mode Figure 6 a shows the output properties of the prepared single functional thread when it is pressed and released repeatedly by a stamp with 10 N of pressing force. The actuation frequency was 3 Hz in this experiment. The open-circuit output voltage was 2.3 V without the FOTS coating. However, by coating the inner surface of the tube with FOTS, we could obtain an output of 27.7 V, which is 12 times greater than the output without the FOTS coating, because FOTS, which is composed of fluorine atoms, provides a higher electron affinity compared to a bare silicone surface. Figure 6 b shows the open-circuit output voltage and short-circuit current from a single thread operated at various actuation frequencies. We used a shaker (LDS V406, Bruel & Kjaer, Naerum, Denmark) to apply a repeated pressurizing force to the thread at various vibration frequencies. Both the output voltage and current increased with the actuation frequency. The maximum output voltage and current were measured as 43 V and 9.9 μA, respectively, at 6 Hz. As shown in the graph, the voltage and current increased with the frequency because the pressurizing force increased as the moving frequency and acceleration increased [ 47 ]. The output properties of a complete device having 5 × 5 row and column threads in a woven structure are shown in Figure 7 . All 10 threads were set in a parallel connection electrically. Figure 7 a shows the open-circuit voltage and short-circuit current. The output voltage did not make a substantial difference compared to the value from a single thread in Figure 6 because the threads were connected in parallel. However, the short-circuit current was increased to 52 μA at 6 Hz, which is five times larger compared to the output from a single thread. Theoretically, this is supposed to produce 10 times more output current than a single thread. However, in this experiment, the output might be decreased because the pressurizing and triboelectric contacts were not perfectly even to all threads. Consequently, the signals from each thread were not perfectly synchronized. Figure 7 b shows the peak-to-peak output voltage and current measured at various load resistances. In this work, we measured the voltage and current across the load using an oscilloscope and a low-noise current preamplifier (SR0570, Stanford Research Systems, Sunnyvale, CA, USA), respectively, to determine the optimal load resistance. As the load resistance increased, the output current decreased because the source current was divided between the internal impedance and load resistor by the magnitude ratio of the parallel-connected impedances. By contrast, the output voltage increased with the resistance. The load resistance was optimal when the product of the voltage and current was at its maximum value. The relationship between the load resistance and the peak-to-peak output power is shown in Figure 7 c for an operating frequency of 3 Hz and applied force of 10 N, respectively, where the maximum power was measured as 201 μW and 0.48 W/m 2 , respectively, at a load resistance of 20 MΩ. 4.2. Capacitive Tactile Sensor Mode As described in Section 2.2 , the proposed device can monitor the location and magnitude of the vertical pressurizing force by measuring the capacitances formed at the crossing positions of the row and column threads. As shown in Figure 8 a, the capacitive sensing cells are formed at the intersections of each row and column functional thread. The initial capacitances of the 5 × 5 capacitive cells in the fabricated device are shown in Figure 8 b. We measured the capacitance value using an LCR meter (4284A, Agilent, Santa Clara, CA, USA). The initial average capacitance of a total of 25 cells was 1.789 pF with a standard deviation of 30 fF. The slight variation between cells might come from the nonuniform thickness of the insulating layer and variations in the gap distance between cells. These are inevitable fabrication errors resulting from manual assembly steps during device fabrication. This initial offset can be minimized by applying a compensation technique to the readout circuit or a signal processing module in practical applications. Figure 8 c shows the capacitance values after applying a vertical force of 2 N to each cell. The capacitance was increased by 12.8% on average with a standard deviation of 40 fF. The sensitivity of a single cell was measured and is depicted in Figure 9 a while applying various vertical forces to a specific cell by using a force gauge (HF-50, JISC, Bristol, UK). The capacitance increases as the applied force increases because the contact area between the row and column threads increases with force. The initial capacitance was measured as 1.79 pF when the device was not pushed. The capacitance increased as the normal force increased, and was measured as 1.97 pF at a force of 2 N. The sensitivity of the capacitive cell was approximately 4.72%/N in a force range of 2 N. However, the capacitive cell became less sensitive to a normal force larger than approximately 2 N because the hollow tubes were almost collapsed and were barely deformed. The sensitivity in this regime is approximately 0.24%/N in a force range from 2 N to 6 N. In this experiment, to determine the hysteresis of the sensor, the capacitance was measured while increasing and decreasing the applied force on the tactile sensor three times repeatedly. The maximum difference between the capacitance values measured during force increment and decrement was 6 fF, which is 0.31% of the initial capacitance. This low hysteresis is a good characteristic for sensor applications. Figure 9 b shows the capacitance values measured from a single cell of the tactile sensor array for the applied force while applying various lateral forces to extend the threads by 2, 4, 6, and 10 mm. In this experiment, the data was obtained from the capacitive cell formed at the intersection point of the third row and third column threads of the device. The initial capacitance, where no vertical force was applied, increased as the lateral stretching length increased because the distance between the top and bottom electrodes decreased as the device was laterally stretched. 4.3. Resistive Strain Sensor Mode The stretching or in-plane strain can be measured by reading the resistance change in each column and row thread. Figure 10 a shows the resistance of a single thread in the releasing state during repeated stretching and releasing actuations of up to 8000 cycles. In this experiment, the stretching length of the thread was fixed to 20% of its initial length. Two different threads were tested. One was a thread coated with APTES on the outer surface of the silicone tube before Ag NWs deposition, and no pretreatment was used for the other thread. As shown in the graph, the initial resistance values with and without APTES were 104 Ω and 890 Ω for each. After we pulled and released the thread for 8000 cycles, the resistance of the device without the APTES coating was increased to 1.65 kΩ, which is an 86% increment. However, little increase in resistance was observed for the APTES-coated thread. In this experiment, it is clear that the stability of the Ag NWs layer on a silicone surface can be dramatically increased by applying APTES as an adhesion layer. The piezoresistive property of an APTES-coated thread was measured using a motorized linear actuator to apply uniaxial tensile strain. Figure 10 b shows the resistance change when the thread is pulled up to 20% and released to its initial length. We can see that the resistance increased and decreased linearly as the strain changed. The gauge factor was about 12.1, which is much larger than that of the conventional metallic strain gauge. In addition to good linearity and sensitivity, the hysteresis of the fabricated thread as a strain sensor was quite low, which was a maximally 4.6% difference during stretching and releasing. The resistance change of the five threads in the fabricated 5 × 5 woven array device was also measured while applying a tensile strain to the threads, of up to 20%. As shown in Figure 10 c, the sensitivity variation between threads was very small, which can be easily compensated in a readout circuit or through signal processing."
} | 4,014 |
35507654 | PMC9067916 | pmc | 1,019 | {
"abstract": "The healthy functioning of the plants’ vasculature depends on their ability to respond to environmental changes. In contrast, synthetic microfluidic systems have rarely demonstrated this environmental responsiveness. Plants respond to environmental stimuli through nastic movement, which inspires us to introduce transformable microfluidics: By embedding stimuli-responsive materials, the microfluidic device can respond to temperature, humidity, and light irradiance. Furthermore, by designing a foldable geometry, these responsive movements can follow the preset origami transformation. We term this device TransfOrigami microfluidics (TOM) to highlight the close connection between its transformation and the origami structure. TOM can be used as an environmentally adaptive photomicroreactor. It senses the environmental stimuli and feeds them back positively into photosynthetic conversion through morphological transformation. The principle behind this morphable microsystem can potentially be extended to applications that require responsiveness between the environment and the devices, such as dynamic artificial vascular networks and shape-adaptive flexible electronics.",
"introduction": "INTRODUCTION Plants have very rich and complex vasculature that transports water and nutrients through their tissues to maintain normal metabolism. For example, in leaves, which represent the main organ of plants for photosynthesis, their veins can deliver the nutrients produced by photosynthesis to the rest of the body while also transporting water throughout the leaflet for transpiration. This vasculature has inspired us to develop artificial systems with embedded fluidic channels, such as biomimetic microfluidic devices ( 1 – 5 ). Moreover, plants have evolved the ability to respond to environmental changes to allow their vasculatures to function more healthily even in an ever-changing natural environment. This ability of individual systems to sense the external environment and adjust themselves accordingly is sometimes known as “environmental adaptability.” In particular, the nastic movement of plants is the fastest way to adapt their shapes to environmental changes in light, temperature, and humidity ( 6 , 7 ). For example, the genus Oxalis deploys the leaflets during the sunny day to promote photosynthesis and folds them at night to retard energy dissipation due to transpiration ( 8 ). However, this plant-like ability to interact with the environment is rarely mentioned in synthetic microfluidic systems ( 9 , 10 ). Endowing microfluidic systems with these stimuli-responsive shape-changing functions can pave the way for sophisticated, multifunctional, or intelligent fluidic systems with dynamic biomimetic design or environmental adaptability ( 3 , 11 – 28 ). Now, microfluidic systems respond to the environment by relying on linked or onboard electronics and computer programming. This can result in a tethered, complex, and overall cumbersome system ( 26 , 29 , 30 ). In contrast, their natural counterparts contain veins as fluid-transporting microchannels and pulvinus as stimuli-responsive actuators in their thin, lightweight, and flexible leaflets ( 6 , 7 ). Besides, along with the nastic movement, the leaflets tend to fold or unfold into a regular geometry that has specific purposes. Two main aspects that restrict the development of stimuli-responsive morphing microfluidics are as follows: First, the mainstream materials used to fabricate microfluidic devices are inert materials without environmental responsiveness, such as polydimethylsiloxane (PDMS) and poly(methyl methacrylate) (PMMA); second, the design of microfluidic devices is usually based on considering the fluidic flow to design the embedded channel shape rather than the overall shape of the microfluidic device. Consequently, most common microfluidic devices do not show environmental responsiveness and targeted shape change, let alone the nastic movement, similar to those in leaves. To gain ability of nastic movement, the microfluidic device has to exhibit stimuli responsiveness at the device level, and the overall shape of the device has to be programmable. These microfluidic devices will enable previously unfeasible applications: For instance, the resultant environmentally adaptive photomicroreactors self-regulate the photosynthetic conversion rate according to weather changes ( 23 , 24 , 26 ). Stimuli-responsive materials—such as shape memory polymers, liquid crystal elastomers, and responsive hydrogels—are capable of changing their shapes in response to external temperature, light, or humidity ( 29 , 31 – 33 ). They have functioned as both sensors and actuators synchronously in many morphable structures. Nonetheless, current microfluidic devices use responsive materials only as localized components, such as light-controlled valves or flow-switching channels, rather than in the overall morphing of a microfluidic device ( 34 – 39 ). To implement a preset overall three-dimensional (3D) morphing, the device’s dimensions, the positions of the responsive materials embedded in the device, and the target operation in response must be engineered specifically ( 40 ). Combining with the ancient art of origami, the desired 3D structures not only can be constructed from precursors via 2D processing techniques but also can mutually convert between a fully deployed 2D plane and a compact 3D form by folding and unfolding ( 29 , 31 , 41 – 44 ). Therefore, by fusing responsive materials into a microfluidic device designed with an origami geometry, it can transform between 2D and 3D states based on the preset structure via stimuli-triggered responses. In this work, the nastic plants inspire our concept of a microfluidic device with environmental responsiveness, and the stimuli-responsive structures offer us the conditions to realize this concept. On the basis of these, we have developed a transformable microfluidic chip by integrating stimuli-responsive materials with a thin and foldable microfluidic chip ( Fig. 1 ). The entire device can respond to changes in temperature, humidity, and light irradiance by transforming along the preset origami folds. Thus, we name this transformable origami microfluidic approach TransfOrigami microfluidics (TOM). We demonstrate that TOM can be applied as an environmentally adaptive photomicroreactor. The transformation reconstructs the reaction channels, changes their light-harvesting capability, and eventually regulates the photosynthetic conversion. Positive feedback control is built into interactions between TOM and the environment. When the external conditions are favorable for the photoreaction, the feedback results in an enhanced photosynthetic conversion and vice versa. As the first of its kind, stimuli-responsive morphing microsystems, our TOM will inspire applications in energy, robotics, or biomedicine that require environmental adaptations, such as artificial vascular networks or flexible electronics with adaptive rhythmic movements ( 40 , 45 , 46 ). Fig. 1. Plant-inspired TOM. O. corniculata at its ( A ) open state during the day time and ( B ) close state during night time. Left insets: Top view of its open and close states, respectively. Right insets: Unfolded and folded states of the corresponding origami structure, respectively. ( C ) Schematic drawing of the TOM at its unfolded state when the temperature is high and light irradiates; folded state when the temperature is low and humidity is high. The insets show the schematic dehydration and hydration in the pNIPAM hydrogel layer of the actuating unit at the unfolded and folded states. ( D ) Bright-field images (channel is filled with blue dye) and dark-field images (channel is filled with fluorescent dye) of TOM’s two states. Scale bars, 10 mm. ( E ) Origami design of the TOM. Solid line indicates the folding line; dashed line indicates the edge line. Scale bar, 10 mm.",
"discussion": "DISCUSSION The reconfiguration and regulation of our morphing microfluidic device can be transformed from 2D to 3D or between different 3D structures. The dynamic switching between different 3D structures adds a temporal dimension, making our device de facto 4D. Inside the 4D microfluidic device, regulation of the fluid behavior is conducted by the reconstruction of microchannels with certain properties, for example, the orientation, mixing efficiency, and flow rate. When fluids interact with the external environment, such as during the photosynthesis reaction in our demonstration, this fluid regulation amplifies the effect of the interaction even more. The trigger for this series of reconfigurations is a change in stimuli in the environment. Therefore, fundamentally, our approach endows the microfluidic system with a plant-inspired way to self-adapt to the environment through morphing. It will be functionally closer to the biological vascular system with enhanced environmental adaptability than traditional microfluidic systems with fixed channel structures and relying on external devices for regulation. This biomimetic concept can be extended to soft microsystems that need to adapt to the environment in the real world, such as wearable electronics that self-adapt to the body’s environment and bionic soft robots that work in a changing environment. Although 4D-printed microfluidics, a concept similar to transformable microfluidics, has been proposed, it relies on 3D printing of stimuli-responsive materials ( 63 , 64 ). Two issues would hamper this: (i) Most of the responsive materials developed using traditional soft lithography lack a set of well-established and compatible micromachining for microfluidics; (ii) 4D printing is still in its infancy stage, where it is not yet involved in morphing microtubing at a resolution that is compatible with soft lithography ( 65 , 66 ). Moreover, our soft devices made by in situ doping and surface modification are all-in-one systems. This avoids not only the errors and matching problems that may occur during the assembly but also the potential detachment caused by instability at the interface of different components during use and eventually increases the robustness of the whole system. The current TOM still has room for improvement. For instance, the resulting 3D microchannel structure can only be formed via the folding of a predesigned 2D channel structure and cannot be transformed into an arbitrary 3D structure. Besides, the response speed of the current TOM is hindered by the slow swelling or shrinkage rate of the hydrogel active layer and the thick PDMS microfluidic layer as the passive layer. These could be improved by developing advanced fabrication techniques and incorporating high-performance responsive components into the material. Thus, further optimization for TOM will be implemented by (i) introducing alternative methods for designing 3D structures, such as krigami; (ii) selecting actuators with higher photothermal efficiency and faster response, such as coupling gold nanorods with liquid crystal elastomer; and (iii) fabricating a thicker active layer and a thinner passive layer in TOM. In general, microfluidics combined with smart materials such as functional hydrogels or elastomers has seen progress during the last decade. However, the study has focused largely on inserting smart materials in the microchannels or modifying the overall matrix material of the device; the programmable morphing of the microfluidic device as a whole is not taken into account. For the TOM system, we propose three potential developments for future research: (i) a better correlation between morphing and fluid in TOM, such as morphing-induced fluid channel switching or fluid-induced morphing; (ii) a wider range of hybrid TOM systems, such as integrating biomaterials into TOM to achieve an organ-on-chip with adaptive rhythmic movements; and (iii) a fully autonomous TOM system, such as a more comprehensive plant-like system combining nastic movement, capillary action, and transpiration. In conclusion, we pioneer plant-inspired morphing origami microfluidics that truly fulfills adaptive photosynthesis. The device coordinates stimuli-responsive morphing materials with a microfluidic chip, which is based on self-actuating elastomer responses to ambient temperature, humidity, and light irradiance, morphing following the preset origami folds. This morphing is further applied to regulate photosynthetic conversion with a built-in positive feedback control in the system. The morphing microfluidics is a smart matter–based intelligent system that could open up a pathway toward the development of intelligent soft devices and artificial vasculature in industrial and biomimetic applications."
} | 3,182 |
33619812 | PMC8252710 | pmc | 1,020 | {
"abstract": "Abstract Anomalous heat waves are causing a major decline of hard corals around the world and threatening the persistence of coral reefs. There are, however, reefs that have been exposed to recurrent thermal stress over the years and whose corals appear to have been tolerant against heat. One of the mechanisms that could explain this phenomenon is local adaptation, but the underlying molecular mechanisms are poorly known. In this work, we applied a seascape genomics approach to study heat stress adaptation in three coral species of New Caledonia (southwestern Pacific) and to uncover the molecular actors potentially involved. We used remote sensing data to characterize the environmental trends across the reef system, and sampled corals living at the most contrasted sites. These samples underwent next generation sequencing to reveal single nucleotide polymorphisms (SNPs), frequencies of which were associated with heat stress gradients. As these SNPs might underpin an adaptive role, we characterized the functional roles of the genes located in their genomic region. In each of the studied species, we found heat stress‐associated SNPs located in proximity of genes involved in pathways well known to contribute to the cellular responses against heat, such as protein folding, oxidative stress homeostasis, inflammatory and apoptotic pathways, and DNA damage‐repair. In some cases, the same candidate molecular targets of heat stress adaptation recurred among species. Together, these results underline the relevance and the power of the seascape genomics approach for the discovery of adaptive traits that could allow corals to persist across wider thermal ranges.",
"conclusion": "5 CONCLUSIONS In this study, seascape genomics allowed us to uncover genetic variants potentially implicated in adaptive processes against different types of heat stress in three coral species of New Caledonia. These variants were located next to genes coding for molecular actors that participate in well‐understood cellular reactions against thermal stress. Of note, some of these potential targets for adaptation recurred in the analyses of different species, supporting the robustness and the power of the seascape genomics. Future studies will focus on performing experimental assays to validate the implication of potentially adaptive genotypes and newly identified genes in the heat stress response and to measure the thermal ranges tolerated by the diverse adaptive genotypes.",
"introduction": "1 INTRODUCTION One of the most dramatic consequences of climate change is the worldwide decline of coral reefs, which are the most biodiverse ecosystems in the marine environment (Hughes et al., 2017 ). Among the main drivers of this decline is coral bleaching, a stress response to anomalous heat waves that eventually causes the death of hard corals (Bellwood et al., 2004 ; Hughes et al., 2017 ). In the most severe episodes, coral bleaching has provoked local coral cover loss of up to 50% (Hughes et al., 2017 , 2018 ), with climate change projections expecting for bleaching conditions to be persistent worldwide by 2050 (Van Hooidonk et al., 2013 ). Despite these alarming perspectives, a glimpse of hope is brought by coral reefs that show resistance after recurrent heat waves (Dance, 2019 ; Hughes et al., 2019 ; Krueger et al., 2017 ; Penin et al., 2013 ; Thompson & van Woesik, 2009 ). One of the mechanisms that might promote heat tolerance in corals is genetic adaptation (Sully et al., 2019 ). Indeed, genetic features potentially involved in thermal tolerance were recently identified in corals from reefs recurrently exposed to heat stress in Japan (Selmoni et al., 2020 ), on the Great Barrier Reef (Fuller et al., 2020 ) and along the western coast of Australia (Thomas et al., 2017 ). In recent years, there has been a growing body of literature investigating how coral thermal adaptation might alter the predictions of reef persistence, and how conservation policies could be modified accordingly (Logan et al., 2014 ; Matz et al., 2018 ; van Oppen et al., 2015 ). Given the crucial role adaptation will play in long‐term reef persistence, there is an urgent need to characterize the adaptive potential of corals (Logan et al., 2014 ; van Oppen et al., 2015 ). For instance, there are still open questions concerning the spatial and temporal scales at which local adaptation operates (Matz et al., 2018 ; Roche et al., 2018 ). Changes in adaptive potential against heat stress have been observed along thermal gradients over hundreds of kilometres (e.g., Thomas et al., 2017 ), but also at reefs with distinct thermal variations located only a few hundred metres apart (e.g., Bay & Palumbi, 2014 ). Furthermore, different coral species are reported to show differential vulnerability against thermal stress, leading to the question of how different life‐history traits (e.g., reproductive strategies, morphology) drive the pace of adaptation (Darling et al., 2012 ; Hughes et al., 2018 ; Loya et al., 2001 ). There are also open questions concerning the molecular mechanisms that might be targeted by heat stress adaptation in corals (Mydlarz et al., 2010 ; van Oppen & Lough, 2009 ; Palumbi et al., 2014 ). Some cellular responses to heat stress are now well characterized, such as DNA repair mechanisms, the activation of the protein folding machinery in the endoplasmic reticulum (ER) or the accumulation of reactive oxygen species (ROS, either endogenous or produced by the symbiont) that progressively elicits inflammatory and apoptotic responses (Maor‐Landaw & Levy, 2016 ; Mydlarz et al., 2010 ; Oakley et al., 2017 ; van Oppen & Lough, 2009 ; Patel et al., 2018 ). However, little is known about which of the many molecular actors participating in these cascades could be targeted by evolutionary processes to increase thermal tolerance. Seascape genomics could contribute to filling these gaps. Seascape genomics is a budding field of population genomics that allows the study of local adaptation in wild populations (Riginos et al., 2016 ). This method combines the environmental characterization of the seascape with a genomic analysis of its population (Rellstab et al., 2015 ). The goal is to identify genetic variants that correlate with environmental gradients and that might underpin an adaptive role (Rellstab et al., 2015 ). Seascape genomics could enhance the characterization of coral adaptive potential because: (i) it requires an extensive sampling strategy that allows the study of adaptation at different geographical scales, and against different types of environmental constraints simultaneously (e.g., mean temperatures, standard deviations, accumulated heat stress; Leempoel et al., 2017 ; Selmoni et al., 2020 ); (ii) its experimental protocol is less laborious in comparison with traditional approaches used for studying coral adaptation (e.g., aquarium experiments, transplantations), and therefore facilitates scaling‐up to a multispecies analysis; and (iii) it is based on genomic data and thus reports candidate molecular targets of adaptation (Rellstab et al., 2015 ; Riginos et al., 2016 ). Moreover, recent work has described how the results of seascape genomics studies on corals can be directly transposed to a conservation perspective and support reef prioritization (Selmoni, Rochat, et al., 2020 ). Here we applied the seascape genomics approach to uncover molecular actors potentially implicated in heat stress adaptation in three bleaching‐prone coral species of New Caledonia, in the southwestern Pacific (Figure 1 ). We first used publicly available satellite data to characterize the seascape conditions for over 1,000 km of the reef system. Coral samples were collected at 20 sites exposed to contrasted environmental conditions. The collected samples underwent a genotype‐by‐sequencing (DArT‐seq) genomic characterization, followed by a seascape genomics analysis accounting for the confounding role of demographic structure. This allowed us to uncover single nucleotide polymorphisms (SNPs) associated with heat stress. We then analysed the functional annotations of genes in proximity of these SNPs and found molecular targets that notably recurred among species and that referred to well‐established heat stress responses in coral cells. Our study highlights the relevance and power of seascape genomics to uncover candidate molecular targets of heat stress adaptation in corals. FIGURE 1 Study area, sampling sites and environmental regions. In (a), the 20 sampling sites around Grande Terre, the main island of New Caledonia (southwest Pacific), are shown in yellow. For every sampling site, the number of genotyped individuals per species ( Acropora millepora : red, Pocillopora damicornis : blue, Pocillopora acuta : green) are given in the corresponding boxes. In the background, coral reefs surrounding Grande Terre are highlighted in five colours representing distinct environmental regions. In (b), environmental characteristics discriminating the five environmental regions are shown",
"discussion": "4 DISCUSSION 4.1 Population structure Prior to the analysis of local adaptation, we evaluated the population structure of the three studied species from New Caledonia. Such preliminary analysis is crucial, since strong structure of neutral genetic variation, for instance due to cryptic species or isolated populations, can cause bias in the investigation of adaptive genetic variants (Rellstab et al., 2015 ; Selmoni, Vajana, et al., 2020 ). However, the fact that we could not identify clear genetic clusters for running the DAPC without a priori information suggested the absence of genetically isolated groups in the three studied populations. This view was supported by the lack of clear differences in the frequency of minor alleles observed across the different sampling locations. The DAPC using sampling locations as the grouping factor indicated the presence of a spatial structure in each of the three studied populations. In Acropora millepora , we observed a weak structure, with substantial variation within sampling locations. This observation was consistent with the weak structure recently observed in an A . millepora population from a section of the Australian Great Barrier Reef with spatial extent similar to the one of New Caledonia (Fuller et al., 2020 ). In comparison, the population structure of the two Pocillopora species appeared to be more stressed. Corroborating these observations, previous work in New Caledonia and northwestern Australia suggested high levels of population differentiation in Pocillopora damicornis (Oury et al., 2020 ; L. Thomas et al., 2014 ), and in Pocillopora acuta in New Caledonia (Gélin et al., 2018 ). Of note, we observed a higher frequency of P . damicornis along the west coast of Grande Terre, and a higher frequency of P . acuta along the east coast. These disproportions are probably due to sampling bias, rather than to a divergent ecological specialization of the two species. Indeed, we often found the two species in sympatry, and previous studies systematically found both species along both coasts of Grande Terre (Gélin, Pirog, et al., 2018 ; Oury et al., 2020 ). 4.2 Different types of local adaptation In each of the studied species, we detected genotype–environment associations that might underpin local adaptation. These associations rarely involved outlier SNPs, but this was not surprising since genotype–environment association methods are more sensitive to small shifts in allele frequencies, compared to outlier tests (Rellstab et al., 2015 ). In general, we observed that the environmental variables associated with the largest number of SNPs were those describing SST averages and standard deviations (80 SNPs across the three species), sea surface salinity (79 SNPs), chlorophyll concentration (65 SNPs) and BAF (54 SNPs). These similar numbers are explainable by the fact that all these variables are partially collinear to each other and some of them might therefore be involved in false associations with SNPs. Furthermore, the number of variables per environmental descriptor is likely to drive the numbers of significant SNPs. For instance, SST and salinity have the largest numbers of environmental descriptors (up to six noncollinear environmental variables per species), while BAF has the lowest (two per species, i.e., BAF1 and BAF5). In fact, BAF is the environmental descriptor with the highest average number of significant SNPs per environmental variable (nine). Coral bleaching is a major threat for coral survival, and bleaching conditions emerge when SST variation exceeds seasonal averages (Hughes et al., 2017 ; Liu et al., 2003 ). BAF descriptors account precisely for this selective constraint (SST variation over average), and this might explain why genotype–environment associations with BAFs were on average more frequent. Previous work on coral seascape genomics also reported a predominance of adaptive signals related to BAF (Selmoni, Rochat, et al., 2020 ). 4.3 Candidate molecular targets for heat stress adaptation Genes located in proximity of the SNPs associated with heat stress displayed some molecular functions that are known to be implicated in coral heat stress responses. In some cases, such functions were found in proximity of the few SNPs significantly associated with heat stress. More often, however, we found such functions as over‐represented among groups of SNPs that are associated with heat stress, but necessarily significantly associated. One of the main examples of such molecular functions concerns molecular chaperones. These are proteins, such as heat shock proteins (Hsp), that intervene in cellular responses to heat stress to assist the folding or unfolding of proteins in the endoplasmic reticulum (Oakley et al., 2017 ). In corals, the role of these proteins in the heat response, as well as their up‐regulation under thermal stress, have been reported in several studies (Desalvo et al., , 2008 , 2010 ; Maor‐Landaw & Levy, 2016 ; Oakley et al., 2017 ; Rosic et al., 2011 ). Here we found the GO terms “molecular chaperones,” “Hsp70 protein binding” and “Hsp90 protein binding” as over‐represented in proximity of SNPs associated with heat stress in the three studied species. Of note, we found one of the SNPs most strongly associated ( q =.08) with heat stress in A . millepora in the coding sequence of chaperone Sacsin, which was recently proposed to be involved in bleaching resistance in A . millepora on the Great Barrier Reef (Fuller et al., 2020 ). Another cellular signature of heat stress is the accumulation of ROS in the cytoplasm (Patel et al., 2018 ). Previous studies showed that corals exposed to heat stress respond to this accumulation by activating the molecular pathways of the oxidative stress response (Louis et al., 2017 ; Nielsen et al., 2018 ; Oakley et al., 2017 ; Voolstra et al., 2009 , 2011 ). In the three studied species, we found SNPs associated with heat stress in proximity of molecular actors of the oxidative homeostasis, such as Quinone oxidoreductase PIG3, Malonyl‐CoA decarboxylase, HMG‐CoA reductase and Enoyl‐[acyl‐carrier‐protein] reductase. One of the proposed sources of ROS accumulation is leakage from the host mitochondrion, even though the underlying mechanisms are poorly known (Dunn et al., 2012 ). It is noteworthy that in A . millepora , the SNP most strongly associated with heat stress was in proximity of the MICOS complex subunit MIC60 gene. MICOS is a key protein in maintenance of the mitochondrial inner membrane architecture, through which ROS are produced, and the outer membrane, through which ROS diffuse into the cytoplasm (Muñoz‐Gómez et al., 2015 ; Zhao et al., 2019 ). Additional molecular signatures of coral heat response were found in proximity of SNPs associated with heat stress, even though such observations were scattered across the three studied species. For instance, oxidative stress is known to trigger inflammatory responses and apoptosis (Courtial et al., 2017 ; Patel et al., 2018 ; Yuan et al., 2019 ), and we observed the over‐representation of GO terms implicated in these functions such as “mitogen‐activated protein kinase binding” (in P . damicornis ) and “p53 binding” (in A . millepora and P . acuta ). Another example concerns the GO term “DNA helicase activity” ( P . acuta ), as this function was previously found as a potential target for heat stress adaptation in a coral population from Japan (Selmoni, Rochat, et al., 2020 ). 4.4 Limitations and future directions In the “Population structure” section we discussed the potential confounding role that neutral genetic variation can have on seascape genomics studies. There are, however, other elements that should be considered when assessing the statistical power of the study. The main element is sample size, as previous work suggested working with sample sizes of at least 200 individuals to secure sufficient statistical power under any demographic scenario (Selmoni, Vajana, et al., 2020 ). We compensated for this potential lack of statistical power with a sampling design strategy maximizing the environmental contrasts between the sampling locations. Nevertheless, partial collinearity persisted between different environmental descriptors and this might have led the false discoveries in the genotype–environment association study (Leempoel et al., 2017 ). A possible solution could be the extension of the study area to the reefs of the neighbouring islands, as this might introduce new combinations of environmental gradients and reduce collinearity. We also encountered important trade‐offs related to the genotyping technique. Compared to traditional RAD‐seq approaches, DArT‐seq loci appeared indeed to be enriched in functional regions of the genome and this facilitated the interpretation of the results (Gawroński et al., 2016 ; Lowry et al., 2017 ). However, some of the genetic variants required substantial imputation (missing rate >20%), and such variants appeared less likely to be detected as associated wityh the environment (Table S8 ). This is not surprising, since rare genotypes (such as adaptive ones) are known to be more difficult to impute (Hoffmann & Witte, 2015 ). An increase in sequencing depth would reduce the need for imputation and consequently increase the statistical power of the study. The next step in the characterization of corals’ adaptive potential is experimental validation. Our work found several genetic variants that might confer selective advantages against thermal stress. For each of the studied species, we can now define multiple‐loci genotypes of heat stress‐resistant colonies and test their fitness under experimental heat stress conducted in aquaria (Krueger et al., 2017 ). As a result, this analysis will allow us to (i) further investigate the role of different heat stress‐associated genotypes and molecular pathways and (ii) provide a concrete measure of the thermal ranges that these coral populations might sustain in the years to come. This information is of paramount importance, as it will allow us to predict the reefs that are expected to already carry heat‐tolerant colonies and to define conservation strategies accordingly (Selmoni, Rochat, et al., 2020 ). For instance, marine protected areas could be established to preserve reefs with higher adaptive potential against heat stress, where such reefs could provide the breeding stock to restore damaged reefs (Baums, 2008 ; van Oppen et al., , 2015 , 2017 )."
} | 4,905 |
35438242 | PMC9400982 | pmc | 1,022 | {
"abstract": "Abstract 2,5‐Furandicarboxylic acid (FDCA) is currently considered one of the most relevant bio‐sourced building blocks, representing a fully sustainable competitor for terephthalic acid as well as the main component in green polymers such as poly(ethylene 2,5‐furandicarboxylate) (PEF). The oxidation of biobased 5‐hydroxymethylfurfural (HMF) represents the most straightforward approach to obtain FDCA, thus attracting the attention of both academia and industries, as testified by Avantium with the creation of a new plant expected to produce 5000 tons per year. Several approaches allow the oxidation of HMF to FDCA. Metal‐mediated homogeneous and heterogeneous catalysis, metal‐free catalysis, electrochemical approaches, light‐mediated procedures, as well as biocatalytic processes share the target to achieve FDCA in high yield and mild conditions. This Review aims to give an up‐to‐date overview of the current developments in the main synthetic pathways to obtain FDCA from HMF, with a specific focus on process sustainability.",
"conclusion": "6 Conclusions and Perspectives This Review tackles the challenge to give a comprehensive overview of the most recent developments for the sustainable transformation of 5‐hydroxymethylfurfural (HMF) into 2,5‐furandicarboxylic acid (FDCA) and portrays an ever‐evolving field for this industrially relevant reaction. The enormous relevance of this synthesis, driven by industrial and academic interests, is testified by the development of a variety of several complementary pathways emerging to combine high production and sustainability. Up to date, the specific design of the heterogeneous metal‐based catalyst architectures helps to develop processes presenting outstanding yields and involving cheap metals even in absence of alkaline conditions, avoiding undesired byproducts and favoring catalyst recycling. Electro‐ and photocatalysis show excellent performance in the laboratory stage. In particular, in a perspective in which renewable energy will gain a predominant role, the electrocatalytic oxidation reaction constitutes a very promising process for clean and green conversion of HMF to FDCA. Moreover, the oxidation process simultaneously produces H 2 that can be used as fuel or reactant. However, the components of electro‐ and photocatalytic devices have not been standardized, and the costs are high. Therefore, efforts should be made in the transfer of such catalytic processes from laboratory to industrial scale, trying to maximize the catalyst recovery in order to minimize the overall production costs. Biotechnological transformations involving cells and enzymes are attracting an increasing interest due to presenting the best performances in extremely mild conditions. Unfortunately, the complexity of the reaction environment, as well as the difficulties in FDCA isolation and purification currently undermine the viability of such an approach for large‐scale production. Finally, it is worth underlining the evolution of new frontier approaches: heterogeneous metal‐free catalysis and continuous HMF oxidation are still poorly explored despite their outstanding potential. Although the technologies hereby discussed objectively differ in key parameters such as reaction conditions, scale, productivity, time of reaction, costs of execution, product isolation, and the technology readiness level (TRL) to be unambiguously compared from a techno‐economical point of view, some relevant trends can be evinced. While the current technological advancement seems to crown metal‐based heterogeneous catalysis as the most suitable industrial approach due to the high productivity and the reliability of the catalytic systems, electrocatalytic processes are gaining tremendous attention, in particular when in correlation with renewable energies exploitation. Indeed, synthetic pathways involving metal‐free systems, light, and biocatalysts currently do not constitute viable alternatives from an industrial perspective, but on the other hand they represent the newborn academic approaches aiming at future safe and sustainable processes for FDCA production.",
"introduction": "1 Introduction The main cause of global warming that scientists and civil society are tackling in recent years is the large amount of anthropogenic CO 2 emissions, which should be urgently reduced. The use of biomass as a monomer source can substantially contribute to solving this concern since biomass removes the atmospheric CO 2 by fixing the carbon during its growth. \n [1] \n In this context, all monomers deriving from biomass are becoming more attractive, and their exploitation represents a challenge for the scientific community. Among all, 5‐hydroxymethylfurfural (HMF), prepared by dehydration of abundant C 6 carbohydrates, \n [2] \n represents a versatile intermediate to obtain active pharmaceutical ingredients, \n [3] \n as well as important bio‐based commodity chemicals for the synthesis of various commercially useful acids, aldehydes, alcohols, and amines. \n [4] \n The most appealing derivatives are 2,5‐dimethylfuran (DMF), a promising bio‐fuel with great energy content, and 2,5‐furandicarboxylic acid (FDCA), which is considered the bio‐based counter‐part of terephthalic acid. Indeed, FDCA has been mainly used to produce poly(ethylene 2,5‐furandicarboxylic acid) (PEF), which could soon replace poly(ethylene terephthalate) (PET) in the packaging market thanks to its greater sustainability and outstanding barrier properties. \n [5] \n The relevance of this key building block is underlined by the new plant for PEF commercialization, planned by the Dutch Company Avantium for the industrial/massive production of FDCA. The world's first flagship plant for the FDCA synthesis is currently under construction at Delfzijl and will be operational in 2024. This facility is expected to produce 5000 tons/year of FDCA, exploiting the YXY technology patented by the Company, which uses fructose as a starting material. \n [6] \n In this context, the development of a fully sustainable route to transform HMF into FDCA is important. This conversion consists of a step‐wise reaction that can follow two different oxidation pathways, through 2,5‐diformylfuran (DFF) or 5‐hydroxymethyl‐2‐furan carboxylic acid (HFCA) (Scheme 1 ), which both lead to the production of FDCA. Scheme 1 General synthetic pathways for FDCA production from HMF Only a few life‐cycle analysis (LCA) studies have been performed to quantify the environmental impact of FDCA synthesis.[ \n 7 \n , \n 8 \n , \n 9 \n , \n 10 \n , \n 11 \n , \n 12 \n ] The reported studies point out that, comparing the production of FDCA with its direct oil‐based competitor, terephthalic acid (TPA), the first one contributes to climate change with 1.60 kg of CO 2 for each kg of FDCA, while the latter one reaches the value of 1.80 kg of CO 2 per equivalent, and the fossil depletion values of FDCA and TPA are 0.44 and 1.17 kg oil equivalent, respectively. \n [10] \n The impact of different scenarios of FDCA purification has been also evaluated, \n [11] \n demonstrating the high effect of energy‐demanding techniques such as flash separation and distillation. Very recently, Avantium with Nova Institute has reported a detailed LCA analysis regarding the production of 250 mL bottles of PEF compared with those of PET, concluding that the PEF‐based products lead to a 33 % decrease in greenhouse gas emission. \n [6] \n However, these studies are characterized by high specificity, strongly influenced by the overall process analyzed, comprehensive of the characteristics of the production site, the presence of existing infrastructures, the transportation impact for feedstock supply, the availability of power supply, and so on. In any case, all the LCAs reported in the literature agree on the conclusion that the largest contribution to the environmental impact derives from the energy demand.[ \n 9 \n , \n 12 \n ] Furthermore, in the sustainability perspective, biomass should be used as feedstock, toxic reagents and/or wastes should be avoided, organic solvents limited in favor of water, and the energy consumption reduced by preferring processes occurring at low temperatures or pressures and for short times. This Review will present an overview of the most recent processes developed to convert HMF into FDCA, focusing on the sustainability of the reaction conditions. Since homogeneous catalysis requires additional steps for catalyst recovery and regeneration, which are energy‐consuming, only heterogeneous and highly efficient catalytic systems will be considered. Particular attention will be paid to the highly efficient metal‐based catalysis and mainly to the innovative base‐free processes. Additionally, this Review will give an update on the most recent developments regarding electrochemical and photochemical catalysis, as well as enzymatic and fermentative processes, which are economic and green techniques, since they need mild reaction conditions and do not require chemical oxidants."
} | 2,247 |
36634144 | PMC9934204 | pmc | 1,024 | {
"abstract": "Significance Bacteria evolve to evade antibiotic pressure, leading to adverse infection outcomes. Understanding the evolutionary dynamics which lead to different antibiotic responses has thus far focused on single-strain bacterial populations, with limited attention to multistrain communities which are more common in nature. Here, we experimentally evolved a simple two-strain community, comprising an antibiotic-resistant strain protecting a susceptible one, and found that susceptible populations evolve tolerance, helping them better survive long antibiotic exposure. Using the interplay between community interactions, antibiotic dynamics, and resource availability, we explain this finding with a simple mathematical model and predict and experimentally verify that an increased resistant strain carrying capacity would render tolerance detrimental. Our results highlight that community interactions can alter bacterial evolutionary responses to antibiotics.",
"discussion": "Discussion In this study, we have shown that when exposed to antibiotics in a community context in which resistant bacteria protect susceptible ones, the susceptible population can evolve tolerance. This tolerance is characterized by a decrease in both the growth and death rates of the bacterial population, which improves their fitness in the community. Moreover, we showed that the evolution of tolerance can be suppressed if we modulate the interactions between resistant and susceptible bacteria, e.g., by increasing the carrying capacity of the resistant strain, which speeds up antibiotic degradation. We achieved this by experimentally evolving a synthetic community in which we could control the environmental conditions, antibiotic concentration, and the strength of interactions between resistant and susceptible bacteria. Finally, using a simple mathematical model of the community, we could also explain the costs and benefits of tolerance by slow growth in various conditions and successfully predict under which conditions it would be expected to emerge in our experimental communities. The strikingly linear relationship between the death rates and growth rates of our evolved isolates hints at a possible causal link between the two, perhaps even indicating an important biological trade-off. While our study did not focus on these possible links, future studies discerning the mechanism behind this pattern are likely to shed light on both the mechanisms and fundamental constraints driving such evolution in response to antibiotics in bacteria. The evolution of mechanisms other than resistance has been observed in other studies with single-species populations exposed to antibiotics above their MIC. One study exposed E. coli populations to different antibiotics during the stationary phase and reported the emergence of persistence (a larger subpopulation of tolerant cells) rather than resistance ( 27 ). Another experimental evolutionary study exposed E. coli populations to different antibiotics during the exponential phase and found single point mutations leading to tolerance (to the specific drug class given) ( 28 ). Interestingly, the line evolved under repetitive ampicillin exposure had a significant decrease in its growth rate. Under the conditions of this study, persistence also evolved in all treatments, while resistance did not evolve. In yet other studies, tolerance by lag evolved ( 4 , 29 ). One crucial difference between these studies and ours is that in our study, the antibiotic concentration is changed naturally by the community itself, instead of having it changed by the experimentalist. In this way, the community changes its own antibiotic landscape over time, and these interactions can make tolerance more beneficial than resistance. While we focused on the evolution of the susceptible strain in our study, it is possible that the resistant auxotroph evolved in our experimental conditions as well. Indeed, it was shown that strains in an obligate cross-feeding system, when exposed to antibiotics, evolved autonomous metabolic activity and weakened the mutualistic interactions ( 12 ). Moreover, the interactions between the two strains in our experiment might change during their evolution. Studying such changes over longer evolutionary timescales is, in our opinion, likely to be a fruitful avenue for future work. Our model included several simplifying assumptions. First, as a proxy for the lysine concentration, we tuned the carrying capacity of the resistant strain, R s a t . We did not explicitly model lysine, its dynamics, and dependence on the abundance of the susceptible strain (say in the form of it producing and secreting lysine). Second, we assumed that lysine concentration affected only the resistant strain’s carrying capacity and not its growth rate, supported by our monoculture measurements ( SI Appendix , Fig. S1 ). However, in some cocultures evolved under high ampicillin and high lysine concentrations, resistant cells survived while the susceptible strain went extinct ( SI Appendix , Fig. S2 ), which does not agree with our model. This could be recapitulated in a variant of our model where the growth rate of the resistant auxotroph increased with the lysine concentration. Third, we assumed that the resistant strain had an arbitrarily large MIC, not dying at any ampicillin concentration. Measurements of the resistant strain “single-cell MIC” ( 30 ) indeed show that it is higher than 800 μg/mL ampicillin. Moreover, we assumed that bacterial growth was not impacted by antibiotic concentrations below the MIC. During the evolutionary experiment, antibiotic concentrations drop continuously, exposing the susceptible strain to sub-MIC concentrations. Sub-MIC concentrations can lead to the evolution and spread of resistance ( 31 , 32 ). Moreover, there is an interplay between resistance and tolerance; indeed, tolerance has already been identified as a stepping stone for resistance ( 29 , 33 , 34 ). We observed a mild increase in the MIC in several of the isolates, albeit with a minor contribution to fitness when compared with tolerance. Nonetheless, in our model, any increase in MIC is beneficial since it reduces the duration of the death phase (in the model, τ M I C ) without any associated cost. To summarize, we have shown that in a synthetic community in which an antibiotic-resistant strain protects a susceptible strain, the evolution of tolerance is restricted by the strength of the protective interaction. This highlights the importance of expanding our knowledge of the intricate interactions between ecology and evolution, especially in the context of antibiotic treatments and their alarming increased failure."
} | 1,674 |
39996083 | PMC11849182 | pmc | 1,025 | {
"abstract": "Introduction The plant restoration and ecological restoration of lead-zinc mines are very important. Methods In this study, we used three local plants to carry out ecological restoration of abandoned lead–zinc mining areas and detected the adaptive mechanisms of soil bacterial diversity and function during the ecological restoration of lead–zinc mines through 16S rRNA sequencing. Results The results revealed that lead-zinc mining significantly reduced the soil bacterial diversity, including the Shannon, Simpson, and observed species indices, whereas the planting of the three ecological restoration plants restored the soil microbial diversity to a certain extent, leading to increases in the Shannon index and Observed species indices. Mining activities significantly reduced the abundances of RB41 and Bryobacter in the bulk soil compared with those in the nonmining areas, whereas the three ecological restoration plants increased the abundances of RB41 and Bryobacter in the rhizosphere soil compared with those in the bulk soil in the mining areas. Following the planting of the three types of ecologically restored plants, the soil bacterial community structure partially recovered. In addition, different plants have been found to have different functions in the lead-zinc ecological restoration process, including iron complex transport system-permitting proteins and ATP binding cassettes. Discussion This study confirms for the first time that plants adapt to the remediation process of abandoned lead-zinc mines by non-randomly assembling rhizosphere bacterial communities and functions, providing a reference for screening microbial remediation bacterial resources and plant microbe joint bioremediation strategies for lead-zinc mines.",
"introduction": "Introduction Lead-zinc ore, a crucial nonferrous metal mineral, is highly important in industry worldwide. Lead–zinc alloys and their derivatives find extensive applications in sectors such as automobiles, batteries, construction, and chemicals, thereby playing a pivotal role in fostering global economic growth ( Li D. et al., 2024 ). However, the mining and smelting of lead-zinc ores have led to substantial environmental challenges ( Chen et al., 2023 ; Chen T. et al., 2022 ). During these processes, considerable amounts of dust and exhaust gases, primarily composed of sulfides and nitrogen oxides, are emitted ( Garry et al., 2018 ; Kan et al., 2021 ). These pollutants not only severely impact the air quality surrounding mining areas but also pose a potential threat to the global atmospheric environment through atmospheric dissemination ( Peng et al., 2023 ). Furthermore, mining and smelting generate wastewater laden with heavy metal ions and acidic substances, particularly lead, zinc, and cadmium ( Guo et al., 2023 ; Ma et al., 2023 ). Direct discharge of this untreated wastewater into water bodies such as rivers and lakes poses risks to aquatic life and threatens human health through the food chain ( Wang et al., 2022 ; Wang et al., 2023 ). Solid waste, including waste rocks and residues from mining, can also contaminate soil ( Junusbekov et al., 2023 ; Sun et al., 2022 ). Prolonged accumulation can degrade soil structure and fertility, adversely affecting crop growth and yield ( Qiao et al., 2022 ). Studies have indicated that lead–zinc mining can systematically affect the surrounding ecosystem, resulting in a decline in biodiversity ( Kastury et al., 2023 ; Larsen et al., 2001 ). Individuals exposed to the mining environment for extended periods are susceptible to heavy metal accumulation and irreversible bodily tissue damage ( Choudhari et al., 2010 ). Therefore, it is imperative to undertake ecological restoration in lead–zinc mining areas and mitigate their ecological risk ( Paluchamy and Mishra, 2022 ; Yang et al., 2023 ). The approaches for the ecological remediation of lead-zinc mine pollution include physical, chemical, and biological methods ( Cai et al., 2021 ; Sha et al., 2023 ). Physical methods involve precipitation, filtration, and adsorption, whereas chemical methods include neutralization, oxidation–reduction, and precipitation, all aimed at reducing lead and zinc concentrations in soil and water ( Jiang S. et al., 2022 ; Lei et al., 2018 ; Luo et al., 2022 ). However, these methods are costly and prone to secondary pollution. Biological methods that leverage the remediation capabilities of plants and microorganisms are cost-effective and environmentally friendly ( Su et al., 2023 ; Diallo et al., 2024 ; Duan et al., 2022 ). Phytoremediation, an in situ technique, avoids secondary ecosystem pollution ( Su et al., 2022 ). By cultivating locally adapted plants, vegetation cover in mining areas can be restored, increasing soil quality ( Li et al., 2018 ). Recently, plant remediation technology has garnered extensive attention and research, with ongoing efforts to screen and cultivate plants with superior remediation capabilities to increase efficiency ( Tang et al., 2019 ; Xiao et al., 2022 ). Additionally, the integration of plant restoration with other techniques, such as plant–microbe combined restoration, has emerged as a research focus ( Xiao et al., 2022 ; Han et al., 2020 ). Through symbiotic relationships, plants and microorganisms efficiently remove pollutants, utilizing the absorption and transformation abilities of plants and the degradation and transformation capabilities of microorganisms ( Chen J. et al., 2022 ; Jiang X. et al., 2022 ; Singh et al., 2022 ). Microorganisms can adsorb harmful substances, removing them from the environment by binding to pollutants through surface viscous substances or extracellular polymeric substances ( Li et al., 2022a ; Li Q. et al., 2024 ). They can also metabolically reduce heavy metal pollutants, converting them into harmless forms ( Dang et al., 2021 ; Bao et al., 2023 ). Moreover, microorganisms secrete growth-promoting substances, regulate plant tolerance and adsorption capacity for heavy metals, and improve soil conditions through long-term coevolution and mutual adaptation with plants ( Jacob et al., 2018 ; Noor et al., 2022 ; Ojuederie and Babalola, 2017 ). Currently, research on in situ phytoremediation of lead–zinc mines is limited, and plant adaptability and remediation capabilities vary across environments. The composition and role of different plant rhizosphere bacterial communities in phytoremediation processes remain unclear ( Tang et al., 2022 ). In this study, we cultivated three lead–zinc accumulator plants— Carex nubigena , Pteris cretica L. var. nervosa , and Neyraudia reynaudiana —in a lead–zinc mining area and adjacent nonmining areas. These plants are known for their ability to accumulate lead and zinc, strong stress resistance, and local adaptability ( Li et al., 2018 ; He et al., 2022 ; He et al., 2023 ; Liu et al., 2020 ). Five years post-planting, the ecological restoration plants demonstrated robust growth. To understand the basis of their efficient environmental adaptability and remediation capabilities, we analyzed changes in the composition and diversity of rhizosphere bacterial communities in mining and nonmining areas via 16S rRNA high-throughput sequencing. Our findings contribute to the screening of plant growth-promoting bacteria and bioremediation strains, providing valuable insights for ecological restoration in lead–zinc mining areas.",
"discussion": "Discussion Abandoned lead–zinc mines can generate waste gas, wastewater, and solid waste ( Cao et al., 2023 ; Fernández-Martínez et al., 2024 ). If left untreated, it may lead to the spread of ecological risks, further causing harmful elements such as heavy metals to harm human and other animal and plant health through the food chain ( Xu et al., 2024 ; Yohannes et al., 2022 ). The characteristics of plant ecological restoration include low energy consumption, low cost, and environmental friendliness ( Hassan et al., 2024 ; Narayanan et al., 2021 ; Thomas et al., 2022 ). Plants can fix or adsorb soil pollutants on their own and cooperate with rhizosphere bacteria, preventing soil erosion and pollutant diffusion in abandoned mining areas ( Doku et al., 2024 ; Duan et al., 2021 ; Xiong et al., 2024 ). In the process of plant ecological restoration, rhizosphere bacteria play crucial roles, including assisting plants in fixing and adsorbing pollutants, enhancing plant stress resistance and environmental adaptability, and promoting plant growth ( Xiao et al., 2022 ; Bennis et al., 2022 ; Xiao et al., 2024 ). In this study, three local plants were selected for the ecological restoration of abandoned lead–zinc mining areas. The results revealed that lead–zinc mining significantly reduced the soil bacterial diversity, whereas after years of plant planting, the soil bacterial diversity was restored to a certain extent, which is consistent with previous research findings ( Deng et al., 2024 ). Research has shown that microbial diversity is closely related to the health level of plants ( Berg et al., 2017 ; Trivedi et al., 2020 ). The greater the microbial diversity is, the greater the ability of the plant to adapt to the environment ( Banerjee and van der Heijden, 2023 ; Cheng et al., 2019 ). The rhizosphere bacterial diversity of Neyraudia reynaudiana and Pteris cretica plants was most significantly restored in the soil of abandoned mining areas, indicating that these two plants can adapt well to the environment of lead–zinc mining and respond to environmental stress by increasing rhizosphere bacterial diversity. In the subsequent ecological restoration process, the planting of these two plants can be increased to improve their role in ecological restoration ( Li et al., 2018 ; He et al., 2022 ). This study revealed that mining activities significantly altered the soil community composition, reducing the abundance of soil RB41 and Bryobacter , whereas the planting of ecologically restored plants significantly increased the abundance of soil RB41 and Bryobacter . RB41 has been found to play an important role in regulating plant health and assisting plants in coping with adverse environmental stress ( Gao et al., 2024 ; Li et al., 2022b ). Bryobacter is an important beneficial bacterium for plants that plays a crucial role in regulating and promoting plant growth ( Contreras et al., 2023 ; Li X. et al., 2022 ; Yang et al., 2022 ). This study confirms for the first time that ecologically restored plants can respond to the environmental stress caused by abandoned lead–zinc mining by enriching beneficial bacterial communities and increasing their environmental adaptability and remediation potential. In addition, mining activities have increased the abundance of Sphingomonas , whereas the planting of ecologically restored plants has reduced the abundance of Sphingomonas . Sphingomonas has also been found to have potential for environmental remediation and the promotion of plant growth ( Asaf et al., 2020 ). These results indicate that plants adapt better to the environment by selectively selecting “matching” bacterial populations through nonrandom selection ( Li et al., 2025 ). RB41 and Thiobacillus were significantly enriched in the rhizosphere soil of plants in mining areas compared with the same type of plant rhizosphere soil from nonmining areas. Thiobacillus has been found to have sulfur oxidation activity and is enriched in the rhizosphere soil of various plants ( Dai et al., 2024 ; Osman et al., 2021 ). These results indicate that RB41 and Thiobacillus are adaptable to ecologically restored plants and have important ecological value in the process of plant ecological restoration. Research has shown that different plants also randomly enrich different microbial communities in the process of ecological restoration, including Blastocatella , Pseudonocardia , and Dongia . Blastocatella has been found to have strong heavy metal tolerance ( Guo et al., 2017 ; Li et al., 2021 ), Pseudonocardia has extensive antibacterial and fungal activity ( Riahi et al., 2022 ), and Dongia has strong environmental adaptability and has been detected in various environments ( Jiang et al., 2024 ; Lu et al., 2022 ). This study is the first to analyze how different plants adapt to the environment of lead–zinc mining by assembling rhizosphere bacterial communities in both common and specific ways. In the subsequent ecological restoration process of lead–zinc mines, we can specifically screen bacterial resources that are suitable for plants for plant bacterial joint ecological restoration. Beta diversity analysis revealed that mining activities had a significant effect on the soil community structure, leading to significant differences. After plants were subjected to ecological restoration, the soil bacterial community structure recovered to a certain degree and was similar to the community structure of the control sample. PCA based on the prediction of bacterial community function also revealed the same phenomenon; that is, mining activities significantly affect the ecological function of bacterial communities, and the planting of ecologically restored plants has a positive effect on the restoration of community function, which is consistent with previous research results ( Deng et al., 2024 ). In addition, we also found that different plants adapt to ecological restoration environments by enriching different functions, including iron complex transport system-permitting proteins and transcription pair coupling factors. The iron complex transport system-permitting protein is related to ion transport ( Raymond et al., 2015 ), and the transcription pair coupling factor plays an important role in bacterial transcriptional regulation ( Mistry et al., 2023 ). This study reveals for the first time that ecologically restored plants adapt to the environment and successfully complete ecological restoration through nonrandom community assembly and functional changes. These research results provide a reference for screening ecological restoration bacterial resources and developing plant bacterial joint ecological restoration strategies."
} | 3,547 |
36927612 | PMC7616385 | pmc | 1,026 | {
"abstract": "Maintaining cohesion between randomly moving agents in unbounded space is an essential functionality for many real-world applications requiring distributed multi-agent systems. We develop a bio-inspired collective movement model in 1D unbounded space to ensure such functionality. Using an internal agent belief to estimate the mesoscopic state of the system, agent motion is coupled to a dynamically self-generated social ranking variable. This coupling between social information and individual movement is exploited to induce spatial self-sorting and produces an adaptive, group-relative coordinate system that stabilises random motion in unbounded space. We investigate the state-space of the model in terms of its key control parameters and find two separate regimes for the system to attain dynamical cohesive states, including a Partial Sensing regime in which the system self-selects nearest-neighbour distances so as to ensure a near-constant mean number of sensed neighbours. Overall, our approach constitutes a novel theoretical development in models of collective movement, as it considers agents who make decisions based on internal representations of their social environment that explicitly take into account spatial variation in a dynamic internal variable.",
"conclusion": "5 Conclusion The incorporation of sociogenesis into models of collective motion using internal agent beliefs gives insight into how social structure can influence, and be influenced by, the motion decisions of individuals. The model presented achieves cohesive collective motion in unbounded space with sensory radii spanning 12.5% of the initial group spread ( k h = 0.125). The success of this model suggests that biasing agent motion using internal variables prevents diffusion in unbounded space, an important problem to solve for future real-world deployment of multi-agent and swarm robotic systems [ 7 , 20 ]. The approach of modelling intelligent agents that use statistical inference methods to locally estimate and adaptively respond to mesoscopic system states opens up avenues for understanding agents who base their decisions on incomplete system information.",
"introduction": "1 Introduction Systems exhibiting self-organised collective motion and spatial sorting are widespread across scales in nature [ 25 ], from flocking [ 22 , 6 ] and separation with respect to social or physical characteristics [ 14 , 21 , 1 ] to territorial segregation [ 19 , 15 ]. Of the existing models of collective movement, little attention has been given to the role of sociogenesis [ 26 ], the theory of how socio-spatial structures form as a result of agent interactions. In nature, however, there is a strong interplay between social dynamics, movement and space-use behaviour. Annular socio-spatial patterns have been widely observed in primate dominance hierarchies, where high ranking individuals occupy central locations in the group, while low ranking ones are found at the peripheries [ 18 , 14 ]. While existing spatially explicit models of dominance hierarchy formation [ 9 , 12 , 8 ] generate correlations between social ranks and spatial centrality, they are restricted to agents that are reactive to social information at the microscopic scale. The work in [ 9 ] analysed the density-driven phase transition for the emergence of condensed clusters of agents, but did not study the mechanisms with which microscopic scale dynamics lead to cluster formation. Work in [ 12 , 8 ] investigated how heterogeneity in agent repulsion generated by micro-scale collisions leads to spatial sorting. A common characteristic of these models is the imposition of periodic boundary conditions to maintain group cohesion. This is not surprising given that, in unbounded domains, cohesive collective random motion is known to require direct [ 24 , 6 , 16 , 29 ] or effective [ 17 ] attractive interactions that share information at long range. Here, we present a sociogenesis framework for self-organised cohesive collective motion in unbounded space. We posit that loss of cohesion occurs due to instabilities arising from agents who base their movement decisions on noisy motion-dependent information at the microscopic scale. As such, we model agents whose behaviour is determined by their belief state - a coarse-grained representation of the local environment. The belief enables agents to respond to socio-spatial information at the mesoscopic scale , which fluctuates at a slower time-scale than the microscopic states it encodes. Importantly, agent beliefs in our model represent spatial variations in internal variables that are not related to motion. In view of our modelling approach, it is noted that while existing models of flocking [ 24 , 22 , 6 ] also leverage mesoscopic-level social interactions to achieve global collective patterns, internal representations have only made use of information corresponding to agent states that are position and motion-related. As a result, the functional form of the agent beliefs used in these models has been restricted to taking an average of positions and velocities of neighbouring agents. Models of spatial coordination in multi-agent systems (MAS) and reinforcement learning (RL) have similarly focused on using internal representations of spatial information, computing Voronoi area partitions [ 5 ] and distributed path selection [ 4 ] based on neighbour positions that are directly sensed or propagated through message passing. Other approaches have focused on coordinating agent motion by optimising coarse-grained, position-related metrics over the agent communication network, such as the number of network connections with neighbours [ 28 ], or using shared neighbour information [ 27 ]. The model presented here offers a generalisation of the mesoscopic coarse-graining procedure by extending the functional form of agent beliefs beyond taking an average. This development is necessary for enabling agents to coordinate using dynamic internal variables that are not motion-related. Since coarse-graining via a simple average leads to a loss of spatial information, coordination of agents using a dynamic internal variable requires a belief which takes into account spatial variation. In this model, a spatially explicit regression model is used for coarse-graining. The form of this generalisation allows for further extension beyond regression to other machine learning approaches. In fact, this generalisability is essential to the functionality of our model, since different functional forms of the belief generate different global socio-spatial configurations. Here, the beliefs are constrained to induce self-organisation into a desired global concave annular state. We consider a system in 1D space, and use the term radial to refer to the line segment extending from the central point to the outer bounds of a given 1D interval, meaning that the desired global state is radially sorted in the social variable."
} | 1,738 |
40332474 | PMC12058968 | pmc | 1,027 | {
"abstract": "Abstract Collective behavior is all around us, from flocks of birds to schools of fish. These systems are immensely complex, which makes it pertinent to study their behavior through minimal models. We introduce such a minimal model for cohesive and aligning self-propelled particles in which group cohesion is established through additive, non-reciprocal torques. These torques cause a particle’s orientation vector to turn toward its neighbor so that it aligns with the separation vector. We additionally incorporate an alignment torque, which competes with the cohesive torque in the same spatial range. By changing the strength and range of these torque interactions, we uncover six states which we distinguish via their static and dynamic properties: a disperse state, a multiple worm state, a line state, a persistent worm state, a rotary worm state, and an aster state. Their occurrence strongly depends on initial conditions and stochasticity, so the model exhibits multistabilities. A number of the states exhibit collective dynamics which are reminiscent of those seen in nature. Graphic abstract) \n \n Supplementary Information The online version contains supplementary material available at 10.1140/epje/s10189-025-00482-7.",
"conclusion": "Conclusions and outlook We have explored the emergent structural and dynamic properties of particles that interact via alignment and non-reciprocal cohesive torques. Depending on their radius of interaction and the strength of their interaction torques, particles exhibit different types of collective behavior, ranging from multiple, short-lived worms in a disperse background to single, long-lived worms, which can exhibit either rotary or persistent dynamics to closely packed asters. By analyzing both static and dynamic properties of the observed collective behaviors, we classified six distinct possible states for this model. The realized state for a given set of parameters is, in many cases, multistable and depends on stochastics and initial conditions. The persistent and rotary worm states that emerge from this model are particularly emblematic of collective behavior because, in these states, particles both move in the same direction and stay together in a single group. Furthermore, these states exhibit continuously changing neighbors and leadership, both of which are observed in animal groups such as flocks of birds [ 48 , 49 ]. Qualitatively, similar worm-like structures can be observed in many examples of collective behavior seen in nature [ 9 , 50 , 51 ]. In future work, it would be interesting to quantitatively compare the dynamics and structures exhibited by this model to those exhibited by specific systems in nature. In addition, this model could also serve as a template for engineering collective behavior in microrobotic systems [ 52 , 53 ], which is desirable for applications in fields such as biomedicine. Our model is particularly conducive to this application because the multistable nature of many states should allow for reconfigurability. We note that our model does not include hydrodynamic flow fields, which are necessary to describe the dynamics of biological microswimmers. Ref. [ 54 ] has recently shown that combining hydrodynamics with Vicsek-like alignment interactions results in a variety of collective behaviors, which differ from the dry case because global polar order does not emerge. The reason is that hydrodynamic interactions between pusher or puller swimmers destroy such a global polar order [ 55 – 57 ]. Whether the worm states observed in our model, which exhibit global polar order, persist when hydrodynamic flow fields are added, certainly also depends on the strength of the torque interactions. This is an interesting course of future work to be pursued based on our previous work in Ref. [ 57 ]. Worm-like structures similar to our persistent worm state have previously been observed for the collective behavior models devised by Couzin et al. [ 16 , 17 ] and Negi et al. [ 37 ]. However, to our awareness, a state analogous to our rotary worm state has not previously been observed. Additionally, Negi et al. analyzed the behavior of their persistent worm-like structure through the lens of railway motion performed by active polymers [ 47 ]. Such a lens is insufficient to describe the worm state exhibited by our model, due to dynamics within the worm itself. A further contrast to the models of Couzin et al. and Negi et al. is that we observe multistabilities in many of our states even when the particles are initialized randomly. Such multistabilites were not observed in these previous models. In this work, we have focused on varying the strength of the torque interaction and the radius of this interaction, which applies to both the alignment and cohesive torques. However, we have always kept the ratio between the alignment and cohesion torques constant. In future work, it would be interesting to vary this ratio as well as to use different spatial ranges for the alignment and cohesive torques. This would enable further exploration of the analogy between the observed collective dynamics in the model and similar emergent behavior in nature. It would also broaden the scope of interactions which could be used to engineer swarms of microrobots for applications in biomedical, among other, fields.",
"introduction": "Introduction Collective behavior is omnipresent in our lives, from the microscale, with examples such as bacterial colonies [ 1 – 4 ] and morphogenesis [ 5 – 7 ], to the macroscale, with examples such as flocks of birds [ 8 – 11 ] and schools of fish [ 12 – 14 ]. Although systems exhibiting collective behavior are commonplace, they are not well understood due to their immense complexity. In order to improve our understanding of these complex systems, minimal models are often used. Such models attempt to qualitatively replicate collective behavior with only a few simple rules of interaction and thus help to isolate the fundamental aspects governing collective behavior. One of the earliest models for collective behavior is the Boid model [ 15 ]. In this model, individuals known as ‘boids’ are subject to interaction rules which promote collision avoidance, alignment, and cohesion. Most subsequent collective behavior models use some variation of these three fundamental rules, with the possible addition of further interaction rules. For example, in the behavioral zonal model [ 16 , 17 ], these three rules are implemented in different spatial areas around each particle. Depending on the parameters of the different interactions, this model can exhibit swarming, milling, or parallel group motion. Although the aforementioned models use both alignment and cohesion to induce collective behavior, it is also possible to achieve collective motion using only alignment [ 18 – 20 ] or cohesion [ 21 , 22 ] interactions. Most famously, in the Vicsek model [ 18 ], particles move together simply by adapting their orientation to align with those of their neighbors. The Vicsek model does not, however, lead to the formation of groups which stay together. In order to achieve this, the alignment interaction must be supplemented by some additional element, such as cohesion interactions [ 23 ]. Oftentimes, in systems where collective behavior is observed, interactions among constituents are non-reciprocal, meaning that Newton’s third law is violated and action-reaction symmetry does not hold [ 24 ]. Indeed, a diverse range of collective behaviors have been seen for systems in which non-reciprocal interactions are present [ 25 – 34 ]. Non-reciprocity can be introduced to a system in many different ways including non-reciprocal force interactions [ 35 , 36 ], interactions in which a vision cone is used [ 16 , 17 , 22 , 37 ], and non-reciprocal torque interactions [ 28 – 31 ]. Here, we specifically focus on non-reciprocity via torque interactions. Zhang et al. [ 31 ] showed that effectively attractive non-reciprocal torques cause active phase separation, which has also been seen for the model of Nilsson and Volpe [ 30 ]. There, the relative position between a particle and its neighbor determines whether the torque acts attractive or repulsive. Models with effectively repulsive non-reciprocal torques [ 28 , 29 ] can exhibit flocking behavior or active phase separation, depending on the exact parameters used. Here, we introduce a model for cohesive and aligning self-propelled particles in which group cohesion is established through additive, non-reciprocal torques. This model combines elements from the models of Couzin et al. [ 16 , 17 ] and Negi et al. [ 37 ]. We use the additive cohesion and alignment torques of Couzin et al. in combination with the hard-core potentials of Negi et al. Unlike both of these models, we do not introduce a vision cone, but rather maintain a full angular vision range. Also in contrast to these models, we keep the interaction range the same for both cohesive and aligning torques. In this paper, we explore the different states which emerge from this model when varying the interaction radius and the strength of the torque interactions. We uncover a rich variety of states: a disperse state, a multiple worm state, a line state, a persistent worm state, a rotary worm state, and an aster state. Although the changes to the previous models of Couzin et al. and Negi et al. are subtle, these changes introduce new dynamics including in the formation of the different states. In particular, we observe multistabilities across many states, a seemingly generic trait which was not observed or discussed in either of the aforementioned models. We note that the model of Couzin et al. exhibits hysteresis; however, it does not exhibit multistabilities emerging from random initializations, as we observe in our model. Furthermore, the rotary worm state exhibited by our model has not been seen in previous models. We begin by introducing our model in Sect. 2 . In Sects. 3.1 and 3.2 , we classify the different observed collective behaviors into distinct states via their static and dynamic properties. We then analyze the resultant state diagram for our chosen range of parameters in Sect. 3.3 . In Sect. 4 , we explore in greater depth the persistent and rotary worm states by analyzing their structural properties as well as the dynamics of individual constituents. We summarize and conclude in Sect. 5 ."
} | 2,599 |
25640575 | PMC4313099 | pmc | 1,028 | {
"abstract": "Understanding the architectural subtleties of ecological networks, believed to confer them enhanced stability and robustness, is a subject of outmost relevance. Mutualistic interactions have been profusely studied and their corresponding bipartite networks, such as plant-pollinator networks, have been reported to exhibit a characteristic “nested” structure. Assessing the importance of any given species in mutualistic networks is a key task when evaluating extinction risks and possible cascade effects. Inspired in a recently introduced algorithm –similar in spirit to Google's PageRank but with a built-in non-linearity– here we propose a method which –by exploiting their nested architecture– allows us to derive a sound ranking of species importance in mutualistic networks. This method clearly outperforms other existing ranking schemes and can become very useful for ecosystem management and biodiversity preservation, where decisions on what aspects of ecosystems to explicitly protect need to be made.",
"discussion": "Discussion In this paper we have presented a novel framework to asses the relative importance of species in mutualistic networks. Inspired by a recent work on economics/econometrics we employ an algorithm, similar in spirit to Google's PageRank but of non-linear character, that we have named MusRank. The algorithm provides two complementary rankings: one for active species (such as insects, birds, fish,…) in terms of their importance and one for passive species (plants and their seeds, anemone, etc) in terms of their vulnerability. We also propose a criterion to assess the quality of any given ranking of species: good rankings lead to a fast break-down of the corresponding mutualistic network when species are progressively removed in decreasing ranking order. In most of the empirical mutualistic networks we have analyzed the use of our novel framework rendered a ranking which clearly outperforms all the alternative ones used as workbench. Results are robust in the sense that different implementations lead to similar rankings. In many cases, the resulting ordering coincides or is very close to the optimal one as found by a -computationally very costly- genetic algorithm. Moreover, MusRank is much faster and finds excellent rankings even for large mutualistic networks for which the genetic algorithm is not able to find optimal solutions in a reasonable computational time. Therefore, the emerging ranking allows for assessing the importance of individual species within the whole system in a meaningful, efficient and robust way. We conclude that rankings of species importance in mutualistic networks should be constructed employing MusRank. Furthermore, as a by-product, the excellent packing of nested matrices provided by this non-linear approach (see Figure 5 ) calls for a redefinition of the way in which nestedness is measured. In particular we suggest that nestedness calculators should use the ranking provided by the present algorithm, which clearly outperforms others in making the nested architecture evident. Indeed, we believe that the nested structure of mutualistic networks is essential for the success of MusRank; it remains to be seen what is the performance of this scheme for bipartite networks without a nested architecture. The novel approach –introduced here for the first time in the context of mutualistic ecological networks– may prove of practical use for ecosystem management and biodiversity preservation, where decisions on what aspects of ecosystems to explicitly protect need to be made."
} | 888 |
29670942 | PMC5903897 | pmc | 1,030 | {
"abstract": "A hydrophilic directional slippery rough surface outperforms conventional liquid-repellent surfaces in water harvesting.",
"introduction": "INTRODUCTION Engineered surfaces with exceptional droplet nucleation and removal ( 1 – 5 ) are of great interest in various energy and water applications, including power generation ( 6 ), thermal management ( 3 , 7 , 8 ), water harvesting ( 2 ), and desalination ( 9 ). However, surfaces that have both exceptional droplet nucleation and removal capabilities are rare. For example, lotus leaf–inspired superhydrophobic surface(s) (SHS) ( 10 – 12 ) can effectively repel water droplets through air-infused surface textures and hydrophobic surface chemistry, but the trapped air can be easily damaged under pressure or high-humidity conditions. Once the trapped air is damaged, the liquids will be in full contact with the surface textures and become highly pinned ( 13 , 14 ), that is, the droplets will be in the Wenzel state ( 15 ). Advanced SHS designs have shown that it is possible to revert the Wenzel state droplets through the jumping droplet departure mechanism ( 14 , 16 , 17 ). Under a large subcooling condition, however, the formation of Wenzel state droplets becomes unavoidable because droplets nucleate within the surface textures. This greatly reduces droplet mobility and leads to surface flooding ( 16 ). As an alternative strategy to resolve the droplet mobility issue of the SHS, new types of liquid-repellent surfaces modeled after the slippery rim of the Nepenthes pitcher plant have been developed. These surfaces, known as slippery liquid-infused porous surface(s) (SLIPS) ( 18 , 19 ), are characterized by a homogeneous and molecularly smooth interface made by fully infusing a textured surface with a hydrophobic liquid lubricant. These surfaces exhibit excellent droplet removal capabilities but suffer from relatively small surface area for droplet nucleation ( 7 , 8 , 20 ). It is energetically more favorable for water vapor and small water droplets to nucleate on hydrophilic surfaces compared to hydrophobic ones ( 2 , 21 – 23 ). The Namib desert beetle, for example, uses patterned hydrophilic surfaces to nucleate and capture tiny water droplets in air for their survival ( 2 ). An ideal surface that sustains efficient droplet nucleation and removal must use hydrophilic surface chemistry and enhanced surface area to maximize droplet nucleation density, as well as must have a pinning-free slippery interface to enhance droplet mobility. Here, we report the design and applications of a bioinspired slippery rough surface(s) (SRS), which combines the salient features of pitcher plants and rice leaves; our surface has hydrophilic surface chemistry, a slippery interface, directional structures, and large surface areas to maximize droplet collection and mobility ( Fig. 1 ). Specifically, our SRS consists of liquid-infused nanotextures on directional microgrooves for enhanced droplet nucleation and transport. The presence of the nanoscale textures on the microgrooves helps retain a thin layer of hydrophilic lubricant to provide a slippery interface for droplet removal. Unlike the beetle-inspired surfaces where droplet nucleation only occurs in the patterned hydrophilic areas ( 22 , 23 ), our surfaces fully use the total surface area for maximized droplet nucleation. Such a surface allows nucleated droplets in the Wenzel state to be removed rapidly, which cannot be achieved by conventional rough surfaces. Here, we have demonstrated that the hydrophilic directional SRS outperforms many of the state-of-the art hydrophobic liquid-repellent surfaces such as SHS and SLIPS in dropwise condensation and fog harvesting owing to the large droplet nucleation density, and fast droplet coalescence and removal. Fig. 1 Hydrophilic directional SRS inspired by pitcher plants and rice leaves. Side view ( top left ) and three-dimensional view ( top right ) of the hydrophilic directional SRS. Photos and schematics showing the pitcher plant–inspired slippery surface ( bottom left ) and rice leaf–inspired directional structured surface ( bottom right ). The photography of rice leaf was reprinted with permission from Bixler and Bhushan ( 39 ).",
"discussion": "DISCUSSION In summary, we have demonstrated that hydrophilic directional SRS can outperform their hydrophobic counterparts and nonlubricated SHS in both droplet nucleation and mobility during dropwise condensation and fog harvesting processes. Many of these water condensates have traditionally pinned to the surface textures—a result of an irreversible transition from the Cassie state to the “sticky” Wenzel state. Our hydrophilic directional SRS resolves this long-standing issue by combining the unique surface functions of pitcher plants (that is, slippery interface) and rice leaves (that is, directional structures and micro/nano hierarchical structures), which can repel liquids regardless of how they wet the surfaces. Note that the longevity and robustness of the hydrophilic directional SRS can be further optimized by engineering the surface textures, tuning the lubricant viscosity, or using lubricant-infused polymers accordingly depending on specific applications (fig. S6). Our work not only demonstrates the first examples of applications using Wenzel state droplets but also shows the possibility of creating a hydrophilic coating with excellent droplet nucleation and water-repellent functions, which may contribute to a broad range of applications related to water and energy, such as antifouling for marine ships, dropwise condensation for power generation and desalination, and water harvesting in arid regions."
} | 1,412 |
32335484 | null | s2 | 1,032 | {
"abstract": "The superhydrophobic lotus leaf has dual-scale surface structures, that is, nano-bumps on micro-mountains. Large hydrophilic particles, due to its high surface energy and weight, have high affility to substrates and tend to precipitate at the bottom of coating films. Small hydrophobic particles, due to its low surface energy and weight, tends to sit on the top of coating films and form porous structures. To mimic the lotus leaf surface, it may be possible to develop dual-sized particle films, in which small particles are decorated on large particles. A one-step spin coating of a mixture of dual-sized silica particles (55/200 nm) was used. Epoxy resin was added to improve the adhesion of particle films. The single-sized and dual-sized particle films were compared. The mechanical robustness of particle films was tested by tape peeling and droplet impact. The novel combination of hydrophobic silica (55 nm) and hydrophilic silica (200 nm) is essential in creating the hierarchical structures. By combining the strong adhesion of hydrophilic silica (bottom of coating film) to polymer substrates and porous structures of hydrophobic silica (top of coating film), we first time report a one-step and versatile approach to create uniform, transparent, robust, and superhydrophobic surface."
} | 324 |
32709258 | PMC7382037 | pmc | 1,033 | {
"abstract": "Background Current understanding of the carbon cycle in methanogenic environments involves trophic interactions such as interspecies H 2 transfer between organotrophs and methanogens. However, many metabolic processes are thermodynamically sensitive to H 2 accumulation and can be inhibited by H 2 produced from co-occurring metabolisms. Strategies for driving thermodynamically competing metabolisms in methanogenic environments remain unexplored. Results To uncover how anaerobes combat this H 2 conflict in situ, we employ metagenomics and metatranscriptomics to revisit a model ecosystem that has inspired many foundational discoveries in anaerobic ecology—methanogenic bioreactors. Through analysis of 17 anaerobic digesters, we recovered 1343 high-quality metagenome-assembled genomes and corresponding gene expression profiles for uncultured lineages spanning 66 phyla and reconstructed their metabolic capacities. We discovered that diverse uncultured populations can drive H 2 -sensitive metabolisms through (i) metabolic coupling with concurrent H 2 -tolerant catabolism, (ii) forgoing H 2 generation in favor of interspecies transfer of formate and electrons (cytochrome- and pili-mediated) to avoid thermodynamic conflict, and (iii) integration of low-concentration O 2 metabolism as an ancillary thermodynamics-enhancing electron sink. Archaeal populations support these processes through unique methanogenic metabolisms—highly favorable H 2 oxidation driven by methyl-reducing methanogenesis and tripartite uptake of formate, electrons, and acetate. Conclusion Integration of omics and eco-thermodynamics revealed overlooked behavior and interactions of uncultured organisms, including coupling favorable and unfavorable metabolisms, shifting from H 2 to formate transfer, respiring low-concentration O 2 , performing direct interspecies electron transfer, and interacting with high H 2 -affinity methanogenesis. These findings shed light on how microorganisms overcome a critical obstacle in methanogenic carbon cycles we had hitherto disregarded and provide foundational insight into anaerobic microbial ecology. Video Abstract",
"conclusion": "Conclusion In methanogenic ecosystems, degradation of organic matter generates H 2 as a central byproduct and necessitates microbial interactions between H 2 -generating organotrophic bacteria and H 2 -consuming methanogenic archaea. However, organotrophic metabolisms have diverse thermodynamic properties and many processes (i.e., HT catabolism) can generate H 2 concentrations much beyond the thermodynamic limit of others (i.e., HS catabolism), which has not been addressed in previous models. Through metagenomic and metatranscriptomic analyses of multiple anaerobic digesters, we predict that uncultured organisms may employ unique strategies to drive thermodynamically competing metabolisms (Fig. 5 )—parallel and broad-range HS and HT metabolism, a shift (often complete) from H 2 to formate as a soluble electron carrier, respiration of low-concentration O 2 , DIET and formate exchange with Methanothrix , and interaction with high H 2 -affinity methanogenesis by Ca. Methanofastidiosum. The observed metabolic behaviors are likely tailored to the thermodynamic conditions in situ and quite distinct from cultured organisms. With such omics-based insights, future cultivation-based studies can be designed to verify and further characterize organisms that perform thermodynamically challenging catabolism under the in situ selective pressures (e.g., enrichment/cultivation of syntrophic degraders in the presence of both methanogens and H 2 -producing fermenters). The newly discovered metabolic strategies and ecology driving organic matter mineralization improve our understanding of carbon cycling in methanogenic ecosystems and foundational knowledge for innovation in biotechnology.\n Fig. 5 Updated scheme of methanogenic organic matter mineralization. Known and novel metabolic interactions and behaviors are shown (black and orange arrows, respectively). For each ecological niche, representative phyla and total number of species (number) associated with these phyla are shown. Ecological niches involving lineages uncultured at the family level or higher are indicated (bold with a gray background). Cascading degradation of polymers to monomers (sugars—red, H 2 tolerant AAs—purple, H 2 sensitive AAs—blue) generates metabolic intermediates whose degradation is H 2 -tolerant or H 2 -independent (black) and H 2 -sensitive (green letters). Novel electron transfer and syntrophic interactions involve formate as a key intermediate (green background) and DIET-mediated electric interactions (yellow arrows facing outward for electrogens and yellow arrows facing inward for electron-consuming species). Abbreviations: Bacteroidota (Bactero), Verrucomicrobiota (Verruco), Fermentibacterota (Ferm), Marinisomatota (Marini), Desulfobacterota (Desulfo), Spirochaetota (Spiro), Halobacterota (Halobac), Euryarchaeota (Euryarc), formate (Fo), acetate (Ac), propionate (PR), butyrate (BT), isobutyrate (IB), isovalerate (IV)",
"discussion": "Results and discussion Metagenomic and metabolic reconstruction Metagenomics analysis of 17 full-scale anaerobic digesters treating wastewater sludge yielded 1343 metagenome-assembled genomes (MAGs) that meet the quality criteria previously proposed (CheckM-estimated completeness—contamination > 50 [ 34 , 35 ]; for Ca. Patescibacteria ≥ 60% and ≤ 5% was used given the inherently low estimated completeness for members of this phylum). These MAGs spanned 66 phyla, as predicted by GTDBtk [ 35 ] (Fig. 2 and S2 , Tables S2 and S3 ). The MAGs retained had estimated completeness and contamination ≥ 85% and ≤ 7.5%, respectively (as predicted by CheckM), except for MAGs affiliated with Ca. Patescibacteria (≥ 60% and ≤ 5%). Out of the obtained MAGs, only 181 were assignable to cultured genera and the remaining belonged to various uncultured genus- (289 MAGs), family- (303), order- (199), class- (110), and phylum-level lineages (261) (Fig. 2 ). MAGs were clustered into 896 species using Mash [ 36 ], based on a pairwise mutation distance of ≤ 0.05 (or ≥ 95% average nucleotide identity), which roughly equates to a 70% DNA-DNA reassociation value that has been proposed as a genome-based species definition [ 37 ]. Based on metatranscriptomes recovered from 9 full-scale anaerobic digesters (all in triplicate; Tables S2 and S3), 176 bacterial species and 16 archaeal species each had transcripts representing ≥ 0.4% and ≥ 0.3% of the mapped transcriptomic reads, respectively, in at least one digester (Table S4 and S5 ). These species with high relative activity are herein referred to as “active” species. Much of the remaining species belong to taxa associated with residual populations (primarily non- Deltaproteobacteria Proteobacteria classes and non -Bacteroidia Bacteroidetes classes) [ 38 , 39 ] carried in as waste from upstream aerobic bioprocesses (Fig. S3 ). Thus, species associated with the above taxa were excluded from following analyses. The “active” species spanned 20 cultured and 11 uncultured phyla, of which Bacteroidota (previously known as Bacteroidetes ), Desulfobacterota ( Deltaproteobacteria ), Firmicutes_A ( Firmicutes ), Spirochaetota ( Spirochaetes ), and Halobacterota ( Euryarchaeota ) were most frequently detected (Fig. 2 ). These taxonomic groups and their nomenclature follow the genome-based phylogeny recently proposed [ 35 ]. Among these, 91.5% and 41.2% of the active bacterial and archaeal species, respectively, belonged to uncultured lineages, clearly implying large knowledge gaps in how bacteria and archaea mineralize organics in situ (Table S4 ).\n Fig. 2 Phylogenetic distribution of MAGs recovered and species (MAG clusters) associated with a high metatranscriptome-based activity. The phylogenetic classification was determined using GTDBtk (left). The number of MAGs associated with a cultured genus or uncultured lineages (at different taxonomic levels) is shown (right). Bacterial and archaeal species respectively associated with metatranscriptome-based activities ≥ 0.4% or ≥ 0.3% of the mapped transcriptomes in at least one reactor are shown To accurately reconstruct the metabolic behavior of individual species, we annotate metabolic pathways with strict criteria by taking advantage of the thermodynamic and energetic restrictions of anaerobic life. Due to the cavernous gap in electron acceptor redox potentials (e.g., O 2 to H 2 O [E°’ of 1.23 V] vs H + to H 2 [E°’ of − 0.42 V]), aerobic degradation is highly exergonic and massive energy recovery occurs from O 2 reduction (e.g . , ~ 32 ATP per glucose), while anaerobic degradation is much more thermodynamically limited and often requires energy investment for disposing electrons. Moreover, certain anaerobic metabolisms can become endergonic (i.e . , ∆G > 0 kJ mol -1 ) with only slight byproduct (e.g . , H 2 ) accumulation and require intimate cross-feeding with byproduct-consuming partners, a symbiosis known as syntrophy [ 3 ]. To thrive at this thermodynamic edge of life, anaerobes must employ unique metabolic strategies for coupling substrate oxidation with electron disposal and optimizing energy input and recovery during this process [ 40 – 42 ]. Pioneering efforts in isolating and characterizing syntrophic metabolizers and their enzymes was paramount for obtaining this foundational knowledge [ 4 , 5 , 43 – 45 ]. Capitalizing on these unique insights into syntrophic metabolism, we identified for each active species metabolic pathways that (i) have electron transfer enzymes that account for all predicted oxidative and reductive reactions, (ii) provided net positive energy conservation either by ATP synthesis and/or by the generation of an ion motive force, (iii) are exergonic in situ, and (iv) have all necessary genes highly expressed (see details in methods section; Fig. S4 as an example; and supplementary tables for a summary of capacities [Table S4 ], summary of H 2 /formate-generating electron transfer capacities [Table S6 ], summary of metabolic behavior across digesters [Table S7 – S9 ], metabolic behavior in individual reactors [Table S10–S18; and Table S19 for all collected together]). The maximum metabolic capacity observed within a species cluster and the total metabolic capacity of that cluster were similar (Table S4 and Fig. S5 ), suggesting consistent ecological roles across digesters. Metabolic reconstruction and metatranscriptome mapping of the 192 species clusters revealed that the phyla contributing most to degradation of polymers (i.e . , expressing multiple extracellular proteases, glycosyl hydrolases, and lipases) were Bacteroidota (42 families on average), Verrucomicrobiota (49), Planctomycetota (38), Acidobacteriota (44), and Marinisomatota (68) (Table S20 ). Phyla contributing most to the degradation of monomers (i.e., expressing multiple sugar and AA degradation pathways) were Bacteroidota (3.6 and 6.9 types of sugar and AA degradation pathways, respectively, on average), Firmicutes_E (0 and 9), Thermotogota (3.3 and 7.6), KSB1 (0 and 9), and Marinisomatota (6 and 8) (Fig. 3 a). Most other phyla also contributed polymer hydrolysis and the subsequent degradation of sugars and AAs but expressed fewer polymer- and monomer-degrading pathways (≤ 36 hydrolase families and ≤ 7 pathways for sugar and AA degradation). The above metabolisms generate FA byproducts such as acetate, propionate, butyrate, isobutyrate, 2-methylbutyrate, and isovalerate whose degradation is highly thermodynamically challenging [ 17 , 19 , 46 ]. Of the 29 active bacterial phyla, only three expressed genes for the oxidation of these FA—Desulfobacterota (12 species), Spirochaetota (3), and Thermotogota (3). To accurately predict capacities to degrade different FAs, each FA degradation pathway (methylmalonyl-CoA pathway [propionate], beta-oxidation [butyrate], mutase + beta-oxidation [isobutyrate], carboxylation + lyase + beta-oxidation [isovalerate]), and hydrogenases (e.g . , FeFe and bidirectional NiFe hydrogenases), formate dehydrogenases (e.g . , Fdh-H and Fdh-N type), electron transfer modules (e.g . , Rnf), and energy conservation that complement each other and allow net energy recovery (e.g . , transcarboxylation for propionate and ETF dehydrogenase for C4 and C5 FAs) were identified. For Desulfobacterota, uncultured members of a Syntrophales family contributed to butyrate, isobutyrate, and isovalerate degradation; the Desulfomonalia order contributed to butyrate and isobutyrate degradation, and a Syntrophobacteraceae genus contributed to propionate degradation. Many Desulfobacterota species also concurrently expressed genes for the degradation of multiple FAs (up to three substrates), a feature that has been unobserved and untested in anaerobic FA-degrading isolates [ 19 ]. Members of an uncultured Spirochaetota class expressed genes for butyrate and isobutyrate degradation. Thermotogota species belonging to an uncultured Thermotogae order was predicted to perform acetate degradation. Members of Halobacterota (i.e., Methanothrix , also formerly known as Methanosaeta ) and other Halobacterota/Euryarchaeota (e.g., Methanoculleus , Methanospirillum , Methanothermobacter , and uncultured Methanomicrobiaceae ), respectively, contribute to the degradation of terminal end products, acetate, H 2 , and formate.\n Fig. 3 Phylum-level overall metabolic activities, the thermodynamics-based H 2 thresholds of the activities, and expression of individual pathways. a For each phylum, the number of species clusters, average number of protease, glycosyl hydrolase, and lipase families expressed across species are shown (normalized to maximum observed average among phyla). Likewise, the average number of sugar-, AA-, and FA-degradation pathways expressed across species is shown. AA degradation pathways are split into those that are H 2 -tolerant (HT) and H 2 -sensitive (HS) based on panel b. b The maximum H 2 concentration that each degradation pathway can tolerate is shown (i.e., ∆G reaction + x *∆G ATPsynthesis = 0, where x is the amount of ATP synthesized per substrate degraded). The ATP yield for each pathway was based on the sum of (i) the ATP consumption/generation in the main carbon transformation pathway and (ii) vectorial H + translocation associated with membrane-based electron transfer (e.g., Rnf, Hyb, Fdn), assuming the shortest electron flow route from substrate oxidation to H 2 /formate generation that involves electron bifurcation and reverse electron transport where possible; all of this was based on pathways that were observed to be expressed in this study. Reactions that would either lose much energy as heat (e.g., cytosolic Fd red -oxidizing H 2 generation) or require energy input under in situ conditions (e.g., cytosolic NADH-oxidizing H 2 generation) were not considered. For substrates whose degradation proceeds through pyruvate or acetyl-CoA, maximum H 2 concentrations for oxidation to acetate are shown (see Supplementary Table S1 for a list of reactions). Note that fermentation pathways (e.g., acetyl-CoA reduction to butyrate) would increase the maximum H 2 but reduce ATP yield. The Gibbs free energy yield at standard conditions and pH 7 (∆G°’) and estimated ATP yields are also shown. See Fig. 1 for details for calculating ATP yield and maximum tolerable H 2 concentration. For each pathway, ∆G reaction was calculated assuming 300 μM acetate, 10 μM for other FAs, 1 mM NH 4 + , 50 kPa CH 4 , 50 mM HCO 3 - , 37 °C, 3.9 × 10 -4 atm H 2 S, and 0.1 μM for all other compounds. ∆G ATPsynthesis was assumed to be 60 kJ/mol. *Although more exergonic alternative pathways exist for these HS AA degradation pathways (e.g., through butyrate fermentation), species only expressing the HS pathway(s) were identified in situ, indicating that HS metabolism of these AAs is relevant in situ. 1 For isovalerate degradation, an ATP synthase ATP:H + ratio of 1:4 was assumed. 2 For H 2 -oxidizing CO 2 -reducing methanogenesis, two H 2 concentrations for two ATP yields assuming different ATP synthase ATP:H + ratios. 3 For propionate and acetate degradation, an ATP synthase ATP:H + ratio of 1:5 was assumed. † Pathways whose directionality cannot be determined by sequence data alone. c For each phylum, the percentage of species expressing individual degradation pathways are shown Thermodynamic conundrum The co-existence of the above processes is puzzling in terms of thermodynamics. Most forms of organotrophy in methanogenic ecosystems are presumed to dispose electrons by reducing H + to H 2 . However, some types of organotrophy can produce H 2 to levels that can thermodynamically inhibit other types if H 2 accumulates to sufficient levels. The question then is how does H 2 -mediated interspecies electron transfer from organotrophic bacteria to methanogenic archaea, which is a core process in methanogenic ecosystems, proceed in these ecosystems? Based on our calculations, the maximum tolerable H 2 concentration varies significantly among substrates and pathways involved (Fig. 3 b). Degradation of sugars and many AAs is highly exergonic and HT, while the degradation of FAs and certain AAs is HS and may require exceptionally low H 2 concentrations (≤ 16 Pa). Despite the rapid H 2 consumption by partner methanogens, the high activity and abundance of organisms performing HT metabolism (44 ~ 70% of mapped metatranscriptomes) may generate localized high H 2 concentrations that can inhibit organisms performing HS metabolism. Moreover, the estimated maximum H 2 concentration thresholds for HS AA degradation (1.1 ~ 10.3 Pa H 2 ) and FA degradation (1.2 ~ 2.8 Pa H 2 ) are very close to the minimum hydrogen threshold that conventional H 2 -utilizing CO 2 -reducing methanogenesis can use (1.7 Pa H 2 ), which is often lower than bulk H 2 concentrations detected in reactors (< 10 Pa) [ 25 , 47 ]. Thus, we expect that the species performing HS metabolism may have unique strategies to circumvent these thermodynamic obstacles. Coupling H 2 -tolerant (HT) and H 2 -sensitive (HS) metabolisms To identify ancillary metabolic pathways supporting HS metabolisms in situ, we compared catabolic capacities across the 192 high-activity species. Pearson correlation revealed correspondence between the number of HS AA metabolisms per species cluster and the number of hydrolytic exoenzyme families (glycosylhydrolases [ p = 6.8 × 10 -5 ] and proteases [ p = 1.3 × 10 -20 ]), pathways for HT AA metabolism ( p = 7.4 × 10 -76 ), sugar degradation pathways ( p = 4.5 × 10 -6 ), and types of both [FeFe] and [NiFe] hydrogenases ( p = 2.3 × 10 -11 and 0.033, respectively) (see Table S4 for categories and values used for all Pearson correlation calculations). This suggests an interaction between hydrolysis of a wide range of polymers, simultaneous catabolism of multiple types of polymer-derived monomers, and diverse H 2 generation pathways. Comparison of phyla showed that Bacteroidota encoded significantly more pathways for HS and HT AA metabolism ( p = 0.016 and 0.018 respectively; Student’s t test), glycolsylhydrolases (0.023), and proteases (0.003) than other phyla. Correlation analyses were not possible for other phyla with fewer species, but the principal component analysis also suggested a qualitative association of Fermentibacterota, Marinisomatota, Verrucomicrobiota, and KSB1 with these features (Fig. 4 a, b). We found that many species of these phyla (25 out of 38 species in Bacteroidota, 1 out of 1 in Fermentibacterota, 1 out of 1 in Marinisomatota, 1 out of 5 in Verrucomicrobiota, and 1 out of 1 in KSB1) expressed genes for both HS and HT metabolism of AAs (e.g., HS and HT AA degradation with H 2 formation) (see Tables S7-S9 for overviews and S10-S18 for individual reactors). Of these, 24 Bacteroidota, 1 Fermentibacterota, 1 Verrucomicrobiota, and 1 KSB1 species were confirmed to consistently perform the above metabolism based on the following criteria: expressing the complete metabolic pathway(s) in at least 50% of the studied reactors where this species comprised ≥ 0.05% of the mapped metatranscriptome (herein referred to as ECM50 species; Table S8 ). We detected HS sensitive pathways for lysine (Bacteroidota), isoleucine/leucine/valine (Bacteroidota and Marinisomatota), arginine (Bacteroidota and KSB1), glutamate (Bacteroidota, KSB1, and Fermentibacterota), glycine (all five phyla), and alanine (all five phyla). (Note that, with rigorous annotation as outlined in the methods [see Fig. S4 as an example], we can determine the directionality of most AA metabolism pathways, exceptions being alanine, cysteine, glutamate, and aspartate metabolism [Table S1 ]). Although HS metabolism would be thermodynamically inhibited by H 2 generated from HT degradative processes in proximal cells or in the same cell, HS, and HT metabolism pathways intersect at shared metabolic intermediates (e.g., NAD[H], NADP[H], and/or ferredoxin) that could potentially be coupled enzymatically to provide for thermodynamically favorable redox reactions.\n Fig. 4 Principal component analysis (PCA) of a metabolic capacities, b expressed pathways, and c individual genes/functions for active species. a PCA of active species and their metabolic capacities: proteases and glycosylhydrolases (GHs) as the number of families encoded in the genome; FA, AA, and sugar degradation as the number of pathways encoded in the genome; electron transfer/energy conservation pathways (i.e., Rnf, Nfn, Fix, Efd, and FloxHdr) as the number of pathways encoded in the genome; H 2 and formate generation as presence/absence; and cytochrome bd oxidase-mediated O 2 respiration as presence/absence. Individual species (points) and metabolic capacities (vectors) are shown. Confidence ellipses (95%) are shown for MAGs belonging to specific phyla. b PCA of active species and the metabolic behavior they expressed: proteases and glycosylhydrolases (GHs) as the number of families expressed in at least one reactor; FA, AA, and sugar degradation as the number of complete pathways expressed in at least one reactor; electron transfer/energy conservation pathways (i.e., Rnf, Nfn, Fix, Efd, and FloxHdr) as the number of pathways expressed in at least one reactor; H 2 and formate generation as the highest hydrogenase/formatted dehydrogenase subunit expression level (calculated as RPKM normalized to specie’s non-zero median expression level); and cytochrome bd oxidase-mediated O 2 respiration as the highest oxidase subunit expression level. Individual species (points) and metabolic capacities (vectors) are shown. c PCA of active species and their functional profiles predicted through eggNOG. Functions that are detected at a significantly higher frequency in Desulfobacterota and Spirochaetota than other phyla ( p < 0.05) are shown as vectors. The functions associated with these vectors are shown in Table S16 Though the hydrolytic organisms could theoretically focus on performing HT metabolism, we suspect, based on our analysis of the pathways present in diverse metagenomes, that these organisms degrade wide ranges of substrates (both HS and HT AA metabolism) to maximize energy recovery from the heterogeneous pool of monomers generated from polymer hydrolysis, thereby compensating for the high energy cost associated with producing extracellular hydrolytic enzymes [ 48 ]. It is important to note that HS and HT AA metabolism generally have similar ATP yields despite thermodynamic differences in substrate degradation. We suspect this energy compensation is important for the above phyla as they express a wide range of hydrolytic enzymes. For protein hydrolysis, many species clusters associated with the above phyla were in the top 30% of all active species for the average number of protease families expressed when active (i.e., in reactors they displayed ≥ 0.5% metatranscriptome-based activity) (> 7.8 families)—35 Bacteroidota (34 of which were ECM50 species), 1 Fermentibacterota (1 ECM50 species), 1 Marinisomatota (1 ECM50 species), 4 Verrucomicrobiota (4 ECM50 species), and 1 KSB1 (1 ECM50 species) respectively) (Table S19 ). Similarly, 21 (21 ECM50 species), 0, 1 (1), 5 (5), and 1 (1) species cluster(s) respectively for carbohydrate hydrolysis (> 7.6 glycosylhydrolase families) and 5 (5 ECM50), 0, 1 (1), 1 (1), and 0 species cluster(s) respectively for lipid hydrolysis (> 1.2 lipase families). In addition, Pearson correlation revealed an association between the numbers of families encoded for each exoenzyme type (all p ≤ 3.0 × 10 -7 ). Thus, these versatile anaerobes hydrolyze a broad range of polymers, generate diverse monomers in the process, and use thermodynamically favorable monomer degradation reactions to drive the concomitant degradation of other monomers whose degradation would be otherwise thermodynamically unfavorable. Nearly all species (96.7% or 88 out of 91) that were predicted to perform HS metabolism couple HS and HT AA degradation in at least one reactor (Table S9 ), suggesting this is the predominant strategy to accomplish HS AA degradation. Shifting to interspecies formate transfer Unlike polymer/monomer catabolism, the number of syntrophic FA degradation pathways encoded in a species cluster had a negative correlation with the number of [FeFe] hydrogenases (Pearson correlation p = 0.044; Fig. 4 a, b). This suggests that the FA-degrading syntrophic metabolizers likely employ an alternative route for the re-oxidation of their reduced carriers. While H 2 exchange is the most well-recognized mode of interspecies electron transfer, CO 2 -reducing formate generation also serves as an important mechanism for electron disposal and transfer [ 33 , 42 , 44 , 49 ]. FA catabolism indeed had a unique positive correlation with both Fdh-H type (cytosolic) and Fdh-N type (membrane-associated) formate dehydrogenases (Pearson correlation p = 2.1 × 10 -12 and 1.5 × 10 -35 ) not observed for AA and sugar metabolism. Nearly all Desulfobacterota species (12 total across uncultured Desulfomonalia order UBA1602, Syntrophales families UBA8958 and UBA2192, and Smithellaceae) actively performing syntrophic FA metabolism in at least one reactor expressed genes for CO 2 -reducing formate generation (11 ECM50 species out of 12 total or 91.7%) and, of these, most only expressed genes for formate generation and not for H 2 generation (82.0% ECM50 species; Table S9 ). In agreement, Desulfobacterota had significantly higher numbers of FA degradation (Student’s t test p = 0.045) and Fdh-H/Fdh-N type formate dehydrogenases ( p = 0.044 and 0.032) compared to other phyla. Most Spirochaeotota (uncultured class UBA4802) populations expressing syntrophic butyrate degradation also expressed genes for formate generation (two out of three FA-degrading Spirochaeotota species ECM50). We also observed a correlation between FA metabolism and Fdh-N type formate dehydrogenases with the number of intracellular energy-conserving electron transport enzyme complexes (see Table S6 for list) (Pearson correlation p = 6.4 × 10 -5 and 2.7 × 10 -4 , respectively), indicating the importance of possessing multiple energy conservation routes for syntrophic FA degradation. Through comparing the presence/absence of individual functions (based on automatic emapper-based annotations) across all active species clusters (Fig. 4 c and Table S20 ), we also identified correlation (Student’s t test; p < 0.05) in Desulfobacterota and Spirochaetota between the FA-degrading enzyme acyl-CoA dehydrogenase, electron transfer flavoprotein:quinone oxidoreductase, and formate dehydrogenases Fdh-H and Fdh-N, which plots out the route of electron flow for the most thermodynamically difficult redox reaction involved in syntrophic FA metabolism—the generation of formate or H 2 from electrons derived from acyl-CoA oxidation. Earlier proteomic studies implied these enzyme systems for H 2 or formate production from electrons derived from acyl-CoA oxidation in Syntrophomonas wolfei [ 45 , 50 ]. The finding that this same enzyme system is used in diverse bacteria suggests that this may be the common mechanism for the difficult redox reaction. Remarkably, many populations lacked hydrogenases (53.3% or 8 out of 15 species; see Table S6 for hydrogenases surveyed). This observation is in stark contrast with what is known about isolated syntrophic organisms, which all possess hydrogenases and employ H 2 as an interspecies electron carrier [ 19 , 51 ]. However, further proteomic studies are necessary to verify the absence of hydrogenases in these novel syntrophic populations. We also identified three putative syntrophic acetate-degrading species clusters in Thermotogota ( Pseudothermotoga and an uncultured Thermotogae order) expressing a previously proposed glycine-mediated acetate degradation pathway (two acetate-degrading Thermotogota ECM50 species) [ 17 ]. Two coupled this with formate generation (no H 2 generation) in at least one reactor (one out of two acetate-degrading Thermotogota species ECM50). Unlike hitherto characterized syntrophs, which are cultured in the absence of other H 2 -generating processes (i.e., only one substrate in the culture medium), these newly discovered uncultured organisms may thrive in the presence of highly thermodynamically favorable H 2 -generating processes. We propose that organisms that perform HS FA catabolism avoid thermodynamic conflict with those that use HT-catabolism by completely or partially forgoing H 2 generation and relying on formate transfer to efficiently transport electrons to physically distant metabolic partners [ 33 , 49 , 50 ]. In contrast with H 2 , formate concentrations are unlikely to accumulate locally in situ as formate-producing activity is absent or low in most polymer/monomer-degrading species (i.e., most of the active community) based on our analyses and formate has a higher diffusion rate than H 2 [ 33 , 49 ]. Although formate is challenging to detect in anaerobic digesters, it is estimated to be at concentrations around 2.5 μM (equivalent to 4.5 Pa H 2 at 37 °C, pH 7, and 50 mM HCO 3 - ) [ 33 ]. Moreover, formate transfer would allow FA degraders to recover additional energy via ion-translocating, formate transporters [ 42 ] (expressed by 88.9% of FA-degrading species). Though uncommon, 18 species were found to couple formate generation with HS AA catabolism (Table S7-S9 ; alanine, glycine, glutamate, isoleucine, leucine, lysine, or valine). These organisms span eight phyla and uncultured lineages that have never been reported to be capable of syntrophic interactions: candidate phylum UBP6, uncultured phylum Krumholtzibacteriota, uncultured phylum Cloacimonadota, Bacteroidota (uncultured Bacteroidales family), Chloroflexota (unc. Anaerolineaceae ), Desulfobacterota (unc. Syntrophorhabdaceae), Firmicutes_E (unc. class DTU015), Firmicutes_G (unc. Limnochordia order DTU010), Myxococcota (unc. class XYA12-FULL-58-9), Spirochaetota (unc. Treponematales family), and Thermotogota (unc. Thermotogae order). Given the thermodynamic sensitivity of the aforementioned AA degradations, these species likely rely on formate generation for the same reasons that the syntrophic FA degraders do. Although we cannot conclude syntrophic capabilities without cultivation, we found that the above species encode enzymes involved in supporting thermodynamically challenging reactions and metabolism: reverse electron transport (NADH:Fd oxidoreductase Rnf [UBP6, Krumholtzibacteriota, Cloacimonadota, Bacteroidota, Firmicutes_E, and Firmicutes_G]) and electron bifurcation (NAD-dependenet Fd:NADP oxidoreductase Nfn [UBP6, Bacteroidota, Desulfobacterota, Firmicutes_E, Spirochaetota, and Thermotogota species]). We also identified several syntroph-associated enzymes in the species’ genomes: monomeric formate dehydrogenase [ 42 ] (UBP6, Desulfobacterota, and Spirochaetota), electron transfer flavoprotein dehydrogenases Fix [ 51 ] or Efd [ 41 ] (Krumholtzibacteriota, Bacteroidota, Desulfobacterota, and Myxococcota), and uncharacterized syntroph-associated redox complex Flox-Hdr [ 52 , 53 ] (Krumholtzibacteriota and Desulfobacterota). In addition, of the 18 identified formate-generating HS AA-degrading species, 7 (4 ECM50) did not couple HS metabolism with HT AA degradation in at least one reactor (UBP6, Chloroflexota, Cloacimonadota, Desulfobacterota, and Thermotogota), suggesting the need for formate-mediated syntrophic interaction to complete HS AA degradation Aerobic respiration by obligate anaerobes Beyond the coupling of HS metabolism with HT metabolism or formate generation, we found a positive correlation between the number of HS AA and FA metabolism pathways with the presence of a cytochrome bd oxidase (Pearson correlation p = 4.3 × 10 -4 and 2.3 × 10 -8 ), a terminal oxidase for aerobic respiration (Fig. 4 a, b). The transcription of these genes was detected in at least one reactor for 36.0% of species actively expressing HS AA degradation belonging to the five versatile hydrolytic phyla reported above (27.8% ECM50 species) and 75.0% of the Desulfobacterota and Spirochaetota species expressing syntrophic formate/H 2 -generating FA degradation (58.3% ECM50 species) (Table S7-S9 ). These organisms possess many O 2 -sensitive enzymes (e.g., pyruvate:ferredoxin oxidoreductase, 2-oxo-glutartate:ferredoxin oxidoreductase, formate dehydrogenases, and FeFe hydrogenases) and lack central O 2 -tolerant enzymes (e.g., pyruvate dehydrogenase and 2-oxoglutarate dehydrogenase), indicating that the organisms are strictly anaerobic and not facultatively aerobic. The association of heme biosynthesis genes (hemACL) with Desulfobacterota and Spirochaetota was also observed (Student’s t test p < 0.05; Fig. 4 c), supporting the functionality of cytochrome bd oxidase. Although the anaerobic digestion ecosystem is considered to be strictly anaerobic, minute amounts of O 2 can enter the system through the influent wastewater [ 54 – 56 ]. This is analogous to gas or water percolation from an aerobic zone to a neighboring anaerobic zone in natural ecosystems. Moreover, cytochrome bd oxidase can function even at nanomolar concentrations of O 2 [ 57 ]. Using this low-concentration O 2 as an alternative electron disposal route can reduce the dependence on H 2 or formate production, which is thermodynamically sensitive to the accumulation of these byproducts and increases the thermodynamic favorability of their overall catabolism. For example, for butyrate oxidation in the presence of nanomolar levels of O 2 , redirecting 1% of the electrons towards O 2 respiration can double H 2 tolerance ([H 2 ] max ; from 2.7 to 5.6 Pa) and increase the thermodynamic favorability by 20% (∆G of − 13.3 to − 15.9 kJ/mol, assuming 10 Pa H 2 , 50 nM O 2 , and other conditions used in Fig. 3 b). Moreover, the terminal oxidase can increase tolerance to oxidative stress by consuming O 2 . Indeed, previous studies have demonstrated that a strictly anaerobic organism can tolerate and benefit from nanomolar concentrations of O 2 [ 58 ] and anaerobic digestion can benefit from controlled microaeration [ 56 , 59 ]. Thus, organisms encountering kinetic and thermodynamic bottlenecks (i.e., hydrolysis and HS AA/FA degradation) of methanogenic organic matter mineralization may depend on O 2 for optimal activity. New routes of electron flow in methanogens To better understand interspecies electron transfer, we investigated the metabolic behavior of methanogenic archaea. As expected, most Euryarchaeota and Halobacterota expressed H 2 - and/or formate-driven CO 2 -reducing methanogenesis genes, syntrophically supporting electron disposal of organotrophic activity (Table S3 ). We also discovered high activity (gene expression) in Ca. Methanofastidiosa (previously known as class WSA2), an archaeon previously proposed to utilize methylated thiols as a carbon source for methanogenesis rather than CO 2 [ 60 ]. Metatranscriptomics provided further evidence that Ca. Methanofastidiosum indeed performs H 2 -oxidation coupled to methylated thiol-reduction to methane in situ (all methyl-reducing Methanofastidiosum were ECM50 species) (Table S7 ). Based on thermodynamics, such methanogens can theoretically tolerate much lower H 2 concentrations than those that use conventional H 2 /CO 2 methanogenesis (0.1 Pa versus 1.7 Pa, respectively; assuming conditions in Fig. 3 ). This would mean that in the presence of methylated thiols (generated from the degradation of methylated compounds such as methionine), Ca . Methanofastidiosum can pull H 2 concentrations to much lower levels than conventional methanogens and more effectively support H 2 generation from HS metabolism. Thus, methylated compounds likely play an important role in overcoming thermodynamically challenging metabolisms in anaerobic digestion and other methanogenic ecosystems. Although interspecies electron transfer in methanogenic ecosystems is often simplified as H 2 exchange, such microbial interactions are clearly more complex. In addition to the exchange of metabolites such as H 2 or formate, microorganisms can also directly transfer electrons to each other, a process called direct interspecies electron transfer (DIET) [ 11 ]. Yet, the prevalence and importance of DIET in anaerobic digestion are unclear. Among methanogens detected in situ, Methanothrix is the only lineage known to be capable of utilizing extracellular electrons to drive CO 2 -reducing methanogenesis [ 11 ], although it is most well known for its capacity to use acetate for methanogenesis. We identified three Methanothrix species expressing DIET-driven CO 2 reduction and acetoclastic methanogenic pathways (all acetate-degrading Methanothrix were ECM50 species; Table S7-S9 ), indicating the presence of “exoelectrogenic” organisms in situ. Inspection of the transcriptomes revealed that 13 and 18 bacterial phyla may perform DIET respectively through multiheme c-type cytochromes (including members of uncultured phyla Omnitrophota, KSB1, and Krumholzibacterota) and conductive pili (including members of uncultured phyla Cloacimonadota, Omnitrophota, Patescibacteria, Krumholzibacterota, and WOR-3). Expression of multiheme c-type cytochromes was observed for Desulfobacterota and Spirochaeota performing syntrophic FA degradation (53.0% of FA-degrading Desulfobacterota and Spirochaetota species; 46.7% ECM50 species) and versatile polymer/monomer-degrading Bacteroidota, Verrucomicrobiota, and KSB1 (54.2%; 40.0% ECM50 species). For conductive pili, we found positive correlation for the presence of conductive pili with syntrophic FA degradation (Pearson correlation p = 8.0 × 10 -7 ) and other capacities associated with in situ FA degraders (Fdh-N type formate dehydrogenase [ p = 1.9 × 10 -8 ], electron transfer complexes [ p = 2.3 × 10 -4 ], and cytochrome bd oxidase [ p = 5.2 × 10 -3 ]), while no correlation was observed with hydrolytic enzymes and AA/sugar degradation. We further confirmed that many FA-degrading Desulfobacterota and Spirochaetota species express conductive pili (60.0%; 46.7% ECM50 species), but only three populations of the versatile hydrolytic HS/HT AA degraders expressed putative conductive pili in at least one reactor (5.7% ECM50 species). Diverse phyla and niches likely take advantage of DIET because exoelectrogenic metabolism can theoretically much more thermodynamically favorable than H 2 generation due to the high reduction potential of c-type cytochromes (E°’ of − 220 to + 180 mV [ 61 ]). The reason for the difference in the distribution of multi-heme cytochromes (all studied niches) and putative conductive pili (preferentially found in syntrophic FA degraders) remains unclear. Though the necessary physical proximity between syntrophs and electron-accepting partners may allow more opportunities for conductive pili to transfer electrons, further investigation is required. In total, hydrolysis, monomer degradation, and FA degradation by uncultured organisms across 20 phyla may rely on Methanothrix species for H 2 -independent extracellular electron transfer, though different niches may use different routes. Further inspection of the transcriptomes revealed the possible involvement of Methanothrix in formate degradation. Although the ability of Methanothrix to degrade formate has been controversial [ 7 , 62 , 63 ], we detected consistent expression of a formate dehydrogenase complex in two out of three Methanothrix species in all reactors they were active (Table S7-S9 ). Based on gene organization of the formate dehydrogenase in the most active Methanothrix species JPASx098 (fdhA with hdrABC and ferredoxins; ≥ 99% similarity to M. soehngenii genes MCON_3277-83), Methanothrix may oxidize formate and funnel electrons into methanogenesis (via HS-CoM/HS-CoB and ferredoxin). We suspect that Methanothrix primarily performs acetate-driven methanogenesis but, in parallel, can uptake formate and electrons from extracellular pili and cytochromes to drive CO 2 -reducing methanogenesis. Therefore, Methanothrix likely plays an essential role in supporting multiple H 2 -independent electron disposal routes for organisms performing HS metabolism. Temperature-based differences The coupling of HS catabolism with hydrolysis/HT catabolism, formate generation, oxygen respiration, and DIET was observed across all reactors, despite variation in temperature (Table S7 ). This indicates that the described phenomena may support HS metabolism at a wide temperature range. Across all studied temperatures, we observed Desulfobacterota and Spirochaetota FA degradation coupled with the expression of formate generation, oxygen respiration, and DIET (with the exception of Desulfobacterota at thermophilic temperature). Strict reliance on formate-generating FA degradation was only observed at mesophilic temperatures. Coupled expression of HS and HT AA degradation (with complementary formate/H 2 generation) was also observed across all temperatures. However, some of these AA-degrading species were only observed to have high activity and express complete pathways at specific temperature ranges—UBP6, KSB1, Cloacimonadota, Fermentibacterota, Marinisomatota, Chloroflexota, Firmicutes_A, Firmicutes (~35 °C); WOR-3_A, Firmicutes_E (> 50 °C), Krumholtzibacterota (≤ 30 °C); Coprothermobacterota (> 40 °C); Myxococcota, Spirochaetota, Planctomycetota (< 50 °C), Caldisericota (~ 35 °C and > 50 °C). For methanogenesis, CO 2 -reducing methanogenesis by Methanothrix (potentially driven by DIET) was detected across all temperatures, but formate oxidation by Methanothrix and methyl reduction by Methanofastidiosa was only observed at temperatures below 50 °C. Thus, based on the available data, strategies for supporting HS metabolism and organisms that perform these challenging reactions differ between anaerobic digesters operated at different temperatures. However, analyses of more samples at non-standard temperatures (~35 °C) are necessary to better characterize temperature-based variation."
} | 10,842 |
35486699 | PMC9170069 | pmc | 1,034 | {
"abstract": "Significance A central problem in evolutionary biology is explaining variation in the organization of task allocation across collective systems. Why do human cells irreversibly adopt a task during development (e.g., kidney vs. liver cell), while sponge cells switch between different cell types? And why have only some ant species evolved specialized castes of workers for particular tasks? Although it seems reasonable to suppose that such differences reflect, at least partially, the different ecological pressures that systems face, there is no general understanding of how a system’s dynamic environment shapes its task allocation. To this end, we develop a general mathematical framework that reveals how simple ecological considerations could potentially explain cross-system variation in task allocation—including in flexibility, specialization, and (in)activity.",
"conclusion": "Conclusion For a collective system that faces a variable environment, flexibility may seem intuitively advantageous. Indeed, if we think of the collective system as an individual in its own right, then this intuition is corroborated by general models for the evolution of phenotypic plasticity, some of which predict that plasticity in labile traits (such as task allocation) should always be favored, at least when not costly ( 80 – 82 ). In contrast, our framework predicts that the evolution of collective flexibility should not necessarily be expected: Flexibility in task allocation may be maladaptive, unachievable, or both. This discrepancy in predictions arises because, instead of presupposing that the optimal phenotype is determined solely by the current environment ( 27 , 80 , 82 ), our framework allows the optimal task allocation to emerge dynamically from the interaction between the system and its environment. This more mechanistic approach accounts for the fact that what allocation is optimal depends not only on the environment but also on what tasks have been recently performed, as the system may, at least to some extent, be able to aggregate its task performance over time. We find that the emergent optimal task allocation may end up being much less variable than the environment itself, thereby limiting the potential for the evolution of flexibility. In particular, flexibility is not favored when the environment changes so slowly that task performance cannot be effectively aggregated across environments, or when the environment changes so quickly that the system cannot adjust to it in time. Flexibility may even be so strongly selected against that it becomes optimal to continue to perform tasks under very unfavorable conditions (when their yields are low or even zero). Whether collective flexibility is selected for, in turn, has ramifications for the internal organization of task allocation. Indeed, whether and to what extent systems may evolve various organizational properties—including task switching, specialization, and inactivity—depends on trade-offs between the potential fitness benefits of collective flexibility and other relevant factors, such as task-switching costs or efficiency benefits to specialization. Through these trade-offs, our framework has the potential to explain substantial variation in the organization of task allocation across ecological conditions and provide theoretical explanations for organizational properties that may, at least at first glance, seem counterintuitive. In Table 1 , we summarize some of these properties that have been observed empirically, together with theoretical interpretations that are possible when one takes into account the variability of the environment (and the associated potential benefits of collective flexibility). Table 1. Puzzling empirical observations on task allocation that can potentially be adaptive in the context of a dynamic environment General observation Illustrative empirical example(s) Possible theoretical interpretation of the general observation Individuals continue to switch tasks in the absence of environmental changes that necessitate a shift in task allocation. Sponge cells ( A. queenslandica ) spontaneously transition between different cell types: As part of normal tissue homeostasis, archeocytes can transdifferentiate to become choanocytes and vice versa ( 52 , 54 ). Frequent task switching may be favored, even when costly, because it facilitates collective flexibility. The incurred costs for “unnecessary” task switching under constant conditions are offset by the benefits of being able to rapidly adjust overall task allocation when environmental conditions do change. (See Costly task switching may evolve to achieve collective flexibility .) Tasks are performed by a mix of generalists and specialists. In A. variabilis cyanobacteria, nitrogen is fixed by specialized heterocysts but also by vegetative cells ( 68 ). The mix of generalists and specialists may represent an intermediate optimum that allows the system to derive some efficiency benefits of specialization (by having some specialists perform the task) while still maintaining the ability to flexibly adjust task allocation to a changing environment (by also allocating generalists to the task when needed). (See Environmental variability constrains the evolution of individual specialization .) In stingless bees ( Tetragonisca angustula ), nest defense is performed by a mix of specialized soldiers and nonspecialists recruited to guarding tasks ( 95 ). Some tasks are performed by specialists, while other tasks are not. During nest construction in Metapolybia wasps, specialists are employed for water foraging but not for other tasks such as building and pulp foraging ( 96 ). Even when their potential efficiency benefits of specialization are the same, tasks can nevertheless differ in whether those efficiency benefits are sufficient to offset the costs of sacrificing some flexibility; in particular, larger efficiency benefits are required to specialize on more environment-sensitive tasks. (See Environmental variability constrains the evolution of individual specialization .) Individuals specialize on a task without deriving efficiency benefits. In Temnothorax albipennis ants, specialization and efficiency do not correlate: Workers that spend more of their time on a certain task are not more efficient at it than workers that spend less time on it ( 64 ). “Inefficient specialists” may be de facto generalists who spend all (or most) of their time on a particular task to complement specialists for other tasks (which could be favored, for example, because they do derive efficiency benefits). (See Environmental variability constrains the evolution of individual specialization .) Individuals are not performing any task, thereby seemingly compromising system productivity. L. allardycei ants appear to spend most (55%) of their time doing nothing ( 74 ). For other examples, see ref. 72 and references therein. If inactive individuals can quickly be recruited to tasks that require additional attention, then having an inactive pool of individuals may be adaptive—even though those individuals do not contribute to productivity—because it enhances collective flexibility: It allows the system to adjust its task allocation more rapidly to changing external circumstances. (See Temporary inactivity can evolve to enhance collective flexibility .) Each row lists a general empirical observation illustrated with a specific example, and a potential theoretical explanation supported by our framework. The proposed explanations illustrate how considering a system’s dynamic environment can provide potential interpretations of unexplained aspects of its task allocation. Beyond the explanation proposed by our framework, many alternative explanations—including nonadaptive ones—deserve consideration as well. We propose that the presented framework is sufficiently versatile to be used broadly to study how ecology shapes the evolution of task allocation, and we outline a few directions for future work. First, while our model accounts for environmental variability by having task yields vary temporally, future models could explore additional ecological pressures, such as unpredictable disturbances that affect individuals performing specific tasks (e.g., predator-induced mortality of foragers in social insects) ( 29 , 30 , 83 – 86 ). Second, future work could relax the implicit assumption that the system has perfect information about its environment, in order to explore the implications of noisy information that is potentially asymmetrically distributed across individuals ( 87 – 90 ). Finally, we have shown that inactivity could (at least in certain systems) be interpreted as an evolutionary innovation that allows systems to partially escape one of the ecological constraints on task allocation—the limited rate at which task allocation can be adjusted. Why and under what conditions similar innovations could have evolved, such as individuals specialized to quickly detect changes in the environment (e.g., sensory cells in multicellular animals), communication mechanisms to quickly spread information about those changes among individuals [e.g., social interactions in insect colonies ( 91 – 93 )], and storage mechanisms that prevent task stocks from being depleted too quickly [e.g., replete ants that act as living reservoirs ( 94 )], are further questions that are ripe for theoretical analysis in the type of ecologically explicit framework developed here.",
"discussion": "Results and Discussion The Evolution of Collective Flexibility. All three ecological time scales impact the optimal task allocation. We start by investigating systems where the yields of only one of the two tasks depend on environmental conditions ( θ 1 > 1 , θ 2 = 1 ). For example, this scenario could represent multicellular cyanobacteria where we can assume that the yields of photosynthesis, but not nitrogen fixation, depend on the time of day, or some ant colonies where we can assume that the yields of foraging, but not of nest-associated tasks such as nursing or nest maintenance, depend on food availability around the colony. For such systems, environmental variability in task yields creates an incentive to adjust task allocation to the environment: Performing the environment-sensitive task specifically under the conditions that favor it can increase task yields and thereby system fitness. At the same time, however, temporarily biasing allocation toward a task comes with the risk that the corresponding neglect of the other task reduces fitness. Therefore, it is not clear a priori whether and when environmental variability renders it optimal to adjust task allocation to the environment. To address this question, we use our model to determine the optimal dynamic task allocation as a function of the three ecological time scales considered: adjustment of task allocation ( Fig. 2 A ), task stock depletion ( Fig. 2 B ), and environmental variability ( Fig. 2 C ). We characterize this optimal task allocation by its equilibrium allocations z 1 , A and z 1 , B (toward task 1 in environments A and B, respectively), as well as the corresponding average allocations a 1 , A ¯ = δ δ + λ ⋅ z 1 , A + λ δ + λ ⋅ z 1 , A + z 1 , B 2 and a 1 , B ¯ = δ δ + λ ⋅ z 1 , B + λ δ + λ ⋅ z 1 , A + z 1 , B 2 ( SI Appendix ), which take into account that the system is not necessarily able to reach its equilibrium allocation. Indeed, the average allocation closely matches the equilibrium allocation when the system is able to adjust its task allocation quickly relative to the time scale of environmental fluctuations ( δ / λ sufficiently large; e.g., scenario iii in Fig. 2 ), but deviates from the equilibrium allocation when the system adjusts too slowly to be able to reach its equilibrium allocation before the environment changes again ( δ / λ not large enough; e.g., scenario ii in Fig. 2 ). We quantify collective flexibility as the difference between the average allocations toward task 1 in environments A and B, i.e., a 1 , A ¯ − a 1 , B ¯ = ( z 1 , A − z 1 , B ) ⋅ δ δ + λ . This measure of collective flexibility equals 0 when a 1 , A ¯ = a 1 , B ¯ , meaning that there is no difference in average task allocation between the two environments, and 1 when a 1 , A ¯ = 1 and a 1 , B ¯ = 0 , meaning that, in environment A, only task 1 is performed, and in environment B, only task 2 is performed. Fig. 2. Optimal task allocation strategies when only one task is sensitive to the environment ( θ 1 = 4 , θ 2 = 1 ). ( A – C ) Optimal equilibrium task allocation ( z 1 , A , z 1 , B ) and the corresponding average task allocations ( a 1 , A ¯ , a 1 , B ¯ ). The level of collective flexibility is quantified as the difference a 1 , A ¯ − a 1 , B ¯ . Parameters δ ( A ), γ ( B ), and λ ( C ) are varied independently, while the other two parameters are kept constant (constant values: γ = 0.2 , δ = 1 , and λ = 1 ). ( D – I ) Evolved task allocation dynamics for three values of δ : δ = 0.1 ( D and E ), δ = 1 ( F and G ), and δ = 10 ( H and I ). Parameters γ = 0.2 and λ = 1.0 are kept fixed. Equilibrium task allocations are ( z 1 , A , z 1 , B ) = ( 0.56 , 0.44 ) ( D and E ), ( z 1 , A , z 1 , B ) = ( 0.84 , 0 ) ( F and G ), and ( z 1 , A , z 1 , B ) = ( 0.85 , 0 ) ( H and I ). ( D , F , and H ) Task allocation dynamics over time, as given by d a i / d t = δ ( z i − a i ) . The same pattern of environmental change is shown for each scenario. ( E , G , and I ) Task stock dynamics over time, as given by d S i / d t = γ ( a i Y i − S i ) (arbitrary units). Changes in the dynamics of S 1 are caused by temporal variation in task yields Y 1 ( θ 1 = 4 ) and task allocation a 1 ; changes in the dynamics of S 2 are caused only by temporal variation in task allocation a 2 because the yields Y 2 for task 2 are constant ( θ 2 = 1 ). We find that the evolution of task allocation depends strongly on all three ecological time scales considered and that diverse task allocation dynamics can be evolutionarily optimal. We first focus on each time scale separately and vary the rate parameters δ , γ , and λ independently (while keeping the other two fixed; see SI Appendix for formal analysis); subsequently, we consider all time scales simultaneously to derive general conditions for the evolution of collective flexibility that depend on the relative magnitudes of δ , γ , and λ . In Fig. 2 D – I , we show the optimal task allocation dynamics for three choices of parameters that differ in the rate δ at which the system can adjust its task allocation to the environment (corresponding to cases i through iii in Fig. 2 A ). For low δ , we find that it is optimal to be unresponsive to the environment and equally allocate individuals to both tasks, thereby ensuring that the system properly balances its attention toward them ( a 1 , A ¯ ≈ a 1 , B ¯ ≈ 1 / 2 ; Fig. 2 A , D , and E ). As δ increases, the system evolves to adjust its task allocation to its environment ( Fig. 2 A and F–I ), preferentially performing the environment-sensitive task under conditions that favor it (i.e., a 1 , A ¯ > 1 / 2 and a 1 , B ¯ < 1 / 2 ). Adjusting task allocation to the environment allows the system to obtain higher task yields, maintain higher task stocks (compare Fig. 2 G and I with E ), and therefore achieve higher fitness ( SI Appendix , Fig. S2 ). The level of collective flexibility a 1 , A ¯ − a 1 , B ¯ that evolves increases with δ , because for higher δ the system evolves more biased equilibrium allocations z 1 , A and z 1 , B ( Fig. 2 A ), and it reaches these equilibrium allocations faster (e.g., compare scenario ii, where the equilibrium allocation is typically not reached before the environment changes [ Fig. 2 F ], to scenario iii, where the system spends most of its time at the equilibrium allocation for the current environment [ Fig. 2 H ]). When we instead vary the rate of task stock depletion γ , we find that adjustments of task allocation fail to evolve when task stocks are depleted too quickly (high γ ; Fig. 2 B ), because then temporarily neglecting tasks is strongly selected against. In contrast, when task stocks are depleted more slowly and the fitness consequences of temporarily neglecting tasks are less dire ( γ → 0 ), it becomes optimal to bias task allocation as much as possible ( z 1 , A → 1 , z 1 , B → 0 ). In this case, the level of collective flexibility approaches δ / ( δ + λ ) and is thus determined by how quickly the system adjusts its task allocation relative to the time scale of environmental fluctuations ( SI Appendix and Fig. 2 B ). Finally, when we vary the rate of environmental fluctuations λ , we find that adjustments of task allocation evolve only if the environment changes sufficiently quickly ( λ sufficiently large; Fig. 2 C ). If the environment changes too slowly, biasing task allocation toward a specific task would again lower fitness, because it causes the other task to go neglected for too long. As λ increases, the equilibrium task allocations z 1 , A and z 1 , B become progressively more biased, although, eventually, adjustments to the environment fail to be realized because the environment changes too quickly for the system to be able to adjust to it (i.e., a 1 , A ¯ and a 1 , B ¯ approach 1/2; Fig. 2 C ). The task allocation dynamics that evolve have some surprising features. For example, as Fig. 2 D shows, it may be optimal to continue to perform tasks when their yields are low (or even zero; SI Appendix , Fig. S3 ): In environment B, task 1 is still being performed by half of the individuals, even though its yield is 4 times lower than in environment A. Thus, the apparent inefficiency of performing tasks under suboptimal conditions can be evolutionarily optimal. Another unexpected finding is that, even though both tasks contribute equally to fitness, the optimal task allocation can be asymmetric. For example, in scenario iii ( Fig. 2 H and I ), individuals spend on average 30% more time performing task 2 than task 1 and only the performance of task 1 is restricted to specific environmental conditions. We can show analytically that if the environment-sensitive task 1 is sufficiently sensitive to the environment (i.e., provided θ 1 > ( 4 γ + λ ) / λ ; SI Appendix ), then for large enough δ it becomes optimal to only perform it in environment A ( z 1 , B becomes 0 and a 1 , B ¯ approaches 0 as δ → ∞ ; Fig. 2 A ). In contrast, for the environment-insensitive task 2, it is always optimal to perform it in both environments (i.e., z 1 , A and a 1 , A ¯ never become 1; Fig. 2 A ), with allocation toward it never dropping below γ / ( 2 γ + λ ) ( SI Appendix ). Thus, only for sufficiently environment-sensitive tasks can it be optimal to temporarily ignore them, and, as a result, total allocation—over the lifespan of the system—may be biased toward environment-insensitive tasks. Collective flexibility evolves only under certain ecological conditions. Having explored the effect of each of the three ecological time scales separately, we now look at the effect of all three varying simultaneously. The level of collective flexibility that evolves in the model varies continuously as a function of λ , δ , and γ , ranging from a regime in which negligible collective flexibility evolves ( a 1 , A ¯ − a 1 , B ¯ ≈ 0 ) to a regime in which near-maximal levels of collective flexibility evolve ( a 1 , A ¯ − a 1 , B ¯ ≈ 1 ), with intermediate levels of collective flexibility in between ( Fig. 3 A , and see SI Appendix , Fig. S4 for the corresponding values of z 1 , A , z 1 , B , a 1 , A ¯ , and a 1 , B ¯ ). We see that the evolution of nonnegligible collective flexibility requires two conditions to be met simultaneously ( Fig. 3 B ). First, the rate of task stock depletion must be sufficiently low relative to the rate at which the environment fluctuates ( γ / λ sufficiently small), so that tasks can be temporarily left unattended without the corresponding task stock being depleted to such low levels that fitness is compromised. Second, the rate of adjustment of task allocation should be sufficiently high relative to the rate at which the environment fluctuates ( δ / λ sufficiently large), so that a biased task allocation can be achieved before the environment changes. Fig. 3. The evolution of collective flexibility. ( A ) The evolution of collective flexibility depends on the relative ratios of three ecological time scales: the time scales of adjustment of task allocation to the environment ( δ parameter), task stock depletion ( γ parameter), and environmental fluctuations ( λ parameter). The two-dimensional visualization takes advantage of the fact that there are three time scales but only two degrees of freedom: Rescaling each time scale by the same amount does not affect the model dynamics. The three axes are interdependent (indeed, δ / λ = ( δ / γ ) ⋅ ( γ / λ ) ); each point in the coordinate space can be orthogonally projected on each axis (in the plane of the figure). Parameters are θ 1 = 4 and θ 2 = 1 . ( B ) Collective flexibility evolves when γ / λ is sufficiently small and δ / λ is sufficiently large. ( C ) Collective flexibility can evolve whenever θ 1 ≠ θ 2 and is maximized for intermediate environment fluctuation rates λ . Parameters γ = 0.2 and δ = 1.0 are kept fixed. These general theoretical conditions for the evolution of collective flexibility comport with empirical observations in desert harvester ants ( P. barbatus ), where collective flexibility—in the form of regulating foraging activity levels in response to day-to-day variation in humidity—has been shown to be adaptive ( 7 , 8 ). Relative to the daily time scale of environmental fluctuations, in harvester ant colonies the adjustment of task allocation indeed takes place on a much faster time scale [within minutes, via interactions with returning foragers or patrollers ( 8 , 36 )], while their task stocks deplete on a much slower time scale [collected seeds are stored for up to months ( 7 )]. Based on our results, we predict that across systems that face similar constraints on task allocation (i.e., have similar values of γ , δ ), such as closely related species of ants, collective flexibility is least likely to evolve in systems living in environments that change very quickly or very slowly and is most likely to evolve in systems living at intermediate degrees of temporal environmental variability ( Fig. 3 C ). Indeed, the evolution of collective flexibility requires the environment to change quickly enough for the system to be able to exploit environmental variability, but not so quickly that the system loses the ability to adjust its task allocation in time. Collective flexibility can evolve whenever tasks differ in how their yields depend on the environment. Our results generalize to scenarios in which not one but both tasks are sensitive to environmental conditions. In this case, either one environment maximizes yields for both tasks (correlated task yields across environments; θ 1 , θ 2 > 1 or θ 1 , θ 2 < 1 ), or the yields for the two tasks are maximized in opposite environments (anticorrelated task yields; θ 1 < 1 < θ 2 or θ 2 < 1 < θ 1 ). As an example of correlated task yields, consider pollen and nectar foraging in a honey bee colony ( 37 ), both of which may result in higher yields during good weather conditions. In contrast, a hypothetical primitive multicellular organism that allocates cells to feeding and motility and navigates a spatially heterogeneous resource environment could be an example of anticorrelated task yields, if feeding (but not motility) is favored when resources are plentiful while motility (but not feeding) is favored when resources are scarce. In this more general setup where both tasks are sensitive to the environment, we find that collective flexibility can evolve as long as the two tasks differ in how they depend on environmental conditions ( θ 1 ≠ θ 2 ; Fig. 3 C ), but that no collective flexibility can evolve when θ 1 = θ 2 (see Fig. 3 C for θ 1 = θ 2 = 4 and SI Appendix for a general mathematical derivation). In particular, collective flexibility can evolve even when task yields are correlated ( Fig. 3 C ; \n θ 1 = 4 , θ 2 = 2 ). In this case, the more environment-sensitive task will be prioritized under conditions favorable to both tasks, increasing the yield for this task to such an extent that it outweighs a concomitant decrease in yield for the less environment-sensitive task (which is now preferentially performed under conditions unfavorable to both tasks). While collective flexibility can evolve even when task yields are correlated, it is most pronounced ( Fig. 3 C ) and provides the largest fitness benefits ( SI Appendix , Fig. S2 ) when task yields are, instead, anticorrelated across environments, so that adjusting task allocation to the environment can simultaneously increase the yields obtained for both tasks. Consequences of Environmental Variability for the Internal Organization of Task Allocation. The framework we have developed to investigate the evolution of collective flexibility additionally presents the opportunity to explore how environmental variability affects the internal organization of task allocation. We do so in three separate extensions of our framework that each focus on a different organizational property: task switching, specialization, and inactivity, respectively. Together, these extensions showcase the versatility of our framework as a platform to investigate the evolution of diverse aspects of task allocation in the context of a dynamic environment. Costly task switching may evolve to achieve collective flexibility. One way in which systems adjust their task allocation to changing conditions is by reallocating existing individuals to different tasks (i.e., task switching). * Task switching may be associated with fitness costs, because adjusting to a new task can require time and/or energy ( 42 – 44 ). So far, we have not considered such costs to flexibility: While in our original model systems are constrained by how quickly they can adjust their task allocation, these adjustments are not costly. We therefore extend our model to incorporate task-switching costs and explore what task-switching strategies evolve when task switching is costly but at the same time required for collective flexibility. Incorporating task-switching costs requires specifying how the net change in the overall task allocation (as given by Eq. 1 ) comes about from individuals switching between tasks ( Fig. 4 A ). For simplicity, we make the minimal assumption that task-switching rates depend only on the task being performed and on the current environment. Consequently, we parametrize our task-switching model by four parameters, r 1 → 2 , A , r 1 → 2 , B , r 2 → 1 , A , and r 2 → 1 , B , that describe the per capita rates of task switching in either direction (from task 1 to task 2 or vice versa) and in either environment (A or B). In SI Appendix , we show that there is a unique way to choose these parameters that is consistent with Eq. 1 : We must have r 1 → 2 , A = δ z 2 , A , r 1 → 2 , B = δ z 2 , B , r 2 → 1 , A = δ z 1 , A , and r 2 → 1 , B = δ z 1 , B . These equations can be interpreted as individuals reevaluating their current task at rate δ (prompted, for example, by observations of the environment) and then adopting task 1 with probability z 1 and task 2 with probability z 2 . While this setup is an oversimplification of the decision-making processes underlying task switching in many real systems, it suffices to explore theoretically the trade-off between the benefits of collective flexibility and the costs of task switching. Fig. 4. The evolution of task allocation strategies under costly task switching. ( A ) The net flow δ ( z 1 − a 1 ) from task 2 to task 1 can be decomposed as the difference between a flow δ z 1 a 2 of individuals switching from task 2 to task 1 at a per capita rate δ z 1 and a flow δ z 2 a 1 of individuals switching from task 1 to task 2 at a per capita rate δ z 2 . Indeed, δ z 1 a 2 − δ z 2 a 1 = δ z 1 ( 1 − a 1 ) − δ ( 1 − z 1 ) a 1 = δ ( z 1 − a 1 ) . ( B ) Evolved reevaluation rates δ , capped at 10 − 3 and at 10 3 (see SI Appendix , Fig. S5 for the corresponding values of z 1 , A and z 1 , B ). ( C – E ) Task-switching dynamics for parameter combinations corresponding to the different qualitative regimes that emerge: no task switching for λ = 0.1 , c = 1 ( C ); deterministic task switching in response to the environment for λ = 0.5 , c = 0.025 ( D ); and stochastic task switching in both environments for λ = 0.1 , c = 0.025 ( E ). Each row corresponds to one individual. In B – E , we have γ = 0.1 , θ 1 = 4 , and θ 2 = 1 . The above assumptions on the rates of task switching allow the net change in task allocation to be decomposed into flows of individuals switching from task 2 to task 1 (at a per capita rate δ z 1 ) and from task 1 to task 2 (at a per capita rate δ z 2 ) ( Fig. 4 A ). The average rate of task switching now equals [2] δ ¯ = δ ⋅ ( a 1 , A ¯ ⋅ z 2 , A + a 2 , A ¯ ⋅ z 1 , A ) + ( a 1 , B ¯ ⋅ z 2 , B + a 2 , B ¯ ⋅ z 1 , B ) 2 < δ ( SI Appendix ). Total (i.e., system-level) task-switching costs can therefore be implemented by subtracting from fitness a cost c δ ¯ , where c is a parameter that controls how costly one task switching event is, regardless of the direction of the switch. To characterize the evolution of task switching, we will calculate what combination of δ ∈ ( 0 , ∞ ) and z 1 , A , z 1 , B ∈ [ 0 , 1 ] maximizes fitness. Thus, δ is now evolvable rather than being a fixed constraint on task allocation. Higher values of δ allow the system to adjust its task allocation more quickly, but may also lead to higher total task-switching costs ( Eq. 2 ). We find that, in the extreme case of no task-switching costs ( c = 0 ), arbitrarily large reevaluation rates δ evolve, as increasing δ allows for quicker collective adjustments of task allocation at no additional cost ( Fig. 4 B ). At the other extreme of very high task-switching costs, arbitrarily small reevaluation rates evolve ( δ → 0 ) because the costs of task switching outweigh any potential benefits of collective flexibility. As a result, individuals rarely switch tasks, and end up being essentially irreversibly committed to their current task ( Fig. 4 C ). This confirms that task-switching costs can suffice to drive the evolution of specialization, a result that has been previously obtained with simulation models ( 43 , 45 ). Our analysis additionally reveals an ecological dimension to this phenomenon: How high task-switching costs need to be to drive the evolution of irreversible specialization depends on the potential benefits of collective flexibility. The higher the benefits of collective flexibility (which are maximized at intermediate environmental fluctuation rates λ ; SI Appendix , Fig. S2 ), the higher the task-switching costs needed to offset those benefits and drive the evolution of irreversible specialization ( Fig. 4 B ). When task-switching costs are not high enough to completely impede the evolution of flexibility, we find that two types of flexible task-switching strategies can evolve ( Fig. 4 B ). In environments that change sufficiently quickly (relative to the time scale of task stock depletion), individuals evolve to instantaneously and deterministically switch tasks ( δ = ∞ ; z 1 , A = 1 , z 1 , B = 0 ) in response to shifts in environmental conditions ( Fig. 4 D ) , akin to undifferentiated cyanobacteria in which all cells switch between photosynthesis during the day and nitrogen fixation at night. While this strategy limits task-switching costs by having individuals switch between tasks only in response to environmental change (indeed, we have δ ¯ = λ / 2 , so the total costs of task switching remain bounded), it is not feasible in environments that change too slowly, where it would cause tasks to go neglected for too long. In those environments, a task-switching strategy evolves instead in which individuals switch tasks stochastically, even in the absence of environmental change, thereby making sure that both tasks receive attention at all times ( Fig. 4 E ). These results provide theoretical corroboration for the empirical observation that in some real systems individuals frequently switch tasks, even in the apparent absence of environmental changes that would necessitate a shift in overall task allocation. In paper wasps ( Polybia ), for example, workers constructing the nest flexibly switch between water collection, pulp collection, and building ( 46 , 47 ). Similarly, while most animal cells are irreversibly assigned a task during development and therefore never switch tasks, some multicellular animals exhibit a more labile (and evolutionarily more ancient) form of cell differentiation in which cells readily transition between different cell types ( 48 – 51 ). For example, in sponges (e.g., Amphimedon queenslandica ), archeocytes can spontaneously transdifferentiate to become choanocytes, and vice versa ( 52 – 54 ). Our results suggest that such individual-level flexibility can be adaptive—even when it leads to unnecessary and potentially costly task switching in the absence of environmental change—because it enables collective adjustments of task allocation when environmental conditions do change. Environmental variability constrains the evolution of individual specialization. So far, we have been agnostic toward which individuals are performing what tasks. In a second extension of our original model, we consider how the system may distribute the performance of different tasks across different individuals (i.e., spatially) in addition to distributing them across different environmental conditions (i.e., temporally). In many extant collective systems, individuals indeed exhibit persistent differences in task performance, although to varying degrees ( 46 , 55 – 63 ). We reason that such specialization may be favored because it has the potential—albeit not universal ( 64 )—intrinsic benefit of improved efficiency (i.e., specialists may be more efficient at a task due to experience or task-specific physiological adaptations) ( 33 , 34 , 65 – 67 ). At the same time, however, specialization limits an individual’s ability to be reallocated to other tasks when needed, thereby constraining the system’s collective flexibility. We therefore use our model to explore how environmental variability impacts the evolution of specialization. We consider the evolution of specialists “one task at a time,” which allows us to capture how the evolution of specialists may potentially differ between tasks. Specifically, we assume that a fraction s of individuals are identical “specialists” that perform only task 1, and the remaining fraction, 1 − s , are identical “generalists” that can perform both tasks and adjust their task allocation at a rate δ without an associated cost, as in the original model. We introduce a parameter α that controls the efficiency gains of specialization: specifically, the yields specialists obtain for task 1 are multiplied by 1 + α . Treating the efficiency parameter α as a fixed constraint, we allow the proportion of specialists s and the task allocation of generalists to evolve, and ask what combination of s ∈ [ 0 , 1 ] and z 1 , A , z 1 , B ∈ [ 0 , 1 ] maximizes fitness ( SI Appendix ). We first determine how the minimum efficiency benefits α * required for the evolution of at least some specialists (i.e., s > 0 ) depend on ecological conditions. We find that specialization is easiest to evolve (i.e., requires the lowest α * ) when the potential benefits of collective flexibility are limited. In particular, when it is optimal to perform a task under all conditions, minimal efficiency benefits suffice: Specialists for such a task can evolve as soon as α > 0 ( Fig. 5 A and SI Appendix ). In contrast, when the benefits of collective flexibility render it optimal to restrict the performance of a task to specific environmental conditions, evolving specialists is more difficult. In this case, the efficiency gains of specialization must outweigh the fitness costs incurred by sacrificing some collective flexibility, and substantial gains in efficiency may be required for specialization to evolve ( Fig. 5 A ). Fig. 5. The evolution of task specialists. ( A ) Minimum efficiency gain α * required for the evolution of specialists that perform only the environment-sensitive task 1. The visualization is two-dimensional, as in Fig. 3 A ; each data point can be orthogonally projected (in the plane of the figure) onto each of the three interdependent axes. ( B ) Optimal fraction of specialists s for α = 0.1 . ( C – E ) Examples of possible task allocation dynamics for the three qualitatively different regimes identified in B : No specialists evolve, and task 1 is performed only by generalists ( C ); task 1 is performed by both generalists and task 1 specialists ( D ); task 1 is performed only by task 1 specialists and the remaining generalists are effectively specialists for task 2 ( E ). Parameters for the displayed task allocation dynamics are 10 γ = δ = λ ( C ), 4 γ = δ = λ ( D ), and γ = δ = λ ( E ). In A – E we have θ 1 = 4 , θ 2 = 1 . See SI Appendix , Fig. S6 for the evolution of specialists for the environment-insensitive task 2. How easy it is to evolve specialists for a task depends on the task’s sensitivity to environmental conditions. Individuals specialize most easily on tasks whose yields are relatively insensitive to environmental fluctuations because, for those tasks, it tends to be evolutionarily optimal to perform them most or all of the time (thereby allowing specialists to evolve even when efficiency benefits are small; SI Appendix , Fig. S6 ). For example, we predict that an undifferentiated cyanobacterium that can photosynthesize only during the day but can fix nitrogen at any time would require smaller efficiency benefits to evolve cells specialized for nitrogen fixation (a less environment-sensitive task) than to evolve cells specialized for photosynthesis (a more environment-sensitive task). Consistent with this, some species of cyanobacteria (e.g., Anabaena variabilis ) have evolved specialized nitrogen-fixing cells, while the remaining cells continue to perform both tasks: They photosynthesize during the day but also engage in environmentally regulated nitrogen fixation, for example at night ( 68 ). Similarly, in the context of reproductive division of labor, we predict that reproductive specialists (e.g., an ant queen whose only task is to lay eggs, or a stem cell in a multicellular tissue whose only task is to produce new cells) evolve most easily in environments where the demand for new individuals is relatively insensitive to environmental fluctuations. In contrast, a more balanced distribution of reproductive tasks [e.g., in ant colonies with worker reproduction ( 69 – 71 )] would be more favored in environments that require flexibility in the production of new individuals, for example, due to recurring but unpredictable disturbances in which substantial numbers of individuals are lost from the system. The optimal fraction of specialists varies across ecological conditions ( Fig. 5 B ). First, when the benefits of collective flexibility outweigh the efficiency gains of specialization, no specialists evolve ( s = 0 ), and all individuals are generalists, dividing their attention over both tasks in response to the environment as in the original model ( Fig. 5 C ). Second, in the absence of substantial benefits to collective flexibility, complete division of labor evolves in which half the individuals become task 1 specialists ( s = 0.5 ), and the remaining generalists perform only task 2—essentially becoming specialists as well, even though they do not derive an efficiency benefit ( Fig. 5 E ). Finally, in between these two extremes, a compromise between improved efficiency and collective flexibility emerges in which a minority of individuals are specialists ( 0 < s < 0.5 ), and the task they specialize on is also performed by generalists under some environmental conditions ( Fig. 5 D ). Thus, it can be evolutionarily optimal to employ a mix of generalists and specialists, thereby reaping some efficiency benefits of specialization while still being able to adjust to the environment. While, for simplicity, we have considered specialists that perform only task 1, we can relax this assumption by allowing specialists to also occasionally perform task 2, albeit less frequently than generalists ( SI Appendix , Fig. S7 ). As before, we find that evolving specialists requires the smallest efficiency gains for tasks whose yields are relatively insensitive to environmental conditions. However, in this more general setup, the evolution of specialists depends not only on gained efficiency for task 1 but also on whether and to what extent specialists incur a reduction in efficiency for task 2 as a result of specializing on task 1. Thus, in general, whether specialists for a task can evolve depends on the interplay between how strongly specialists bias their task performance toward it, how sensitive the task is to environmental conditions, and how gains in efficiency for the task trade-off with a loss in efficiency for other tasks ( SI Appendix , Fig. S7 ). Temporary inactivity can evolve to enhance collective flexibility. As a final application of our original model, we consider the possibility of inactivity. While we have so far assumed that individuals are continually active, individuals could also be temporarily inactive (i.e., not performing any task). Low levels of inactivity are not necessarily surprising (they could be explained by constraints on activity, e.g., the need to rest), but some systems show perplexingly high levels of inactivity ( 72 , 73 ). For example, Leptothorax allardycei ants appear to spend most (55%) of their time doing nothing ( 74 ). This raises the question of whether temporary inactivity could, in fact, be adaptive, despite inactive individuals not contributing to productivity. Multiple hypotheses have been suggested for why high levels of temporary inactivity could be beneficial ( 72 , 75 – 78 ). We use our model to explore one such hypothesis, which proposes that inactivity may be adaptive because it increases the system’s flexibility ( 72 , 77 ). Specifically, temporarily inactive individuals could enhance collective flexibility if they can be more quickly recruited (e.g., due to being more attentive to increased demand for any task) to tasks that require additional attention than individuals that are actively performing other tasks. We implement inactivity by letting active individuals quit their task at a rate τ , upon which they join an inactive pool. As in our task-switching model, we assume that individuals actively performing the other task are recruited to task i at rate δ z i ( Fig. 4 A ), but that inactive individuals are recruited to task i at rate κ δ z i ( SI Appendix , Fig. S8 A ). Setting κ > 1 incorporates the assumption that inactive individuals can be more quickly recruited to a task than active individuals performing a different task. Because individuals switch from being active to being inactive at rate τ and switch from being inactive to active at total rate κ δ , each individual spends, on average, a fraction τ / ( τ + κ δ ) of its time inactive. Thus, while the composition of the inactive pool will change over time, its total size (as a fraction of the total number of individuals) stabilizes at τ / ( τ + κ δ ) . In response to the environment, the system can adjust what tasks active individuals perform, but not what fraction of individuals are (in)active. To determine whether inactivity can evolve even though temporarily inactive individuals do not contribute to productivity, we calculate what combination of task quitting rate τ ∈ [ 0 , ∞ ) and z 1 , A , z 1 , B ∈ [ 0 , 1 ] maximizes fitness. Although task quitting is not required to achieve collective flexibility, we find that having a positive τ (and thereby a pool of temporarily inactive individuals) may indeed be evolutionarily advantageous ( SI Appendix , Fig. S8 B ). Moreover, high levels of inactivity (up to 25% of each individual’s time, in some cases) can evolve, resulting in a correspondingly sizable pool of inactive individuals at any one time (up to 25% of individuals; see SI Appendix , Fig. S8 D ). Thus, inactivity can indeed evolve to enhance collective flexibility, allowing the system to more rapidly adjust its task allocation to changing circumstances ( SI Appendix , Fig. S8 C and E ). We expect inactivity to evolve for this reason when collective flexibility is adaptive but not fully realizable due to constraints on the rate at which the system can adjust its task allocation ( SI Appendix , Fig. S8 B and D ). These results provide theoretical support for the previously proposed hypothesis that observed worker inactivity in social insects could be a component of an adaptive colony-level task allocation strategy ( 72 , 77 ). Specifically, they confirm that high levels of inactivity could have evolved to allow colonies to adjust more rapidly what task is being performed by active workers in response to environmental fluctuations, even when the number of active workers stays constant over time. This hypothesis is consistent with empirical evidence from some species of social insects, where temporarily inactive workers have been confirmed to act as a “reserve” labor force [i.e., Temnothorax rugatulus ants ( 76 )], although other explanations might be required for inactivity in other species ( 61 , 76 – 79 )."
} | 11,551 |
37106823 | PMC10135469 | pmc | 1,035 | {
"abstract": "Simple Summary In order to counter the increased deposition of greenhouse gases in the atmosphere, which has resulted in significant climatic changes, the production of alternative fuels to replace conventional fossil fuels has become necessary due to the rapidly diminishing concentration of fossil fuels and the rising global demand for energy. This investigation focused on two different samples that could be used as anolytes to produce energy in single- and double-chamber microbial fuel cells with a graphite electrode. In microbial fuel cells’ energy production, the microbes consume organic substrates, use them in their metabolic processes, and produce valuable products that can be used as fuel to produce energy. The highest voltage outputs from the investigated bacterial strains were generated by strains A1 and A2, at 402 mV and 350 mV, respectively. Strain A6 of the ten different bacterial strains produced the least electricity, with a measurement of 35.03 mV. Abstract Natural resources are in short supply, and the ecosystem is being damaged as a result of the overuse of fossil fuels. The creation of novel technology is greatly desired for investigating renewable and sustainable energy sources. Microorganisms have received a lot of interest recently for their potential to transform organic waste into sustainable energy and high-value goods. New exoelectrogens that can transmit electrons to electrodes and remove specific wastewater contaminants are expected to be studied. In this study, we examined three distinct samples (as determined by chemical oxygen demand and pH) that can be used as anolytes to generate power in single-chamber and double-chamber microbial fuel cells using graphite electrodes. Wastewater from poultry farms was studied as an exoelectrogenic anolyte for microbial fuel cell power generation. The study examined 10 different bacterial strains, numbered A1 through A10. Due to their highly anticipated capacity to metabolize organic/inorganic chemicals, the diverse range of microorganisms found in poultry wastewater inspired us to investigate the viability of generating electricity using microbial fuel cells. From the investigated bacterial strains, the highest voltage outputs were produced by strains A1 ( Lysinibacillus sphaericus) and A2 ( Bacillus cereus ), respectively, at 402 mV and 350 mV. Among the 10 different bacterial strains, strain A6 generated the least amount of electricity, measuring 35.03 mV. Furthermore, a maximum power density of 16.16 1.02 mW/m 2 was achieved by the microbial fuel cell using strain A1, significantly outperforming the microbial fuel cell using a sterile medium. The strain A2 showed significant current and power densities of 35 1.12 mA/m 2 and 12.25 1.05 mW/m 2 , respectively. Moreover, in the two representative strains, chemical oxygen demand removal and Coulombic efficiency were noted. Samples from the effluent anode chamber were taken in order to gauge the effectiveness of chemical oxygen demand removal. Wastewater had an initial chemical oxygen demand content of 350 mg/L on average. Strains A1 and A2 decomposed 94.28% and 91.71%, respectively, of the organic substrate, according to the chemical oxygen demand removal efficiency values after 72 h. Strains A1 and A2 had electron donor oxidation efficiencies for 72 h of 54.1% and 60.67%, respectively. The Coulombic efficiency increased as the chemical oxygen demand decreased, indicating greater microbial electroactivity. With representative strains A1 and A2, Coulombic efficiencies of 10% and 3.5%, respectively, were obtained in the microbial fuel cell. The findings of this study greatly advance the field as a viable source of power technology for alternative energy in the future, which is important given the depletion of natural resources.",
"conclusion": "5. Conclusions In this study, a single-chamber MFC containing poultry wastewater was used to isolate putative electroactive microorganisms. Two Gram-positive bacterial strains, Lysinibacillus sphaericus (A1) and Bacillus cereus (A2), were isolated and their electrogenic capacities were investigated. To be more precise, this study used a double-chambered MFC to enrich graphite anodes with different electroactive bacteria. The isolated strain A1 (Lysinibacillus sphaericus) was electrochemically characterized in the study as a facultative anaerobic electrogenic bacterium. The COD removal efficiency over 72 h, according to the Lysinibacillus sphaericus A1 findings, was 94.28%. A maximum power density of 16.16 mW/m 2 and a CE of 10% were obtained in the MFC. Bacillus cereus A2, the second novel electroactive isolate, likewise produced promising outcomes. Bacillus cereus had a 72 h COD removal efficacy of 91.71%, and a maximum power density of 12.25 ± 1.05 mW/m 2 . As a result, the results of this study point to Lysinibacillus sphaericus A1 as a strong candidate for the design and development of MFCs for energy production. These isolated electroactive bacteria were placed in long-term storage for further, more detailed studies. To improve the efficiency of isolated electrogens, other carbon sources included in artificial or real wastewater can be evaluated. Finding more effective electrogenic organisms requires further research into the diversity of accessible electrogenic microbes as well as the methods employed to transfer extracellular charge to electrodes. This study will contribute to a better understanding of the electrogenic potential of pathogens present in the avian microbiome, which is currently largely uncharacterized. The findings of this study also broaden the knowledge of exoelectrogens for energy generation, and the vast range of substrates used by the strains raises the possibility for MFC applications in waste management and renewable energy production.",
"introduction": "1. Introduction The need for electricity has risen during the last few years. Fossil fuels and nuclear power are two nonrenewable energy sources which are widely employed worldwide. Fossil fuels are the source of energy that cause the most environmental harm since they continuously release carbon dioxide into the atmosphere, which becomes hazardous when there is an excessive amount of it [ 1 ]. Through air pollution and global warming, the quick depletion of fossil fuels has significantly impacted human life [ 2 ]. However, several countries have made remarkable efforts to discover a workable solution to the energy problem by concentrating on renewable energy sources, such as solar energy, water energy, and wind energy [ 3 , 4 , 5 ]. Through the use of highly valuable metal catalysts, these experiments have revealed a novel method for producing energy using a fuel cell [ 6 ]. In actuality, there are numerous advantages to adopting fuel cells over other energy providers, including improved efficiency, the absence of mobile components that cause less sonic “pollution”, and the release of zero environmentally harmful gases, such as CO 2 , CO, NO x , and SO x . On the other hand, these new energy sources have two drawbacks: high cost and low mass production. This is where the concept of a microbial fuel cell (MFC) becomes significantly useful [ 7 ]. It is important to remember that the microbial fuel cells (MFCs) approach is a novel bioelectrochemical method that tries to generate energy by utilizing the electrons obtained from biological reactions facilitated by bacteria [ 8 ]. It is also worth noting that electroactive microorganisms have sparked a lot of interest in the creation of novel biotechnological systems with minimal environmental impact. They can be applied to the creation of value-added goods, the bioremediation of ecosystems contaminated with metals, and sustainable energy production [ 9 ]. It is also important to highlight that due to their extensive freshwater use for the continuous operations of cutting up, washing, and packaging meat, poultry slaughterhouses release enormous amounts of wastewater into the environment. Additional processes used in poultry slaughterhouses, such as scalding, de-feathering, evisceration, and bird washing, also consume a lot of water and produce a lot of wastewater. According to the literature, a 2.3 kg bird will typically drink 26.5 L of water each day [ 10 , 11 ]. The biochemical oxygen demand (BOD) and chemical oxygen demand (COD) measurements show that the effluent from poultry slaughterhouses is significantly polluted with organic materials. Blood, fats, oils, grease, and proteins are among the other components that have significant nitrogen and phosphorus content in poultry slaughterhouse wastewater [ 12 ]. Hence, there is a significant risk of the pollution of freshwater sources when inadequately treated poultry wastewater is discharged. The deoxygenation of rivers, groundwater contamination, eutrophication, and the development of water-borne diseases are just a few of the significant environmental and health problems this can lead to [ 13 ]. There is a high likelihood of finding novel electrogenic bacteria since poultry wastewater has a high pollutant loading level. Electrogenic bacteria are a class of microorganisms that can transfer electrons extracellularly through the cell envelope to or from electron acceptors such as electrodes, oxide minerals, and other bacteria under anaerobic or microaerobic conditions [ 14 ]. The early substrates utilized in the lab were mostly glucose, acetate, or other straightforward substrates to ascertain the behavior of electrode materials, membranes, and other such things, as well as the reactor architecture or microbial activity [ 15 , 16 , 17 ]. Just recently, investigations employing actual wastewater as a substrate have been carried out. The energy savings from sludge treatment and wastewater aeration were the biggest benefit. The output of sludge by MFCs is also lower than that of anaerobic digesters and aerobic-activated sludge (AS) treatment systems. These have reduced temperature sensitivity, limited electrical installations at sludge treatment facilities, and no energy used for aeration [ 18 ]. Fundamentally, wastewater is the most widely used substrate for MFC operations because of its large proportion of organic load and lack of cost. In particular, agro-food wastewater is ideal due to its high biodegradability [ 19 , 20 , 21 ]. The numerous electroactive and complementary non-electroactive microorganisms convert the chemical energy stored in the chemical components of wastewater or biomass into electrical energy [ 22 ]. With its low thermal efficiency, the Carnot thermodynamic cycle in an ideal thermal machine is avoided by this direct conversion of chemical energy to electrical energy. Theoretically, MFCs are comparable to traditional fuel cells in that they can achieve higher efficiency. Additionally, because wastewater is used, it is a renewable energy source, and its “fuel” supply can be controlled relatively more easily than that of wind turbines, where the wind cannot be controlled at all, and photovoltaics, where the sun’s rays cannot be controlled but are predictable [ 23 ]. A swine wastewater treatment facility was suggested in the study by Ding et al. [ 24 ], based on single-chamber air-cathode MFCs with a solution capacity of 340 mL on a laboratory scale, as well as a separate low-cost flocculation process. According to the findings, an energy recovery rate of roughly 0.664 kWh/m 3 wastewater mixture was attained. Additionally, 96.6% removal efficiency of COD, 60% removal of ammonia, 37.5 W/m 3 power density, and 21.6% Coulombic efficiency were attained. However, despite its potential, the best configuration for MFCs is still being researched, and efforts are now being made to improve its performance by developing more selective proton exchange membranes and alternative electrode materials. Small cells connected in series appear to offer higher potentials than larger reactor volumes [ 25 ]. The expense of materials and residential wastewater’s poor buffering capability are currently the main obstacles to the full-scale application of MFCs. Due to this, MFCs have not yet found use in the industry [ 25 ]. It is also important to emphasize that the composition of the wastewater and the type of electrode materials utilized can have a significant impact on the effectiveness of these systems. In the MFC, wastewater-growing microorganisms consume the organic substrate and release electrons that are then used to enhance the process of generating energy [ 26 ]. In order to break down the substrate and produce energy, the type of microbial population in the biofilm is essential. Several sources of microbial inocula, such as bagasse-based paper mill wastewater [ 27 ], fresh sediment [ 28 ], dye processing wastewater [ 29 ], marine sediment [ 30 ] as well as sludge [ 31 ] have been used to successfully generate microbial colonies that can transfer their electrons. The choice and appropriateness testing of the inoculum source utilized in the MFC is critical because the kind of microbial population in the biofilm is crucial to both substrate breakdown and energy production. Despite being one of the largest producers of highly polluted wastewater, the poultry industry has unfortunately not been thoroughly examined in relation to MFC energy production. Excreta, feathers, spilled feeds and water, dead birds, cracked eggs, wastewater, litter or bedding materials, and waste from the slaughterhouse and hatcheries are all included in poultry wastes or manure [ 32 ]. In this work, attention is paid to enriched biofilm communities that contain bacteria capable of donating electrons to the anode as their final electron acceptor, termed anode-respiring bacteria (ARB). To realize the full potential of MFC technologies, it is important to study the different organisms and mixed communities capable of anodic respiration so that a wider range of metabolic processes can be found, understood, and exploited. This can only be achieved if we expand our search for a new ARB beyond the locations investigated so far. To be more specific, this study assesses the effects of isolated bacteria from anode plates of single-chambered MFC with poultry wastewater samples on electricity generation of the double-chambered MFC.",
"discussion": "4. Discussion The environmental load of wastewater and other waste streams has increased due to an increase in untreated environmental outputs from industries and an increase in the human population. The activated sludge process, which requires aeration and is consequently energetically and financially expensive, is one of the most widely used wastewater treatment techniques [ 42 ]. The recovery of waste materials as resources is made easier by the present movement toward a circular economy. The complexity of wastewater is also growing daily, in addition to its volume. As a result, technologies for wastewater treatment that are both affordable and sustainable must be developed. Moreover, electrochemical technologies, such as fuel cells, exhibit tremendous future promise as power technologies for alternative energy sources due to the depletion of natural resources. The MFC, which generates bio-electricity from various organic fuel sources, is one such promising invention. In order to create power through waste treatment, MFCs use electroactive bacteria to extract chemical energy from used organic molecules [ 43 ]. This study worked on the electrochemical characterization of bacteria isolated from poultry wastewater. Possible electroactive cultures were isolated by inoculating a smear collected from the surface of a graphite anode onto Petri dishes with MH medium after the best anolyte, which displayed the highest voltage, was identified. As a consequence, 10 isolates of the bacterial cultures (A1 A2, A3, A4, A5, A6, A7, A8, and A10) were discovered. These isolates underwent a 72 h test for electrical activity. Of the 10 isolated pure cultures of microorganisms, 7 were facultative anaerobes. The remaining three species of bacteria of the genus Arthrobacter, Rhodococcus, Pseudomonas were obligate aerobes. Among the seven facultative anaerobic bacteria, bacteria of the genera Lysinibacillus and Bacillus cereus, which are the closest in the phylogenetic tree and have similar metabolic processes, showed electroactivity. Both genera of bacteria are capable of high consumption of acetate. Lysinibacillus sphaericus comprises a group of motile Gram-positive spore-forming bacilli. Members of this group are characterized by their terminal endospore, the capability to utilize acetate as the sole carbon source, and the presence of lysine and aspartic acid in their cell wall peptidoglycan. According to literature sources, bacteria Lysinibacillus sphaericus are capable of adsorbing toxic metals (cadmium, lead, arsenic, mercury, chromium) and precious metals. The metal ions become attached to the functional groups or the protein layer, followed by the binding of metal ions to the reactive groups present on the bacterial cell wall; the internalization of metal ions occurs inside the cell. Transcriptomic, proteomic, and electrochemical analyses show that the electrode respiration of Lysinibacillussphaericus mainly depends on electron mediators, and c-type cytochromes may be involved in its respiration. Extracellular electron transport (EET) is a key driving force in biogeochemical element cycles and microbial chemical–electrical–optical energy conversion on Earth. Gram-positive bacteria are ubiquitous and even dominant in EET-enriched environments. However, attention and knowledge of their EET pathways are largely lacking. The Gram-positive bacterium Lysinibacillus has extremely long cells (>1 mm) and conductive nanowires, promising a unique and enormous role in the microenvironments where it lives. Furthermore, the isolation of the potential electrogene makes it possible to develop an electrochemical strategy for connecting and forming the surrounding microbial community on a minimal scale. Spore-forming bacteria belonging to the genus Bacillus have the ability to release flavins. These flavins allow Bacillus sp. to mediate electron transfer to electrodes in MFCs and provide increased power generation in microbial consortia with Gram-negative bacteria or yeast. The incorporation of Bacillus cells into anaerobic sludge has also had a significant impact on power generation in MFCs. By promoting the formation of an electroactive biofilm and inhibiting methanogenesis, Bacillus cereus enhances current production in the MFC. As noted from the results, in the first 17 to 19 days of operation, the performance of a single-chambered MFC using three poultry wastewater anolytes reached a maximum voltage of 420 mV and a maximum current density of 41.8 mA/m 2 . As said before, an MFC may produce power directly from a wide range of organic or inorganic chemicals by acting as a catalyst and using microbes. Traditionally, fuel cells use an oxidant at the cathode and a fuel at the anode to transform chemical energy into electrical energy. Electricity is generated as a result of the liberated electrons and protons moving through an external circuit. MFCs use an organic matter and microbial fuel solution, with the anode and cathode separated by an ion exchange membrane [ 44 ]. However, the direct passage of electrons from the bacterium to the anode significantly reduces efficiency. Therefore, exogenous mediators such as thionine, methyl viologen, and humic acid are utilized in electrochemically inactive microbial cells. These serve as electron shuttles, diffusing electrons to the anode, allowing them to discharge, and then diffusing them back to the bacterial cells. However, these mediators are extremely expensive and harmful to microorganisms [ 45 ]. In addition, tests of two isolated bacterial cultures designated strains A1 and A2, produced the greatest voltage outputs (402 mV and 350 mV, respectively) and were significantly different from the strains of the other eight bacterial isolates and controls. Ten different bacterial strains were tested, and strain A6 produced the least electricity at 35.03 mV. In addition, a parallel investigation of the current and power densities was carried out in order to characterize electrical generation completely. When compared to MFCs using a sterile medium, strain A1 dramatically enhanced performance, with a maximum power density of 16.16 1.02 mW/m 2 . In terms of current density and power density, strain A2 displayed high values of 35 ± 1.12 mA/m 2 and 12.25 ± 1.05 mW/m 2 , respectively. Koffi and Okabe [ 46 ] claimed that an MFC may commonly provide power densities between 1 and 2000 mW m- 2 . So far, the low voltage or power issue has been solved by simply connecting numerous MFCs in series or parallel. Although a unit of serially stacked MFCs may deliver a larger voltage, it has frequently been found to be challenging and unsuccessful because individual the tendency of MFC units to switch polarity due to fuel shortages causes a sizable overall voltage decay. A stacked polarized capacitor has recently been charged using individual MFC units connected to an MPPT system in an effort to manage and reduce voltage reversal occurrence. However, only voltages between 2 and 3 V could be increased using this technical method [ 46 ]. According to the results of COD removal efficiency after 72 h, strains A1 and A2 converted 91.71% and 94.28% of the organic substrate, respectively. The electron donor oxidation efficiencies of strains A1 and A2 for 72 h were 54.1% and 60.67%, respectively. Additionally, the results showed that Coulombic efficiency increased when COD was decreased, indicating an increase in microbial electroactivity. The representative strains A1 and A2 were then selected for further investigation. It should be underlined that COD is a measurement of the amount of oxygen used during the oxidation of oxidizable organic matter when a potent oxidizing agent is present. The number of organic compounds in wastewater are typically estimated indirectly using this method. High COD indicates the presence of all organic matter types, both biodegradable and nonbiodegradable, and subsequently the level of pollution in the water. Because of this, COD can be used to detect organic pollution in surface waters [ 47 ]. The results of COD removal were found to be comparable to those of Li et al. [ 48 ], whose work sought to identify and characterize a COD-degrading bacterium that can efficiently break down slaughter wastewater. In their investigation, six COD-degrading bacteria were found in the sludge and wastewater from the Hunan meat industry’s slaughterhouses. Bacillus velezensis was found and categorized as the strain with the highest COD degradation rate through morphological observation and 16S rDNA sequence analysis, reaching 11.80%. It should be mentioned that the potassium permanganate method was used to assess the COD breakdown rate of each strain. The acquired pure bacterial cultures were characterized phenotypically after being inoculated on Mueller–Hilton agar. A variety of non-picky organisms can be grown on the Mueller–Hinton agar, which is a non-selective, non-differential medium. It is referred to as a “loose” agar, which works better than other forms of media to mediate the rate of antibiotic diffusion. On a nutritional medium, isolate A1 colonies are opaque, dark yellow, smooth, and shiny. The bacteria from the A2 isolate are rod-shaped and Gram-positive. Bacteria that are both sporogenous and non-sporogenous make up the diverse group known as Gram-positive rods. The genus Bacillus is made up of aerobic, sporogenous organisms that are Gram-positive. There are several species in this, but Bacillus anthracis, the agent of anthrax, is the most significant from a medical and veterinary standpoint. Isolate A2 colonies are rounded and rough, have wavy edges, are opaque, and are a fuzzy white or slightly yellow tint. Rod-shaped and Gram-positive bacteria make up the A2 isolate. It is also crucial to emphasize that the production of alternative fuels to replace conventional fossil fuels has become necessary due to the rapidly diminishing concentration of fossil fuels and the increasing global demand for energy [ 49 ]. This is carried out in order to counter the increased deposition of greenhouse gases in the atmosphere, which has resulted in significant climatic changes. Rising temperatures and sea levels are just two of the potentially disastrous effects of these changes. Additionally, the findings of the double-chambered MFC tests on the isolates’ energy performance were contrasted with those of well-known electroactive bacterial strains ( Table 2 ). Based on the results, it is clear that the density of the Lysinibacillus sphaericus A1 and Bacillus cereus A2 strains were several times less than that of the Shewanella oneidensis MR-1 production strain and less than the well-known Lysinibacillus spharicus VA5, Lysinibacillus sphaericus D-8. However, the Coulombic efficiency was higher than that of the Lysinibacillus sphaericus VA5, which was explained by the effectiveness of the removal of the COD from the Lysinibacillus sphaericus A1 strain. It should be noted that the electrochemical and Coulombic efficiency of the selected Lysinibacillus sphaericus A1 strain was higher than that of the well-known Corynebacterium sp. MFC03 strain. Furthermore, the Bacillus cereus A2 strain showed severe electrical activeness and was not inferior in producing current density and power compared to such an electric active strain as Corynebacterium sp. strain MFC03. These comparative indicators of the above strains indicate a high electrical potential of the Lysinibacillus sphaericus A1 and Bacillus cereus A2."
