pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
29170549 | null | s2 | 8,771 | {
"abstract": "Biomaterials engineered with specific bioactive ligands, tunable mechanical properties, and complex architectural features have emerged as powerful tools to probe how cells sense and respond to the physical properties of their material surroundings, and ultimately provide designer approaches to control cell function."
} | 79 |
36447563 | PMC9663969 | pmc | 8,773 | {
"abstract": "Many geckos have the remarkable ability to reversibly adhere to surfaces using a hierarchical system that includes both internal and external elements. The vast majority of studies have examined the performance of the adhesive system using adults and engineered materials and substrates (e.g., acrylic glass). Almost nothing is known about how the system changes with body size, nor how these changes would influence the ability to adhere to surfaces in nature. Using Tokay geckos ( Gekko gecko ), we examined the post-hatching scaling of morphology and frictional adhesive performance in animals ranging from 5 to 125 grams in body mass. We quantified setal density, setal length, and toepad area using SEM. This was then used to estimate the theoretical maximum adhesive force. We tested performance with 14 live geckos on eight surfaces ranging from extremely smooth (acrylic glass) to relatively rough (100-grit sandpaper). Surfaces were attached to a force transducer, and multiple trials were conducted for each individual. We found that setal length scaled with negatively allometry, but toepad area scaled with isometry. Setal density remained constant across the wide range in body size. The relationship between body mass and adhesive performance was generally similar across all surfaces, but rough surfaces had much lower values than smooth surfaces. The safety factor went down with body mass and with surface roughness, suggesting that smaller animals may be more likely to occupy rough substrates in their natural habitat.",
"introduction": "Introduction \n Animals attach to surfaces in numerous ways, including claws, suction, and both wet and dry adhesion. In fact, some animals can utilize multiple attachment mechanisms [ 1 – 2 ], leading to multifunctionality across surfaces of varying roughness. Dry adhesion is found in many invertebrates and squamate reptiles, and has been a focus of both engineering and biological studies [ 3 ]. Models are frequently used to describe adhesion, such as the Johnson–Kendall–Roberts (JKR) model [ 4 ]. In this case, the force required to pull an elastic sphere from a flat surface is determined using the radius of the sphere and the adhesion energy between a sphere and the surface. More recent studies use the JKR model to determine the role of setal density in adhesion from insects to geckos [ 5 ]. Despite many advancements in our understanding of adhesion across organisms, few studies have incorporated ecologically relevant variables. \n The ability of geckos to adhere to smooth surfaces has fascinated scientists since Aristotle, and has been followed by countless studies focused on uncovering the mechanisms of gecko adhesion, from as early as the 1800’s [ 6 – 10 ] to modern investigations (reviewed in [ 3 ]). Like the Lotus Effect [ 11 ], the Gecko Effect has seen a surge in attention over the past couple of decades [ 12 ]. There are over 1000 species of geckos with adhesive capabilities, with multiple origins of the system [ 13 – 14 ]. However, much of what is known about gecko adhesion and its associated structures is based on studies of a single species, the Tokay gecko ( Gekko gecko ) [ 15 – 19 ]. Additionally, the primary focus has been on adult geckos, likely given their larger size. Two key questions have received very little attention. First, how does adhesive performance vary across surfaces of different roughness? Second, how does adhesive performance and morphology vary with body size? Although the former has been the subject of a few studies, the latter has received almost no attention. \n Surfaces in nature are rarely smooth and geckos are found on all types of surfaces from rough rocks to undulant tree bark [ 20 – 25 ] ( Figure 1 ). Recent studies have begun to explore the role of surface roughness on frictional adhesion in geckos [ 1 , 21 , 25 – 26 ], and performance typically declines as roughness increases. For example, Vanhooydonck and colleagues examined the effects of substrate structure on speed and acceleration capacity in climbing geckos, and they found that acceleration was greatest on the smoothest surface (wood) where the most contact between the adhesive system and the surface could be made [ 27 ]. This illustrates that the main issue faced by geckos that are attaching via adhesive pads is the contact area between the setae and the surface. With increasingly rough surfaces, the area for contact decreases, leading to decreased adhesive performance. In a modeling framework, the force of adhesion can be related