} | 6,455 |
37535207 | PMC10400695 | pmc | 1,036 | {
"abstract": "Main conclusion Nicotiana attenuata’s capacity to interact with arbuscular mycorrhizal fungi influences its intraspecific competitive ability under field and glasshouse conditions, but not its overall community productivity. Abstract Arbuscular mycorrhizal (AM) fungi can alter the nutrient status and growth of plants, and they can also affect plant–plant, plant–herbivore, and plant–pathogen interactions. These AM effects are rarely studied in populations under natural conditions due to the limitation of non-mycorrhizal controls. Here we used a genetic approach, establishing field and glasshouse communities of AM-harboring Nicotiana attenuata empty vector (EV) plants and isogenic plants silenced in calcium- and calmodulin-dependent protein kinase expression (ir CCaMK ), and unable to establish AM symbioses. Performance and growth were quantified in communities of the same (monocultures) or different genotypes (mixed cultures) and both field and glasshouse experiments returned similar responses. In mixed cultures, AM-harboring EV plants attained greater stalk lengths, shoot and root biomasses, clearly out-competing the AM fungal-deficient ir CCaMK plants, while in monocultures, both genotypes grew similarly. Competitive ability was also reflected in reproductive traits: EV plants in mixed cultures outperformed ir CCaMK plants. When grown in monocultures, the two genotypes did not differ in reproductive performance, though total leaf N and P contents were significantly lower independent of the community type. Plant productivity in terms of growth and seed production at the community level did not differ, while leaf nutrient content of phosphorus and nitrogen depended on the community type. We infer that AM symbioses drastically increase N. attenuata ’s competitive ability in mixed communities resulting in increased fitness for the individuals harboring AM without a net gain for the community. Supplementary Information The online version contains supplementary material available at 10.1007/s00425-023-04214-z.",
"introduction": "Introduction Interactions between plants of the same or different genotype(s) are influenced by their microbial partners, in particular, arbuscular mycorrhizal (AM) fungi belonging to the subphylum Glomeromycotina (Spatafora et al. 2016 ). This plant–AM fungal interaction is thought to be based on an exchange of nutrients. Due to their thin hyphae and extensively branched networks, AM fungi improve a host plant’s acquisition of fitness-limiting mineral nutrients, in particular phosphorus [up to 90% of its total phosphorous (P) uptake (Smith and Read 2008 )], but also other minerals such as nitrogen (N), manganese, potassium, magnesium, zinc, and water (González-Guerrero et al. 2016 ; Garcia et al. 2016 ; Wipf et al. 2019 ). The plants provide up to 20% of their photoassimilates in the form of carbohydrates (Smith and Read 2008 ) and lipids to the fungus (Bravo et al. 2017 ). Mycorrhiza-specific phosphate transporters are induced during fungal colonization, enabling an indirect route of P uptake that complements direct P uptake via roots (Smith et al. 2004 ). Furthermore, recent studies show that not only active inoculum but also the AM fungal necromass increases plant growth (Jansa et al. 2020 ). AM fungi can also improve a plant’s resistance against biotic and abiotic stressors, such as drought, heat, pathogen or herbivore attack (Santander et al. 2017 ; Rivero et al. 2018 ; Sanmartin et al. 2020 ), providing additional routes to altered plant performance from the interaction. The vast majority of land plants interact with AM fungi (Brundrett and Tedersoo 2018 ). The earliest described interaction dates to more than 400 million years ago (Strullu-Derrien et al. 2018 ), and the interaction may have been crucial for the colonization of land by plants (Brundrett and Tedersoo 2018 ). The molecular dialogue between plants and AM fungi starts before their physical interaction and is governed by a conserved set of genes for fungal recognition and signaling that are required for the establishment of a functional symbiosis (e.g., MacLean et al. 2017 ; Wipf et al. 2019 ; Xue and Wang 2020 ). One key element is a calcium- and calmodulin-dependent protein kinase (CCaMK) located in the nucleus and thought to decode a Ca 2+ signal to establish a complex with IPD3 (INTERACTING PROTEIN OF DOES NOT MAKE INFECTIONS 3, MacLean et al. 2017 ). Plants impaired in the expression of CCaMK do not establish a functional symbiosis (Levy et al. 2004 ). The costs/benefits of AM fungal inoculation for a host plant are commonly evaluated by growing plants with and without AM fungi to calculate the differences in growth and fitness, such as biomass, flower and seed numbers, between inoculated and non-inoculated plants (Chaudhary et al. 2016 ). The difference is also termed “response” or “responsiveness” (Wilson and Hartnett 1998 ; Konvalinková and Jansa 2016 ) and many studies have revealed a large variation in the response of plants, with outcomes not always positive, but neutral or negative, depending on the environment (Johnson 2010 ). Environmental stress factors (low light, high P), but also the plant and fungal genotype, can determine if harboring AM fungi is beneficial for plants (Berger and Gutjahr 2021 ; Bennett and Groten 2022 ). Differences in response are not only found at the species level, but also within the same species. Natural plant populations consist of many different genotypes, often in the same area. These genotypes may vary in their response to AM fungal colonization as shown for many crop varieties, e.g., wheat, barley, maize (Sawers et al. 2017 ; Elliott et al. 2020 ; Thirkell et al. 2020 ). For example, for maize, it was shown that different mycorrhiza-associated maize lines vary in P acquisition efficiency and in their expression of P-transporter genes (Sawers et al. 2017 ). In addition to these direct effects at the individual plant level, AM fungi may alter plant–plant competition and facilitation of niche partition (Bever et al. 2012 ; Hodge and Fitter 2013 ). AM fungi colonize more than 72–80% of all land plant species (Brundrett and Tedersoo 2018 ). Their fungal external hyphae spread out in the surrounding soil and can colonize neighboring plants of the same or different species at the same time, thus creating a so-called common mycorrhizal network (CMN) (Barto et al. 2012 ). The CMN may modify the distribution of limiting mineral nutrients among plants (Lekberg et al. 2010 ; Weremijewicz et al. 2016 ), but also exchange information, e.g., about pathogen or herbivore attack of neighboring plants (Babikova et al. 2014 ; Song et al. 2014 , 2019 ). In vitro experiments have revealed that AM fungi move P from P-rich to P-poor nutrient patches (Whiteside et al. 2019 ). Differences in sink strengths and changes in nutrient transfer belowground may influence the symmetry of competition among networked plants (Montesinos-Navarro et al. 2018 ; Whiteside et al. 2019 ; van’t Padje et al. 2021 ). It is assumed that under limiting conditions, plants of different species are more likely to require different resources than those of the same species (Mayfield and Levine 2010 ), and different plant species may differ in the efficiency with which they interact with the same fungal partner (van’t Padje et al. 2021 ), thus reducing competition. If an individual in a population or a species in a diverse community performs better in terms of growth and fitness than its neighbors, it may alter the plant community dynamics (Bray et al. 2003 ), plant diversity, and/or ecosystem productivity (van der Heijden et al. 2015 ). Over time, these differences in response can increase or decrease the abundance of plants. Thus, AM fungi may alter the structure of plant communities (Bennett and Cahill Jr 2016 ; Tedersoo et al. 2020 ). A number of studies investigated how resource distributions and plant–plant competition are modified when different plant species are connected to the same CMN (Walder et al. 2012 ; Weremijewicz and Janos 2013 ; Milkereit et al. 2018 ). For example, AM fungal inoculation enhanced the competitive ability of perennial species competing with annual species (Lin et al. 2015 ). Using isotopically labeled P, N, and carbon with a flax–sorghum CMN, the C 3 flax was found to realize the largest benefit compared with the C 4 sorghum, investing little, but gaining the most P and nitrogen provided by the CMN (Walder et al. 2012 ). Less is known about how AM colonization and connections to a CMN affect intraspecific competition. Most studies examined the consequences for unequal competition between different-sized individuals. Some studies focused on CMNs and their effect on seedling growth of the same species: larger individuals benefit at the expense of smaller neighbors (Janoušková et al. 2011 ; Weremijewicz and Janos 2013 ). Large Andropogon gerardii plants in the sun received more mineral nutrients than smaller, shaded conspecifics, leading to asymmetric belowground competition (Weremijewicz et al. 2016 ). These studies looked at the costs and benefits of an individual plant compared to a neighbor within a CMN. However, effects might scale-up beyond the individual plant to influence population-level yields (McGale et al. 2020 ). Furthermore, using different-sized plants with different growth traits or even different species will complicate the assessment of mycorrhizal effects. A major limitation of previous work has been the sole reliance on glasshouse studies to experimentally control AM colonization. The glasshouse enables comparisons between inoculated and non-inoculated plants and the ability to separate roots of different plants, but it has numerous disadvantages for the analysis of this highly context-dependent interaction (Bennett and Groten 2022 ). First, plants are commonly grown in pots in glasshouse studies where they rapidly become pot-bound, unable to freely forage for nutrients, and unnaturally increase competition and other root–root interactions. Moreover, many studies use only a single fungal species for inoculation (Koricheva et al. 2009 ; Hoeksema et al. 2010 ). Plants grown in nature can usually explore a much larger soil volume with fewer direct root–root interactions (e.g., see root biomass of glasshouse and field-grown plants in Groten et al. 2015 ), and roots interact with a high diversity of different AM fungal species. For most field experiments, non-inoculated controls are either missing or confounded by the treatments used to kill the fungal partner (fungicidal fumigation, heat treatment) which alter nutrient or microbiome compositions in addition to the AM fungal association. This problem can be solved using (near)isogenic lines with different capacities to interact with AM fungi, and some have been used in a paired plant in a pot comparison (Facelli et al. 2014 ; Groten et al. 2015 ; Bowles et al. 2016 ; Fabiańska et al. 2020 ), but rarely have these pot-bound approaches been scaled-up to rigorously test the effects of AM fungi under natural conditions. Here we elucidate the importance of AM fungi for individual plants connected or not connected by a CMN to neighboring plants of the same species. We hypothesized that: (i) AM increase plant–plant competition so that size inequalities would be amplified when plants differ in responsiveness to AM fungi, and (ii) the presence of non-mycorrhizal plants of the same species would change the growth and fitness outcomes for populations. We used a wild tobacco species, Nicotiana attenuata , as a model plant, because its natural history, as well as its interaction with bacterial and fungal microbes, is well-described (Santhanam et al. 2014 , 2015 ; Groten et al. 2015 ; Schuman and Baldwin 2016 ). After wildfires, the plants naturally occur in large monocultures (Lynds and Baldwin 1998 ) of different accessions differing in their chemical composition (Kallenbach et al. 2012 ; Li et al. 2016 ) and their capacity to interact with AM fungi (Wang et al. 2018b ); hence, intraspecific competition is very high. This system allows us to study the effect of AM within one species under natural conditions. We established microcosms in the glasshouse (Song et al. 2019 ) and small plant communities in the plant’s natural habitat where plants differed in their capacity to create a CMN. Nicotiana attenuata plants impaired in the interaction with AM fungi due to the RNAi silencing of a calcium- and calmodulin-dependent protein kinase (Groten et al. 2015 ; Wang et al. 2018b ) were compared against isogenic transformation control plants using a fully functional empty vector (EV). The RNAi silenced plants harbor only a single insert, and three independently transformed lines were evaluated in the glasshouse and field in an earlier study giving similar results (Groten et al. 2015 ). Plants were grown either in initially size-matched monoculture communities or communities consisting of the two lines. We measured growth (rosette size, stalk length, branch lengths and numbers, photosynthesis and nutrient contents) and fitness parameters (flower, capsule, and seed numbers) of plants from the three community types in the glasshouse and field. While individual plants were affected by capacity of their neighbors to interact with AM, growth performance at community level was similar for monocultures and mixed cultures.",
"discussion": "Discussion We manipulated a plant’s capacity to connect to a mycorrhizal network to evaluate whether AM colonization modulates neighbor competition for nutrients and if an asymmetry in neighbor competition within a community translates into differences in the productivity of plant communities of the same species. We used two isogenic wild tobacco lines to answer these questions: one line was impaired in interacting with AM fungi due to silencing a key gene of the symbiotic pathway, CCaMK , while the second line was fully functional. Both were shown previously to have a similar growth when grown in single pots without inoculum, a prerequisite for this type of comparative study (Rillig et al. 2008 ). Our system mimics the extremes of nature—the same species, either interacting and connecting or not to a CMN. In natural populations, gradients of AM root colonization and plant responses are commonly found (Wang et al. 2018a ; Sawers et al. 2017 ; Savary et al. 2020 ). For N. attenuata, it was shown that two different accessions (Utah, UT and Arizona, AZ) and the resulting crossing population differ in their mycorrhizal colonization and in the accumulation of the arbuscule-marker carboxyblumenol glucoside in the glasshouse and in the field (Wang et al. 2018a ), underlining the ecological importance of our study. Here, we used two experimental systems—an experimental field and a microcosm system—to test our hypotheses. In contrast to our previous study (Wang et al. 2018b ), we not only focused on two plants growing in the same small pot, but we created small communities. In the field, the roots could forage unconstrained by pot dimensions for nutrients, connect with neighboring plants and they were exposed to the natural AM fungal community. This is important, as population-level responses that reveal how AM affects plant coexistence and community structure can only be fully addressed in the field (Hart et al. 2003 ). The additional microcosm approach enabled us to rule out direct effects of root–root competition, and we increased the volume of the pots (4 × for a single plant) to have less pot-constrained root growth. Furthermore, neighbor recognition that might be independent of the AM fungal association is minimized due to the root-free inoculum compartment separating the root systems and thus strongly reducing information about their neighbor’s identity and performance based on belowground root exudates and volatiles (Semchenko et al. 2018 ; Kong et al. 2018 ). Similarly, specific bacterial communities which may also alter growth performance (Chen et al. 2012 ; Gfeller et al. 2019 ) and microRNAs produced by plants that may act as signaling molecules affecting gene expression in other, nearby plants (Betti et al. 2021 ), are unlikely to be exchanged. The results revealed that AM colonization modulates the competitive relationship between plants of the same species differing only in their capacity to interact with AM fungi. In both the field and the glasshouse, AM associations led to an asymmetry in growth and fitness of plants connected to a functional network compared to those not connected. ir CCaMK plants grown adjacent to an AM-colonized neighbor had significantly smaller roots, shoots and stalk lengths, and a corresponding lower rate of photosynthesis and chlorophyll content. The loss of access to fungal partners in ir CCaMK plants was reflected in a performance gain of neighboring fully functional empty vector plants. These results corroborate and extend the findings of a recent pot study growing the same lines together in a single pot (Wang et al. 2018b ). In the previous study, growth differences were much more pronounced, most likely due to higher competition in the same pot with highly limited amounts of nutrients, while in the field and in larger pots without direct root–root interactions, competition is more relaxed. In accordance with this assumption, the plants unable to interact with AM fungi and competing with fully functional plants were clearly P starved (Wang et al. 2018b ). Similarly, in tomato and Medicago , fully functional wild-type plants outcompeted mutant lines that acted as non-AM fungal surrogate hosts when grown in the same pot in the glasshouse (Facelli et al. 2010 , 2014 ). In all these studies, the AM-harboring partner pre-empted limiting resources. The difference in water content was not significant, but there was a tendency for reduced water content in ir CCaMK plants within mixed communities. Recently, it was shown that mycorrhizal fungi can transport water to their host plants, even from sources that cannot be accessed by plant roots (Wang et al. 2022 ). It was estimated that AM may contribute more than 40% of the water transpired by plants (Kakouridis et al. 2022 ). Hence, the interaction with AM fungi seems to enable plants to extract the limited water resources more efficiently compared to neighboring ir CCaMK plants. In contrast to the mixed cultures, we did not find significant growth and fitness differences among the lines grown in monoculture. This means that as long as the neighbors have the same “handicap” or advantage, the outcome in terms of growth and fitness is similar. The similarity in growth among the two monocultures in the field and a similar result in the glasshouse also indicates that the adjustments in the field setups between the two field seasons do not affect the questions addressed in this study. Furthermore, in contrast to the monocots rice and maize that show CCaMK expression in leaves and have a role in antioxidant defense and ABA signaling (Ma et al. 2012 ; Shi et al. 2012 ) in N. attenuata , CCaMK is not expressed in leaves (see http://nadh.ice.mpg.de/NaDH/ , NIATv7_g40298; Brockmöller et al. 2017 ). The lack of growth differences between the two lines under non-competitive conditions is in accordance with previous results using the same lines in the field and glasshouse (Groten et al. 2015 ) and corroborates that intraspecific competition only occurs when different isogenic lines compete for nutrients, while we could not find neighbor effects within the same genotype. However, in our setup, we did not vary the number of plants per pot or plant density in the field, meaning that we may have missed a stronger effect of AM fungal colonization on intraspecific competition. A number of studies have shown that intraspecific density modifies a plant’s response to mycorrhizal colonization (e.g., Schroeder-Moreno and Janos 2008 ; Collins et al. 2010 ). For example, increasing densities significantly reduced the biomass of individual mycorrhizal Trifolium subterraneum plants but the biomass of individual non-mycorrhizal plants was only slightly reduced (Facelli et al. 1999 ). Increasing intraspecific competition also decreased growth differences in Plantago lanceolata (Ayres et al. 2006 ). Though we did not find differences in growth and fitness when the two genotypes were grown in monocultures, P, N, K, and Cu levels were significantly higher in EV plants compared to ir CCaMK plants in the field. Consistent with the results presented here, fruits of tomato mutants with reduced mycorrhizal colonization (rmc) compared to their wild-type progenitor were also not altered in growth when grown in monocultures or under non-inoculated conditions, but they still differed in nutrient contents (Cavagnaro et al. 2004 , 2006 ). For these experiments, it is not known if the lower nutrient content resulted in further fitness consequences. Higher nutrient levels may, on the one hand, facilitate the production of defense compounds and make the plants more resistant against necrotrophic pathogens, while, on the other hand, leaves and fruits may become more attractive to herbivores and biotrophic pathogens, as shown for AM-colonized tree saplings (Ferlian et al. 2021 ). In contrast to our study, field and glasshouse studies with maize plants defective in mycorrhiza-specific Pi transporter Pht1;6 gene and their wild-type progenitor showed lower nutrient levels and less biomass production, not only when plants were competing for nutrients, but also when plants were grown in monocultures (Fabiańska et al. 2020 ). We can only speculate about these different results, but one hypothesis is that additional signaling pathways that may govern, e.g., P-transporter expression, had already been initiated in the maize mutants. P transport via the direct root pathway is known to be attenuated during mycorrhizal colonization (Smith et al. 2011 ). Pht1;6 acts downstream of a symbiosis receptor kinase (SYMRK—the mutation found to be responsible for the reduced mycorrhizal colonization in rmc tomatoes; Nair and Bhargava 2012 ) and of CCaMK (MacLean et al. 2017 ). Hence, the direct P pathway may have already been altered by contact with AM fungi in the maize mutants. As plant size and reproduction are often correlated, size inequalities can result in reproductive inequalities (Weiner 1988 ). Here, we also show that ir CCaMK plants in mixed communities were not only smaller but also produced significantly fewer flowers and capsules, and at least in the glasshouse also less seeds per capsule than neighboring EV plants. Being connected to a hyphal network promotes the performance of genotypes interacting with a mycorrhizal partner, thus providing genotype-specific benefits. Considering our findings in the light of the distribution of genotypes within plant populations, the lower number of capsules and seeds per capsule found in the field for ir CCaMK plants growing in competition with EV could lead to changes in the composition of the long-term seed bank in the soil and the frequency of the genotypes would be altered. Data on long-term community composition of different accessions of the same species in relation to the efficiency of interacting with AM fungi are not yet available. Studies based on different species provide contradictory results ranging from strong (van der Heijden and Scheublin 2007 ) to weak (Milkereit et al. 2018 ) or no effect (Koch et al. 2012 ) of AM fungi on plant community productivity and composition. In this context, it might be important to consider that N. attenuata germinates in nature after fires, and some studies have shown that fire initially reduces the number of AM fungal spores, while more nutrients are available (Hartnett et al. 2004 ) which may give plants with no or weak AM interactions a competitive chance. Given the controversy surrounding CMN benefits, we were not only interested in the performance of individual plants compared to their neighbors, but also whether the productivity of a community is altered when individual plants have different capacities to interact with AM fungi. Focusing on growth parameters, we did not find a net benefit for a specific community type. In mixed cultures, the growth depression of ir CCaMK plants was compensated by the performance gains of EV plants, so that mixed communities realized no net benefit. An independent study used the same Na-ir CCaMK lines for a different question, and they also found that population-level effects were independent of belowground AM fungal interactions (McGale et al. 2020 ). However, results for fitness traits (total flower and capsule numbers) were not conclusive—we observed a net benefit in the field but a decline in the glasshouse. We can only speculate about these dissimilar results. One assumption is that the strictly limited amounts of nutrients in the glasshouse study impaired further benefits and only allowed for an exploitation of resources at the expense of the neighbor, while in the field, additional resources were mobilized leading to a net benefit. Across most parameters measured, we observed in the field that EV in mixed communities performed significantly better compared to ir CCaMK plants and the single genotype communities, while the major differences in the glasshouse were growth depressions and reduced fitness of ir CCaMK and over-performance of EV in mixed communities compared to the two monoculture communities. In addition to pot-bound nutrient limitations of the glasshouse, additional root–root interactions in the field cannot be ruled out. Based on our results, we are unable to disentangle if individual plant–plant competition drives the net fitness benefit in the field or if the connection to the CMN plays an additional role. In conclusion, these experiments highlight the importance of AM fungi in intraspecific plant–plant competition in natural plant communities, allowing AM-harboring plants to outperform non-harboring conspecifics when competing on a small spatial scale. AM associations provided few demonstrable performance benefits at the community level, consistent with recent critical reviews of the data for CMN benefits in forests (Karst et al. 2023 ). Mycorrhizae may have profound effects on long-term plant population dynamics—altering the genetic contribution of individuals from one generation to the next. Author contribution statement KG and ITB conceived and designed the research. ITB prepared both field experiments, FY generated the data of the field experiment in 2016. KG analyzed the data. KG wrote the first version of the manuscript with further contributions from all authors. All authors read and approved the manuscript."
} | 6,723 |
31263777 | PMC6588552 | pmc | 1,037 | {
"abstract": "Nitrogen availability often restricts primary productivity in terrestrial ecosystems. Arbuscular mycorrhizal fungi are ubiquitous symbionts of terrestrial plants and can improve plant nitrogen acquisition, but have a limited ability to access organic nitrogen. Although other soil biota mineralize organic nitrogen into bioavailable forms, they may simultaneously compete for nitrogen, with unknown consequences for plant nutrition. Here, we show that synergies between the mycorrhizal fungus Rhizophagus irregularis and soil microbial communities have a highly non-additive effect on nitrogen acquisition by the model grass Brachypodium distachyon . These multipartite microbial synergies result in a doubling of the nitrogen that mycorrhizal plants acquire from organic matter and a tenfold increase in nitrogen acquisition compared to non-mycorrhizal plants grown in the absence of soil microbial communities. This previously unquantified multipartite relationship may contribute to more than 70 Tg of annually assimilated plant nitrogen, thereby playing a critical role in global nutrient cycling and ecosystem function.",
"introduction": "Introduction Nitrogen (N) is a limiting nutrient in many natural and managed ecosystems 1 , 2 . Arbuscular mycorrhizal (AM) fungi form symbioses with the majority of terrestrial plants and can substantially enhance plant N acquisition from soil, thereby potentially alleviating plant N limitation and playing an important role in plant productivity and soil nutrient cycling 3 – 9 . Although they appear to lack the genetic machinery necessary for decomposition, AM fungi can acquire a substantial quantity of mineral N from organic matter 6 – 14 . A growing body of literature implicates other soil biota with decomposer capabilities as key players in AM fungal N acquisition and transfer to plants 15 . However, it is not clear whether multipartite AM-microbial interactions result in competitive versus synergistic nutrient acquisition 11 , 14 , 16 , 17 and how these interactions respond to global environmental changes such as N enrichment. Terrestrial ecosystems experience substantial N enrichment due to atmospheric deposition and fertilizer applications, with consequences for soil organic matter dynamics, microbial biodiversity, plant community composition, and primary productivity 16 – 21 . Long-term N enrichment of grassland soils results in substantial changes in microbial community structure and functional gene representation 17 – 20 . Although it is recognized that these changes may have important implications for ecosystem function, the particular mechanisms through which long-term N enrichment influences plant-biotic interactions and plant productivity are not fully understood. In order to account for these relationships in Earth system models and predict ecosystem response to increasing N enrichment, it is necessary to understand the extent to which AM-microbial interactions mediate plant N acquisition and associated biogeochemical processes 22 – 25 . Here we show that multipartite synergies between AM fungi and soil microbial communities substantially enhance plant and fungal N acquisition from organic matter and microbial acquisition of plant photosynthates. Long-term N enrichment disrupts these synergies, resulting in diminished mycorrhizal N acquisition from organic matter. These results have implications for terrestrial nutrient cycling models, agricultural management, and our understanding of ecosystem response to global change.",
"discussion": "Discussion Our results demonstrate that emergent synergies between plants, mycorrhizal fungi, and free-living soil microbes have a highly non-additive effect on plant N acquisition from organic matter. Although these relationships have been explored in previous studies, the role of the synergies between AM fungi and free-living microorganisms in plant N acquisition and soil organic matter cycling has not been quantified directly 3 , 6 . Here we show that more than half of the N that AM plants derive from organic matter may be attributed to a synergistic relationship between AM plants and soil microbial communities and that this synergy is disrupted by a history of N enrichment. Applied to estimates of global plant N uptake, these results suggest that more than 70 Tg of annually assimilated plant N can be attributed to interactions between AM plants and soil microbes, but that these relationships are sensitive to environmental change 35 . These findings can be used to constrain Earth system models and improve agricultural management, where organic inputs provide an important supply of N to plants. Since terrestrial ecosystems are often N-limited, this also has implications for global N cycling and net primary productivity 36 , 37 ."
} | 1,187 |
35208911 | PMC8878055 | pmc | 1,038 | {
"abstract": "Turfgrass landscapes have expanded rapidly in recent decades and are a major vegetation type in urbanizing ecosystems. While turfgrass areas provide numerous ecosystem services in urban environments, ecological side effects from intensive management are raising concerns regarding their sustainability. One potentially promising approach to ameliorate the ecological impact and decrease the use of agricultural chemicals is to take advantage of naturally evolved turfgrass-associated microbes by harnessing beneficial services provided by microbiomes. Unfortunately, especially compared to agricultural crops, the microbiomes of turfgrasses are not well understood. Here, we analyzed microbial communities inhabiting the leaf and root endospheres as well as soil in two bermudagrass cultivars, ‘Latitude 36’ and ‘TifTuf’, which exhibit distinct tolerance to nematode damage, with the goal of identifying potential differences in the microbiomes that might explain their distinct phenotype. We used 16S rRNA gene V4 and ITS2 amplicon sequencing to characterize the microbiomes in combination with microbial cultivation efforts to identify potentially beneficial endophytic fungi and bacteria. Our results show that Latitude 36 and TifTuf showed markedly different fungal microbiomes, each harboring unique taxa from Ascomycota and Glomeromycota, respectively. In contrast, less difference was observed from bacterial and archaeal microbiomes, which were dominated by Bacteroidetes and Thaumarchaeota, respectively. The TifTuf microbiomes exhibited lower microbial diversity compared to Latitude 36. Many sequences could not be classified to a higher taxonomic resolution, indicating a relatively high abundance of hitherto undescribed microorganisms. Our results provide new insights into the structure and composition of turfgrass microbiomes but also raise important questions regarding the functional attributes of key taxa.",
"conclusion": "4. Conclusions In this study, we showed that two bermudagrass cultivars that have varying levels of tolerance towards PPN do have significantly different fungal microbiomes but relatively similar bacterial and archaeal communities. Many of them belong to unknown taxa, suggesting there is still a lack of information about the microbial communities associated with turfgrasses. While fungal and bacterial species with previously reported nematicidal activity were either absent or not detected in high numbers in our datasets, we were able to identify organisms that were much more abundant in the more tolerant than in the less tolerant cultivar, albeit with identical turfgrass management practices. Functional bioassays that screen for the nematicidal activities of novel isolates are necessary to test the hypothesis on whether they could be responsible for the observed phenotype and might detect a multitude of novel strains with nematicidal activities. Many factors other than microbiomes can potentially affect the host tolerance towards nematode damage, such as differences in root morphology, host innate immune system, and host stress responses. A challenge for assessing the role of other biotic and abiotic factors is the lack of robust baseline information on the diversity of microbiomes that might influence host physiology and metabolic functions. The datasets produced in this study will help to address this issue and can guide more targeted studies on the structure and function of microbiomes in turfgrasses.",
"introduction": "1. Introduction Urban ecosystems are expanding globally at a rate that is unprecedented in human history and are increasingly important in terms of climate change and ecosystem functionality worldwide [ 1 , 2 ]. Turfgrass areas have become an integral component of modern urban and suburban landscapes with expanding urbanization [ 3 ]. Next to their unique roles in providing recreational, aesthetic, and health benefits to humans, turfgrasses provide multiple ecosystem services, including controlling soil erosion, water runoff, and improving soil quality [ 4 ]. Turf also helps in sequestering carbon and ameliorating urban heating, noise, glare, and visual pollution [ 5 ]. Turfgrass ecosystems are strongly influenced by intense management practices, including frequent mowing, fertilization, and irrigation, and are rich in organic matter due to extensive root growth and the continuous addition of clippings following mowing [ 6 ]. Thus, turf ecosystems represent one of the large terrestrial carbon pools in urban ecosystems. Despite having high potential for increased microbial activity [ 7 ], little is known about the microbial communities inhabiting turfgrasses, especially when compared to those in economically important crops [ 8 ]. Given the importance of plant-associated microbes and microbiomes in ecosystem services [ 9 , 10 , 11 ], understanding the resident microbes in turfgrasses is an essential first step towards promoting healthy and sustainable turf ecosystems. Turfgrasses are prone to damage by diverse pathogens and diseases. One of the most prevalent and serious causes of damage in turfgrasses is plant-parasitic nematodes (PPN). In turfgrasses in the US, sting nematodes ( Belonolaimus spp.), spiral nematodes ( Helicotylenchus spp.), stubby-root nematodes ( Paratrichodorus spp.), stunt nematodes ( Tylenchorhynchus spp.), lance nematodes ( Hoplolaimus spp.), and root-knot nematodes ( Meloidogyne spp.) are among the most commonly encountered PPN [ 12 ]. The damage caused by PPN can vary among nematode species and is dependent on their population density [ 13 ]. In general, PPN impair normal root growth and cause symptoms, including root necrosis and galling, which restricts water and nutrient uptake, resulting in leaf chlorosis and patchy turf growth [ 14 ]. Root damage caused by PPN can serve as entry points for fungal and bacterial pathogens as well, thus exacerbating the initial symptoms in turfgrasses [ 13 ]. A wide range of nematode susceptibility was observed from different turfgrass cultivars [ 15 ], partly driven by inherent, different degrees of tolerance to stress and pathogens. Moreover, fungal and bacterial members of turfgrass microbiomes have been shown to possess activities that can contribute to distinct tolerance towards PPN. For example, the infection of grasses with the endophytic fungus Epichloë coenophiala (previously described as Acremonium coenophialum or Neotyphodium coenophialum ) [ 16 ] can reduce PPN populations in soils [ 17 , 18 ]. In addition, nematophagous fungi or nematode-trapping fungi within the family Orbiliaceae can prey on PPN [ 19 ]. Among bacteria, several strains of plant growth-promoting rhizobacteria (PGPR), such as Bacillus and Pseudomonas species, have been identified to suppress PPN [ 20 , 21 , 22 ], and nematophagous bacteria, such as Pasteurella punctata , B. thuringiensis , and B. nematocida, can infect and kill nematodes using insecticidal toxins or unique mechanisms, such as ‘Trojan-horse’-like interactions [ 23 ]. However, many questions remain regarding potentially beneficial microorganisms and their mode of action in turfgrass microbiomes. Successfully addressing fundamental questions on the structure of turfgrass microbiomes is important for understanding the link between microbial diversity and ecosystem functioning and may help to harness and exploit microbes for sustainable turfgrass management. In this study, we analyze microbiomes from two bermudagrass cultivars that show distinct tolerance towards nematode damage and identify differences in the composition of their microbiomes that might explain the observed phenotypes.",
"discussion": "3. Results and Discussion 3.1. A Unique System for the Study of Turfgrass Microbiomes and Host Functions The structure and function of the microbial communities associated with turfgrasses are mostly unknown [ 43 ]. Recent studies analyzing the microbial communities from the root endosphere and rhizosphere demonstrated distinct community composition across different turfgrass species and across different regions, suggesting a broad host range for specific microbial taxa [ 44 , 45 ], and indicated their potential benefits for turfgrasses to cope with environmental stressors [ 46 ]. Genetically similar bermudagrass cultivars that exhibit distinct phenotypes are an ideal system to test the potential impact of beneficial microbes. Below, we analyze and discuss the fungal, bacterial, and archaeal microbiomes, with a focus on potentially beneficial microorganisms that might explain the observed increased tolerance of TifTuf (T+) to damage by PPN. 3.2. General Patterns in Microbiome Structure between Cultivars and Microhabitats The taxonomic characterization of bacterial and archaeal communities was performed by 16S rRNA gene V4 amplicon sequencing and fungal communities by ITS2 amplicon sequencing, with a particular emphasis on identifying distinctive differences in the microbiomes of the two cultivars that might explain the observed different tolerance to nematode infection. A total of 11,703 bacterial, 249 archaeal, and 1043 fungal ASVs were identified from four replicate samples of leaves, roots, and soil, recovered from Latitude 36 (T−) and TifTuf (T+). We first performed hierarchical clustering on the ASV frequencies to characterize the overall patterns in the microbial community structure ( Figure 1 ), and further compared the community structure to associated environmental parameters ( Table 2 ). This resulted in distinctly different groupings of fungal ( Figure 1 a), bacterial ( Figure 1 b), and archaeal ( Figure 1 c) microbiomes. Latitude 36 (T−) and TifTuf (T+) showed markedly different fungal microbiomes, while no similar distinct difference was observed from the bacterial and archaeal microbiomes between these cultivars. Instead, two statistically supported groupings were revealed from the bacterial communities based on their microhabitat, each represented by the community from leaves and the rest. No apparent clustering was found from the archaeal microbiomes, but the archaeal community from the TifTuf (T+) leaves was divergent and separated from all the other archaeal communities. To further investigate the patterns observed from the fungal, bacterial, and archaeal communities, we analyzed the α and β diversity of the microbiomes. Non-metric multidimensional scaling analysis (NMDS) and the Shannon diversity index (H′) combined with multivariate analysis of variance revealed differences between Latitude 36 (T−) and TifTuf (T+), and among the microhabitat on microbial composition ( Figure 2 and Figure 3 ). Overall, the fungal microbiomes from Latitude 36 (T−) and TifTuf (T+) were distinctly different, and a clear separation between communities from the leaves, roots, and soil was observed ( Figure 2 a). In contrast, the differences in the bacterial communities were based on the microhabitat, i.e., the leaves and the rest, roots, and soil of the two cultivars clustered ( Figure 2 b), whereas the archaeal communities from these cultivars did not exhibit distinct differences ( Figure 2 c). The fungal microbiomes from Latitude 36 (T−) exhibited generally higher α-diversity than the ones from TifTuf (T+) ( Figure 3 a, p < 0.01, two-tailed Mann–Whitney U test), and the soil showed the highest fungal diversity, followed by the roots and leaves ( p < 0.001, Kruskal–Wallis with Dunn’s post-hoc test). Similar to the fungal communities, the bacterial and archaeal diversity were also higher in Latitude 36 (T−) compared to TifTuf (T+) ( Figure 3 b,c, p < 0.0001, two-tailed Mann–Whitney U tests). The soil bacterial communities exhibited the highest diversity, followed by the ones from roots and leaves ( Figure 3 b, p < 0.0001, Kruskal–Wallis with Dunn’s post-hoc test). For the archaeal communities, both the soil and root communities showed higher diversity than the leaf communities ( Figure 3 c, p < 0.005, Kruskal–Wallis with Dunn’s post-hoc test), but no statistical difference was observed between the soil and root archaeal communities. 3.3. Detailed Microbial Community Structure in Bermudagrasses Exhibiting Different Susceptibility to Nematode Infection: Fungal Microbiomes The fungal microbiomes were mostly represented by three phyla. Ascomycota was the dominant fungal phylum across all the samples by representing 68.5 ± 11.5% to 90.7 ± 11.2% of the total fungal amplicons, followed by Basidiomycota (2.8 ± 2.1% to 31.3 ± 11.5%) and Glomeromycota (0.03 ± 0.04% to 14.5 ± 12.6%) ( Figure 4 a). Within Ascomycota, two classes, Dothideomycetes and Sordariomycetes, represented the majority of the sequences. A higher contribution from Dothideomycetes was observed in the leaves (leaves; 70.4 ± 19.6%, roots and soil; 30.5 ± 19.7%, p < 0.001, two-tailed Mann–Whitney U test), while Sordariomycetes dominated the roots and soil (leaves; 19.0 ± 10.9%, roots and soil; 60.6 ± 21.4%, p < 0.0001, two-tailed Mann–Whitney U test) ( Figure 4 b). Within Dothideomycetes, the order Pleosporales represented most of the sequences (95.6 ± 4.5% of the total Dothideomycetes amplicons, Figure 4 c), while Sordariomycetes were represented by several orders ( Figure 4 d). Within Basidiomycota, the class Agaricomycetes dominated across all the samples by representing 96.6 ± 5.0% of the total Basidiomycota amplicons. Among Agaricomycetes, the order Russulales was highly abundant in the leaves (leaves; 61.4 ± 29.1%, roots and soil; 16.6 ± 23.6%, p < 0.01, two-tailed Mann–Whitney U test), and the order Agaricales was highly abundant in the roots and soil (leaves; 5.2 ± 8.5%, roots and soil; 47.4 ± 27.4%, p < 0.001, two-tailed Mann–Whitney U test) ( Figure 4 e). Glomeromycota were barely found in the leaves (<0.1%) and were mostly observed from the soil. They were represented by two classes, Glomeromycetes and Paraglomeromycetes. Patterns in relative amplicon abundance are strongly influenced by the reciprocal interplay between a taxon’s own abundance and the changing abundances of other taxa. To further characterize the differences in fungal communities that might link to the different tolerance to nematode infection, indicator species analysis was performed to determine the distinct taxa from each cultivar. A total of 57 and 31 unique ASVs were revealed from Latitude 36 (T−) and TifTuf (T+), respectively ( Supplementary Table S1 ). From the taxa showing higher indicator values, Latitude 36 (T−) contained more indicator species from Ascomycota, mostly from the leaves and soil rather than the roots, while TifTuf (T+) had more indicator species from Glomeromycota, which were mostly retrieved from the roots and soil ( Table 3 ). Under higher taxonomic resolution, more pronounced differences between the cultivars were observed, such as Fusarium spp. and Mariannaea sp. were found from TifTuf (T+), while Myrothecium sp. and other taxa were found from Latitude 36 (T−) within the same taxonomic hierarchy of the order (Hypocreales; Sordariomycetes; Ascomycota). In contrast, the indicator taxa from Glomeromycota were mostly unknown at higher taxonomic resolution, in particular from TifTuf (T+), preventing us from comparing the unique taxa for each cultivar. Consistent with the results from the diversity analysis showing reduced microbiome diversity from TifTuf (T+) compared to Latitude 36 (T−), fewer taxa unique to TifTuf (T+) were recovered. However, our results also revealed that many of them were unknown at the genus level. Specific fungal groups within the dominant phylum, Ascomycota, which are known to suppress the populations of plant-parasitic nematodes and can be coupled to the observed differences in nematode susceptibility, were further examined. The systemic clavicipitaceous fungal endophytes, such as Epichloë, which are commonly found in cool-season grasses (Clavicipitaceae; Hypocreales; Ascomycota), were recovered from our samples as well. Although they were not highly represented, we found more from the root endosphere in TifTuf (T+) compared to Latitude 36 (T−) (Latitude 36; 2.1%, TifTuf; 4.3% of the total Ascomycota amplicons). Therefore, possibly, these fungal endophytes could confer enhanced tolerance against nematode infection. Another group of microorganisms that could be responsible for distinct phenotypes related to nematode damage is nematophagous fungi or nematode-trapping fungi belonging to the family Orbiliaceae (Orbiliales; Orbiliomycetes; Ascomycota) [ 19 ]. While present in our datasets, Orbiliales made up less than 1% of the amplicons within Ascomycota and showed no significant difference between Latitude 36 (T−) and TifTuf (T+). Unexpected results from our studies include the higher representation of the order Russulales in the leaves compared to the roots and soil since Russulales are known as ectomycorrhizal fungi [ 47 ]. All of the Russulales amplicons in our dataset were unknown, even at the family level, suggesting they might represent new taxa within this order. The top BLAST hits indicate that these sequences might be related to Peniophora , a genus that was recently described from the leaves of wild grass [ 48 ]; therefore, the ASVs retrieved here might represent unknown Peniophora species. However, our results indicate that endophytic fungi were mostly represented by Ascomycota, in line with previous studies on fungal endophytes in grasses [ 44 , 49 , 50 , 51 ]. A similar composition between Latitude 36 (T−) and TifTuf (T+), and across leaf and root endospheres and soil, was observed under lower taxonomic resolution, with leaf communities dominated by Dothideomycetes and root and soil communities by Sordariomycetes. However, under higher taxonomic resolution, distinct differences in taxonomic composition and distribution based on their microhabitat could be revealed. For example, ASVs of the order Hypocreales in TifTuf (T+) were mostly represented by Fusarium spp. And recovered from the roots and soil, while different taxa from Hypocreales, especially Myrothecium sp., was found in Latitude 36 (T−) and recovered from the leaves and soil. 3.4. Similar Bacterial Microbiomes in Latitude 36 and TifTuf The bacterial communities from the leaves, roots, and soil in Latitude 36 (T−) and TifTuf (T+) were dominated by Bacteroidetes by representing 31.6 ± 11.0% and 52.7 ± 22.7% of the total 16S rRNA gene sequences, respectively. The Bacteroidetes sequences were comprised mostly of four classes (Chitinophagales, Cytophagales, Flavobacteriales, and Sphingobacteriales) ( Figure 5 ). Different contributions based on their specific microhabitat were observed, for example, for Cytophagales, which were represented more in roots and soil compared to leaves (leaves; 3.5 ± 1.5%, roots and soil; 9.6 ± 3.1%, p < 0.0001, two-tailed Mann–Whitney U test), as indicated from the NMDS plot ( Figure 2 b). However, in general, similar distributions from these four Bacteroidetes classes were observed between Latitude 36 (T−) and TifTuf (T+). To capture the potential differences in bacterial communities between these cultivars, indicator species analysis was performed and a total of 333 unique ASVs were found from Latitude 36 (T−), while only 16 unique ASVs were identified from TifTuf (T+) ( Supplementary Table S2 ). From the taxa showing higher indicator values, Latitude 36 (T−) contained more indicator species, representing diverse taxa from Bacteroidetes and Proteobacteria, and most of them were found across leaves, roots, and soil. In contrast, the unique taxa from TifTuf (T+) were mostly observed from the roots and soil, and many of them were unknown taxa that could not even be assigned at the phylum level ( Table 4 ). Similar to the findings from the fungal microbiomes, additional differences in microbial composition were apparent under higher taxonomic resolution. For example, each cultivar contained different genera within the same family (Chitinophagaceae; Chitinophagales; Bacteroidia; Bacteroidetes). Similar to other turfgrass microbiome studies showing Proteobacteria as the dominant bacterial phylum [ 44 , 45 ], Proteobacteria were highly represented in our samples as well (Latitude 36; 18 ± 9%, TifTuf; 14 ± 11% of the total 16S rRNA gene sequences). Contrasting results compared to previous studies were that Bacteroidetes dominated across all the samples, although a recent study also showed the dominance of Bacteroidetes from the grass soil community [ 52 ]. The observed differences between those studies could be caused by other physicochemical factors, such as soil nutrients and pH, temperature, water content, or by primer bias towards certain taxonomic groups derived from different primer sets [ 53 ]. Additionally, there are several biotic and abiotic factors in nature that might have key impacts on microbiome composition and dynamics, such as morphological differences from Latitude 36 (T−) and TifTuf (T+) ( Supplementary Figure S1 ), different density in plant-parasitic nematodes ( Table 1 ), and soil nutrients ( Table 2 ). Our results indicate more sampling efforts are required towards establishing microbial baselines in turfgrass microbiomes to accurately characterize and assess microbial community changes. 3.5. Archaeal Microbiomes in Bermudagrasses: Potential Link to Nitrification The archaeal communities were mostly represented by two phyla, Nanoarchaeota and Thaumarchaeota ( Figure 6 a), comprising the class Woesearchaeia within Nanoarchaeota (95.8 ± 15.9% of the total Nanoarchaeota) and two families, Nitrosopumilaceae and Nitrososphaeraceae, within the class Nitrososphaeria in the phylum Thaumarchaeota ( Figure 6 b). Compared to their bacterial counterpart, archaeal communities were far less abundant, mostly representing less than 1% of the total 16S rRNA gene amplicons. However, increased archaeal presence was observed in the soil by representing 4.2 ± 3.8% and 4.7 ± 3.1% of the total 16S rRNA gene sequences from Latitude 36 (T−) and TifTuf (T+), respectively ( Figure 5 a). Although not highly abundant, distinctly different archaeal community composition was observed between Latitude 36 (T−) and TifTuf (T+): Woesearchaeia within Nanoarchaeota were more abundant in TifTuf (T+), while ASVs relating to Nitrosophaeraceae within Thaumarcheaota were present in higher numbers in Latitude 36 (T−). For the soil communities, where the most archaeal 16S rRNA gene sequences were recovered, a similar contribution from Nanoarchaeota was observed from Latitude 36 (T−) and TifTuf (T+), each representing 8.9 ± 12.3% and 7.3 ± 4.9% of the total archaeal sequences, respectively. Nitrosopumilaeceae represented the most archael ASVs from TifTuf (T+) (75.5 ± 14.1%), while a similar contribution from Nitrosopumilaceae (47.8 ± 12.4%) and Nitrosophaeraceae (39.7 ± 22.3%) was observed from Latitude 36 (T−). The family Nitrosopumilaceae was mostly represented by Candidatus Nitrosotenuis sp., while the Nitrososphaeraceae family was more represented by an unknown genus, followed by Candidatus Nitrosocosmicus sp. ( Figure 6 b). Both Nitrosopumilaceae and Nitrosophaeraceae are families within the class Nitrososphaeria, which is a taxonomic group that is considered to have a dominant role in the oxidation [ 54 , 55 ]. Ammonia-oxidizing archaea (AOA) are ubiquitously detected in natural environments, including soils, where they have an active role in the nitrogen cycle [ 56 ]; thus, it is not surprising to recover AOA from the soil samples. No archaeal indicator species unique to each cultivar was found, suggesting that most archaeal members were observed from both cultivars. A few ASVs specific to soil were observed, possibly because more archaeal sequences were recovered in the soil compared to the leaves and roots ( Supplementary Table S3 ). However, our results clearly showed distinct AOA community composition between Latitude 36 (T−) and TifTuf (T+) and across leaves, roots, and soil as well. Many of them were unknown taxa and their potential role in nitrification is uncertain. Interestingly enough, Latitude 36 (T−) was shown to outperform TifTuf (T+) under reduced water and N inputs [ 57 ], which could be driven by distinct AOA members. Although it might not directly explain the observed differences towards nematode damage, our limited results suggest that other benefits from microbiomes can influence host physiology, health, function, and eventually link to nematode susceptibility. 3.6. Culturable Endophytic Microbes Isolated from the Bermudagrass Cultivars Microbiomes can provide benefits other than tolerance to nematode infection for grasses as well. For example, Epichloë and other clavicipitaceous fungal endophytes have been shown to contribute to enhanced nutrient uptake, drought tolerance, disease resistance, and deterrence of insect herbivores [ 58 ]. Similarly, bacterial endophytic communities associated with grass roots could also promote host fitness and improved tolerance towards different abiotic stresses [ 59 ]. Despite these known potentially beneficial functional traits of microbiomes, surprisingly little is known about turfgrass endophytes, in particular in warm-season grasses. As an exploratory effort to test cultivation efficiency, fungal endophytes from root tissues were isolated. A total of 13 pure cultures comprising ten different taxa were isolated. Seven strains were closely related to known taxa Fusarium concolor , F. fujikuroi , F. proliferatum , Cochliobolus lunatus , Aspergillus niger , Nigrospora sphaerica , and one bacterial species, Bacillus velezensis . The other three cultures represented uncultured Pleosporales species ( Table 5 ). To identify whether these cultured taxa were highly represented in the amplicon dataset, the ITS2 sequences from the cultured fungal taxa were compared to the fungal ITS2 amplicon dataset, and seven matching 100% ASVs were identified ( Table 5 ). One amplicon sequence matching 100% to B. velezensis was also found from the 16S rRNA gene amplicon dataset. Although these matching ASVs were not abundant across our samples, two taxa, F. concolor and unknown Pleosporales sp., were significantly more observed from TifTuf (T+) than Latitude 36 (T−) ( p < 0.001, two-tailed Mann–Whitney U test, Figure 7 and Table 5 ). Fusarium are cosmopolitan phytopathogenic fungi known to produce diverse toxic secondary metabolites (mycotoxins) [ 60 ]. In contrast to other Fusarium species, F. concolor does not have biosynthetic gene clusters for fumonisins [ 61 ] and was found to be nonpathogenic to a susceptible spring wheat cultivar, but it was able to produce other mycotoxins, moniliformin and enniatin B toxins, in vitro [ 62 ]. Surprisingly enough, mycotoxins enniatin B and moniliformin showed significant nematicidal activities against root-knot nematode Meloidogyne javanica [ 63 ], possibly explaining the lower number of root-knot nematodes from TifTuf (T+) ( Table 1 ), where more F. concolor sequences were recovered compared to Latitude 36 (T−) ( Figure 7 ). Unexpectedly, a bacterial isolate representing Bacillus velezensis was also recovered using media designed to isolate endophytic fungi ( Table 5 ). Bacillus species are known as plant growth-promoting rhizobacteria (PGPR) since they stimulate plant growth through the synthesis of plant growth hormones and suppress plant pathogens through secondary metabolites [ 64 ]. Although not highly represented in our amplicon dataset, B. velezensis was shown to have strong nematicidal effects on egg hatching and the second-stage juvenile (J2) survival of root-knot nematode M. incognita [ 65 ], which might explain the observed different nematode density between Latitude 36 (T−) and TifTuf (T+) ( Table 1 ). While bioassays under various conditions are needed to test the potential nematidical effects from these cultured fungal and bacterial endophytes, and the need exists to optimize the sampling scheme and culture media to recover more beneficial microorganisms, our microbial cultivation efforts to harness beneficial endophytic microbes indicate the potential to develop microbial biocontrol agents to suppress and treat nematode infection in turfgrasses."
} | 7,050 |
22509341 | PMC3324498 | pmc | 1,039 | {
"abstract": "As atmospheric levels of CO 2 increase, reef-building corals are under greater stress from both increased sea surface temperatures and declining sea water pH. To date, most studies have focused on either coral bleaching due to warming oceans or declining calcification due to decreasing oceanic carbonate ion concentrations. Here, through the use of physiology measurements and cDNA microarrays, we show that changes in pH and ocean chemistry consistent with two scenarios put forward by the Intergovernmental Panel on Climate Change (IPCC) drive major changes in gene expression, respiration, photosynthesis and symbiosis of the coral, Acropora millepora , before affects on biomineralisation are apparent at the phenotype level. Under high CO 2 conditions corals at the phenotype level lost over half their Symbiodinium populations, and had a decrease in both photosynthesis and respiration. Changes in gene expression were consistent with metabolic suppression, an increase in oxidative stress, apoptosis and symbiont loss. Other expression patterns demonstrate upregulation of membrane transporters, as well as the regulation of genes involved in membrane cytoskeletal interactions and cytoskeletal remodeling. These widespread changes in gene expression emphasize the need to expand future studies of ocean acidification to include a wider spectrum of cellular processes, many of which may occur before impacts on calcification.",
"introduction": "Introduction Coral reefs are highly productive and biologically diverse ecosystems despite the oligotrophic waters that surround them [1] . They are important to millions of coastal dwelling people across the planet, underpinning industries such as fishing and tourism [2] . Coral reefs appear to be facing a significant increase in local and global stressors [1] , [3] . Global warming and ocean acidification have recently emerged as key threats to the long-term survival of coral reefs. Rapidly warming oceans are driving an increase in the frequency and intensity of mass bleaching events [3] , while steadily acidifying oceans have caused large decreases in the concentration of carbonate ions and potentially the ability of marine calcifiers to precipitate calcium carbonate [4] . High levels of atmospheric CO 2 ([CO 2 ] atm ) and subsequent ocean acidification have been implied as a major factor in several extinction events on coral reefs in geological time [5] . The ocean uptake of [CO 2 ] atm produces carbonic acid (HCO 3 \n − ) as the carbon dioxide reacts with water. Protons (H + ), which are formed due to the resulting dissociation of carbonic acid to bicarbonate ions (CO 3 \n 2− ), react with carbonate ions, forming more HCO 3 \n − and thus reducing carbonate ions available for marine organisms [6] . This decrease in [CO 3 \n 2− ] leads to a reduction in the saturation state of calcium carbonate forms such as aragonite, calcite and high magnesium calcite and thus a reduction in calcification by marine organisms [4] , [7] . To date, most studies of ocean acidification have focused on its impact on calcification rates [4] , as opposed to targeting the physiological processes that lead to the biological deposition of calcium carbonate in these organisms and/or sustain organism health (fitness). It is now clear that overall the predicted reduction in ocean pH and [CO 3 \n 2− ] can be correlated with a decrease in calcification for a diverse range of marine calcifiers, however the response is variable, often non linear and there are inter and intra specific differences [4] , [8] , [9] . In addition, for studies conducted in the field, ocean acidification effects can be compounded by ocean warming [10] . Calcification is clearly important, but many other physiological processes may be affected in marine organisms [11] , [12] , [13] . By assessing these impacts we can commence unraveling cellular and physiological processes that eventually lead to a decrease in calcification rates. This in turn can provide information to explain currently observed discrepancies in calcification rates, which is important if we are to understand the full ramifications of rapid ocean acidification for coral reefs. Here, we investigate what physiological processes in Acropora millepora are affected by changes in ocean pH, both at the level of the phenotype and gene expression level and show that exposure to high CO 2 drive major changes in gene expression, respiration, photosynthesis and symbiosis for the reef building coral.",
"discussion": "Results and Discussion In a study of 8606 unigenes from the coral Acropora millepora exposed to ambient, mid and high CO 2 conditions as predicted by the IPCC ( Table 1 ), we report that increases in dissolved CO 2 after 1 and 28 days affected processes including: metabolism, membrane-cytoskeleton interactions, signaling, translation, transport, calcification, protein folding and apoptosis ( Figure 1 , Table S1 ). In total, acidification resulted in 643 differentially expressed transcripts (FDR, 5%); the largest number of these differentially expressed genes are up or down regulated in the high CO 2 treatment compared to the control at day 28. This was also reflected in principal component analysis which showed that high CO 2 corals at day 28 where separated from the other samples implying the greatest variation ( Figure S1 ). Differentially expressed genes were subjected to K-means clustering in order to group genes with similar temporal expression patterns and we identified 6 major synexpression clusters (I–VI) ( Figure 1 ). Transcripts with homology to known genes (352 transcripts, Blastx, E-score cutoff 10 −6 ) were assigned to gene ontology (GO) categories and subjected to classification analysis to identify enriched GO groups ( Figure 2 ). From the pie charts in Figure 1 which show what major GO categories genes in the synexpression clusters belong to, it is apparent that more changes in cytoskeleton interactions occur in cluster IV, more changes in signaling and catalysis occur in clusters I–III and large changes in transport occur in cluster II. Quantitative real-time PCR of ten representative genes supported the results, where each candidate gene in the qPCR followed the trends found in the microarray data with expression levels either increasing or decreasing in response to high CO 2 conditions ( Figure 3 , Table S2 , S3 ) compared to control corals at day 28. Changes in response to high and mid CO 2 conditions for day 1, where less gene expression changes occurred, contained many changes in heat shock proteins and signaling which differed from changes at day 28 ( Table S1 ). However, this study only had a single time point at a shorter time scale. It would be useful in future studies to better define changes in gene expression levels within the first few days of exposure, which would require an experiment with several time points within these first few days. It should be noted that this study used small sample sizes (n = 3 for microarray analysis and n = 4 for physiology and qPCR) and that future studies would benefit from greater sample sizes, perhaps a greater range of differentially expressed genes would be detected, and more robust conclusions drawn from the physiological data. 10.1371/journal.pone.0034659.g001 Figure 1 Graphical representation of differentially expressed genes in Acropora millepora across pCO 2 treatments (control, medium and high) at day 1 and 28. K-means clustering was applied to group genes (synexpression clusters I–VI) by common temporal expression patterns. Yellow represents upregulation and blue represents downregulation, scale bar is on a log 2 ratio. Each row corresponds to a transcript and each column represents the mean expression (n = 3). For each cluster average log 2 fold changes (±SE) at day 28 are indicated and pie charts classify genes into major biological processes according to enriched GO categories. 10.1371/journal.pone.0034659.g002 Figure 2 Classification analysis for Acropora millepora transcripts that were differentially expressed across pCO 2 treatments (control, medium and high) at day 1 and 28. Gene enrichments (P<0.05) across GO categories are shown. The program GOEAST was used to test for enriched GO categories among differentially expressed genes. Color scheme indicates parent categories (binding, actin cytoskeleton, catalytic activity, metabolic processes and transporter activity) and individual pie segments are annotated for more specific GO categories. The sizes of the pie segments are proportional to the total number of genes enriched. The proportion of differentially expressed genes which were assigned to gene ontology categories was 55%. 10.1371/journal.pone.0034659.g003 Figure 3 Log 2 relative expression of selected genes using quantitative real-time PCR. Expression levels of genes are plotted as ratio of relative expression of high CO 2 corals versus control corals at day 28. The relative expression for these selected genes was normalized to AdoHcyase and Rbl7. Bars represent standard error of the mean (n = 4). 10.1371/journal.pone.0034659.t001 Table 1 Carbonate chemistry parameters a across experimental conditions. IPCC pH * \n ALK * (µM/kgSW) DIC * (µM/kgSW) (Aragonite) pCO 2 (matm) CO 3 \n 2− (µmol kg −1 ) Control (present) 8.0–8.2 2281.9±15.8 1832.4±59.4 3.93–5.21 260–440 253.8±17.9 A1B (medium) 7.8–7.9 2260.0±12.6 2165.4±51.0 1.14–3.71 600–790 145.3±33.7 A1FI (high) 7.6–7.7 2283.3±13.5 2345.5±214.4 0.77–2.85 1010–1350 89.5±13.0 * \n Measured values. \n a \n Parameters were calculated from measured values of pH, total alkalinity (ALK), dissolved inorganic carbon (DIC), temperature (25°C) and salinity (35 ppm), using the program CO2SYS. \n Changes at the mRNA level, where the majority of differentially expressed genes were found at day 28 in the high CO 2 treatment, were confirmed by physiological measurements ( Fig. 4 ). Acropora millepora branches lost Symbiodinium cells in response to changes in ocean chemistry, ( Figure 4A ; Kruskal Wallis test, H 2,12 = 7.54,p = 0.023), where after a 28 day exposure, Symbiodinium populations in the high CO 2 treatment were reduced (1.02×10 6 ±5.34×10 4 ) to less than half the density compared to control branches (2.3×10 6 ±4.68×10 5 ) ( Figure 4A ). The remaining symbiont cells also became less productive and the photosynthetic capacity (as measured by P net max cell −1 and P gross max cell −1 ) was reduced. There was a 60% reduction (Kruskal Wallis test, H 2,12 = 8.34, p = 0.015) in P net max cell −1 and a 50% reduction (Kruskal Wallis test, H 2,12 = 7.73, p = 0.021) in P gross max cell −1 (P net max−LEDR) in the high CO 2 treatment compared to the control ( Figure 4B ). Decreasing rates of gross photosynthesis per Symbiodinium cell, compounded by reduced Symbiodinium populations, may lead to a reduction in photoassimilates translocated to the host coral. These changes are likely to have long-term negative effects on host growth and fecundity, with the prospect of increased susceptibility to disease and mortality, especially if Symbiodinium populations fail to recover rapidly [14] . The observed decrease in P gross max is consistent with previous acidification studies [11] , [15] ; however, in our study there was no change in LEDR per remnant Symbiodinium cell among CO 2 conditions (Kruskal Wallis test, H 2,12 = 1.65, p = 0.437). This may be due to the application of very different light conditions to Crawley et al [15] which exposed coral to sub-saturation light intensities and only had a short experimental time scale. More importantly, there was a 3-fold downturn in dark respiration per coral surface area ( Figure 4C ), (Kruskal-Wallis test, H 2, 12 = 6.71, p = 0.035), which is typically associated with a decline in host maintenance and/or growth [16] . Rapid growth, either as tissue growth or calcium carbonate deposition necessitates high respiration rates, but the observed reductions in the rate of respiration can suggest suppression of growth rates and/or metabolism. Physiological changes in this study preceded any observable changes in calcification/growth as determined by changes in buoyant weight, as there was no difference in branch calcification/growth rates between the 3 treatments after the 28 day incubation (Kruskal Wallis test, H 2,12 = 0.50, p = 0.778) ( Figure 4D ), despite the downturn in both energy production and respiration observed in the high CO 2 treatment. This result may reflect that in this case, observable effects on calcification/growth rates require longer experimental incubation, as the buoyant weight technique may be too insensitive to measure the potential small changes in calcification/growth that may have occurred. 10.1371/journal.pone.0034659.g004 Figure 4 The effect of increasing CO 2 in seawater (control, medium and high) after 28 days on coral-algal physiology. (A) Symbiodinium cell number in reef-building coral, Acropora millepora (B) photosynthetic capacity per symbiont cell measured as P net max, light enhanced dark respiration (LEDR), and P gross max (P net max−LEDR) (C) dark respiration(R dark ) and (D) relative calcification/growth as % change in weight (g) of coral branches over the 28 day experimental period. Error bars represent the standard error of the mean (n = 4). Metabolism Changes to metabolic rates are a common outcome of environmental stress [17] . Changes in gene expression suggest that Acropora millepora may have reduced its metabolism under high CO 2 conditions at day 28 ( Figure 1 cluster IV–VI, Figure 2 , Table S1 ), mirroring the oxygen flux change ( Figure 4 ). There was an overall down-regulation of genes involved in the tricarboxylic acid (TCA) cycle and the mitochondrial electron transport chain ( Table S1 ), indicating reduced oxidative metabolism and capacity to generate ATP and NADPH. There was also an upregulation of triglyceride lipase and Acyl-CoA dehydrogenase ( Figure 1 cluster I, II, Table S1 ), which may point to an increase in the breakdown of lipids for energy use [18] , [19] . Interestingly, there was an increase in mitochondrial transcripts for ATPase ( Figure 1 cluster II, Table S1 ). Cellular apoptosis is often preceded by an increase in mitochondrial ATPase activity, resulting in an influx of potassium, the activation of caspases and ultimately cell death [20] . Metabolic suppression has been shown in a range of marine organisms in response to CO 2 fluctuations [13] , [21] . The majority of energy needs in tropical reef building corals are supplied by the photosynthetic endosymbionts [22] , but host heterotrophy can occasionally meet host requirements [23] . Depressions in aerobic metabolic activity due to mitochondrial disruptions can undermine the viability of host cells regardless of the trophic source of organic carbon supplied into the TCA cycle. In this particular case, metabolic suppression due to acidosis is likely to have long-term fitness costs. Acid-Base Regulation and Ion/Macromolecule Transport Maintaining pH homeostasis is critical to a range of cellular functions [24] . Studies of acid-base regulation and hypercapnia suggest significant physiological challenges for marine fish and worms [13] , [25] . There are cases where mitochondrial energy production is tied to acid-base regulation through HCO 3 \n − transport [26] , bi-direction H + pumping by F 0 F 1 ATPase [20] , or Na + /H + and Cl − /HCO 3 \n − transporters on the cell membrane [25] . Membrane proteins play an integral role in: pH homeostasis of the cell, membrane lipid composition and cell shape maintenance [27] . For A. millepora , 28 days of high CO 2 conditions resulted in changes in membrane transporters ( Figure 1 , 2 , Table S1 ). Specifically, there was downregulation of proton channels (V-type proton ATPases), phosphate transport and protein transport at the cell membrane ( Figure 1 cluster IV,V, Table S1 ). At the same time, sodium and potassium transporters, cell membrane receptors and an ABC transporter were upregulated ( Figure 1 cluster I, Table S1 ). In eukaryotes, ABC type transporters are involved in the export of unwanted molecules, such as toxins [28] from the cell. V- type proton ATPases at the cell membrane serve to acidify the extracellular environment which in turn activates a series of signaling cascades [29] . In the cnidarian ectoderm, plasma membrane proton ATPase activity has been tied to CO 2 uptake [30] . A decrease in this transporter may indicate a decrease in CO 2 uptake under acidification stress. Due to concurrent increases in energy saving ion gradient transporters such as Na + /H + exchangers, the decrease in V- type ATPases for proton transport may also be the result of an active suppression of the more costly ATP dependent ion transporters [31] . In addition, a lipid transporter was upregulated in the high CO 2 treated corals at day 28 ( Figure 1 cluster II, Table S1 ), a change not found in acid base regulation of other marine organisms [13] , [21] , [25] , perhaps implying changes to the lipid configuration of the cell membrane as a response to ocean acidification [32] . Stress Response Mechanisms Abiotic changes are likely to elicit a cellular stress response (CSR), a universally conserved mechanism to protect macromolecules within cells from the potential damage that physical, chemical or biological stressors may cause. The CSR can increase the tolerance temporarily to the stressor, and remove already damaged cells through apoptosis [33] . Transcripts of A. millepora that encode a number of cellular defenses, and transcripts involved in maintenance of protein integrity (molecular chaperones) were downregulated ( Figure 1 cluster V, Table S1 ), whilst genes, protecting the cells against oxidative stress through oxidoreductase activity (eg. Catalase, FAD linked oxidase and selenoprotein [34] , [35] , [36] ) and involved in apoptosis (caspase 3, TRAF3, p53 inducible protein 11 and programmed cell death protein 4 [37] , [38] , [39] ), were upregulated in high CO 2 treated corals at day 28 ( Figure 1 cluster I, II, III, Table S1 ). Bcl-2, MALT1 and API-5, potential inhibitors of apoptosis [40] , [41] , [42] were downregulated ( Figure 1 cluster V, Table S1 ). The upregulation of apoptotic transcripts is consistent with the upregulation of mitochondrial ATPase described above, which together point to disruption in the mitochondrion leading to cell death [20] . An increase in apoptosis may reflect that prolonged environmental stress, and either a lack of cell pH homeostasis or elevated maintenance costs, has resulted in cell damage. The loss of Symbiodinium cells and an increase in transcripts alleviating oxidative stress may point to impairment in the photosynthetic apparatus in the dinoflagellate symbiont or an impairment of the coral mitochondria [14] . This, in turn, would increase the presence of oxygen radicals in the host tissues and imply cell damage potential. The fact that high CO 2 conditions resulted in overall downregulation of protein folding transcripts, may be a sign that the coral tissue may no longer have the capacity to maintain these integral services. Interestingly, at day one of the high CO 2 treatment there was an upregulation of Heat shock protein 40, a change not found at Day 28 ( Figure 1 cluster II, Table S1 ). It must be noted however, that the two other main heat shock proteins (hsp) were not differentially expressed between treatments (hsp 70 and 90) but were maintained at a high expression levels, and their presence may be sufficient for the integrity of newly made proteins. Calnexin and alpha mannosidase transcripts were upregulated ( Figure 1 cluster II, Table S1 ) which would increase the quality control and protein folding ability in the endoplastic reticulum for proteins that will then be further transported to the golgi complex [43] , [44] . It is possible that the other downregulated chaperones could be temporarily reduced while awaiting more favorable environmental conditions. A coral c-type lectin, which is involved in innate immunity in corals [45] was downregulated ( Figure 1 cluster V, Table S1 ) under high CO 2 conditions, indicating that this cell stress response may not be responding appropriately, and this decrease may compromise the coral further as in a stressed holobiont, susceptibility to pathogens may increase [46] . Ca 2+ Ion Binding/Transport and Cell Communication Several calcium (Ca + ) ion binding proteins were downregulated in high CO 2 treatments at day 28 ( Figure 1 cluster IV, V, Table S1 ). Transcripts for calcium-binding receptors that are potentially involved in innate immunity [47] , [48] were also suppressed, implying an adverse change in signaling potential at the cell membranes. Downregulation of calmodulin, FKBP12 and EGF-hand proteins also implies potential disruption in cell calcium homeostasis [49] , [50] , [51] , as these calcium binding proteins control the Ca 2+ release from ryanodine receptors (RyR) within the endoplasmic reticulum (ER), which is an intracellular Ca 2+ storing organelle [49] , [50] , [51] , [52] . Calpain, an important Ca 2 \n + activated protease that has roles in membrane-cytoskeleton interactions, signal transduction, cell differentiation and apoptosis [53] , was upregulated. Changes in these calcium binding proteins indicate that certain signaling pathways may have been altered. Membrane-Cytoskeleton Interactions Exposure to high seawater CO 2 concentrations for 28 days resulted in several differentially expressed genes involved in membrane-cytoskeleton interactions and cytoskeletal remodeling ( Figure 1 , 2 , Table S1 ). It is possible that the change in regulation of these transcripts reflects a change in proteins involved in cytoskeletal interactions, cytoskeletal organization, intracellular transport, cell shape integrity and cell motility [54] , [55] \n [55] . Specifically, there was downregulation of cytoskeletal actin 1, centractin, radixin and coatomer epsilon subunit and radixin ( Figure 1 cluster IV, V, Table S1 ), whilst there was upregulation of tubilin and Lgl tumor suppressor unit ( Figure 1 cluster I, II, Table S1 ). The actin cytoskeleton is important in a diverse range of processes such as cell motility, contractibility, mitosis and cytokinesis, intracellular transport, endocytosis and secretion. In addition, it has been suggested that actin is also involved in regulation of gene transcription through changes in the cytoskeletal actin dynamics or assembly of transcriptional regulatory complexes [55] . Actin is also an important part of the nuclear complex being required for the transcription of RNA polymerases and is also involved in the export of RNAs and proteins from the nucleus [55] . It is possible that the downregulation of cytoskeletal actin in high CO 2 conditions reflects a change in the regulation of gene transcription of proteins involved in cytoskeletal interactions. In addition this downregulation can imply changes in the intracellular transport, plasma membrane interactions and cell shape/integrity. There was also an upregulation of alpha tubilin, which forms a constituent of the microtubule filaments, involved in cytoskeletal organization and vesicle transport. Downregulation of coatomer epsilon subunit implies changes in protein trafficking between the endoplasmic reticulum and the Golgi complex, while upregulation of Lgl tumor suppressor unit indicates changes to events controlling cell polarity [56] , [57] . Cell volume control changes have been recorded in other marine organisms such as crabs in response to hypercapnia [58] , and similar changes may be occurring in the stressed coral cells. The downregulation of Radixin, an important protein involved in linking the plasma membrane to the cytoskeleton using actin rich surfaces [59] , supports the downregulation of cytoskeletal actin. Centractin, or Actin Related Protein 1 (ARP1), was also downregulated under higher CO 2 stress, and this is an important activator of cytoplasmic vesicle movement [60] . In contrast, at day one in high CO 2 stressed corals, there was an upregulation of Radixin and Centractin, or Actin Related Protein 1 (ARP1) ( Figure 1 cluster VI, Table S1 ), indicating that different changes in cytoskeletal interactions were occurring at this stage. The cytoskeleton has profound effects on the plasma membrane. At times, there may be uninhibited lateral diffusion of lipids and proteins across the plasma membrane; the influx of these molecules can be regulated by the membrane-cytoskeleton links. These become obstacles to free diffusion through diffusion-limited lipid domains [54] . It may be that changes in these membrane-cytoskeleton links in this study reflect changes in transport across the membrane. Rab/Ras GTPases Exposure to increased CO 2 concentrations for 28 days lead to an up ( Figure 1 cluster II, Table S1 ) and downregulation ( Figure 1 cluster V, Table S1 ) of transcripts that resemble members of the Rab/Ras GTPases families (small G-proteins). The Ras GTPase superfamily is a small monomeric group of GTPases, which are involved in cell proliferation and cell signaling events in response to external stimuli. Disruption of the Ras signaling pathway is a key component in the progression of tumor growth [61] . The Rab GTPase family is part of the Ras GTPase superfamily and plays a key role in many membrane-trafficking events in eukaryotic cells, such as exocytosis. This group of proteins which tightly associates with the cell membrane is involved in transport vesicle formation, actin and tubilin motility, docking and membrane fusion. Rab proteins are active when bound to GTP and are inactive when bound to GDP [62] , [63] . In its active state, the Rab protein regulates the transport of lipids and proteins between distinct membrane bound organelles through interactions with downstream effector proteins which are recruited onto the membranes [63] . Small G-proteins are implicated in most cellular events where plasma membrane-cytoskeleton interactions or plasma membrane shape changes (plasma membrane deformations) occur. The observed upregulation in members of these small monmeric GTPases most likely reflects changes in the cell membrane and cytoskeletal interactions to accommodate changes in external seawater chemistry. Extracellular matrix Changes in the extracellular matrix (ECM) have previously been implied to potentially affect calcification [34] , [36] . Our expression patterns indicate that only two transcripts encoding previously described ECM proteins changed after 28 days under mid CO 2 exposure. SEC13L1 was upregulated ( Figure 1 cluster III, Table S1 ) while peroxidasin was downregulated ( Figure 1 cluster IV, Table S1 ). At day 28 under high CO 2 exposure, there was also a downregulation to a predicted protein in the extracellular matrix ( Table S1 ). This implies that small changes to calcification may have started occurring and that perhaps with a longer experimental incubation time more ECM and calcification related transcripts would have been differentially expressed. This overall supports our findings at the phenotype level where no change in calcification/growth was found ( Figure 4 ). Overall in this study there were fewer changes in transcripts which may be involved in calcification, in response to ocean acidification, compared to gene expression studies with corals exposed to thermal stress where changes to the following transcripts were observed; collagen α-1, ECM matrix metalloprotease, papilin, carboxypetidase inhibitor SmC1, procollagen, galaxin and SCP-like extracellular protein [34] , [36] . Cell-wide Responses by Corals to Ocean Acidification: A Model To highlight the differences between acidosis which may be a factor in this study from the impact of hypercapnia seen in other marine organisms, we purpose a model ( Figure 5 ) of cell-wide, coral host response to high CO 2 stress. This model attempts to account for the classic acidosis response (acid-base regulation and metabolic depression) and the novel responses observed in the studied coral (apoptosis, signaling events, calcium homeostasis, cytoskeletal remodeling, cytoskeletal-membrane interactions and oxidative stress). The coral specific responses may result from increased reactive oxygen species (ROS) and/or increased reactive nitrogen species (RNS) created from a disturbance in the Symbiodinium cell, the host mitochondria, or both [14] , [34] . Upregulation of catalase, FAD-linked oxidase and selenoprotein indicates that there may be an increased amount of ROS in the coral cells [64] , [65] . Increased ROS/RNS can result in a disruption to the calcium homeostasis [34] . The role of internal [Ca 2+ ] increase in coral bleaching has been suggested previously [34] . The downregulation of calmodulin (CaM), FKBP12 and EF-hand proteins under high CO 2 stress indicates that there may be a disruption to the Ca 2+ homeostasis [49] , [50] , [51] . Modifications of the actin cytoskeleton, membrane-cytoskeleton interactions and cell receptor/adhesion properties will be affected by a disruption in Ca 2+ homeostasis and metabolic suppression [34] , [66] . Both oxidative stress and an increase in intracellular Ca 2+ can lead to apoptosis and changes in transcripts indicate that both the NF-kB and p53 apoptotic pathways [67] , [68] were upregulated. Changes in predicted proteins in the extracellular matrix may imply changes in or restructuring of the extracellular matrix. Our model suggests that similar cellular events are occurring under acidosis induced bleaching to those reported for thermally induced bleaching [14] , [34] , [36] , but with the addition of changes to acid-base regulation and mitochondrial ATPase activity. 10.1371/journal.pone.0034659.g005 Figure 5 A proposed model of cellular events occurring as a result of ocean acidification. These changes lead to compromised health in Acropora millepora (reduction in symbiont cells and decreased photosynthesis and respiration). The schematic depicts an endodermal cell which contains the symbiont cell. Cellular events depicted here are likely to also occur in other cell types which do not contain symbionts, especially the acid base changes at the cell membrane. Changes in carbonate chemistry lead to changes in acid base regulation and cell membrane transporters. Acid base regulation may not be sufficient leading to acidosis within the cell. This could increase reactive oxygen species (ROS) due to a disruption (⋆) in the Symbiodinium cell (S) and/or in the coral host mitochondrion (M), which may also produce reactive nitrogen species (RNS). The overall oxidative stress and a disruption to calcium stores at the endoplasmic reticulum (ER) can lead to calcium imbalance. This in turn leads to events such as changes in the extracellular matrix, cytoskeletal remodeling, changes in cytoskeletal interactions, disruption to cell reception and signaling potential, and an increase in cell death. Moreover disruption in both the host mitochondrion and Symbiodinium cell leads to a decrease in metabolism and a decrease in metabolite transfer from the symbiont cell. In addition the disruption in the host mitochondrion can also lead to cell death. For the cell membrane transporters black arrows indicate upregulation and white arrows indicate downregulation. On our present greenhouse trajectory, we are likely to use all the >4000 Gt of carbon present in the global fossil fuel reserves by 2400. This will significantly acidify the oceans for thousands of years [69] and take them to a point not seen in tens of millions of years [70] . Our study highlights the imperative to investigate the impacts of ocean acidification on processes other than those involved in biomineralisation. Also, there is a need for more studies investigating the effects of naturally occurring changes in pCO 2 on marine calcifiers in situ . This is a priority, if we are to understand the fate of the many supporting roles that corals contribute to the maintenance of coral reefs."
} | 8,088 |
34124018 | PMC8193722 | pmc | 1,040 | {
"abstract": "Polyethylene terephthalate (PET) is globally the largest produced aromatic polyester with an annual production exceeding 50 million metric tons. PET can be mechanically and chemically recycled; however, the extra costs in chemical recycling are not justified when converting PET back to the original polymer, which leads to less than 30% of PET produced annually to be recycled. Hence, waste PET massively contributes to plastic pollution and damaging the terrestrial and aquatic ecosystems. The global energy and environmental concerns with PET highlight a clear need for technologies in PET “upcycling,” the creation of higher-value products from reclaimed PET. Several microbes that degrade PET and corresponding PET hydrolase enzymes have been successfully identified. The characterization and engineering of these enzymes to selectively depolymerize PET into original monomers such as terephthalic acid and ethylene glycol have been successful. Synthetic microbiology and metabolic engineering approaches enable the development of efficient microbial cell factories to convert PET-derived monomers into value-added products. In this mini-review, we present the recent progress of engineering microbes to produce higher-value chemical building blocks from waste PET using a wholly biological and a hybrid chemocatalytic–biological strategy. We also highlight the potent metabolic pathways to bio-upcycle PET into high-value biotransformed molecules. The new synthetic microbes will help establish the circular materials economy, alleviate the adverse energy and environmental impacts of PET, and provide market incentives for PET reclamation.",
"conclusion": "Conclusion and Perspectives Notably, the titer, yield, and rate (TYR) of monomers’ bioproduction from PET-derived substrates ( Figure 1 ) need to be improved via metabolic engineering and process design approaches to enable commercial production. Development of in silico computational and machine learning programs to assist the design–build–test–learn cycle (high throughput screening of enzymes and design metabolic pathways) enables rational engineering of commercially applicable superior microbial biocatalyst to upcycle PET. ALE enables the strain to optimize further the engineered genome and fine-tuning of the desired metabolic pathway ( Kim et al., 2013 ; Lee et al., 2014 ; Oh et al., 2016 ). We could deploy ALE to improve the PET conversion TYR of the engineered strain. The waste PET may carry toxic compounds such as emerging contaminants (ECs) and poly organic pollutants (POPs); thus, the process requires a priori detoxification steps. For instance, we could use efficient chemical-based metal-organic frameworks or enzyme-based laccase or peroxidase to detoxify the PET-associated ECs and POPs ( Pi et al., 2018 ; Mishra et al., 2019 ; Morsi et al., 2020 ). Indeed, laccase can be expressed on PET upcycling microbes to enable in situ detoxification ( Chen et al., 2016 ). Given that most of the substrates, intermediates, and targeted products are toxic to the host microbes, engineering multiple toxicity tolerance mechanisms will be necessary. For instance, overcoming the aldehyde tolerance and acid tolerance during EG metabolism could be achieved by alleviating the metabolic bottlenecks and engineering the protein quality control machineries ( Franden et al., 2018 ; Jayakody et al., 2018 ; Guan and Liu, 2020 ). One exciting area of study is adapting the architecture of cellulosomes to develop a multi-enzyme complex capable of efficiently degrading PET. The advanced synthetic biology techniques enable the formulation of large cellulosomes and facilitate superior activity toward the recalcitrant cellulose ( Anandharaj et al., 2020 ). The same concept could be adopted to tailor microbial cell factories to the degradation of high-crystalline PET via developing PETsome ( Supplementary Figure 1 ). It is vital to discover the component that would act as a PET binding domain, analogous to the cellulose-binding domain ( Ribitsch et al., 2013 ; Weber et al., 2019 ). An efficient cell surface expressing system for bacteria has recently been developed ( Chen et al., 2019 ; Dvořák et al., 2020 ). Together, those approaches can be implemented to design a consolidated bioprocessing system, a microbial system that can efficiently degrade and upcycle PET into advanced chemicals simultaneously. It will be beneficial to overcome techno-economic challenges such as end-product toxicity on degradation enzymes and the overall operating costs. In summary, PET upcycling via synthetic microbial biocatalyst or hybrid biochemical approaches has shown great promise to sustainable large-scale solutions for PET waste management in terms of end-of-life to PET. It is essential to perform a comprehensive life cycle and techno-economic analysis to identify the upcycle process’ industrial and environmental feasibility using the engineered biocatalyst. We envision that innovative synthetic microbiology and metabolic engineering approaches may enable the microbial biocatalyst to reach the commercial scale from laboratory bioreactor to upcycle PET, create a circular material economy, and help protect our environment from PET waste.",
"introduction": "Introduction Plastic, a synthetic polymer, plays a vital role in modern life due to its versatility, advantageous material properties, and low production cost. It has been estimated that about 5–13 million tons of plastic is ended up in the ocean annually, and 5 trillion plastic particles are estimated to float in Earth’s oceans, which injures and kills marine life ( Eriksen et al., 2014 ; Law et al., 2020 ). Although plastic is less prone to biodegradability, it can be partially fragmented to microplastic (5 mm to 1 μm) particularly by ultraviolet radiation, and microplastics have invaded not only terrestrial and marine ecosystem, but atmospheric ecosystems as well ( Eriksen et al., 2014 ; de Sá et al., 2018 ; Allen et al., 2019 ). Microplastics enters the food chain, spreads toxins, and poses a potential threat to human health ( Wang et al., 2018 ). The systematic presence of synthetic micro polymers is threatening to create a global-scale environmental crisis ( Thompson et al., 2009 ; Jambeck et al., 2015 ; Geyer et al., 2017 ). Polyethylene terephthalate (PET) is a thermoplastic polyester of terephthalic acid (TPA) and ethylene glycol (EG) monomers ( Kim and Lee, 2012 ). The wide applicability of PET in various industries such as in packaging, textiles, electrical and electronics, and automotive industry is related to its properties such as high mechanical strength, light weight, electrical insulating properties, chemical inertness, and gas and moisture barrier properties ( Webb et al., 2013 ). Global PET fiber and resin production was estimated to be around 77 million tons in 2015 ( Fiber Economics Bureau, 2015 ; Plastic Insight, 2016 ). Mismanagement of PET waste contributes to the plastic pollution and demanding techno-economically feasible end-of-life or circular economy options for PET, including recycling post-consumer PET back to the original material ( Hamid et al., 2018 ; Meys et al., 2020 ). Mechanical conversion of PET to the same use often results in PET polymer with poorer mechanical and structural properties, and thus, lower value by 33% ( Awaja and Pavel, 2005 ). PET can be chemically recycled via full breakdown to monomers and repolymerized back to PET; however, chemical-based recycling costs of PET to remake the same polymer is not economically feasible ( Rahimi and García, 2017 ; Vollmer et al., 2020 ). However, It has been predicted that recycled PET over virgin PET has remarkable energy and environmental impact on reducing greenhouse gas (GHG) emissions by 1.5 CO 2 -eq-ton/recycle PET and energy input over virgin PET by >20 MJ/ton ( Rorrer et al., 2019 ). The novel discovery of PET depolymerizing enzymes has transformed the field to develop a techno-economically feasible bio-based PET recycling process ( Wei and Zimmermann, 2017 ; Tournier et al., 2020 ; Zimmermann, 2020 ). Researchers have uncovered novel PET hydrolases from the microorganism in plastic ecosystems (Plastisphere) and investigated them to establish bio-based PET recycling approaches ( Mueller, 2006 ; Kawai et al., 2020 ). The CARBOIS, a green chemistry company, developed industrially applicable enzyme-based recycling technology to remake PET bottles with similar material properties only using recycled PET monomers ( Tournier et al., 2020 ). Comprehensive review articles have been published on bio-based PET recycling techniques ( Wei and Zimmermann, 2017 ; Blank et al., 2020 ; Ru et al., 2020 ; Wei et al., 2020 ; Kawai, 2021 ). With the advances in synthetic microbiology, the development of sustainable microbial-based “PET upcycling” toward a green route of the circular economy becomes attractive. Upcycling is achieved by adding value to the PET waste by providing a path for utilizing PET-derived compounds to manufacture high-value chemicals and materials ( Kenny et al., 2008 ; Rorrer et al., 2019 ; Blank et al., 2020 ; Sohn et al., 2020 ). Microbial cell factories have been tailored to the deconstruction of PET in concert with the chemical processes (i.e., hybrid biochemical process). PET-derived monomers can be biotransformed into high-value platform chemicals and biomaterials, including bioplastic PET alternatives. It enables the creation of a circular material economy for PET ( Sohn et al., 2020 ; Tiso et al., 2020 ). Hence, this mini-review highlights the current progress on microbial-based PET upcycling."
} | 2,403 |
36874912 | PMC9982732 | pmc | 1,043 | {
"abstract": "Heavy metals (HMs) contamination and vegetation destruction in the mining area caused by mining activities are severely increasing. It is urgent to restore vegetation and stabilize HMs. In this study, we compared the ability of HMs phytoextraction/phytostabilization of three dominant plants, including Artemisia argyi (LA) , Miscanthus floridulus (LM), and Boehmeria nivea (LZ) in a lead-zinc mining area in Huayuan County (China). We also explored the role of the rhizosphere bacterial community in assisting phytoremediation using 16S rRNA sequencing technology. Bioconcentration factor (BCF) and translocation factor (TF) analysis showed that LA preferred accumulating Cd, LZ preferred accumulating Cr and Sb, and LM preferred accumulating Cr and Ni. Significant ( p < 0.05) differences were found among the rhizosphere soil microbial communities of these three plants. The key genera of LA were Truepera and Anderseniella , that of LM were Paracoccus and Erythrobacter , and of LZ was Novosphingobium . Correlation analysis showed some rhizosphere bacterial taxa (e.g., Actinomarinicola , Bacillariophyta and Oscillochloris ) affected some soil physicochemical parameters (e.g., organic matter and pH) of the rhizosphere soil and enhanced the TF of metals. Functional prediction analysis of soil bacterial community showed that the relative abundances of genes related to the synthesis of some proteins (e.g., manganese/zinc-transporting P-type ATPase C, nickel transport protein and 1-aminocyclopropane-1-carboxylate deaminase) was positively correlated with the phytoextraction/phytostabilization capacity of plants for heavy metals. This study provided theoretical guidance on selecting appropriate plants for different metal remediation applications. We also found some rhizosphere bacteria might enhance the phytoremediation of multi-metals, which could provide a reference for subsequent research.",
"conclusion": "5 Conclusion According to our investigation of BCF and TF for metals, LA preferred accumulating Cd, LZ preferred accumulating Cr and Sb, and LM preferred accumulating Cr and Ni. The bioaccumulating capacity for multi-metals of Artemisia argyi Levl was more potent, while the translocating capacity was weaker than that of Miscanthus floridulus (Lab.) and Boehmeria nives (L.). The dominant bacterial groups, ecological networks and soil properties (e.g., available phosphorus and moisture content) were different in the rhizosphere soil of these three plants. A higher proportion of positive links in the ecological network might enhance the metal uptake of plants. PICRUSt analysis and correlation tests indicated that genes related to the synthesis of proteins (e.g., manganese/zinc-transporting P-type ATPase C, nickel transport protein and ACC deaminase) could also promote phytoremediation. The rhizosphere bacterial community contained genera, such as Truepera , Anderseniella , Paracoccus , Erythrobacter and Novosphingobium , which may represent potential bacterial strain resources for plant-microbes combined remediation of HM-contaminated soils.",
"introduction": "1 Introduction Due to the continuous mining activities, heavy metals (HMs) contamination and vegetation destruction are severely increasing. It is a severe threat to ecological biodiversity and human health. Therefore, soil restoration and governance need urgently be solved ( Xiang et al., 2021 ; Zerizghi et al., 2022 ). The primary remediation methods for metal-contaminated soils include physical, chemical and biological strategies. In general, physical and chemical remediation methods are costly, unfriendly to the environment, and can easily lead to secondary pollution, so that many researchers now focus on bioremediation strategies ( González Henao and Ghneim-Herrera, 2021 ; Azhar et al., 2022 ). Phytoremediation has been widely used due to its advantages of being cost-effective, economical, eco-friendly, and sustainable ( Ahemad, 2019 ; Shrivastava et al., 2019 ). Currently, As hyperaccumulators such as Pteris vittata and Pteris cretica , Zn hyperaccumulator Sedum alfredii , Cd hyperaccumulator Viola baoshanensis , and Mn hyperaccumulator Phytolacca acinose have been found in China for the phytoremediation of mine tailings ( Niu et al., 2021 ). The effectiveness of phytoremediation depends on the chemical and physical properties of the plant, the bioavailability of metals in the soil, and the ability of rhizosphere soil microorganisms to absorb, transfer and detoxify metals ( Subpiramaniyam, 2021 ). The removal of HMs by microorganisms has the advantages of easy use, low cost, large adsorption capacity, and high efficiency. Among these, bacteria, fungi and algae are widely used ( Yin et al., 2019 ). In general, heavy metal ions can be adsorbed and combined by some functional groups of bacterial polysaccharide mucus layer, e.g., carboxyl, amino, phosphate, and sulfate ( Yue et al., 2015 ). The adsorbed heavy metal ions can enter the microbial cells by metal-related enzymes or proteins, changing the redox state of heavy metal ions and thereby reducing their toxicity. In practical applications, plants and microorganisms for bioremediation are more efficient ( Vaid et al., 2022 ). Usually, microorganisms can improve plant extraction by increasing the availability of HMs in plants and increasing plant biomass ( Wood et al., 2016 ; Mishra et al., 2017 ). Previous studies have shown that Patescibacteria increased the availability of HMs in the rhizosphere and promoted the remediation of heavy metal-contaminated soil by Sedum alfredii ( Tian et al., 2022 ). Moreover, the plant preference for some HMs is influenced by the rhizosphere ecological characteristics, including the specific hormones, root exudates, soil nutrients, soil properties and rhizosphere soil microbes ( Bais et al., 2006 ). The phytoremediation efficiency was regulated by soil enzyme activity and beneficial rhizosphere-associated microorganisms Trifolium repens L. ( Lin et al., 2021 ). Therefore, the combined plant-microbes method can improve the heavy metal resistance of plants and achieve an ideal remediation effect. In recent years, it has also been found the limited efficiency of phytoremediation with a single plant ( Buscaroli et al., 2017 ; Rizwan et al., 2019 ; Hosseinniaee et al., 2022 ) and the co-planting pattern of complementary plants for metals enrichment may be more efficient. The co-plantation of Solanum nigrum with Quercus nuttallii or Quercus pagoda effectively improved the enrichment of cadmium (>40%) and zinc (>30%) ( Qu et al., 2021 ). The environmental problem is severe in the lead-zinc mining area in Huayuan County, Hunan Province, China. This area is a super large ore deposit with many tailings ponds. Wastewater and ore sand was discharged through tunnels and village ditches, accumulating in the surrounding crops and villagers. Barbaric mining and poor management caused environmental pollution, damaged mine landforms, reduced vegetation coverage, soil erosion, and other problems. Therefore, we aim to explore the relationships among metals uptake of the dominant plants, rhizosphere bacterial community and the soil environment under the long-term HMs stress, analyze the ecological characteristics of the rhizosphere (physicochemical properties, soil enzyme activities and bacterial community structure), and the major microbial groups and metabolisms responsible for the oxidoreduction/detoxification/tolerance of HMs in the rhizosphere and absorption/translocation of HMs of plants.",
"discussion": "4 Discussion In recent decades, our demand for metal resources has been increasing, which accelerated the mining of metal mines. The continuous expansion of mining areas led to more and more arable land facing environmental pollution ( Diaz-Morales et al., 2021 ; Li et al., 2022 ). The increase in HMs will stimulate the production of a large number of reactive oxygen species, which will seriously impact the quality of crops and vegetables ( Clemens and Ma, 2016 ) and lead to human health risks ( Xiang et al., 2021 ). Phytoremediation is a sustainable approach to remediating contaminated sites ( Hasnaoui et al., 2020 ). Thus, the study of the interaction between the hyperaccumulators and microorganisms can help formulate effective remediation strategies for mining areas ( Xuan et al., 2021 ). In this study, we investigated the rhizosphere ecological characteristics of the dominant plants around the mining area and further explored their potential for multi-metal(loid)s phytoremediation. It was found that As, Cd, Cu, Pb and Zn have reached potentially hazardous levels ( \n Table 1 \n ). We found that Artemisia argyi, Miscanthus floridulus , Boehmeria nivea are the dominant enrichment plants in this area, and the application of dominant plants combined with microbial communities for the bioremediation of soil pollution was explored ( \n Figure 6 \n ). Figure 6 Schematic plot of the interaction between rhizosphere bacteria and plants in multi-HMs contaminated soil. Microorganisms can directly or indirectly change the availability of metals and promote plant growth and absorption of HMs by regulating soil physicochemical properties or secretion of secondary metabolites. Our study found that Cyanobacteria/Chloroplast, Chloroflexi and Acidobacteria were key phyla for the plant to accumulate HMs. Different plants had different dominant rhizosphere flora, which affected the remediation of different HMs. The dominant phylum Cyanobacteria/Chloroplast in LZ was significantly ( p< 0.05) positively correlated with TF of Cd, Cu, Mn, Pb and Zn ( \n Figure S2 \n ). However, few studies reported the responses of Cyanobacteria/Chloroplast and their roles in phytoremediation under HMs stress, and it might need further study to verify. Chloroflexi was the dominant phylum in the LA group ( \n Figure 1B \n ), significantly ( p < 0.05) higher than in LM and LZ ( \n Figure 2B \n ). Studies have found that Chloroflexi had an obvious advantage in polluted soil ( Hemmat-Jou et al., 2018 ; Koner et al., 2022 ), which could rapidly adapt to heavy metal stress and affect the content of organic matter to control the practical availability of Cr and Pb for plants ( Tang et al., 2019 ). It is consistent with this study, where Chloroflexi was also significantly ( p < 0.05) positively correlated with the organic matter ( \n Figure S2 \n ). Acidobacteria was the dominant phylum in the LM group ( \n Figure 1B \n ) and was significantly ( p < 0.05) positively correlated with BCF of Cr ( \n Figure S2 \n ). Previous research found that the relative abundance of Acidobacteria was inversely related to pH ( Debnath et al., 2016 ). Sun et al. (2022) found that the relative abundance of Acidobacteria was positively correlated to the contents of total Cr and available Cr, and Acidobacteria could cut down soil pH. The key genera of LA were Truepera and Anderseniella , that of LM were Paracoccus and Erythrobacter , and of LZ was Novosphingobium . The oxidation-reduction and nitrogen fixation activities of Truepera allow it to support plant growth ( Zhou et al., 2022 ). Anderseniella can eliminate metabolic wastes, heavy metals, and aromatic chemicals ( Karimi et al., 2019 ). Paracoccus can produce acid, alter the chemical and physical parameters in the soil and encourage plants to absorb heavy metals ( Carvalho et al., 2018 ). Erythrobacter is an iron metabolism bacterium that can secrete iron carriers and promote plant growth ( Li L. et al., 2020 ). Novosphingobium has strong antioxidant activity and can slow down the toxic effect of heavy metals on plants ( Petruk et al., 2018 ). Except for the dominant flora of different plants, the study found that they shared the same vital microorganisms ( \n Figure 3D \n ). These microorganisms (e.g., Proteobacteria, Acidobacteria and Firmicutes) could assist plants in accumulating HMs. Studies have found that in the harsh tailings environment, the colonies mentioned above belonged to the dominant flora and could benefit the expression of metal resistance genes ( Guo et al., 2017 ; Jiang et al., 2021 ; Koner et al., 2022 ). HMs in the soil can affect the growth of microorganisms through protein denaturation, cell membrane damage and inhibition of RNA expression and metabolism ( Wang et al., 2020 ; Duan et al., 2022 ). It has also been studied that HMs stress enhanced the functional fitness of endophytic bacterial communities ( Yao et al., 2022 ). Most microorganisms were positively correlated to the contents of soil organic matter so that we might stabilize soil microbial communities and increase the abundance of beneficial bacteria by improving soil fertility. Notably, the research found that the abundance of Bacteroidetes increased in extremely harsh soils, and Bacteroidetes transferred and enriched a large amount of Ni ( \n Figure S2 \n ), which is consistent with Jiao et al. (2022) findings. Besides, Tang et al. (2019) found that Bacteroidetes also affected the availability of Cu and Zn. Okkeri and Haltia ( 1999 ; 2006 ) found that the zinc-transporting P-type ATPase C was not only related to the transport of Zn but also to Cd and Pb, which is consistent with our genetic prediction. Similarly, nickel transport protein is a novel metal regulatory protein associated with heavy metal Ni transport ( Dosanjh and Michel, 2006 ; Wang et al., 2014 ). In addition, there are functional genes secreting siderophore ( Carroll and Moore, 2018 ), indoleacetic acid ( Estenson et al., 2018 ) and ACC deaminase ( Glick, 2005 ) in rhizosphere soil microorganisms. It is indicated that many plant growth-promoting bacteria (PGPB) resist HMs in rhizosphere soil. They significantly improved plant growth in heavy metal-contaminated soils and could enhance heavy metal phytoremediation by binding super-enriched plants ( Ahemad, 2019 ). Xu et al. (2022) found that Actinobacteriota and Gemmatimonadota promoted plant growth and fixed Cd in rhizosphere soil. Li et al. (2022) found that Sphingomonas promoted plant growth and degraded organometallic compounds to remediate soil HMs contamination. Similarly, Ham et al. (2022) found that Pseudopyroactor promoted plant growth and improved antioxidant capacity. PGPB significantly improve plant growth in heavy metal-contaminated soils ( Ahemad, 2019 ) and may represent potential bacterial strain resources in plant-PGPR combined remediation of HM-contaminated soils. The presence of HMs significantly destabilizes network structure ( Jiang et al., 2020 ; Caili et al., 2022 ). A more complex network represents that microbial activities and interactions are more active and intensive, which might have beneficial functions in the phytoremediation of HMs in soils ( Hou et al., 2019 ). Plants consistently interact with a core set of microbes contributing to plant performance ( Luo et al., 2022 ). We found that the network of LZ was the most complex ( \n Figure 3C \n ), and Boehmeria nives (L.) showed higher TFs of all ten metals. Thus, a good rhizosphere ecological network might be helpful for plants to absorb heavy metals."
} | 3,788 |
37752236 | PMC10522574 | pmc | 1,044 | {
"abstract": "All biology happens in space, and spatial structuring plays an important role in mediating biological processes at all scales from cells to ecosystems. However, the metabolomic structuring of the coral holobiont has yet to be fully explored. Here, we present a method to detect high-quality metabolomic data from individual coral polyps and apply this method to study the patterning of biochemicals across multiple spatial (~1 mm - ~100 m) and organizational scales (polyp to population). The data show a strong signature for individual coral colonies, a weaker signature of branches within colonies, and variation at the polyp level related to the polyps’ location along a branch. Mapping metabolites to either the coral or algal components of the holobiont reveals that polyp-level variation along the length of a branch was largely driven by molecules associated with the cnidarian host as opposed to the algal symbiont, predominantly putative sulfur-containing metabolites. This work yields insights on the spatial structuring of biochemicals in the coral holobiont, which is critical for design, analysis, and interpretation of studies on coral reef biochemistry.",
"introduction": "Introduction Spatial patterns in natural communities are an illustration of the processes that shape them. These patterns in biological systems emerge due to a combination of both biotic and abiotic factors. Just as advances in remote sensing allow for the exploration of increasingly larger scales, advances in molecular methods now facilitate the investigation of decreasingly smaller scales. Molecular variation at the atomic level can now be revealed from single cells to organisms 1 to whole ecosystems 2 . Despite a relatively small spatial footprint (~280,000 km 2 ), coral reefs are one of the most diverse and productive ecosystems 3 . Processes such as dispersal, community interactions, and disturbances act together with environmental factors to create spatial signatures on the reef landscape 4 . At the macroscale, spatial dimensions vary from 10 s to 1000 s of meters, defining reef-wide patterns of organization 5 . At the mesoscale (i.e., meters to centimeters) corals often form their own local patterning 6 – 8 . The patterns seen on the macro and meso-scale in coral reefs are often a product of microscale structuring on a single coral colony 9 , where the coral holobiont creates micro-environments which host unique viral, microbial, and biochemical assemblies 2 , 7 , 10 , 11 . However, the extent to which coral biochemistry changes within and between scales has yet to be thoroughly addressed. Here, we developed a metabolomics approach to investigate coral biochemistry starting from the fundamental organizational unit of a coral—the polyp. We analyzed the metabolomes of individual polyps from multiple branches across multiple colonies within a reef, to assess the spatial distribution of biochemicals across several spatial (~1 mm–100 m) and organizational (polyp, branch, colony, population) scales. Understanding the variability and spatial distribution of biochemicals across scales on coral reefs provides insight into the spatial ecology of the coral holobiont, which is critical for experimental design and data interpretation in future research. For this study, three branches were collected from each of 19 Montipora capitata colonies on a patch reef (21.451, −157.795) in Kāneʻohe Bay, Oʻahu, Hawaiʻi. Two branches were sampled from opposite sides of the colony, and one from the center (Fig. 1a ). Six single-polyp biopsies were removed from each branch with a 16-gauge, blunt-tipped probe needle (Grainger) by sampling coral tissue directly surrounding an individual corallite to ensure the isolation of a single-polyp. The sample was then removed from the needle by pushing air through a syringe directly into a 1.5-ml, amber glass vial containing 100 μl of 70% methanol (Fig. 1b–f ). The branch sampling scheme was as follows: Polyp 1 : 1 cm above the base of branch; Polyps 2 and 3 : the next consecutive polyps from polyp 1 toward the tip; Polyp 4 : ¼ distance between Polyp 3 and tip of branch; Polyp 5 : ½ distance between Polyp 3 and tip; Polyp 6 : tip of branch). This yielded a total sample set of 342 individual polyps from 57 branches across 19 coral colonies (Fig. 1 ). These samples were randomized and assigned arbitrary labels prior to mass spectrometry analysis. Fig. 1 Single-polyp sampling scheme. a Sampling scheme where 19 Montipora capitata colonies were selected from Reef 13 (21.451, −157.795) in Kāneʻohe Bay, Oʻahu, Hawaiʻi. Three branches were clipped from each colony, and six single polyps were sampled from each branch for a total sample size of 342 polyps. b – e Each branch was sampled at a single-polyp resolution using a 16-gauge probe needle. f Each single-polyp sample was ejected from the 16-gauge needle directly into an amber, glass vial containing 100 µl of 70% methanol. Samples were processed for untargeted metabolomics analysis via liquid chromatography-tandem mass spectrometry (LC-MS/MS) as previously described in Roach et al. 12 . Data files were converted to mzXML format for being processed with MZmine 2.53, the Global Natural Product Social Molecular Networking (GNPS) web-based platform, and SIRIUS 13 – 15 . These files were then compared to samples of bleached corals and symbiont isolates using molecular mapping 12 to identify the putative source of metabolites. (i.e., metabolites from the coral host, algal symbiont, or shared). Additionally, raw data files were analyzed through Compound Discover for putative molecular annotation 16 . For detailed methodology of the workflow and analysis, please see the “Methods” section below.",
"discussion": "Results and discussion This single-polyp method produces a robust, high-quality metabolomics data signal LC-MS/MS analysis of single-polyps collected with our approach produced a robust metabolome profile, similar to that of the more traditionally sampled, larger coral nubbins (Supplementary Figs. S1 and S2 ). The 342 individual polyps collected from 19 coral colonies produced metabolome data with a total of 555 unique metabolite features not found in blanks of which 67 (12.07%) had an MS/MS spectral match to a known compound in the GNPS database (details of GNPS library hits available in Supplementary Data 1 ). These were then manually inspected for good MS/MS alignment and curated to remove non-biological compounds resulting in 52 compound annotations that are at level two according to the Metabolomics Standards Initiative in Sumner et al. 15 (Supplementary Data 1 ). We also searched this data against the mzCloud database and found 50 reliable annotations (level two annotations, above an alignment score of 90%, Supplementary Data 2 ). To further classify metabolite features, we used the molecular family classifier software, CANOPUS , to assign compounds to molecular classes and found that 75.2% of the MS/MS spectra detected could be assigned to the Class level of the ClassyFire molecular taxonomy 17 , which are considered level three according to Sumner et al. 15 . This demonstrated that while most of the MS/MS spectra in our coral metabolomes did not have direct hits to the GNPS libraries, they could be more readily assigned to classes of compounds. Variation in coral metabolomes across spatial and organizational scales To assess the differences in general metabolomic profiles we compared the richness and Shannon Entropy of samples. There were no significant differences in richness (ANOVA p = 0.11) or entropy (ANOVA p = 0.84) by the different sampling areas on a branch (i.e., polyp number). There was, however, a significant difference in both richness (ANOVA p < 0.0001) and entropy (ANOVA p < 0.0001) between the different colonies. Within a single colony there were significant differences (ANOVA p < 0.05) in richness and entropy by branch in 8 out of 19 colonies (42.1%). PERMANOVA analysis demonstrated significant effects of colony ( p ≤ 0.001) and polyp location ( p = 0.015). The data were visualized in a principal component plot (Fig. 2a ) displaying a strong signature driven by colony. Discriminant analysis supervised by colony validated this signature (Fig. 2b ) with 100% of the samples being classified correctly. Colonies 962 and 983 were notable outliers in both analyses compared to the other colonies which clustered more tightly. Average within-colony variance (7.52 × 10 −4 ) was less than half the average between-colony variance (1.52 × 10 −3 ; p < 0.05) (Fig. 1f ). In addition to the colony signature, there was a distinction between individual branches within a colony, with the average variance within a branch being significantly less than the average variance between branches from the same colony (Fig. 2c ). Independent discriminant analyses for each colony supervised by branch demonstrated a range of misclassification rates (Min = 0%; Max = 55.56%) with an average misclassification rate of 27.19% (Std. Dev. = 21.01%). Despite the higher average misclassification for branches within a colony, it is notable that there was 0% misclassification for 6 out of the 19 colonies (31.58%). This indicates a highly significant signature of branches within a colony for some colonies, while less so for others. Fig. 2 Single-polyp statistical analyses. a Principal component plot and ( b ) canonical plot from discriminant analysis supervised by colony using all metabolites. Color legend for genotypes is the same for ( a , b ). c Box plots showing the mean standard deviation of the abundance of every compound in all samples (“across colonies”), between branches within a colony (“between branches”), and between polyps within branches (“within branches”); *** p < 0.05. Boxplots are median with quartiles and whiskers extending 1.5 IQR beyond quartiles. d Canonical plot from discriminant analysis supervised by polyp number (i.e., position on the branch with lower numbers being closer to the base of the branch). e Pairwise Bray–Curtis dissimilarity matrix for polyps on the same branch. Shared letters represent no statistically significant differences between groups (pairwise Kruskal–Wallace; α = 0.95). f Heat map of ClassyFire SubClass chemical groups that were significantly correlated with the distance from the base of the branch with Pearson’s r values shown on the right of the map. Signature of polyp location within a single coral branch In addition, we found a significant signature of the polyp location (PERMANOVA p = 0.015), which was strengthened when the colony was considered (PERMANOVA p = 0.001). A canonical plot generated via discriminant analysis supervised by polyp number (i.e., sampling location on a branch) revealed that the samples formed three distinct clusters (base, middle, and tip of the branches; Fig. 2d ). This analysis was even able to discriminate between adjacent polyps with high rates of accuracy (93.33%) (Fig. 2d ). Furthermore, adjacent polyps 1 and 2 were found to be significantly dissimilar to one another (Fig. 2e ). To provide a general assessment of the type of molecules changing with distance from the base of a coral branch, the ClassyFire classifications of each metabolite were compared to the distance from the base of the coral branch using linear regression. Only two molecular SubClass families were significantly negatively correlated with distance to base after Bonferroni p value correction (triterpenoids and hydroxysteroids) while numerous SubClasses increased in relative abundance as sample distance from the base increased including sesquiterpenoids, amino acids and derivatives, organosulfonic acids and derivatives, carbamate esters, and others (Fig. 2f , Supplementary Figs. S3 and S 4 and Supplementary Data 2 ). Within branches, there was minimal correlation between pairwise sample dissimilarity (Bray–Curtis) and physical distance between samples (mean R 2 adj = 0.0295), indicating little support for isolation by distance at this scale in the coral metabolome. However, a large portion of the variance among polyps within a branch was explained by the distance of the polyp from the base of the branch (artificial neural network regression analysis R 2 = 0.83). Independent linear regressions of distance to the base with all biochemical features in the dataset were conducted, and seven biochemicals were significantly correlated ( p < 0.05) with the distance to the base of the branch with an R 2 > 0.10. Though none of these biochemical features had GNPS annotations, five of these seven compounds belonged to a single MS/MS network in GNPS (Fig. 3a, b ). SIRIUS 4.8.2 14 was used to calculate in silico molecular formulas and structures for all five molecules of interest in the GNPS network. Many of the molecular formulas predicted were sulfur containing, thus, we analyzed the MS/MS spectra of one of the more abundant molecules ( m / z = 458.2452, C 25 H 35 O 3 N 3 S) and were able to identify fragments in the low mass range that contained at least a single sulfur atom, providing further support for this molecular formula (Supplementary Fig. S5 ). Searching of the mzCloud library with Thermo® Compound Discoverer software (see “Methods”) also predicted the same molecular formula but had no annotation for the spectrum. Because the various cheminformatic approaches used to identify these molecules did not reveal plausible candidates, they remain structurally unknown (level-3 according to the metabolomics standards initiative 15 and level 4 according to identification confidence levels 18 ). MASST searching against GNPS public data 19 with the MS/MS spectrum of the compound listed above revealed that this molecular spectrum (m/z458.2452) was found in four other datasets on GNPS, all coral-associated (see Supplementary Materials for link to these publicly available datasets), supporting its existence as a coral metabolite of interest. Fig. 3 Metabolites significantly correlated with the distance from the base of a coral branch. a Linear regressions of all metabolites with a significant R 2 > 0.10. Metabolite IDs shown in the figure panel represent the GNPS ID, followed by the mass charge ratio, and retention time. b Network and putative molecular formulas for the five highly correlated metabolites all in the same GNPS subnetwork. All masses listed are exact masses without adducts. A high-definition full page version of ( b ) is included in the Supplementary (Supplementary Fig. S6 ). c Linear regressions of the sum of all host-associated and algal-associated metabolites with the distance from the base of a coral branch. Metabolomic signatures within a branch are largely driven by host-derived molecules To better describe the source of variation along the length of a branch we applied our holobiont metabolome mapping approach (originally developed in ref. 12 ) where LC-MS/MS data from bleached M. capitata and purified algal symbionts were co-networked on GNPS with this single-polyp data. MS/MS spectral matching across the two datasets enabled assignment of compounds as “host” if they were 10x more abundant in the bleached corals and as “symbiont” if they were at least 10x more abundant in algal pellets. Molecular mapping revealed all seven of the sulfur containing metabolites that correlated with branch length (Fig. 3a ) were significantly enriched (Kruskal–Wallace p < 0.05) in the coral host relative to the algal symbionts. Furthermore, the sum of all host-associated metabolites was significantly correlated with the distance from the base of the branch ( p < 0.001) with relatively strong predictive power ( R 2 = 0.159); whereas algal-associated metabolites were significantly less predictive of the distance along the branch with the R 2 being an order of magnitude lower than for host-associated metabolites ( p = 0.038, R 2 = 0.009) (Fig. 3c ). Further discussion and conclusion This pattern may reflect the differentiation of growing apical polyps in Acroporidae or may be due to gradients in abiotic factors, such as light or flow. These findings provide important insight into the spatial variability and organization within and between coral polyps, branches, colonies, and populations. Understanding the amount of variability across micro- and macro-scales directly impacts our understanding of spatial structuring within and between scales of the coral holobiont 20 , 21 . As this work provide evidence for non-random structuring of the metabolome at multiple spatial and organizational levels, it offers valuable insight into the current debates concerning the variability and heterogeneity of metabolomes across scales, which is a critical component to consider when designing the approach to large scale ecological sampling schemes and interpreting data in future experiments.\n\nFurther discussion and conclusion This pattern may reflect the differentiation of growing apical polyps in Acroporidae or may be due to gradients in abiotic factors, such as light or flow. These findings provide important insight into the spatial variability and organization within and between coral polyps, branches, colonies, and populations. Understanding the amount of variability across micro- and macro-scales directly impacts our understanding of spatial structuring within and between scales of the coral holobiont 20 , 21 . As this work provide evidence for non-random structuring of the metabolome at multiple spatial and organizational levels, it offers valuable insight into the current debates concerning the variability and heterogeneity of metabolomes across scales, which is a critical component to consider when designing the approach to large scale ecological sampling schemes and interpreting data in future experiments."
} | 4,462 |
34815411 | PMC8611031 | pmc | 1,045 | {
"abstract": "Living cells have the capability to synthesize molecular components and precisely assemble them from the nanoscale to build macroscopic living functional architectures under ambient conditions. The emerging field of living materials has leveraged microbial engineering to produce materials for various applications but building 3D structures in arbitrary patterns and shapes has been a major challenge. Here we set out to develop a bioink, termed as “microbial ink” that is produced entirely from genetically engineered microbial cells, programmed to perform a bottom-up, hierarchical self-assembly of protein monomers into nanofibers, and further into nanofiber networks that comprise extrudable hydrogels. We further demonstrate the 3D printing of functional living materials by embedding programmed Escherichia coli ( E. coli ) cells and nanofibers into microbial ink, which can sequester toxic moieties, release biologics, and regulate its own cell growth through the chemical induction of rationally designed genetic circuits. In this work, we present the advanced capabilities of nanobiotechnology and living materials technology to 3D-print functional living architectures.",
"introduction": "Introduction 3D bioprinting technology, which is relatively well-established for printing mammalian cells in the context of tissue engineering, has more recently been applied to print microbial cells for biotechnological and biomedical applications 1 – 8 . Although inkjet printing, contact printing, screen printing, and lithographic techniques have been explored to print microbes, extrusion-based bioprinting has become one of the most widely used techniques due to its simplicity, compatibility with a variety of bioinks, and cost-effective instrumentation 2 , 9 – 11 . In an early example of this concept, a mixture of alginate and E. coli was extruded onto a printing surface consisting of calcium chloride, upon which the alginate molecules crosslink to form a solidified gel 7 . A similar ionic crosslinking strategy was exploited to generate photocurrent with 3D printed cyanobacteria 12 . In another approach, a multi-material bioink comprised of hyaluronic acid, κ -carrageenan, fumed silica, and a photo-initiator was employed to 3D-print Pseudomonas putida and Acetobacter xylinum . Also, photo-crosslinked pluronic F127 acrylate-based bioinks have been utilized to print living, responsive materials/devices, and catalytically active living materials 4 , 6 , 13 . An alternative strategy made use of freeze-dried Saccharomyces cerevisiae as the primary component of a bioink formulation consisting of nanocellulose, polyethylene glycol dimethacrylate, and a photoinitiator 3 . The latter approach yielded remarkably high cell densities of 10 9 cells ml −1 , but the need for freeze-drying could significantly affect the survival rate of other microbial species as well as their thixotropic behavior. In an interesting approach, the viscoelastic gel-like characteristics of Bacillus subtilis ( B. subtilis ) biofilms facilitated direct printing. However, the wild-type biofilms were unable to maintain the printed line widths (as they expanded three-fold in width after printing), while the engineered variants had lower storage modulus and viscosity that restricted their printing in multiple layers 8 . In yet another strategy, a fused deposition modeling was adapted to deposit molten agarose (75 °C) containing B. subtilis spores onto a cold substrate (16 °C), resulting in hardened patterns upon cooling 5 . Here, the high-temperature processing works well for spores, but limits applicability to a wide range of cell types. Although the above examples demonstrate that many bioink designs have already been explored, none so far have fully leveraged the genetic programmability of microbes to rationally control the mechanical properties of the bioink. This would be advantageous for several reasons, including the possibilities of more sustainable manufacturing practices, raw material fabrication in resource-poor environments (terrestrial or extra-terrestrial), and enhanced material performance through bio-inspired design and the precision of genetic programming. In contrast to the examples described above, we envisioned to (1) design an extrudable bioink that had high print fidelity, (2) produce the bioink entirely from engineered microbes by a bottom-up approach and (3) create a programmable platform that would enable advanced functions for the macroscopic 3D living architectures, and thereby push the emerging field of living materials to unexplored frontiers 3 – 10 , 12 – 28 . In this work, we present microbial ink that is produced entirely from the genetically engineered E. coli biofilms. We show the detailed characterization of the microbial ink and demonstrate its structural and shape integrity. Further, by embedding genetically programmed E. coli cells in the microbial ink, we demonstrate the 3D printing of therapeutic living material, sequestration living material, and regulatable living material."
} | 1,260 |
36506141 | PMC9730456 | pmc | 1,046 | {
"abstract": "The emergence of ionotronic materials has been recently\nexploited\nfor interfacing electronics and biological tissues, improving sensing\nwith the surrounding environment. In this paper, we investigated the\nsynergistic effect of regenerated silk fibroin (RS) with a plant-derived\npolyphenol ( i.e. , chestnut tannin) on ionic conductivity\nand how water molecules play critical roles in regulating ion mobility\nin these materials. In particular, we observed that adding tannin\nto RS increases the ionic conductivity, and this phenomenon is accentuated\nby increasing the hydration. We also demonstrated how silk-based hybrids\ncould be used as building materials for scaffolds where human fibroblast\nand neural progenitor cells can highly proliferate. Finally, after\nproving their biocompatibility, RS hybrids demonstrate excellent three-dimensional\n(3D) printability via extrusion-based 3D printing to fabricate a soft\nsensor that can detect charged objects by sensing the electric fields\nthat originate from them. These findings pave the way for a viable\noption for cell culture and novel sensors, with the potential base\nfor tissue engineering and health monitoring.",
"conclusion": "4 Conclusions Designing electronics exploiting\nthe mechanisms of biological tissues\nand improving sensing with the surrounding environment is challenging\nmainly when turning toward using natural rather than synthetic materials.\nSilk fibroin is used in this context. In this study, we developed\na novel ionotropic silk-based material that serves as a substrate\nfor cell seeding and proliferation, mimicking the physiologically\nperformed functions by the extracellular matrix as close as possible.\nWe reported results showing that the hybrid silk formation can be\napplied as anchorage to many cell types, including fibroblasts and\niPSc-derived neural precursor cells. It was demonstrated how adding\ntannin to the RS increases the ionic conductivity and allows the cells\nto migrate and connect. We demonstrated the use of silk fibroin in\nsoluble plant-derived polyphenol as a biomaterial ink to prepare a\n3D grid that senses an electric field originating from static charges\non the surface of an object. We report a proof of concept on using\nthese ionic sensors to detect the electrical amount produced by cell\ndivision by a non-invasive method. The results enlarged materials\nengineering, envisioning applications in noncontact sensors for implantable\ndevices.",
"introduction": "1 Introduction Mimicking the sensing\nmechanisms of some animals that use electroreceptors\non their skin to catch prey without physical contact, 1 − 3 ionotronic materials can be designed to sense a variety of stimuli,\nsuch as check noncontact spatial perception. 4 − 6 Regenerated\nsilk is a biopolymer derived from Bombyx\nmori with tunable processability and physical properties,\nwhose unique advantages rely on biocompatibility, biodegradability,\nand bioabsorption, 7 − 10 as well as ionic conductivity due to the presence of Ca 2+ ions that have strong capability to capture water molecules. 11 , 12 These merits promote regenerated silk fibroin (RS) as a biocompatible\nionotronic material. Among the plant-derived polyphenols, tannins\noffer important renewable\nresources due to their relative abundance, ease and sustainability\nof the extraction process, and their good reactivity. 13 , 14 Tannins can be divided into two main classes, condensed and hydrolyzable,\nwhich have different chemical structures and, thus, properties. 15 In the current research, we focused our attention\non hydrolyzable chestnut extract for its abundance at the national\nlevel (Italy) as well as for the higher antioxidant properties of\nhydrolyzable tannin than the condensed one. 16 Tannins are the second natural resource of polyphenols behind\nonly\nlignin, 14 which would be an interesting\nresource for this kind of application. On the other hand, although\nresearch is looking for useful ways to exploit lignin, industrially,\nit is not yet possible to rely on raw lignin for value-added application. 17 Indeed, the high heterogeneity due to the origin\nand extraction method limits the industrial development, 17 and nowadays, most of the raw material is still\ndirectly burnt. 18 Notably, the coassembly\nof tannins with proteins has been considered\na facile method for developing a bioinspired RS-based hydrogel sealant. 19 In this regard, stimuli-responsive conductive\nhydrogels have been exploited as sensors for health monitoring, biocompatible\nelectronic devices, and wound healing in tissue engineering. 20 − 22 Our previous work observed that tannins play a crucial role in the\none-step method to generate an elastomer-like material based only\non RS and calcium chloride (CaCl 2 ). 23 Such materials present adequate mechanical properties\nto follow\nthe displacement and deformation of soft biological tissues, when\nshaped in holey structures, as we previously showed. 24 In this work, inspired by ion electronics, Ca 2+ -tannin-RS\nfilms (RS/T), with different tannin concentrations, were prepared\nvia CaCl 2 dissolution of degummed silk and the subsequent\naddition of tannin. The introduction of tannin to Ca 2+ -RS\nfilms endowed the sample with higher ionic conductivity. More importantly,\nthese solutions were used as biomaterial ink for 3D-printing multilayered\ngrid-like sensors that could detect the spatial localization of charged\nobjects. We report as a proof of concept on using such 3D-printed\nsensors to detect the bioelectrical activity of live microorganisms.\nFinally, we developed a process based on such hybrid RS protein, which\ncan self-assemble into biocompatible scaffolds at room temperature.\nThis biobased method can provide a valuable environment for cell culture\nas an alternative to Matrigel, severely limited by the variability\nin its composition and the presence of xenogenic contaminants. 25 − 27 Mimicking the conduction mechanism of organs by conducting electricity\nvia ions trapped inside, 28 , 29 hybrid RS promotes\nthe differentiation and proliferation of fibroblasts and neural progenitor\ncells (NPCs) at rates that match its use in biomedical implants.",
"discussion": "3 Results and Discussion We considered\nthe RS-based materials as structural and supporting\nsubstrates possessing ionic conductivity due to the presence of Ca 2+ ions. 35 Hence, we focused on\ninvestigating the ionic conductive behavior of RS and RS/T samples.\nAs reported in Figure 1 A, when the Ca 2+ mass ratio increases from 90/10 to 60/40,\nthe conductivity of the RS increases. This increase in ionic conductivity\nrelies on the content of both Ca 2+ ions and trapped water\nin the final materials. Thus, we focus on the 60/40 composition for\nthe subsequent investigation and 3D printing. As reported in Figure 1 A, the conductivity\nof RS/T significantly increases also as a response to the presence\nof tannin. For example, when the tannin content passes from 1 to 5\nwt %, the ionic conductivity increases by two orders of magnitude.\nBased on these experimental results, we can propose a similar model\nof the effect of tannin addition on the conductivity enhancement of\npolymer films; 36 a large number of hydroxyl\ngroups on tannin provided opportunities for RS to interact with tannin\nthrough van der Waals forces (such as hydrogen bonding interactions).\nIn contrast, the interactions between RS chains remain the same, which\nresults in samples with higher conductivities. Figure 1 Evaluation of the relative\nhumidity and temperature on the ionic\nconductivity of the prepared samples. (A) The ionic conductivity of\nRS films with different RS/CaCl 2 weight ratios and 60/40\nRS with additional tannin content was measured at 25 °C and environmental\nhumidity ( i.e. , 50%RH). (B) Conductivity–humidity\ncurve of 60/40 RS, 60/40 RS/T1, and 60/40 RS/T5. (C) Conductivity\nchanges of the 60/40 RS/T5 film at 25 and −15 °C. The mutual interactions of ions, tannin, and silk\nfibroin are humidity-dependent,\nthus allowing the tuning of ionic conductivity. Indeed, as shown in Figure 1 B, the current intensity\nof RS, RS/T1, and RS/T5 samples varies in step, with RH changes from\n0 to 80%. Fitting the current–voltage curves and assuming valid\nOhm law, current data can be converted to the conductivity change–RH\ncurve as shown in Figure 1 B. The slope of the linear region of the data reported in Figure 1 B allows us to estimate\nthe sensitivity value ( i.e. , δ(Δ S / S 0 )/ΔRH, where Δ S , S 0 , and ΔRH are the\nconductivity change, initial conductivity, and RH change, respectively),\nwhich is 4.9%RH –1 . It should be observed that the\nconductivity of the RS/T5 sample exhibits a trend of decrease and\nthen increase with increasing humidity values; this effect could be\nattributed to a slight swelling of the sample with increasing thickness\nas recently reported elsewhere. 33 We then\nexamined the effect of temperature change from −15 to 25 °C\non the 60/40 RS/T5 film ( Figure 1 C and Figure S1 ). The film\nshows evident conductivity variation amplitude in three cycles at\nthese two temperatures. By redispersing RS and RS/T samples\nin DMEM, the resulting material\nhas the aspect of a gel ( Figure 2 A). Upon gelation, fibroin yielded an opaque color\ndepending on the different compositions of the solutions. OD changes\nover time for RS and RS/T solutions are reported in Figure 2 B. The observed OD changes\nare associated with the shift in the fibroin structure reported below.\nThe OD increase observed with increasing tannin content up to 5 wt\n% suggests a saturation effect typical of gelation. 34 This behavior is less marked when lower tannin contents\nare used. Figure 2 (A) Photograph of RS and RS/T gels. (B) Optical density changes\n(at 550 nm) of the RS and RS/T solutions as a function of time. The XRD technique was used to characterize the\ncrystalline structure\nof the prepared films. XRD patterns of RS in FA solution, after drying,\nand bulk silk fibroin are reported in the inset of Figure 3 A. The regenerated silk exhibited\na broad, amorphous peak centered at around 2θ = 25°, which\nmainly arises from the convolution of different bands at 19.4°,\n20.3°, 24.6°, and 29.3°, corresponding to silk I crystalline\nspacing values of 0.44, 0.41, 0.35, and 0.30 nm, respectively, in\nagreement with literature data. 37 − 40 This pattern differs from bulk fibroin, mainly ascribed\nto the silk II crystalline structure. Formic acid dissolves silk,\ndisrupting the hydrogen bonds within the antiparallel β-sheet\nstructures. In RS, antiparallel β-structures are thus substituted\nby β-turns, which confer a different toughness to the silk fibers. Figure 3 (A) XRD\npatterns for RS and RS/T samples for different weight concentrations\nof tannin regenerated in FA solution and (B) after the redispersion\nwith DMEM. Inset of panel (A): XRD curves for bulk fibroin and RS.\nCurves are vertically shifted for clarity. (C) The asterisk indicates\nthe FTIR spectra of neat RS and RS/T composites starting from the\n60/40 composition obtained from FA solution and after redispersion\nin DMEM. (D) Structure composition of the prepared specimens with\nboth procedures. The asterisks indicate the samples redispersed in\nDMEM. Moreover, superimposed on an amorphous silk band,\na pattern of\nlow-intensity Bragg peaks is visible, whose linewidth indicates the\npresence of crystals of micrometric size in the silk matrix. According\nto the literature, these crystals can be ascribed to a mixture of\ndifferent hydrated forms of CaCl 2 . 41 In particular, the peak at 2θ = 16°, not present in the\nXRD pattern of anhydrous CaCl 2 , is a clear marker of the\nhydration of CaCl 2 . Samples obtained by adding tannin\nshow a silk XRD curve with a\nbroad band at 2θ = 25°, very similar to RS, proving that\ntannin preserves the crystalline structure of regenerated silk. However,\nBragg peaks with different relative intensities can be observed in\nthese samples, suggesting a different orientation of the corresponding\nCaCl 2 microcrystals. Although not very high, we also noted\nthat the power of the Bragg peaks varies on repeated measurements\nof the same sample without a clear temporal trend (see Figure S2 ). We speculate that these variations\ncan be ascribed to rearrangements on the micrometric scale, probably\ndue to the modification of the sample hydration degree. XRD\ncurves relative to the samples after DMEM treatment are reported\nin Figure 3 B for tannin\nat 0 wt %, RS/T1, and RS/T2. The silk II crystalline structure is\npreserved, but a different and more intense Bragg pattern, ascribed\nto the presence of microcrystals of varying composition and size,\nis observed. Also, in this case, XRD patterns of every sample slightly\nchange with time: additional measurements performed by rotating the\nsample (data not shown) exhibit a complex and very intense pattern\nof several Bragg peaks, indicating that the microcrystals’\npositions and orientations change due to the mechanical micrometric\nmovements and rearrangements of the gel form. Indeed, samples dissolved\nin DMEM, even after solvent evaporation, present a gel phase in which\nmicrocrystals may rearrange their orientation. Finally, from the XRD\nspectra, we did not record peaks related to the structure functions\nof liquid formic acid; 42 from this evidence,\nwe can assume a negligible amount of residual formic acid in our samples. Fourier transform infrared attenuated total reflection spectroscopy\n(FTIR-ATR) is a widely acknowledged method to investigate the chain\nconformation ( i.e. , β-sheets, random coils,\nα-helices, and turns) of silk fibroin cast films. 30 , 43 , 44 The comparison of the FTIR spectra\nbetween new RS and RS/T samples and those dispersed in DMEM ( Figure 3 C and Figure S3 ) shows the signature bands associated\nwith RS, such as amide I (1642 cm –1 ) and amide II\n(1515 cm –1 ), as previously reported. 30 The results reporting the secondary protein\nstructure content are reported in Figure 3 D. From these data, adding tannin to RS generally\nincreased the amorphous structure when the RS/T samples were redispersed\nin DMEM. MTT assay was performed on RS/T2 and RS/T5 samples.\nAfter 24 h,\nall the compounds with a concentration in the range of 7.8 μg/mL\nto 0.25 mg/mL resulted in safe results for the CaCo-2 cell line (see Figure S4 ); a very low cytotoxic effect (viability,\n<80%) was observed with RS/T5 at 0.5 mg/mL and with RS/T5 at the\nhighest concentration (1 mg/mL). On the contrary, RS/T2 demonstrated\nsafe results for all concentrations assayed. After 48 h of treatment,\nthe highest concentration resulted in cytotoxicity for cells: 0.5\nand 1 mg/mL. In particular, RS/T5 showed a viability <70% with\n0.5 mg/mL and a viability around 60% with 1 mg/mL. The best result\nwas obtained with RS/T2 with a safe profile after 24 h (up to 1 mg/mL)\nand 48 h (up to 0.5 mg/mL). The bioresorbability of the RS and RS/T\nsamples was also investigated ( Figure S4 ). The RS sample largely dissolves within 4 days, with a difference\nfor the RS/T samples that show degradation within 7 days. Fibroblast\nand NPC morphology, confluence, and growth indicate\nthe good biocompatibility of RS hybrids ( Figure 4 A,B). As described in the literature, we\nobserved that low doses of tannin were not cytotoxic to fibroblasts 45 and NPCs, confirming their antioxidant and cytoprotective\nproperties. In particular, tannin appears to positively influence\nfibroblast and NPC adhesion ( Figure 4 A,B) and fibroblast proliferation by the increase in\ncell confluence in Figure 4 A . The cell viability assay was evaluated\nbased on the ratio of live (>80%) and dead cells (automated cell\ncounting\nby Trypan Blue Viability Assay) and based on the fibroblast and NPC\nmorphology, growth, and confluence that were evaluated daily by an\ninverted microscope. Figure 4 (A) Fibroblast and (B) NPC adhesion and growth on RS and\nRS/T hybrids.\nBright-field (left) and confocal microscopy (right) images of human\nfibroblasts and NPCs seeded on modified glass surfaces covered with\nRS Hybrid films after 7 days of incubation in standard conditions.\nCells were labeled with the fluorescent Hoechst (blue channel) and\nWGA568 (red channel) dyes, targeting DNA and sialic acid—component\nof the plasma membrane—respectively. Confocal images correspond\nto maximum intensity Z axis projections over 10 μm.\nScale bars indicate 100 μm for bright-field microscopy and 20\nμm for confocal microscopy. Then, we explored the capabilities of these biomaterials\nto be\nused as inks for extrusion-based 3D printing ( Figure 5 A). The involved solutions possess a liquid-like\nbehavior, with G ″ higher than G ′ for all the tested solutions, as we previously showed. 24 Moreover, they exhibit a Newtonian behavior,\nwith a viscosity approximately equal to 4 mPa s and the absence of\na yield stress. As clearly stated in the literature, 46 , 47 those rheological properties do not favor the printability of complex\n3D shapes using the selected material. However, since the aim of this\npaper is the microfabrication of four-layer grids, the rheological\nproperties should only favor the extrudability of the materials, granted\nby the low viscosity, and the shape retention, connected to the surface\ntension. Figure 5 3D-printed sensing grids based on the RS/T5 sample. (A) Photos\nof the structures of RS, RS/T1, and RS/T5 grids after the 3D printing\nprocess and acetate foil removal. The scale bars indicate 10 mm. (B)\nLinewidth for each RS-based solution. (C) Photographs showing the\nRS/T grid transferred on the Teflon substrate and PET and S. cerevisiae fermenting yeast objects positioned\nto a distance of 1 cm. The scale bar indicates 10 mm. (D) Current–voltage\ncharacteristics of charged PET and fermenting yeast placed at distances\nof 1 and 3 cm, respectively. (E) Electrical resistivity of the RS/T5\ngrid measured by positioning charged PET and fermenting yeast at distances\nof 1 and 3 cm, respectively. (F) The induced open-circuit voltage\nwas measured, while the PET object’s initial distance varied\nbetween the sample and the object. The red line indicates the oscillating\nlength between 3 and 1 cm. (G) Scheme illustrating the sensing mechanism\nof the RS/T ionotronic sample; nearby things that usually have static\ncharges on their surfaces cause the model to be charged through noncontact\nelectrification, thus inducing a voltage change in the RS/T electric\nfield receiver. (H) Graph of induced open-circuit voltage across the\nexternal load when the skin was used as a charged object (see Movie S1 ). Indeed, the dimension of the grid lines depends\non the interaction\nbetween the RS hybrid inks and the substrate. The final linewidth\nis larger than the needle size used for printing ( i.e. , internal diameter of 210 μm) and increases with the content\nof tannins (428, 514, and 546 μm, respectively, for RS, RS/T1,\nand RS/T5) due to more pronounced surface wettability, as previously\nshown. 24 Statistical analysis revealed\na significant difference ( p < 0.0001) when the\ntannins were added to the solution regardless of the tannin concentration.\nIn contrast, no significant differences arose between the grids made\nof different tannin-laden RS solutions ( Figure 5 B). Printing performance in terms of line\ndimensions could be improved in the future, taking into consideration\nthe increase in line dimensions during the grid design and the printing\nparameter selection ( e.g. , decreasing the volumetric\nflow). Thus, we will be able to fabricate grids with narrower line\ndimensions until the limit imposed by the Plateau–Rayleigh\ninstability. 46 Thanks to the design\nfreedom given by 3D printing, we focused on\nholey structures, which provide several mechanical advantages compared\nwith casted solid ones. 24 In particular,\nthe decrease in the unwanted lateral displacement, global stiffness,\nand the stress of the single grid line allow the grids to more easily\ncomply with movements and deformations of substrates on which they\ncan be transferred to act as sensors. The bioelectrical environment\nof the gut is regulated by the ionic\nconductance of live bacterial strains; 48 our 3D grid was\nutilized as a proof of concept to investigate the bioelectrical processes\nof a living yeast cell during its division ( Figure 5 C–E). Considering that this process\noccurs in the environment of the gut, we investigate the degradation\nin terms of weight loss of the prepared samples ( Figure S5 ). From these results, we observed that tannin reduces\nthe degradation. We choose the S. cerevisiae yeast cell because its genome is completely sequenced and not pathogens;\ntherefore, it can be handled without precautions. The bioelectrical\nsignal of S. cerevisiae was found to\ninduce electrical resistivity variations by moving the vial between\ntwo fixed positions. These findings pave the way for using such 3D-printed\nmaterials to investigate the gut microenvironment. We also test the\nsensing capability of the RS/T5 grid of a charged object ( i.e. , PET) that was fixed between two fixed distances ( Figure 5 C–E). When\nthe charged object vertically oscillated, the induced voltage generated\nby the relative position was measured ( Figure 5 F). This effect can be explained based on\nelectrostatic induction: when a charged object is approaching ( Figure 5 G), it originates\nan electric field that causes an electric current and potential drop\nacross the external load between the receiver and the ground ( Figure S6 ). Thus, by measuring the open-circuit\nvoltage, we can estimate the intensity of the object’s electric\nfield and, as a consequence, its distance from the receiver. These\nfindings prove the potentiality of the RS/T grid to detect a hand\nmovement, thus catching skin as a charged object in a noncontact manner\n( Figure 5 H and Movie S1 )."
} | 5,430 |
29492458 | PMC5821488 | pmc | 1,047 | {
"abstract": "A bioinorganic hybrid system based on bacterial surface display and biomimetic silicification for hydrogen production.",
"introduction": "INTRODUCTION The immoderate consumption of fossil fuels has caused many serious environmental problems, such as the greenhouse effect, global climate change, acid rain, and ozone depletion, which may greatly restrict the economic and social development of humans. The utilization of renewable and clean energy resources is highly desirable to address these global problems. Solar energy is the most important renewable energy resource on Earth, but it is difficult to harness. Thus, strategies exploiting the efficiency of photosynthesis for solar energy capture to enable the sustainable production of chemicals, such as hydrogen or other fuels, are urgently needed ( 1 , 2 ). Using H 2 production as an example, several innovative inorganic-biological hybrid systems that combine the light-harvesting capabilities of photosystem I or a semiconductor and the catalytic power of the hydrogenase enzyme or a metal catalyst have been developed in the past two decades ( 3 – 5 ). Notably, King and colleagues ( 6 – 8 ) developed a series of highly efficient biohybrid hydrogen production systems by using CdTe nanocrystals/cadmium sulfide (CdS) nanorods and purified [Fe-Fe]-hydrogenase from Clostridium acetobutylicum . However, because of the oxygen-sensitive nature of these reactions, the low yield of the isolated hydrogenase, and the high cost of precious metals, the efficiency of photocatalytic H 2 production by these bioinorganic hybrid systems requires further improvement. Whole bacterial cells have been used as biohybrid catalysts for H 2 production to overcome the limitations of both enzymes and synthetic catalysts in bioinorganic hybrid systems. Using this approach, several research groups have produced H 2 by hybridizing a titanium dioxide (TiO 2 ) semiconductor with wild-type bacterial cells ( 9 ). Recently, the significant breakthrough reported by Honda et al . ( 10 ) showed that recombinant Escherichia coli cells expressing both hydrogenase and maturase genes enabled whole-cell photocatalytic H 2 production with TiO 2 . However, the low biocompatibility of semiconductors and the oxygen intolerance of hydrogenases limit the use of this whole-cell system for practical H 2 production. Researchers have been working to optimize hybrid systems to protect the catalytic activity from oxygen stress and thus overcome these limitations. Notably, Douglas et al . ( 11 ) reported a smart self-assembling biomolecular catalyst for H 2 production using the bacteriophage P22 coat protein to encapsulate and protect the oxygen-tolerant [NiFe]-hydrogenase. Meanwhile, inspired by biomimetic mineralization, Tang et al . ( 12 ) developed a powerful silicification-induced green algae system for sustainable photobiological H 2 production under aerobic conditions. Inspired by these pioneering studies, we propose the development of an ideal whole-cell photocatalytic hydrogen production system that incorporates the following components: (i) a biocompatible light-harvesting inorganic semiconductor, (ii) active engineered E. coli cells as a biocatalyst, and (iii) a reliable shell to protect the reaction from oxygen. Here, we aim to develop engineered E. coli cells that combine the biosynthetic capability of CdS nanoparticles with surface-displayed heavy metal–binding proteins and the biocatalysis of oxygen-tolerant [NiFe]-hydrogenase. Combined with the biomimetic silicon encapsulation of whole E. coli cells, the bioinorganic hybrid system achieves photocatalytic hydrogen production under aerobic conditions ( Fig. 1 ). Fig. 1 Proposed surface-display biohybrid approach to light-driven hydrogen production in air.",
"discussion": "DISCUSSION Efficient solar-to-chemical conversion strategies are greatly needed to address global energy and environmental problems. One important challenge is the development of simple and practical photosynthetic systems for the production of hydrogen or other fuels by exploring the fundamental chemistry of photosynthesis. Here, we developed a photocatalytic hydrogen biosynthesis strategy based on a semiconductor–engineered E. coli biohybrid system encapsulated with biomimetic polymers that function under aerobic conditions. Unlike reported microorganisms that induce nanoparticle precipitation through natural metabolic pathways, the surface-displaying bacterial system applied here facilitates the controllable biosynthesis of biocompatible CdS semiconductors by the model organism E. coli under mild conditions. In addition, our results further verified the light-harvesting capability of biosynthesized CdS semiconductors. This strategy could be applied to another well-established biological chassis, such as Bacillus or yeast, to expand its applications. The self-aggregated E. coli cells protect the activity of an oxygen-intolerant enzyme, ensuring that the enzyme performs efficient biological whole-cell catalysis under simple and mild conditions in air. Together, this bioinorganic hybrid system based on bacterial surface display and biomimetic silica encapsulation technologies will likely become an alternative approach for the convenient utilization of solar energy."
} | 1,317 |
36748607 | PMC9837558 | pmc | 1,048 | {
"abstract": "Comparing obligate endosymbionts with their free-living relatives is a powerful approach to investigate the evolution of symbioses, and it has led to the identification of several genomic traits consistently associated with the establishment of symbiosis. ‘ Candidatus Nebulobacter yamunensis’ is an obligate bacterial endosymbiont of the ciliate Euplotes that seemingly depends on its host for survival. A subsequently characterized bacterial strain with an identical 16S rRNA gene sequence, named \n \n Fastidiosibacter lacustris \n \n , can instead be maintained in pure culture. We analysed the genomes of ‘ Candidatus Nebulobacter’ and \n \n Fastidiosibacter \n \n seeking to identify key differences between their functional traits and genomic structure that might shed light on a recent transition to obligate endosymbiosis. Surprisingly, we found almost no such differences: the two genomes share a high level of sequence identity, the same overall structure, and largely overlapping sets of genes. The similarities between the genomes of the two strains are at odds with their different ecological niches, confirmed here with a parallel growth experiment. Although other pairs of closely related symbiotic/free-living bacteria have been compared in the past, ‘ Candidatus Nebulobacter’ and \n \n Fastidiosibacter \n \n represent an extreme example proving that a small number of (unknown) factors might play a pivotal role in the earliest stages of obligate endosymbiosis establishment.",
"introduction": "Introduction Intracellular bacterial symbionts, or ‘endosymbionts’, are commonly found associated with a wide variety of hosts, but are largely investigated only in arthropods [ 1, 2 ] and a few other macroorganisms [ 3–5 ]. Most of the biological diversity of these symbiotic systems is actually found in the smallest of hosts: microbial eukaryotes (protists) [ 6 ]. Prokaryote-protist symbioses are especially well-known in ciliates [ 7 ], with the widespread genus Euplotes emerging as a model system that hosts many symbionts with unknown (but sometimes essential) roles [ 8, 9 ]. ‘ Candidatus Nebulobacter yamunensis’ ( \n \n Thiotrichales \n \n , \n \n Gammaproteobacteria \n \n ) is one of the so-called ‘accessory’ symbionts of Euplotes , inhabiting the cytoplasm of Euplotes aediculatus ( Fig. 1 ) [ 8, 10, 11 ] from geographical areas as distant as Italy and India. Like other accessory symbionts, ‘ Ca . Nebulobacter’s’ association with Euplotes appears stable under laboratory conditions, even if the bacterium is not proven to be essential for the host like the more well-known \n \n Polynucleobacter \n \n [ 12 ]. The reverse is not true: ‘ Ca . Nebulobacter’ seemingly cannot survive outside its host [ 10 ], which is also in line with most symbionts of Euplotes . ‘ Ca . Nebulobacter’ however stands out among accessory symbionts because it does not fall within any clade of specialized intracellular bacteria, such as \n \n Rickettsiales \n \n or \n \n Holosporales \n \n ( \n \n Alphaproteobacteria \n \n ), which are known for their reduced genomes and infectious features (one representative of these more ubiquitous symbionts, ‘ Ca . Cyrtobacter zanobii’, is curiously always detected co-occurring with ‘ Ca . Nebulobacter’ [ 8 ]). The closest relatives of ‘ Ca . Nebulobacter’ are free-living, and one in particular, \n \n Fastidiosibacter lacustris \n \n [ 13 ], shares an identical 16S rRNA gene sequence with it. Fig. 1. Fluorescence in situ hybridization on Euplotes aediculatus Eae1. Fluorescent signals from species-specific ‘ Ca . Nebulobacter yamunensis’ probe NebProb203 [ 10 ] highlights the presence of the symbiont within the host cytoplasm. Bar length corresponds to 10 µm. Very close symbiotic/free-living pairs of organisms are invaluable since the most common approach to investigating changes related to the onset of symbiosis is to compare symbionts with free-living relatives. The longer the time since their divergence, the noisier such data become, which is what makes genera with very close symbiont/free-living pairs such as \n \n Polynucleobacter \n \n [ 14, 15 ], \n \n Serratia \n \n [ 16 ], and \n \n Sodalis \n \n [ 17, 18 ] such important model systems. We sequenced the genome of the host-restricted ‘ Ca . Nebulobacter yamunensis’ and compared it with that of the free-living \n \n Fastidiosibacter lacustris \n \n [ 13 ] to see if key differences related to the transition to endosymbiosis could be observed. Instead, we found the two genomes to show extreme molecular and functional similarities, despite their apparently contrasting lifestyles, which we confirmed with a new attempt to cultivate both organisms in parallel. Overall, these two bacteria display completely different ecologies without any of the corresponding genomic features that are almost universally observed in other endosymbiotic systems, however recent. This intriguing puzzle may open a new window into understanding the earliest stages in the evolution of intracellular symbioses.",
"discussion": "Results and discussion No typical signatures of obligate symbiosis in the genome of ‘ Ca . Nebulobacter yamunensis’ A 2.1 Mbp-long draft genome was assembled for ‘ Ca . Nebulobacter yamunensis’ from the metagenome of Euplotes aediculatus strain Eae1, including 1986 predicted coding sequences (191 of which are potentially pseudogenes). The general characteristics of the genome are reported in Table 1 . Table 1. General genomic features and assembly statistics of ‘Ca . Nebulobacter yamunensis’ and \n \n Fastidiosibacter lacustris \n \n . “Completeness, %” refers to the values calculated by BUSCO [ 49 ]; for insertion sequences, only hits with a bit-score higher than 500 were selected. n/a, not applicable; ORFs, open reading frames \n ‘ Ca. Nebulobacter yamunensis’ \n \n \n \n \n Fastidiosibacter lacustris \n \n \n \n \n Host species \n \n \n \n Euplotes aediculatus (ciliate) \n \n \n n/a (free-living) \n \n Genome size, bp \n \n \n 2 098 416 \n \n \n 2 068 591 \n \n Contigs, no. \n \n \n 71 \n \n \n 62 \n \n Contigs longer than 1000 bp, no. \n \n \n 56 \n \n \n 51 \n \n Longest contig, bp \n \n \n 222 531 \n \n \n 256 557 \n \n N50, bp \n \n \n 89 098 \n \n \n 89 533 \n \n Average coverage \n \n \n 46 × \n \n 100 × \n \n Completeness, % \n \n \n 97.6 \n \n \n 99.2 \n \n G+C content, % \n \n \n 36.57 \n \n \n 36.33 \n \n Total ORF, no. \n \n \n 1986 \n \n \n 1914 \n \n of which, predicted pseudogenes, no. \n \n \n 191 \n \n \n 129 \n \n Coding density, % \n \n \n 83.2 \n \n \n 83.5 \n \n Average gene length, bp \n \n \n 934 \n \n \n 942 \n \n tRNA-coding genes, no. \n \n \n 38 \n \n \n 38 \n \n Insertion sequences, no. \n \n \n 25 \n \n \n 25 \n \n Accession number \n \n \n JAMBMW000000000 \n \n \n QLIR00000000 \n \n Reference \n \n \n This study \n \n \n Xiao et al 2018 \n In spite of its relatively small genome size, the reconstruction of ‘ Ca . Nebulobacter yamunensis’ functional and metabolic traits revealed little depletion ( Fig. 2a , Table S1, available in the online version of this article), especially compared with the extensive metabolic impairment reported for other essential [ 15 ] and accessory [ 19 ] symbionts of Euplotes . Nevertheless, while the central carbon/energy metabolism is virtually complete, the lack of the phosphofructokinase-1 gene suggests that the Embden-Meyerhof-Parnas glycolytic pathway is absent, as is also the case with every symbiont of Euplotes whose genome has been sequenced to date [ 9, 14, 15, 19, 35–37 ]. Complete biosynthetic pathways were also identified for purines, pyrimidines, fatty acids, glycerophospholipids (except phosphatidyl-inositol and cardiolipin), cell envelope components (peptidoglycan and lipopolysaccharides), and most l -amino acids. Exceptions include l -methionine and l -cysteine, where at least half the biosynthetic pathway-specific genes are missing. Only one (shared) gene, ilvE , was detected for the l -valine, l -leucine, and l -isoleucine pathways, while none was retrieved for l -histidine biosynthesis. Finally, l -aspartate biosynthesis may be absent due to the pseudogenization of the aspartate aminotransferase gene (Table S2). Co-factor biosynthesis is also mostly intact, with only the pantothenate, pyridoxine, and thiamine pathways likely absent. No corresponding transporter was annotated, suggesting that these co-factors are either not strictly required, or are imported from the host using non-specific (or otherwise annotated) transporters. Corroborating the idea that ‘ Ca . Nebulobacter’ is an unusually self-sustained symbiont, previous ultrastructural studies showed that stress conditions on the Euplotes host have a more conspicuous effect on its essential symbiont, \n \n Polynucleobacter \n \n , than on ‘ Ca . Nebulobacter’ [ 11 ]. Fig. 2. Comparative analyses of ‘ Ca . Nebulobacter yamunensis’ and \n \n Fastidiosibacter lacustris \n \n genomes. (a): selected predicted metabolic functions (see also Table S1). (b): NUCmer pairwise alignment dotplot: purple - long aligned single-copy regions, blue - repeats (mostly at contig ends); contigs are delimited by dashed lines. (c): MAUVE pairwise alignment; contigs are delimited by thin red vertical lines. An oddly complementary pattern was found between the transporter sets of ‘ Ca . N. yamunensis’ and the co-occurring accessory symbiont ‘ Candidatus Cyrtobacter zanobii’ ( \n \n Rickettsiales \n \n ) [ 19 ] (Table S3). For instance, all subunits for ribose ( rbsABC ) and putrescine ( potFGHI ) transporters appear only in ‘ Ca . Nebulobacter’, while none was detected for ‘ Ca . Cyrtobacter’. Conversely, malate ( yflS ) and riboflavin ( ribN ) transporters are exclusively found in ‘ Ca . Cyrtobacter’. In the absence of other clear signs of co-dependency, this pattern alone is unlikely to be a result of co-evolution. Rather, it might be suggestive of a form of niche separation in which the coexistence of ‘ Ca . Nebulobacter yamunensis’ and ‘ Ca . Cyrtobacter zanobii’ is facilitated by depleting different resources from the host cytoplasm. Analogous patterns of symbiont co-existence have been previously investigated in bacteriocyte-sharing bacteria in whiteflies [ 38 ]. High similarities in sequence and functional features between an obligately symbiotic and a free-living bacterium The 16S rRNA gene sequence extracted from the ‘ Ca . Nebulobacter yamunensis’ genome is identical to those of other known symbiotic ‘ Ca . Nebulobacter’ strains, as well as that of the free-living \n \n Fastidiosibacter lacustris \n \n strain SYSU HZH-2 T (=NBRC 112274 T ) [ 13 ]. On a genomic scale, the average nucleotide identity (ANI) between the genomes of ‘ Ca . Nebulobacter yamunensis’ and \n \n F. lacustris \n \n is 99.38 % (avg. aligned length=1 417 738 bp; genome coverage of 68.67 and 69.64% respectively), well above the conventional 95 % threshold for bacterial species. Pairwise alignments between the two assemblies confirmed their high degree of sequence similarity ( Fig. 2b ) and synteny ( Fig. 2c ). Even their overall genomic characteristics appear extremely alike, and many (129) ORFs were predicted to be pseudogenes in \n \n Fastidiosibacter \n \n as well (Table 1); in fact, 79 pseudogenes are shared by both bacteria investigated here and hence likely predate their divergence, rather than being linked to their ecological differences (Table S2). At least 1665 orthologous sequences are shared between the two genomes; among the remaining coding sequences unique either to ‘ Ca . Nebulobacter yamunensis’ (278) or \n \n F. lacustris \n \n (207), the vast majority encode for unknown hypothetical proteins (82.4 and 73.9% respectively; Table S4). The aforementioned metabolic traits described for the symbiotic ‘ Ca . Nebulobacter’ are almost identical in the free-living \n \n Fastidiosibacter \n \n ( Fig. 2a , Table S1). Minor differences, such as potentially incomplete synthetic pathways for aspartate and biotin (due to pseudogenization of the aspartate aminotransferase gene and bioF respectively) in ‘ Ca . Nebulobacter yamunensis’ ( Fig. 2a ), do not reflect the expected extensive loss-of-function and genome erosion predicted for even the most recently evolved obligate symbionts. In fact, some gene losses seem to have occurred in the free-living, rather than in the symbiotic lineage, as indicated by the lack of the asparagine synthase gene in \n \n Fastidiosibacter \n \n . Comparative analyses on genes involved in secretion systems also provided no clear indication of different life strategies. The lack of genes involved in substrate recruitment ( virD4 ) and target attachment ( virB5 ) of type IV secretion systems (T4SSs) in \n \n F. lacustris \n \n might speculatively indicate a more relaxed selective pressure in maintaining the functionality of the corresponding structures, which ‘ Ca . N. yamunensis’ might still use to mediate important interactions with its host. This remains however largely speculative, since virtually nothing is known about the molecular interactions between bacterial symbionts and ciliate hosts. Finally, \n \n F. lacustris \n \n also possesses a larger set of tra genes putatively involved in the F-like type IV conjugative system [ 39 ]. No proliferation of mobile elements was found in either genome. Parallel growth experiments under the same conditions gave the predicted different results for ‘ Ca . N. yamunensis’ and \n \n F. lacustris \n \n . The former, obtained from the lysate of its host (where it was confirmed to be present by FISH), never produced any colonies, whereas the latter did so 3–4 days after the inoculation. What does it take to become a symbiont? ‘ Ca . Nebulobacter yamunensis’ and \n \n Fastidiosibacter lacustris \n \n , contrary to what their names suggest, are closely related strains of the same species, with little to no differentiating genomic features. Nevertheless, one is an obligate endosymbiont found in multiple strains of the ciliate Euplotes aediculatus , stable under laboratory culture conditions but unable, or at least very fastidious, to grow outside its host; the other seems instead to be a typical free-living bacterium, growing under relatively simple conditions and surviving in culture collections without the need for a symbiotic partner. Any interpretation of this scenario comes with some difficulties. Based on the original environmental isolation of \n \n F. lacustris \n \n , it cannot be completely excluded that it was originally harboured by some host, possibly even Euplotes . This however would still not explain why \n \n F. lacustris \n \n is able to grow in isolation and ‘ Ca . Nebulobacter’ is not. ‘ Ca . Nebulobacter’/ \n \n Fastidiosibacter \n \n might both be opportunistic symbionts, like a gammaproteobacterium relative, \n \n Francisella \n \n [ 40, 41 ]. However, opportunistic \n \n Francisella \n \n spp. exhibit a broad host range, while ‘ Ca . Nebulobacter’ is only known from a single Euplotes host species. Moreover, members of \n \n Francisella \n \n can usually be grown in isolation [ 42 ], so this explanation alone does not account for the differential cultivability of the strains studied here. Based on current data, we provisionally lean towards the simpler interpretation, i.e. that ‘ Ca . Nebulobacter’ and ‘ \n \n Fastidiosibacter lacustris \n \n ’ are two lineages whose ecology diverged so recently that none of the changes usually associated with early stages of endosymbiosis are yet apparent in the genome of ‘ Ca . Nebulobacter’, making them ‘ecological variants’ of the same microbial species. Changes, genomic or otherwise, might well have occurred but remain undetected because we simply don’t know what to look for. If this is indeed the case, which should be confirmed by assessing phylogeny and trait diversity among multiple free-living ( \n \n Fastidiosibacter \n \n ) and symbiotic (‘ Ca . Nebulobacter’) strains, it is a sign of how much more we need to learn about how bacteria transition from a free-living to an endosymbiotic lifestyle – something that is particularly true if we suspect that drastic changes might occur in the context of regulatory networks [ 43 ], which are considerably less understood than changes directly affecting protein structures or metabolic pathways. At present, we cannot link any obvious molecular feature to such a fundamental and striking phenotypic difference as the requirement to live inside a host cell for survival. A well-known example of bacteria with marked phenotypic differences despite high genetic similarities is Shigella/Escherichia coli, which can display some morphological, biochemical, serological, and pathological differences, but are indistinguishable from a phylogenetic perspective [ 44, 45 ]. For symbiotic/free-living pairs, the arthropod-infecting \n \n Serratia symbiotica \n \n encompasses strains ranging from cultivable extracellular parasites to intracellular host-restricted mutualists [ 16 ]. However, while the range of their pairwise ANI values approaches 99 % [ 46 ], genomes of \n \n S. symbiotica \n \n ’s cultivable and host-restricted strains still differ substantially in size and structural organization [ 16 ], contrary to ‘ Ca . Nebulobacter’/ \n \n Fastidiosibacter \n \n . The genus \n \n Sodalis \n \n similarly displays a broad ecological diversification [ 18 ]. The host-restricted \n \n Sodalis pierantonius \n \n also shares enough of its genomic sequences with the opportunistic non-symbiotic species, Sodalis praecaptivus, to exceed the conventional threshold for species delimitation [ 18 ], consistently with its recent symbiotic engagement [ 17 ]. However, \n \n Sodalis pierantonius \n \n already displays diverging genomic features, such as an explosive proliferation of insertion sequences [ 17 ], while no such difference was recorded in the models investigated here. Another ciliate host, Heterometopus , also hosts an archaeal endosymbiont which is not dramatically different from its closest known free-living relative [ 47 ]. While several closely related free-living/endosymbiotic microorganisms are known, the ‘Ca. Nebulobacter’ /Fastidiosibacter pair stands out as laying at the extreme end of the spectrum of examples, and highlights the sometimes-overlooked possibility that, at least in certain cases, the changes required to become an obligate endosymbiont may be very small [ 48 ], and perhaps yet unpredictable, despite the apparently cataclysmic nature of the resulting ecological change."
} | 4,571 |
20526435 | PMC2880544 | pmc | 1,049 | {
"abstract": "A critical need for enhancing usability and capabilities of microfluidic technologies is the development of standardized, scalable, and versatile control systems 1 , 2 . Electronically controlled valves and pumps typically used for dynamic flow regulation, although useful, can limit convenience, scalability, and robustness 3 – 5 . This shortcoming has motivated development of device-embedded non-electrical flow-control systems. Existing approaches to regulate operation timing on-chip, however, still require external signals such as timed generation of fluid flow, bubbles, liquid plugs or droplets, or an alteration of chemical compositions or temperature 6 – 16 . Here, we describe a strategy to provide device-embedded flow switching and clocking functions. Physical gaps and cavities interconnected by holes are fabricated into a three-layer elastomer structure to form networks of fluidic gates that can spontaneously generate cascading and oscillatory flow output using only a constant flow of Newtonian fluids as the device input. The resulting microfluidic substrate architecture is simple, scalable, and should be applicable to various materials. This flow-powered fluidic gating scheme brings the autonomous signal processing ability of microelectronic circuits to microfluidics where there is the added diversity in current information of having distinct chemical or particulate species and richness in current operation of having chemical reactions and physical interactions."
} | 373 |
23146305 | PMC3503607 | pmc | 1,050 | {
"abstract": "Background The inherent recalcitrance of lignocellulosic biomass is one of the major economic hurdles for the production of fuels and chemicals from biomass. Additionally, lignin is recognized as having a negative impact on enzymatic hydrolysis of biomass, and as a result much interest has been placed on modifying the lignin pathway to improve bioconversion of lignocellulosic feedstocks. Results Down-regulation of the caffeic acid 3- O -methyltransferase (COMT) gene in the lignin pathway yielded switchgrass ( Panicum virgatum ) that was more susceptible to bioconversion after dilute acid pretreatment. Here we examined the response of these plant lines to milder pretreatment conditions with yeast-based simultaneous saccharification and fermentation and a consolidated bioprocessing approach using Clostridium thermocellum , Caldicellulosiruptor bescii and Caldicellulosiruptor obsidiansis . Unlike the S. cerevisiae SSF conversions, fermentations of pretreated transgenic switchgrass with C. thermocellum showed an apparent inhibition of fermentation not observed in the wild-type switchgrass. This inhibition can be eliminated by hot water extraction of the pretreated biomass, which resulted in superior conversion yield with transgenic versus wild-type switchgrass for C. thermocellum , exceeding the yeast-based SSF yield. Further fermentation evaluation of the transgenic switchgrass indicated differential inhibition for the Caldicellulosiruptor sp. strains, which could not be rectified by additional processing conditions. Gas chromatography–mass spectrometry (GC-MS) metabolite profiling was used to examine the fermentation broth to elucidate the relative abundance of lignin derived aromatic compounds. The types and abundance of fermentation-derived-lignin constituents varied between C. thermocellum and each of the Caldicellulosiruptor sp. strains. Conclusions The down-regulation of the COMT gene improves the bioconversion of switchgrass relative to the wild-type regardless of the pretreatment condition or fermentation microorganism. However, bacterial fermentations demonstrated strain-dependent sensitivity to the COMT transgenic biomass, likely due to additional soluble lignin pathway-derived constituents resulting from the COMT gene disruption. Removal of these inhibitory constituents permitted completion of fermentation by C. thermocellum , but not by the Caldicellulosiruptor sp. strains. The reason for this difference in performance is currently unknown.",
"conclusion": "Conclusions In general, the reduction in recalcitrance drastically improved the susceptibility to hydrolysis and bioconversion for yeast-based SSF, and after removal of water soluble inhibitors, high levels of fermentation products were also produced by C. thermocellum. The Caldicellulosiruptor sp. strains yielded only lower levels of fermentation products under these conditions with the transgenic feedstocks. The differential between bacterial fermentation inhibition may, in part, be explained by different aromatic constituents in the fermentation broth. Additionally, it may also be explained by the microorganisms having varying degrees of tolerance to these compounds. Overall, it may be concluded that biomass sources with reduced recalcitrance resulting from lignin pathway modification are a valuable resource for producing economical biofuels. However, during characterization of new biomass sources, in vitro assays such as sugar release assays should be supplemented with in vivo fermentation tests which we have shown can detect inhibitory compounds present in the biomass hydrolysate. The exact source and nature of these inhibitory compounds impacting the fermentation performance of our CBP candidate microorganisms warrants further investigation.",
"discussion": "Discussion The combination of a feedstock with increased enzymatic digestibility in combination with the CBP approach, which will eliminate the need for exogenous hydrolytic enzymes, has the potential to further reduce the cost of biofuels. Therefore we examined the fermentation performance of both wild-type and transgenic switchgrass lines using Clostridium thermocellum , Caldicellulosiruptor obsidiansis , and Caldicellulosiruptor bescii. Using three lines of switchgrass down-regulated in the COMT gene \n[ 12 ], we have shown that a milder pretreatment process does not impact the improved product yield generated by fermentation of the COMT down-regulated switchgrass biomass during yeast-based SSF. However, when a CBP-capable bacterium is tested, a significant differential of fermentation inhibition is detected, as judged by product yield on carbohydrate. In the case of C. thermocellum fermentations of dilute acid pretreated feedstock, the cellulosome and/or free carbohydrolases appear functional, as indicated by high levels of liberated unfermented glucose and cellobiose in the fermentation broth. At the same time, COMT transgenic feedstock lines clearly generate greater inhibition compared to the wild-type switchgrass, in the case of C. thermocellum fermentation. The inhibition of fermentation was shown to be removed after hot water extraction was applied to the dilute acid pretreated feedstock lines, suggesting that the inhibition is caused by water-soluble constituents. The picture is quite different for the Caldicellulosiruptor sp. strains tested. Fermentation of dilute acid pretreated and hot water extracted biomass that was readily fermented by C. thermocellum caused significant reduction in fermentation yield for T1-2-TG and T1-3-TG substrates with both Caldicellulosiruptor sp. strains. In addition, there were only low levels of unconsumed sugar remaining in the broth at the end of fermentation, indicating that both fermentation and hydrolysis were negatively impacted for the two highly down-regulated COMT feedstock lines. Moreover, the apparent differential of fermentation inhibition between the three CBP microorganisms, measured by unconsumed carbohydrates or low product yields, was readily detected when a less severe hot water pretreatment was used to prepare the feedstock lines. The apparent differential of inhibition between bacterial fermentations was particularly interesting because it was not seen in yeast-based SSF, and was an unexpected result. We hypothesize that the reduction in fermentation yield could be a biomass, microbe, or a biomass-microbe combined effect. A result that supports the hypothesis of a biomass effect contributing to the apparent inhibition is the significant reduction in yield of the Caldicellulosiruptor sp. strains’ fermentation of dilute acid pretreated, highly down-regulated COMT T1-2 and T1-3 lines, which is not present in the moderately down-regulated T1-12 transgenic line or the wild-type lines. Another possible reason for the apparent differential of inhibition is the various modes of interaction and hydrolysis employed by the hydrolytic system used by the microorganisms. As a result, they may release different or varying concentrations of inhibitory aromatic constituents, including mono-phenolic acids and sugar-aromatic conjugates. It is also not unreasonable to expect that the three microorganisms have different levels of tolerance for various inhibitory compounds. We analyzed the fermentation broth and appropriate controls with GC-MS based metabolite profiling in an attempt to determine if mono-phenolic acids or other aromatic constituents were causing the observed inhibition. We showed temperature, media components, and fungal enzymes alone did not generate aromatic constituents or mono-phenolics, which are components of plant cell walls and known to inhibit bacterial fermentation \n[ 26 , 27 ]. The aromatic constituents, including mono-phenolic acids found in the fermentation broth for hot water versus dilute acid pretreatment are different. The variation in lignin derived constituents may be explained by the difference in pretreatment severity affecting the lignin structure and content \n[ 28 ]. In the case of hot water pretreatment, there was a mild biomass effect. Of specific interest was the increased relative abundance of aromatic constituents in the Caldicellulosiruptor sp. strains in comparison to C. thermocellum . This indicates that C. thermocellum ’s hydrolytic system (cellulosome and free enzymes) might be producing a cleaner (less aromatic constituents) carbohydrate hydrolysate from the hot water pretreated switchgrass feedstocks than the Caldicellulosiruptor sp. strains. In contrast to the hot water pretreated feedstock results, dilute acid pretreated feedstocks did not show a notable difference in aromatic or mono-phenolic acid content between the different types of biomass or microorganisms. However, results showed that a tentatively identified compound, coumaroyl-benzaldehyde, was present in statistically different levels for both the biomass and microbe effect. The minimal biomass effect for either pretreatment was surprising, because our original hypothesis was based on the premise that the modification of the lignin pathway altered the lignin composition and content of the transgenic feedstock lines, and therefore, the concentration or composition of lignans generated and or released during pretreatment and bacterial hydrolysis and fermentation would appear quite different in comparison to the wild-type feedstock. The differential of bacterial fermentation inhibition may, in part, be explained by the aromatic constituents in the fermentation broth. Additionally, it may also be explained by the microorganisms having varying degrees of tolerance to these compounds. In general, the reduction in recalcitrance drastically improved the susceptibility to bioconversion for yeast-based SSF and, after the inhibition was removed; high levels of fermentation products were produced by C. thermocellum. As a result, biomass sources with reduced recalcitrance resulting from lignin pathway modification are a valuable resource for producing economical biofuels, but the impact of the lignin modification on the three bacteria’s fermentation performance needs to be further studied to determine the cause of reduction in fermentation yield."
} | 2,552 |
37307467 | PMC10288593 | pmc | 1,051 | {
"abstract": "Significance The brain is a recurrent network of neuronal cells, which are modeled by, but different from, simplified mathematical representations used in typical artificial neural networks. How does the characteristic connectivity in the recurrent network, such as the modular organization, interact with the biophysical properties of the cells to formulate the dynamics and function of biological neuronal networks? To address this question, we used the framework of reservoir computing to readout the stimulus responses of cultured neuronal networks and evaluated their computational properties in pattern classification and timer tasks. We show that reservoirs of biological neurons filter input signals, which can be classified by a linear decoder and that modularity in the recurrent connectivity facilitate the classification performance.",
"discussion": "Discussion In this study, we constructed mBNNs with modular topology on engineered substrates and demonstrated the effectiveness of the modular architecture and biological signal transduction in pattern classification tasks under the reservoir computing framework. Our experiments demonstrated that the accuracy of pattern classification was positively correlated with the functional modularity of BNNs in the lower range of the modularity Q . Conventionally, a reservoir computing system based on artificial neural networks takes advantage of a sparsely connected random network as the reservoir layer to exploit its high-dimensional dynamics ( 20 , 52 ). BNNs grown on uniform substrates similarly form random networks, although some metric dependence exists due to the finite length of the axons. However, the connectivity between neurons is usually dense, and spontaneous and evoked activity patterns of BNNs in uniform culture are typically governed by the so-called network bursts, i.e., repetitive firing of neurons that occur coherently across a large population of neurons in a network ( 53 – 55 ). Inducing modular topology in BNNs has been shown to break the excessive coherence in cultured BNNs and increase the variety of intrinsic activity patterns ( 47 , 48 , 56 ), which is probably the underlying mechanism behind the positive correlation between functional modularity and classification accuracy in mBNNs. We investigated the short-term memory of the mBNN reservoir using a timer task and showed that mBNNs retain a short-term memory of several 100 ms. This value is mostly in agreement with the report by Dranias et al., which showed that the short-term memory in uniformly cultured BNNs in a stimulus-specific memory task was ~1 s ( 57 ). Kubota et al. reported a memory capacity of 20 ms in uniformly cultured BNNs ( 58 ), but the difference of that value from that in Dranias et al. is most likely caused by the task design. Despite the differences in the reports, characteristic timescales of tens of milliseconds to several seconds, which are comparable to physical reservoirs based on soft materials ( 28 ), organic molecules ( 29 ), or water in a bucket ( 59 ), have been consistently observed in BNNs. This timescale matches that of many real-world signals, such as visual and auditory signals, making BNNs a suitable substrate for physical reservoirs in practical applications. The mBNN coupled with an external linear decoder was successfully demonstrated to be capable of classifying not only static spatial patterns but also spatiotemporal signals, such as human spoken digits. This contrasts with a previous pioneering study in which a nonlinear classifier (support vector machine) was used to classify spatiotemporal data such as jitter spike train and random music delivered to a uniform BNN ( 32 ). While it remains to be investigated whether modularity in BNNs increases performance in other tasks in reservoir computing, such as motor control ( 33 , 60 ), signal generation ( 21 ), and time-series prediction tasks ( 24 ), previous computational studies on artificial and spiking neural networks bearing modular structures ( 34 , 36 , 38 ) strongly imply that modularity in mBNN can improve the performance of these tasks as well. In certain tasks, a higher classification accuracy could be achieved when the output layer was trained to decode directly from the input photostimulation patterns bypassing the mBNN ( Fig. 5 B and E and SI Appendix , Fig. S4 B ). The performance of such a decoder, however, was strongly task-dependent. As shown in SI Appendix , Fig. S4 B , the performance suddenly dropped to a chance level when the decoder was tasked to classify signals with identical spatial components. The mBNN reservoir, in contrast, could be trained to distinguish temporally reversed “zero” from the original “zero” ( SI Appendix , Fig. S4 C ). Furthermore, since the output generated from the photostimulation patterns abruptly decreased to ~0 within 1 s ( SI Appendix , Fig. S7 ), it would be difficult to train the decoder to generate an output with a delay of more than 1 s. The computational benefit of filtering the signal through BNNs could potentially become evident in such “delayed-output” tasks as well. Most importantly, our experiments demonstrated that mBNNs act as a generalization filter by reducing the trajectory distance between common classes and increasing the distance between separate classes, enabling classification by readout layers pretrained on different genders or digits. The ability to generalize input signals and encode categorized information is one of the characteristics of information processing in the brain ( 45 ), while being a major challenge for machine learning systems ( 61 ). Such “categorical learning” was impossible when the input signal was directly decoded without an mBNN. In machine learning, data augmentation techniques such as noise addition, scaling, and rotation have been used to improve the generalization ability by reducing overfitting and expanding the decision boundary of the trained model ( 62 ). Although data augmentation has mostly been used for image classification in artificial neural networks, more recently, the technique has been reported to improve the generalization capability in time-series classification ( 63 , 64 ). We hypothesize that the inherent noise in biological neurons intrinsically performs data augmentation to improve the generalization capability of the BNN reservoir. Reservoir computing is an elegant framework that enables the establishment of a link between the high-dimensional dynamics of complex systems and their computational functions. By applying the concept of physical reservoirs, we constructively investigated how BNNs transform input signals and how their input responses can be utilized for computing. The mBNN reservoir could pave the way toward providing a platform to investigate structure-function relationships in BNNs, develop brain-like physical reservoirs, and understand the mechanisms of diseases with abnormalities in network modularity and population dynamics, such as autism ( 65 ) and epilepsy ( 66 )."
} | 1,756 |
36212815 | PMC9539880 | pmc | 1,052 | {
"abstract": "In seafloor sediments, the anaerobic oxidation of methane (AOM) consumes most of the methane formed in anoxic layers, preventing this greenhouse gas from reaching the water column and finally the atmosphere. AOM is performed by syntrophic consortia of specific anaerobic methane-oxidizing archaea (ANME) and sulfate-reducing bacteria (SRB). Cultures with diverse AOM partners exist at temperatures between 12°C and 60°C. Here, from hydrothermally heated sediments of the Guaymas Basin, we cultured deep-branching ANME-1c that grow in syntrophic consortia with Thermodesulfobacteria at 70°C. Like all ANME, ANME-1c oxidize methane using the methanogenesis pathway in reverse. As an uncommon feature, ANME-1c encode a nickel-iron hydrogenase. This hydrogenase has low expression during AOM and the partner Thermodesulfobacteria lack hydrogen-consuming hydrogenases. Therefore, it is unlikely that the partners exchange hydrogen during AOM. ANME-1c also does not consume hydrogen for methane formation, disputing a recent hypothesis on facultative methanogenesis. We hypothesize that the ANME-1c hydrogenase might have been present in the common ancestor of ANME-1 but lost its central metabolic function in ANME-1c archaea. For potential direct interspecies electron transfer (DIET), both partners encode and express genes coding for extracellular appendages and multiheme cytochromes. Thermodesulfobacteria encode and express an extracellular pentaheme cytochrome with high similarity to cytochromes of other syntrophic sulfate-reducing partner bacteria. ANME-1c might associate specifically to Thermodesulfobacteria, but their co-occurrence is so far only documented for heated sediments of the Gulf of California. However, in the deep seafloor, sulfate–methane interphases appear at temperatures up to 80°C, suggesting these as potential habitats for the partnership of ANME-1c and Thermodesulfobacteria .",
"conclusion": "Conclusion Here we cultured a thermophilic AOM consortium at 70°C consisting of methane-oxidizing archaea from the ANME-1c clade with Thermodesulfobacteria as sulfate-reducing partner bacteria. Our study bridges the temperature gap between AOM activity previously observed by pore water profiles and tracer experiments, and its in vitro demonstration in cultures. ANME-1c is a basal lineage to the ANME-1a/b clade. Interestingly, ANME-1c MAGs encode a hydrogenase operon that is not present in the ANME-1a/b clades. ANME-1c neither produces nor consumes hydrogen. The hydrogenase genes have low expression and this enzyme is likely a remnant of the ancestor of the Ca. Syntrophoarchaeia, an organism that was likely a multi-carbon alkane oxidizer. The function of this hydrogenase in ANME-1, but also other members of the Syntrophoarchaeia is unresolved. Based on indirect evidence ANME-1 were repeatedly suggested to be facultative methanogens ( Seifert et al., 2006 ; Treude et al., 2007 ; Orcutt et al., 2008 ; Lloyd et al., 2011 ; Kevorkian et al., 2021 ). Here, we demonstrated that even ANME-1c that encode a hydrogenase are not able to reverse their metabolism toward net methanogenesis. Cultivation-based approaches should be used to test whether the ANME-1a/b that encode hydrogenases are capable of methanogenesis. Most likely, ANME and their partners interact via DIET. The partner Thermodesulfobacterium encodes an extracellular pentaheme c-type cytochrome with high expression. This cytochrome is highly similar to Ca. Desulfofervidus auxilii cytochromes that have high expression during AOM at 60°C. This evidence suggests a central role of this pentaheme cytochrome of Thermodesulfobacteria in DIET. In addition, multiheme cytochromes and flagella from ANME-1c are highly expressed under AOM conditions. However, the role of these cytochromes with unknown subcellular location in DIET needs further investigation. Our study provides culture-based evidence for the feasibility of AOM at high-temperature conditions and the versatility and evolutionary diversity of the organisms mediating AOM in marine environments. To our knowledge, this is the first report of syntrophic consortia of ANME and Thermodesulfobacteria. ANME-1c and Ca. T. torris co-occur in environmental samples from the Guaymas Basin and Pescadero Basin (Gulf of California; Laso-Pérez et al., 2022 ; Speth et al., 2022 ). This limited geographical distribution is likely a result of undersampling of heated methane-rich environments. Apart from the rather rare methane-rich hydrothermal vents, consortia of these phylotypes might inhabit deep sulfate–methane interfaces. Indeed, these sulfate–methane interfaces occur at depths up to 150 m below the seafloor and at temperatures of up to 80°C ( Teske et al., 2021 ; Beulig et al., 2022 ). The question of whether these sulfate–methane interfaces are a habitat for AOM needs to be addressed. Future studies should aim to search for the ANME-1c and Thermodesulfobacteria co-occurring in such deep-sea sediments.",
"introduction": "Introduction In anoxic deep-sea sediments, the greenhouse gas methane is produced abiotically by thermocatalytic decay of buried organic matter or biotically by methanogens ( Whiticar, 1999 ). Anaerobic oxidation of methane (AOM) mitigates the flux of methane to the water column and eventually to the atmosphere by consuming 90% of the methane produced in the deep sediments ( Hinrichs and Boetius, 2002 ; Reeburgh, 2007 ; Regnier et al., 2011 ). In marine sediments, AOM primarily couples to sulfate reduction in a 1:1 stoichiometry: \n (1) \n C H 4 + SO 4 2 − → H S − + H C O 3 − + H 2 O . \n AOM is mediated by anaerobic methanotrophic archaea (ANME) that oxidize methane to CO 2 by reversing the methanogenesis pathway ( Hallam et al., 2004 ; Meyerdierks et al., 2010 ; Wang et al., 2013 ). ANME do not encode respiratory pathways, but they pass the reducing equivalents liberated during AOM to sulfate-reducing partner bacteria (SRB), forming characteristic consortia ( Boetius et al., 2000 ; Michaelis et al., 2002 ; Orphan et al., 2002 ; McGlynn et al., 2015 ; Wegener et al., 2015 ). The nature of this syntrophic association and the mechanisms involved in the transfer of reducing equivalents from ANME toward SRB are not completely resolved at the molecular level. Originally, it was proposed that the archaeal partners produce molecular hydrogen that is consumed by the bacterial partners ( Hoehler et al., 1994 ). However, most ANME do not code for hydrogenases ( Chadwick et al., 2022 ). Previous studies support the hypothesis of direct interspecies electron transfer (DIET) involving multiheme cytochromes and pilus proteins ( Meyerdierks et al., 2010 ; McGlynn et al., 2015 ; Wegener et al., 2015 ). The partner SRBs use the AOM-derived electrons for anaerobic respiration with sulfate as final electron acceptor ( Boetius et al., 2000 ; Wegener et al., 2015 ; Laso-Pérez et al., 2016 ). The limited energy yield of sulfate-dependent AOM [equation (1) , ΔG°′ = −16.67 kJ mol −1 at standard conditions and ΔG = −20 to −40 kJ mol −1 in marine AOM habitats] needs to be shared between ANME and their syntrophic partner SRB ( Thauer, 2011 ). ANME inhabit a variety of marine habitats including cold seeps ( Boetius et al., 2000 ; Orphan et al., 2001 ), mud volcanoes ( Niemann et al., 2006 ), gas hydrates ( Lanoil et al., 2001 ; Orcutt et al., 2004 ), hydrothermal vents ( Inagaki et al., 2006 ; Biddle et al., 2012 ) and deep subsurface sediments ( Roussel et al., 2008 ). ANME are polyphyletic and fall into three distinct phylogenetic groups (ANME-1, ANME-2, and ANME-3). ANME-3 often dominate AOM at mud volcanoes, where they form consortia with Desulfobulbus -related bacteria ( Niemann et al., 2006 ; Lösekann et al., 2007 ). Cultivation attempts of ANME-3 have not been successful so far. ANME-2 are globally distributed in a variety of benthic habitats and are typically found associated with Desulfosarcina / Desulfococcus bacteria (DSS, Seep-SRB1 and Seep-SRB2 clades; Knittel et al., 2003 , 2005 ; Boetius and Knittel, 2010 ). ANME-2 are dominant at cold seeps with high methane fluxes and temperatures below 20°C ( Knittel et al., 2005 ). Cultivation attempts at temperatures ≤ 20°C resulted in the enrichment of ANME-2c ( Holler et al., 2009 ; Wegener et al., 2016 ). ANME-1 prevail in most deep sulfate–methane transition zones (SMTZs; Ishii et al., 2004 ; Niemann et al., 2005 ; Treude et al., 2005 ), in hydrothermally heated sediments in the Guaymas Basin ( Teske et al., 2002 ; Schouten et al., 2003 ; Ruff et al., 2015 ; Dombrowski et al., 2018 ), and in the Auka vent field, in the Pescadero Basin (Gulf of California; Speth et al., 2022 ). Meso- and thermo-philic AOM cultures have been obtained from Guaymas Basin sediments at 37°C, 50°C, and 60°C ( Holler et al., 2011 ; Wegener et al., 2016 ). These cultures consisted of ANME-1a and HotSeep-1 ( Ca. Desulfofervidus) as partner bacteria. Ca. Desulfofervidus sequences are also found in situ at these sites ( McKay, 2014 ; Dowell et al., 2016 ). Previous short-term incubations revealed AOM activity at temperatures up to 75°C or 85°C, but the microorganisms performing AOM under these conditions were not assessed ( Kallmeyer and Boetius, 2004 ; Holler et al., 2011 ; Adams et al., 2013 ). Here, we obtained an active AOM culture at 70°C (AOM70) from Guaymas Basin hydrothermal sediments consisting of a previously uncultured ANME-1 subgroup ( Teske et al., 2002 ) and an apparently obligate syntrophic Thermodesulfobacterium partner. We describe their function and interaction based on physiological experiments and molecular data.",
"discussion": "Results and discussion AOM enrichment cultures at 70°C A slurry produced from hydrothermally-heated sediments from the Guaymas Basin and sulfate-reducer medium was supplemented with a methane:CO 2 headspace and incubated at 70°C. These incubations showed methane-dependent sulfate reduction, as measured by an increase of sulfide in the medium ( Figure 1B ; Supplementary Figure 1 ). These incubations produced sulfide 12 to 15 mM sulfide within about 100 days. The slurries were diluted 1/10 (v/v) in fresh SRB medium, and a fresh methane:CO 2 headspace was added and the incubation was proceeded. After four additional incubation and dilution steps, the produced AOM70 cultures were virtually sediment-free, contained microbial aggregates visible with the naked eye and produced approximately. 100 μmol sulfide L –1 d –1 . To our knowledge, this is the first long-term cultivation of AOM-performing microorganisms above 60°C. The culture showed strongly decreased sulfide production at 60°C. It tolerated a transfer to 75°C, but became inactive at 80°C, confirming prior results on the upper-temperature limit of AOM made in short-term incubations with Guaymas Basin sediments ( Holler et al., 2011 ; McKay et al., 2016 ). Figure 1 Microbial composition and growth of thermophilic AOM cultures. (A) CARD-FISH on AOM aggregates with the probes EUB388 I-III (green) and ANME-1-389 (red). Bacteria (green) and archaea (red) form shell-type consortia. (B) Methane-dependent sulfide production in AOM cultures. The red arrows indicate when culture medium was replaced. (C) 16S rRNA gene relative abundance in long-read metagenomic reads and shotgun metatranscriptomic reads (triplicate metatranscriptomes). The enrichment is dominated by ANME-1 and sulfate reducers of the class Thermodesulfobacteria. Other bacteria and archaea such as Acetothermia and Bathyarchaeia are side community members from original sediment samples. Thermophilic AOM community at 70°C To resolve the microbial community composition of the AOM70 culture, we obtained a long-read metagenome and triplicate short-read metatranscriptomes. The community consisted mainly of ANME-1 (~50% relative abundance of mapped reads) and Thermodesulfobacteria (~20% relative abundance) based on 16S rRNA gene fragments recruited from the metagenome ( Figure 1C ) that were rare in the original sediment samples ( Supplementary Figure 2 ). Metatranscriptomic samples were also dominated by ANME-1 (~70% relative abundance) and Thermodesulfobacteria (~15% relative abundance), forming the active AOM community at 70°C. Both the metagenome and metatranscriptome revealed noticeable populations of Bathyarchaeota and Acetothermia (<10% relative abundance of mapped reads). Yet, we were not able to reconstruct MAGs of these organisms; hence, their potential functions are unknown. Previous studies suggest that these organisms ferment or oxidize biomolecules produced by the AOM community ( Kellermann et al., 2012 ; Dombrowski et al., 2017 ; Hao et al., 2018 ; Zhu et al., 2022 ). After long-read metagenome assembly and binning, we obtained two high-quality MAGs of the two members of the AOM consortium ( Table 1 ). The Thermodesulfobacterium MAG has a size of 1.7 Mbp and GC content of 29%. The bin is almost complete (98.6%) and has no contamination based on the presence of 104 bacterial single-copy marker genes (CheckM; Parks et al., 2015 ). The ANME-1 bin has a size of 1.5 Mbp and a GC content of 47.8%. The bin is 90.8% complete and has contamination of 7.9% based on 149 archaeal marker genes (CheckM; Parks et al., 2015 ). Table 1 Metagenome-assembled genomes retrieved from AOM70 enrichment cultures. Ca. Thermodesulfobacterium ANME-1c ( Ca. Methanophagales) No. of contigs 4 16 Genome size 1.702 Mbp 1.493 Mbp L50/N50 2/808,565 bp 5/109,767 bp GC content 29.0% 47.8% Completeness * 98.6% 90.8% Contamination * <1% 7.89% * Completeness and contamination were calculated with CheckM. We attempted to visualize the enriched ANME-1 using a previously established probe targeting the whole ANME-1 clade (ANME-1-350, Supplementary Table 2 ; Boetius et al., 2000 ). In situ hybridization with the ANME-1-350 failed, because the probe has two mismatches with the 16S rRNA sequence of the enriched ANME-1. The 16S rRNA gene of this ANME-1 belong to a clade ancestral to all ANME-1a/b, namely ANME-1c ( Supplementary Figure 4 and Discussion below; Laso-Pérez et al., 2022 ). A newly developed ANME-1-389 probe specifically binds to ANME-1c cells. Because the probes available for partner SRB do not target the 16S rRNA sequence of Thermodesulfobacteria , we designed three candidate probes to target this clade ( Supplementary Table 2 ). Unfortunately, none of these probes hybridized the 16S rRNA of this organism after various CARD-FISH attempts ( Supplementary Table 2 ). Thermodesulfobacteria are likely the partner bacteria of ANME-1c during AOM at 70°C based on the abundance of bacterial cells and their gene content (see Discussion below). Furthermore, all genes coding for dissimilatory sulfate reductase ( dsr ) in the metagenome belong to the Thermodesulfobacterium MAG. Double hybridization with the ANME-1c and the general bacterial probes ( Supplementary Table 2 ) revealed a dominance of “shell-type” aggregates consisting of ANME-1c and partner bacteria ( Figure 1A ; Supplementary Figure 3 ). These consortia consist of clumps of ANME-1c cells, surrounded by smaller rod-shaped bacterial cells. These shell-type aggregates differ from the predominantly mixed-type aggregates of moderately thermophilic consortia growing at 50°C–60°C ( Holler et al., 2011 ; Wegener et al., 2015 ). A shell-type growth morphology is often observed in cold-adapted ANME ( Knittel et al., 2005 ). The reason for the different association types is unknown. Phylogeny of deep-branching ANME-1c On the basis of whole genome comparison, the ANME-1 population detected in the AOM70 culture falls into the recently named ANME-1c clade ( Figure 2A ; Laso-Pérez et al., 2022 ; Speth et al., 2022 ).The ANME-1c group is basal to its sister groups ANME-1a and ANME-1b within the order ANME-1 ( Ca. Methanophagales). The 16S rRNA gene phylogenetic tree supports this phylogenetic placement ( Supplementary Figure 4 ). ANME-1c belong to the class Syntrophoarchaeia with the ANME-1, Ca. Syntrophoarchaeales and Ca. Alkanophagales. Considering an average nucleotide identity (ANI) of <83% for distinct species and >95% for the same species ( Jain et al., 2018 ) the ANME-1c clade consists of two distinct species clusters ( Supplementary Figure 5 ). The ANME-1c MAG from the AOM70 culture belongs to the cluster of Ca. Methanoxibalbensis ujae from Pescadero Basin ( Laso-Pérez et al., 2022 ). ANME-1c 16S rRNA gene sequences have been detected in hydrothermal sediments of the Guaymas Basin and the Juan de Fuca Ridge ( Supplementary Figure 4 ; Teske et al., 2002 ; Merkel et al., 2013 ; McKay et al., 2016 ), and a MAG of the ANME-1c clade (accessions: SAMN09215218, GCA_003661195.1) was derived from Guaymas Basin hydrothermal sediments ( Dombrowski et al., 2018 ). ANME-1c are also present in rock samples from hydrothermal fields in Pescadero Basin (Gulf of California; Speth et al., 2022 ). The ANME-1c clade was originally named “ANME-1b” by Teske and coworkers to differentiate this lineage from previously described cold-seep ANME-1 ( Teske et al., 2002 ) and later renamed to ANME-1Guaymas because it was predominantly recovered from Guaymas Basin ( Biddle et al., 2012 ; Merkel et al., 2013 ; Dowell et al., 2016 ). These sequences originate from sediment cores with sulfate-reducing activity at temperatures between 65°C and 90°C, showing that these archaea are likely all thermophiles ( Biddle et al., 2012 ). Furthermore, the high GC content (>60%) of ANME-1c 16S rRNA genes indicates that these archaea might have temperature optima in the upper range of thermophily above 70°C ( Merkel et al., 2013 ). Figure 2 ANME-1 phylogenomic tree and hydrogenase phylogeny. (A) Phylogeny of ANME-1 order ( Ca. Methanophagales) with related Ca. Alkanophagales, Ca. Syntrophoarchaeales, and Ca. Santabarbaracales. Maximum likelihood phylogenomic tree based on an alignment of 38 archaeal conserved genes from 55 genomes ( Supplementary Table 3 ). Geoglobus sequences were the outgroup to set the tree root (not shown). (B) Hydrogenase phylogeny. ANME-1, Ca. Alkanophagales and Ca. Syntrophoarchaeales hydrogenases are located at the base of groups 1 g and 1 h of NiFe hydrogenases. A complete hydrogenase tree is shown in Supplementary Figure 11 . MAGs from cultured ANME are depicted in bold. Shading in both trees indicates the three subdivisions of the ANME-1: ANME-1c, ANME-1a and ANME-1b. ANME-1 AOM70 is the genome discussed in the main text. Scales indicate nucleotide substitution per site. Bootstrap support is based on 100 iterations above 70% and above 90%. Genomic and metabolic features of ANME-1c ANME-1c codes for a complete methanogenesis pathway including a canonical methane-active Mcr ( Figure 3 ). The mcr ABC genes in ANME-1c have the highest expression (CLR > 7) among all genes in the dataset. This high expression of mcr confirms previous transcriptomic work in ANME ( Haroon et al., 2013 ; Krukenberg et al., 2018 ). The activation of methane is the rate-limiting step of AOM, and ANME would promote this reaction by producing large amounts of Mcr ( Scheller et al., 2010 ; Thauer, 2011 ). Similar to other ANME-1 archaea, ANME-1c does not encode a N 5 , N 10 -methylene-H 4 MPT reductase ( mer ). This gene might be substituted by a 5,10-methylenetetrahydrofolate reductase ( met; \n Stokke et al., 2012 ; Krukenberg et al., 2018 ). The function of this bypass has not been verified yet. All other genes of the methanogenesis pathway show a relatively high expression with CLR values between 0.1 and 3.4 ( Supplementary Table 4 ), supporting a catabolic function of the encoded genes. ANME-1c encodes and expresses the methanogenesis-related membrane-bound complex H + -translocating F 420 :quinone oxidoreductase ( fqo ) that catalyzes the transfer of electrons from reduced cofactors to the quinone pool ( Pereira et al., 2011 ). ANME-1c encodes an ATP synthase, which is a common feature in ANME to enable the oxidative phosphorylation of ATP, coupled to the influx of protons. ANME-1c encodes a sulfate adenylyltransferase ( cys N; low expression, CLR = −0.02) and an adenylylsulfate kinase ( cys C; high expression, CLR = 2.21) that could be used for assimilatory sulfate metabolism, but it lacks the key genes for dissimilatory sulfate reduction. The ANME-1c MAG lacks a complete nitrogenase operon, suggesting it is incapable of nitrogen fixation. The capability for nitrogen fixation has been shown only in ANME-2 archaea but not in ANME-1 ( Dekas et al., 2009 , 2015 ; Orphan et al., 2009 ; Krukenberg et al., 2018 ). The nitrogenase subunits nif DH detected in ANME-1c and other ANME-1 genomes ( Meyerdierks et al., 2010 ) are likely paralogs of cfb CD because they are located in an operon with genes encoding the biosynthetic pathway of coenzyme F 430 ( Zheng et al., 2016 ; Moore et al., 2017 ). Coenzyme F 430 functions as a prosthetic group that binds to the active site of McrA, and is therefore a key molecule for methanogens and methanotrophs ( Friedmann et al., 1990 ; Ermler et al., 1997 ; Shima et al., 2012 ). Figure 3 Key metabolic pathways in ANME-1c and Ca. Thermodesulfobacterium torris and metatranscriptomic expression during AOM. Gene expression values were normalized to centered-log ratios (CLR). A CLR value of 0 represents the mean expression of all genes in a genome. The asterisk next to the ANME-1c cytochrome indicates unknown cell localization. H 4 MPT, tetrahydromethanopterin; MF, methanofuran; Fd, ferredoxin; Mcr, methyl-coenzyme M reductase; Mtr, tetrahydromethanopterin S-methyltransferase; Met, 5,10-methylenetetrahydrofolate reductase; Mtd, methylenetetrahydromethanopterin dehydrogenase; Mch, methenyltetrahydromethanopterin cyclohydrolase; Ftr, formylmethanofuran-tetrahydromethanopterin formyltransferase; Fmd, formylmethanofuran dehydrogenase; Cdh/Acs, CO dehydrogenase/acetyl-coenzyme A synthase complex; rTCA, reductive tricarboxylic acid cycle; Sat, sulfate adenylyltransferase; Apr, adenylylsulfate reductase; Dsr, dissimilatory sulfate reductase; Fqo, ferredoxin: quinone oxidoreductase; Atp, ATP synthase; Hyd, hydrogenase; Qmo, quinone-modifying oxidoreductase; Nuo, NADH:ubiquinone oxidoreductase; Cyt c , multiheme cytochrome c -like protein; PilA, bacterial pilus protein; FlaB, archaeal flagellum protein (archaellum). ANME-1c likely performs autotrophic carbon fixation via the carbon monoxide dehydrogenase/acetyl-CoA synthase complex (Cdh/Acs; Kellermann et al., 2012 ). All the cdh transcripts are highly abundant (CLR between 0.4 and 2.0, Supplementary Table 4 ), supporting the use of this pathway for autotrophy. ANME-1c does not encode other complete carbon fixation pathways. The reductive tricarboxylic acid (rTCA) cycle is incomplete, lacking the key enzyme pyruvate carboxylase. The rTCA cycle genes have relatively low expression (CLR −0.7 to 1.8, Supplementary Table 4 ). Enzymes of this pathway may play a role in the biosynthesis of cell building blocks ( Meyerdierks et al., 2010 ). Like all ANME-1, ANME-1c contains a β-oxidation pathway. The phylogenetically related multi-carbon alkane oxidizers, Ca. Syntrophoarchaeales, Ca. Alkanophagales and Ca. Santabarbaracales harbor several copies of the β-oxidation genes and use the encoded pathway to split alkane-derived acyl-CoA into acetyl-CoA units ( Laso-Pérez et al., 2016 ; Wang et al., 2021 ). However, the expression of β-oxidation genes in ANME-1c is relatively low, especially the first two reactions (CLR −0.2 to 0.5, Supplementary Table 4 ). Furthermore, ANME-1c lacks the electron transfer flavoprotein ( etf AB) needed to oxidize acyl-CoA to enoyl-CoA. Hence, β-oxidation may not serve a catabolic function in ANME-1c, but play a role in biosynthesis of cell compounds. Wang and colleagues suggested that the ancestor of Syntrophoarchaeia (family including ANME-1, Ca. Syntrophoarchaeales and Ca. Alkanophagales) activated multi-carbon alkanes with their multi-carbon-alkane specific Mcr (Acrs) forming the corresponding alkyl-CoM as intermediate ( Laso-Pérez et al., 2016 ; Wang et al., 2021 ). It was proposed that ANME-1 acquired a methane-activating Mcr from methylotrophic methanogens, likely from the clade Ca. Methanofastidiosa/ Ca. Nuwarchaeia, and later lost the acr genes ( Borrel et al., 2019 ; Wang et al., 2021 ). Phylogeny and environmental distribution of AOM-associated Thermodesulfobacteria We compared the Thermodesulfobacterium MAG in AOM70 cultures with the Thermodesulfobacteria MAGs from Pescadero Basin and to MAGs retrieved from databases (NCBI and JGI). Our AOM70 Thermodesulfobacterium shares >95% ANI with a MAG of a Thermodesulfobacterium from Pescadero Basin ( Laso-Pérez et al., 2022 ; Speth et al., 2022 ; Supplementary Figure 7 ). Based on 16S rRNA phylogeny ( Supplementary Figure 6 ), the Thermodesulfobacterium AOM70 sequences form a cluster with sequences originating from Guaymas Basin and Pescadero Basin hydrothermal seeps ( McKay et al., 2016 ; Lagostina et al., 2021 ; Pérez Castro et al., 2021 ; Speth et al., 2022 ). Several species of Thermodesulfobacterium have been isolated from hot springs ( Zeikus et al., 1983 ; Sonne-Hansen and Ahring, 1999 ; Hamilton-Brehm et al., 2013 ), petroleum reservoirs ( Rozanova and Khudiakova, 1974 ) and hydrothermal vents ( Jeanthon et al., 2002 ; Moussard et al., 2004 ). The 16S rRNA gene sequence of our AOM70 Thermodesulfobacterium is 96% identical to the closest cultured representative, Thermodesulfobacterium geofontis , isolated from Obsidian Pool, Yellowstone National Park ( Hamilton-Brehm et al., 2013 ). Considering an ANI <83% for distinct species and >95% for the same species ( Jain et al., 2018 ) the Thermodesulfobacteria MAG from the AOM70 culture metagenome and the Pescadero MAG are a new candidate species in the genus Thermodesulfobacterium ( Supplementary Figure 8 ). We propose the taxon name Candidatus Thermodesulfobacterium torris (torris “firebrand” referring to the thermophilic lifestyle and the formation of black aggregates in the cultures). Metabolism of the partner bacteria Thermodesulfobacteria Members of the Thermodesulfobacteria family have not been previously reported as partner bacteria in AOM. All Thermodesulfobacteria isolates are sulfate-reducing (hyper) thermophiles with growth optima between 65°C and 90°C. They differ in the range of electron donors or carbon sources they use, which include molecular hydrogen, formate, lactate, and pyruvate ( Zeikus et al., 1983 ; Sonne-Hansen and Ahring, 1999 ; Jeanthon et al., 2002 ; Moussard et al., 2004 ). Similar to other members, Ca. T. torris encodes a complete dissimilatory sulfate reduction pathway, including sulfate adenylyltransferase ( sat ), adenylylsulfate reductase ( apr ), and dissimilatory sulfite reductase ( dsr ). In Ca. T. torris this pathway is highly expressed during AOM (average CLR values between 4.0 and 6.5, Supplementary Table 4 ). In addition, Ca. T. torris contains and expresses the Dsr-associated membrane complex ( dsr KMOP) which takes up electrons from the periplasmic cytochrome c pool to reduce a disulfide bond in the cytoplasmic DsrC ( Pereira et al., 2011 ; Venceslau et al., 2014 ). The quinone-modifying oxidoreductase ( qmo ABC) genes are present in an operon together with the Apr genes. In fact, the Qmo membrane complex interacts with Apr through a third unknown protein and channels electrons from the membrane ubiquinones via electron confurcation ( Ramos et al., 2012 ). Both the dsr KMOP and the qmo ABC transcripts have high expression ( Supplementary Table 4 ). Other cytoplasmic enzymes commonly associated with heterodisulfide reductases, such as the methylviologen reducing hydrogenase (Mvh/Hdr), were not found in the dataset. For energy conservation, Ca. T. torris uses a membrane-bound NADH:ubiquinone oxidoreductase (Nuo) and an ATP synthase (Atp). Nuo couples the reduction of NAD + by reduced ubiquinones in the cytoplasmic membrane to the translocation of protons to the periplasmic space. The proton gradient generated enables oxidative phosphorylation in the ATP synthase. The reductive acetyl-CoA pathway (Wood-Ljungdahl pathway) for carbon fixation is incomplete in the genome. Ca. T. torris does not encode a formate dehydrogenase ( fdh ) or a carbon monoxide dehydrogenase/acetyl-CoA complex ( cdh / acs ), but it encodes the enzymes catalyzing C 1 -tetrahydrofolate transformations. These reactions are necessary for several cell processes including nucleic acid biosynthesis ( Ducker and Rabinowitz, 2017 ). Instead, Ca. T. torris likely fixes carbon via the rTCA cycle. The genome codes for an almost complete rTCA cycle, lacking a succinyl-CoA synthetase. This enzyme is likely substituted by a putative acetyl-CoA synthetase encoded in the genome and highly expressed (CLR = 3.39, locus MW689_000791). Acetyl-CoA synthetases have sequence homology with succinyl-CoA synthetases and are also active toward succinate with reduced affinity ( Sánchez et al., 2000 ). Similarly, the thermophilic partner bacterium Ca. Desulfofervidus auxilii and other non-symbiotic thermophilic SRB fix carbon via the rTCA cycle ( Schauder et al., 1987 ; Krukenberg et al., 2016 ). By contrast, meso- and psychrophilic AOM partner bacteria fix carbon using the Wood-Ljungdahl pathway ( Skennerton et al., 2017 ). We aimed to enrich Ca. T. torris by incubating aliquots of the AOM70 culture with H 2 , formate, lactate, or pyruvate as electron donors. None of the substrates resulted in immediate sulfide production ( Supplementary Figure 9 ). Pyruvate caused sulfide production after 15 days, which likely indicates the growth of originally rare microorganisms, similar as shown for mesophilic AOM cultures ( Zhu et al., 2022 ). These incubations suggest that Ca. T. torris is an obligate syntrophic bacterium that fully depends on the transfer of reducing equivalents in AOM. Transfer of reducing equivalents between ANME-1c and Thermodesulfobacteria Because ANME have no own respiratory pathways, they need to transfer the reducing equivalents liberated during AOM to their sulfate-reducing partners. Multiple mechanisms have been proposed for syntrophic fermentation, including interspecies hydrogen transfer ( Schink, 1997 ). A canonical syntrophy based on interspecies hydrogen transfer would require membrane-bound hydrogenases in both partners. Notably, the ANME-1c MAGs code for a complete nickel-iron hydrogenase, a feature that is rare in other ANME-1 genomes. The hydrogenase database (HydDB) annotation classifies this hydrogenase within the group 1 g of hydrogenases that are typically found in thermophilic organisms ( Brock et al., 1972 ; Fischer et al., 1983 ; Huber et al., 2000 ; Laska et al., 2003 ). Yet this hydrogenase is only poorly expressed (CLR < −0.3, Supplementary Table 4 ). In contrast, the Ca. T. torris MAG lacks hydrogenases. The addition of molecular hydrogen to the culture did not stimulate sulfide production in the AOM culture, which confirms that Ca. T. torris cannot grow on hydrogen. Based on these observations we exclude hydrogen as electron carrier from ANME-1c toward Ca. T. torris. Our results confirm thermodynamic models which excluded hydrogen exchange in AOM consortia ( Sørensen et al., 2001 ). Most other AOM partner bacteria such as SeepSRB-1a and SeepSRB2 are also obligate syntrophs and do not encode hydrogenases ( Nauhaus et al., 2002 ; Wegener et al., 2016 ; Krukenberg et al., 2018 ). Ca. D. auxilii , performs DIET when growing as partner in AOM or short-chain alkane oxidation at 50°C–60°C ( Wegener et al., 2015 ; Laso-Pérez et al., 2016 ; Krukenberg et al., 2018 ; Hahn et al., 2020 ), but it also shows growth on hydrogen ( Krukenberg et al., 2016 ). It has been shown that DIET allows more efficient growth than interspecies hydrogen transfer ( Summers et al., 2010 ). In AOM and short-chain alkane-oxidizing consortia, cells are densely packed and the intercellular space contains cytochromes and nanowire-like structures ( McGlynn et al., 2015 ; Wegener et al., 2015 ; Laso-Pérez et al., 2016 ; Krukenberg et al., 2018 ). In these mesophilic and thermophilic consortia, both partners express cytochrome and pil A genes ( Laso-Pérez et al., 2016 ; Krukenberg et al., 2018 ). The genes pil A (bacterial pilin) in Ca. T. torris and fla B (archaeal flagellin) in ANME-1c show a high expression in the metatranscriptomes (CLR values of 1.8 and 3.2, respectively, Supplementary Table 4 ). The archaeal flagellum (archaellum) is highly similar to bacterial type IV pili ( Albers and Jarrell, 2015 ) and might also be involved in electron transfer over longer distances. The conductivity of the archaellum from the methanogen Methanospirillum hungatei was demonstrated, yet its possible role in interspecies electron transfer is unclear ( Walker et al., 2019 ). Conductive filaments that enable the transport of electrons across long distances toward extracellular electron acceptors have been widely studied in Geobacter ( Reguera et al., 2005 ; Malvankar et al., 2011 ; Shrestha et al., 2013 ; Adhikari et al., 2016 ). Reguera et al. (2005) showed that pilus-deficient Geobacter mutants could not transfer electrons to extracellular electron acceptors, suggesting an involvement of pilin proteins in this process ( Reguera et al., 2005 ). The molecular basis of DIET in sulfate-dependent AOM has been intensively discussed in the past years ( McGlynn et al., 2015 ; Wegener et al., 2015 ; Chadwick et al., 2022 ; Yu et al., 2022 ). McGlynn et al. (2015) proposed a model based on direct interspecies electron transfer via multiheme cytochromes for AOM consortia, using evidence from single-cell activities, microscopic observations and genomics ( McGlynn et al., 2015 ). Krukenberg et al. (2018) showed that both partners highly express cytochromes with a low number of heme groups (3–5 heme binding motifs) during thermophilic AOM ( Krukenberg et al., 2018 ). In AOM consortia at 60°C, it was observed that SRB Ca. Desulfofervidus auxilii expressed pili genes and that the intercellular space was filled with nanowire structures similar to syntrophic consortia of Geobacter ( Wegener et al., 2015 ). Indeed it was recently shown that the filaments in Geobacter sulfurreducens are not formed by pilin proteins (PilA), but rather by stacked OmcS hexaheme cytochromes ( Wang et al., 2019 ). Instead, PilA might be involved in secretion of OmcS cytochromes ( Gu et al., 2021 ). Both ANME-1c and Ca. T. torris encode several multiheme cytochromes. ANME-1c codes for several proteins with 2 to 8 heme-binding motifs ( Figure 4 ; Supplementary Table 4 ). A cytochrome c7 and a protein without annotation, both with 3 heme groups, are among the top expressed genes (CLR values of 5.9 and 2.96, respectively, Supplementary Table 4 ). However, the predicted subcellular localizations of these putative cytochromes are unknown, as reported by PSORTb. Interestingly, these cytochromes are highly similar to extracellular cytochromes that were highly expressed in ANME-1 during AOM at 60°C ( Supplementary Table 4 ; Krukenberg et al., 2018 ). Ca. T. torris also contains numerous multiheme cytochromes, with up to 26 heme-binding motifs. A cytochrome-like gene with five heme-binding motifs and predicted extracellular localization shows high expression levels similar to the dsr A (CLR value of 4.0, Figure 4 ). This putative pentaheme cytochrome shares high sequence identity (<40% identity) with a Ca. Desulfofervidus auxilii OmcS-like protein ( Wegener et al., 2015 ; Krukenberg et al., 2018 ) (locus tag HS1_000170, Supplementary Figure 10 ). We hypothesize that the extracellular pentaheme cytochromes from Ca. T. torris are likely involved in receiving electrons derived from methane oxidation. ANME-1c cytochromes with undetermined cell localization might also be involved in interspecies electron transfer. Figure 4 Expression and subcellular localization of multiheme cytochromes and cellular appendages in ANME-1c (top) and Thermodesulfobacterium (bottom) in AOM70 cultures. Gene expression is noted as centered-log ratio values, with 0 as the mean expression of all genes in each genome. mcr A and dsr A mean CLR values are displayed as a reference (red and green lines, respectively). Symbols show the predicted subcellular localization of cytochromes (PSORTb). Three symbols in a vertical line correspond to CLR values of a specific locus in triplicate metatranscriptomes. Alternative roles of the membrane-bound hydrogenase in ANME-1c Some ANME-1c MAGs code for a NiFe membrane-bound hydrogenase, which is an uncommon feature in most ANME genomes ( Stokke et al., 2012 ; Wegener et al., 2015 ; Krukenberg et al., 2018 ). This hydrogenase forms a clade with those of Ca. Syntrophoarchaeum and Ca. Alkanophagales ( Figure 2B ; Supplementary Figure 11 ; Laso-Pérez et al., 2016 ; Wang et al., 2021 ). In Ca. Syntrophoarchaeum, the hydrogenase is highly expressed during anaerobic propane and butane oxidation, albeit its function is also unknown ( Laso-Pérez et al., 2016 ). In contrast, in ANME-1c the NiFe-hydrogenase has low expression ( Figure 3 ; Supplementary Table 4 ). A recent study described the ANME-1c as an ancestral clade at the base of ANME-1a/1b and as a sister branch of Ca. Syntrophoarchaeales ( Laso-Pérez et al., 2022 ). ANME-1c were proposed to be facultative methanogens based on the encoded hydrogenase and the unclear association with partner bacteria in environmental samples ( Laso-Pérez et al., 2022 ). The capability of ANME-1 to perform methanogenesis has been repeatedly suggested ( Seifert et al., 2006 ; Treude et al., 2007 ; Orcutt et al., 2008 ). ANME-1 16S rRNA and mcr A genes and transcripts were found in methanogenic sediment horizons at White Oak River estuary and in the sulfate–methane transition zone ( Lloyd et al., 2011 ; Kevorkian et al., 2021 ). The authors weighted this as an argument for a potential role of ANME-1 in methanogenesis. There is no genomic evidence that ANME-1 from these environment contain hydrogenases, which are required to perform methanogenesis from CO 2 , and this question should be addressed in future metagenomic studies. To test whether ANME-1c are capable of methanogenesis, we transferred AOM70 culture aliquots to sulfate-free medium and exchanged the methane in the headspace with an H 2 /CO 2 atmosphere. Hydrogen addition did not stimulate production of methane in AOM70 cultures in the course of 4-month incubations ( Supplementary Figure 12 ). According to these results, ANME-1c are incapable of hydrogenotrophic methanogenesis. The habitable zones of the hydrothermally-heated sediments in Guaymas Basin are rich in sulfate due to hydrothermal circulation and seawater advection ( Ramírez et al., 2021 ). This provides additional evidence for ANME-1c being obligate methane oxidizers that depend on syntrophic partnerships with sulfate-reducers. The hydrogenase in ANME-1c might be a remnant from the common ancestor of the Ca. Syntrophoarchaeum/ Ca. Alkanophagales/ANME-1 clade. These ancestral alkanotrophic archaea might have performed interspecies electron transfer based on hydrogen transfer. In the course of evolution that capability was replaced by an apparently more efficient DIET mechanism via extracellular multiheme cytochromes ( Summers et al., 2010 )."
} | 9,892 |
32301469 | null | s2 | 1,055 | {
"abstract": "Fe is a critical nutrient to the marine biological pump, which is the process that exports photosynthetically fixed carbon in the upper ocean to the deep ocean. Fe limitation controls photosynthetic activity in major regions of the oceans, and the subsequent degradation of exported photosynthetic material is facilitated particularly by marine heterotrophic bacteria. Despite their importance in the carbon cycle and the scarcity of Fe in seawater, the Fe requirements, storage and cytosolic utilization of these marine heterotrophs has been less studied. Here, we characterized the Fe metallome of Pseudoalteromonas (BB2-AT2). We found that with two copies of bacterioferritin (Bfr), Pseudoalteromonas possesses substantial capacity for luxury uptake of Fe. Fe : C in the whole cell metallome was estimated (assuming C : P stoichiometry ∼51 : 1) to be between ∼83 μmol : mol Fe : C, ∼11 fold higher than prior marine bacteria surveys. Under these replete conditions, other major cytosolic Fe-associated proteins were observed including superoxide dismutase (SodA; with other metal SOD isoforms absent under Fe replete conditions) and catalase (KatG) involved in reactive oxygen stress mitigation and aconitase (AcnB), succinate dehydrogenase (FrdB) and cytochromes (QcrA and Cyt1) involved in respiration. With the aid of singular value decomposition (SVD), we were able to computationally attribute peaks within the metallome to specific metalloprotein contributors. A putative Fe complex TonB transporter associated with the closely related Alteromonas bacterium was found to be abundant within the Pacific Ocean mesopelagic environment. Despite the extreme scarcity of Fe in seawater, the marine heterotroph Pseudoalteromonas has expansive Fe storage capacity and utilization strategies, implying that within detritus and sinking particles environments, there is significant opportunity for Fe acquisition. Together these results imply an evolved dedication of marine Pseudoalteromonas to maintaining an Fe metalloproteome, likely due to its dependence on Fe-based respiratory metabolism."
} | 523 |
40328774 | PMC12056181 | pmc | 1,056 | {
"abstract": "Flexible thermoelectric devices enable direct energy conversion between heat and electrical energy, making them ideal for wearable electronics and personal thermal management. Yet, current devices lack functional module expansion, which limits the customization for diverse energy-harvesting heat sources and complicates their assembly to meet the specific power requirements of electrical appliances. Moreover, existing devices cannot be stacked to enhance thermoelectric cooling performance while maintaining flexibility and self-healing capabilities. Here, by selectively encapsulating liquid metal electrodes with carbon nanotube-doped self-healing materials with increased thermal conductivity, we substantially improve heat transfer across thermoelectric legs, thereby maximizing energy conversion efficiency. The device achieves a normalized power density of 3.14 μW⋅cm −2 ⋅ K −2 , setting a benchmark for self-healing thermoelectric devices. Benefiting from self-healing materials and liquid metal, the device demonstrates both self-healing capabilities and modular assembly, greatly expanding the application scenarios of flexible thermoelectric devices in wearable power generation and refrigeration.",
"introduction": "Introduction The rapid advancement of wearable electronic health and exercise monitoring devices has created a demand for efficient power sources that can adapt to diverse applications 1 – 6 . Among the various power supply technologies 7 – 12 , thermoelectric devices (TEDs) have received significant attention due to their ability to directly convert heat into electricity via the Seebeck effect 13 , 14 . As TEDs can harvest environmental thermal energy without mechanical components, they are considered promising candidates for powering wearable electronics 15 – 19 . Recently, the development of flexible thermoelectric devices has marked a significant step in improving both performance and conformability. Various approaches, such as utilizing organic thermoelectric materials 20 , 21 , embedding thermoelectric fibers 22 , 23 or bulks 15 – 17 , 24 into elastic polymers, and printable thermoelectric inks 25 , 26 have shown promise in creating flexible TEDs capable of adapting to different shapes and applications. Despite extensive efforts, several challenges impede the broader application of flexible TEDs. A primary issue is the lack of modularity and expandability of flexible TEDs, which restricts the customization for different energy-harvesting scenarios and hinders the convenient assembly for specific power requirements 27 , 28 . Another limitation lies in the use of polydimethylsiloxane (PDMS) as the encapsulating material in flexible TEDs. While PMDS endows inorganic TEDs with enhanced flexibility, it also leads to high thermal resistance, which decreases the temperature difference between the two ends of the thermoelectric legs and compromises the thermoelectric efficiency of devices. Furthermore, the lack of self-healing capabilities in many flexible TED designs poses durability concerns, particularly in wearable applications where mechanical stresses are prevalent. In this work, we developed a self-healing and modularized flexible TED by integrating the Bi 2 Te 3 -based TE legs, liquid metal EGaIn interconnects, and disulfide crosslinked polyurethane (DSPU) encapsulation. To improve the output power of this flexible TED, we created a selective encapsulation strategy to minimize the parasitic heat loss of the thermoelectric elements. The liquid metal interconnects were selectively encapsulated by using carbon nanotube-doped DSPU (CD). As a result, this thermoelectric device (CD-TED) demonstrates a high normalized power density of 3.14 μW cm −2 K −2 . This value outperforms recently reported self-healing flexible thermoelectric generators and is among the highest levels compared with other flexible TEDs 15 – 17 , 19 , 25 , 28 – 33 . The good performance in conjunction with the flexibility and self-healing ability of the CD-TED greatly advances its application in wearable power generation and low-power electronics. As our CD-TED presents exceptional self-healing properties coupled with the innovative design of LEGO-like shapes and wire circuits, it enables users to modularly assemble and expand the TEDs in both flat and vertical directions. This modular design allows customization based on the heat source or power consumption of the electronics. Moreover, implementing a multi-layered tower-like structure enables optimal cooling for thermoelectric devices, achieving a maximum cooling temperature of 6.2 K at a current level of 0.8 A. In comparison with single-layer thermoelectric devices, this technology has led to a larger cooling temperature difference by a factor of 1.6.",
"discussion": "Discussion This study introduces an innovative thermoelectric device (CD-TED) that is not only flexible and self-healing but also modularly assembled, offering exceptional performance. The selective encapsulation of liquid metal electrodes with carbon nanotube-doped self-healing materials enhances the heat transfer to the TE legs while maintaining its self-healing properties and exhibiting a thermal conductivity of 0.9 W m −1 K −1 . The CD-TED maximizes the Δ T TE /Δ T Applied ratio by up to 50%, leading to a remarkable 171% increase in power output compared to the D-TED. The finite element analysis further validates the improved heat transfer ability and TE performance, attributable to the incorporation of carbon nanotube-doped self-healing materials. Enhanced by self-healing materials and liquid metals, this thermoelectric device exhibits both self-healing capabilities and modular assembly, while also adapting to various cylindrical curvatures. These characteristics broaden the versatility of flexible thermoelectric devices, opening the avenues for wider applications in wearable energy harvesting and temperature regulation."
} | 1,473 |
37810281 | null | s2 | 1,057 | {
"abstract": "Engineered living materials (ELMs) combine living cells with polymeric matrices to yield unique materials with programmable functions. While the cellular platform and the polymer network determine the material properties and applications, there are still gaps in our ability to seamlessly integrate the biotic (cellular) and abiotic (polymer) components into singular material, then assemble them into devices and machines. Herein, we demonstrated the additive-manufacturing of ELMs wherein bioproduction of metabolites from the encapsulated cells enhanced the properties of the surrounding matrix. First, we developed aqueous resins comprising bovine serum albumin (BSA) and poly(ethylene glycol diacrylate) (PEGDA) with engineered microbes for vat photopolymerization to create objects with a wide array of 3D form factors. The BSA-PEGDA matrix afforded hydrogels that were mechanically stiff and tough for use in load-bearing applications. Second, we demonstrated the continuous "
} | 245 |
37887584 | PMC10604192 | pmc | 1,058 | {
"abstract": "Bio-inspired (biomimetic) materials, which are inspired by living organisms, offer exciting opportunities for the development of advanced functionalities. Among them, bio-inspired superhydrophobic surfaces have attracted considerable interest due to their potential applications in self-cleaning surfaces and reducing fluid resistance. Although the mechanism of superhydrophobicity is understood to be the free energy barrier between the Cassie and Wenzel states, the solid-surface technology to control the free energy barrier is still unclear. Therefore, previous studies have fabricated solid surfaces with desired properties through trial and error by measuring contact angles. In contrast, our study directly evaluates the free energy barrier using molecular simulations and attempts to relate it to solid-surface parameters. Through a series of simulations, we explore the behavior of water droplets on surfaces with varying values of surface pillar spacing and surface pillar height. The results show that the free energy barrier increases significantly as the pillar spacing decreases and/or as the pillar height increases. Our study goes beyond traditional approaches by exploring the relationship between free energy barriers, surface parameters, and hydrophobicity, providing a more direct and quantified method to evaluate surface hydrophobicity. This knowledge contributes significantly to material design by providing valuable insights into the relationship between surface parameters, free energy barriers, and hydrophilicity/hydrophobicity.",
"conclusion": "5. Conclusions We performed coarse-grained molecular simulations to investigate the effects of droplets on solid surfaces with bio-inspired nanostructures with the aim of contributing to the advancement of solid-surface engineering. Our primary aim was to gain a deeper insight into the properties of these surfaces, focusing in particular on the nanostructure inspired by the compound eyes of mosquitoes, known as the mosquito eye structure. Through this investigation, we aimed to explore the influence of the physical properties of this nanostructure on the free energy barrier. Our results can be summarized as follows: The magnitude of the free energy barrier serves as an indicator of the hydrophilicity/hydrophobicity of the surface. In general, higher free energy barriers correspond to more hydrophobic surfaces, while lower free energy barriers correspond to more hydrophilic surfaces. When considering the physical properties alone without taking into account the surface chemistry, we observed that increasing the distance between the surface pillars results in a lower free energy barrier. Similarly, when considering only the physical properties and not the surface chemistry, we found that higher surface pillar heights lead to higher free energy barriers. The w − Δ G and h − Δ G curves obtained from our simulations provide predictive and evaluative capabilities for the free energy barrier of the surface. This allows us to predict and evaluate the hydrophilicity/hydrophobicity of the surface. Our study provides valuable insights into the behavior of droplets hitting solid surfaces with bio-inspired nanostructures. By considering the physical properties of the nanostructure, we gain a better understanding of how these surfaces affect the free energy barrier. These findings have implications for the field of solid-surface engineering and offer potential applications in the design of surfaces with specific hydrophilic or hydrophobic properties.",
"introduction": "1. Introduction Bio-inspired materials, which are inspired by living organisms, have emerged as a promising avenue for the development of advanced functions in various scientific disciplines [ 1 , 2 , 3 ]. Among these materials, bio-inspired superhydrophobic surfaces are gaining attention for their potential applications in self-cleaning surfaces and reducing fluid drag [ 4 ]. The compound eyes of mosquitoes [ 5 , 6 ], the tiny hairs of lotus leaves [ 7 , 8 ], and the leg structures of water striders [ 9 ] are typical examples of superhydrophobic structures found in nature that exhibit significant hydrophobic properties. In this context, scientists aim to explore the nanostructural intricacies of these bio-inspired materials to understand the underlying mechanisms responsible for their unique properties. The ability to mimic the hydrophobic properties of these natural structures opens up new possibilities for a wide range of practical applications, from anti-fouling coatings to efficient fluid transport systems [ 1 , 10 ]. A thorough understanding of the Wenzel [ 11 ] and Cassie [ 12 ] states is essential when studying surface hydrophobicity. These two states are crucial in describing the wetting behavior of droplets on rough or textured surfaces. In the Wenzel state, the droplet wets the surface completely, spreading over the entire roughness and making contact with the bottom of the surface structure. On the other hand, the Cassie state is characterized by the droplet resting on top of the surface structure, with only the tips of the surface features in contact, and there may be air pockets trapped between the droplet and the bottom of the surface structure. Although the Cassie state is highly hydrophobic, the Wenzel state has a lower equilibrium free energy than the Cassie state, and there is a large free energy barrier between the two states, so that a drop in the Wenzel state cannot spontaneously return to the Cassie state. This free energy barrier is therefore the dominant factor in the design of superhydrophobic materials. Leroy and Müller-Plathe’s pioneering work proposed that surface hydrophilicity/ hydrophobicity could be tuned by controlling the length and depth of the roughness pattern [ 13 ]. Their research was instrumental in elucidating the relationship between surface parameters and the wetting behavior of materials. However, a limitation of their study was the lack of a quantified evaluation standard to accurately measure these effects. Colin and Parkin suggested that the design of superhydrophobic surfaces should focus on two main features: low surface energy to achieve a contact angle above 90 ∘ on a flat surface and high surface roughness to increase the hydrophobicity of the surface [ 14 ]. For hydrophobic polymer films, hydrophobicity is evaluated using contact angle and surface energy [ 15 ]. The current study relies on contact angle and surface energy as the primary evaluation criteria. The contact angle serves as a common indicator to measure the contact state of liquid droplets on the surface of a material, providing insight into the interaction between the surface of the material and the liquid. If the surface of the material is hydrophobic, the droplets exhibit a high contact angle, indicating a preference for sliding across the surface rather than making direct contact. On the other hand, surface energy plays a pivotal role in characterizing the surface properties of a material, with low surface energy typically correlating with hydrophobic behavior [ 16 ]. As a result, these evaluation criteria have been widely used in research and application to assess the hydrophobicity of various materials. Quéré et al. introduced an analytical model to investigate the impact of surface parameters on hydrophobicity and contact angles [ 17 ]. Their work highlighted the significant correlation between surface hydrophobicity and surface free energy. However, they did not establish a direct link between surface free energy barriers, surface parameters, and surface hydrophobicity. However, as mentioned above, while the free energy barrier is the essence of the phenomenon, it is also a quantity that is difficult to measure directly. In recent years, there have been numerous instances of employing molecular simulations to investigate droplet behavior on surfaces [ 18 , 19 , 20 , 21 ]. Molecular simulations offer a comprehensive view of the entire droplet–surface interaction process, allowing for the precise and quantitative adjustment of surface parameters. Consequently, this approach facilitates a more in-depth exploration of the intricate interplay between surface parameters and droplet behavior on the surface. While many researchers have traditionally focused on investigating the relationship between droplet contact angles and surface energy, our research takes a unique perspective. We delve into the connection between surface parameters, free energy barriers, and material properties. This departure from the conventional methodology is characterized by the utilization of the free energy barrier as a fundamental reference standard to evaluation surface hydrophobicity. Therefore, our study goes beyond traditional approaches by exploring the relationship between free energy barriers, surface parameters, and hydrophobicity, providing a more direct and quantified method to evaluate surface hydrophobicity. Through a series of simulations, we investigate the behavior of water droplets on surfaces with varying surface parameters. We establish a profound relationship between surface parameters, free energy barriers, and material properties. By bypassing the reliance on conventional contact angle measurements, our research opens new avenues for the accurate assessment of surface hydrophobicity and provides valuable insights for material design and engineering. This innovative approach allows more direct determination of the free energy barrier on the surface parameter and provides a deeper understanding of how to optimize materials for desired hydrophilic/hydrophobic properties, thereby streamlining material design processes.",
"discussion": "4. Discussion From Figure 4 , it is clear that the free energy barrier increases as the ratio of w / r droplet decreases, indicating a smaller distance between the surface nanobumps. This suggests that a higher energy is required for the transition from the Cassie state (where the droplet is on top of the surface bumps) to the Wenzel state (where the droplet penetrates into the gaps between the bumps), making the transformation of droplets to the Wenzel state on the surface more difficult. In our study, we constructed the model based on the compound eyes of mosquitoes, which have very small distances between nanobumps. This finding is consistent with the research of Bhushan [ 7 ], as they observed that decreasing the distance between surface bumps increases the probability of air pocket formation and leads to an increase in the contact angle, indicating a more hydrophobic surface. These results support the simulation results in our study, where smaller values of w correspond to larger free energy barriers, indicating greater surface hydrophobicity. The results obtained in this study show that Δ G increases quadratically as the space between the columnar structures decreases. It is important to note that our simulation model represents an idealized state; it models a perfectly uniform pattern in a mosquito eye structure with a narrow extra space. In real-world scenarios, the surface would have a more complex structure with varying column heights. In practical situations, reducing the spacing between the nanobumps would likely have a more pronounced effect on the free energy barrier of the surface than shown by the curve resulting from our simulations. Note that although these relationships held for ranges of w and h close to the droplet’s radius, it is not clear whether they hold for a narrower w or a higher h. This will be investigated in the future. According to Extrand’s theory, the surface forces should exceed gravity (or other body forces) and act in an upward direction. In addition, the surface columns must be high enough to prevent the liquid from coming into contact with the bottom of the surface, which could cause the liquid to be pulled down and can lead to collapse. Figure 5 shows a significant increase in the free energy barrier as the height of the surface pillars ( h ) increases, particularly when the pillar height exceeds the radius of the droplet. This indicates that increasing the pillar height effectively increases the free energy barrier. Based on Figure 6 , it is evident that the slope undergoes a significant change (approximately 3.85 times) before and after the surface pillar height becomes equal to the droplet radius. This observation may lead us to hypothesize that when the surface pillar height exceeds the liquid diameter, there is also a sudden increase in the free energy barrier. This is expected to be related to the position of the center of mass of the droplet and the height of the column. Specifically, the slope of Line 1 is 0.64 × 10 6 , while the slope of Line 2 is 2.49 × 10 6 . The slope is approximately 3.85 times higher. This indicates that the height of the surface pillar has a more pronounced effect on the surface free energy barrier. Furthermore, the effect of changing the height of the surface pillars appears to be more substantial than that of changing the distance between the pillars. The results show that both w and h influence the magnitude of the free energy barrier. In the study by Burton and Bharat [ 31 ], it is observed that introducing a pattern on a flat polymer surface reduces adhesion and the coefficient of friction while increasing the contact angle, indicating a more hydrophobic surface. Similarly, Kwon et al. [ 32 ] demonstrate that the incorporation of hierarchical nanotextures on the surface increases its hydrophobicity. Both studies emphasize the positive effect of surface roughness and hierarchy on hydrophobicity. An intermediate state, more typical in nature, becomes increasingly likely when we adjust the surface pillar height, h , to match the diameter of the droplet. We have calculated the relationship between the droplet center coordinates with time in Figure 7 , and it can be seen that after a long period of time after the droplet impacts on the surface, the droplet center coordinates are almost unchanged; this can be basically recognized as a stable state. In this intermediate state, more than 70% of the droplet is usually fully immersed in the pillar structure, and sometimes even 100% of the droplet is fully immersed in the column structure, yet the droplet does not touch the lower end of the surface. This phenomenon aligns with David’s suggestion of a transitional state between the Cassie and Wenzel states [ 33 ]. We postulate that the frequently observed intermediate state in our superhydrophobic interface simulations represents this transitional phase. Our conjecture regarding the energy landscape of this transition state is illustrated in Figure 8 . In this situation, a droplet in the intermediate state has the potential to transition to either the Cassie or the Wenzel state. By providing the necessary energy stimulus to activate the droplet’s motion, it may be possible to manipulate the droplet in the intermediate state, making it more amenable to transitioning into either the Cassie or Wenzel state. We attempted to integrate the functions w – G and h – G into a three-dimensional coordinate system depending on the function ( 5 ), with the x - and y -axes representing the ratios of w and h to r droplet , and the z -axis representing Δ G 2 , as shown in Figure 9 Upon analysis, we observed that the highest point on the surface occurs when w / r droplet is at its minimum and h / r droplet is at its maximum. It is also evident that the rate of change of Δ G along the h / r droplet axis is greater than the rate of change along the w / r droplet axis. This observation suggests that designing the material to emphasize the height of the surface pillars is a more promising approach to increasing the hydrophobicity of the surface rather than focusing solely on adjusting the distance between the pillars. This approach not only adds depth to our research but also visualizes the intricate connection between surface parameters and the associated energy barriers, aiding with a comprehensive understanding of our study’s outcomes.\n (5) Δ G 2 ( w , h ) = Δ G ( w ) · Δ G ( h ) Here, Δ G ( w ) = a w 2 + b w + c , and Δ G ( h ) = d h 2 + e h + f ; a, b, c, d, e, and f are fitting parameters. A unique aspect of our study is the ability to calculate the free energy barrier based solely on surface parameters. In previous research, surface experiments have typically been carried out to measure the contact angle of droplets on the surface, allowing an assessment of the hydrophilic or hydrophobic nature of the surface. In contrast, our study focuses on calculating the magnitude of the free energy barrier using only surface parameters, eliminating the need for experimental measurements. This provides significant convenience for material design, as it allows the direct determination of surface hydrophilicity or hydrophobicity based on the free energy barrier and without the need to perform experiments."
} | 4,249 |
39090985 | PMC11388927 | pmc | 1,059 | {
"abstract": "Abstract Chain elongating bacteria are a unique guild of strictly anaerobic bacteria that have garnered interest for sustainable chemical manufacturing from carbon-rich wet and gaseous waste streams. They produce C 6 –C 8 medium-chain fatty acids, which are valuable platform chemicals that can be used directly, or derivatized to service a wide range of chemical industries. However, the application of chain elongating bacteria for synthesizing products beyond C 6 –C 8 medium-chain fatty acids has not been evaluated. In this study, we assess the feasibility of expanding the product spectrum of chain elongating bacteria to C 9 –C 12 fatty acids, along with the synthesis of C 6 fatty alcohols, dicarboxylic acids, diols, and methyl ketones. We propose several metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and utilize constraint-based metabolic modelling to predict pathway stoichiometries, assess thermodynamic feasibility, and estimate ATP and product yields. We also evaluate how producing alternative products impacts the growth rate of chain elongating bacteria via resource allocation modelling, revealing a trade-off between product chain length and class versus cell growth rate. Together, these results highlight the potential for using chain elongating bacteria as a platform for diverse oleochemical biomanufacturing and offer a starting point for guiding future metabolic engineering efforts aimed at expanding their product range. One-Sentence Summary In this work, the authors use constraint-based metabolic modelling and enzyme cost minimization to assess the feasibility of using metabolic engineering to expand the product spectrum of anaerobic chain elongating bacteria.",
"conclusion": "Conclusion Developing chain elongating bacteria as a bioproduction platform for oleochemical synthesis could play a crucial role in sustainable manufacturing from carbon-rich waste streams. Our analyses demonstrate that the synthesis of C 9 –C 12 FAs, as well as conversion to their corresponding fatty alcohols, dicarboxylic acids, diols, and MKs, is thermodynamically feasible and generates sufficient ATP production for growth coupling. In the case of C 9 –C 12 FAs, enzyme cost analyses point towards a trade-off between growth rate and ATP yield with increasing chain length, which may explain why chain elongation in nature terminates at chain lengths of C 8 . While some experimental data support our results for selectivity for hexanoate over butyrate production in C. lactatifermentans and C. kluyveri at lower growth rates (Kenealy & Waselefsky, 1985 ; Wang et al., 2022 ), investigating the effect of chain length on growth rate and ATP yield considering other substrates across the diversity of chain elongating bacteria is still needed. Other aspects that need to be investigated to understand limitations in chain length include the effect of product toxicity and enzyme–substrate affinity for longer chain FAs. This includes strategies chain elongators employ to protect against MCFA toxicity and evaluation of the substrate range that can be accommodated by native RBO cycle enzymes. To address these problems, genetic engineering tools need to be developed, including genetic cargo delivery systems, the construction of regulatory element libraries, and gene knock-in/knock-out methods to delete native chain elongating bacteria genes and to enable heterologous gene expression. These genetic modifications should allow the production of a broad range of oleochemicals through chain elongation using the pathways proposed in this work. The production of longer chain length FAs and their subsequent conversion into alcohols are ideal targets for initial demonstrations of oleochemical production using engineered chain elongating bacteria as these products are not redox limited in our analyses. However, the functionalization of the aliphatic end of the FAs required for dicarboxylate and diol production will be a significant challenge due to the anaerobic requirements of chain elongating bacteria. While we present an avenue to accomplish this through nitrite reduction and the use of microcompartments, screening of other enzymes that can functionalize FAs without oxygen will be important to chain elongating bacteria metabolic engineering efforts. Additionally, the production of MKs was found to be feasible, though not growth coupled as it requires carboxylate co-production for ATP production. Given previous demonstrations of MK production in E. coli under microaerobic conditions (Lan et al., 2013 ), other conditions and production pathways should be investigated to improve yields in chain elongators. As a proof of concept, our modelling work focused on the conversion of single carbon substrates to various oleochemicals using chain elongating bacteria in isolation. The deployment of such engineered strains will likely require investigation into how mixed substrates and potential interactions with other microbes will affect product profiles. For example, the presence of more electron-rich substrates, such as glucose, may push fermentation towards less desirable, shorter chain products (e.g. acetate, butyrate) as it is advantageous for growth rather than chain elongation based on the insights from our resource allocation modelling. Overall, our analyses indicate that chain elongating bacteria are a promising biomanufacturing chassis for accessing several oleochemical product classes. This work provides a basis for which bioproducts could be accessed through metabolic engineering of chain elongating bacteria and highlights potential barriers, including trade-offs between growth and product yield, which need to be further investigated. Developing genetic tools for chain elongating bacteria will be critical for validating our modelling predictions, understanding the fundamental physiology of this functional guild, and expanding their product spectrum, in isolation as pure cultures, and in the context of self-assembled and synthetic microbiomes.",
"introduction": "Introduction Circular economies require the recycling of societal “wastes” into new bioproducts to support sustainable human activity. Anaerobic fermentation processes can upcycle carbon from agricultural residues, food waste, and industrial off-gases into useful fuels, chemicals, and materials. One promising process is microbial chain elongation, which uses anaerobic microbiomes in open-culture systems to synthesize medium-chain fatty acids (MCFAs) from complex organic and gaseous waste streams (Angenent et al., 2016 ; Holtzapple et al., 2022 ; Scarborough et al., 2022 ). The process works by producing key intermediates, such as lactate and/or ethanol, through either organic waste fermentation (breaking down) or gas fermentation (building up), which subsequently undergo a secondary fermentation to produce MCFAs via chain elongation. Several studies have demonstrated stable MCFA production from organic waste (Grootscholten et al., 2014 ; Stamatopoulou et al., 2020 ) and gaseous feedstocks (Bäumler et al., 2022 ; Diender et al., 2016 ; Fernández-Blanco et al., 2022 ) via chain elongation at the bench scale and pilot scale, and more recently, a demonstration plant has been built in the Netherlands by the Dutch company ChainCraft ( https://www.chaincraft.nl/ ). At the heart of the process is a functional guild of obligate anaerobes called “chain elongating bacteria” that use a native reverse beta-oxidation (RBO) pathway (see Fig. 1B ) to ferment lactate, ethanol, and other electron-rich organic substrates (e.g. sugars and glycerol) into C 4 –C 8 carboxylates (i.e. butyrate, hexanoate, and octanoate) as part of their growth. The pathway is unique because it allows redox balancing, while also conserving energy through a novel flavin-based electron bifurcation mechanism (Buckel & Thauer, 2018 ; Li et al., 2008 ). Almost all known chain elongators belong to the phylum Bacillota (previously Firmicutes) and are a phylogenetically and physiologically diverse group. While most chain elongators remain uncultivated, over 15 strains have been isolated and sequenced to date, with most new isolates being reported in the last 10 years (Candry & Ganigué, 2021 ). These isolates, particularly Clostridium kluyveri , have been used to establish synthetic co-cultures to convert gaseous or sugar-based feedstocks to MCFAs and their corresponding alcohols (Bäumler et al., 2022 ; Diender et al., 2016 ; Haas et al., 2018 ; Lynd et al., 2022 ; Otten et al., 2022 ) which could be expanded to more complex feedstocks by selecting lactate- and/or ethanol-producing partners with improved hydrolytic capabilities. Fig. 1. (A) Anaerobic conversion of organic waste to oleochemicals using synthetic co-cultures inspired by natural systems. Complex polymers are degraded by lactic acid bacteria (LAB), which in turn provide lactic acid to chain elongating bacteria for medium-chain fatty acid synthesis. (B) The reverse beta-oxidation pathway in chain elongating bacteria and proposed product spectrum expansion (brown boxes). EMP = Embden–Meyerhof–Parnas glycolysis; LDH = electron-confurcating lactate dehydrogenase/electron-transferring flavoprotein; ADH = alcohol dehydrogenase; ADA = acetaldehyde dehydrogenase; PFOR = pyruvate ferredoxin oxidoreductase; PFL = pyruvate formate lyase; PTA = phosphate acetyltransferase; ACK = acetate kinase; ACACT = acetyl-CoA C acetyltransferase; HACD = 3-hydroxyacyl-CoA dehydrogenase; ECOAH = enoyl-CoA hydratase; EBACD = electron bifurcating acyl-CoA dehydrogenase; CoAT = CoA transferase; Fd = ferredoxin; RNF = proton translocating ferredoxin: NAD + oxidoreductase complex; HYD = Fe–Fe hydrogenase. Created with BioRender.com . The development of synthetic co-cultures, inspired by mixed culture chain elongation processes, could represent a platform for improving titres, rates, and yields of MCFA production (Fig. 1A ). Moreover, there is growing interest in genetically modifying the native RBO cycle in chain elongators to expand the product spectrum beyond C 4 –C 8 monocarboxylates, including alcohols, diols, dicarboxylates, and other bulk chemicals (Agena et al., 2023 ; Guss & Riley, 2021 ; Strik et al., 2022 ). This would enable the anaerobic synthesis of diverse medium-chain oleochemicals from wet and gaseous waste streams, which is expected to have improved economic feasibility compared to sugar-based aerobic fermentations (Holtzapple et al., 2022 ). However, the technical feasibility of making products other than C 4 –C 8 monocarboxylates via metabolic engineering of chain elongators remains unexplored. As tools to genetically modify chain elongators emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), methodologies to rationally design these new biocatalysts are needed. In this study, we use constraint-based metabolic modelling along with thermodynamic analyses to evaluate the feasibility of synthesizing diverse medium-chain oleochemicals (C 6 –C 12 fatty acids [FAs], primary alcohols, dicarboxylates, diols, and methyl ketones [MKs]) from key intermediate substrates (lactate, ethanol, sugars, and glycerol) using anaerobic chain elongating bacteria (Fig. 1 ). We first analyse MCFA production scenarios beyond natural C 8 production up to C 12 , highlighting trade-offs between ATP yield and expected growth rate using a combination of metabolic modelling and enzyme cost minimization (ECM) analyses. Subsequently, we propose several modifications to the RBO cycle to synthesize target medium-chain oleochemical products and use constraint-based metabolic modelling to determine overall pathway stoichiometry, thermodynamic feasibility, and theoretical product yields. Our results indicate that metabolic engineering of chain elongating bacteria could enable the anaerobic synthesis of diverse medium-chain oleochemicals at industrially relevant yields. Moreover, we identify challenges with engineering the RBO pathway in chain elongators and offer potential solutions to overcome them via metabolic engineering. We anticipate that these results will provide a useful starting point for engineering microbial chain elongation to serve as a platform for sustainable chemical manufacturing.",
"discussion": "Results and Discussion Impact of MCFA Chain Length on Growth Rate and ATP Yield Modelling Chain Elongation Beyond Octanoic Acid Currently, the products of chain elongation are restricted to butyric and hexanoic acids (C 4 –C 6 ), with limited synthesis of octanoic acid (C 8 ) (Nelson et al., 2017 ; Zhu et al., 2017 ). Genetically modifying chain elongating bacteria to improve selectivity and increase MCFA chain length could benefit product yields, while also expanding the process to produce MCFAs with larger markets (C 8 –C 12 ). To assess the feasibility of MCFA production beyond C 8 , we predicted overall pathway stoichiometry, redox balance, free energy change, and theoretical ATP and product yields of C 4 –C 12 FAs via pFBA with ATP yield set as the objective function using a simplified metabolic model describing core chain elongation metabolism (iFermCell193, see the “Methods” section) (Table 1 ). We focused on lactate utilizing chain elongating bacteria (e.g. Pseudoramibacter alactolyticus, Megasphaera hexanoica, Caproicibacterium lactatifermentans ) as lactate is the primary substrate in mixed culture processes converting organic wastes (Fig. 1 ) (Contreras-Dávila et al., 2020 ). Model simulations with different electron donors, including ethanol, glucose, xylose, and glycerol, were also evaluated ( Supplemental Material S1 , Tables S1 – S5 ) to demonstrate process feasibility for a range of other substrates known to be consumed by chain elongating bacteria (e.g. ethanol by Clostridium kluyveri ). Table 1. Predicted Stoichiometry for the Synthesis of C 4 –C 12 FAs from Lactate Chain length Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol lactate) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) C 4 Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.500 0.250 −14.99 −59.96 0.489 C 6 Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 C 8 Lactate + 0.75 H + + 0.375 ADP → CO 2 + 0.5 H 2 + 0.5 H 2 O + 0.25 octanoate + 0.375 ATP 0.250 0.375 −35.10 −93.60 0.402 C 10 Lactate + 0.8 H + + 0.4 ADP → CO 2 + 0.2 decanoate + 0.4 H 2 + 0.6 H 2 O + 0.4 ATP 0.200 0.400 −40.11 −100.29 0.385 C 12 Lactate + 0.833 H + + 0.417 ADP → CO 2 + 0.167 dodecanoate + 0.333 H 2 + 0.667 H 2 O + 0.417 ATP 0.167 0.417 −43.45 −104.29 0.373 The model predicted that ATP yield and overall reaction free energy change increase with MCFA chain length from C 4 to C 12 (Table 1 ), indicating that MCFA synthesis beyond C 8 could be theoretically possible. Moreover, the production of C 4 –C 12 chain lengths from lactate released ≥50 kJ/mol per ATP, indicating that these reactions produce sufficient energy for ATP synthesis (Thauer et al., 1977 ). C 9 –C 12 carboxylate production with other electron donors, including ethanol, glucose, xylose, and glycerol, was also found to be feasible ( Supplemental Material S1 , Tables S1 – S5 ). This observation agrees with past modelling results that suggested hexanoic and octanoic acid production should improve ATP yields from lactate compared to butyric acid production (Scarborough et al., 2018 ). The predicted increase in net ATP production is attributed to a greater flux through the proton translocating ferredoxin: NAD + oxidoreductase complex (RNF) for each turn of the RBO cycle, which leads to greater ATP production via the ion motive force (IMF). In the RBO cycle, each successive elongation generates reduced ferredoxin from the electron-bifurcating acyl-CoA dehydrogenase (EBACD). This ferredoxin goes on to drive RNF to recover some NADH and/or is used by a soluble ferredoxin hydrogenase (HYD1) to evolve H 2 as a terminal electron sink. As confirmed in the predicted flux distributions, elongation to longer chain lengths increases the total NADH demand required to reduce the acyl chains (HACD and EBACD in Fig. 1 ). Thus, with increasing chain length, a greater proportion of the reduced ferredoxin produced by EBACD is used to regenerate NADH via RNF, rather than driving HYD1. This ultimately leads to increased ATP production through the IMF along with decreased H 2 evolution for longer chain lengths. These results also indicate that past a certain chain length (>C 12 ), H 2 evolution will cease as all electron equivalents will be required solely for chain elongation. Intracellular Thermodynamic Landscape and Resource Allocation Our initial modelling analysis suggested that the improved ATP yield with longer chain lengths should naturally select for MCFAs beyond C 8 . However, all pure and mixed culture chain elongation studies have only observed C 4 –C 8 MCFA products. This points to a potential trade-off between growth yield (or ATP yield) and growth rate (or ATP production rate), resulting from optimal resource allocation. In a model based on resource allocation theory ( Supplementary Material S1 ; inspired by a similar model developed by Flamholz et al., 2024 ), where catabolic ATP production and anabolic ATP consumption rates are linearly related to the biomass fractions allocated to either function, RBO enzyme cost per unit ATP production flux and chain elongating bacteria growth rate are inversely correlated (Fig. 2 ). This simplified model captures the fact that a higher growth rate is expected to necessitate both a larger pool of anabolic enzymes and ribosomes (Basan, 2018 ) and a higher ATP production flux, requiring that catabolic machinery produces ATP faster with less enzyme. Fig. 2. ECM results of RBO for production of C 4 –C 12 carboxylates from lactate and the estimated relative effect on growth rate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Each reaction carries a cost which accounts for the pathway's stoichiometry (bottom bar), increased cost incurred due to a reaction's proximity to equilibrium (middle bar), and the cost due to sub-saturation reactant concentrations (top bar). Bottom right panel: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate, as derived from a simplified resource allocation model (Supplemental Material S1) (Top). Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate (Middle). ATP yield on lactate for different chain elongation simulated by pFBA (Bottom). To evaluate the enzyme investment per ATP flux required to produce carboxylates of different chain lengths, we performed ECM analysis. ECM places a lower bound on a pathway's enzyme demand per unit flux [g/(mol/h)] by accounting for the pathway's length and stoichiometry, reaction thermodynamics, and saturation effects following from the optimal set of metabolite concentrations, given appropriate bounds on those concentrations and parametrization of enzyme kinetics (Flamholz et al., 2013 ). This analysis coupled with ATP yield results from our model showed that minimal enzyme cost per unit ATP flux increases with chain length (Fig. 2 ). This suggests that chain elongating bacteria making longer length products invest more of their proteome into high yield catabolism, at the trade-off of having less catalytic capacity for growth. These modelling predictions align with experimental results indicating that increased selectivity and production for hexanoic acid over butyrate is associated with a decrease in growth rate in C. lactatifermentans when grown on lactate as compared to glucose (Wang et al., 2022 ). Moreover, caproate selectivity has been shown to increase with hydraulic retention time (reduced growth rate) in continuous cultures of C. kluyveri and a chain elongation microbiome (Grootscholten et al., 2013 ; Kenealy & Waselefsky, 1985 ). According to the optimized free energy changes predicted by ECM, thiolase reactions (ACACT) are the least exergonic in RBO ( Supplemental Material S1 , Fig. 1 ). Since these reactions operate near equilibrium, they have a smaller forward-to-reverse flux ratio (Noor et al., 2014 ) and therefore require a larger enzyme pool to sustain a given net flux through the pathway. The set of metabolite concentrations that minimizes pathway enzyme cost fixes the thiolase products (oxoacyl-CoAs) at low concentrations to drive these reactions ( Supplemental Material S2 ). This heightens the demand for the proceeding enzymes in the pathway, 3-hydroxyacyl-CoA dehydrogenases (HACDs) since they are far from saturation (Fig. 2 ). Similarly, enoyl-CoA hydratase (ECOAH) reactions are thermodynamically constrained, meaning their reactant and product concentrations are kept high and low, respectively ( Supplemental Material S2 ). This further increases the protein burden of HACD as well as the proceeding electron bifurcating acyl-CoA dehydrogenase (EBACD) (Fig. 2 ). At longer chain lengths, ECOAH reactions are predicted to be more favourable ( Supplemental Material S1 , Fig. 1 ), causing the rate at which enzyme cost increases with chain length to decrease. Interestingly, when oxoacyl-CoA concentrations are constrained by a lower bound of 1 µM, as is the case for all other metabolites, products beyond C 4 are deemed infeasible due to endergonic subsequent thiolase reactions. To reflect observations of chain elongation beyond C 4 , oxoacyl-CoAs can fall to sub-micromolar levels in this model. An alternative solution, as previously suggested in a similar instance with the citric acid cycle (Noor et al., 2014 ), is to posit enzymatic channelling. Channelling between ACACT and HACD would reduce the protein cost of both reactions by effectively merging the two into one favourable reaction. Indeed, channelling between ACACT, HACD, and ECOAH has been described with crystal structures of bacterial and human beta-oxidation complexes (Ishikawa et al., 2004 ; Xia et al., 2019 ). So long as a similar mechanism is active from C 4 to C 12 , the trend of enzyme cost per ATP flux increasing with chain length is expected to hold. This has major implications for engineering chain elongating bacteria to produce MCFAs beyond C8, as the production of longer chain length would provide more ATP per mole of electron donor, but at the cost of a lower growth rate. As a result, compensatory efforts to bolster growth rate, for example through adaptive laboratory evolution (Sandberg et al., 2019 ), may be needed to achieve higher productivity. Expanding and Controlling Chain Elongation Products with Metabolic Engineering Metabolic engineering strategies for the expansion of RBO-derived products have been reviewed for non-chain elongating, model organisms with established genetic tools, such as Escherichia coli, Saccharomyces cerevisiae , and select acetogenic Clostridium species (Tarasava et al., 2022 ). The approach for these strains relies on the engineered reversal of β-oxidation accomplished through extensive strain engineering such as knockout or repression of native β-oxidation regulators and overexpression of genes from chain elongating bacteria and other oleaginous hosts (Tarasava et al., 2022 ). Most of these hosts do not natively rely on the RBO cycle for redox balancing or energy conservation, which is what makes the RBO cycle growth coupled in the chain elongating bacteria. Further, installing orthogonal systems for the engineered reversal of β-oxidation imposes an increased metabolic burden on non-chain elongating hosts as high expression is likely needed to ensure sufficient flux through the pathway (Liu et al., 2018 ). This can lead to redox limitations or inefficient substrate use for chain elongation in these non-chain elongating hosts, and approaches such as two-phase production are likely needed to maximize product yields (Lange et al., 2016 ). Production with non-chain elongating hosts will likely require further strain engineering to improve product yields and selectivity. As an alternative to engineering model organisms, tools to genetically modify chain elongating bacteria are beginning to emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), which will open the door to expanded opportunities to produce oleochemicals via anaerobic fermentation. Chain elongating bacteria are ideal hosts for RBO-based bioproduction as the RBO cycle is innately growth coupled in these strains (no reconstruction needed) and as members of anaerobic communities can utilize solid and gaseous waste feedstocks rather than relying on refined substrates (e.g. sugars). Further, product yields in anaerobes are typically much greater as they do not synthesize as much biomass compared to aerobic systems and most electrons end up in products (Cueto-Rojas et al., 2015 ). However, the feasibility of pathways for the production of other oleochemicals, including, fatty alcohols, dicarboxylates, diols, and MKs, has not been explored in anaerobic chain elongating bacteria. Here, we propose potential production pathways and enzymes required to generate a wide range of oleochemicals (Fig. 3 ) and use pFBA and ECM to assess the feasibility and potential effects on ATP yield and growth rate. Fig. 3. Proposed pathways for chain elongating bacteria product spectrum expansion. (A) Fatty alcohols pathway. AdhE = bifunctional aldehyde-alcohol dehydrogenase; AOR = aldehyde: ferredoxin oxidoreductase; FadM = acyl-CoA thioesterase. (B) Diol and dicarboxylic acids pathways. NIR = nitrite reductase; NOD = nitric oxide dismutase; AlkBGT = alkane 1-monooxygenase complex; yjgB = cinnamyl alcohol dehydrogenase; chnE = 6-oxohexanoate dehydrogenase; Rd = rubredoxin. (C) Methyl ketone pathway. FadM: thioesterase. Created with Biorender.com . iFermCell193 was further modified to capture the reactions required for hexanoic acid conversion to other products (iFermCell356). Table 2 summarizes the potential pathway stoichiometries for the conversion of hexanoic acid (C 6 ) to the corresponding fatty alcohol, dicarboxylate, and diol using lactate as an electron donor, along with the net ATP yields and overall Gibbs free energy change of reaction for each substrate–product pair as predicted by our model. All predicted pathway stoichiometries have overall favourable standard Gibbs free energies of reaction and release ≥50 kJ/mol per ATP required for ATP production. The model also predicts that net ATP yield is not negatively impacted by conversion into these other bioproducts, which indicates that their production can be growth coupled, similar to MCFAs. In the following sections, we describe potential metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and discuss specific shifts in redox requirements that occur predicted by the model, along with the impact on the metabolic cost of the additional enzymes estimated by ECM (Fig. 4 ). The production of methyl ketones was also modelled, but it was found to negatively impact ATP yield (Fig. 5 ). Fig. 4. ECM results of RBO for production of novel chain elongation products from lactate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Fig. 5. Production profiles for the co-production of 2-pentanone and butyrate from lactate and standard Gibbs free energy of reaction. Molar yields of ATP are negatively related to increasing production of 2-pentanone. Co-production of butyrate is required with increasing 2-pentanone production. Table 2. Predicted Pathway Stoichiometry for Different C 6 Oleochemicals Produced from Lactate Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol Lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Carboxylate; RBO Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 Alcohol; RBO + AdhE Lactate + H + + 0.333 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.333 ATP 0.333 0.333 −45.18 −135.53 0.382 Alcohol; RBO + AOR Lactate + H + + 0.5 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.5 ATP 0.333 0.500 −45.18 −90.35 0.382 Dicarboxylate; RBO + AlkBGT Lactate + H + + 0.667 NO 2 − + 0.333 ADP → 0.333 adipate + CO 2 + 0.667 H 2 + H 2 O + 0.333 N 2 + 0.333 ATP 0.333 0.333 −228.21 −684.62 0.539 Diol; RBO + AlkBGT Lactate + 0.5 NO 2 − + 1.25 H + + 0.5 ADP → 0.25 1,6-hexanediol + 0.25 acetate + CO 2 + H 2 O + 0.25 N2 + 0.5 ATP 0.250 0.500 −191.63 −383.26 0.332 RBO = reverse beta-oxidation, AdhE = bifunctional alcohol dehydrogenase, AOR = aldehyde:ferredoxin oxidoreductase, AlkBGT = alkane 1-monooxygenase complex Fatty Alcohol Production Medium-chain fatty alcohols (FAOHs) could be produced using MCFAs as building blocks. Even though FAOH production pathways (Fig. 3A ) exist in anaerobic bacteria, such as acetogens and butanol-producing clostridia (Moon et al., 2016 ), this has yet to be observed in most chain elongating bacteria. In nature, acetogenic bacteria use aldehyde:ferredoxin oxidoreductase (AOR) to convert FAs into fatty aldehydes in monoculture or when they are co-cultured with chain elongating bacteria (Benito-Vaquerizo et al., 2020 ; Diender et al., 2016 ). In another instance, aldehyde-alcohol dehydrogenase (AdhE) can convert acyl-CoAs into aldehydes (Mehrer et al., 2018 ). In these two scenarios, specific electron carriers [one reduced ferredoxin and one NAD(P)H in the former and two NAD(P)H in the latter] are necessary to obtain the final product. Final conversion relies on the alcohol dehydrogenase (Adh) activity or the bifunctionality of AdhE (Liew et al., 2017 ). Ultimately, both AOR- and AdhE-based FAOH production should result in reduction in hydrogen evolution if implemented in the chain elongating bacteria as 4 electron equivalents are required for the additional reduction reactions. We incorporated both AOR and AdhE FAOH synthesis pathways in our model and assessed their effects on FAOH production, redox balances, and ATP yield. In chain elongation to FAs, H 2 is evolved as an electron sink to maintain redox balance. We hypothesized that instead of producing H 2 , these electron equivalents could be used to reduce the carboxylate to alcohol. Consistent with this, the model predicted that hexanol (C 6 FAOH) production does not result in H 2 production (Table 2 ), and instead, flux is redirected from HYD1 to RNF to produce NADH required by AOR and AdhE pathways. However, in the AOR pathway, the model predicted a net increase in ATP flux yield through the combined action of IMF and substrate-level phosphorylation as compared to hexanoic acid production ( Supplementary Material S3 ). ECM results show that hexanol production via AOR carries an enzyme cost per ATP flux of 2.1 times that of butyrate (Fig. 4 ), whereas hexanoic acid production requires 2.7 times more than butyrate (Fig. 2 ). This reflects the increased ATP yield from FAOH production with AOR and the prediction that the additional protein needed to reduce hexanoate to hexanol represents a small fraction of the pathway's total protein demand. In the AdhE case, ECM results show that hexanoic acid and hexanol production require approximately equal enzyme investments of 2.7 and 2.6 times the cost of butyrate production, due to bypassing the use of phosphate acetyltransferase (PTA) and acetate kinase (ACK) for alcohol production. In both AOR and AdhE cases, it can be expected that chain elongating bacteria engineered for FAOH production will not experience significant growth rate penalties compared to production of MCFA of the same chain length. In fact, growth rate could theoretically increase if the AOR pathway is implemented due to a reduction of enzyme cost per unit ATP flux (Fig. 6 ). A controlled metabolic engineering intervention would be required to avoid overexpression of heterologous FAOH production genes, which could be detrimental to growth rate, while successfully diverting flux from HYD1 to RNF to simultaneously balance the increased NADH demand and generate a proton motive force. Fig. 6. Top: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate for bioproducts evaluated in this study, as derived from a simplified resource allocation model ( Supplemental Material S1 ). Middle: Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate. Bottom: ATP yield on lactate for different chain elongation simulated by pFBA. Dicarboxylate and Diol Production: Pathways Requiring Oxygen The anaerobic oxidation of MCFAs into diols and dicarboxylic acids poses a significant metabolic engineering challenge in chain elongators, as the lack of oxygen as an electron acceptor results in a considerable decrease in reaction energetics. Methane oxidation is a clear example where the anaerobic pathway needs to be coupled to nitrite or sulphate reduction (Knittel & Boetius, 2009 ). Alkane-degrading microbes can anaerobically oxidize C–H bonds by fixing CO 2 and converting alkanes into FAs. It is hypothesized that this mechanism is initiated by an ethylbenzene dehydrogenase and the whole pathway happens at the great expense of 6 ATP equivalents, suggesting that an unexplored energy-coupling mechanism may take place in these organisms given the low ATP yields in anaerobes ( Supplemental Material S1 , Fig. 2 ) (Heider et al., 2016 ; Shou et al., 2021 ). Alternatively, alkane monooxygenase complexes such as alkane 1-monooxygenase complex (AlkBGT) have been used in the hydroxylation of terminal ω-carbons of different compounds, including FAs (Clomburg et al., 2015 ). However, implementing AlkBGT in anaerobes is currently infeasible due to its requirement for O 2 . To solve these problems, we envision making use of another hypothesized hydroxylation mechanism in anaerobes: the intra-aerobic pathway (Ettwig et al., 2010 ). Achieving such conditions requires “oxygen donors”, such as nitrite, hydrogen peroxide, or perchlorate, whose decomposition provides the required oxygen for AlkBGT. Since oxygen presence can be problematic in this scenario, it is important to maintain its intracellular concentrations low, which could be achieved by tuning nitrite reductase/nitric oxide dismutase (NIR/NOD) and AlkBGT expression or even improving AlkBGT rates through protein engineering. If oxygen levels are still high enough to disrupt cell function, a metabolic strategy that could be explored would be to confine the oxygen generation reaction within microcompartments, pseudo-organelles common in various organisms, including anaerobes, which have been used in several applications, such as protecting cells from toxic intermediates and oxygen (Heinhorst & Cannon, 2020; Kennedy et al., 2021 ). In our proposed pathway, the first module comprises oxygen generation, where nitrite is reduced to nitric oxide by NIR (Wang et al., 2019 ), followed by the action of NOD to yield N 2 and O 2 . The second module encompasses the oxidation reactions, where AlkB would be used to oxidize FAs and FAOHs to ω-hydroxyacids and diols, respectively, using rubredoxin as an electron carrier (Fig. 3B ). The remaining portion of the AlkBGT complex is responsible for regenerating reduced rubredoxin by consuming NADH. Further oxidation of ω-hydroxyacids to dicarboxylic acids, however, would require additional enzymes, such as cinnamyl alcohol dehydrogenase (yjgB) and 6-oxohexanoate dehydrogenase (chnE) (Clomburg et al., 2015 ), even though it is reported that AlkBGT complex can further oxidize hydroxyls to carboxyls and even ester groups, a phenomenon called overoxidation (van Nuland et al., 2017 ). Protein discovery and engineering will play a key role in bringing these pathways to life in the chain elongating bacteria. For example, understanding anaerobic carboxylation mechanisms and discovering other related pathways would circumvent the need of using oxygen and microcompartments. Additionally, to the best of our knowledge, no FAOHs have been used as substrates for AlkBGT, which would require protein characterization and engineering to achieve diol production if intra-aerobic conditions are necessary. The reactions described above were incorporated into our model and the effect on redox balance and ATP yield was assessed for the production of adipic acid and 1,6-hexanediol from hexanoic acid. NO 2 − was supplied in our simulations as a substrate to source O 2 for the hydroxylation step. Nitrite reduction via NIR requires an investment of electron equivalents from NADH. However, the model predicted that the production of adipic acid does not require additional electron equivalents from H 2 nor lactate as H 2 evolution and molar yield of adipic acid from lactate was the same as in the hexanoic acid case. For adipic acid, the ω-hydroxyacid intermediate that is produced by AlkBGT is subsequently oxidized, releasing an NADH that is required for the reduction of NO 2 − and balancing redox. However, in the case of 1,6-hexanediol, an additional electron source is required for the initial reduction of hexanoic acid to hexanol. Thus, diol production is not as carbon efficient because additional electron equivalents from lactate are required to reduce both terminal ends into alcohols, leading to the co-production of acetate to balance carbon, yielding a net increase in ATP yield due to substrate-level phosphorylation through PTA/ACK. ECM found that 1,6-hexanediol carries an enzyme cost lower than that of hexanol produced via AOR, 1.4 times the cost of butyrate. Cost decreases since both pathways share an ATP yield of 0.5, while diol production does away with the CoA transferase (CoAT) reaction. This reaction is itself enzymatically expensive due to its low driving force, and it increases the cost of PTA and ACK reactions as CoAT requires a higher acetate concentration ( Supplemental Material S2 ). Adipic acid production requires 5.4 times more enzyme investment per ATP flux than butyrate (Fig. 4 ). The heightened enzyme cost is largely due to the unfavourable YJGB reaction, in which 6-hydroxyhexanoate reduces NAD + to form 6-oxohexanoate. Minimal enzyme cost is achieved when 6-hydroxyhexanoate concentration is high, driving the YJGB reaction but indirectly increasing the demand for termination enzymes (CoAT, PTA, and ACK) through compensatory heightening of hexanoate and acetate concentrations ( Supplemental Material S2 ). This suggests that it may be more challenging to achieve high adipic acid productivity due to a reduction in growth rate. On the other hand, 1,6-hexanediol production could even increase growth rate vs. hexanoic acid, similarly to hexanol production via AOR (Fig. 6 ). MK Production MKs are another chemical class that can be produced using MCFAs as precursors. Currently, in the engineered RBO cycle, MKs are obtained from the hydrolysis of oxoacyl-CoA molecules by the thioesterase FadM. This reaction initially generates ketoacids, which in turn spontaneously decarboxylate to result in the final MK (Fig. 3C ) (Yan et al., 2020 ). In addition to FadM introduction, other metabolic engineering strategies are necessary to achieve MK production in chain elongating bacteria. The first strategy would be to substitute the native CoATs by FadM, translocating termination from acyl-CoA to oxoacyl-CoA. Since the RBO cycle is essential to redox balance and ATP production in chain elongating bacteria, sole acetone and 2-butanone production would not be possible, as FadM would terminate before one complete turn of the RBO cycle. Therefore, to optimize MK production, it is necessary to allow MCFA with MK co-production. FadM was incorporated into our model to assess the potential to produce 2-pentanone from 6-oxohexanoyl-CoA. The model predicted that 2-pentanone production is not optimal for ATP yield, which aligns with the early termination initiated by FadM at 6-oxohexanoyl-CoA, where the ATP from the complete elongation to hexanoic acid is not produced. To assess the effect of 2-pentanone production on ATP yield, we fixed 2-pentanone production at various values and determined the overall pathway stoichiometry at each point using pFBA with ATP yield set as the objective function (Table 3 ). It was found that while 2-pentanone production is thermodynamically feasible for the scenarios modelled, its production is negatively coupled with butyrate and ATP production (Fig. 5 ). CO 2 production also increases with increasing 2-pentanone production due to the spontaneous decarboxylation of the ketoacid. In isolation, this loss of CO 2 severely impacts the carbon efficiency of this production pathway. However, it could be alleviated if MK production was conducted in the context of a microbial community, where other functional guilds could recycle the CO 2 (Baleeiro et al., 2023 ). Table 3. Selected Predicted Pathway Stoichiometries for 2-Pentanone Production Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Methyl ketone; RBO + FadM Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.000 0.250 −14.99 −59.96 0.000 Lactate + 0.545 H + + 0.183 ADP → 0.03 2pentanone + 0.455 butyrate + 1.03 CO 2 + 1.03 H 2 + 0.183 ATP 0.030 0.183 −12.13 −66.48 0.029 Lactate + 0.575 H + + 0.138 ADP → 0.05 2-pentanone + 0.425 butyrate + 1.05 CO 2 + 1.05 H 2 + 0.138 ATP 0.050 0.138 −10.23 −74.38 0.048 Lactate + 0.605 H + + 0.093 ADP → 0.07 2-pentanone + 0.395 butyrate + 1.07 CO 2 + 1.07 H 2 + 0.093 ATP 0.070 0.093 −8.32 −89.97 0.067 Lactate + 0.665 H + + 0.003 ADP → 0.11 2-pentanone + 0.335 butyrate + 1.11 CO 2 + 1.11 H 2 + 0.003 ATP 0.110 0.003 −4.51 −1804.79 0.105 RBO = reverse beta-oxidation, FadM = oxoacyl-CoA thioesterase"
} | 10,557 |
39090985 | PMC11388927 | pmc | 1,059 | {
"abstract": "Abstract Chain elongating bacteria are a unique guild of strictly anaerobic bacteria that have garnered interest for sustainable chemical manufacturing from carbon-rich wet and gaseous waste streams. They produce C 6 –C 8 medium-chain fatty acids, which are valuable platform chemicals that can be used directly, or derivatized to service a wide range of chemical industries. However, the application of chain elongating bacteria for synthesizing products beyond C 6 –C 8 medium-chain fatty acids has not been evaluated. In this study, we assess the feasibility of expanding the product spectrum of chain elongating bacteria to C 9 –C 12 fatty acids, along with the synthesis of C 6 fatty alcohols, dicarboxylic acids, diols, and methyl ketones. We propose several metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and utilize constraint-based metabolic modelling to predict pathway stoichiometries, assess thermodynamic feasibility, and estimate ATP and product yields. We also evaluate how producing alternative products impacts the growth rate of chain elongating bacteria via resource allocation modelling, revealing a trade-off between product chain length and class versus cell growth rate. Together, these results highlight the potential for using chain elongating bacteria as a platform for diverse oleochemical biomanufacturing and offer a starting point for guiding future metabolic engineering efforts aimed at expanding their product range. One-Sentence Summary In this work, the authors use constraint-based metabolic modelling and enzyme cost minimization to assess the feasibility of using metabolic engineering to expand the product spectrum of anaerobic chain elongating bacteria.",
"conclusion": "Conclusion Developing chain elongating bacteria as a bioproduction platform for oleochemical synthesis could play a crucial role in sustainable manufacturing from carbon-rich waste streams. Our analyses demonstrate that the synthesis of C 9 –C 12 FAs, as well as conversion to their corresponding fatty alcohols, dicarboxylic acids, diols, and MKs, is thermodynamically feasible and generates sufficient ATP production for growth coupling. In the case of C 9 –C 12 FAs, enzyme cost analyses point towards a trade-off between growth rate and ATP yield with increasing chain length, which may explain why chain elongation in nature terminates at chain lengths of C 8 . While some experimental data support our results for selectivity for hexanoate over butyrate production in C. lactatifermentans and C. kluyveri at lower growth rates (Kenealy & Waselefsky, 1985 ; Wang et al., 2022 ), investigating the effect of chain length on growth rate and ATP yield considering other substrates across the diversity of chain elongating bacteria is still needed. Other aspects that need to be investigated to understand limitations in chain length include the effect of product toxicity and enzyme–substrate affinity for longer chain FAs. This includes strategies chain elongators employ to protect against MCFA toxicity and evaluation of the substrate range that can be accommodated by native RBO cycle enzymes. To address these problems, genetic engineering tools need to be developed, including genetic cargo delivery systems, the construction of regulatory element libraries, and gene knock-in/knock-out methods to delete native chain elongating bacteria genes and to enable heterologous gene expression. These genetic modifications should allow the production of a broad range of oleochemicals through chain elongation using the pathways proposed in this work. The production of longer chain length FAs and their subsequent conversion into alcohols are ideal targets for initial demonstrations of oleochemical production using engineered chain elongating bacteria as these products are not redox limited in our analyses. However, the functionalization of the aliphatic end of the FAs required for dicarboxylate and diol production will be a significant challenge due to the anaerobic requirements of chain elongating bacteria. While we present an avenue to accomplish this through nitrite reduction and the use of microcompartments, screening of other enzymes that can functionalize FAs without oxygen will be important to chain elongating bacteria metabolic engineering efforts. Additionally, the production of MKs was found to be feasible, though not growth coupled as it requires carboxylate co-production for ATP production. Given previous demonstrations of MK production in E. coli under microaerobic conditions (Lan et al., 2013 ), other conditions and production pathways should be investigated to improve yields in chain elongators. As a proof of concept, our modelling work focused on the conversion of single carbon substrates to various oleochemicals using chain elongating bacteria in isolation. The deployment of such engineered strains will likely require investigation into how mixed substrates and potential interactions with other microbes will affect product profiles. For example, the presence of more electron-rich substrates, such as glucose, may push fermentation towards less desirable, shorter chain products (e.g. acetate, butyrate) as it is advantageous for growth rather than chain elongation based on the insights from our resource allocation modelling. Overall, our analyses indicate that chain elongating bacteria are a promising biomanufacturing chassis for accessing several oleochemical product classes. This work provides a basis for which bioproducts could be accessed through metabolic engineering of chain elongating bacteria and highlights potential barriers, including trade-offs between growth and product yield, which need to be further investigated. Developing genetic tools for chain elongating bacteria will be critical for validating our modelling predictions, understanding the fundamental physiology of this functional guild, and expanding their product spectrum, in isolation as pure cultures, and in the context of self-assembled and synthetic microbiomes.",
"introduction": "Introduction Circular economies require the recycling of societal “wastes” into new bioproducts to support sustainable human activity. Anaerobic fermentation processes can upcycle carbon from agricultural residues, food waste, and industrial off-gases into useful fuels, chemicals, and materials. One promising process is microbial chain elongation, which uses anaerobic microbiomes in open-culture systems to synthesize medium-chain fatty acids (MCFAs) from complex organic and gaseous waste streams (Angenent et al., 2016 ; Holtzapple et al., 2022 ; Scarborough et al., 2022 ). The process works by producing key intermediates, such as lactate and/or ethanol, through either organic waste fermentation (breaking down) or gas fermentation (building up), which subsequently undergo a secondary fermentation to produce MCFAs via chain elongation. Several studies have demonstrated stable MCFA production from organic waste (Grootscholten et al., 2014 ; Stamatopoulou et al., 2020 ) and gaseous feedstocks (Bäumler et al., 2022 ; Diender et al., 2016 ; Fernández-Blanco et al., 2022 ) via chain elongation at the bench scale and pilot scale, and more recently, a demonstration plant has been built in the Netherlands by the Dutch company ChainCraft ( https://www.chaincraft.nl/ ). At the heart of the process is a functional guild of obligate anaerobes called “chain elongating bacteria” that use a native reverse beta-oxidation (RBO) pathway (see Fig. 1B ) to ferment lactate, ethanol, and other electron-rich organic substrates (e.g. sugars and glycerol) into C 4 –C 8 carboxylates (i.e. butyrate, hexanoate, and octanoate) as part of their growth. The pathway is unique because it allows redox balancing, while also conserving energy through a novel flavin-based electron bifurcation mechanism (Buckel & Thauer, 2018 ; Li et al., 2008 ). Almost all known chain elongators belong to the phylum Bacillota (previously Firmicutes) and are a phylogenetically and physiologically diverse group. While most chain elongators remain uncultivated, over 15 strains have been isolated and sequenced to date, with most new isolates being reported in the last 10 years (Candry & Ganigué, 2021 ). These isolates, particularly Clostridium kluyveri , have been used to establish synthetic co-cultures to convert gaseous or sugar-based feedstocks to MCFAs and their corresponding alcohols (Bäumler et al., 2022 ; Diender et al., 2016 ; Haas et al., 2018 ; Lynd et al., 2022 ; Otten et al., 2022 ) which could be expanded to more complex feedstocks by selecting lactate- and/or ethanol-producing partners with improved hydrolytic capabilities. Fig. 1. (A) Anaerobic conversion of organic waste to oleochemicals using synthetic co-cultures inspired by natural systems. Complex polymers are degraded by lactic acid bacteria (LAB), which in turn provide lactic acid to chain elongating bacteria for medium-chain fatty acid synthesis. (B) The reverse beta-oxidation pathway in chain elongating bacteria and proposed product spectrum expansion (brown boxes). EMP = Embden–Meyerhof–Parnas glycolysis; LDH = electron-confurcating lactate dehydrogenase/electron-transferring flavoprotein; ADH = alcohol dehydrogenase; ADA = acetaldehyde dehydrogenase; PFOR = pyruvate ferredoxin oxidoreductase; PFL = pyruvate formate lyase; PTA = phosphate acetyltransferase; ACK = acetate kinase; ACACT = acetyl-CoA C acetyltransferase; HACD = 3-hydroxyacyl-CoA dehydrogenase; ECOAH = enoyl-CoA hydratase; EBACD = electron bifurcating acyl-CoA dehydrogenase; CoAT = CoA transferase; Fd = ferredoxin; RNF = proton translocating ferredoxin: NAD + oxidoreductase complex; HYD = Fe–Fe hydrogenase. Created with BioRender.com . The development of synthetic co-cultures, inspired by mixed culture chain elongation processes, could represent a platform for improving titres, rates, and yields of MCFA production (Fig. 1A ). Moreover, there is growing interest in genetically modifying the native RBO cycle in chain elongators to expand the product spectrum beyond C 4 –C 8 monocarboxylates, including alcohols, diols, dicarboxylates, and other bulk chemicals (Agena et al., 2023 ; Guss & Riley, 2021 ; Strik et al., 2022 ). This would enable the anaerobic synthesis of diverse medium-chain oleochemicals from wet and gaseous waste streams, which is expected to have improved economic feasibility compared to sugar-based aerobic fermentations (Holtzapple et al., 2022 ). However, the technical feasibility of making products other than C 4 –C 8 monocarboxylates via metabolic engineering of chain elongators remains unexplored. As tools to genetically modify chain elongators emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), methodologies to rationally design these new biocatalysts are needed. In this study, we use constraint-based metabolic modelling along with thermodynamic analyses to evaluate the feasibility of synthesizing diverse medium-chain oleochemicals (C 6 –C 12 fatty acids [FAs], primary alcohols, dicarboxylates, diols, and methyl ketones [MKs]) from key intermediate substrates (lactate, ethanol, sugars, and glycerol) using anaerobic chain elongating bacteria (Fig. 1 ). We first analyse MCFA production scenarios beyond natural C 8 production up to C 12 , highlighting trade-offs between ATP yield and expected growth rate using a combination of metabolic modelling and enzyme cost minimization (ECM) analyses. Subsequently, we propose several modifications to the RBO cycle to synthesize target medium-chain oleochemical products and use constraint-based metabolic modelling to determine overall pathway stoichiometry, thermodynamic feasibility, and theoretical product yields. Our results indicate that metabolic engineering of chain elongating bacteria could enable the anaerobic synthesis of diverse medium-chain oleochemicals at industrially relevant yields. Moreover, we identify challenges with engineering the RBO pathway in chain elongators and offer potential solutions to overcome them via metabolic engineering. We anticipate that these results will provide a useful starting point for engineering microbial chain elongation to serve as a platform for sustainable chemical manufacturing.",
"discussion": "Results and Discussion Impact of MCFA Chain Length on Growth Rate and ATP Yield Modelling Chain Elongation Beyond Octanoic Acid Currently, the products of chain elongation are restricted to butyric and hexanoic acids (C 4 –C 6 ), with limited synthesis of octanoic acid (C 8 ) (Nelson et al., 2017 ; Zhu et al., 2017 ). Genetically modifying chain elongating bacteria to improve selectivity and increase MCFA chain length could benefit product yields, while also expanding the process to produce MCFAs with larger markets (C 8 –C 12 ). To assess the feasibility of MCFA production beyond C 8 , we predicted overall pathway stoichiometry, redox balance, free energy change, and theoretical ATP and product yields of C 4 –C 12 FAs via pFBA with ATP yield set as the objective function using a simplified metabolic model describing core chain elongation metabolism (iFermCell193, see the “Methods” section) (Table 1 ). We focused on lactate utilizing chain elongating bacteria (e.g. Pseudoramibacter alactolyticus, Megasphaera hexanoica, Caproicibacterium lactatifermentans ) as lactate is the primary substrate in mixed culture processes converting organic wastes (Fig. 1 ) (Contreras-Dávila et al., 2020 ). Model simulations with different electron donors, including ethanol, glucose, xylose, and glycerol, were also evaluated ( Supplemental Material S1 , Tables S1 – S5 ) to demonstrate process feasibility for a range of other substrates known to be consumed by chain elongating bacteria (e.g. ethanol by Clostridium kluyveri ). Table 1. Predicted Stoichiometry for the Synthesis of C 4 –C 12 FAs from Lactate Chain length Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol lactate) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) C 4 Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.500 0.250 −14.99 −59.96 0.489 C 6 Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 C 8 Lactate + 0.75 H + + 0.375 ADP → CO 2 + 0.5 H 2 + 0.5 H 2 O + 0.25 octanoate + 0.375 ATP 0.250 0.375 −35.10 −93.60 0.402 C 10 Lactate + 0.8 H + + 0.4 ADP → CO 2 + 0.2 decanoate + 0.4 H 2 + 0.6 H 2 O + 0.4 ATP 0.200 0.400 −40.11 −100.29 0.385 C 12 Lactate + 0.833 H + + 0.417 ADP → CO 2 + 0.167 dodecanoate + 0.333 H 2 + 0.667 H 2 O + 0.417 ATP 0.167 0.417 −43.45 −104.29 0.373 The model predicted that ATP yield and overall reaction free energy change increase with MCFA chain length from C 4 to C 12 (Table 1 ), indicating that MCFA synthesis beyond C 8 could be theoretically possible. Moreover, the production of C 4 –C 12 chain lengths from lactate released ≥50 kJ/mol per ATP, indicating that these reactions produce sufficient energy for ATP synthesis (Thauer et al., 1977 ). C 9 –C 12 carboxylate production with other electron donors, including ethanol, glucose, xylose, and glycerol, was also found to be feasible ( Supplemental Material S1 , Tables S1 – S5 ). This observation agrees with past modelling results that suggested hexanoic and octanoic acid production should improve ATP yields from lactate compared to butyric acid production (Scarborough et al., 2018 ). The predicted increase in net ATP production is attributed to a greater flux through the proton translocating ferredoxin: NAD + oxidoreductase complex (RNF) for each turn of the RBO cycle, which leads to greater ATP production via the ion motive force (IMF). In the RBO cycle, each successive elongation generates reduced ferredoxin from the electron-bifurcating acyl-CoA dehydrogenase (EBACD). This ferredoxin goes on to drive RNF to recover some NADH and/or is used by a soluble ferredoxin hydrogenase (HYD1) to evolve H 2 as a terminal electron sink. As confirmed in the predicted flux distributions, elongation to longer chain lengths increases the total NADH demand required to reduce the acyl chains (HACD and EBACD in Fig. 1 ). Thus, with increasing chain length, a greater proportion of the reduced ferredoxin produced by EBACD is used to regenerate NADH via RNF, rather than driving HYD1. This ultimately leads to increased ATP production through the IMF along with decreased H 2 evolution for longer chain lengths. These results also indicate that past a certain chain length (>C 12 ), H 2 evolution will cease as all electron equivalents will be required solely for chain elongation. Intracellular Thermodynamic Landscape and Resource Allocation Our initial modelling analysis suggested that the improved ATP yield with longer chain lengths should naturally select for MCFAs beyond C 8 . However, all pure and mixed culture chain elongation studies have only observed C 4 –C 8 MCFA products. This points to a potential trade-off between growth yield (or ATP yield) and growth rate (or ATP production rate), resulting from optimal resource allocation. In a model based on resource allocation theory ( Supplementary Material S1 ; inspired by a similar model developed by Flamholz et al., 2024 ), where catabolic ATP production and anabolic ATP consumption rates are linearly related to the biomass fractions allocated to either function, RBO enzyme cost per unit ATP production flux and chain elongating bacteria growth rate are inversely correlated (Fig. 2 ). This simplified model captures the fact that a higher growth rate is expected to necessitate both a larger pool of anabolic enzymes and ribosomes (Basan, 2018 ) and a higher ATP production flux, requiring that catabolic machinery produces ATP faster with less enzyme. Fig. 2. ECM results of RBO for production of C 4 –C 12 carboxylates from lactate and the estimated relative effect on growth rate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Each reaction carries a cost which accounts for the pathway's stoichiometry (bottom bar), increased cost incurred due to a reaction's proximity to equilibrium (middle bar), and the cost due to sub-saturation reactant concentrations (top bar). Bottom right panel: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate, as derived from a simplified resource allocation model (Supplemental Material S1) (Top). Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate (Middle). ATP yield on lactate for different chain elongation simulated by pFBA (Bottom). To evaluate the enzyme investment per ATP flux required to produce carboxylates of different chain lengths, we performed ECM analysis. ECM places a lower bound on a pathway's enzyme demand per unit flux [g/(mol/h)] by accounting for the pathway's length and stoichiometry, reaction thermodynamics, and saturation effects following from the optimal set of metabolite concentrations, given appropriate bounds on those concentrations and parametrization of enzyme kinetics (Flamholz et al., 2013 ). This analysis coupled with ATP yield results from our model showed that minimal enzyme cost per unit ATP flux increases with chain length (Fig. 2 ). This suggests that chain elongating bacteria making longer length products invest more of their proteome into high yield catabolism, at the trade-off of having less catalytic capacity for growth. These modelling predictions align with experimental results indicating that increased selectivity and production for hexanoic acid over butyrate is associated with a decrease in growth rate in C. lactatifermentans when grown on lactate as compared to glucose (Wang et al., 2022 ). Moreover, caproate selectivity has been shown to increase with hydraulic retention time (reduced growth rate) in continuous cultures of C. kluyveri and a chain elongation microbiome (Grootscholten et al., 2013 ; Kenealy & Waselefsky, 1985 ). According to the optimized free energy changes predicted by ECM, thiolase reactions (ACACT) are the least exergonic in RBO ( Supplemental Material S1 , Fig. 1 ). Since these reactions operate near equilibrium, they have a smaller forward-to-reverse flux ratio (Noor et al., 2014 ) and therefore require a larger enzyme pool to sustain a given net flux through the pathway. The set of metabolite concentrations that minimizes pathway enzyme cost fixes the thiolase products (oxoacyl-CoAs) at low concentrations to drive these reactions ( Supplemental Material S2 ). This heightens the demand for the proceeding enzymes in the pathway, 3-hydroxyacyl-CoA dehydrogenases (HACDs) since they are far from saturation (Fig. 2 ). Similarly, enoyl-CoA hydratase (ECOAH) reactions are thermodynamically constrained, meaning their reactant and product concentrations are kept high and low, respectively ( Supplemental Material S2 ). This further increases the protein burden of HACD as well as the proceeding electron bifurcating acyl-CoA dehydrogenase (EBACD) (Fig. 2 ). At longer chain lengths, ECOAH reactions are predicted to be more favourable ( Supplemental Material S1 , Fig. 1 ), causing the rate at which enzyme cost increases with chain length to decrease. Interestingly, when oxoacyl-CoA concentrations are constrained by a lower bound of 1 µM, as is the case for all other metabolites, products beyond C 4 are deemed infeasible due to endergonic subsequent thiolase reactions. To reflect observations of chain elongation beyond C 4 , oxoacyl-CoAs can fall to sub-micromolar levels in this model. An alternative solution, as previously suggested in a similar instance with the citric acid cycle (Noor et al., 2014 ), is to posit enzymatic channelling. Channelling between ACACT and HACD would reduce the protein cost of both reactions by effectively merging the two into one favourable reaction. Indeed, channelling between ACACT, HACD, and ECOAH has been described with crystal structures of bacterial and human beta-oxidation complexes (Ishikawa et al., 2004 ; Xia et al., 2019 ). So long as a similar mechanism is active from C 4 to C 12 , the trend of enzyme cost per ATP flux increasing with chain length is expected to hold. This has major implications for engineering chain elongating bacteria to produce MCFAs beyond C8, as the production of longer chain length would provide more ATP per mole of electron donor, but at the cost of a lower growth rate. As a result, compensatory efforts to bolster growth rate, for example through adaptive laboratory evolution (Sandberg et al., 2019 ), may be needed to achieve higher productivity. Expanding and Controlling Chain Elongation Products with Metabolic Engineering Metabolic engineering strategies for the expansion of RBO-derived products have been reviewed for non-chain elongating, model organisms with established genetic tools, such as Escherichia coli, Saccharomyces cerevisiae , and select acetogenic Clostridium species (Tarasava et al., 2022 ). The approach for these strains relies on the engineered reversal of β-oxidation accomplished through extensive strain engineering such as knockout or repression of native β-oxidation regulators and overexpression of genes from chain elongating bacteria and other oleaginous hosts (Tarasava et al., 2022 ). Most of these hosts do not natively rely on the RBO cycle for redox balancing or energy conservation, which is what makes the RBO cycle growth coupled in the chain elongating bacteria. Further, installing orthogonal systems for the engineered reversal of β-oxidation imposes an increased metabolic burden on non-chain elongating hosts as high expression is likely needed to ensure sufficient flux through the pathway (Liu et al., 2018 ). This can lead to redox limitations or inefficient substrate use for chain elongation in these non-chain elongating hosts, and approaches such as two-phase production are likely needed to maximize product yields (Lange et al., 2016 ). Production with non-chain elongating hosts will likely require further strain engineering to improve product yields and selectivity. As an alternative to engineering model organisms, tools to genetically modify chain elongating bacteria are beginning to emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), which will open the door to expanded opportunities to produce oleochemicals via anaerobic fermentation. Chain elongating bacteria are ideal hosts for RBO-based bioproduction as the RBO cycle is innately growth coupled in these strains (no reconstruction needed) and as members of anaerobic communities can utilize solid and gaseous waste feedstocks rather than relying on refined substrates (e.g. sugars). Further, product yields in anaerobes are typically much greater as they do not synthesize as much biomass compared to aerobic systems and most electrons end up in products (Cueto-Rojas et al., 2015 ). However, the feasibility of pathways for the production of other oleochemicals, including, fatty alcohols, dicarboxylates, diols, and MKs, has not been explored in anaerobic chain elongating bacteria. Here, we propose potential production pathways and enzymes required to generate a wide range of oleochemicals (Fig. 3 ) and use pFBA and ECM to assess the feasibility and potential effects on ATP yield and growth rate. Fig. 3. Proposed pathways for chain elongating bacteria product spectrum expansion. (A) Fatty alcohols pathway. AdhE = bifunctional aldehyde-alcohol dehydrogenase; AOR = aldehyde: ferredoxin oxidoreductase; FadM = acyl-CoA thioesterase. (B) Diol and dicarboxylic acids pathways. NIR = nitrite reductase; NOD = nitric oxide dismutase; AlkBGT = alkane 1-monooxygenase complex; yjgB = cinnamyl alcohol dehydrogenase; chnE = 6-oxohexanoate dehydrogenase; Rd = rubredoxin. (C) Methyl ketone pathway. FadM: thioesterase. Created with Biorender.com . iFermCell193 was further modified to capture the reactions required for hexanoic acid conversion to other products (iFermCell356). Table 2 summarizes the potential pathway stoichiometries for the conversion of hexanoic acid (C 6 ) to the corresponding fatty alcohol, dicarboxylate, and diol using lactate as an electron donor, along with the net ATP yields and overall Gibbs free energy change of reaction for each substrate–product pair as predicted by our model. All predicted pathway stoichiometries have overall favourable standard Gibbs free energies of reaction and release ≥50 kJ/mol per ATP required for ATP production. The model also predicts that net ATP yield is not negatively impacted by conversion into these other bioproducts, which indicates that their production can be growth coupled, similar to MCFAs. In the following sections, we describe potential metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and discuss specific shifts in redox requirements that occur predicted by the model, along with the impact on the metabolic cost of the additional enzymes estimated by ECM (Fig. 4 ). The production of methyl ketones was also modelled, but it was found to negatively impact ATP yield (Fig. 5 ). Fig. 4. ECM results of RBO for production of novel chain elongation products from lactate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Fig. 5. Production profiles for the co-production of 2-pentanone and butyrate from lactate and standard Gibbs free energy of reaction. Molar yields of ATP are negatively related to increasing production of 2-pentanone. Co-production of butyrate is required with increasing 2-pentanone production. Table 2. Predicted Pathway Stoichiometry for Different C 6 Oleochemicals Produced from Lactate Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol Lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Carboxylate; RBO Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 Alcohol; RBO + AdhE Lactate + H + + 0.333 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.333 ATP 0.333 0.333 −45.18 −135.53 0.382 Alcohol; RBO + AOR Lactate + H + + 0.5 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.5 ATP 0.333 0.500 −45.18 −90.35 0.382 Dicarboxylate; RBO + AlkBGT Lactate + H + + 0.667 NO 2 − + 0.333 ADP → 0.333 adipate + CO 2 + 0.667 H 2 + H 2 O + 0.333 N 2 + 0.333 ATP 0.333 0.333 −228.21 −684.62 0.539 Diol; RBO + AlkBGT Lactate + 0.5 NO 2 − + 1.25 H + + 0.5 ADP → 0.25 1,6-hexanediol + 0.25 acetate + CO 2 + H 2 O + 0.25 N2 + 0.5 ATP 0.250 0.500 −191.63 −383.26 0.332 RBO = reverse beta-oxidation, AdhE = bifunctional alcohol dehydrogenase, AOR = aldehyde:ferredoxin oxidoreductase, AlkBGT = alkane 1-monooxygenase complex Fatty Alcohol Production Medium-chain fatty alcohols (FAOHs) could be produced using MCFAs as building blocks. Even though FAOH production pathways (Fig. 3A ) exist in anaerobic bacteria, such as acetogens and butanol-producing clostridia (Moon et al., 2016 ), this has yet to be observed in most chain elongating bacteria. In nature, acetogenic bacteria use aldehyde:ferredoxin oxidoreductase (AOR) to convert FAs into fatty aldehydes in monoculture or when they are co-cultured with chain elongating bacteria (Benito-Vaquerizo et al., 2020 ; Diender et al., 2016 ). In another instance, aldehyde-alcohol dehydrogenase (AdhE) can convert acyl-CoAs into aldehydes (Mehrer et al., 2018 ). In these two scenarios, specific electron carriers [one reduced ferredoxin and one NAD(P)H in the former and two NAD(P)H in the latter] are necessary to obtain the final product. Final conversion relies on the alcohol dehydrogenase (Adh) activity or the bifunctionality of AdhE (Liew et al., 2017 ). Ultimately, both AOR- and AdhE-based FAOH production should result in reduction in hydrogen evolution if implemented in the chain elongating bacteria as 4 electron equivalents are required for the additional reduction reactions. We incorporated both AOR and AdhE FAOH synthesis pathways in our model and assessed their effects on FAOH production, redox balances, and ATP yield. In chain elongation to FAs, H 2 is evolved as an electron sink to maintain redox balance. We hypothesized that instead of producing H 2 , these electron equivalents could be used to reduce the carboxylate to alcohol. Consistent with this, the model predicted that hexanol (C 6 FAOH) production does not result in H 2 production (Table 2 ), and instead, flux is redirected from HYD1 to RNF to produce NADH required by AOR and AdhE pathways. However, in the AOR pathway, the model predicted a net increase in ATP flux yield through the combined action of IMF and substrate-level phosphorylation as compared to hexanoic acid production ( Supplementary Material S3 ). ECM results show that hexanol production via AOR carries an enzyme cost per ATP flux of 2.1 times that of butyrate (Fig. 4 ), whereas hexanoic acid production requires 2.7 times more than butyrate (Fig. 2 ). This reflects the increased ATP yield from FAOH production with AOR and the prediction that the additional protein needed to reduce hexanoate to hexanol represents a small fraction of the pathway's total protein demand. In the AdhE case, ECM results show that hexanoic acid and hexanol production require approximately equal enzyme investments of 2.7 and 2.6 times the cost of butyrate production, due to bypassing the use of phosphate acetyltransferase (PTA) and acetate kinase (ACK) for alcohol production. In both AOR and AdhE cases, it can be expected that chain elongating bacteria engineered for FAOH production will not experience significant growth rate penalties compared to production of MCFA of the same chain length. In fact, growth rate could theoretically increase if the AOR pathway is implemented due to a reduction of enzyme cost per unit ATP flux (Fig. 6 ). A controlled metabolic engineering intervention would be required to avoid overexpression of heterologous FAOH production genes, which could be detrimental to growth rate, while successfully diverting flux from HYD1 to RNF to simultaneously balance the increased NADH demand and generate a proton motive force. Fig. 6. Top: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate for bioproducts evaluated in this study, as derived from a simplified resource allocation model ( Supplemental Material S1 ). Middle: Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate. Bottom: ATP yield on lactate for different chain elongation simulated by pFBA. Dicarboxylate and Diol Production: Pathways Requiring Oxygen The anaerobic oxidation of MCFAs into diols and dicarboxylic acids poses a significant metabolic engineering challenge in chain elongators, as the lack of oxygen as an electron acceptor results in a considerable decrease in reaction energetics. Methane oxidation is a clear example where the anaerobic pathway needs to be coupled to nitrite or sulphate reduction (Knittel & Boetius, 2009 ). Alkane-degrading microbes can anaerobically oxidize C–H bonds by fixing CO 2 and converting alkanes into FAs. It is hypothesized that this mechanism is initiated by an ethylbenzene dehydrogenase and the whole pathway happens at the great expense of 6 ATP equivalents, suggesting that an unexplored energy-coupling mechanism may take place in these organisms given the low ATP yields in anaerobes ( Supplemental Material S1 , Fig. 2 ) (Heider et al., 2016 ; Shou et al., 2021 ). Alternatively, alkane monooxygenase complexes such as alkane 1-monooxygenase complex (AlkBGT) have been used in the hydroxylation of terminal ω-carbons of different compounds, including FAs (Clomburg et al., 2015 ). However, implementing AlkBGT in anaerobes is currently infeasible due to its requirement for O 2 . To solve these problems, we envision making use of another hypothesized hydroxylation mechanism in anaerobes: the intra-aerobic pathway (Ettwig et al., 2010 ). Achieving such conditions requires “oxygen donors”, such as nitrite, hydrogen peroxide, or perchlorate, whose decomposition provides the required oxygen for AlkBGT. Since oxygen presence can be problematic in this scenario, it is important to maintain its intracellular concentrations low, which could be achieved by tuning nitrite reductase/nitric oxide dismutase (NIR/NOD) and AlkBGT expression or even improving AlkBGT rates through protein engineering. If oxygen levels are still high enough to disrupt cell function, a metabolic strategy that could be explored would be to confine the oxygen generation reaction within microcompartments, pseudo-organelles common in various organisms, including anaerobes, which have been used in several applications, such as protecting cells from toxic intermediates and oxygen (Heinhorst & Cannon, 2020; Kennedy et al., 2021 ). In our proposed pathway, the first module comprises oxygen generation, where nitrite is reduced to nitric oxide by NIR (Wang et al., 2019 ), followed by the action of NOD to yield N 2 and O 2 . The second module encompasses the oxidation reactions, where AlkB would be used to oxidize FAs and FAOHs to ω-hydroxyacids and diols, respectively, using rubredoxin as an electron carrier (Fig. 3B ). The remaining portion of the AlkBGT complex is responsible for regenerating reduced rubredoxin by consuming NADH. Further oxidation of ω-hydroxyacids to dicarboxylic acids, however, would require additional enzymes, such as cinnamyl alcohol dehydrogenase (yjgB) and 6-oxohexanoate dehydrogenase (chnE) (Clomburg et al., 2015 ), even though it is reported that AlkBGT complex can further oxidize hydroxyls to carboxyls and even ester groups, a phenomenon called overoxidation (van Nuland et al., 2017 ). Protein discovery and engineering will play a key role in bringing these pathways to life in the chain elongating bacteria. For example, understanding anaerobic carboxylation mechanisms and discovering other related pathways would circumvent the need of using oxygen and microcompartments. Additionally, to the best of our knowledge, no FAOHs have been used as substrates for AlkBGT, which would require protein characterization and engineering to achieve diol production if intra-aerobic conditions are necessary. The reactions described above were incorporated into our model and the effect on redox balance and ATP yield was assessed for the production of adipic acid and 1,6-hexanediol from hexanoic acid. NO 2 − was supplied in our simulations as a substrate to source O 2 for the hydroxylation step. Nitrite reduction via NIR requires an investment of electron equivalents from NADH. However, the model predicted that the production of adipic acid does not require additional electron equivalents from H 2 nor lactate as H 2 evolution and molar yield of adipic acid from lactate was the same as in the hexanoic acid case. For adipic acid, the ω-hydroxyacid intermediate that is produced by AlkBGT is subsequently oxidized, releasing an NADH that is required for the reduction of NO 2 − and balancing redox. However, in the case of 1,6-hexanediol, an additional electron source is required for the initial reduction of hexanoic acid to hexanol. Thus, diol production is not as carbon efficient because additional electron equivalents from lactate are required to reduce both terminal ends into alcohols, leading to the co-production of acetate to balance carbon, yielding a net increase in ATP yield due to substrate-level phosphorylation through PTA/ACK. ECM found that 1,6-hexanediol carries an enzyme cost lower than that of hexanol produced via AOR, 1.4 times the cost of butyrate. Cost decreases since both pathways share an ATP yield of 0.5, while diol production does away with the CoA transferase (CoAT) reaction. This reaction is itself enzymatically expensive due to its low driving force, and it increases the cost of PTA and ACK reactions as CoAT requires a higher acetate concentration ( Supplemental Material S2 ). Adipic acid production requires 5.4 times more enzyme investment per ATP flux than butyrate (Fig. 4 ). The heightened enzyme cost is largely due to the unfavourable YJGB reaction, in which 6-hydroxyhexanoate reduces NAD + to form 6-oxohexanoate. Minimal enzyme cost is achieved when 6-hydroxyhexanoate concentration is high, driving the YJGB reaction but indirectly increasing the demand for termination enzymes (CoAT, PTA, and ACK) through compensatory heightening of hexanoate and acetate concentrations ( Supplemental Material S2 ). This suggests that it may be more challenging to achieve high adipic acid productivity due to a reduction in growth rate. On the other hand, 1,6-hexanediol production could even increase growth rate vs. hexanoic acid, similarly to hexanol production via AOR (Fig. 6 ). MK Production MKs are another chemical class that can be produced using MCFAs as precursors. Currently, in the engineered RBO cycle, MKs are obtained from the hydrolysis of oxoacyl-CoA molecules by the thioesterase FadM. This reaction initially generates ketoacids, which in turn spontaneously decarboxylate to result in the final MK (Fig. 3C ) (Yan et al., 2020 ). In addition to FadM introduction, other metabolic engineering strategies are necessary to achieve MK production in chain elongating bacteria. The first strategy would be to substitute the native CoATs by FadM, translocating termination from acyl-CoA to oxoacyl-CoA. Since the RBO cycle is essential to redox balance and ATP production in chain elongating bacteria, sole acetone and 2-butanone production would not be possible, as FadM would terminate before one complete turn of the RBO cycle. Therefore, to optimize MK production, it is necessary to allow MCFA with MK co-production. FadM was incorporated into our model to assess the potential to produce 2-pentanone from 6-oxohexanoyl-CoA. The model predicted that 2-pentanone production is not optimal for ATP yield, which aligns with the early termination initiated by FadM at 6-oxohexanoyl-CoA, where the ATP from the complete elongation to hexanoic acid is not produced. To assess the effect of 2-pentanone production on ATP yield, we fixed 2-pentanone production at various values and determined the overall pathway stoichiometry at each point using pFBA with ATP yield set as the objective function (Table 3 ). It was found that while 2-pentanone production is thermodynamically feasible for the scenarios modelled, its production is negatively coupled with butyrate and ATP production (Fig. 5 ). CO 2 production also increases with increasing 2-pentanone production due to the spontaneous decarboxylation of the ketoacid. In isolation, this loss of CO 2 severely impacts the carbon efficiency of this production pathway. However, it could be alleviated if MK production was conducted in the context of a microbial community, where other functional guilds could recycle the CO 2 (Baleeiro et al., 2023 ). Table 3. Selected Predicted Pathway Stoichiometries for 2-Pentanone Production Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Methyl ketone; RBO + FadM Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.000 0.250 −14.99 −59.96 0.000 Lactate + 0.545 H + + 0.183 ADP → 0.03 2pentanone + 0.455 butyrate + 1.03 CO 2 + 1.03 H 2 + 0.183 ATP 0.030 0.183 −12.13 −66.48 0.029 Lactate + 0.575 H + + 0.138 ADP → 0.05 2-pentanone + 0.425 butyrate + 1.05 CO 2 + 1.05 H 2 + 0.138 ATP 0.050 0.138 −10.23 −74.38 0.048 Lactate + 0.605 H + + 0.093 ADP → 0.07 2-pentanone + 0.395 butyrate + 1.07 CO 2 + 1.07 H 2 + 0.093 ATP 0.070 0.093 −8.32 −89.97 0.067 Lactate + 0.665 H + + 0.003 ADP → 0.11 2-pentanone + 0.335 butyrate + 1.11 CO 2 + 1.11 H 2 + 0.003 ATP 0.110 0.003 −4.51 −1804.79 0.105 RBO = reverse beta-oxidation, FadM = oxoacyl-CoA thioesterase"
} | 10,557 |
39090985 | PMC11388927 | pmc | 1,060 | {
"abstract": "Abstract Chain elongating bacteria are a unique guild of strictly anaerobic bacteria that have garnered interest for sustainable chemical manufacturing from carbon-rich wet and gaseous waste streams. They produce C 6 –C 8 medium-chain fatty acids, which are valuable platform chemicals that can be used directly, or derivatized to service a wide range of chemical industries. However, the application of chain elongating bacteria for synthesizing products beyond C 6 –C 8 medium-chain fatty acids has not been evaluated. In this study, we assess the feasibility of expanding the product spectrum of chain elongating bacteria to C 9 –C 12 fatty acids, along with the synthesis of C 6 fatty alcohols, dicarboxylic acids, diols, and methyl ketones. We propose several metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and utilize constraint-based metabolic modelling to predict pathway stoichiometries, assess thermodynamic feasibility, and estimate ATP and product yields. We also evaluate how producing alternative products impacts the growth rate of chain elongating bacteria via resource allocation modelling, revealing a trade-off between product chain length and class versus cell growth rate. Together, these results highlight the potential for using chain elongating bacteria as a platform for diverse oleochemical biomanufacturing and offer a starting point for guiding future metabolic engineering efforts aimed at expanding their product range. One-Sentence Summary In this work, the authors use constraint-based metabolic modelling and enzyme cost minimization to assess the feasibility of using metabolic engineering to expand the product spectrum of anaerobic chain elongating bacteria.",
"conclusion": "Conclusion Developing chain elongating bacteria as a bioproduction platform for oleochemical synthesis could play a crucial role in sustainable manufacturing from carbon-rich waste streams. Our analyses demonstrate that the synthesis of C 9 –C 12 FAs, as well as conversion to their corresponding fatty alcohols, dicarboxylic acids, diols, and MKs, is thermodynamically feasible and generates sufficient ATP production for growth coupling. In the case of C 9 –C 12 FAs, enzyme cost analyses point towards a trade-off between growth rate and ATP yield with increasing chain length, which may explain why chain elongation in nature terminates at chain lengths of C 8 . While some experimental data support our results for selectivity for hexanoate over butyrate production in C. lactatifermentans and C. kluyveri at lower growth rates (Kenealy & Waselefsky, 1985 ; Wang et al., 2022 ), investigating the effect of chain length on growth rate and ATP yield considering other substrates across the diversity of chain elongating bacteria is still needed. Other aspects that need to be investigated to understand limitations in chain length include the effect of product toxicity and enzyme–substrate affinity for longer chain FAs. This includes strategies chain elongators employ to protect against MCFA toxicity and evaluation of the substrate range that can be accommodated by native RBO cycle enzymes. To address these problems, genetic engineering tools need to be developed, including genetic cargo delivery systems, the construction of regulatory element libraries, and gene knock-in/knock-out methods to delete native chain elongating bacteria genes and to enable heterologous gene expression. These genetic modifications should allow the production of a broad range of oleochemicals through chain elongation using the pathways proposed in this work. The production of longer chain length FAs and their subsequent conversion into alcohols are ideal targets for initial demonstrations of oleochemical production using engineered chain elongating bacteria as these products are not redox limited in our analyses. However, the functionalization of the aliphatic end of the FAs required for dicarboxylate and diol production will be a significant challenge due to the anaerobic requirements of chain elongating bacteria. While we present an avenue to accomplish this through nitrite reduction and the use of microcompartments, screening of other enzymes that can functionalize FAs without oxygen will be important to chain elongating bacteria metabolic engineering efforts. Additionally, the production of MKs was found to be feasible, though not growth coupled as it requires carboxylate co-production for ATP production. Given previous demonstrations of MK production in E. coli under microaerobic conditions (Lan et al., 2013 ), other conditions and production pathways should be investigated to improve yields in chain elongators. As a proof of concept, our modelling work focused on the conversion of single carbon substrates to various oleochemicals using chain elongating bacteria in isolation. The deployment of such engineered strains will likely require investigation into how mixed substrates and potential interactions with other microbes will affect product profiles. For example, the presence of more electron-rich substrates, such as glucose, may push fermentation towards less desirable, shorter chain products (e.g. acetate, butyrate) as it is advantageous for growth rather than chain elongation based on the insights from our resource allocation modelling. Overall, our analyses indicate that chain elongating bacteria are a promising biomanufacturing chassis for accessing several oleochemical product classes. This work provides a basis for which bioproducts could be accessed through metabolic engineering of chain elongating bacteria and highlights potential barriers, including trade-offs between growth and product yield, which need to be further investigated. Developing genetic tools for chain elongating bacteria will be critical for validating our modelling predictions, understanding the fundamental physiology of this functional guild, and expanding their product spectrum, in isolation as pure cultures, and in the context of self-assembled and synthetic microbiomes.",
"introduction": "Introduction Circular economies require the recycling of societal “wastes” into new bioproducts to support sustainable human activity. Anaerobic fermentation processes can upcycle carbon from agricultural residues, food waste, and industrial off-gases into useful fuels, chemicals, and materials. One promising process is microbial chain elongation, which uses anaerobic microbiomes in open-culture systems to synthesize medium-chain fatty acids (MCFAs) from complex organic and gaseous waste streams (Angenent et al., 2016 ; Holtzapple et al., 2022 ; Scarborough et al., 2022 ). The process works by producing key intermediates, such as lactate and/or ethanol, through either organic waste fermentation (breaking down) or gas fermentation (building up), which subsequently undergo a secondary fermentation to produce MCFAs via chain elongation. Several studies have demonstrated stable MCFA production from organic waste (Grootscholten et al., 2014 ; Stamatopoulou et al., 2020 ) and gaseous feedstocks (Bäumler et al., 2022 ; Diender et al., 2016 ; Fernández-Blanco et al., 2022 ) via chain elongation at the bench scale and pilot scale, and more recently, a demonstration plant has been built in the Netherlands by the Dutch company ChainCraft ( https://www.chaincraft.nl/ ). At the heart of the process is a functional guild of obligate anaerobes called “chain elongating bacteria” that use a native reverse beta-oxidation (RBO) pathway (see Fig. 1B ) to ferment lactate, ethanol, and other electron-rich organic substrates (e.g. sugars and glycerol) into C 4 –C 8 carboxylates (i.e. butyrate, hexanoate, and octanoate) as part of their growth. The pathway is unique because it allows redox balancing, while also conserving energy through a novel flavin-based electron bifurcation mechanism (Buckel & Thauer, 2018 ; Li et al., 2008 ). Almost all known chain elongators belong to the phylum Bacillota (previously Firmicutes) and are a phylogenetically and physiologically diverse group. While most chain elongators remain uncultivated, over 15 strains have been isolated and sequenced to date, with most new isolates being reported in the last 10 years (Candry & Ganigué, 2021 ). These isolates, particularly Clostridium kluyveri , have been used to establish synthetic co-cultures to convert gaseous or sugar-based feedstocks to MCFAs and their corresponding alcohols (Bäumler et al., 2022 ; Diender et al., 2016 ; Haas et al., 2018 ; Lynd et al., 2022 ; Otten et al., 2022 ) which could be expanded to more complex feedstocks by selecting lactate- and/or ethanol-producing partners with improved hydrolytic capabilities. Fig. 1. (A) Anaerobic conversion of organic waste to oleochemicals using synthetic co-cultures inspired by natural systems. Complex polymers are degraded by lactic acid bacteria (LAB), which in turn provide lactic acid to chain elongating bacteria for medium-chain fatty acid synthesis. (B) The reverse beta-oxidation pathway in chain elongating bacteria and proposed product spectrum expansion (brown boxes). EMP = Embden–Meyerhof–Parnas glycolysis; LDH = electron-confurcating lactate dehydrogenase/electron-transferring flavoprotein; ADH = alcohol dehydrogenase; ADA = acetaldehyde dehydrogenase; PFOR = pyruvate ferredoxin oxidoreductase; PFL = pyruvate formate lyase; PTA = phosphate acetyltransferase; ACK = acetate kinase; ACACT = acetyl-CoA C acetyltransferase; HACD = 3-hydroxyacyl-CoA dehydrogenase; ECOAH = enoyl-CoA hydratase; EBACD = electron bifurcating acyl-CoA dehydrogenase; CoAT = CoA transferase; Fd = ferredoxin; RNF = proton translocating ferredoxin: NAD + oxidoreductase complex; HYD = Fe–Fe hydrogenase. Created with BioRender.com . The development of synthetic co-cultures, inspired by mixed culture chain elongation processes, could represent a platform for improving titres, rates, and yields of MCFA production (Fig. 1A ). Moreover, there is growing interest in genetically modifying the native RBO cycle in chain elongators to expand the product spectrum beyond C 4 –C 8 monocarboxylates, including alcohols, diols, dicarboxylates, and other bulk chemicals (Agena et al., 2023 ; Guss & Riley, 2021 ; Strik et al., 2022 ). This would enable the anaerobic synthesis of diverse medium-chain oleochemicals from wet and gaseous waste streams, which is expected to have improved economic feasibility compared to sugar-based aerobic fermentations (Holtzapple et al., 2022 ). However, the technical feasibility of making products other than C 4 –C 8 monocarboxylates via metabolic engineering of chain elongators remains unexplored. As tools to genetically modify chain elongators emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), methodologies to rationally design these new biocatalysts are needed. In this study, we use constraint-based metabolic modelling along with thermodynamic analyses to evaluate the feasibility of synthesizing diverse medium-chain oleochemicals (C 6 –C 12 fatty acids [FAs], primary alcohols, dicarboxylates, diols, and methyl ketones [MKs]) from key intermediate substrates (lactate, ethanol, sugars, and glycerol) using anaerobic chain elongating bacteria (Fig. 1 ). We first analyse MCFA production scenarios beyond natural C 8 production up to C 12 , highlighting trade-offs between ATP yield and expected growth rate using a combination of metabolic modelling and enzyme cost minimization (ECM) analyses. Subsequently, we propose several modifications to the RBO cycle to synthesize target medium-chain oleochemical products and use constraint-based metabolic modelling to determine overall pathway stoichiometry, thermodynamic feasibility, and theoretical product yields. Our results indicate that metabolic engineering of chain elongating bacteria could enable the anaerobic synthesis of diverse medium-chain oleochemicals at industrially relevant yields. Moreover, we identify challenges with engineering the RBO pathway in chain elongators and offer potential solutions to overcome them via metabolic engineering. We anticipate that these results will provide a useful starting point for engineering microbial chain elongation to serve as a platform for sustainable chemical manufacturing.",
"discussion": "Results and Discussion Impact of MCFA Chain Length on Growth Rate and ATP Yield Modelling Chain Elongation Beyond Octanoic Acid Currently, the products of chain elongation are restricted to butyric and hexanoic acids (C 4 –C 6 ), with limited synthesis of octanoic acid (C 8 ) (Nelson et al., 2017 ; Zhu et al., 2017 ). Genetically modifying chain elongating bacteria to improve selectivity and increase MCFA chain length could benefit product yields, while also expanding the process to produce MCFAs with larger markets (C 8 –C 12 ). To assess the feasibility of MCFA production beyond C 8 , we predicted overall pathway stoichiometry, redox balance, free energy change, and theoretical ATP and product yields of C 4 –C 12 FAs via pFBA with ATP yield set as the objective function using a simplified metabolic model describing core chain elongation metabolism (iFermCell193, see the “Methods” section) (Table 1 ). We focused on lactate utilizing chain elongating bacteria (e.g. Pseudoramibacter alactolyticus, Megasphaera hexanoica, Caproicibacterium lactatifermentans ) as lactate is the primary substrate in mixed culture processes converting organic wastes (Fig. 1 ) (Contreras-Dávila et al., 2020 ). Model simulations with different electron donors, including ethanol, glucose, xylose, and glycerol, were also evaluated ( Supplemental Material S1 , Tables S1 – S5 ) to demonstrate process feasibility for a range of other substrates known to be consumed by chain elongating bacteria (e.g. ethanol by Clostridium kluyveri ). Table 1. Predicted Stoichiometry for the Synthesis of C 4 –C 12 FAs from Lactate Chain length Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol lactate) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) C 4 Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.500 0.250 −14.99 −59.96 0.489 C 6 Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 C 8 Lactate + 0.75 H + + 0.375 ADP → CO 2 + 0.5 H 2 + 0.5 H 2 O + 0.25 octanoate + 0.375 ATP 0.250 0.375 −35.10 −93.60 0.402 C 10 Lactate + 0.8 H + + 0.4 ADP → CO 2 + 0.2 decanoate + 0.4 H 2 + 0.6 H 2 O + 0.4 ATP 0.200 0.400 −40.11 −100.29 0.385 C 12 Lactate + 0.833 H + + 0.417 ADP → CO 2 + 0.167 dodecanoate + 0.333 H 2 + 0.667 H 2 O + 0.417 ATP 0.167 0.417 −43.45 −104.29 0.373 The model predicted that ATP yield and overall reaction free energy change increase with MCFA chain length from C 4 to C 12 (Table 1 ), indicating that MCFA synthesis beyond C 8 could be theoretically possible. Moreover, the production of C 4 –C 12 chain lengths from lactate released ≥50 kJ/mol per ATP, indicating that these reactions produce sufficient energy for ATP synthesis (Thauer et al., 1977 ). C 9 –C 12 carboxylate production with other electron donors, including ethanol, glucose, xylose, and glycerol, was also found to be feasible ( Supplemental Material S1 , Tables S1 – S5 ). This observation agrees with past modelling results that suggested hexanoic and octanoic acid production should improve ATP yields from lactate compared to butyric acid production (Scarborough et al., 2018 ). The predicted increase in net ATP production is attributed to a greater flux through the proton translocating ferredoxin: NAD + oxidoreductase complex (RNF) for each turn of the RBO cycle, which leads to greater ATP production via the ion motive force (IMF). In the RBO cycle, each successive elongation generates reduced ferredoxin from the electron-bifurcating acyl-CoA dehydrogenase (EBACD). This ferredoxin goes on to drive RNF to recover some NADH and/or is used by a soluble ferredoxin hydrogenase (HYD1) to evolve H 2 as a terminal electron sink. As confirmed in the predicted flux distributions, elongation to longer chain lengths increases the total NADH demand required to reduce the acyl chains (HACD and EBACD in Fig. 1 ). Thus, with increasing chain length, a greater proportion of the reduced ferredoxin produced by EBACD is used to regenerate NADH via RNF, rather than driving HYD1. This ultimately leads to increased ATP production through the IMF along with decreased H 2 evolution for longer chain lengths. These results also indicate that past a certain chain length (>C 12 ), H 2 evolution will cease as all electron equivalents will be required solely for chain elongation. Intracellular Thermodynamic Landscape and Resource Allocation Our initial modelling analysis suggested that the improved ATP yield with longer chain lengths should naturally select for MCFAs beyond C 8 . However, all pure and mixed culture chain elongation studies have only observed C 4 –C 8 MCFA products. This points to a potential trade-off between growth yield (or ATP yield) and growth rate (or ATP production rate), resulting from optimal resource allocation. In a model based on resource allocation theory ( Supplementary Material S1 ; inspired by a similar model developed by Flamholz et al., 2024 ), where catabolic ATP production and anabolic ATP consumption rates are linearly related to the biomass fractions allocated to either function, RBO enzyme cost per unit ATP production flux and chain elongating bacteria growth rate are inversely correlated (Fig. 2 ). This simplified model captures the fact that a higher growth rate is expected to necessitate both a larger pool of anabolic enzymes and ribosomes (Basan, 2018 ) and a higher ATP production flux, requiring that catabolic machinery produces ATP faster with less enzyme. Fig. 2. ECM results of RBO for production of C 4 –C 12 carboxylates from lactate and the estimated relative effect on growth rate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Each reaction carries a cost which accounts for the pathway's stoichiometry (bottom bar), increased cost incurred due to a reaction's proximity to equilibrium (middle bar), and the cost due to sub-saturation reactant concentrations (top bar). Bottom right panel: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate, as derived from a simplified resource allocation model (Supplemental Material S1) (Top). Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate (Middle). ATP yield on lactate for different chain elongation simulated by pFBA (Bottom). To evaluate the enzyme investment per ATP flux required to produce carboxylates of different chain lengths, we performed ECM analysis. ECM places a lower bound on a pathway's enzyme demand per unit flux [g/(mol/h)] by accounting for the pathway's length and stoichiometry, reaction thermodynamics, and saturation effects following from the optimal set of metabolite concentrations, given appropriate bounds on those concentrations and parametrization of enzyme kinetics (Flamholz et al., 2013 ). This analysis coupled with ATP yield results from our model showed that minimal enzyme cost per unit ATP flux increases with chain length (Fig. 2 ). This suggests that chain elongating bacteria making longer length products invest more of their proteome into high yield catabolism, at the trade-off of having less catalytic capacity for growth. These modelling predictions align with experimental results indicating that increased selectivity and production for hexanoic acid over butyrate is associated with a decrease in growth rate in C. lactatifermentans when grown on lactate as compared to glucose (Wang et al., 2022 ). Moreover, caproate selectivity has been shown to increase with hydraulic retention time (reduced growth rate) in continuous cultures of C. kluyveri and a chain elongation microbiome (Grootscholten et al., 2013 ; Kenealy & Waselefsky, 1985 ). According to the optimized free energy changes predicted by ECM, thiolase reactions (ACACT) are the least exergonic in RBO ( Supplemental Material S1 , Fig. 1 ). Since these reactions operate near equilibrium, they have a smaller forward-to-reverse flux ratio (Noor et al., 2014 ) and therefore require a larger enzyme pool to sustain a given net flux through the pathway. The set of metabolite concentrations that minimizes pathway enzyme cost fixes the thiolase products (oxoacyl-CoAs) at low concentrations to drive these reactions ( Supplemental Material S2 ). This heightens the demand for the proceeding enzymes in the pathway, 3-hydroxyacyl-CoA dehydrogenases (HACDs) since they are far from saturation (Fig. 2 ). Similarly, enoyl-CoA hydratase (ECOAH) reactions are thermodynamically constrained, meaning their reactant and product concentrations are kept high and low, respectively ( Supplemental Material S2 ). This further increases the protein burden of HACD as well as the proceeding electron bifurcating acyl-CoA dehydrogenase (EBACD) (Fig. 2 ). At longer chain lengths, ECOAH reactions are predicted to be more favourable ( Supplemental Material S1 , Fig. 1 ), causing the rate at which enzyme cost increases with chain length to decrease. Interestingly, when oxoacyl-CoA concentrations are constrained by a lower bound of 1 µM, as is the case for all other metabolites, products beyond C 4 are deemed infeasible due to endergonic subsequent thiolase reactions. To reflect observations of chain elongation beyond C 4 , oxoacyl-CoAs can fall to sub-micromolar levels in this model. An alternative solution, as previously suggested in a similar instance with the citric acid cycle (Noor et al., 2014 ), is to posit enzymatic channelling. Channelling between ACACT and HACD would reduce the protein cost of both reactions by effectively merging the two into one favourable reaction. Indeed, channelling between ACACT, HACD, and ECOAH has been described with crystal structures of bacterial and human beta-oxidation complexes (Ishikawa et al., 2004 ; Xia et al., 2019 ). So long as a similar mechanism is active from C 4 to C 12 , the trend of enzyme cost per ATP flux increasing with chain length is expected to hold. This has major implications for engineering chain elongating bacteria to produce MCFAs beyond C8, as the production of longer chain length would provide more ATP per mole of electron donor, but at the cost of a lower growth rate. As a result, compensatory efforts to bolster growth rate, for example through adaptive laboratory evolution (Sandberg et al., 2019 ), may be needed to achieve higher productivity. Expanding and Controlling Chain Elongation Products with Metabolic Engineering Metabolic engineering strategies for the expansion of RBO-derived products have been reviewed for non-chain elongating, model organisms with established genetic tools, such as Escherichia coli, Saccharomyces cerevisiae , and select acetogenic Clostridium species (Tarasava et al., 2022 ). The approach for these strains relies on the engineered reversal of β-oxidation accomplished through extensive strain engineering such as knockout or repression of native β-oxidation regulators and overexpression of genes from chain elongating bacteria and other oleaginous hosts (Tarasava et al., 2022 ). Most of these hosts do not natively rely on the RBO cycle for redox balancing or energy conservation, which is what makes the RBO cycle growth coupled in the chain elongating bacteria. Further, installing orthogonal systems for the engineered reversal of β-oxidation imposes an increased metabolic burden on non-chain elongating hosts as high expression is likely needed to ensure sufficient flux through the pathway (Liu et al., 2018 ). This can lead to redox limitations or inefficient substrate use for chain elongation in these non-chain elongating hosts, and approaches such as two-phase production are likely needed to maximize product yields (Lange et al., 2016 ). Production with non-chain elongating hosts will likely require further strain engineering to improve product yields and selectivity. As an alternative to engineering model organisms, tools to genetically modify chain elongating bacteria are beginning to emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), which will open the door to expanded opportunities to produce oleochemicals via anaerobic fermentation. Chain elongating bacteria are ideal hosts for RBO-based bioproduction as the RBO cycle is innately growth coupled in these strains (no reconstruction needed) and as members of anaerobic communities can utilize solid and gaseous waste feedstocks rather than relying on refined substrates (e.g. sugars). Further, product yields in anaerobes are typically much greater as they do not synthesize as much biomass compared to aerobic systems and most electrons end up in products (Cueto-Rojas et al., 2015 ). However, the feasibility of pathways for the production of other oleochemicals, including, fatty alcohols, dicarboxylates, diols, and MKs, has not been explored in anaerobic chain elongating bacteria. Here, we propose potential production pathways and enzymes required to generate a wide range of oleochemicals (Fig. 3 ) and use pFBA and ECM to assess the feasibility and potential effects on ATP yield and growth rate. Fig. 3. Proposed pathways for chain elongating bacteria product spectrum expansion. (A) Fatty alcohols pathway. AdhE = bifunctional aldehyde-alcohol dehydrogenase; AOR = aldehyde: ferredoxin oxidoreductase; FadM = acyl-CoA thioesterase. (B) Diol and dicarboxylic acids pathways. NIR = nitrite reductase; NOD = nitric oxide dismutase; AlkBGT = alkane 1-monooxygenase complex; yjgB = cinnamyl alcohol dehydrogenase; chnE = 6-oxohexanoate dehydrogenase; Rd = rubredoxin. (C) Methyl ketone pathway. FadM: thioesterase. Created with Biorender.com . iFermCell193 was further modified to capture the reactions required for hexanoic acid conversion to other products (iFermCell356). Table 2 summarizes the potential pathway stoichiometries for the conversion of hexanoic acid (C 6 ) to the corresponding fatty alcohol, dicarboxylate, and diol using lactate as an electron donor, along with the net ATP yields and overall Gibbs free energy change of reaction for each substrate–product pair as predicted by our model. All predicted pathway stoichiometries have overall favourable standard Gibbs free energies of reaction and release ≥50 kJ/mol per ATP required for ATP production. The model also predicts that net ATP yield is not negatively impacted by conversion into these other bioproducts, which indicates that their production can be growth coupled, similar to MCFAs. In the following sections, we describe potential metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and discuss specific shifts in redox requirements that occur predicted by the model, along with the impact on the metabolic cost of the additional enzymes estimated by ECM (Fig. 4 ). The production of methyl ketones was also modelled, but it was found to negatively impact ATP yield (Fig. 5 ). Fig. 4. ECM results of RBO for production of novel chain elongation products from lactate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Fig. 5. Production profiles for the co-production of 2-pentanone and butyrate from lactate and standard Gibbs free energy of reaction. Molar yields of ATP are negatively related to increasing production of 2-pentanone. Co-production of butyrate is required with increasing 2-pentanone production. Table 2. Predicted Pathway Stoichiometry for Different C 6 Oleochemicals Produced from Lactate Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol Lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Carboxylate; RBO Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 Alcohol; RBO + AdhE Lactate + H + + 0.333 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.333 ATP 0.333 0.333 −45.18 −135.53 0.382 Alcohol; RBO + AOR Lactate + H + + 0.5 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.5 ATP 0.333 0.500 −45.18 −90.35 0.382 Dicarboxylate; RBO + AlkBGT Lactate + H + + 0.667 NO 2 − + 0.333 ADP → 0.333 adipate + CO 2 + 0.667 H 2 + H 2 O + 0.333 N 2 + 0.333 ATP 0.333 0.333 −228.21 −684.62 0.539 Diol; RBO + AlkBGT Lactate + 0.5 NO 2 − + 1.25 H + + 0.5 ADP → 0.25 1,6-hexanediol + 0.25 acetate + CO 2 + H 2 O + 0.25 N2 + 0.5 ATP 0.250 0.500 −191.63 −383.26 0.332 RBO = reverse beta-oxidation, AdhE = bifunctional alcohol dehydrogenase, AOR = aldehyde:ferredoxin oxidoreductase, AlkBGT = alkane 1-monooxygenase complex Fatty Alcohol Production Medium-chain fatty alcohols (FAOHs) could be produced using MCFAs as building blocks. Even though FAOH production pathways (Fig. 3A ) exist in anaerobic bacteria, such as acetogens and butanol-producing clostridia (Moon et al., 2016 ), this has yet to be observed in most chain elongating bacteria. In nature, acetogenic bacteria use aldehyde:ferredoxin oxidoreductase (AOR) to convert FAs into fatty aldehydes in monoculture or when they are co-cultured with chain elongating bacteria (Benito-Vaquerizo et al., 2020 ; Diender et al., 2016 ). In another instance, aldehyde-alcohol dehydrogenase (AdhE) can convert acyl-CoAs into aldehydes (Mehrer et al., 2018 ). In these two scenarios, specific electron carriers [one reduced ferredoxin and one NAD(P)H in the former and two NAD(P)H in the latter] are necessary to obtain the final product. Final conversion relies on the alcohol dehydrogenase (Adh) activity or the bifunctionality of AdhE (Liew et al., 2017 ). Ultimately, both AOR- and AdhE-based FAOH production should result in reduction in hydrogen evolution if implemented in the chain elongating bacteria as 4 electron equivalents are required for the additional reduction reactions. We incorporated both AOR and AdhE FAOH synthesis pathways in our model and assessed their effects on FAOH production, redox balances, and ATP yield. In chain elongation to FAs, H 2 is evolved as an electron sink to maintain redox balance. We hypothesized that instead of producing H 2 , these electron equivalents could be used to reduce the carboxylate to alcohol. Consistent with this, the model predicted that hexanol (C 6 FAOH) production does not result in H 2 production (Table 2 ), and instead, flux is redirected from HYD1 to RNF to produce NADH required by AOR and AdhE pathways. However, in the AOR pathway, the model predicted a net increase in ATP flux yield through the combined action of IMF and substrate-level phosphorylation as compared to hexanoic acid production ( Supplementary Material S3 ). ECM results show that hexanol production via AOR carries an enzyme cost per ATP flux of 2.1 times that of butyrate (Fig. 4 ), whereas hexanoic acid production requires 2.7 times more than butyrate (Fig. 2 ). This reflects the increased ATP yield from FAOH production with AOR and the prediction that the additional protein needed to reduce hexanoate to hexanol represents a small fraction of the pathway's total protein demand. In the AdhE case, ECM results show that hexanoic acid and hexanol production require approximately equal enzyme investments of 2.7 and 2.6 times the cost of butyrate production, due to bypassing the use of phosphate acetyltransferase (PTA) and acetate kinase (ACK) for alcohol production. In both AOR and AdhE cases, it can be expected that chain elongating bacteria engineered for FAOH production will not experience significant growth rate penalties compared to production of MCFA of the same chain length. In fact, growth rate could theoretically increase if the AOR pathway is implemented due to a reduction of enzyme cost per unit ATP flux (Fig. 6 ). A controlled metabolic engineering intervention would be required to avoid overexpression of heterologous FAOH production genes, which could be detrimental to growth rate, while successfully diverting flux from HYD1 to RNF to simultaneously balance the increased NADH demand and generate a proton motive force. Fig. 6. Top: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate for bioproducts evaluated in this study, as derived from a simplified resource allocation model ( Supplemental Material S1 ). Middle: Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate. Bottom: ATP yield on lactate for different chain elongation simulated by pFBA. Dicarboxylate and Diol Production: Pathways Requiring Oxygen The anaerobic oxidation of MCFAs into diols and dicarboxylic acids poses a significant metabolic engineering challenge in chain elongators, as the lack of oxygen as an electron acceptor results in a considerable decrease in reaction energetics. Methane oxidation is a clear example where the anaerobic pathway needs to be coupled to nitrite or sulphate reduction (Knittel & Boetius, 2009 ). Alkane-degrading microbes can anaerobically oxidize C–H bonds by fixing CO 2 and converting alkanes into FAs. It is hypothesized that this mechanism is initiated by an ethylbenzene dehydrogenase and the whole pathway happens at the great expense of 6 ATP equivalents, suggesting that an unexplored energy-coupling mechanism may take place in these organisms given the low ATP yields in anaerobes ( Supplemental Material S1 , Fig. 2 ) (Heider et al., 2016 ; Shou et al., 2021 ). Alternatively, alkane monooxygenase complexes such as alkane 1-monooxygenase complex (AlkBGT) have been used in the hydroxylation of terminal ω-carbons of different compounds, including FAs (Clomburg et al., 2015 ). However, implementing AlkBGT in anaerobes is currently infeasible due to its requirement for O 2 . To solve these problems, we envision making use of another hypothesized hydroxylation mechanism in anaerobes: the intra-aerobic pathway (Ettwig et al., 2010 ). Achieving such conditions requires “oxygen donors”, such as nitrite, hydrogen peroxide, or perchlorate, whose decomposition provides the required oxygen for AlkBGT. Since oxygen presence can be problematic in this scenario, it is important to maintain its intracellular concentrations low, which could be achieved by tuning nitrite reductase/nitric oxide dismutase (NIR/NOD) and AlkBGT expression or even improving AlkBGT rates through protein engineering. If oxygen levels are still high enough to disrupt cell function, a metabolic strategy that could be explored would be to confine the oxygen generation reaction within microcompartments, pseudo-organelles common in various organisms, including anaerobes, which have been used in several applications, such as protecting cells from toxic intermediates and oxygen (Heinhorst & Cannon, 2020; Kennedy et al., 2021 ). In our proposed pathway, the first module comprises oxygen generation, where nitrite is reduced to nitric oxide by NIR (Wang et al., 2019 ), followed by the action of NOD to yield N 2 and O 2 . The second module encompasses the oxidation reactions, where AlkB would be used to oxidize FAs and FAOHs to ω-hydroxyacids and diols, respectively, using rubredoxin as an electron carrier (Fig. 3B ). The remaining portion of the AlkBGT complex is responsible for regenerating reduced rubredoxin by consuming NADH. Further oxidation of ω-hydroxyacids to dicarboxylic acids, however, would require additional enzymes, such as cinnamyl alcohol dehydrogenase (yjgB) and 6-oxohexanoate dehydrogenase (chnE) (Clomburg et al., 2015 ), even though it is reported that AlkBGT complex can further oxidize hydroxyls to carboxyls and even ester groups, a phenomenon called overoxidation (van Nuland et al., 2017 ). Protein discovery and engineering will play a key role in bringing these pathways to life in the chain elongating bacteria. For example, understanding anaerobic carboxylation mechanisms and discovering other related pathways would circumvent the need of using oxygen and microcompartments. Additionally, to the best of our knowledge, no FAOHs have been used as substrates for AlkBGT, which would require protein characterization and engineering to achieve diol production if intra-aerobic conditions are necessary. The reactions described above were incorporated into our model and the effect on redox balance and ATP yield was assessed for the production of adipic acid and 1,6-hexanediol from hexanoic acid. NO 2 − was supplied in our simulations as a substrate to source O 2 for the hydroxylation step. Nitrite reduction via NIR requires an investment of electron equivalents from NADH. However, the model predicted that the production of adipic acid does not require additional electron equivalents from H 2 nor lactate as H 2 evolution and molar yield of adipic acid from lactate was the same as in the hexanoic acid case. For adipic acid, the ω-hydroxyacid intermediate that is produced by AlkBGT is subsequently oxidized, releasing an NADH that is required for the reduction of NO 2 − and balancing redox. However, in the case of 1,6-hexanediol, an additional electron source is required for the initial reduction of hexanoic acid to hexanol. Thus, diol production is not as carbon efficient because additional electron equivalents from lactate are required to reduce both terminal ends into alcohols, leading to the co-production of acetate to balance carbon, yielding a net increase in ATP yield due to substrate-level phosphorylation through PTA/ACK. ECM found that 1,6-hexanediol carries an enzyme cost lower than that of hexanol produced via AOR, 1.4 times the cost of butyrate. Cost decreases since both pathways share an ATP yield of 0.5, while diol production does away with the CoA transferase (CoAT) reaction. This reaction is itself enzymatically expensive due to its low driving force, and it increases the cost of PTA and ACK reactions as CoAT requires a higher acetate concentration ( Supplemental Material S2 ). Adipic acid production requires 5.4 times more enzyme investment per ATP flux than butyrate (Fig. 4 ). The heightened enzyme cost is largely due to the unfavourable YJGB reaction, in which 6-hydroxyhexanoate reduces NAD + to form 6-oxohexanoate. Minimal enzyme cost is achieved when 6-hydroxyhexanoate concentration is high, driving the YJGB reaction but indirectly increasing the demand for termination enzymes (CoAT, PTA, and ACK) through compensatory heightening of hexanoate and acetate concentrations ( Supplemental Material S2 ). This suggests that it may be more challenging to achieve high adipic acid productivity due to a reduction in growth rate. On the other hand, 1,6-hexanediol production could even increase growth rate vs. hexanoic acid, similarly to hexanol production via AOR (Fig. 6 ). MK Production MKs are another chemical class that can be produced using MCFAs as precursors. Currently, in the engineered RBO cycle, MKs are obtained from the hydrolysis of oxoacyl-CoA molecules by the thioesterase FadM. This reaction initially generates ketoacids, which in turn spontaneously decarboxylate to result in the final MK (Fig. 3C ) (Yan et al., 2020 ). In addition to FadM introduction, other metabolic engineering strategies are necessary to achieve MK production in chain elongating bacteria. The first strategy would be to substitute the native CoATs by FadM, translocating termination from acyl-CoA to oxoacyl-CoA. Since the RBO cycle is essential to redox balance and ATP production in chain elongating bacteria, sole acetone and 2-butanone production would not be possible, as FadM would terminate before one complete turn of the RBO cycle. Therefore, to optimize MK production, it is necessary to allow MCFA with MK co-production. FadM was incorporated into our model to assess the potential to produce 2-pentanone from 6-oxohexanoyl-CoA. The model predicted that 2-pentanone production is not optimal for ATP yield, which aligns with the early termination initiated by FadM at 6-oxohexanoyl-CoA, where the ATP from the complete elongation to hexanoic acid is not produced. To assess the effect of 2-pentanone production on ATP yield, we fixed 2-pentanone production at various values and determined the overall pathway stoichiometry at each point using pFBA with ATP yield set as the objective function (Table 3 ). It was found that while 2-pentanone production is thermodynamically feasible for the scenarios modelled, its production is negatively coupled with butyrate and ATP production (Fig. 5 ). CO 2 production also increases with increasing 2-pentanone production due to the spontaneous decarboxylation of the ketoacid. In isolation, this loss of CO 2 severely impacts the carbon efficiency of this production pathway. However, it could be alleviated if MK production was conducted in the context of a microbial community, where other functional guilds could recycle the CO 2 (Baleeiro et al., 2023 ). Table 3. Selected Predicted Pathway Stoichiometries for 2-Pentanone Production Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Methyl ketone; RBO + FadM Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.000 0.250 −14.99 −59.96 0.000 Lactate + 0.545 H + + 0.183 ADP → 0.03 2pentanone + 0.455 butyrate + 1.03 CO 2 + 1.03 H 2 + 0.183 ATP 0.030 0.183 −12.13 −66.48 0.029 Lactate + 0.575 H + + 0.138 ADP → 0.05 2-pentanone + 0.425 butyrate + 1.05 CO 2 + 1.05 H 2 + 0.138 ATP 0.050 0.138 −10.23 −74.38 0.048 Lactate + 0.605 H + + 0.093 ADP → 0.07 2-pentanone + 0.395 butyrate + 1.07 CO 2 + 1.07 H 2 + 0.093 ATP 0.070 0.093 −8.32 −89.97 0.067 Lactate + 0.665 H + + 0.003 ADP → 0.11 2-pentanone + 0.335 butyrate + 1.11 CO 2 + 1.11 H 2 + 0.003 ATP 0.110 0.003 −4.51 −1804.79 0.105 RBO = reverse beta-oxidation, FadM = oxoacyl-CoA thioesterase"