to surface energy of the substrate, the area of the adhering pad, and the compliance of the system [ 28 ]. However, most studies use widely varying surfaces [ 1 ] or uniform 3D printed surfaces [ 29 ] that do not capture the random fine-scale roughness that is likely apparent in natural habitats. For example, Huber et al. [ 26 ] measured the shear adhesive force of a single spatula on surfaces with asperities ranging from 100–300 nm, and Gillies et al. [ 29 ] manipulated the surface roughness of a macroscopic engineered rough surface in which they manipulated the wavelength and amplitude of peaks that were on the same length scale of the subdigital lamellae. Both studies found that shear adhesion was significantly reduced (up to 95% reduction of force produced on acrylic glass) on surfaces where the surface structure was close to matching the animal’s adhesive structure, highlighting the importance of considering length-scale and the impact it has on gecko adhesion when testing the effects of roughness. \n \n Figure 1 \n \n Images of a Tokay gecko in its natural habitat in Vietnam (photo courtesy of Lee Grismer. This content is not subject to CC BY 4.0.) (A), a Tokay gecko in the lab on a glass surface (photo by Timothy Higham, and has not been published previously) (B), an SEM image of the distal portion of the digit (C), an SEM image of the 2000 grit sandpaper surface (D), and a 3D image of the 2000 grit sandpaper surface using confocal laser scanning microscopy (E). \n \n \n \n Surface roughness can be qualitatively characterized in different ways (rough vs smooth), but it is a complex parameter to quantify as real surfaces in nature vary over many length scales and can have significant effects on the efficacy of an adhesive system [ 30 – 31 ]. At this point in time, the ability to quantify the topography of surfaces of varying roughness [ 21 ], and to replicate them [ 32 – 36 ], rather than using vague categorizations, allows for the possibility to test fine-scale interactions of animal adhesion and traction with more accuracy [ 35 – 37 ]. A key factor that is relatively unexplored is that the morphology of the adhesive system likely changes with body size, which can then impact the amount of contact made between the adhesive system and the surface on which it is clinging or moving. For example, setal length and toepad area have been found to increase with body size in the southern African gecko Chondrodactylus bibronii [ 38 ]. Beyond intraspecific scaling, a recent study found extreme positive allometry in toepad area among animals that have adhesive pads, from mites to geckos [ 39 ]. Thus, it is likely that animals of different size will have varying clinging and locomotor performance on rough surfaces. \n Allometry plays a significant role in natural systems by imparting physical constraints of supporting different body sizes, but also in the mechanical consequences in relation to locomotion [ 40 – 43 ]. Scaling becomes increasingly important when structures on the surface of the animal must support the body through adhesion on vertical or near-vertical surfaces [ 39 , 44 ]. In the case of dry adhesives, studies have focused on the scaling of toepad morphology because of the inherent signficance for adhesive locomotion (i.e., larger toepad area leads to greater area for surface contact). In a study of geckos, skinks, and anoles, Irschick et al. [ 45 ] found a strong correlation between shear adhesive force and toepad area. However, the slope of the relationship between toepad area and body mass was lower than that of clinging ability. This suggests that there are other underlying mechanisms that contribute to clinging ability apart from pad area. Adhesive pad area across climbing taxa spans seven orders of magnitude and scales with positive allometry [ 44 ] but, after accounting for size and phylogeny, toepad area scaled with isometry or sometimes negative allometry within certain clades. A key question is how larger organisms that use adhesion will support their body weight when climbing vertical surfaces. \n The scaling of adhesive components has been addressed by Webster et al. [ 38 ] and, to a lesser extent, by Delannoy [ 46 ]. In a study of Chondrodactylus bibronii , setal density, setal basal diamter and setal spacing did not change significantly throughout ontogeny, but pad area and setal length increased with body size [ 38 ]. Despite these increases, estimated adhesive force capacity, relative to body size, decreased with ontogeny. However, there is a mismatch between morphological measurements and measurements of adhesive force in that morphology is often used to estimate force-generating capabilities on surfaces of varying roughness. What is missing is a study that examines both the scaling of adhesion on different surfaces and the changes in morphology throughout with body size. It is predicted that increasing roughness will decrease adhesive performance due to the limited area of contact islands [ 21 , 47 ]. \n