} | 10,557 |
39090985 | PMC11388927 | pmc | 1,060 | {
"abstract": "Abstract Chain elongating bacteria are a unique guild of strictly anaerobic bacteria that have garnered interest for sustainable chemical manufacturing from carbon-rich wet and gaseous waste streams. They produce C 6 –C 8 medium-chain fatty acids, which are valuable platform chemicals that can be used directly, or derivatized to service a wide range of chemical industries. However, the application of chain elongating bacteria for synthesizing products beyond C 6 –C 8 medium-chain fatty acids has not been evaluated. In this study, we assess the feasibility of expanding the product spectrum of chain elongating bacteria to C 9 –C 12 fatty acids, along with the synthesis of C 6 fatty alcohols, dicarboxylic acids, diols, and methyl ketones. We propose several metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and utilize constraint-based metabolic modelling to predict pathway stoichiometries, assess thermodynamic feasibility, and estimate ATP and product yields. We also evaluate how producing alternative products impacts the growth rate of chain elongating bacteria via resource allocation modelling, revealing a trade-off between product chain length and class versus cell growth rate. Together, these results highlight the potential for using chain elongating bacteria as a platform for diverse oleochemical biomanufacturing and offer a starting point for guiding future metabolic engineering efforts aimed at expanding their product range. One-Sentence Summary In this work, the authors use constraint-based metabolic modelling and enzyme cost minimization to assess the feasibility of using metabolic engineering to expand the product spectrum of anaerobic chain elongating bacteria.",
"conclusion": "Conclusion Developing chain elongating bacteria as a bioproduction platform for oleochemical synthesis could play a crucial role in sustainable manufacturing from carbon-rich waste streams. Our analyses demonstrate that the synthesis of C 9 –C 12 FAs, as well as conversion to their corresponding fatty alcohols, dicarboxylic acids, diols, and MKs, is thermodynamically feasible and generates sufficient ATP production for growth coupling. In the case of C 9 –C 12 FAs, enzyme cost analyses point towards a trade-off between growth rate and ATP yield with increasing chain length, which may explain why chain elongation in nature terminates at chain lengths of C 8 . While some experimental data support our results for selectivity for hexanoate over butyrate production in C. lactatifermentans and C. kluyveri at lower growth rates (Kenealy & Waselefsky, 1985 ; Wang et al., 2022 ), investigating the effect of chain length on growth rate and ATP yield considering other substrates across the diversity of chain elongating bacteria is still needed. Other aspects that need to be investigated to understand limitations in chain length include the effect of product toxicity and enzyme–substrate affinity for longer chain FAs. This includes strategies chain elongators employ to protect against MCFA toxicity and evaluation of the substrate range that can be accommodated by native RBO cycle enzymes. To address these problems, genetic engineering tools need to be developed, including genetic cargo delivery systems, the construction of regulatory element libraries, and gene knock-in/knock-out methods to delete native chain elongating bacteria genes and to enable heterologous gene expression. These genetic modifications should allow the production of a broad range of oleochemicals through chain elongation using the pathways proposed in this work. The production of longer chain length FAs and their subsequent conversion into alcohols are ideal targets for initial demonstrations of oleochemical production using engineered chain elongating bacteria as these products are not redox limited in our analyses. However, the functionalization of the aliphatic end of the FAs required for dicarboxylate and diol production will be a significant challenge due to the anaerobic requirements of chain elongating bacteria. While we present an avenue to accomplish this through nitrite reduction and the use of microcompartments, screening of other enzymes that can functionalize FAs without oxygen will be important to chain elongating bacteria metabolic engineering efforts. Additionally, the production of MKs was found to be feasible, though not growth coupled as it requires carboxylate co-production for ATP production. Given previous demonstrations of MK production in E. coli under microaerobic conditions (Lan et al., 2013 ), other conditions and production pathways should be investigated to improve yields in chain elongators. As a proof of concept, our modelling work focused on the conversion of single carbon substrates to various oleochemicals using chain elongating bacteria in isolation. The deployment of such engineered strains will likely require investigation into how mixed substrates and potential interactions with other microbes will affect product profiles. For example, the presence of more electron-rich substrates, such as glucose, may push fermentation towards less desirable, shorter chain products (e.g. acetate, butyrate) as it is advantageous for growth rather than chain elongation based on the insights from our resource allocation modelling. Overall, our analyses indicate that chain elongating bacteria are a promising biomanufacturing chassis for accessing several oleochemical product classes. This work provides a basis for which bioproducts could be accessed through metabolic engineering of chain elongating bacteria and highlights potential barriers, including trade-offs between growth and product yield, which need to be further investigated. Developing genetic tools for chain elongating bacteria will be critical for validating our modelling predictions, understanding the fundamental physiology of this functional guild, and expanding their product spectrum, in isolation as pure cultures, and in the context of self-assembled and synthetic microbiomes.",
"introduction": "Introduction Circular economies require the recycling of societal “wastes” into new bioproducts to support sustainable human activity. Anaerobic fermentation processes can upcycle carbon from agricultural residues, food waste, and industrial off-gases into useful fuels, chemicals, and materials. One promising process is microbial chain elongation, which uses anaerobic microbiomes in open-culture systems to synthesize medium-chain fatty acids (MCFAs) from complex organic and gaseous waste streams (Angenent et al., 2016 ; Holtzapple et al., 2022 ; Scarborough et al., 2022 ). The process works by producing key intermediates, such as lactate and/or ethanol, through either organic waste fermentation (breaking down) or gas fermentation (building up), which subsequently undergo a secondary fermentation to produce MCFAs via chain elongation. Several studies have demonstrated stable MCFA production from organic waste (Grootscholten et al., 2014 ; Stamatopoulou et al., 2020 ) and gaseous feedstocks (Bäumler et al., 2022 ; Diender et al., 2016 ; Fernández-Blanco et al., 2022 ) via chain elongation at the bench scale and pilot scale, and more recently, a demonstration plant has been built in the Netherlands by the Dutch company ChainCraft ( https://www.chaincraft.nl/ ). At the heart of the process is a functional guild of obligate anaerobes called “chain elongating bacteria” that use a native reverse beta-oxidation (RBO) pathway (see Fig. 1B ) to ferment lactate, ethanol, and other electron-rich organic substrates (e.g. sugars and glycerol) into C 4 –C 8 carboxylates (i.e. butyrate, hexanoate, and octanoate) as part of their growth. The pathway is unique because it allows redox balancing, while also conserving energy through a novel flavin-based electron bifurcation mechanism (Buckel & Thauer, 2018 ; Li et al., 2008 ). Almost all known chain elongators belong to the phylum Bacillota (previously Firmicutes) and are a phylogenetically and physiologically diverse group. While most chain elongators remain uncultivated, over 15 strains have been isolated and sequenced to date, with most new isolates being reported in the last 10 years (Candry & Ganigué, 2021 ). These isolates, particularly Clostridium kluyveri , have been used to establish synthetic co-cultures to convert gaseous or sugar-based feedstocks to MCFAs and their corresponding alcohols (Bäumler et al., 2022 ; Diender et al., 2016 ; Haas et al., 2018 ; Lynd et al., 2022 ; Otten et al., 2022 ) which could be expanded to more complex feedstocks by selecting lactate- and/or ethanol-producing partners with improved hydrolytic capabilities. Fig. 1. (A) Anaerobic conversion of organic waste to oleochemicals using synthetic co-cultures inspired by natural systems. Complex polymers are degraded by lactic acid bacteria (LAB), which in turn provide lactic acid to chain elongating bacteria for medium-chain fatty acid synthesis. (B) The reverse beta-oxidation pathway in chain elongating bacteria and proposed product spectrum expansion (brown boxes). EMP = Embden–Meyerhof–Parnas glycolysis; LDH = electron-confurcating lactate dehydrogenase/electron-transferring flavoprotein; ADH = alcohol dehydrogenase; ADA = acetaldehyde dehydrogenase; PFOR = pyruvate ferredoxin oxidoreductase; PFL = pyruvate formate lyase; PTA = phosphate acetyltransferase; ACK = acetate kinase; ACACT = acetyl-CoA C acetyltransferase; HACD = 3-hydroxyacyl-CoA dehydrogenase; ECOAH = enoyl-CoA hydratase; EBACD = electron bifurcating acyl-CoA dehydrogenase; CoAT = CoA transferase; Fd = ferredoxin; RNF = proton translocating ferredoxin: NAD + oxidoreductase complex; HYD = Fe–Fe hydrogenase. Created with BioRender.com . The development of synthetic co-cultures, inspired by mixed culture chain elongation processes, could represent a platform for improving titres, rates, and yields of MCFA production (Fig. 1A ). Moreover, there is growing interest in genetically modifying the native RBO cycle in chain elongators to expand the product spectrum beyond C 4 –C 8 monocarboxylates, including alcohols, diols, dicarboxylates, and other bulk chemicals (Agena et al., 2023 ; Guss & Riley, 2021 ; Strik et al., 2022 ). This would enable the anaerobic synthesis of diverse medium-chain oleochemicals from wet and gaseous waste streams, which is expected to have improved economic feasibility compared to sugar-based aerobic fermentations (Holtzapple et al., 2022 ). However, the technical feasibility of making products other than C 4 –C 8 monocarboxylates via metabolic engineering of chain elongators remains unexplored. As tools to genetically modify chain elongators emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), methodologies to rationally design these new biocatalysts are needed. In this study, we use constraint-based metabolic modelling along with thermodynamic analyses to evaluate the feasibility of synthesizing diverse medium-chain oleochemicals (C 6 –C 12 fatty acids [FAs], primary alcohols, dicarboxylates, diols, and methyl ketones [MKs]) from key intermediate substrates (lactate, ethanol, sugars, and glycerol) using anaerobic chain elongating bacteria (Fig. 1 ). We first analyse MCFA production scenarios beyond natural C 8 production up to C 12 , highlighting trade-offs between ATP yield and expected growth rate using a combination of metabolic modelling and enzyme cost minimization (ECM) analyses. Subsequently, we propose several modifications to the RBO cycle to synthesize target medium-chain oleochemical products and use constraint-based metabolic modelling to determine overall pathway stoichiometry, thermodynamic feasibility, and theoretical product yields. Our results indicate that metabolic engineering of chain elongating bacteria could enable the anaerobic synthesis of diverse medium-chain oleochemicals at industrially relevant yields. Moreover, we identify challenges with engineering the RBO pathway in chain elongators and offer potential solutions to overcome them via metabolic engineering. We anticipate that these results will provide a useful starting point for engineering microbial chain elongation to serve as a platform for sustainable chemical manufacturing.",
"discussion": "Results and Discussion Impact of MCFA Chain Length on Growth Rate and ATP Yield Modelling Chain Elongation Beyond Octanoic Acid Currently, the products of chain elongation are restricted to butyric and hexanoic acids (C 4 –C 6 ), with limited synthesis of octanoic acid (C 8 ) (Nelson et al., 2017 ; Zhu et al., 2017 ). Genetically modifying chain elongating bacteria to improve selectivity and increase MCFA chain length could benefit product yields, while also expanding the process to produce MCFAs with larger markets (C 8 –C 12 ). To assess the feasibility of MCFA production beyond C 8 , we predicted overall pathway stoichiometry, redox balance, free energy change, and theoretical ATP and product yields of C 4 –C 12 FAs via pFBA with ATP yield set as the objective function using a simplified metabolic model describing core chain elongation metabolism (iFermCell193, see the “Methods” section) (Table 1 ). We focused on lactate utilizing chain elongating bacteria (e.g. Pseudoramibacter alactolyticus, Megasphaera hexanoica, Caproicibacterium lactatifermentans ) as lactate is the primary substrate in mixed culture processes converting organic wastes (Fig. 1 ) (Contreras-Dávila et al., 2020 ). Model simulations with different electron donors, including ethanol, glucose, xylose, and glycerol, were also evaluated ( Supplemental Material S1 , Tables S1 – S5 ) to demonstrate process feasibility for a range of other substrates known to be consumed by chain elongating bacteria (e.g. ethanol by Clostridium kluyveri ). Table 1. Predicted Stoichiometry for the Synthesis of C 4 –C 12 FAs from Lactate Chain length Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol lactate) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) C 4 Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.500 0.250 −14.99 −59.96 0.489 C 6 Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 C 8 Lactate + 0.75 H + + 0.375 ADP → CO 2 + 0.5 H 2 + 0.5 H 2 O + 0.25 octanoate + 0.375 ATP 0.250 0.375 −35.10 −93.60 0.402 C 10 Lactate + 0.8 H + + 0.4 ADP → CO 2 + 0.2 decanoate + 0.4 H 2 + 0.6 H 2 O + 0.4 ATP 0.200 0.400 −40.11 −100.29 0.385 C 12 Lactate + 0.833 H + + 0.417 ADP → CO 2 + 0.167 dodecanoate + 0.333 H 2 + 0.667 H 2 O + 0.417 ATP 0.167 0.417 −43.45 −104.29 0.373 The model predicted that ATP yield and overall reaction free energy change increase with MCFA chain length from C 4 to C 12 (Table 1 ), indicating that MCFA synthesis beyond C 8 could be theoretically possible. Moreover, the production of C 4 –C 12 chain lengths from lactate released ≥50 kJ/mol per ATP, indicating that these reactions produce sufficient energy for ATP synthesis (Thauer et al., 1977 ). C 9 –C 12 carboxylate production with other electron donors, including ethanol, glucose, xylose, and glycerol, was also found to be feasible ( Supplemental Material S1 , Tables S1 – S5 ). This observation agrees with past modelling results that suggested hexanoic and octanoic acid production should improve ATP yields from lactate compared to butyric acid production (Scarborough et al., 2018 ). The predicted increase in net ATP production is attributed to a greater flux through the proton translocating ferredoxin: NAD + oxidoreductase complex (RNF) for each turn of the RBO cycle, which leads to greater ATP production via the ion motive force (IMF). In the RBO cycle, each successive elongation generates reduced ferredoxin from the electron-bifurcating acyl-CoA dehydrogenase (EBACD). This ferredoxin goes on to drive RNF to recover some NADH and/or is used by a soluble ferredoxin hydrogenase (HYD1) to evolve H 2 as a terminal electron sink. As confirmed in the predicted flux distributions, elongation to longer chain lengths increases the total NADH demand required to reduce the acyl chains (HACD and EBACD in Fig. 1 ). Thus, with increasing chain length, a greater proportion of the reduced ferredoxin produced by EBACD is used to regenerate NADH via RNF, rather than driving HYD1. This ultimately leads to increased ATP production through the IMF along with decreased H 2 evolution for longer chain lengths. These results also indicate that past a certain chain length (>C 12 ), H 2 evolution will cease as all electron equivalents will be required solely for chain elongation. Intracellular Thermodynamic Landscape and Resource Allocation Our initial modelling analysis suggested that the improved ATP yield with longer chain lengths should naturally select for MCFAs beyond C 8 . However, all pure and mixed culture chain elongation studies have only observed C 4 –C 8 MCFA products. This points to a potential trade-off between growth yield (or ATP yield) and growth rate (or ATP production rate), resulting from optimal resource allocation. In a model based on resource allocation theory ( Supplementary Material S1 ; inspired by a similar model developed by Flamholz et al., 2024 ), where catabolic ATP production and anabolic ATP consumption rates are linearly related to the biomass fractions allocated to either function, RBO enzyme cost per unit ATP production flux and chain elongating bacteria growth rate are inversely correlated (Fig. 2 ). This simplified model captures the fact that a higher growth rate is expected to necessitate both a larger pool of anabolic enzymes and ribosomes (Basan, 2018 ) and a higher ATP production flux, requiring that catabolic machinery produces ATP faster with less enzyme. Fig. 2. ECM results of RBO for production of C 4 –C 12 carboxylates from lactate and the estimated relative effect on growth rate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Each reaction carries a cost which accounts for the pathway's stoichiometry (bottom bar), increased cost incurred due to a reaction's proximity to equilibrium (middle bar), and the cost due to sub-saturation reactant concentrations (top bar). Bottom right panel: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate, as derived from a simplified resource allocation model (Supplemental Material S1) (Top). Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate (Middle). ATP yield on lactate for different chain elongation simulated by pFBA (Bottom). To evaluate the enzyme investment per ATP flux required to produce carboxylates of different chain lengths, we performed ECM analysis. ECM places a lower bound on a pathway's enzyme demand per unit flux [g/(mol/h)] by accounting for the pathway's length and stoichiometry, reaction thermodynamics, and saturation effects following from the optimal set of metabolite concentrations, given appropriate bounds on those concentrations and parametrization of enzyme kinetics (Flamholz et al., 2013 ). This analysis coupled with ATP yield results from our model showed that minimal enzyme cost per unit ATP flux increases with chain length (Fig. 2 ). This suggests that chain elongating bacteria making longer length products invest more of their proteome into high yield catabolism, at the trade-off of having less catalytic capacity for growth. These modelling predictions align with experimental results indicating that increased selectivity and production for hexanoic acid over butyrate is associated with a decrease in growth rate in C. lactatifermentans when grown on lactate as compared to glucose (Wang et al., 2022 ). Moreover, caproate selectivity has been shown to increase with hydraulic retention time (reduced growth rate) in continuous cultures of C. kluyveri and a chain elongation microbiome (Grootscholten et al., 2013 ; Kenealy & Waselefsky, 1985 ). According to the optimized free energy changes predicted by ECM, thiolase reactions (ACACT) are the least exergonic in RBO ( Supplemental Material S1 , Fig. 1 ). Since these reactions operate near equilibrium, they have a smaller forward-to-reverse flux ratio (Noor et al., 2014 ) and therefore require a larger enzyme pool to sustain a given net flux through the pathway. The set of metabolite concentrations that minimizes pathway enzyme cost fixes the thiolase products (oxoacyl-CoAs) at low concentrations to drive these reactions ( Supplemental Material S2 ). This heightens the demand for the proceeding enzymes in the pathway, 3-hydroxyacyl-CoA dehydrogenases (HACDs) since they are far from saturation (Fig. 2 ). Similarly, enoyl-CoA hydratase (ECOAH) reactions are thermodynamically constrained, meaning their reactant and product concentrations are kept high and low, respectively ( Supplemental Material S2 ). This further increases the protein burden of HACD as well as the proceeding electron bifurcating acyl-CoA dehydrogenase (EBACD) (Fig. 2 ). At longer chain lengths, ECOAH reactions are predicted to be more favourable ( Supplemental Material S1 , Fig. 1 ), causing the rate at which enzyme cost increases with chain length to decrease. Interestingly, when oxoacyl-CoA concentrations are constrained by a lower bound of 1 µM, as is the case for all other metabolites, products beyond C 4 are deemed infeasible due to endergonic subsequent thiolase reactions. To reflect observations of chain elongation beyond C 4 , oxoacyl-CoAs can fall to sub-micromolar levels in this model. An alternative solution, as previously suggested in a similar instance with the citric acid cycle (Noor et al., 2014 ), is to posit enzymatic channelling. Channelling between ACACT and HACD would reduce the protein cost of both reactions by effectively merging the two into one favourable reaction. Indeed, channelling between ACACT, HACD, and ECOAH has been described with crystal structures of bacterial and human beta-oxidation complexes (Ishikawa et al., 2004 ; Xia et al., 2019 ). So long as a similar mechanism is active from C 4 to C 12 , the trend of enzyme cost per ATP flux increasing with chain length is expected to hold. This has major implications for engineering chain elongating bacteria to produce MCFAs beyond C8, as the production of longer chain length would provide more ATP per mole of electron donor, but at the cost of a lower growth rate. As a result, compensatory efforts to bolster growth rate, for example through adaptive laboratory evolution (Sandberg et al., 2019 ), may be needed to achieve higher productivity. Expanding and Controlling Chain Elongation Products with Metabolic Engineering Metabolic engineering strategies for the expansion of RBO-derived products have been reviewed for non-chain elongating, model organisms with established genetic tools, such as Escherichia coli, Saccharomyces cerevisiae , and select acetogenic Clostridium species (Tarasava et al., 2022 ). The approach for these strains relies on the engineered reversal of β-oxidation accomplished through extensive strain engineering such as knockout or repression of native β-oxidation regulators and overexpression of genes from chain elongating bacteria and other oleaginous hosts (Tarasava et al., 2022 ). Most of these hosts do not natively rely on the RBO cycle for redox balancing or energy conservation, which is what makes the RBO cycle growth coupled in the chain elongating bacteria. Further, installing orthogonal systems for the engineered reversal of β-oxidation imposes an increased metabolic burden on non-chain elongating hosts as high expression is likely needed to ensure sufficient flux through the pathway (Liu et al., 2018 ). This can lead to redox limitations or inefficient substrate use for chain elongation in these non-chain elongating hosts, and approaches such as two-phase production are likely needed to maximize product yields (Lange et al., 2016 ). Production with non-chain elongating hosts will likely require further strain engineering to improve product yields and selectivity. As an alternative to engineering model organisms, tools to genetically modify chain elongating bacteria are beginning to emerge (Agena et al., 2023 ; Cheng et al., 2019 ; Guss & Riley, 2021 ), which will open the door to expanded opportunities to produce oleochemicals via anaerobic fermentation. Chain elongating bacteria are ideal hosts for RBO-based bioproduction as the RBO cycle is innately growth coupled in these strains (no reconstruction needed) and as members of anaerobic communities can utilize solid and gaseous waste feedstocks rather than relying on refined substrates (e.g. sugars). Further, product yields in anaerobes are typically much greater as they do not synthesize as much biomass compared to aerobic systems and most electrons end up in products (Cueto-Rojas et al., 2015 ). However, the feasibility of pathways for the production of other oleochemicals, including, fatty alcohols, dicarboxylates, diols, and MKs, has not been explored in anaerobic chain elongating bacteria. Here, we propose potential production pathways and enzymes required to generate a wide range of oleochemicals (Fig. 3 ) and use pFBA and ECM to assess the feasibility and potential effects on ATP yield and growth rate. Fig. 3. Proposed pathways for chain elongating bacteria product spectrum expansion. (A) Fatty alcohols pathway. AdhE = bifunctional aldehyde-alcohol dehydrogenase; AOR = aldehyde: ferredoxin oxidoreductase; FadM = acyl-CoA thioesterase. (B) Diol and dicarboxylic acids pathways. NIR = nitrite reductase; NOD = nitric oxide dismutase; AlkBGT = alkane 1-monooxygenase complex; yjgB = cinnamyl alcohol dehydrogenase; chnE = 6-oxohexanoate dehydrogenase; Rd = rubredoxin. (C) Methyl ketone pathway. FadM: thioesterase. Created with Biorender.com . iFermCell193 was further modified to capture the reactions required for hexanoic acid conversion to other products (iFermCell356). Table 2 summarizes the potential pathway stoichiometries for the conversion of hexanoic acid (C 6 ) to the corresponding fatty alcohol, dicarboxylate, and diol using lactate as an electron donor, along with the net ATP yields and overall Gibbs free energy change of reaction for each substrate–product pair as predicted by our model. All predicted pathway stoichiometries have overall favourable standard Gibbs free energies of reaction and release ≥50 kJ/mol per ATP required for ATP production. The model also predicts that net ATP yield is not negatively impacted by conversion into these other bioproducts, which indicates that their production can be growth coupled, similar to MCFAs. In the following sections, we describe potential metabolic engineering strategies to accomplish these conversions in chain elongating bacteria and discuss specific shifts in redox requirements that occur predicted by the model, along with the impact on the metabolic cost of the additional enzymes estimated by ECM (Fig. 4 ). The production of methyl ketones was also modelled, but it was found to negatively impact ATP yield (Fig. 5 ). Fig. 4. ECM results of RBO for production of novel chain elongation products from lactate. Enzyme cost per unit ATP flux is reported as a multiple of butyrate production's cost. Fig. 5. Production profiles for the co-production of 2-pentanone and butyrate from lactate and standard Gibbs free energy of reaction. Molar yields of ATP are negatively related to increasing production of 2-pentanone. Co-production of butyrate is required with increasing 2-pentanone production. Table 2. Predicted Pathway Stoichiometry for Different C 6 Oleochemicals Produced from Lactate Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol Lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Carboxylate; RBO Lactate + 0.667 H + + 0.333 ADP → CO 2 + 0.667 H 2 + 0.333 H 2 O + 0.333 hexanoate + 0.333 ATP 0.333 0.333 −26.75 −80.24 0.431 Alcohol; RBO + AdhE Lactate + H + + 0.333 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.333 ATP 0.333 0.333 −45.18 −135.53 0.382 Alcohol; RBO + AOR Lactate + H + + 0.5 ADP → CO 2 + 0.667 H 2 O + 0.333 hexanol + 0.5 ATP 0.333 0.500 −45.18 −90.35 0.382 Dicarboxylate; RBO + AlkBGT Lactate + H + + 0.667 NO 2 − + 0.333 ADP → 0.333 adipate + CO 2 + 0.667 H 2 + H 2 O + 0.333 N 2 + 0.333 ATP 0.333 0.333 −228.21 −684.62 0.539 Diol; RBO + AlkBGT Lactate + 0.5 NO 2 − + 1.25 H + + 0.5 ADP → 0.25 1,6-hexanediol + 0.25 acetate + CO 2 + H 2 O + 0.25 N2 + 0.5 ATP 0.250 0.500 −191.63 −383.26 0.332 RBO = reverse beta-oxidation, AdhE = bifunctional alcohol dehydrogenase, AOR = aldehyde:ferredoxin oxidoreductase, AlkBGT = alkane 1-monooxygenase complex Fatty Alcohol Production Medium-chain fatty alcohols (FAOHs) could be produced using MCFAs as building blocks. Even though FAOH production pathways (Fig. 3A ) exist in anaerobic bacteria, such as acetogens and butanol-producing clostridia (Moon et al., 2016 ), this has yet to be observed in most chain elongating bacteria. In nature, acetogenic bacteria use aldehyde:ferredoxin oxidoreductase (AOR) to convert FAs into fatty aldehydes in monoculture or when they are co-cultured with chain elongating bacteria (Benito-Vaquerizo et al., 2020 ; Diender et al., 2016 ). In another instance, aldehyde-alcohol dehydrogenase (AdhE) can convert acyl-CoAs into aldehydes (Mehrer et al., 2018 ). In these two scenarios, specific electron carriers [one reduced ferredoxin and one NAD(P)H in the former and two NAD(P)H in the latter] are necessary to obtain the final product. Final conversion relies on the alcohol dehydrogenase (Adh) activity or the bifunctionality of AdhE (Liew et al., 2017 ). Ultimately, both AOR- and AdhE-based FAOH production should result in reduction in hydrogen evolution if implemented in the chain elongating bacteria as 4 electron equivalents are required for the additional reduction reactions. We incorporated both AOR and AdhE FAOH synthesis pathways in our model and assessed their effects on FAOH production, redox balances, and ATP yield. In chain elongation to FAs, H 2 is evolved as an electron sink to maintain redox balance. We hypothesized that instead of producing H 2 , these electron equivalents could be used to reduce the carboxylate to alcohol. Consistent with this, the model predicted that hexanol (C 6 FAOH) production does not result in H 2 production (Table 2 ), and instead, flux is redirected from HYD1 to RNF to produce NADH required by AOR and AdhE pathways. However, in the AOR pathway, the model predicted a net increase in ATP flux yield through the combined action of IMF and substrate-level phosphorylation as compared to hexanoic acid production ( Supplementary Material S3 ). ECM results show that hexanol production via AOR carries an enzyme cost per ATP flux of 2.1 times that of butyrate (Fig. 4 ), whereas hexanoic acid production requires 2.7 times more than butyrate (Fig. 2 ). This reflects the increased ATP yield from FAOH production with AOR and the prediction that the additional protein needed to reduce hexanoate to hexanol represents a small fraction of the pathway's total protein demand. In the AdhE case, ECM results show that hexanoic acid and hexanol production require approximately equal enzyme investments of 2.7 and 2.6 times the cost of butyrate production, due to bypassing the use of phosphate acetyltransferase (PTA) and acetate kinase (ACK) for alcohol production. In both AOR and AdhE cases, it can be expected that chain elongating bacteria engineered for FAOH production will not experience significant growth rate penalties compared to production of MCFA of the same chain length. In fact, growth rate could theoretically increase if the AOR pathway is implemented due to a reduction of enzyme cost per unit ATP flux (Fig. 6 ). A controlled metabolic engineering intervention would be required to avoid overexpression of heterologous FAOH production genes, which could be detrimental to growth rate, while successfully diverting flux from HYD1 to RNF to simultaneously balance the increased NADH demand and generate a proton motive force. Fig. 6. Top: Enzyme cost per ATP flux relative to butyrate versus specific growth rate relative to butyrate for bioproducts evaluated in this study, as derived from a simplified resource allocation model ( Supplemental Material S1 ). Middle: Anabolic fraction of biomass corresponding to specific growth rate and catabolic fraction of biomass remaining to supply ATP at a given growth rate. Bottom: ATP yield on lactate for different chain elongation simulated by pFBA. Dicarboxylate and Diol Production: Pathways Requiring Oxygen The anaerobic oxidation of MCFAs into diols and dicarboxylic acids poses a significant metabolic engineering challenge in chain elongators, as the lack of oxygen as an electron acceptor results in a considerable decrease in reaction energetics. Methane oxidation is a clear example where the anaerobic pathway needs to be coupled to nitrite or sulphate reduction (Knittel & Boetius, 2009 ). Alkane-degrading microbes can anaerobically oxidize C–H bonds by fixing CO 2 and converting alkanes into FAs. It is hypothesized that this mechanism is initiated by an ethylbenzene dehydrogenase and the whole pathway happens at the great expense of 6 ATP equivalents, suggesting that an unexplored energy-coupling mechanism may take place in these organisms given the low ATP yields in anaerobes ( Supplemental Material S1 , Fig. 2 ) (Heider et al., 2016 ; Shou et al., 2021 ). Alternatively, alkane monooxygenase complexes such as alkane 1-monooxygenase complex (AlkBGT) have been used in the hydroxylation of terminal ω-carbons of different compounds, including FAs (Clomburg et al., 2015 ). However, implementing AlkBGT in anaerobes is currently infeasible due to its requirement for O 2 . To solve these problems, we envision making use of another hypothesized hydroxylation mechanism in anaerobes: the intra-aerobic pathway (Ettwig et al., 2010 ). Achieving such conditions requires “oxygen donors”, such as nitrite, hydrogen peroxide, or perchlorate, whose decomposition provides the required oxygen for AlkBGT. Since oxygen presence can be problematic in this scenario, it is important to maintain its intracellular concentrations low, which could be achieved by tuning nitrite reductase/nitric oxide dismutase (NIR/NOD) and AlkBGT expression or even improving AlkBGT rates through protein engineering. If oxygen levels are still high enough to disrupt cell function, a metabolic strategy that could be explored would be to confine the oxygen generation reaction within microcompartments, pseudo-organelles common in various organisms, including anaerobes, which have been used in several applications, such as protecting cells from toxic intermediates and oxygen (Heinhorst & Cannon, 2020; Kennedy et al., 2021 ). In our proposed pathway, the first module comprises oxygen generation, where nitrite is reduced to nitric oxide by NIR (Wang et al., 2019 ), followed by the action of NOD to yield N 2 and O 2 . The second module encompasses the oxidation reactions, where AlkB would be used to oxidize FAs and FAOHs to ω-hydroxyacids and diols, respectively, using rubredoxin as an electron carrier (Fig. 3B ). The remaining portion of the AlkBGT complex is responsible for regenerating reduced rubredoxin by consuming NADH. Further oxidation of ω-hydroxyacids to dicarboxylic acids, however, would require additional enzymes, such as cinnamyl alcohol dehydrogenase (yjgB) and 6-oxohexanoate dehydrogenase (chnE) (Clomburg et al., 2015 ), even though it is reported that AlkBGT complex can further oxidize hydroxyls to carboxyls and even ester groups, a phenomenon called overoxidation (van Nuland et al., 2017 ). Protein discovery and engineering will play a key role in bringing these pathways to life in the chain elongating bacteria. For example, understanding anaerobic carboxylation mechanisms and discovering other related pathways would circumvent the need of using oxygen and microcompartments. Additionally, to the best of our knowledge, no FAOHs have been used as substrates for AlkBGT, which would require protein characterization and engineering to achieve diol production if intra-aerobic conditions are necessary. The reactions described above were incorporated into our model and the effect on redox balance and ATP yield was assessed for the production of adipic acid and 1,6-hexanediol from hexanoic acid. NO 2 − was supplied in our simulations as a substrate to source O 2 for the hydroxylation step. Nitrite reduction via NIR requires an investment of electron equivalents from NADH. However, the model predicted that the production of adipic acid does not require additional electron equivalents from H 2 nor lactate as H 2 evolution and molar yield of adipic acid from lactate was the same as in the hexanoic acid case. For adipic acid, the ω-hydroxyacid intermediate that is produced by AlkBGT is subsequently oxidized, releasing an NADH that is required for the reduction of NO 2 − and balancing redox. However, in the case of 1,6-hexanediol, an additional electron source is required for the initial reduction of hexanoic acid to hexanol. Thus, diol production is not as carbon efficient because additional electron equivalents from lactate are required to reduce both terminal ends into alcohols, leading to the co-production of acetate to balance carbon, yielding a net increase in ATP yield due to substrate-level phosphorylation through PTA/ACK. ECM found that 1,6-hexanediol carries an enzyme cost lower than that of hexanol produced via AOR, 1.4 times the cost of butyrate. Cost decreases since both pathways share an ATP yield of 0.5, while diol production does away with the CoA transferase (CoAT) reaction. This reaction is itself enzymatically expensive due to its low driving force, and it increases the cost of PTA and ACK reactions as CoAT requires a higher acetate concentration ( Supplemental Material S2 ). Adipic acid production requires 5.4 times more enzyme investment per ATP flux than butyrate (Fig. 4 ). The heightened enzyme cost is largely due to the unfavourable YJGB reaction, in which 6-hydroxyhexanoate reduces NAD + to form 6-oxohexanoate. Minimal enzyme cost is achieved when 6-hydroxyhexanoate concentration is high, driving the YJGB reaction but indirectly increasing the demand for termination enzymes (CoAT, PTA, and ACK) through compensatory heightening of hexanoate and acetate concentrations ( Supplemental Material S2 ). This suggests that it may be more challenging to achieve high adipic acid productivity due to a reduction in growth rate. On the other hand, 1,6-hexanediol production could even increase growth rate vs. hexanoic acid, similarly to hexanol production via AOR (Fig. 6 ). MK Production MKs are another chemical class that can be produced using MCFAs as precursors. Currently, in the engineered RBO cycle, MKs are obtained from the hydrolysis of oxoacyl-CoA molecules by the thioesterase FadM. This reaction initially generates ketoacids, which in turn spontaneously decarboxylate to result in the final MK (Fig. 3C ) (Yan et al., 2020 ). In addition to FadM introduction, other metabolic engineering strategies are necessary to achieve MK production in chain elongating bacteria. The first strategy would be to substitute the native CoATs by FadM, translocating termination from acyl-CoA to oxoacyl-CoA. Since the RBO cycle is essential to redox balance and ATP production in chain elongating bacteria, sole acetone and 2-butanone production would not be possible, as FadM would terminate before one complete turn of the RBO cycle. Therefore, to optimize MK production, it is necessary to allow MCFA with MK co-production. FadM was incorporated into our model to assess the potential to produce 2-pentanone from 6-oxohexanoyl-CoA. The model predicted that 2-pentanone production is not optimal for ATP yield, which aligns with the early termination initiated by FadM at 6-oxohexanoyl-CoA, where the ATP from the complete elongation to hexanoic acid is not produced. To assess the effect of 2-pentanone production on ATP yield, we fixed 2-pentanone production at various values and determined the overall pathway stoichiometry at each point using pFBA with ATP yield set as the objective function (Table 3 ). It was found that while 2-pentanone production is thermodynamically feasible for the scenarios modelled, its production is negatively coupled with butyrate and ATP production (Fig. 5 ). CO 2 production also increases with increasing 2-pentanone production due to the spontaneous decarboxylation of the ketoacid. In isolation, this loss of CO 2 severely impacts the carbon efficiency of this production pathway. However, it could be alleviated if MK production was conducted in the context of a microbial community, where other functional guilds could recycle the CO 2 (Baleeiro et al., 2023 ). Table 3. Selected Predicted Pathway Stoichiometries for 2-Pentanone Production Product; pathway Overall equation Product yield (mol P/mol lactate) ATP yield (mol ATP/mol lactate) Δ G r °′ (kJ/mol) Δ G r °′ per ATP (kJ/mol/mol ATP) Yield (g P/g lactate) Methyl ketone; RBO + FadM Lactate + 0.5 H + + 0.25 ADP → 0.5 butyrate + CO 2 + H 2 + 0.25 ATP 0.000 0.250 −14.99 −59.96 0.000 Lactate + 0.545 H + + 0.183 ADP → 0.03 2pentanone + 0.455 butyrate + 1.03 CO 2 + 1.03 H 2 + 0.183 ATP 0.030 0.183 −12.13 −66.48 0.029 Lactate + 0.575 H + + 0.138 ADP → 0.05 2-pentanone + 0.425 butyrate + 1.05 CO 2 + 1.05 H 2 + 0.138 ATP 0.050 0.138 −10.23 −74.38 0.048 Lactate + 0.605 H + + 0.093 ADP → 0.07 2-pentanone + 0.395 butyrate + 1.07 CO 2 + 1.07 H 2 + 0.093 ATP 0.070 0.093 −8.32 −89.97 0.067 Lactate + 0.665 H + + 0.003 ADP → 0.11 2-pentanone + 0.335 butyrate + 1.11 CO 2 + 1.11 H 2 + 0.003 ATP 0.110 0.003 −4.51 −1804.79 0.105 RBO = reverse beta-oxidation, FadM = oxoacyl-CoA thioesterase"
} | 10,557 |
29460958 | PMC5947824 | pmc | 1,063 | {
"abstract": "Abstract Harnessing the metabolic potential of uncultured microbial communities is a compelling opportunity for the biotechnology industry, an approach that would vastly expand the portfolio of usable feedstocks. Methane is particularly promising because it is abundant and energy‐rich, yet the most efficient methane‐activating metabolic pathways involve mixed communities of anaerobic methanotrophic archaea and sulfate reducing bacteria. These communities oxidize methane at high catabolic efficiency and produce chemically reduced by‐products at a comparable rate and in near‐stoichiometric proportion to methane consumption. These reduced compounds can be used for feedstock and downstream chemical production, and at the production rates observed in situ they are an appealing, cost‐effective prospect. Notably, the microbial constituents responsible for this bioconversion are most prominent in select deep‐sea sediments, and while they can be kept active at surface pressures, they have not yet been cultured in the lab. In an industrial capacity, deep‐sea sediments could be periodically recovered and replenished, but the associated technical challenges and substantial costs make this an untenable approach for full‐scale operations. In this study, we present a novel method for incorporating methanotrophic communities into bioindustrial processes through abstraction onto low mass, easily transportable carbon cloth artificial substrates. Using Gulf of Mexico methane seep sediment as inoculum, optimal physicochemical parameters were established for methane‐oxidizing, sulfide‐generating mesocosm incubations. Metabolic activity required >∼40% seawater salinity, peaking at 100% salinity and 35 °C. Microbial communities were successfully transferred to a carbon cloth substrate, and rates of methane‐dependent sulfide production increased more than threefold per unit volume. Phylogenetic analyses indicated that carbon cloth‐based communities were substantially streamlined and were dominated by Desulfotomaculum geothermicum . Fluorescence in situ hybridization microscopy with carbon cloth fibers revealed a novel spatial arrangement of anaerobic methanotrophs and sulfate reducing bacteria suggestive of an electronic coupling enabled by the artificial substrate. This system: 1) enables a more targeted manipulation of methane‐activating microbial communities using a low‐mass and sediment‐free substrate; 2) holds promise for the simultaneous consumption of a strong greenhouse gas and the generation of usable downstream products; and 3) furthers the broader adoption of uncultured, mixed microbial communities for biotechnological use.",
"conclusion": "4 CONCLUSIONS The availability of methane makes it an attractive target for industrial use, and its high greenhouse warming potential makes its oxidation appealing for both environmental and economic purposes. Here we provide an approach to harness the methane‐oxidizing, sulfide‐producing capabilities of microbial communities from marine methane seeps in the service of these dual aims. Microscopic imaging revealed that ANME and SRB were predominant on the carbon cloth substrate, adopted a novel, surface‐associated, aggregate‐free arrangement, and were selectively enriched compared with native seep sediment (a finding that may be attributable to the electrical conductivity of carbon cloth). Harnessing complex natural assemblages of organisms offers important advantages over reductionist attempts to engineer model organisms, including metabolic redundancy and an expanded functional space unconstrained by cultivability requirements. Achieving a methane‐activating, sulfide‐producing mixed culture in a sediment‐free context is a key development that presents compelling biotechnological and greenhouse gas mitigation opportunities. On the carbon cloth artificial substrate, sulfide production rates per unit volume increased more than threefold, making the system presented here promising for bioremediating or PHA‐producing SRB, or downstream sulfide‐utilizing processes. More broadly, this work shows the feasibility of shifting complex microbial communities to tractable systems devoid of the physicochemical challenges of natural substrates. The incorporation of uncultivated, mixed microbial communities into reactor systems will enable practitioners to access an enormous range of biological function for industrial aims, opening a new realm of biotechnological potential.",
"introduction": "1 INTRODUCTION The use of methane in the production of infrastructure‐compatible, industrially relevant products is an attractive prospect with benefits of energy generation, chemical production, and climate regulation. The production of natural gas, which is composed primarily of methane, has reached historically high rates (U.S. Energy Information Administration, 2016 ). This trend provides affordable feedstock on a global scale, much of which is produced by small, distributed units that are not well‐served by energy‐intensive chemical plants that require high capital infrastructure expenditures (Clomburg, Crumbley, & Gonzalez, 2017 ; Emerson et al., 2017 ). In addition, leakage associated with natural gas production delivers tens of teragrams of methane to the atmosphere each year (Brandt et al., 2014 ). Given methane's strength as a greenhouse gas, minimizing this loss is a key priority of climate change mitigation strategies (Howarth, 2014 ; Stocker, 2014 ). While synthetic catalysts have shown promise in methane reforming (York, Xiao, & Green, 2003 )—particularly through high‐temperature transformations (Takenaka, Kobayashi, Ogihara, & Otsuka, 2003 ; Tang, Zhu, Wu, & Ma, 2014 )—biological processes are particularly compelling given their more moderate operating conditions, highly selective methane‐activating biochemistry, and self‐sustaining maintenance. A microbial system's ability to perform several transformations within a small volume minimizes complexity and cost of multi‐step conversions (Clomburg, Crumbley, & Gonzalez, 2017 ). The anaerobic oxidation of methane (AOM) is an energy‐ and carbon‐efficient process with a near‐catalytic processivity: only 1–3% of methane‐carbon enters cell biomass (Nauhaus, Albrecht, Elvert, Boetius, & Widdel, 2007 ; Treude et al., 2007 ). In contrast, aerobic methanotrophy proceeds at faster rates, but sequesters a higher proportion of methane‐derived carbon as biomass (Hanson & Hanson, 1996 ) and has a lower energy efficiency and higher reactant costs (Haynes & Gonzalez, 2014 ; Trotsenko & Murrell, 2008 ). AOM with sulfate as an electron acceptor is typically mediated by syntrophic consortia of anaerobic methanotrophic archaea (ANME) and sulfate reducing bacteria (SRB) that are abundant in anoxic sediments at marine methane seeps (Boetius et al., 2000 ; Ruff et al., 2015 ). Their slow growth (Girguis, Cozen, & DeLong, 2005 ; Krüger, Wolters, Gehre, Joye, & Richnow, 2008 ) and remarkable resistance to laboratory culture isolation make ANME‐SRB consortia challenging constituents of industrial processes. Despite these apparent obstacles, AOM metabolism holds substantial promise for the efficient conversion of a strong greenhouse gas into high‐value products. Attempts to incorporate AOM into a cultivable methanogen have produced a M. acetivorans strain with an ANME‐1 mcr gene that uses ferric iron as an electron acceptor (Soo et al., 2016 ) and whose modeled capabilities include methanol, ethanol, butanol, and isobutanol production (Nazem‐Bokaee, Gopalakrishnan, Ferry, Wood, & Maranas, 2016 ). The addition of a 3‐hydroxybutyryl‐CoA dehydrogenase gene from C. acetobutylicum enabled lactate formation (McAnulty et al., 2017 ). These important contributions would be complemented by sulfate‐coupled methane‐oxidizing industrial capabilities, whereby diffusible, environmentally abundant electron acceptors could help generate easily transportable liquid‐phase products. Moreover, there are clear advantages of working with mixed communities; in comparison to pure cultures, diverse assemblages of organisms are more resilient to environmental perturbations and offer a substantially wider range of functional capabilities (Briones & Raskin, 2003 ; Girvan, Campbell, Killham, Prosser, & Glover, 2005 ). Here, we report a method for sustaining methane‐oxidizing, sulfide‐generating communities on artificial substrates that can be maintained without frequent re‐inoculation from deep‐sea sediments. This approach offers several industrial benefits, both in its current configuration and as an enabling technology for future enhancements. By converting methane to dissolved inorganic carbon, greenhouse warming potential is reduced more than an order of magnitude (Stocker, 2014 ), and engineered autotrophs could be introduced downstream to generate high‐value liquid fuels (Claassens, Sousa, dos Santos, de Vos, & Van der Oost, 2016 ; Lan & Liao, 2013 ). In addition to building a symbiosis with ANME to enable methane consumption, sulfate‐reducing bacteria can remediate heavy metals (García, Moreno, Ballester, Blázquez, & González, 2001 ; Joo, Choi, Kim, Kim, & Oh, 2015 ) and produce plastic precursor storage molecules (Hai, Lange, Rabus, & Steinbüchel, 2004 ; Wang, Yin, & Chen, 2014 ). Sulfide generated through this process can be used as feedstock for genetically tractable chemoautotrophs (Kernan, West, & Banta, 2017 ; Nybo, Khan, Woolston, & Curtis, 2015 ), which can make high‐value chemical products (Kernan et al., 2016 ). Finding a conductive physical scaffold to enable industrially‐relevant reactions within the context of an ecophysiologicaly sustainable system—whereby microbial communities can take advantage of natural electrical and chemical gradients (e.g., Pfeffer et al., 2012 )—would decrease transport and scale‐up costs while building a robust, customizable system. Furthermore, methane seep microbial communities can withstand relatively high concentrations of sulfide, a prominent contaminant of natural gas reservoirs (Faramawy, Zaki, & Sakr, 2016 ; Sublette & Sylvester, 1987 ), obviating the costly process of feedstock purification and opening large quantities of “dirty methane” to bioindustrial use. To inform the conditions needed for the initial biochemical transformations in this multipartite system, we identify salinity and temperature conditions needed for optimized methane consumption in natural sediment mesocosms, and evaluate the changes in microbial community structure associated with heightened methane consumption and sulfide production rates. Ultimately, we demonstrate robust sediment‐free AOM activity on a conductive, high‐surface area artificial substrate that promotes enhanced methane activation and sulfide production via a streamlined microbial community. These methanotrophic biocatalytic systems are poised to provide economically viable solutions for liquid fuel or chemical production.",
"discussion": "3 RESULTS AND DISCUSSION 3.1 Medium composition and influence of salinity on methanotrophic activity ANME‐SRB consortia are most abundant and active in marine sediments perfused with upward‐advecting reduced fluids and downward circulating sulfate‐rich seawater (Joye et al., 2004 ; Treude et al., 2003 ). We sought to determine if AOM activity could be retained under low‐salt conditions, as reduced salinity may be desirable in select biotechnology applications. Triplicate mesocosm incubations of seep sediment were established at 4 °C under conditions of varying salinity but constant sulfate concentrations (to maintain a consistent Gibbs free energy value; Supplementary Table S1). Methane oxidation was measurable only in samples with salinities greater than ∼2.2 wt% salt (22 g/kg, or ∼63% seawater salinity; Figure 1 ). Full‐salinity samples oxidized methane at a rate of 138 (+/− 18.4) nmol/cm 3 d after 245 days, while the 63% salinity incubations exhibited 32% (+/− 8.5%) as much activity. These values are consistent with previously characterized seep substrates (Marlow, Steele, Ziebis, et al., 2014 ; Treude et al., 2003 ; Wegener et al., 2008 ). The absence of detectable methane oxidation in other conditions—samples with 17–39% salinity levels and full‐salinity killed control incubations—suggests that a threshold of >∼40% seawater salinity (e.g., 15.8 ms/cm) is required for AOM activity. This threshold is likely closer to 45% if the observed decrease in methane oxidizing activity with lower salinity is governed by a linear relationship. The use of constant sulfate concentrations (∼28 mM) means that the energetic driving force of AOM was not a distinguishing factor. Figure 1 Concentrations of methane oxidized and sulfide produced in triplicate incubations of Gulf of Mexico seep sediment at distinct salinity levels and constant sulfate concentrations. The sulfate concentrations and salinity levels of each incubation are provided in Supplementary Table S1. Full lines and circles represent methane data; dotted lines and triangles signify sulfide data; error bars show standard deviations Copyright © 2017 Wiley Periodicals, Inc. ANME‐SRB‐mediated AOM results in a 1:1 stoichiometry of methane oxidation and sulfate reduction (Boetius et al., 2000 ), but alternative electron donors can facilitate sulfate reduction in seafloor sediments (Kleindienst et al., 2014 ) and decouple this 1:1 relationship. In our experiments, sulfide production exceeded methane consumption: using the 245‐day time points of the 100% and 63% salinity incubations, AOM accounted for 74% and 71% of sulfide production, respectively. Sulfide production was not observed at 20% salinity but did occur in the absence of methane activation at 39% and 28% salinity (Figure 1 ). These data demonstrate that sulfate reducing organisms in our mesocosm incubations can use non‐methane electron donors and are less affected by—but still subject to—challenges associated with low salinity. Our data demonstrate that Gulf of Mexico seep sediment‐derived communities require full seawater levels of salinity to attain optimal activity. While this finding is consistent with the marine environment from which the inoculum was collected, a physiological explanation remains to be determined. Freshwater AOM has typically been associated with non‐sulfate electron acceptors such as nitrate (Deutzmann & Schink, 2011 ; Smemo & Yavitt, 2007 ), and previous observations of freshwater sulfate‐reducing AOM reported substantially lower rates (∼8 nmol/cm 3 day; (Timmers et al., 2016 )). Some methanogens require seawater‐level salinity, potentially to maintain cell ultrastructure (Kadam, Ranade, Mandelco, & Boone, 1994 ). Energetics‐based effects of salinity on metabolism—potentially associated with electrical conductivity (Wegener, Krukenberg, Riedel, Tegetmeyer, & Boetius, 2015 ) or osmoregulation (Csonka & Hanson, 1991 )—warrant further investigation. 3.2 Influence of temperature on sulfate reduction activity by seep sediment communities To determine if AOM‐coupled sulfide production could be enhanced at elevated temperatures, we measured sulfide production under high‐methane, full seawater salinity conditions at 5, 20, 35, and 50 °C. Triplicate incubations of Gulf of Mexico sediment (recovered from ∼5 °C sediments) were evaluated for 72 days, and measured sulfide concentrations were used as a proxy for methane oxidation (Figure 2 ). As described above, sulfide production rates were consistently proportional to methane oxidation rates in experiments using co‐localized seep sediment inoculum (0.71–0.74 methane oxidized: sulfate reduced, Figure 1 ). Figure 2 Concentrations of sulfide produced by triplicate batch incubations of methane seep sediment at 5.5, 20, 35, and 50 °C. Killed control values were subtracted from each treatment, and error bars show standard deviations Copyright © 2017 Wiley Periodicals, Inc. Sulfide production rates were highest at 35 °C: this condition yielded ∼18.6 nmol/cm 3 day over the first 44 days of the experiment, corresponding to 13.8 nmol methane oxidation/cm 3 day (calculated using a sulfide production‐to‐methane oxidation scaling factor of 0.74 from the full‐salinity incubations). This value is substantially lower than the methane oxidation rate of the full‐salinity condition described above, likely due to localized heterogeneity of seep sediments (Marlow, Steele, Case, et al., 2014 ; Orphan et al., 2004 ) and the “dilution” effect of using a broader horizon of sediment beyond the standard sulfate‐methane transition zone (Joye et al., 2004 ; Lloyd et al., 2010 ). Sulfide production rates in the 50 °C incubations were also substantial, reaching an average of 5.3 nmol/cm 3 day after 44 days, or ∼28% that generated by the 35 °C samples. Despite the thermal regime (∼5 °C seafloor sediment) of the inoculum's origin, previous studies have revealed temperature optima for AOM approximately 5–10 °C above in situ temperatures (Boetius et al., 2009 ), and, for marine sediment sulfate reduction, >20 °C above in situ temperatures (Arnosti, Jørgensen, Sagemann, & Thamdrup, 1998 ; Sagemann, Jørgensen, & Greeff, 1998 ). In our experiments, heightened rates of overall metabolism associated with higher temperatures (Price & Sowers, 2004 ) appear to dominate up to 35 °C. Above this temperature, methanotrophic or sulfate reduction‐associated enzymes may be compromised, and/or thermophilic constituents of the community may out‐compete ANME/SRB for key nutrients (Levén, Eriksson, & Schnürer, 2007 ). Our observations could also indicate a decoupling of methane consumption and sulfate reduction (see below), which has been noted at sites of hydrothermal AOM (Kallmeyer & Boetius, 2004 ). Between days 44 and 61, sulfide levels of the high‐temperature mesocosm samples diminished substantially (Figure 2 ). This decrease was likely caused by abiotic loss through oxidation with ferric minerals and pyrite formation (Peckmann et al., 2001 ; Wilkin & Barnes, 1997 ), and/or by a proliferation of anaerobic sulfide‐oxidizing organisms (see “ Changes in Microbial Community Composition ” section 3.3). 3.3 Changes in microbial community composition To evaluate changes in the microbial community that may account for the observed sulfide concentrations—and possibly provide prescriptive insight into shaping complex communities for heightened metabolic activity —16S rRNA gene surveys of five samples (original inoculum sediment at T \n 0 , 72‐day time point samples from 5, 20, 35, and 50 °C seep sediment incubations) were conducted. The 16S rRNA gene data suggest that, compared with native seep sediment microbial assemblages, more diverse communities exhibit heightened levels of sulfide production at elevated temperatures. Alpha diversity metrics Chao‐1 (which estimates species richness) and Inverse Simpson (which incorporates richness and evenness) reveal that heightened sulfide production rates in seep sediment incubations may correlate with both higher relative richness and a greater degree of evenness (Supplementary Table S2), corresponding to a functionally diverse community with many low‐abundance constituents. ANME were relatively scarce in the 16S rRNA gene sequence data (<0.1% in all samples), which may seem surprising given the prevalence of sulfide production coupled with methane oxidation at the incubation scale. However, the Earth Microbiome Project (EMP) primers—used here in order to enable direct comparison with thousands of other studies that have used EMP protocols—have been shown to be relatively ineffective in capturing archaeal 16S sequences in general (Raymann, Moeller, Goodman, & Ochman, 2017 ), and ANME‐2 16S sequences in particular (Case et al., 2015 ), and future studies targeting these lineages should consider alternate primer sets (Parada, Needham, & Fuhrman, 2016 ; Timmers, Widjaja‐Greefkes, Plugge, & Stams, 2017 ). To confirm ANME‐SRB presence, FISH was performed on 35 °C incubation sediment collected on day 72. Numerous canonical clumped aggregates of ANME‐1/2 (Boetius et al., 2000 ) and Desulfosarcina / Desulfococcus (Manz, Eisenbrecher, Neu, & Szewzyk, 1998 ) were abundant (e.g., field of view provided in Figure 5 b). By evaluating the relative abundances of known sulfate reducing and sulfide oxidizing lineages (Supplementary Table S3), we can use community sequencing data to provide insight into the sulfide production and consumption dynamics observed in the sediment incubations (Figure 2 ). The presence and abundance of genomic DNA do not directly reveal metabolic activity levels (Maier, Güell, & Serrano, 2009 ; Muyzer & Stams, 2008 ). Nonetheless, an organism's abundance does reveal potentially prominent metabolic activities, as higher abundances provide more opportunities for genes to be expressed and associated metabolic reactions—inferred from close cultured representatives or detailed environmental studies—to be enacted. This approach is frequently used to model community function (Bowman, Amaral‐Zettler, Rich, Luria, & Ducklow, 2017 ) and is particularly promising for specific metabolic reactions with few branch points, such as methane and sulfur catabolism (Fierer, 2017 ; Preheim et al., 2016 ). The high relative abundance of sequences derived from sulfide oxidizing bacteria (Supplementary Table S2) in the 35 and 50 °C sediment communities, which accounted for 22% (+/− 4% SE) and 28.5% of recovered sequences, respectively (compared with 6.2% from the T 0 inoculum), helps explain the decrease in sulfide concentration between days 44 and 72. The 35 °C condition also had a substantial proportion of sulfate reducing bacteria sequences (12.4% +/− 1.7% SE compared with 7.2% from the T \n 0 inoculum), consistent with its elevated sulfide production. 3.4 Harnessing AOM on artificial substrates Establishing a sediment‐free, methane oxidizing, sulfide producing microbial community is a strategic priority for the bioenergy industry. Such a capability would enhance the quantity of product per unit volume, lower maintenance costs, and provide a physical scaffold to enrich for specific interspecies communities. Seep sediments are frequently comprised of phyllosilicates transported by fluvial erosion, which can minimize permeability (thereby limiting nutrient inflow and waste removal), adhere to cells, and restrict movement (Hong et al., 2014 ). AOM also promotes the formation of authigenic carbonates, which may ultimately entomb methane oxidizing constituents (Marlow, Peckmann, & Orphan, 2015 ). Here, two artificial substrates were used as potential scaffolds for sediment‐free anaerobic methanotrophic communities grown with CH 3 D. Carbon cloth is a permeable, high surface area material shown to promote microbial attachment (Zhang et al., 2012 ; Zhao et al., 2013 ) and stimulate direct electron transfer in co‐cultures (Chen et al., 2014 ). Zircon beads offer a non‐conductive but topologically suitable surface for microbial colonization (Scarano, Piattelli, Caputi, Favero, & Piattelli, 2004 ), and their spherical shapes enable a sediment‐like packing structure. Sulfide production in carbon cloth and zircon bead incubations was monitored over the course of 44 days at 5.5, 20, 35, and 50 °C (Figures 3 a and 4 a). Carbon cloth incubations produced sulfide at higher rates than corresponding seep sediment incubations. During the period of peak sulfide production (15 days, Figure 3 a), the D/H ratios of the 35 °C incubations' water indicated that 24.2 nmol methane/cm 3 day had been oxidized, accounting for 42% of sulfide production (using the 1:1 stoichiometry of sulfate‐associated AOM). Subtracting this component—as well as sulfide produced in the methane‐free control (154 µM, or 17%, consistent with other AOM enrichments (Holler et al., 2011 ) − 41% of overall sulfide production was unaccounted for (Figure 3 b). This sulfide was methane‐dependent, but did not correspond to a concomitant oxidation of methane; it may have resulted from sulfate reduction linked to oxidation of non‐methane organics generated by ANME primary producers. The result signifies an amplification of methane‐driven sulfide production useful for downstream technologies that incorporate genetically engineered sulfide oxidizers (Kernan et al., 2016 ). Figure 3 (a) Concentrations of sulfide produced by triplicate batch incubations of carbon cloth at 5.5, 20, 35, and 50 °C. Killed control values were subtracted from each treatment, and error bars show standard deviations. * = water for D/H analysis was collected. (b) Analysis of methane consumption and sulfide levels of the isothermal methane‐free nitrogen control reveal the balance between methane‐dependent sulfide production, methane oxidation‐linked sulfide production, and non‐methane‐dependent sulfide production Copyright © 2017 Wiley Periodicals, Inc. Peak sulfide production on zircon bead substrates was substantially lower than in both sediment‐ and carbon cloth‐hosted incubations (Figure 4 a). Over the first 13 days of incubation, sulfide production rates in zircon bead‐hosted incubations were 15.3 nmol/cm 3 d (35 °C) and 4.5 nmol/cm 3 day (50 °C), or 24% and 32% that of their carbon cloth‐hosted counterparts, respectively. In contrast to the carbon cloth incubations, the 35 °C zircon bead‐hosted sample exhibited negligible amounts of methane‐independent sulfide production (Figure 4 b). Figure 4 (a) Concentrations of sulfide produced by triplicate batch incubations of zircon beads at 5.5, 20, 35, and 50 °C. Killed control values were subtracted from each treatment, and error bars show standard deviations. * = water for D/H analysis was collected. (b) Proportions of sulfide at day 15 accounted for by methane‐dependent sulfide production, methane oxidation‐linked sulfide production, and non‐methane‐dependent sulfide production Copyright © 2017 Wiley Periodicals, Inc. Cell abundance counts revealed decreased biomass on both carbon cloth and zircon beads in comparison to sediment‐based incubations (Supplementary Table S4): cell abundances in zircon bead incubations were 46.2% +/− 15.3% SD that of their sediment‐hosted inoculating communities, while carbon cloth incubations hosted 43.2% +/− 14.1% SD as many cells. For each temperature condition, zircon bead samples had the highest per‐cell sulfide production levels (using maximum sulfide levels attained during the incubations; Supplementary Figure S1). If longer incubation times could enhance biomass while retaining the same per‐cell activity, zircon beads may become competitive with carbon cloth as an artificial substrate. However, carbon cloth incubations had a higher proportion of surface‐associated cells (69.3% +/− 5.1% SD compared with 21.8% +/− 8.7% SD for zircon bead incubations; Supplemenatary Table S4), suggesting a specific adherence that enables easier transferability via substrate rather than a liquid phase. In addition, zircon beads are 11.3 times more dense than carbon cloth, offering diminished sulfide production on a per‐mass basis. To understand which microbial constituents could be responsible for the increased levels of methane oxidation and sulfide production under carbon cloth‐associated conditions, 16S rRNA gene surveys were conducted on three 35 °C carbon cloth samples and one 35 °C zircon bead sample. Compared with the 35 °C sediment incubations, the overall community diversity decreased in both carbon cloth (Inv. Simpson from 26.5 to 3) and zircon bead (10.5) incubations (Supplementary Table S2). The relative proportions of known sulfate reducers increased from 12.1% +/− 0.9% SE in sediment samples to 66% +/− 0.1% SE in carbon cloth‐ and 19.4% in zircon bead‐hosted communities. These results suggest that abstraction of a community from the sediment matrix reduces available niches and promotes dissolved substrate (e.g., sulfate) catabolism. The 35 °C carbon cloth incubations exhibited a particularly streamlined community: sequences corresponding to Desulfotomaculum geothermicum accounted for 55.6% +/− 1.5% SE of all recovered sequences, signifying a clear selection for a specific lineage. D. geothermicum can utilize a number of organic molecules as electron donors (Daumas, Cord‐Ruwisch, & Garcia, 1988 ), and members of the genus are associated with AOM at the Lost City hydrothermal field (Aullo, Ranchou‐Peyruse, Ollivier, & Magot, 2013 ) and Gulf of Mexico hydrate mounds (Mills, Hodges, Wilson, MacDonald, & Sobecky, 2003 ). The second most prevalent lineage was an unclassified Marinobacter representative, present at 7.9% +/− 1.7% SE; these organisms have been implicated in denitrifying hydrocarbon oxidation (Rontani, Mouzdahir, Michotey, & Bonin, 2002 ). Microscopic imaging and cell counts using a general DNA stain and an ANME‐1/2‐specific FISH probe revealed that ANME‐1 and/or ANME‐2 were relatively prevalent, accounting for an estimated 26.2% +/− 8% SE of all microbial constituents in the 35 °C carbon cloth incubations. 3.5 Microscopic analyses reveal distinct ANME‐SRB arrangements Microscopic imaging of sediment and carbon cloth samples using FISH demonstrated distinct spatial associations among SRB and ANME‐1 and ‐2 constituents. In sediment samples, ANME and SRB cells form intermixed consortia in which syntrophy is likely maintained by cytochrome‐mediated direct electron transfer (McGlynn, Chadwick, Kempes, & Orphan, 2015 ). These aggregates were prevalent in 35 °C sediment incubations (Figure 5 b). In 35 °C carbon cloth incubations, constituents were observed in association with the surfaces of individual fibers (Figures 5 e and 5 f). A selected area of carbon cloth with a relatively high microbial load revealed an abundance of SRB and a patchy distribution of ANME‐1/2 cells (Figures 5 g and 5 h). Un‐incubated carbon cloth subjected to identical hybridization conditions produced no visible fluorescence signal (data not shown). Figure 5 Cell abundances and representative images of microbial communities in seep sediment and adhered to artificial substrates. (a) Cell counts of seep sediment inoculum and from polyurethane filter and carbon cloth surface‐associated communities. Incubations lasted 27 days and were conducted with or without methane, as indicated, at 35 °C. Cell abundance values indicate the number of DAPI‐visible cells per cm 3 ; percentages and partition lines indicate the proportion of DAPI‐active cells that were illuminated by the ANME‐932 or DSS‐658 FISH probes. (b–h) Representative FISH/reflectance images of the respective substrates. (b) Sediment from a 35 °C batch incubation recovered after 72 days. (c–d) Polyurethane filter recovered after 27 days, exhibiting substantial particulate but minimal biomass adherence. (e–h) Carbon cloth from a 35 °C batch incubation recovered after 44 days. Panels e and f show single strands of carbon cloth, with a moderate white light reflectance component in panel f to show the strand outline. Panels g and h show composite images from a 94 µm thick z ‐stack, with vertical steps of 0.8 µm. Panel G includes a reflectance channel to show the topology of the carbon cloth; the reflectance channel is removed in panel h to reveal sites of successful hybridization. A control sample of unincubated carbon cloth exhibited no probe hybridization (images not shown). See text for microscope parameters. In all images, the FISH probes ANME‐932‐Cy3 (red stain), and DSS‐658‐6FAM (green stain) were used; regions of apparent yellow coloration indicate nearly co‐localized red and green stains. The scale bar in panel c applies to panels c and d, the scale bar in panel f applies to e and f, and the scale bar in panel h applies to g and h Copyright © 2017 Wiley Periodicals, Inc. The carbon fibers may provide conductive, electron‐transferring contact between ANME and SRB, which would preserve the redox‐tuned symbiosis but obviate the need for direct contact, thereby enabling more streamlined nutrient delivery and waste removal. The partial segregation between lineages on the carbon cloth, the detection of direct electron transfer between ANME and SRB (McGlynn, Chadwick, Kempes, & Orphan, 2015 ), as well as previous work revealing carbon cloth as an electron‐carrying intermediary for G. metallireducens and M. barkeri (Chen et al., 2014 ), make this an intriguing possibility worthy of continued investigation. 3.6 Exploring the properties of artificial substrate adherence and enrichment While microscopic analysis of the carbon cloth revealed a substantial surface‐associated constituency, zircon bead incubations showed abundant organisms in the filtered supernatant but no surface‐associated organisms. This observation revealed the importance of (a) textural and/or (b) electrical properties in facilitating ANME‐SRB partnerships on artificial substrates. To assess the relative importance of these two variables, seep sediment was incubated (+/− methane) with carbon cloth and polyurethane filter, which mimics the high porosity and surface area of carbon cloth but is non‐conductive. After 27 days of incubation at 35 °C, cell counts revealed a more than fivefold increase in microbial abundance on carbon cloth compared with polyurethane filter under methane‐replete conditions (Figure 5 a), and negligible quantities of ANME and SRB cells were visible on polyurethane (Figures 5 c and 5 d). Approximately half of this increase is likely methane‐independent (the carbon cloth hosted 1.7 times as many cells than the filter under methane‐free conditions), but most of the enhancement was attributable to the addition of methane. Furthermore, carbon cloth incubations in general, and the methane‐rich carbon cloth incubation in particular, favored a selective, substrate‐associated enrichment of ANME (ANME‐1/2) and SRB ( Desulfosarcina / Desulfococcus ) constituents. While 40% of observed biomass in the inoculum seep sediment consisted of these AOM‐implicated lineages—a proportion roughly maintained on the polyurethane filters—51% and 73% of attached cells consisted of these lineages on carbon cloth under methane‐free and methane‐rich conditions, respectively (Figure 5 a). Electrical resistance values confirmed that under incubation conditions, carbon cloth was more than four orders of magnitude more conductive than sediment, zircon beads, and polyurethane filter (Supplementary Table S5). These findings suggest that the electrical conductivity of carbon cloth is a key factor in its capacity to enrich ANME/SRB and enhance methane‐fueled sulfide production. Recent microbial fuel cell studies are consistent with this interpretation, as conductive carbon substrates have been used to stimulate electricity generation from methane in ANME enrichments (Chen & Smith, 2018 ; Ding et al., 2017 ; Zhao, Ji, Li, & Ren, 2017 ). Previous efforts with sediment‐free ANME‐SRB enrichments have focused largely on the physiological relationships between consortia members (Krukenberg et al., 2016 ; Laso‐Pérez et al., 2016 ). The resulting mixed cultures can take years to establish and have diminished methane‐dependent sulfate reduction rates (Wegener, Krukenberg, Ruff, Kellermann, & Knittel, 2016 ). Our effort prioritizes enhanced function (methane consumption and methane‐dependent sulfide production) and sediment removal over community purity, and offers compelling progress on these fronts. Our analysis shows not only that abstraction from the sediment matrix is possible with a complex AOM‐enacting community over the course of weeks rather than years, but that methane‐dependent sulfide production per unit volume increases by a factor of 3.5 on the carbon cloth artificial substrate. (This value was calculated using 44‐day and 19‐day time points and methane‐dependent sulfide production coefficients of 0.74 (Figure 1 ) and 0.83 (Figure 3 b), for the 35 °C sediment and carbon cloth incubations, respectively). Harvesting of such uncultivable communities onto artificial substrates is a promising method for expanding the possible functionalities of biotechnological systems. 3.7 Applications of microbially mediated methane activation The findings presented above signify important developments in the use of methane for bioindustrial practices. Optimized salinity and temperature conditions inform reactor conditions and set bounds on operational costs. Abstracting AOM activity from the sediment matrix obviates the need for a continuous supply of costly deep‐sea inoculum and avoids obstacles to nutrient delivery and fluid flow. Developing an inexpensive medium, devoid of costly organic amendments or electron shuttles, is also beneficial to reducing operating costs. Continued microbial “auto‐refinement,” taking advantage of the natural tendency of microbial communities to assemble based on the most efficient metabolism of provided energy sources, will likely yield a highly predictable—and more productive—assemblage for methane bioprocessing. Once mobilized by uncultivated ANME, methane‐derived electrons can be incorporated into a range of systems. SRB have been used for bioremediation (Chardin et al., 2002 ; García et al., 2001 ; Hard, Friedrich, & Babel, 1997 ) and can accumulate intracellular storage molecules (e.g., polyhydroxyalkanoates, or PHAs, Hai, Lange, Rabus, & Steinbüchel, 2004 ), which are attractive precursors for plastics (Verlinden, Hill, Kenward, Williams, & Radecka, 2007 ; Wang, Yin, & Chen, 2014 ). Existing knowledge of PHA metabolic pathways (Aldor & Keasling, 2003 ; Poblete‐Castro et al., 2013 ) and the genetic tractability of several SRB lineages make enhanced bioplastic production a compelling opportunity, and the direct transfer of methane‐derived electrons to SRB would likely increase rates of metabolic activity. Using metabolic products of the ANME‐SRB consortium, dissolved inorganic carbon can be mobilized by engineered autotrophs into alcohol fuels (Lan & Liao, 2013 ) while sulfide could be used as an electron donor to power engineered, chemical‐synthesizing strains of sulfide oxidizing bacteria (Kernan et al., 2016 ). These applications would also limit the atmospheric release of methane, a greenhouse gas with a global warming potential 21 times that of carbon dioxide over a 100‐year period. Despite these opportunities, challenges remain in the design and scale‐up of distributed bioreactors. Corrosion associated with unreacted sulfide is a longstanding challenge that has attracted substantial research (e.g., Revie, 2008 ); in addition to incorporating appropriate reactor materials, avoiding corrosion will require careful tuning of flow rates and metabolic activity within secondary reactors. Initial analysis has demonstrated that using natural gas wells as the reactor can overcome mass transfer obstacles and reach 100% conversion with sufficient reaction rates (Emerson et al., 2017 ), though such rates were not attained in this study, and the prospects of deploying high‐surface area carbon cloth at scale requires further investigation. Nonetheless, the work presented here offers an initial avenue toward rate enhancement per unit volume, and we anticipate that continued improvements of the electrical and physical properties of artificial substrates, as well as incorporation of recently discovered microbial communities oxidizing methane at substantially higher rates ( Marlow et al., in prep ; Raineault et al., 2017 ) will bridge the gap."
} | 9,862 |
28059105 | PMC5216391 | pmc | 1,064 | {
"abstract": "A benthic microbial electrochemical systems (BMES) of 195 L (120 cm long, 25 cm wide and 65 cm height) was constructed for sediment organic removal. Sediment from a natural river (Ashi River) was used as test sediments in the present research. Three-dimensional anode (Tri-DSA) with honeycomb structure composed of carbon cloth and supporting skeleton was employed in this research for the first time. The results demonstrated that BMES performed good in organic-matter degradation and energy generation from sediment and could be considered for river sediments in situ restoration as novel method. Community analysis from the soil and anode using 16S rDNA gene sequencing showed that more electrogenic functional bacteria was accumulated in anode area when circuit connected than control system.",
"discussion": "Results and Discussion TOC and TN variation in the sediment The initial total nitrogen 13 and the TOC of the sediment was 3.1 and 33.7 g kg −1 in dry sediment base ( Table S1 ). For the BMES reactor, the TOC was decreased by 5.0% during the first 15 days of operation, and then the TOC removal was 5.6% during the second 15 days, a relative lower removal of 3.6% was obtained during the third 15 days, while it only decreased by 0.3% during the last 15 d (Total decreased by 14.5%) ( Fig. 1 ). The BMES kept a high TOC removal during the first 45 days before the removal rate sharply declined (there was no obvious change during the last 15 days) ( Table S2 ). During the whole operation, the TOC removal efficiency of BMES (14.5%) was 1.2- and 6.9-fold in comparison with the S Control (11.6% removal) and W Control (only 2.1% removal). It indicated that the flushing method (W Control), which was usually used for sediment restoration has little effects on the organics removal. From the experimental data obtained here, it is obviously that BMES demonstrated good capability of simultaneous organic matter removal as reported 14 15 . TN decreased by 3.2% during the first 15 days of the BMES operation, and then decreased by 1.6% during the second 15 days, a removal of 12.7% was obtained during the third 15 days, while it only decreased by 1.0% during the last 15 d (Total decreased by 18.5%). But, during the same period, there were no significant changes in S Control (only 1.9% removal) and W Control (1.0% removal) during the whole operation ( Table S3 ). The TN removal rate of the BMES was 8.7− and 18.2− fold than that of S Control and W Control. PAHs removal performances In total 12 kinds of PAHs were detected in the present sediments ( Table 1 ). The concentrations of both Benzo(a) pyrene (BaP) and Benzo (k) fluoranthene (BkF) (five-ringed PAHs) was higher than 12 mg/kg, accounting for 60.35% of total PAHs (TPAHs) in the initial sediments ( Fig. 2 ). It is probably because low molecular weight PAHs, containing two-to three-ringed PAHs were less toxic to microbes in the soil and can serve as a carbon source involved in the microbial metabolisms and accumulated more in the sediment 8 . Also high molecular weight PAHs, containing four-to six-ringed PAHs were hard to be decomposed by microbes. The total contents of BaP during the 60 days’ operation were decreased by 50%, 37.4% and 30.8% in BMES, S Control and W Control respectively and BkF decreased by 50%, 28% and 21.8% in BMES, S Control and W Control. More important, BMES has higher PAHs removal efficiencies than that the S Control and W Control (1.4 fold PAH removal than S Control and 1.8 fold than W Control). Complex organic compounds, such as PAHs decomposition might need a long time leading to inhibition of TOC further removal in the sediment 16 ( Fig. 1 ). Sorption of PAHs on natural sorbents, like sands, sandy loam soils, and silt loam soils are usually normal process and is regarded as dynamic fast step compared to the biodegradation process 17 . Former research also showed that dissolved organic matter has an positive effect of PAHs sorption and induced PAHs contents are usually proportion to the dissolved organic matter 18 . PAHs in the present system was expected to be attached on the anode surface firstly and then decomposed by the anode bacteria in BMES. But in BMES, with the microbes enrichment on the anode, PAH sorption/desorption hysteresis declined, due to EPS production of anode bacteria 19 . It is obvious that the degradation was accelerated in BMES owing to the promotion by the current generated in BMES. It can also be concluded that washing flushing has no extra effects on the PAHs removal. The attached PAHs on anode was then further degraded by the electrogensis bactetria. TOC and TN changes in the water layer of BMES Total 135 L tap water (TOC 1.9 mg/L) was used for the water layer, the TOC contained in the sediment began to be flushed and released into the water layer. The initial TOC concentration in water was 60 ~70 mg L −1 , yet was reduced to <30 mg L −1 in the first 3 days in the three research systems ( Fig. 3 ). Resettling of pollutants to sediments 20 21 might be the main reason for the declination of TOC in the first 3 days, yet some released soluble organics from sediments remained in the water layers. In the followed 5 d to 25 d, the TOC of the BMES and S Control had similar trend. After 60 days’ operation, the TOC of the BMES was approximately half than that of the S Control. The results indicated that the deployment of electrode as electron acceptor in BMES was helpful to better removal of organic pollutants than natural biodegradation in S control. The lowest TOC contained in water layer was achieved by W Control which was washed by fresh water in a continuous flow rate of 400 L h −1 . After 135 L tap water (TN 1.9 mg/L) was used for the water layer, the initial concentration of TN in the water was 39 to 40 mg L −1 for the flushing and also the desorption. During the first 3 days, TN content in BMES and W Control dropped sharply, and in the later 30 days experiments, TN removal rates in BMES and W Control slowed down and kept at 7.47~13.3 mg L −1 and 1.1 ~3.2 mg L −1 respectively. In the first 3 days, TN in S control was at 39–45 mg L −1 and then at 17–28 mg L −1 in the following experimental period. DO, pH and EC changes in the water layer of BMES As the reaction proceeded, the DO in the BMES, S Control and W Control maintained a relatively stable range (4.3 ± 0.082 mg L −1 , 2.6 ± 0.21 mg L −1 and 6.5 ± 0.1 mg L −1 ). Higher DO in W Control should be related to the flushing water and keep the DO concentration similar to the DO in normal water ( Fig. 4 ). The DO in BMES was 2 times higher than that in S control. In this study, changes in pH were observed for the entire duration. The experimental results clearly indicate that the pH of three reactors showed similar performance, and all were stable at 6.8–7.4. For the BMES, water conductivity was relatively constant at 135 μs cm −1 , which was similar to that in the SMFC systems using fresh water as the electrolyte. This indicates a relatively low conductivity compared to the BMFC reported previously 22 23 24 . The decomposition of organic would usually have robbed the dissolved oxygen from water which might deteriorate water quality. It is significant that the higher DO in BMES meant less DO was consumed in TOC or TN removal in BMES. When the EC of the three reactors were steady, W Control was 4- and 5-fold lower than the BMES and Control, respectively. It is significant that EC had the same trend with the TOC in the surface water. The difference on conductivity reduced with the decrease of TOC 25 . Power generation in BMES Under the closed circuit (CC) mode, an average voltage of 90 mV was initially achieved in the BMES, although some voltage fluctuations were observed ( Fig. 5A ). The voltage decreased to 65 mV on day 11. In the present study, the internal resistance evaluated by the polarization slope method was about 40 Ω. This showed that the scale of mass transport limitations influenced the BMES performance, predominantly at the anode electrode. The potential and power density, as a function of current density, were obtained as shown in Fig. 5B . Based on the power curve, the maximum power density for the BMES was 81 mW m −2 at day 20, which is relatively higher compared with that in previous studies 26 . For the BMES, the anode potential varied over a narrow range of −410 to −440 mV, while the cathode potential covered a much wider range of 240 to 180 mV, possibly indicating that its current was limited by the cathode or external resistance ( Fig. 5C ). A power management system (PMS) was designed that enables BMES to drive 9 red LEDs in parallel 27 . Six independent capacitor-based circuits were used to harvest electrical energy from the BMES. Each circuit consisted of capacitors (3.3 F, Panasonic Corporation, Japan) and relays controlled by programmable microcontroller (XD-J16H, Xunda Corporation, China). The capacitors in each circuit were charged in parallel by the corresponding module (1 min) and then charged capacitors connected in series to discharge (1 min) ( Movie S1 ). Our results demonstrate that BMES can be a viable alternative renewable power source ( Fig. S2 ). There are also strong correlations between the current output of a simple anode-resistor-cathode device and rates of anaerobic microbial activity (TOC content) in a diversity of anoxic sediments 28 . At the initial stage of deployment, the BMES could obtain high voltage outputs, but over a long operational period, nutrient depletion, mass transfer resistance, and anode/cathode fouling, could lower the system performance 29 30 . Several factors, such as spatial and temporal variability, natural fluctuations, temperature changes, pressure variations, water flow changes, salinity and conductivity changes, and dissolved oxygen affect the BMES functions 29 . The kinetics of electron transfer from microorganisms to the anode were mainly restricted by the anode potential. Overall, the BMES revealed a high cathode potential, the lowest anode potential, and the highest cell potential (189–591 mV) 7 . Community analysis Analysis of the anodic bacterial community using 16S rDNA gene pyrosequencing revealed the enrichment of genera with potential exoelectrogenic capability 31 . And this has significant implications in determining the microbial community structure in SMFCs in each aquatic environment and consequently the fate and removal kinetics of organic pollutants in contaminated sediments 32 . The Shannon diversity index provided the species richness (i.e., the number of species present) and distribution of each species (i.e., the evenness of the species) among all the species in the community 33 . BMES had the highest diversity (Shannon = 7.07) that was slightly larger than that of Initial (Shannon = 6.87) and S Control (Shannon = 6.50) among the 4 communities, while S Control had the lowest diversity (Shannon = 6.09). Based on the sequencing results ( Fig. 6 ), the communities analyzed Initially composed of Gammaproteobacteria (50.38%), Bacteroidetes (15.33%), and TM7 (7.37%). Bacteroidia (18.8%), Proteobacteria (32.05%), Chloroflexi (21.4%), and Firmicutes (20.75%) were detected in the S Control. W Control was colonized by Firmicutes (37.93%), Proteobacteria (28.27%), and Chloroflexi (13.9%). For the BMES, Bacteroidia (14.1%), Proteobacteria (38.4%), Chloroflexi (21.04%), and Firmicutes (9.22%) were detected. Deltaproteobacteria was also believed to be responsible for the electron transfer to the electrode and were also detected in Initial, BMES, S Control, and W Control with 1.95%, 12.26%, 4.89% and 11.89% dominance, respectively, which differs slightly from the values reported in two previous studies at 10% and 70%, respectively 12 34 . Geobacter was also detected abundantly in these soil samples, representing approximately 0.07%, 4.94%, 0.16% and 0.12% of the total bacteria in Initial, BMES, S Control, and W Control respectively. The ratios of Geobacteraceae sequences were previously reported to have increased from 0.13 to 0.74% in response to an increase in current density in rhizosphere MFC 35 . Here, Geobacter sequences were more abundant in anode-associated soil than in bulk soil, suggesting that they were involved in electricity generation in the BMES. Earlier studies have observed that Gammaproteobacteria and Bacteroidetes occurred at high numbers within libraries from electrode biofilms, in which the Gammaproteobacteria were believed to be responsible for electron transfer and consequently power generation 23 34 . Flavobacteria and Bacteroidia have earlier been identified as the dominant bacteria for electricity generation 23 36 ( Fig. 6 ). The microbial community composition observed in this study shows that the microbial community in BMES was dominated by exoelectrogenic bacteria 37 38 . A number of less prevalent bacteria were also detected ( Fig. 6 ) 39 . The variety of the bacterial colonies in sediments is attributed to the process of enrichment of electricity bacteria in anodes 36 ."
} | 3,263 |
34081399 | PMC8313253 | pmc | 1,065 | {
"abstract": "Summary Environmental and host‐associated microbial communities are complex ecosystems, of which many members are still unknown. Hence, it is challenging to study community dynamics and important to create model systems of reduced complexity that mimic major community functions. Therefore, we developed MiMiC, a computational approach for data‐driven design of simplified communities from shotgun metagenomes. We first built a comprehensive database of species‐level bacterial and archaeal genomes ( n = 22 627) consisting of binary (presence/absence) vectors of protein families (Pfam = 17 929). MiMiC predicts the composition of minimal consortia using an iterative scoring system based on maximal match‐to‐mismatch ratios between this database and the Pfam binary vector of any input metagenome. Pfam vectorization retained enough resolution to distinguish metagenomic profiles between six environmental and host‐derived microbial communities ( n = 937). The calculated number of species per minimal community ranged between 5 and 11, with MiMiC selected communities better recapitulating the functional repertoire of the original samples than randomly selected species. The inferred minimal communities retained habitat‐specific features and were substantially different from communities consisting of most abundant members. The use of a mixture of known microbes revealed the ability to select 23 of 25 target species from the entire genome database. MiMiC is open source and available at https://github.com/ClavelLab/MiMiC .",
"conclusion": "Conclusion Synthetic communities are very important experimental models to study microbiomes but very few approaches have been developed for their design. We share a bioinformatic tool to create synthetic community compositions that mimic the functional prokaryotic potential within input metagenomes, which we hope will facilitate experimental studies in many fields of microbiology and biotechnology.",
"introduction": "Introduction Microbial communities are ubiquitous and influence many fundamental processes ranging from carbon and nitrogen cycles in water and soil to health and disease regulation in host‐associated habitats (Thompson et al ., 2017 ; Vujkovic‐Cvijin et al ., 2020 ). A major bottleneck for the study of these communities is the vast number of microbes that are still unknown (Hug et al ., 2016 ). This prevents accurate assessment of community dynamics and interactions with the environment. Moreover, the tremendous complexity of these communities, due to hundreds of members and the possible interactions between them, renders the task of understanding how they establish and function very difficult. Hence, being able to design simplified communities of microbes as proxy for the native community of interest is important, albeit not an easy task. Such simplified (or synthetic) communities can be used in modelling or experimental approaches (e.g. continuous culture or gnotobiology) to highlight fundamental concepts underlying relationships between community members, evolutionary processes, or mechanisms of interactions with environmental factors or host species (Payne et al ., 2012 ; Brugiroux et al ., 2016 ; Bauer et al ., 2017 ; Noronha et al ., 2019 ; Tanoue et al ., 2019 ; Streidl et al ., 2021 ). A variety of examples of such synthetic communities have been published for several habitats, including plant roots (Armanhi et al ., 2017 ; Niu et al ., 2017 ; Vorholt et al ., 2017 ; Herrera Paredes et al ., 2018 ; Zhang et al ., 2019 ), soils (Kleyer et al ., 2017 ; Puentes‐Tellez and Falcao Salles, 2018 ; Zhalnina et al ., 2018 ) and gastrointestinal tracts (Schaedler et al ., 1965 ; Becker et al ., 2011 ; Petrof et al ., 2013 ; Brugiroux et al ., 2016 ; Calatayud Arroyo et al ., 2018 ). Experimental approaches to build minimal communities include incremental, function‐driven selection of taxa in vitro or in vivo (functional enrichments) or tailored assemblage of previously isolated axenic strains (community assembly) (Clavel et al ., 2017 ). Functional enrichment has the advantage of directly providing communities that carry out the desired function, as published already in the context of plant fitness or the induction of specific immune responses (Atarashi et al ., 2013 ; Herrera Paredes et al ., 2018 ; Stein et al ., 2018 ), but they are experimentally demanding because of the necessity to test the given function after each round of enrichment. In contrast, the drawback of community assembly, the approach that is most commonly followed, is that the selection of strains is knowledge‐driven, i.e . based on ease of cultivation, availability of genomic information, an educated opinion on phylogenetic diversity, known functions and occurrence of taxa in the ecosystem of interest. Developing methods towards data‐driven design of synthetic communities of microbes would open new avenues by providing tailored synthetic community compositions fitted to the specific need of individual studies. Considering functional redundancy within complex microbial ecosystems, i.e. several community members carry out the same given function (Tian et al ., 2020 ), favouring minimal communities that best represent the array of functions expressed rather than the original taxonomic profile of complex microbial communities is a sound objective (Johns et al ., 2016 ; Eng and Borenstein, 2019 ; McCarty and Ledesma‐Amaro, 2019 ). In this context, the present study aimed at creating and benchmarking a bioinformatic pipeline, called MiMiC, for automated prediction of synthetic community compositions mimicking the functional repertoire of the input microbial ecosystems.",
"discussion": "Results and discussion Overall concept of MiMiC The rationale behind MiMiC is to infer the composition of a synthetic community of prokaryotes based on individual metagenomic profiles. Therefore, the pipeline processes (meta)genomic data into vectors of protein families (Pfams) used as the foundation for an iterative scoring process to determine a short list of best matching genomes from a comprehensive database. A schematic overview of the pipeline and tools used can be seen in Fig. 1A , experimental details are provided in the methods section, and all scripts and data are accessible via the project‐specific repository: https://github.com/ClavelLab/MiMiC . The current genome database consists of 22 627 species‐level genomes from bacteria and archaea spanning a total of 53 phyla and representing an average of 2523 ± 452 Pfams each (Fig. 1B ). Ecosystem‐specific genome databases can also be used, as currently included for the human ( n = 803) (Zou et al ., 2019 ), mouse ( n = 104) ( www.dsmz.de/miBC ) (Lagkouvardos et al ., 2016 ) and pig intestine ( n = 111) ( www.dsmz.de/pibac ) (Wylensek et al ., 2020 ). After selecting a minimal number of genomes (either pre‐set by the user or determined in silico as detailed in the methods) from the database by maximizing the ratio of matches‐to‐mismatches to the input metagenome, MiMiC returns their NCBI RefSeq genome accession numbers along with various genome‐derived statistics. Running MiMiC with 50 iterations against the entire reference genome database using a metagenomic Pfam binary vector as input returned results within an average of 11 min by a computer system with 32GB RAM and 12 cores operating on linux (x86_64‐pc‐linux‐gnu) and using R version 3.6.3 (2020‐02‐29). Fig. 1 Schematic overview of the architecture and content of MiMiC. See methods section for all details. A. (Meta)genomic reads are quality‐checked and processed into binary vectors of protein families (Pfams). The resulting metagenomic profile (MG‐1) is used for iterative selection of a minimal number of genomes (calculated by a knee point approach) from the database (B‐1 to B‐n) that best cover the functional potential of the input data (highest number of matches; least number of mismatches). B. The genome database currently consists of 22 627 species‐level genomes of archaea and bacteria spanning a total of 53 phyla, with an average of approximately 2500 Pfams per genome. Functional metagenomic profiles to infer synthetic community composition In order to test MiMiC, we processed a set of 937 shotgun metagenomes from six different microbial habitats: marine water, soil, human tongue and the intestine of humans, mice and pigs (see the methods for details). The rationale was to select shotgun profiles representing very distinct (e.g. environmental vs. host‐associated) but also more closely related ecosystems (host‐specific microbiomes) obtained from published studies. Expectedly, multidimensional analysis of functional profiles from these metagenomes showed a clear distinction between environmental and host‐associated microbiomes (Fig. 2A ). The generated binary Pfam‐based profiles also allowed to differentiate gut microbiomes from the three host species analysed (Fig. 2B ). A prerequisite for inferring the composition of simplified consortia from complex native communities is the ability to cover a sufficiently high fraction of metagenomic functions using reference genomes. For all ecosystems, the median metagenomic Pfam coverage by the entire database was close to 100% (Fig. 2C ). While this covered fraction decreased when using the host‐specific database for gut microbiomes, all median values remained > 90% and all single values > 75%. Fig. 2 Input metagenomes from six different types of complex microbial communities. A. Multi‐dimensional plot based on the presence/absence of Pfams (Jaccard index) in each individual metagenome (dots) from two environmental (marine water and soil) and four host‐associated metagenomes: the tongue and gut of humans (H) and the gut of mice (M) and pigs (P). The number of computed metagenomes per environment is indicated next to the corresponding cluster of samples. P < 0.01 as tested by permutational multivariate analysis of variance using distance matrices with the function adonis in R. B. Same as in A showing further host species delineations between gut microbiomes. C. Coverage of all metagenomes per habitat category (i.e. percentage of metagenomic Pfams also present in the reference MiMiC‐processed genomes) by the entire database (DB) or host‐specific collections of genomes for the human, mouse and pig intestine. Altogether, these data show that binary Pfam‐based annotation of shotgun metagenomes retained enough resolution to distinguish even closely related ecosystems and the established genome databases can be used as a robust foundation for synthetic community design. Features of the inferred minimal consortia MiMiC predictions were performed on all 937 individual metagenomes aforementioned. Knee point calculation (see methods) to determine the optimal number of species within the inferred synthetic communities after 50 iterations revealed lowest ( n = 7) and highest ( n = 10) median diversity for soil and pig gut microbiomes respectively (Fig. 3A ). The cumulative metagenomic coverage by the inferred synthetic communities was > 80% for all habitats, with a sharp increase in function coverage (up to 75% of metagenomic Pfams) observed already for the first four species selected (Fig. 3B ). Each individually generated synthetic community was then compared with 100 sets of randomly selected species for each metagenome, resulting in a total of 7200–27 100 random sets for each ecosystem. The cumulative functional coverage was always significantly higher (+10–15%) for MiMiC predictions vs. the random sets (Fig. 3C ). In contrast, the fraction of mismatches was significantly lower for MiMiC predictions in two cases (human tongue and gut), while equal for soil and mouse gut and higher for marine water and pig gut. Fig. 3 Output of MiMiC analysis on complex communities. A. Distribution of the number of species per minimal consortium according to knee point‐based calculation after 50 iterations. The number of samples considered is indicated below the name of each community category. B. Coverage of metagenomic profiles ( y ‐axis) according to incremental selection of genomes by MiMiC until sample‐specific knee points ( x ‐axis = rank of genome selection). C. Performance (% of matches and mismatches) of MiMiC outputs compared with 100 sets of an equal number of species randomly selected from the entire database for each individual metagenome. P ‐values were calculated using the Wilcoxon rank‐sum test; *** P < 0.001. The synthetic communities created were then assessed in terms of composition respective to their ecosystem of origin. Multidimensional binary plots of the presence/absence of selected species per individual community showed that the main difference between environmental (marine water and soil) and host‐associated ecosystems previously observed with the native metagenomic profiles was conserved after MiMiC predictions (Fig. 4A ). The distinction among host‐associated communities was less clear, with marked inter‐individual differences especially in the case of human tongue samples. Nonetheless, simplified communities from the pig intestine were most distinct from the others (Fig. 4B ). Ecosystem‐specific species were identified by reporting the 10 most prevalent species per habitat, e.g . those species that were most often selected across all input metagenomes for the given habitat (Fig. 4C ). This analysis showed no overlap between environmental and host‐associated ecosystems. A few common species were observed between habitats within each of these two main ecosystem types, while each type of communities was characterized by four to seven uniquely selected species. Strikingly, soil samples were represented by species with an overly low prevalence, indicating marked inter‐sample differences in composition, as illustrated in the native metagenomic profiles (Fig. 2A ) and the wide range of species diversity within the synthetic communities (Fig. 3A ). Fig. 4 Ecosystem‐specific synthetic communities. A. Multi‐dimensional plot based on the presence/absence of species (Jaccard index) selected by MiMiC within individual minimal consortia (dots), each corresponding to one input metagenome from six ecosystem types. The centroid of each type is indicated directly by the label or else by grey arrowheads within the label in case of overlaps. P < 0.01 as tested by permutational multivariate analysis of variance using distance matrices with the function adonis in R. B. Same as in A showing further host species delineations between gut microbiomes. C. Species identity (with sequence accession in brackets) of the 10 most prevalent genomes selected by MiMiC for each habitat type, prevalence being defined as the percentage of minimal microbial consortia containing the given species. The total number of metagenomes/minimal consortia considered per habitat category is shown below the x ‐axis. Species identities are written in habitat‐specific colours if unique and in black if shared between habitats. Comparison with alternative strategies for genome selection We further evaluated the relevance of the MiMiC selection strategy using the metagenomic data from the pig intestine ( N = 271), as we have recently established a comprehensive collection of gut bacterial isolates and corresponding genomes from pigs (Wylensek et al ., 2020 ), allowing host‐specific analysis and taxonomic annotation of the data. The standard MiMiC procedure described above was compared with synthetic communities selected based on (i) most abundant members (Most abund.), at a number of species equal to that selected by MiMiC (knee point method); (ii) genomes with the greatest number of additional Pfam matches with the input metagenome, i.e. mismatches were not considered (m‐only); and (iii) excluding the first selected genome (skip 1st). The metagenomic Pfam fraction covered by the respective synthetic communities was highest with the ‘m‐only’ approach, albeit at the expense of a substantial proportion of functions not present in the original metagenomes (> 20%), justifying the consideration of mismatches in the selection process (Fig. 5A ). The fraction of mismatches was also significantly higher for the ‘Most abund.’ and ‘Skip 1st’ strategies, although the magnitude of differences to MiMiC was much lower than for ‘m‐only’ and the metagenomic coverage was slightly higher for the ‘Most abund.’ strategy. These subtle changes in terms of percentages of Pfams translated into more evident differences when looking at the diversity of top‐10 most prevalent species selected across all 271 metagenomes by each method (Fig. 5B ). MiMiC selected only 6 of the 10 most abundant taxa, highlighting the functional input of lower abundant taxa. Species selected on the basis of matches only were drastically different from the other approaches (in particular, no species in common with MiMiC), likely due to the preferred selection of functionally pluripotent species. Skipping the species most often ranked no. 1 (i.e. for which the genome was most often selected first), namely Streptococcus alactolyticus , favoured the selection of the species Roseburia porci within the top‐10 species. The little differences observed between these two methods (MiMiC Vs. Skip 1st) in terms of coverage values (Fig. 5A ) highlight that overall Pfam profiles can be compensated between several species with a synthetic community. This is further discussed below in the section ‘Recommendations and outlooks’. Fig. 5 Comparison with alternative strategies for genome selection. Data were generated using the 271 pig metagenomes (Xiao et al ., 2016 ) and genomes from the pig intestinal bacterial collection (Wylensek et al ., 2020 ). A. Fraction of matches (m) and mismatches (mi) within the synthetic communities as a percentage of metagenomic Pfams for each of the following four methods: (i) MiMiC (see detailed description in the methods; (ii) synthetic communities based on most abundant members (Most abund.), at a number of species equal to that selected by MiMiC (knee point method); (ii) calculated using only Pfam matches between any genome and the input metagenome, i.e. mismatches were not considered (m‐only); (iii) excluding the first selected genome (skip 1st). B. List of species selected by the different methods. The 10 most prevalent species, i.e. most often selected by the given method across all input metagenomes, are shown. The bars indicate their respective prevalence (% of 271). The relative abundance of each species is shown as logarithmic values of RPKM (reads per kilo base per million mapped reads). Benchmarking against a reference community To assess the performance of MiMiC further, we used metagenomic data from the mock community MBARC (Singer et al ., 2016 ), which consists of 23 bacterial and 3 archaeal species. Iterations were run against the entire genome database ( n = 22 627; including the target genomes) until full functional coverage was reached, which happened at a number of 68 species (Fig. 6A ). Knee point determination returned a number of 25 target species, which indirectly confirms the relevance of this approach, as close to the number of reference taxa within this community. Of note, one of the 26 target species, namely Nocardiopsis dassonvillei , was absent from the input metagenomic reads and can thus not be considered as a community member within the dataset tested. Comparison of cumulative functional coverage and numbers of mismatches at the theoretical number of species ( n = 26) between the MiMiC prediction and 100 sets of randomly selected species confirmed the superiority of targeted community design (Fig. 6B ). Interestingly, from the MiMiC species‐rank (i.e. the order at which genomes are selected) 19 onwards, we observed an increased amount of both matches and mismatches in the MiMiC prediction (blue dots) compared with the random sets (black box plots). Four of the seven species selected during the late iterations corresponded to species characterized by a low relative abundance within the MBARC metagenome analysed (Fig. 6C ), suggesting that such taxa contribute to higher mismatch values due to their incomplete occurrence within metagenomic reads. Altogether, of the 25 target species present within the dataset, 23 were selected by MiMiC (Fig. 6C ). Fig. 6 Mock community analyses. A. Functional coverage and number of mismatches across the total iterations required to cover all Pfams from the input reference metagenome. The dashed red line represents the knee point determined on functional coverage. B. MiMiC output (red dots) against 100 sets of randomly picked genomes until the number of 26 taxa. C. Strain identification (with sequence accession in brackets) of the target species within the mock community sorted in decreasing relative abundance within the shotgun sequencing data ( x ‐axis, in %) as reported in the original paper (Singer et al ., 2016 ). The rank (i.e. iteration) at which each species was selected by MiMiC is indicated on top of the bars. Recommendations and outlooks Despite its usefulness by enabling individualized synthetic community design based on functional profiles, the proposed approach has some limitations that we transparently present here and will be addressed in further versions of the tool. MiMiC is reference‐based; i.e. predictions are dependent on the quality of the genome database. However, the pool and diversity of genomes publicly available are growing exponentially and the reference dataset will be regularly updated, including microbes other than prokaryotes such as fungi (Richard and Sokol, 2019 ). Moreover, using ecosystem‐specific genome databases can help restricting predictions to those taxa most relevant for that ecosystem. Users may also modify the provided databases by applying their own taxonomic filters to restrict the search to relatively high taxonomic ranks contained in an input metagenome, e.g. families. This would narrow the pool of reference genomes to ecosystem‐relevant taxa while retaining enough diversity to not antagonize the concept of this tool: providing consortia mimicking ecosystem functions and not taxonomic profiles. This will, however, come at the cost of additional computation time required for the taxonomic annotation of metagenomic reads. The current genome selection approach is intrinsically limited by the generally low standard annotation power of genomes (without manual curation) and the functional resolution of Pfams, i.e. core functions outnumber specific functional features, inflating the possible functional redundancy between selected genomes. A future approach that considers the rareness of functions among microbes may better recapitulate the individuality of ecosystems. Determination of the number of species to be contained in the output synthetic community is not an easy task. For instance, the output of knee point calculation is influenced by the number of scoring iterations, which we recommend keeping below 50, as the steep increase in metagenomic coverage occurs across the first 10 selected taxa. Additional parameters beyond metagenomic coverage will also be considered in the future to determine the number of species within synthetic communities. As an alternative to sample‐specific determination of the number of species to be included in a synthetic community, users can also run the MiMiC script by setting a defined number of iterations corresponding to the wished number of species without considering knee point calculation. Over the last few years, two algorithms for designing minimal microbial communities have been published. The first aimed for desired metabolic capacities based on integer linear programming (Eng and Borenstein, 2016 ). While a step‐forward, this approach is limited due to complete metabolic pathways being targeted, which may be prevented by genome completeness and functional annotation quality (Parks et al ., 2015 ; Karp et al ., 2018 ). Moreover, communities of surprisingly low complexity were inferred using random selection of metabolite pairs (substrate product), and no information was provided on the complexity of usage and computation time. The second method aimed to generate communities that would be stable within a chemostat based on a user‐provided list of strains with known quorum sensing and bacteriocin production and sensitivity values (Karkaria et al ., 2021 ). This method was designed for application to engineered strains, for which such data would be known due to having been designed into the strains genome, and not for environmental isolates. Neither of the methods aim to generate ecosystem representative minimal communities, making MiMiC a unique addition to the existing suite of bioinformatic tools currently available. Future incorporation of systems biology approaches into MiMiC will provide in silico validation of community structure and species interactions prior in experimental testing (Bauer et al ., 2017 ; Venturelli et al ., 2018 )."
} | 6,271 |
35696568 | PMC9231504 | pmc | 1,066 | {
"abstract": "Significance Canonical quorum sensing is characterized as a cell–cell communication process that bacteria use to coordinate group behaviors. Diffusible signal molecules induce gene expression in response to cell density. Here, we describe a mechanism in the acyl-homoserine lactone signaling pathway of Pseudomonas aeruginosa that ensures this sensing and response at the group level. We show that two accessory proteins termed antiactivators prevent self-sensing, the cell-autonomous and density-independent reception of signals produced by the same cell. Self-sensing in turn generates population-wide bimodality in gene expression that may explain the response heterogeneity observed in other quorum-sensing bacteria. The ability to experimentally tune the sensing mode adds functionality to the design of cell–cell signaling circuits in synthetic biology.",
"discussion": "Discussion In this study, we have experimentally described self-sensing in a QS bacterial population, and we have characterized antiactivation as a governing mechanism that prevents it. We employed the well-understood AHL signaling system of the opportunistic pathogen P. aeruginosa as our experimental model. Self-sensing has been shown in the peptide-based QS system of B. subtilis , although the mechanism is not fully understood ( 13 ). Peptide signals are secreted into the environment and then sensed by a membrane receptor ( 33 ). How the signal producer gains preferential access to the signal in this system is unclear. In acyl-HSL signaling systems, the signals are produced intracellularly and generally freely diffuse across the cell envelope, although active transport is sometimes involved ( 4 ). The AHL signal is sensed intracellularly by a cytoplasmic receptor and transcriptional regulator, such that the potential for self-sensing is evident. Of note, the concept of “diffusion sensing,” in which an individual cell would perceive self-produced QS signals that accumulate in a confined extracellular space, is not considered self-sensing in this context ( 34 ). The P. aeruginosa acyl-HSL QS circuitry contains three antiactivator proteins that sequester the QS receptor LasR, thereby increasing the signal threshold necessary for induction. In our study, we focused on two antiactivators, QteE and QslA, that we previously found to exert the largest effect on QS gene expression ( 17 ). QteE reduces LasR protein stability, likely through direct protein–protein interaction ( 21 ). QslA in turn binds LasR to disrupt its dimerization and subsequent DNA binding but does not appear to reduce LasR protein stability ( 19 ). Moreover, qteE is activated by acyl-HSL QS, permitting negative feedback control ( 28 ). Using gfp reporter fusions, we initially measured the expression of four QS-controlled genes in the WT and the qteE qslA double mutant at the population level ( Fig. 1 ). We found that these genes respond differently to the lack of antiactivation. The reasons for this are not entirely clear and likely involve multiple factors during transcription initiation at target promoters, besides LasR binding affinity alone ( 27 ). As mentioned above, these factors include activation by the second acyl-HSL system ( rhl ) in the case of lasB ( 28 ) and negative feedback regulation by RsaL in the case of lasI ( 29 , 35 ). Next, we quantified the expression of lasI′-gfp in antiactivator single and double mutants at the single-cell level. To enable measurements at very low cell densities, we implemented a sampling procedure that concentrated large volumes of culture and a cultivation scheme that reduced preexisting GFP expression to background levels. Our flow cytometry data showed that successive deletion of antiactivators greatly reduces the induction threshold (with approximately 10-fold reduction in the qteE and qslA single mutants and approximately 100-fold reduction in the qteE qslA double mutant) and increases the proportion of constitutively active cells ( Figs. 2 and 3 ). Despite apparent mechanistic differences, qteE and qslA single mutants both showed similar QS responses. To prove that the observed patterns are caused by antiactivator-dependent self-sensing, we designed a cocultivation procedure of signal-proficient and -deficient cells distinguishable by different red fluorescent tags. While the antiactivator-proficient strain pair showed a near-simultaneous response, the antiactivator-deficient strain pair showed a substantial difference: Signal-proficient cells showed a much higher and earlier response than signal-deficient cells ( Fig. 4 ). Self-sensing has been proposed in several theoretical studies ( 8 – 12 ). If network parameters are such that the intracellular signal concentration exceeds the induction threshold, self-sensing occurs. One modeling study in particular investigated the transition between group-sensing and self-sensing in QS populations ( 11 ). Fujimoto and Sawai showed that cell-to-cell variability in gene expression can lead to heterogeneous populations in which some cells self-activate and others do not, if the intracellular signal concentrations are near the induction threshold. They further predicted that the proportion of self-activating cells gradually increases with cell density as the accumulation of extracellular signal contributes to activation. The similarities to our experimental results are striking (although their modeling results were obtained at steady state, whereas our gene expression measurements were obtained in dynamic batch culture). Our flow cytometry data show a heterogeneous, bimodal induction pattern for the antiactivator double mutant, with a rather gradual increase in the proportion of induced cells, in contrast to the unimodal and rapid induction pattern for the WT ( Fig. 2 ). Bimodality can emerge from a bistable QS system, which is what Fujimoto and Sawai assumed for their model, and which is plausible for the las system of P. aeruginosa ( 10 , 32 ). Intriguingly, heterogeneity in QS gene expression has recently been observed in several other bacterial species ( 7 , 36 – 40 ). While the underlying mechanisms are not clear, our work suggests that self-sensing may be involved. These observations also indicate that self-sensing is physiologically relevant. The emergent phenotypic diversity at the population level offers potential benefits, including bet-hedging in dynamic environments, or a division of labor in biofilm communities ( 41 , 42 ). Likewise, group-to self-sensing transitions may be relevant to the physiology of P. aeruginosa itself. There may be environmental conditions, for example nutrient stress, that promote self-sensing by enhancing signal production or lowering the induction threshold ( 43 ). Regulation may be at the level of LasI, LasR, or any of the antiactivator proteins. These mechanisms may also lead to a scenario where all cells in the population “self-sense,” which we did not observe here but which would occur if the intracellular signal concentration was significantly above the induction threshold. Other components that could contribute to limiting self-sensing include the third antiactivator, QscR, the transcriptional repressor, RsaL, or an as-yet-uncharacterized, fourth antiactivator. Our work suggests a function—in an evolutionary sense—for antiactivators in QS systems. Antiactivators not only tune the QS induction threshold but they also enable canonical group-sensing by suppressing self-sensing. From this perspective, it may become clearer why P. aeruginosa harbors three, rather than just one, antiactivator protein. All three might provide a partially redundant “fail-safe” mechanism to ensure group-level QS. Other, not necessarily mutually exclusive, functions are plausible. For example, the acquisition of one or multiple antiactivators could result in a “cheater” phenotype ( 44 ). These cells would exploit neighboring cells with no or fewer antiactivators that express QS-controlled secretions at a higher level. In certain contexts, antiactivation might be preferred over other mechanisms that could provide the same effect, such as a decrease in the affinity of the receptor to its signal, which in turn would be accompanied by a loss in signal specificity ( 45 ). Nevertheless, it is evident that not all QS systems require antiactivation to enable group-level signaling. The antiactivators thus far identified and characterized are QteE, QslA, and QscR in P. aeruginosa , as well as TraM and TrlR (TraS) in A. tumefaciens ( 12 , 22 , 46 ). In some strains of A. tumefaciens , two TraM-type antiactivators function in parallel QS systems ( 47 , 48 ). QscR and TrlR are LuxR homologs, whereas the other antiactivator proteins are unique, without any sequence similarity to each other. TraM homologs are found in Rhizobiaceae and Bradyrhizobiaceae ( 18 ). The structural and mechanistic diversity of antiactivators allows for different functional roles and suggests that novel antiactivator proteins remain to be discovered in other QS systems and species as well. Regardless of its prevalence, antiactivation has provided us with a tool to investigate the balance between self- and group-sensing in a QS population. This knowledge is of fundamental importance to our understanding of the functional capacity of QS systems and may find application in synthetic biology to design QS circuits with specific properties."
} | 2,358 |
29062949 | PMC5625728 | pmc | 1,067 | {
"abstract": "Microbial polyhydroxyalkanoates (PHA) have been produced as bioplastics for various purposes. Under the support of China National Basic Research 973 Project, we developed synthetic biology methods to diversify the PHA structures into homo-, random, block polymers with improved properties to better meet various application requirements. At the same time, various pathways were assembled to produce various PHA from glucose as a simple carbon source. At the end, Halomonas bacteria were reconstructed to produce PHA in changing morphology for low cost production under unsterile and continuous conditions. The synthetic biology will advance the PHA into a bio- and material industry.",
"introduction": "1 Introduction Polyhydroxyalkanoates (PHA), polylactic acid (PLA), poly(butylene succinate) (PBS), polyethylene (PE), poly(trimethylene terephthalate) (PTT), polypropylene (PPP), polyethylene terephthalate (PET) and poly(propylene carbonate) (PPC) are biodegradable or biobased polymers with at least one monomers produced by microbial conversions or microbial industrial biotechnology [1] . The weaknesses of microbial or enzymatic processes compared with the chemical processing make industrial biotech products less competitive with the chemical ones. However, taking advantages of the molecular biology and synthetic biology methods as well as changing process patterns, bioprocesses could be developed as competitive as chemical ones, these including the minimized cells, open and continuous fermentation processes et al. [2] This review aims to report progresses made by the China National Basic Science Research Project 973 during 2012–2016 on synthetic biology of PHA. Many bacteria have been found to produce various polyhydroxyalkanoates (PHA) biopolyesters. In many cases, it is not easy to control the structures of PHA including homopolymers, random copolymers and block copolymers as well as ratios of monomers in the copolymers. However, the weakening of beta-oxidation cycle in Pseudomonas putida and Pseudomonas entomophila led to controllable synthesis of all kinds of PHA structures including monomer ratios in random and/or block copolymers when fatty acids are used as PHA precursors. Introduction of functional groups into PHA polymer chains in predefined proportions has become a reality provided fatty acids containing the functional groups are taken up by the bacteria for PHA synthesis. This allows the formation of functional PHA for further grafting ( Fig. 1 ). The PHA diversity is further widened by the endless possibility of controllable homopolymerization, random copolymerization, block copolymerization and grafting on functional PHA site chains( Fig. 1 ) [3] . Fig. 1 Introduction of functional groups into PHA polymer chains in predefined proportions has become a reality provided fatty acids containing the functional groups are taken up by the bacteria for PHA synthesis. The PHA diversity is further widened by the endless possibility of controllable homopolymerization, random copolymerization, block copolymerization and grafting on functional PHA site chains [3] . Fig. 1 PHA diversity is generated by engineering the three basic synthesis pathways including the acetoacetyl-CoA pathway (pathway I), in situ fatty acid synthesis (pathway II), and/or beta-oxidation cycles (pathway III), as well as PHA synthase specificity and process control. It is now possible to tailor the PHA structures via genome editing or process engineering. The increasing PHA diversity and maturing PHA production technology should lead to more focused research into their low-cost and/or high-value applications [4] (see Fig. 2 ). Fig. 2 PHA diversity is generated by engineering the three basic synthesis pathways including the acetoacetyl-CoA pathway (pathway I), in situ fatty acid synthesis (pathway II), and/or beta-oxidation cycles (pathway III), as well as PHA synthase specificity [4] , [5] . Fig. 2 Similarly to the genome, transcriptome, and proteome, the PHA spectrum exhibits diverse and dynamic modifications – the term 'PHAome' has been created to reflect not only the diversity of monomers, homopolymers, random and block copolymers, functional and graft polymers, molecular weights, and combinations of the above, but also the ranges of PHAs with various molecular weights and monomer ratios that are present at a particular time point in a bacterial cell. It has become very important to understand the PHAome and ensuring an ample supply of PHAs to promote the discovery of new properties and applications of this family of advanced materials [5] ."
} | 1,146 |
37018404 | PMC11318656 | pmc | 1,070 | {
"abstract": "Corals are imminently threatened by climate change–amplified marine heatwaves. However, how to conserve coral reefs remains unclear, since those without local anthropogenic disturbances often seem equally or more susceptible to thermal stress as impacted ones. We disentangle this apparent paradox, revealing that the relationship between reef disturbance and heatwave impacts depends upon the scale of biological organization. We show that a tropical heatwave of globally unprecedented duration (~1 year) culminated in an 89% loss of hard coral cover. At the community level, losses depended on pre-heatwave community structure, with undisturbed sites, which were dominated by competitive corals, undergoing the greatest losses. In contrast, at the species level, survivorship of individual corals typically declined as local disturbance intensified. Our study reveals both that prolonged heatwaves projected under climate change will still have winners and losers and that local disturbance can impair survival of coral species even under such extreme conditions.",
"introduction": "INTRODUCTION Marine heatwaves threaten the persistence of tropical scleractinian corals ( 1 – 3 ) and, with them, the biologically diverse ecosystems that these foundational reef-building species support. Corals are particularly vulnerable to temperature anomalies, with increases of only 1°C capable of disrupting their obligate symbiosis with the photosynthetic dinoflagellate microalgae (family Symbiodiniaceae) that normally fuel them, causing the coral animal to expel its symbionts and bleach ( 4 , 5 ). Prolonged bleaching typically leads to coral starvation and mortality ( 5 ). Although climate change has long been recognized as a serious threat to tropical corals ( 6 – 8 ), the recent preponderance of marine heatwaves—persistent anomalously warm ocean temperatures ( 9 )—has shifted focus from the threats posed by gradually rising temperatures and ocean acidification ( 8 ) to these punctuated disturbances ( 1 , 3 ). Already, three global coral bleaching events triggered by El Niño–fueled marine heatwaves (1997–1998, 2010, and 2014–2017) have caused devastating coral losses ( 10 , 11 ). Climate change models project that both the intensity and frequency of marine heatwaves will increase in the coming decades ( 1 , 3 ), such that many of the world’s coral reefs are predicted to undergo annual bleaching events by midcentury ( 12 ). These events will, however, not occur in isolation. On almost all reefs, climate change is superimposed on a suite of chronic local anthropogenic disturbances ( 13 )—ranging from coastal development and associated pollution and reef sedimentation to overexploitation, destructive fishing practices, and disease—that have already substantially altered coral communities through reductions in coral cover and changes to community composition, with largely unknown consequences for species and ecosystem resilience to thermal stress. Given the intensification of marine heatwaves and the ubiquity of local anthropogenic disturbances on coral reefs, there is an urgent need to understand how these stressors interact ( 14 ). However, to date, few coral bleaching studies have explicitly examined multiple stressors ( 15 ). Chronic local anthropogenic disturbance might mediate coral reef responses to thermal stress, either increasing susceptibility—as documented for massive corals on the Mesoamerican Reef following the 1998 El Niño ( 16 , 17 )—or conversely enhancing resilience—if disturbances have already eliminated the most vulnerable coral species, leaving behind only the hardiest ones ( 18 )—as documented on Kenyan reefs ( 19 ). Alternatively, exposure to chronic local disturbance may have no effect, with thermal stress affecting corals irrespective of underlying protection, as found recently on the Great Barrier Reef ( 20 ). The degraded state of most modern reefs is widely acknowledged ( 8 , 13 , 21 , 22 ), and as managers seek to understand how to manage coral reefs under climate change, one might expect that examination of multiple stressors would be common practice in modern coral reef research. However, when we systematically reviewed studies reporting on the effects of recent marine heatwaves (2014–2021)—a period that included six of the seven hottest years on record at the time of study—we found that only 10% ( n = 20 of 194) had explicitly tested if local anthropogenic disturbance influenced heat stress effects on corals (fig. S1 and data file S1). The most common comparisons in these studies were between sites exposed to local disturbance and ones actively managed inside a marine protected area, although we note that coral reefs may also be de facto protected from local disturbance by the absence of impacts in remote locations. Almost half ( n = 9) of the 20 studies reported a positive effect of protection (i.e., decreased coral bleaching and/or mortality) on corals during heat stress events, 40% ( n = 8 of 20) reported no effect, and 15% ( n = 3 of 20) concluded that protection was detrimental to corals during heat stress. Conflicting evidence among the few bleaching studies that have tested for the effects of local anthropogenic disturbance and the overall lack of attention to this fundamental aspect of modern coral reefs impedes understanding of how best to manage these ecosystems in a warming world. How coral reefs are transformed by climate change this century will depend not only on their exposure to thermal stress and local anthropogenic disturbance but also on the sensitivity and response capacity of individual coral species to these stressors ( 23 – 25 ). Coral sensitivity to thermal stress is determined by biological traits, such as tissue thickness ( 26 ), and physiological tolerance, which is influenced by factors including the type and abundance of the coral colony’s obligate algal endosymbionts ( 27 – 30 ). Response capacity, in turn, may reflect species-specific propensity for acclimatization (e.g., the flexibility to switch or shuffle symbionts or to up-regulate host thermal stress responses) and adaptation (e.g., selection for traits of either the host or symbionts that confer a fitness advantage under stressful conditions) ( 28 , 31 , 32 ). Interspecific differences in sensitivity to thermal stress have long been recognized ( 23 , 26 , 33 , 34 ), with “winners” generally able to either avoid bleaching during thermal stress or recover from it after warming subsides, and “losers” tending to bleach and die quickly in response to warming ( 34 ). Because environmental filtering is stronger under stressful conditions ( 35 – 37 ), reefs increasingly stressed by marine heatwaves may lose diversity and converge toward simpler assemblages, as losers are eliminated from the species pool. However, because repeated heatwaves may turn some winners into losers and vice versa ( 38 ), questions remain about how corals with different sensitivities will respond to heatwaves of increasing frequency, duration, and intensity. Which corals will endure in communities will also depend on whether species exhibit positive or negative cotolerance to thermal stress and local anthropogenic disturbance ( 35 , 39 ). Predicting future reef states thus requires not only accounting for underlying anthropogenic disturbances but also understanding interspecific variability in survivorship through heatwaves. To date, however, most heatwave studies have focused on quantifying coral bleaching, a symptom of thermal stress, rather than coral mortality, a fundamental parameter required to assess the demographic effects of such events. This disconnect reflects the challenge of quantifying coral mortality, which, under the strictest standards, requires following individual colonies over time and, at minimum, requires documenting coral cover before and after a heatwave, as opposed to bleaching assessments that require only a single site visit. Although bleaching may be an accurate proxy for mortality in short heatwaves, during prolonged events that are becoming the norm, for corals that either bleach quickly and die (and hence are unlikely to be recorded in the bleached state) or those that can persist in a bleached state for prolonged periods, it is not ( 40 ). Here, we took advantage of the ecosystem-scale natural experiment that occurred at the epicenter of the 2015–2016 El Niño, the central equatorial Pacific Ocean, where prolonged heat stress blanketed a spatial gradient of chronic local anthropogenic disturbance on the world’s largest atoll, Kiritimati. We quantified thermal stress around the atoll using high-precision in situ temperature loggers and satellite data, and human disturbance using a combined metric of local human population and fishing pressure. Our primary objective was to evaluate how exposure to local disturbance modulates the impacts of heat stress on corals at both the community (i.e., among coral species) and species (i.e., for individual coral species) levels. In addition, we sought to assess whether coral bleaching, the most commonly recorded reef metric during heatwaves, accurately predicts coral mortality. Thus, over the course of nine expeditions before (2013–2015), during (2015–2016), and after (2016–2017) the heatwave, at sites exposed to consistent heat stress but varying levels of local disturbance ( Fig. 1, A and D , and figs. S2 and S3), we quantified coral community composition and bleaching ( n > 250,000 points from 94 photo surveys) and, in one of the largest longitudinal studies of individual corals to date ( 41 ), tracked the fate of >850 individual coral colonies. Fig. 1. Reef communities across a gradient of chronic human disturbance before the 2015-2016 El Niño. ( A ) Reef sites on Kiritimati (central equatorial Pacific Ocean) at which coral community structure and individually tagged coral colonies (sites encircled in black) were tracked over the course of the 2015–2016 El Niño. ( B ) Parameter estimates and 95% confidence intervals for factors examined [Dist., local human disturbance; NPP, net primary productivity ( l , linear; q , quadratic); Temp., maximum monthly mean (MMM) temperature; Expo., wave exposure] for their influence on hard coral cover before thermal stress (table S4). ( C ) Mean community composition (CCA, crustose coralline algae; turf, turf algae; abiotic, sediment, sand, and rubble) of the forereef benthos among sites, classified by their exposure to chronic local human disturbance. VL, very low; L, low; M, medium; H, high; VH, very high. ( D ) Photos of the coral reef communities before the El Niño at sites representing each of the atoll’s levels of local human disturbance. Photo credits: D. Claar (very low and very high), University of Victoria (UVic); M. Watson (medium), UVic; and The Baum Lab (low and high), UVic.",
"discussion": "DISCUSSION Despite the urgent need to leverage all available solutions for enhancing coral reef resilience to thermal stress in the face of escalating climate change, it has been unclear whether managing local anthropogenic disturbances on reefs helps or hinders in this regard. Our study sheds new light on this debate. Tracking whole coral communities and individual species exposed to consistent thermal stress, but different levels of underlying local anthropogenic disturbance, clarified that the relationship between local anthropogenic disturbance and climate resilience varies qualitatively across biological scales. At the community level, we found that reefs exposed to high levels of chronic local anthropogenic disturbance fared better through a prolonged heatwave than those shielded from local disturbance, an outcome driven by differences in the coral community composition among sites. In contrast, when comparing survival rates within individual coral species, the predominant relationship for those species that were not eradicated by the heatwave was one of declining survivorship as local disturbance increased. These findings have implications for managing and restoring coral reefs, as climate change–driven marine heatwaves continue to intensify in frequency and duration. Winners and losers in prolonged heatwaves Overall, we documented mass coral mortality (89% hard coral cover loss) arising from a heatwave that persisted for a remarkable 10 months unabated. At its peak, heat stress accumulated to 25°C-weeks (DHWs), a level that had not previously been anticipated to occur on any reef until midcentury ( 26 ). Its occurrence 35 years earlier than predicted underscores how rapidly climate change is advancing ( 2 ). Although corals typically exhibit high interspecific variability in their sensitivity to thermal stress ( 33 ), a heatwave this extreme might have been expected to overwhelm the tolerance of even the most resilient species, resulting in high mortality rates across the board. Instead, we found that interspecific coral cover losses still varied widely, from the complete loss of tabular Acropora and foliose Montipora to a loss of only 65% in the stress-tolerant mounding coral Platygyra spp. Mortality rates for the individual colonies that we tracked were also highly variable across species, but with lower mortality rates for the stress-tolerant species, because many of the colonies of these species had surviving corallites, thus potentially paving the way for their recovery. Only on nearby tiny Jarvis Island was thermal stress more extreme (maximum of 31.58°C-weeks) during this heatwave ( 42 , 43 ). There, reefs underwent an estimated >98% decline in hard coral cover, with severe losses of Montipora spp. (100%), Pocillopora spp. (>90%), and Pavona spp. (~85%) recorded in the surveyed areas ( 43 , 44 ). Together, these results illustrate that extremely prolonged heatwaves still have winners and losers, such that these events will cause not only enormous coral losses but also substantial changes in community composition. Influence of local anthropogenic disturbance on coral survival In species-specific models testing the influence of exposure to chronic local anthropogenic disturbance on coral resilience to thermal stress, we detected an inverse relationship for those species with stress-tolerant life histories. Survivorship of all stress-tolerant species was at least twice as high at sites absent local disturbance compared to sites with the highest exposure and was more than 10 times as high for the species with the strongest relationship, P. ryukyuensis . Increased coral sensitivity to thermal stress with greater human disturbance is, we expect, most likely attributable to the diminished water quality at the most highly disturbed sites. On Kiritimati, raw sewage and pollution inputs have resulted in increased turbidity and sedimentation (fig. S3) as well as greater concentrations of bacteria, virus-like particles, and potential pathogens in the water column at these sites ( 45 , 46 ). Previous studies have shown that low water quality can change the coral microbiome ( 47 – 49 ), and microbiome analyses of subsets of our tracked corals before the heatwave showed increased bacterial diversity at highly disturbed sites ( 50 ). Such changes can have knock-on effects for coral physiology and survivorship even in the absence of warming ( 51 ), which may then be exacerbated during heatwave events. Poor water quality, as with other environmental stressors, can also lead to changes in coral-algal symbioses ( 52 – 54 ), which may then influence coral resilience to subsequent thermal stress. Analyses of our tracked P. ryukyuensis colonies, and two of the three cryptic coral lineages in tracked P. lobata colonies, revealed symbioses with distinct Symbiodiniaceae across the disturbance gradient that were linked to coral survival during heat stress ( 55 , 56 ). Although the mechanism underlying the relationship between coral survival and local disturbance remains unclear for the other stress-tolerant corals that we tracked—and may differ across species—we suggest that it likely results from distinct coral microbiomes and their associated physiological traits found across the disturbance gradient. We attribute the failure to detect an effect of local disturbance on coral resilience to thermal stress in our two competitive coral species to their extremely high mortality. A recent study of a less severe heatwave on Moorea, French Polynesia showed that bleaching severity was significantly increased by local nitrogen pollution in the competitive coral genera Acropora and Pocillopora ( 57 ). Overall, these results suggest that impairment of coral resilience to thermal stress by local stressors may be a general phenomenon and at least deserves increased research and management attention. Notably, however, these species-level results stand in contrast to our own community-level results and previous studies, which have suggested that local disturbance enhances coral reef resilience to thermal stress. Over a decade ago, Côté and Darling ( 58 ) argued that this would be the case for coral reefs exposed to—and altered by—local stressors, because the most sensitive coral species would already have been eliminated, leaving behind a more stress-tolerant community. That is, if organisms exhibit cotolerance to stressors (or, conversely, cosensitivity) such that they respond similarly to them, then the combined effects of the stressors may be antagonistic, resulting in a response that is less than the sum of their individual effects ( 14 , 39 , 59 ). Although antagonistic effects due to stressor cotolerance are not a given—as multiple stressors may also exhibit synergistic or additive effects—such interactions appear to be fairly common on reefs exposed to local stressors and global climate change. Darling et al . ( 19 ) found that while the stress-tolerant corals that dominated fished reefs in Kenya before the 1998 El Niño were barely affected by the bleaching event, reefs in no-take reserves had more diverse coral assemblages, including many corals with competitive life history traits that exhibited cosensitivity to fishing and bleaching, and incurred heavy losses. More recently, Cannon et al . ( 60 ) showed that central Pacific reefs in the Gilbert Islands that were exposed to higher levels of chronic local pressure were dominated by a coral species tolerant of nutrient loading and turbidity and were subsequently less affected during a bleaching event than nearby reefs with fewer local pressures. At Kiritimati’s highest disturbance sites, we found that of the competitive coral species, Acropora were completely absent, and while some encrusting Montipora persisted, only a few colonies of the foliose form (common in less disturbed sites) were recorded. Thus, the “positive” effect of local disturbance reflects different community compositions and the variable thermal sensitivities of the coral species that dominate disturbed reef communities, rather than there being a mechanism by which local disturbance itself enhances coral resilience to thermal stress. More difficult to reconcile with either our species- or community-level findings are studies reporting that coral responses to thermal stress occur irrespective of local protection or remoteness, influenced only by the reef’s exposure to thermal stress. In surveys of the Great Barrier Reef Marine Park during the 2016 marine heatwave, for example, Hughes et al . ( 20 ) documented severe bleaching on reefs in each of the park’s types of management zones and concluded that local management of water quality (assessed using long-term chlorophyll a concentration) and fishing pressure had little to no influence on coral resistance to extreme heat. Similarly, a study in one of Indonesia’s oldest marine parks during the same heatwave found that management zone made no difference to coral losses ( 61 ). More recently, Baumann et al .’s ( 62 ) global meta-analysis tested the relationship between human influence and coral resilience and concluded that reefs isolated from human pressures are not more resilient to climate change, noting that even the world’s most remote reefs bear the impacts of intense marine heatwaves. We concur that, at broad spatial scales, exposure to thermal stress will be highly variable across reefs, and this may well be the primary determinant of reef impacts; remote reefs are not immune to high thermal stress exposure levels. Such an emphasis on current and future thermal stress exposures has proven useful when considering future thermal refugia for coral reefs, as in the “50 Reefs” conservation prioritization ( 63 , 64 ). At finer spatial scales, however, where thermal stress exposure is the same (or very similar) across reefs, a corollary of the conclusion that coral responses are the same irrespective of protection is that coral sensitivities and response capacities to thermal stress must be the same across the protection levels. Because this outcome seems unlikely, we posit that, in these studies: (i) corals were exposed to similar conditions inside and outside the protected area, such that the communities did not differ, or (ii) exposures to thermal stress across protection levels were not actually equal; or (iii) different impacts were not detected because of insufficient power, or bleaching was measured at only a single time point such that the full ecological impacts of the event were not quantified. Considering our results together across scales suggests that, although local anthropogenic disturbance can result in the loss of sensitive coral species such that the remaining community is more tolerant to subsequent thermal stress, when comparing “apples with apples”—that is, the same species across different levels of local anthropogenic disturbance—there is clear evidence that local disturbance can impair survival. Thus, while there is compelling recent evidence that coral reef recovery following bleaching events may not be aided by minimizing human disturbances to reefs ( 18 ), our study suggests that previous conflicting results pertaining to coral community resilience to thermal stress may be resolved through consideration of biological scale. Coral bleaching does not foretell demographic impacts of prolonged heatwaves Our repeated reef surveys during an extended bleaching event also provide an empirical test of the relationship between coral bleaching and mortality. Given the many challenges associated with conducting in situ assessments of coral bleaching events—including the need to marshal resources quickly when heatwaves arise, the limited reef area that can be assessed by divers, and the complexities of accessing remote reefs—rapid reef surveys at a single time point during a heatwave are often used to assess ecological impacts. However, whereas bleaching incidence can lead to decreased coral growth and reproduction, the capacity of corals to recover from bleaching means that it may not accurately foretell coral mortality, and hence the overall demographic impacts of heatwaves. We found no relationship between bleaching prevalence and subsequent mortality levels in any of our tagged coral species. Instead, we found that the species with the highest bleaching incidence early in the event ( P. ryukyuensis and Favites pentagona ) had among the lowest mortality, while a species with very low bleaching incidence ( P. grandis ) suffered near complete mortality ( Fig. 6A and figs. S12 and S13). Mismatches between bleaching and mortality could arise if certain coral species can resist the onset of bleaching more than others but then only persist in a bleached state for a short period ( 40 , 65 ). Such mismatches will be more likely in the prolonged heatwaves that are projected to become more common under climate change ( 40 , 66 ), thus highlighting the need for increased sampling during these events to accurately gauge demographic impacts. As the capacity to use satellite-derived data to accurately monitor coral bleaching increases, these sources could help to overcome this challenge. Coral reef recovery from prolonged heatwaves unlikely under climate change We posit that coral reef recovery from prolonged heatwaves is increasingly unlikely because of long ecosystem recovery times and the diminishing interval between successive heatwaves under climate change ( 67 , 68 ). On Kiritimati, our sampling up until 3 years after the end of the heatwave (2019, before the onset of coronavirus disease 2019) revealed juvenile corals and regrowth of colonies that had experienced partial mortality, which together resulted in some increase in overall coral cover but still left the ecosystem a long way from full recovery. Long-term studies of coral reefs from the Indian and Pacific Oceans following the major 1998 El Niño found that recovery of hard coral cover typically took more than a decade and involved substantial turnover of community composition, with “recovered” reefs tending to have lower coral diversity and be dominated by fast-growing corals ( 69 – 72 ). Recovered reefs in Moorea, for example, are now dominated by “fields” of Pocillopora ( 73 ), while recovering reefs in the Seychelles became dominated by fast-growing, branching Acropora corals ( 74 ). Reef recovery following mass bleaching events is also not guaranteed. Following the 1998 El Niño, more than 40% of surveyed reefs in the Seychelles underwent regime shifts to fleshy macroalgae ( 72 ). Those that were on a recovery trajectory, which had high coral cover before the 1998 El Niño, still had not fully recovered by 2014, and although full recovery was projected to be complete within 17 to 29 years ( 74 ), progress was nullified by Seychelles’ 2016 bleaching event ( 75 ). Such outcomes are increasingly likely with climate change ( 67 ). Thus, as with many reefs, the probability of full recovery of Kiritimati’s reefs now seems slim. Persistence of coral reefs throughout the 21st century will be dictated almost entirely by the extent to which greenhouse gas emissions are reduced ( 66 ). Our study shows that prolonged heatwaves under climate change will not only substantially reduce coral cover but also transform the remaining coral community composition. All of Kiritimati’s reefs suffered staggering losses during this study, including those exposed to very low disturbance, which had arguably been among the most pristine remaining on the planet before the heatwave. Diminishing intervals between recurrent heatwaves will leave most of the world’s reefs with insufficient time to recover after such events ( 67 ). Emissions reductions that only limit warming to 2°C are projected to result in the loss of virtually all coral reefs (99%), whereas if warming is limited to 1.5°C, then losses could be limited to between 70 and 90% ( 66 , 76 ). Under such dire conditions, strategies additional to greenhouse gas emissions reductions that can reliably enhance coral resistance to, or recovery from, marine heatwaves should be broadly deployed. However, the efficacy and scalability of the potential options remains uncertain. Our study provides evidence that, at least for some coral species, resilience to thermal stress is enhanced as local anthropogenic disturbances are reduced. These findings imply that alleviating local disturbances—such as by improving water quality, which is likely one of the most tractable options for reef managers—could not only benefit natural coral reefs but also aid coral restoration efforts, improving the odds of success for the individual coral species that are out-planted on reefs. With much still to learn about the interactions between multiple stressors on coral reefs, we encourage researchers to explicitly incorporate local disturbances into future studies of marine heatwave impacts on reefs. In addition to urgent reductions in greenhouse gas emissions, evidence-based local management actions that are both scalable and durable are urgently needed as a means of increasing the odds of persistence for these imperiled ecosystems under climate change."
} | 7,022 |
40166204 | PMC11957125 | pmc | 1,071 | {
"abstract": "Efforts towards microbial conversion of lignin to value-added products face many challenges because lignin’s methoxylated aromatic monomers release toxic C 1 byproducts such as formaldehyde. The ability to grow on methoxylated aromatic acids (e.g., vanillic acid) has recently been identified in certain clades of methylotrophs, bacteria characterized by their unique ability to tolerate and metabolize high concentrations of formaldehyde. Here, we use a phyllosphere methylotroph isolate, Methylobacterium extorquens SLI 505, as a model to identify the fate of formaldehyde during methylotrophic growth on vanillic acids. M. extorquens SLI 505 displays concentration-dependent growth phenotypes on vanillic acid without concomitant formaldehyde accumulation. We conclude that M. extorquens SLI 505 overcomes potential metabolic bottlenecks from simultaneous assimilation of multicarbon and C 1 intermediates by allocating formaldehyde towards dissimilation and assimilating the ring carbons of vanillic acid heterotrophically. We correlate this strategy with maximization of bioenergetic yields and demonstrate that formaldehyde dissimilation for energy generation rather than formaldehyde detoxification is advantageous for growth on aromatic acids. M. extorquens SLI 505 also exhibits catabolite repression during growth on methanol and low concentrations of vanillic acid, but no diauxie during growth on methanol and high concentrations of vanillic acid. Results from this study outline metabolic strategies employed by M. extorquens SLI 505 for growth on a complex single substrate that generates both C 1 and multicarbon intermediates and emphasizes the robustness of M. extorquens for biotechnological applications for lignin valorization.",
"introduction": "Introduction Lignin, a major component of woody plant cell walls, is one of Earth’s most abundant renewable carbon sources. It comprises a complex network of polycyclic aromatic polymers that provide rigidity for growing plants and acts as a barrier against harsh weather and grazing herbivores 1 , 2 . Microbial degradation of lignin is interesting both from ecological and biotechnological perspectives, as increased understanding of these processes relate to efficient carbon cycling in natural ecosystems as well as the exploitation of lignin as a feedstock for petrochemical production 2 , 3 . Microbial degradation of lignin-derived methoxylated aromatic acids (e.g., vanillic acid, ferulic acid, protocatechuic acid) is distributed across soil and plant microorganisms 2 , 4 , 5 . While variations of aromatic acid degradation modules exist 6 , vanillic acid is commonly used as a model for investigating aromatic acid degradation 7 – 9 . Aerobic growth on vanillic acid proceeds through the following enzymatic reactions: ferulic acid is oxidized to vanillic acid ( ech ), which is demethylated to protocatechuic acid ( vanAB ) and produces formaldehyde as an obligate byproduct. Protocatechuic acid serves as a substrate for the β-ketoadipate pathway. It undergoes a series of ring cleavage steps ( pcaHG, pcaB, pcaC, pcaD ) to generate β-ketoadipate, which is converted to succinyl-CoA and acetyl-CoA ( pcaIJ, pcaF ), common building block metabolites for the TCA cycle and other assimilatory cycles ( Figure 1A , pathway indicated in blue) 10 , 11 . The obligate production of formaldehyde in the conversion of vanillic acid to protocatechuic acid has been hypothesized to be a major bottleneck in the efficient degradation of methoxylated aromatic acids 12 – 15 . Formaldehyde is a highly toxic electrophile that can crosslink with proteins and DNA, inhibiting cell growth 16 . Previous studies have shown that Pseudomonas putida KT2440 17 , 18 and Rhodococcus jostii RHA1 19 , both model organisms for aromatic acid metabolism, release appreciable amounts of formaldehyde during growth on vanillic acid. Formaldehyde excretion correlates with cell death in these organisms despite both encoding native formaldehyde detoxifying mechanisms 18 . To overcome the rate-limiting step of C 1 byproduct toxicity, there have been efforts to engineer microbes such as Escherichia coli , Corynebacterium glutamicum , and Burkholderia cepacia for enhanced aromatic acid degradation through substrate channeling, compartmentalization, and introduction of novel formaldehyde oxidation pathways 5 , 12 . The majority of microorganisms are highly susceptible to formaldehyde toxicity, but methylotrophic bacteria have evolved robust formaldehyde detoxification mechanisms. Methylotrophs are defined by their ability to utilize reduced carbon compounds lacking carbon-carbon bonds (e.g., methane, methanol, methylamine) as their sole source of carbon and energy 20 , 21 . Recently, the ability to degrade methoxylated aromatic acids has been identified in specific clades of methylotrophic bacteria. Still, the extent to which methoxydotrophy–the metabolism of methoxy groups of aromatic compounds– occurs is not widely understood 19 . During methylotrophic growth in the model, pink-pigmented facultative methylotroph, Methylobacterium extorquens , coupled redox reactions in the periplasm oxidize primary C 1 substrates (methanol, methylamine, etc) to formaldehyde and transfer the resulting electrons to cytochromes directly linked to the electron transport chain for energy conservation 20 , 22 . Formaldehyde is transported from the periplasm to the cytoplasm, where it is immediately covalently attached to a tetrahydromethanopterin (H 4 MPT) carbon carrier and oxidized to formate through a series of steps that generate NAD(P)H as reducing power 20 , 22 . Formate serves as a branchpoint between assimilation and dissimilation. Formate can be oxidized to CO 2 with the concomitant generation of NADH via formate dehydrogenases or assimilated via a partial reduction facilitated by a tetrahydrofolate carbon carrier to enter the serine cycle for assimilation ( Figure 1A , pathway in black) 23 . Methylotrophic growth is considered limited by NADPH, as the primary source of NADPH necessary to assimilate formate into the serine cycle is the oxidation of formaldehyde to formate in the preceding steps 20 , 23 , 24 . Degradation of methoxylated aromatic acids by methylotrophs yields both C 1 (formaldehyde) and multicarbon (acetyl-CoA, succinyl-CoA) intermediates – an example of a single complex substrate that produces two distinct intermediates that can be assimilated into metabolic pathways via different entry points ( Figure 1A ). Many methylotrophs, including members of the extorquens clade, are facultative organisms capable of utilizing multicarbon compounds as a carbon and energy source by feeding intermediates directly into the TCA cycle for assimilation 20 , 21 . Because the serine cycle, used for the assimilation of C 1 intermediates, and the TCA cycle, used for the assimilation of multicarbon intermediates, share enzymatic reactions but run in opposite directions, simultaneous assimilation of C 1 and multicarbon compounds cannot occur 21 , 24 . Previous studies have demonstrated that when methylotrophs are grown on succinate (multicarbon) and methanol (C 1 ) simultaneously, they allocate the C 1 substrate to dissimilation and the multicarbon substrate to assimilation 25 – a strategy that overcomes diauxic growth while maximizing bioenergetic yields during co-substrate growth. In this study, Methylobacterium extorquens SLI 505, a recent isolate from the soybean phyllosphere 26 , is used as a model to understand how methylotrophs overcome bottlenecks surrounding methylotrophic and heterotrophic pathway operation in their metabolism of aromatic acids. M. extorquens SLI 505 upregulates its methylotrophic and heterotrophic machinery in response to increasing concentrations of vanillic acid, yet formaldehyde is not assimilated during growth. Further, formaldehyde detoxification only enables optimization of carbon utilization at the level of growth rate and not yield, despite that formaldehyde detoxification is efficient even during growth on high concentrations of vanillic acid. Finally, catabolite repression is observed when methanol and low concentrations of vanillic acid are available. Our data suggests that currency metabolites such as NAD(P)H may play a role in defining the distribution of carbon between assimilation and dissimilation.",
"discussion": "Discussion How complex substrates are converted to various intermediates that can be assimilated through two drastically different modes of metabolism is a fundamental gap in our understanding of microbial physiology 34 , 35 . Aromatic acid metabolism is a model by which to understand how methylotrophic bacteria maximize bioenergetic yields during growth on multiple intermediates and can bolster our understanding of how facultative methylotrophs balance methylotrophy and heterotrophy. Here, we establish M. extorquens SLI 505 as a model organism for vanillic acid metabolism. Genes involved in methoxylated aromatic acid catabolism are reported and shown to be organized similarly to what has been found in other extorquens clade members capable of aromatic acid metabolism. Growth phenotypes on a wide range of vanillic acid concentrations are starkly concentration-dependent, with reduced growth at higher substrate concentrations that we hypothesize is related to NADPH limitation rather than formaldehyde accumulation and mixed utilization patterns when grown on vanillic acid with methanol. To date, much of the literature surrounding aerobic aromatic acid metabolism in bacteria has focused on mechanisms by which aromatic rings are converted into building blocks for assimilation and how bacteria can salvage carbon from complex substrates 8 , 9 , 9 , 36 . The toxicity of aromatic acids, whether due to their inherent chemical properties or C 1 intermediates that arise in their degradation, has been considered a biochemical inevitability during the metabolism of these compounds and/or a target for metabolic engineering for biotechnological applications 12 , 18 . However, the recent focus on the ability of particular clades of methylotrophic bacteria to robustly grow on aromatic acids has expanded our understanding of how lignin-derived aromatic compounds might influence microbial physiology in natural environments and how this might be co-opted for biotechnology 19 , 26 , 37 . Methylotrophic bacteria are especially attractive organisms for investigating aromatic acid metabolism because they have evolved elegant mechanisms by which not only to detoxify but also assimilate high concentrations of formaldehyde as a routine consequence of their metabolism 21 , 23 . Non-methylotrophic bacteria that natively encode formaldehyde detoxification mechanisms, such as Pseudomonas putida KT2440, display a marked decrease in growth due to formaldehyde accumulation even at low concentrations of vanillic acid and efforts to engineer P. putida KT2440 to overcome this limitation are ongoing 18 , 38 . In contrast, M. extorquens SLI 505 is an environmental isolate naturally capable of robust aromatic acid metabolism without concomitant formaldehyde accumulation, even during growth on concentrations as high as 12 mM ( Table 1 ). 13 C fingerprinting with methoxy-labeled 13 C-vanillic acid ( Figure 3A ) was used to track labeled carbon in amino acids as a proxy for determining whether formaldehyde generated from vanillic acid was assimilated towards biomass. Of the detectable proteinogenic amino acids, labeled carbon was absent from all except for methionine, which was roughly 40% labeled ( Figure 3B ). Methionine biosynthesis requires abstraction of methyl groups and the partial label incorporation is likely a reflection of methionine biosynthesis alone and not indicative of assimilation of formaldehyde, as labeled carbon is absent from all other amino acids. Other studies using labeled vanillic acid for investigating aromatic acid metabolism in the non-methylotrophic bacterium, Sphingobium sp . SYK-6 also noted labeling of methionine and links between methionine and C 1 metabolism 39 . Despite the minor labeling of methionine, we still conclude that formaldehyde generated during growth on vanillic acid is not assimilated towards biomass. When methylotrophic bacteria grow on methanol, 100% of the initial substrate must be converted to formaldehyde before it is shunted towards assimilatory or dissimilatory pathways 21 , 24 . In contrast, the growth of methylotrophic bacteria on methoxylated aromatic acids mainly produces heterotrophic intermediates and only results in an eighth of the total carbon in the substrate being converted to formaldehyde ( Figure 1A , 3E ). Thus, it is not surprising that formaldehyde toxicity is not a major constraint during this metabolism based solely on differences in the relative amounts of formaldehyde generated by vanillic acid. Although we hypothesize that the more substantial constraints on metabolism are at the level of currency metabolites, we cannot rule out the possibility that other metabolic or non-metabolic processes contribute to the concentration-dependent phenotypes we report here. We see no evidence for stress response in our transcriptomic data set for high vanillic acid concentrations in comparison to other substrate conditions, and the growth media is sufficiently buffered to prevent vanillic acid itself from causing acute acid stress. A recent study suggested that M. extorquens experiences membrane depolarization as a result of excess vanillic acid diffusion across membranes that can lead to disruption of proton motive force for energy production 37 . Our preliminary analysis of membrane permeability ( Supplementary Figure 2 ) of M. extorquens SLI 505 in a variety of substrates and vanillic acid concentrations did not indicate a correlation with growth on high concentrations of vanillic acid. However, a link between membrane depolarization and energy metabolism could relate to the lack of growth of M. extorquens SLI 505 at vanillic acid concentrations higher than 15 mM and warrants further investigation in this system. The fate of formaldehyde in methylotrophic bacteria is unique in that formaldehyde detoxification is necessarily coupled to formaldehyde dissimilation. In literature discussing methylotrophy, the terms detoxification and dissimilation are often used interchangeably as they relate to this metabolism. Here, we take advantage of mutants deficient in formaldehyde assimilation (Δ ftfL ) and formaldehyde detoxification and dissimilation (Δ mptG ) to disentangle these processes. Because aromatic acid metabolism does not require formaldehyde assimilation for growth, it is a useful model to interrogate these processes. M. extorquens SLI 505 Δ ftfL is incapable of assimilating formaldehyde yet does not accumulate formaldehyde ( Figure 4C ) and grows faster than wild-type at most vanillic acid concentrations ( Figure 3C ). The latter was a surprising growth phenotype, as loss of ftfL has no obvious advantage if assimilation of C 1 units does not naturally occur during this metabolism. Growth advantages due to loss of ftfL have been reported in M. extorquens AM1 evolved in succinate 40 , but physiological benefits during growth on vanillic acid remain unknown. In contrast, a Δ mptG mutant incapable of synthesizing the carbon carrier necessary for formaldehyde oxidation to formate has substantially slower growth rates than wild-type at all vanillic acid concentrations ( Figure 4A ) yet manages to reach the same final OD 600 values as wild-type ( Figure 4B ). Interestingly, addition of lanthanum chloride significantly improves growth rates of M. extorquens SLI 505 Δ mptG during growth on vanillic acid (data not shown). We hypothesize that this is due to lanthanum chloride inducing expression of lanthanide-dependent alcohol dehydrogenases that might play a role in formaldehyde oxidation, 41 and investigations into the role of lanthanides during growth on vanillic acid is the subject of ongoing work. Growth phenotypes of M. extorquens SLI 505 Δ mptG sharply contrast what has recently been reported for a strain of M. extorquens PA1 engineered to metabolize vanillic acid, where a Δ mptG mutation is fatal 37 . The differences in the essentiality of mptG in our natural isolate M. extorquens SLI 505 versus the engineered M. extorquens PA1 strain could highlight differences in evolution or metabolic regulation of formaldehyde-related processes in these two systems that are worth investigating further. Here, we hypothesized that the growth rate defect of our Δ mptG strain was not due to formaldehyde accumulation, as intracellular formaldehyde levels of this mutant were below toxic levels for this strain ( Figure 4C ) 30 , but rather due to the importance of formaldehyde dissimilation and nucleotide pools in the form of NADP + and NADPH. This is not unprecedented, as tight regulation between NAD(P)H pools, cofactor selectivity, and vanillic acid metabolism have been reported in Pseudomonas putida KT2440 29 and Sphingobium sp . SYK-6 39 . Methylotrophic metabolism has been shown to be NADPH-limited 24 , and this appears to hold true even for substrates such as vanillic acid that do not require formaldehyde assimilation. Future work will investigate nucleotide pools in Δ mptG strains to further substantiate our hypothesis about NADPH limitation generating a partial bottleneck for carbon assimilation during growth on high concentrations of vanillic acid. Growth on methanol and low concentrations of vanillic acid alleviates some of the growth phenotypes exhibited during growth on low concentration of vanillic acid alone ( Figure 6A , B ). Several important conclusions can be drawn about how M. extorquens SLI 505 copes with growth on low concentrations of vanillic acid with methanol: (i) consumption follows a hierarchy, where growth on methanol occurs before growth on vanillic acid with a marked diauxic shift, (ii) growth on vanillic acid in the second phase of the growth curve occurs without the lag characteristic of growth on vanillic acid alone, (iii) methanol itself induces expression of methylotrophic pathways 42 , yet despite both methylotrophic and heterotrophic pathways being operational, vanillic acid is still metabolized as shown in Figure 4E . Here, the presence of methanol with low concentrations of vanillic acid would theoretically “prime” the cells for methylotrophy; this would shift the expression profile in this condition to reflect growth on high concentrations of vanillic acid, where both methylotrophic and heterotrophic pathways are expressed ( Figure 5 ). Yet, growth phenotypes do not necessarily correlate with predictions made from transcriptomic analyses alone, as diauxic growth as a result of operation of both methylotrophic and heterotrophic pathways does not occur during growth on methanol and high concentrations of vanillic acid ( Figure 6H ). The vanillic acid used in this experiment is not fully labeled. Thus we cannot rule out the possibility that the unlabeled ring carbons of vanillic acid is being metabolized and/or assimilated in combination with methanol during the first stage of growth. However, this scenario would require simultaneously operating methylotrophic and heterotrophic pathways which is unlikely to occur. Traditionally, sequential utilization of substrates occurs if a preferred substrate supports a higher growth rate, assuming both substrates are in excess 34 . Limitation of substrates eliminates sequential or preferential substrate utilization in favor of co-utilization to maximize growth 35 , 43 . In contrast to this paradigm, growth on methanol and high concentrations of vanillic acid exacerbates the growth phenotypes exhibited during growth on high concentrations of vanillic acid alone ( Figure 6H ). Further labeling and 13 C fingerprinting studies will be required to validate if a single or both substrates are being consumed but catabolic repression via diauxic growth is no longer present. In recent decades, methylotrophs have emerged as promising model organisms for biotechnological manipulation due to their genetic tractability, multi-omics characterizations, and demonstrated flux through pathways directly linked to the production of value-added chemicals 20 , 44 . However, much of this work has been done with renewable C 1 feedstocks such as methane or methanol 20 , 45 – 47 . There are vast biotechnological implications for robust aromatic acid metabolism in methylotrophs that does not result in formaldehyde accumulation 3 , 7 , 48 – 50 . We report growth of M. extorquens SLI 505 on vanillic acid concentrations as high as 15 mM, substantially higher than what has been reported by many other model organisms with well-characterized aromatic acid metabolisms 9 , 39 . It has been shown that the inclusion of M. extorquens PA1 (non-aromatic acid utilizer) in a lignin-degrading consortium is sufficient to detoxify formaldehyde 17 ; similar studies have not been replicated for M. extorquens strains naturally capable of aromatic acid metabolism. Additionally, we have demonstrated that growth on high concentrations of vanillic acid is natively coupled to the accumulation of the bioplastic polyhydroxybutyrate ( Supplementary Figure 1C ), providing an attractive starting point for future metabolic engineering efforts."
} | 5,362 |
30837333 | PMC6401477 | pmc | 1,072 | {
"abstract": "Specific recognition of cognate signals is considered fundamental to cell signaling circuits as it creates fidelity in the communication system. In bacterial quorum sensing (QS), receptor specificity ensures that bacteria cooperate only with kin. There are examples, however, of QS receptors that respond promiscuously to multiple signals. “Eavesdropping” by these promiscuous receptors can be beneficial in both interspecies competition and cooperation. Despite their potential significance, we know little about the prevalence of promiscuous QS receptors. Further, many studies rely on methods requiring receptor overexpression, which is known to increase apparent promiscuity. By systematically studying QS receptors in their natural parent strains, we find that the receptors display a wide range of selectivity and that there is potential for significant cross talk between QS systems. Our results provide a basis for hypotheses about the evolution and function of promiscuous signal receptors and for predictions about interspecies interactions in complex microbial communities.",
"introduction": "INTRODUCTION Many bacteria use quorum sensing (QS) to communicate with kin and coordinate group behaviors ranging from antibiotic production to virulence factor secretion and biofilm formation ( 1 ). In many proteobacteria, QS is mediated by acyl-homoserine lactone (AHL) signals. AHL QS systems consist of a signal synthase and a dimeric cytosolic receptor that serves as a transcriptional activator or repressor ( Fig. 1A ). AHLs can diffuse through cellular membranes ( 2 , 3 ) and are comprised of a homoserine lactone core with an acyl tail ( Fig. 1B ). Most known AHLs possess fatty acyl tails that vary in length from 4 to 20 carbons and in modifications, particularly at the third carbon, which can be unsubstituted or have a hydroxy or oxo modification. To date, roughly 20 different naturally produced fatty AHLs have been identified among hundreds of quorum sensing organisms ( 4 , 5 ). Thus, there is some degeneracy whereby QS systems from different organisms produce and respond to the same signal. Despite the very similar structures of natural AHL signals, receptors are believed to be highly specific for and sensitive to their cognate signal ( 6 ). There are, however, reported exceptions to this paradigm. For example, Chromobacterium violaceum is frequently used as a tool for AHL detection due to its receptor’s promiscuous response to multiple AHLs ( 7 ). Furthermore, “eavesdropping” through promiscuous receptors has been shown to affect both interspecies competition ( 8 ) and cooperation ( 9 ) in laboratory experiments. In in vivo settings, interspecies cross talk via degenerate signals and/or promiscuous receptors have both been shown to modulate bacterial virulence to the benefit or detriment of a plant host ( 9 – 11 ), and similar interactions have been hypothesized to occur during human infections ( 12 ). Given that most bacteria are found in mixed polymicrobial communities, it is tempting to speculate that cross talk between QS systems mediates numerous interspecies interactions ( 13 , 14 ). FIG 1 Diagram of a generic AHL QS circuit and structures of AHLs used in this study. (A) AHL QS systems generally contain a synthase (I) that produces an AHL signal, depicted here as a star. The signal acyl chains vary in length from 4 to 20 carbons, with potential hydroxy or oxo modification on the 3rd carbon, double bonds, branching, and/or terminal aryl moieties. At low cell densities, signals diffuse away from cells. At high cell densities, signals accumulate and can bind the QS receptor (R), which is a cytosolic transcription factor that regulates genes involved in group behaviors. (B) Chemical structures of AHLs used. Non-IUPAC descriptions of compounds are as follows: (1) C4-HSL; (2) 3OHC4-HSL; (3) C6-HSL; (4) 3OC6-HSL; (5) 3OHC6-HSL; (6) C8-HSL; (7) 3OC8-HSL; (8) 3OHC8-HSL; (9) C10-HSL; (10) 3OC10-HSL; (11) 3OHC10-HSL; (12) C12-HSL; (13) 3OC12-HSL; (14) 3OHC12-HSL; (15) C14-HSL; (16) 3OC14-HSL; (17) 3OHC14-HSL; (18) C16-HSL; (19) 3OC16-HSL. Despite their potential importance, we know little about the prevalence, function, and evolution of promiscuous QS receptors. There have been prior studies aimed at comprehensively profiling the responses of a set of receptors to large sets of natural and synthetic ligands ( 15 , 16 ), and multiple studies have measured the selectivity of individual receptors against smaller sets of AHLs ( 7 , 17 – 23 ). These studies were limited, however, in that many of them used heterologous expression of the receptors in Escherichia coli . Due to tractability, signal preferences and receptor selectivity are frequently studied using E. coli engineered to report receptor activity ( 24 ). Such methods require artificial expression of the AHL receptor, likely to a higher degree than the receptor’s natural expression level. Importantly, increased AHL receptor expression has been linked to increased sensitivity and promiscuity ( 19 , 25 ), and a previous study comparing LasR receptor activity in its parent species, Pseudomonas aeruginosa , with heterologous expression in E. coli found significant discrepancies between these two methods ( 26 ). Previous reports may, therefore, overestimate receptor promiscuity and the potential for cross talk. We sought to systematically study QS receptor selectivity in the receptors’ natural parent strains, thereby avoiding overexpression and enabling more robust predictions of how bacteria would respond to nonself signals in nature. We selected seven receptors for our characterization: LuxR from Vibrio fischeri , CviR from C. violaceum , LasR, RhlR, and QscR from P. aeruginosa , and BtaR1 and BtaR2 from Burkholderia thailandensis ( Table 1 ). These organisms range from soil saprophytes ( C. violaceum and B. thailandensis ) to a squid symbiont ( V. fischeri ) to human ( P. aeruginosa and C. violaceum ) and plant ( P. aeruginosa ) pathogens and, with the exception of V. fischeri , are frequently members of polymicrobial communities ( 27 – 32 ). Their receptors control a variety of processes, including antibiotic production (CviR and BtaR2), extracellular enzyme production (LasR, RhlR, and CviR), and luminescence (LuxR). Critically, the selected QS systems are well described, enabling study of their activation in the bacteria that naturally express them. TABLE 1 Sensitivity of AHL receptors to cognate signals Organism Receptor Signal EC 50 a \n E. coli EC 50 b \n P. aeruginosa LasR 3OC12-HSL 593 ± 128 nM 12.9 ± 3.6 nM RhlR C4-HSL >100 µM d \n 122 ± 17 µM QscR 3OC12-HSL c \n 1.90 ± 0.27 µM 53.4 ± 11.3 nM B. thailandensis BtaR1 C8-HSL 50.5 ± 4.6 nM 10.5 ± 3.6 nM BtaR2 3OHC10-HSL 15.0 ± 5.3 nM 60.6 ± 16.0 nM V. fischeri LuxR 3OC6-HSL 272 ± 15 nM NT C. violaceum CviR C6-HSL 83.4 ± 24.5 nM NT a EC 50 is the concentration required for half-maximal activity of the receptor in its native host. b EC 50 values for activation of receptors heterologously expressed in E. coli (DH5α). NT, not tested. c QscR is an orphan/solo receptor and does not have a paired signal synthase, but it does respond to 3OC12-HSL produced by LasI. d RhlR activity was not saturated at 1 mM C4-HSL. To measure selectivity, we quantified receptor responses to a panel of synthetic AHL signals, calculating both percent activation and concentration of half-maximal activation (EC 50 ) for each signal. To better compare our results to previous studies, we also made the same measurements using heterologous expression of the receptors in E. coli . The E. coli reporters consistently overestimated sensitivity and promiscuity for our selected receptors. We determined that overexpression of the receptors is sufficient to account for the differences between E. coli and native reporters. Surprisingly, we also found that overexpression of the receptors can lead to AHL-independent activity in P. aeruginosa . By using our activation data, we developed a quantitative selectivity score for each receptor. We found that the receptors display a wide range of signal preferences and selectivity. Some receptors, such as RhlR, are highly specific for their cognate signal, while others, such as BtaR2, are very promiscuous. The remaining receptors are on a continuum, with many displaying intermediate levels of selectivity and the ability to respond strongly and sensitively to at least one noncognate signal. These results suggest the potential for significant AHL-mediated interspecies interactions in nature and are a prelude to understanding the evolution of signal and receptor diversity.",
"discussion": "DISCUSSION AHL QS has long been recognized as a form of intraspecies bacterial communication. There are, however, examples of “promiscuous” AHL signal receptors, which are able to respond to signals other than their self-produced cognate AHL ( 8 , 9 ). Existing data on QS selectivity are limited and often generated by E. coli reporter methods in which overexpression of the receptor may artificially enhance promiscuity. To better understand the prevalence and potential function and evolution of promiscuous QS receptors, we systematically studied the selectivity of AHL receptors in their native host organisms. To compare our results with previous studies, we also constructed reporters of receptor activity using heterologous expression in E. coli . The E. coli reporters consistently overestimated receptor sensitivity and promiscuity. Further, we found that overexpression of the AHL receptors in P. aeruginosa was sufficient to increase receptor sensitivity and promiscuity to levels equal to or greater than those of the E. coli reporters. Transcription of the target DNA in our reporter assays is a reflection of many processes, including AHL receptor stability, receptor dimerization, and, ultimately, receptor binding to DNA. AHL binding both stabilizes receptors by promoting proper folding and protecting them against proteolysis and promotes receptor dimerization and binding to DNA ( 55 – 57 ). When considering activity in our reporter assays as a reflection of a binding reaction, receptor + AHL ⇌ receptor · AHL, the concentration of the receptor-AHL complex, and therefore the activity of the reporter, is dependent on both the concentration of AHL ([AHL]) and [receptor] in addition to the affinity of the receptor for the AHL. This fundamental principle of protein-ligand interactions can explain how increased expression of the receptor increases the sensitivity of the activity reporter to both cognate and noncognate AHLs. Indeed, changes in receptor expression and/or stability have been linked to generalized changes in receptor sensitivity previously ( 19 , 25 , 26 ). Importantly, AHL receptor expression is typically affected by complex regulatory systems, and receptor expression levels can vary between strains and between environmental conditions ( 58 , 59 ). It is likely that this variability in expression level leads to variable AHL receptor responses to both self and nonself signals in natural systems. Surprisingly, we also found that receptor overexpression in P. aeruginosa results in AHL-independent activity. Although it is possible that a non-AHL small molecule is responsible for the observed activity ( 54 , 60 ), it is also possible that artificially high expression of the receptors could drive ligand-free DNA binding in our system. Because PqsE was recently suggested to be the synthase of an alternative RhlR ligand ( 54 ), we tested the effect of pqsE deletion on RhlR activity. In our pqsE deletion strain, rhlR overexpression still resulted in AHL-independent RhlR activity. Given this finding and given that all three receptors (RhlR, LasR, and QscR) displayed AHL-independent activity when overexpressed in P. aeruginosa , we favor ligand-free activation as an explanation for our observed AHL-independent activity. AHL receptors are typically unstable in the absence of an AHL and require AHL for binding to promoters in vitro ( 20 , 55 , 57 , 61 ). However, some receptors, such as RhlR ( 54 ), are more stable in their parent strains than when expressed in E. coli or purified, and further, some orphan/solo AHL receptor homologs are able to exert AHL-independent control over their regulons ( 38 , 62 ). Perhaps increased expression of the P. aeruginosa receptors produces sufficient quantities of stable receptor to promote some degree of ligand-free DNA binding in the parent strain. E. coli reporter methods, of course, have important applications. First, the QS systems of many bacteria have not been studied well enough to construct reporters of receptor activity in the natural host organism. E. coli methods can also remove confounding factors that arise from the complex natural regulation of QS systems and their products. Some caution must be applied, however. The molecular mechanisms that modulate the activity of QS receptors and their regulons in natural host organisms are sometimes present and functional in E. coli as well ( 63 ). Additionally, E. coli has an AHL receptor, SdiA, which can interfere with receptor activity studies by activating transcription from the target promoter ( 64 , 65 ). Although some researchers use sdiA deletion strains ( 43 , 66 ), it is common to use readily available chemically competent cells such as TOP10 or, as we have, DH5α which have intact sdiA and the potential for artifacts associated with this receptor. We note, however, that in our E. coli experiments, activity of the reporter required expression of the receptor of interest and was, therefore, unlikely to be affected by SdiA. Finally, our findings highlight that results from any study using artificial expression of QS receptors should be interpreted with their limitations in mind, namely, the artifacts of increased receptor sensitivity and promiscuity, and the potential for ligand-independent activity. These findings may also inform the design of engineered cell circuits where it is important to limit cross talk between receptors and where AHL-independent activation of receptors may have detrimental effects on applications in engineered biosensors and targeted therapeutic delivery systems ( 67 ). By systematically and quantitatively measuring receptor responses in their natural backgrounds, we found that AHL QS receptors display a wide range of signal preferences and selectivity. Some AHL receptors, such as RhlR, are highly specific for their cognate signal. Because it is highly conserved across P. aeruginosa clinical isolates and is essential for virulence in animal models, RhlR has emerged as a potential antivirulence therapeutic target for P. aeruginosa ( 60 , 66 , 68 , 69 ). Encouragingly, RhlR’s specific detection of C4-HSL may be advantageous for the development of selective RhlR inhibitors. Our finding that RhlR activity is not saturated even at 1 mM C4-HSL is somewhat surprising, but it is consistent with previous studies where high concentrations of C4-HSL were required for maximal activity ( 23 , 66 , 69 ). The relative shallowness of RhlR’s dose-response curve to C4-HSL could allow for a greater ability to modulate activity of the RhlR regulon in nature. In laboratory cultures, clinical isolates of P. aeruginosa can make as much as fivefold more C4-HSL than the laboratory strain PAO1 and can also produce larger amounts of various RhlR-regulated products ( 68 ), possibly due to altered gene regulation and/or increased C4-HSL production. On the other end of the spectrum, certain receptors, such as BtaR2, are very promiscuous, responding to nanomolar concentrations of several signals. The selectivity of the rest of the AHL receptors lies on a continuum, with most receptors responding strongly and sensitively to at least one noncognate signal. Using these data, we can begin to make testable hypotheses about cross talk between QS systems in natural polymicrobial communities. In the context of a QS proficient strain, promiscuous activation by a noncognate signal often results in early activation of the QS receptor (i.e., activation at lower cell densities) ( 8 , 9 , 11 ). It is also important to consider that noncognate AHLs can inhibit receptor activation by cognate AHLs through various mechanisms, including partial agonism, receptor destabilization, and stabilization of the receptor in an inactive conformation ( 7 , 15 , 16 , 18 , 19 , 22 , 44 ). As with activation, the receptors in our study were variably sensitive to inhibition. We and others have found that LuxR and CviR are sensitive to inhibition by noncognate AHLs ( 7 , 15 , 16 , 18 ). In our study, LasR and QscR were less sensitive to inhibition. Previous studies have reported more significant inhibition of LasR activity by noncognate AHLs, with some AHLs acting as inhibitors at lower concentrations and as agnoists at higher concentrations ( 26 , 46 ). Therefore, we may have missed the inhibitory activity of some AHLs by measuring at a single concentration. In any case, it is clear that AHL receptors are susceptible to inhibition by noncognate AHLs and that there are likely a multitude of complex positive and negative interactions mediated by AHLs in natural polymicrobial communities. Our quantitative scoring of receptor selectivity also serves as a basis for hypotheses about the benefits of specific versus promiscuous QS activation and about how the diversity in QS signals and receptors evolved. With the exception of QscR, which has no cognate signal, all receptors tested were most sensitive (i.e., they respond with the lowest EC 50 ) to their self-produced cognate signal. By amino acid sequence, QscR is more closely related to AHL receptors from other organisms than to the other P. aeruginosa receptors, LasR and RhlR ( 70 ). QscR has, therefore, been hypothesized to have been introduced to P. aeruginosa via horizontal gene transfer ( 20 ). It is possible that QscR originates from a C10-HSL-responsive receptor and has evolved additional 3OC12-HSL recognition. Alternatively, promiscuous QscR activation could have arisen due to some selective advantage for P. aeruginosa . Interestingly, the two most promiscuous receptors, CviR and BtaR2, control the synthesis of potent antibiotics, violacein ( 7 ) and bactobolin ( 21 ), respectively. We have previously shown that promiscuous activation of CviR can confer a competitive advantage to C. violaceum due to QS-controlled antimicrobial production ( 8 ). We speculate that the selectivity of each AHL receptor may be dictated by its regulon, i.e., by a selective advantage gained from either specific or promiscuous receptor activation, and/or by the evolutionary history of the receptor. We are just beginning to understand the large diversity in AHL receptor structure and function. To gain insight into the origin and function of promiscuous signaling, it will be necessary to determine what, if any, impact exists for the promiscuous activation of each receptor. Our comprehensive data set on AHL receptor selectivity gives us a rational basis to begin to address longstanding questions about how the diverse array of AHL signal synthase-receptor pairs has evolved in proteobacteria and about how signaling systems interact in nature."
} | 4,831 |
38551008 | PMC10973203 | pmc | 1,073 | {
"abstract": "Summary Building machines that can augment or replace human efforts to accomplish complex tasks is one of central topics for humanity. Especially, nanomachines made of discrete numbers of molecular components can perform intended mechanical movements in a predetermined manner. Utilizing free energies of Watson-Crick base pairing, different types of DNA nanomachines have been invented to operate intended stepwise or autonomous actions with external stimuli, and we here summarized the motive forces that drive DNA-based nanomachineries. DNA tweezers, DNA origami actuators, DNA walkers, and DNA machine-enabled bulk sensing are discussed including structural motif design, toehold creations for strands displacement reactions, and other input forces, as well as examples of biological motor-driven hybrid nanomachines. By addressing these prototypical artificial nanodevices, we envision future focuses should include developing various input energies, host cell-assisted structure self-replication, and nonequilibrium transportations.",
"conclusion": "Conclusion Humans always want to build a human-like machine to do all kinds of work instead of humans. Man-made uncomplicated energy-conversion systems such as water mill and windmill can transform the kinetic energy of wind or water into mechanical energy, assisting humans to accomplish hard jobs. Other than energy converters, robots are decidedly more complex and can replace humans to perform repetitive and dangerous tasks with high accuracy and reliability. Guided by external controls or preset programs, robots operate autonomously and exhibit predetermined behaviors in response to their environment. In the field of synthetic biology, it shares a similar objective to build smart entities that are comparable to beings than to machines. As summarized here, the particular example of DNA nanomachineries has exploited tremendous efforts to assemble nanostructures conducting a complex series of predesigned actions. Despite many obstacles remaining to be surmounted, DNA-based smart and functional motor systems operated by desired inputs ( Table 1 ) have been established. To further advance DNA nanomachines, self-replication to reproduce itself is another aim. 118 In modern factory automation, robotic machines such as drilling, shaper, welding, and painting can work coordinately to manufacture designed machines. There have been several profound studies on bioproduction of ssDNA oligonucleotides. 119 Combining isothermal assembly assisted by enzymes 120 and kinetics control, it may be feasible to harvest designed DNA nanomachines with microbial production in a prescribed manner. With the fast developed computer-aided sequence design with cross-linked reactions and a sufficient supply of DNA materials, we believe DNA-enabled synthetic nanomachines with spatial and temporal controls can accomplish specified assignments in real-life biomedical and clinical applications. Table 1 Different motive forces that drive DNA nanomachine Interaction type Specific method Intermolecular DNA interactions Strands displacement reactions pH-responsive triplex DNA Base pair stacking Intramolecular DNA interactions B-Z transitions Entropic elasticity Hairpins and quadruplex Hybrid driving forces Enzyme assistance DNA aptamer Motor proteins Non-DNA interactions Hydrophobic effect Host-guest recognition Peptide-peptide interactions Macroscopic manipulation Light, electric and magnetic field",
"introduction": "Introduction During a bike transportation from home to office, it follows the bike lane and safely carries the passenger to the directed place. As an example, cycling can be used to interpret the term machine which converts energy and does work. 1 The feet on pedals transmit the power of downstroke to forces that rotate wheels forward through roller chain and series of gears ( Figure 1 A). Such mechanical power transmission illustrates the basic endeavors that people have challenged to create machines and make jobs less difficult. Apart from macroscopic man-made machineries, life evolution has also developed tremendous amounts of molecular motors that conduct multi-level sophisticated tasks to drive vital activities. For instance, the small rotary motor ATP synthase catalyzes the formation of ATP—the energy currency—from ADP and orthophosphate. The transmembrane proton flow propels the c-ring subunit to spin and promote the ATP synthesis with the binding-change mechanism ( Figure 1 B). Figure 1 Bicycle and ATPase as macroscopic man-made machine and the smallest biomolecular motor, respectively (A) Bicycle converts the inputted man power to the rotation of wheels. The forces (red arrow) that pushed down on the pedals rotate the crankset (yellow arrow), generating a torque which is transmitted to the wheel through the bicycle chain (green arrow) and rear cassette (blue arrow). The forces of the wheels against the ground cause the backward forces (F rear -F fwd ) to wheels, propelling the bicycle forward. (B) The electron-transport chain generates a proton gradient across the inner mitochondrial membrane. Protons then flow back into the matrix to equalize the distribution, creating a proton-motive force which drives the spinning movement of c-ring and the rotational catalytical phosphorylation of ADP to ATP. To imitate the macroscopic and biological counterparts and expand the machinery design principle to nanoscale, it inspires to create synthetic nanomachines. Energies from small motions or coupled chemical reactions can be harnessed to supply powers for nanoscale actuations that can perform enormous roles in biomedical, mechanical, sensing, and monitoring applications. Resembling the real machines at the macroscopic level, synthetic nanomachines receive input energy and transform it into the work output, producing quasi-mechanical movement in artificial or biological microenvironment. Surrounded by various species, nanomachines encounter multi-level interactions and thermodynamic fluctuations and thus demand an accurate design with high energy-conversion efficiency. Moreover, the intuited creation of miniature-sized nanomachines is usually aimed to mimic the biological motors. Learning from coupled cascade reactions in biological systems, synthetic nanomachines arrange chemical stimulus one by one in orders so that it can operate according to the inputted information. Though far less complicated, nowadays molecular switches and shuttles from cis-trans isomerism to “rotaxanes” and “catenanes,” as well as cross-linked responsive chemical reactions chains at nanoscale, have been built up. Here in this review, we focus on DNA-enabled synthetic nanomachines which either employ DNA structures as machine body or are driven by DNA inputted energies. Firstly, DNA nanotechnology has demonstrated the capacity of using DNA, omitting its biological functions but as programmable polymers, for numerous nanostructures fabrication. Through sequence design to achieve discrete hybridization events between certain segments of different DNA strands, many assembly strategies have been invented to implement the structural DNA nanotechnology. Owing to the highly specified DNA sequences, each DNA strand that weaved into nanostructures can be precisely identified at addressable locations. Hence the design and assembly of DNA nanostructures can be fine-tuned in a controllable manner to realize structures with defined shape, sizes, and curvatures. Though the reported sequence-averaged persistence length of single DNA duplex is merely ∼50 nm, multiple DNA duplex can be bundled up through crossover connections (e.g., in origami) to form larger rigid entities not subjected to this limitation. Moreover, the mature nucleic acid chemistry can introduce various functional groups on DNA strands such as amino and thiol, dyes, and click chemistry tags to enable DNA-templated assembly for many nanomaterials. Secondly, in the classic Watson-Crick duplex model, the highly structured double-stranded DNA (dsDNA) is cooperatively stabilized by hydrogen bonding between A=T or G≡C and π-stacking of neighboring base pairs. DNA duplex is only a marginally stable system which depends on many factors such as temperatures, salt and buffer concentrations, GC content (the proportion of Guanine and Cytosine in a sequence), and sequence length. Even at the physiological environment, it is consistently undergoing thermally driven structural fluctuations (also known as “breathing”), the temporarily unwinding of a limited region of duplex to form a small single-stranded DNA (ssDNA) bubble. Such thermodynamic window allows for not only many fundamental biological reactions but also the sequence-specific hybridization and low-energy-favored strands displacement reactions (SDRs). In addition, DNA strands that are encoded with desired sequences possess several secondary structures, from duplex of A-form, B-form, and Z-form, to triplex and quadruplex which offer various means to realize regulatory dynamics. As organized in the current paper, we will first summarize state-of-the-art motive forces involved in DNA nanomachines. We termed intermolecular interactions as the engagement of different objects such as hybridization between two oligonucleotides and stacking interactions among structures, while intramolecular interactions are via formation of intramolecular structures such as hairpins. Non-DNA interactions as well as macroscopic manipulation techniques such as light and electric and magnetic field are later included. Then two unique research fields DNA walker and DNA machine-enabled sensing are discussed."
} | 2,401 |
33195040 | PMC7596381 | pmc | 1,074 | {
"abstract": "Bioinspired superhydrophobic surfaces are an artificial functional surface that mainly extracts morphological designs from natural organisms. In both laboratory research and industry, there is a need to develop ways of giving large-area surfaces water repellence. Currently, surface modification methods are subject to many challenging requirements such as a need for chemical-free treatment or high surface roughness. Laser micro-nanofabrications are a potential way of addressing these challenges, as they involve non-contact processing and outstanding patterning ability. This review briefly discusses multiple laser patterning methods, which could be used for surface structuring toward creating superhydrophobic surfaces.",
"conclusion": "Conclusion and Outlook This minireview has introduced three laser patterning methods. These methods could be used to make geometrical modifications to various materials in realizing bioinspired superhydrophobic surfaces. Fast scan by DLW enables micro–structuring (such as microgratings or microgrids), which is convenient for constructing hierarchical surface structures, and can be easily combined with the other two methods discussed above. DLIP can directly acquire periodical 3D micro-nanostructures and simply manipulate surface wettability. Moreover, LIPSS can be self-organized during the structuring process of either DLW or DLIP with an even smaller scale (a spatial period of hundreds of nanometers). Since the three laser-structuring methods compensate for one another in either processing speed or structure scale, combined laser-surface-patterning is more advantageous in large-area-processing, particularly bioinspired superhydrophobic surfaces. Future studies might consider using more light sources (besides the infrared) for processing. The phase modulation of laser beams might also further increase the structuring speed. Comprehensive further exploitation of these laser-processing techniques will most likely continue to improve current surface modification methods, enabling bio-functionalization.",
"introduction": "Introduction Bioinspired surfaces are surfaces with artificial micro-nanostructures that mimic functional biological structures in natural organisms. They exploit the diverse functions of the natural counterparts, who have acquired rich biological features and advantages during their long-term evolution and adaptation to nature. Among these biological features, surface wettability is a key property due to its extensive possible applications. There are a number of natural surfaces that show properties such as superhydrophobicity as well as the ability to self-clean, including rose petals, reed leaves, and even the skin of some animals (Martinez-Calderon et al., 2016a ). It has been fully demonstrated that such superhydrophobicity is decided by both surface chemistry and unique surface micro-nanostructures (Florian et al., 2018 ), which provides the basic strategy for developing bioinspired superhydrophobic surfaces. In rose petals, the surfaces consist of densely packed microhills covered by a large number of microtrenches, whose wetting regime corresponds to the Wenzel state (Feng et al., 2011 ). In another example, lotus leaves, both papilla-like microstructures and ceraceous-covered nanoscale textures form hierarchical micro-nanostructures and result in water-repellent behavior, whose wetting regime corresponds to the Cassie-Baxter state (Fan et al., 2015 ). During recent decades, manipulation of surface wettability, especially obtaining superhydrophobic surfaces, has proved to have potential applications in a series of fields, including for self-cleaning materials (Vorobyev and Guo, 2015 ), corrosion improvement (Su and Yao, 2014 ), and oil-water separation (Liu et al., 2017 ). Recently, it has also been found that such superhydrophobic surfaces may shed light on cutting-edge fields such as developing anti-bacterial properties (Bremus-Koebberling et al., 2012 ; Schieber et al., 2017 ) and de-icing aircraft (Jung et al., 2011 ; Heydari et al., 2013 ). Based on this research, both chemical and geometrical modifications have been adopted and used to obtain superhydrophobic surfaces. This can involve coating target surfaces with a hydrophobic layer (Kam et al., 2012 ; Weisensee et al., 2014 ), or treating them by wet etching (Weibel et al., 2010 ). The main problem with chemical surface treatment is the unavoidable chemical residue that sometimes remains on target surfaces, which may be toxic or even negatively affect the device performance under certain temperature and pressure conditions. In pursuit of extreme surface roughness, a variety of micro-nanofabrication techniques can be adopted, including molding (Zhao et al., 2008 ), electrodeposition (Liu et al., 2016 ), and photolithography (Limongi et al., 2015 ). Among these techniques, laser structuring is known for its non-contacting, wide material-adaptability, and advanced 3D patterning ability. The first work to demonstrate laser-structured superhydrophobic surfaces dates back to 2006 when silicon surface was textured with microspikes by 800-nm femtosecond (fs) lasers under the flow of a reactive gas SF 6 (Baldacchini et al., 2006 ). Together with a 3-h chemical modification by (heptadecafluoro-1,1,2,2-tetrahydrodecyl) trichlorosilane [CF 3 (CF 2 ) 7 CH 2 CH 2 SiCl 3 ], the two-step treatment turned the silicon surface to superhydrophobic. With this increased laser fluence, hierarchical structures were also obtained on a silicon substrate by similar one-step processing via fs lasers under a reactive atmosphere (Zorba et al., 2008 ). The resulting microconicals and nanoprotrusions appeared all over the microstructure surfaces and successfully mimicked the surface structures of a lotus–leaf, realizing similar water–repellent properties. The formation of the microstructures was ascribed to the cooperation of capillary waves and laser-induced etching (Vorobyev and Guo, 2013 ). With more explorations in this scheme, it has been found that fs laser direct-writing can induce four kinds of self-organized structures: ripples, microgrooves, microspikes, and hierarchical complex structures (Stratakis et al., 2020 ). These micro-nanostructures give rise to surface textures as well as hydrophobicity/superhydrophobicity, and morphologies are tunable by laser parameters such as laser fluence, pulse duration, and repetition rate (Vorobyev and Guo, 2013 ; Stratakis et al., 2020 ). A series of bioinspired surfaces that highly mimic functional natural surfaces have been realized by flexibly utilizing multiple laser-processing methods (Stratakis et al., 2020 ). Surfaces have been created by fs laser direct-writing, which have also proved to have applications in a variety of fields, including self-cleaning, stimuli-responsive surfaces (light, electro, pH, etc.) (Papadopoulou et al., 2009b ; Stratakis et al., 2010 , 2011 ), cell adhesion (Ranella et al., 2010 ), tissue engineering (Papadopoulou et al., 2009a ), etc. (Vorobyev and Guo, 2013 ; Stratakis et al., 2020 ). Recently, it has also been demonstrated that tuning fs-laser polarization could induce hierarchical surface structures on Ni, due to the unique donut-like profile of radically- and azimuthally- polarized fs lasers (Skoulas et al., 2017 ). In order to avoid chemical pollution to target surfaces, one can focus on direct laser-structuring methods, which are laser-induced periodic surface structures (LIPSS), direct laser-interference patterning (DLIP), and direct laser writing (DLW). Herein, we have briefly reviewed recent achievements in obtaining superhydrophobic surfaces via the laser-structuring methods. According to the diverse surface properties of the target materials as well as the varied requirements of different applications, a single technique or even multi-techniques can be used to define functional surface micro-nanostructures for manipulating surface wettability. This covered the technological features of the three distinctive laser-structuring methods and the resulting surface structure morphologies due to laser-enabled geometrical modifications. Especially, the flexible method combinations involved in obtaining complex hierarchical surface structures for extreme surface roughness as well as water-repellent property."
} | 2,079 |
35516104 | PMC9059740 | pmc | 1,075 | {
"abstract": "Biological fouling, where marine microorganisms attach densely to various submerged surfaces, has been a serious economic problem worldwide. Different from most antifouling approaches based on stiff and solid materials or coatings, a soft and wet coating composed of zwitterionic polymer was prepared in this paper. With the combination of the anti-polyelectrolyte effect of poly- N -(3-sulfopropyl)- N -(methacryloxyethyl)- N , N -dimethylammonium betaine (PSBMA) and the typical polyelectrolyte effect of polyacrylic acid (PAA), a bicomponent hydrogel coating with weak swelling in saline solution was achieved, which could avoid peeling from solid substrates. The bicomponent hydrogel coating showed strong tensile properties and good compression performance and slipperiness. Although the large Young's modulus of the coating relatively weakens the drag reduction effect, entering the mixed lubrication region in low sliding rate is easy and a low friction coefficient at a high rate could thus be obtained. With the aid of silane coupling agent and weak deformation in water and saline solution, the hydrogel coating could be bound tightly on solid surfaces. After strong sandy water abrasion, the bicomponent hydrogel coating could maintain its original state without any cracks and peeling. The hydrogel coating exhibits good anti-bacterial adhesion and anti-protein adsorption. The bicomponent zwitterionic hydrogel coating reported here provides a new strategy for marine antifouling and drag reduction studies.",
"conclusion": "4. Conclusions Exhibiting the combination of the anti-polyelectrolyte effect of PSBMA and the typical polyelectrolyte effect of PAA, the bicomponent hydrogel coatings with weak swelling in saline solution were achieved, which could avoid the peeling from solid substrates. The bicomponent hydrogel coatings exhibited good mechanical properties and relatively small swelling degree and were slippery. Although the large Young's modulus of the coatings weakens the drag reduction effect relatively, entering the mixed lubrication region in a low sliding rate and thus a low μ at high rate could be easily obtained. With the aid of silane coupling agent together with the weak deformation in water and saline solution, the hydrogel coatings could be bound tightly on the solid surface. After undergoing strong sandy water abrasion, the bicomponent hydrogel coatings maintained their original state without any cracks owing to their slippery surface and tough adhesion. Furthermore, the hydrogel coatings exhibited good AF properties. The bicomponent zwitterionic hydrogel coating reported in this paper provides a new idea for the marine AF and drag-reduction study.",
"introduction": "1. Introduction Ships, oil platforms, and other facilities inevitably encounter marine biofouling problems, which is a global problem for marine industry and activities 1,2 and has been a serious economic problem worldwide. 3,4 Marine biofouling usually occurs on an immersed surface as a result of several successive steps from the formation of a conditioning film followed by the attachment of macroalgae, fungi, protozoa and the last invertebrate larvae. 4,5 This phenomenon increases the surface roughness and the weight, causing reduced speed and manoeuvrability of ships, frequent dry-docking cleaning, and high fuel consumption. 6 Antifouling (AF) compounds and methods have been developed to inhibit fouling by marine organisms. Tributyltin (TBT), 6 which exhibits high AF activity, is the most popular AF compound at one time. However, TBT was banned worldwide since 2008 due to its high endocrine disruptive effect (IMO 1999). Exploring novel alternative AF systems is urgently necessary to protect submerged surfaces and marine environment. Several environmentally benign AF approaches have been developed, such as surface coating with natural AF compounds 7,8 and fouling-release coatings based on silicones and fluorocarbon polymers that have low surface energy. 9,10 Most of these AF approaches are based on stiff and solid materials or coatings, which are effective only on vessels moving at speeds greater than 14 knots ( e.g. Intersleek 900). 11 Biofouling process is affected by many physical–chemical factors, such as surface tension, 12 wettability, 13 elastic modulus, 14 surface chemistry, 15 surface roughness 16 and topography. 17,18 Considering that the formation of biofilm is a nonspecific and reversible process, highly hydrated surface is proposed as a feasible approach to develop promising materials for fouling resistance applications. 19–21 Recently, zwitterionic compounds, including 2-methacryloyloxyethyl phosphorylcholine, 22–24 carboxybetaine methacrylate (CBMA), sulfobetaine methacrylate (SBMA) 25–30 and so on, have been found to exhibit ultra-low-fouling, indicating that the surfaces coated with these polymers allow less than 5 ng cm −2 of protein adsorption. 27–29 The surfaces coated with zwitterionic poly-CBMA highly resist non-specific protein adsorption even from undiluted blood plasma and serum 27 and prohibits long-term bacterial colonization by Pseudomonas aeruginosa for up to 10 days at room temperature 31 and the attachment of green marine alga, diatoms 32,33 and Amphibalanus amphitrite. 34 The ultra-low-fouling performance of zwitterionic materials is due to the high hydration around the opposing charges and high energetics required to remove that hydration layer. 27,28 However, zwitterionic compounds are usually used in a form of nanofilms by means of superficial atom transfer radical polymerization or layer-by-layer self-assembly, that requires fastidious synthetic conditions 35 and is unavailable for large-scale construction of marine AF coatings. As far as highly hydrated surface is concerned, hydrogel materials, being soft, wet and exhibiting many fascinating properties which cannot be found in solid and liquid materials, followed into our spotlight. The surface properties of hydrogels can be easily modulated by selecting monomer species. Many synthetic polymer hydrogels, including neutral, positively charged and negatively charged hydrogels, exhibit remarkable AF performance against algae 15 and barnacle cyprid larvae in vitro . 16 Moreover, hydrogels exhibit extremely low surface frictional forces against themselves or solid substrates. 10,11,36 Therefore, hydrogels have been investigated as AF material due to its soft surface, drag-reducing effect and low surface energy, just like fish epidermal mucus. Gong et al. reported that one of the key factors of hydrogel AF performance is a low elastic modulus. 37 Researchers have successfully developed many new kinds of cross-linked polymer hydrogels for effective marine AF. 38–44 Jiang et al. prepared hydrogels with hierarchical surfaces displaying superoleophobicity, which show the potential for AF application. 45 All of the hydrogel samples prepared are independent bulk hydrogels that ignore the adhesion between hydrogels and solid substrates. As fouling occurs most in solid surface, such as ship's hulls, aquaculture cages, cooling water intake channels of power plants and so on, the applicability of bulk hydrogels is limited due to the processing difficulty. Another method for hydrogel coating preparation is to blend gel polymer with coating resin, and then a layer of hydrogel is formed on solid surface through water absorption and leaching out of the polymer. 46 In addition, Hansen M. R. et al. synthesized a cross-linked zwitterionic thin film to the PDMS surface by the ultraviolet photopolymerization method. The paper paid more attention to the potential of coatings in biological applications, and the film only worked under mild cell and protein adsorption conditions. 47 Rosenhahn et al. prepared low fouling thin hydrogel coatings by synthesizing sulfobetaine- and sulfabetaine-bearing zwitterionic copolymers containing a photo-cross-linker. 48 These method are not in the preliminary stage of research and is far from actual application in the sea. 46,49 Developing absolute hydrogel coatings seems extremely difficult due to its poor adhesive strength to substrates and the strong swelling or shrinking in pure water or saline solution. In this paper, a zwitterionic hydrogel coating with hydrophilic, soft and slippery submerged surface was prepared on solid surface with strong cohesiveness. SBMA was used to construct the AF and slippery hydrogel coating. Simple SBMA hydrogel coating showed excellent AF and drag-reducing performances; however, the swelling or shrinking behaviors during processing hindered its practical application in harsh marine environment. Poly-SBMA (PSBMA) exhibits antipolyelectrolyte effect, implying that PSBMA hydrogel would swell when transferred from water to saline solution. Conventional water-soluble polymers, especially charged polymers with typical polyelectrolyte effect, were introduced to form bicomponent hydrogel coating to suppress such swelling that would make hydrogel coating be peeled off from the solid substrate. On the other hand, the single PSBMA hydrogel coating is fragile due to its stiff macromolecular chain. The design of such bicomponent hydrogel could improve simultaneously the mechanical properties of hydrogel based on the double network (DN) or interpenetrating network method 40 to some extent. Building on the ideas above and the aid of silane coupling agent, tough, antifouling and slippery bicomponent hydrogel coatings adhered on solid substrate were prepared in this paper. The hydrogel swelling behaviors during processing, mechanical properties, anti-bacterial adhesion performance, anti-protein adsorption performance and drag-reduction properties of various coatings (single or bicomponent) were explored in detail.",
"discussion": "3. Results and discussion 3.1. Swelling behavior during hydrogel coating preparation Hydrogels are usually prepared in an aqueous circumstance and will shrink or swell to some extent when immersed into sea water with high salt content. When one side of the hydrogels as a coating is adhered on a solid surface, the swelling or shrinking is anisotropic. Such heterogeneous deformations would cause the coating to partially or completely peel off from the solid substrates. During synthesis of bicomponent hydrogel coating, the first hydrogel network must be immersed into the precursor solution of the second monomer with high concentration. In this case, the first network that has adhered on the substrate might be peeled off due to the shrinking or swelling. Thus, swelling behaviours of the first network in the second monomer precursor solution and the ultimate hydrogel coating in saline solution were first investigated. According to DN hydrogel method, 51 the first hydrogel network is composed of crosslinked stiff macromolecules, whereas the second hydrogel network is uncrosslinked flexible polymer chain. In this paper, both the mechanical and the anti-fouling properties of the hydrogel should be considered simultaneously. Although the chains of PSBMA are stiff and the corresponding hydrogel is fragile, the hydrogel coatings with PSBMA as either the first or second network were prepared in this paper. Moreover, the two networks were all crosslinked. Thus, the ultimate hydrogels were termed as bicomponent hydrogel. In the case of PSBMA as the second network, Fig. 1a presents the swelling ratio of the four first hydrogel networks, PAA, polyacrylamide (PAM), polymethacrylic acid (PMAA) and poly(2-acrylamido-2-methylpropanesulfonic acid) (PAMPS), in SBMA monomer solutions with a concentration of 1 M. Here, the swelling ratio is the volume of the hydrogel in the second monomer ( i.e. SBMA) solution ( V 2 ) to that the volume of the hydrogel just synthesized by using a UV radiation ( V 1 ). After transferring from the glass mould into SBMA solution, several hydrogels were all swollen. Possessing large methyl as side group compared with PAA, PMAA hydrogel exhibits a swelling ratio close to 1. PAM hydrogel, as a neutral polymer hydrogel, shows a swelling ratio of 1.8. PAMPS, as a typical anionic polyelectrolyte with sodion as contra-ion, exhibits much large swelling ratio of ∼10. Note that PAA, PAM and PMAA hydrogels show lower swelling ratio than that of PAMPS hydrogel, and the superabsorbent capacity comes from the strong intrinsic polyelectrolyte macromolecules. The swollen balance of these hydrogels in SBMA solution is dependent on the hydrophilicity of the polymer chain, the degree of crosslinking, ion interaction and the osmotic pressure. PMAA chains show relatively weak hydrophilicity owing to the large methyl group and thus low swelling ratio. Although SBMA does contain both positively and negatively charged hydrophilic groups in the same molecules, SBMA is formally uncharged, which causes weak ion interaction between SBMA and the chains of hydrogels. Furthermore, these opposing charges lead to large dipole moments for SBMA molecules with hydrophilicity intermediate between the ionic and conventional nonionic chemicals. In this paper, the concentration of SBMA solution is 1 M, being much higher than its critical micelle concentration, 52 implying that most SBMA molecules are in the form of micelles, which generate lower osmotic pressure. All of the mentioned factors above make PAA, PAM and PMAA hydrogels display low swelling ratio in SBMA aqueous solution. Peeling off from the solid substance is difficult if these three hydrogels act as the first network. Fig. 1 Swelling ratio of the first hydrogel network in the second monomer solution. (A) PAA, PAM, PMAA and PAMPS hydrogels in SBMA aqueous solution with a concentration of 1 M; (B) PSBMA hydrogel in AA, AM and AMPS aqueous solution with a concentration of 1 M. In the case of SBMA as the first network, Fig. 1b presents the swelling ratio of the PSBMA hydrogel networks in the second monomer (AA, AM or AMPS) aqueous solution with a concentration of 1 M. After being transferred from the glass mould into AA aqueous solution, PSBMA hydrogel slightly shrank. PSBMA hydrogel could maintain its original shape in AM solution and would neither shrink nor swell. In AMPS solution, the swelling ratio is approximately 3.0, which is the largest one. Note that the swelling ratio of PSBMA hydrogel in the second monomer aqueous solution was much lower than that of PAA, PAM, PMAA and PAMPS hydrogel in SBMA solution owing to the intrinsic zwitterionic character of PSBMA chain. The result of Fig. 1 reveals that AA, MAA and AM could be selected as the other component to construct the bicomponent hydrogel coatings whether SBMA is used as the first or second monomer. Considering that most fouling algae are negatively charged together with the weak swelling during preparation processing, more flexible and negatively charged PAA was selected to construct the bicomponent hydrogel coating with SBMA. Exhibiting the combination of the anti-polyelectrolyte effect of PSBMA and the typical polyelectrolyte effect of PAA, the bicomponent hydrogel coatings would show weak swelling in saline solution and avoid the peeling from solid substrates. The swollen ratios of PAA, PSBMA, PAA/PSBMA ( i.e. PAA/PSBMA-i in Table 1 ) and PSBMA/PAA hydrogels in NaCl solution at various concentrations and artificial sea water (ASW, sea salt aqueous solutions with a concentration of 3.3 wt%) are shown in Fig. 2 . Here, the swollen ratio is the volume of hydrogel in solution ( V ) to that in deionized water ( V 0 ). PAA hydrogel shrinks while PSBMA hydrogel swells in saline solution. PAA hydrogel coating, as a typical anionic polyelectrolyte, exhibits a relatively larger volume in low concentration of saline solution due to electrostatic shielding effect. 53 When the concentration of NaCl solution is higher than a certain value, the swollen ratio decreases with concentration, that is, volume collapse occurs and causes shrinkage. The swelling degree of PAA hydrogel in ASW is only 0.3. By contrast, PSBMA hydrogel, as a zwitterionic polymer, exhibits significant antipolyelectrolyte effect, that is, the molecular chains are more stretchable in NaCl solution causing volume expansion and the swollen ratio increases with the concentration of NaCl solution. The swollen ratio of PSBMA hydrogel in ASW is 2.9 or so. Combining the above two polymers with opposite electrolyte effect in saline solution, the hydrogel coating with weak swelling could be achieved. As shown in Fig. 2 , both the PAA/PSBMA-i and PSBMA/PAA bicomponent hydrogels exhibit lower swollen ratio than that of single PSBAM and PAA hydrogel in NaCl solution and in ASW (1.5–1.8). Moreover, the swollen ratio for the bicomponent hydrogels changed slightly after varying the concentration of NaCl solution. Increasing the crosslinker loading of PSBMA that acted as the second network (such as PAA/PSBMA 0.5% ) leads to the smaller swollen ratio and weaker sensitivity to the concentration of NaCl solution. Fig. 2 Swollen ratio of PAA, PSBMA, PAA/PSBMA and PSBMA/PAA hydrogel in saline solution and ASW. The samples correspond to those shown in Table 1 . 3.2. Tough bicomponent hydrogel \n Fig. 3 presents the tensile and compressive stress–strain curves of PAA, PSBMA, PAA/PSBMA-i and PSBMA/PAA hydrogel in water and saline solution. Both PSBMA and PAA hydrogels exhibit low tensile strength ( σ ) and elastic modulus ( E ). PSBMA hydrogel is especially fragile in water and its σ and elongation at break ( ε ) are extremely low, only 2 kPa and 30% respectively, which is not enough to resist water scour as the coating of an immersed surface. After combination with PAA, such as in PAA/PSBMA-i and PSBMA/PAA hydrogels, the mechanical properties were increased remarkably ( Fig. 3A ). The hydrogel with more crosslinker shows higher E and σ . To further evaluate the marine applicability of PAA/PSBMA-i hydrogel, we performed tensile tests after immersing in 1 M NaCl solution and ASW ( Fig. 3B ). PAA/PSBMA 0.5% with moderate mechanical strength was chosen for the further studies. Both σ and ε of PAA/PSBMA 0.5% hydrogel in saline solution decreased due to the slight swelling compared with that in water. Although the mechanical strength decreases, the strength of the bicomponent hydrogel was still high enough to be used in a form of submerged coating and could match the mechanical requirements of many applications for marine field. Fig. 3 (A) Tensile stress–strain curves for PAA, PSBMA, PAA/PSBMA and PSBMA/PAA hydrogel in deionised water. (B) Tensile stress–strain curves for PAA/PSBMA 0.5% hydrogel under deionised water, 1 M NaCl solution and ASW. (C) Compressive stress–strain curves for PAA, PSBMA, PAA/PSBMA-i and PSBMA/PAA hydrogel. (D) Compressive properties of several hydrogels. Besides the tensile mode, the mechanical properties under compressive mode were also measured as shown in Fig. 3C and D . Similar to that under tensile mode, the bicomponent hydrogels exhibited high compressive modulus, strength and failure strain. Given the similar flexibility (or toughness) between PAA and PSBMA, slight difference in mechanical properties was observed for PSBMA/PAA and PAA/PSBMA-i. Swollen degree q (w/ w 0 ) is roughly the inverse of polymer volume fraction, E ∝ q −9/4 and E ∝ q −1 , which respectively prevail for neutral and polyelectrolyte hydrogels in water at equilibrium swelling state regardless of their preparation conditions. 54 To elucidate whether the bicomponent hydrogels behave like typical neutral or polyelectrolyte hydrogels, we plotted the relationship between E and q of the PAA/PSBMA-i hydrogel synthesised at different cross-linker loading and monomer concentrations of the precursor solution, as shown in Fig. 4 . We found that the scaling relationship of E and q follows E ∝ q −1.7 . The exponent −1.7 is between the theoretical value of −9/4 for a neutral hydrogel and −1 for a polyelectrolyte hydrogel. For a polyzwitterions hydrogel, such as poly( N -(carboxymethyl)- N , N -dimethyl-2-(methacryloyloxy)ethanaminium, inner salt) hydrogel, the exponent is −2.15 because of the presence of both the negatively charged –COO − group and the positively charged R 3 N + group and thus behaving as a hydrophilic neutral hydrogel in water. 40 The exponent −1.7 of PAA/PSBMA-i hydrogel is lower than −2.15 due to the introduction of polyelectrolyte PAA. Fig. 4 Relationship between q and E of PAA/PSBMA hydrogel. 3.3. Anti-bacterial adhesion and anti-protein adsorption Marine biofouling usually occurs on the immersed surfaces as a result of several successive steps and originates from the formation of a conditioning film, followed by the attachment of macroalgae, fungi, and protozoa to the last invertebrate larvae. The fundamental reason for the fouling of marine organisms is that protein adsorption leads to the growth of bacteria and thus the enrichment of marine organisms. To evaluate the AF property of the as-obtained bicomponent hydrogel coatings, anti-bacterial adhesion and anti-protein adsorption experiments were carried out. Fig. 5 and 6 presents the results of the adsorption tests of Escherichia coli ( E. coli ) and Staphylococcus aureus ( S. aureus ) on the four hydrogels, that is, two kinds of single component hydrogels and two kinds of bicomponent hydrogel coatings. PSBMA hydrogel exhibited fewest bacterial adhesion amount attributed to its strong hydrophilicity arising from specific zwitterionic structure, whereas weak polyelectrolyte PAA hydrogel exhibited a certain amount of bacterial adhesion due to its negatively charged surface. The anti-bacterial adhesion of the two bicomponent hydrogel was intermediate between the two single network hydrogels. PSBMA/PAA hydrogel with PSBMA as the first network showed better anti-bacterial adhesion than PAA/PSBMA-i with PAA as the first network, perhaps because of the slightly higher E of PAA/PSBMA-i. 55 In addition, the E. coli adhere more than the S. aureus to the four hydrogels. This can be explained in two ways. One is the shape of bacteria. Compared with the spherical S. aureus , the rod-shaped E. coli exhibits larger contact area on the surface of the hydrogel coatings, which makes it easier to be adhered. The second is the thickness and structure of the peptidoglycan layer of bacterial cell wall. E. coli is a kind of Gram-negative bacteria, while S. aureus is Gram-positive bacteria. Gram-negative bacteria E. coli has thinner cell wall as well as more abundant lipid and less peptidoglycan content in cell wall than Gram-positive bacteria— S. aureus . Thus, E. coli is softer than S. aureus which makes it easier to deform and adhere to sample surfaces. Fig. 5 Fluorescence microscope graphs of Escherichia coli (upper) and Staphylococcus aureus (lower) adhesion to hydrogel. (A)/(E) PAA coating; (B)/(F) PSBMA coating; (C)/(G) PAA/PSBMA-i coating; (D)/(H) PSBMA/PAA coating. Fig. 6 Quantitative results of standard plate counting assay. Protein adsorption on underwater equipment surfaces is thought to be the first step of many undesired marine biofouling, creating a conditioning film on the substrates, followed by the adhesion of cells and bacteria and colonisation by micro/macroalgae and other macro foulants. The amount of protein adsorbed on the surface is one of the most important factors in evaluating the anti-biofouling property of materials. As shown in Fig. 7A , PSBMA hydrogel showed an excellent anti-protein adsorption performance as expected. The anti-protein adsorption performance of the two bicomponent hydrogel was poorer than that of PSBMA, which was consistent with the results of anti-bacterial adhesion. The other bicomponent hydrogels were obtained by changing the type of the second network using PSBMA as the first network, such as PSBMA/PAM, PSBMA/polyhydroxyethyl methacrylate (PHEMA) and PSBMA/polydimethyl diallyl ammonium chloride (PDMDAAC), and their anti-protein adsorption results are shown in Fig. 7B . Compared with single component hydrogels, the introduction of PSBMA could play a certain effect on inhibiting protein adsorption, in which PSBMA/PAM showed excellent anti-protein adsorption performance. However, the mechanical properties of PSBMA/PAM are too poor to act as a submerged coating. Fig. 7 (A) BSA adsorption of PAA, PSBMA, PAA/PSBMA-i and PSBMA/PAA hydrogel coatings. (B) BSA adsorption of PAM, PSBMA/PAM, PSBMA/PDMDAAC, PHEMA and PSBMA/PHEMA hydrogel coatings. 3.4. Friction reduction \n Fig. 8 shows the friction behaviour of the hydrogels in water and saline solution. The friction coefficient ( μ ) of PU plate, usually as the surface component of ship, did not considerably change with sliding velocity ( v ) in water. Polydimethylsiloxane (PDMS), with low surface energy, showed a typical Stribeck behaviour, that is, high μ in low v region and then low μ in large v region. The μ ∼ v curves of PAA and PSBMA hydrogels that reached the swollen equilibrium state in deionised almost overlapped with that of PU despite their high water content and did not considerably change with v . The normal strain ( λ = P / E , P is the normal pressure) during friction for PAA and PSBMA hydrogel was approximately 8.4%. Under a very large λ , the hydrogels were forced to deform severely and led to large real contact area with the substrate and therefore increased friction. The mixed friction was also difficult to form in this case. The bicomponent hydrogel coating, PAA/PSBMA-i and PSBMA/PAA, exhibited similar friction behaviour to that of PDMS, where a weak mixed friction region could be observed. Usually, a hydrogel with high E could form mixed lubrication in relatively low v due to its large elastic coherence length. 56 Thus, high E is effective in reducing the friction at high v region. Note that all the hydrogels in water exhibited friction as high as that of PU. Fig. 8 (A) Friction coefficient curves of hydrogels in deionised water. (B) Friction coefficient curves of hydrogels in 1 M NaCl solution. (C) Friction coefficient curves of hydrogels in ASW. (D) Friction coefficient curve of hydrogels in 1 M NaCl solution. The friction of the hydrogel coatings in 1 M NaCl solution and ASW was shown in Fig. 8B and C . The friction behaviour of PU and PDMS in ASW was the same as that in pure water due to their bulk nonionic character and anhydration. However, the μ of PAA hydrogel was even higher than the μ of PU. In saline solution, PAA chain severely shrunk, and its E increased. Furthermore, PAA hydrogel immersed in saline solution showed a weak attractive interaction with glass surface. By contrast, PSBMA hydrogel coating exhibited much lower μ than that in pure water due to its swelling in saline solution coming from the zwitterionic group on the side of PSBMA chain. The smallest μ was as low as 0.003. Reasonably, PAA/PSBMA-i with SBMA as the second monomer showed lower μ than PSBMA/PAA with AA as the second monomer. For PAA/PSBMA-i hydrogel, the surface bound a large amount of water molecules to form a water layer and PSBMA existed as the second network so that the surface of the hydrogel was mostly composed of PSBMA molecules, which can reduce the friction coefficient through electrostatic hydration. The second monomer played an important role in friction behavior. To validate the surface property difference of these four hydrogels, the contact angle test in air under conditions of deionized water, NaCl solution and ASW was measured. As shown in Fig. 9 , the contact angles of water on the four hydrogels in different mediums exhibited the same trend, and PSBMA hydrogel showed a very hydrophilic surface after reaching swelling equilibrium in saline solution and became superhydrophilic, which correlated with the observations in the tribological test. Fig. 9 Water contact angle of hydrogels. By changing the crosslinking degree of the second network, we obtained hydrogels with different modulus. Even the hydrogel coatings with high E exhibited a large μ in boundary friction region, entering the mixed lubrication region in a low sliding rate and thus a low μ at a higher rate of 10 −2 to 10 −1 m s −1 could be easily obtained ( Fig. 8D ). 3.5. Adhesive and wear The adhesive state on solid substrates is important to the durability of the soft hydrogel coatings. Silane coupling agent surface treatment allowed easy and tight adherence of the soft coating on conventional solid substrate, such as glass, Al and PU plates, by binding with hydrogel through free radical polymerisation. A test has been performed to examine the robustness of the hydrogel coating. The glass with hydrogel coatings was adhered on the inner wall of a beaker in which sandy water was poured into. With a stirrer, the sandy water was stirred at a rate of 700 rpm for 1 h. Fig. 10 shows the polarizing optical micrograph of the hydrogel coatings before and after sand abrasion. PAA hydrogel coating cracked after sand abrasion and a part edge of the coating was even peeled off. No apparent cracks and splits were observed for the PSBMA and PAA/PSBMA hydrogel coatings, indicating that the surfaces were slippery and the hydrogel coatings were sufficiently tough, and the coatings could maintain the original state even after strong sand abrasion. Unexpectedly, PSBMA/PAA hydrogel coating showed a slight rupture under the microscope which was invisible under naked eye, probably because the surface of the hydrogel coating was mainly composed of the second polymer, PAA. The results of Fig. 10 are consistent with Fig. 8 , where PAA hydrogel coating exhibited large μ , whereas the others showed relatively small μ , especially in high sliding velocity. This result indicates that the more slippery the surface is, the less sand abrasion occurs. Furthermore, even under intense scour of sandy water, the hydrogel coatings could not be peeled off the glass surface, indicating the strong cohesiveness between the coating and solid surface. Fig. 10 Polarizing optical micrograph of hydrogels before and after sand abrasion with rotating rate = 700 rpm, rotating time = 1 h."
} | 7,547 |
24567077 | PMC3933906 | pmc | 1,076 | {
"abstract": "Recent empirical and theoretical works on collective behaviors based on a topological interaction are beginning to offer some explanations as for the physical reasons behind the selection of a particular number of nearest neighbors locally affecting each individual's dynamics. Recently, flocking starlings have been shown to topologically interact with a very specific number of neighbors, between six to eight, while metric-free interactions were found to govern human crowd dynamics. Here, we use network- and graph-theoretic approaches combined with a dynamical model of locally interacting self-propelled particles to study how the consensus reaching process and its dynamics are influenced by the number k of topological neighbors. Specifically, we prove exactly that, in the absence of noise, consensus is always attained with a speed to consensus strictly increasing with k . The analysis of both speed and time to consensus reveals that, irrespective of the swarm size, a value of k ~ 10 speeds up the rate of convergence to consensus to levels close to the one of the optimal all-to-all interaction signaling. Furthermore, this effect is found to be more pronounced in the presence of environmental noise.",
"discussion": "Discussion As mentioned by Vicsek & Zafeiris in their recent review 11 , very few exact results about collective motion are actually available. Most relevant to the present study is the exact formulation of the convergence to consensus in a population of autonomous agents achieved by Cucker and Smale based on their own model 38 . Note that this powerful result was established under the strong assumption of a weighted all-to-all signaling connectivity between agents. The new exact result embodied by Theorem A relies on a much weaker assumption of limited local interactions corresponding to a SSN of the k -nearest neighbor digraph type. It is worth adding that the SSN considered in our model independently switches at each time instant with the characteristic time scale τ = 1. In other words, it follows a Markovian process of order “zero”. This simple yet tractable model represents a very first step in understanding swarm dynamics from the network science standpoint. To allow for a more realistic treatment, the SSN should mathematically be modeled as a continuous-time Markovian process, which would embody the coherent evolution of the signaling network (see Supplementary Information – Continuous-time Markovian process). The multiagent dynamical systems driven by such Markovian switching networks aimed at generating consensus behaviors (see detailed in Supplementary Information ) have been extensively studied in control theory over the past few years; see e.g. 39 40 41 . One of the common restrictive assumptions in these works turns out to be the balance condition as used in Young et al. 21 . Consequently, these results do not apply here and a further thorough investigation is needed to fully understand the consensus reaching process ruled by topologically interacting neighbors under Markovian SSN. Corollary B is quite remarkable in the sense that the rate of convergence to consensus C S is shown to strictly increase with the addition of edges (by means of increasing k ) in the G ( N , k ) model. In general, when dealing with directed networks, this is however not the case—simply adding edges may not necessarily lead to faster convergence; only those edges contributing to a larger algebraic connectivity contribute to a faster convergence to consensus (see e.g. 36 ). The results associated with Fig. 1A and Fig. 1B , relating to the influence of k on the speed to consensus C S and its rate of variation with k , have a far-reaching consequence from the SSN design standpoint: adding more edges to the SSN does accelerate the convergence to consensus but this acceleration is very rapid when going from k = 3 to 10, but quickly becomes negligible when even higher values of k are being considered. Practically, adding more edges by increasing the number of topological agents with whom one is interacting is feasible but only up to a certain extent as there is always a cost associated with information exchange and also due to inherent limits in terms of signaling mechanisms, sensory and cognitive capabilities—for instance, see 42 for such biological considerations with pigeons. Therefore, when accounting for the cost of adding new edges, a trade-off value k t for the number of topologically interacting agents emerges from the competition between, on the one hand, faster consensus and higher interaction cost on the other hand. This fact is in complete agreement with the results by Young et al. 21 obtained using a fixed static sensing graph corresponding to the steady state. Our numerical analysis of the dependence of the time to consensus T C with respect to the number of topological neighbors is in complete agreement with those for the consensus speed, namely: (i) for a given value of k , increases moderately with N , (ii) the time to consensus decreases very rapidly in the range 3 ≤ k ≤ 10 for all values of N as is well illustrated by the sharp increases in in that specific range of values for k , (iii) the rate of variation of the time to consensus with respect to k , , is found to be almost independent on N similarly to what was observed for . It appears therefore that the effectiveness of the consensus reaching process is seriously impeded for the smallest values of k ~ 3, with either no convergence or a very slow one. On the other side of the spectrum, having k ~ O ( N ) most likely brings along overwhelming communication costs either for the living organisms or for the resources, with very limited gain in the consensus dynamics. Having the number of topological neighbors in the narrow interval k ∈ [8,12] not only speeds up the consensus reaching process significantly compared to the smallest possible values of k , but in addition, appears to be even more effective in the presence of noise. As mentioned earlier and further detailed in the Methods section below, in the present swarming problem, consensus means the convergence to a common state asymptotically or in a finite time among all group members through local interactions. Hence, our results only apply to this particular type of consensus and not to more sophisticated emergent behaviors such as rendez-vous in space or collision avoidance for instance 26 . From an analytical standpoint, our specific synchronization or consensus protocol is embedded in Eqs. (10) and (11). It consists of two components: (i) the actual operation performed on the local (here topological) information gathered by a given agent—a linear averaging of the relative agents' states with respect to the actual agent receiving the information. The linearity conveniently enables us to formulate the dynamical problem as in Eq. (11) and to use the powerful results of graph theory. However, nonlinear protocols could be considered as in the case of nonlinearly coupled oscillators on complex networks 43 . The linearity allows for a complete and insightful analytical investigation of the consensus reaching process; (ii) the set of local neighbors formally represented by the set of indices in Eq. (10) that defines the network connectivity. Both components are embedded into the graph Laplacian L ( t ). It is worth adding that if one considers more complex collective collectives such as human dynamics on social networks, the more intricate nature of the local interactions gives rise to social networks with properties vastly different from the SSN considered here. Standard social networks are known to be heterogeneous scale-free networks 44 while the SSN in our case is homogeneous 22 . Therefore, the methodology presented here can be applied to more complex social interactions but there is no doubt that the conclusions, somehow, have to be different from those obtained in our study. In conclusion, all the results reported above shed a completely new light on the physical reasons behind the selection of a particular number of nearest neighbors locally affecting each individual's dynamics. These results are also quantitatively consistent with the number of topological neighbors reported for flocks of starlings."
} | 2,080 |
39974892 | PMC11838272 | pmc | 1,077 | {
"abstract": "Microbial communities experience environmental fluctuations across timescales from rapid changes in moisture, temperature, or light levels to long-term seasonal or climactic variations. Understanding how microbial populations respond to these changes is critical for predicting the impact of perturbations, interventions, and climate change on communities. Since communities typically harbor tens to hundreds of distinct taxa, the response of microbial abundances to perturbations is potentially complex. However, while taxonomic diversity is high, in many communities taxa can be grouped into functional guilds of strains with similar metabolic traits. These guilds effectively reduce the complexity of the system by providing a physiologically motivated coarse-graining. Here, using a combination of simulations, theory, and experiments, we show that the response of guilds to nutrient fluctuations depends on the timescale of those fluctuations. Rapid changes in nutrient levels drive cohesive, positively correlated abundance dynamics within guilds. For slower timescales of environmental variation, members within a guild begin to compete due to similar resource preferences, driving negative correlations in abundances between members of the same guild. Our results provide a route to understanding the relationship between functional guilds and community response to changing environments, as well as an experimental approach to discovering functional guilds via designed nutrient perturbations to communities.",
"introduction": "Introduction Natural microbial communities, in contexts ranging from human hosts to soils, are buffeted by perturbations due to changes in host physiology, moisture, nutrients, pH, and temperature. Understanding how these environmental fluctuations impact community composition, interactions ( 1 ), and ultimately metabolic processes ( 2 ) is of critical importance. For example, changes in moisture, temperature, nutrients, and pH in soils impact the production of greenhouse gasses ( 3 – 8 ). However, understanding the response of these consortia to environmental fluctuations remains a challenge because they harbor hundreds of distinct taxa. In principle, the abundance of each taxon might respond to perturbations via many distinct mechanisms. For example, changes in nutrient availability can impact interactions by relieving competition, while changes in moisture or pH can alter oxygen and nutrient availability respectively ( 2 , 9 , 10 ). Due to the diversity of taxa and mechanisms that can impact their abundances, one might expect that the response of a community environmental perturbations should be complex and high-dimensional. Despite the potential complexity of community responses to perturbations, empirical and theoretical results suggest that communities are comprised of groups of taxa, or functional guilds, that perform similar functional roles. For example, in anaerobic digesters groups of taxa specialize in performing distinct steps in the fermentation cascade ( 11 , 12 ). Similarly, in marine snow-degrading communities, groups of taxa perform polysaccharide degradation, oligomer uptake, and cross-feeding of excreted metabolites ( 13 ). These functional guilds are comprised of multiple coexisting taxa with similar metabolic preferences. As a result, while communities often retain hundreds of members, in many cases, these taxa can be grouped into functional guilds, with members of a guild exhibiting similar metabolic preferences. Functional guilds are thought to emerge, at least in part, from correlations in microbial traits such as strains specializing in sugar or acid catabolism ( 14 ) but not both. Therefore, rather than dissecting how each strain responds to environmental perturbations, it might be much simpler to understand how guilds respond collectively to environmental perturbations ( 15 ). Since members of a single guild participate in similar metabolic transformations, it might be reasonable to assume guilds respond cohesively to environmental perturbations. Indeed, there is empirical evidence that functional guilds respond collectively to changing environmental conditions. For example, complex soil communities in bioreactors exhibit reproducible transitions between denitrification and dissimilatory reduction to ammonia as the carbon-to-nitrogen ratio is varied, reflecting changing dominant functional guilds in the system. This transition is thought to arise from the stoichiometric differences between denitrification and DNRA ( 16 ). Similar patterns are observed in large-scale metagenomic surveys of soil microbiomes ( 1 ). Likewise, distinct functional guilds in the cow rumen microbiome change in abundance in response to changes in lactate production during fiber fermentation ( 17 ). Theoretically, the notion of functional guilds has been described using a modular structure in the traits that members of a community possess ( 18 ). Therefore, functional guilds potentially provide a route to understanding how communities collectively respond to environmental change. Specifically, rather than dissect the response of each strain in a system to a perturbation, it might be sufficient to understand how functional guilds respond to environmental changes ( 19 , 20 ). In this sense, functional guilds might enable a more coarse-grained view of the response of communities to environmental change ( 21 ). Here we investigate this idea using simulations and experiments. Counterintuitively, we find that the response of functional guilds depends on the timescale of environmental changes. Rapid changes in nutrient levels drive correlated dynamics between guild members, with members of the same guild increasing or decreasing their abundances cohesively across the guild. In this fast fluctuation regime, the community-level response reflects the guild structure. In contrast, when environmental fluctuations occur on a slow timescale, abundance dynamics are dominated by intra-guild competition. In this regime, the abundance of members of the same guild exhibit negatively correlated dynamics due to competitive interactions, and the community-level response does not reflect the guild structure.",
"discussion": "Discussion Here, we have characterized the response of microbial communities composed of metabolic guilds to environmental fluctuations. We demonstrate that fast environmental fluctuations excite a cohesive response within metabolic guilds and that cohesion is lost for slow environmental fluctuations. The changing cohesion of metabolic guilds with timescales of fluctuations has important implications for understanding community dynamics in the wild. For example, it is routine to study correlations in abundances of microbes across time and space ( 31 ). Interpreting correlated abundance dynamics is a long-term challenge in microbial ecology ( 32 , 33 ). Here we show that the sign of these correlations can vary strongly with the timescale of environmental variation driving abundance dynamics. This means that care must be used in interpreting when abundance correlations might reflect ecological associations ( 34 ). The observation also offers an opportunity to understand metabolic guild structure in communities and to expose underlying structural properties that drive community function. Impact of timescale on collective properties This result illustrates the importance of dynamics in the collective properties of microbial communities. In this work, we define timescale in two different ways: (1) the rate of environmental fluctuation ( Fig. 2 , 3 ) and (2) time after environmental perturbation ( Fig. S2 , Fig. 4 ). In both cases, the community response is the same, cohesive on short timescales, and not cohesive on long timescales. In this context, we should expect that on short timescales the collective properties of a community (e.g. nutrient flux) should contain contributions from each organism within a functional guild. On long timescales, however, the collective properties may be dominated by a single strain that is the most effective competitor within the guild in that environment. As a result, we might expect that the collective flux of metabolites (e.g. ( 2 )) might reflect the activity of all members in a guild on short timescales and be driven by the dominant competitor within the guild on longer timescales. It remains open to test this hypothesis in microcosm experiments with controlled environmental forcing ( 28 , 35 – 37 ). Implications for experimental determination of functional guilds In addition to providing insight into natural ecological processes, this work informs experimental design. In particular, the inference of functional guilds has recently been a topic of interest in microbial ecology and microbiology ( 38 – 40 ). Identification of functional guilds in microbial communities is valuable because it can enable coarse-graining communities using effective groups of taxa ( 2 , 41 , 42 ). This coarse-graining provides a low-dimensional description of communities that can enable predictions of community dynamics and function. A common approach to inferring guilds is to perform serial dilution enrichments to identify strains that perform well in an environmental condition of interest ( 28 ). Our theoretical results suggest that the observed correlations between strains in such an experiment depend on the timescale over which those correlations are observed. Thus, ideally, one would observe correlated responses of groups over short timescales. We proposed and tested an experimental procedure for identifying functional guilds via serial-dilution experiments. One can imagine this approach being scaled up by performing short-term enrichment experiments on complex communities in diverse environments ( 43 ). In these contexts, our approach might help define metabolically cohesive units within the community. Extension to cross-feeding interactions Here, we largely restricted ourselves to consideration of resource competition. Recent work suggests that such interactions dominate in many contexts ( 44 , 45 ). However, it is the case that metabolic byproducts from one organism can be used by another, and these cross-feeding interactions are qualitatively distinct from competition ( 1 , 17 ). Intriguingly, we find that simulations with cross-feeding interactions give rise to positive inter -guild correlations for all fluctuation timescales ( Supplmentary Information ). This raises the possibility that different classes of interactions can be identified by studying correlations across timescales of environmental fluctuation. The role of cross-feeding in driving dynamics between guilds is an important avenue for future work. Interpretation and effect of private resources In this study, we provide each strain with a fluctuating “private resource” that ensures that it maintains some significant biomass ( Fig. 1A ). This ensures that the community remains relatively stable over time, which is generally the case in naturally-occuring communities ( 31 , 46 – 48 ). Ecologically, such private resources need not be interpreted literally as nutrients available only to a single guild member of the community. Instead, these resources may be viewed more abstractly as some mechanism that is orthogonal to the block resources that maintains diversity within functional guilds, such as a spatial structure or phage predation ( 49 – 52 ). For instance, the background fluctuations introduced by the private resources may arise from migration from different environments or adherence to surfaces that provide unique access to resources ( 53 , 54 ). Although it is in principle possible for block resources alone to support coexistence, identifying a community that coexists across an ensemble of fluctuating resources requires significant fine-tuning. In particular, we expect that even in silico communities chosen such that members coexist at average values of the block resources will experience extinctions in fluctuating environments. Although it is in principle possible that such fine-tuning will arise from ecological and evolutionary processes, it necessarily biases the community composition and implicitly makes an assumption about the source of community stability. For this reason, we chose to maintain coexistence using private resources, allowing us to choose an unbiased distribution of traits within functional guilds. While extensive theoretical progress has been made regarding the source of community stability and strain coexistence ( 55 – 57 ), the effect of such processes on community cohesion dynamics has not been considered. Our results suggest that this is an exciting avenue of future research, raising questions about how assumptions about mechanisms of coexistence influence the dynamics of intraguild cohesion."
} | 3,205 |
39974892 | PMC11838272 | pmc | 1,077 | {
"abstract": "Microbial communities experience environmental fluctuations across timescales from rapid changes in moisture, temperature, or light levels to long-term seasonal or climactic variations. Understanding how microbial populations respond to these changes is critical for predicting the impact of perturbations, interventions, and climate change on communities. Since communities typically harbor tens to hundreds of distinct taxa, the response of microbial abundances to perturbations is potentially complex. However, while taxonomic diversity is high, in many communities taxa can be grouped into functional guilds of strains with similar metabolic traits. These guilds effectively reduce the complexity of the system by providing a physiologically motivated coarse-graining. Here, using a combination of simulations, theory, and experiments, we show that the response of guilds to nutrient fluctuations depends on the timescale of those fluctuations. Rapid changes in nutrient levels drive cohesive, positively correlated abundance dynamics within guilds. For slower timescales of environmental variation, members within a guild begin to compete due to similar resource preferences, driving negative correlations in abundances between members of the same guild. Our results provide a route to understanding the relationship between functional guilds and community response to changing environments, as well as an experimental approach to discovering functional guilds via designed nutrient perturbations to communities.",
"introduction": "Introduction Natural microbial communities, in contexts ranging from human hosts to soils, are buffeted by perturbations due to changes in host physiology, moisture, nutrients, pH, and temperature. Understanding how these environmental fluctuations impact community composition, interactions ( 1 ), and ultimately metabolic processes ( 2 ) is of critical importance. For example, changes in moisture, temperature, nutrients, and pH in soils impact the production of greenhouse gasses ( 3 – 8 ). However, understanding the response of these consortia to environmental fluctuations remains a challenge because they harbor hundreds of distinct taxa. In principle, the abundance of each taxon might respond to perturbations via many distinct mechanisms. For example, changes in nutrient availability can impact interactions by relieving competition, while changes in moisture or pH can alter oxygen and nutrient availability respectively ( 2 , 9 , 10 ). Due to the diversity of taxa and mechanisms that can impact their abundances, one might expect that the response of a community environmental perturbations should be complex and high-dimensional. Despite the potential complexity of community responses to perturbations, empirical and theoretical results suggest that communities are comprised of groups of taxa, or functional guilds, that perform similar functional roles. For example, in anaerobic digesters groups of taxa specialize in performing distinct steps in the fermentation cascade ( 11 , 12 ). Similarly, in marine snow-degrading communities, groups of taxa perform polysaccharide degradation, oligomer uptake, and cross-feeding of excreted metabolites ( 13 ). These functional guilds are comprised of multiple coexisting taxa with similar metabolic preferences. As a result, while communities often retain hundreds of members, in many cases, these taxa can be grouped into functional guilds, with members of a guild exhibiting similar metabolic preferences. Functional guilds are thought to emerge, at least in part, from correlations in microbial traits such as strains specializing in sugar or acid catabolism ( 14 ) but not both. Therefore, rather than dissecting how each strain responds to environmental perturbations, it might be much simpler to understand how guilds respond collectively to environmental perturbations ( 15 ). Since members of a single guild participate in similar metabolic transformations, it might be reasonable to assume guilds respond cohesively to environmental perturbations. Indeed, there is empirical evidence that functional guilds respond collectively to changing environmental conditions. For example, complex soil communities in bioreactors exhibit reproducible transitions between denitrification and dissimilatory reduction to ammonia as the carbon-to-nitrogen ratio is varied, reflecting changing dominant functional guilds in the system. This transition is thought to arise from the stoichiometric differences between denitrification and DNRA ( 16 ). Similar patterns are observed in large-scale metagenomic surveys of soil microbiomes ( 1 ). Likewise, distinct functional guilds in the cow rumen microbiome change in abundance in response to changes in lactate production during fiber fermentation ( 17 ). Theoretically, the notion of functional guilds has been described using a modular structure in the traits that members of a community possess ( 18 ). Therefore, functional guilds potentially provide a route to understanding how communities collectively respond to environmental change. Specifically, rather than dissect the response of each strain in a system to a perturbation, it might be sufficient to understand how functional guilds respond to environmental changes ( 19 , 20 ). In this sense, functional guilds might enable a more coarse-grained view of the response of communities to environmental change ( 21 ). Here we investigate this idea using simulations and experiments. Counterintuitively, we find that the response of functional guilds depends on the timescale of environmental changes. Rapid changes in nutrient levels drive correlated dynamics between guild members, with members of the same guild increasing or decreasing their abundances cohesively across the guild. In this fast fluctuation regime, the community-level response reflects the guild structure. In contrast, when environmental fluctuations occur on a slow timescale, abundance dynamics are dominated by intra-guild competition. In this regime, the abundance of members of the same guild exhibit negatively correlated dynamics due to competitive interactions, and the community-level response does not reflect the guild structure.",
"discussion": "Discussion Here, we have characterized the response of microbial communities composed of metabolic guilds to environmental fluctuations. We demonstrate that fast environmental fluctuations excite a cohesive response within metabolic guilds and that cohesion is lost for slow environmental fluctuations. The changing cohesion of metabolic guilds with timescales of fluctuations has important implications for understanding community dynamics in the wild. For example, it is routine to study correlations in abundances of microbes across time and space ( 31 ). Interpreting correlated abundance dynamics is a long-term challenge in microbial ecology ( 32 , 33 ). Here we show that the sign of these correlations can vary strongly with the timescale of environmental variation driving abundance dynamics. This means that care must be used in interpreting when abundance correlations might reflect ecological associations ( 34 ). The observation also offers an opportunity to understand metabolic guild structure in communities and to expose underlying structural properties that drive community function. Impact of timescale on collective properties This result illustrates the importance of dynamics in the collective properties of microbial communities. In this work, we define timescale in two different ways: (1) the rate of environmental fluctuation ( Fig. 2 , 3 ) and (2) time after environmental perturbation ( Fig. S2 , Fig. 4 ). In both cases, the community response is the same, cohesive on short timescales, and not cohesive on long timescales. In this context, we should expect that on short timescales the collective properties of a community (e.g. nutrient flux) should contain contributions from each organism within a functional guild. On long timescales, however, the collective properties may be dominated by a single strain that is the most effective competitor within the guild in that environment. As a result, we might expect that the collective flux of metabolites (e.g. ( 2 )) might reflect the activity of all members in a guild on short timescales and be driven by the dominant competitor within the guild on longer timescales. It remains open to test this hypothesis in microcosm experiments with controlled environmental forcing ( 28 , 35 – 37 ). Implications for experimental determination of functional guilds In addition to providing insight into natural ecological processes, this work informs experimental design. In particular, the inference of functional guilds has recently been a topic of interest in microbial ecology and microbiology ( 38 – 40 ). Identification of functional guilds in microbial communities is valuable because it can enable coarse-graining communities using effective groups of taxa ( 2 , 41 , 42 ). This coarse-graining provides a low-dimensional description of communities that can enable predictions of community dynamics and function. A common approach to inferring guilds is to perform serial dilution enrichments to identify strains that perform well in an environmental condition of interest ( 28 ). Our theoretical results suggest that the observed correlations between strains in such an experiment depend on the timescale over which those correlations are observed. Thus, ideally, one would observe correlated responses of groups over short timescales. We proposed and tested an experimental procedure for identifying functional guilds via serial-dilution experiments. One can imagine this approach being scaled up by performing short-term enrichment experiments on complex communities in diverse environments ( 43 ). In these contexts, our approach might help define metabolically cohesive units within the community. Extension to cross-feeding interactions Here, we largely restricted ourselves to consideration of resource competition. Recent work suggests that such interactions dominate in many contexts ( 44 , 45 ). However, it is the case that metabolic byproducts from one organism can be used by another, and these cross-feeding interactions are qualitatively distinct from competition ( 1 , 17 ). Intriguingly, we find that simulations with cross-feeding interactions give rise to positive inter -guild correlations for all fluctuation timescales ( Supplmentary Information ). This raises the possibility that different classes of interactions can be identified by studying correlations across timescales of environmental fluctuation. The role of cross-feeding in driving dynamics between guilds is an important avenue for future work. Interpretation and effect of private resources In this study, we provide each strain with a fluctuating “private resource” that ensures that it maintains some significant biomass ( Fig. 1A ). This ensures that the community remains relatively stable over time, which is generally the case in naturally-occuring communities ( 31 , 46 – 48 ). Ecologically, such private resources need not be interpreted literally as nutrients available only to a single guild member of the community. Instead, these resources may be viewed more abstractly as some mechanism that is orthogonal to the block resources that maintains diversity within functional guilds, such as a spatial structure or phage predation ( 49 – 52 ). For instance, the background fluctuations introduced by the private resources may arise from migration from different environments or adherence to surfaces that provide unique access to resources ( 53 , 54 ). Although it is in principle possible for block resources alone to support coexistence, identifying a community that coexists across an ensemble of fluctuating resources requires significant fine-tuning. In particular, we expect that even in silico communities chosen such that members coexist at average values of the block resources will experience extinctions in fluctuating environments. Although it is in principle possible that such fine-tuning will arise from ecological and evolutionary processes, it necessarily biases the community composition and implicitly makes an assumption about the source of community stability. For this reason, we chose to maintain coexistence using private resources, allowing us to choose an unbiased distribution of traits within functional guilds. While extensive theoretical progress has been made regarding the source of community stability and strain coexistence ( 55 – 57 ), the effect of such processes on community cohesion dynamics has not been considered. Our results suggest that this is an exciting avenue of future research, raising questions about how assumptions about mechanisms of coexistence influence the dynamics of intraguild cohesion."
} | 3,205 |
39974892 | PMC11838272 | pmc | 1,078 | {
"abstract": "Microbial communities experience environmental fluctuations across timescales from rapid changes in moisture, temperature, or light levels to long-term seasonal or climactic variations. Understanding how microbial populations respond to these changes is critical for predicting the impact of perturbations, interventions, and climate change on communities. Since communities typically harbor tens to hundreds of distinct taxa, the response of microbial abundances to perturbations is potentially complex. However, while taxonomic diversity is high, in many communities taxa can be grouped into functional guilds of strains with similar metabolic traits. These guilds effectively reduce the complexity of the system by providing a physiologically motivated coarse-graining. Here, using a combination of simulations, theory, and experiments, we show that the response of guilds to nutrient fluctuations depends on the timescale of those fluctuations. Rapid changes in nutrient levels drive cohesive, positively correlated abundance dynamics within guilds. For slower timescales of environmental variation, members within a guild begin to compete due to similar resource preferences, driving negative correlations in abundances between members of the same guild. Our results provide a route to understanding the relationship between functional guilds and community response to changing environments, as well as an experimental approach to discovering functional guilds via designed nutrient perturbations to communities.",
"introduction": "Introduction Natural microbial communities, in contexts ranging from human hosts to soils, are buffeted by perturbations due to changes in host physiology, moisture, nutrients, pH, and temperature. Understanding how these environmental fluctuations impact community composition, interactions ( 1 ), and ultimately metabolic processes ( 2 ) is of critical importance. For example, changes in moisture, temperature, nutrients, and pH in soils impact the production of greenhouse gasses ( 3 – 8 ). However, understanding the response of these consortia to environmental fluctuations remains a challenge because they harbor hundreds of distinct taxa. In principle, the abundance of each taxon might respond to perturbations via many distinct mechanisms. For example, changes in nutrient availability can impact interactions by relieving competition, while changes in moisture or pH can alter oxygen and nutrient availability respectively ( 2 , 9 , 10 ). Due to the diversity of taxa and mechanisms that can impact their abundances, one might expect that the response of a community environmental perturbations should be complex and high-dimensional. Despite the potential complexity of community responses to perturbations, empirical and theoretical results suggest that communities are comprised of groups of taxa, or functional guilds, that perform similar functional roles. For example, in anaerobic digesters groups of taxa specialize in performing distinct steps in the fermentation cascade ( 11 , 12 ). Similarly, in marine snow-degrading communities, groups of taxa perform polysaccharide degradation, oligomer uptake, and cross-feeding of excreted metabolites ( 13 ). These functional guilds are comprised of multiple coexisting taxa with similar metabolic preferences. As a result, while communities often retain hundreds of members, in many cases, these taxa can be grouped into functional guilds, with members of a guild exhibiting similar metabolic preferences. Functional guilds are thought to emerge, at least in part, from correlations in microbial traits such as strains specializing in sugar or acid catabolism ( 14 ) but not both. Therefore, rather than dissecting how each strain responds to environmental perturbations, it might be much simpler to understand how guilds respond collectively to environmental perturbations ( 15 ). Since members of a single guild participate in similar metabolic transformations, it might be reasonable to assume guilds respond cohesively to environmental perturbations. Indeed, there is empirical evidence that functional guilds respond collectively to changing environmental conditions. For example, complex soil communities in bioreactors exhibit reproducible transitions between denitrification and dissimilatory reduction to ammonia as the carbon-to-nitrogen ratio is varied, reflecting changing dominant functional guilds in the system. This transition is thought to arise from the stoichiometric differences between denitrification and DNRA ( 16 ). Similar patterns are observed in large-scale metagenomic surveys of soil microbiomes ( 1 ). Likewise, distinct functional guilds in the cow rumen microbiome change in abundance in response to changes in lactate production during fiber fermentation ( 17 ). Theoretically, the notion of functional guilds has been described using a modular structure in the traits that members of a community possess ( 18 ). Therefore, functional guilds potentially provide a route to understanding how communities collectively respond to environmental change. Specifically, rather than dissect the response of each strain in a system to a perturbation, it might be sufficient to understand how functional guilds respond to environmental changes ( 19 , 20 ). In this sense, functional guilds might enable a more coarse-grained view of the response of communities to environmental change ( 21 ). Here we investigate this idea using simulations and experiments. Counterintuitively, we find that the response of functional guilds depends on the timescale of environmental changes. Rapid changes in nutrient levels drive correlated dynamics between guild members, with members of the same guild increasing or decreasing their abundances cohesively across the guild. In this fast fluctuation regime, the community-level response reflects the guild structure. In contrast, when environmental fluctuations occur on a slow timescale, abundance dynamics are dominated by intra-guild competition. In this regime, the abundance of members of the same guild exhibit negatively correlated dynamics due to competitive interactions, and the community-level response does not reflect the guild structure.",
"discussion": "Discussion Here, we have characterized the response of microbial communities composed of metabolic guilds to environmental fluctuations. We demonstrate that fast environmental fluctuations excite a cohesive response within metabolic guilds and that cohesion is lost for slow environmental fluctuations. The changing cohesion of metabolic guilds with timescales of fluctuations has important implications for understanding community dynamics in the wild. For example, it is routine to study correlations in abundances of microbes across time and space ( 31 ). Interpreting correlated abundance dynamics is a long-term challenge in microbial ecology ( 32 , 33 ). Here we show that the sign of these correlations can vary strongly with the timescale of environmental variation driving abundance dynamics. This means that care must be used in interpreting when abundance correlations might reflect ecological associations ( 34 ). The observation also offers an opportunity to understand metabolic guild structure in communities and to expose underlying structural properties that drive community function. Impact of timescale on collective properties This result illustrates the importance of dynamics in the collective properties of microbial communities. In this work, we define timescale in two different ways: (1) the rate of environmental fluctuation ( Fig. 2 , 3 ) and (2) time after environmental perturbation ( Fig. S2 , Fig. 4 ). In both cases, the community response is the same, cohesive on short timescales, and not cohesive on long timescales. In this context, we should expect that on short timescales the collective properties of a community (e.g. nutrient flux) should contain contributions from each organism within a functional guild. On long timescales, however, the collective properties may be dominated by a single strain that is the most effective competitor within the guild in that environment. As a result, we might expect that the collective flux of metabolites (e.g. ( 2 )) might reflect the activity of all members in a guild on short timescales and be driven by the dominant competitor within the guild on longer timescales. It remains open to test this hypothesis in microcosm experiments with controlled environmental forcing ( 28 , 35 – 37 ). Implications for experimental determination of functional guilds In addition to providing insight into natural ecological processes, this work informs experimental design. In particular, the inference of functional guilds has recently been a topic of interest in microbial ecology and microbiology ( 38 – 40 ). Identification of functional guilds in microbial communities is valuable because it can enable coarse-graining communities using effective groups of taxa ( 2 , 41 , 42 ). This coarse-graining provides a low-dimensional description of communities that can enable predictions of community dynamics and function. A common approach to inferring guilds is to perform serial dilution enrichments to identify strains that perform well in an environmental condition of interest ( 28 ). Our theoretical results suggest that the observed correlations between strains in such an experiment depend on the timescale over which those correlations are observed. Thus, ideally, one would observe correlated responses of groups over short timescales. We proposed and tested an experimental procedure for identifying functional guilds via serial-dilution experiments. One can imagine this approach being scaled up by performing short-term enrichment experiments on complex communities in diverse environments ( 43 ). In these contexts, our approach might help define metabolically cohesive units within the community. Extension to cross-feeding interactions Here, we largely restricted ourselves to consideration of resource competition. Recent work suggests that such interactions dominate in many contexts ( 44 , 45 ). However, it is the case that metabolic byproducts from one organism can be used by another, and these cross-feeding interactions are qualitatively distinct from competition ( 1 , 17 ). Intriguingly, we find that simulations with cross-feeding interactions give rise to positive inter -guild correlations for all fluctuation timescales ( Supplmentary Information ). This raises the possibility that different classes of interactions can be identified by studying correlations across timescales of environmental fluctuation. The role of cross-feeding in driving dynamics between guilds is an important avenue for future work. Interpretation and effect of private resources In this study, we provide each strain with a fluctuating “private resource” that ensures that it maintains some significant biomass ( Fig. 1A ). This ensures that the community remains relatively stable over time, which is generally the case in naturally-occuring communities ( 31 , 46 – 48 ). Ecologically, such private resources need not be interpreted literally as nutrients available only to a single guild member of the community. Instead, these resources may be viewed more abstractly as some mechanism that is orthogonal to the block resources that maintains diversity within functional guilds, such as a spatial structure or phage predation ( 49 – 52 ). For instance, the background fluctuations introduced by the private resources may arise from migration from different environments or adherence to surfaces that provide unique access to resources ( 53 , 54 ). Although it is in principle possible for block resources alone to support coexistence, identifying a community that coexists across an ensemble of fluctuating resources requires significant fine-tuning. In particular, we expect that even in silico communities chosen such that members coexist at average values of the block resources will experience extinctions in fluctuating environments. Although it is in principle possible that such fine-tuning will arise from ecological and evolutionary processes, it necessarily biases the community composition and implicitly makes an assumption about the source of community stability. For this reason, we chose to maintain coexistence using private resources, allowing us to choose an unbiased distribution of traits within functional guilds. While extensive theoretical progress has been made regarding the source of community stability and strain coexistence ( 55 – 57 ), the effect of such processes on community cohesion dynamics has not been considered. Our results suggest that this is an exciting avenue of future research, raising questions about how assumptions about mechanisms of coexistence influence the dynamics of intraguild cohesion."
} | 3,205 |
39974892 | PMC11838272 | pmc | 1,078 | {
"abstract": "Microbial communities experience environmental fluctuations across timescales from rapid changes in moisture, temperature, or light levels to long-term seasonal or climactic variations. Understanding how microbial populations respond to these changes is critical for predicting the impact of perturbations, interventions, and climate change on communities. Since communities typically harbor tens to hundreds of distinct taxa, the response of microbial abundances to perturbations is potentially complex. However, while taxonomic diversity is high, in many communities taxa can be grouped into functional guilds of strains with similar metabolic traits. These guilds effectively reduce the complexity of the system by providing a physiologically motivated coarse-graining. Here, using a combination of simulations, theory, and experiments, we show that the response of guilds to nutrient fluctuations depends on the timescale of those fluctuations. Rapid changes in nutrient levels drive cohesive, positively correlated abundance dynamics within guilds. For slower timescales of environmental variation, members within a guild begin to compete due to similar resource preferences, driving negative correlations in abundances between members of the same guild. Our results provide a route to understanding the relationship between functional guilds and community response to changing environments, as well as an experimental approach to discovering functional guilds via designed nutrient perturbations to communities.",
"introduction": "Introduction Natural microbial communities, in contexts ranging from human hosts to soils, are buffeted by perturbations due to changes in host physiology, moisture, nutrients, pH, and temperature. Understanding how these environmental fluctuations impact community composition, interactions ( 1 ), and ultimately metabolic processes ( 2 ) is of critical importance. For example, changes in moisture, temperature, nutrients, and pH in soils impact the production of greenhouse gasses ( 3 – 8 ). However, understanding the response of these consortia to environmental fluctuations remains a challenge because they harbor hundreds of distinct taxa. In principle, the abundance of each taxon might respond to perturbations via many distinct mechanisms. For example, changes in nutrient availability can impact interactions by relieving competition, while changes in moisture or pH can alter oxygen and nutrient availability respectively ( 2 , 9 , 10 ). Due to the diversity of taxa and mechanisms that can impact their abundances, one might expect that the response of a community environmental perturbations should be complex and high-dimensional. Despite the potential complexity of community responses to perturbations, empirical and theoretical results suggest that communities are comprised of groups of taxa, or functional guilds, that perform similar functional roles. For example, in anaerobic digesters groups of taxa specialize in performing distinct steps in the fermentation cascade ( 11 , 12 ). Similarly, in marine snow-degrading communities, groups of taxa perform polysaccharide degradation, oligomer uptake, and cross-feeding of excreted metabolites ( 13 ). These functional guilds are comprised of multiple coexisting taxa with similar metabolic preferences. As a result, while communities often retain hundreds of members, in many cases, these taxa can be grouped into functional guilds, with members of a guild exhibiting similar metabolic preferences. Functional guilds are thought to emerge, at least in part, from correlations in microbial traits such as strains specializing in sugar or acid catabolism ( 14 ) but not both. Therefore, rather than dissecting how each strain responds to environmental perturbations, it might be much simpler to understand how guilds respond collectively to environmental perturbations ( 15 ). Since members of a single guild participate in similar metabolic transformations, it might be reasonable to assume guilds respond cohesively to environmental perturbations. Indeed, there is empirical evidence that functional guilds respond collectively to changing environmental conditions. For example, complex soil communities in bioreactors exhibit reproducible transitions between denitrification and dissimilatory reduction to ammonia as the carbon-to-nitrogen ratio is varied, reflecting changing dominant functional guilds in the system. This transition is thought to arise from the stoichiometric differences between denitrification and DNRA ( 16 ). Similar patterns are observed in large-scale metagenomic surveys of soil microbiomes ( 1 ). Likewise, distinct functional guilds in the cow rumen microbiome change in abundance in response to changes in lactate production during fiber fermentation ( 17 ). Theoretically, the notion of functional guilds has been described using a modular structure in the traits that members of a community possess ( 18 ). Therefore, functional guilds potentially provide a route to understanding how communities collectively respond to environmental change. Specifically, rather than dissect the response of each strain in a system to a perturbation, it might be sufficient to understand how functional guilds respond to environmental changes ( 19 , 20 ). In this sense, functional guilds might enable a more coarse-grained view of the response of communities to environmental change ( 21 ). Here we investigate this idea using simulations and experiments. Counterintuitively, we find that the response of functional guilds depends on the timescale of environmental changes. Rapid changes in nutrient levels drive correlated dynamics between guild members, with members of the same guild increasing or decreasing their abundances cohesively across the guild. In this fast fluctuation regime, the community-level response reflects the guild structure. In contrast, when environmental fluctuations occur on a slow timescale, abundance dynamics are dominated by intra-guild competition. In this regime, the abundance of members of the same guild exhibit negatively correlated dynamics due to competitive interactions, and the community-level response does not reflect the guild structure.",
"discussion": "Discussion Here, we have characterized the response of microbial communities composed of metabolic guilds to environmental fluctuations. We demonstrate that fast environmental fluctuations excite a cohesive response within metabolic guilds and that cohesion is lost for slow environmental fluctuations. The changing cohesion of metabolic guilds with timescales of fluctuations has important implications for understanding community dynamics in the wild. For example, it is routine to study correlations in abundances of microbes across time and space ( 31 ). Interpreting correlated abundance dynamics is a long-term challenge in microbial ecology ( 32 , 33 ). Here we show that the sign of these correlations can vary strongly with the timescale of environmental variation driving abundance dynamics. This means that care must be used in interpreting when abundance correlations might reflect ecological associations ( 34 ). The observation also offers an opportunity to understand metabolic guild structure in communities and to expose underlying structural properties that drive community function. Impact of timescale on collective properties This result illustrates the importance of dynamics in the collective properties of microbial communities. In this work, we define timescale in two different ways: (1) the rate of environmental fluctuation ( Fig. 2 , 3 ) and (2) time after environmental perturbation ( Fig. S2 , Fig. 4 ). In both cases, the community response is the same, cohesive on short timescales, and not cohesive on long timescales. In this context, we should expect that on short timescales the collective properties of a community (e.g. nutrient flux) should contain contributions from each organism within a functional guild. On long timescales, however, the collective properties may be dominated by a single strain that is the most effective competitor within the guild in that environment. As a result, we might expect that the collective flux of metabolites (e.g. ( 2 )) might reflect the activity of all members in a guild on short timescales and be driven by the dominant competitor within the guild on longer timescales. It remains open to test this hypothesis in microcosm experiments with controlled environmental forcing ( 28 , 35 – 37 ). Implications for experimental determination of functional guilds In addition to providing insight into natural ecological processes, this work informs experimental design. In particular, the inference of functional guilds has recently been a topic of interest in microbial ecology and microbiology ( 38 – 40 ). Identification of functional guilds in microbial communities is valuable because it can enable coarse-graining communities using effective groups of taxa ( 2 , 41 , 42 ). This coarse-graining provides a low-dimensional description of communities that can enable predictions of community dynamics and function. A common approach to inferring guilds is to perform serial dilution enrichments to identify strains that perform well in an environmental condition of interest ( 28 ). Our theoretical results suggest that the observed correlations between strains in such an experiment depend on the timescale over which those correlations are observed. Thus, ideally, one would observe correlated responses of groups over short timescales. We proposed and tested an experimental procedure for identifying functional guilds via serial-dilution experiments. One can imagine this approach being scaled up by performing short-term enrichment experiments on complex communities in diverse environments ( 43 ). In these contexts, our approach might help define metabolically cohesive units within the community. Extension to cross-feeding interactions Here, we largely restricted ourselves to consideration of resource competition. Recent work suggests that such interactions dominate in many contexts ( 44 , 45 ). However, it is the case that metabolic byproducts from one organism can be used by another, and these cross-feeding interactions are qualitatively distinct from competition ( 1 , 17 ). Intriguingly, we find that simulations with cross-feeding interactions give rise to positive inter -guild correlations for all fluctuation timescales ( Supplmentary Information ). This raises the possibility that different classes of interactions can be identified by studying correlations across timescales of environmental fluctuation. The role of cross-feeding in driving dynamics between guilds is an important avenue for future work. Interpretation and effect of private resources In this study, we provide each strain with a fluctuating “private resource” that ensures that it maintains some significant biomass ( Fig. 1A ). This ensures that the community remains relatively stable over time, which is generally the case in naturally-occuring communities ( 31 , 46 – 48 ). Ecologically, such private resources need not be interpreted literally as nutrients available only to a single guild member of the community. Instead, these resources may be viewed more abstractly as some mechanism that is orthogonal to the block resources that maintains diversity within functional guilds, such as a spatial structure or phage predation ( 49 – 52 ). For instance, the background fluctuations introduced by the private resources may arise from migration from different environments or adherence to surfaces that provide unique access to resources ( 53 , 54 ). Although it is in principle possible for block resources alone to support coexistence, identifying a community that coexists across an ensemble of fluctuating resources requires significant fine-tuning. In particular, we expect that even in silico communities chosen such that members coexist at average values of the block resources will experience extinctions in fluctuating environments. Although it is in principle possible that such fine-tuning will arise from ecological and evolutionary processes, it necessarily biases the community composition and implicitly makes an assumption about the source of community stability. For this reason, we chose to maintain coexistence using private resources, allowing us to choose an unbiased distribution of traits within functional guilds. While extensive theoretical progress has been made regarding the source of community stability and strain coexistence ( 55 – 57 ), the effect of such processes on community cohesion dynamics has not been considered. Our results suggest that this is an exciting avenue of future research, raising questions about how assumptions about mechanisms of coexistence influence the dynamics of intraguild cohesion."
} | 3,205 |
31923325 | null | s2 | 1,079 | {
"abstract": "The mutualistic symbiosis between forest trees and ectomycorrhizal fungi (EMF) is among the most ubiquitous and successful interactions in terrestrial ecosystems. Specific species of EMF are known to colonize specific tree species, benefitting from their carbon source, and in turn, improving their access to soil water and nutrients. EMF also form extensive mycelial networks that can link multiple root-tips of different trees. Yet the number of tree species connected by such mycelial networks, and the traffic of material across them, are just now under study. Recently we reported substantial belowground carbon transfer between Picea, Pinus, Larix and Fagus trees in a mature forest. Here, we analyze the EMF community of these same individual trees and identify the most likely taxa responsible for the observed carbon transfer. Among the nearly 1,200 EMF root-tips examined, 50%-70% belong to operational taxonomic units (OTUs) that were associated with three or four tree host species, and 90% of all OTUs were associated with at least two tree species. Sporocarp "
} | 268 |
36559798 | PMC9788077 | pmc | 1,080 | {
"abstract": "Superhydrophobic materials have recently attracted great interest from both academia and industry due to their promising applications in self-cleaning, oil–water separation, etc. Here, we developed a facile method to prepare hybrid PDMS/TiO 2 fiber for superhydrophobic coatings. TiO 2 could be uniformly distributed into PDMS, forming a hierarchical micro/nano structure on the surface of the substrate. The contact angle of the superhydrophobic coating could reach as high as 155°. The superhydrophobic coating possessed good self-cleaning performance, corrosion resistance, and durability. It was found that gravity-driven oil–water separation was achieved using stainless steel mesh coated with the PDMS/TiO 2 coating. More importantly, the coated filter paper could not only separate oil and pure water but also corrosive solutions, including the salt, acid, and alkali solution.",
"conclusion": "4. Conclusions (1) The TiO 2 nanofibers prepared by the hydrothermal method are uniform in size. The TiO 2 nanofibers are added to the PDMS to form a superhydrophobic coating, which is sprayed on the stainless steel mesh to obtain good superhydrophobic performance. At the same time, the superhydrophobic coating has excellent corrosion resistance and can still protect the stainless steel mesh after 720 h of immersion; (2) When the superhydrophobic SSM is used as oil–water separation material, the separation efficiency of different oil–water mixtures is above 94%, and the separation efficiency is basically unchanged in acid, alkali, and salt environments; (3) After 50 times of continuous separation, the separation efficiency is still above 92%. After 40 cycles of polishing, the coating still has a high water contact angle, indicating that the superhydrophobic coating has good wear durability; (4) The PDMS/TiO 2 composite coating has excellent self-cleaning performance.",
"introduction": "1. Introduction With the development of the economy and industries and the improvement of people’s living standards, environmental issues are increasingly of concern. Especially the oily wastewater produced by industrial emissions, accident leakage, domestic waste, etc., not only pollutes water but also threatens human life. Furthermore, the discharge of oils also leads to a great loss of energy and resources. Thereby removal of oil from wastewater is challenging work for the conservation of ecosystems. Various cleanup methods based on bioremediation, chemical treatment, and mechanical recovery have been applied to the separation of oil–water mixtures [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ]. The design and preparation of superhydrophobic surfaces with water contact angles >150° have garnered widespread attention in terms of basic research and industrial applications [ 9 , 10 , 11 , 12 ]. Nature gives good examples, such as the lotus leaf with self-cleaning properties due to a bumpy waxy surface with hierarchical micro/nanostructures. High surface roughness and low surface free energy are necessary prerequisites for obtaining stable superhydrophobic characteristics (water contact angle of more than 150° and sliding angle of less than 10°). Metal mesh shows great potential due to its high thermal and mechanical stability. In some studies, nanoparticles are endowed with superhydrophobic properties on the mesh surface by etching and deposition [ 13 , 14 ], but such methods will significantly reduce the mechanical properties of the substrate under certain harsh environments, thus limiting their application. In other studies, nanoparticles are combined with polymers, and polymer resins are used as binders to prepare superhydrophobic and superhydrophilic coatings so as to bind nanoparticles to the mesh surface [ 15 , 16 , 17 , 18 ]. Cao [ 16 ] spray-modified a PU/SiO 2 composite coating to make the grid superhydrophobic to hot water. In another study, PMMA was used to improve the strength between carbon nanotubes and steel mesh surfaces [ 17 ]. In recent years, the fabrication of superhydrophobic surfaces based on semi-conductor oxide materials such as ZnO [ 18 ], CuO [ 19 ], and TiO 2 [ 20 ] has been conducted. TiO 2 nanostructures have become a focus of tremendous interest due to their unique physicochemical properties and good chemical stability related to applications including photocatalysis [ 21 ], anti-ultraviolet protection [ 22 ], and chemical sensors [ 23 ]. Nanosized TiO 2 powder is now commercially available and can be used as a raw material for the preparation of superhydrophobic surfaces. Bolvardi et al. [ 24 ] reported commercialized TiO 2 nanoparticles with different sizes forming a coating consistent with PDMS. Qing et al. [ 25 ] fabricated superhydrophobic surfaces of TiO 2 /PDMS by dip coating, and the as-prepared superhydrophobic surface had an excellent self-cleaning ability. Compared with the spherical structure, the top size of the cluster-like structure was larger, and the size of the bottom was smaller. Capillary pressure, namely, the required pressure yielded by the air–water interfacial area to resist water into the micro–nano structure gaps, decreased with increases in the nanocrystal spacing. The cluster-like TiO 2 nanofiber spacing was smaller than the TiO 2 spherical structure. Therefore, when undergoing oil–water separation recycling, TiO 2 nanofibers had better separation performance [ 26 ]. In this paper, TiO 2 nanofibers prepared by the hydrothermal method were combined with PDMS to prepare a composite coating, which was sprayed on the surface of stainless steel mesh to form a fiber cluster. Moreover, the effects of mechanical friction, the pH of the mixture, and continuous oil–water separation performance of superhydrophobic stainless steel mesh were studied. This method can rapidly and effectively prepare oil–water separation materials on a large scale and can be used in harsh environments.",
"discussion": "3. Results and Discussion 3.1. Morphology and Composition of TiO 2 Nanofibers Figure 2 shows the TEM morphology and XRD diagram of TiO 2 . It can be seen from Figure 2 a,b that the TiO 2 powder was prepared into smooth and uniform nanofibers with a diameter of 50–80 nm and a length of 400 nm–600 nm. Figure 2 c is the XRD diagram before and after the modification of TiO 2 . The diffraction peaks of the samples were matched with the standard card of TiO 2 (JCPDS Card No. 21-1272), which indicates that the TiO 2 nanofibers were mainly the anatase type. The diffraction peaks at 25.5°, 37.7°, 47.9°, 53.7°, 55.1°, and 62.8°, respectively, corresponded to the (101), (004), (200), (105), (211), and (204) crystal planes of anatase TiO 2 crystals [ 27 , 28 ]. 3.2. PDMS/TiO 2 -Coated Stainless Steel Mesh SEM analysis was carried out to examine the surface morphology. The change in morphology after coating is shown in Figure 3 . As seen in Figure 3 a, the uncoated stainless steel mesh had a relatively smooth surface. After coating with the PDMS/TiO 2 composite coating ( Figure 3 c,d), the stainless steel wires were covered uniformly and completely by the coating, and a relatively rough surface was observed, while the structure of the stainless steel mesh was well preserved. The diameter of the uncoated stainless steel wire was calculated to be 26.9 ± 2.4 μm, while the diameter of the coated stainless steel wire was 31.3 ± 4.7 μm. The thickness of the composite coating was then calculated to be about 2.2 ± 1.2 μm. The composite coating covered completely the stainless steel wires, and almost no coating was observed in the pores of the coated stainless steel mesh, which guaranteed the free flow of the water–oil mixture. 3.3. Contact Angle Measurement The surface wettability of the coated SSM was characterized by water contact angle, as shown in Figure 4 . The contact angle of the pretreated SSM was almost 0 ( Figure 4 a), and the water droplets rapidly penetrated. The composite coating gave the SSM superhydrophobic propertidx, and the contact angle reached 155° ± 0.5° ( Figure 4 b). Figure 4 c,d show the influence of the pH on contact angle. The contact angle decreases slightly in the range of pH value from 1 to 14, but is greater than 150°. The water droplets with pH of 1, 7, and 14 on the superhydrophobic coating are perfectly spherical. When different liquids were dropped onto the coated SSM, the red n-hexane droplets rapidly permeated, leaving only the red oily imprint. Other liquids showed the morphology of spherical droplets on the surface (see the photograph of Figure 4 e), demonstrating that the materials treated with composite coating have good hydrophobic property to different aqueous solutions. 3.4. Corrosion Resistance Test It is necessary to consider the anti-corrosion property of stainless steel mesh during use. Electrochemical impedance spectroscopy (EIS) was used to study the corrosion resistance of the SSM. Figure 5 a shows the Nyquist plots of the pretreated stainless steel electrode and the coated stainless steel electrode after soaking in 3.5% NaCl solution for 1 h. In the initial immersion, the aggressive ions unevenly penetrated into the coating and did not reach the coating/steel interface; both electrodes had a capacitance arc, and the electrode after coating had a larger capacitance arc than the blank electrode, indicating that the coated stainless steel electrode had higher corrosion resistance. As time progressed, the corrosive ions in the coating became homogeneous. There was an electrocoupling layer capacitor on the corrosive interface, which formed a parallel circuit with the interface resistance. At this stage, the coating had not been completely destroyed and it could still exhibit anticorrosion performance. After 30 days of immersion ( Figure 5 b), it was found that the capacitive resistance was reduced, but it was still far greater than that of the blank stainless steel electrode, indicating that the prepared composite coating has a long-term protective effect on stainless steel. With the prolongation of the immersion time, the accumulation of corrosion products at the interface between the coating and mild steel resulted in the appearance of Warburg diffusion resistance. Warburg impedance appeared in the low-frequency region, indicating that the diffusion of ions appeared in the coating, and the protective effect of the coating was weakened. The results showed that the surface of the composite coating had good anti-corrosion performance and could be used in corrosive environments. 3.5. Oil–Water Separation Performance In order to study the oil–water separation performance of the coated SSM, several oils such as chloroform, n-hexane, vegetable oil, and mineral oil were used for separation tests. As shown in Figure 6 , the separation efficiency of the coated SSM for different oils was above 94%. The separation efficiency of n-hexane (94.8%) and chloroform (94.1%) was slightly lower than that of other oils due to the high volatility; in addition, viscosity was another factor affecting the separation efficiency. Therefore, vegetable oil (96.3%) remained on the inner wall and the mesh of the separation device, resulting in a low separation efficiency [ 29 ]. When the water phase was changed to aqueous solutions with different pH values, the separation efficiency basically did not change; thus, it was shown that the coated SSM could effectively separate different oil–water mixtures. Table 1 shows a comparison of the separation between the TiO 2 /PDMS-coated SSM and other separation materials reported in the references. 3.6. Durability Measurement One of the most important aspects of oil–water separating meshes is their durability against harsh mechanical conditions. In order to investigate the durability of the coated SSM, a wear resistance test was carried out. As shown in Figure 7 a, after 40 cycles of sanding, the surface still maintained its superhydrophobicity, which verified its excellent wear resistance and durability. The reusability performance of the coated SSM was studied according to the separation efficiency of the n-hexane/water mixture, as shown in Figure 7 b. It was clear that even after 30 times of separation, the superhydrophobic SSM still showed excellent separation efficiency of more than 94% and a contact angle of 150.4°. However, with the increase of the number of separation cycles, the separation efficiency decreased, but it was still above 92%. By observing the separation device, it was found that the main reason for the decrease in efficiency was that a small amount of n-hexane volatilized, which led to the reduction of the collected n-hexane and of the separation efficiency. The results showed that the superhydrophobic SSM had stable oil–water separation efficiency and good reusability. 3.7. Self-Cleaning Performance The self-cleaning performance of the composite coating was measured by cleaning soil. A small amount of soil was sprinkled on the coated glass slide, which was placed in a petri dish to form a small inclination, as shown in Figure 8 a. The water dripped onto the tested glass slide through the syringe. During the water drop falling test, the soil was wrapped and quickly removed by the falling water drops so as to obtain a clean surface, as shown in Figure 8 b,c. This is because the prepared superhydrophobic coating had a very small sliding angle. Under the impact pressure and injection direction of the water droplets, it was easy for the water droplets roll down and carry away the pollutants on the surface. Even if the pollutants were wrapped, the spherical form of the liquid water droplets could be maintained. Here, the sliding angle of the coated glass is small, and accordingly the soil on the coated glass is easy to be removed as the water drops roll down, as shown in Figure 8 d. The reason why keeping the superhydrophobic coatings stable is important is that the powerful water jet could be damaging to the coating of a surface as much as the former could make the latter’s wettability decline [ 37 , 38 ]. According to this fact, this study revealed how to use a water jet impact as a way to test its ability to keep hybrid microsphere coatings stable. With an angle of incidence and a high speed, a blue colored water jet, about 3 cm above the sample surface, was syringed toward the superhydrophobic coating surface. Reflected in a direct way at an angle of about 20° from the contact point, Figure 9 a revealed that the water jet didnot spread on the surface, neither did the superhydrophobic coating change. The coating could keep the appearance original as much as 10 min after the water jet test. What’s more, as it could be observed from the image of SEM in Figure 9 b, it was the hierarchical micro/nano structure of the coating that remained almost intact. The CA keep 150° unchanged even after the test. Thus, we could conclude that the superhydrophobic coating composed of the TiO 2 fiber held high stability toward the water jet."
} | 3,734 |
39161452 | PMC11332185 | pmc | 1,081 | {
"abstract": "Self-healing polymers are extensively researched for the sustainability of materials. The introduction of dynamic networks instead of traditional cross-linkers for an autonomous healing mechanism in elastomers is a promising strategy for improving rubber properties. However, exchangeable covalent bonds in a dynamic network generally rely on external stimulants and fillers, which can compromise the material's performance. Herein, we introduce a mechanically strong yet resilient and independent self-healing polymeric network by dual cross-linking of bonds based on covalent and non-covalent dual interaction. The thiourea-based polymer polyether thiourea ethylene glycol (PTUEG) was blended with natural rubber (NR) and epoxidized natural rubber (ENR) to strengthen the mechanical characteristics of the material NR-ENR-PTUEG 3 . In the material, the thermoplastic polymer PTUEG 3 applied the thiourea linkage as a hydrogen bonding and dynamic covalent motif together to enhance mechanical adaptability in a self-healing polymer network exhibiting stiffness, toughness, and resilience, thereby extending its longevity. The resulting mechanical characteristics of the NR-ENR-PTUEG 3 with 25 phr PTUEG 3 exhibited tensile stress 4.8 ± 0.3 MPa and high elongation at break 833 ± 0.1%, demonstrating far better performance than that of pristine NR, and 85% recovery of its original strength at ambient temperature. The healing behaviour is strongly influenced by thiourea-based polymer contents, enabling autonomous self-healing at ambient temperature, exhibiting in situ load-bearing efficiency in the repaired material, and maintaining their mechanical characteristics.",
"conclusion": "4. Conclusions In this study, a thiourea-based rubber material, NR-ENR-PTUEG 3 , was successfully synthesized through a dynamic covalent network and hydrogen bond interaction, enabling self-healing without external stimulants, additives, or fillers. Pristine NR has poor self-healing and mechanical properties and required additives to vulcanize the rubber and enhance its strength. However, this compromises its healing efficiency due to irreversible crosslinking after vulcanization. Therefore, thiourea was formulated to cross-link the NR-ENR network, curing the elastomer and improving its mechanical strength and inherent self-healing efficiency. The blending exhibited autonomous self-healing in rubber with good mechanical strength due to the presence of a dual network mechanism. Spectra findings revealed that the incorporation of PTUEG 3 served as a multifunctional linkage, facilitating the formation of covalent bonds between the C and S groups in the epoxy and thiourea, as well as non-covalent bonds between the H and O groups in the thiourea and oxygenous epoxy group, respectively. This formed a reversible network that increased toughness and durability by establishing a hybrid rubber network. While the TGA results indicate that the blended material exhibits good thermal stability, it is important to note that the specific contribution of molecular motion to this stability is not fully understood. Given the lack of detailed research on molecular motion, these conclusions should be considered preliminary. Further studies are needed to elucidate the molecular dynamics that contribute to the observed thermal behavior. The NR-ENR-PTUEG 3 at 15 phr exhibited good tensile strength (4.8 ± 0.3 MPa) before healing and tensile strength (3.92 ± 0.6 MPa) after healing without any external stimulation or additives at room temperature. However, it was observed that the network also exhibited limited efficiency at 25 phr PTUEG 3 loading, slightly reducing elongation at break and healing efficiency. On the other hand, toughness and mechanical stress increases with the increase in PTUEG 3 loading up to 25 phr. The blended rubber material demonstrated the ability to heal when broken surfaces were in contact at room temperature, regaining mechanical strength and exhibiting healing efficiency of up to 85% without any triggering agent. This approach accelerated the self-healing process, eliminated the dependency of traditional additives and external stimulation such as heat, light, radiation, etc. and its inherent self-healing efficiency shows promising outcomes in advanced material applications where safety, performance, and longer fatigue life are major concerns.",
"introduction": "1. Introduction Natural rubber (NR) stands out as a significant and extensively used bio-based elastomer due to its mechanical strength and resilience properties. In its uncured condition, it is a biodegradable resource, making it renewable. 1–3 However, to achieve desired mechanical, thermal, and chemical resistance properties like other rubbers, it typically undergoes vulcanization, additive cross-linking, or chemical processing. This treatment makes it suitable for extensive industrial applications such as tires, tubing, sealing joints, shock-absorbing layers, and insulation. 4–8 The traditional vulcanization process typically involves the use of fillers such as carbon black, silica, sulphur or peroxide, and similar materials as curing agents to establish a permanent, irreversible cross-linking network, resulting in a mechanically strong material. 9 However, they pose challenges such as filler dispersion, energy consumption, and environmental pollution due to the formation of an irreversible network. 10,11 Additionally, these products are susceptible to damage from various causes, including mechanical stress, chemical exposure, heat, light, radiation, or a combination of these factors. 12,13 Such damage can lead to the formation of micro-cracks or deep fractures within the material, making detection and external intervention by conventional methods challenging or impossible. The internal damage not only reduces the material's performance but also serves as a catalyst for further catastrophes like macro-cracks, moisture swelling, and de-bonding. 14 Traditional repair methods such as welding, gluing, or patching can extend the material's lifespan but often compromise its mechanical strength in the repaired region. 15 Addressing this need, Cordier has initiated the development of self-healing rubber using a dynamic supramolecular self-assembly approach involving multiple hydrogen bonds in the network. 16 These relatively weak bonds collectively form a network at room temperature, allowing the material to autonomously repair damage without requiring external stimulation. This innovative approach opens up new possibilities for the development of widespread applications of self-healing materials. However, the healing process can only operate once the capsules are broken, and the mechanism is no longer present to participate in the same region. Subsequently, various methods have been developed to achieve self-healing properties in rubber for instance by introducing dynamic covalent bonds, such as TEMPO-oxidized cellulose nanocrystals and dicarboxylic acid, into the rubber network. Dynamic covalent cross-linking involves establishing exchangeable covalent bonds in a dynamic network that may break and reconstruct under specific circumstances. This enables rubber to repair damage under certain stimulations, resulting in excellent self-healing efficiency. 17 These modifications allow rubber to efficiently repair damage under stimulation, thereby significantly enhancing its durability and demonstrating its potential for high-temperature environments. However, while these approaches demonstrated good efficacy with external triggers such as heat, 17 ultraviolet radiation, 18 light, 19,20 current, 21 and catalysts 22 to initiate the self-healing process, these stimuli often compromised mechanical strength and restricted the mobility of molecular chains. This limitation hinders mechanical performance and limits potential applications. 15,23 Thermoplastics materials are often weak in toughness and resilience, while having strong mechanical stiffness and strength. Although elastomers are very resilient, their stiffness and strength are constrained therefore utilizing the dynamic covalent bond with a non-covalent network in materials aids in addressing the challenge of growing demand for sustainable rubber-based materials that exhibit resilient like elastomers and mechanically strong like thermoplastics. It also eliminates the dependency on external stimulants and additives or fillers for a wide range of desirable rubber-based materials. In a recent study, Yanagisawa reported a series of poly (ether thiourea)s incorporating ethylene glycol (TUEG) as a spacer. 24 This novel material exhibits remarkable self-healing properties and robustness without relying on external stimulation. Notably, it possesses low glass transition temperatures ( T g ) and sufficient chain segment mobility, enabling macroscopic flow that aids the self-healing process. Consequently, damaged regions rapidly reconnect through the reshuffling of chemical interactions, leading to partial or complete material recovery. Presently, TUEG has been found to improve the performance of numerous material properties, including ion-conducting, shape memory, antibacterial, Li batteries, and solar cells, endowing them with both mechanically robust and self-healing properties. 25–32 However, polymeric binders lacking three-dimensional cross-linking have the potential to dissolve due to their low cross-linking density. Introducing PTUEG 3 into the rubber network may form a dual network containing hydrogen bonds and dynamic covalent networks, and improve significant properties of the rubber. However, to our knowledge, there have been no reports of high-performance self-healing rubber materials obtained by blending PTUEG3 with rubber. A material blended with NR, epoxidized natural rubber (ENR), and PTUEG 3 was prepared. The thiourea moieties of PTUEG 3 in the material may form dense non-crystalline hydrogen bonding interactions and undergo dynamic covalent bond exchange with the epoxy moieties of ENR. 33 This interaction will facilitate the bond exchange between thermosets and thermoplastics that improve the mechanical performance and allow the recovery of the material. The autonomous recovery under ambient conditions eliminates the need for external stimuli and conventional vulcanizing agents or additives. Additionally, it helps to improve mechanical properties, such as high elasticity, resilience, and damping characteristics, making it more convenient for various application scenarios.",
"discussion": "3. Result and discussion 3.1. Preparation of the rubber NR-ENR-PTUEG 3 PTUEG 3 can be blended with various polymers to establish a dual crosslinked polymer network based on hydrogen bonding and dynamic covalent crosslinking, enabling materials to achieve self-healing properties. Blending PTUEG 3 with NR and ENR may also have a similar promoting effect. Initially curing of material tested from 120 °C to 180 °C varying temperatures and times, and at 160 °C for 10 minutes (including 7 min under compression) provided the best balance of the material performance, including tensile strength, elasticity, and hardness, while preventing over-curing or thermal degradation. Further the effect of PTUEG 3 concentration was assessed through analysis of swelling, mechanical properties, and self-healing efficiency. The light brown NR-ENR-PTUEG 3 obtained by blending NR, ENR, and PTUEG 3 has a rubber based texture ( Fig. 1a and b ). The thiourea moieties in NR-ENR-PTUEG 3 help activating the epoxy ring, leading to nucleophilic attack and playing an important role in self-healing ( Fig. 1c ). In addition, the hydrogen on the nitrogen in the PTUEG 3 structure can form hydrogen bonds with sulfur, oxygen, etc. , providing additional attraction and further stabilizing the physical and chemical properties of the material. These interconnected networks will promote significant self-healing and mechanical strength within the rubber material structure. Fig. 1 Mechanism of self-healing rubber NR-ENR-PTUEG 3 . (a) Chemical structure of the component used in the synthesis of NR-ENR-PTUEG 3 . (b) Self-healing mechanism of prepared NR-ENR-PTUEG 3 . (c) Chemical reaction between epoxy and thiourea motif. 3.2. Characterisation of the rubber NR-ENR-PTUEG 3 To confirm the potential structure in the NR-ENR-PTUEG 3 , the precursor blending of pristine PTUEG 3 , NR and ENR were analysed by FT-IR using the ATR model ( Fig. 2a ). In the NR-ENR-PTUEG3 spectra, the epoxy ring undergoes a ring-opening reaction with the amine, forming a three-dimensional polymer network. Unsaturated hydrocarbons, such as alkenes have C–H bonds that absorb at higher frequencies, usually above 3000 cm −1 . The absence of C–H stretching above 3000 cm −1 indicates the absence of unsaturated carbon, while the presence of three peaks suggests the presence of methyl and methylene groups. Specifically, these peaks are assigned as follows: the CH 3 asymmetric stretch at 2963 cm −1 , the CH 2 asymmetric stretch at 2917 cm −1 , and the CH 3 symmetric stretch at 2853 cm −1 . Additionally, characteristic absorption peaks of ENR are observed at 1663, 1447, 1377, and 840 cm −1 , corresponding to the cis -1,4-polyisoprene structure. 35 The peak at 1663 cm −1 represents C \n \n\n<svg xmlns=\"http://www.w3.org/2000/svg\" version=\"1.0\" width=\"13.200000pt\" height=\"16.000000pt\" viewBox=\"0 0 13.200000 16.000000\" preserveAspectRatio=\"xMidYMid meet\"><metadata>\nCreated by potrace 1.16, written by Peter Selinger 2001-2019\n</metadata><g transform=\"translate(1.000000,15.000000) scale(0.017500,-0.017500)\" fill=\"currentColor\" stroke=\"none\"><path d=\"M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z\"/></g></svg>\n\n C stretching, while the peak at 1377 cm −1 corresponds to the methyl group umbrella mode. The intense peak at 840 cm −1 signifies the C–H wagging vibration of an alkene group, which diminished further in the blended material due to crosslinking between hydrogen and the oxygenous group of ENR and the H-bond group of thiourea. 36 A small peak is observed at 1017 cm −1 , which shifts towards 1020 cm −1 in the NR-ENR-PTUEG 3 structure ( Fig. 2a ). This shift is attributed to the involvement of the 1133 cm −1 peak in the ENR stretching C–O–C from the epoxy ring, indicating the formation of the β-hydroxyl ring-opening reaction. The opening of the epoxy ring is important indication that reveals the cross-linking mechanism between PTUEG 3 and NR-ENR. Furthermore, the ring-opening reaction promotes the possibility of hydrogen bonding between ENR and PTUEG 3 molecules, allowing more active sites for further dynamic cross-link network, which is favourable for the self-healing of the blended material. Fig. 2 Characterization of NR-ENR-PTUEG 3 . (a) FTIR spectra of NR, ENR, PTUEG 3 and NR-ENR-PTUEG 3 . (b) Toluene swelling for crosslinking density. (c) TGA curves, and (d) DTG curves. In PTUEG 3 , the broad vibrational bands appearing around 3289 and 2920 cm −1 are attributed to –NH bending. The symmetric and asymmetric C S stretching vibrations at 940 cm −1 , characteristic of thiourea, are observed to shift to a lower frequency of 933 cm −1 in the NR-ENR-PTUEG 3 . The former is diagnostic of NH deformation vibration of nonlinearly H-bonded thiourea units, while the latter is characteristic of NH stretching vibration. 37 This shift suggests that densely located thiourea units tightly cross-link the polymer main chains via H-bonding interactions without inducing crystallization in the blended material, as the resultant H-bonded arrays are nonlinear and less ordered. 24 Additionally, peaks observed around 1552 cm −1 , 1093 cm −1 , and 822 cm −1 are attributed to –CN stretching and –CS groups present in PTUEG 3 . Consequently, PTUEG 3 with multiple reactive groups on its surface served as a multifunctional linkage, facilitating the formation of covalent bonding between the C and S group present in thiourea and epoxy groups, thus forming a hybrid rubber network. The presence of multiple reactive groups in PTUEG 3 enhances its interfacial compatibility with the NR-ENR matrix, as the β-hydroxyl ester linkages regulate bond exchange reactions and activate copolymerization. These findings provide evidence supporting the confirmation of PTUEG 3 grafting with ENR and NR. Furthermore, the crosslinking equilibrium swelling experiment provided confirmation of the crosslinking status of the NR-ENR-PTUEG 3 containing NR, ENR, and PTUEG 3 . The total crosslinking density of the vulcanized rubber sample was determined by swelling experiment using a toluene solution. As anticipated, the control NR (pristine) completely dissolved within 12 h after immersed in toluene for 48 h due to a lack of structural linkage holding the rubber chain together. 34 On the other hand, the materials prepared with ENR and PTUEG 3 exhibited swelling, indicating the formation of cross-links to varying degrees. The degree of crosslinking depends on the concentration of the epoxy and thiourea group formulations present in the NR network. Therefore, various formulations of NR-ENR-PTUEG3 with different compositions were investigated ( Table 1 ). The NR-ENR-PTUEG 3 absorbed less toluene compared to the NR-ENR and exhibited reduced swelling. The crosslinking density analysis results ( Fig. 2b ) demonstrates that the extent of crosslinking density increases with the presence of ENR and PTUEG 3 in the formulations. The dynamic covalent linkage and epoxy group inhibited chain mobility upon contact with toluene, resulting in decreased toluene absorption, reduced swelling volume, and crosslinking density significantly increased. Compared to pristine NR, NR-ENR and NR-ENR-PTUEG 3 form a more cross-linked network. The TGA curves of the blended rubber material NR-ENR-PTUEG 3 indicate a single-stage degradation process characterized by well-defined initial and final degradation temperatures. Within the temperature range of 100–270 °C, there is no discernible difference in the percentage mass loss. The initial degradation temperatures for NR and NR-ENR are noted at 300 and 310 °C, respectively ( Fig. 2c ). Subsequently, the maximum initial degradation temperature of NR-ENR-PTUEG 3 is notably observed at 370 °C. Although there is no disparity in the percentage mass loss within the temperature range of 370–420 °C in all the material, the incorporation of PTUEG 3 into the rubber impacted its thermal property. In contrast to the NR-ENR-PTUEG 3 , the NR and NR-ENR blend initiate degradation earlier at high temperatures, primarily due to the presence of C C bonds in the backbone. Despite PTUEG 3 being thermally less stable than NR-ENR, blending it with the NR-ENR slightly improved the thermal stability of PTUEG 3 due to changes in its chain mobility. It is reasonable to assume that the presence of a soft molecular segment can also contribute to improving the thermal stability of the material. Thus, thiourea emerged as our choice for executing our concept, given its recognition as a soft molecular chain suitable for NR vulcanization due to the presence of sulphur atoms and low glass transition temperature that assist sufficient chain segment mobility of thiourea that enable macroscopic flow to assist the inherent self-healing process in rubber. The DTG curves of the NR-ENR network improve the thermal stability comparatively to pristine NR shows a maximum weight loss temperature at 389 °C, while the incorporation of PTUEG 3 into rubber makes maximum weight loss towards higher temperature at 391 °C ( Fig. 2d ). The improved thermal stability in ENR-NR network may due to the formation of hydrogen bonding strengthen the interaction between polymeric chain in the network. These findings indicate that while the incorporation of PTUEG 3 into rubber moderately affects their thermal stability, it also helps to improve the overall thermal stability of the rubber. In the rubber network, the chemical interaction between NR-ENR and PTUEG 3 involves thiourea S C NH 2 carbonyl groups in thermoplastic polymers linking with the thermosets of ENR, thereby opening the epoxy ring along with the rubber chains. This prepared thermoplastic polymer serves as a synthetic intermediate for polymers, and forming a cross-linked network. 38–40 Consequently, PTUEG 3 , with its reactive carbonyl sites, acts as a multifunctional linkage and contributes to the formation of a dual network mechanism. The combined action of covalent bonding and hydrogen bonding results in a blended structure that reconfigures the network topology. This configuration endows the material with the ability to rapidly fill rubber cracks at room temperature without requiring thermally activation. Additionally, it enhances its rheological, mechanical, morphological, and maturing characteristics. 3.3. Mechanical analysis of NR-ENR-PTUEG 3 Mechanical characteristics obtained through tensile testing provide significant insights into the effects of the prepared materials at different formulations. Fig. 3a–d illustrate the mechanical performance of NR-ENR and various NR-ENR-PTUEG 3 formulations compared to NR, which typically exhibits low tensile strength and high elastic modulus, as is typical for elastomeric materials. In the comparative study, pristine NR exhibited a stress of 0.19 MPa at 336% elongation at break, notably poor performance compared to the NR-ENR blend, which exhibited a stress of 2.27 MPa at 470% elongation at break. This improvement maybe due to the chain entanglement resulting from the non-covalent network, as demonstrated in the swelling analysis. On the other hand, the blended rubber network shows potential as a multi-functional cross-linking network, facilitated by the presence of multiple reactive sites on PTUEG 3 . These sites contribute to the formation of a tough and resilient network. Consequently, an increase in PTUEG 3 concentration in the network led to improvements in maximum stress, toughness, and elongation at break. Fig. 3 Typical stress–strain curves of NR, NR-ENR and NR-ENR-PTUEG 3 with different concentrations of PTUEG 3 (a) maximum stress, (b) elongation at break, (c) elastic modulus, (d) toughness, and (e) comparative study between this work and reported self-healing rubber materials. Detailed information is summarized in Table S1. † Significant results observed for the NR-ENR-PTUEG 3 at various formulations of 5 phr, 15 phr, and 25 phr of PTUEG 3 demonstrated that tensile strength increases with higher concentrations, reaching 3.91, 4.59, and 4.83 MPa, respectively. Corresponding, elongations at break were 600%, 934%, and 833%, showing significant improvements compared to NR-ENR, attributed to the enhanced crosslinking with the thiourea group to form dynamic networks. The combination of NR-ENR reinforced with PTUEG 3 is noted for simultaneously enhancing strength and elongation. The presence of short ENR chains linking PTUEG 3 together provides additional stress resistance during stretching, while longer NR chains contribute to the elasticity of the material. The network formed by hydrogen bonding in NR is considerably weaker compared to the robust sulfur covalent crosslink network in PTUEG 3 . When the material is subjected to external stretching, the weaker hydrogen bonding network is the first to be disrupted. This preferential breakage of hydrogen bonds allows the material to absorb and dissipate energy efficiently, preventing immediate failure. As a result, this mechanism helps to enhance the overall mechanical properties of NR-ENR- PTUEG 3 , such as its toughness and resilience. The energy dissipation through the breaking of hydrogen bonds allows the sulfur covalent crosslink network to maintain its integrity, contributing to the durability and elasticity of the material under stress. 41 This hybrid rubber network, coupled with the covalent network, significantly improves the strength of the NR-ENR-PTUEG 3 . The maximum improvements in tensile strength and toughness were observed at higher concentration of PTUEG 3 , indicating increased crosslinking within the thiourea network. In contrast, the decrease in elastic modulus observed in each blended material with increasing concentration of PTUEG 3 and ENR compared to pristine NR network exhibiting their elastomeric characteristics. This dual reaction mechanism, involving both covalent and non-covalent networks, significantly contributes to the mechanical properties of NR-ENR-PTUEG 3 , which exhibit superior mechanical performance compared to the pristine NR material. Moreover, mechanical strength plays a crucial role in the self-healing properties of the blended material. The optimum rubber demonstrates enhanced tensile strength and high strain percentage compared to previously reported works ( Fig. 3e ). Comparative analysis with other reported studies also highlights typical tensile strength, elongation at break, and self-healing efficiency over time (Table S1 † ). 3.4. Self-healing property of NR-ENR-PTUEG 3 To demonstrate the self-healing ability of the blended rubber, pristine NR, NR-ENR, and various formulations of NR-ENR-PTUEG 3 were initially cut in the middle using a clean knife. The damaged samples were immediately re-joined, and manual pressure was applied for 30 seconds. The specimens were then left at room temperature for 12 h to self-heal. Tensile testing was conducted to evaluate the mechanical characteristics of the repaired material ( Fig. 4a ). The results from the tensile test showed that the healed samples preserved the mechanical characteristics of the original samples, mostly recovering compared to the pristine NR and NR-ENR material. Formulations with 5 phr, 15 phr, and 25 phr of PTUEG 3 exhibited significant strengths of 1.07, 3.92, and 3.36 MPa, respectively, after 12 h. The corresponding elongations at break were 325%, 724%, and 599%, respectively, due to the further bond exchange reaction with the thiourea group forming dynamic networks. In comparison, NR and NR-ENR exhibited strength of 0.07 and 0.26 MPa, with corresponding elongation at break of 818% and 325%, respectively, indicating poor recovery compared to the blended rubber material. Although some healing effects were observed in NR and NR-ENR blend, indicating the broken hydrogen bonds cannot completely recover in a short time, thus fewer weak bonds participate and exhibited the contribution of polar epoxy groups that limited in the chain diffusion of NR-ENR, resulting in poor self-healing. The interaction of epoxy groups with thiourea in the healing process of the NR-ENR-PTUEG 3 was highly effective. For a more intuitive comparison, the stress–strain curve of the healed NR-ENR-PTUEG 3 with 15 phr was recorded from 0 h to 12 h at 3 h intervals to observe the significant time for maximum self-healing ( Fig. 4b ). The healing efficiency initially exhibited 5%, 20%, 36%, and 85% MPa at 0 h, 3 h, 6 h, and 12 h, respectively, compared to its original strength. This indicates that the healing efficiency and crosslinking network improved with prolonged healing time, aligning with findings in self-healing materials with different healing times. 42 Fig. 4 Self-healing property of NR-ENR-PTUEG 3 . (a) Typical stress–strain test of the self-healing rubber. (b) Self-healing time study of NR-ENR-PTUEG 3 . (c) Strength recovery of the repaired material. (d) The tensile test performance of the self-healing NR-ENR-PTUEG 3 . (e) Image of the NR-ENR-PTUEG 3 material performing in situ load-lifting test, and the different colors of the samples are due to varying lighting conditions at the time of photography. (f) Optical microscopy image showing the healing of the repaired material. The initial 5% MPa at 0 h indicates relatively inferior self-healing due to weak crosslinking with the thiourea and hydrogen networks. However, the substantial improvement to 85% MPa at 12 h demonstrated extensively improved self-healing at room temperature. This indicates that the healing efficiency and crosslinking network improved with prolonged healing time, aligning with findings in self-healing materials with different healing times. 42 And the reformation of the hydrogen-bond network plays a primary role in the recovery process, as these bonds can quickly re-establish themselves, contributing to the initial stages of healing. In contrast, the dynamic covalent bonds take a longer time to reform due to their more complex nature, which involves the gradual reorganization of molecular structures. This disparity in the recovery times of different bond types results in a slower overall elasticity recovery of the rubbers at room temperature. The slower reformation of covalent bonds suggests that while initial strength and integrity are regained relatively quickly, full mechanical restoration and elasticity take longer to achieve. 43 The self-healing efficiency was calculated using eqn (2) , comparing the stress and strain curves before and after 12 h of healing at room temperature ( Fig. 4c ). As a result, the self-healing NR-ENR-PTUEG 3 exhibited excellent recovery (27%, 85%, and 69%) at PTUEG 3 concentration of 5 phr, 15 phr, and 25 phr in the rubber. However, a tensile test indicated that 15 phr PTUEG 3 loading resulted in improved elongation at break and healing efficiency, while an increase in PTUEG 3 loading to 25 phr led to a decline in strain and healing efficiency. Nonetheless, more detailed studies are needed to understand the mechanism of the self-healing process. To further examine the mechanical strength of the repaired material, in situ load-bearing tests were conducted at three different loads (250 g, 500 g, and 750 g) in pristine NR. These tests exhibited an inability to support an initial weight of 250 g, breaking from the repaired region. Similarly, NR-ENR was unable to support a load of 500 g from the attach region; however, it withstood the load test at 250 g for a few seconds due to the chain entanglement of the rubber. In contrast, the repaired NR-ENR-PTUEG 3 after 12 h significantly sustained a 750 g load without breaking from the damaged region for several hours ( Fig. 4e ), demonstrating the presence of a dual healing network based on dynamic covalent and hydrogen bonds, which enhances the self-healing ability of NR-ENR-PTUEG 3 . In this network, the PTUEG 3 chains maintain high mobility at room temperature, resulting in network rearrangement. Overall, the NR-ENR-PTUEG 3 exhibited excellent self-healing capabilities, with mechanical strength attributed to the dynamic covalent and non-covalent networks compared to pristine natural rubber. The healing process was monitored at intervals of every 3 h using optical microscopy to visualize the self-healing lines of the blended material on the surface ( Fig. 4f ). In both pristine NR and NR-ENR, a sharp crack line between the cuts is visibly apparent from the top surface within the cut and re-joined regions. Interestingly, at room temperature, the healed NR-ENR-PTUEG 3 gradually fades the damaged region over a span of 12 h, affirming the ability to self-heal without any external stimulation and showcasing the excellent stiffness of the material after being healed. Consequently, this clearly illustrated the significant role of the thiourea network in the rubber network."
} | 7,856 |
26953605 | PMC4989324 | pmc | 1,083 | {
"abstract": "Microbes are well-recognized members of the coral holobiont. However, little is known about the short-term dynamics of mucus-associated microbial communities under natural conditions and after disturbances, and how these dynamics relate to the host's health. Here we examined the natural variability of prokaryotic communities (based on 16S ribosomal RNA gene amplicon sequencing) associating with the surface mucus layer (SML) of Porites astreoides, a species exhibiting cyclical mucus aging and shedding. Shifts in the prokaryotic community composition during mucus aging led to the prevalence of opportunistic and potentially pathogenic bacteria ( Verrucomicrobiaceae and Vibrionaceae ) in aged mucus and to a twofold increase in prokaryotic abundance. After the release of aged mucus sheets, the community reverted to its original state, dominated by Endozoicimonaceae and Oxalobacteraceae . Furthermore, we followed the fate of the coral holobiont upon depletion of its natural mucus microbiome through antibiotics treatment. After re-introduction to the reef, healthy-looking microbe-depleted corals started exhibiting clear signs of bleaching and necrosis. Recovery versus mortality of the P. astreoides holobiont was related to the degree of change in abundance distribution of the mucus microbiome. We conclude that the natural prokaryotic community inhabiting the coral SML contributes to coral health and that cyclical mucus shedding has a key role in coral microbiome dynamics.",
"introduction": "Introduction Corals live in a well-described mutualism with photoautotrophic endosymbiotic dinoflagellates of the genus Symbiodinium , frequently referred to as zooxanthellae (reviewed by Muscatine, 1990 ). More recently, the concept of the coral holobiont ( Rohwer et al. , 2002 ) has been proposed to describe the association of the coral host with a diverse microbial community including representatives of fungi, endolithic algae, bacteria and archaea, of which certain associations are species-specific ( Ritchie and Smith, 1997 ; Rohwer et al. , 2002 ; Koren and Rosenberg, 2006 ; Carlos et al. , 2013 ). The sum of these microorganisms and their combined genetic material forms the coral microbiome, whose core composition is determined by the host but whose presence allows for metabolic adaptations to local environmental conditions by selection of beneficial genes ( Kelly et al. , 2014 ; Ainsworth et al. , 2015 ). The putative functions of the coral microbiome comprise, among others, the protection against pathogens ( Rohwer et al. , 2002 ; Shnit-Orland and Kushmaro, 2009 ) and the supply and cycling of nutrients ( Lesser et al. , 2004 ; Wegley et al. , 2007 ). Gates and Ainsworth (2011) propose that all taxonomic components should be considered as important because of their potential to stimulate the holobiont's functioning. However, the extent of the microbiome's contribution to the health of the coral holobiont and to coral reef resilience remains largely unknown. Whereas the association between the coral and its eukaryotic symbionts of the genus Symbiodinium is fairly well characterized ( Rowan, 2004 ; Frade et al. , 2008 ), relatively little is known about the spatial and temporal variation of the coral's prokaryotic microbiome. Recent studies have shown shifts in the microbiome from healthy to diseased corals and under altered environmental conditions ( Bourne et al. , 2007 ; Mao-Jones et al. , 2010 ), supporting the idea that both the disturbance of the native microbiota and the direct infection by specific pathogens threaten the well-being of corals. The coral host provides several distinct habitats for its microbial inhabitants—the tissue ( Lesser et al. , 2004 ), the gastrovascular cavity ( Herndl and Velimirov, 1985 ), the skeleton ( Sharshar et al. , 1997 ) and the surface mucus layer (SML) ( Rohwer et al. , 2002 ; Kooperman et al. , 2007 ), each harboring a distinct microbial community ( Sweet et al. , 2011a ). The SML is particularly important for the biology of corals not only as a habitat for a distinctive suite of coral-associated microbes but also due to its nutritional, protective and cleansing roles ( Brown and Bythell, 2005 ). Consisting of polymeric glycoproteins and lipids ( Bythell and Wild, 2011 ), coral mucus provides a nutritious medium on which a diverse assemblage of microbes thrives, many of which are highly host specific ( Rohwer et al. , 2002 ). Although coral mucus is in constant contact with the adjacent seawater, their prevalent microbial communities exhibit almost no overlap ( Rohwer et al. , 2001 , 2002 ; Frias-Lopez et al. , 2002 ). Furthermore, it has been hypothesized that the microbial community in the coral SML operates as a defense barrier and therefore protects the coral against invasive microbes either because of the production of antimicrobial substances or simply because of the occupation of this interface niche ( Rohwer et al. , 2002 ; Reshef et al. , 2006 ; Rosenberg et al. , 2007 ; Shnit-Orland and Kushmaro, 2009 ). The SML is a very dynamic system whereby its molecular organization and composition vary between coral species and over time ( Brown and Bythell, 2005 ). In addition, the SML experiences sloughing, either continuously or periodically, from the coral surface into the reef environment ( Bythell and Wild, 2011 ). This cycle of structural changes creates habitat dynamics to which microbial communities are likely responding ( Nelson et al. , 2013 ). Corals of the genus Porites sp. provide an ideal model system to study natural short-term dynamics of the mucus microbiota because of a clearly recognizable and well-described aging process that precedes periodical sloughing of the entire mucus layer and its reformation ( Coffroth, 1991 ). The SML initially appears as a transparent surface film whose visual appearance slowly changes over the course of a few days into a conspicuous aged mucus sheet. After the release of this aged sheet of mucus into the water column, new fluidic mucus is produced at the surface of the coral leading to a new cycle, suggested to follow a lunar periodicity ( Coffroth, 1991 ). Our rationale is that the temporal transformation and periodical release of mucus from the surface of poritid corals may trigger or relate to the establishment of a microbial succession analogous to the one shown for bacterioplankton after phytoplankton blooms ( Teeling et al. , 2012 ). To better understand the role of the mucus microbiota on the health and resilience of coral holobionts, we apply an indicator species approach aiming (1) to unravel the natural short-term dynamics of mucus-associated prokaryotic communities coupled to the cyclical aging and release of SML of Porites astreoides and (2) to follow the successional steps taking place after disruption of the coral's microbiome.",
"discussion": "Discussion Natural dynamics of coral mucus-associated prokaryotes Although the cyclic aging and shedding of the SML in colonies of Porites sp. is fairly well documented (see Coffroth, 1991 and Supplementary Information ), the associated dynamics of its mucus-dwelling prokaryotic community remained to be resolved. We demonstrate that the prokaryotic community undergoes significant changes throughout the mucus aging cycle, both in terms of cell abundance and community composition. Generally, the SML of P. astreoides was dominated by the bacterial families Oxalobacteraceae and Endozoicimonaceae , the latter of which has commonly been found in healthy corals ( Apprill et al. , 2013 ; Lema et al. , 2014 ; Meyer et al. , 2014 ) and suggested to have coevolved with specific host species ( Bayer et al. , 2013 ). In contrast, aged mucus sheets exhibited a high relative abundance of Verrucomicrobiaceae , Flammeovirgaceae, Rhodobacteraceae and Vibrionaceae (see Figure 5 and Supplementary Figure S5 ). The latter two bacterial families include coral pathogens and are commonly associated with coral diseases ( Ben-Haim et al. , 2003 ; Sunagawa et al. , 2010 ). After the detachment of the aged mucus sheet, however, the microbial community reverted to its original composition within 3–5 days (see Figure 2 ), supporting the idea that periodical mucus shedding in poritid corals generates a natural, rather deterministic fluctuation of the mucus-dwelling prokaryotic community, taking place within a temporal scale of weeks. As the prokaryotic indicator assemblages associated with aged mucus sheets showed surprisingly high similarity to the community associated with mucus of disturbed coral colonies ( Figures 4 and 5 ), some of which died off, we propose that periodic mucus shedding ( Coffroth, 1991 ) is an important mechanism supporting coral health by periodically removing undesirable prokaryotes from the surface of the colony leading to the maintenance of a beneficial mucus microbiome. Once coral mucus is detached from the colony, it functions as particle trap before sinking to the seafloor, where it acts as energy and nutrient source for benthic organisms in coral reefs ( Wild et al. , 2004 ; Naumann et al. , 2009 ). Alongside, aged mucus sheets frequently harbor sediment particles ( Figure 1a ), suggesting that prokaryotes colonizing upper sediment layers in coral reefs could ‘hitch a ride' to the nutrient-rich mucus layer via sediment resuspension. Thus, sediments may serve as a ‘seed-bank' for coral mucus-associated microbes as proposed by Carlos et al. (2013) . Both, the resuspension of sediment particles (and its associated microbes) onto the coral's surface and the rapid sedimentation of detached coral mucus would lead to a high connectivity between the microbial communities of aged mucus and coral reef sediments, and consequently, may explain the similarity in their prokaryotic community composition ( Figure 5 ). In contrast, microbial communities of the ambient seawater exhibited only minor overlap with those found in mucus. This confirms earlier findings ( Rohwer et al. , 2001 ; Frias-Lopez et al. , 2002 ) and stresses the importance of the SML as a selective medium for the microbial pool in the adjacent seawater. Prokaryotic mucus re-colonization after antibiotics disturbance Antibiotic treatment led to a significant reduction of prokaryotic abundance in the SML and concurrent changes in community composition. Although the full extent of the influence of antibiotics on the holobiont's fitness remains elusive, no visual cues of a negative impact on coral health were noted. This confirms previous reports of minimal effect of antibiotic treatment on zooxanthellae photosynthetic efficiency and host tissue protein content ( Gilbert et al. , 2012 ). These findings suggest that coral hosts are not strictly dependent on their mucus-associated prokaryotic symbionts, at least for such short periods of up to 8 days and in the absence (or deactivation) of pathogens. Control colonies kept in the aquaria without the addition of antibiotics exhibited obvious signs of bleaching and necrosis and underwent a significant shift in their mucus prokaryotic community ( Figure 4 ). The bacterial family Endozoicimonaceae , which showed the highest relative abundances in newly produced SMLs and is associated with mucus of healthy P. astreoides colonies ( Meyer et al. , 2014 ), showed a significant decrease in its relative abundance within the control group. Concomitant with this decrease, many bacterial groups such as Rhodobacteraceae , Verrucomicrobiaceae , Colwelliaceae , Oceanospirillaceae and Flavobacteriaceae increased in their (relative and absolute) abundance (see Figure 3 and Supplementary Figure S8). These findings are consistent with a previous study attributing the expression of lesions in P. astreoides colonies to the loss of Endozoicimonaceae and the proliferation of an opportunistic bacterial community ( Meyer et al. , 2014 ). Compositional shifts in the microbial community associated with the SML have been observed under stressful environmental conditions ( Thurber et al. , 2008 ). Based on these findings, the observed health deterioration of the untreated control group may have been caused by aquaria conditions leading to a destabilization of the natural mucus community. This interpretation is in agreement with the hypothesis that disturbances in the dynamic equilibrium of the coral's native microbiota may result in health deterioration ( Lesser et al. , 2007 ). Both, antibiotic-treated and control colonies, once brought back to their natural habitat, exhibited rapid changes in their mucus-associated prokaryotic community ( Figure 4 and Supplementary Figure S8 ). For the microbe-depleted (treated) colonies, the increase in Vibrionaceae was dominated by Vibrio sp., a genus harboring well-known coral pathogens such as Vibrio shilonii ( Kushmaro et al. , 1997 ) and Vibrio corallilyticus ( Ben-Haim et al. , 2003 ; Garren et al. , 2014 ). The latter is reported to use coral-produced sulfur compounds as a cue to target stressed corals ( Garren et al. , 2014 ). Curiously, a recent study has shown that P. astreoides (among other coral species) produces copious amounts of the organic sulfur compound dimethylsulfoniopropionate, particularly upon stress ( Frade et al. , 2015 ). Furthermore, Verrucomicrobiaceae, Flammeovirgaceae and Rhodobacteraceae , other families associated with necrotic and diseased P. astreoides colonies, have been found to be overrepresented in poritid corals suffering from White Band Disease ( Séré et al. , 2013 ; Roder et al. , 2014 ). Although non-treated colonies slowly recovered from tissue lesions and regained a mucus community very similar to their original community, antibiotic-treated colonies suffered from mortality (from day 3 onward) and exhibited increased relative abundances of bacterial families described as early colonizers of marine biofilms, such as Rhodobacteraceae and Oceanospirillacae ( O'Toole et al. , 2000 ; Sweet et al. , 2011b ). Although we cannot exclude other synergistic effects on the host's health, these results suggest that microbe-depleted SML provides an open niche, which gets rapidly colonized by opportunistic bacteria. Thus, we hypothesize that the re-establishment of both coral health and the native prokaryotic community after disturbance of the P. astreoid es holobiont depends on the initial degree of disruption of the microbiome. Total recovery of control colonies under in situ environmental conditions in contrast to mortality of microbe-depleted colonies suggests that, upon disturbance, the remnant prokaryotic community in the mucus/tissue may act as ‘seed-bank'. The mucus microbiome and its influence on coral health and survival Host-associated bacteria form unique microbiomes highly adapted to a particular host niche ( Ainsworth et al. , 2015 ). Although being generally considered to comprise commensals, microbiomes can fulfill important biological needs of their hosts, for example, immune development or nutrient acquisition ( Round and Mazmanian, 2009 ; Shin et al. , 2011 ). We identified Endozoicimonaceae and Oxalobacteraceae as significant indicators for the mucus microbiome of healthy P. astreoides colonies ( Figure 5 ). The reduction in the relative abundance of these microbiome members observed before coral necrosis and bleaching suggests that the loss of beneficial bacteria can result in a serious health threat for the holobiont, often associated with an increase in opportunistic and potentially pathogenic bacteria ( Meyer et al. , 2014 ). Although our study does not aim at determining the metabolic function of particular microbiome members, it reveals that an intact mucus microbiome may function as a barrier against potentially harmful bacteria. This defense barrier could be based on commensal-like prokaryotes, which prevent harmful colonization by successfully outcompeting pathogens ( Reid et al. , 2001 ) and/or depend on stimulating host immune response via the recognition of commensal-derived signals such as microbial-associated molecular patterns ( Mackey and McFall, 2006 ). Our results suggest that the mucus microbiome acts as defense barrier against pathogenic microbes, therefore facilitating homeostasis and contributing to the survival of the coral holobiont. In summary, we demonstrate that the previously documented periodical mucus aging and shedding cycle in P. astreoides is provoking predictable shifts in the mucus microbiome, leading to changes between a beneficial community and a potentially opportunistic/pathogenic one. The periodical release of mucus seems to be part of a life strategy that supports the maintenance of a beneficial mucus microbiome and the resilience of coral health in shallow water habitats characterized by frequent sediment resuspension. However, severe disruption of the natural microbial community upon external stress could negatively and irreversibly affect the fate of the coral holobiont. Further investigations on the functional capacities of the mucus prokaryotic community are warranted to better understand the role of the mucus microbiome in the dynamic equilibrium of the coral holobiont. Finally, we have shown that specific bacterial members can be used as indicators of coral microbiome disruption, paving the way to the development of early diagnostic tools to monitor the health status of corals."