The efficacy of adhesives that mimic a gecko’s system depends upon knowing the natural interactions between the animal and the substrate. All else being equal, longer and softer setal shafts are predicted to result in better adhesion on rough surfaces due to a reduction in the effective elastic modulus [ 48 ]. However, it is currently unclear whether this translates into higher forces, relative to body mass, under whole-organism experimental studies. In order to fully understand how performance is influenced by roughness, incorporating variation in body size is important. Tokay geckos are ideal for investigating the role of body size variation given that they reach very large body sizes and they live in rainforests that likely exhibit variation in roughness (see Figure 1 for example). \n Here we will measure Tokay gecko adhesive structures and compare theoretical force estimates to actual performance values to address several questions: 1) how does pad area scale with adhesive force over a significant range in body size, 2) do more compliant setae translate to higher adhesive force, and 3) do larger geckos exhibit greater adhesive force, relative to body mass? Additionally, to better understand how gecko adhesive structures interact with surface asperities on the setal level, we tested the impact of surface roughness on shear adhesion over a large size range of geckos, given that even the smallest change in setal length and compliance could impact the capacity to make local adjustments to rough surfaces [ 49 ]. To test how asperity size impacts adhesion, we generated 3D surface profiles of seven different sandpaper grits and carried out adhesion trials over a large size range of G. gecko . This research is significant given that most studies are limited to using 2D profiles, missing an entire axis of surface structure variation. Additionally, no study to date has quantified the effect of body size on adhesive performance.",
"discussion": "Discussion \n Our integrative approach, combining morphology, performance, and 3D surface topography, revealed key aspects of scaling that have significant impacts on our understanding of gecko adhesion. Setal diameter and density did not change with body size, whereas toepad area, and setal length, and therefore setal aspect ratio, increased with body size. Frictional adhesion, measured experimentally, increased with body size across all surfaces. However, adhesive safety factor was not only lower on rougher surfaces, but also lower for larger animals. This has implications for ecology, especially habitat use through ontogeny, but also biomimetics. If we are attempting to mimic the adhesive system, would it be beneficial to reconstruct the adhesive system of a small gecko or a large one? Can we create an adaptable system that could achieve a constant level of adhesive performance, relative to body mass, across surfaces of varying roughness? \n \n Scaling of morphology and estimated frictional adhesion \n Toepad area is important for gecko adhesion given than more points of contact will be made as area increases, assuming a constant density and a smooth surface. Indeed, density did not change with body mass in our study, but toepad area increased isometrically with body mass (slope of 2.03; Table 1 ). This is almost identical to that found for Chondrodactylus bibronii by Webster et al. [ 38 ]. Body mass increased relative to SVL 3.34 , which is not different from isometry. This suggests that adhesive force is greater, relative to body mass, for smaller individuals. Indeed, adhesive force estimates, based on toepad area and density, increased with body mass 0.61 . This, again, is supported by the work of Webster et al. [ 38 ]. However, this differs from the interspecific study by Irschick et al. [ 45 ], in which a positively allometric relationship between pad area and body mass was observed. As noted by Webster and colleagues [ 38 ], this is like attributable to the fact that they included many types of subdigital pad design [ 50 ]. Two explanations for our observation are outlined by Webster et al. [ 38 ]. First, the lack of a positively allometric relationship may be the result of physical constraints. Having much larger toepads, relative to body size, in larger animals would potentially lead to overlapping pads and, ultimately, disruption of adhesion. Second, smaller animals may benefit from relatively larger toepads since few contact islands are likely to be encountered in any given footfall [ 38 ]. Ecological consequences are discussed below. \n \n \n Scaling of frictional adhesion \n In addition to the morphological analyses and the estimates of frictional adhesion, we experimentally measured the latter. Adhesive force increased with body size on the acrylic glass surface, with peak values approximating those in other studies. The scaling exponent was 0.54, and this was negatively allometric if we assume adhesive force is directly proportional to toepad area. This is in contrast to previous studies that examine multiple species of pad-bearing lizards that found a scaling exponent not significantly different from 1 [ 45 , 51 ]. One explanation for the difference between the current studies and the two former studies is that those included multiple types of toepads [ 50 ]. \n \n \n Predicted versus measured frictional adhesive force \n In addition to the morphological analyses and the estimates of frictional adhesion, we experimentally measured the latter. This, to our knowledge, is the first study to directly compare estimates of adhesive force (from morphology) to experimental measurements. Although the slopes and the strength of the regression are comparable between the predicted and measured frictional adhesive forces ( Figure 4 ), there are significant differences in the actual values. The predicted forces are much higher than the measured, with the experimental values averaging 50% of the predicted values. This may not be surprising. Theoretical estimates, based on density, pad area, and the average force per seta, are reliant upon the assumption that every single seta makes contact. Our results suggest that, even on incredibly smooth surfaces such as acrylic glass, quite a few setae are not in contact with the surface. However, there are other reasons for this mismatch. The manus and pes of the gecko, when pulled across a surface, do not have all of the toes aligned parallel to the direction of movement. In a study of Tokay geckos, Stewart and Higham [ 16 ] used an apparatus to pull individuals across an acrylic glass platform while obtaining high-speed video in ventral view. They found that the average digit angle started at approximately 30 degrees and decreased to approximately 10 degrees throughout a pulling trial. At the same time, overlap among toepads within a single foot also increased, reducing the area of contact from almost 100% to approximately 75% near the end of the pulling trial [ 16 ]. The angles of the toes, never becoming completely parallel, coupled with the decreased contact area of the toepads, likely decrease the force generating capacity of the setal fields. Beyond these macroscopic factors, it is unclear whether all of the setae, in a region of the pad that is seemingly in contact with the surface, are actually engaged with the surface. There could be interactions among setae that preclude their attachment. Damage to some setae may also decrease the efficacy of setal attachment. Future work that visualizes the actual contact across the entire setal field will reveal whether this is an important factor. \n \n \n The influence of surface roughness \n The roughness of the sandpaper surfaces in our study ranged from Sq = 34.8 μm (100 grit sandpaper) to Sq = 3.16 μm (3000 grit sandpaper). Acrylic glass is considered perfectly smooth. The roughness of the substrates had a large impact on the adhesive ability in Tokay geckos. Despite the similar slopes (with the exception of acrylic glass, which was negatively allometric), adhesive force on 100 grit sandpaper was, on average, only 13% of that on acrylic glass. The likely explanation for this result is that the available contact area was reduced with increasing roughness. This has been observed in other studies that model the contact between a gecko toe and surfaces that vary in asperity size [ 25 , 49 , 52 – 53 ]. Frictional adhesive force increased isometrically with body mass across all of the sandpaper surfaces, but this led to a decrease in safety factor (see below) since force is related to toepad area, and toepad increases at a slower rate compared to body mass. \n \n \n Ecological consequences of reduced safety factor \n Safety factor (SF), as measured experimentally, decreases with body size in Tokay geckos, and it also decreases with increasing roughness. The former result aligns with the previous work on C. bibronii by Webster et al. [ 38 ]. They found, using estimates of frictional adhesive force, that SF decreased with body mass −0.47 . We found that SF decreased with body mass −0.46 on acrylic glass and with body mass −0.45 on 100 grit sandpaper. Despite the similar slopes, SF on 100 grit sandpaper was, on average, only 13% of that on acrylic glass. Thus, large animals on rough surfaces will face the largest challenges due to lower safety factor. Indeed, the lowest SF was 9.7, which was a large gecko on the 100 grit surface. The largest value was 502, which was the smallest gecko on the acrylic glass surface. \n What does this mean for animals in nature? We predict that these drastic differences in SF will likely dictate, to some extent, habitat use. Wright et al., 2021 examined how clinging performance in geckos and anoles on natural surfaces might predict the surfaces used in nature [ 24 ]. They found that performance on natural substrates predicted which texture (rough vs smooth) was most often used by each species. Translating this interspecific study to our intraspecific experiments, we predict that larger animals should occupy