} | 4,365 |
27554786 | PMC5009624 | pmc | 1,084 | {
"abstract": "Leaf vascular patterns are the mechanisms and mechanical support for the transportation of fluidics for photosynthesis and leaf development properties. Vascular hierarchical networks in leaves have far-reaching functions in optimal transport efficiency of functional fluidics. Embedding leaf morphogenesis as a resistor network is significant in the optimization of a translucent thermally functional material. This will enable regulation through pressure equalization by diminishing flow pressure variation. This paper investigates nature’s vasculature networks that exhibit hierarchical branching scaling applied to microfluidics. To enable optimum potential for pressure drop regulation by algorithm design. This code analysis of circuit conduit optimization for transport fluidic flow resistance is validated against CFD simulation, within a closed loop network. The paper will propose this self-optimization, characterization by resistance seeking targeting to determine a microfluidic network as a resistor. To advance a thermally function material as a switchable IR absorber.",
"conclusion": "Conclusion of Methods This iterative procedure will determine optimization of vasculature employing resistance seeking targeting by hierarchical channel succession, as a resistor networks. Hence this optimization process is determined by selection of a known hydraulic resistor value R 0 , to enable evaluation that determines vascular channel network geometry, to give flow equalization as a development mechanism as denoted by: The pressure drop across the outermost longitudinal channel is given by We can then determine the required resistances of the other longitudinal succession channels R 1 , R 2 , R 3 etc that can be determined as a recursive pattern: Feed in principle upstream and down stream channels, manifold resistance, Rm 1 , Rm 2 , Rm 3 is determined by N side channels ( N denoting the total number of channels is 2 N + 1): This computation design methodology of the device predicted pressure drops results in the manifold are in good agreement with CFD results. Results indicate resistances of the longitudinal microchannels are similar to the theoretical results. The tapered sections resistances are in good agreement, except for the tapered section involving the inlet port. This is expected since the computation solution cannot take into account the flow expanding away from the circular inlet port. Computational analysis conclusively demonstrates being able to design microfluidic networks using a theoretical approach, to achieve optimization of circuit resistance of transport fluidic flow. Optimization is achieved through pressure equalization by diminishing flow pressure variation. The two-step algorithm approach enables self-organized channel resistance with its own independency for optimum potential for pressure drop regulation. Methodology of successive conduit sequences of hierarchical formations perform regulatory roles. This is functionally significant in the analysis of hydraulic resistance to compute simulations of flow rates in microvasculature channel networks. Predicted pressure drop and flow analysis within channel network is in agreement with the analytical solution for fully-developed laminar flows, giving validity to the algorithm code as a iterative procedure. The analytical results using theoretical resistance are based on R0. The weakness of this approach is represented by R1 and R0, as CFD simulations already indicate that flow rates through R4, R3 and R2 are almost identical. If the network is constrained to device 148 mm width, multi microchannel sequence succession could start from the outermost channel and work inwards for R3, R2, R1, and R0. The attraction of this is the footprint of the network stays constrained to a 148 mm width with the outermost edge 3 mm microchannel. This method is a reverse analysis as the outermost channel width is unknown to begin with and so the width of the device cannot be known in advance. The optimized resistance condition could be based on theoretical manifold resistance. If design channel widths uses Rm manifold resistance obtained from CFD simulation to estimate the desired flow resistance in longitudinal channels, as given by: Flow rate through the manifold is denoted by Qm . This would lead to greater optimized microfluidic design, by code modification to obtain uniform flow across the polymer pane. The morphogenesis of leaf vasculature sets an underlying process of flow distribution, pressure, fluidic transport and resistance. This notion of precise hydrodynamic control of microfluidic’s will progresses thermal material characterization; to advance a polymer device, into a switchable IR absorber. This is achieved by hierarchical succession of branching sequence patterns that conform to rules of minimum effective power flow rates in the transportation of fluidics within fractal networks. That is determined by computational theoretical analysis of resistance seeking targeting.",
"discussion": "Results and Discussion Vein formations of primary (stem), secondary (mid, parallel, polar–circulative boundary vein) and minor veins (tertiary for localized fluidic flow) deal with specific leaf material regions 8 . This is functionality significant for material characterization, as it represents vein conduit sequences succession 9 10 . This working formation hierarchy is in response to tolerance to damage, water stress conditions and redundancy. The polar vein completes the network of nested conduit loops to maintain fluidic flow from stem and mid vein vasculature. All veins diminish in size distally from fluidic input supply. This arrangement of diminishing vein size order by primary veins and secondary is the relationship to leaf apex scale 11 . Vasculature patterns are linked to material scale in the formation of conduit network geometry as they perform regulatory roles. The fluidic input and export flow within these hierarchical networks are subject to flow resistance and flow rate. Hydraulic resistance in fluidic conduits channels conform to minimum fluidic flow to achieved reduced pressure drop for fluidic flow efficiency. This is determined by hierarchical structure to minimize resistance R for optimal fluidic transport. The resistance is determined by mechanical energy when a flowing liquid is subjected to a change in direction. The parameter coefficient of this resistance is: Pressure drop between inlet and export stem vasculature Q Fluidic Flow R Resistance Vasculature as a Resistor Optimal transport efficiency in natural fluidic pattern formations can be defined as a resistor. This is flow resistance evaluation in determining channel conduit scaling of vasculature branching networks. Channels that are distally positioned from fluidic input are affected by pressure drop in fully developed laminar flow. Veins will be subjected to increasing resistance or rather pressure drop for any given flow rate. To evaluate this hydrodynamic question essentially rules by Hagen-Poiseuille’s law, which suggests a constant flow resistance, a pressure loss linearly increasing with flow rate. Poiseuille number. (Po) can be applied to vascular leaf formations and represented as a resistor conductance circuit. Fig. 3 . A range of electrical potential can determine leaf vasculature optimization by conductance circuit increases. This is a relationship to channel length, radius and hierarchy, Fig. 4 . (a) Defined a network hierarchical multiplication order with a circulative loop network in connection to flow pressure. (b) Represents a non- hierarchical loop network in relationship to flow pressure. (c) Colour plot bar denotes fluidic pressure. (d) Y denotes (dissipation) hierarchical network loop efficiency of flow and conductance distribution. This is called a sink model fractal network of branching sequential order. Figure 4 , denotes the importance of establishing a fixed pressure (or electrical potential) by optimum resistance from fluidic source input. In leaf venation networks this is concentrated in the outermost edges of the leaf of high resistance conductance. Each channel within leaf vasculature is self-organized with its own independency for optimum potential. This hierarchical fluidic transport efficiency for optimal channel networks, Fig. 5 , is defined by y = 0.75. When y is small enough, y = 0.25, there is an increasing breaking down of fluidic flow and flow resistance in connection to fluidic source input 12 13 . When y > 1.0 the network has no hierarchy with uniform order with nonzero conductance to leaf edges 14 . This represents increased pressure, resistance and concentration of channel cross sectional area focused on fluidic input into the network 15 16 . Fluctuations changes in optimum structured networks is a correlationship of laminar fluid flow and resistance 17 . Varying independent channel optimum resistance will determine fluidic transport hierarchy and minimization of pressure drop. This pressure drop will vary resistance in the network between fluidic multi micro channels. To determine pressure drop in longitudinal channels R,L0 to R,L3, Fig. 5 , is dependent on upstream (Rp1 to Rp4) and downstream (R_cp1 to R_cp4) micro channels. Analysis for both the upstream and downstream channels can be determined, to evaluate if the upstream and downstream resistances are different. Knowing the pressure drops (delta P) will allow an estimate of the actual flow resistances. Analysis of delta P will determine optimization of resistance by microchannel sequence succession. Multi micro channels widths are significant as longitudinal microchannels Length and microchannel Depth is determined by material scale. If we assume flow rate is equal within multi microfluidic channels, we can evaluate flow; to predict pressure variation by analysis. Volumetric flow is evaluated by; u_bar (velocity coefficient,) Dh (hydraulic channel diameter), Re (Reynolds number based on hydraulic diameter), Po (Poiseuille number), tau (mean wall channel shear stress) and delta P, along each channel. Resistance is then evaluated from R = delta P/Q flow rate. Fluidic inlet flow to feed distally channels by optimization, is achieved through pressure equalization by diminishing flow pressure variation. This equalization of resistance transport flow is resistance-seeking targeting that can be presented as a resistor network, Fig. 5 . Material channel resistor network Evaluation of resistance optimization is centered on a single microchannel that all other channels succession sequence will emulate. This resistance-seeking targeting is defined as having, R0 and length L0. This microchannel presents the path as least resistance for fluidic flow through the vasculature network. The greatest resistance to flow is presented by R4 as it is distally removed from fluidic input. Other channels are then labeled R1, R2 etc. in sequence from the target resistance channel. If we assume that we know R0 the aim is to design the channel geometry to give equal flow rates through all the microchannels. Fluidic feed in and out flow manifold micro channels are denoted by Rm1, Rm2 etc. Resistances of each longitudinal channel is evaluated and designed to an individual particular resistance function. The equation for the resistances follows a recursive pattern: Consider the flow channel R1: Fluidic flow rate Q, can be cancelled out, as flow rate will be constant, assumed, if equal flow resistance is achieved. Hence analysis of target resistance for the following channels can be considered: Channel R2: Channel R3: Hence the sequence will continue for R4. An algorithm is used by increasing the width of the channel in fixed increments of delta w (delta w to be equal to 0.1 microns, to enable 1 micron accuracy). If the resistance of the channel is greater than the “target” hydraulic resistance (of the central channel), then the program increases the width by delta w. Hence a ratio of the maximum resistance to minimum resistance can be optimized. This approach of defining individual channel resistance is the same approach applied in leaf vasculature. This analysis will then automatically feed into calculating the target resistance of the various side upstream and downstream distribution manifold Rm channels. Target resistance is required to give the same flow rate through each channel with the resistance of the outermost channel, R4, needs to be approximately 25% lower than R0. Once optimized widths are achieved the flow rate can be estimated to test whether the proposed method of optimization is successful against CFD simulation in a polymer device. Microfluidics Translucent Device Design The channel vascular geometry design in the polymer device was set at longitudinal channels equal spacing pattern formation of 15.575 mm, with channel widths of: R0–2.0 mm, R1–2.3 mm, R3–2.6 mm, R4–2.8 mm and the outermost channel R4–3.0 mm. Hence target hydraulic resistance of the central channel is determined, as indicated below. This experimentation by an algorithm is a analytical solution for resistance, by channel width succession for sequence optimization. Hydraulic resistance of central channel (2.0 mm wide × 1.0 mm deep). Assume a constant temperature of 25 °C. At this temperature, the dynamic viscosity coefficient is equal to The pressure drop for fully-developed flow along a section of channel of length, L , can be determined by balancing the pressure forces and the wall shear forces: where is mean wall shear stress and P is the wetted perimeter of the channel, mean flow velocity in the channel. The Reynolds number of the flow is defined by where D h is the hydraulic diameter of the channel, defined as Assuming the flow is laminar, we can then use the Poiseuille number to calculate the average wall shear stress. The mean wall shear stress, , can be related to the Fanning friction factor , f , which in turn can be expressed as the ratio of the Poiseuille number, Po, and Reynolds number, Re: Substituting for Re in Eqn. (5) gives Substituting the shear stress into Eqn. (2) gives Finally, the hydraulic resistance , R , of the channel is given by Hence, This is identical to Eqn. (18) in Emerson et al . 18 . Consider a 2 mm wide by 1 mm deep channel, 186.243 mm long. The hydraulic diameter of the channel is defined as From Table 42 in Shah and London 19 , the Poiseuille number, Po, is 15.54806 for a square duct 2:1 aspect ratio. Thus, The algorithm follows the same procedures detailed above, with the exception that the Poiseuille number, Po, for a given, channel height and aspect ratio, α , is found using the analytical solution involving an infinite series summation (obtained by combining eqns 333 and 340 in Shah and London): where α = h / w . Once R0 is known, choosing the optimized value of resistance to achieve equal flow rate through all channels can be determined. This mathematical design procedure predicted pressure drop results of the outermost channel are in good agreement for a fully developed laminar flow that gives validity of the algorithm code. However the ratio of the maximum resistance/minimum resistance = 0.29477840E + 09/0.15429990E + 09 = 1.91. Of the initial channel sequence was high. Resistance variations in the channel sequence are far to high as a optimized solution. If a fixed pressure was to be applied across the test device, then the flow rates would also vary by the same factor of 1.91. High resistance of the central channel would imply that the volumetric flow rate in the central channel is lower, than that in any of the other channels. The issue is, Fig. 5 , the pressure at the start of the longitudinal channels vary considerably by fluidic input channel resistance for the tapered channel sections Rm1 to Rm4. These distributions input and export manifolds, will change the fluidic pressure to each individual longitudinal channel R0 to R4 entrance and exit. Hence this variation can be predicted in the pressure applied to the individual channels. If the pressure variation in Rm, manifolds channels, is great in comparison to the pressure drop along the channels, the optimization strategy must account for this. Hence the formulated design procedure in terms of the mathematical algorithm can be determined by analysis in a two-stage approach., The second stage program calculates the widths (longitudinal channels) that have been considered in eqns (1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12) . Determine the widths of the manifold at the location of each longitudinal channel: The change in manifold width between each successive longitudinal channel is given by Thus, Generalizing: Determine the average width of each of the tapered manifold channels: (i.e. determine the width of the manifold half-way along each section) Generalizing: Determine the hydraulic resistances of the manifold channels: From the previous analysis, we know that the hydraulic resistance, R , of a rectangular channel having a constant cross-sectional area is given by If we assume that the angle of the manifold taper is small, then the individual flow resistances in the manifold can be calculated using the cross-section at the mid-point. Thus, where A i and D hi are the area and hydraulic diameter at the mid-point of the manifold channel, i . e . It should be noted that Po in Eqn. (23) is a function of the aspect ratio ( ) of the channel. It is informative to calculate the individual pressure drops along the manifold, as these can be checked against CFD results. Fundamentally, the pressure drop along a section is simply the product of the flow rate and the hydraulic resistance, i . e . ∆ p = QR . Thus, Generalizing: Equation (29) is for the case with four channels either side of the central channel. In the case of N side channels: Finally, we calculate the required hydraulic resistances of the longitudinal channels: The pressure drop across the central longitudinal channel is given by We can then determine the required resistances of the other longitudinal channels: Consider channel 1: Hence, or Consider channel 2: Hence, or Substituting Eqn. (34) into Eqn. (37) gives Similarly, it can be shown that and The resistances therefore can be determined recursively using: or In the case of N side channels: Once we have determined the required resistances, R 1 to R 4 , we can then calculate the optimized channel widths. The methodology for this has already been considered to optimize the channels so they each had the same resistance, R 0 . An identical procedure is used here but the “target resistance” needs to change slightly for each individual channel to compensate for the additional pressure drop in the distribution manifolds. Flow rates will never be equal to each other, as there are other influences effecting flow rate, such as entrance effects and flow curvature at the start and end of the longitudinal channels. The experimentation analytical solution analysis for the resistance (1–43) requires validation against CFD simulation. Computational Fluidic Dynamics–Translucent Device Design Results of a CFD simulation of the inlet manifold for 2 mm deep set of channels sequences has been undertaken. CFD analysis of the manifold, Fig. 6 , focused on the symmetry boundary condition along the centerline of the channel. Prescribing an arbitrary inlet pressure of zero ran the algorithm. A mass flow rate was specified across the outlet boundaries (mdot = 1.66166 × 10 −5 kg/s) for the full-sized outlets and half this mass flow rate at the central outlet to account for symmetry. The CFD results show that the pressures at the start of the longitudinal channels vary quite considerably. Hence the pressure at the entrance to the central channel is approximately −0.2 N/m 2 whereas the pressure at the entrance to the outermost channel is about −0.75 N/m 2 . All pressures are relative to the inlet of the modeled flow polymer device. Hence the pressure difference of 0.55 N/m 2 exists between the various channels. This compares with a theoretical pressure drop of 4.91 N/m 2 along the central channel. The pressure drop in the inlet manifold may be up to ~10% of the pressure drops in the main channels. CFD modeling of the device with optimized widths will estimate the flow rate variation between the channels Fig. 7 . CFD simulation gives comparable pressure drops in the four tapered sections and comparable estimates for Rm1 to Rm4. Analysis indicates the approximating the taper by using cross-section, half way along the section is valid for this device. These are the new target resistances; Fig. 8 , that gives channel succession steady state flow targets Note that the resistance of the outermost channel, R4, needs to be approximately 25% lower than R0. Second stage is calculating the target resistance of the various channels. Hence the outermost channel R4 was undertaken to assess flow resistance to the target central channel. Below is the CFD simulation of 3.0 mm × 2.0 mm outermost channel: CFD model data:- Channel width = 3.0 mm Channel depth = 2.0 mm Channel length = 203.635 mm ρ = 997 kg/m 3 density of the fluid (water) μ = 8.9 × 10 −4 Ns/m 2 viscosity of the fluid A computational mesh composed of 30 × 20 × 400 cells (=240,000 cells) has been used for the CFD simulations. The volumetric flow rate, Q , is specified to be However, the CFD code (CFD-ACE+) needs the flow rate to be specified as a mass flow rate, m in kg/s. Thus, The mass flow rate was specified at the inlet boundary whilst the pressure was set to zero (atmospheric conditions) at the downstream boundary. Results from the analytical solution: Input channel width 0.003 (m), channel length 0.203635 (m) alpha = 0.66667 Po = 14.71183884 u_mean = 0.00277778 m/s Dh = 0.00240000 m Re = 7.46816 tau = 0.00757728 N/m 2 deltap = 2.57167 N/m 2 Resistance = 0.15429990E + 09 The CFD simulation was setup to estimate the overall pressure drop along the channel for a mass flow rate of 1.66166 × 10 −5 kg/s (corresponding to 1 ml/min). Figures 9 and 10 , illustrates that the predicted pressure drop along the channel R 4 is 2.5836 N/m 2 , which is in very good agreement with the analytical solution of 2.57167 N/m 2 . Computations indicate the hierarchical optimized sequence of longitudinal channel widths should be in the following sequential order: R0–2.000 mm, R1–2.095 mm, R2–2.202 mm, R3–2.317 mm and R4–2.431 mm, Fig. 11 . We could also use the CFD results, Tables 1 and 2 , to compute the actual hydraulic resistances in the manifold, i.e. Rm1 to Rm4. Analytical resistances are generally in good agreement with the CFD results Tables 1 and 2 , with a maximum error of ~15% for Rm1, a 7% error for Rm2 and Rm3, and a 1% error for Rm4. This procedure obeys the rules of minimum flow velocity rates, minimum resistance to achieve the lowest pressure drop within a microfluidic network. Each micro channel within this artificial network will influence thermal conductance if subjected to impact heat load. The fluidic medium is therefore acting as a heat sink. Hence fluidic laminar flow rate regulation will determine and influence thermal interface transport exchange to a translucent material device. Experimental Testing Method The laboratory testing of the prototype is not focused upon thermal conductivity but the adsorption of solar (ie non-thermal) IR, which then will heat up the polymer structural assembly. Transition temperature of the polymer, will be characterized by volumetric based steady state flow. This capture of energy by solar modulation will progress a thermal function polymer as a IR radiation stop band with lower phase transition temperature. Fabrication consisted of two plates of 5 mm polymer to create the structural assembly. The base plate contained the microchannel network that is fabricated by laser cutting into the surface of the base plate. This channel geometry will contain the microfluidic based flows. The polymer counterplate acts as the solar radiation absorber pane. These two plates have been bonded together to form the structural assembly-testing device, Fig. 12 . The optically clear polymer is subjected to an artificial solar (incandescent light) source that emitted IR wavelength 1000 watts per m 2 . Solar heat load increase the surface temperature of the polymer surface pane. Distilled water is be pumped through the channel network that directed the structural assembly of the polymer. The fluidic input and extract temperature into the manifolds channels was monitor by thermocouples. Heating of the fluid under flow gave a temperature profile. Sensors monitored material–fluid thermal flow across this interface by extract water temperature. This analysis will enable assessment of thermal switchable IR absorber by water flow. Figure 13 illustrates experimental testing design. The heating effect from the polymer surface pane and switchability IR absorber was observed by experimentation. Heat transport across the fluid and polymer material interface was evaluated for energy capture. Fluidic interface with polymer material regions, under uniform flow, acted as an IR absorber to lower device phase transition temperature. A load radiation density applied was 1000 W/m 2 to the device. The counter plate acted as a partially absorbing 210 W/m 2 pane, the fluid adsorbed 707 W/m 2 and the remaining 83 W/m 2 was transmitted by radiation through the polymer. The length of a longitudinal channel in volumetric steady state flow is an individual heat linear absorber of IR within a multi microchannel network. The energy balance of solar radiation is dependent on hydrodynamic behaviour of fluids in steady state pressure driven flows. Results indicated heating of water connected to a partially absorbing pane by passage through the microfluidic based flow gave thermal switching characterization. By modulating volumetric flow rates in the device enabled a temperature difference to decrease roughly inversely with flow rate. Tailored flow rates gave a controlled processing of a thermally functional polymer by microfluidics for desired solar absorber characterization. Future Translucent Device Application Using microfluidic based flows into a structural assembly of a polymer will advance materials desired energy capture and storage functionality. This steady state flow network of continuously circulating a fluid within it, through it and out of it, by microfluidic based flows to direct the structural assembly of a polymer 20 . This uniform parabolic flow will remove stored liquid temperature out of the polymer for solar energy modulation efficiency. If this liquid is replaced with incoming fluid, this creates a photoabsorptive system. This approach enables thermal switching selectivity of a polymer device in response to heat load, IR. This research is not focused on thermal conductivity but the absorption of solar (non-thermal) IR by heat built up. This represents a thermal exchange transfer cycle of fluidic absorption through vascular channels, Fig. 14 . The micro vascular network will determine thermal switching optimization to material temperature regions. Multi microchannel network will regulate material temperature by management of: Resistance Optimization. Radiative/Convective heat interface transfer. These parameters will give optimization of a thermally functional material in relationship to surface temperature fluctuations. The heating effect from a surface material pane is regulated by water uniform parabolic and laminar flow profile for transition temperature decrease, Fig. 15 . Management of thermal flow would progress building translucent facades internal and external surfaces, as the polymer device will act as a thermal flow bridge. Present envelope glazing systems depend on reducing the g-value with solar radiation shading, for minimizing internal thermal load transmission. However these component systems cannot adapt to changing environmental conditions, as they are designed to static boundaries as determined by U-value. These performance modes must change it role from a static element to an energetic façade. Climatic global warming requires performance change by the hour, season and weather conditions. Nature has developed functional materials of complex hierarchy to regulate thermal conductance by venation 21 . Greater demands have been placed to minimize operational building energy use by maximizing generated energy and day lighting that are integrated within the building envelope 22 . A thermally functional polymer will advance these aims as a permanent IR absorber, to adapt to changing environmental conditions. To enable transition temperature decrease as a heat flow cycle, for regulation of thermal interface transport exchange to material regions Fig. 16 . This thermal management may also enable dehumidification of translucent facades by convective cooling by air to external surfaces. This is a dynamic heat seeking system to progress current static facade to a thermally functional adaptive layer 23 24 . The integration of artificial microfluidic networks of solar absorbing fluid in active flow is a new methodology. Future progression is to determine thermal characterization of the device by thermal switching for heat flow targeting. This may impact on the vascular resistor computational process via measuring thermal conductance heat flow fluxes, in relationship to laminar thermal flow absorption. Conclusion of Methods This iterative procedure will determine optimization of vasculature employing resistance seeking targeting by hierarchical channel succession, as a resistor networks. Hence this optimization process is determined by selection of a known hydraulic resistor value R 0 , to enable evaluation that determines vascular channel network geometry, to give flow equalization as a development mechanism as denoted by: The pressure drop across the outermost longitudinal channel is given by We can then determine the required resistances of the other longitudinal succession channels R 1 , R 2 , R 3 etc that can be determined as a recursive pattern: Feed in principle upstream and down stream channels, manifold resistance, Rm 1 , Rm 2 , Rm 3 is determined by N side channels ( N denoting the total number of channels is 2 N + 1): This computation design methodology of the device predicted pressure drops results in the manifold are in good agreement with CFD results. Results indicate resistances of the longitudinal microchannels are similar to the theoretical results. The tapered sections resistances are in good agreement, except for the tapered section involving the inlet port. This is expected since the computation solution cannot take into account the flow expanding away from the circular inlet port. Computational analysis conclusively demonstrates being able to design microfluidic networks using a theoretical approach, to achieve optimization of circuit resistance of transport fluidic flow. Optimization is achieved through pressure equalization by diminishing flow pressure variation. The two-step algorithm approach enables self-organized channel resistance with its own independency for optimum potential for pressure drop regulation. Methodology of successive conduit sequences of hierarchical formations perform regulatory roles. This is functionally significant in the analysis of hydraulic resistance to compute simulations of flow rates in microvasculature channel networks. Predicted pressure drop and flow analysis within channel network is in agreement with the analytical solution for fully-developed laminar flows, giving validity to the algorithm code as a iterative procedure. The analytical results using theoretical resistance are based on R0. The weakness of this approach is represented by R1 and R0, as CFD simulations already indicate that flow rates through R4, R3 and R2 are almost identical. If the network is constrained to device 148 mm width, multi microchannel sequence succession could start from the outermost channel and work inwards for R3, R2, R1, and R0. The attraction of this is the footprint of the network stays constrained to a 148 mm width with the outermost edge 3 mm microchannel. This method is a reverse analysis as the outermost channel width is unknown to begin with and so the width of the device cannot be known in advance. The optimized resistance condition could be based on theoretical manifold resistance. If design channel widths uses Rm manifold resistance obtained from CFD simulation to estimate the desired flow resistance in longitudinal channels, as given by: Flow rate through the manifold is denoted by Qm . This would lead to greater optimized microfluidic design, by code modification to obtain uniform flow across the polymer pane. The morphogenesis of leaf vasculature sets an underlying process of flow distribution, pressure, fluidic transport and resistance. This notion of precise hydrodynamic control of microfluidic’s will progresses thermal material characterization; to advance a polymer device, into a switchable IR absorber. This is achieved by hierarchical succession of branching sequence patterns that conform to rules of minimum effective power flow rates in the transportation of fluidics within fractal networks. That is determined by computational theoretical analysis of resistance seeking targeting."
} | 8,367 |
31156601 | PMC6533897 | pmc | 1,087 | {
"abstract": "The process of anaerobic digestion in which waste biomass is transformed to methane by complex microbial communities has been modeled for more than 16 years by parametric gray box approaches that simplify process biology and do not resolve intracellular microbial activity. Information on such activity, however, has become available in unprecedented detail by recent experimental advances in metatranscriptomics and metaproteomics. The inclusion of such data could lead to more powerful process models of anaerobic digestion that more faithfully represent the activity of microbial communities. We augmented the Anaerobic Digestion Model No. 1 (ADM1) as the standard kinetic model of anaerobic digestion by coupling it to Flux-Balance-Analysis (FBA) models of methanogenic species. Steady-state results of coupled models are comparable to standard ADM1 simulations if the energy demand for non-growth associated maintenance (NGAM) is chosen adequately. When changing a constant feed of maize silage from continuous to pulsed feeding, the final average methane production remains very similar for both standard and coupled models, while both the initial response of the methanogenic population at the onset of pulsed feeding as well as its dynamics between pulses deviates considerably. In contrast to ADM1, the coupled models deliver predictions of up to 1,000s of intracellular metabolic fluxes per species, describing intracellular metabolic pathway activity in much higher detail. Furthermore, yield coefficients which need to be specified in ADM1 are no longer required as they are implicitly encoded in the topology of the species’ metabolic network. We show the feasibility of augmenting ADM1, an ordinary differential equation-based model for simulating biogas production, by FBA models implementing individual steps of anaerobic digestion. While cellular maintenance is introduced as a new parameter, the total number of parameters is reduced as yield coefficients no longer need to be specified. The coupled models provide detailed predictions on intracellular activity of microbial species which are compatible with experimental data on enzyme synthesis activity or abundance as obtained by metatranscriptomics or metaproteomics. By providing predictions of intracellular fluxes of individual community members, the presented approach advances the simulation of microbial community driven processes and provides a direct link to validation by state-of-the-art experimental techniques.",
"conclusion": "Conclusion We are only at the beginning to understand the complex interplay at work in microbial communities in natural or engineered systems. Community systems biology and quantitative modeling are instrumental in advancing our understanding ( Zengler and Palsson, 2012 ; Hanemaaijer et al., 2015 ). Here, for anaerobic digestion which is harnessed in biogas reactors to produce methane from waste streams, we provide a first step to switch from process-based gray box models to species-based community models. As these models provide flux predictions on the enzyme level down to the genome-scale, they provide a convenient common reference to which multi OMICS data can be related to. Likewise, such data can be used to refine the model by constraining fluxes to observed ranges. Integrating multi OMICS data with modeling is a promising strategy to elucidate microbial interactions in complex communities ( Zuñiga et al., 2017 ). While FBA models have been employed before to determine optimal community compositions at steady state ( Koch et al., 2016 ), we here expand this approach to the dynamic situation. Our coupled models provide a convenient tool to interpret time series data from operational biogas plants, to explore theoretically possible maximal yields and process efficiency, to identify early warning signals indicating looming reactor breakdowns and to test intervention strategies to avoid costly reactor breakdowns. To make optimal use of microbial community driven processes, the need for active management has been recognized ( Carballa et al., 2015 ). In this context, mechanistic community models resolving intracellular activity are a crucial component, which prospectively could be integrated into a model-based online monitoring and control scheme.",
"introduction": "Introduction Anaerobic digestion is a naturally occurring process driven by a microbial community which is harnessed in biogas plants to convert organic waste material to methane and CO 2 . Being a suitable building block in a renewable energy landscape, biogas production is a popular topic in research and development ( Lora Grando et al., 2017 ). The elucidation of the four sequential steps of anaerobic digestion – hydrolysis, acidogenesis, acetogenesis, and methanogenesis – has allowed for the development of mathematical models that describe the full process. Anaerobic Digestion Model No. 1 (ADM1) is one of the most prominent models and has been in use for more than 16 years ( Batstone et al., 2002 ). The model captures the four process steps of anaerobic digestion by a set of ordinary differential equations (ODEs). Biochemical conversions are expressed as first order kinetics for hydrolysis, and as Monod-type kinetics with additional inhibitory terms for the remaining steps. Microbial activity is resolved on the functional level, with seven state variables indicating the abundance of sugar degraders, amino acid degraders, long chain fatty acid degraders, valerate and butyrate degraders, propionate degraders, and acetoclastic and hydrogenotrophic methanogens. While attempts have been made to include microbial diversity on individual process steps in ADM1 ( Ramirez et al., 2009 ), the inclusion of individual species with their unique metabolic potential has not been attempted before. As experimental data on the genomic level becomes more and more readily available, models are desirable that can take advantage of such data ( Kreft et al., 2017 ). While ADM1 is able to model reactor performance and other abiotic data well, it’s predictions regarding the microbial community, for example total biomass, are less certain. Including microbial community data is hence expected to improve the performance of current models drastically ( Lauwers et al., 2013 ). For sequenced species with annotated genomes, constraint-based techniques are today a standard tool to predict a species’ phenotype from its genotype ( Lewis et al., 2012 ). In particular, Flux-Balance-Analysis (FBA) can be used to predict a population’s growth rate, cross-membrane compound fluxes as well as intracellular metabolic flux distribution, given the metabolic network as defined by the totality of its enzymatic repertoire. These predictions rely on the assumption that intracellular metabolites are at steady state, i.e., their production rates match their consumption rates, and that the cell orchestrates its metabolic flux distribution in order to maximize its growth rate. In dynamic FBA, the steady-state assumption is restricted to consecutive short time intervals, so that dynamic trajectories can be simulated as with regular ODE based models ( Antoniewicz, 2013 ). The FBA approach has been successfully coupled to reactive transport models, increasing model predictive power by reducing the need for empirical calibration ( Scheibe et al., 2009 ). Such models provide quantitative predictions of intracellular activity which are compatible with measured OMICS data targeting transcription activity of enzyme-encoding genes, or enzyme abundance directly. Experimental data can either serve as a benchmark, or can be used to refine the model ( Reed, 2012 ). Dynamic FBA modeling provides predictions on the temporal evolution of up to 1,000s of enzymatically catalyzed metabolic fluxes per species if genome-scale metabolic network models are employed. Such models become available for a steadily increasing number of microbial species ( King et al., 2016 ; Magnúsdóttir et al., 2016 ). In this work, we evaluate whether biogas process modeling can benefit from these model advancements by coupling selected FBA models of methanogenic archaea of varying complexity to ADM1. For this purpose, we replace the acetoclastic and/or hydrogenotrophic methanogenesis pathways in ADM1 by FBA models of Methanosarcina barkeri and Methanococcus maripaludis and compare both steady-state simulation results and a pulsed feeding scenario.",
"discussion": "Discussion Opening the Gray Box: Making Models OMICS-Data-Ready After 16 years of process-based modeling of anaerobic digestion with ADM1, we here demonstrate the feasibility of replacing individual process steps in ADM1 – here the two methanogenesis pathways – by microbial species-centric FBA models detailing intracellular metabolic activity in down to genome-scale detail. Tracking the dynamics of complex microbial communities in terms of compositional changes over time has become feasible by cultivation-independent techniques focusing on next-generation-sequencing [amplicon sequencing targeting the 16S rRNA gene ( Pace, 1997 )] or single-cell based methods ( Liu et al., 2018 ). Although the absolute quantification of species-specific cell numbers or biomass in complex communities remains a challenge ( Bonk et al., 2018a ), relative compositional data have become readily available. To make use of such data, mathematical models need to contain state variables which refer to these populations. ADM1 focuses on process steps which are mathematically described by ODEs and which are carried out not by a single species but typically by a community of varying diversity, depending on the process step. ADM1 has been extended before to include a number of hypothetic species which catalyze the same process step but with different kinetic parameters to account for this diversity ( Ramirez et al., 2009 ). This approach provided insights into the effect of diversity on reactor performance and response to perturbations in terms of resistance and resilience. However, as this approach sticks to ADM1’s ODE formulation to describe metabolic activity, many more difficult to measure parameters were necessary in the model. These had to be chosen randomly, making this approach difficult to fit to experimental data on microbial community dynamics. Instead of increasing diversity, our model focusses on improving the description of biochemical activity for individual process steps. In this first step, we replace the ODE formulation subsuming the activity of all species performing a process step by a FBA model of one representative species. In the next step, additional competing species can be included as FBA models to capture the diversity as in the model by Ramirez et al. (2009) . Individual FBA models detail how substrate and other compounds taken up from the environment by a particular microbial species are transformed into biomass, maintenance energy, and secreted products by the species’ metabolic network. Enzymatically catalyzed reactions are the building block of these networks, and models provide quantitative predictions for all intracellular fluxes. Such data can be compared to measured enzyme activity obtained by metatranscriptomics or metaproteomics. Taking the opposite route, such experimental data can also be used to constrain the model ( Reed, 2012 ), although the correlation between transcriptomic data, proteomic data, and actual enzyme activity is still under debate ( Haider and Pal, 2013 ). Nevertheless, this enhanced compatibility with state-of-the-art experimental techniques is a great asset of the presented coupled model. Model Coupling and Comparison Coupling FBA models to ADM1 using the direct approach proved to be easy to implement and feasible in terms of the additionally required computational demand. The complexity and stiffness of ADM1’s ODEs likely make the static or dynamic optimization approach, as alternatives to the direct approach, infeasible due to their excessive computational demand ( Mahadevan et al., 2002 ). The coupled model inherits the flexible parametrization available in ADM1, for example allowing for the easy implementation of changing compositions of the reactor input and dynamic feeding regimes. Simulation results regarding overall process performance agreed well between the standard ADM1 and the novel coupled models. Differences were observed regarding the biomass concentrations and their dynamics for the microbial species now simulated by FBA models. Both methanogenic populations of M. barkeri and M. maripaludis had lower predicted biomass concentrations in the coupled model. Taken together with the unaltered prediction of methane production, the novel model hence predicts a higher per cell activity of the respective populations. Under a pulsed feeding scenario, the reference ADM1 model predicted identical trajectories of the biomass concentration for both methanogens, while the coupled model predicted the growth dynamics of the hydrogenotrophic population to fluctuate more, leading to stronger shifts of the contribution of both pathways, yet overall leading to lower amplitudes in gas and methane production. Non-continuous feeding had been shown before to favor more robust microbial communities ( De Vrieze et al., 2013 ; Bonk et al., 2018b ). Exploring the limits of flexible feeding regimes is additionally an important aspect that needs more consideration within the context of the flexibilization of biogas production. This would allow to offset the fluctuating output of other renewable, weather-dependent energy supplies such as wind and solar energy ( Mauky et al., 2017 ). Yields, which are specified as constant parameters in ADM1, do no longer need to be specified in the coupled model, where they are determined by the available substrate, the current demand for cellular maintenance, and the topology of the metabolic network. Yields become a model output, allowing for a more realistic incorporation of the non-constant relationship between yield and growth rate ( Lipson, 2015 ) in the coupled models. Maximization of microbial growth was used as the optimization criterion during FBA computations. Except for a fixed flux constraint for considering the non-growth associated maintenance demand, no other constraints were imposed on internal fluxes, so that predicted flux distributions were also optimal with respect to biomass yield. This would not be the case for overflow metabolism situations. To consider such cases, protein allocation can additionally be considered and biomass yield instead of growth rate can be optimized for, requiring more sophisticated techniques ( Adadi et al., 2012 ; Klamt et al., 2018 ). To model M. barkeri , we used both a minimal (103 reactions) and a genome-scale FBA model (816 reactions). Simulation results agreed very well qualitatively ( Figure 2 – 4 ), and only minor quantitative differences were observed, for example a minimal acetate concentration of 17 mg/L, allowing for growth for the genome-scale model vs. 22 mg/L for the minimal model. In light of these minor differences, the minimal model seems to capture the essential growth features of M. barkeri very well and is sufficient to model this organism as part of an anaerobic digestion microbial community. Non-growth Associated Maintenance For the coupled model, we observed predicted reactor breakdowns for NGAM values beyond 1.4 mmol ATP/gDW/h, when simulated under constant conditions. For the dynamic simulation, we selected a value of 0.5 mmol ATP/gDW/h as this value led to steady-state predictions close to those of the standard ADM1. While higher values have been estimated before, including 0.9 mmol ATP/gDW/h for M. maripaludis , and 3.6 mmol ATP/gDW/h for acetate converting M. barkeri [collected in ( Koch et al. (2016) ], maintenance energies have also been reported to be lower than theoretical predictions under methanogenic conditions ( Scholten and Conrad, 2000 ). A combination of experimental and modeling approaches has recently suggested even smaller values of below 0.116 mmol ATP/gDW/h for short hydraulic retention times, likely being applicable to also longer hydraulic retention times ( Bonk et al., 2019 ). Fitting with this observation, maintenance demands in a binary propionate utilizing syntrophic methanogenic culture were experimentally determined to be 0.14 mmol ATP/gDW/h for Syntrophobacter fumaroxidans and 0.025 mmol ATP/gDW/h for Methanospirillum hungatei ( Hamilton et al., 2015 ). Under slow growing conditions such as in anaerobic digestion, care must hence be taken when selecting NGAM values, as they strongly impact predicted community composition ( Koch et al., 2016 ). Choice of Species to Model To add species-specific state variables to ADM1, we started by replacing ODE descriptions of both methanogenic pathways by species-specific FBA models. Methanogenesis is the archaea-driven last of four steps in anaerobic digestion. It is associated with a low microbial diversity, which increases toward the first steps of the process ( Campanaro et al., 2016 ). As only few currently known archaea are capable of methanogenesis and as they are well-described, the choice of methanogenesis as our first target was both natural and the most straight-forward. We only considered one species per pathway, while in typical biogas microbiomes, more than one methanogenic species will compete for acetate and hydrogen. Our coupled model can easily be extended to include such competition. Somewhat more challenging is the replacement of the other three steps preceding methanogenesis in anaerobic digestion. Not only is the microbial diversity higher in those steps, but also the functional roles of only few species have been elucidated. For acetogenesis during which volatile fatty acids are transformed to acetate for example, only three syntrophic propionate oxidizers and three syntrophic butyrate oxidizers have been described ( Worm et al., 2014 ). A further bottleneck is the availability of FBA models. For example, up to date no FBA model for any of the three known butyrate oxidizers is publicly available. This requires additional efforts in generating the required models, although current computational pipelines brought model development times from years down to weeks ( Latendresse et al., 2012 ). And even for not yet cultivated species, growing genomic resources focusing on anaerobic digestion hold the promise to reconstruct at least draft models for such species ( Campanaro et al., 2016 ), ultimately allowing for FBA-based coupled models to capture the natural diversity of anaerobic digestion. We selected the methanogenic species in this study based on their relevance in anaerobic digestion and the availability of respective FBA models. Once models become available for methanogens often found in biogas reactors including for example Methanoculleus marisnigri and Methanosaeta concilii , it will be possible to include exactly those species in the model, which have been found to be dominant in the anaerobic digestion process to be simulated, for example based on 16S rRNA gene- and/or mcrA -based community analysis. Such tailor-made models offer the most faithful representation of the actual microbial community at hand and are expected to surpass current modeling approaches in their predictive power for simulating industrial biogas processes."
} | 4,842 |
33690896 | PMC8244046 | pmc | 1,089 | {
"abstract": "Abstract Variation among functionally similar species in their response to environmental stress buffers ecosystems from changing states. Functionally similar species may often be cryptic species representing evolutionarily distinct genetic lineages that are morphologically indistinguishable. However, the extent to which cryptic species differ in their response to stress, and could therefore provide a source of response diversity, remains unclear because they are often not identified or are assumed to be ecologically equivalent. Here, we uncover differences in the bleaching response between sympatric cryptic species of the common Indo‐Pacific coral, Pocillopora . In April 2019, prolonged ocean heating occurred at Moorea, French Polynesia. 72% of pocilloporid colonies bleached after 22 d of severe heating (>8 o C‐days) at 10 m depth on the north shore fore reef. Colony mortality ranged from 11% to 42% around the island four months after heating subsided. The majority (86%) of pocilloporids that died from bleaching belonged to a single haplotype, despite twelve haplotypes, representing at least five species, being sampled. Mitochondrial (open reading frame) sequence variation was greater between the haplotypes that experienced mortality versus haplotypes that all survived than it was between nominal species that all survived. Colonies > 30 cm in diameter were identified as the haplotype experiencing the most mortality, and in 1125 colonies that were not genetically identified, bleaching and mortality increased with colony size. Mortality did not increase with colony size within the haplotype suffering the highest mortality, suggesting that size‐dependent bleaching and mortality at the genus level was caused instead by differences among cryptic species. The relative abundance of haplotypes shifted between February and August, driven by declines in the same common haplotype for which mortality was estimated directly, at sites where heat accumulation was greatest, and where larger colony sizes occurred. The identification of morphologically indistinguishable species that differ in their response to thermal stress, but share a similar ecological function in terms of maintaining a coral‐dominated state, has important consequences for uncovering response diversity that drives resilience, especially in systems with low or declining functional diversity.",
"introduction": "Introduction Rapid human‐induced climate change has led to an urgent need to identify the genetic, species, and trait diversity associated with the ability of ecosystems to avoid regime shifts (Webster et al. 2017 ). An important element of diversity that maintains ecosystem states is response diversity (sensu Elmqvist et al. 2003 ), where species sharing similar ecological functions differ in their response to perturbations (Chapin et al. 1997 , Yachi and Loreau 1999 ). If functionally similar species respond negatively and in the same way to common stressors, the capacity for biological communities to withstand these effects and avoid shifts to different states is reduced (Baskett et al. 2014 ). However, overlooking cryptic species within functional groups may miss an important source of variation contributing to response diversity, and limit the success of management and intervention approaches aimed at protecting biological systems (Bickford et al. 2007 , Stat et al. 2012 ). Therefore, the identification of response diversity within groups of closely related and morphologically indistinguishable taxa is essential to complement existing trait‐based approaches based on functional diversity determined by morphological variation (Madin et al. 2016 ). There is widespread evidence that cryptic genetic species (evolutionarily distinct genetic lineages that are “hidden” by morphological similarity and plasticity) exist within many recognized “species” (Bickford et al. 2007 , Stat et al. 2012 , Arrigoni et al. 2019 , Cowman et al. 2020 ). Yet studies on how biological diversity contributes to the ability of an ecosystem to maintain a desired state following disturbance (resilience; Holling 1973 ) often do not identify cryptic species, or assume they are ecologically equivalent (Elmqvist et al. 2003 ). Response diversity among cryptic species could increase resilience by increasing the capacity for one species to compensate for the temporary loss of another in terms of the function it provides (Chapin et al. 1997 , Elmqvist et al. 2003 , Baskett et al. 2014 ). Resilience via response diversity, and, the continued coexistence of cryptic species, depends on niche partitioning or a lack of a competitive hierarchy among cryptic species, as well as disturbances varying over spatial scales that are smaller than the expected spatial scale of dispersal (Baskett et al. 2014 ). Corals in the genus Pocillopora , which dominate reefs throughout much of the Indo‐Pacific, are well known for exhibiting morphological plasticity and overlapping morphological phenotypes that frequently make species identification based on gross morphology unreliable (Pinzón et al. 2013 , Marti‐Puig et al. 2014 , Paz‐García et al. 2015 , Edmunds et al. 2016 ). As a result, it is often unknown if analyses of purportedly distinct populations in fact contain multiple species (Bickford et al. 2007 , Stat et al. 2012 ). While the existence of morphologically similar yet genetically divergent lineages complicates the study of population and community biology in the field, it also provides a previously unrecognized source of diversity potentially important for uncovering differences in responses to environmental perturbations. Furthermore, thermal regimes causing coral bleaching can vary over horizontal and vertical scales of meters to kilometers, and timescales of hours to days (Safaie et al. 2018 , Wyatt et al. 2020 ). Importantly, such fine‐scale variation is not accurately characterized by sea surface temperature (SST) data collected remotely and averaged over several kilometers (Leichter et al. 2006 ). Spatial variability in thermal stress and mortality over scales of a few kilometers increases the potential for larval dispersal over scales greater than a few kilometers to supply recruits from less impacted sites to more impacted sites (Baskett et al. 2014 ). Few studies have adequately linked the magnitude and timescales of thermal variability in shallow tropical waters that occurs over scales of meters to kilometers to bleaching impacts on coral reefs (though see Penin et al. 2007 , Safaie et al. 2018 , Wyatt et al. 2020 ) and response diversity. Furthermore, bleaching can also affect larger colonies more than smaller colonies (Nakamura and van Woesik 2001 , Shenkar et al. 2005 , McClanahan et al. 2008 , Brandt 2009 , van Woesik et al. 2011 ). Response diversity could therefore manifest as differences among functionally similar species in their size‐dependent response to thermal stress, in addition to the recognized differences among morphologically identified species (Glynn 1993 , McClanahan et al. 2005 , van Woesik et al. 2011 , Pratchett et al. 2013 , Guest et al. 2016 ). In 2019, a severe bleaching event occurred on the reefs at Moorea, French Polynesia. Prior to bleaching, the cover of hard coral on the fore reef was dominated by broadcast spawning colonies in the genus Pocillopora (Adjeroud et al. 2018 , Holbrook et al. 2018 ). Since at least the 1970s, Pocillopora colonies have been very abundant on the fore reefs of Moorea, but have come to dominate reefs in the last two decades (Berumen and Pratchett 2006 , McWilliam et al. 2020 ). The relatively recent dominance of Pocillopora highlights the importance of studying this genus to better understand why it has been relatively more successful than other genera (e.g., Acropora ) that have declined in abundance at this location (Berumen and Pratchett 2006 , McWilliam et al. 2020 ). Hidden diversity (Souter 2010 , Schmidt‐Roach et al. 2013 ) in this genus, field observations of size‐dependent bleaching (see also K. E. Speare et al., unpublished manuscript ), and spatial variation in bleaching prevalence provided the impetus to study response diversity among cryptic genetic species. Our goals were to (1) quantify the source of the stress by quantifying the regime of temperature variability and how it differed between sites several kilometers apart, (2) quantify the impact of the stress on the Pocillopora community in terms of bleaching prevalence, severity, and mortality, and (3) quantify response diversity within the Pocillopora community as the extent to which bleaching mortality differed among cryptic genetic species. We uncovered differences in the bleaching response of cryptic genetic species occurring in sympatry (response diversity). We also found differences in bleaching mortality among sites that related to local differences in the thermal regime, providing a way for response diversity to influence resilience.",
"discussion": "Discussion Understanding variability in the response of functionally similar species to environmental stress is critical to understand how diversity increases community or ecosystem‐level resilience to disturbances, such as those linked to global change. Previous studies on a range of taxa, such as plants (Laliberte et al. 2010 ), microbes (Wohl et al. 2004 ), and fishes (Nash et al. 2016 ), have shown how response diversity stabilizes ecosystems. In most cases however, species are easily identified. Overlooking cryptic species in population and community studies may therefore miss an important source of variation that could lead to such stability and resilience. For example, instead of changes in population size‐structure that would be expected from not recognizing cryptic genetic species, our results show that bleaching has shifted the relative abundance of species within the Pocillopora assemblage, driven by a decline in haplotype 11. By genetically identifying colonies in the field and censusing them before and after a major bleaching event, we have uncovered previously unrecognized response diversity among cryptic coral species that dominate coral cover on the fore reefs of Moorea (Elmqvist et al. 2003 , Baskett et al. 2014 ). Response diversity within Pocillopora corals may have contributed to this system being able to maintain coral dominance, despite losing diversity in functional traits associated with declines in acroporid corals (Adjeroud et al. 2018 , McWilliam et al. 2020 ). Response diversity was further enhanced by differential heat accumulation, measured as degree heating days using in situ seawater temperature records (and not SST, which is the most common approach), among sites separated by several kilometers. Mortality at the genus level was lower at sites with less heat accumulation, providing the opportunity for less impacted sites to supply recruits to impacted sites in the coming years. Importantly, regardless of whether symbiont identity or composition could offer an explanation for the causes of mortality differences among species, our results have important consequences for understanding how response diversity can maintain coral dominance. Response diversity enhances ecological resilience when species that can withstand thermal stress occupy different habitats, or are not competitively inferior, compared to species that are susceptible to thermal stress but can recover quickly (Baskett et al. 2014 ). The three taxa of Pocillopora that dominated the fore reef at Moorea in February 2019 ( P. meandrina , haplotype 10, and haplotype 11; Fig. 5c ) differed in their bleaching mortality. These three taxa also appear to exhibit differences in their relative abundance across different depths (E. C. Johnston et al., unpublished manuscript ), as well as similarities in the net outcome of survival, growth, and fecundity. An outbreak of crown‐of‐thorns sea stars (ending around 2009) followed by Cyclone Oli in 2010 resulted in the near complete loss of live coral (to <5% cover at all fore reef sites around Moorea; Adam et al. 2011 ). The near complete absence of live coral after these disturbances suggests that the relative abundance of Pocillopora lineages recorded in February 2019 (Fig. 5c ) reflects possible lineage differences in the net outcome of recruitment, growth, and survival between 2010 and 2019. If so, P. meandrina , haplotype 10, and haplotype 11 may have exhibited the highest recruitment, growth, or survival compared to other lineages, thereby rapidly increasing in abundance between 2010 and 2019 to become some of the dominant lineages on the north shore by 2019 (Fig. 5c ). We speculate that the high recruitment of pocilloporid corals that occurred around 2011 on the fore reef of Moorea (Edmunds 2017 ) was dominated by P. meandrina , haplotype 10, and haplotype 11. Furthermore, the larger sizes of haplotype 11 colonies compared to colonies from all other lineages in 2019 (Fig. 5a ) also suggests that haplotype 11 colonies grew faster than colonies belonging to the other haplotypes, or dominated recruitment for the first few years of recovery. These hypotheses, of course, require testing using methods that estimate population vital rates (Edmunds and Riegl 2020 ). A conclusion that bleaching mortality varied among species is not a new discovery (Glynn 1993 , McClanahan et al. 2005 , van Woesik et al. 2011 , Pratchett et al. 2013 , Guest et al. 2016 ). What is novel, however, is that bleaching differed among common species that are closely related, living in sympatry, are members of a relatively young (<3 Mya) monophyletic radiation (Johnston et al. 2017 ), and cannot be reliably identified in the field based on gross morphology (Pinzón et al. 2013 , Marti‐Puig et al. 2014 , Schmidt‐Roach et al. 2014 , Paz‐García et al. 2015 ). The latter, in particular, has forced researchers in the past to pool corals within a genus, preventing a full understanding of the impacts of bleaching in prior events, both in Moorea (Pratchett et al. 2013 ) and in other locations (van Woesik et al. 2011 , Guest et al. 2016 ). More generally, response diversity has received much attention because it underlies ecosystem stability and resilience (Elmqvist et al. 2003 , Baskett et al. 2014 ), but empirical examples in natural populations are exceedingly rare. Available evidence leads to the hypothesis that haplotype 11 suffered the greatest mortality in the recent bleaching in Moorea because it is a “cooler water”‐adapted species. Haplotype 11 has so far only been documented at Moorea and surrounding islands (~17°–18° S; Forsman et al. 2013 , Edmunds et al. 2016 ). Haplotype 11 is closely related to haplotype 6a ( P. ligulata ), differing by only two base pair substitution (mtORF sequences 99.8% identical over 935 bp), but is morphologically most similar to P. eydouxi . Haplotype 6a ( P. ligulata ) has so far only been documented in the Hawaiian Islands (~20°–30° N), and is most common at subtropical islands in the northernmost section of this island chain (Johnston et al. 2018 ). Very little is known about haplotype 6a ( P. ligulata ), or whether it is a distinct species to haplotype 11. In our photographs from March and May 2019, we do not know the identity of bleached and unbleached Pocillopora colonies, or what proportion of colonies that bleached subsequently died or recovered (the colonies surveyed in March and May were different to the colonies surveyed between February and August). We do not, therefore, know if colonies from all lineages bleached and only haplotype 11 subsequently died, or if only haplotype 11 bleached. We speculate that that size‐dependent bleaching affecting Pocillopora in both 2007 (Pratchett et al. 2013 ) and 2019 (Fig. 3 ) was driven by higher susceptibility to bleaching in haplotype 11. Heat accumulation and mortality from bleaching varied among sites separated by several kilometers, and larval dispersal greater than this scale increases the potential for dispersal to connect locations that experience different levels of thermal stress (Baskett et al. 2014 ). The potential for spatial variation in coral bleaching over meters to kilometers is well recognized (Hoegh‐Guldberg and Salvat 1995 , Penin et al. 2007 , Lenihan et al. 2008 ). In particular, the shallow back reef habitat at Moorea has a different regime of thermal variability than adjacent the fore reef, notably between night and day (Putnam and Edmunds 2011 ), and the threats from bleaching are quite different in the back reef than on the fore reef. Our results suggest that spatial variation in bleaching mortality in Pocillopora corals on the fore reef, specifically, could be driven by spatial differences in heat accumulation, spatial differences in the relative abundance of host haplotypes, or both. However, our sampling of genetic material in February 2019 was too limited to directly estimate the interactive effects of site and haplotype identify on the probability of colony mortality. Heat accumulation was highest on the north shore (Site 1) and lowest on the east shore (Site 4) in 2019. The cause of differences in heat accumulation among sites is possibly related to local differences in the internal wave climate and the extent to which internal waves, which normally bring cooler water to the reef, were reduced at each site during 2019 (Leichter et al. 2012 , Wyatt et al. 2020 ). Similarly, mortality at the genus level was highest on the north shore and lowest on the east shore by August. Mortality did not increase with size on the east shore (Site 3 and 4), but all colonies sampled at Site 3 were <30 cm diameter and only 4% of the colonies at Site 4 were >30 cm diameter. Furthermore, site 3 was dominated by haplotype 10 and the relative abundance of haplotypes did not change before and after bleaching. The advent of genomic methods to advance the understanding of species boundaries and evolutionarily distinct units has important implications beyond better estimates of biodiversity (Knowlton 1993 , Bickford et al. 2007 , Stat et al. 2012 ). There is a long history of identifying corals in the field based on morphology, but it is becoming increasingly clear that morphology may not be a sufficient or reliable means of delineating many closely related coral species for the purpose of studying population and community dynamics (Stat et al. 2012 , Arrigoni et al. 2019 , Cowman et al. 2020 ). As shown here, delineating cryptic genetic species in field studies using verified genetic markers is essential to uncover ecological and evolutionary processes that are hidden by analyses at the genus, morphological, or trait level. For example, all the Pocillopora lineages we focus on are thought to reproduce via broadcast spawning (Schmidt‐Roach et al. 2012 ). Reproductive mode is a common way to categorize corals for the purpose of predicting changes in community structure (Hughes et al. 2019 ). However, not all of the common lineages sampled here responded to bleaching in 2019 in the same way. Without delineating cryptic genetic species in the field, studies will be limited to descriptions of decline in percent cover at coarse taxonomic levels. The results presented here have important implications for identifying ecological portfolios (Webster et al. 2017 ) at Moorea and surrounding islands, where diversity in environmental conditions, habitat types, phenotypes, and genotypes would buffer the capacity for Pocillopora to absorb impacts of climate change (van Woesik 2017 , Webster et al. 2017 ). Since at least the 1970s, coral cover at Moorea has seen large declines and recovery on multiple occasions, where coral cover has become increasingly dominated by corals in the genus Pocillopora after each recovery period (Berumen and Pratchett 2006 , Adjeroud et al. 2018 , Holbrook et al. 2018 , McWilliam et al. 2020 ). The presence of multiple co‐occurring cryptic species of Pocillopora that differ in bleaching mortality, and differences in heat accumulation over kilometer scales driven by differences in local oceanographic phenomena, may underlie the seemingly unusual patterns of resilience known for Pocillopora at Moorea and perhaps explain at least some of the success of Pocillopora relative to Acropora (Edmunds 2017 , Adjeroud et al. 2018 , Holbrook et al. 2018 )."
} | 5,110 |
38029083 | PMC10658910 | pmc | 1,090 | {
"abstract": "Production of methane by methanogenic archaea, or methanogens, in the rumen of ruminants is a thermodynamic necessity for microbial conversion of feed to volatile fatty acids, which are essential nutrients for the animals. On the other hand, methane is a greenhouse gas and its production causes energy loss for the animal. Accordingly, there are ongoing efforts toward developing effective strategies for mitigating methane emissions from ruminant livestock that require a detailed understanding of the diversity and ecophysiology of rumen methanogens. Rumen methanogens evolved from free-living autotrophic ancestors through genome streamlining involving gene loss and acquisition. The process yielded an oligotrophic lifestyle, and metabolically efficient and ecologically adapted descendants. This specialization poses serious challenges to the efforts of obtaining axenic cultures of rumen methanogens, and consequently, the information on their physiological properties remains in most part inferred from those of their non-rumen representatives. This review presents the current knowledge of rumen methanogens and their metabolic contributions to enteric methane production. It also identifies the respective critical gaps that need to be filled for aiding the efforts to mitigate methane emission from livestock operations and at the same time increasing the productivity in this critical agriculture sector.",
"introduction": "1. Introduction Livestock production in the US emitted close to 200 million metric tons of CO 2 -equivalent (MMT CO 2 –e) of methane, mainly originating from enteric fermentation in beef and dairy cattle representing 72 and 25% of emissions from livestock, respectively ( EPA, 2022 ). The corresponding value at the global scale is approximately 2,500 MMT CO 2 -e ( EPA, 2023a ), and it is estimated to rise substantially due to an increase in demand for milk and meat to feed the 9.8 billion global population by 2050 ( FAO, 2018 ; Henchion et al., 2021 ). Methane is 28 times more potent greenhouse gas (GHG) with a much shorter shelf-life than CO 2 ( EPA, 2023b ). In the rumen, it is produced as a by-product of microbial fermentation, and methanogenic archaea or methanogens are the only microorganisms that are known to produce methane anaerobically ( Smith and Hungate, 1958 ; Ramanathan et al., 1985 ; Wolfe, 1992 ). In addition to contributing to global warming, methane emission from the rumen causes a loss of 2–12% of the energy provided by the feed ( Johnson and Johnson, 1995 ; Janssen, 2010 ). Hence, a reduction of methane emission from cattle would have a greater near-term contribution to the effort toward mitigating global climate change and improving animal productivity ( Janssen, 2010 ; Beauchemin et al., 2020 ). For the above-mentioned importance, the metabolism of rumen microbes including methanogens has been investigated for almost eight decades ( Barker, 1936 ; Elsden, 1945 ; Hungate, 1950 ; Beijer, 1952 ; Hungate, 1966 ; Henderson et al., 2015 ; Seshadri et al., 2018 ). These studies yielded a plethora of basic and applied science information about rumen methanogens including their role in facilitating microbial fermentation in the rumen ( Hungate, 1966 ; Beauchemin et al., 2020 ). These details have been leveraged for developing tools for mitigating methane emission in the livestock industry and some of these can provide an average of 30% reduction in methane production with acceptable safety in both beef and dairy cattle ( Yu et al., 2021 ). However, the outcome varies greatly ( Patra et al., 2017 ; Arndt et al., 2022 ). What causes such variabilities? Which methanogens escape such intervention and how could one target them effectively? What factors drive the composition of a rumen methanogen community over another, spatially and temporally? Answering these questions requires a deeper understanding of the metabolic diversity and in situ physiology of rumen methanogens, which sorely remains incomplete even after close to eight decades of interrogation. It is because the current knowledge base for this field has mostly been built on studies with pure culture isolates from the rumen, which are a few, and inferences from the properties of non-rumen methanogen isolates ( Jeyanathan, 2010 ; Seshadri et al., 2018 ). The technical hurdles of working with strict anaerobes and the absence of clues to specific auxotrophies have limited the isolation efforts, which could have allowed useful in vitro studies. The culture-independent approaches leveraging high throughput omics are beginning to fill the above-mentioned gap in terms of phylogenetic diversity and metabolic potentials. The discovery of species from the Methanomassiliicoccales order that provide an additional route for removing the hydrogen-based thermodynamic block on ruminal fermentation ( Borrel et al., 2013 ) and key genomic features that allow rumen methanogens to associate with other organisms ( Leahy et al., 2010 ; Ng et al., 2016 ) and battle the toxicity of plant product (i.e., tannin) are examples of such advances ( Kelly et al., 2016c ; Loh et al., 2020 ). However, the absence of information on the metabolic and physiological properties of individual rumen methanogens that are generally obtained from studies on pure culture isolates or even low complexity enrichments has prevented making a clear sense of physiological data originating from in vivo or whole animal-based measurements. This review presents a summary and analysis of the past and evolving knowledge of rumen methanogens ( Figure 1 ) including the ongoing and upcoming research that would fill the above-mentioned gaps and help the efforts to mitigate enteric methane emissions while bringing sustainability to the livestock industry. Figure 1 Ecophysiology and metabolic adaptation of rumen methanogens. A schematic diagram illustrating functional roles of methanogens that facilitate the continuation of rumen microbial fermentation by removal of H 2 from microbial fermentation to generate methane. In the process, methanogens interact with different functional guilds via syntrophic associations and cross-feedings. Uptakes of nutrients and genetic materials via horizontal gene transfer (HGT) are shaping rumen methanogen metabolism, physiology, and lifestyle resulting in better adaptations and competitiveness in the rumen environment. Interactions between methanogens and other rumen microbiota are diverse and complex where methanogens are found as free-living, in a physical association or syntrophic relationship with other microbes, attach to the rumen epithelial cells as part of rumen epimural community, or ecto−/endosymbiosis with protozoa (right panel). Metabolic adaptation of methanogens in rumen environment (lower panel) results in loss of biosynthetic genes generating oligotrophy, acquisition of new functions through HGT, and physiological adaptation to methanogenic substrate fluctuations in the rumen (i.e., high and low \n p H 2 \n conditions following feeding) that have significant impacts on the emergence of CO 2 - and methyl-reducing hydrogenotrophs (i.e., K s and the deployment of different Mcr isozymes)."
} | 1,796 |
25505375 | null | s2 | 1,091 | {
"abstract": "The need for dynamic, elastomeric polymeric biomaterials remains high, with few options with tunable control of mechanical properties, and environmental responses. Yet the diversity of these types of protein polymers pursued for biomaterials-related needs remains limited. Robust high-throughput synthesis and characterization methods will address the need to expand options for protein-polymers for a range of applications. To address this need, a combinatorial library approach with high throughput screening is used to select specific examples of dynamic protein silk-elastin-like polypeptides (SELPs) with unique stimuli responsive features, including tensile strength, and adhesion. Using this approach 64 different SELPs with different sequences and molecular weights are selected out of over 2,000 recombinant E. coli colonies. New understanding of sequence-function relationships with this family of proteins is gained through this combinatorial-screening approach and can provide a guide to future library designs. Further, this approach yields new families of SELPs to match specific material functions."
} | 278 |
28966438 | null | s2 | 1,092 | {
"abstract": "Algae are ubiquitous in the marine environment, and the ways in which they interact with bacteria are of particular interest in marine ecology field. The interactions between primary producers and bacteria impact the physiology of both partners, alter the chemistry of their environment, and shape microbial diversity. Although algal-bacterial interactions are well known and studied, information regarding the chemical-ecological role of this relationship remains limited, particularly with respect to quorum sensing (QS), which is a system of stimuli and response correlated to population density. In the microbial biosphere, QS is pivotal in driving community structure and regulating behavioral ecology, including biofilm formation, virulence, antibiotic resistance, swarming motility, and secondary metabolite production. Many marine habitats, such as the phycosphere, harbour diverse populations of microorganisms and various signal languages (such as QS-based autoinducers). QS-mediated interactions widely influence algal-bacterial symbiotic relationships, which in turn determine community organization, population structure, and ecosystem functioning. Understanding infochemicals-mediated ecological processes may shed light on the symbiotic interactions between algae host and associated microbes. In this review, we summarize current achievements about how QS modulates microbial behavior, affects symbiotic relationships, and regulates phytoplankton chemical ecological processes. Additionally, we present an overview of QS-modulated co-evolutionary relationships between algae and bacterioplankton, and consider the potential applications and future perspectives of QS."
} | 420 |
26834315 | null | s2 | 1,093 | {
"abstract": "Supramolecular hydrogels have the advantages of stimuli responsiveness and self-healing compared to covalently crosslinked hydrogels. However, the existing supramolecular hydrogels are usually poor in mechanical properties especially in extensibility. In addition, these supramolecular hydrogels need a long self-healing time and have low self-healing efficiency. In this manuscript, we report a novel strategy to develop highly extensible and fast self-healing supramolecular hydrogels by using pre-coordinated mussel-inspired catechol-Fe"
} | 134 |
35650184 | PMC9160091 | pmc | 1,095 | {
"abstract": "Carbon-negative synthesis of biochemical products has the potential to mitigate global CO 2 emissions. An attractive route to do this is the reverse β-oxidation (r-BOX) pathway coupled to the Wood-Ljungdahl pathway. Here, we optimize and implement r-BOX for the synthesis of C4-C6 acids and alcohols. With a high-throughput in vitro prototyping workflow, we screen 762 unique pathway combinations using cell-free extracts tailored for r-BOX to identify enzyme sets for enhanced product selectivity. Implementation of these pathways into Escherichia coli generates designer strains for the selective production of butanoic acid (4.9 ± 0.1 gL −1 ), as well as hexanoic acid (3.06 ± 0.03 gL −1 ) and 1-hexanol (1.0 ± 0.1 gL −1 ) at the best performance reported to date in this bacterium. We also generate Clostridium autoethanogenum strains able to produce 1-hexanol from syngas, achieving a titer of 0.26 gL −1 in a 1.5 L continuous fermentation. Our strategy enables optimization of r-BOX derived products for biomanufacturing and industrial biotechnology.",
"introduction": "Introduction Current extractive industrialization processes annually release 9 gigatons (Gt) of CO 2 (total anthropogenic CO 2 emission >24 Gt), while only fixing approximately 120 megatons 1 . This considerable imbalance in the global carbon cycle is a leading cause of climate change, motivating the need for both new CO 2 waste gas recycling strategies and carbon, as well as energy-efficient routes to chemicals and materials 2 . Industrial biotechnology is one promising carbon-recycling approach to address this need. For example, Clostridium autoethanogenum has emerged as a cellular factory to convert carbon oxides in the atmosphere (e.g., CO, CO 2 ) and green hydrogen into sustainable products, like ethanol 3 , 4 . Unfortunately, designing, building, and optimizing biosynthetic pathways in cellular factories remains a complex and formidable challenge. A key issue is that past metabolic engineering efforts have chiefly focused on linear heterologous pathways that limit the co-development of multiple biochemical products. Cyclic and iterative pathways offer the unique advantage of providing access to hundreds of molecules with different chemistries and chain lengths from one core pathway. Reverse β-oxidation (r-BOX) is one such cyclic pathway that is highly modular and has been demonstrated as a promising route to many sustainable molecules with different chemistries and carbon chain length 5 , 6 . For example, r-BOX has been successfully used in Escherichia coli to produce high titers of C4-C10 saturated 7 and α,β-unsaturated carboxylic acids 8 (Fig. 1A ), adipic acid, or tiglic acid 9 . In addition to E. coli , r-BOX has been shown to work in a range of platforms that offer process advantages and can access alternative feedstocks 5 , 10 . While promising for manufacturing diverse products from a single iterative pathway, r-BOX studies have mostly shown the production of biochemical mixtures. This necessitates extensive downstream purification 11 and presents a major challenge of r-BOX in engineering and controlling product selectivity. Selection of unique combinations of enzyme homologs and expression levels is required to achieve product selectivity, adding development time, cost, and optimization bottlenecks. Fig. 1 Establishing a cell-free r-BOX platform. A General scheme of r-BOX. A thiolase (1) adds an acetyl-CoA unit every cycle to a growing acyl-CoA chain. The 3-keto group is then successively reduced (2), dehydrated (3) and reduced (4) again to form a two-carbon elongated acyl-CoA group. Termination enzymes (5) generate products from the acetyl-CoA intermediates of this iterative process. B Initial screen of r-BOX enzymes to establish a functional base case for in vitro hexanoic acid synthesis. All assays contained 0.3 μM of each cycle enzyme and 0.15 μM Ec_tesA. Conditions (i) Ec_fadB, Eg_ter; (ii) Ck_hbd1, Ca_crt, Eg_ter; (iii) Ec_fadB, Td_ter; (iv) Ck_hbd1, Ca_crt, Td_ter. The highest hexanoic acid production was observed for Cn_bktB, Ck_hbd1, Ca_crt, Td_ter, which was used in all future optimizations. C Thioesterase background activity from the extract strain used for iPROBE was assessed by omitting thioesterase Ec_tesA. D Carbon utilization assays using the base case of r-BOX enzymes with 120 mM 13 C glucose. The assay contains approximately 120 mM 12 C acetate from extract preparation and addition as CFME salts. % of 13 C label in the hexanoic acid product was determined by GC-MS. 13 C labeled carbons are shown in black, 12 C labeled carbons in gray. E Buffer, concentration of glycolytic enzymes, and cofactor optimization for the initial r-BOX set. Detailed results and full information on enzyme abbreviations in Source Data file. Data represent n = 3 independent experiments, with the standard deviation and mean shown. We have recently established a cell-free framework for in vitro prototyping and rapid optimization of biosynthetic enzymes (iPROBE) 12 . This workflow is able to accelerate the testing of biochemical pathways from months-to-years in non-model organisms (e.g., C. autoethanogenum ), or weeks in model organisms (e.g., E. coli ), to just a few days, while also increasing the number of pathways that can be tested 12 . The platform uses combinatorial assembly of enzyme homologs produced in vitro using cell extract and DNA templates to study and tune pathway performance. For example, iPROBE allowed us to screen 54 different linear pathways for 3-hydroxybutyrate production, 205 variants of a butanol pathway, and 580 unique linear pathway combinations for limonene synthesis starting with glucose as a carbon source 13 – 17 . iPROBE additionally enabled the optimization of 15 competing reactions for acetone biosynthesis in C. autoethanogenum 18 and pathway performance in each study correlated well with in vivo yields. Adapting iPROBE to study cyclic and iterative pathways like r-BOX could allow engineering product specificity but has not previously been attempted. Here, we adapt iPROBE to optimize r-BOX for the specific production of butanol, butanoic acid, hexanol, and hexanoic acid. In addition, we develop an automated liquid-handling workflow, allowing us to screen 440 unique enzyme combinations and 322 assay conditions. A key feature of the workflow was the application of high-throughput characterization of CoA metabolite concentrations in reactions containing variable sets of r-BOX enzymes. We use self-assembled monolayers for matrix-assisted laser desorption/ionization-mass spectrometry (SAMDI-MS) for this analysis. 17 We show that the source strain used for iPROBE extracts influences the yields and specificity of the screened pathways and use previous in vivo optimizations in E. coli to find an optimal strain for r-BOX cell-free prototyping 9 , 11 . We identify specific pathways for all four target products and demonstrate—for the first time—that pathway performance correlates well across three platforms: a cell-free system, a heterotrophic model organism ( E. coli ), and an autotrophic organism ( C. autoethanogenum ) capable of using syngas as the sole carbon and energy source. We were able to implement our optimized r-BOX pathways and generate strains with the highest to date reported titers of 3.06 ± 0.03 gL −1 hexanoic acid and 1.0 ± 0.1 gL −1 hexanol in E. coli , and 0.26 gL −1 hexanol in C. autoethanogenum . The direct correlation between prototyping in vitro and in vivo implementation in two metabolically different organisms forms a new blueprint for the generation and optimization of biochemical pathways for metabolic engineering and synthetic biology.",
"discussion": "Discussion Here, we use high-throughput in vitro enzyme prototyping (iPROBE) to optimize r-BOX product specificity and compare r-BOX pathway performance across three different platforms—a cell-free system, E. coli as a heterotrophic model organism, and autotrophic C. autoethanogenum capable of using syngas as the sole carbon and energy source. Using the in vitro iPROBE system as a guide, we were able to significantly improve r-BOX production, achieving the highest reported titers and/or selectivities towards C6 r-BOX products in E. coli and demonstrating r-BOX for the first time in C. autoethanogenum from syngas 11 , 31 , 32 . Our approach has several key features. First, we automated iPROBE using an acoustic liquid handling robot to increase the throughput power of this approach. We used this automated approach to show that extracts from knockout strains are critical to increase starting substrate pools (i.e., acetyl-CoA), reduce side product formation, control carbon flux (i.e., enable theoretical yield calculations), and remove competing background reactivities. The developed JST07 extract can be adapted for use for iPROBE prototyping of a wide range of biosynthetic pathways that use CoA ester intermediates and start with acetyl-CoA or pyruvate (e.g., isoprenoids, fatty acids, cannabinoids, polyketides). Additionally, it has the potential to enable faster optimization of recently established CoA ester-dependent new-to-nature/synthetic pathways like CETCH 33 and FORCE 34 . Second, by using this approach, we were able to tailor r-BOX to produce a single product at high selectivity. As is generally the case for iterative pathways, r-BOX typically generates a product mixture 11 , 31 , 32 . Here we identified designs that specifically favor either C4 or C6 products with the termination step playing a key role. Another key step in the r-BOX pathway is the initial priming step, which is a thermodynamically challenging carbon-carbon bond forming thiolase reaction. Efficient removal of the thiolase product by HBD additionally helps to overcome this bottleneck. Third, we compared designs across three platforms. Despite the distinct differences between the three systems (e.g., no cofactor, salt, inorganic phosphate homeostasis in vitro, low activity of the bifunctional alcohol-aldehyde dehydrogenases AdhE in E. coli , utilization of the oxygen-sensitive ferredoxin dependent Bcd-Etf and Aor enzymes in C. autoethanogenum ), we show that iPROBE can accelerate cellular design especially when prototyping complex pathways for non-model organisms. The r-BOX variants identified as good candidates for butanol, butanoic acid, hexanol, and hexanoic acid synthesis translated well to the two in vivo systems and enabled the E. coli strains with the highest and most specific demonstrated titers of hexanol and hexanoic acid products achieved in E. coli to date as well as the synthesis of hexanol in the acetogenic C. autoethanogenum . This sets a good starting point for applying iPROBE to other metabolic pathways, while keeping in mind that inherent differences between in vitro and in vivo systems may not always allow direct correlation across these systems. For example, anaerobicity of C. autoethanogenum and dependence on oxygen-sensitive enzymes (that cannot be prototyped in in vitro systems so far). Looking forward, our strategy of in vitro prototyping, optimization in model organisms, and implementation into production hosts as well as the observed correlations between the systems can be widely utilized to reduce the development time of new industrial strains and enable optimization a wide range of r-BOX derived products."
} | 2,848 |
25709609 | PMC4285865 | pmc | 1,096 | {
"abstract": "Hyperaccumulators are plants that can extract heavy metal ions from the soil and translocate those ions to the shoots, where they are sequestered and detoxified. Hyperaccumulation depends not only on the availability of mobilized metal ions in the soil, but also on the enhanced activity of metal transporters and metal chelators which may be provided by the plant or its associated microbes. The rhizobiome is captured by plant root exudates from the complex microbial community in the soil, and may colonize the root surface or infiltrate the root cortex. This community can increase the root surface area by inducing hairy root proliferation. It may also increase the solubility of metals in the rhizosphere and promote the uptake of soluble metals by the plant. The bacterial rhizobiome , a subset of specialized microorganisms that colonize the plant rhizosphere and endosphere, makes an important contribution to the hyperaccumulator phenotype. In this review, we discuss classic and more recent tools that are used to study the interactions between hyperaccumulators and the bacterial rhizobiome , and consider future perspectives based on the use of omics analysis and microscopy to study plant metabolism in the context of metal accumulation. Recent data suggest that metal-resistant bacteria isolated from the hyperaccumulator rhizosphere and endosphere could be useful in applications such as phytoextraction and phytoremediation, although more research is required to determine whether such properties can be transferred successfully to non-accumulator species.",
"conclusion": "Conclusions The composition of the bacterial rhizobiome coupled with the genomic, transcriptomic, and proteomic analysis of plant–microbe interactions may help us to understand in more detail the associations between hyperaccumulators and the surrounding bacterial communities of the endosphere and rhizosphere. It will be interesting to compare the rhizobiome of different facultative metallophytes, such as N. caerulescens adapted to grow in different metalliferous and non-metalliferous soils (Pollard et al., 2014 ), because this will help to isolate the bacteria that contribute to the hyperaccumulator phenotype. However, the rhizosphere is a dynamic environment with the community undergoing rapid spatiotemporal changes in response to external factors. The metabolic profiling of microbial colonies by in situ mass spectrometry (Traxler and Kolter, 2012 ) should therefore be integrated with omics-based profiling methods in a systems biology approach (Figure 1F ) to facilitate the investigation of interactions between the rhizobiome and hyperaccumulator plants, thus providing an advanced toolkit for phytotechnology applications. Conflict of interest statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest."
} | 733 |
39413365 | PMC11615781 | pmc | 1,097 | {
"abstract": "Abstract Neuromorphic computing, a promising solution to the von Neumann bottleneck, is paving the way for the development of next‐generation computing and sensing systems. Axon‐multisynapse systems enable the execution of sophisticated tasks, making them not only desirable but essential for future applications in this field. Anisotropic materials, which have different properties in different directions, are being used to create artificial synapses that can mimic the functions of biological axon‐multisynapse systems. However, the restricted variety and unadjustable conductive ratio limit their applications. Here, it is shown that anisotropic artificial synapses can be achieved on isotropic materials with externally localized doping via electron beam irradiation (EBI) and purposefully induced trap sites. By employing the synapses along different directions, artificial neural networks (ANNs) are constructed to accomplish variable neuromorphic tasks with optimized performance. The localized doping method expands the axon‐multisynapse device family, illustrating that this approach has tremendous potentials in next‐generation computing and sensing systems.",
"conclusion": "3 Conclusion In summary, we have successfully induced anisotropic properties in an isotropic material using a localized EBI method. The anisotropic transistor composed 2D MoS 2 nanoflakes is fabricated after localized EBI treatment, exhibiting diverse responses when stimulated by the same either optical or electrical stimulus along different directions. Benefiting from modulation of area doses during EBI process, the multiterminal synaptic transistor with adjustable anisotropy is further realized. Axon−multisynapse system with connection heterogeneity in neural networks is simulated. Artificial synapses in axon−multisynapse system demonstrate variable synaptic plasticity along different directions. Eventually, based on the synaptic plasticity of multiterminal synaptic transistor, the ANN and ONN are constructed to execute neuromorphic tasks, including image recognition and colored‐digit recognition. Responding to various neuromorphic tasks, the synapses in variable directions are employed to enhance recognition performance. These findings provide feasible approaches to simply structure and accomplish complicated functions for neuromorphic computing system.",
"introduction": "1 Introduction In the era of big data, conventional computing systems, constrained by von Neumann architecture, face challenges in energy efficiency due to the segregation of computing and storage units. [ \n \n 1 \n , \n 2 \n , \n 3 \n \n ] Inspired by the brain, neuromorphic computing offers a promising and energy‐efficient approach for developing advanced intelligent systems. [ \n \n 4 \n , \n 5 \n , \n 6 \n \n ] The axon‐multisynapse architecture, where each axon connects to multiple synapses, plays a crucial role in neural information processing and complicated brain function within the neural network. [ \n \n 7 \n , \n 8 \n \n ] This architecture enables neurons to transmit information simultaneously to multiple target neurons or regions, facilitating the integration and coordination of diverse neural signals for more efficient processing. Connection heterogeneity induced by the intrinsic heterogeneity in axon‐multisynapse system allows diverse responses to the same signal, increasing the structural complexity and functional diversity of neural network. [ \n \n 9 \n , \n 10 \n , \n 11 \n , \n 12 \n \n ] This diversity enables sophisticated tasks to be accomplished through cooperation with other properties in brain such as perception, motor control, and cognitive processes. Furthermore, connection heterogeneity optimizes the performance of synaptic plasticity, improving the adaptability and learning capacity of neural networks in responding to dynamic external environments. [ \n \n 13 \n , \n 14 \n , \n 15 \n \n ] Thus, realizing artificial synapses with connection heterogeneity is a crucial step for achieving neuromorphic computing with high complexity and realizing sophisticated brain functions. Recent developments in axon‐multisynapse system based on anisotropic 2D materials exhibit its great potential to mimic the human brain behaviors. Tian et al. first achieved the axon‐multisynapse network on anisotropy 2D black phosphorus with compact device structure (BP). [ \n \n 9 \n \n ] Qin et al. further illustrated that synaptic transistor based on anisotropic selenium (t‐Se) can act as a filter with low power consumption. [ \n \n 10 \n \n ] These works provide feasible approaches to construct axon‐multisynapse system with desired connection heterogeneity. However, in the context of prior research, the construction of the axon‐multisynapse system has predominantly depended on the inherent anisotropy of the channel material. The inflexible anisotropy ratio and band structure have imposed constraints on its applicability. Furthermore, a significant proportion of anisotropic 2D materials either lack scalability in their preparation or exhibit instability in atmospheric conditions. [ \n \n 16 \n , \n 17 \n \n ] These challenges have curtailed the potential applications of artificial anisotropic synapses. In this work, we employ a localized electron beam irradiation (EBI) technique to induce anisotropic properties in an isotropic MoS 2 synaptic device. The precisely controlled n‐type doping, a consequence of localized EBI, results in the MoS 2 ‐based transistor demonstrating a variety of photoelectric characteristics in response to identical stimuli along different directions. With this approach, we can precisely tune the anisotropic synaptic behavior by modulating the EBI intensity. We further investigate connection heterogeneity in the MoS 2 ‐based multiterminal synaptic transistor to realize an axon‐multisynapse system. By harnessing the synaptic plasticity in our artificial anisotropic synapse, we successfully execute tasks of image recognition and colored‐digit identification through the use of artificial neural networks (ANNs). The axon‐multisynapse system with varying orientations is fine‐tuned to maximize both the recognition speed and rate for their respective tasks. Our results offer a universal approach for the design and fabrication of axon‐multisynapse system.",
"discussion": "2 Results and Discussion 2.1 Memory Features and Synaptic Functions of UVO‐Treated MoS 2 Transistor We first investigate the electrical and optical memory behavior of the MoS 2 ‐based synaptic device with axon‐singlesynapse structure, which will be turned into axon‐multisynapse later in the manuscript. Figure \n \n 1 a presents a schematic structure of a back‐gate transistor constructed from MoS 2 . The introduction of trap sites on the MoS 2 surface is facilitated through the use of ultraviolet/ozone (UVO). Figure S1 (Supporting Information) illustrates the Raman characteristics of MoS 2 films before and after the UVO treatment. Figure 1b presents the double sweep transfer curves of the MoS 2 ‐based transistor with various swept ranges of gate voltage ( V gs \n ) after UVO treatment. The threshold voltage ( V th \n ) for the forward sweep is lower as compared to reverse sweep, resulting in clockwise hysteresis. The hysteresis window can be tuned from 13 to 64 V with the increase of maximum V gs \n from 20 to 50 V. This suggests that the charge trapping behavior can be modulated by adjusting the maximum value of the V gs \n . Figure S2 (Supporting Information) explores double sweep transfer characteristics of the device with different irradiation time of UVO treatment. Figure S3 (Supporting Information) presents the transfer characteristics of UVO‐treated MoS 2 transistor under various drain voltage ( V ds \n ) and corresponding output characteristics under varying V gs \n . Statistic characteristics of 30 devices are given in Figure S4 (Supporting Information). Continuous changes of drain current under different gate pulse and retention properties are given in Figure S5 (Supporting Information). Subsequently, we investigate the optical modulation of the UVO‐treated MoS 2 transistor under illumination with wavelengths of 650, 480, and 400 nm (30 mW cm −2 ). Figure 1c shows the photoresponse of MoS 2 transistor during a 110 s illumination period, followed by a 120 s post‐illumination recovery phase. Under continuous illumination, the drain current exhibits a gradual increase over time, indicative of a current excitation characteristic. The drain current increases from 40 nA to ≈ 97, 129, and 195 nA under the V ds \n of 0.1 V, corresponding to light wavelengths of 650 nm, 480 nm and 400 nm, respectively. Then the light‐induced current will drop steeply and stabilize at ≈ 63 (650 nm), 77 (480 nm), and 118 nA (400 nm) after light illumination, indicating the nonvolatility responding to optical signals. The memory characteristics of the MoS 2 ‐based transistor is attributed to the trap sites generated after UVO treatment. [ \n \n 18 \n , \n 19 \n \n ] Figure S6 (Supporting Information) elucidates the band diagram and the underlying mechanism of this memory behavior. Figure 1 The memory properties and synaptic behaviors of MoS 2 ‐based transistor after UVO treatment. a) Structure schematic of the MoS 2 ‐based transistor. b) Transfer characteristics of the MoS 2 ‐based transistor with the different V gs‐max \n (the maximum value of V gs \n sweeping range) at V ds \n of 0.5 V. c) The different responses of drain current under illumination with different wavelengths. d) Schematic illustration of the biological synapse and the MoS 2 ‐based artificial synaptic device. The V ds \n is set as 0.1 V. e) EPSC generated under negative pulses with different amplitudes of −10, −20, −30, and −40 V. f) IPSC generated under positive pulses with different amplitudes of 10, 20, 30, and 40 V, respectively. g) Reproducible potentiation and depression processes under continuous negative pulses and positive pulses. h) The fitted nonlinearity coefficient of potentiation and depression processes. i) The change of PSC under optical and electric signals. The drain electrode is applied reading pulses of 0.1 V with a time interval of 3 s. The distinct memory behaviors exhibited by the UVO‐treated MoS 2 transistor inspire further investigation into the artificial synaptic characteristics of this device. The artificial synapse mimics biological synapses [ \n \n 20 \n , \n 21 \n , \n 22 \n \n ] by generating a corresponding postsynaptic current (PSC) in response to pulse signals applied at the gate as presynaptic stimulation (Figure 1d ). Figure 1e,f illustrate the synaptic response to gate pulses of varying amplitudes. The MoS 2 ‐based artificial synapses simulate excitatory PSC (EPSC) under negative presynaptic signals and inhibitory PSC (IPSC) under positive ones. Figure S7a,b (Supporting Information) show EPSC behaviors in response to pulses with different widths and frequencies. Continuous electrical pulse trains are used to modulate synaptic weight in Figure 1g . The PSC is progressively enhanced or suppressed under negative or positive pulse trains (Figure S7c , Supporting Information), mirroring biological long‐term potentiation (LTP) and long‐term depression (LTD), respectively. The PSC is adjusted over five cycles of negative and positive pulse sequences, demonstrating the high repeatability of these artificial synapses. The nonlinearity factor (NLF) [ \n \n 23 \n , \n 24 \n \n ] is fitted according to the behavioral model (See Note S1 , Supporting Information), as indicated in Figure 1h . The MoS 2 ‐based artificial synapse presents the NLF of 3.19 for LTP and 4.34 for LTD. The retention property after pulse sequence is investigated in Figure S7d (Supporting Information). Based on the long‐term plasticity of MoS 2 artificial synapses, the ANN model is constructed to conduct pattern recognition, as shown in Figures S8 and S9 (Supporting Information). Figure 1i presents the wavelength dependent optical potentiation and electrical habitation characteristics of the MoS 2 synaptic device. During the illumination, drain current increases more quickly with a shorter wavelength. The light‐induced PSC are 337.64%, 183.79%, and 119.96% for 400, 480, and 650 nm, respectively. The light with shorter wavelength generates more additional electrons into the channel, larger electrical pulse is needed to reset the PSC to initial state. In order to balance the potentiation and habitation behavior of the optical and electrical pulse, the electrical pulse amplitude is set to be 25 (400 nm), 20 (480 nm), and 18 V (650 nm). These results illustrate the potential of the UVO‐treated MoS 2 transistor in neuromorphic computing systems, paving the way for the development of axon‐multisynapse system later in this study. 2.2 Localized EBI Method to Realize Anisotropic Conductivity We subsequently use localized EBI to achieve anisotropic conductivity of MoS 2 for axon‐multisynapse system. Figure \n \n 2 a illustrates the experimental scheme while Figure S10 (Supporting Information) details device fabrication procedures. A polymethyl methacrylate (PMMA) film is spin‐coated onto the device as a protective layer to prevent the deposition of impurities and electron‐beam‐generated reactions on the MoS 2 surface. Figure S11 (Supporting Information) presents the comparison of transfer curves of the devices before and after the spin‐coating process. Doping is accomplished by EBI treatment in e‐beam lithography system with an energy of 10 keV. [ \n \n 25 \n \n ] We first investigate the carrier concentration tuning ability of the EBI doping method. Figure 2b shows the transfer characteristics of MoS 2 transistor after EBI across the entire device. Initially, the MoS 2 ‐based transistor is undoped with a V th \n of ≈−6.93 V. V th \n gradually shifts negatively with increasing e‐beam dose. Eventually, V th \n shifts to −10.49 V and the overall current increases over 4 times after EBI with an area dose of 150 µC cm −2 . Further increasing the e‐beam dose to 200 µC cm −2 shows little influence to the carrier and current density. The carrier concentration n is calculated as: [ \n \n 26 \n \n ] \n \n (1) \n n = C 0 V g s − V t h q \n where C 0 \n is the gate capacitance per unit area and is 34.5 nF cm −2 for 100 nm SiO 2 , | q | = 1.6 × 10 −19 C is the elementary charge and V th \n is the threshold voltage. The carrier concentration of MoS 2 at the V gs \n of 0 V is tuned from 1.49 × 10 12 cm −2 to 2.37 × 10 12 cm −2 with 200 µC cm −2 EBI treatment. After being treated by EBI, the electron‐hole pairs are generated in the SiO 2 layer. Subsequently, electrons will be trapped at the interface between MoS 2 and SiO 2 . Holes will accumulate at the interface between Si and SiO 2 . As a result, an electrostatic field is building up to form the n‐type doping effect on the MoS 2 . The above results indicate that carrier concentration can be well controlled by the EBI treatment. Utilizing the precisely control of the EBI with e‐beam lithography, the EBI doping method not only allows for precisely control of the carrier concentration but also the doping area. We use photoluminescence (PL) and Raman mapping to image the pattern we predefined. Figure 2c shows PL mapping image of a stripe pattern drawn using the e‐beam lithography. The PL image clearly reveals a striped modulation resulting from the different localized carrier density. [ \n \n 27 \n , \n 28 \n \n ] Due to the resolution limitation of the PL mapping, the width of the stripe is limited to 1 µm. We remark that the feature size of this doping method can be even smaller due to the high resolution of the e‐beam lithography. The uniform FWHM of Raman mapping before and after EBI indicates that this doping method doesn't introduce defects or disorder into MoS 2 lattice (Figure S12 , Supporting Information). Figure 2 The anisotropic behaviors and mechanisms of MoS 2 ‐based transistor treated by localized EBI. a) Schematic of the device treatment procedures by electron beam. b) Transfer characteristics of primordial MoS 2 ‐based transistor after EBI treatment on the whole channel with different area doses at V ds \n of 0.5 V. c) Mapping image of PL peak center of the MoS 2 nanoflake after localized EBI. d) Schematic of the anisotropic multiterminal synaptic device irradiated by localized electron beam. Synapse 1 and synapse 2 are represented by devices along direction 1 and direction 2, respectively. The scale bar is 10 µm. The channel length and width are 6 and 3 µm, respectively. e) Transfer characteristics of the synapses 1 and 2 at the V ds \n of 0.5 V. f) Schematic diagrams to illustrate the n‐type doping induced by electron beam irradiation. g) Energy band diagrams of the MoS 2 transistor before and after EBI. Benefiting from the localized precisely modulation of MoS 2 carrier density, we achieve anisotropic characteristics by fabricating periodic highly conductive stripe [ \n \n 29 \n \n ] on isotropic MoS 2 device with EBI. As depicted in Figure 2d , a UVO‐treated MoS 2 transistor is subjected to a periodically localized EBI. The irradiated direction is denoted by the red line (1 µm line width). Figure 2e displays the double sweep transfer characteristics with an on/off ratio of ≈ 10 6 along direction 1 (synapse 1) and along direction 2 (synapse 2) after the localized EBI treatment with an area dose of 150 µC cm −2 . The UVO‐induced memory properties are preserved after EBI. Compared to synapse 1, the drain current of synapse 2 is 2 times higher than synapse 1, illustrating that anisotropic features are achieved by the periodically localized EBI. The mechanism of the anisotropic characteristics is proposed in Figure 2f . The regions exposed and unexposed to the electron beam exhibit distinct doping concentrations, which is caused by the localized electron‐hole pairs generated in the SiO 2 layer. [ \n \n 30 \n , \n 31 \n \n ] To further illustrate the anisotropic properties, the energy band diagrams are proposed in Figure 2g . After EBI treatment, n‐type doping will enable the Fermi level (E F ) shift closer to the conduction band in the irradiated area. In synapse 1, these two regions constitute periodically n‐n + junction, while in synapse 2, they are arranged in parallel. The highly conductive stripes increase the conductivity of synapse 2. However, the n‐n + junction and the less conductive unexposed area limit the conductivity of synapse 1. 2.3 Adjustable Anisotropy Ratio Achieved by Modulating EBI Intensity We further investigate the anisotropic electrical properties of MoS 2 ‐based multiterminal transistor with various EBI intensities. Figure \n \n 3 a,b shows the double sweep transfer characteristics along two directions with the increasing area doses from 0 to 200 µC cm −2 . For n‐channel MoS 2 device, electron trapped at absorptive sites is suppressed at large negative V gs \n , which refers to the forward sweep starting from negative side. For the forward sweep in Figure 3a , the V th \n shows a little negative shift of 1.23 V from ‐6.20 V to ‐7.43 V after 200 µC cm −2 EBI for synapse 1. During forward sweep, the traps are not occupied, and the conductivity is closer to the characteristics without charge trapping behavior. As discussed previously, the conductivity of synapse 1 is mainly determined by the unexposed area, although the EBI continuously increases the conductivity of the highly doped stripe, the high conductive region shows little influence to the conductivity of this direction. On the contrary, for the forward sweep in Figure 3b , the V th \n shows negative shift of 4.24 V from −6.44 to −10.68V for synapse 2 with the same EBI intensity which is much larger than synapse 1. As discussed previously, the conductivity of synaptic 2 is mainly determined by the exposed area. With the EBI intensity increasing, the conductivity of exposed area increases, along with the conductivity of synapse 2. Furthermore, the hysteresis window is 14.5 V for synapse 1 before EBI treatment and decreases to 13.8 V after EBI with an area dose of 200 µC cm −2 . In contrast, the hysteresis window of synapse 2 is reduced from 13 to 10 V under the same condition, demonstrating a more significant change compared to synapse 1. This anisotropic behavior is attributed to the nonuniform distribution of trap sites in the MoS 2 channel, which affects the memory characteristics of the MoS 2 ‐based transistor. EBI treatment induces n‐type doping, which results that trap sites are occupied by generated electrons. The variation in electron distribution between direction 1 and direction 2 results in different numbers of occupied traps. Figure 3c is the I – V characteristics of both synapses with different EBI intensities under 20 V V gs \n . The device shows a slight increase by only 0.87 µA for synapse 1. The increase is caused by the reduction of resistance in exposed area, and the small value is attributed to the large resistance of the unexposed area. The current increases from 4.13 to 11.43 µA by ≈2.77 times for synapse 2, which matches the transfer curve well. The drain current for both synapses under 0 V V gs \n of forward curves and reverse curves are plotted in Figure 3d , exhibiting the different tendencies with increasing dose. The current of synapse 1 shows little changes under all doses and the current of synapse 2 shows the highest value under 150 µC cm −2 . In particular, the reverse current of synapse 2 is enhancing from 2.42 −11 A to 9.79 −9 A at 0 V V gs \n . The current isn't increasing while the dose is further raised. The current ratio of these two directions and V th \n with different e‐beam doses are given in Figure 3e . The current ratio increases with the e‐beam dose. The current ratios range from 0.96 to 3.28 for forward curves and range from 1.24 to 406.40 for reverse curves. Figure 3f demonstrates the comparison of anisotropic MoS 2 ‐based transistor with other 2D anisotropic device, [ \n \n 9 \n , \n 10 \n , \n 11 \n , \n 32 \n , \n 33 \n , \n 34 \n , \n 35 \n , \n 36 \n , \n 37 \n , \n 38 \n , \n 39 \n , \n 40 \n , \n 41 \n , \n 42 \n \n ] exhibiting high on/off ratio and adjustable anisotropic ratio. Figure 3 The anisotropic properties of MoS 2 ‐based multiterminal transistor after EBI treatment with different area doses. Properties of the transfer characteristics of MoS 2 ‐based transistor after EBI treatment with different area doses at V ds \n of 0.5 V for a) synapse 1 and b) synapse 2. c) Output characteristics after EBI treatment with different area doses at V gs \n of 20 V for synapse 1 and synapse 2. d) Drain current extracted from transfer curves of both synapses under zero gate bias. e) The current ratio and threshold voltage of both synapses with different EBI doses. f) Comparison of anisotropic ratio and on/off ratio of 2D anisotropic device. 2.4 Simulation of Connection Heterogeneity in Axon‐Multisynapse System According to the anisotropic electrical properties of the MoS 2 ‐based multiterminal transistor after EBI, axon−multisynapse system is constructed to simulate the connection heterogeneity, which is significant feature to achieve diversity of neurological events. As shown in Figure \n \n 4 a , external environmental signals are perceived and then transmitted through the axon to the synapse. The identical signal from the axon generates various responses at postsynapses because of connection heterogeneity. These responses are eventually processed to accomplish complicated functions like perception, learning, memory, oblivion, and so forth along with other properties in the brain. Before measurement, both synapse 1 and synapse 2 are adjusted to the respective initial state by applying the same sequence of 50 pulses with the amplitude 20 V. Figure S13 (Supporting Information) presents the anisotropic EPSC and IPSC responses of synapse 1 and synapse 2. Figure 4b demonstrates the various changes of PSC (ΔPSC/PSC) along directions 1 and 2 under negative and positive pulse trains. After the same 20 negative pulses, the change of PSC will gradually increase to 103.98% in synapse 1 and 28.11% in synapse 2. Then PSC of both synapse 1 and synapse 2 is restored to the initial state under 20 positive pulses. Figure S14 (Supporting Information) shows corresponding fitted nonlinearity coefficient of synapse 1 and synapse 2. These results indicate that the change of synaptic weight in synapse 1 is significantly larger than that in synapse 2. The anisotropic synaptic plasticity results from the difference in drain current along the two directions induced by EBI. Before applying stimulus signal at gate, the initial current of synapse 2 is larger comparing to that of synapse 1. Due to the identical voltage spike from gate, the number of trapped charges is approximately equal, resulting the similar ΔPSC but different change of PSC in the two artificial synapses. Figure 4 The anisotropic synaptic behaviors of the MoS 2 ‐based multiterminal transistor. a) Biological axon‐multisynapse system with connection heterogeneity. b) Conductance modulation of the device for synapse 1 and synapse 2. Anisotropic STDP behaviors for c) synapse 1 and d) synapse 2. e) The change of PSC under different pulse widths for synapse 1 and synapse 2. The anisotropic change of PSC under optical and electric signals of f) 650 nm and 16 V, g) 480 nm and 18 V, h) 400 nm and 22 V. Spiking timing dependent plasticity (STDP) is an essential learning characteristic of long‐term plasticity, in which the synaptic weight can be modulated by the time intervals ( Δt ) between pre‐ and post‐synaptic spikes. [ \n \n 43 \n \n ] As shown in Figure S15 (Supporting Information), spike pairs are designed to realize STDP characteristics. Figure 4c,d demonstrates the change of synaptic weight responding to the varied Δt in synapse 1 and synapse 2, which is consistent with the asymmetric Hebbian STDP rule in biology. According to the relationship between W Change \n and Δt in asymmetric Hebbian STDP (See Note S2 , Supporting Information), A + , A ‐ , τ + , and τ ‐ are ≈125.92%, ≈67.21%, ≈28.70 ms, and ≈30.26 ms for synapse 1 and ≈34.43%, ≈26.67%, ≈32.94 ms and ≈36.32 ms for synapse 2, which indates the dissimilarity of STDP behaviors in these two synapses. Figure 4e depicts the anisotropic PSC change reponding to variable pulse widths with fixed amplitude of ‐15 V. Equally, the change of drain current along direction 1 is larger than that along direction 2. The optical responses for the two synapses under illumination with various optical wavelengths are presented in Figure S16 (Supporting Information). Subsequently, photonic potentiation and electric habituation are further investigated for the anisotropic synapses, as shown in Figure 4f–h . Also, reading pulses of 0.1 V are applied to the drain electrode with a time interval of 3 s when illuminated. The PSC will increase gradually responding to optical stimulation with different wavelengths. And the change of PSC in synapse 1 is also larger comparing to synapse 2. After light illumination, positive voltage pulses with the amplitude of 16 V, 18 V, and 22 V are applied to restore the light‐induced PSC, responding to the light wavelengths of 650, 480, and 400 nm, respectively. The anisotropic response to optical stimulation also results from the differing initial currents in the two directions. When exposed to light, photoinduced carriers are generated in the MoS 2 ‐based channel. Synapses 1 and 2 exhibit comparable ΔPSC as a result of the MoS 2 layer's essentially constant quantity of photoinduced carriers in response to light illumination in both directions. Whereas due to the difference of initial currents in these two directions, the multiterminal synaptic transistor exhibits anisotropic change of PSC in response to light illumination. These findings provide the feasible approach constructing axon−multisynapse system to accomplish various complicated functions in brain. [ \n \n 44 \n , \n 45 \n , \n 46 \n \n ] \n 2.5 Image Recognition and Colored‐Digit Recognition Based on Anisotropic Synaptic Features Based on the anisotropic synaptic plasticity of axon−multisynapse system, an ANN is established to carry out image recognition, as illustrated in Figure \n \n 5 a . The 96 × 96 pixels recognized image is from STL‐10 dataset after greyscale processing. [ \n \n 47 \n \n ] Subsequently, feature maps are extracted through an 18‐layer convolutional neural network (ResNet18) [ \n \n 48 \n \n ] and then passed to the fully connected network for classification. Schematic diagram of this convolutional neural network is shown in Figure S17 (Supporting Information). In emulation process, weight updates are simulated by the LTP/LTD properties of synapse 1 and synapse 2, respectively. Figure 5b demonstrates the progression of the output image after different epochs during the learning process. It is observed that the ANN learned by the synaptic plasticity of synapse 1 displays more rapid learning effects, exhibiting the emergence of object outline after fewer learning epochs. Nevertheless, the ANN learned by the LTP/LTD features of synapse 2 shows improved accuracy, delivering clearer object details in the ultimate state. As presented in Figure 5c , the recognition accuracy of ANN based on synapse 1 is superior up to 56 epochs, but the recognition accuracy of ANN based on synapse 2 surpasses it after 56 epochs. These results suggest that the ANN based on synapse 1 enables faster image recognition, while the ANN based on synapse 2 achieves higher accuracy in the long run. Figure 5 Neuromorphic computing by the neural networks based on the axon−multisynapse system. a) The schematic of ANN to accomplish image recognition. b) Mapping synaptic weights of the input image for synapse 1 and synapse 2 after different learning epochs. c) The comparison of image recognition accuracy of synapse 1 and synapse 2. d) The schematic of ONN to recognize colored digits. e) MNIST handwritten datasets for colored‐digit recognition. Set 1 consists of colored digits without distracting background, and Set 2 is composed of colored digits with distracting background. f) The recognition accuracy of the synapse 1 and synapse 2 to recognize colored digits without distracting background. g) Mapping synaptic weights of colored digit for synapse 1 and synapse 2 at initial state and after 50 epochs trained by Set 1. h) The recognition accuracy of the synapse 1 and synapse 2 to recognize colored digits with distracting background. i) Mapping synaptic weights of colored digit for synapse 1 and synapse 2 at initial state and after 200 epochs trained by Set 2. Moreover, a dual‐layer optoelectronic neural network (ONN) with colored identification function is constructed to effectively accomplish pattern recognition tasks according to the photoelectric synaptic behaviors of anisotropic MoS 2 ‐based transistor, as presented in Figure 5d . The input layer is formed by 784 cone cell sets, and each set comprises 3 input neurons, exhibiting distinctive responses to red (R), blue (B), and purple (P) light. The colored lines are the artificial synapses with synaptic plasticity in various R/B/P weight regions. These sets of neurons generate varied synaptic dynamics in response to the input colored‐digit image and the 30 output neurons are used to export the results of R0‐R9, B0‐B9 and P0‐P9. MNIST handwritten dataset [ \n \n 49 \n \n ] is employed for colored‐digit recognition after some modifications, as depicted in Figure 5e . The first set consists of colored digits without distracting background (Set 1) and the other is composed of colored digits with distracting background (Set 2). More details about the datasets’ origin, the number of images, the classification categories, and preprocessing steps are illustrated in Figures S18 and S19 (Supporting Information). During simulation, colored backgrounds are considered as visual distractions to interfere with the recognition of digits. Figure 5f shows recognition accuracy of the ONN for colored digits in Set 1. Following 50 learning epochs, the ONN associated with synapse 1 achieves the recognition accuracy of 83.11%, whereas the recognition accuracy of ONN linked with synapse 2 is 69.75%. In Figure 5g , the visualization of synaptic weights for both synapse 1 and synapse 2 is presented. The clearer delineation of digital patterns after trained by Set 1 illustrates the improved proficiency in recognizing colored digits. Furthermore, the visualization of synaptic weights of synapse 1 exhibits greater clarity compared to synapse 2, suggesting better performance of the ONN associated with synapse 1 for colored digit recognition without visual distractions originating from the background. Meanwhile, the recognition of colored digits in Set 2 is also carried out. Figure 5h demonstrates the ONN connected by synapse 1 and synapse 2 acquires the accuracy rates of 69.86% and 85.31% after 200 learning epochs, respectively. Post‐training visualization with synaptic weights of synapse 1 and synapse 2 by Set 2 confirms the enhance of recognition capability for colored digits, as shown in Figure 5i . Particularly, the mapping image of synapse 2 displays more distinguishable digit patterns against the background, indicating the superior performance for the colored‐digit recognition with visual distractions from the background. The above results indicate that synapses in both directions of axon−multisynapse system have enormous potential for neuromorphic computing responding to different demands."
} | 8,406 |
37638314 | PMC10448768 | pmc | 1,098 | {
"abstract": "Recent developments in artificial neural networks and their learning algorithms have enabled new research directions in computer vision, language modeling, and neuroscience. Among various neural network algorithms, spiking neural networks (SNNs) are well-suited for understanding the behavior of biological neural circuits. In this work, we propose to guide the training of a sparse SNN in order to replace a sub-region of a cultured hippocampal network with limited hardware resources. To verify our approach with a realistic experimental setup, we record spikes of cultured hippocampal neurons with a microelectrode array ( in vitro ). The main focus of this work is to dynamically cut unimportant synapses during SNN training on the fly so that the model can be realized on resource-constrained hardware, e.g., implantable devices. To do so, we adopt a simple STDP learning rule to easily select important synapses that impact the quality of spike timing learning. By combining the STDP rule with online supervised learning, we can precisely predict the spike pattern of the cultured network in real-time. The reduction in the model complexity, i.e., the reduced number of connections, significantly reduces the required hardware resources, which is crucial in developing an implantable chip for the treatment of neurological disorders. In addition to the new learning algorithm, we prototype a sparse SNN hardware on a small FPGA with pipelined execution and parallel computing to verify the possibility of real-time replacement. As a result, we can replace a sub-region of the biological neural circuit within 22 μs using 2.5 × fewer hardware resources, i.e., by allowing 80% sparsity in the SNN model, compared to the fully-connected SNN model. With energy-efficient algorithms and hardware, this work presents an essential step toward real-time neuroprosthetic computation.",
"introduction": "1. Introduction In the field of systems neuroscience, studies on brain-machine interface (BMI) to replace semi-permanent functions of the human brain have been conducted for the treatment of neurological disorders or the use of neuroprosthetics (Zhang et al., 2020 ). For example, Song et al. ( 2007 ) have replaced the function of damaged hippocampal neurons with a mathematical model. The model predicts the electrical transmission between neurons so that similar electrical functionality can be artificially generated for damaged neurons. Recently, the authors in Hampson et al. ( 2018 ) have demonstrated that electrical stimulation to the biological neuron improves memory function in human subjects by predicting electrical transmission between neurons. However, most studies on BMI are based on traditional offline learning, making it challenging to actively cope with biological learning such as neuroplasticity. Moreover, the complexity of a mathematical model becomes intractable as a biological neural circuit to be replaced becomes larger (Song et al., 2016 ; She et al., 2022 ). Recently, artificial neural networks (ANNs) were used in explaining how the brain learns to perform perceptual and cognitive tasks (Richards et al., 2019 ). Specifically, brain-inspired spiking neural networks (SNNs) were utilized to understand activity patterns of neural circuits (Doborjeh et al., 2019 ; Lee et al., 2019 ; Kumarasinghe et al., 2021 ). Several recent studies have shown promising results on the capability of understanding a high-level brain functionality using SNN models, e.g., decoding neuro-muscular relationships (Kumarasinghe et al., 2021 ) or establishing a peripheral nervous system (Lee et al., 2019 ). Owing to the biological interpretability of the SNN model, it is even possible to mimic the microscopic behaviors of neural circuits, i.e., spike timings, firing rates, and burst patterns (Sun et al., 2010 ; Dominguez-Morales et al., 2021 ). In addition to biological plausibility, SNNs are energy efficient because they only compute when spikes are present ( event-driven ). Therefore, many studies have focused on improving the training accuracy of SNNs by introducing surrogate gradient descent (Fang et al., 2021 ; Zheng et al., 2021 ) or converting pre-trained ANNs into SNNs (Han and Roy, 2020 ; Han et al., 2020 ) even for tasks that are mainly used for ANNs such as computer vision. In addition to the algorithmic improvement, neuromorphic hardware chips have been designed, either analog (Benjamin et al., 2014 ) or digital (Akopyan et al., 2015 ; Davies et al., 2018 ), to process large-scale asynchronous SNNs efficiently. The main objective of neuromorphic hardware is to simulate the behavior of a large number of neurons in real-time with low power consumption. However, prior works suffer from the inability to support, or partially support, biologically plausible neuron models, or synaptic learning rules. To address these challenges, Lee et al. ( 2018 ) and Baek et al. ( 2019 ) have presented programmable SNN hardware that supports a wide range of neuron models and synaptic learning rules. Another approach is to use an FPGA platform, which allows flexible modification of neuron models and network structures by reconfiguring the hardware architecture (Cheung et al., 2016 ; Sripad et al., 2018 ). To efficiently process large-scale SNNs on multiple FPGA chips, SNN hardware with novel routing algorithms for energy-efficient computation of nonlinear neuron models have been proposed (Yang et al., 2019 ). Moreover, efficient implementations and algorithms have been proposed to support the mechanisms of various biological brain regions, such as the cerebellum and hippocampus, in large-scale SNNs (Yang et al., 2021a , b ). In short, SNNs can imitate biological neural networks (BNNs) more closely than other ANN counterparts with higher energy efficiency. Therefore, SNN is an ideal option in neuroprosthetics modeling to increase energy efficiency and biological plausibility (Li et al., 2021 ). In order to predict precise spike timings, several supervised learning rules have been proposed (Wang et al., 2020 ), and are typically divided into gradient descent learning (Bohte et al., 2002 ) and STDP-based learning (Ponulak and Kasiński, 2010 ). Although gradient descent learning can solve complex tasks, it is unsuitable for online learning because of its higher parameter dependence and slower learning speed than the synaptic plasticity learning (Lobo et al., 2020 ). STDP-based supervised learning is more suitable for online learning. To minimize the complexity of STDP-based supervised learning, we present a simple yet effective learning method called STDP-assisted spike-timing learning (SA-STL). With the help of our SA-STL rule, we can aggressively remove less important synapses dynamically in the SNN model with a little loss in the learning capability. In this work, we focus on reproducing the target spike train with a limited number of synapses in the SNN model. It is validated using both synthetic data and our cell culture data . Then, this paper provides an initial set of experiments to understand the possibility of replacing a sub-region of a neural circuit by training a recurrent SNN. To directly replace the sub-region of the neural circuit, we map each artificial neuron in our SNN model to each cultured biological neuron being monitored by a single probe in a microelectrode array (MEA). Connectivity between artificial neurons is trained by STDP-assisted supervised learning to generate a spike train that is identical to the desired spike train of the MEA. To demonstrate the real-time replacement, we implemented our SNN model on a hardware platform, i.e., Xilinx PYNQ-Z2 board, running at 50 MHz with pipelined execution. Overall, the key contributions of this work can be summarized as: Dataset Collection : We cultured a hippocampal neuronal network to collect spike activities of biological neurons for more realistic experiments. The data is collected every 12 h over 10 days, which provides 20 sessions in total. Learning Algorithm : We replaced the sub-region of the biological neural network by predicting spikes based on input spikes through an online STDP-based supervised learning rule. We proposed a novel learning method that reliably removes synapses in the SNN model, which leads to a more efficient hardware implementation. This results in the hardware design occupying less area and consuming less power. Hardware Implementation : We implemented a sparse SNN hardware on FPGA that predicts spikes of biological neurons in the replaced region in real-time (i.e., < 1 ms). The remainder of this paper is organized as follows. Section 2.1 introduces various neuron models and synaptic learning rules. Section 2.2 presents our SA-STL rule that dynamically selects important synapses to be connected when training an SNN model. In Section 2.3, we provide an experimental setup for replacing a sub-region of a neural circuit with the trained SNN model. Section 3.2 presents the details of SNN hardware architecture and analyzes the spike prediction accuracy using the actual hardware for real-time replacement. Then, we conclude the paper in Section 4.",
"discussion": "4. Discussions In this paper, we presented a novel learning algorithm, SA-STL, to efficiently remove synapses in an SNN model that replaces a sub-region of a biological neural circuit. The proposed SA-STL rule dynamically selects synapses that have more relevance to predicting spike timings of the target neural circuit. Then, the hardware prototype was designed on a small FPGA to reproduce spikes at the replaced region in real-time. To demonstrate the effectiveness of our software-hardware co-design approach, we collected neural recording data to conduct more realistic experiments. This work can be seen as an initial step for multidisciplinary research to replace a brain function with SNN hardware. Compared to the fully-connected SNN, our sparse SNN hardware could infer the spikes of the replaced sub-region in 22 μs with 2.5 × fewer hardware resources. It will have a more significant impact when we replace the brain functionality of a larger region in real-time using an implantable chip. Based on this initial set of experiments, our future work is to implement a closed-loop system where real-time spike communication happens between the main neural circuit (BNN) and the replaced SNN via electrical stimulation. This can be done by developing a precise electrical stimulation system that stimulates biological neurons connected to an SNN. Currently, our work assumes that such a stimulation system is available, and we allow inferred/measured spikes to convey data without any loss across BNN-SNN boundaries. Developing a precise electrical stimulation system along with low-impedance electrodes is one of our future works and is a fundamental challenge for repairing damaged neural circuits. Another challenge in processing spike trains in real-time is spike sorting, a process to identify the location of a neuron that has generated the spike at each electrode. Therefore, hardware for real-time spike classification across a large number of electrodes becomes another future research direction. Despite these limitations, this work presents an essential step toward real-time computation for neural prosthetics. Beyond the cultured hippocampus, we could replace a neural function at an impaired sub-region of the human brain with SNNs."
} | 2,852 |
38793178 | PMC11123252 | pmc | 1,099 | {
"abstract": "Resistive random access memory (RRAM) holds great promise for in-memory computing, which is considered the most promising strategy for solving the von Neumann bottleneck. However, there are still significant problems in its application due to the non-uniform performance of RRAM devices. In this work, a bilayer dielectric layer memristor was designed based on the difference in the Gibbs free energy of the oxide. We fabricated Au/Ta 2 O 5 /HfO 2 /Ta/Pt (S3) devices with excellent uniformity. Compared with Au/HfO 2 /Pt (S1) and Au/Ta 2 O 5 /Pt (S2) devices, the S3 device has a low reset voltage fluctuation of 2.44%, and the resistive coefficients of variation are 13.12% and 3.84% in HRS and LRS, respectively, over 200 cycles. Otherwise, the bilayer device has better linearity and more conductance states in multi-state regulation. At the same time, we analyze the physical mechanism of the bilayer device and provide a physical model of ion migration. This work provides a new idea for designing and fabricating resistive devices with stable performance.",
"conclusion": "4. Conclusions In conclusion, we prepared double oxide layers with different Gibbs free energies as functional layers and compared them with single functional layer devices. The Au/Ta 2 O 5 /HfO 2 /Ta/Pt devices have a larger switching ratio of 58.7, V Set and V Reset as low as −0.55 V and 0.46 V, respectively, and operating voltages that are smaller than those of S1 and S2 devices. Analysis of the statistical distributions of the switching voltage and resistance values shows that the δ/μ values of the V Set , V Reset , HRS, and LRS are only 8.22%, 2.44%, 13.12%, and 3.84%, respectively, which are smaller than the corresponding relative fluctuations of the single-layer devices. This indicates that the uniformity of the device is improved. The interface effect of the functional layer in the S3 device makes its multi-state modulation more linear. We present a detailed physical mechanism of resistive switching to explain the device’s performance enhancement. High yields were obtained in the verification of the device’s microscale performance. The Au/Ta 2 O 5 /HfO 2 /Ta/Pt RRAM devices proposed in this study show great potential for nonvolatile memory applications, in-store computing, and micro-shrinkage integration and provide a new idea for the design and fabrication of resistor devices with stable performance.",
"introduction": "1. Introduction As the limits of Moore’s Law are approached, computers using the Von Neumann architecture are limited by a storage wall and a power wall, and there is an urgent need to develop new memory-device solutions to meet the requirements of modern society for big data, artificial intelligence, and emerging industries [ 1 , 2 ]. Compared with the current mainstream charge-based flash memory, resistive switching random access memory (RRAM) has been considered one of the most promising prospects for next-generation non-volatile memory (NVM) devices owing to its simple structure, high integration density, high-speed operation, low power consumption, and good compatibility with conventional CMOS processes [ 3 , 4 , 5 ]. The structure of RRAM devices is similar to the traditional sandwich structure, consisting of a top electrode, a dielectric layer, and a bottom electrode. Pt, Au, Ti, Cu, Ag, or TiN are usually used as electrode materials [ 6 , 7 ]. Organics [ 8 ], transition metal oxides [ 7 , 9 ], perovskites [ 10 ], and two-dimensional materials [ 11 ] can be used as dielectric layers. Among the different materials, binary transition metal oxides are used for resistive device preparative studies owing to their simple chemical compositions [ 12 , 13 ], polymorphic switching properties, and compatibility with complementary metal oxide semiconductor (CMOS) fabrication processes [ 8 ]. Memory resistors can be used for storage and synapse mimicry [ 14 , 15 ]. Traditional methods of simulating neurons require dozens of conventional electronics, transistors, capacitors, etc. [ 16 ] This results in a huge challenge for power consumption and integration of the chip. The conductance state of the memristor is continuously adjustable under an applied electric field, but the uniformity of the memristor and the linearity of the polymorphic regulation are important performance metrics for its applicability, which has become a key parameter to be optimized [ 5 , 8 , 16 ]. Due to the existence of only a metal-semiconductor interface in the single-layer device, the concentration of oxygen ions and oxygen vacancies cannot be regulated, which increases the formation and breakage of conducting channels randomly and makes the device performance unstable [ 17 , 18 ]. Scientists have proposed many ways to improve stability, such as introducing nanocrystals in the functional layer [ 19 , 20 , 21 ], impurity doping [ 22 , 23 ], and integrating a layer of pinpoint electrodes [ 24 , 25 , 26 ]. However, these solutions require the addition of additional microstructure processing, sacrificing the scalability of micro-miniaturization and increasing production costs. The common types of thin film growth are chemical vapor deposition, reactive sputtering, atomic layer deposition, magnetron sputtering, and sol-gel. Among these, chemical vapor deposition lacks stability in the process of growing thin films. Reactive sputtering needs to maintain a high-temperature atmosphere during growth, which makes the method incompatible with CMOS processes. Atomic layer deposition is suitable for growing uniform films on substrates with gradients, but it is costly. The sol-gel method is less costly, but its homogeneity is poor. In contrast, magnetron sputtering can grow homogeneous films in a lower-temperature atmosphere, which is favorable for film growth [ 27 ]; therefore, in this study, the magnetron sputtering technique was used to prepare dielectric films. Therefore, we need to investigate simple and efficient methods to regulate the formation and breakage of conducting channels to improve the stability of the devices. Different Gibbs free energies lead to the varying simplicity of binding of oxygen ions to oxygen vacancies [ 28 ]. Therefore, we designed bilayer dielectric devices with different Gibbs free energies to improve the performance homogeneity of the devices. In this work, we fabricated and investigated Au/HfO 2 /Pt (S1), Au/Ta 2 O 5 /Pt (S2), and Au/Ta 2 O 5 /HfO 2 /Ta/Pt (S3) devices. Compared with single functional layer devices, S3 devices have enhanced stability, lower switching voltages, and more linear regulation of multiple states. The film roughness was characterized using atomic force microscopy. Importantly, we provide a detailed mechanistic explanation of the S3’s superior performance and ultimately validate the device’s microscopic performance.",
"discussion": "3. Results and Discussion We designed the memory resistor device of this work based on the differences in the oxide Gibbs free energy transitions. As shown in Figure 1 , we use the resistive transfer mechanism to determine the reasons for the superior performance of Au/Ta 2 O 5 /HfO 2 /Ta/Pt devices. The initial state of the device is shown in Figure 1 a, where more oxygen vacancies exist in the hafnium oxide layer near the tantalum side because tantalum is more capable of absorbing oxygen than the tantalum–oxygen interface [ 29 ]. The oxygen vacancy content of the hafnium oxide layer was characterized as shown in Figure S1 , with an oxygen vacancy content of 42.23%. As shown in Figure 1 b, when a negative bias is applied on the top electrode, oxygen in the dielectric layer will undergo the reaction in Equation (1), producing oxygen vacancies and oxygen ions, which migrate toward the bottom electrode, and oxygen vacancies move toward the top electrode under the action of the electric field [ 30 ]. The device completes the setup process when oxygen vacancies are connected to the top and bottom electrodes, as shown in Figure 1 c.\n (1) O + 2 e − = V o • • + O 2 − The lower Gibbs free energy means that the oxidation process is more likely to occur [ 31 ]. The magnitude of the Gibbs free energy transitions for oxide formation in Ta 2 O 5 and HfO 2 are −1903.2 kJ/mol and −1010.8 kJ/mol, respectively [ 32 , 33 ]. Hence, oxygen ions are more likely to recombine with oxygen vacancies in the tantalum oxide layer. Furthermore, the migration activation energy of oxygen ions at the interface is lower than that of the bulk phase [ 28 ]. As a consequence, oxygen ions at the interface between HfO 2 and Ta 2 O 5 are more likely to migrate under the proper electric field strength. As shown in Figure 1 d and e, when a positive bias voltage is applied to the top electrode of the device, the oxygen ions at the interface migrate and react with the oxygen vacancies in the tantalum oxide layer in a complex reaction, as shown in Equation (2), and a reset process occurs, resulting in the formation of the HRS [ 30 ].\n (2) O 2 − + V o • • = O + 2 e − Therefore, the connection and breaking of the conductive channel of the device occur at the Ta 2 O 5 /HfO 2 interface, which results in a more regular change in the conductive path and thus a more uniform distribution of high and low resistance values and operating voltages of the device. Oxygen ion migration at the Ta 2 O 5 /HfO 2 interface of S3 devices requires only a smaller voltage to drive compared with single-layer functional layer devices, resulting in a smaller switching voltage. The lower operating voltage results in less heat build-up during the reset process [ 33 ], which makes the multi-state regulation of S3 devices more linear. Figure 2 a shows that we fabricated a 64 × 64 crossbar array using photolithography and lift-off processes. More details of the crossbar array are shown under the 5× optical microscope image in the upper right corner of Figure 2 a. The line width of the array is 2 μm and the spacing is 10 μm, as seen in the 100× optical microscope image in the bottom right of Figure 2 a. The surface morphologies of the functional layers of S1, S2, and S3 devices were characterized by AFM, as shown in Figure S2 , and the surface roughnesses of the functional layers of S1 and S2 devices were 1.052 nm and 1.175 nm, respectively. as shown in Figure 2 b, the surface roughness of the S3 device film was 1.355 nm, which indicates that the fabricated films are relatively flat and suitable for the preparation of memristor devices. From Figure S3 , we can see that the electroforming voltage of the S3 device is higher than that of the S1 and S2 devices, which is due to the fact that the bilayer device requires a larger voltage to drive the oxygen vacancies to form a conductive channel during the electroforming process [ 7 ]. We investigated the S1 and S2 devices. As illustrated in Figure 3 a, when a voltage from 0 to −2 V is applied to the S1 device, the SET process occurs at −0.92 V, and the current changes abruptly from 1.5 to 5 mA. When a reverse voltage of 0 to 2.5 V is applied, a RESET process occurs at 0.92 V, and the current fades from 4.3 to 1.4 mA. The S1 device is capable of over 50 DC cycles. The I – V curve of the S2 device is shown in Figure 3 b. When a negative voltage of −2.5 V is applied to the Au electrode, the SET process can be observed at −1.8 V, where the current changes abruptly from 1.8 to 5 mA. When a positive voltage of 3 V is applied, the device switches to the RESET process, and the current changes gradually from 8 mA to 4 mA. The curves were repeated over 70 times. As shown in Figure 3 c, by applying a sweep voltage from 0 to −1.0 V to the S3 device, the SET process occurs at −0.54 V and the current suddenly increases from 0.1 to 3 mA. With a reverse positive sweep from 0 to 1.4 V, the device can return to the initial OFF state and the current gradually decreases from 4 to 0.1 mA in one integration cycle. By the same operation method, the S3 device can run steadily for over 200 cycles. This indicates that the S3 device has higher stability than S1 and S2 devices during C2C operation, with significantly lower V Set and V Reset for S3 compared with S1 and S2, respectively. Figure 3 d shows the cumulative distribution of the high and low resistance values of S1, S2, and S3 at 0.5 V. The switching ratio of the devices is calculated by reading the average of the high and low resistance values of the S1, S2, and S3 devices at 0.5 V. The HRS/ LRS ratio of the S3 device is 58.7, and the ratios of the S1 and S2 devices are 7.2 and 55.2, respectively, which indicates that the S3 device has a larger switching ratio. The results show that the ON/OFF ratio of the S3 device is sufficient for RRAM devices to be used for storing data [ 24 ]. Here, relative fluctuations are defined by δ/μ, where δ is the standard deviation and μ is the mean value. The relative HRS volatilities of S1, S2, and S3 devices are 14.59%, 57.46%, and 13.12%, respectively, and the relative LRS volatilities are 7.85%, 22.00%, and 3.84%, respectively. Both the high and low resistance fluctuation coefficients of the S3 device are smaller than those of the S1 and S2 devices, indicating that the S3 device has excellent uniformity. This high degree of homogeneity is due to the different Gibbs free energies of the bilayer devices, as well as the smaller migration energy of oxygen ions at the HfO 2 /Ta 2 O 5 interface [ 27 , 29 ], which limits the disruption and restoration of the conductive channels to the vicinity of the Ta 2 O 5 interface where the Gibbs free energies are lower, reduces the randomness of the conductive channel disconnection, and increases the uniformity of the high- and low-resistance states. Figure 3 e shows the cumulative distribution of V Set and V Reset for S1, S2, and S3. We can see that the δ/μ values of the V set of S1, S2, and S3 are 10.76%, 12.57%, and 8.22%, respectively, and the δ/μ values of the V reset of S1, S2, and S3 are 12.87%, 11.51%, and 2.44%, respectively. The S3 device has significantly decreased δ/μ compared with the operating voltages corresponding to S1 and S2. Comparative results show that the S3 device is more stable and requires a smaller driving voltage to connect and disrupt the conductive channels of the device, as the oxygen ion mobility energy at the interface is lower than that of the bulk phase. The lower operating voltage assists in reducing power consumption [ 28 ]. Figure 3 f shows the retention performance of the devices. The S1 and S2 devices have good retention performance in both the high- and low-resistance states with very little fluctuation. In comparison, the S3 device has better stability with almost no fluctuation in HRS and LRS over 10 4 s. As shown in Figure S4 , we performed programming endurance tests on S1, S2, and S3 devices. During the voltage pulse fatigue tests, the resistance of S1 and S2 devices changed significantly within 10 5 pulses, whereas the high and low resistances of S3 devices did not fluctuate significantly within 10 6 pulses, which indicates that the fatigue resistance of S3 devices is better than that of single-layer devices. These observations suggest that the S3 device has superior storage characteristics. Figure 4 shows the temperature change curve of the S3 device in the low-resistance state, and the on-current increases with temperature, which is consistent with the trend of the oxygen vacancy conductive mechanism. This proves that the conductive channel of the device consists of oxygen vacancies, which is consistent with our proposed conductive mechanism [ 7 ]. Temperature has a large impact on the performance of the device; therefore, in this work, the I – V performance of the S3 device was tested in an 85 °C environment, and the results are shown in Figure 5 , where the δ/μ values of V Set , V Reset , HRS, and LRS are statistically calculated to be 11.55%, 6.72%, 22.35%, and 8.95%, respectively. Compared with the performance of the S3 device at room temperature, the volatility of the test results conducted at 85 °C is increased, which is due to the increase in temperature, which decreases the stability of the oxygen vacancies in the device and leads to an increase in the fluctuations. However, it is clear from the I – V performance of the devices that the S3 device is still able to function properly in an 85 °C environment. Multiple conductance states in memristors have a wide range of applications in areas such as ultrahigh-density information storage, logic storage circuits, and neural networks, and the higher the linearity of the conductance states, the more favorable it is to improve the accuracy of the device in the application [ 34 , 35 , 36 ]. The polymorphic regulation was obtained by utilizing DC voltage scanning during the device reset process, starting from the voltage at the beginning of the reset and increasing the cut-off voltage in steps of 0.02 V until the end of the reset process. The conductance values obtained from each cut-off voltage regulation were read, and five points were selected for each of the S1, S2, and S3 devices to be modulated. The results of the statistical multistate regulation are shown in Figure 6 , where we can see that the S1 and S2 devices have 20 conductance states and 18 conductance states, respectively, adjusted under the control of the cut-off voltage, and the resulting conductance states are slightly less linear. Compared with the S1 and S2 devices, our S3 device can regulate up to 32 conduction states with higher linearity than the S1 and S2 devices. This is because the resetting process of the stacked structure of the S3 device occurs at the interface of hafnium oxide and tantalum oxide, which reduces the randomness of the conductive channel changes and improves the linearity of the multiple conductive states of the S3 device. Finally, the yield of the S3 device in the array shown in Figure 2 a was tested, as shown in Figures S5 and S6 . The yield of the device reached 79/81 × 100% ≈ 97.5%, and the device-to-device uniformity of the S3 device is 92.37%, which indicates that it has good micro-miniaturization potential. As summarized in Table 1 , in comparison with other literature on the same device structure, the present work has a lower switching voltage and 32 adjustable conductance states, which are important for the optimization of the device performance. Due to the good stability of the S3 device, its conductance was regulated using a pulse voltage. As shown in Figure 7 a, we applied a pulse voltage with an amplitude of 0.6 V and a pulse width of 3 μs to regulate the conductance of the S3 device by changing the period of the pulse. After applying 32 pulse voltages, the maximum change in current was achieved by pulse regulation with a period of 23 μs, and the minimum change in current was achieved by pulse regulation with a period of 63 μs. The conductance change rate obtained by pulse regulation is shown in Figure 7 b. It can be seen that for the same number of pulses, the conductance change rate of the pulse voltage regulation with a pulse period of 23 μs is more than 60%, while the conductance change rate of the pulse voltage regulation with a pulse period of 63 μs is only 10%. A good frequency-dependent property is shown, and this property can be used for frequency-dependent synaptic learning behavior [ 8 , 40 ]."
} | 4,842 |
39832126 | PMC11745301 | pmc | 1,101 | {
"abstract": "Hardware neural networks could perform certain computational tasks orders of magnitude more energy-efficiently than conventional computers. Artificial neurons are a key component of these networks and are currently implemented with electronic circuits based on capacitors and transistors. However, artificial neurons based on memristive devices are a promising alternative, owing to their potentially smaller size and inherent stochasticity. But despite their promise, demonstrations of memristive artificial neurons have so far been limited. Here we demonstrate a fully on-chip artificial neuron based on microscale electrodes and halide perovskite semiconductors as the active layer. By connecting a halide perovskite memristive device in series with a capacitor, the device demonstrates stochastic leaky integrate-and-fire behavior, with an energy consumption of 20 to 60 pJ per spike, lower than that of a biological neuron. We simulate populations of our neuron and show that the stochastic firing allows the detection of sub-threshold inputs. The neuron can easily be integrated with previously-demonstrated halide perovskite artificial synapses in energy-efficient neural networks.",
"conclusion": "Conclusion In conclusion, we have demonstrated the first fully on-chip halide perovskite artificial neuron. The neuron consists of only two components, which lends itself well to high-density integration, and shows clear leaky-integrate-and-fire behavior, important for integration in neuromorphic hardware. The spiking of the neuron is stochastic, similar to biological neurons, yet with a lower energy consumption per spike between 20 to 60 pJ. The stochastic spiking of the neuron is beneficial for detecting sub-threshold input, similar to biological neurons. The energy consumption of the neuron could be further reduced by lowering the capacitance of the capacitor. The similarity in device architecture of this artificial neuron to the downscaled artificial synapses of MAPbI 3 that we have shown before, 19 allows easy implementation of energy-efficient all-halide perovskite neuromorphic hardware.",
"introduction": "Introduction Artificial intelligence-based systems have seen a rapid increase in their capabilities in a wide range of tasks, such as natural language processing, 1 image recognition, 2,3 and strategizing. 4,5 The increase in the performance of these systems is accompanied by an exponential increase in the computational power, and thus the energy consumption. 6 Neuromorphic computing addresses this issue by implementing neural networks in hardware, lowering the required energy by orders of magnitude compared to conventional computers. 7 Neuromorphic chips rely on two main components for computation: artificial neurons, which integrate incoming signals and fire a voltage pulse upon reaching a threshold, and artificial synapses, which determine the connection strength between neurons. Ideally, both components can be integrated into a single chip in a dense arrangement to enable large-scale artificial neural networks. Both the neurons and synapses are typically implemented with electronic circuits composed of transistors and capacitors. 8 On the other hand, implementations that use memristive elements, which change their resistance based on an applied voltage, can be more compact and highly energy efficient, making them an attractive alternative. 9 Much research has gone into developing artificial synapses that directly use the resistance change of a memristive element as a proxy for connection strength. 9–12 Memristive elements also show promise for use in artificial neurons, because of the inherent stochasticity in their resistance changes. 13 This inherent stochasticity of memristive neurons can be leveraged for better signal representation, 14,15 or more efficient probabilistic computing than would be possible with deterministic neurons. 16 Nonetheless, applying memristive elements in artificial neurons is more complex and has been much less explored compared to their application in synapses. Here, we demonstrate a simple memristive neuron based on a halide perovskite memristive element. Metal halide perovskites are semiconducting compounds that efficiently conduct both electronic and ionic charge carriers. 17 The efficient ion conduction in halide perovskites readily induces hysteresis, which was previously exploited to make energy-efficient artificial synapses. 18–20 While various halide perovskite artificial synapses have been reported, only one halide perovskite neuron has been experimentally demonstrated before. 21 However, this previous implementation used off-chip circuitry to implement signal integration and neuron-like spiking, making scaling difficult. We connect a microscale volatile halide perovskite memristive device in series with a capacitor. The series capacitor applies a reverse bias on the memristive element after spiking of the neuron, which aids in resetting the memristive element after each spike. This makes our neuron design more robust against non-reversible resistance changes of the memristive component than designs with a series resistor, 22,23 or capacitor connected in parallel. 16,24,25 Because our design consists of only two components, the neuron is also more easily scalable than implementations that require more complex circuitry besides the memristive element. 14,26,27 Moreover, the efficient ion conduction of halide perovskites allows an operating voltage of hundreds of millivolts, lower than in previous memristive neurons which is favorable for low energy consumptions. We fabricate our crosspoint neurons with a previously developed procedure that prevents degradation of the halide perovskite layer during lithography. 19 Our neuron is integrated fully on-chip without the need for external circuitry to emulate neuron functionality. In that way, the device architecture of our halide perovskite memristive device lends itself to further downscaling and the neuron could be easily integrated with halide perovskite artificial synapses that we have demonstrated before to form artificial neural networks with ultralow-energy consumption. 19",
"discussion": "Results and discussion Artificial neurons can be fabricated from a resistive switch that shows rapid, highly volatile switching connected in series with a capacitor. 28 Thereby, successive voltage pulses eventually switch the memristive element to the low resistance state, charging the capacitor (firing). Then, the charged capacitor reverse-biases the memristive element, switching it off again. We use a resistive switch that comprises of methylammonium lead triiodide (MAPbI 3 ) as the active layer, and a gold and silver contact as the bottom and top contact respectively ( Fig. 1a and Methods section). The 2.5 μm wide contacts are arranged in an overlapping back-contact geometry, where the two contacts are orthogonally placed on top of each other with an insulating spacer layer of SiO 2 in between. All lithographic processing steps are therefore performed before the perovskite deposition. The compact, dense structure lends itself to downscaling. 19 This resistive switch shows a unipolar behavior with a clear threshold voltage of about 0.3 V, where the resistance rapidly changes by four orders of magnitude from approximately 1 GΩ to 100 kΩ ( Fig. 1b ). This resistance change is maintained for a short period only after switching off the voltage pulse, about 125 ms in the case of Fig. 1c , a requirement for the fabrication of an artificial neuron. A histogram of retention times based on 40 measurements is given in Fig. S2 (ESI † ). In no case is the retention time more than 500 ms. The resistance changes of the resistive switch are stochastic in nature, as is apparent from the histograms of the time to switch after applying the voltage pulse in Fig. S3a–c (ESI † ) and their corresponding fit with a Poisson distribution. Such a Poisson distribution for the switching time is expected for resistive switches that change their resistance due to stochastic formation and destruction of conductive filaments. 13 We note that resistance change can also occur for the same device but without the MAPbI 3 layer, as illustrated by Fig. S4 (ESI † ). The switching then happens at about 10× higher voltages. It has previously been shown that silver filaments can form in SiO 2 layers, 29 and the resistance changes therefore likely occur due to filament formation through the SiO 2 spacer between the Ag and Au electrodes. Thus, the role of the halide perovskite layer in the final device is to strongly facilitate the formation of these Ag filaments, enabling lower voltage operation and thereby reducing the energy consumption of the device. Fig. 1 A volatile halide perovskite resistive switch. (a) Optical microscopy image of the cross-point formed by the gold and silver electrodes before deposition of the halide perovskite layer, with a schematic image of the full resistive switching device. A gold bottom electrode and silver top electrode sandwich an SiO 2 insulating layer. Halide perovskite is spin-coated over the electrodes and forms the active layer of the device. (b) I – V curve of the device, measured between −0.5 and 0.5 V. The measured current increases by approximately 4 orders of magnitude at 0.3 V. The device returns to the initial high-resistive state as soon as the voltage is reduced to 0 V again and shows symmetric resistive switching properties in the negative poling direction. (c) Retention time measurement of the resistive switch. The resistance increases to that of the device in the high resistance state after approximately 125 ms. The full measurement is given in Fig. S1 (ESI † ). To turn this resistive switch into an artificial neuron, it needs to be connected to a capacitor. We implement this on-chip by connecting the resistive switch in series with a 300 pF capacitor that is formed by the Au bottom contact, the thermal SiO 2 layer and the highly-doped Si substrate, as shown in Fig. 2a . With such a connection, the operation of the neuron follows three key steps, depicted in Fig. 2b . In the first step, stimulation, the input voltage pulse experiences a resistive switch with high resistance. Therefore, every voltage pulse deposits only a small amount of charge on the capacitor, insufficient to build up significant voltage. After several pulses, the resistance of the resistive switch will promptly change to the low resistive state. At that point, the second step (firing) is initiated. The capacitor is quickly charged and the charge on the capacitor sets up a voltage that opposes the input voltage. The third step (resetting) is initiated when the applied voltage is removed. The capacitor discharges through the resistive switch, causing the resistive switch to return to the high resistive state, and the cycle can restart. Fig. 2 Operation of the artificial spiking neuron. (a) The neuron is constructed by connecting the memristive part of the device, consisting of the gold bottom electrode, silver top electrode and the MAPbI 3 layer, with the capacitor formed by the gold electrode and contact pad, the 100 nm thermal SiO 2 layer and the highly doped Si substrate in series. (b) Schematic representation of the three stages of the operation of the neuron. Upon application of a voltage, the device first undergoes a “stimulation” phase, where there is no significant voltage build-up on the capacitor due to the high resistance of the memristive part of the device. After enough voltage has been applied to the device, the memristive device switches to the low-resistance state and the capacitor is rapidly charged, causing a voltage buildup on the capacitor, i.e. “firing” of the neuron. When the applied voltage is removed, the capacitor discharges. This reverse-biases the resistive switch, aiding the disruption of the conductive filament, called the “resetting” process. (c) A pulsed measurement of the artificial neuron. A pulse train of 5 ms, 0.75 V pulses are applied with a 33 Hz frequency, resulting in firing spikes on the capacitor. \n Fig. 2c shows the experimental realization of the spiking of the artificial neuron. A 33 Hz, 750 mV pulse train is applied to the device and the voltage across the capacitor is measured. We observe firing pulses on the capacitor after one to three applied pulses. Fitting of the charging and discharging of the capacitor in Fig. S5a and b (ESI † ) reveals that the resistance of the resistive switch is reduced to 1 to 4 MΩ during most firing steps. The resistance obtained from the fit is higher than the 100 kΩ obtained in the voltage sweep in Fig. 1b , indicating that the device has not fully switched to the low resistance state. The voltage drop over the resistive switch is gradually reduced as the filament is forming and the capacitor is charged, leading to only partial formation of the filament. This partial formation of the filament further aids the volatility and energy efficiency of the device. During discharging of the capacitor in the resetting step, a resistance of approximately 10 MΩ is extracted, which corresponds to the input impedance of the oscilloscope. Assuming that the resistive switch is brought back to its 1 GΩ high resistance state during the resetting step, the oscilloscope provides a lower resistance discharge path for the capacitor, which is a limitation of our current measurement setup (see Fig. S5c, ESI † ). The capacitive discharge fit immediately corresponds to the oscilloscope impedance (Fig. S5a, ESI † ), from which we conclude that the resistive switch is reset as soon as the bias is removed, at least on the timescale of the measurement. No firing pulses were measured if the halide perovskite layer was omitted, as shown in Fig. S6 (ESI † ). The resistance changes that underlie the spiking behavior of the neuron, therefore, occur through the halide perovskite layer at these low applied voltages. Fig. S7 (ESI † ) shows that the stochastic spiking of the neuron was reproducible over multiple measurements. The firing pattern of the neuron is stochastic in nature, which is expected from the underlying stochastic switching mechanism of the resistive switch. Similar to the resistive switch itself, Fig. S8a (ESI † ) shows that the time under bias before spiking of the neuron follows a Poisson distribution, with a mean of 6.9 ms for the 0.75 V pulses. This stochastic switching is also observed in biological neurons and can have advantages compared to purely deterministic neurons. To demonstrate this advantage we use the experimentally obtained mean switching time and resistances to model the behavior of the stochastic neuron. We compared the simulated stochastic neuron to a hypothetical deterministic neuron with a deterministic threshold of the same time constant (6.9 ms) to determine the ability of stochastic and deterministic neurons to represent the input voltage pulse train. Modeling of the neuron is discussed in more detail in Supplementary note S1 (ESI † ). \n Fig. 3a shows the simulated spiking behavior of a stochastic and a deterministic neuron. The spiking of the simulated stochastic neuron is similar to that in the measurement shown in Fig. 2c . The simulated deterministic neuron, on the other hand, spikes at regular intervals. Fig. 3 Simulations comparing the stochastic spiking of the neuron with a hypothetical deterministic version of the neuron. (a) Comparison of a simulated stochastic and deterministic spiking neuron, with the same input as in Fig. 2c . Similar spiking behavior is obtained for the simulated and experimentally measured stochastic neurons. The deterministic neuron always spikes after a cumulative 6.9 ms of bias has been applied. (b) Simulated spiking behavior of populations of 100 stochastic and deterministic neurons. Ten voltage pulses are applied in the simulation with the same pulse duration, length, and magnitude as (a). Blue-shaded regions indicate the application of the 750 mV pulses, while the red marks indicate spiking by the neuron. While the deterministic neurons all spike at the same time, spiking by the stochastic neurons is distributed more evenly throughout the applied pulses. (c) The population codes obtained for each applied pulse in (b). We define the population code as the cumulative amount of spikes output by the population. For the deterministic population, the population code increases with each even number of applied pulses, while the stochastic population shows a more gradual increase with each applied pulse. (d) The representation error of deterministic and stochastic populations as a function of the population size, averaged over 1000 simulations. Deterministic populations have the same representation error regardless of their size. The representation error of the stochastic neurons decreases as the population size increases. The representation error of the stochastic populations is lower for population sizes of 11 or more neurons. The blue shaded region indicates one standard deviation. To achieve more biologically plausible, robust, and accurate spiking neural networks, neurons are typically implemented in populations. 14,15 In these networks, input signals are fed into the neurons in the populations and their collective output is collected as a population code. Fig. 3b shows a simulation of populations of 100 stochastic or deterministic neurons. While the spikes of the stochastic neurons are distributed over all input voltage pulses, the deterministic neurons spike uniformly roughly each second input pulse. From the simulations of the stochastic and deterministic neuron populations, we calculate the population code as the cumulative amount of spikes output by the total population after each successive input pulse, Fig. 3c . The population code for the deterministic populations increases stepwise, showing that the stochastic neurons can better distinguish different numbers of applied pulses, i.e. , they can better encode or represent the input. This process by which stochastic neurons can pick up on sub-threshold signals is called “stochastic resonance”. Biological neurons, which are also stochastic, rely on stochastic resonance to detect otherwise sub-threshold signals. 30 To study the effect of population size on the reliability of signal detection, we simulated population codes for populations of 1 and up to 100 neurons and computed a signal representation error for each population size, see Fig. 3d . Supplementary note S1 (ESI † ) explains how the representation error was determined. This representation error measures how well the population can encode and distinguish between different inputs. The representation error is initially larger for small populations of stochastic neurons compared to deterministic ones. However, the error rapidly decreases as the population size increases and drops below that of the deterministic neurons for relatively small population sizes of 11 or more stochastic neurons. These results are in line with previous work where the same benefit was found for stochasticity in artificial neuron populations. 14,15 Experimentally, the neurons are stochastic, but the stochasticity is tunable. The spiking behavior of the neuron can be tuned by changing the parameters of the input voltage pulses. As shown in Fig. 4a , the neuron outputs spikes with a higher probability for each input pulse if the frequency of the incoming pulses is increased. On the other hand, a lower input pulse frequency in Fig. 4b leads to no spiking of the neuron, which is a clear demonstration of the leaky behavior of the neuron. Another demonstration of the leaky-integrate-and-fire behavior of the neuron is given in Fig. S10 (ESI † ). Increasing the pulse duration to 7.5 ms leads to firing with each applied voltage pulse, whereas 2 ms pulses applied with the same frequency do not lead to spiking of the neuron. Fig. 4 Tunability of the firing of the neuron. (a) Increasing the frequency of the incoming voltage pulses to 50 Hz leads to a higher firing probability with each input pulse. (b) At a lower frequency of incoming voltage pulses of 20 Hz the neuron does not fire. (c) A lower input voltage of 400 mV, corresponding to connection of the neuron through high resistance synapses, leads to no firing of the neuron. Changing the voltage also provides a way to change the firing pattern of the neuron. When the neuron is integrated in full networks, this would be equivalent to connecting the neuron through synapses with a low connection strength, i.e. a high resistance. The measurement in Fig. 4c illustrates that a lower voltage drop over the neuron due to a resistive artificial synapse leads to no spiking of the neuron. Our spiking neuron therefore shows the leaky-integrate-and-fire behavior and synaptic strength-dependent spiking properties required for constructing neuromorphic hardware with the synapse. The energy consumption of the firing pulses can be calculated by , with C the capacitance of the on-chip capacitor and V the voltage of the firing pulse, which yields an energy consumption per firing pulse between 20 to 60 pJ. This is already lower than the energy consumed by a biological neuron (on the order of 100 pJ), 31 and artificial neurons that have been implemented in hardware spiking neural networks before, 32 even in this early adaptation. More energy-efficient silicon artificial neurons that were demonstrated before have not yet been implemented in full networks. 33 In addition, neurons based on electronic circuits of traditional transistors and capacitors require a large number of these components, 8,33 making the circuits bulky and therefore limiting the maximum density that can be reached on the final chip. In contrast, our design consists of only two components and could therefore be incorporated in higher densities more easily. Moreover, there is no detectable voltage build-up on the capacitor during the stimulation step before firing, meaning that the energy consumption per spike can be reduced by reducing the capacitance of the capacitor without negatively influencing the functioning of the neuron. We discuss further scaling effects in Supplementary note S2 in the ESI. † Biological neurons are sensitive to input signals of similar frequencies that we use in this work. 34 Although these frequencies are significantly lower than that of conventional computers, the different way that information is processed in neuromorphic networks still allows for efficient computation. In fact, neuromorphic networks require synapses and neurons that have time constants that are well-matched to their input for efficient computation. Thus, interfacing with the natural world, e.g. for learning from visual input, requires operating frequencies similar to those we use here. 7,35 These time constants can be difficult to achieve with CMOS-based neuromorphic hardware. 36 Our neuron therefore provides a convenient alternative that is natively capable of operating at these frequencies. The ability to incorporate these neurons and the corresponding artificial synapses on flexible substrates could allow for novel application areas, including soft robots or even in combination with biological tissue. In addition, ion conductivity and corresponding resistance changes of halide perovskites can be tuned by light stimulation. 37 Perovskite neurons could therefore also open up new possibilities of hybrid electronic-photonic neuromorphic hardware, such as low-power smart sensors."
} | 5,892 |
25887753 | PMC4381663 | pmc | 1,103 | {
"abstract": "Background The symbiosis between corals and the dinoflagellate alga Symbiodinium is essential for the development and survival of coral reefs. Yet this fragile association is highly vulnerable to environmental disturbance. A coral’s ability to tolerate temperature stress depends on the fitness of its resident symbionts, whose thermal optima vary extensively between lineages. However, the in hospite population genetic structure of Symbiodinium is poorly understood and mostly based on analysis of bulk DNA extracted from thousands to millions of cells. Using quantitative single-cell PCR, we enumerated DNA polymorphisms in the symbionts of the reef-building coral Pocillopora damicornis , and applied a model selection approach to explore the potential for recombination between coexisting Symbiodinium populations. Results Two distinct Symbiodinium ITS2 sequences (denoted C100 and C109) were retrieved from all P. damicornis colonies analysed. However, the symbiont assemblage consisted of three distinct Symbiodinium populations: cells featuring pure arrays of ITS2 type C109, near-homogeneous cells of type C100 (with trace ITS2 copies of type C109), and those with co-dominant C100 and C109 ITS2 repeats. The symbiont consortia of some colonies consisted almost entirely of these putative C100 × C109 recombinants. Conclusions Our results are consistent with the occurrence of sexual recombination between Symbiodinium types C100 and C109. While the multiple-copy nature of the ITS2 dictates that the observed pattern of intra-genomic co-dominance may be a result of incomplete concerted evolution of intra-genomic polymorphisms, this is a less likely explanation given the occurrence of homogeneous cells of the C109 type. Conclusive evidence for inter-lineage recombination and introgression in this genus will require either direct observational evidence or a single-cell genotyping approach targeting multiple, single-copy loci. Electronic supplementary material The online version of this article (doi:10.1186/s12862-015-0325-1) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusion While the results presented in this study do not provide unequivocal evidence of recombination between divergent Symbiodinium lineages, they provide an initial ‘proof of principle’ for its occurrence. In doing so, this study draws attention to the important evolutionary implications that may accompany the generation of new genetic diversity in Symbiodinium , including the potential for rapid symbiont adaptation through introgression. Progress in this area has been hindered by a lack of available methodology, an obstacle that is addressed here through the development of new molecular and statistical methods focused on the individual Symbiodinium cell. Additional development of this research may help to characterize and predict the evolutionary response of the coral-algal symbiosis to the many anthropogenic impacts currently threatening the world’s coral reefs. Data accessibility Amino acid sequence data is deposited in GenBank (accession number KJ530690) Quantitative PCR (qPCR) cycling threshold values and model parameters accompany the manuscript as supplemental information.",
"discussion": "Discussion Assessing the incidence of recombination between divergent Symbiodinium lineages is made difficult by their apparent haplontic life cycle, a lack of amenability to culture in many types (particularly in clade C Symbiodinium ), and the paucity of high-resolution single-copy genetic markers. This study attempts to circumvent these obstacles by developing protocols to isolate and extract DNA from individual Symbiodinium cells, establish and quantify the dominant ribotype(s) within each genome, and test competing hypotheses explaining the observed pattern of intra-genomic variation. Using these techniques, a population of putative inter-lineage recombinants is identified inhabiting the reef building coral Pocillopora damicornis at the isolated, high-latitude reef of Lord Howe Island, Australia. Method development The single-cell isolation and extraction method described here facilitated the rapid preparation of individual Symbiodinium cells prior to PCR (around 20 per hour), with the potential to be further improved with the application of flow-cytometry and fluorescence activated cell sorting (FACS). The protocol also showed good efficiency, with around 85% of isolated cells undergoing successful PCR amplification. The downstream application of DGGE and DNA sequencing successfully revealed the dominant ribotype(s) within individual cells, providing a reliable assessment of inter-genomic ITS2 diversity within the P. damicornis symbiont consortium. Used in conjunction with plasmid cloning, this method could be used to evaluate levels of intra-genomic variation in other genetic markers, providing an important assessment of their phylogenetic utility. The qPCR assay developed in this study offers sufficient sensitivity to quantify ITS2 ratios at the sub-clade level. This represents a significant improvement in resolution from earlier clade-level assays [ 71 - 76 ], since the sub-clade presents a more ecologically-relevant taxonomic unit [ 19 ]. This assay is also the first to quantify polymorphic rDNA sequences within individual Symbiodinium cells, and the second to do so in dinoflagellates (see also [ 51 ]). This provides an important insight into the level of ITS2 variation within the Symbiodinium genome, underscoring concerns about its utility in establishing diversity estimates [ 46 ], and its suitability for quantifying the dynamics of mixed infections [ 74 ]. In particular, substantial differences in rDNA copy numbers observed between Symbiodinium types C100 and C109 highlight the perils of using ITS2-qPCR to estimate abundance ratios of coexisting symbionts without single-cell validation. Finally, the statistical methodology developed here can identify potential admixture in symbiont populations based on intra-genomic ITS2 ratios. Conflicting hypotheses of one, two and three coexisting populations were formulated, corresponding to the existence of a single symbiont clone harbouring a non-diagnostic polymorphism (NDP), the coexistence of two ‘pure’ (homogeneous) ribotypes, and mixed populations of genetically homogeneous and heterogeneous Symbiodinium cells, respectively. The model consistent with the latter hypothesis received unambiguous statistical support in all six P. damicornis colonies analysed. However, the model selection approach relies on forming a set of candidate models that are representative of the biological processes under investigation [ 70 ]. While the mixture model representing H 2 is consistent with a population of recombinant genotypes coexisting with parental populations (progenitors), it cannot explicitly prove this scenario. This is because a similar pattern could arise from the incomplete concerted evolution of ancestral polymorphisms (ICEAP). Recombination or incomplete concerted evolution of ancestral polymorphisms? The existence of both C100 and C109 ribotypes in the homogeneous condition affirms their status as diagnostic of separate Symbiodinium sub-clades (i.e. neither sequence represents a degenerating pseudo-gene). Furthermore, these two ribotypes differ at five variable nucleotide sites in the ITS2 region (2% divergence), while NDPs typically feature a single nucleotide substitution or insertion/deletion (indel) that distinguishes them from the dominant sequence variant [ 19 , 36 ]. However, if both ribotypes were present within the genome of the most recent common ancestor of Symbiodinium C100 and C109, processes of concerted evolution may not have had sufficient time to homogenize the rDNA arrays of both taxa. Hence copies of the ribotype that is now diagnostic of the sister taxon may remain in the genome of one or both lineages. The Symbiodinium genome routinely hosts a diverse assemblage of ITS2 sequences [ 46 ], and several putative cases of ICEAP appear in the literature. For example, the ITS2 sequence diagnostic of Symbiodinium glynni (type D1) also occurs within the genome of S. trenchii (type D1a), with the incomplete displacement of a vestigial polymorphism invoked to explain their intra-genomic coexistence [ 46 ]. However, several features of the data presented here suggest that an alternative explanation of recombination is feasible. First, the C100 and C109 sequences coalesce at the ancestral type C3, as opposed to either representing an intermediate evolutionary step toward the other (e.g. C103 and C118 in P. damicornis and C3hh and C3n in Seriatopora hystrix ; see Figure 2 b). If concerted evolution has not had sufficient time to homogenize all C109 rDNA repeats in the C100 genome, then vestigial copies of the intermediate C3 sequence would also likely persist as a non-dominant intra-genomic variant. Rather, the C3 sequence was not detected in any of the cells analysed, despite its characteristic DGGE band pattern (see supplementary material in [ 53 ]). Second, concerted evolutionary processes rapidly homogenize intra-genomic co-dominance, either completely displacing a non-dominant polymorphism or leaving only background traces [ 44 , 77 , 78 ]. This is inconsistent with the similar proportional abundance of ITS2 polymorphisms within many of the genetically heterogeneous cells observed here, with more than a third of all symbionts featuring C C100 :C TOTAL ratios of between 0.25 and 0.75. Finally, frequency ‘dips’ along the C C100 :C TOTAL spectrum depict a degree of genetic isolation between genetically heterogeneous Symbiodinium cells and either of the ‘pure’ genotypes (i.e. homogeneous C100 and C109 cells), consistent with the substantial fitness loss often experienced by F 2 and later-generation backcross genotypes (as a result of processes such as ‘hybrid breakdown’; see [ 79 , 80 ]). While recombination represents a plausible explanation for the intra-genomic codominance of the C100 and C109 ribotypes, there remains a possibility that this pattern resulted from ICEAP. Addressing this question will likely require a significant investment of resources, including the development of a suite of single-copy markers, the generation of isoclonal cultures or the application of whole genome amplification (WGA; in order to facilitate multi-locus genotyping analysis on individual cells), and/or continued attempts to induce the sexual life cycle, both within and between cultured Symbiodinium lineages. Another area requiring investigation is the morphological, physiological and ecological characterization of putative Symbiodinium recombinants. Concerted evolution operates via a series of stochastic processes that occur independently of natural selection [ 81 ]. By contrast, recombination between lineages is often accompanied by drastic changes in morphology, performance and fitness [ 79 , 82 - 84 ], even involving diversification into new habitats [ 85 ]. Investigating the form, function, distribution and ecology of genetically heterogeneous Symbiodinium cells may therefore provide further insight into the incidence and potential evolutionary effects of recombination within and between Symbiodinium lineages. Background symbiont populations The results of this study indicate that at least three ITS2 genotypes can coexist within the symbiont consortium of P. damicornis (C100, C100/C109 and C109). While homogeneous Symbiodinium C109 cells were only ever detected at background levels (constituting less than 7% of the symbiont population), the biological relevance of this population may extend well beyond providing a presumably minor contribution to the overall productivity of the symbiosis. Genetically heterogeneous Symbiodinium cells outnumbered ‘pure’ genotypes in more than half of the colonies sampled, suggesting that rare sexual reproduction events between C100 and C109 may facilitate asexual proliferation of the F 1 generation, with potentially important functional implications for the coral colony. The evolutionary contribution of rare Symbiodinium types may be more important still, if recombinants create a ‘bridge’ for the migration of genetic material to the dominant lineage (i.e. introgression; see Figure 1 b). A small number of genetically heterogeneous symbionts featured C C100 :C TOTAL ratios near 0.75, and thus potentially represent F 1 × C100 backcross genotypes. However, this pattern could equally have arisen from ICEAP, differential rDNA inheritance in the F 1 generation (arising from dissimilar copy-numbers between parent taxa; e.g. [ 51 ]), or even concerted evolution acting to homogenize rDNA variability in the recombinant genome (e.g. [ 77 ]). Establishing the incidence of introgression would initially require the identification of individual F 1 - and backcross classes. This in turn requires the genotyping of a large number of individuals, and the analysis of at least 13–50 ancestry-informative loci per individual [ 86 , 87 ]. This study was not sufficiently resourced to carry out such a comprehensive task; however it does serve to highlight the perils of dismissing symbionts that persist in low abundance as biologically-irrelevant or simply representing surface contamination."
} | 3,334 |
36425175 | PMC9670684 | pmc | 1,104 | {
"abstract": "Superhydrophobic surfaces can be derived from roughening hydrophobic materials. However, the superhydrophobic surfaces with various micro/nano morphologies present variations of chemical and mechanical durability, which limits their practical applications. Very little actually is known about comparing durability and corrosion resistance of concave and convex superhydrophobic surface structures systematically. In this paper, two kinds of superhydrophobic AlNiTi amorphous coatings with concave and convex surfaces were obtained by chemical etching and hydrothermal methods, respectively. Benefiting from nanoscale sheet structure, the convex superhydrophobic coating displays higher water-repellence (contact angle = 157.6°), better self-cleaning performance and corrosion resistance. The corrosion current density of the convex superhydrophobic surface is approximately one order of magnitude smaller than the concave superhydrophobic surface. Besides, the long-term chemical stability and mechanical durability of both superhydrophobic surfaces were also investigated. The formation and damage mechanisms of these two kinds of superhydrophobic surfaces were proposed. It is hoped that these investigations could provide clear guidance for the real-world applications of superhydrophobic amorphous coatings.",
"conclusion": "5. Conclusion In this work, two kinds of superhydrophobic AlNiTi amorphous coatings were prepared by chemical etching and hydrothermal methods. Both coatings show typical micro/nano hierarchical structures: the concave pit structure with an average size of 1.68 μm for the ES coating and the convex sheets with a size range from 49 to 95 nm for the HS coating. The water contact angle value of the concave pit structure is 153.1°, while that of the convex sheet structure is 157.6°. Compared with the ES superhydrophobic coating, the HS superhydrophobic coating presents better water repellency, self-cleaning performance and corrosion resistance. The I corr value of the HS coating is approximately two orders of magnitude lower than that of the AS coating and one order of magnitude smaller than that of the ES coating. The corrosion inhibition efficiency of the HS superhydrophobic coating reaches 99.1%. Additionally, the HS superhydrophobic coating shows better chemical durability under a harsh environment including long-term immersing in salt, acid and alkali solutions. Similarly, after 300 cycles of abrasion, the HS coating still remains superhydrophobicity, which is significantly higher than the 200 cycles of wear for the ES coating. The nano-convex superhydrophobic amorphous coating prepared by hydrothermal methods shows a promising application potential, due to its excellent self-cleaning, corrosion resistance, chemical and mechanical durability.",
"introduction": "1. Introduction For the lotus leaf, there are hydrophobic wax tubules located on convex cell papillae on the surface, forming the hierarchically micro-nano structure. The micro-nano structure and hydrophobic molecules diminish the contact area of droplets with the surface, which leads to the formation and stability of air cushions. 1 It results in high contact angle (CA), low adhesive force and excellent self-cleaning properties. 2 Inspired by the lotus leaf effect, the superhydrophobic surface (water contact angle >150 °C) has generated widespread attention due to its excellent performance, for example, corrosion inhibition, self-cleaning and drag reduction, etc. 3–6 Meanwhile, the superhydrophobic characteristic also contains potential value in technical and industrial applications. 7 The strategy for constructing superhydrophobic surfaces by using micro-nano structures and low surface energy materials is indispensable conditions. 8,9 Numerous methods have been successfully developed in the past few decades, such as anodic oxide, 10,11 chemical etching, 12–14 electrodeposition, 15,16 and hydrothermal method. 17–19 Among them, chemical etching and hydrothermal techniques are facile methods to prepare superhydrophobic coatings on metal surfaces. Chen et al. 12 designed a crater-like structured superhydrophobic Al alloy with HCl and stearic acid mixture solutions. The CA values of droplets all exceed 150° at different pH values, indicating the remarkable chemical and mechanical stability of the crater-like structure under acidic and alkaline conditions. Kumar et al. 13 reported that the superhydrophobic Al coating with rough rectangular pits-like microstructures was fabricated by chemical etching technique. The coating shows a high water static CA of 162.0° with excellent stability and self-cleaning performances. Guo et al. 14 synthesized superhydrophobic coating on a 7055-Al alloy surface with micron petal-shaped structures. The corrosion inhibition efficiency of the coating reaches 99.67%. The water repellency avoids the corrosion ions infiltrating the substrate and the growth of corrosion products. The superhydrophobic Al foils with a micro-nano sheet layer were fabricated using a one-step hydrothermal method. 17 The superhydrophobic surface shows drag reduction, self-cleaning and anti-icing properties. Zhang et al. 18 developed a one-step hydrothermal strategy to prepare superhydrophobic 5083-Al alloys with a high CA of 167.2°. The micro flower-like structures contribute to the eventual anti-wetting performance and favorable corrosion resistance. An irregular conical structured superhydrophobic surface was formed on Al substrates by Lan et al. 19 The CA of the superhydrophobic surface reaches 162°, and the alloy exhibits a large charge transfer resistance ( R ct ) and prominent anti-icing performances. To improve corrosion resistance of arc sprayed Al coating, the superhydrophobic surfaces with CA of 153.4° were obtained by ultrasonic etching and modification processing in our previous investigation. 20 The corrosion current density of superhydrophobic surface decreases by 2 orders of magnitude due to the formation of typical micro/nano-scale structures on the etched Al coating. The chemical and mechanical durability is a vital factor limiting superhydrophobic surfaces practical applications. 21 Though many recent advances in this region, the durability of superhydrophobic surface by different methods varies greatly. Nevertheless, extremely scarce reports concentrated on comparing corrosion resistance and durability of superhydrophobic surfaces prepared by etching and hydrothermal methods. It has remained elusive in micro-nano superhydrophobic structure of etched concave surface and hydrothermal convex surface. Therefore, it is highly desirable to clarify the corrosion resistance and durability of different micro/nanostructures superhydrophobic coatings. Prompted by this, an attempt was initiated to synthesize two kinds of superhydrophobic structures on arc-sprayed AlNiTi amorphous coating surfaces by chemical etching and hydrothermal methods. The wettability and corrosion resistance of both superhydrophobic structures was compared. The chemical and mechanical durability were analyzed based on the liner friction and long-term corrosion experiments. Besides, the formation process and impairing mechanisms were further proposed.",
"discussion": "4. Discussions 4.1 Formation mechanism of the coatings Chemical etching and hydrothermal methods are the simplest strategies to prepare superhydrophobic coatings. In this work, the ES coating is fabricated by a two-step process. Firstly, the AlNiTi amorphous coating is roughed with NaOH solution. Fig. 11(a) and (b) are SEM images and etching process schematic diagrams of the ES coating, respectively. At the initial stage of etching, corrosive pitting appears on the micro-protrusion surfaces, as seen in Fig. 11(b) . As the etching progresses, the size of the pit increases gradually to a microscale corrosive crater, and then numerous nanoscale corrosion pits are formed inside them. It promotes the formation of micro/nanoscale hierarchical structures, which enhances the coating roughness. Secondly, the etched coatings are modified by trimethoxy (1 H ,1 H ,2 H ,2 H -heptadecafluorodecyl) silane to reduce the surface energy. Fig. 11 (a) and (b) SEM image and etching process schematic diagrams of the ES coating; (c) and (d) SEM image and hydrothermal reaction process schematic diagrams of the HS coating. After modification, the XPS spectra of the ES coating surface contain five elements: the F 1s at 689.1 eV, C 1s at 292.0 eV, O 1s at 532.8 eV, Si 2p at 104.1 eV and Al 2p at 74.8 eV, as shown in Fig. 12(a) . The Al 2p spectrum consists of Al 3+ ( Fig. 12(b) ), confirming the chemical reactions on the coating surface. For the C 1s spectrum in Fig. 12(c) , all characteristic groups of trimethoxy (1 H ,1 H ,2 H ,2 H -heptadecafluorodecyl) silane molecules are discovered including C–C at 285.3 eV, C–O at 286.5 eV, C–O–Si at 288.9 eV, CF 2 at 291.7 eV and CF 3 at 293.9 eV. From Fig. 12(d) , the peaks of Al–O at 530.8 eV and C–O at 532.9 eV are also remarked in the O 1s spectrum, respectively. The Al–O peak indicates that the low surface energy material trimethoxy (1 H ,1 H ,2 H ,2 H -heptadecafluorodecyl) silane molecules are successfully grafted onto the coating surface through the formation of Al[CF 3 (CF 2 ) 7 CH 2 CH 2 SiO] 3 . The reaction process is as follows: 18,28 2 Al 3+ + CF 3 (CF 2 ) 7 CH 2 CH 2 Si(OCH 2 CH 3 ) 3 → Al[CF 3 (CF 2 ) 7 CH 2 CH 2 SiO] 3 Fig. 12 XPS spectra of the ES and HS coatings: (a) survey spectra, (b) Al 2p, (c) C 1s, and (d) O 1s. The CA value of the ES coating is up to 153.1°, as shown in Fig. 4 . The superhydrophobicity and water repellent of the ES coating profit from ultralow surface energy groups of CF 2 and CF 3 . \n Fig. 11(d) shows the hydrothermal reaction process schematic diagrams of the HS coating. Numerous nanoscale sheet structures are formed on the coating surface by hydrothermal reaction. These nanoscale sheets grow gradually as a function of reaction time. Finally, the micro/nano hierarchical structures are formed and the surface roughness is improved, leading to its CA value of 157.6° ( Fig. 3 and 4 ). Similarly, there are three signals of the C 1s (284.9 eV), O 1s (531.9 eV) and Al 2p (74.8 eV) on XPS spectra of the HS coating surface, as seen in Fig. 12(a) . According to the illustration in Fig. 12(b) , the high-resolution Al 2p consists of only one component Al 3+ in the hydrolysis reaction. The peaks of high-resolution C 1s spectrum ( Fig. 12(c) ) at 284.4 eV, 284.8 eV and 288.5 eV correspond to C–C, C–O and O–C \n \n\n<svg xmlns=\"http://www.w3.org/2000/svg\" version=\"1.0\" width=\"13.200000pt\" height=\"16.000000pt\" viewBox=\"0 0 13.200000 16.000000\" preserveAspectRatio=\"xMidYMid meet\"><metadata>\nCreated by potrace 1.16, written by Peter Selinger 2001-2019\n</metadata><g transform=\"translate(1.000000,15.000000) scale(0.017500,-0.017500)\" fill=\"currentColor\" stroke=\"none\"><path d=\"M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z\"/></g></svg>\n\n O, respectively. There are two peaks of Al–O at 531.3 eV and C–O at 531.8 eV in high-resolution O 1s spectra ( Fig. 12(d) ). The presence of the Al–O peak indicates that the Branching reaction occurs on the coating surface: 3 Al 3+ + CH 3 (CH 2 ) 16 COOH → Al[CH 3 (CH 2 ) 16 COO] 3 XPS results reveal the triumphant grafting of stearic acid on the surface. Therefore, the nanosheets are formed on the surface by hydrolysis of Al to Al 3+ and neutralization reaction to obtain Al[CH 3 (CH 2 ) 16 COO] 3 with low surface energy. Similar results were also reported for the superhydrophobic AA6061 alloys and superhydrophobic surfaces on aluminum foil. 17,29 4.2 Durability mechanism of the coatings From Fig. 7 and 9 , the superhydrophobic coatings show better corrosion resistance in comparison to the AS coating. Furthermore, the HS coating presents long-term chemical durability. The improving mechanisms of chemical stability for the superhydrophobic coatings are illustrated by the model schematics diagrams, as shown in Fig. 13 . For the AS coating surface, when water droplets drop on its surface, hydrophobic microscale Cassie state is formed due to numerous microscale protrusions. After being immersed in corrosive solutions, the coating surface is covered with a thin air film. The air film is easily destroyed under hydrostatic pressure, resulting in an irreversible transition from the microscale Cassie to the Wenzel state. As a result, the corrosive media (such as O 2 , Cl − , H + and OH − ) directly contact the coating surface and react with it. 30 With the extension of soaking time, corrosion media severely damage the coating surface and penetrates into the coating to form corrosion pits. Fig. 13 Schematic diagrams of the chemical durability mechanisms for the coatings. Based on the kinetic process of wetting transitions (WTs), the Cassie state will collapse when the pressure exceeds the critical pressure for the breakdown of the Cassie state. 31 For typical microscale pillar structures, the hydrophobic surface is derived from capillary pressure by the curvature of the droplets on the contact surface. Therefore, it is a crucial role to obtain a high critical pressure to resist further penetration of droplets by inducing nanostructure on the surface. For the ES coating, a large number of micro/nano corrosion pits are detected on the microscale protrusions surface by etching, as seen in Fig. 11(a) . It leads to the transformation of microscale Cassie state to nanoscale Cassie state in corrosion solutions. The direct contact between corrosive media and the coating is inhibited by air cushions and closely spaced hydrophobic alkyl molecules. As corrosive media continuously erodes, some hydrophobic molecules will detach from the coating surface due to the partial decomposition of chemical interface bond. Subsequently, partial superhydrophobic film will be permeated by the formation of nano/micro pores and the coating surface will be wet. However, air pockets still exist in the structures. The ES coating shows an intermediate state, namely the Marmur state. 32 In this case, although the superhydrophobicity of the coating deteriorates, it still has excellent hydrophobicity in corrosive solutions, as shown in Fig. 9 . For the HS coating, numerous nanoscale sheets exist on the micro-pillar surface, forming micro/nano hierarchical structures ( Fig. 11(c) and (d) ). Nano-Cassie state is triggered on the HS coating surface by droplets. Compared with micro-concave pits, the convex sheets have better superhydrophobic and durable properties due to their smaller size. This phenomenon can be explained by surface free energy curves, in which the interface of a stable solid–liquid–gas system needs to be convex rather than concave. 31,33 The schematic diagrams of the mechanical durability mechanism for the coatings are depicted in Fig. 14 . During the abrasion process, the micro protrusion parts at the top of the AS coating are easily worn under mechanical pressure ( Fig. 14 ). The CA of the AS coating value decreases rapidly due to the reduction of roughness, as seen in Fig. 10(b) . Comparatively, micro/nano hierarchical structures provide a valid method to strengthen the mechanical stability of superhydrophobic surfaces. In this case, the microstructures can bear the load and provide robustness to protect nanostructures. The remaining nanostructures ensure the superhydrophobicity of the coatings. 34 For the microscale concave structures, the micro/nano hierarchy is easily damaged by sandpaper, leading to the decrease of roughness and the removal of low-energy materials, thus worsening the superhydrophobic performance. However, the convex structures still hold the micro/nano hierarchical rough structures on their surfaces after abrasion. Hence, the nanoscale convex superhydrophobic surfaces have better mechanical durability. Fig. 14 Schematic diagrams of the mechanical durability mechanisms for the coatings. The theory of energy thermodynamics can be used to reveal the superior chemical stability and durability of superhydrophobic surfaces with micro/nano convex structures. For the solid–liquid–air system, the surface free energy ( W ) is expressed as follows: 35 4 W = A SL γ SL + A SA γ SA + A LA γ LA where A SL and γ LA , A SA and γ SL and A LA and γ SA are the areas and interfacial tensions of the solid/liquid interfaces, solid/air interfaces and liquid/air interfaces, respectively. To ensure the stability of the superhydrophobic surface, the local free energy needs to reach the minimum value, that is, d W = 0 and d 2 W > 0. Through the Lagrange method including multiple variable constraints, the stability criterion formula is obtained as follows: 35 5 d A SL d θ < 0 where d θ is a small change of local CA at the solid/liquid interface. In addition, it is also equal to the change in the slope of the surface. Consequently, the stability of the superhydrophobic surface depends on the sign of curvature, and a stable surface requires a convex surface, not a concave one. 35,36 Therefore, the prepared micro/nano convex surfaces possess excellent stable superhydrophobicity even after chemical or physical damage."
} | 4,317 |
40128235 | PMC11933384 | pmc | 1,108 | {
"abstract": "Systems metabolic engineering is facilitating the development of high-performing microbial cell factories for producing chemicals and materials. However, constructing an efficient microbial cell factory still requires exploring and selecting various host strains, as well as identifying the best-suited metabolic engineering strategies, which demand significant time, effort, and costs. Here, we comprehensively evaluate the capacities of various microbial cell factories and propose strategies for systems metabolic engineering steps, including host strain selection, metabolic pathway reconstruction, and metabolic flux optimization. We analyze the metabolic capacities of five representative industrial microorganisms as cell factories for the production of 235 different bio-based chemicals and suggest the most suitable host strain for the corresponding chemical production. To improve the innate metabolic capacity by constructing more efficient metabolic pathways, heterologous metabolic reactions, and cofactor exchanges are systematically analyzed. Additionally, we present metabolic engineering strategies, which include up- and down-regulation target reactions, for the improved production of chemicals. Altogether, this study will serve as a comprehensive resource for the systems metabolic engineering of microorganisms in the bio-based production of chemicals.",
"introduction": "Introduction Systems metabolic engineering 1 , 2 , which integrates the strategies and tools of synthetic biology, systems biology, and evolutionary engineering with traditional metabolic engineering, allows more efficient development of microorganisms for the sustainable production of various chemicals, including bulk chemicals 3 – 6 , fine chemicals 7 , 8 , fuels 9 – 11 , polymers 12 – 14 , and natural products 15 – 19 from renewable resources instead of fossil resources. Starting with project design, systems metabolic engineering aims to optimize host strain selection, metabolic pathway construction, and metabolic fluxes, while considering fermentation and downstream processes 20 . However, exploring the vast metabolic space, represented by the combinations of the metabolic networks of different host strains and strain engineering strategies, still demands significant time, effort, and costs. Model microorganisms such as Escherichia coli and Saccharomyces cerevisiae have been the primary workhorses for metabolic engineering due to the availability of the most abundant knowledge on their genetic and metabolic characteristics and also the gene manipulation tools. However, selecting a host strain requires consideration of the most suitable metabolic characteristics for the production of a target chemical. These include the presence of a native biosynthetic pathway for the target chemical, or the potential to produce it effectively when a heterologous or new biosynthetic pathway is introduced, capacity to produce the target chemical, the safety of the microorganism, and the environmental conditions in which the microorganism can thrive 21 . Recent advancements in bioengineering tools, such as clustered regularly interspaced short palindromic repeats (CRISPR) 22 and serine recombinase-assisted genome engineering (SAGE) 23 , have enabled the metabolic engineering of non-model organisms that naturally produce target chemicals more amenable. Obviously, performing metabolic engineering on a host strain that possesses the highest biosynthetic capacity toward the target product is a promising strategy, as the strain has a potential to more efficiently produce chemicals compared to the other strains with lower biosynthetic capacity. The production performance is defined by three key metrics: titer (the amount of product per volume), productivity (specific productivity, the rate of production per unit of biomass, or volumetric productivity, the rate of production per volume), and yield (the amount or mole of product per amount or mole of consumed substrate) 24 . Among these key metrics, yield determines the required raw material costs, significantly affecting the overall bioprocess costs. Thus, selecting a host strain with a biosynthetic pathway that maximizes the yield of chemical production is crucial. Genome-scale metabolic models (GEMs), which represent gene-protein-reaction associations in organisms through mathematical models, have been used to analyze the biosynthetic capacities and engineering strategies for developing microbial cell factories 20 , 25 , 26 . For example, gene knockout targets for the improved production of l -valine in E. coli were identified at the systems level by performing in silico knockout simulations for each gene in the strain, which would otherwise require considerable time, effort, cost for real experiments 27 . GEM-based approaches have not only identified gene targets for engineering but also characterized strain variations 28 , constructed biosynthetic pathways toward desired chemicals 29 , 30 , analyzed metabolic resource allocations in host strains 31 , and predicted metabolic interactions between microbial communities 32 . Although GEMs have been utilized to optimize host strain selection, metabolic pathway construction, and metabolic fluxes, a comprehensive exploration of the processes at the systems level still demands significant effort. In this study, we aim to provide resources for host strain selection, metabolic pathway construction, and metabolic flux optimization. To support host strain selection, we provide the metabolic capacities for 235 chemicals that have been produced, even if only minimally, in representative industrial microorganisms by calculating the maximum theoretical yield (Y T ), the maximum production of the target chemical per given carbon source when resources are fully used for the target chemical production, and maximum achievable yield (Y A ), the maximum production of the target chemical per given carbon source, considering cell growth and maintenance. For further improvement of metabolic pathway reconstruction, we have also systematically analyzed the expansion of innate metabolic capacity through the addition of heterologous reactions and cofactor exchanges in native metabolic reactions, and rewiring of innate metabolism to improve target chemical production. Furthermore, metabolic engineering strategies, which include the target reactions to be up- and down-regulated, are suggested for the improved production of chemicals. To demonstrate the versatility and applicability of these resources, we selected various products, including amino acids ( l -lysine and l -glutamate) and ornithine used as nutritional supplements; precursors for biopolymers (sebacic acid and putrescine); a bulk chemical (propan-1-ol); and a key precursor for various natural products (mevalonic acid) as case studies. The resources presented in this study can also be employed for analyzing the other 229 chemicals (Supplementary Data 1 – 23 ) and also for other chemicals not described here using similar approaches.",
"discussion": "Discussion Planning a metabolic engineering project necessitates an extensive search through the entire decision-making process, including the selection of target chemicals, host strains, and pathways to be engineered. Finding an optimal strategy for the project is challenging without systematically exploring the vast metabolic space. To aid the initiation of a metabolic engineering project, we provide a comprehensive evaluation of the capabilities of microbial cell factories. The maximum yield of a chemical indicates how efficiently a biosynthesis pathway can transform a carbon source into the target chemical, thereby guiding metabolic engineers in selecting the most optimal host strain by comparing the maximum yields across different strains. In this study, we calculated the Y T and Y A of 235 bio-based chemicals in five representative host strains for metabolic engineering (i.e., B. subtilis , C. glutamicum , E. coli , P. putida , and S. cerevisiae ). It should be noted that the development of genome engineering tools for non-model organisms increasingly enables the use of non-traditional strains for metabolic engineering 58 – 60 . Therefore, expanding the analysis of this study to encompass all available genome data could inspire the use of less-explored organisms for metabolic engineering as well. Conventional constraint-based modeling approaches (e.g., FBA) do not account for gene expression, regulatory network, or allocation of macromolecules within an organism 61 – 63 . Although our study is limited to calculating the maximum yields of bio-based chemicals using FBA, integrating multi-scale mechanisms will allow for a more accurate calculation of the biotechnologically achievable maximum yields of these chemicals. Moreover, since we neglected the target chemical-specific transport reactions in our analysis, further characterization and inclusion of these exporters will enable more accurate yield calculations. Recent advancements in metabolic modeling have highlighted the importance of integrating enzyme kinetics and proteome constraints to understand the metabolism of microorganisms. Although proteome-integrated GEMs offer valuable insights under specific experimental conditions, their applicability is limited when exploring diverse environmental and substrate conditions. To assess the potential impact of such constraints on maximum yields, we compared the maximum yields obtained from iML1515, a GEM of E. coli , and its enzyme-constrained counterpart, eciML1515 (Supplementary Note 1 ) 64 . While the magnitude of maximum yields from the enzyme-constrained model can vary, the overall trends and distributions were statistically indistinguishable. These findings demonstrate that while enzyme constraints could provide a more realistic value of maximum yields, the general trends across conditions remain consistent. Thus, the current GEMs provide robust results for exploring metabolic capacities over a wide range of conditions, although further incorporation of enzymatic information continues to be a valuable tool when more precise, condition-specific yield calculations are required. We also proposed engineering strategies to enhance the innate metabolic capacities of microbial cell factories or to rewire their metabolism toward target chemical production. To identify heterologous reactions for introduction into a microbial cell factory, we utilized and curated a universal model that accounts for all reported metabolic reactions. It is evident that the search of heterologous reactions is limited by the quality and quantity of the universal model. As most currently available GEMs have been reconstructed using highly curated reference GEMs or established reaction databases, the universal model is constrained by limited knowledge of biological reactions rather than reflecting the extensive metabolic space of nature. Developing GEM reconstruction pipelines that directly extract specific metabolic reactions from genomes would enable the exploration of more diverse and plausible metabolic engineering strategies. Altogether, the resources showcase 42,976 cases detailing the capacities of host strains for 235 bio-based chemicals under different aeration conditions using different carbon sources (5440 cases for B. subtilis , 9792 cases for C. glutamicum , 11,424 cases for E. coli , 3264 cases for P. putida , and 13,056 cases for S. cerevisiae ), alongside 1,925,500 cases detailing engineering strategies (784,774 heterologous reaction targets + 32,867 cofactor exchange targets + 613,863 targets identified by FVSEOF + 493,996 targets identified by iBridge). These resources provided in this study will be useful for selecting a host strain, improving innate metabolic capacity by constructing more efficient metabolic pathways through the introduction of heterologous metabolic reactions and cofactor exchanges, and identifying target reactions for up- and down-regulation to enhance the bio-based production of chemicals. Selecting the production strain needs to consider various factors, such as growth rate, maximum achievable or optimal/desirable cell concentration, culture condition (e.g., temperature, pH, nutritional requirement, and medium cost), ease of product purification, GRAS status, and others, in addition to the maximum theoretical and achievable yields presented in this study. While the presented maximum yields alone do not capture all dynamic aspects such as growth kinetics and process-specific conditions, they provide a valuable approximation for assessing the inherent metabolic capacity of different strains. As such, this resource serves as an essential reference for narrowing down candidate strains for further experimental validation. Moreover, when combined with additional criteria that reflect the conditions of interest, such as the high productivity achieved in fed-batch fermentations driven by rapid cell growth, the resource can guide strain selection and further cell factory design. Although our approach does not offer a complete solution, it will play an essential role in advancing towards the development of high-performing microbial cell factories."
} | 3,303 |
37754180 | PMC10526311 | pmc | 1,109 | {
"abstract": "With the continuous integration of material science and bionic technology, as well as increasing requirements for the operation of robots in complex environments, researchers continue to develop bionic intelligent microrobots, the development of which will cause a great revolution in daily life and productivity. In this study, we propose a bionic flower based on the PNIPAM–PEGDA bilayer structure. PNIPAM is temperature-responsive and solvent-responsive, thus acting as an active layer, while PEGDA does not change significantly in response to a change in temperature and solvent, thus acting as a rigid layer. The bilayer flower is closed in cold water and gradually opens under laser illumination. In addition, the flower gradually opens after injecting ethanol into the water. When the volume of ethanol exceeds the volume of water, the flower opens completely. In addition, we propose a bionic Venus flytrap soft microrobot with a bilayer structure. The robot is temperature-responsive and can reversibly transform from a 2D sheet to a 3D tubular structure. It is normally in a closed state in both cold (T < 32 °C) and hot water (T > 32 °C), and can be used to load and transport objects to the target position (magnetic field strength < 1 T).",
"conclusion": "4. Conclusions In this research, we studied the response characteristics of PNIPAM and PEGDA hydrogels. Since the PNIPAM hydrogel is temperature-sensitive and solvent-responsive, and the PEGDA hydrogel is insensitive to changes in both temperature and solvent composition, the bilayer structure can produce a bidirectional response to temperature and solvent composition. This can not only achieve a bidirectional transition of 2D-3D, but also maintain the 3D structure without continuous stimulations. Based on the bilayer structure, we fabricated a bionic Venus flytrap soft robot. The robot appeared as a normally closed tubular structure in either cold or hot water. With the addition of CNTs and Fe 3 O 4 , the robot possessed a photothermal conversion capability and magnetic response characteristic. The soft robot opened gradually when the active layer was illuminated, and the robot closed again after removing the laser. In addition, driven by an external magnetic field, the robot moved along a specific trajectory to the target position. We illustrated the intelligent transport capacity of the soft robot in both cold and hot water, and demonstrated the fixed-point transportation of the soft robot in the stomach model. We believe that this robot will promote the development of soft robots and provide a reference value for future targeted drug delivery technology in biomedicine.",
"introduction": "1. Introduction Intelligent robot technology has made great progress, and more and more robot technologies have leapfrogged traditional industrial applications and entered many fields, such as healthcare [ 1 , 2 , 3 , 4 , 5 , 6 , 7 ], education and entertainment, and intelligent transportation [ 8 , 9 , 10 ]. The development of flexible cordless robots has become crucial, as the high requirements for robots to operate in complex environments increase. Inspired by nature, researchers are working on smart materials with reversible deformation properties, such as stimulus-responsive hydrogels [ 11 , 12 , 13 , 14 , 15 ] and liquid crystal elastomers [ 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 ]. Among these several smart soft materials, hydrogels stand out because of their wonderful biocompatibility and deformation reversibility and have been being used by more and more researchers for the manufacture of soft robots [ 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. Poly (N-isopropyl acrylamide) (PNIPAM) hydrogel is a type of temperature-sensitive hydrogel [ 14 , 32 , 33 , 34 ]. The volume of PNIPAM hydrogel can mutate near its critical dissolution temperature (LCST = 32 °C). When the temperature is below the LCST (<32 °C), its volume expands rapidly, and when the temperature is above the LCST (>32 °C), it shrinks rapidly. The PNIPAM hydrogel prepolymer is cured to 2D sheets with a non-uniform expansion field inside by ultraviolet (UV) curing technology. These 2D sheets can deform into a 3D structure upon specific external stimulations. To maintain the 3D structure, even in the absence of stimulation, we cure a polyethylene glycol diacrylate (PEGDA) hydrogel layer on the PNIPAM layer. PEGDA hydrogel does not exhibit significant changes in its swelling ratio upon a change in temperature. Hence, for this bilayer structure, the PNIPAM layer acts as the active layer, while the PEGDA layer acts as the rigid layer. Additionally, the swelling ratio of PNIPAM hydrogel can respond not only to a change in temperature, but also to a change in the solvent composition of solutions ( Figure 1 ). The volume of PNIPAM reaches its maximum only in pure water or pure ethanol. In a mixed solution of ethanol and water, PNIPAM shrinks. In contrast, the swelling ratio of PEGDA hydrogel does not change significantly regardless of whether it is in either solution. In this research, we fabricated PNIPAM sheets with microchannels and took advantage of the temperature response characteristics of PNIPAM to achieve self-folding. Subsequently, an additional PEGDA layer was cured on the surface of the PNIPAM sheet to obtain the PNIPAM–PEGDA bilayer structure. We fabricated bionic flowers and bionic soft robot Venus flytraps based on this bilayer structure, which can be used to wrap and transport objects under the effect of an external magnetic field. This study provides a reference for the programmable deformation of smart soft materials, which has great prospects in intelligent transportation, bionics, and biomedicine fields."
} | 1,418 |
34683592 | PMC8537208 | pmc | 1,110 | {
"abstract": "The construction of superhydrophobic surfaces necessitates the rational design of topographic surface structure and the reduction of surface energy. To date, the reported strategies are usually complex with multi-steps and costly. Thus, the simultaneous achievement of the two indispensable factors is highly desired, yet rather challenging. Herein, we develop a novel structure engineering strategy of realizing the fabrication of a functionally integrated device (FID) with a superhydrophobic surface via a one-step spraying method. Specifically, silica nanoparticles are used to control the surface roughness of the device, while polydimethylsiloxane is employed as the hydrophobic coating. Benefitting from the adopted superhydrophobicity, the as-fabricated FID exhibits a continuous, excellent oil-water separating performance (e.g., 92.5% separating efficiency) when coupled with a peristaltic pump. Notably, a smart design of incorporating a gas switch is adopted in this device, thereby effectively preventing water from entering the FID, realizing thorough oil collection, and avoiding secondary pollution. This work opens up an avenue for the design and development of the FID, accessible for rapid preparation and large-scale practical application.",
"conclusion": "4. Conclusions In summary, a one-step spraying strategy is successfully demonstrated for constructing a superhydrophobic FID for efficient light oil-water separation. The superhydrophobicility of the fabricated FID is realized by virtue of a delicate surface structure modification by SiO 2 NPs and a coating of a low-surface-energy surface of PDMS. The evaluating tests prove that our one-step spraying strategy has excellent substrate and oil type versatility. Furthermore, an oil-water separation apparatus with a smart gas switch is set up using the as-prepared FID and a peristaltic pump. Benefitting from the adopted superhydrophobicity and designed gas switch, the apparatus shows controlled oil collection, preventing water from entering the FID, avoiding secondary pollution, and realizing thorough oil collection (i.e., separation efficiency of 92.5%). Although our FID shows excellent performance for light oil absorption, it is limited when absorbing heavy oil with high viscosity. The heavy oil absorption process suspends after a few minutes as the pores of our FID is blocked by the initial adsorbed heavy oil due to its poor fluidity. This work opens up an avenue for the design and development of the FID accessible for rapid preparation and large-scale applications.",
"introduction": "1. Introduction Oil spills occurring occasionally during oil exploration and transportation pose great threats to the marine ecosystem, which would cause catastrophe for ocean lives by poisoning animals/plants, damaging their habitat and slowing the reproductive rate [ 1 , 2 ]. The oily water also endangers human health in many ways, such as contaminated sea food and polluted freshwater resources. In this regard, oil-water separation techniques are in great demand for oil spill clean-up. Among them, the functionally integrated device (FID) that enables simultaneous oil absorption and separation has attracted intensive attention due to its merits of cost-efficiency, energy-saving, and environmental benignity [ 3 , 4 ]. A Functionally integrated device (FID) refers to a miniaturized system with specific function and structural design in one device, which can function in a sequential manner to complete designated missions. The reported FID is mainly focused on the miniaturized devices with a sealed or fully open design [ 5 , 6 ]. These devices have been demonstrated for applications in advanced research fields or for practical uses, including biomimicking devices, macroscopic supramolecular assembly, mini-generators, active oil absorption and separation, etc. [ 7 , 8 ]. Besides, most reported FIDs are only effective to light oils with low viscosity, yet do not work well for viscous heavy oil. The strong demand for absorbing heavy oil with high viscosity drives researchers to further improve the structure and functions of the FIDs. To enhance the absorption/separating efficiency of the FID, it is necessary for constructing superhydrophobic/superoleophilic surfaces realizing the thorough oil separation from oily water [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 ]. Typically, superhydrophobic surfaces can be achieved by surface modification and construction of topographic surface structure [ 21 , 22 ]. For example, a variety of top-down approaches (e.g., lithography [ 23 ], laser beam treatment [ 24 ]) and bottom-up approaches (e.g., self-assembly [ 25 ], 3D printing [ 26 ]) have been developed to construct micro- and nanostructures on surfaces. In addition, the adoption of coatings with low surface free energy is a popular means for surface modification and enhancement of hydrophobicity. Despite research progress, the current fabrication methods are limited due to the involvement of multiple step treatments and noble-metal raw materials (e.g., Au and Ag), making them inaccessible for rapid preparation (i.e., within several minutes) and large-scale practical application [ 27 , 28 , 29 ]. Therefore, the simultaneous achievement of the rationally designed surface structure and reduced surface energy are highly required, yet rather challenging. In this article, we demonstrate a facile, cost-effective spraying strategy for fabricating a superhydrophobic FID for light oil absorption using the polydimethylsiloxane-coated silica nanoparticles (SiO 2 @PDMS) composite supported by a copper foam substrate ( Figure 1 ). Specially, the SiO 2 NPs with high thermal stability and good mechanical strength are used to control the surface roughness of the device [ 30 ], while the PDMS with hydrophobic property, high chemical stability and good weatherability is employed as the low-surface-energy coating [ 31 , 32 ]. When connected to a peristaltic pump using a tube, the FID performs as a collector, realizing continuous and thorough oil collection from water. The oil-water separation ability of our FID under some simulated real conditions of use (i.e., salt water, droplet oil on the water, and oil underwater) as well as the oil-water separation recycling tests of our FID were evaluated to ascertain its performance stability under practical application environment."
} | 1,591 |
32807967 | PMC7502555 | pmc | 1,111 | {
"abstract": "Multifunctional living materials are attractive due to their powerful ability to self-repair and replicate. However, most natural materials lack electronic functionality. Here we show that an electric field, applied to electricity-producing Geobacter sulfurreducens biofilms, stimulates production of previously unknown cytochrome OmcZ nanowires with 1,000-fold higher conductivity (30 S/cm), and 3-fold higher stiffness (1.5 GPa), than the cytochrome OmcS nanowires that are important in natural environments. Using chemical imaging-based multimodal nanospectroscopy, we correlate protein structure with function, and observe pH-induced conformational switching to β-sheets in individual nanowires, which increases their stiffness and conductivity by 100-fold due to enhanced π-stacking of heme groups; this was further confirmed by computational modelling and bulk spectroscopic studies. These nanowires can transduce mechanical and chemical stimuli into electrical signals to perform sensing, synthesis and energy production. These findings of biologically-produced, highly-conductive protein nanowires may help to guide the development of seamless, bidirectional interfaces between biological and electronic systems.",
"discussion": "Discussion Our studies show that the OmcZ nanowires confer conductivity to electricity-producing biofilms that could explain high biofilm conductivity even in the absence of OmcS nanowires 7 . Lowering the pH of OmcS and OmcZ nanowires to pH 2 caused permanent structural transition to a more conductive β-sheet form. Denaturation typically destroys protein function. However, we found that at pH 2, the conductivity of OmcZ nanowires increases 100-fold at pH 2 ( Fig. 5b ) and nanowires maintain similar morphology ( Extended Data Fig. 9 ). Therefore, it is unlikely that the protein is denatured. Furthermore, the structural change was irreversible, confirming permanent transition to beta sheets at pH 2. Recent studies have emphasized the need for understanding the mechanism of environmentally-triggered conformational changes in proteins for the design of protein-based functional materials 38 . Our finding that coil/turns transform into β-sheets in microbial nanowires provides a mechanism for pH-induced conformational changes that reduce their diameter and enhance stiffness and conductivity. Lowering the pH can induce the formation of β-sheets in synthetic peptides 39 – 41 and the interaction of cytochrome c with SDS causes a transition to β sheets 42 . A similar effect could explain the pH-induced formation of β-sheets that reduces the diameter of microbial nanowires. Another mechanism that could account for the pH-induced reduction of nanowire diameter is a dehydration effect that causes proteins and other polymers to adopt a more compact structure at lower pH 43 – 45 . In particular, distal histidine coordinating to heme can stabilize the water molecule within the heme pocket by accepting a hydrogen bond 44 . At low pH, the histidine becomes protonated and swings out of the heme pocket, thus destabilizing the water occupancy and leading to dehydration 44 . Additional structural studies will help evaluate these possibilities. Our findings that highly-conductive microbial nanowires contain β-sheets contrast with all prior studies that have assumed that nanowires of both G. sulfurreducens WT and W51W57 strains are made up of PilA with a purely α-helical structure 28 . Solution NMR studies of the PilA monomer have indicated helical structure even at pH 5 28 . Therefore, combined with immunogold labelling and cryo-EM, our structural and functional studies suggest that highly-conductive microbial nanowires are composed of c -type cytochromes and not PilA. In summary, we have demonstrated the feasibility of manipulating the production and structure of protein nanowires to control their conductivity and stiffness. Our study established the capabilities of multimodal nanospectroscopic approaches for visualizing and quantifying the large-scale conformational changes in biomolecules. Precise control of electronic and mechanical properties of nanowires can be achieved via targeted environmental changes such as changing the pH which alters heme stacking or applying an electric-field. Previous studies have shown that an electric field can activate a synthetic gene circuit by creating an oxidizing environment 46 . Additional studies are required to evaluate if a similar mechanism plays a role in these natural systems. Our quantitative method to visualize conformation-induced functional changes is likely applicable to a variety of molecular systems. With OmcZ nanowires displaying a million-fold higher conductivity than synthetic biodegradable materials 1 , we anticipate these new materials will introduce several new features urgently needed for the next generation of bioelectronics, including low cost, ease of synthesis, lack of toxicity, mechanical flexibility, as well as scalable and facile processing with controlled biological properties 1 ."
} | 1,255 |
34075730 | PMC8336620 | pmc | 1,112 | {
"abstract": "Abstract For the practical applications of wearable electronic skin (e‐skin), the multifunctional, self‐powered, biodegradable, biocompatible, and breathable materials are needed to be assessed and tailored simultaneously. Integration of these features in flexible e‐skin is highly desirable; however, it is challenging to construct an e‐skin to meet the requirements of practical applications. Herein, a bio‐inspired multifunctional e‐skin with a multilayer nanostructure based on spider web and ant tentacle is constructed, which can collect biological energy through a triboelectric nanogenerator for the simultaneous detection of pressure, humidity, and temperature. Owing to the poly(vinyl alcohol)/poly(vinylidene fluoride) nanofibers spider web structure, internal bead‐chain structure, and the collagen aggregate nanofibers based positive friction material, e‐skin exhibits the highest pressure sensitivity (0.48 V kPa −1 ) and high detection range (0–135 kPa). Synchronously, the nanofibers imitating the antennae of ants provide e‐skin with short response and recovery time (16 and 25 s, respectively) to a wide humidity range (25–85% RH). The e‐skin is demonstrated to exhibit temperature coefficient of resistance (TCR = 0.0075 °C −1 ) in a range of the surrounding temperature (27–55 °C). Moreover, the natural collagen aggregate and the all‐nanofibers structure ensure the biodegradability, biocompatibility, and breathability of the e‐skin, showing great promise for practicability.",
"conclusion": "3 Conclusion An intelligent integrated e‐skin was developed, which was self‐powered and able to detect pressure, temperature, and humidity. Moreover, the application of biocompatible CA, and all NFs structure provided good biocompatibility and air permeability (air permeation 20%). In the detection range of 0–0.5 and 0.5–30 kPa, the detection sensitivity was 0.48 and 0.25 V kPa −1 , respectively, and 0.17 V kPa −1 within 30–135 kPa, owing to its spider‐web structure and internal bead‐chain fiber structure. Furthermore, the energy collected by the triboelectric (pressure‐sensing) layer was transferred to the e‐skin to allow for the temperature and humidity detection. In the dynamic range of 25–55 °C, the e‐skin presented a good thermal response (0.0075 °C −1 , R \n 2 = 0.99). Similarly, the water absorption performance of the CA material and the ant antennae mimicking structure indicated that the addition of an “antenna” to the humidity‐sensing layer of the e‐skin could provide a good humidity response at 25–85% RH (response time is 16 s, recovery time is 25 s). Based on the abovementioned multiple advantages, the excellent self‐powered, multifunctional, and low‐cost integrated intelligent e‐skin can be widely used in the fields of intelligent robots, interactive wearable devices, artificial prostheses, and disabled patients.",
"introduction": "1 Introduction As the largest organ of the human body, skin not only protects the human body from environmental hazards, but also ensures its timely perception of temperature, pressure, and vibration of the external environment. [ \n \n 1 \n , \n 2 \n \n ] In the era of Internet of Things, electronic skin (e‐skin) can even surpass the sensory functions of human skin, making it a basic data collection device, with wide range of applications in artificial prostheses, smart robots, wearable devices, health monitoring systems, and other fields. [ \n \n 3 \n , \n 4 \n , \n 5 \n , \n 6 \n \n ] Nonetheless, the development of a multifunctional, intelligent, and integrated e‐skin remains a key challenge. [ \n \n 7 \n , \n 8 \n , \n 9 \n , \n 10 \n \n ] However, mimicking biological structures (such as human skin, bird feathers, and plant leaves) may aid in the development of materials with excellent performance. [ \n \n 11 \n , \n 12 \n \n ] For the practical applications of an e‐skin material, the sensitivity, self‐power capability, biocompatibility, breathability, flexibility, lightness, and cost effectiveness are needed to be assessed and tailored simultaneously. [ \n \n 13 \n , \n 14 \n , \n 15 \n , \n 16 \n , \n 17 \n \n ] However, only a few e‐skins endowed with these integrated features have been reported till date. [ \n \n 18 \n , \n 19 \n \n ] \n E‐skins are complex arrays of soft sensors that achieve information by monitoring the conversion of various environmental stimuli, including temperature, humidity, and pressure, into real‐time and visualized electronic pulses. [ \n \n 20 \n , \n 21 \n , \n 22 \n \n ] Recently, in order to improve the overall performance of e‐skins, they have been endowed with special functions, such as electroluminescence, self‐healing, shape memory effect, fireproofing, waterproofing, and heat transfer. [ \n \n 14 \n , \n 23 \n , \n 24 \n , \n 25 \n , \n 26 \n \n ] Despite the continuous improvement and optimization of the abovementioned multiple functions, e‐skins that can truly imitate human skin and its multiple functions to achieve optimal integration are extremely rare; although this forms the basis for e‐skin to be truly intelligent with widespread applications. [ \n \n 7 \n , \n 27 \n , \n 28 \n \n ] Most e‐skins can detect only one type of external stimulus, significantly restricting their practical applications. Some e‐skins are able to detect more than one stimulus but have insufficient sensitivity, as manifested in their small detection range, slow response, and long recovery time. [ \n \n 7 \n , \n 27 \n , \n 28 \n , \n 29 \n \n ] For example, in a certain position of an artificial prosthesis such as fingertip, in order to achieve the multistimulation response of human skin, several sensors with different functions must work together. However, the planar array often causes a slight deviation of the sensing position, which is obviously fatal in high‐precision applications, and incorporation of more number of sensors leads to higher manufacturing cost. To address this issue, new conductive or luminescent materials have been developed to increase the sensitivity of e‐skin to pressure, temperature, or humidity, with the need for a complicated synthesis process. [ \n \n 24 \n , \n 30 \n \n ] In previous studies, an increased sensitivity was achieved by designing and modifying the structure of the material, such as imitating pyramid or skin‐inspired materials, to obtain a larger contact area under same pressure. [ \n \n 31 \n , \n 32 \n , \n 33 \n \n ] However, this method often requires etching or mold inversion, leading to a complicated production process and expensive manufacturing. Therefore, there is an urgent need to develop a multifunctional, sensitive, and truly intelligent e‐skin that can effectively detect pressure, humidity, and temperature, and can provide a basis for the true industrialization of flexible sensing. Moreover, practically, such a multifunctional e‐skin cannot be completely driven using a traditional battery because the pollution caused by battery electrolyte often causes health hazards and inconvenience associated with its replacement, charging, and recycling. [ \n \n 3 \n , \n 34 \n , \n 35 \n , \n 36 \n \n ] Therefore, it is critical to develop a self‐powered sensor array that can harvest human biomechanical energy based on the electrostatic coupling effect. [ \n \n 37 \n , \n 38 \n \n ] Although several self‐powered e‐skins have been reported that are tightly attached to the skin to efficiently collect biomechanical energy, most of them are produced with airtight or slightly toxic polymeric membranes such as fluororubber, polydimethylsiloxane, or other compact semiconductor membranes (such as gallium arsenide, titanium dioxide, and zinc oxide). [ \n \n 39 \n , \n 40 \n , \n 41 \n , \n 42 \n \n ] These materials may cause skin discomfort and even induce itching and inflammation, in particular, after long‐term contact with human skin. [ \n \n 15 \n , \n 43 \n , \n 44 \n \n ] Therefore, it is highly desirable to construct e‐skins using materials with high air permeability and biocompatibility as substrates. Notably, nanofiber (NF) materials with an inherently high surface area and high porosity can provide e‐skin with high sensitivity and excellent air permeability as priority objects. [ \n \n 15 \n , \n 45 \n \n ] Besides, considering that the e‐skin should operate normally within a period and then degrade into biocompatible and harmless bio‐electronic products, selection of a substrate material with good biocompatibility is urgently required. [ \n \n 46 \n , \n 47 \n , \n 48 \n \n ] Biomaterials such as collagen are used as friction material in triboelectrification because of their ability to lose electrons. [ \n \n 49 \n , \n 50 \n \n ] Furthermore, in order to apply biomaterials to e‐skin to significantly harvest biomechanical energy generated by the human body, a reasonable and sophisticated mechanism design is required. Herein, a multifunctional, self‐powered, sensitive, flexible, breathable, biodegradable, intelligent, and integrated e‐skin with multihierarchical and all‐NFs structure was successfully developed. Further, the e‐skin was used to effectively collect biomechanical energy and monitor whole‐body physiological signals of pressure, temperature, and humidity. The as‐fabricated intelligent integrated e‐skin consists of four parts: 1) a triboelectric (pressure‐sensing) layer composed of collagen aggregate nanofibers (CA NFs) and bead‐chain‐net poly(vinyl alcohol)/poly(vinylidene fluoride) nanofibers (B‐C‐N PVA/PVDF NFs). This layer mimics a spider web and bead‐chain structure and can simultaneously generate electricity by collecting biomechanical energy and detect pressure; 2) Multiwalled carbon nanotubes (MWNTs) doped in poly(3,4‐ethylenedioxythiophene): poly (styrenesulfonate) (PEDOT:PSS) as a conductive material is added to CA, and then spun to obtain CA‐PEDOT‐MWNTs NFs (CA‐P‐M NFs) as a temperature‐sensing layer; and 3) CA and acidified MWNTs are compounded and electro spun to prepare NFs with ant antenna structure (CA‐M NFs) for humidity detection. 4) The triboelectric collection system (LTC3588‐1) was used to collect biomechanical energy and provide energy for temperature and humidity sensing. The pressure sensitivity of the prepared e‐skin reached 0.48 V kPa −1 in the range 0–0.5 kPa, 0.25 V kPa −1 (0.5–30 kPa), and 0.17 V kPa −1 (30–135 kPa). Moreover, it exhibited short response time (2.1 s), recovery time (3.9 s), and excellent temperature coefficient of resistance (TCR = 0.0075 °C −1 , R \n 2 = 0.99) and sensitivity to temperature in the range of 27–55 °C. Even under 25–85% relative humidity (RH), the output signals showed a fabulous linear relationship. Owing to the complete NF structure, e‐skin exhibited an air permeation of 20.87 mm s −1 , which guarantees comfort during wearing. This study combined multifunctional, sensitive, and self‐powered material to ensure an intelligent and integrated e‐skin in order to detect various physiological or environmental signals. It could even ensure permeability and biodegradability, thereby helping to promote more practical and environmentally friendly applications of e‐skin in human–machine interfaces and artificial intelligence."
} | 2,765 |
38313553 | PMC10832036 | pmc | 1,113 | {
"abstract": "The operation of aerospace equipment is often affected\nby icing\nand frosting. In order to reduce the loss caused by icing in the industrial\nfield, it is an effective method to prepare superhydrophobic coatings\nby modifying nanoparticles with low surface energy materials. In order\nto explore a method of preparing a superhydrophobic surface that can\nbe popularized, a two-step spraying method was employed to create\na superhydrophobic coating. The surface was characterized by Fourier\ntransform infrared spectroscopy (FTIR) and field emission scanning\nelectron microscopy (SEM). The optimal preparation process was obtained\nby analyzing the surface contact angle data. The results showed that\nstearic acid was grafted onto the surface of TiO 2 by esterification\nreaction. The existence of long methyl and methylene hydrophobic groups\nin the tail of the stearic acid molecule made the modified TiO 2 hydrophobic. It is verified that water molecules have strong\nadsorption on the surface of unmodified TiO 2 . Stearic acid\nmolecules can reduce the interfacial energy in the system.",
"conclusion": "4 Conclusions Organic modification of\nnano-TiO 2 by SA was carried\nout by a two-step spraying method. TiO 2 was successfully\nmodified to superhydrophobicity, and the optimum preparation process\nwas 0.3 g TiO 2 . The hydrophobic mechanism was analyzed\nthrough the characterization of the modified nanoparticles and molecular\ndynamics simulation: SA reacted with the free hydroxyl on the surface\nof TiO 2 by the carboxyl group in the head, and the hydrophobic\neffect was achieved by the extremely long carbon chain of SA and the\nmethyl and methylene groups in the tail. The concentration changes\nof the water layer on the surface of SiO 2 and TiO 2 were analyzed, and the strong adsorption of water molecules on the\ntwo surfaces was verified. The hydrophilicity of the SiO 2 surface was slightly stronger than that of the TiO 2 surface.\nSA molecules can reduce the interfacial interaction energy in the\nsystem. The adsorption of water molecules on the TiO 2 surface\nis affected by van der Waals force and electrostatic force. van der\nWaals forces on the surface of SA acidification contributed most.",
"introduction": "1 Introduction Various superhydrophobic\nphenomena widely exist in our life. Generally,\ncontact angles higher than 150° and rolling angles lower than\n5° are called superhydrophobic surfaces. 1 , 2 Inspired\nby the superhydrophobic phenomenon in life, in recent years, more\nand more scientific researchers have carried out research on superhydrophobic\nsurfaces. 3 − 5 Because of its special contact behavior, it is widely\nused in various fields, including corrosion prevention, 6 , 7 self-cleaning, 8 , 9 antifouling, 10 , 11 and deicing. 12 , 13 Research shows that surface heterogeneity\nand roughness are two important factors that constitute superhydrophobic\nproperties. 14 , 15 Generally speaking, it is the\nmost common method to combine low surface energy materials and micro/nano\nstructures to obtain superhydrophobic surfaces. Up to now, many\nresearchers have used different methods to construct\nsuperhydrophobic surfaces, including vapor deposition, 16 electronic etching, 17 sol–gel process, 18 , 19 and phase separation. 20 In these methods, the surface roughness is first\nobtained by using tools and then modified by using low surface energy\nsubstances. He et al. 21 used silane coupling\nagent KH-570 to organically modify the inorganic filler titanium dioxide\n(TiO 2 ) on the glass substrate. The hydrophobicity and dispersion\nof the modified TiO 2 particles were enhanced, and the hydrophobic\neffect was the best when the mass fraction of KH-570 was 15%. Yang\net al. 22 used stearic acid as a low surface\nenergy modifier to modify the surface of TiO 2 and cellulose\nnanofiber composites (CNF). The surface of the CNF was coated with\nTiO 2 . Because the surface of CNF was rich in hydroxyl groups,\nit could be connected with the hydroxyl groups on the surface of TiO 2 by hydrogen bonding. TiO 2 promoted the nucleation\nand further growth of the CNF surface. Stearic acid (SA) removed hydrophilic\nhydroxyl groups in the TiO 2 /CNF system and introduced hydrophobic\nalkyl groups. The TiO 2 /CNF composite has a nano scale rough\nstructure on the surface, which makes it superhydrophobic. However,\nmost of the above methods for preparing superhydrophobic surfaces\nare complex and have a variety of materials, which is not conducive\nto wide development of the preparation process. Limited by many\nharsh requirements of experimental research, with\nthe increasing progress of computer development, researchers began\nto apply molecular dynamics to the field of superhydrophobic and gradually\nbecame a new trend, 23 , 24 that is, to study the interaction\nbetween interfaces at the micro level through molecular dynamics simulation\nand analyze the interaction mechanism between droplets and superhydrophobic\nsurfaces to understand the wettability of surfaces. 25 , 26 Werder et al. 27 simulated the behavior\nof liquid droplets in different carbon nanotube (CNT) radii. The contact\nangle of liquid droplets on the CNT surface in the simulation system\nwas calculated as a function and compared with the experimental observation\nresults of Gogotsi et al., 28 and it was\nconcluded that the interface has non wetting behavior. Rui et al. 29 studied the diffusion effect of water molecules\non the coal surface with different degrees of coalification. In the\nsimulation, the change of the contact angle is consistent with the\nexperimental results. The research results show that the increase\nof oxygen-containing functional groups leads to a decrease of the\ncontact angle. The hydrogen bond formed between water molecules and\ncoal provides an important contribution to the diffusion effect of\nwater droplets on the coal surface. Although these studies have analyzed\nthe superhydrophobic phenomenon at the microscopic level, there is\nless analysis of the specific units that play a role in this hydrophobic\neffect and the functional groups in the molecular chain are ignored. Nanomaterials ensure the construction of micronano structures on\nsuperhydrophobic surfaces. In recent years, research on superhydrophobic\ncoatings based on TiO 2 nanoparticles has attracted much\nattention. As a common nanomaterial, TiO 2 nanoparticles\nare one of the ideal superhydrophobic materials for constructing nanoscale\nrough structures. It is low cost, is easy to obtain, and has a wide\nrange of applications. Therefore, in this paper, superhydrophobic\nsurfaces were prepared by a two-step spraying method using SA and\nnano-TiO 2 particles with low surface energy as raw materials.\nThe simulation model was constructed based on the experimental materials,\nand the crystal structure was constructed based on the first principle\nof molecular dynamics. The bonding relationship between TiO 2 and SA was analyzed by density functional theory under the conditions\nof classical Newtonian mechanics, which was compared with the experimental\nresults of infrared characterization. At the same time, the superhydrophobic\nphenomenon of water molecules on the surface of TiO 2 molecules\ngrafted with SA molecules was studied at the molecular level, which\nprovided a basis for the study of microwettability and the analysis\nof macrowettability of superhydrophobic coatings and explored a scalable\npreparation method of superhydrophobic coatings.",
"discussion": "3 Results and Discussion 3.1 Surface Wettability Analysis Under\nthe dosage of 3 wt % SA in Table 1 , the CA of group A sprayed in one step increases with\nthe increase of the mass fraction of TiO 2 . When the addition\namount of TiO 2 is 0.3 g, the CA reaches the highest point,\nwhich is 152.85°, and when the addition amount increases to 0.4\ng, the CA is 106.70°. Consequently, in group B of two-step spraying,\nthe addition amount of TiO 2 is 0.3 and 0.4 g. The results\nshow that the optimal experimental scheme is when the addition amount\nof TiO 2 is 0.3 g in two-step spraying. Table 1 Sample Preparation and Contact Angle group TiO 2 dosage (g) dry CA (°) A 0.1 no 99.93 0.2 no 124.99 0.3 no 152.85 0.4 no 106.70 B 0.3 yes 153.15 0.4 yes 108.00 The hydrophobic effect of the prepared surface was\nexplored by\nchanging the amount of TiO 2 (0.1, 0.2, 0.3, 0.4 g). The\nchange trend and imaging diagram of CA are shown in Figure 1 . It can be seen from Figure 1 b that the blank\ngroup showed strong hydrophilicity due to the smoothness of the glass\nand the large number of hydroxyl groups attached to the glass surface\nin the air. The CA was 59.04°, and the surface was completely\nwetted after a period of time. When the surface is sprayed with a\nlayer of modified nano-TiO 2 particles, SA reacts with the\n−OH and −COOH groups on the surface of nanoparticles,\nthat is, esterification reaction, leaving a long −CH 2 chain as the tail, making the surface hydrophobic. With the increase\nof TiO 2 addition, the surface is completely covered by\nhydrophobic long chains. After the addition amount of TiO 2 was increased to 0.4 g again, SA was not enough to modify the nanoparticles,\nso the hydrophobic effect decreased. The experimental results of 0.4\ng of TiO 2 addition in group B also began to decrease the\nCA, which proved that too much TiO 2 addition will affect\nthe hydrophobicity. For the sake of experimental efficiency, in the\ntwo groups, it is more appropriate to add 0.3 g of TiO 2 in group B and bake for 1 h. Figure 1 Schematic diagram of CA change (a), CA\nimaging diagram of blank\ngroup (b), and optimal group CA imaging (c). (d) Comparison of an\nunmodified TiO 2 surface and a modified TiO 2 surface. As shown in Figure 1 d, the droplets directly penetrate and spread on the\nsurface of the\nunmodified TiO 2 surface while the droplets on the modified\nTiO 2 surface exhibit a nearly circular shape with a high\nCA. The results showed that TiO 2 without SA modification\nshowed strong hydrophilicity while the surface of the modified TiO 2 showed good hydrophobicity after SA modification. This also\nindirectly proves that SA successfully modified the nano-TiO 2 particles. 3.2 Prepared Surface Characterization 3.2.1 Micro Morphology Analysis of Nano-TiO 2 Particles In order to explore the hydrophobic mechanism\nof the prepared superhydrophobic surface, the micro morphology of\nnano-TiO 2 particles was characterized by SEM in Figure 2 . Spherical TiO 2 particles were analyzed under a scanning electron microscope\nwith magnification of 5 and 1 μm. In this experiment, the particle\nsize of TiO 2 was 20 nm. However, it can be found in the\nfigure that the nanoparticles were not uniform and there was agglomeration.\nThis is because the TiO 2 particles exposed to the air have\nhigh surface energy and the surface area is large under the elliptical\nshape, which promotes the aggregation. Figure 2 (a) Micromorphology of\nTiO 2 nanoparticles. (b) High-rate\nmicrostructure of TiO 2 nanoparticles. The roughness section depth of the coating surface\nis shown in Figure 3 a, with the highest\npeak value of 3.37 μm and the lowest valley value of −2.24\nμm. The section Ra value measured within the length of 4 mm\nof the surface is 0.744 μm. The surface morphology of nano-TiO 2 particles modified by SA is shown in Figure 3 b,c. Compared with Figure 2 , the shape of TiO 2 has not changed\nbut the surface is more uniform as a whole. Most particles are about\n25 nm in diameter, and there are a lot of gaps between the particles,\nso air can fill them. Under the action of SA, a low surface energy\nmaterial, the agglomeration of TiO 2 particles is prevented\nand the surface free energy is reduced. Figure 3 Depth roughness of the\ncoating surface (a). Micromorphology of\nthe superhydrophobic TiO 2 surface (b). High-rate microstructure\nof superhydrophobic TiO 2 surface (c). 3.2.2 Fourier Transform Infrared Spectroscopy\n(FTIR) Analysis Characteristic groups on the TiO 2 surface were analyzed by FTIR. In Figure 4 , the absorption peaks of SA appeared at\n2915 and 2848, corresponding to the stretching vibration peaks of\nthe C–H bond in −CH 3 and −CH 2 , respectively. The absorption peak also appeared in the TiO 2 /SA spectrum at the same position, indicating that SA had\nbeen linked to the surface of the TiO 2 particles. The absorption\npeaks at 1698 on these two spectral curves correspond to the stretching\nvibration of the C=O bond in the −COOH group, indicating\nthat SA was grafted onto the surface of TiO 2 by esterification\nreaction between the carboxyl group of the SA molecule and the hydroxyl\ngroup on the surface of TiO 2 . The existence of long methyl\nand methylene hydrophobic groups in the tail of SA makes the modified\nTiO 2 hydrophobic. Figure 4 Fourier infrared spectrum curve. 3.2.3 Delayed Icing Experiment on a Superhydrophobic\nSurface The delayed icing test tested the icing of 5 μL\ndroplets on two surfaces at −10 and −15 °C. Because\nicing is affected by ambient temperature and humidity, the ambient\nair humidity is maintained at 68%, each group of tests is set five\ntimes, and the average value is taken as the final experimental data. Figure 5 records\nthe delayed icing process of droplets on two surfaces. It can be found\nthat no matter what kind of surface the droplet is on, it is completely\ndivided into two stages, from the beginning to the icing. For the\nfirst stage, first, the droplet exchanges heat with the surrounding\natmosphere under the action of the ambient temperature. When the droplet\nsurface is cooled to 0 °C, due to the crystal nucleus being more\nlikely to appear on the solid/liquid/gas three-phase contact surface,\nthe droplet begins to undergo heterogeneous nucleation in a very short\ntime. For the second stage, the ice layer of nucleation and crystallization\ngradually grows up from the bottom contact interface. Finally, the\ndroplet completely freezes and expands in volume. When the ambient\ntemperature is −10 °C, the droplet cooling time on the\nsurface of the ordinary glass slide is 17s and the time required for\ncomplete icing is 10 s. On the TiO 2 /SA superhydrophobic\nsurface, the droplet cooling time is 721 s, the time required for\ncomplete icing is 33 s, and the icing is delayed by 17 times. Figure 5 Icing process\nof an ordinary glass carrier (a) and a TiO 2 /SA superhydrophobic\nsurface (b) at −10 °C. Icing process\nof an ordinary glass carrier (c) and a TiO 2 /SA superhydrophobic\nsurface (d) at −15 °C. In order to study the influence of different temperatures\non the\nicing time, an icing test at −15 °C was set up. It can\nbe found that after many tests, when the ambient temperature is reduced\nto −15 °C, the droplets on the surface of the ordinary\nglass slide are frozen, the cooling time is very short, and the freezing\ntime is 13 s; the delayed icing time of the TiO 2 /SA superhydrophobic\nsurface is 225 s. The above results show that the TiO 2 /SA\nsuperhydrophobic coating exhibits fairly good anti-icing ability. 3.3 Molecular Dynamics Simulation Analysis 3.3.1 Hydrogen Bond Analysis and SA Adsorption\nModel The mechanism of SA-modified TiO 2 nanoparticles\nwas determined by molecular dynamics simulation. In the simulation\nsystem, hydrogen bonds are generated among water molecules in the\nwater layer and between water molecules and TiO 2 crystals.\nStudies have shown that the hydrogen bond length is 1.1–2.5\nÅ, 30 , 31 As shown in Figure 6 a, the motion process of the water molecular\nlayer on the TiO 2 surface is simulated. Hydrogen bonds\nformed between water molecules and water molecules and between water\nmolecules and TiO 2 after the 300 ps simulation step. The\naverage hydrogen bond length between water molecules is 2.16 Å,\nand the average hydrogen bond length between water molecules and TiO 2 is 2.41 Å, both within a reasonable range, which proves\nthe correctness of the model in this paper. Figure 6 (a) Hydrogen bonds in\nthe model. SA adsorption simulation: (b)\nenergy changes. (c) Initial state. (d) Adsorption state. The adsorption behavior of three SA molecules on\nthe TiO 2 crystal was simulated in ethanol solution. Figure 6 b shows that the\nenergy of the model system\ntends to be minimized after 500 ps of simulation and the solution\nis in equilibrium. Random distribution of SA molecules in solution\nat the initial state is shown in Figure 6 c. Under the COMPASS force field, due to\nthe adsorption of TiO 2 on the surface molecules, SA molecules\nchange from being in a free state to being adsorbed on the TiO 2 surface, the alkyl chain in the molecule is nearly vertical,\nand the tail is far from the surface, as shown in Figure 6 d. 3.3.2 Analysis of Water Molecule Concentration\nDistribution on the SiO 2 /TiO 2 Surface Figure 7 demonstrates\nthe concentration distribution of the water molecular layer on the\nsurface of SiO 2 and TiO 2 and analyzes the influence\nof adsorption on the movement characteristics of the water molecular\nlayer. On the Z axis, it can be seen from Figure 7 a-1 that the initial\nwater molecules of the model were distributed in the range of 20–50\nÅ from the surface of SiO 2 . Under the nonbonding force,\nthe water molecular layer was adsorbed to 20–30 Å from\nthe surface after 300 ps movement ( Figure 7 a-2) and the maximum concentration was 8.70.\nIn Figure 7 b-1,b-2,\nwhen the water layer on the surface of TiO 2 reaches the\nequilibrium state, the maximum concentration curve is located at 25\nÅ away from the surface and there are still water molecules at\n24–40 Å, which proves that the hydrophilicity of the SiO 2 surface is slightly stronger than that of the TiO 2 surface. At the same time, on both surfaces, water molecules in\nthe X and Y axes fluctuate only\nin a fixed, small range. Figure 7 Concentration distribution of the water molecular\nlayer on the\nSiO 2 surface (a)/TiO 2 surface (b). 3.3.3 Root Mean Square Displacement and Diffusion\nCoefficient Water droplets with a radius of 1 nm were further\nconstructed to simulate the wetting process on the TiO 2 surface and the SA molecular acidification surface. After the dynamic\nsimulation of 300 ps time step, the energy of the two models is stable\nand no violent fluctuation is observed, which proves that the model\nreaches the equilibrium state. The mean square displacement of nanowater\nmolecular clusters is used to illustrate the molecular trajectory.\nAs shown in Figure 8 , during the process from the initial state to the equilibrium state,\nthe mean square displacement of water molecular clusters on the TiO 2 surface is almost twice larger than that on the acidification\nsurface of the SA molecule, indicating that water molecular clusters\nhave high activity and a wide displacement range on their surfaces.\nSecond, the diffusion coefficient of water molecular clusters can\nbe expressed by the fitted function of root-mean-square displacement,\nand the formula is as follows: 1 In the formula, D is the diffusion coefficient, Na denotes the number\nof molecules, r denotes the position of molecules\nat a certain moment, and t denotes the time step. Figure 8 Energy\nchange of acidified surface composed of the TiO 2 surface\n(a), SA molecule (b), and mean square displacement (c). After fitting, the function of water molecular\nclusters on the\nTiO 2 surface and acidified surface is Y 1 = 1.95 X – 21.47, Y 2 = 1.58 X + 28.86, respectively. D 1 and D 2 are 3.25\n× 10 –9 and 2.63 × 10 –9 m 2 /s, respectively. The smaller diffusion coefficient\nof the acidified surface means a weaker diffusion effect and smaller\nmobility of water molecular clusters. Figure 9 shows the\nspecific conformation of water molecules on the TiO 2 surface\n(a) and acidified surface (b) during the simulation process. Within\na fixed time step, the droplets on the TiO 2 surface quickly\nspread and cover the surface while the droplets on the acidified surface\nstill remain elliptical after a series of bounces, which indicates\nthat SA molecules successfully modify TiO 2 from hydrophilicity\nto hydrophobicity. Figure 9 Wetting process of water droplets on Ti O 2 ( a ) and SA (b) acidified surfaces. 3.3.4 Interaction Energy between SA Layer and\nWater Cluster, TiO 2 Surface, and Water Cluster Model System The acidification surface was constructed by optimized SA molecules.\nThe adsorption layer was a water molecular cluster. After the adsorption\nlayer reached equilibrium with the substrate surface, the interaction\nenergy of the two model systems of the SA layer-water cluster and\nthe TiO 2 surface water cluster was calculated, respectively.\nThe calculation formula is as follows: 2 Among them, E total , E surface ,\nand E polymer represent the total energy\nof the system, the surface energy of the substrate, and the energy\nof the adsorbate in the system, respectively, S represents\nthe contact area, and E expresses the interfacial\ninteraction energy. The smaller E indicates a more\nstable adsorption. SA, as a low surface energy material, can reduce\nthe surface free energy. After calculation, the interaction energy\nof the acidified surface (−0.0317 kcal/mol/Å 2 ) is lower than that of the TiO 2 surface (−0.297\nkcal/mol/Å 2 ). This means that the modification of\nSA molecules significantly reduces the adsorption of water molecular\nclusters on the surface, which also confirms the hydrophobic phenomenon.\nIn addition, as shown in Figure 10 , van der Waals force (−0.138 kcal/mol/Å 2 ) and electrostatic force (−0.159 kcal/mol/Å 2 ) jointly dominate the adsorption of water molecules on the\nTiO 2 surface. For the SA-acidified surface, the contribution\nof van der Waals force (−0.0227 kcal/mol/Å 2 ) played a major role. Figure 10 Interaction energy of different surface systems."
} | 5,424 |
31492882 | PMC6731289 | pmc | 1,114 | {
"abstract": "Anaerobic degradation (AD) of heterogeneous agricultural substrates is a complex process involving a diverse microbial community. While microbial community composition of a variety of biogas plants (BPs) is well described, little is known about metabolic processes and microbial interaction patterns. Here, we analyzed 16 large-scale BPs using metaproteomics. All metabolic steps of AD were observed in the metaproteome, and multivariate analyses indicated that they were shaped by temperature, pH, volatile fatty acid content and substrate types. Biogas plants could be subdivided into hydrogenotrophic, acetoclastic or a mixture of both methanogenic pathways based on their process parameters, taxonomic and functional metaproteome. Network analyses showed large differences in metabolic and microbial interaction patterns. Both, number of interactions and interaction partners were highly dependent on the prevalent methanogenic pathway for most species. Nevertheless, we observed a highly conserved metabolism of different abundant Pseudomonas spp . for all BPs indicating a key role during AD in carbohydrate hydrolysis irrespectively of variabilities in substrate input and process parameters. Thus, Pseudomonas spp . are of high importance for robust and versatile AD food webs, which highlight a large variety of downstream metabolic processes for their respective methanogenic pathways.",
"conclusion": "Conclusions Metaproteome analyses of 16 agricultural large-scale BPs enable a deeper insight into AD. Temperature, pH, substrate and VFA concentration were identified as main drivers for metaproteomic profiles. Comprehensive correlation analyses enabled the identification of potential marker organisms for defined process conditions, such as Petrotogaceae for high temperatures. Moreover, monitoring the MCR of Methanosarcinales could be a suitable biomarker to recognize ongoing acidification and avoid process failures. BPs clustered similar on both, species and protein level but differed on functional level, indicating a high resilience and flexibility of the microbial community. BPs were classified to acetoclastic, hydrogenotrophic or a mixture of both pathways, while methylotrophic methanogenesis was of minor importance. Such classifications could be meaningful for upcoming studies. Network analysis of each methanogenic pathway revealed that microbial interaction patterns widely differed among BPs reflecting large differences in metabolic processes in the prevalent methanogenic pathway. However, all methanogenic pathways have in common that Pseudomonas spp . were main drivers in hydrolytic processes indicating their versatile metabolism in wide-ranging process conditions and substrate variabilities. Although no interactions between hydrogenotrophic methanogens and SAOBs were observed in the networks, our data emphasize the importance of syntrophic acetate-oxidation for hydrogenotrophic methanogenesis. In addition, network analysis underline the role of D. tunisiensis and P. bacterium 1109 during AD of various agricultural substrates. Further research is required to obtain a deeper understanding of anaerobic degradation and to include this knowledge in control technology to make biogas production more flexible and efficient.",
"introduction": "Introduction Climate change and rising energy consumption trigger innovation in the global energy market. New approaches and expansion of existing renewable energy technologies such as biogas, solar, water and wind are needed to facilitate goals in reducing carbon dioxide emissions 1 . In contrast to other renewable energy sources, microbial biogas production is based on anaerobic fermentation of a wide variety of substrates. Different members of a microbial community accomplish the basic steps of anaerobic degradation (AD), which are hydrolysis, acidogenesis, acetogenesis and finally methanogenesis 2 – 4 . Hydrogenotrophic, acetoclastic and/or methylotrophic methanogenesis occur in biogas plants (BPs) alone or in combination with each other, mainly depending on abiotic parameters such as substrate, pH, temperature, ammonia content or reactor type 5 – 10 . The majority of BPs in Germany are fed mainly with agricultural products such as corn, grass and few other crop plants, as well as manure and other agricultural waste. Although food residues are less common, they are of economic interest as these substrates are propitious alternative substrates or substrate additions. Therefore, microbial communities in BPs have to overcome variations of each process parameter and substrate by metabolic readjustment to enable a stable and high biogas production. Hence, detailed knowledge of involved microorganisms as well as their metabolic processes is crucial for future optimization of AD. In previous studies, mainly species composition and abundances of BPs were characterized on 16S rRNA gene level 11 – 16 , which allows phylogenetic affiliation followed by predicting their potential metabolic functions. Metatranscriptomics or metaproteomics are capable tools to link present transcript or gene expression level to both, metabolic functions and phylogenetic affiliations in complex microbial community compositions 17 . Recently, high metabolic activity of members of the kingdom Archaea in a BP were obtained by 16S rRNA gene amplicon and metatranscriptomics approach 18 or a combination of metagenome and metaproteome analyses 19 . Metaproteome analyses of different BPs also provided important information how different process parameters shape proteomic profiles 5 . Methodologies for analyzing and interpreting omics data have rapidly changed in the last decade. Tools for network analyses such as MENA 20 , SparCC 21 or CoNet 22 are frequently used to predict interactions between microorganisms. Such network calculations are mainly based on 16S rRNA gene amplicon abundance data and therefore anaerobic degradation-based findings have to be interpreted with caution, as metabolic functions are difficult to predict. Other tools like STRING 23 focus on discovery of protein-protein interaction networks for explanation of microbial interactions. Unfortunately, the relatively low number of reliable database entries limits those tools for AD. In contrast to most other studies, we analyzed not only the metaproteome of a single BP but sixteen large-scale BPs. Furthermore, we measured the metaproteome of five independent replicates for each BP (same time point) to produce robust results. Main goals of this study are (i) to identify most important parameters driving the AD on protein level, (ii) to group the BPs according to their metaproteome (taxonomic and functional profiles), (iii) to arrange the BPs corresponding to their prevalent methanogenesis pathway and (iv) to identify microbial key players and their interaction patterns by a metabolic and microbial network analysis. The overall aim was to gain a better understanding of the metabolic processes during AD with a focus on methanogenesis as well as explore possibilities of metaproteomics for practical applications.",
"discussion": "Discussion We observed for almost all metabolic steps of AD high protein abundances for transport, glycolysis/gluconeogenesis and methanogenesis. Proteins for all different methanogenic pathways were present in each BP, indicating that a mixture of methanogenic pathways simultaneously convert agricultural substrates to biogas in large-scale BPs. Nevertheless, majority of BPs was dominated by either hydrogenotrophic or acetoclastic methanogenesis as prevalent pathway. In comparison with other approaches like stable isotope probing combined with a nucleic acid approach (DNA-SIP or RNA-SIP), metaproteomics enable accurate access to microbial phylogeny, function and its abundances in any scale of the bioreactors. In contrast, RNA- or DNA-SIP approaches are well established for small-scale bioreactors 27 – 30 , but expensive and artificial as results cannot easily be upscaled. Robustness and reproducibility of our approach is supported by the results for the five replicates of a BP: in only two cases one replicate was assigned to another main methanogenic pathway compared to remaining four replicates. This approach could also be attractive for plant operators. If they know the main methanogenic pathway, they could adapt relevant process parameters to optimize the AD processes in their digester to increase biogas production. For instance pH adjustment is biogas rate limiting, if hydrogenotrophic methanogenesis prevails 31 . Nevertheless, for further critical review of our approach, more metaproteome data sets and its associated process parameters from large-scale BPs have to be analyzed. Numbers and compositions of BP clusters differed on taxonomic and functional level (Fig. 2 ), suggesting a highly adaptive metabolic network of various active key members in a complex microbial community. Even if similar microbial communities were present, the members fulfilled different ecological processes resulting in defined interaction patters. This in turn suggests a high grade of specialization, which is dependent on a variety of factors, such as process parameters, presence and absence of potential interaction partners or bioavailability of substrates. Therefore, to gain a preferably comprehensive understanding of the AD processes, the need for holistic approaches like metaproteomics are profitable. Results of microbial community composition on 16S rRNA gene for instance has the risk of underestimating the importance and activity of especially methanogens compared to proteomics based approaches 10 . Environmental variables, such as substrate or temperature, are known to affect community composition during AD as revealed by nucleic acid based analyses 6 , 10 , 32 . Similarly, a variety of parameters (e.g. temperature, pH, feedstock, VFA) were found to significantly affect the metaproteome profiles of BPs what is in accordance with other studies 5 , 33 . Some already suggested suitable marker organisms for different types of BPs, as well as potential biomarkers for AD in BPs could be approved 5 . In addition, some potential new biomarkers were described here. In our study is D. tunisiensis a promising candidate for a marker organism of thermophilic BPs. This species is known to be a key player for hydrolysis in BPs, and its genome encode a variety of genes associated with complex polysaccharide degradation 34 . Positive correlations of glycolytic proteins for D. tunisiensis indicate a high metabolic activity of Petrotogaceae during degradation of complex carbohydrates under thermophilic conditions, as observed in other studies 35 . Peptococcaceae bacterium 1109 was highly abundant in most BPs and positively correlated to pH. Therefore, a decrease in pH (as observed during acidification) could possibly be detected by a decrease of proteins from this species. In addition, observed correlations of pH and different species were solely positive. This is surprising as all BPs were single-stage fermenters and negative correlations between hydrolyzing organisms and pH were expected, due to their lower pH optima. This findings indicate a highly adapted bacterial community, which is able to hydrolyze substrates efficiently even at high pH ≥ 7.7 (see Supplementary Table S2 ) and possibly overcome bottlenecks during rate-limiting hydrolysis. Many different pathways seem to be influenced by VFA concentration (Supplementary Table S4 ), which is in line with previous results as high VFA concentrations inhibit both, hydrolysis and methanogenesis 36 . Nevertheless, it has been reported that high acetate (i.e. about 2400 mg/L) and butyrate concentrations (about 1,800 mg/L) seem to have no effect on methanogenesis, while high propionate concentrations (900 mg/L) are more critical 37 . Higher proportions of propionate have been observed for three BPs (BP 10, BP 12 and BP 16), that were assigned to HyMe. Although there are more hydrogenotrophic BPs, it is noticeable that these three BPs showed lower abundances of known SAOB, such as T. phaeum or S. schinkii (Supplementary Table S1 ). Acetate oxidation by syntrophs can be a rate limiting step of hydrogenotrophic methanogenesis 38 , and therefore a lack of SAOB during hydrogenotrophic methanogenesis is likely to lead to an accumulation of VFAs, which in turn can inhibit methanogenic activities. Therefore, monitoring the abundance of proteins from SAOBs seems to be a suitable way to check the stability and performance of hydrogenotrophic BPs. In addition, our results suggest that MCR from Methanosarcinales is as promising biomarker candidate for acidification. Acetoclastic organisms in combination with the presence of mixed acid fermentation enabled a fast metabolization of different VFAs. Acetoclastic microorganisms have to metabolize more substrate to obtain the same energy as hydrogenotrophic microorganisms 39 , 40 . As differences in process parameters as well as used substrates could also strongly influence VFA concentrations 41 , 42 , future studies should evaluate whether VFA monitoring is useful for all types of BPs. Network analyses approach revealed distinct metabolic and microbial interaction patterns for each methanogenic pathway. Different number of nodes and edges for each network indicated highly complex microbial interactions patterns, with HyMe being the most complex one. In contrast, BoNe showed the fewest number of nodes and edges and could be considered to be the least complex network. Possibly, the prevalence of two methanogenesis pathways prevent the organisms from a too deep specialization and consequently, less organisms and interactions are necessary for a stable AD process. In addition, network analyses support the importance of Defluviitoga tunisiensis , as well as P. bacterium 1109 for AD. As no common interactions of both species with other microorganisms could be observed among all networks, it can be assumed that their high flexibility for different interaction partners, may lead to their generalistic behavior. In contrast, highly specialized organisms can substantially contribute to AD such as 15 different Pseudomonas spp ., which shared 37 edges among themselves over all networks and BPs. Stable expression for most proteins of the different Pseudomonas spp . (Supplementary Fig. S6 ), indicating a highly conserved carbohydrate metabolism of those species, even if process parameters are different (Supplementary Fig. S2 ). Additionally, Pseudomonas spp . seems to prefer the Entner-Doudoroff (ED) pathway instead of the Embden-Meyerhof (EM) pathway (Supplementary Fig. S5 ) for glucose degradation, even if ED is energetically unfavorable. This finding may be linked to the lack of phosphofructokinase for most Pseudomonas spp ., which is the key enzyme of EM 43 . Different Pseudomonas spp . metabolize glucose through ED which have been previously shown in lab based experiments of glucose metabolism 44 , 45 . Results of this study can be used to optimize process conditions to facilitate metabolic activity of Pseudomonas spp . in agricultural BPs. Even if no direct interactions between SAOBs and hydrogenotrophic methanogens could be observed, the numbers of interactions of SAOBs in each of the three networks clearly indicate the great importance of SAOBs for HyMe, while they seem to play a minor role during AcMe. During AcMe a big proportion of produced acetate is metabolized by acetoclastic methanogens, and therefore SAOBs have to compete with them. These findings were supported by higher mean protein abundances for T. phaeum (AcMe: 0.2%, HyMe: 1.0%, BoMe: 0.6%), as well as S. schinkii (AcMe: 0.3%, HyMe: 4.7%, BoMe: 1.8%)."
} | 3,931 |
36133247 | PMC9419703 | pmc | 1,115 | {
"abstract": "The triboelectric effect is one of the most trending effects in energy harvesting technologies, which use one of the most common effects in daily life. Herein, an impervious silicone elastomer-based triboelectric nanogenerator (SE-TENG) is reported with a micro roughness-created silicone elastomer film and Ni foam as triboelectric layers with opposite surface charges. The surface roughness modification process was performed via a cost-effective soft lithography technique using sandpaper. The replicated film was then used as the negative triboelectric layer and porous Ni foam was used as the positive triboelectric layer. The device exhibited the advantage of high stability due to the porous nature of the Ni foam, which could not damage the roughness pattern of the elastomer film. The device generated a maximum electrical output of ∼370 V/6.1 μA with a maximum area power density of ∼17 mW m −2 at a load resistance of 1 GΩ. Furthermore, the SE-TENG device was packed using polyethylene to protect it from humidity and made to be a water-resistant SE-TENG (WR-SE-TENG). The device was stable under different percentages of relative humidity, showing a uniform electrical output in the range of 10% RH to 99% RH. This proves that the packing is highly resistant against moisture and humidity. The device was also used for demonstrating its capability in powering small electronic components such as charging commercial capacitors, glowing LEDs and powering wrist watches. Further, the WR-SE-TENG device was used to scavenge bio-mechanical energy from human motions and also used for a real-time application of zero power consuming/self-powered pressure sensors. As an active sensor, the device showed linear sensing behavior and a sensitivity of 0.492 μA kPa −1 .",
"conclusion": "4. Conclusions In summary, a highly reliable impervious SE-TENG was fabricated successfully using a roughness-created silicone elastomer and Ni foam as active layers. The cost-effective soft lithography technique was introduced for creating micro-roughness on the silicone elastomer film. With a similar configuration, two other TENG devices made of Al-SE and Cu-SE were fabricated and the electrical output was analyzed. Among the three devices, SE-TENG made of Ni foam showed a higher electrical output ∼370 V/6.1 μA with a maximum area power density of ∼17 mW m −2 at 1 GΩ load resistance. The device performance in real-time was demonstrated by charging commercial capacitors, glowing LEDs and powering a digital watch. The durability of the device was analyzed by performing a stability test for a period of 2000 s. To make the device to work actively in the harsh and humid environmental conditions, the device was fabricated as a water-resistant SE-TENG by packing it with polyethylene films and laminated tightly and demonstrated its water-resistant capability by actuating it inside a water tub. The electrical output comparison with respect to the packed and unpacked SE-TENG device was analyzed and performed a humidity test of actuating the device under various relative humidity (10% RH to 99% RH). The test showed that the performance of the TENG device was not affected due to humidity as well as the packing gives a stable protection to the SE-TENG device from humidity. Finally, the device was used to harness the bio mechanical energy from daily human motions such as finger, hand and foot tapping. Also, the device was demonstrated for the potential real-time application of a self-powered/zero power consuming pressure sensor, showing a good sensitivity of 0.492 μA kPa −1 and the correlation coefficient of 0.9985. The above experiments and demonstrations prove that SE-TENG is a promising candidate to be used for measuring variable pressures in harsh environments such as fluid pressure, gas pressure and water level indications.",
"introduction": "1. Introduction Energy harvesting technologies have recently been the trending topic, which have potential as solutions to the global energy crisis. In recent years harvesting energy from sustainable energy sources such as wind, 1 water waves, 2,3 and bio-mechanical 4,5 and vibrational 6,7 sources has gained immense attention due to their simple energy conversion ability. Among them, energy harvesting from mechanical motions has expanded due to the abundant mechanical energy in daily life. Triboelectric nanogenerators (TENGs) are considered to be promising candidates for harvesting mechanical energy from various sources such as tides, 3 ocean waves, 8,9 human motions, 10 and strain. 11 Because of their simple design, cost-effective fabrication, highly reliable output performance and durability, many researchers are working towards their commercialization. 4 The first ever TENG was reported by Z. L. Wang's group in 2012 for harvesting mechanical energy 12 and utilized for small-scale energy harvesting. Currently, TENGs have been improved further with technological advancements, and used for energy harvesting and various applications in self-powered sensors. 13–15 In addition, many researchers are working to improve the output performance of TENGs by modifying the device structure, improving the surface contact area and doping. 16,17 In general, TENGs function via two major effects called triboelectrification and electrostatic induction. When two dissimilar triboelectric materials come into contact with each other either in vertical contact and separation or lateral sliding motion, a surface charge develops on the layers and drives the electrons to flow through an external load with the production of a potential drop. 18 This shows that the proper selection of active layers is the major paradigm in designing high performance TENG devices. Next, to design the high performance TENG devices, two key factors needs to be considered: (i) surface charge generation increases with an increase in applied pressure and (ii) surface charge increases with an increase in contact points according to the Volta-Helmholtz hypothesis. 19 The most suitable route to multiply the contact point is to create surface roughness on the active layers. There are several techniques to create surface roughness such as photolithography, reactive ion etching, inductively coupled plasma etching, and thermal imprint lithography. 20–22 However, these techniques involve high cost, careful operation, and are time consuming. 23 Similarly, a major drawback of TENGs is their stability under moist and humid environmental conditions. 24 There is strong evidence proving that the output performance of TENGs is reduced drastically due to humidity, which hampers their full-fledged commercialization for daily use. 25–27 Recently, TENGs have been designed to humidity resistant by developing their active layers via super hydrophobic techniques. 28 However, still, the selection of materials, cost and time consuming processes make this a complicated process. To overcome the above limitations, herein, we report a water-resistant silicone elastomer-based triboelectric nanogenerator (WR-SE-TENG), which is impervious in nature. The device was configured with a metal-dielectric configuration using nickel (Ni) foam and a micro roughness-created silicone elastomer film as the active triboelectric layers. The silicone elastomer film was fabricated via an easy and cost-effective soft lithography technique using commercial micro roughness sandpaper. Ni foam, which has a porous network structure, contributes to high surface roughness, and thus was used directly as the positive triboelectric layer. Ni foam having a porous nature could not damage the roughness on the silicone elastomer layer during the contact and separation process, making the output stable for a prolonged duration. The device was laminated using a polyethylene sheet with the help of a pouch laminator inside a glove box. The packing resists water and humidity, which affects the performance of the TENG. The output performances of the silicone elastomer (SE-TENG) device with different positive layers such as aluminum (Al), copper (Cu), and nickel (Ni) foam were also compared. Among them, SE-TENG made of Ni foam as the positive layer generated a maximum voltage and current of ∼370 V/6.1 μA with a maximum area power density of ∼17 mW m −2 at 1 GΩ load resistance. The stability of SE-TENG was analyzed for 2000 s and the stability of WR-SE-TENG was analyzed under various relative humidity ranging from 10% RH to 99% RH. This approach proves that the device was protected from humidity as well as stable in its output performance without a decrease in its efficiency. The device was used successfully for charging commercial capacitors, glowing light emitting diodes (LEDs) and powering up electronic wrist watches. Further, the WR-SE-TENG device was used for scavenging bio-mechanical energy from human motions such as finger tapping, palm tapping and foot tapping and LED glowing under bio-mechanical motions. The above experiments and tests prove that the WR-SE-TENG device is a promising sensor device to work under harsh and humid weather conditions.",
"discussion": "3. Results and discussion \n Fig. 1a shows the layer-by-layer schematic of the SE-TENG device and its digital photographic image. The triboelectric layers are made of nickel foam and roughness-created PDMS film. The FE-SEM image shows the structure of the Ni foam and the micro roughness created via the soft lithography technique. Fig. 1b shows the step-by-step fabrication of the SE-TENG device. The device fabrication started with the preparation of supporting substrates for the contact and separation-based TENG, followed by placing electrodes on either sides. The electrodes themselves act as positive triboelectric layers, whereas the negative triboelectric layer was fabricated via a soft lithography technique. By using this technique the micro structured roughness was transferred to the silicone elastomer film from sandpaper. The detailed fabrication process is explained in the Experimental section. Fig. S1 † shows the triboelectric series chart showing the triboelectric materials used in the fabrication of the SE-TENG device. Fig. 1c represents a schematic chart of the preparation of the silicone elastomer film via the cost effective soft lithography technique. Silicone elastomer part A and part B were mixed equally in a 1 : 1 ratio in a beaker. The solution was then poured dropwise on a piece of sandpaper and spread evenly through the spin coating technique. The silicone elastomer coated sandpaper was then dried at 30 °C for 6 h. After drying, the film was peeled off from the sandpaper and utilized as an active layer in the SE-TENG. The detailed film preparation process is explained in the experimental section. The working mechanism of SE-TENG is depicted in Fig. 2a–d , which is because of the triboelectric and electrostatic effects. In the initial condition, the top electrode is in contact with the bottom dielectric layer with no flow of electrons in the electrodes. Due to the mechanical motion applied on the device, the layers separate from each other, leading to the occurrence of a charge difference across the electrodes. This phenomenon induces the electrons to move from the top electrode to the bottom electrode through the external circuit. This action is responsible for the first half cycle of the electrical output signal (alternating current (AC)). With further actuating motion on the device, the layers again move close to each other, which induces the electrons to flow in the reverse direction, leading to the generation of a second half cycle. The potential distribution of the device was theoretically analyzed via a simple finite element simulation using the COMSOL Multiphysics software. Fig. 2e–g show the surface potential distribution from the contact state until the maximum release state. Fig. 1 SE-TENG device schematic and fabrication. (a) Layer-by-layer schematic of the SE-TENG device and the inset shows the FE-SEM morphology of the porous Ni foam and roughness-created silicone elastomer film with a photograph of the device. (b) Step by step schematic showing the SE-TENG device fabrication with every working layer used in the device. (c) Step-by-step fabrication of the silicone elastomer film from liquid silicone via the soft lithography technique using micro-roughness sandpaper. Fig. 2 Working mechanism of SE-TENG. (a–d) Contact and separation mode working mechanism of SE-TENG device with pressing and releasing motion and the respective electron flow directions. (e–h) Potential distribution of SE-TENG at various separation distances using the COMSOL software. The electrical output analysis was performed for the TENG devices made of Al, Cu and Ni foam as positive electrodes and silicone elastomer remained as the negative triboelectric layer. Fig. 3a–c show the electrical output comparison of the TENG devices made of Al, Cu and Ni foam positive electrodes. The output was higher in case of the TENG device made of Ni foam. As is known, the roughness in the triboelectric layers plays a major role in the enhancement of the electrical output. Similarly, Ni foam, which has a porous nature, has a good surface roughness as well as a positive triboelectric property, which proved to be a suitable material for contact electrification. On the other hand, the Al and Cu films do not have any surface roughness or porosity on their surface. The drawback with a plain positive side and a micro roughness negative layer is the damage of the roughness after few actuations. However, the Ni foam and roughness-created negative layer led to an enhancement in the contact area as well as the triboelectric charge generation. The porous layer in the positive side under contact and separation working mode did not damage the roughness present in the negative side, which made the output stable for a prolonged period. The maximum electrical output obtained from the SE-TENG with an Ni electrode was ∼370 V/6.1 μA. In contrast, the SE-TENG made of Cu and Al as the positive layer generated a maximum electrical output of ∼180 V/2.5 μA and ∼270 V/3.6 μA, as shown in Fig. 3a and b , respectively. Fig. 3c shows the performance of the electrical output and the difference with respect to the positive layers. The silicone elastomer layer showed a maximum charge quantity of 160 nC. Further analysis and applications were performed using the Ni foam SE-TENG device due to its high electrical output performance. To confirm the electrical output, which is purely from the triboelectric effect, a switching polarity test was carried out by switching the polarity of the TENG device during its measurement. The device showed an exact 180° phase shift in its generated electrical output signal. This clearly confirms that the electrical output is purely from the SE-TENG device, as shown in Fig. 3d . To prove the contribution of surface roughness and its role in enhancing the electrical performance, the electrical measurement was carried out with two SE-TENG devices, one made of the roughness-created elastomer film and the other with a plain film. The electrical output clearly evidenced that the SE-TENG made of the surface roughness-created elastomer film showed a higher output than the plain elastomer film, as shown in Fig. 3e and f , respectively. Fig. 3 Electrical output analysis of SE-TENG device. (a–c) Voltage and current output comparison of SE-TENG device showing the maximum electrical output with Ni foam as the positive triboelectric material. (d) Voltage signal showing the polarity test of SE-TENG with forward and reverse connection characteristics. (e and f) Electrical output performance comparison of silicone elastomer film with roughness and without surface roughness. To validate the electrical output of the SE-TENG device, various confirmatory tests were performed, such as charging commercial capacitors, glowing LEDs, and powering up a digital wristwatch. Fig. 4a shows the charging characteristics of various rating commercial capacitors such as 0.22 μF, 1 μF, 10 μF and 22 μF for a period of 150 s. The inset in Fig. 4a shows the charging pattern during the device contact and separation with respect to external motion using a linear motor. Fig. S2 † shows the circuit diagram, which was used for performing the capacitor charging. The capacitor with the lowest rating 0.22 μF charged quickly to 31 V in 150 s, storing a maximum energy of 99 μJ. Similarly, the capacitor with the highest rating stored a maximum energy of 11 μJ, which was charged to 0.9 V in 150 s. The other two capacitors, 1 μF and 10 μF, charged to 5 V and 2 V with the maximum energy storage of 125 μJ and 18.05 μJ, respectively as shown in Fig. 4b . Fig. 4c shows the charging and discharging cycle of a 10 μF commercial capacitor under the mechanical force of 10 N. The capacitor stored 5 V in 150 s and the force was removed. Consequently, this made the capacitor discharge the stored potential to 3.5 V. The cycle was repeated two more times to show the cyclic stability of the commercial capacitor charging and discharging. The SE-TENG device showed a maximum area power density of ∼17 mW m −2 at 1 GΩ load resistance, as shown in Fig. 4d . This indicates that 1 GΩ resistance was the exact load-matching resistance for the SE-TENG device, which can be used further for focusing in the real-time applications. To show the durability of the SE-TENG device, a stability test was carried out for a period of 2000s, as shown in Fig. 4e and the stability pattern is shown in Fig. S5. † The inset shows the uniform peak pattern (every 600 s) from the start of the test until 2000 s. This shows that the device can work for a long period of time with a stable electrical output. Further, the SE-TENG was used to glow 60 green LEDs in series connection, as shown in Fig. 4f and Video S4. † These tests validate that the SE-TENG is a promising candidate for powering low power electronic devices. Fig. 4 Real-time output analysis and durability test for SE-TENG device. (a) Commercial capacitor charging characteristics with various capacitors such as 0.22 μF, 1 μF, 10 μF and 22 μF for a period of 150 s. (b) Energy storage analysis of the capacitors charged using the SE-TENG device. (c) Charging and discharging cyclic characteristics of 1 μF capacitor. (d) Impedance matching analysis and instantaneous area power density of SE-TENG device upon various resistance values; the device shows the maximum area power density of ∼17 mW m −2 at 1 GΩ load resistance. (e) Stability analysis of SE-TENG showing its stable power delivering nature for a period of 2000 s. Inset shows the output peak pattern with the interval of 600 s. (f) 60 green LEDs glowing using SE-TENG upon applying force by pressing and releasing the device. Recently, one of the drawbacks faced by TENGs is their performance under the influence of humidity. The stability and performance of TENGs decrease due to a change in humidity. This is due to the interaction of moisture with the triboelectric layers, which leads to a reduction in surface charge generation. Thus, to overcome this issue, the SE-TENG was packed completely with a polyethylene pouch and sealed using a pouch laminator. The packing does not allow humidity, air or water droplets into the layers of the device. The water-resistance capability and the electrical response behavior of WR-SE-TENG is shown in Video S1. † A schematic and digital photograph of the water-resistant SE-TENG is shown in Fig. 5a . The water resistance capability of SE-TENG was tested by dipping the TENG device in a box full of water, as shown in Fig. S3. † The electrical output performance of the packed and unpacked SE-TENG is shown in Fig. 5b and c , respectively. The device shows a similar output performance with respect to packing in case of both voltage and current, proving that the performance was not reduced due to the packing. Fig. 5d and e show the electrical output of the water-resistant SE-TENG by finger pressing under water. The output showed the exact phase shift in signal after being packed and actuated inside the water tub. The inset in Fig. 5d shows the device immersed in water and the inset in Fig. 5e shows the LED glowing by pressing the device by placing it in water. Further, the water-resistant SE-TENG was tested under various relative humidity (% RH), as shown in Fig. 5f . The device worked stably in the humidity range of 10% RH to 99% RH. Fig. S3 † shows the homemade humidity chamber used for performing the humidity test. This also confirms that the device was packed perfectly, and the influence of humidity does not affect the performance of the SE-TENG device. This proves that the packed TENG devices can be used in harsh weather and environmental conditions. Fig. 5 Water-resistant SE-TENG device fabrication and its electrical performance. (a) Layer-by-layer schematic diagram of WR-SE-TENG device and digital photograph showing the device packed with polyethylene and placed inside water. (b and c) Voltage and current behavior of the packed and unpacked SE-TENG device, respectively. (d and e) Voltage and current behavior and polarity configurations of the WR-SE-TENG device when pressed inside a water tub, respectively. (f) Humidity test of the WR-SE-TENG device under various relative humidity of 10%, 42%, 68%, 80% and 99% RH, where the device showed a stable output for the entire period. After undergoing various confirmatory tests and performance analysis experiments using the SE-TENG and water-resistant SE-TENG, the device was used for demonstrating a few real-time applications, which are mandatory for any type of energy harvester. Fig. 6a–f show the capability of SE-TENG for scavenging bio-mechanical energy from human hand and leg motions. The water-resistant SE-TENG device was attached on the floor using Scotch tape to avoid unwanted movements and the LEDs glowed under human hand and leg motions. The biomechanical energy scavenging was demonstrated by analyzing the electrical signal by applying force by palm and leg tapping. This demonstration is shown in Videos S2 and S3. † The voltage and current were maximum during leg tapping, which is due to the weight and large pressure acting on the device, leading to the generation of a higher electrical output as shown in Fig. 6g and h . This demonstration gives clear evidence of utilizing SE-TENG as a bio-mechanical energy harvester to harness biomechanical motions in daily life. Similarly, a digital wristwatch was powered using the water-resistant SE-TENG using a commercial 22 μF capacitor, as shown in Fig. 6i . In addition to that the water resistant SE-TENG was demonstrated as a zero power consuming pressure sensor. The device was placed on a flat surface and different ranges of force applied initially and its current profile recorded. The current profile increased with an increase in the applied force, as shown in Fig. 6j . The sensing characteristics were studied based the relationship between the changes in pressure with respect to the increase in current. The sensor showed a linear response upon an increase in pressure with a sensitivity of 0.492 μA kPa −1 and the correlation coefficient of 0.9985, as shown in Fig. 6k . Fig. 6l shows the electric output response of the active pressure-sensing TENG device with its ability to detect the pressure of different weight objects falling from a height of 10 cm. This shows the ability of the device to detect pressure in low ratings. From the results and studies, it is clear that the water-resistant TENG can effectively be used as a self-powered force sensor in harsh and humid environments, such as in water tanks, water pipes and infusion pumps. Fig. 6 Bio-mechanical energy harvesting and zero power consuming pressure sensor applications. (a–h) Digital photographs of the WR-SE-TENG device under hand and foot tapping motions and LED glowing under human motions. (i) Powering up an electronic wristwatch with the help of a 22 μF capacitor. (j) Force analysis of the WR-SE-TENG device upon various force from 0.4 N to 10 N and its current output profile. (k) Self-powered/zero power consuming pressure sensor with difference pressure level and the linear behavior of its current value showing a correlation coefficient of 0.9985 and good sensitivity of 0.492 μA kPa −1 . (l) Real-time response of different light-weight items (paper clip, coin, key, bolt and battery) dropped on the device and its corresponding electrical output response."
} | 6,140 |
37234610 | PMC10205513 | pmc | 1,116 | {
"abstract": "Microbial consortia under anaerobic conditions are involved in oxidizing organic matter in the sludge to produce methane gas. However, in developing countries like Kenya, these microbes have not been fully identified to target them for the efficient harnessing of biofuel. This study collected wet sludge from two anaerobic digestion lagoons 1 and 2 that were operational during sampling at Kangemi Sewage Treatment Plant, in Nyeri County, Kenya. DNA was extracted from samples using commercially available ZymoBIOMICS™ DNA Miniprep Kit and sequenced using Shotgun metagenomics. Samples were analyzed using MG-RAST software (Project ID: mgp100988), which allowed for identifying microorganisms directly involved in various stages of methanogenesis pathways. The study found hydrogenotrophic methanogens, such as Methanospirillum (32%), Methanobacterium (27%), Methanobrevibacter (27%), and Methanosarcina (32%), being predominant in the lagoon communities, whereas acetoclastic microorganisms such as the Methanoregula (22%) and the acetate oxidazing bacteria such as Clostridia (68%) were the key microbes for that pathway in the sewage digester sludge. Furthermore, Methanothermobacter (18%), Methanosarcina (21%), Methanosaeta (15%), and Methanospirillum (13%) carried out the methylotrophic pathway. In contrast, Methanosarcina (23%),Methanoregula (14%), methanosaeta (13%), and methnanoprevibacter (13%) seemed to play an important role in the final step of methane release. This study concluded that the sludge produced from the Nyeri-Kangemi WWTP harbors microbes with significant potential for biogas production. The study recommends a pilot study to investigate the efficiency of the identified microbes for biogas production.",
"conclusion": "5 Conclusion The approach presented in this study allowed exploration in detail of complex microbial communities coming from methane-producing environments. Microbial communities in methane-producing sludge environments would be expected to contain a high abundance of genes of different steps of hydrogenotrophic, acetoclastic, and methylotrophic pathways, which optimally are encoded by a few microbes. This view was true for the two lagoons under investigation, especially in the freshly created lagoon 1. Nevertheless, we observed different levels of methanogenesis genes and their dispersion amongst various microorganisms. This was especially apparent for the acetoclastic pathway suggesting that the syntrophic acetate oxidation bacteria are reservoirs of metagenomic genes that contribute to the methane cycle. There was no significant difference in the diversity of acetoclastic and the hydrogenotrophic pathways. However, the methylotrophic pathways were significantly different between the lagoons. The organisms responsible for the last step release of methane gas were significantly different among the four samples. There is a need to quantify the methane produced by the different pathways of methanogenesis. An investigation on the effect of physio chemical factors such as heavy metals and antibiotics on microbes responsible for the different stages of methane production at the plant will also be prudent. The study also recommends a pilot experiment on the efficiency of the main methanogens identified in this study to produce biogas, individually and as a consortium.",
"introduction": "1 Introduction The microbial community in sludge possesses great potential as a bio-energy source. Despite such favourable industrial opportunities, appropriate microbial identification techniques have not been applied to exploit it; therefore, the sludge has been considered a ‘black box′ with possible unexploited biotechnological reactions [ 1 ]. Culture and isolation-dependent methods do not present the actual methanogenesis reaction in sludge since most microorganisms within the sludge communities cannot be cultured in vitro [ 2 ]. Great advancements in microbial studies have been made in recent years, and nucleic acid-based molecular methods can identify methanogenic microorganisms by DNA sequencing of their ribosomal RNA (r RNA) genes without the need to isolate the microorganisms [ 3 ]. Currently, hydrolysis, acidogenesis, acetogenesis, and the methanogenesis are the main steps involved in converting organic matter into methane [ 4 ]. The hydrolysis, acidogenesis, acetogenesis, are driven by a wide spectrum of microbiomes except for the last step, methanogenesis which are exclusively driven by a specific group of archaea known as methanogens. These steps form the three main pathways involved in methanogenesis and the final step of the release of methane gas; (i) the hydrogenotrophic pathway, which uses the H 2 /CO 2 substrate where the CO 2 binds to methanofuran, and it is broken down to formyl-methanofuran in the presence of H 2 . This process is catalyzed by the enzyme formyl-methanofuran dehydrogenase [ 5 ]. The formal part of formyl-methanofuran is transferred to coenzyme tetrahydromethanopterin forming formyl-tetrahydromethanopterin catalyzed by enzyme formyl transferase [ 6 ]. The formyl-tetrahydro-methanofuran is broken down to methyl-tetrahydromethanopterin catalyzed by coenzyme F420. A methyltransferase-catalyzed reaction allows the methyl group of methyl-tetrahydromethanopterin to be transferred to coenzyme M [ 7 ]. (ii) The acetoclastic pathway, where acetate is activated to acetyl-CoA by the action of acetate kinase or activity of acetyl-CoA synthetase. The acetyl-CoA molecule is then dismutated using the enzyme acetyl-CoA decarbonylase, where the carbonyl group is oxidized to carbon dioxide while the methyl group is reduced to methane [ 8 ]. (iii) Methylotrophic pathway utilizes methanol and methylated amines as substrates. The methyl group is transferred to the corrinoid protein by the methyltransferase. The coronoid protein is then channeled through the methanogenic pathways in the methyl -CoM stage, where they are finally reduced to methane [ 9 ]. (iv) The final step of methane production involves methyl-coenzyme M reductase and two coenzymes: N-7 mercapto heptanoyl threonine phosphate (HS-HTP) and coenzyme F430 [ 10 ]. Many microorganisms responsible for the methanogenesis process have been reported in the literature. The archaea such as Methanobacteria, Methanocella, Methanococcus, Methanomicrobia, Methanopyra, Methanosarcina and Methanomassiliicoccus has been identified as the main methanogens [ 11 ]. Among these, Methanosarcina has been proposed to be capable of carrying out the final step in methanogenesis of the release of methane [ 12 ]The advancement in metagenomics sequencing has favourably enabled a better understanding of the sludge microbiome community and microorganisms responsible for converting organic matter to methane. Even though metagenomics potential has been in the past restricted to archaea, metagenomics sequencing studies have shown that this conclusion underestimates the methanogens potential and the microbes involved [ 13 ]. With the growing technology, recent studies have revealed that we are beginning to understand the methanogenesis process and the microbes involved in the process [ 11 ]. In Kenya, the biotechnological production of energy from wastewater sludge is a potential venture for a relatively efficient, low-cost wastewater-sludge treatment system. The Nyeri Water and Sewerage Company (NYEWASCO)-Kangemi wastewater treatment plant is one of Kenya's modern and best-managed WWTP. The treatment plant has simple processes designed as trickling filters, sedimentation tanks, anaerobic lagoons, and maturation ponds. The sludge treatment process contains; (i) the desludging chamber, a tank that separates sludge and the liquid components through hydraulic pressure and desludges after every three (3) hours. In the wastewater treatment process, the raw sludge from the different desludging chambers is then pumped into the (ii) sludge well, where sludge is allowed to settle before it is pumped to (iii) the sludge lagoons. The lagoons have a volume of 25000 m 3 with an Organic Loading Rate (OLR) averaging 2.01 kg/ m3. d and a hydraulic retention time (HRT) of 90 days. The sludge lagoons are digestion tanks where anaerobic digestion takes place for three to four months. Here the vegetation and scum are allowed to accumulate over time as part of the biological treatment of sludge and later on are removed. The treated sludge is then allowed through the underground valve to the drying beds by gravity. The dry beds are fitted with concrete slabs with spacing between the water from dewatering to infiltrating to the ground. The sludge is allowed to dry for a month during the wet season and fourteen days during the dry season before being sold to farmers for agricultural land application. The plant produces between 75 and 250 tons of dried-up sludge per month. The sludge is sold at USD 5 per ton, making a profit of around USD 330 to USD 1100 per month from sewage sludge (NYEWASCO, 2007). Its digester system can be upgraded to include biogas production, which is currently lacking. However, it is important to have prior baseline information on the profile of the microbial composition of the sludge from the WWTP. In addition, identifying the microorganisms which metabolize the organic compounds in the wastewater sludge to produce the energy (methane) is vital [ 13 ]. This will provide tangible evidence for a cheaper alternative energy source for the Nyeri-Kangemi WWTP and provide information on the biological properties and possible application of biotechnology, including genetic modification of methanogenic organisms for technical applications. This study identified microorganisms such as protists, bacteria, and archaea in the wastewater sludge from the Nyeri-Kangemi WWTP using the metagenomics method. The shotgun metagenomics techniques was used to characterize the microorganisms. The genes were predicted using the de novo gene prediction pathways [ 3 ] and provide microbial diversity and helped to detect their abundances in the sludge samples. The functional methanogenic annotation was performed by classifying predicted metagenomics proteins into protein families using sequence or hidden Markov models (HMM) databases [ 14 ].",
"discussion": "4 Discussion The whole-genome metagenomic analysis is a useful approach for comprehensively describing complex microbial communities [ 11 ] Various tools are available for metagenomic analysis to enable different insights into the environmental community function and performance. This study applied a commonly used metagenomic analytical tool (MG-RAST) to describe and compare four composite sample sequences through deep shotgun metagenomics. The MG-RAST pipeline features allowed the functional structure of the representative samples' phylogenetic placement of methanogenesis-related genes as described in [ 18 ]. The metagenomic analysis with the MG-RAST pipeline offered an insight into the metagenomic community structure and the abundance of genes involved in methanogenesis. However, it should be noted that this is a general approach and, therefore, difficult to determine interactions between microorganisms involved in each pathway [ 11 ]. MG-RAST provides an analysis platform where KO, SEED subsystems, and the RefSeq databases are explored to allow for a more detailed view and identify a specific function with the simultaneous assignment to a taxonomic group. The abundance of archaea in current study corresponded well with the proportions of the different samples' functional annotations related to methanogenesis. Even though higher hits were recorded in lagoon 1compared to the aged lagoon 2 for all the genes responsible for the different methanogenesis pathways, this is probably attributed to the fact that the sludge in lagoon 2 has stabilized over time, and therefore more microbes can now thrive. Lagoon 2 recorded a higher diversity of the profiles except in the methylotrophic pathway, where lagoon 1 recorded both higher numbers of hits and diversity of organisms ( Fig. 16 ). This can be attributed to more algal blooms on the liquid surface of lagoon one because some of the one-carbon compounds used by methylotrophs, such as methanol and Trimethylamine N oxide (TMAO), are produced by phytoplankton [ 19 ]. All the pathways recorded higher diversity with Shannon Wiener indices above 2.5 except for the hydrogenotrophic pathway, which had a lower diversity of organisms but recorded the highest hits compared to the other pathways. Similar results were recorded by Ref. [ 11 ]. The metagenome had a relatively high abundance of genes of the hydrogenotrophic pathway despite a low abundance of Archaea in the samples analyzed. The results revealed Metharnosarcina as the most abundant archaea in the fresh lagoon one, while Methanospirillum was abundant in the aged lagoon 2. Similar trends are recorded in other studies ([ 5 , 20 ]) [ 20 ]. reported the Firmicutes and Nitrospira genera as the predominant bacteria while Methanosaeta, Methanosarcina , and the Methanospirillum dominated the archaea. On the other hand [ 5 ], recorded a shift in the composition of archaea from Methanosaeta to Mycobacterium. The majority of the annotated mcr sequences were assigned taxonomically to the genera Methanoregula and Methanospirilum in both lagoons suggesting these genera play a dominant role in the last step of methane production in the sludge. There is no information linking the acidophilic Methanoregula with the mcr genes as encountered in this study. Other studies suggest Methanospirillum [ 21 ]; Methanocorpusculum, Methanobacterium [ 22 ]; and Methnanosaeta [ 23 ]as the major taxonomic groups assigned to the mcr genes sequences. The Methanospirillum identified in this study is well adapted with a large genome suggesting the presence of unrecognized biochemical/physiological properties that likely extend to the other Methanospirillaceae and include the ability to form the unusual sheath-like structure and to successfully interact with syntrophic bacteria [ 21 ]. The phylogenetic placement of hydrogenotrophic pathway organisms was annotated to Methanobrevibacter and Methanosarcina genera in lagoon one, while in lagoon 2, the Methanoregula and Methanospirillum were the dominant hydrogenotrophic genera [ 11 ]. reported Methanobrevibacter , Methanomassiliicoccales, Methanoregula, and Methanoculleus as the major contributors to methane production in sewage sludge. The study proved that the hdr genes are found in the methanogenic archaea and acid and thiosulphate-reducing bacteria such as Halanaerobium, sulfate-reducing Desulfovibrio, and the alkene degrading Desulfatibacilum. Other studies ([ 24 , 25 ]) have supported this finding where “methanogenic” genes are also present in other archaea and bacteria [ 25 ]. reported the sulfate-reducing Archaeoglobus fulgidus using many enzymes and coenzymes in anaerobic lactic acid oxidation to produce CO2, also used by methanogenic archaea in the reduction of CO2 to methane. Desulfobacterium autotrophicum contains gene clusters for the heterodisulfide reductase [ 24 ]. The acetoclastic pathway is the most active and important methanogenesis pathway, especially in sludge anaerobic digesters where acetate contributes two-thirds of the total methane production [ 8 ]. The Methanosarcina and Methanosaeta have been described as the genera where acetoclastic methanogenesis occurs [ 26 , 27 ]. The phylogenetic placement of the cdr genes was assigned to Clostridium and Methanoregula as the major taxonomic groups in this category in lagoon one and the aged lagoon two, respectively. The abundance of Clostridium in lagoon 1 (one) acetoclastic pathway consortium compared to methanogens is an interesting phenomenon suggesting the possibility that they play a role in the production of acetate [ 28 ]. in their work supported the involvement of some bacteria, such as clostridia, when studying metabolic reconstruction of metagenome-assembled genomes (MAGs) from a thermophilic sewage waste biowaste digester covering the basic functions of the biogas microbial community; consistently identified the uncultured Dethiobacteraceae together with Syntrophaceticus, Tepidanaerobacter, and unclassified Clostridia as members of a potential acetate-oxidizing core community in nine full-scale digesters, whereas acetoclastic methanogens were barely detected. This may be annotated to the fact that acetoclastic methanogens and syntrophic acetate-oxidizing bacteria (SAOB) compete for acetate, a major intermediate in the mineralization of organic matter. The results presented in this study may provide new insights into a remarkable anaerobic digestion ecosystem where members of the Bacteria domain possibly realize acetate catabolism [ 28 ]. further demonstrated this by metagenomics and enrichment cultivation, revealing a core community of diverse and novel uncultured acetate-oxidizing bacteria and concluding that their genomic repertoire suggests metabolic plasticity besides the potential for syntrophic acetate oxidation [ 29 ]. found that contaminants such as antibiotics limit acetoclastic methanogens, and the resistant syntrophic acetate bacterial oxidants take over from the methanogens. A suggestion has been given that there might be a shift where syntrophic acetate oxidation replaces acetoclastic methanogenesis during thermophilic digestion of bio-waste [ 30 ] Nevertheless, there is a need for more studies to quantify syntrophic acetate oxidation versus acetoclastic methanogenesis. The methylotrophic pathway was dominated by Methanosaeta and Methernothermobacter genera in the lagoons 1. Methanoregula and Methanospirillum were the dominant methylotrophic methane producers in the aged lagoons [ 31 ]. reported a similar result in anaerobic digesters with Methanomassiliicoccus, Methanosarcina, metanospirillum, and methanosaeta in the list of organisms representing the methylotrophic pathway . Nitrosococcus, Methylococcus, and Methylobacterium were the abundant bacteria in this freshly prepared lagoon 1, while the Thermincola, Mycobacterium, and Methylococcus were abundant bacteria in the methylotrophic annotated consortium. Kaster et al. [ 9 ] suggested the existence of the methylotrophic bacteria that use methanogenic enzymes and coenzymes in their energy metabolism. According toGilmore et al. [ 32 ], Methanobacterium, Methanosarcina, Methanosphaera, and Methanocorpusculum are suggested to be capable of methylotrophic, acetoclastic, and hydrogenotrophic methanogenesis. In this study, Methanoregula seemed to be dominating all the three methanogenesis pathways. This may be due to their ability that trends toward energy conservation in genome composition [ 9 ]. 4.1 Limitation of the study The study is without limitations, which may affect our results. This study only describes microorganisms of only one treatment plant with an aim of providing a baseline data for possible application of biogas production. Biases may have been introduced during the DNA extraction process but this was minimised by the use of an optimal extraction kit specifically for metagenomics compared to other kits. High throughput illumina sequencing has also been proven to produce good repeatability. The sequencing depth of 3.2 GB may not be deep enough to satisfactorily explore rare species within sludge in lagoons 1 and 2, however relatively high number of sequences ranging between 149, 542,217 and 485,547,217 per sample. The other advantage is that metagenomics sequencing used in this work is neither low throughput nor PCR based. Lastly, an average base pairs ranging from 157 to 172 bps used in this study may be short. For most metagenomics pipelines, including MR-RAST, 100 bps is long enough to identify a species but an optimal of 200 bps id recommended for a trade-off between the rate of under prediction and the production of such reads."
} | 4,974 |
29844349 | PMC5974012 | pmc | 1,117 | {
"abstract": "Coral reefs face many stressors associated with global climate change, including increasing sea surface temperature and ocean acidification. Excavating sponges, such as Cliona spp., are expected to break down reef substrata more quickly as seawater becomes more acidic. However, increased bioerosion requires that Cliona spp. maintain physiological performance and health under continuing ocean warming. In this study, we exposed C. orientalis to temperature increments increasing from 23 to 32 °C. At 32 °C, or 3 °C above the maximum monthly mean (MMM) temperature, sponges bleached and the photosynthetic capacity of Symbiodinium was compromised, consistent with sympatric corals. Cliona orientalis demonstrated little capacity to recover from thermal stress, remaining bleached with reduced Symbiodinium density and energy reserves after one month at reduced temperature. In comparison, C. orientalis was not observed to bleach during the 2017 coral bleaching event on the Great Barrier Reef, when temperatures did not reach the 32 °C threshold. While C. orientalis can withstand current temperature extremes (<3 °C above MMM) under laboratory and natural conditions, this species would not survive ocean temperatures projected for 2100 without acclimatisation or adaptation (≥3 °C above MMM). Hence, as ocean temperatures increase above local thermal thresholds, C. orientalis will have a negligible impact on reef erosion.",
"introduction": "Introduction Increasing global temperatures are requiring organisms to acclimate to greater thermal extremes, migrate, or suffer reduced fitness and, potentially, local extirpation. The earth’s climate is already estimated to be 0.85 °C warmer than it was in 1880, which is affecting both terrestrial and marine ecosystems 1 . Much of the thermal energy (~60%) associated with warming has been absorbed by the oceans, resulting in melting sea ice, rising sea levels 1 , and record temperatures in tropical waters 2 , 3 . Ocean warming has already resulted in extensive coral mortality 3 , 4 , as evidenced in 2015/2016, when extreme temperatures led to consecutive mass coral bleaching events around the world 2 , 3 . Corals contain photosynthetic dinoflagellates (genus Symbiodinium ) that provide them with organic carbon 5 . However, the symbiosis is thermally sensitive and exposure to elevated temperature disrupts Symbiodinium photosynthesis 6 and causes coral bleaching 7 , 8 . Some coral species and Symbiodinium ‘types’ are more thermally tolerant than others 9 – 11 , but even tolerant genotypes can be overwhelmed by severe temperature stress 3 . Nonetheless, in some cases, previous exposure to high temperature or association with tolerant Symbiodinium can lead to greater thermal tolerance of the coral symbiosis 9 , 10 , 12 , 13 and accelerate recovery following bleaching 9 , 14 , 15 . In comparison to reef-building corals, sponges are thought to be relatively tolerant of increasing sea surface temperatures 16 , 17 . In particular, some bioeroding sponges can tolerate temperatures that induce bleaching in sympatric corals 18 – 22 . Bioeroding sponges, principally the genus Cliona , are important members of coral reef communities as they erode the limestone substratum by reducing the pH at the sponge:substratum interface 23 , dissolving the substratum, and extracting microscopic ‘chips’ of calcium carbonate 24 , 25 . Like corals, many bioeroding sponge species form symbioses with photosynthetic Symbiodinium and photosynthesis enhances their growth and bioerosion 26 – 28 . However, while dependence on Symbiodinium may increase the thermal sensitivity of Cliona , little is known about how these sponges will tolerate predicted incremental temperature increases or whether they can recover from extreme thermal stress 29 , 30 . Experimental research combining elevated temperature and reduced pH has shown that sponge bioerosion rates will likely increase under conditions of ocean acidification 31 – 35 . However, warming can have negative effects on bioeroding sponges, including bleaching or necrosis; and it is likely that these negative effects will override all other environmental factors 29 , 30 , 32 . For instance, under temperature and pH conditions predicted for 2100, the bioeroding sponge Cliona orientalis bleaches, and the associated reduction in photosynthetic productivity results in a negative energy budget for the sponge despite accelerated rates of erosion 29 , 34 . Warming was subsequently identified as the primary stressor inducing bleaching in a bioeroding sponge 30 . However, temperature tolerance appears to vary among bioeroding sponge species as bleaching or mortality was not observed in all studies 31 , 36 . Therefore, the net effect of climate change on bioeroding sponges, and on their erosion rates, appears to be species-specific. Identifying thermal thresholds under near-future warming requires measurement of performance across a broad range of incremental temperature changes. This incremental approach enables a holistic understanding of temperature effects by allowing quantification of the optimal temperature for peak physiological performance, along with derivation of sub-lethal and lethal temperature thresholds 37 , 38 . A similar approach has been applied to corals to quantify adaptation to local thermal regimes 39 and to determine how coral respiration and photosynthesis varies with temperature 40 . Here, we experimentally assessed the ability of C. orientalis to tolerate incrementally increasing sea surface temperatures between 23–32 °C, which represents the annual temperature range for the studied sponges (22–30 °C) and warmer temperatures predicted for 2100 (31, 32 °C) 1 . In addition, we monitored recovery from temperature stress to evaluate the impact of thermal exposure on sponge survival. Photosynthetic performance of Symbiodinium and energy reserves of the sponge were quantified to identify temperature optima and define thermal thresholds. To contextualize the laboratory experiment, we assessed bleaching severity for C. orientalis during the 2017 mass coral bleaching event.",
"discussion": "Discussion Bioeroding sponges are generally thought to be tolerant of environmental stressors, including elevated temperature, ocean acidification, and eutrophication, raising concerns about increased reef erosion under future projected climate scenarios 42 , 43 . Here, we show that incremental increases in ocean temperature up to 30 °C have negligible effects on C. orientalis , but C. orientalis bleaches when exposed to 32 °C, and exhibits little potential for recovery. At the collection site, 32 °C represents an increase of 3 °C above the maximum monthly mean temperature and corresponds to the increase expected under very high greenhouse gas emissions by 2100, but could represent the mean temperature as soon as 2078 (RCP 8.5) 1 , suggesting that C. orientalis could bleach regularly by the end of this century. The results of this study do not support the hypothesis that bioeroding sponges (particularly those species with photosynthetic symbionts) will play a larger role in structuring future reefs. Temperature exposure in the laboratory revealed a narrow thermal threshold for C. orientalis , with sponges appearing visibly healthy after 10 days at 31 °C, but bleaching after only 3 days at 32 °C. This narrow threshold is similar to several sympatric coral species that bleached following 1 °C temperature increases between 31 and 33 °C 44 . In thermally sensitive corals, bleaching coincides with reduced condition and growth 7 and C. orientalis exhibited similar negative responses, including a 75% reduction in Symbiodinium density, 17% reduction in organic matter and 44% reduction in protein content of C. orientalis . Few bleached cores exhibited necrosis which had been previously reported in C. orientalis from Orpheus Island after exposure to only 2 °C above MMM. The discrepancy between studies likely results from the faster temperature increases or acute exposures (3–72 h) used in previous research 32 , 45 . The thermal threshold identified here is consistent with findings for C. orientalis in the southern GBR, which tolerates exposure to +2.0 °C above MMM (27.3 °C)MMM = 27.3 °C 34 , but bleaches at +2.7 °C 30 , and dies at +3.5 °C 34 . Taken together, these experimental and field results suggest that C. orientalis can tolerate current ocean temperatures, but will have little capacity to cope with the warmer oceans projected for 2100. The primary cause of coral bleaching is exposure to extreme ocean temperature, although longer exposure to moderate increases in temperature can also induce bleaching 46 , 47 . Cumulative thermal exposure is a product of the amount and the duration of stress, which is incorporated into the degree heating weeks (DHW) index, which can be used to accurately predict bleaching 3 . In the laboratory, C. orientalis bleached after 2.5 DHW, similar to corals that bleach after 2 DHW under natural conditions 3 . Consistent with our field observations at Orpheus Island, there are few reports of C. orientalis bleaching under natural conditions. In most cases, other Cliona species ( C. aprica , C. caribbaea, C. varians , and C. vermifera ) have tolerated periods of elevated temperature better than neighbouring corals 18 – 20 , including exposures exceeding 31 °C 18 and even 33 °C 20 . In our surveys, C. orientalis did not bleach although temperatures did not exceed 31 °C, which is below the 32 °C thermal threshold identified during our experiment. A 32 °C threshold is consistent with other Cliona - Symbiodinium symbioses, as C. varians was recently reported to bleach when mean temperatures exceeded 31 °C for 10 days 48 . The combination of a 32 °C laboratory bleaching threshold with the lack of bleaching during the 2017 coral bleaching event suggests that current summer temperatures could lead to faster local erosion rates in the near future. Coral bleaching is often preceded by disruption of Symbiodinium photosynthesis 6 , 49 which leads to the production of toxic oxygen radicals, which must be neutralized to prevent damage to lipids, proteins and DNA 50 . The mechanisms of bleaching in sponges may be similar, however, if damage to the photosystems was responsible for triggering bleaching in C. orientalis , the response must have been very rapid: when C. orientalis bleached, Symbiodinium still retained ~66% of F v /F m which had only declined for 3 days. After eight days of exposure to 32 °C, the photosynthetic capacity of the symbiosis was diminished, coinciding with a loss of Symbiodinium and chlorophyll. Similar effects have previously been observed in bleached C. orientalis 30 , 34 , and scleractinian corals, where a loss of Symbiodinium coincides with a loss of lipids, proteins, and organic matter 51 , 52 . In addition, bleaching can disrupt the bacterial symbioses in C. orientalis 53 and scleractinian corals 54 . Prior to C. orientalis bleaching, there was some evidence that respiration rates increased (29–31 °C) and that energy reserves were reduced (31 °C), suggesting that the sponges expend resources to maintain their symbiosis at sub-bleaching temperatures. Respiration in C. orientalis was fastest at intermediate temperatures, likely contributing to the significant decline in the productivity of the symbiosis. Nevertheless, bleached C. orientalis had similar respiration rates to control sponges despite their reduced condition 30 . The absence of an effect of bleaching (i.e., absence of Symbiodinium ) on respiration rates highlights the need to separate measurement of host and Symbiodinium respiration 55 . Based on their low biomass relative to the biomass of the sponge tissue, it is likely that Symbiodinium makes a minor contribution to overall respiration, and other factors such as pumping or feeding rates may dictate energetic demand and respiration in thermally stressed sponges 56 . The ability to persist in warming oceans will depend upon recovery of symbionts and energy reserves, before exposure to any subsequent bleaching-inducing temperatures 52 . After C. orientalis bleached at 32 °C, the sponges did not recover during four weeks at *30 °C, with no recovery of the symbiosis or sponge condition. The only parameter that changed during recovery was photosynthesis, where the rates of oxygen production and photochemical efficiency were higher in sponges returned to *30 °C than in sponges at 32 °C. However, based on visual observations and the lack of recovery of Symbiodinium , relatively high photochemical efficiency was likely due to fouling by photosynthetic epibionts rather than a re-establishment of the Symbiodinium population. In corals, recovery can take between 1.5 and 10 months and some species do not recover within 12 months 9 , 51 , 52 , 57 , 58 . Our experiment indicated that C. orientalis did not recover Symbiodinium under aquarium conditions, but the availability of Symbiodinium may have limited recovery. In other laboratory studies, C. orientalis have recovered Symbiodinium following irradiance-induced bleaching 59 , 60 , but further study is necessary to determine whether C. orientalis can regain symbionts following thermal bleaching under natural conditions. Observations in the Florida Keys, USA indicate that C. varians can recover from thermal bleaching (M. Hill pers. comm.), although some Symbiodinium likely remained within the sponge 48 . Association with tolerant Symbiodinium , especially multiple types of Symbiodinium , can aid recovery from coral bleaching 14 . Here, all C. orientalis cores harboured the same symbiont, S. endoclionum 41 , and exhibited little flexibility in their symbiotic association before or after bleaching. This may make C. orientalis more vulnerable to warming than reef taxa that can associate with multiple Symbiodinium clades 10 , 12 , as C. orientalis harbours S. endoclionum over a large geographic range 41 . Little is known about the genetic diversity or physiology of clade G Symbiodinium , which have only have been found in bioeroding sponges 61 – 63 , foraminifera 64 , and one octocoral species 65 . A physiological comparison of Clade G to other Symbiodinium has suggested that the Clade G from C. orientalis are more thermally tolerant than the clade C or D inhabiting scleractinian corals 45 . However, here we have refined the thermal threshold, showing that while the clade G symbiont in C. orientalis can tolerate current summer temperatures (<32 °C), photosynthesis is impaired at predicted future temperatures (≥32 °C). Recent mass coral bleaching events are a clear indication that ocean warming is a primary threat to reef corals 3 and accelerated bioerosion by Clionaid sponges under ocean acidification would further compound the adverse outcomes of climate change. However, here we show that while the symbiosis between C. orientalis and its associated Symbiodinium tolerates current maximum sea surface temperatures, the partnership breaks down as sea surface temperatures reach 32 °C. A relatively high tolerance of present day temperature extremes may benefit C. orientalis via coral mortality and increased substratum availability in the short term 22 , 66 , however bioeroding sponges with Symbiodinium will be severely affected by ocean temperatures expected by 2100."
} | 3,867 |
37109870 | PMC10144407 | pmc | 1,118 | {
"abstract": "The development and utilization of new energy sources is an effective means of addressing the limits of traditional fossil energy resources and the problem of environmental pollution. Triboelectric nanogenerators (TENG) show great potential for applications in harvesting low-frequency mechanical energy from the environment. Here, we propose a multi-cylinder-based triboelectric nanogenerator (MC-TENG) with broadband and high space utilization for harvesting mechanical energy from the environment. The structure consisted of two TENG units (TENG I and TENG II) assembled by a central shaft. Both an internal rotor and an external stator were included in each TENG unit, operating in oscillating and freestanding layer mode. On one hand, the resonant frequencies of the masses in the two TENG units were different at the maximum angle of oscillation, allowing for energy harvesting in a broadband range (2.25–4 Hz). On the other hand, the internal space of TENG II was fully utilized, and the maximum peak power of the two TENG units connected in parallel reached 23.55 mW. In contrast, the peak power density reached 31.23 Wm −3 , significantly higher than that of a single TENG unit. In the demonstration, the MC-TENG could power 1000 LEDs, a thermometer/hygrometer, and a calculator continuously. Therefore, the MC-TENG will have excellent application in the field of blue energy harvesting in the future.",
"conclusion": "4. Conclusions In summary, this work designed a multi-cylinder-based triboelectric nanogenerator with broadband and high space utilization based on the conventional oscillating triboelectric nanogenerator. The MC-TENG mainly consisted of two TENG units, each unit containing a pair of an internal rotor and external stator. The relative sliding between the rotor and the stator, driven by the external energy, forms a TENG in the freestanding layer mode. The difference in the size of the two TENG units and the mass block masses makes the maximum angle of mass block oscillation in each TENG unit correspond to different resonant frequencies. The resonant frequency of TENG I was f r = 4 Hz, which is a relatively high frequency, while the resonant frequency of TENG II was f r = 2.25 Hz, which is a relatively low frequency. Therefore, MC-TENG can collect mechanical energy in a complex external environment with a broad frequency range. The maximum power density of two TENG units connected in parallel could reach 31.23 Wm −3 , making full use of the device’s internal space and significantly improving the space utilization. The MC-TENG can also collect energy in over 240° directions, offering the advantage of multi-directional energy harvesting. The demonstration illustrates that MC-TENG can power 1000 LEDs continuously and power a temperature/humidity meter and a calculator that work intermittently.",
"introduction": "1. Introduction With the development of global industries, energy demand has increased dramatically [ 1 ]. Fossil energy accounts for a significant part of energy use, but burning fossil fuels has many negative aspects such as the greenhouse effect and harmful particles [ 2 , 3 ]. This problem is compounded by fossil energy being non-renewable and finite. The energy crisis is becoming increasingly severe, so more attention has been given to the development and use of renewable energy sources [ 4 , 5 , 6 ]. Solar, wind, hydro, and nuclear energy are green energy sources that have long been in the public eye. However, the oceans, which cover the most significant part of the Earth’s surface, also contain massive energy [ 7 , 8 , 9 , 10 ]. Tidal and ocean currents have been exploited to generate electricity, but water wave energy is not still well-utilized [ 11 , 12 , 13 , 14 ]. It is, therefore, of great interest to develop an energy harvesting device that can efficiently harvest water wave energy and is environmentally friendly. Most modern electronics collect energy primarily by electromagnetic induction [ 15 , 16 ]. Nevertheless, their primary drawback is that they require sophisticated procedures such as manual charging and maintenance, particularly at low-frequencies where energy conversion is less efficient [ 17 , 18 , 19 ]. In 2012, a novel energy conversion technology, called a triboelectric nanogenerator, which was first proposed by Prof. Z. L. Wang, attracted widespread attention, as its main purpose is to collect low amounts of mechanical energy [ 20 , 21 , 22 , 23 ]. Because of the advantages of small size, low cost, and high energy conversion efficiency, TENG has great advantages in the field of energy harvesting and self-powered sensing [ 24 ]. Triboelectric nanogenerators are based on the coupling effects of triboelectrification and electrostatic induction [ 25 ]. They use two triboelectric materials with different polarities rubbing against each other to produce two charges of opposite polarity on their respective surfaces. Due to its own unique working mechanism, the triboelectric nanogenerator is significantly better than the conventional electromagnetic generator in collecting low-frequency energy, especially the water wave energy in oceans [ 26 , 27 , 28 ]. Currently, the pendulum type P-TENG proposed by Lin et al. has good electrical output performance around 2 Hz, but a large part of the space inside its spherical shell is wasted [ 29 ]. Similarly, Rui et al. reported a cylindrical pendulum shaped triboelectric nanogenerator (CP-TENG) whose main structure consisted of two cylinders assembled to collect the wave energy inside the ocean by the internal rotor oscillation. However, the disadvantage of this structure is that there is still a great deal of unused space inside, and the efficiency of space utilization could be higher. Since the structure has good output performance in a minimal frequency range and almost stops working when the oscillation frequency is far from the optimal frequency, the structure cannot adapt to the complex and changing marine environment [ 30 ]. In this paper, we proposed a multi-cylinder-based triboelectric nanogenerator with broadband and high space utilization to address the problems of narrowband range and low space utilization in the previously reported work. The two TENG units contained in the device both consisted of a rotor and a stator using arched FEP films and aluminum electrodes as the friction materials. The design of the multi-cylinder structure improved the utilization of internal space and electrical output performance. On the other hand, the characteristic of different oscillating parts with different resonant frequencies was used to achieve broadband energy harvesting. First, the effect of the perimeter and thickness of the arched FEP films on the electrical output performance of each MC-TENG unit was investigated. Next, the mass size of the mass block and the motion amplitude of the linear motor was adjusted to find the optimal resonant frequency of the two TENG units, in addition to illustrating the ability of the TENG to generate energy harvesting in multiple directions. Finally, the ability of MC-TENG to power miniature electrical devices was demonstrated. It was further demonstrated that MC-TENG is essential for future mechanical energy harvesting in the environment and for powering microelectronic equipment.",
"discussion": "3. Results and Discussion 3.1. Structure and Working Principle of the MC-TENG The structure diagram of MC-TENG and the working schematic are shown in Figure 1 . The MC-TENG consisted of two TENG units ( Figure 1 a), where each TENG unit consisted of an internal rotor and an external stator. TENG I mainly collects high-frequency mechanical vibration energy, and TENG II collects low-frequency mechanical vibration energy. The difference between the two TENG units is in their sizes. Two acrylic hollow cylinders were snapped together to form the rotor and stator of each TENG unit. Fifteen arched FEP films were attached to the external surface of the rotor, and lead wires were attached to the internal surface as mass blocks. Thirty interdigitated electrodes were affixed to the internal surface of the stator. The acrylic center shaft was fixed to the stator cover after passing through the bearing in the middle of the rotor cover, which allows the rotor to move relative to the stator when it oscillates. Since the contact between the arched FEP film and the electrode is soft contact, compared with the point contact or wire contact form of the previous TENG, the friction is significantly reduced [ 31 , 32 , 33 ]. The contact area between the arched FEP films and the aluminum electrode was larger, and the electrical output performance of the TENG was better. In addition, the arched FEP films allowed the rotor to rotate back and forth in both the clockwise and counterclockwise directions during operation, improving the energy harvesting efficiency. The operation of a TENG unit in the freestanding layer mode is shown in Figure 1 b. After soft contact, the arched FEP film slides relative to the aluminum electrode, and the negative charge will be concentrated in the arched FEP film. Meanwhile, the positive charge will be concentrated in the aluminum electrode part due to the position of the aluminum ahead of the FEP in the frictional electric sequence [ 30 , 34 ]. As shown in Figure 1 (bI), in the initial state, if both the FEP and aluminum electrodes are not charged after soft contact, an electrostatic charge is generated due to triboelectrification. As the arched FEP film is in complete contact with aluminum electrode A, the negative charge on the surface of the FEP membrane and the positive charge on the surface of the aluminum electrode A should be equal and there is no current generation in the external circuit due to the charge balance. In Figure 1 (bII), when TENG I is subjected to the external vibration energy, the mass block on the inner wall of the rotor swings to the right, and the arched FEP film is partially in contact with electrode B. Currently, the charge of electrode A flows into electrode B, while a counterclockwise current is generated in the external circuit. In Figure 1 (bIII), as the mass block continues to swing so that the arched FEP films are in complete contact with electrode B, the negative charge on the arched FEP films is wholly neutralized with the positive charge on electrode B. In Figure 1 (bIV), the mass block reaches the highest position and then swings to the left, making the arched FEP film come into contact with electrode A again [ 35 ]. The direction of the external current changes, and finally, the arched FEP films again returns to its initial position. Figure 1 (bI–IV) describes the complete working process of a TENG unit. The mass block’s periodic swinging back and forth generates an alternating current in the external circuit. The potential distribution in two electrodes in a TENG unit is simulated by the COMSOL finite element method, as shown in Figure 1 c, which can explain the principle of the current generation in the external circuit more clearly. 3.2. Performance of the MC-TENG As shown in Figure 2 a, only one arched FEP film was applied to the rotor surface to save material and simplify the procedure. By adjusting the perimeter ( L ) and thickness ( H ) of the arched FEP films, the film parameters best suited for the operation of the TENG unit were optimized. In the first experiment, the effect of the perimeter of the arched FEP films on the electrical output performance of MC-TENG was investigated by fixing the thickness of the arched FEP films to 12.5 μm and then varying the perimeter of the arched FEP films in two TENG units (TENG I: 16–26 mm, TENG II: 24–34 mm). Figure 2 b–e shows the peak currents and transferred charges of TENG I and TENG II at different arched FEP film perimeters. At an L value of 20 mm, the maximum peak current in the TENG I unit was 0.11 μA, and the corresponding transfer charge was 2.24 nC. In contrast, the TENG II unit could generate a maximum peak current of 0.41 μA and a transfer charge of 2.73 nC at an L value of 28 mm. The electrical output performance of the TENG decreased when the perimeter of the arched FEP films exceeded the optimum L value because the spacing between the rotor and the stator was limited. The perimeter of the arched FEP films will contact the two adjacent electrodes simultaneously when the perimeter of the arched FEP films is too large, and the output performance is negatively affected. Therefore, the L value of TENG I was determined to be 20 mm and that of TENG II to be 28 mm. We continued to investigate the effect of the thickness of the arched FEP film on the electrical output of MC-TENG. Figure 2 f,g shows the output performance of TENG I under different thicknesses of arched FEP film, and it was found that the maximum peak current value was 1.1 μA, and the transferred charge was 6 nC at H = 30 μm. As shown in Figure 2 h,i, TENG II had a maximum peak current value of 1.5 μA and a transfer charge of 13 nC at H = 30 μm. At 12.5 μm, the thickness of the arched FEP films was small, and the repulsion of the electrostatic force resulted in insufficient contact between the arched FEP films and the aluminum electrode. The output performance of TENG was negatively affected. The thickness of the arched FEP film was larger than 50 μm, and the hardness was more prominent after the roll, which increased the frictional resistance during contact with the electrode, thus accelerating the wear of the device. Therefore, the thickness of the FEP arch film in TENG was set as 30 μm. The above data were acquired on a linear motor with excitation parameters of A = 30 mm, f = 2 Hz, and m = 120 g. Since only one arched FEP film was applied and the frictional resistance between the rotor and the stator was low, the excitation parameters were chosen to ensure that the rotor oscillated periodically and that the arched FEP films could cross one electrode. Next, we continued to study the effects of different external conditions on the electrical output performance of the MC-TENG. The direction of motion of the linear motor in Figure 1 a was ensured to be perpendicular to the central shaft. The mass block’s mass ( m ) inside the MC-TENG rotor and the linear motor’s motion frequency ( f ) were varied. The swing angle ( α ) of the mass block was measured using an angle sensor, and the average short-circuit current ( I asc ) value in each TENG unit cycle was also calculated. The mass block was chosen between 30 g and 210 g for TENG I and 60–240 g for TENG II. As in Figure 3 b,d, the oscillation angle ( α ) of the mass block increased with the increase in the mass block, but the frequency ( f ) corresponding to the maximum angle of each mass block appeared to decrease. The increase in frequency led to a shortening of the periodical oscillation time of the mass block. Therefore, the resonant frequency ( f r ) corresponding to the maximum oscillation angle of TENG was lower for the larger mass block. In this work, we hoped that TENG I had the maximum swing angle under high frequency vibration, while TENG II had the maximum swing angle under low frequency vibration. The maximum swing angle corresponded to the best resonant frequency ( f r ) of each TENG unit, and the purpose was to make the MC-TENG meet the broadband operating performance. As shown in Figure 3 b, the maximum swing angle α = 60° corresponded to the mass block m = 30 g and the maximum swing angle α = 96° corresponded to m = 60 g for TENG I near f = 3.75 Hz, which was about 1.5 times that of the former. The I asc value after m = 60 g in Figure 3 c tended to be stable, so we chose m = 60 g for the mass block of TENG I. The mass block of TENG II in Figure 3 d had a mass of m = 210 g, and the corresponding maximum oscillation angle α = 70° occurred close to f = 2.25 Hz. However, when f = 2 Hz, the mass block hardly oscillated, the arched FEP films could not cross a whole electrode, and the TENG II unit had almost no electrical output at this time. As shown in Figure 3 e, the corresponding I asc values stabilized after the swing angle α reached its maximum value for different mass blocks in TENG II. Considering the limited volume of MC-TENG, the mass block of TENG II was determined as m = 210 g. The excitation amplitude ( A ) of the linear motor motion in the above work was fixed as A = 20 mm. After determining the mass block parameters of each unit in the MC-TENG, the continued optimization of the operating conditions was investigated by varying the linear motor of the excitation amplitude A . Three parameters of A = 20 mm, 25 mm, and 30 mm were used in the following experiments. As shown in Figure 3 f,g, the deflection angle α and I asc of each unit’s mass block in MC-TENG increased with the excitation amplitude A . When A = 30 mm, the motion of the mass block in TENG I changed from oscillation to irregular rotation after f = 4 Hz because of the excessive excitation energy. Therefore, the output performance of TENG I after f = 4 Hz is not further discussed. Therefore, we later chose A = 25 m as the excitation amplitude of the linear motor. Figure 3 h,i mainly shows the relationship between the mass block swing angle α and I asc of each TENG unit in MC-TENG and the external excitation frequency when A = 25 mm. The maximum swing angle α = 102° of the mass block in TENG I corresponded to the optimum resonant frequency ( f r ) size of 4 Hz. In contrast, the maximum swing angle α = 96° of TENG II corresponded to the optimum resonant frequency ( f r ) size of 2.25 Hz. In Rui’s work [ 29 ], the CP-TENG had relatively good electrical output, which was only around f r = 1.75 Hz. In contrast, in this work, through reasonable optimized parameters (mass block m and excitation amplitude A ), the two TENG units in the MC-TENG could ensure that at least one TENG unit was working after f = 2.25 Hz. TENG II had the best output when f r = 2.25 Hz, and TENG I had the best output when its f r was equal to 4 Hz, and both TENG units could work simultaneously in the frequency interval above-mentioned. Therefore, the goal of collecting a broadband energy range was achieved. Figure 4 demonstrates the performance of MC-TENG in multi-directional energy harvesting. In Figure 4 a, we measured the short-circuit current ( I sc ) and transferred charge ( Q sc ) of each TENG unit of MC-TENG by adjusting the deflection angle ( β ) of the linear motor movement direction concerning the center shaft of MC-TENG. With f = 4 Hz as the TENG I excitation frequency and f = 2.25 Hz as the TENG II excitation frequency, the oscillation angle α of the mass block in each unit was the maximum at β = 0°. It was found in Figure 4 b–e that as the angle β increased, the short-circuit current generated by each TENG unit gradually decreased, and the transferred charge tended to be stable. The arched FEP films could always cross a whole electrode when β was less than 60°, resulting in a constant amount of transferred charge. At β = 90°, the vibration direction of the linear motor was parallel to the center shaft of the MC-TENG, and the mass block basically did not oscillate, so there was almost no electrical output. Figure 4 f shows the effective electrical output range of the MC-TENG, which could still produce a sound output even if β exceeded 240°. This indicates that MC-TENG can also be applied to harvest multi-directional vibration energy in the future [ 30 , 36 ]. 3.3. Demonstration The output power as well as the charging capacity of the MC-TENG were finally measured in this work. The peak current ( I max ) [ 36 ] and peak power ( P p ) of the respective units in the MC-TENG were first measured at low-frequency ( f = 2.25 Hz) and high-frequency ( f = 4 Hz) conditions under different resistances, respectively. The peak power was calculated by using the formulation of P p = I max 2 R . The output power of the two TENG units, which were rectified and then connected in parallel, was also measured and compared with the output power of the single TENG unit. The I max decreased with increasing load resistance, while the P p increased with increasing load resistance. As shown in Figure 5 a, the swing angle ( α ) of the mass block in TENG I under low-frequency conditions was slight, so there was only a 0.115 mW maximum P p at a load resistance of 70 MΩ. The maximum P p in TENG II was 5.02 mW at 40 MΩ. The maximum P p of the two TENG units connected in parallel was 10.62 mW at 40 MΩ, roughly twice the sum of the power values of the two independent TENG units. In Figure 5 b, the P p of TENG I at 30 MΩ was 2.07 mW. The maximum P p of TENG II at 20.16 mW occurred at 4 × 10 2 MΩ. Moreover, when the load resistance was 3 × 10 2 MΩ, the P p of the TENG after its parallel connection reached 23.55 mW. The volume of TENG I was 2.278 × 10 −4 m 3 , and the volume of TENG II was 7.54 × 10 −4 m 3 , so the power density of each TENG unit corresponding to the peak power under different conditions could be calculated. As shown in Figure 5 c, the power density of TENG I was 0.05 Wm −3 and 9.07 Wm −3 , and the power density of TENG II was 6.66 Wm −3 and 26.74 Wm −3 at low frequency ( f = 2.25 Hz) and high frequency ( f = 4 Hz), respectively. The power density of the TENG units connected in parallel was 14.08 Wm −3 and 31.23 Wm −3 , respectively. The power density of MC-TENG was significantly better than that in the previous work [ 37 , 38 , 39 , 40 ]. After rectification, the power density of the TENG units connected in parallel was significantly higher than that of a single TENG unit, which proves that the MC-TENG makes full use of the space inside the cylinder and improves the space utilization to a large extent [ 29 ]. To study the charging capability of the MC-TENG, a 100 μF capacitor was connected to the circuit of MC-TENG. Since the output of TENG was poor at low-frequency and the peak power of MC-TENG was more significant at high-frequency, high-frequency ( f = 4 Hz) was chosen to observe the charging capability of the MC-TENG. As shown in Figure 5 d, after 60 s, TENG I charged the 100 μF capacitor to 2.73 V, TENG II could charge it to 3.07 V, and the 100 μF capacitor could be charged to 4.7 V by the two TENG units after rectification. To demonstrate the charging capability more visually, several other parameters of the capacitors were chosen to be connected in the MC-TENG. As shown in Figure 5 e, the larger the capacitor value, the longer the charging time. TENG II was tested for endurance because the device is subject to wear and tear. As shown in Figure 5 f, the transfer charge of the TENG II could still be maintained at about 250 nC through 100,000 cycles of continuous operation, and the transfer charge decayed less with an increasing cycle time. This output performance of MC-TENG could be used to drive some miniature electronic devices. Figure 5 g and Supplementary Video S1 show that the MC-TENG can continuously light up 1000 LEDs. Figure 5 h and Supplementary Video S2 used the MC-TENG to power a temperature/humidity meter for some time when the capacitor was charged to 2.18 V. Figure 5 i and Supplementary Video S3 show that the calculator can work continuously after charging to 1.5 V. It is worth noting that the charging curve kept rising. However, the curve’s slope decreased slightly, indicating that the harvested vibration energy can ensure the continuous operation of the calculator and the meter. These experiments allow for the future application of the TENG in the ocean, where the waves must lap, causing the mass block of the MC-TENG to oscillate, thus generating an electrical output to power some miniature electronic devices."
} | 5,960 |
30002369 | PMC6043547 | pmc | 1,119 | {
"abstract": "In contrast to AI hardware, neuromorphic hardware is based on neuroscience, wherein constructing both spiking neurons and their dense and complex networks is essential to obtain intelligent abilities. However, the integration density of present neuromorphic devices is much less than that of human brains. In this report, we present molecular neuromorphic devices, composed of a dynamic and extremely dense network of single-walled carbon nanotubes (SWNTs) complexed with polyoxometalate (POM). We show experimentally that the SWNT/POM network generates spontaneous spikes and noise. We propose electron-cascading models of the network consisting of heterogeneous molecular junctions that yields results in good agreement with the experimental results. Rudimentary learning ability of the network is illustrated by introducing reservoir computing, which utilises spiking dynamics and a certain degree of network complexity. These results indicate the possibility that complex functional networks can be constructed using molecular devices, and contribute to the development of neuromorphic devices.",
"introduction": "Introduction Brain-inspired computing has attracted considerable attention in recent years for its potential to perform intelligent, robust and low-power computations in situations in which conventional algorithm-based computing on Neumann-based computers may falter 1 . Neumann- and silicon-based AI accelerators, such as MIT and NVIDIA’s Eyeriss processors 2 , a machine-learning supercomputer (DaDianNao) 3 and commercial general-purpose graphic processing units (GP-GPUs), are used to create intelligent artefacts with learning and cognitive abilities. The silicon-based neural accelerators mentioned above provide such intelligent functions; however, they are based on advanced computer science and engineering, and are not based on contemporary neuroscience, which causes their applications to be limited to those such as pattern classification and inference. Neuromorphic hardware, on the other hand, is based on neuroscience, and provides excellent opportunities to replicate higher-level brain functions. In contemporary neuromorphic hardware (e.g. IBM’s neurosynaptic chip (TrueNorth) 4 , analogue or digital neuromorphic integrated circuits 5 , 6 , etc.), artificial spiking neurons that mimic nerve impulse (spike) generation and the construction of their dense and complex networks are essential. Coding neuronal information using spikes is functionally important upon transmitting actions on neuronal membranes (active transmission lines) in noisy and unreliable environments 7 . The usefulness of spiking neural networks in practical applications has not become completely clear; however, it has recently been demonstrated that complex and spontaneous dynamics generated by large-scale spiking neural networks are useful for blind source separation 8 , reservoir computing 9 and so on. In present neuromorphic systems, both the integration density and wiring complexity, which directly represent the potential intelligent information processing ability, are much lower than those of human brains 10 , because present major neuromorphic hardware is only composed of silicon complementary metal–oxide semiconductor (CMOS) devices. In this report, we present an extremely dense, molecular neuromorphic network device, composed of a network of single-walled carbon nanotubes (SWNTs) complexed with polyoxometalate (POM) 11 as an alternative to the present silicon CMOS analogue and digital neural processors. At least two types of devices are required to construct analogue neuromorphic hardware: synaptic devices and neuronal membranes. The synaptic device lies at the intersection between the axonal and dendritic wires of the neuron devices and acts as a memristive junction whose coupling strength is stored. A synaptic device consisting of a network of carbon nanotubes (CNTs) has been proposed 12 . The neuronal membrane (neuron) device emits spikes (nerve impulses) and transmits the impulses to other neurons via axonal and dendritic wires 13 . Such neural membrane devices have not been further explored in spite of recent significant advances in materials science. SWNTs are good candidate materials for neural membrane devices because metallic CNT-based conductors generate large electrical noise with rich dynamics 14 – 16 . Moreover, it has been observed that the electronic state and conduction mode of an SWNT vary enormously with the type of molecular species adsorbed 17 , 18 . Phosphododecamolybdic acid, (H 3 PMo 12 O 40 ; PMo 12 hereafter), is one of the POMs exhibiting reversible multi-electron redox properties 19 – 21 , electronic versatility 22 , 23 and negative differential resistance (NDR) on highly ordered pyrolytic graphite (HOPG) 24 . In this report, we present a complexed SWNT/POM network molecular neuromorphic device consisting of a dense and complex network of spiking molecules (PMo 12 particles 25 ) that imitates a large-scale spiking neural network. Based on experimental studies of the fabricated SWNT/POM network, we first discuss its NDR and noise properties and then demonstrate its collective impulse generation. To reveal the possible impulse generation mechanism, we propose an abstract model of the network by assuming a two-dimensional (2D) structure of molecular junctions and demonstrate that the model yields results in good agreement with the experimental results. In Supplementary Note 2 , we illustrate the potential of the SWNT/POM network model for neuromorphic reservoir computing 26 , 27 by demonstrating basic learning ability of the network."
} | 1,404 |
26584513 | PMC4673831 | pmc | 1,120 | {
"abstract": "Geckos have the extraordinary ability to prevent their sticky feet from fouling while running on dusty walls and ceilings. Understanding gecko adhesion and self-cleaning mechanisms is essential for elucidating animal behaviours and rationally designing gecko-inspired devices. Here we report a unique self-cleaning mechanism possessed by the nano-pads of gecko spatulae. The difference between the velocity-dependent particle-wall adhesion and the velocity-independent spatula-particle dynamic response leads to a robust self-cleaning capability, allowing geckos to efficiently dislodge dirt during their locomotion. Emulating this natural design, we fabricate artificial spatulae and micromanipulators that show similar effects, and that provide a new way to manipulate micro-objects. By simply tuning the pull-off velocity, our gecko-inspired micromanipulators, made of synthetic microfibers with graphene-decorated micro-pads, can easily pick up, transport, and drop-off microparticles for precise assembling. This work should open the door to the development of novel self-cleaning adhesives, smart surfaces, microelectromechanical systems, biomedical devices, and more.",
"discussion": "Discussion Hansen and Autumn suggested the contact self-cleaning mechanism, in which self-cleaning occurs when F w-p > F s-p (ref. 7 ). In other words, small solid particles may bind more strongly to the substrate than to the toe pads. Consequently, when pressing a contaminated toe pad against a clean surface multiple times, the particles are removed from the toe pads via a force imbalance. Interestingly, their results show ∼40% force recovery for gecko setal arrays when simulating gecko walking 7 , the value of which is almost the same as the results of our probability experiments at the low pull-off velocity regime ( Fig. 2c ). However, for those particles that bind more strongly to the toe pads than to the substrate surface, that is, F w-p < F s-p , according to the contact self-cleaning mechanism 7 , self-cleaning cannot take place at quasistatic states (low pull-off speeds). In our experiment, we show that even if F w-p < F s-p , at the quasistatic state, self-cleaning can still occur at high pull-off velocities ( Fig. 2c ). This newly identified dynamic self-cleaning mechanism increases the efficiency of the self-cleaning by a factor of 2, compared with that reported by Autumn and others 7 20 . In fact, it is found in our previous experiment that a live gecko sheds dirt from its toes, recovering nearly 80% of its original stickiness in only four steps 8 . This whole animal experiment is consistent with our results for a single seta tests under dynamic loadings (that is, probability of particle detachment: ∼80% at high pull-off speed). This proposed dynamic self-cleaning mechanism, stemming from the distinctive dynamic contact characteristics of spatula, explains the higher self-cleaning rate of live geckos very well 8 . In addition, our results show that the shear velocity also plays a role in enhancing the self-cleaning efficiency. During animal locomotion, active digital hyperextension generates high normal pull-off speed as well as shearing speed before each step, effectively and efficiently dislodging the dirt from its toes in a progressive manner, which keeps the gecko's feet sticky yet clean. In summary, this study has provided direct evidence that the unique shape of nanoscale spatula pads plays a crucial role in generating robust and stable adhesion while permitting efficient self-cleaning capabilities in dynamic regimes. Furthermore, emulating the gecko self-cleaning principle has led to the design of novel artificial setae as powerful micromanipulation tools, which can viably manipulate and precisely assemble micro-/nano-particles by tuning pull-off velocity. It is foreseeable that this self-cleaning mechanism has important impacts on the development of many biologically inspired technologies, including smart and antifouling surfaces, micro-/nano-assemblies, under water cell manipulation technologies, and microelectromechanical systems devices."
} | 1,022 |
35746877 | PMC9629503 | pmc | 1,122 | {
"abstract": "Abstract Tropical coral reefs are hotspots of marine productivity, owing to the association of reef-building corals with endosymbiotic algae and metabolically diverse bacterial communities. However, the functional importance of fungi, well-known for their contribution to shaping terrestrial ecosystems and global nutrient cycles, remains underexplored on coral reefs. We here conceptualize how fungal functional traits may have facilitated the spread, diversification, and ecological adaptation of marine fungi on coral reefs. We propose that functions of reef-associated fungi may be diverse and go beyond their hitherto described roles of pathogens and bioeroders, including but not limited to reef-scale biogeochemical cycles and the structuring of coral-associated and environmental microbiomes via chemical mediation. Recent technological and conceptual advances will allow the elucidation of the physiological, ecological, and chemical contributions of understudied marine fungi to coral holobiont and reef ecosystem functioning and health and may help provide an outlook for reef management actions.",
"introduction": "Introduction The coral reef: a microbially driven ecosystem Tropical coral reefs are highly diverse and productive ecosystems shaped by their main ecosystem engineers, reef-building corals. Corals are “holobionts,” multicellular animal hosts associated with a diverse suite of prokaryotes and microeukaryotes (Rohwer et al. 2002 ). The best studied host–microbe interaction in these holobionts is the coral–dinoflagellate symbiosis, a reciprocal nutrient-exchange relationship (Muscatine and Porter 1977 ). This symbiosis has formed the very foundation of the ecological success of coral reefs over hundreds of millions of years, and its breakdown can rapidly lead to host morbidity and death (Rädecker et al. 2021 ). While our understanding of the coral–dinoflagellate symbiosis builds upon decades of extensive research (Davy et al. 2012 ), other coral and reef-associated microbiota are presumed to be of importance for holobiont and ecosystem functioning as well, but their roles remain poorly understood. Recently, functional studies on coral-associated prokaryotes and their role in host–microbe interactions have gained traction focusing on nitrogen cycling pathways (Rädecker et al. 2015 ), sulfur cycling, specifically in the context of dimethylsulfoniopropionate (DMSP; Glossary) transformations (Raina et al. 2010a ), antioxidant (Dungan et al. 2021 ), and antibiotic activities (Ritchie 2006 ). Increasingly, genomic studies elucidating the functional diversity of coral bacteria suggest some prokaryotes may be drivers of coral holobiont functioning, resilience, and ecological adaptation (Vega Thurber et al. 2009 , Neave et al. 2017 , Pogoreutz et al. 2022 ), and marine probiotic applications are currently being explored for bioremediation and reef restoration purposes (Rosado et al. 2018 , Doering et al. 2021 ). Corals and other reef holobionts are also home to members of Archaea and other microeukaryotes, including fungi (Wegley et al. 2007 , Ainsworth et al. 2017 ). The enigmatic Kingdom of Fungi is considered an ecological driving force that shapes terrestrial ecosystems (including some of the harshest ecosystems on Planet Earth; Coleine et al. 2022 ) and global biogeochemical cycles by interconnecting different levels of biological and ecological organization (Bahram and Netherway 2022 ). Yet, on coral reefs, studies on the taxonomic and functional diversity of fungi have been rare and far between. Aims of this review Fungi in the marine realm, and on coral reefs in particular are understudied compared to terrestrial and freshwater ecosystems (Bärlocher and Boddy 2016 ). We here interpret the available knowledge on coral reef-associated fungi in the light of fungal functional traits and ecological niches in different ecosystems to propose a conceptual perspective of fungal interactions on coral reefs. We cover a spectrum of putative functions and ecological interactions based on fungal cellular, physiological, metabolic, and molecular traits to illustrate their manifold ecological potential. Based on this, we conceptualize how these functional traits may have facilitated the spread, diversification, and ecological adaptation of fungi in coral reef environments. We propose that reef-associated fungi are functionally and metabolically diverse and might contribute to coral reef biogeochemical cycles potentially impacting multiple levels of biological organization, ranging from the cellular to holobiont and, ultimately, the reef scale via benthic–pelagic coupling (Glossary). We further discuss the potential spectrum of interactions of fungi with other organisms on and around reefs, ranging from mutualism and commensalism to parasitism. Based on these comparisons, we hypothesize that fungi may play a pivotal role in the health and ecological functioning of coral reefs, and in reef-building coral holobionts in particular. Finally, we conclude our work with future research directions that we hope will stimulate the advancement of research of fungi on coral reefs. Abundance and microhabitats of marine fungi and their diversity on coral reefs Fungal abundance and microhabitats Environmental substrate availability is a major driver of abundance and biomass of marine fungi (Clipson et al. 2006 ). Not surprisingly, fungal cell numbers and biomasses are much lower in the open ocean compared to sediments and terrestrial ecosystems, their occurrence likely being restricted to association with particles (Wurzbacher et al. 2010 ). Yeasts in the pelagic zone of oligotrophic lakes or coastal environments exhibit low cell densities typically below one colony-forming unit (CFU) ml –1 up to 47 CFU ml –1 in hypertrophic systems (Woollett and Hedrick 1970 , Libkind et al. 2003 ). In highly productive coastal upwelling systems, fungi can exhibit similar biomasses as heterotrophic prokaryotes, thereby significantly contributing to the living microbial carbon (C) and nitrogen (N; Gutiérrez et al. 2011 ). While no information on fungal abundances on coral reefs are currently available, the typically oligotrophic conditions would suggest low environmental abundances of pelagic fungi, which may increase under eutrophication or dissolved organic carbon enrichment, as observed for copiotrophic bacteria (known as “microbialization’ of coral reefs; Haas et al. 2016 ; Glossary). In oligotrophic aquatic systems, expected to support only low to moderate fungal biomass, fungal contributions to ecosystem nutrient cycling may be of lesser significance than that of prokaryotes. Fungal metabolism becomes potentially relevant in very specific scenarios (summarized in Wurzbacher et al. 2010 ), such as in stagnant microhabitats, biofilms, on surfaces, and environments characterized by steep spatial gradients (Sampaio et al. 2007 ); on algae (Kagami et al. 2007a ); in aggregates and extracellular polysaccharides (Masters 1971 ); in the presence of highly recalcitrant (Glossary) nutrient sources that require specific enzymes to be metabolized (Reisert and Fuller 1962 , Fischer et al. 2006 ); as part of symbiotic associations (Whisler et al. 1975 , Gimmler 2001 , Ibelings et al. 2004 ) and predator–prey relationships (Barron 1996 ). On coral reefs, multiple if not all these scenarios may apply at varying spatial scales ranging from the cellular over the holobiont to, potentially, the ecosystem scale (Fig. 1 ). While pelagic fungi might not be very abundant on coral reefs, reefs harbor diverse and abundant benthic substrata suitable for fungal colonization such as the reef framework and rubble along with reef sediments. Coral skeletons underneath the living tissues constitute stagnant microhabitats characterized by (micro)surfaces, porous matrices, and steep gradients of light, oxygen, pH, and nutrients (Risk and Muller 1983 , Schlichter et al. 1997 , Venn et al. 2011 , Wangpraseurt et al. 2012 ). Further, coral reefs harbor a diversity of potential uni- and multicellular hosts fungi could associate with (Fig. 1 ). Coral tissues and skeletons are populated by microalgae such as the Symbiodiniaceae (Dinoflagellata; Davy et al. 2012 ) and Ostreobium (Chlorophyta), respectively, as well as by prokaryotes and fungi (Bentis et al. 2000 , Rohwer et al. 2002 ), so that diverse mutualistic, antagonistic, and/or synergistic microbe–microbe interactions could arise. Finally, corals constantly secrete mucus containing high levels of recalcitrant dissolved organic C and extracellular polysaccharides (Nelson et al. 2013 ), resulting in aggregate formations in the water column that contribute to reef energy transfer and nutrient cycles (Wild et al. 2004 ). Taken together, we hypothesize that coral reefs potentially harbor a diversity of fungi that might exhibit numerous functions in the pelagic and benthic communities, which is discussed in detail in the following sections. Figure 1. Potential microhabitats of fungi on coral reefs. Marine fungi likely inhabit diverse microhabitats on coral reefs, such as substrates (rock, rubble, and interstitial spaces in sediments) and biofilms that have formed on reef substrates (Sampaio et al. 2007 ), but also the water column, where fungi may be predominantly associated with particles or planktonic organisms (Wurzbacher et al. 2010 ). Mucosal spaces of reef invertebrates, in particular Cnidaria, but also vertebrates such as fish could potentially attract fungi (Reisert and Fuller 1962 , Fischer et al. 2006 , Nelson et al. 2013 ). Furthermore, any animal epithelial or macrophyte surface (e.g. seagrasses and macroalgae) may harbor fungal communities (Kagami et al. 2007a , Ettinger and Eisen 2019 ). Aside from surface colonization, endolithic fungi have been reported from calcium carbonate skeletons of corals and the reef framework (Risk and Muller 1983 , Priess et al. 2000 ). Finally, fungi may be associated with the tissues or cells of macro-organisms (Nadal et al. 2008 , Trofa et al. 2008 ). The center represents different stages and/or forms of fungal cells. The outer circles represent potential microhabitats of fungi on the coral reef. Diversity Little is known about the diversity, ecology, and evolution of animal-associated (Bahram and Netherway 2022 ) and marine fungi, including coral- and reef-associated fungi (Golubic et al. 2005 , Amend et al. 2019 , Gladfelter et al. 2019 ). The few fungal diversity studies available for coral holobionts (collated in Table S1 , Supporting Information ) represent an appreciable geographic spread of sampling locations (Fig. 2 ). At the same time, they reflect the well-known constraints of phylogenetic markers and/or genomic databases available for fungi (Frau et al. 2019 , Rabbani et al. 2021 ), and hence do not permit definitive statements on the specificity of the coral-associated fungal community at lower taxonomic ranks ( Table S2 , Supporting Information ). However, these sampling efforts so far provide a valuable first glimpse into coral-associated fungal communities. Of these, most studies assessed entire corals without separation into surface mucus layer, tissues, and skeleton (e.g. Chavanich et al. 2022 ). A subset of studies characterized fungal communities in coral mucus and tissues separately from the coral skeleton (Bonthond et al. 2018 , Rabbani et al. 2021 ), while others focused entirely on the skeleton, and/or limestone reef substrates (Kohlmeyer and Volkmann-Kohlmeyer 1989 , Kohlmeyer et al. 2000 ; Góes-Neto et al. 2020 , Cárdenas et al. 2022 ; Table S2 , Supporting Information ). Consequently, we cannot currently extract more specific information regarding potential compartmentalization of fungal communities within the coral holobiont. However, we highlight notable taxa consistently reported in association with coral holobionts between these studies. Figure 2. Overview of fungal diversity studies in corals. (A) Geographic distribution of sampling sites and investigated coral genera (created using maps package in R ). (B) Pruned phylogenetic trees displaying consistently reported fungal phyla (and classes for Ascomycetes and Basidiomycetes) across studies (NCBI taxonomy; generated using phyloT v2 , Letunic 2015 ). High proportions of Ascomycota in the culturable fraction of coral fungal isolates (Lifshitz et al. 2020 , Paulino et al. 2020 ) and in fungal sequencing data are apparent, sometimes in excess of 85% relative abundance (Wegley et al. 2007 , Góes-Neto et al. 2020 , Cárdenas et al. 2022 ), although dominances of Basidiomycota or Chytridiomycota (Glossary) sequences were reported for endolithic communities (Glossary) of some corals (Góes-Neto et al. 2020 ; Fig. 2 ). The most commonly reported ascomycetes in corals are Sordariomycetes, notably Lindra (Lulworthiales), Hyalorhinocladiella (Ophiostomatales), and Physalospora spp. (Xylariales; Vega Thurber et al. 2009 , Amend et al. 2012 , Bonthond et al. 2018 , Góes-Neto et al. 2020 ). Further, Hypocreales, as well as Dothideomycetes, Eurotiomycetes, and Saccharomycetes (Wegley et al. 2007 , Lifshitz et al. 2020 , Paulino et al. 2020 , Rabbani et al. 2021 , Cárdenas et al. 2022 ) are consistently reported across coral species and ocean basins ( Table S1 , Supporting Information ). Most notably, sequences affiliated to Hortaea spp. including H . werneckii in the order Dothideomycetes were consistently reported (Amend et al. 2012 , Bonthond et al. 2018 , Rabbani et al. 2021 , Cárdenas et al. 2022 ), the latter being an emerging model organism for osmotolerance studies (please refer to “Fungal Traits” ). Within the Eurotiomycetes, notable representatives are Aspergillus spp. or Penicillium spp. (Wegley et al. 2007 , Lifshitz et al. 2020 , Paulino et al. 2020 , Rabbani et al. 2021 , Chavanich et al. 2022 ). Members of the Basidiomycota are commonly reported from corals at low relative abundances, and include Ustilaginomycetes, Agaricomycetes, Microbotryomycetes, and Malasseziomycetes (Wegley et al. 2007 , Bonthond et al. 2018 , Lifshitz et al. 2020 , Paulino et al. 2020 , Rabbani et al. 2021 ). Yet, while “truly” marine fungi are considered those able to grow and/or sporulate in marine environments, to form symbiotic relationships with marine organisms, to adapt and evolve at the genetic level, and/or be metabolically active in marine environments (Pang et al. 2016 ), disentangling true marine indwellers from fungi stemming from terrestrial input or laboratory contamination remains a challenge (Amend 2014 ). Few studies have investigated the community dynamics of coral-associated fungal communities in response to environmental change. While it appears that coral-associated fungal communities might be host-specific (Cárdenas et al. 2022 , Chavanich et al. 2022 ), they are extremely diverse and heterogeneous, which may mask further subtle community differences shaped by the environment (Amend et al. 2012 , Bonthond et al. 2018 , Rabbani et al. 2021 ). As such, no geographical patterns of coral-associated fungal communities have been apparent so far (Rabbani et al. 2021 ), but a greater phylogenetic diversity and heterogeneity of fungi in acroporid corals was reported for reefs in warmer compared to cooler waters (Amend et al. 2012 ) as well as in corals exhibiting tissue lesions (Lifshitz et al. 2020 ). Further, increased abundances of sequences affiliated to Saccharomycetes and Malasseziomycetes (Chavanich et al. 2022 ) and reduced abundances of Sordariomycetes and Agaricomycetes were reported in bleached or heat-stressed) corals, respectively (Cárdenas et al. 2022 ). Finally, coral-associated fungal metagenomic sequences were shown to increase and/or shift toward zoosporic members under environmental stress suggesting fungal proliferation (Wegley et al. 2007 , Vega Thurber et al. 2009 , Góes-Neto et al. 2020 ). Importantly, while information on the diversity of associated fungal communities can be considered scarce, even less information is available regarding fungal functional traits and their interactions on coral reefs and with(in) (coral) holobionts (Ainsworth et al. 2017 , Gladfelter et al. 2019 ). In this light, two major fungal groups have received attention in the past: first, putative pathogens and opportunists such as Aspergillus sydowii , a fungus linked to sea fan aspergillosis resulting in large scale mortality (Smith et al. 1996 ); second, endolithic, i.e. skeleton-associated fungi of reef-building corals (Kendrick et al. 1982 , Golubic et al. 2005 , Fig. 3 ). The state of knowledge on these two most widely studied groups of reef-associated fungi is briefly summarized below (refer to fungal parasites, pathogens , and opportunists ). Figure 3. Established and proposed holobiont–fungal interactions in reef-building corals. Corals are complex holobionts comprised of distinct functional compartments: the tissues (gastrodermis and ectodermis), algal endosymbionts (Symbiodiniaceae; localized in host gastrodermal cells), and the skeleton, which harbors a diverse microbiome including the filamentous alga Ostreobium . Under unperturbed conditions (left panel), coral tissues receive large amounts of organic carbon (C org ) from Symbiodiniaceae (1). Endolithic fungi coreside with the filamentous algae Ostreobium inside the coral skeleton where they form visible banding patterns (2). Endolithic fungi may “attack” Ostreobium cells but are unable to penetrate healthy host tissues, which form skeletal protuberances around fungal hyphae (Bentis et al. 2000 ) (3). Endolithic fungi may perform organic matter remineralization and mineral weathering, resulting in high inorganic nutrient concentrations in skeletal pore water (Risk and Muller 1983 ) (3). Under environmental stress (right panel), coral tissues experience a disruption of algal endosymbiont C org translocation and subsequently expel Symbiodiniaceae, resulting in tissue paling (“coral bleaching”), altered holobiont nutrient cycling, and impaired host immunity (Rädecker et al. 2021 ) (4). Coral bleaching results in transparent host tissues and allows more light to penetrate into the skeleton (5), resulting in blooms of endolithic phototrophs and C org translocation from the endoliths to the coral tissues (Fine and Loya 2002 ) (6). Environmental stress may increase diversity and proliferation of coral-associated fungi (Vega Thurber et al. 2009 ; Amend et al. 2012 ) (6), and increased skeletal erosion (Yarden et al. 2007 ). Thermal stress and weakening of host immune responses may result in opportunistic growth and saprobic activity of coral-associated fungi, which may be accompanied by fungal lifestyle switching (7). In a severely immunocompromised host, fungal infection of tissues and remaining algal endosymbionts may exacerbate holobiont health (Strake et al. 1988 ), leading to host mortality (Alker et al. 2001 ) (8). Arrows without dash represent established fluxes, dashed arrows represent hypothesized fluxes. Blue fluxes refer to hypothesized C org fluxes to fungi. Black text refers to established activity and interactions. Red text refers to proposed interactions. Blue dashed arrows refer to the proposed diverting of C org to fungi. Fungal traits and potential relevance on coral reefs The ocean environment is starkly different from terrestrial ecosystems. In order to better understand and resolve the putative roles of fungi in marine environments and reef-associated holobionts, it is hence critical to consider the cellular, physiological, metabolic, and genomic traits fungi are equipped with ( Table S3 , Supporting Information ). These traits have allowed fungi to conquer diverse niches and ecosystems, including some of the harshest known environments (Coleine et al. 2022 ). Fungi can go airborne via spore dispersal, utilize airborne compounds for growth, or stimulate plant growth via volatile compounds (Vespermann et al. 2007 , Naznin et al. 2013 ). In aquatic ecosystems, fungi occupy a diversity of ecological niches, and can exhibit a diversity and biomass comparable to that of prokaryotes (Gutiérrez et al. 2011 ). Fungi are found in rather contrasting environments from sewage treatment plants to ultraoligotrophic conditions, such as in water distillation apparatuses (summarized in Wurzbacher et al. 2010 ). Ultimately, the ability for lichenization (not further discussed in this review), i.e. fungal partners engaging in photosymbiosis with algae, render fungi the ultimate pioneers of barren, harsh environments (Spribille et al. 2016 ). This versatility of fungi regarding dispersal and habitat colonization can be attributed to a range of functional traits and adaptations as described below, some of which may be fundamental to help them thrive in marine environments including coral reefs, and to potentially engage in complex symbioses. Fungal cell wall properties, cellular integrity, and osmotolerance Fungal cells have peculiar characteristics, which may be relevant for survival in the ocean. Their cell walls consist of multiple layers of polysaccharides (Szaniszlo and Mitchell 1971 , Durán and Nombela 2004 ), which render the cell highly stable and water absorbent. Interestingly, differences in cell wall compositions between ecologically restricted terrestrial and marine species exhibit quantitative, but not qualitative differences in carbohydrate, amino sugar, amino, and fatty acid composition (Szaniszlo and Mitchell 1971 , Ravishankar et al. 2006 , Plemenitaš et al. 2014 , Danilova et al. 2020 ). Further, marine fungi were shown to produce enzymes involved in fatty acid modifications to maintain cell wall fluidity and integrity (Turk et al. 2004 , Kogej et al. 2007 , Gostincar et al. 2009 ). Another feature of cell walls in aquatic fungi is the incorporation of melanin (Cordero and Casadevall 2017 ), which increases cell strength, rigidity, and tolerance to hydrostatic pressure, high UV radiation, and osmotolerance (Casadevall et al. 2017 , Cordero and Casadevall 2017 ), properties relevant for adaptation to marine environments. Ocean salinity (0.6 M NaCl) is considered a potential barrier to fungal growth (El Baidouri et al. 2021 ). Hence, osmolytes likely play a central role in fungal adaptation to marine environments (Danilova et al. 2020 , Gonsalves and Nazareth 2020 ). Across marine environments, osmotolerance in fungi is variable, with some species exhibiting strong local adaptation to (hyper)saline environments (Kohlmeyer and Kohlmeyer 2013 , Buchalo et al. 2019 ). Particularly high halotolerance has been reported for yeasts such as H. werneckii (Hohmann et al. 2007 ), an emerging fungal model organism able to grow in environments up to 5 M NaCl (Plemenitaš et al. 2014 ). Adaptations of H. werneckii to high osmolarity beyond the discussed cell wall properties include ion accumulation (Kogej et al. 2005 ) and modifications in the high-osmolarity-glycerol (HOG) signaling pathway (Turk and Plemenitas 2002 ), which controls the regulation of the osmolyte glycerol (Muzzey et al. 2009 ). The HOG signal transduction pathway is rapidly activated following cell shrinking under hyperosmotic shock, initiating inorganic ion export (Proft and Struhl 2004 ), cell cycle arrest (Escoté et al. 2004 ), diminished translation (Bilsland-Marchesan et al. 2000 ), closure of glycerol export channels (Tamás et al. 2003 ), and activation of glycolysis to counteract cell shrinking (Dihazi et al. 2004 ). Hog1 is deactivated once the cell commences reswelling due to glycerol accumulation (Hohmann et al. 2007 ). In addition to glycerol, halophilic and halotolerant fungi produce diverse pools of osmolytes such as saccharides, polyols, melanin, mycosporine-like amino acids, and unidentified UV-absorbing compounds (Kogej et al. 2006 , Ravishankar et al. 2006 , Danilova et al. 2020 ). Often, multiple osmolytes are produced in response to hyperosmotic shock, and the composition of osmolyte pools differs with fungal identity, growth phase (Kogej et al. 2007 ), and environmental pH (Gonsalves and Nazareth 2020 ). Finally, the genomes of aquatic fungi and yeasts encode for high numbers of major enzymes involved in cellular oxidative stress responses such as superoxide dismutases, catalases, and peroxiredoxins (Gostinčar and Gunde-Cimerman 2018 ). These enzymes are central to mounting antioxidant responses in high Na+, low K+ environments (Gostinčar and Gunde-Cimerman 2018 ), which may in part explain the “Phoma pattern” (Ritchie 1959 ), the correlation of osmotolerance with thermotolerance (Prista et al. 2005 ). Some marine fungi genomes are further characterized by a high G + C content (e.g. Emericellopsis atlantica ; Hagestad et al. 2021 ), a feature previously linked to complex environmental adaptation and horizontal gene transfer (HGT; Mann and Chen 2010 ) and halotolerance in prokaryotes (Jacob 2012 ). Unicellularity and dimorphic switching facilitating the aquatic and host-associated lifestyle Plasticity of morphological and lifestyle traits has allowed fungi to colonize a variety of environmental niches through different strategies (Větrovský et al. 2019 ). While multicellular, filamentous hyphal networks are common in terrestrial habitats, fungi that colonize sugar-rich plant-associated microhabitats such as fruit and nectar, aquatic environments, or uni- and multicellular eukaryotic hosts including intestinal environments tend to rely on unicellular, yeast-like lifestyles, and/or dormant spores (Andrews et al. 1994 , Nagy et al. 2017 ). Even more successful are fungi with the ability to reversibly switch between the hyphal multicellular and unicellular form (Boyce and Andrianopoulos 2015 ). Dimorphic switching (Glossary) has been observed in many terrestrial ascomycetes that are known pathogens in insect or mammalian hosts including humans, but are also able to survive in free-living forms. Dimorphic switching in pathogens is triggered by environmental cues, primarily temperature (Pasricha et al. 2017 , Francisco et al. 2019 ). The switch to a unicellular lifestyle typically involves the compositional remodeling of the hyphal cell wall (characterized by mannoproteins, glucans, and chitin) to evade detection by the host's immune system (Klis et al. 2009 , Nagy et al. 2017 ). By these means and via nitric oxide reductases and other antioxidants, many dimorphic fungi can modulate the host immune response and proliferate intracellularly within host phagocytes (Holbrook et al. 2011 , Nagy et al. 2017 , Chandrasekar et al. 2022 ). While many dimorphic fungi are infectious in their unicellular stage, others, such as human-associated commensal Candida spp., can invade and harm immunosuppressed hosts through switching from unicellular to their hyphal stage, which causes damage by penetrating tissues through filamentous growth (Trofa et al. 2008 ). Similarly, most plant pathogens become pathogenic during their hyphal stage, which enables the fungus to invade the plant tissues (Nadal et al. 2008 ). In corals, endolithic fungi seemingly attempt invasion of coral tissues from the calcareous skeleton underneath (Bentis et al. 2000 ), and hyphae-like cells co-occur with morbid, disease-like host phenotypes (Strake et al. 1988 , Work et al. 2008 ). These fungi potentially spread and infect corals as free-swimming yeast-like forms or may switch to hyphal growth to opportunistically invade immunosuppressed, stressed corals. Finally, dimorphic switching might not only facilitate opportunistic or parasitic interactions, but potentially the establishment of mutualistic symbioses, as observed during lichenization of Umbilicaria muhlenbergi (Wang et al. 2020 ). Overall, the ability of fungi to switch between unicellular and multicellular forms facilitates potentially numerous strategies to survive diverse environments and to engage in interkingdom interactions. While such strategies remain to be discovered in the marine realm, the capacity for dimorphic switching to modulate immune responses for host invasion could explain the prevalence and ubiquity of fungi across marine hosts and habitats. As such, fungal characteristics involved in dimorphic switching will be an interesting trait to investigate to reveal yet unknown mechanisms of coral–fungal interactions. Motility, chemotaxis, and attachment Some unicellular aquatic fungi exhibit a degree of motility. Members of the ancestral lineage Chytridiomycota, or chytrids have adapted to “the life aquatic” via active swimming, specifically targeting new substrates and hosts by producing high numbers of motile asexual zoospores (van Hannen et al. 1999 ). Motility in chytrids is mediated by chemotactic behavior (Glossary: chemotaxis) toward specific amino acids and carbohydrates (Muehlstein et al. 1988 , Scholz et al. 2017 ). Specific cell structures, including the chytrid rumposome , a complex of interconnecting tubules connecting the cell surface with the flagellar apparatus, are implicated in the zoospore response to environmental signals (Powell 1983 ). In contrast to motile zoospores, most fungi including yeasts are nonmotile and typically require substrates to grow on. Attachment strategies to such substrates are hence important and facilitated by the production of mucilaginous sheaths, expression of surface proteins, called flocculins (Ogawa et al. 2019 ), or spore walls (Jones 2006 ), as reported from some red and black yeasts (Andrews et al. 1994 ). Extracellular polysaccharides are associated with enhanced growth under oligotrophic conditions and may bind both ionic and nonionic nutrients (Kimura et al. 1998 ). Lectins, a group of carbohydrate binding proteins are primarily present in the cell wall of aquatic yeasts and implicated in aggregation and adhesion to substrates (Singh et al. 2011 ), specifically attaching to polysaccharides on the cell walls of hosts, or to detritus (summarized in El Baidouri et al. 2021 ). A diversity of adhesion strategies allows for the direct connection between filamentous fungi with yeasts resulting in the formation of structures, so called biocapsules, in the liquid environment (Ogawa et al. 2019 ), which could help facilitate attachment in the ocean. Nutrient acquisition strategies Diverse and highly effective nutrient acquisition strategies are one of the major hallmarks of fungal metabolism. These include exceptional enzymatic capabilities. Major groups of enzymes produced by fungi are relevant for the decomposition and degradation of recalcitrant organic matter, thereby playing an important role in ecosystems via the regeneration of C and N sources. Marine and freshwater chytrids are widely assumed to employ a range of extracellular enzymes as part of their diverse secretome, including carbohydrate-active enzymes (CAZymes; Glossary; Lange et al. 2019 ). A range of fungal enzymes target humic acids or polymers such as lignin, (hemi)celluloses, tunicin (Kohlmeyer and Kohlmeyer 2013 , Castaño et al. 2021 ), or chitin (Tang et al. 2006 ). The latter occurs in high abundances not only in terrestrial but also aquatic ecosystems, e.g. as part of arthropod exoskeletons and fungal cell walls (Reisert and Fuller 1962 ). The marine realm is home to many unique substrates either not found in terrestrial environments, or subject to modifications such as sulfation, i.e. the addition of sulfate groups, the removal of which is necessary for substrate utilization (Janusz et al. 2017 , Schultz-Johansen et al. 2018 , Barbosa et al. 2019 , Kappelmann et al. 2019 ). This includes algal-derived complex polysaccharides, including but not limited to laminarin, fucoidan, porphyrin, and chitin. The broad substrate range observed in some marine fungi (Thomas et al. 2022 ) is likely related to the diverse battery of CAZymes they harbor, such as glycoside hydrolases, which render fungi capable of degrading otherwise recalcitrant polysaccharides. Generalists such as Emericellopsis atlantica tend to harbor a higher diversity of CAZymes than specialists permitting the degradation of a greater range of substrates (Zhao et al. 2014b , Hagestad et al. 2021 ). Importantly, a high diversity of CAZymes and broad substrate range may convey high adaptive capacity to different hosts or substrates, are likely implicated in the diversification of nutritional modes (Janusz et al. 2017 ) and suggest marine fungi may act as vectors of organic matter transfer within marine food webs (Thomas et al. 2022 ). Ultimately, a broad substrate range may be beneficial for adaptation to oligotrophic marine environments, such as coral reefs. In oligotrophic environments, most marine fungi may seek out and adapt to specific niches where nutrients and/or organic matter are “concentrated,” such as the environment of uni- and multicellular hosts of the coral reef benthos. Sponges for instance are filter feeders that efficiently remove particulate and dissolved organic matter from tons of cubic meters of seawater per hour, and are known hosts to marine fungi (Anteneh et al. 2019 ). Pelagic systems, however, are likely inhabited by parasitic and saprobic fungi such as chytrids, which infect phytoplankton hosts and draw from their pool of photosynthetic organic carbon (Klawonn et al. 2021 ) or degrade particulate organic matter (Roberts et al. 2020 ). Of note, the expression of fungal chitinases, peptidases, and relatives of β-N-acetylglucosaminidases has been reported in reef-building corals (Amend et al. 2012 ), suggesting similar lifestyles as in the water column. Fungi as secondary metabolite producers Fungi produce a plethora of structurally unique bioactive compounds that have evolved as key molecules in fungal chemical communication, defense, and competition, facilitating interactions with hosts and other microorganisms (Kusari et al. 2012 , Bahram et al. 2018 , Keller 2019 ). Fungal metabolites exhibit numerous antibacterial, antifungal, antiviral, and anticancer bioactivities, which have long attracted interest in fungi as a source of new drugs (Keller 2019 ). In fact, the first antibiotic molecule in history, penicillin, was discovered nearly a century ago from the culturable fungus Penicillium notatum (Wong 2003 ). Another prominent example includes the potent anticancer compound paclitaxel (taxol), which is widely used in the treatment of different types of cancer, and which was initially isolated in 1993 from an endophytic fungus ( Taxomyces adrenae ) associated with Pacific yew trees ( Taxus brevifolia ; Stierle et al. 1993 ). An increasing interest in the untapped chemical diversity of marine fungi has arisen during the last years (Agrawal et al. 2018 , Liu et al. 2019 ). Many marine fungi associated with algae and marine invertebrates such as sponges and corals have been shown to produce a broad diversity of metabolites with varied bioactivities (El-Gendy et al. 2018 , Bovio et al. 2019 , Kamat et al. 2020 , Peng et al. 2021 ). However, despite the increasing number of studies investigating marine fungal metabolites and bioactivities, their biological and ecological roles remain largely unknown. Secondary metabolite synthesis often relies on primary metabolite pools (i.e. initial building blocks), which feed into specialized biosynthetic pathways involving large multimodular enzymes such as polyketide synthases PKSs, nonribosomal peptide synthetase NRPSs, prenyltransferases, and terpene cyclases (Brakhage and Schroeckh 2011 , Keller 2019 ). Genes encoding these enzymes are arranged in Biosynthetic Gene Clusters (BGCs; Glossary; Brakhage and Schroeckh 2011 ). Given the high energy cost of secondary metabolite production, fungi have evolved effective strategies to control the expression of BGCs (Shostak et al. 2020 ). Many BGCs in monocultured fungi are often silent, and their expression is highly dependent upon environmental and biotic stimuli (Brakhage and Schroeckh 2011 , Netzker et al. 2015 ). For example, the phytopathogenic fungus Sclerotinia sclerotiorum activates different BGCs when infecting different hosts (Allan et al. 2019 ) and the fungal BGC encoding the production of the antibacterial compound bikaverin is only activated when exposed to metabolites from the bacterial competitor Ralstonia solanacearum (Spraker et al. 2018 ). The remarkable flexibility of fungal metabolism has hindered the understanding of their biological roles and modes of action, especially in marine fungi. However, the development of new tools allowing the study of metabolites in situ (e.g. MALDI-tof; Glossary) and the use of genome mining approaches to identify BGCs has tremendously increased our knowledge in recent years (Boya et al. 2017 , Medema et al. 2021 ). Although this blooming field has so far focused on unsilencing BGCs for drug discovery purposes (Brakhage and Schroeckh 2011 ), much can be learned and applied for the ecological study of fungi and will without doubt provide new opportunities to better understand the roles that fungal secondary metabolites play in coral reefs and holobionts. Rapid adaptive evolution of fungal genomes Fungal genomes vary greatly regarding their organization, composition, and ploidy levels. While typically small and dynamic, genome sizes range from around 2 Mb (similar to those of many bacteria) in the unicellular parasitic Microsporidia to around 2 Gb in Pucciniales (rust fungi), in the same order of magnitude as the human genome (Stajich 2017 ). Fungal genomes (those of pathogens in particular) have an extraordinary capacity for rapid evolution reflected in distinct genome compositions and compartmentalization, extensive sequence divergence, and distinct chromosome organization (Möller and Stukenbrock 2017 , Stajich 2017 , summarized in Feurtey and Stukenbrock 2018 ), along with an abundance of transposable elements (Hess et al. 2014 , Miyauchi et al. 2020 , Gluck-Thaler et al. 2022 ), evidence for diversifying selection linked to environmental adaptation, niche specialization, and host–microbe interactions (Sperschneider et al. 2015 ). Further, there is increasing evidence for interspecific gene exchange through hybridization or frequent HGT and viral transfer (HVT; Bian et al. 2020 , Wang et al. 2021b , Gluck-Thaler et al. 2022 ). These mechanisms are poorly explored in fungi, may occur between highly distinct species of fungi (Soanes and Richards 2014 ) and nonfungal organisms including hosts, and have been predominantly studied in pathogenic terrestrial lineages (Friesen et al. 2006 , Menardo et al. 2016 ). Gene exchange via hybridization occurs sexually or asexually (Roper et al. 2011 , Stukenbrock 2016 ), typically during secondary contact of fungal propagules, and can give rise to novel adaptive traits and adaptive capacity with new ecological niches and hosts (Soanes and Richards 2014 , Feurtey and Stukenbrock 2018 ). This includes the rapid evolution of host specificities and virulence phenotypes (Stukenbrock et al. 2012 , Menardo et al. 2016 , Silva et al. 2018 ). Similarly, HGT/HVT between fungi, other eukaryotes, bacteria, and viruses may not only drive rapid adaptive fungal evolution, but also mediate switches from pathogenic to nonpathogenic lifestyles (Zhou et al. 2021 ). Rates of prokaryotic HGT differ between fungal lineages, with proportions of prokaryotic HGT events ranging from none in the Saccharomycetales up to 65% of investigated cases in the Pezizomycotina (Marcet-Houben and Gabaldón 2010 ). HGT with nonfungal eukaryotes include interactions with insect and plant hosts (Zhao et al. 2014a ), but are unexplored in the marine realm. Importantly, as genes involved in the same metabolic pathways are often physically clustered in the genome (Wisecaver and Rokas 2015 ), the acquisition of (partial) gene clusters via HGT/HVT can extend the physiological repertoire of a recipient organism by providing complete, novel metabolic pathways (Feurtey and Stukenbrock 2018 ). While more research is required, the here described capacity for rapid adaptive evolution may not be limited to pathogenic lineages and may help facilitate the adaptation and radiation of fungi to new niches in the marine realm, such as pelagic or interstitial environments including sediments, coral skeletons, and different hosts on coral reefs. Coordinated efforts to increase the availability of genomic sources of coral reef-associated fungi will help elucidate the genetic underpinnings of marine fungal adaptation. Genome functional gene content in marine fungi The survival of microorganisms in oligotrophic marine environments requires the evolution of diverse transporters and catalysts capable of functioning under an alkaline pH and ionic stress (Moran et al. 2004 , Bonugli-Santos et al. 2015 ). While little information is available for marine fungi, a similar observation was made in the model yeast Dendryphiella hansenii . Compared to terrestrial yeasts, its genome is particularly enriched with genes for C and N transport, but also for multidrug resistance (Lépingle et al. 2000 ). Dendryphiella hansenii has numerous examples of gene duplications in conjunction with reductions in the proportion of noncoding DNA and the shortening of overall gene lengths. This results in similar genome sizes, but different genomic coding densities in marine and nonmarine yeasts (coding densities of 79.2% and 70.3% in genome sizes of around 13 Mb in D . hansenii and S . cerevisiae , respectively). The observed gene duplications may reflect the requirements of a more demanding environment, such as a marine habitat, which selects for the retention of duplicated genes even when resulting changes in encoded protein activities are very slight (Dujon et al. 2004 ). Transcriptional features of fungi Changes to transcriptional activity (Glossary: transcription) during certain stages in the fungal life cycle may produce phenotypic variation in response to fluctuating or changing environments, which may be conducive to survival and acclimation. Conidiation, the formation of conidiophores from vegetative hyphae, is one such critical stage. Wang et al. ( 2021a ) reported that conidia in Aspergillus nidulans, A . fumigatus , and Talaromyces marneffei exhibited transcriptional activity while still in the conidiophore, and synthesized mRNA until their release and dormancy was established. Conidia exhibit environment-specific transcriptional responses to temperature shock, osmotic shock, or zinc deficiency, which affects conidial content (mRNAs, proteins, and secondary metabolites). This in turn affects the fitness and capabilities of fungal cells after germination, stress and antifungal resistance, mycotoxin and secondary metabolite production, and virulence (Wang et al. 2021a ). Thereby, the conidia synthesize and store transcripts according to prevalent environmental conditions. Some freshwater fungal lineages were proposed to have evolved from terrestrial fungi in part due to their sticky drifting, branched conidiospores which may easily attach to submerged substrates (Grossart et al. 2019 ). While this remains yet to be confirmed, it may be plausible that not only drifting dormant spores, but entire conidiophores of terrestrial or freshwater fungi may be relocated into the ocean via run-off. Maintaining transcriptional activity, displaced, drifting conidia still developing may be able to attain acclimation via “front-loading” of conidial content before being released from the conidiophore. Such physiological peculiarity could help explain the activity, acclimatization, and in the long run adaptation and diversification of fungal lineages in marine environments, including coral reefs. Finally, little is known about host–fungi symbioses and their underlying molecular mechanisms of symbiosis establishment and maintenance on coral reefs. Major changes in host–symbiont gene (co-) expression reflecting genetic reprogramming and modulation of molecular crosstalk may facilitate novel associations, as reported for arbuscular mycorrhizal fungi symbioses (Glossary; Mateus et al. 2019 ). While no mycorrhiza-like fungal symbioses on coral reefs are currently known, molecular tools such as dual RNA-seq technology may help elucidate the nature of marine host–fungi relationships and identify putative key genes associated with symbiosis establishment, as previously employed for other poorly understood coral–microbe associations (Mohamed et al. 2018 ). Fungal ecology in the context of coral reefs Fungi are recognized for their role as major conduits mediating the transfer of energy and nutrients through terrestrial food webs (Azam 1998 , Moore et al. 2004 ). While fungi-mediated organic matter transformation and nutrient cycling processes in the ocean are less understood (Amend et al. 2019 ), we know different ecological guilds (Glossary) of fungi occur in the ocean, such as saprotrophs (Cunliffe et al. 2017 , Hagestad et al. 2021 ), parasites (Laundon et al. 2021 ), and putative pathogens (Smith et al. 1996 , Yarden et al. 2007 ). At the land–ocean interface, endophytic and mycorrhizal associations with plants are known from salt marshes (Newell 1996 , Clipson et al. 2006 ). On coral reefs however, no comparable examples of mutualistic host–microbe or microbe–microbe interactions of fungi have been reported yet. Rather, studies are skewed toward opportunistic, pathogenic, or parasitic interactions due to their environmental impact (Bentis et al. 2000 , Alker et al. 2001 , Sweet et al. 2013 ). Fungal interactions on coral reefs are hypothesized to include interspecies (fungal–fungal; Bärlocher and Kendrick 1974 ) or interkingdom interactions (fungal–prokaryote and fungal–eukaryote; Golubic et al. 2005 ). The extent of these interactions likely varies with functional traits of the interacting partners as well as abiotic factors (Cheeke et al. 2017 , Francisco et al. 2019 ). In the light of functional traits of fungi from different ecosystems including marine fungi, we argue that fungi may be relevant for coral reef ecosystem functioning at different levels of biological organization and spatial scales. In the following sections, we discuss potential scenarios in which fungi could exert beneficial functions in the context of mediating biogeochemical cycles, and potential mutualistic organismal interactions on coral reefs. This is followed by examples of known and hypothesized pathogenic, opportunistic, and parasitic interactions. Fungal contributions to biogeochemical cycling in the ocean: a metabolic black box Marine fungi likely contribute to the remineralization of recalcitrant organic matter and processes significant for the cycling of C, N, phosphorus (P), and sulfur (S) in marine systems (Gutiérrez et al. 2011 , 2020 ). Marine fungi harbor an extensive battery of suitable exoenzymes, which may result in high substrate affinity and broad substrate range (Newell 1996 , Zhao et al. 2014b , Hagestad et al. 2021 , Thomas et al. 2022 ; see Section Fungal functional traits—Nutrient acquisition strategies , where the fungal secretome is introduced). Thereby, marine fungi may help mobilize organic C in the ocean via the remineralization of recalcitrant high molecular weight detritus, thereby diverting energy to higher trophic levels through saprobic (Gutiérrez et al. 2011 , 2020 , Thomas et al. 2022 ) and parasitic routes (Klawonn et al. 2021 ). Beyond C, however, our knowledge on marine fungal biogeochemical cycling remains obscure. Of particular interest is N, a major limiting element in the oligotrophic ocean, including coral reefs (Cardini et al. 2015 , Rädecker et al. 2015 , Pogoreutz et al. 2017 ). N is essential for the growth and activity of marine fungi (Clipson et al. 2006 ), which may satisfy much of their N requirements from the degradation of photosynthates (Dring and Dring 1992 ) and recalcitrant polymeric compounds including chitin in the molts and carapaces of marine crustaceans (Kirchner 1995 , Tang et al. 2006 ). Endophytic and mycorrhizal fungi likely account for nearly all N present in the decaying standing plant biomass on salt marshes (Newell 1996 , Clipson et al. 2006 ), while fungal rather than bacterial denitrification is a major driver of N 2 O production in redox-dynamic coastal sediments in the German Wadden Sea (Wankel et al. 2017 ). Although intertidal Wadden Sea and subtidal coral reef sediments will starkly differ in their (a)biotic properties, reef sediments are a place of significant microbial turnover of organic matter such as partially recalcitrant coral mucus aggregates. Mineralization of coral mucus fuels benthic and pelagic productivity on coral reefs via the release of limiting inorganic nutrients such as N and P (Wild et al. 2004 , 2005 ). Fungi may contribute to such coral mucus remineralization processes via the turnover of other recalcitrant organic matter in reef sediments (Fig. 4 ). In corals, fungal N metabolism was suggested to help prevent N loss from the holobiont (Rädecker et al. 2015 ). Indeed, fungal genes associated with N metabolism and transport are well represented in coral-associated metagenomes and fungal mRNA transcripts. These genes are related to the metabolism of nucleic acids, amines, and cellular nitrogen compounds, as well as enzymes involved in urea, glutamate, glutamine, and ammonification pathways (Wegley Kelly et al. 2007 ; Amend et al. 2012 ). It was further proposed that fungal N metabolism might partially account for the high levels of inorganic N concentrations in the interstitial pore water in coral skeletons, where septate fungi can be abundant (Le Campion-alsumard et al. 1995 ). Figure 4. Synthesis of known (black arrows) and proposed (red arrows) interactions and functions of fungi associated with the coral holobiont and the coral reef ecosystem. The most obvious and best-studied fungal interactions on coral reefs include putative parasitism, pathogenesis, and bioerosion. Based on fungal functions in terrestrial and other aquatic ecosystems, we propose that reef-associated fungi may further play roles in the structuring of holobiont- and reef-associated microbial communities and biogeochemical cycling. Fungal P and S cycling properties in the ocean remain largely unknown. The macronutrient P commonly occurs at very low concentrations in the open ocean and in oligotrophic coastal ecosystems, such as tropical coral reefs. P concentrations limit oceanic bacterial productivity (Van Wambeke et al. 2002 ), help maintain marine photosymbioses (Wiedenmann et al. 2012 , Rädecker et al. 2015 ) and are a primary driver of pelagic marine thraustochytrid distribution and biomass across space and time (Bongiorni and Dini 2002 ). Fungi could potentially contribute to P cycling via remineralization processes in reef sediments and coral skeletons (Risk and Muller 1983 , Wild et al. 2004 ) or by primary mineral weathering as observed in terrestrial ectomycorrhizal fungi (Landeweert et al. 2001 ). Both could in part explain comparatively high phosphate levels in the pore water of coral skeletons (Risk and Muller 1983 , Fig. 3 ). Importantly, fungal nutrient release could help alleviate nutrient limitation at small spatial scales (mm to cm) for other organisms within the coral skeleton ( Ostreobium ; prokaryotes) or potentially even the tissues (coral host, Symbiodiniaceae, prokaryotes; Fig. 3 ). Different (in)organic S compounds (including sulfides and methanethiol) are readily metabolized by different marine fungal isolates, suggesting a tentative contribution to coral reef S cycling (Wainwright 1989 , Faison et al. 1991 , Phae and Shoda 1991 , Bacic and Yoch 1998 ). Of particular interest may be DMSP transformations, as reflected in the degradation of DSMP from algae and salt-marsh grass Spartina alterniflora by Fusarium lateritium (Bacic and Yoch 1998 ) and the presence and activity of a DMSP lyase implicated in DMSP catabolism (Glossary) in the coral pathogen A. sydowii (Kirkwood et al. 2010 ). Considering the potential ecological relevance of DMSP as osmolyte and antioxidant in corals (Raina et al. 2009 ), fungal DMSP transformations could be of importance in the holobiont. In conclusion, fungi may contribute to coral reef biogeochemical cycling, albeit at likely varying spatial scales and levels of biological organization (Figs 3 and 4 ). Experimental approaches aiming to elucidate fungal biogeochemical cycling on coral reefs from the cellular to the ecosystem level may draw from a diversity of novel analytical tools enabling the study of cell-to-cell level interactions to broad ecological questions (cf. Challenges, open questions and future directions ). Fungal decomposition activity as driver of microbiome structure and function Fungi and bacteria share numerous microhabitats where they form dynamic, coevolving assemblages (Deveau et al. 2018 ). Such close spatial coexistence in the environment or within complex holobionts gives rise to a spectrum of interactions ranging from antagonistic to synergistic (Glossary; Bengtsson 1992 , Mille-Lindblom et al. 2006 ). Antagonistic interactions may be based on interference competition involving allelochemicals (i.e. chemicals produced by living organisms that affect physiological processes in other organisms), such as in early stages of host infection and substrate colonization (Mille-Lindblom et al. 2006 ). Synergistic interactions may include the provisioning of “public goods:” fungi may release resources via the generation of more accessible, intermediate decomposition products of recalcitrant organic matter, which bacteria (or other organisms) cannot access on their own (Tang et al. 2006 , Schneider et al. 2012 , Roberts et al. 2020 ). Metaproteomic analysis of microbial leaf litter decomposing communities showed that the majority of proteins affiliated to extracellular hydrolytic enzymes were related to fungi, and none to bacterial hydrolases (Schneider et al. 2012 ). Strong positive correlation of bacterial abundances with fungal extracellular enzymes suggested bacterial “cheating behavior” (Velicer 2003 ) by exploiting low molecular weight carbohydrates from fungal decomposition (Boer et al. 2005 , Schneider et al. 2012 ). Such fungal–bacteria interactions can affect host and ecosystem health and functioning by structuring microbiome community composition (Boer et al. 2005 , McFrederick et al. 2014 , Bahram et al. 2018 ). Indeed, saprobic chytrids decomposing chitin particles were found to alter the associated bacterial community structure and diversity (Roberts et al. 2020 ), which may be related to fungal processing of recalcitrant organic matter into more readily accessible C (Tang et al. 2006 , Cunliffe et al. 2017 , Roberts et al. 2020 , Thomas et al. 2022 ). Coral reef-associated pelagic and benthic fungi may exhibit similar roles as decomposers, thereby contributing to bacterial community structuring, colonization and succession in complex holobionts and the environment, ultimately shaping ecosystem biogeochemical cycling (as suggested for pelagic chytrids: Roberts et al. 2020 , Klawonn et al. 2021 , Fig. 4 ). Manipulative studies leveraging metaproteomic and metabolomic approaches combined with coculturing and next generation sequencing applications (Glossary) may help elucidate this intriguing prospect of fungal interactions in pelagic and benthic coral reef environments. Ecological interactions through chemical mediation Chemical communication between organisms is one of the most primitive and widespread languages in nature, and a major driver of biological complexity. Bacteria and fungi are widely recognized for their roles in shaping ecosystems through the production of semiochemicals (i.e. chemical substances released by an organism that affect the behavior of other organisms; Davis et al. 2013 , Ditengou et al. 2015 ). These metabolites govern many intra- and interspecific interactions, and while some provide collective benefits (i.e. public goods, such as in biofilm formation or quorum sensing), others shape communities through antagonism (Hogan 2006 , Schoenian et al. 2011 ). Despite the importance of microbial chemical mediation in diverse, complex ecosystems such as coral reefs, this remains a highly underexplored research area. In this section, we provide different examples of fungal chemical mediation that may be of relevance to coral holobiont and coral reef ecosystem functioning, with the aim to identify research gaps and new potential research avenues. Quorum sensing Quorum sensing (QS), a concerted, density-dependent cell-to-cell signaling process, is one of the most widely studied chemical communication strategies in fungi (Hogan 2006 , Barriuso et al. 2018 , Tian et al. 2021 ). Unicellular fungi produce QS molecules that accumulate in the surrounding environment during the growth of the population. When cell concentrations exceed a threshold, the QS molecules trigger coordinated population processes such as virulence/pathogenesis, morphological differentiation, sporulation, secondary metabolite production, and enzyme secretion (Barriuso et al. 2018 ). While several QS molecules have been characterized (e.g. terpenes, lactones, alcohols, peptides, and oxylipins), the structure, specificity, and mode of action of many fungal QS molecules remain puzzling (Affeldt et al. 2012 , Barriuso et al. 2018 ). For example, farnesol (12-carbon sesquiterpene alcohol) has been considered a broad-spectrum QS molecule capable of eliciting dimorphic switching and hyphal growth in phylogenetically diverse fungi including Candida albicans, Ophiostoma piceae , and Penicillium decumbens (Hornby et al. 2001 , Guo et al. 2011 , de Salas et al. 2015 ). However, in other Ophiostoma species, O. ulmi and O. floccosum , dimorphic switching is mediated by 2-methyl-1-butanol and cyclic sesquiterpenes (Berrocal et al. 2012 , 2014 ). Butyrolactones have also been identified in different fungi (i.e. Penicillium sclerotiorum and Aspergillus terreus ) as QS molecules capable of regulating the production of antimicrobial metabolites (Raina et al. 2010b , 2012 ). In other Aspergillus species, oxylipins (i.e. lipids created from fatty acid oxidation; Glossary) regulate morphological differentiation and mycotoxin production (Tsitsigiannis et al. 2005 , Horowitz Brown et al. 2008 , Affeldt et al. 2012 ). In addition to population cooperation, QS molecules also play important roles in interspecific interactions and cross-kingdom signaling (Sztajer et al. 2014 , Dixon and Hall 2015 ). For example, farnesol acts as an antimicrobial agent in both bacteria and fungi (Derengowski et al. 2009 ), and can disrupt bacterial QS communication (Cugini et al. 2007 ). In coral reefs, the mechanisms underpinning fungal QS and the potential role of these molecules in other ecosystem interactions and processes remain poorly understood. Interestingly, coral fungal endophyte extracts were found to inhibit bacterial QS (Martín-Rodríguez et al. 2014 ), suggesting that fungi are likely to play important ecological roles such as influencing microbiome assembly, structuring, and antifouling protection in coral holobionts (Fig. 4 ). Oxylipins Oxylipins (Glossary) are ubiquitous in a wide range of organisms, play major roles in biological processes, e.g. regulating inflammation and cellular homeostasis in metazoans. Oxylipins are major mediators of cross-kingdom talk and host–fungal interactions that encompass predator–prey, mutualistic, and pathogenic and biological processes (Holighaus and Rohlfs 2019 , Niu and Keller 2019 ). For example, the fungal volatile oxylipin 1-octen-3-ol is found in the feces of Aspergillus -infested beetle larvae, which in turn is used by parasitoid wasps as a cue to detect unfavorable environments (i.e. mold-infested beetles; Steiner et al. 2007 ). This repellant response was found to be innate in the wasps, suggesting that host-associated fungi may be important in parasitoid host-finding strategies (Steiner et al. 2007 ). Fungal oxylipins can also alter oxylipin production in plant and mammalian hosts in order to modulate or attenuate responses of the host immune system, and hence facilitate infection (Brodhagen et al. 2008 , Patkar et al. 2015 ). In symbiotic cnidarians (Glossary: Cnidaria), oxylipins of algal endosymbionts (Symbiodiniaceae) are presumed to suppress host oxylipin expression to facilitate symbiont persistence, thereby helping maintain the cnidarian–algal symbiosis (Matthews et al. 2017 , Lawson et al. 2019 ). So far, these studies have focused on the cnidarian–dinoflagellate relationship, but the putative role of fungi and bacteria in coral holobiont oxylipin signaling remains unexplored. Similarly, the production of volatile oxylipins in coral reef organisms and their ecological roles remain unknown but warrant further investigation. Volatile organic compounds Volatile Organic Compounds (VOCs; Glossary) are small compounds that diffuse easily through water and gas and play a critical role in biosphere–atmosphere interactions, in plant signaling, and as infochemicals in multitrophic interactions (Yuan et al. 2009 , Kegge and Pierik 2010 ). The importance of fungal VOCs in microbe–microbe and host–microbe interactions has long been overlooked, but recent studies suggest they have a significant role in long-distance signaling in bacterial–fungi interactions (Effmert et al. 2012 , Schmidt et al. 2015 , Jones et al. 2017 ). Fungal VOCs can have a wide a range of antagonistic (e.g. growth and virulence inhibition and antimicrobial properties) and mutualistic (e.g. growth promotion and secondary metabolite production) effects on other microbes (Strobel et al. 2001 , Vespermann et al. 2007 , Minerdi et al. 2008 ). However, fungi themselves are also exposed to microbial VOCs that modify their behavior. For instance, Pseudomonas aeruginosa produces VOCs that stimulate the growth of the opportunistic fungus A. fumigatus , favoring its invasion of lung tissue (Briard et al. 2016 ). Although the role of VOCs in marine systems remain highly unexplored, the study of “volatilomes” (Glossary), i.e. the collection of all VOCs emitted by an organism, and their putative role in organism welfare and holobiont interactions is gaining traction in coral reef studies (Swan et al. 2016 , Lawson et al. 2020 , Olander et al. 2021 ). Environmental factors such as temperature stress drive the composition and diversity of VOC emissions in coral holobionts (Lawson et al. 2021 ). Furthermore, the study of volatilomes from different holobiont partners suggest a complex multipartite metabolic interplay between the algal endosymbionts, their associated bacteria, and coral hosts (Lawson et al. 2021 ). Whilst the diversity of VOCs produced by coral-associated fungi remains yet to be charted, studies have shown that fungi can also produce important VOCs previously identified in coral holobionts, such as DMS (Bacic et al. 1998 ), an abundant catabolite of DMSP (Curson et al. 2011 ). For example, the coral pathogen A. sydowii is known to catabolize DMSP into DMS (Kirkwood et al. 2010 ). These examples highlight the importance of uncovering the diversity and putative functional roles of coral-associated fungal VOCs, which could help understand the functioning of the coral holobiont in unperturbed and stressful environments. Cooperative metabolite synthesis As the different partners in a symbiosis often work collaboratively, the synthesis of some metabolites (VOCs, QS molecules or secondary metabolites) highly depends upon these complex multipartite interactions (Partida-Martinez and Hertweck 2005 , Shao et al. 2020 ). Therefore, not only the cross-communication can affect the production of secondary metabolites by modulating gene expression or unsilencing silent BGCs (Netzker et al. 2015 ), but distinct organisms can also provide different substrates, enzymes and pathways that lead to cooperative metabolite biosynthesis (Cavaliere et al. 2017 ). Although this topic has recently been explored in microbial coculturing approaches under what is known as the OSMAC (One Strain MAny Compounds; Glossary) approach, cooperative biosynthesis in hospite still remains a black box (Cavaliere et al. 2017 ). Therefore, research efforts aiming to understand the functional role of fungi in coral holobionts should focus on understanding the contribution of the different symbiotic or commensal organisms to the “holometabolome,” i.e. the net metabolome of the holobiont, instead of individual holobiont members in isolation. Fungal mutualism in the ocean Evidence from the fossil record to natural and experimental cocultures Little is known about mutualistic and commensal fungal relationships in the ocean, but new symbiotic relationships are continuously being discovered (Zhang et al. 2021b , Schvarcz et al. 2022 ). This suggests that marine mutualistic fungi may not be absent but may have been simply overlooked until now. Fossil and extant records indicate that putative chemoautotrophic mutualistic fungi–bacterial consortia may have existed over geological timescales in the oceanic crust (Ivarsson 2012 , Bengtson et al. 2014 ). Most notably, a few exciting examples are showcasing the high potential of fungi to form (mutualistic) symbioses with other organisms in the marine realm. This includes the recently described association of a marine sediment-dwelling fungus with its bacterial endosymbiont, the latter of which modulates the antimicrobial (polyketide) biosynthetic activity of its fungal host (Shao et al. 2020 ). This behavior is thought to convey a protective mechanism against other microbial competitors, which may also contribute to microbiome structuring in the sediment. A second example has described the experimental, forced coculture of the marine alga Nannochloropsis oceanica with the soil fungus Mortierella elongata . This coculturing effort not only demonstrated changes in the productivity and growth (Du et al. 2019 ), but also the induction of reciprocal C and N translocation between the two organisms as well as the eventual incorporation of viable algal cells within the fungal mycelium. Notably, this exciting artificially induced endosymbiosis remained stable over months of cocultivation, demonstrating a latent capacity for fungal–algal mutualism in the marine realm and providing a unique opportunity for the study of evolution of endosymbioses and fungal adaptations to novel marine hosts and environments. Such unique observations imply there is a high likelihood of coral reefs, widely known for their great functional, ecological, and taxonomic diversity, harboring a plethora of similar relationships between fungi and other reef biota. This example also impressively highlights the importance of experimental coculturing efforts for the discovery and study of marine microbial interactions. The quest for probiotic potential of fungi on coral reefs The poor outlook of coral reefs persisting in the Anthropocene (Hughes et al. 2017 ) is currently driving significant research efforts to explore the potential of “microbial” strategies to mitigate some of the detrimental effects of rapid global climate change (Peixoto et al. 2021 ). These efforts include the development of microbiome modification protocols and probiotic consortia (Glossary: probiotics) to physiologically “augment” coral holobionts with the goal of increasing their resilience to environmental perturbation, in particular ocean warming. One main goal is to maintain the coral–algae symbiosis, which can be rapidly destabilized by frequent heat wave events resulting in coral reef degradation across the world. While probiotics are already employed in marine food production (Parata et al. 2021 ), the development of such applications for coral reef conservation is much more recent (Rosado et al. 2018 , Buerger et al. 2020 , Doering et al. 2021 , Dungan et al. 2021 , Zhang et al. 2021a ). The probiotic potential of reef-associated fungi remains unexplored, but many fungal traits, in particular metabolic capabilities and bioactivities, align well with desired beneficial functions for coral probiotics as outlined by (Peixoto et al. 2021 ). For instance, the mitigation of cellular stress via antioxidant properties of probiotics is considered a potential strategy to alleviate the effects of heat stress in coral holobionts (Rosado et al. 2018 , Dungan et al. 2021 ). Indeed, fungi are known to contribute to oxidative homeostasis (Glossary) in mycorrhizal symbioses (Nath et al. 2016 , Huang et al. 2017 ). The putative roles of coral-associated fungal products, such as antioxidant enzymes (Gostinčar and Gunde-Cimerman 2018 ), photo-protective compounds (Sinha et al. 2007 ), or detoxifying enzymes (Massaccesi et al. 2002 ), hence warrant further investigation for their suitability in stress mitigation in corals. Furthermore, fungal antibiotics and QS molecules could be beneficial for their marine hosts via protection from pathogen entry (Ritchie 2006 ) and/or via fungal antifouling activity (Xu et al. 2015 ), which could help maintain the host-associated microbial community in times of stress. Finally, nutritional benefits provided by probiotics to sustain the marine holobiont during times of low environmental nutrient availability or environmental stress (Cardini et al. 2015 ) may be desirable. This is particularly important for corals that have undergone “bleaching,” a morbid state following mass expulsion of the coral's intracellular endosymbiotic algae (Strake et al. 1988 ). “Bleached” corals are vulnerable as they are starved of their major energy source, their algal symbionts’ photosynthate, while simultaneously being weakened by environmental stress (Rädecker et al. 2021 ). Here, different N cycling pathways have been implicated in the maintenance or destabilization of the coral–algae symbiosis depending on prevailing environmental conditions (Rädecker et al. 2015 , Pogoreutz et al. 2017 ). While potential beneficial traits of endolithic fungi are yet to be characterized, pathways involved in the cycling of N (Wegley et al. 2007 , Amend et al. 2012 ), a major element limiting productivity on oligotrophic coral reefs, could potentially help maintain nutritional homeostasis under environmental stress (Rädecker et al. 2015 ). In these (and potentially other) contexts, fungi could potentially play crucial roles in the health and microbiome structuring of reef-dwelling holobionts, such as corals. These potential functions warrant the exploration of coral- and reef-associated fungi in the quest for coral probiotics, which until now has focused on the algal endosymbionts (Buerger et al. 2020 ) and bacteria associated with corals (Rosado et al. 2018 ). We hence propose to include reef-associated fungi into future efforts to elucidate the untapped probiotic microbial potential, which may help deliver novel solutions for coral reef management and restoration. Fungal pathogens, opportunists, and parasites in the coral reef environment Sea fan aspergillosis Fungal pathogens in humans and commercially important crops have been, and still are receiving significant attention (Feurtey and Stukenbrock 2018 ). On coral reefs, different disease-like phenotypes have been linked to fungi (Ravindran et al. 2001 , Yarden et al. 2007 , Sweet et al. 2013 , Soler-Hurtado et al. 2016 ). The probably best studied among them is a putative epizootic causing “sea fan” aspergillosis (Glossary) in octocorals (Alker et al. 2001 ; from here on referred to as “aspergillosis”). The first aspergillosis outbreak followed by mass mortality of Gorgonia flabellum and G . ventalina in the Caribbean was described in the 1980s (summarized in Smith and Weil 2004 ). Similar outbreaks affecting other octocorals in the tropical Atlantic and the Tropical Eastern Pacific were observed in the 1990s, and the 2000s, respectively (Smith and Weil 2004 , Barrero-Canosa et al. 2013 ). Symptoms of aspergillosis include tissue lesioning and recession, followed by discoloration (“purpling”) of affected tissues and gall formation. Affected tissue samples contained high loads of septate fungal hyphae, and culturing efforts identified the putative agent of the disease, a soil-dwelling saprobic fungus affiliated to A. sydowii and a “pollutogen” from terrestrial run-off onto coral reefs (Smith et al. 1996 ). Transfection experiments from infected onto healthy sea fans initially suggested aspergillosis is transmissible (Smith et al. 1996 ). However, recent evidence suggests A . sydowii may not be the cause of aspergillosis (Toledo-Hernández et al. 2008 , 2013 ), but rather an opportunist invading the tissues due to declining host health (Rypien et al. 2008 , Toledo-Hernández et al. 2008 ). Outbreaks were often linked with increases in ambient temperature (Kim and Harvell 2004 ), which also promotes the growth of A. sydowii isolates in vitro (Alker et al. 2001 ). Similarly, elevated temperatures in situ likely weaken coral immune defenses, permitting the proliferation of fungal opportunists (Ward et al. 2007 ). The mechanisms of virulence associated with aspergillosis remain poorly understood. Phenotypic characterization of A. sydowii demonstrated a correlation of the secondary metabolites sydowinol, sydowinin A and B, and sydowic acid with strain pathogenicity (Smith and Weil 2004 ). These metabolites adversely affected the photochemical efficiency of coral-associated algal endosymbiont cultures (Symbiodiniaceae) with different symbiont types being differentially susceptible (Hayashi et al. 2016 ). It remains unknown whether these secondary metabolites also affect algal symbiont physiology in hospite , and whether this mechanism is ultimately linked to aspergillosis. Finally, pathogenic coral-associated A . sydowii strains harbor DMSP lyase dddP , which catalyzes the catabolism of DMSP (Kirkwood et al. 2010 ), a compatible solute produced in high abundances by Symbiodiniaceae, the coral (Raina et al. 2010b ) and associated bacteria (Lawson et al. 2018 ). Coral tissues typically contain high concentrations of this compound (Raina et al. 2010b ) and could, therefore, provide abundant substrate for DMSP catabolizing fungi such as A . sydowii (Kirkwood et al. 2010 ). It remains unclear whether the ability to catabolize DMSP confers any selective advantages to A . sydowii , such as in its ability to colonize host corals, whether it affects its pathogenicity, detoxification, or chemical signaling (Kirkwood et al. 2010 ), or whether the fungus merely utilizes DMSP as an environmental cue or C source. Overall, these studies suggest that fungal invaders leverage on chemical cues and crosstalk to interact with other members of the coral holobiont. Further studies will be required for a better understanding of the environmental and biotic drivers and mechanisms of fungal pathogenicity and opportunistic infection on coral reefs. Enigmatic endolithic fungi of coral reefs Coral endolithic fungal communities associated with the calcareous skeleton have been studied for decades, yet their functions remain to be fully characterized (Pernice et al. 2019 , Ricci et al. 2019 ). Coral endolithic fungi were proposed to contribute to nutrient cycling via remineralization of organic matter, such as dead cells (Risk and Muller 1983 , Priess et al. 2000 ). Endolithic fungi are largely viewed as bioeroders, parasites, or opportunistic pathogens (Yarden et al. 2007 , Gleason et al. 2017 ) as they seemingly “attack” the filamentous, skeleton-dwelling algae Ostreobium , but also attempt to penetrate the live tissue layer of corals (Bentis et al. 2000 ). The coral host thwarts these fungal attacks by continuously accreting layers of “repair aragonite,” forming characteristic perl- or cone-like structures around the ever-probing hyphae (Bentis et al. 2000 , Fig. 3 ). Isolates of the coral associated basidiomycete Cryptococcus , a genus implicated in human cryptococcosis, were shown to selectively prolong short-term survival of skeletogenic coral cell types in coculture, which was interpreted as stimulation of coral defense reactions by the presence of the fungus (Domart-Coulon et al. 2004 ). Interestingly, the tissues of fire corals ( Millepora complanata ; Hydrozoa) were laden with fungal hyphae following a marine heat wave resulting in “coral bleaching.” Opportunistic coral-associated saprobic fungi may be able to overcome weakened immune defenses of their vulnerable host, analogous to human secondary fungal infection in the aftermath of viral disease (Baddley et al. 2021 ; Fig. 3 ). While metabolic interactions of endolithic fungi with other members of the coral holobiont (coral host cells, Symbiodiniaceae, or associated prokaryotes) have yet to be characterized, it may be possible they divert photosynthate from coral-associated algae ( Ostreobium , Symbiodiniaceae), as observed in phytoplankton–chytrid pathosystems (Kagami et al. 2007b , Klawonn et al. 2021 , Fig. 3 ). It remains to be determined whether this proposed interaction indeed occurs, and whether it is ecologically relevant in healthy corals under unperturbed conditions. During times of prolonged environmental stress however, when organic C translocation from endolithic algal communities to the coral host may become physiologically relevant (Fine and Loya 2002 ), depletion of such alternative C supplies by parasitic fungi could further exacerbate the health of the impaired coral host (Fig. 3 ). In addition, a mechanistic understanding of the potential for virulence of skeleton- or tissue-dwelling fungi under environmental perturbation is required. Increased seawater surface temperatures have been linked to the reemergence of sea fan aspergillosis (Kim and Harvell 2004 ), and heat or excess nutrient stress are known to induce virulence in coral bacterial pathogens ( Vibrio shilonii and V . corallilyticus ; Rosenberg and Falkovitz 2004 , Kimes et al. 2012 ) and reef bacterioplankton (Cárdenas et al. 2018 ), respectively. Combined molecular, metabolomic and (cryogenic) imaging applications in hospite may help shed light on fungal disease and opportunism in the tissues of coral holobionts. Parasitic interactions in the water column and implications for reef benthic–pelagic coupling Reef-building corals, the main ecosystem engineers of tropical coral reefs, are mixotrophic photosymbiotic holobionts. Corals ingest pelagic organisms ranging from mesozooplankton to phyto- and bacterioplankton, which can contribute significantly to the corals’ C and N budgets and maintain coral health and resilience under environmental stress (Grottoli et al. 2006 ). Coral reef and pelagic food webs are thereby inevitably and intimately linked via benthic–pelagic coupling. This implies that changes in pelagic food web dynamics may have cascading effects on coral reef food webs and nutrient cycling, and vice versa . One of the best studied examples of aquatic fungal interactions are parasitic associations of chytrids with phytoplankton (Kagami et al. 2007a , Klawonn et al. 2021 ). Despite the often-high proportion of infected host cells during chytrid outbreaks in some lakes (up to 90%; Kagami et al. 2006 , Rasconi et al. 2012 , Gerphagnon et al. 2017 ), fungal parasitism has been rarely considered in food web or nutrient cycling studies in other aquatic systems. However, fungal parasitism may be an integral part of aquatic food webs, as it affects fluxes of energy, nutrients, and elements (Kagami et al. 2014 ) and may help maintain the overall health of phytoplankton populations by selectively removing moribund and senescent cells (Laundon et al. 2021 ). Furthermore, fungal parasitism establishes novel trophic links by taking up organic matter from large “inedible” phytoplankton and by subsequently being consumed by zooplankton, which increases the efficiency of trophic transfer by drawing energy and nutrients up to higher trophic levels (Kagami et al. 2007b , Agha et al. 2016 , Sánchez Barranco et al. 2020 ). Klawonn et al. ( 2021 ) demonstrated the significance of this “fungal shunt” (originally described as “mycoloop;” Kagami et al. 2006 ) in a model freshwater diatom–chytrid pathosystem, where fungal infection affected holobiont organic C partitioning. Diatom-associated chytrid sporangia and free-swimming zoospores met their metabolic requirements by diverting C and N from their diatom host, while organic C retained in the host cell and transfer efficiencies of C and N to associated and free-living bacteria decreased significantly. Assuming an infection prevalence of 54% in a lake phytoplankton population, up to 20% of total diatom-derived photosynthetic C would be diverted to chytrids, bypassing the microbial loop (Klawonn et al. 2021 ). Parasitic chytrid outbreaks can thus shape microbially mediated C and N flows at the base of aquatic food webs and accelerate biogeochemical cycles. While currently little information is available on the prevalence, severity, and ecological importance of phytoplankton–chytrid infections in the marine realm, let alone the coral reef water column, potential fungal parasite-driven changes in food web dynamics could affect the energetics of hetero- and mixotrophic coral reef filter feeders via benthic–pelagic coupling (Fig. 4 ). Specifically, parasitic chytrids bypass the microbial loop and transfer energy from phytoplankton to grazing zooplankton (Kagami et al. 2007b , Klawonn et al. 2021 ). Considering that tropical coral reefs typically thrive in oligotrophic waters, even minor increases in reef-associated zooplankton biomass and/or nutrient content could result in significant ecological feedback, with (potentially beneficial) nutritional effects for coral reef benthic filter feeders (Fig. 4 ). While speculative at this point, controlled in vitro and mesocosm studies combined with metabarcoding approaches characterizing pelagic fungal communities on coral reefs may help elucidate the ecological significance of (pelagic) parasitic fungi in coral reef food webs. Challenges, open questions, and future directions Synthesis Fungi have conquered terrestrial, freshwater, and marine environments owing to unique sets of functional traits thought to facilitate fungal spread, diversification, and ecological adaptation. We here provided a conceptual perspective on the putative roles of understudied fungi on coral reefs by integrating our knowledge of these traits from other ecosystems and hosts with the current knowledge of fungal interactions on coral reefs, a brief summary of which is provided below. In the open oligotrophic ocean, fungal biomass is low compared to phytoplankton and bacteria, and likely particle associated. This may also apply to oligotrophic tropical coral reef waters, where pelagic parasitic and saprobic fungi may affect the reef-food web via benthic–pelagic coupling. While fungal contributions to reef biogeochemical cycling may be of smaller magnitude compared to terrestrial environments, it could be of ecological relevance at different scales of biological organization on oligotrophic tropical coral reefs. On the reef, fungi may be ecologically relevant in specific scenarios, including in reef substrata, in the reef framework, and in complex holobionts. Fungi engaging in microbe–microbe or host–microbe associations may be a driver of nutrient cycling and microbial community structuring via chemical mediation. Fewer examples of fungal interactions are known from marine compared to terrestrial environments, most of which are of pathogenic or opportunistic nature. On coral reefs, this includes disease-like phenotypes linked to Aspergillus spp. as well the seemingly opportunistic behavior of coral skeleton-associated fungi. Parasitic associations with coral reef phytoplankton affecting the pelagic food web and benthic–pelagic coupling may occur but remain yet to be confirmed. Based on our literature examination of fungal traits from various ecosystems and marine coculturing studies, we propose that there is appreciable potential for mutualistic fungal interactions on coral reefs (and potentially other marine environments) yet to be discovered. The lack of examples of such mutualistic associations of marine fungi with other biota may be due to a bias toward the study of fungal pathogens and opportunists, a trend noticeable also in the study of terrestrial crop and human pathogen emergence. Taken together, we anticipate that functions of reef-associated fungi are likely diverse, spanning a spectrum of interkingdom interactions that may mediate processes at different levels of biological organization, from cell–cell interactions to ecosystem-scale effects. These may be facilitated via fungal chemical communication and defense, microbiome structuring as well as biogeochemical cycling, thereby extending beyond their previously reported roles as pathogens and opportunists. In the following, we briefly summarize major challenges and open questions in the study of reef-associated fungi and propose a multidisciplinary toolbox to help address these questions. Open key questions and future directions Diverse knowledge gaps remain regarding the ecology of marine fungi in coral reef environments. Broadly, outstanding questions include: how does the taxonomic and functional diversity of fungal communities in corals and on coral reefs distribute across space and time; which ecological guilds and types of (symbiotic) interactions do occur in coral holobionts and on the reef; what are the mechanistic underpinnings of microbe–microbe and host–microbe interactions that fungi engage with in coral holobionts and on coral reefs; which (a)biotic drivers govern, maintain, and alter fungal communities and interactions; vice versa , how do fungal interactions on coral reefs shape their (a)biotic environment from the cellular to the ecosystem level? These questions could be addressed by employing a multipronged approach which targets different levels of biological organization for an integrated view of fungal functions in complex holobionts and the ecosystem (Fig. 5 ): Resolving the fundamental technical challenges and streamlining of fungal community assessment workflows will aid in addressing the key question of fungal diversity, distribution, and dynamics on coral reefs. The choice of molecular markers and DNA extraction tools for fungal community characterization is inherently biased (Frau et al. 2019 ), highlighting the importance of workflow optimization and standardization. Developments of metabarcoding markers for marine fungi should not only aim to increase marker specificity to reduce cross-amplification (Scholz et al. 2016 ), but could employ a greater diversity and/or combination of markers, for instance the internal transcribed spacer (ITS; ITS2) rDNAs in conjunction with the small and large ribosomal subunits (SSU and LSU rRNA, respectively) and/or different protein-encoding regions (Tekpinar and Kalmer 2019 ). Such efforts could further benefit from or be complemented by the application of long-read and hybrid sequencing applications for marker genes and/or entire metagenomes (Lücking et al. 2020 , Furneaux et al. 2021 ). Further, novel analytical frameworks to resolve genetic delineation in complex (marine) fungal communities may help solve shortcomings around the high intragenomic variance of some target regions commonly used for fungal metabarcoding, such as the ITS rDNA (Lindner and Banik 2011 ). A prominent example for such a framework is SymPortal , which resolves “defining” intragenomic variants for ITS2-type profiles of coral-associated endosymbiotic algal communities (Hume et al. 2019 ). Finally, the development of specific markers for fungal functional genes encoding for metabolic pathways perceived to be potentially relevant for coral and reef health could further be explored, including but not limited to selected CAZymes, DMSP lyases or genes associated with major fungal N cycling pathways. Such developments in conjunction with increased sequencing efforts and availability of genomic information may help improve community characterization at higher taxonomic levels, and ultimately allow for a more accurate assignment of functional groups (Bahram and Netherway 2022 ). At the same time, current efforts toward database optimization and expansion are critical for the meaningful interpretation of phylogenetic and diversity data (e.g. Martorelli et al. 2020 ), and will ultimately aid the prioritization to include marine fungal diversity in conservation efforts (Vatova et al. 2022 ). Functional work on coral reef-associated fungi (and prokaryotes) is challenged by e.g. the pervasive amount of host-derived nucleic acids, which constitute a major hurdle for culture-independent sequencing approaches, as well as well-known limits to microbial cultivability (Alain and Querellou 2009 , Robbins et al. 2019 , Pogoreutz et al. 2022 ). Here, a range of technologies could be accessed to address these challenges. Laser-capture microdissection approaches could be employed to target selected fungal cells in different holobiont compartments, i.e. (coral) tissues or microenvironments such as the ectodermis, gastrodermis, mesoglea, gastric cavity, skeleton, or mucus. Such obtained samples could be used for low-input omics techniques such as single-cell genomics, transcriptomics, proteomics, or metabolomics to elucidate fungal activities and putative interactions in the intact symbiosis (Hughes et al. 2022 ). Further, customized microfluidics platforms (Glossary) could aim to accommodate a range of eukaryotic cell shapes (including filamentous and branched structures; Millet et al. 2019 ) and sizes, potentially in combination with high-throughput microbial culturing approaches to increase isolation and culturing success of slow-growing or viable, but not (currently) culturable reef-associated fungi. Such platforms may permit the application of novel coculturing or microcosm approaches mimicking small scale environments, such as the coral host environment or the phycosphere of algal symbionts, as recently established for phytoplankton-associated bacterial communities (Raina et al. 2022 ). Ultimately, the proposed approach may lead to the discovery of new marine symbioses, will increase the availability of genomic and functional data, may aid (co-)cultivation efforts (fungal–bacterial, fungal–fungal, fungal–microalgal, and host–fungal) for experimental interrogation (Millet et al. 2019 ), as detailed below, and may ultimately lead to the discovery of new marine symbioses. The nature and specific mechanisms underpinning microbe–microbe interactions (fungal–bacterial and fungal–algal) could be addressed in a combination of different culture-dependent and -independent applications. High-throughput OSMAC applications on microbial cocultures (Cavaliere et al. 2017 ) coupled with metatranscriptomics, -proteomics, or metabolomics would not only permit the comparison of metabolic profiles, but enable an integrated assessment of chemical crosstalk in synthetic microbe–microbe associations in a range of different environmental conditions and substrates ( sensu Presley et al. 2020 ). Further, anabolic turnover (Glossary: anabolism) and exchange of metabolites in these microbe–microbe associations could be mapped and quantified at (sub)cellular resolution by combining cocultures grown on isotopically labeled substrates with stable isotope probing (SIP; Glossary) and Nanoscale or Time of Flight Secondary Ion Mass Spectrometry (Nano- and ToF-SIMS; Glossary; Raina et al. 2017 ). Specifically, cultures of algae, bacteria, and fungi originally isolated from coral and other reef holobionts could be labeled separately with distinct isotopes (e.g. 13 C, 15 N, and 34 S) prior to unlabeled coculture, and subsequently preserved and prepared for correlative electron microscopy and SIMS analysis to visualize (sub)cellular assimilation and distribution of isotopic labels and metabolic interactions between cells. Such approaches, especially in combination with correlative fluorescence in situ hybridization (FISH; Glossary) and omics applications as outlined above, would allow for the assessment of interactions in natural microbe–microbe assemblages or in clearly defined, synthetic communities of reduced complexity by targeting specific microbial functional groups or combinations of taxa. Notably, cultures of major coral-associated algal symbionts are available at dedicated microbial culture collections and can be readily maintained in stable cultures. Similarly, the availability of well-characterized coral bacterial isolates is steadily increasing, facilitating functional studies of microbe–microbe interactions in vitro (Sweet et al. 2021 , Pogoreutz et al. 2022 ). Another interesting venue is the study of coral reef host–fungal interactions, which can be approached at either the cellular or the organismal level. In recent years, cell lines of corals, anemones and sponges have become available (Domart-Coulon et al. 2001 , Ventura et al. 2018 ), which can be leveraged for functional laboratory model-system approaches in nonmodel reef organisms at the cellular level. Such an approach employing a methodological toolbox as described in (3) would have multiple benefits: the characterization and visualization of real-time host–microbe interactions in simplified, defined holobionts (one host, one or multiple selected microbes) and the study of metabolic interactions without the confounding effects of host–host cell interactions. The use of stable isotope labeling approaches to target specific metabolisms and trace the fate of specific molecules or substrates (e.g. fungal assimilation, fungal–host translocation, or vice versa ), as detailed for the study of complex multipartite host–microbe interactions (Lê Van et al. 2016 , Rädecker et al. 2021 ) may here be particularly suitable. At the organismal level, understanding the roles of different fungi in the coral holobiont (e.g. mutualists, opportunists, and pathogens) will be central for our understanding of coral holobiont health and resilience to different environmental stresses. For example, mutualistic fungal strains could potentially be used in novel probiotic and reef restoration applications, while an understanding of identity, drivers and mechanisms of fungal pathogenicity may inform the diagnosis, mitigation, or possibly even prevention of coral disease outbreaks via meaningful management tools. Here, host inoculation experiments with potential mutualistic and/or pathogenic strains could be used for functional interrogation. Specifically, a combination of in-depth physiological phenotyping assays (e.g. photophysiological parameters, algal symbiont densities in hospite , oxygen fluxes, nutrient uptake and release dynamics, oxidative stress, and so on; Rosado et al. 2018 , Doering et al. 2021 , Rädecker et al. 2021 ) and (meta)transcriptomic, (meta)proteomic, or (holo)metabolomic assessments ( sensu Mohamed et al. 2018 , Santoro et al. 2021 ) of the host or holobiont could be employed. Another perspective which warrants exploration is the notion that the microbial community may be able to modulate coral host phenotypic responses via epigenetic modification (Barno et al. 2021 ). Fungi exhibit epigenetic interactions with plant hosts (de Palma et al. 2019 ). It will, hence be of interest to characterize whether coral-associated fungi are capable of causing epigenetic changes in immune or environmental response genes in their host, whether such epigenetic changes result in distinct host phenotypes, and whether they may affect host resilience to environmental stress. Together, the proposed approaches may help elucidate not only the nature of the association of diverse fungi with their hosts, but also of the fungal potential to mitigate holobiont stress in global change scenarios. Finally, assessing the roles that fungi play at the community and ecosystem-levels and their spatio-temporal dynamics will be critical to advance the knowledge of fungal functions in coral reef ecosystems. Understanding and predicting how changes in coral reef fungal communities might translate into community-scale cascade effects and shifts should become a major goal. These include but are not limited to the (a) determination of (a)biotic drivers of fungal distribution and abundance in space and time, (b) quantification of fungi-specific activities at the community level such as their involvement in nutrient and energy transfer within and between the benthic and the pelagic reef environment, and (c) modeling of community-level budgets (e.g. C or N budgets) as well as ecological networks under consideration of fungal activity. In terrestrial environments, climate has been often reported as one of the strongest drivers affecting fungal community composition (Cavicchioli et al. 2019 , Egidi et al. 2019 ), therefore, a considerable challenge will be to understand the dynamics and biogeography of reef fungal communities. Disentangling how anthropogenic impacts and climate change affect both the structure and functioning of reef fungal communities will be paramount and will inevitably require modeling the effects of changing environmental conditions on biotic interactions (e.g. changes in use of resources, competition, or pathogenicity). To tackle such a challenging endeavor, it will be necessary to identify and quantify rates of specific fungal metabolisms of interest across different benthic and pelagic coral reef microhabitats. This can for instance be achieved by systematic and high-resolution sampling campaigns at smaller spatial and temporal scales, which can be generated through in vitro (on cultures), ex situ (in holobionts, such as individual corals or polyps), or in situ (on structurally complex benthic communities) experiments, as recently established for rate measurements of selected prokaryotic or holobiont-level metabolisms, such as dinitrogen fixation or oxygen fluxes (Cardini et al. 2016 , Roth et al. 2019 ). In situ set-ups could be equipped with multiple sampling ports for the controlled and reproducible sampling of different biological and chemical variables (Roth et al. 2019 ), and employed in either natural (e.g. comparison of metabolic rates between coral reef sites with different degrees of anthropogenic disturbance or natural variation, such as thermal fluctuations) or manipulative experiments (e.g. in situ simulation of major environmental stressors, such as global warming, ocean acidification, or eutrophication). Such approaches may hold the promise to address the work outlined above in a) to c), to help elucidate fungal contributions to reef-scale biogeochemical cycling. Figure 5. Schematic summary of key research questions and topics to improve our understanding of fungal communities and roles in coral reefs ecosystems ranked by spatial scales and biological complexity. Glossary Anabolism vs. catabolism The former is the synthesis of complex molecules (i.e. catabolites) from simpler ones, which requires energy, whereas the latter is the breakdown of complex molecules into simpler ones (i.e. catabolites) releasing energy. Antagonistic vs. synergistic interactions The former describes a cumulative effect, i.e. less than additive, i.e. less than the sum of effects (for instance, by stressors or organisms acting in isolations). The latter defines a cumulative effect greater than the additive sum of effects. Aspergillosis A disease caused by fungi of the genus Aspergillus . These fungi can infect a wide range of hosts, ranging from coral to humans. Benthic–pelagic coupling Processes that connect the sea floor (i.e. the benthic zone) and the water column (i.e. the pelagic zone) through the exchange of energy, mass, or nutrients. It plays a prominent role in nutrient cycling and energy transfer in aquatic food webs, and thereby ecosystem processes. BGC: tightly linked sets of (mostly) nonhomologous genes participating in a common, discrete metabolic pathway. The genes are arranged in physical proximity to each other on the genome, and their expression is often coregulated. Common in bacterial and fungal genomes and most widely known for the production of secondary metabolites. CAZymes Carbohydrate-active enzymes which build and break down complex carbohydrates and glycoconjugates for a large body of biological roles (collectively studied under the term of Glycobiology). Chemotaxis (or chemotactic behavior) Movement of a cell or organism in response to an environmental diffusible chemical substance. Chytridiomycota (chytrids) Unicellular or mycelic, aerobic zoosporic fungi that operate as saprotrophs and pathogens in freshwater, brackish, and marine habitats. Cnidaria A phylum within the animal kingdom which includes jellyfish, anemones, and reef-building corals. Cnidaria are simple, multicellular organisms characterized by two main cell layers (ectodermis and endodermis) and an apparatus consisting of stinging cells (cnidocytes) for prey capture and defense. Dimorphic switching The ability of several fungi to switch between a multicellular hyphal and unicellular yeast morphology and growth form. The mechanism underlying this biological reorganization process depends on external (environmental/chemical) triggers. Dimethylsulfonioproprionate (DMSP) An organosulfur compound produced in vast quantities by phytoplankton and seaweeds, and known to have osmoprotectant and antioxidant function. DMSP is an important carbon source for many marine bacteria, which can break it down via the DMSP demethylation or DMSP cleavage pathways. Ecological guilds Any group of species that exploit the same resources, or that exploit different resources in related ways. Among fungi, common guilds are decomposers (saprobic), pathogens, endophytes, and mycorrhiza. Endolithic (microbial) communities A group of organisms including cyanobacteria, fungi, algae, and bacteria that dwell in the pore space of rocks and similar substrates, such as coral skeletons. FISH A molecular technique that uses fluorescent probes that bind to only particular parts of a nucleic acid (DNA or RNA) sequence with a high degree of sequence complementarity. Widely used in the field of microbial ecology to identify taxa and to visualize the distribution and proportion of specific taxa within environmental samples. MALDI-TOF Matrix Assisted Laser Desorption Ionization coupled to Time-Of-Flight mass spectrometry. Microbialization The observed shift in ecosystem trophic structure toward higher microbial biomass and energy use. On coral reefs, causes of microbialization include overfishing and eutrophication. Microfluidics Refers to the behavior, precise control, and manipulation of fluids geometrically constrained to a small scale. In (micro)biology, it offers a powerful approach to control the complete cellular environment, thereby enabling the study of microbial community microscale organization, cellular behavior, adaptation, or gene expression. Mycorrhizal fungi Mycorrhizae are soil-borne fungi closely associated with the roots of terrestrial plants. Arbuscular mycorrhiza colonize the intercellular spaces of plant roots (in contrast to ectomycorrhizal fungi). Arbuscular mycorrhizae are considered vital endosymbionts of plant holobionts, as they enhance productivity. Next generation sequencing Sequencing is the process of determining the order of nucleotides in entire genomes or targeted regions of DNA or RNA. Next-generation sequencing (NGS) is a technology that offers ultrahigh throughput, scalability, and speed, and includes applications such as metabarcoding (deep sequencing of target regions) for whole community studies, whole genome sequencing, RNA sequencing, or the assessment of genome-wide DNA methylation and DNA–protein interactions. Oxidative stress and oxidative stress responses The former is the imbalance between the systemic manifestation of reactive oxygen species (ROS) and a biological system's ability to readily detoxify the reactive intermediates or to repair the damage resulting in cellular components, including proteins, lipids, and the DNA. Oxidative stress responses encompass the production of antioxidant enzymes (including but not limited to superoxide dismutase, catalases, and peroxiredoxins), which aim to strike a balance between ROS production and consumption. Oxylipins Lipids, often bioactive, generated by the oxidation of polyunsaturated fatty acids (PUFAs). VOCs Compounds with a high vapor pressure and low water solubility. Although VOCs gas can be emitted from different solids and liquids, in this manuscript we only refer to VOCs of biogenic origin (i.e. produced by living organisms). Volatilome Study of all the VOCs that are produced by a biological matrix (organism and ecosystem). OSMAC One Strain MAny Compounds is an approach, which by altering cultivation parameters (e.g. medium composition, physical properties, or strain coculture), aims to activate silent BGCs and expand or modify the metabolite production fingerprints of microbial strains. Probiotics Live microorganisms with beneficial qualities for a host/recipient organism. Probiotic microorganisms help to restore recipient health by antagonistic action against pathogenic microbes, or enhance performance and growth by providing nutritional benefits to the recipient. Recalcitrant vs. labile Describes the bioavailability for or timescales of degradation of organic matter by organisms, which is reflected in the timescales by which this matter is respired. Organic matter follows a spectrum of recalcitrant (degraded slowly over years to decades, or resistant to degradation by microbes) to labile (rapidly degraded, within minutes to hours). Secondary ion mass spectrometry (SIMS) A technique used to analyze the composition of solid surfaces and thin films, which permits the spatial mapping of atoms or molecules. In correlation with electron microscopy increasingly used in biological research to create isotopic maps of histological sections, which can be used to visualize the assimilation and translocation of nutrients within complex symbiotic systems, such as corals. SIP A technique in microbial ecology for tracing uptake of nutrients by microbes. A substrate is enriched with a heavier stable isotope, i.e. consumed by the organisms to be studied. Biomarkers with the heavier isotopes incorporated into them can be separated from biomarkers containing the more naturally abundant lighter isotope by buoyant density centrifugation. As an example, 15 N 2 can be used to find out which microbes are active nitrogen fixers. Transcription Process of making an RNA copy (mRNA) of a gene's DNA."
} | 26,884 |
37992173 | PMC10664982 | pmc | 1,124 | {
"abstract": "Cross-linked elastomers are stretchable materials that typically are not recyclable or biodegradable. Medium-chain-length polyhydroxyalkanoates (mcl-PHAs) are soft and ductile, making these bio-based polymers good candidates for biodegradable elastomers. Elasticity is commonly imparted by a cross-linked network structure, and covalent adaptable networks have emerged as a solution to prepare recyclable thermosets via triggered rearrangement of dynamic covalent bonds. Here, we develop biodegradable and recyclable elastomers by chemically installing the covalent adaptable network within biologically produced mcl-PHAs. Specifically, an engineered strain of Pseudomonas putida was used to produce mcl-PHAs containing pendent terminal alkenes as chemical handles for postfunctionalization. Thiol-ene chemistry was used to incorporate boronic ester (BE) cross-links, resulting in PHA-based vitrimers. mcl-PHAs cross-linked with BE at low density (<6 mole %) affords a soft, elastomeric material that demonstrates thermal reprocessability, biodegradability, and denetworking at end of life. The mechanical properties show potential for applications including adhesives and soft, biodegradable robotics and electronics.",
"introduction": "INTRODUCTION Polyhydroxyalkanoates (PHAs) are natural polyesters synthesized by microorganisms for carbon and energy storage ( 1 – 3 ). PHAs have been long studied as thermoplastics that are biodegradable and biocompatible, making these bio-based polyesters inherently performance advantaged ( 4 ). To date, biotechnological efforts have resulted in the capability to design and scale tailor-made PHAs produced in various organisms ( 2 , 3 , 5 ) and from expanded feedstocks such as lignocellulose-derived sugars ( 6 ), aromatic compounds ( 7 ), and plastics ( 8 , 9 ). Besides microbial PHA production, chemocatalytic synthesis offers rapid and precision polymerization of designer monomers to PHAs ( 10 ). PHA properties are usually modulated by varying the chemistry of the aliphatic side chains [primarily based on the poly(-3-hydroxyalkanoate) backbone, β-PHAs]. In general, short-chain-length PHAs (scl-PHAs) (C 4 -C 5 monomers, corresponding to C 1 -C 2 side chains) are rigid, semicrystalline plastics with glass-transition temperatures ( T g ) from ~0° to −20°C and high melting temperatures ( T m ) ~100° to 180°C. Medium chain length (mcl, C 6 -C 12 monomers, ≥C 3 side chains) PHAs are soft materials exhibiting T g values ca. −30° to −50°C and elastomeric tensile profiles (no yield point). After a threshold side-chain length of ~C 5 , mcl-PHAs become semicrystalline with a helical backbone and ordered side chains ( 11 – 13 ). Longer side chains of >C 11 are less common and exhibit further side-chain crystallinity ( 5 , 14 ). An even greater variety of copolymer compositions and properties can be achieved with scl- co -mcl PHAs, enabled by engineering the PHA synthase to accept a wide substrate range ( 2 , 5 , 15 ). Sustainable thermosets based on cross-linked PHAs remains an area rich for innovation ( 16 – 20 ). To enable thermoset PHAs, chemical handles for postfunctionalization can be introduced to PHAs during biosynthesis. As a prominent example, alkene-containing comonomers (derived, for example, from unsaturated fatty acid feedstocks) can be incorporated to yield unsaturated PHAs (usPHAs), thus providing a convenient platform for materials design ( 16 , 18 , 21 – 24 ). Since the 1990s, usPHAs have been explored as substrates for grafting of small molecules ( 18 , 25 ) and polymers ( 23 , 25 ) as well as for cross-linking reactions ( 16 – 18 , 20 , 26 ), to produce water-soluble polymers, latex suspensions, hydrogels, tissue scaffolds, drug delivery polymers, and conductive nanocomposites. Cross-linking usPHAs can be achieved by a variety of methods, including thiol-ene coupling, epoxidation and ring-opening, and sulfur vulcanization; out of these, the thiol-ene coupling (or click) reaction is straightforward, well-controlled, and commonly applied to alkene-containing polymers ( 16 , 27 ). Given the variety of chemical motifs that can be incorporated during postfunctionalization, there is also an opportunity for cross-linked PHAs to be circular, specifically through dynamic covalent cross-linking. Covalent adaptable (or dynamic covalent) networks (CANs) have emerged as a strategy to prepare recyclable thermosets ( 28 – 32 ), as damaged materials can be thermally treated to reconfigure the network. Vitrimers are classified as CANs with an associative exchange mechanism, in which cross-link density remains constant during exchange events. The boronic ester (BE) is an associative linkage with broad utility, as dioxaborolane-linked networks undergo metathesis at mild temperatures and without a catalyst ( 33 – 37 ). BEs are convenient to install and have been introduced through thiol-ene coupling of BE-containing linkers. A variety of commodity polymers have been made into BE vitrimers ( 36 ), including polybutadiene and butadiene-styrene rubbers ( 35 , 38 ). In this work, we use mcl-usPHA, thiol-ene, and dynamic covalent chemistries to produce networks that are both reprocessable and biodegradable. Beginning with the soft and semicrystalline usPHA poly-( R )-(3-hydroxydecanonate- co -3-hydroxyundecenoate) (PHDU) produced by our engineered strain of Pseudomonas putida , we target BE networks with low and high cross-linking densities from PHDU with the molar ratio of undecenoate (%U) = 6 and 22 mol%, respectively, as illustrated in Fig. 1 . We aimed to understand the effects of cross-linking on thermomechanical and rheological material properties, as well as end-of-life scenarios. The structure-property relationships of these PHA-based vitrimers can inform future material design toward targeted applications in the field of sustainable elastomers. Fig. 1. PHDU vitrimer life cycle. Production of PHDU in engineered P. putida from fatty acid substrates, postfunctionalization into BE–cross-linked vitrimers (PHDU-BE), and multiple end-of-life scenarios. RT, room temperature; DMPA, 2,2-Dimethoxy-2-phenylacetophenone; THF, tetrahydrofuran.",
"discussion": "DISCUSSION To develop performance-advantaged bioproducts, baseline structure-property relationships measured in early research can inform next-generation materials that are optimized for specific applications ( 4 ). Once candidate materials are identified, aspects of the entire life cycle can be further optimized for sustainability and cost. Starting from the bioprocess component of this work, sodium decanoate and 10-undecenoic acid can be prepared from vegetable oil and were used here to generate precise and reproducible polymers. Alternatively, the precursors could also be derived from central metabolite acetyl-CoA and synthesized via the FAB pathway ( 3 ). As acetyl-CoA is the starting metabolite in FAB and can be produced from various renewable feedstocks, this provides an opportunity to expand the carbon sources ( 7 , 8 , 50 , 51 ). Another important consideration for any bioprocess is the separation of the final product. Efficient separation technologies of PHAs from cells are an active research area ( 52 , 53 ). For the cross-linking methodology, a solventless melt-state process would be ideal, but requires good mixing of several species and this can be challenging at a small scale. On a larger scale, a UV-assisted 3D printing approach could be developed for cross-linking mcl-usPHAs using either dynamic or permanent cross-links ( 54 ). The thermomechanical properties of PHDU-BE materials exhibit a dependence on morphology and cross-link density, which are interconnected properties affected by the moisture sensitivity of the BE linkage ( 29 , 33 ). As BE linkages hydrolyze in equilibrium with the environment, there is greater mobility for crystallization, and crystallization kinetics likely increase as the cross-link density decreases ( Fig. 8 ). As revealed by WAXS, the proposed crystal structure is like the linear PHA with some distortion. Overall, lower crystallinity contributes to lower σ B and moduli of the cross-linked PHDU. Reports of more stable BE vitrimers are based on more hydrophobic materials [e.g., polybutadiene ( 38 ) and polyethylene ( 36 )] and/or with higher thermal transitions [e.g., polystyrene; ( 36 )], whereas PHDU is amphiphilic ( Fig. 8 ). Fortunately, recent research has demonstrated BE linkers with enhanced stability ( 46 ). Fig. 8. PHDU-BE network characteristics. Cartoon representation of a PHDU-BE network ( middle ) with chain entanglements, closed and open BE linkages, and crystallites ( left ); ( right ) the interplay among moisture, cross-link density, crystallization, and mechanical properties. The dependency of material properties on cross-link density is also exemplified by the weak and brittle quality of PHDU-BE-22. Study of PHDU-BE-22 by HR-MAS and ssNMR revealed different domains of rigidity in the material, perhaps consistent with the brittle physical properties. In addition, these studies revealed that the grafting density from thiol-ene coupling is reliably quantified by HR-MAS only when the network is fully penetrable by solvent. The differences between PHDU-BE-6 and PHDU-BE-22 invite further study into cross-link density effects in a controlled fashion, by using a more stable dynamic linkage. We also emphasize that many elastomers are composites with fillers (e.g., carbon black, silica, and biomass) that toughen the material and can increase moduli values by greater than one order of magnitude. Thus, effects of filler on mcl-PHA elastomers also merit further investigation. The observed range of properties provides insights into the property space for mcl-PHA–based elastomers. These structure-property relationships facilitate future material design through judicious selection of linear mcl-usPHA, cross-link density, and cross-link type. For example, beginning with less crystalline, hexanoate-based usPHAs could suppress crystallization even at low cross-linking densities, while longer side chains may promote crystallinity. Notably, the properties of semicrystalline PHDU-BE-6 are similar to elastomeric BE vitrimers prepared from polybutadiene and polybutadiene- co -styrene, including metrics for T g , E′ , E , ɛ B , and σ B ( 35 , 38 ). Semicrystalline PHDU-BE-6 properties also showed similarity to a silicone rubber control. While the elastic recovery of the PHDU elastomer could be improved, perhaps by mixed static and dynamic linkages, these similarities suggest that mcl-PHA elastomers may be a suitable replacement for butadiene and silicone-based rubbers when degradability is desired. Our PHDU-BE networks compare well to other biodegradable and/or self-healing polymeric networks reported in the literature ( 26 , 55 – 57 ). Sustainable soft materials are desired in the fields of soft robotics and soft electronics for both environmental and human-facing applications ( 55 , 58 – 61 ). In soft robotics applications, there is an array of design parameters and challenges including tissue-like character (e.g., E = 0.1 to 10 MPa), elasticity, melt processability into unique shapes, self-healing ability, and circularity via bio-derived materials with biodegradability and/or recyclability ( 55 ). Elastomers from mcl-PHAs can meet each of these, while dynamic cross-links solve the conflicting nature of melt processing and elasticity requirements ( 54 , 56 , 62 – 64 ). Our examination of PHDU-BE-6 viscosity showed that extrusion-based processing techniques for thermoplastics, including 3D printing, would be possible. Triggered self-healing or reprocessability is also ideal for device repair and prototyping, while denetworking allows recovery of other components ( 55 , 56 , 64 ). The balance of material stability with biodegradability, as demonstrated by the freshwater biodegradation test, is another advantage for biodegradable robotics or electronics, such as for soil or marine sensors ( 59 , 65 ). We note that our aerobic freshwater biodegradation results are consistent with the literature for mcl-PHAs, which show markedly slower degradation kinetics than scl-PHAs ( 66 ). Another potential application for mcl-PHA elastomers is removable adhesives or fugitive glue, often used in advertising. Amorphous PHDU-BE-6 is tacky and elastic and showed good adhesion to plastic, metal, and glass surfaces while easily removed with no residue. In summary, we developed elastomeric vitrimers based on mcl-usPHAs. The BE linkage can hydrolyze during ambient storage, but this labile linkage allowed us to study the structure-property relationships between the cross-link density and network morphology, which is critical to future material development. The soft elastomer PHDU-BE-6 shows mechanical and processing properties that are desirable for soft, biodegradable robotics and electronics, biomedical devices, and removable adhesives, with several sustainable end-of-life routes: reprocessing, denetworking, and biodegradation."
} | 3,261 |
32098815 | PMC7042694 | pmc | 1,125 | {
"abstract": "The initial pivotal phase of bacterial biofilm formation known as reversible attachment, where cells undergo a period of transient surface attachment, is at once universal and poorly understood. What is more, although we know that reversible attachment culminates ultimately in irreversible attachment, it is not clear how reversible attachment progresses phenotypically, as bacterial surface-sensing circuits fundamentally alter cellular behavior. We analyze diverse observed bacterial behavior one family at a time (defined as a full lineage of cells related to one another by division) using a unifying stochastic model and show that our findings lead to insights on the time evolution of reversible attachment and the social cooperative dimension of surface attachment in PAO1 and PA14 strains.",
"introduction": "INTRODUCTION Biofilms are surface-adhered communities or suspended aggregates of bacteria that have increased tolerance to environmental stresses and antibiotics and that impact human health and the environment in complex ways. These biofilms can be harmful by causing diseases ( 1 , 2 ) and can be beneficial by serving as commensals in various hosts; they also have applications in bioremediation and energy production ( 3 ). A critical step in forming a bacterial biofilm is surface sensing ( 4 ), where free-swimming planktonic cells detect, attach to, and physiologically respond to a surface. Recent work has shown that different appendages or extracellular structures, such as flagella ( 5 , 6 ) or type IV pili (TFP) ( 7 , 8 ), are involved in activating cellular responses (e.g., protein production, motility, and biofilm formation) during surface sensing. In many bacterial species, these responses are controlled primarily by intracellular secondary messenger molecules, such as cyclic diguanylate (c-di-GMP) ( 9 – 16 ) and cyclic AMP (cAMP) ( 8 , 17 , 18 ). For Pseudomonas aeruginosa , a clinically relevant model system ( 19 ), there are at least two well-studied but distinct surface-sensing circuits, the Wsp and the Pil-Chp systems, which can contribute to initiating biofilm formation. In our current understanding, the Wsp system senses through the membrane-bound, chemosensory-like Wsp protein complex, which localizes laterally along the cell body ( 10 ), activating the diguanylate cyclase WspR and c-di-GMP synthesis via a mechanism that requires WspR clustering ( 20 ). On the other hand, the Pil-Chp system senses a surface through polarly localized TFP, which activate the adenylate cyclase CyaB and result in cAMP synthesis. Increased cAMP levels then induce the production and secretion of PilY1, which in turn activates the diguanylate cyclase SadC and results in c-di-GMP synthesis ( 17 ). Downstream consequences of c-di-GMP synthesis include exopolysaccharide (EPS) production and motility suppression. Different strains of P. aeruginosa , such as PAO1 and PA14, utilize these surface-sensing mechanisms to various extents. The PAO1 strain uses predominantly the Wsp system ( 21 ), leading to the surface deposition of the EPS Psl ( 22 , 23 ), while PA14 uses predominantly the Pil-Chp system, leading to the suppression of surface motility ( 17 ) and production of a Pel-dominant biofilm matrix ( 24 ). Despite the existence of diverse machineries to sense, adhere to, and proliferate on surfaces, it is commonly observed that bacteria initially seem to have a difficult time attaching to a surface, as indicated by typical flow cell studies where P. aeruginosa often takes >20 h before attaching to the surface in large numbers ( 25 , 26 ). This phenomenon was first reported in the 1930s ( 27 , 28 ). Using high-speed microscopy to measure the distribution of surface residence times, it was previously observed that the overwhelming majority of cells that land on the surface eventually detach, and it is only after a prolonged and variable time lag that cells begin to rapidly cover the surface ( 8 ). We stress that the low apparent probability of successful attachment is not simply a matter of cells “bouncing” off the surface. For example, during reversible attachment, it is not uncommon for cells to attach and stay long enough to divide but then subsequently detach. Moreover, the unpredictability of reversible attachment cannot be circumvented with better measurement statistics; the duration of reversible attachment of individual cells and populations always appears random and does not converge to a specific duration for the same initial conditions. The foundational question that we address here is what bacteria are doing during this period of “reversible attachment” besides attaching themselves to the surface. For example, can an attaching cell help any other cell attach to the surface? If so, does it help all cells or employ a more selective strategy to help either nearby cells (spatial neighbors) or their progeny (temporal neighbors) to attach? The combination of defining characteristics in reversible attachment, a low probability of success, intrinsic time dependence, and structurally random outcomes suggests that the use of a stochastic model may lead to interesting answers. Here, we show that the use of an exactly solvable “divide-detach” stochastic model, designed to examine the reversible attachment behaviors of P. aeruginosa PAO1 and PA14 lineages in the form of family trees, reveals differences in their reversible attachment behaviors that suggest contrasting surface colonization strategies. Within this model, reversible attachment is described by two parameters: effective division rate and effective detachment rate. We find that reversible attachment can be understood if we analyze behavior using lineage time (the time that a lineage stays continually on the surface) rather than an experiment time, defined by the time from inoculation. Specifically, reversible attachment comprises two regimes of behavior, defined by whether cells of the lineage stay on the surface long enough to divide, or not, before detaching. For lineages that detach before dividing at all, both PAO1 and PA14 behave similarly with nearly certain lineage “extinction,” wherein the entire lineage detaches. For lineages that stay long enough to divide, PAO1 and PA14 show surprisingly different behaviors. Our model provides a framework wherein time-dependent division and detachment rates and distributions of lineages can be extracted from our experiments. For PAO1, individual lineages commit relatively quickly to a surface compared to what occurs with PA14, resulting in PAO1 displaying a steadily progressive increase of a surface cell population that is irreversibly attached (i.e., committed to forming a biofilm). For the PAO1 strain, as reported previously ( 21 ), the Wsp-based surface sensing results in early c-di-GMP-mediated EPS production that can promote attachment of a cell’s spatial neighbors. In contrast, PA14 lineages exhibit high rates of cell detachment from surfaces. However, Pil-Chp surface sensing modulates motility via cAMP and allows progeny cells to retain a memory of the surface ( 8 ), so that PA14 lineages ultimately form a planktonic population that is primed for improved surface attachment; this process thus ultimately promotes irreversible attachment and biofilm formation. Our model provides a framework for understanding the cooperative and social nature of surface attachment and for categorizing different surface colonization strategies that lead to biofilm formation, each presumably with its own advantages under different circumstances.",
"discussion": "DISCUSSION Clearly, the application of stochastic models can be quite powerful in understanding microbiological systems that involve strong fluctuations. The behavior of each lineage is a record of how a specific cell and its progeny managed to stay and proliferate on the surface during cellular changes induced by surface sensing, which has multigenerational consequences. Even though the probability of a specific cell attaching to a surface and proliferating successfully is initially vanishingly small, surface sensing can modify outcomes by changing the structure of family trees, as we can see from the evolution of reversible attachment from the nonprocessive to processive regimes, for example. Interestingly, that the process of reversible attachment can be described by a stochastic model is telling: whether a bacterium encountering a surface makes it to irreversible attachment and eventually participates in biofilm formation may be quantitatively cognate to the description of whether patient zero’s disease will die out after a few infections or take hold and become an epidemic. The fact that biofilm formation seems to inevitably happen is due to factors such as the large number of lineages that encounter the surface and the existence of multigenerational memory, which can mitigate against initial failure to attach by conditioning a planktonic population primed for improved subsequent attachment. Indeed, a recent study applied a variation of our approach to antibiotic treatment of bacteria ( 39 ). In fact, the quantitative evolution of bacterial populations in early biofilm formation is analogous to a time-reversed version of antibiotic treatment: the nonprocessive regime of reversible attachment behaves like bacterial population dynamics for antibiotic treatment well above the MIC. In the present study, however, we are able to perform an unprecedented level of longitudinal comparison between theory and experiment. Because we have information on the fates for every cell in a large number of bacterial lineages that occur during early biofilm formation, we can directly measure and analyze the time evolution of the system. This analysis provides a conceptual framework for understanding the taxonomy of surface colonization strategies and reveals an unanticipated difference between PAO1 and PA14 behavior. One of the old questions about biofilm formation is whether it is the newly landed cells or the dividing cells on the surface that contribute more to the biomass increase in the biofilm. Our results suggest that not only is the answer species and strain dependent, the question itself is misleading because of the assumed either/or format of the answer. Surface sensing can evolve progenitor cells which land on a surface and commit almost their entire division lineage to the surface, thereby drastically increasing biomass. Furthermore, our results suggest that surface attachment and reversible attachment can have a social dimension; when bacteria attach to the surface, they can help other cells in the population to attach and remain adhered, as described below. Attachment of PAO1 promotes attachment of neighboring cells; attachment of PA14 promotes attachment of progeny cells. The divide-detach stochastic model highlights two distinct but complementary strategies for surface colonization that are illustrated by PAO1 and PA14. For PAO1, surface population increase takes the form of the few families that are more successful in retaining surface progeny. PAO1 families generally stay on the surface during biofilm formation, likely due to the Wsp surface-sensing system and early Psl EPS secretion, which facilitates surface attachment of a cell’s spatial neighbors. Previous work has shown that early surface attachment behavior depends on EPS production via the Wsp system ( 9 , 23 ). In contrast, for PA14, surface population increase takes the form of many families that are less successful in retaining surface progeny due to surface detachment. However, PA14 cells and their progeny can “remember” the surface due to the Pil-Chp system and multigenerational cAMP-TFP memory ( 8 ), which primes them for biofilm formation whether they are currently on the surface or not and eventually leads to progressive suppression of motility appendage activity. Both strategies are viable for surface colonization. PAO1 cells tend to attach, increase their surface population more quickly, and persist longer on a surface than PA14 cells, which suggests that PAO1 can potentially attach to surfaces even in ecologically crowded environments or successfully form biofilms by outgrowing competing species. Indeed, this has been experimentally observed: EPS-producing P. aeruginosa strains tend to persist on surfaces better than non-EPS-producers, despite possible exploitation by “cheaters” that can potentially use the communal good of EPS ( 40 ). In contrast, PA14 cells exposed to a surface do not initially stay on the surface and slowly increase surface coverage. Rather, they and their progeny form a surface-sensing-activated planktonic population that can quickly attach and colonize the surface later in time, which may be better adapted for overwhelming host defense (i.e., a naive surface) than for microbial competition. Moreover, it is interesting to note that EPS secretion is extracellular and can be shared both with neighbors from different lineages and with descendants in close proximity to help them attach and remain attached ( 41 ). On the other hand, memory is intracellular and can be passed down only temporally through division, thus allowing cells to help only their progeny to remain attached. It is possible that our observations and results with PAO1 and PA14 may be generalizable to other P. aeruginosa strains. The majority of strains in the International Pseudomonas Consortium Database (IPCD) can be identified as either PAO1-like or PA14-like based on their phylogeny (i.e., have the same phylogenetic subgroup as either PAO1 or PA14) ( 42 – 45 ). Consistently with our results, crystal violet biofilm assays show that the PAO1-like strains seem to produce early biofilms faster than the PA14-like strains ( Fig. S6 ). Although it is clear from the data spread that there is more to explaining differences in biofilm behavior than pseudomonad phylogenetic diversity, this observation suggests that the phylogenetic distance from either PAO1 or PA14 may be incorporated into a metric for categorizing a P. aeruginosa strain’s biofilm formation behavior as either PAO1-like or PA14-like. It is likely that these bacterial strategies have their own advantages under different circumstances. Furthermore, our model can be applied to other bacterial systems to understand how they utilize their cellular machinery for orchestrating different types of social cooperativity during surface attachment and for their implicit surface colonization strategies. 10.1128/mBio.02644-19.6 FIG S6 Crystal violet biofilm assay results for 35 P. aeruginosa strains (25 PAO1-like and 10 PA14-like strains, including PAO1 and PA14 strains) in the International Pseudomonas Consortium Database (IPCD). These strains are identified as either PAO1-like or PA14-like based on their phylogeny (i.e., in the same phylogenetic subgroup as either PAO1 or PA14) ( 42 – 45 ). The OD 550 values are proportional to the amount of biofilm stained by crystal violet. Circles represent individual biological replicates, each of which is the mean of results from 4 technical replicates. Longer horizontal lines represent the mean OD 550 values. Vertical lines and error bars indicate the 95% confidence interval calculated from the bootstrap sampling distribution of the mean OD 550 values. A comparison of these distributions shows that the mean OD 550 values for the PAO1-like strains are higher than the mean OD 550 values for PA14-like strains ( P value of 0.02). Download FIG S6, EPS file, 0.6 MB . Copyright © 2020 Lee et al. 2020 Lee et al. This content is distributed under the terms of the Creative Commons Attribution 4.0 International license ."
} | 3,919 |
22227949 | null | s2 | 1,126 | {
"abstract": "We study a range of neural dynamics under variations in biophysical parameters underlying extended Morris-Lecar and Hodgkin-Huxley models in three gating variables. The extended models are implemented in NeuroDyn, a four neuron, twelve synapse continuous-time analog VLSI programmable neural emulation platform with generalized channel kinetics and biophysical membrane dynamics. The dynamics exhibit a wide range of time scales extending beyond 100 ms neglected in typical silicon models of tonic spiking neurons. Circuit simulations and measurements show transition from tonic spiking to tonic bursting dynamics through variation of a single conductance parameter governing calcium recovery. We similarly demonstrate transition from graded to all-or-none neural excitability in the onset of spiking dynamics through the variation of channel kinetic parameters governing the speed of potassium activation. Other combinations of variations in conductance and channel kinetic parameters give rise to phasic spiking and spike frequency adaptation dynamics. The NeuroDyn chip consumes 1.29 mW and occupies 3 mm × 3 mm in 0.5 μm CMOS, supporting emerging developments in neuromorphic silicon-neuron interfaces."
} | 301 |
28569746 | PMC5461491 | pmc | 1,127 | {
"abstract": "Communication provides the basis for social life. In ant colonies, the prevalence of local, often chemically mediated, interactions introduces strong links between communication networks and the spatial distribution of ants. It is, however, unknown how ants identify and maintain nest chambers with distinct functions. Here, we combine individual tracking, chemical analysis and machine learning to decipher the chemical signatures present on multiple nest surfaces. We present evidence for several distinct chemical ‘road-signs' that guide the ants' movements within the dark nest. These chemical signatures can be used to classify nest chambers with different functional roles. Using behavioural manipulations, we demonstrate that at least three of these chemical signatures are functionally meaningful and allow ants from different task groups to identify their specific nest destinations, thus facilitating colony coordination and stabilization. The use of multiple chemicals that assist spatiotemporal guidance, segregation and pattern formation is abundant in multi-cellular organisms. Here, we provide a rare example for the use of these principles in the ant colony.",
"discussion": "Discussion Social insects are fascinating because they manage to coordinate their actions without central control. In the ant colony this coordination is tightly linked to the ants' spatial distribution. Combining a methodology for detecting chemicals adsorbed onto the nest surfaces with individual ant tracking techniques and machine learning, we have revealed that ants utilize complex chemical patterning within the nest. These chemical signatures constitute a new form of stigmergy and serve as ‘road-signs' that assist ant orientation, colony stabilization and spatial segregation between different task groups. Our work provides the first experimental evidence of a navigation mechanism of any possible type that aids ant orientation within the nest. While it was previously known that ant hydrocarbons are found on nest surfaces 25 27 28 , their function, if any, remained unclear. Here we show that chemical signatures assist navigation within the nest. We identify six different chemical profiles that signify different chamber functions within and around the nest (queen's chamber, worker chambers with or without brood, entrance chamber, refuse area and proximal foraging arena) and more probably exist. Moreover, we present direct behavioural evidence that the ants recognize and follow at least three of these road-signs into specific nest areas. Despite the richness of these chemical road-signs we hypothesize that this stigmergic mechanism is not exclusive and complemented by other subterranean navigational tools such as spatial memory and quorum sensing. The description of the social insect colony as a superorganism (or social-organism 48 ) has proven to be a useful analogy of exceptional breadth 49 . Our work suggests the addition of yet another layer to this correspondence by relating the nest chambers to different organs, and the ants to migrating cells 50 51 52 . Using the principles of stigmergy, passively or actively generated chemicals serve as the superorganism's chemokynes 53 . Multiple chemical signals serve to differentially guide and stabilize trafficking and organization of the mobile agents that make up the whole. Future work will be required to further explore the validity and usefulness of this compelling analogy."
} | 858 |
35889733 | PMC9322026 | pmc | 1,129 | {
"abstract": "Superhydrophobic materials have been widely applied in rapid removal and collection of oils from oil/water mixtures for increasing damage to environment and human beings caused by oil-contaminated wastewater and oil spills. Herein, superhydrophobic materials were fabricated by a novel polypyrrole (PPy)/ZnO coating followed by hexadecyltrimethoxysilane (HDTMS) modification for versatile oil/water separation with high environmental and excellent reusability. The prepared superhydrophobic surfaces exhibited water contact angle (WCA) greater than 150° and SA less than 5°. The superhydrophobic fabric could be applied for separation of heavy oil or light oil/water mixtures and emulsions with the separation efficiencies above 98%. The coated fabric also realized highly efficient separation with harsh environmental solutions, such as acid, alkali, salt, and hot water. The superhydrophobic fabric still remained, even after 80 cycles of separation and 12 months of storage in air, proving excellent durability. These novel superhydrophobic materials have indicated great development potentials for oil/water separation in practical applications.",
"conclusion": "4. Conclusions In summary, a novel PPy/ZnO/HDTMS coated superhydrophobic surface has been presented which was capable of oil/water separation even in harsh environments. The PPy/ZnO coating was prepared by in situ polymerization and deposition method, which was an effective method for constructing nanometer-scale grains on different substrates. The as-prepared materials exhibited superhydrophobic and superoleophilic properties, which were applied for immiscible oil/water mixtures and emulsions with separation efficiencies above 98% with a high purity of collected oils. Moreover, the superhydrophobic material also exhibited outstanding reusability and stability through 80 repeated separations of tetrachloromethane/water mixture and 12 months of exposure to air. The reported superhydrophobic surface has displayed strong potential for versatile oil/water separation needed for practical applications. The superhydrophobic material prepared in this work displayed outstanding performance for removing/collecting oil contaminants. This strategy for achieving superhydrophobic material is simple, feasible, and highly efficient, offering a wide range of potential applications in reclaiming the oil/water mixtures from chemical leakages and oil spills.",
"introduction": "1. Introduction Increasing water pollution from offshore oil spills and industry activities has threatened the human beings and environment seriously [ 1 , 2 ]. The traditional treatment methods, including microbial degradation, oil containment fences, and in situ burning, have disadvantages of low efficiency, low selectivity, and cumbersome operation. Thus, finding an effective method to realize rapid collection of organic solvents and oils from water has become urgent. The superhydrophobic material, which shows different interfacial behaviors towards oil and water, has attracted significant attention in oil/water separation. The superhydrophobic material is water-repellent, with a water contact angle (WCA) greater than 150° and a sliding angle (SA) smaller than 10°. The superhydrophobic surface could be realized by the following approaches: construction of a suitable roughness structure and application of low surface energy substances on rough surface. Various methods have been developed to prepare superhydrophobic surfaces, including sol–gel method [ 3 ], hydrothermal treatment [ 4 ], layer by layer assembly [ 5 ], electrochemical methods [ 6 ], chemical vapor deposition [ 7 ], plasma treatment [ 8 ], chemical etching [ 9 ], and self-assembly method [ 10 ]. The superhydrophobic materials were usually constructed on textile [ 11 , 12 , 13 ], fabrics [ 14 ], meshes [ 15 ], and so on. In particular, the inorganic micro/nanoparticles have been used to create the roughness structure for superhydrobicity preparation, such as TiO 2 [ 16 ], SiO 2 [ 17 ], Fe 3 O 4 [ 18 ], ZnO [ 19 ], and carbon nanotubes [ 20 ]. ZnO nanoparticles have significantly attracted interests due to their low cost, accessibility, and stability in application [ 21 ]. Barthwal et al. [ 22 ] introduced a superhydrophobic coating, which is composed of multi-walled carbon nanotubes/ZnO composite modified with polydimethylsiloxane. The coating showed excellent performance in self-cleaning, anti-fouling, and oil/water separation. Wei et al. [ 23 ] prepared superhydrophobic material for oils purification. The ZnO nanoflower modified SiC composite ceramic membranes were fabricated by chemical bath deposition method. Sun et al. [ 24 ] fabricated a superhydrophobic polyurethane sponge-based coating on a ZnO/epoxy resin solution followed by modification with stearic acid. Pal et al. [ 25 ] achieved superhydrophobic substrates based on Ag/ZnO/Ag hybrid structure. These methods for fabrication of superhydrophobic coating always require a time-consuming and tedious process. Most methods faced common drawbacks, such as air pollution, poor bonding strength between substrate and coating, specialized reagent, and harsh condition requirements [ 26 , 27 , 28 , 29 ]. In addition, most of the superhydrophobic materials could be damaged under corrosive solutions and harsh conditions, which has greatly limited their application in practical separation of real oil with harsh aqueous solutions [ 30 ]. Therefore, developing a novel superhydrophobic material with simple process, high environmental ability, and durability is of great importance for its practical applications in oil/water separation from chemical leakages and oil spills. For realizing the low surface energy on rough surface, hexadecyltrimethoxysilane (HDTMS) is highly used for preparation of superhydrophobic material in oil/water separation [ 31 , 32 ]. Chen et al. [ 33 ] coated the flax fiber with ZnO-HDTMS using a plasma-grafted poly (acrylic acid) as the binding agent for efficient oil/water separation. Zhang et al. [ 34 ] bonded ZnO nanoparticles to the surface of polyurethane/polysulfone with hydrophobically modification of HDTMS to prepare a new waterproof and moisture permeable membrane. In this paper, we report a facile, inexpensive, but versatile method for preparation of superhydrophobic surfaces with high durability. The samples were integrated with PPy and ZnO nanoparticles by in situ polymerization and deposition method. PPy/ZnO was loaded onto different substrates, which constituted an appropriate hierarchical structure. After further modification with HDTMS, the samples became superhydrophobic and superoleophilic, which could be used for oil/water mixtures separation. This method had the following apparent advantages: (1) There were no restrictions on the shape or type of the substrates. Fabric, sponge, copper mesh, forestry, and agricultural residues could all be applied without destroying the intrinsic appearance; (2) This method was economical and efficient because it was conducted at low temperature without special materials or equipment; (3) The superhydrophobic coating showed low-term durability and mechanical stability in oil/water separation; (4) This method could prepare superhydrophobic materials on a large scale. Apparently, a simple and feasible strategy has been put forward to fabricate superhydrophobic surfaces, which provides a feasible solution for oil/water separation.",
"discussion": "3. Results and Discussion 3.1. Surface Morphology and Wettability Micro/nanostructures’ creation and modification of low surface energy material are of importance to realize superhydrophobicity. The surface morphologies of raw and coated samples were investigated. Figure 1 a,d,g,j,m,p showed the typical images of original corn stalk, sawdust, cotton, fabric, sponge, and copper mesh. After the coating treatment, the surface of all samples became hierarchical and rough. From high magnifications, ZnO nucleated on the surface and grew in addition to precipitation or in lieu of precipitation. The irregular distribution of particles covered the surfaces of samples uniformly, which created micro/nanostructures required by superhydrophobicity. The PPy/ZnO/HDTMS coating was analyzed with the Cassie–Baxter equation [ 35 ]: (4) c o s θ c = f s ( c o s θ + 1 ) − 1 In this equation, θ c and θ are the apparent WCA of a water droplet on rough and smooth solid surface, respectively. f s is the apparent area fraction of the solid surface in contact with water. Thus, 1− f s is apparent area of trapped air contacting with water. Table 1 showed the calculated data with the Cassie–Baxter equation according to the wettability of PPy/ZnO/HDTMS coated samples. The WCA of original corn stalk, sawdust, cotton, fabric, sponge, and copper mesh were 0°, 0°, 0°, 0°, 90.9°, and 85.2°, respectively. The WCA of PPy/ZnO/HDTMS coated samples were 152.7°, 153.9°, 153.3°,157.6°, 163.1°, and 162.1°, respectively. Specifically, the fraction of air (1− f s ) in contact with water of all coated samples was above 0.88, illustrating that the air occupied more than 88% of the contact area when contacting with water droplets on the coated rough surface. Thus, the air preserved in hierarchical texture mainly supported the sphere-shaped water droplet. The photographs of water droplets on original and PPy/ZnO/HDTMS coated samples were shown in Figure 2 . All coated samples presented dark surface because of the first step of PPy modification. From Figure 2 , the water droplets could keep a spherical shape on surface of coated samples, which was consistent with the calculated data in Table 1 . Therefore, it is proved that the coated samples became superhydrophobic. The modification of ZnO nanoparticles and HDTMS are of importance to the endowment of superhydrophobicity on different substrates. 3.2. Formation Mechanism As shown in Figure 3 , the FTIR spectra were tested in order to ensure the chemical compositions of the prepared coating. For raw fabric, the strong absorption peak at 3424 cm −1 was attributed to -OH stretching vibration. The -OH stretching vibration was greatly reduced after coating treatment due to the modification of HDTMS. The FTIR spectrum for PPy [ 36 , 37 , 38 ] and HDTMS [ 39 ] is shown in Table 2 . The characteristic peaks at 1685, 1558, and 1315 cm −1 in coated fabric were assigned to the C = N, C = C, C–N in PPy. Two absorption peaks at 2851 and 2921 cm −1 stemmed from C–H stretching vibration of –CH 2 – and –CH 3 . The adsorption at 2918 cm −1 (C–H stretching vibration) in original fabric split into two peaks (2921 and 2851 cm −1 ) after coating, which was attributed to HDTMS modification. Furthermore, the stretching vibration of Si–C was observed at 781 cm −1 , demonstrating that the HDTMS was successfully grafted onto the PPy/ZnO coated fabric. The XRD diffraction patterns of raw and coated fabric were shown in Figure 4 . The raw fabric had typical diffraction peaks of cellulose at 2θ = 14.9°, 16.5°, and 22.9°, which also existed in coated fabric. The diffraction peaks of the coated fabric at 32.45°, 34.76°, 36.82°, and 47.65° corresponded to the (100), (002), (101), and (102) lattice planes of ZnO nanoparticles, proving that the ZnO nanoparticles were successfully loaded on fabric surface. ZnO nanoparticles played a significant role in constituting the hierarchical structure for realizing the superhydrophobic properties. The reaction steps of the PPy/ZnO/HDTMS coating are shown in Figure 5 a. PPy combined with the fabric in the first by in situ polymerization reaction. ZnO nanoparticles were deposited on PPy surface by precipitation method. From Figure 5 b, the hydrolysis of HDTMS generated –OH groups which replaced the –O–CH 3 groups. The –OH groups on ZnO nanoparticles or formed by hydrolysis of HDTMS were both reactive, which resulted in the condensation reaction between them. Furthermore, the –OH groups of hydrolyzed HDTMS reacted with each other, forming the Si–O–Si bond. Analogously, HDTMS could be chemically bonded with PPy through the condensation reaction between the –OH groups in hydrolyzed HDTMS and N–H bond in PPy to further reduce the surface energy. 3.3. Oil/Water Separation The PPy/ZnO/HDTMS coated fabric with superhydrophobic and superolephilic properties was used for oil/water mixtures separation. The separation process driven by gravity is shown in Figure 6 . Heavy oil like tetrachloromethane and light oil like petroleum ether were used for separation. The oil/water mixtures were poured from the top crossing the coated fabric which was fixed between two tubes. For tetrachloromethane/water separation, the separation device was placed with a vertical type ( Figure 6 a and Video S1 ). The water flowed into the tube first but could not permeate through the coated fabric for its superhydrophobicity. The tetrachloromethane, arriving late, passed through the water phase and penetrated into the coated fabric because of the higher density of tetrachloromethane and the superoleophilicity of the fabric, enabling the heavy oil/water mixture to be completely and successfully separated. The oil filtrate fraction and flux of the coated fabric for tetrachloromethane/water mixture were calculated to be 99.3% and 782.4 L m −2 h −1 , respectively. For petroleum ether/water separation, the separation device was placed with an inclined type. As shown in Figure 6 b and Video S2 , the petroleum ether encountered and passed through the coated fabric first, while the water could not submerge the coated fabric, which resulted in a complete separation of light oil/water mixture. With further evaluation and calculation, the separation and flux of the coated fabric for petroleum ether/water mixture were 98.4% and 426.7 L m −2 h −1 , respectively. Other mixtures could also be separated using the coated fabric efficiently, for example, dichloromethane/water mixture (99.4%, 855.6 L m −2 h −1 ), the diesel/water mixture (99.3%, 432.9 L m −2 h −1 ), and ethyl acetate/water mixture (98.6%, 598.4 L m −2 h −1 ). Moreover, the coated fabric also had an outstanding separation effect for water-in-oil emulsion, and the filtration procedure and effects were also investigated. As shown in Figure 6 c and Video S3 , the water-in-tetrachloromethane emulsion was slowly filtered owing to gravity. The coated fabric was wet with the oil phase which fell into the measuring beaker. Finally, the foggy cloudy emulsion became transparent and clarified. The micron water droplets in emulsion before filtration was evaluated to be 2–5 μm, which completely disappeared after separation ( Figure 6 d). The oil filtrate fraction and flux of the coated fabric for the water-in-tetrachloromethane emulsion were found to be 99.1% and 736.9 L m −2 h −1 . Except for the water-in-tetrachloromethane emulsion, other emulsions could also be separated efficiently, for example, the water-in-chloroform emulsions (99.2%, 853.6 L m −2 h −1 ), methylbenzene-in-water emulsion (99.1%, 822.4 L m −2 h −1 ). In conclusion, the PPy/ZnO/HDTMS coated fabric exhibited an outstanding separating performance towards different water/oil mixtures, which could satisfy different needs in practical usage, showing its great potential. The oil absorption capacity of the coated fabric was analyzed, which is shown in Table 3 . The oil absorption capacity towards different oils differed for the physical and chemical properties of oils. The original fabric had absorption capacity for ethyl acetate of 4.83 g/g, while the oil absorption capacity of the coated fabric was 9.61 g/g, which was almost two times that of the original fabric. The as-prepared fabric could absorb oils up to many times its own mass. The fabric after coating treatment could absorb more oil. The roughness provided by PPy/ZnO and the low surface obtained by HDTMS obviously enhanced the absorption capacity of fabric for oils. 3.4. Durability Evaluation Corrosive solutions, such as salt, acidic, basic, or hot solutions, are the existing challenges for practical application of superhydrophobic material in oil/water separation. The wettability of the coated fabric toward corrosive solutions was examined and shown in Figure 7 a. The WCA of hot water (100 °C), 1M of HCl solution, 1M of NaOH solution, and 1M of NaCl solution were all above 150°, showing excellent superhydrophobic property of the coated fabric. The separation capacity of the coated fabric for oil/water separation under harsh conditions was subsequently evaluated. The mixtures of tetrachloromethane and various corrosive solutions were used for evaluation. The oil phase passed through the coated fabric quickly, leaving the corrosive solutions above the coated fabric. As shown in Figure 7 b, the separation of these tetrachloromethane/corrosive solution mixtures was all completely achieved with the oil filtrate fraction above 98%. Thus, the coated fabric could realize highly efficient separation even in harsh environments, indicating its excellent capability in oil/water separation. The superhydrophobic fabric also showed an outstanding durability. As shown in Figure 8 a, the reusability of the coated fabric was evaluated by duplicating the oil/water separation for 80 times. The oil filtrate fraction of tetrachloromethane/water mixture after 80 separation cycles remained above 98%. The surface morphology of the coated fabric after being reused for 80 cycles is depicted in Figure 8 b, and no changes were observed with this fabric, proving excellent reusability of the superhydrophobic fabric for oil/water separation. Figure 8 d showed the abrasion resistance of the coated fabric. The superhydrophobic fabric was placed on a sandpaper with 1000 meshes, which was pulled forth and back under a weight of 100 g for a distance of 10 cm. The WCA of the coated fabric after 80 abrasion cycles was still above 150°, demonstrating outstanding mechanical robustness. Figure 8 c showed the surface morphology of the coated fabric after 80 abrasion cycles. The surface was still rough and hierarchical, which was critical for keeping superhydrophobic properties. The firm modification of the superhydrophobic coating was responsible for the respectable durability of the coating. The storage stability test of the coated fabric in air was conducted. As depicted in Figure 8 e, the coated fabric showed a stable superhydrophobicity with the WCA higher than 155° after being exposed in air for 12 months. All these characterizations demonstrated durable stability of the coated fabric for oil/water separation. 3.5. Comparison of the As-Prepared Fabric with Other Reported Superhydrophobic Fabrics A comparison of the as-prepared fabric with other reported superhydrophobic fabrics in preparation and separation was made to demonstrate the significance of this work ( Table 4 ). The other methods for fabrication of superhydrophobic coating always require a time-consuming and tedious process. The method proposed in this work showed obvious advantages of low cost and easy to operate. Additionally, both heavy and light oils could be separated from oil/water mixtures. As for separation under harsh environments, the as-prepared fabric could maintain its superhydrophobicity and oil filtrate fraction under specific harsh conditions, demonstrating outstanding performance for highly efficient separation in harsh environments. Generally, the as-prepared fabric showed superior property compared with other superhydrophobic fabrics."
} | 4,870 |
36432230 | PMC9697845 | pmc | 1,130 | {
"abstract": "The development of novel materials with coexisting volatile threshold and non-volatile memristive switching is crucial for neuromorphic applications. Hence, the aim of this work was to investigate the memristive properties of oxides in a Hf–Nb thin-film combinatorial system deposited by sputtering on Si substrates. The active layer was grown anodically on each Hf–Nb alloy from the library, whereas Pt electrodes were deposited as the top electrodes. The devices grown on Hf-45 at.% Nb alloys showed improved memristive performances reaching resistive state ratios up to a few orders of magnitude and achieving multi-level switching behavior while consuming low power in comparison with memristors grown on pure metals. The coexistence of threshold and resistive switching is dependent upon the current compliance regime applied during memristive studies. Such behaviors were explained by the structure of the mixed oxides investigated by TEM and XPS. The mixed oxides, with HfO 2 crystallites embedded in quasi amorphous and stoichiometrically non-uniform Nb oxide regions, were found to be favorable for the formation of conductive filaments as a necessary step toward memristive behavior. Finally, metal–insulator–metal structures grown on the respective alloys can be considered as relevant candidates for the future fabrication of anodic high-density in-memory computing systems for neuromorphic applications.",
"conclusion": "4. Conclusions Following the screening results of the anodic memristors grown on a Hf–Nb combinatorial library, the coexistence of threshold and non-volatile memristive switching was observed for metal–insulator–metal structures grown on Hf-45 at.% Nb and Hf-50 at.% Nb alloys. At the same time, improved performances regarding resistive state ratios, low voltage switching ranges, or multi-level switching characteristics, were demonstrated for the respective MIMs. In addition, the memristive switching in the bipolar or unipolar mode was identified to be dependent upon the current compliance applied during the I–U switching. Unipolar switching characteristics were found at a higher current compliance range, while bipolar ones were evidenced at a lower current compliance range. This was explained by Joule heating typically released for the higher current compliance range. The mixed switching behavior for the MIMs grown on the Hf-45 at.% Nb and Hf-50 at.% Nb alloys was also justified by HfO 2 crystallites embedded in the mixed oxides, facilitating the formation of CFs and was related to the non-uniformity of the Nb oxides’ stoichiometry. Finally, a regular trend in resistive switching could not be recognized by the high-throughput screening of the anodized Hf–Nb system. Nevertheless, the screening outcome indicated that MIMs grown on Hf-45 at.% Nb and Hf-50 at.% Nb alloys may be considered for the fabrication of high-density in-memory computing systems for neuromorphic applications.",
"introduction": "1. Introduction The development of computing and information technology is progressing rapidly considering that an extensive amount of daily generated data has to be processed and stored [ 1 , 2 ]. Nowadays, computing systems rely on von Neumann computing architectures with separated processing and memory units [ 1 , 3 ]. Evidently, this can justify the fact that conventional memory technology has already reached its scaling and processing speed limits [ 4 ]. Memristors are foreseen as the most promising candidates for the next generation of non-volatile memories due to their simple structure, high switching speed, scalability, and low power consumption [ 2 , 5 ]. The selection of bottom and top electrodes plays a crucial role in resistive switching based on the formation of conductive filaments (CFs) inside of an active layer, this being the basis of the metal-insulator–metal (MIM) structure [ 6 , 7 ]. Memristors based on transitional metals have shown reliable volatile threshold and non-volatile switching characteristics [ 8 , 9 , 10 ]. While the non-volatile switching mechanism is dominated by CF formation due to the movement of O and electrolyte species inside of the oxide, threshold switching relies on the thermal formation/dissolution of metallic CFs [ 11 , 12 ]. The diffusive dynamics of metallic species mimic the dynamics of biological synapses in terms of synaptic plasticity regulated by Ca 2+ dynamics. An increase or decay of the current response by applying threshold voltage values is similar to the decay of the ion concentration while releasing neurotransmitters in a neuron [ 13 ]. This emphasizes the relevance of threshold switching for a neuromorphic application but also as a selector device due to I – V nonlinearity [ 5 , 14 , 15 , 16 , 17 , 18 ] Synaptic weight and storage possibilities can be obtained by the implementation of non-volatile multi-level resistive switching [ 11 , 12 , 13 ]. Clearly, devices with co-existing memristive and threshold switching characteristics are desirable for the relevant applications of memristive devices [ 8 , 19 , 20 , 21 , 22 , 23 ]. Considering that anodic memristors have shown promising resistive switching characteristics, the further development and optimization of such devices are highly relevant [ 24 , 25 , 26 ]. Hence, the non-volatile properties of Hf memristors and threshold switching behavior, observed in Nb-based devices, were the motivation for the current study [ 27 , 28 ]. Additionally, it has already been observed that NbOx-based devices are forming-free since the stability of CFs can be affected by applying an electric field controlling the alignment of O species in the oxide [ 29 , 30 ]. Both materials, Hf and Nb, were mixed to produce bottom electrodes based on their alloys. It has been already proven that alloy-based memristors have the potential for the large-scale implementation of memristors integrated into crossbar arrays due to their stable and controllable performance. This benefits programming capabilities while consequently reducing the cost of device fabrication and broadening the range of memristive materials for neuromorphic applications [ 31 , 32 ]. Moreover, the selection of the active memory layer is crucial considering that it is responsible for the switching characteristics and device performance. It should be also noted that the possible future applications of these devices have to be considered when selecting the materials for an active layer [ 32 ]. Dynamic synaptic characteristics have been demonstrated for transition metal oxide (TMO) devices, including those based on HfO 2 , with these devices consuming low power and switching at numerous resistive levels [ 33 ]. Oxide-based synapses have also been used as 3D vertical-structured parallel devices meeting low-cost criteria and a high integration density [ 34 ]. Hence, this work focused on mixed oxides grown anodically on Hf–Nb alloys as active layers. Electrochemical oxidation is not only an advantageous oxide fabrication method due to its simplicity and low cost but also due to the possibility of already forming in situ CFs during oxide growth. In this way, forming-free memristors can be fabricated and defect engineered by the simple tuning of the electrochemical parameters [ 35 ]. This agrees with the described requirements for synaptic device optimization, while allowing controllable resistive switching in the memory layer grown on alloys [ 32 ]. Finally, the properties of anodic memristors grown on a Hf–Nb library were investigated for the first time in the current study. This is of high relevance for the development of neuromorphic systems applied for neuromorphic vision, sensors, wearable electronics, or similar [ 36 , 37 , 38 ]."
} | 1,920 |
35529743 | PMC9073238 | pmc | 1,131 | {
"abstract": "In the recovery of rare earth elements (REE) microbial biosorption has shown its theoretical ability as an extremely economically and environmentally friendly production method in the last few years. To evaluate the ability of two cyanobacterial strains, namely Anabaena spec. and Anabaena cylindrica to enrich dissolved trivalent REE, a simple protocol was followed. The REE tested in this study include some of the most prominent representatives, such as europium (Eu), samarium (Sm) and neodymium (Nd). Within the experiments, a fast decrease of the REE 3+ concentration in solution was tracked by inductively coupled plasma mass spectrometry (ICP-MS). It revealed an almost complete (>99%) biosorption of REE 3+ within the first hour after the addition of metal salts. REE 3+ uptake by biomass was checked using laser-induced breakdown spectroscopy (LIBS) and showed that all three selected REE 3+ species were enriched in the cyanobacterial biomass and the process is assigned to a biosorption process. Although the biomass stayed alive during the experiments, up to that, a distinction whether the REE 3+ was intra- or extracellularly sorbed was not possible, since biosorption is a metabolism independent process which occurs on living as well as non-living biomass. For europium it was shown by TEM that electron dense particles, presumably europium particles with particle sizes of about 15 nm, are located inside the vegetative cyanobacterial cells. This gave clear evidence that Eu 3+ was actively sorbed by living cyanobacteria. Eu 3+ biosorption by cell wall precipitation due to interaction with extracellular polysaccharides (EPS) could therefore be excluded. Finally, with XRD analysis it was shown that the detected europium particles had an amorphous instead of a crystalline structure. Herein, we present a fast biosorptive enrichment of the rare earth elements europium, samarium and neodymium by Anabaena spec. and Anabaena cylindrica and for the first time the subsequent formation of intracellular europium particles by Anabaena spec.",
"conclusion": "Conclusion The two cyanobacterial strains Anabaena spec. and Anabaena cylindrica both are capable to actively biosorb the rare earth elements europium, samarium, and neodymium from highly diluted (μM range) aqueous solutions. The uptake by biomass proceeds very fast and almost complete. Within the first hour after REE addition more than 99% of the initial REE concentration is biosorbed by the cyanobacteria. Moreover, it was shown that Anabaena spec. accumulates and incorporates europium particles. These particles were detected by TEM only inside the heterocysts, neither outside the heterocysts nor in or outside the vegetative cells. TEM revealed maximum particle sizes of about 15 nm. The particles are evenly spread over the interior heterocyst and show no preferred localization ( e.g. at the thylakoid membranes) within the cells. With XRD analysis it was shown that the europium particles had a non-crystallite structure. Since the samples exhibited no Bragg reflections, no determination of the particles' average size via XRD was possible. However, Anabaena spec. and Anabaena cylindrica have been proven to be efficient and environmental friendly biosorbents for the enrichment of rare earth elements, europium, samarium and neodymium from diluted aqueous solutions. Nevertheless, further research is needed in the area of “harvesting” the accumulated REE particles and a subsequent sustainable recycling option. On the other hand, possible implementations of cyanobacteria-supported REE-collecting steps in municipal wastewater treatment plants or recovery stations in running waters and rivers are future perspectives.",
"introduction": "Introduction In recent years, the use of rare earth elements (REE) has tremendously increased and is thus accompanied by a rising worldwide demand. REE are spent in various modern electronic high tech devices like mobile phones, computers, LCDs and screens as well as for the emerging green technologies, e . g. in wind energy converters, electric cars, energy saving lamps and catalytic converters. Besides, also medicinal applications are known e . g. gadolinium compounds as contrast agent for magnetic resonance imaging and REE as crop fertilizers, which are widely used in agriculture. As a consequence, REE are heavily exposed from these anthropogenic products to the environment in metallic or even ionic form thus finding an entry to rivers and waste waters. As an example, Kulaksız and Bau 1,2 recently showed that the River Rhine carries considerable amounts of the REE lanthanum, samarium, and gadolinium. As a source, industrial effluents were identified. This means that tons of these REE are likely to reach the North Sea unexploited every year. The increasing demand for REE thus provides impetus to the development of efficient and environment-friendly recovery methods. Since REE usually are present in small concentrations in rivers and in waste waters, methods especially for separating and accumulating these valuable metals from very diluted solutions are necessary. Actually, several procedures like precipitation, ion exchange, electrochemical methods, reverse osmosis and adsorber resins are applied. 3,4 However, high process costs, environmental impact due to toxic resins or inefficient recovery of highly diluted REE are disadvantageous properties which come along with these common procedures. 3,5,6 In contrast, the use of (micro-)biological methods gains more and more importance, since they provide alternative, efficient and environment-friendly methods of resource recovery. 3,5–7 Especially, the uptake of metal ions by microbial biomass (biosorption) is already known, with several bacteria strains even being able to biosorb rare earth metals from diluted solutions. In the area of biological waste water treatment biosorption is already discussed as an efficient process for the bio-removal of heavy metals, e . g. lead and cadmium. 8 Especially, the recovery of REE metal ions from aqueous solutions with the help of microorganisms is in the focus of current research. For a number of organisms such as bacteria, yeasts and algae this ability was already shown. 6,9 Adsorption of REE onto the cell walls of Gram-positive bacteria like Bacillus subtilis , Gram-negative bacteria like Escherichia coli , Saccharomyces cerevisiae , several brown algae ( sargassum ) and many more has been reported. 3,10–14 Among the various microorganisms studied, also cyanobacterial strains proved to be highly capable to biosorb and accumulate dissolved (heavy) metals as well as REE. Kim et al. 15 used Phormidium , a genus of filamentous cyanobacteria, to adsorb various REE (La, Pr, Nd, Sm, Gd, Dy) from an ore leachate. With the dead biomass used in this study they found a fast biosorption which occurred for a short time at an early stage. The interaction of micromolar solutions of [UO 2 (CO 3 ) 2 ] 2− with the cyanobacterial strain Anabaena torulosa was investigated by Acharya et al. 16 They showed the importance of cell viability for optimal uranyl binding. Heat killed cells or extracts of the extracellular polysaccharides exhibited only limited binding of uranyl. Adsorption of ionic species is mainly attributed to the binding to functional groups on the outer cell wall or more precisely the exocellular polysaccharidic layers (sheath, capsule and mucilage). Especially, exopolysaccharide (EPS) producing cyanobacteria are very efficient in binding heavy metal ions, thus being promising candidates for the removal of positively charged metal ions from aqueous solutions. 17 Okajima et al. 18,19 for instance, extracted the polysaccharide sacran from cyanobacterium Aphanothece sacrum . They investigated the sorption efficiency of sacran-containing hydrogels towards Nd 3+ and observed that the gels sorbed excessive amounts of Nd 3+ , in addition being more efficient than conventional alginate-containing gels. Besides, bioaccumulation, i.e. the active intracellular accumulation of the ionic metal species may occur in cyanobacteria. As this is a plasma membrane mediated transport of metal ions into cellular compartments, 8 it requires living cells for the intracellular enrichment and is often associated with a defense mechanism of the cell towards a toxic metal. 20 The “detoxification” of the mobile metal species may than proceed by chemical conversion in the cytoplasma via transformation of the oxidation state, precipitation of metal ions by cell contents (biomacromolecules), enzymatically driven redox reactions or a combination of them. Especially cyanobacteria are known to produce metal-binding proteins in the cytoplasma, e.g. cysteine-rich metallothioneins, in response to the presence of metal ions, which sequester them in biologically inactive, non-cell-toxic forms. 21 Although the interaction of metallothioneins with heavy metals such as cadmium, zinc and copper are widely described in the literature, 22–25 these proteins are also capable of binding trivalent REE metal ions such as La 3+ . 26 Interestingly, the interaction of metal ions with metallothioneins can furthermore lead to a nanocluster formation as described for in vitro experiments with gold by Mercogliano and DeRosier. 27 Some cyanobacteria are able to reduce the biosorbed metal ions to their zero-valent state (bioreduction) by enzymatic redox reactions and finally to accumulate them intra- or extracellular as crystalline nanoparticles in vivo . This is frequently shown for noble metals like gold, silver, palladium or platinum. Brayner et al. 28 examined three different cyanobacteria strains, namely Anabaena flos-aquae , Calothrix pulvinata , and Leptolyngbya foveolarum , with respect to their ability to biosynthesize Au-, Ag-, Pd-, and Pt-nanoparticles. All three strains studied, succeeded in the extracellular formation of gold- and silver-nanoparticles, respectively, with final particle sizes from 5–12 and 15–40 nm. The size as well as the amount of nanoparticles formed was found to be dependent on the strain used. Due to a high number of heterocysts, Anabaena flos-aquae has shown a high concentration of the enzyme nitrogenase and produced therefore in a few minutes well-defined nanoparticles. Dahoumane et al. 29 stated a high gold adsorption rate for Anabaena flos-aquae that however comes along with a wide size distribution of nanoparticles produced as well as a fast cell death. In recent studies Anabaena spec. 30 and Anabaena cylindrica 31 were used to study the time-dependent growth of crystalline Au 0 nanoparticles from diluted Au 3+ solutions. In contrast to Anabaena flos-aquae , both Anabaena spec. and Anabaena cylindrica , do not produce any toxic anatoxin-a, 32 which is a major advantage featuring environmentally friendly metal bio-recovery as well as nanoparticle biosynthesis. Gold nanoparticles with average sizes of 9 and 10 nm, respectively, in both species mainly located at the thylakoid membranes of the vegetative cells rather than in heterocysts were already found after two hours and four hours, respectively. One of them even shows selective biosorption and biosynthesis of gold nanoparticles from multimetal solutions. 32 In contrast to the above mentioned nanoparticle formation of noble metals, a conversion of mobile REE (ions) into insoluble forms, in the manner of REE particles or REE nanoparticles by living cyanobacteria was not presented yet. Up to now, only REE biosorption, as described above, is known for some cyanobacteria. Herein, we present a fast biosorptive enrichment of the rare earth elements europium, samarium and neodymium by Anabaena spec. and Anabaena cylindrica and for the first time the subsequent formation of intracellular europium particles by Anabaena spec.",
"discussion": "Discussion This study adds valuable results on the ongoing work in the research field on Anabaena spec. and Anabaena cylindrica and their biosorptive capabilities towards the rare earth elements Eu, Sm and Nd. The cyanobacteria strain Anabaena spec. is able to biosorb in vivo the trivalent REE Eu 3+ from aqueous solutions and to form intracellularly europium particles. Similar features for Anabaena spec. as well as Anabaena cylindrica and the addition of gold have recently been reported by some of the authors. 30,31 Furthermore, Anabaena cylindrica was found to be highly selective towards the formation of gold nanoparticles in the presence of rhodium and iridium. 32 Although, these and further studies reported on the subsequent nanoparticle formation of precious metals also using other cyanobacterial strains ( e.g. Anabaena flos-aquae 28,42 ), the formation of REE particles, however, by living cyanobacteria has never been reported yet. The second cyanobacterial strain used within this study, Anabaena cylindrica , is capable to efficiently extract the rare earth elements samarium and neodymium from (highly) diluted aqueous solution, in the meaning of a biosorption process. Hence, this partial outcome is in line with various other studies, also describing solely the biosorption capability of biomass towards diluted REE (see introduction). In contrast to Anabaena flos-aquae , the presently used strains Anabaena spec. (SAG 12.82) and Anabaena cylindrica (SAG 1403.2) do not produce and release anatoxin-a, which is a major advantage. 43 Furthermore, these two cyanobacteria representatives are undemanding organisms and they are easy to handle in terms of cultivation as well as biosorption and possibly bioreduction. They can even withstand to some extend malnutrition for several days and weeks, whilst still being able to biosorb and intracellularly retain REE. The biosorption process, i.e. the decrease of REE 3+ concentration in the aqueous media due to the uptake by biomass, was tracked by ICP-MS and LIBS. From ICP-MS analysis one can estimate that upon exposure to a REE 3+ solution biosorption starts immediately. Within the first hour of incubation almost the entire (>99%) REE 3+ amount is removed from the media. LIBS was used to prove the biomass as the sorbent of the REE. Especially for the europium series, ICP-MS analysis shows that biosorption starts immediately upon exposure to an overall concentration of 0.111 mM of Eu 3+ . After the first hour of incubation only 0.6% of the initial Eu 3+ concentration was still detectable in the supernatant ( i.e. only 76 ppb left). This, in turn, means that within this time almost the entire Eu 3+ amount ( i.e. up to 99.4%) is removed from the media. Experiments with concentrations approximately twice as high ( i.e. 0.191 mM Eu 3+ ) were tolerated by the biomass as well and gave similar results. A reduction of Eu 3+ concentration by about 75% after 40 minutes was determined and an almost total uptake of dissolved europium ions was determined after 3 h. These results are quite impressive, since these experiments in this higher concentration range were performed in the absence of any nutrient medium. However, biosorption is more effective when experiments are performed in culture media rather than in a starving culture, proving biosorption is an active in vivo process. The results imply for both experiments a very fast biosorption process for Anabaena spec. within the low Eu 3+ concentration range observed in this study. This corroborates the general idea of describing biosorption as a process with rapid kinetics. 4,23 Furthermore, the ICP-MS measurements confirm that no significant release of europium occurs once it is biosorbed. The accumulation of europium in the biomass was qualitatively proven by LIBS. Altogether six characteristic peaks in the LIBS spectra of the europium incubated biomass sample could uniquely be assigned to europium (compare NIST data base). Additionally, these signals are in accordance with LIBS spectra obtained from the europium stock solution used for the experiments (See Fig. 3 ). However, LIBS experiments performed on Eu 3+ incubated biomass samples confirmed the deduction from ICP-MS analysis: the vast majority of biosorption proceeds within the first hour after europium addition. A potential temporal development meaning an increase of Eu concentration in the biomass during the experiment cannot be deduced from LIBS data. Since the signal intensities are not necessarily connected to a concentration value, it is not possible to conclude any (absolute or relative) concentration value of the analyte of interest. A quantification of the biosorbed europium with LIBS was not possible at this point and is generally still challenging up today for this kind of experiments. Characterization of the Eu deposits was done ex situ by TEM and XRD. With TEM electron dense particles were localized only inside the cells of Anabaena spec., in particular only inside their heterocysts. Furthermore, these particles were statistically distributed inside the entire heterocyst instead of being preferentially located at the thylakoids. These results are somehow surprising and in sharp contrast to our previous findings, 30,31 where the thylakoid membranes of the vegetative cells were preferred deposits of biosynthesized gold nanoparticles. They even differ from the results found by Rochert et al. 32 who identified the HEP as the preferred region for gold nanoparticle formation from multimetall solutions. Recently, Dahoumane et al. 44 also demonstrated the key role of the thylakoid membranes during the biosynthesis of noble metal nanoparticles. With the help of TEM they confirmed on the favored localization (the so-called place of birth) of the first produced gold nanoparticles within the thylakoids of a micro-algae. From TEM micrographs the maximum size of the europium particles was determined to be approximately 15 nm for all the times observed in our study. Needless to say, also smaller particles were detected, which leads to a broad size distribution throughout the whole experiment, similar to a potential law than a Gaussian distribution. As the maximum size is not increasing with time, the final maximum size of the biosynthesized particles must have been achieved before the first sampling after 10 h. It is likely that the nucleation starts immediately after incubation and biosorption and moreover might be finished in the first few hours since all ions were removed from supernatant within one hour. Besides the size also the shape of the europium particles was characterized by TEM. The images revealed a more or less irregular shape of the particles with a diffuse, scarcely defined rim (see Fig. 8B and E ). This leads to the assumption of the presence of agglomerated smaller particles and to the result which was finally found by XRD. With XRD analysis it was shown that the detected europium particles had an amorphous instead of a crystalline structure. Since the samples exhibited no Bragg reflections, no calculation of the particles' average size using the Scherrer-equation was possible. Therefore, only a roughly estimated maximum size from TEM images instead of a precisely calculated average particle size from XRD was obtained. Although we had no device to analyze the oxidation state of the europium ( e.g. by XPS, EDX), to reveal whether if it was reduced to zero or still has the initial oxidation state “III” or changed else, we assume the particles inside the heterocysts to be amorphous agglomerated europium particles. Via interactions of biomacromolecules present inside the HC an agglomeration of Eu 3+ and finally stabilization of amorphous particles might proceed. Microbial products, e . g. metallothionein-like proteins are known to bind and sequester rare earth metals 45 and are also found in cyanobacteria. 22,25 Mercogliano and DeRosier 27 showed that in vitro synthesized metallothionein-gold nanoclusters are visible in TEM and reveal varying sized clusters. They stated more than 20 gold atoms to be bound to a single metallothionein molecule. Hence, it is likewise to assume a strong complexation of the biosorbed Eu 3+ for the cyanobacterial strains used here. Cheng et al. 46 recently demonstrated the adsorption and subsequent mineralization of lanthanum by Bacillus licheniformis , a Gram-positive bacterium commonly found in the soil. Conversion of amorphous La( iii ) adsorbed on cell surface phosphate groups to precipitated crystalline nanomineral monazite(La) (LaPO 4 ) occurred within 30 days. Prior to this conversion no crystalline phases were detected with XRD. Likewise, biomineralization of Sm( iii ) into monazite(Sm) on the outer cell-surface of Saccharomyces cerevisae and Pseudomonas fluorescens was shown by Jiang et al. 47 In case of an ionic origin of the europium particles, means a salt-like precipitate ( e.g. europiumhydroxide) inside the heterocysts, one would at least expect an evidence of large crystalline structures from XRD, what, however, was not confirmed in our study. In fact, no crystalline structure of the TEM detected europium particles can be approved. Due to the non-crystallinity and the missing information in terms of the oxidation state of the europium particles, we consciously avoid calling them nanoparticles in this context. In our understanding, at least crystallinity of the metal particles is a prerequisite feature for this designation. Though the experiments with Anabaena cylindrica revealed, that the biosorption of the REE samarium and neodymium occurs, a particle formation could not be proven within these series. With ICP-MS it was clearly shown that samarium is removed by up to 99% within the first 20 minutes after the addition of an overall concentration of 0.139 mM Sm 3+ . After one hour 99.5% and moreover after 12 h more than 99.9% of the added samarium was removed from the media. A renewed increase of the samarium concentration, due to a possible re-release, was not detected, much like in the europium series. Similar results were obtained for the biosorption of neodymium. In comparable timescales almost the same amount of neodymium was sorbed. The LIBS measurements gave in addition a clear evidence for the biosorption capabilities of Anabaena cylindrica towards Sm 3+ and Nd 3+ . In the observed timescale of up to 200 h, samarium and neodymium were unambiguously detected in each of the respective samples. Although no information about the REE concentration can be gathered from the LIBS data, one result can be stated without a doubt: samarium and neodymium are efficiently recovered from highly diluted aqueous solutions by the biomass. In combination with the ICP-MS measurements which reveal no re-release of already sorbed species also the cyanobacteria strain Anabaena cylindrica can be referred to as an efficient and reliable biosorbent towards REE, especially for Sm 3+ and Nd 3+ . Within this study, we obtained in a first instance comparable results for both, Anabaena spec. and Anabaena cylindrica concerning their ability to biosorb the dissolved trivalent REE species, i.e. Eu 3+ , Sm 3+ and Nd 3+ . The biosorption process proceeds in all cases quite fast within a few minutes after metal salt addition. Impressively, within such a short time, almost the entire REE 3+ amount was removed from the surrounding media by the biomass. The exact mechanism responsible for the REE entrapment is not clear, and not the subject of interest here either. Nevertheless, one can state, that there has to be a quite effective metal binding towards these REE, since no leaching and re-release of these species was detected even after more than several weeks. Moreover, it was shown that at least the accumulation of europium must proceed via an active, metabolism-dependent process, since the intracellular uptake and the internalization of metal, which was demonstrated by TEM, requires microbial activity 4,48 in the short time span within this experiment. The observed intracellular uptake of europium and formation of particles is highly interesting and somehow surprising, since biosorption is commonly attributed to proceed on the outer cell wall and particularly via the extracellular polymeric substances. In case of the cyanobacteria and especially their heterocysts the internalization at least for europium has to be a privileged route over cell wall adsorption, since europium particles are solely detected inside the HCs but never seen on the outside. As already mentioned, we assume the particles to be biomolecule-complexed europium species in their initial ionic state. This is based on many respects: First, since a rather random distribution all over the interior of the HC instead of a favored localization next to the thylakoids (as expected from our previous study and other studies, which call them nanoparticles' place of birth 44 ) is observed, we cannot claim a possible bioreduction leading to Eu 0 as well as a subsequent formation of europium nanoparticles. Second, the absence of characteristic reflections in XRD, provides convincing evidence for an amorphous origin of the observed europium particles. The presence of crystalline europium nanoparticles is therefore excluded. However, due to the detection by TEM, the particles (have to) exhibit a much higher electron density, which means a lot of heavy metal atoms, in comparison to the surrounding biomass. Besides the already excluded possible presence of metallic nanoparticles, a high electron density may also originate from an agglomeration and therefore densification of a high amount of europium by biomacromolecules, most likely proteins, present in the cells. Caused by the inherent size and structure of these organic chelators, the assumption may also explain the irregular shape and diffuse rim of the particles visualized by TEM. Moreover, also the random distribution of the particles inside the HC justifies this agglomeration theory, since the macromolecules are freely available inside the cytoplasma. Furthermore, due to the high redox potential generally known for (trivalent) REE species ( e.g. Eu 3+ → Eu 0 ( E 0 = −1.991 V), Sm 3+ → Sm 0 ( E 0 = −2.304 V) and Nd 3+ → Nd 0 ( E 0 = −2.323 V)) 49 a bioreduction in the sense of an enzyme mediated electron transfer and therefore the formation of zero-valent REE species is hardly reasonable. At least for the previously shown gold nanoparticle formation by these two cyanobacteria 30 a differing behavior was not experienced. If one assumes the two strains to show the same behavior towards the REE, the resulting differences in particle formation have to be caused by the single REE itself or its particular and individual interaction inside the cell. The three REE species are all of trivalent state and originate all from the corresponding REE nitrate hexahydrate compounds REE(NO 3 ) 3 ·6H 2 O. Differing transport processes inside the cell, various ionic radii of the REE 3+ as well as their hydrated equivalents and slightly differing redox potentials (see above) may be possible explanations but lack for further investigation."
} | 6,786 |
23332119 | null | s2 | 1,133 | {
"abstract": "Many bacterial and archaeal lineages have a history of extensive and ongoing horizontal gene transfer and loss, as evidenced by the large differences in genome content even among otherwise closely related isolates. How ecologically cohesive populations might evolve and be maintained under such conditions of rapid gene turnover has remained controversial. Here we synthesize recent literature demonstrating the importance of habitat and niche in structuring horizontal gene transfer. This leads to a model of ecological speciation via gradual genetic isolation triggered by differential habitat-association of nascent populations. Further, we hypothesize that subpopulations can evolve through local gene-exchange networks by tapping into a gene pool that is adaptive towards local, continuously changing organismic interactions and is, to a large degree, responsible for the observed rapid gene turnover. Overall, these insights help to explain how bacteria and archaea form populations that display both ecological cohesion and high genomic diversity."
} | 263 |
36636576 | PMC9831110 | pmc | 1,135 | {
"abstract": "Event-based dynamic vision sensors provide very sparse output in the form of spikes, which makes them suitable for low-power applications. Convolutional spiking neural networks model such event-based data and develop their full energy-saving potential when deployed on asynchronous neuromorphic hardware. Event-based vision being a nascent field, the sensitivity of spiking neural networks to potentially malicious adversarial attacks has received little attention so far. We show how white-box adversarial attack algorithms can be adapted to the discrete and sparse nature of event-based visual data, and demonstrate smaller perturbation magnitudes at higher success rates than the current state-of-the-art algorithms. For the first time, we also verify the effectiveness of these perturbations directly on neuromorphic hardware. Finally, we discuss the properties of the resulting perturbations, the effect of adversarial training as a defense strategy, and future directions.",
"introduction": "1. Introduction Compared to the neural networks commonly used in deep learning, Spiking Neural Network resemble the animal brain more closely in at least two main aspects: the way their neurons communicate through spikes, and their dynamics, which evolve in continuous time. Aside from offering more biologically plausible neuron models for computational neuroscience, research in the applications of Spiking Neural Network is currently blooming because of the rise of neuromorphic technology. Neuromorphic hardware is directly compatible with Spiking Neural Network and enables the design of low-power models for use in battery-operated, always-on devices. Adversarial examples are an “intriguing property of neural networks” (Szegedy et al., 2013 ) by which the network is easily fooled into misclassifying an input which has been altered in an almost imperceptible way by the attacker. This property is usually undesirable in applications: it was proven, for example, that an adversarial attack may pose a threat to self-driving cars (Eykholt et al., 2018 ). Because of their relevance to real-world applications, a large amount of work has been published on this subject, typically following a pattern where new attacks are discovered, followed by new defense strategies, in turn followed by proof of other strategies that can still break through them (see Akhtar and Mian, 2018 for a review). With the advent of real-world applications of spiking networks in neuromorphic devices, it is essential to make sure they work securely and reliably in a variety of contexts. In particular, there is a significant need for research on the possibility of adversarial attacks on neuromorphic hardware used for computer vision tasks. In this paper, we make an attempt at modifying event-based data, by adding and removing events, to generate adversarial examples that fool a spiking network deployed on a convolutional neuromorphic chip. This offers important insight into the reliability and security of neuromorphic vision devices, with important implications for commercial applications. 1.1. What is event-based sensing? Event-based Dynamic Vision Sensor share characteristics with the mammalian retina and have several advantages over conventional, frame-based cameras: Camera output in the form of events and thus power consumption are directly driven by changes in the visual scene, omitting output completely in the case of a static scene. Pixels fire independently of each other which results in a stream of events at microsecond resolution instead of frames at fixed intervals. This enables very low latency and high dynamic range. The sparse, asynchronous Dynamic Vision Sensor output does not suit current high-throughput, synchronous accelerators such as GPUs. To process event-based data efficiently, neuromorphic hardware is being developed, where neurons are only updated whenever they receive an event. Spiking neuromorphic systems include large-scale simulation of neuronal networks for neuroscience research (Furber et al., 2012 ) and low-power real-world deployments of machine learning algorithms. Spiking Convolutional Neural Network as well as conventional Convolutional Neural Network have been run on neuromorphic chips such as IBM's TrueNorth and HERMES (Esser et al., 2016 ; Khaddam-Aljameh et al., 2021 ), Intel's Loihi (Davies et al., 2018 ) and SynSense's Speck and Dynap-CNN (Liu et al., 2019 ) for low-power inference. The full pipeline of event-based sensors, stateful spiking neural networks, and asynchronous hardware—which is present in SynSense's Speck—allows for large gains in power efficiency compared to conventional systems. 1.2. Adversarial attacks on discrete data The history of attack strategies against various kinds of machine learning algorithms pre-dates the advent of deep learning (Biggio and Roli, 2018 ), but the phenomenon received widespread interest when adversarial examples were first found for deep convolutional networks (Szegedy et al., 2013 ). In general, given a neural network classifier C and an input x which is correctly classified, finding an adversarial perturbation means finding the smallest δ such that C ( x +δ)≠ C ( x ). Here, “smallest” refers to minimizing ||δ||, where the norm is chosen arbitrarily depending on the requirements of the experiment. For example, using the L ∞ norm (maximum norm) will generally make the perturbation less noticeable to a human eye, while the use of the L 1 norm will encourage sparsity, i.e., a smaller number of perturbed pixels. There are two main challenges in transferring existing adversarial algorithms to event-based vision: The presence of a continuous time dimension, as opposed to frames taken at fixed intervals; The binary discretization of input data and SNN activations, as opposed to traditional image data (at least 8 bit) and floating point network activations. Event-based sensors encode information in the form of events that have a timestamp, location ( x, y ) and polarity (lighting increased or decreased). Because at any point in time an event can either be triggered or not, one can simply view event-based inputs as binary data by discretizing time ( Figure 1 ). In this view, the network's input is a three-dimensional array whose entries describe the number of events at a location ( x, y ) and in time bin t ; an additional dimension, of size 2, is added due to the polarity of events. If the time discretization is sufficiently precise, and no more than one event appears in each voxel, the data can be treated as binary. Figure 1 Schematic of the attack procedure on Dynamic Vision Sensor data. In this work, we present new algorithms that adapt the adversarial attacks SparseFool (Modas et al., 2018 ), and adversarial patches (Brown et al., 2017 ), to work with the time dynamics of spiking neural networks, and with the discrete nature of event-based data. We focus on the case of white box attacks, where the attacker has full access to the network and can backpropagate gradients through it. We test our attacks on the Neuromorphic MNIST (Orchard et al., 2015 ) and IBM Gestures (Amir et al., 2017 ) datasets, which are the most common benchmark datasets within the neuromorphic community. Importantly, for the first time, we also test the validity of our methods by deploying the attacks on neuromorphic hardware. Our contributions can be summarized as follows: We contribute algorithms that adapt several adversarial attacks strategies to event-based data and Spiking Neural Network, with detailed results to quantify their effectiveness and scalability. We show that these adapted algorithms outperform current state-of-the-art algorithms in the domain of Spiking Convolutional Neural Network. We show targeted universal attacks on event-based data in the form of adversarial patches, which do not require prior knowledge of the input. We validate the resulting adversarial examples on an Spiking Neural Network deployed on a convolutional neuromorphic chip. To the best of our knowledge, this is the first time the effectiveness of adversarial examples is demonstrated directly on neuromorphic hardware.",
"discussion": "5. Discussion We studied the possibility of fooling Spiking Neural Network through adversarial perturbations to Dynamic Vision Sensor data, and verified these perturbations on a spiking convolutional neuromorphic chip. There were two main challenges to this endeavor: the discrete nature of event-based data, and their dependence on time. Dynamic Vision Sensor attacks also have different sparsity requirements, because the magnitude of the perturbation is measured in terms of number of events added or removed. For this purpose, we adopted a surrogate-gradient method and backpropagation-through-time to perform white-box attacks on spiking networks. We presented SpikeFool, and adapted version of SparseFool, which we compared to current state-of-the-art methods on well-known benchmarks. We find that SpikeFool achieves near perfect success rates at lowest perturbation magnitudes on time-discretized samples of the Neuromorphic MNIST and IBM Gestures datasets. In the best cases, the attack requires the addition of less than a hundred events over 200 ms. To the best of our knowledge, we were also the first to show that the perturbation is effective on a network deployed on a neuromorphic chip, implying that the method is resilient to the small but non-trivial mismatch between simulated and deployed networks. Additionally, since SpikeFool computes perturbations offline and not on a live stream of Dynamic Vision Sensor events, we also investigated a more realistic setting, where an adversary can inject spurious events in the form of a patch inserted into the visual field of the Dynamic Vision Sensor camera. We demonstrated that we can generate patches for different target labels. Although these patches require a much higher amount of added events, they do not require prior knowledge of the input sample and therefore offer a realistic way of fooling deployed neuromorphic systems. A natural next step would be to understand whether it is possible to build real-world patches that can fool the system from a variety of distances and orientations, as Eykholt et al. ( 2018 ) did for photographs. Moreover, it will be interesting to see how important knowledge about the architecture is and if one can generate patches by having access to a network that differs from the one deployed."
} | 2,612 |
36730197 | PMC9963355 | pmc | 1,136 | {
"abstract": "Significance Bacteria live in multispecies, spatially structured communities ubiquitously in the natural world. These communities, or biofilms, have a strong impact on microbial ecology, but we often do not know how cellular scale interactions determine overall biofilm structure and community dynamics. Here we explore this problem in the context of predator–prey interaction, with two prey species— Vibrio cholerae and Escherichia coli —being attacked by the bacterial predator Bdellovibrio bacteriovorus . We find that when V. cholerae and E. coli grow together in biofilms, the architectures that they produce change in ways that cannot be predicted from looking at each prey species alone, and that these changes in cell group structure impact the community dynamics of predator–prey interaction in biofilms.",
"discussion": "Discussion Exploration of multispecies biofilm communities using live high-resolution imaging is crucial to understanding microbial ecology at the spatial scale on which cell–cell interactions occur ( 11 , 20 , 23 , 82 – 91 ). Here we tracked the spatial population dynamics of the bacterial predator B. bacteriovorus in dual-species prey biofilms of V. cholerae and E. coli , finding that the survival rates of both prey species are altered, but in opposite directions, when they are growing together. V. cholerae produces biofilm cell clusters that reach a cell packing density threshold past which B. bacteriovorus cannot enter, protecting the prey within. E. coli can become enveloped along the basal layers of these highly packed structures, co-opting predator protection from V. cholerae and increasing E. coli survival relative to when growing on its own. By contrast, in dual-species biofilms, a fraction of V. cholerae becomes entangled with E. coli early during biofilm growth, leading to an alternate structure that is more homogeneously mixed, disordered, and loosely packed. These disordered cell groups do not block predator cell entry, and all prey within them are killed by B. bacteriovorus . As a result of these biofilm structural dynamics, V. cholerae survival decreases in co-culture with E. coli relative to when growing on its own. At any given location, which of these two alternative cell group structures emerge depends on the initial distance between V. cholerae and E. coli cells that have attached to the underlying surface. Surface colonization patterns therefore determine the relative occurrence of predation-protected cell groups versus susceptible cell groups and the overall rates of B. bacteriovorus predator survival for each prey species. This study makes explicit that the cellular arrangement and tightly packed structure of clonal V. cholerae groups can operate as a type of public good ( 39 , 92 ) that confers predator protection to the cells within [among many other benefits ( 30 , 33 – 35 , 55 , 74 , 93 – 97 )]. Other species—here, E. coli , whose mono-species biofilms are susceptible to B. bacteriovorus —can take advantage of this protective architecture when small groups of them become enveloped by expanding, highly packed biofilms of V. cholerae . By contrast, if too many E. coli cells are present in close enough proximity to V. cholerae at the start of biofilm growth, then V. cholerae cannot initiate its normal cell group structure, and the public good benefit of predation protection completely breaks down in that location. It is notable that the spatial architecture of biofilm-producing bacteria can manifest as a public good that is exploitable across species in this manner. In this case, the stability of V. cholerae cooperative architecture depends on the initial surface population density, which determines whether V. cholerae cell lineages have enough space to nucleate the highly packed core regions of expanding biofilm clusters before encountering cells of other species. Though distinct in mechanistic detail, this example should fall under related social evolution principles as other kinds of microbial cooperation that provide benefits in a distance-dependent manner. Recent work has highlighted in detail how the population dynamics and evolutionary stability of this class of cooperative behavior depend on the spatial range of cooperative sharing, the population/community composition, and spatial cell arrangements during early biofilm growth ( 33 , 34 , 39 , 93 , 96 , 98 ). The interplay of V. cholerae , E. coli , and B. bacteriovorus in co-culture emphasizes that the population dynamics of different species in a community can depend quite strongly on the cellular resolution details of biofilm structure, which in turn can differ in unexpected ways between mono-species and multispecies systems. In recent years microbiologists have made tremendous strides in understanding the cellular and molecular nuances of biofilm architecture and their relationship to microbial ecology and evolution. By necessity for tractability in many cases, much of this work has focused on one species at a time. Our experiments here highlight how new and interesting questions about the drivers of biofilm structure, and the relationship between biofilm structure and community ecology, can arise from modest increases in complexity with multispecies systems. Here, it appears as though E. coli —if adjacent to V. cholerae at the start of biofilm growth—may interfere with normal localization of at least one component of the V. cholerae matrix, which could then contribute to the qualitative differences in ordered versus disordered architectures that appear later during biofilm growth. The connection between initial surface coverage and multispecies prey biofilm architecture, and the additional connection between biofilm architecture and predator exposure, together lead to an interesting dependence between early biofilm growth conditions and predator–prey ecology. It would be fruitful to explore how and when these relationships generalize to other species combinations and biofilm environmental growth conditions with increasing ecological realism. Where prior studies have analyzed multispecies biofilms at high resolution, they have also indicated important consequences for community structure and environmental impacts ( 11 , 12 , 22 , 47 , 48 , 99 – 103 ). A notable recent example examined the detailed structure of multispecies biofilm communities growing as plaque in dental caries ( 48 ). Kim et al. showed that Streptococcus mutans forms consistent spatial arrangements in biofilm co-culture with other oral microbiota species. In this case, S. mutans consistently produces core clonal regions, around which form layers of non- mutans streptococci followed by non-streptococci. The metabolic activity of S. mutans within the inner regions of these multispecies biofilms caused low local pH that could recapitulate the rapid demineralization of enamel that occurs during development of caries in vivo. Our work here highlights how the details of early surface colonization conditions can cascade into qualitative differences in subsequent biofilm architecture and ecological dynamics. This result points to several goals for future work. Any phenotypes that alter surface exploration or settling patterns, including gliding and twitching motility, as well as any positive or negative interactions within and between species—for example, via shared adhesin production, metabolite trophic interaction, or toxin secretion—could also cascade to major differences in biofilm spatial architecture. Differences in environmental topography and the orientation of nutrient supply, which may often derive in natural environments from the underlying surface rather surrounding liquid, should also be pursued to gain a fuller picture of how the subtleties of biofilm growth in realistic environments impact community structure. Novel experimental systems that implement these increases in ecological realism while maintaining access by high-resolution imaging will be important platforms for further study."
} | 2,019 |
32825332 | PMC7569806 | pmc | 1,137 | {
"abstract": "A molecular robot is a microorganism-imitating micro robot that is designed from the molecular level and constructed by bottom-up approaches. As with conventional robots, molecular robots consist of three essential robotics elements: control of intelligent systems, sensors, and actuators, all integrated into a single micro compartment. Due to recent developments in microfluidic technologies, DNA nanotechnologies, synthetic biology, and molecular engineering, these individual parts have been developed, with the final picture beginning to come together. In this review, we describe recent developments of these sensors, actuators, and intelligence systems that can be applied to liposome-based molecular robots. First, we explain liposome generation for the compartments of molecular robots. Next, we discuss the emergence of robotics functions by using and functionalizing liposomal membranes. Then, we discuss actuators and intelligence via the encapsulation of chemicals into liposomes. Finally, the future vision and the challenges of molecular robots are described.",
"conclusion": "5. Conclusions and Future Outlook In this review, we focused on recent advances in liposome-based molecular robots. We described microfluidic methods of GUV generation and the emergence of sensors, actuators, and intelligence systems via the functionalization of liposomal membranes and the encapsulation of chemicals into GUVs. To date, the efficient generation of GUVs with a size-monodispersity and chemical encapsulation ability has been achieved by using microfluidic techniques. In addition, DNA capsules and polymersomes offer a potential as the compartments of molecular robots. Regarding the sensors, synthetic ion channels and nanopore sensing technologies are useful for sensing external chemicals. In particular, recent developments of de novo -designed peptide, protein, and DNA channels can provide stable structures with good size controllability, and these may attract attention for use as sensors of molecular robots. Nanopore sensing has recently offered specific chemical detection by combining DNA computing techniques, and this system has also provided both sensory and intelligence functions. Furthermore, viscosity measurements have been achieved by using hairpin DNA-integrated αHL nanopores. Regarding the actuators, encapsulated molecular motors can facilitate the deformation of GUVs, with the actuation being controlled by specific stimulations. In the future, we will look towards a mechanism in which the actuation of molecular motors will be efficiently converted into migratory motion. It is proposed that the intelligent control of molecular robots can be achieved with DNA computing technologies and the isothermal amplification of DNA. Because intelligence plays a role as the interface between sensors and actuators, an intelligence system should be designed and optimized for applications in molecular robots. Finally, molecular robots are microorganism-imitating robots aimed towards a practical usage in the fields of medicine, drug discovery, environmental science, food science, and energy science in a manner distinct from artificial cells. However, it is still challenging to integrate all the robotics systems into a single compartment because it is necessary to design and construct each of the functions, including sensors, actuators, and intelligence, so that they do not interfere with each other and can act in a single GUV. In the present state, molecular robots have not shown a clear distinction from artificial cells. As a field, we should aim to develop robotic systems that can provide superior functions and abilities to cells by combining various technologies that encompass not only synthetic biology but also mechanical, electrical, chemical, and information engineering. We believe that the establishment of molecular robotics methods would offer a game-changing technology in these various fields.",
"introduction": "1. Introduction An artificial cell is a cell-imitating artificial system that exhibits characteristics of living cells, including evolution, self-reproduction, metabolization, and communication [ 1 , 2 , 3 ]. The development of artificial cells is one of the main objectives in the field of synthetic biology, and numerous studies have been reported. For example, gene expression [ 4 ], metabolic networks [ 5 , 6 ], growth and division [ 7 , 8 ], adaptivity [ 9 ], communications [ 10 , 11 ], and motility [ 12 ] of artificial cells have been reported. Although the recent developments of artificial cells successfully provided the potential for practical applications [ 13 , 14 ], the functioning of artificial cells has not yet reached a practical level. Hence, another concept called “molecular robotics” has been proposed, which aims to offer practical uses in the fields of medicine, drug discovery, environmental science, food science, and energy science [ 15 , 16 , 17 ]. In this concept, the essential robotics components that are intelligence systems [ 18 ], sensors [ 19 ], and actuators [ 20 , 21 , 22 , 23 ] are developed by integrating various technologies, such as DNA nanotechnologies, synthetic biology, polymer chemistry, and robotics, and these components are implemented into a microcompartment using bottom-up approaches. Although these elements have been individually developed, with performances that potentially offer practical uses, the integration of these elements into a single system is still challenging. To integrate these three elements into a single robotics system, a compartment that can encapsulate these components is required. Giant unilamellar vesicles (GUVs) [ 24 , 25 ], DNA capsules [ 26 , 27 ], gels [ 28 , 29 ], and polymersomes [ 30 ] have been proposed as structures that may act as such compartments. The compartment of a molecular robot must be able to undergo membrane functionalization with membrane proteins and be able to facilitate the encapsulation of molecules in order to implement intelligence systems, sensors, and actuators. GUVs can, in particular, satisfy these requirements. Here, we review recent advances in GUV-based molecular robot technologies, including methods of GUV production, the emergence of sensors by the functionalization of liposomal membranes, and the emergence of actuation and intelligence by encapsulating macromolecules into GUVs. In addition, we summarize the future outlook towards the integration of individual elements into a single compartment and the manufacture of a liposome-based molecular robot."
} | 1,629 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.