lower regions of a tree to avoid the negative consequences of falling a large distance. This is assuming that they occupy the same type of substrate as a smaller gecko. If they do fall from a high position, the impact force is likely to exceed the adhesive force capacity of their feet [ 54 ]. In terms of surface texture, they might be expected to avoid very rough surfaces in order to preserve a modest SF. In contrast, smaller geckos will have larger values of SF, and will experience lower impact forces from a fall, suggesting that they might safely occupy higher regions of a tree. Future field observations that determine the potential for ontogenetic habitat shifts are needed. Additionally, an ecomechanical model [ 55 ], incorporating contact area information from surfaces in nature, would help to understand the mechanisms underlying any shifts. \n \n \n Future directions and biomimetics \n Our results detail the changes in both morphology and adhesive performance in relation to body size in a single species of gecko. These results generally align with other intraspecific studies, but not with interspecific studies. This mismatch is interesting, and requires further investigation. Specifically, there are different subdigital pad designs across geckos [ 50 ], but only one type has been investigated thus far. Do these patterns hold across geckos with different types of pads? If not, how and why? \n Our results suggest important changes that occur throughout ontogeny. However, it is also clear that there are constraints in what changes are possible, such as setal width and density. These constraints may simply be due to spacing, which does not appear to change with body size. If spacing among setae does not change, it follows that the diameter of each seta is also likely to remain constant. If not, there might be negative interactions among the setal shafts (e.g., clumping). How can this be applied to biomimetics? Can we construct adaptable adhesive devices that accommodate different surfaces, or that can change depending on the need? Should robots be fitted with different systems depending on size? Although setal shaft diameter and density do not change in living animals, how does changing these conditions in artificial systems alter function? Not only might this assist with biomimetic initiatives, but it might also help us to understand why these do not occur in living animals. Regardless, we often use animals of a specific size when making connections to biomimetics, including the construction of artificial adhesives. Body size should be included given the large differences across individuals. This could be achieved by using an array of species that vary in size or, as in the case of our study, a series of individuals that span a wide range in body size. The benefits of the latter are that the general form of the toepad, and likely other aspects of the integrated adhesive system, are kept constant. However, the diversity among species will clearly offer other tantalizing insights into the possible solutions to adhesion."
} | 5,656 |
38380107 | PMC10877099 | pmc | 8,774 | {
"abstract": "Highlights • Microbe-meditated decomposition of agricultural residues and their utilization for enhancing nutrient cycling in the soil and crop productivity. • Integration of emerging recycling technologies for crop residue management and enzyme mediated biotransformation protocols for recovery of high value molecules from crop residues. • Advances in microbial biotechnology have made it possible to transform waste materials into valuable resources."
} | 113 |
38248584 | PMC10813543 | pmc | 8,776 | {
"abstract": "The robot task sequencing problem and trajectory planning problem are two important issues in the robotic optimization domain and are solved sequentially in two separate levels in traditional studies. This paradigm disregards the potential synergistic impact between the two problems, resulting in a local optimum solution. To address this problem, this paper formulates a co-optimization model that integrates the task sequencing problem and trajectory planning problem into a holistic problem, abbreviated as the robot TSTP problem. To solve the TSTP problem, we model the optimization process as a Markov decision process and propose a deep reinforcement learning (DRL)-based method to facilitate problem solving. To validate the proposed approach, multiple test cases are used to verify the feasibility of the TSTP model and the solving capability of the DRL method. The real-world experimental results demonstrate that the DRL method can achieve a 30.54% energy savings compared to the traditional evolution algorithm, and the computational time required by the proposed DRL method is much shorter than those of the evolutionary algorithms. In addition, when adopting the TSTP model, a 18.22% energy reduction can be achieved compared to using the sequential optimization model.",
"conclusion": "6. Conclusions This work studied the simultaneous optimization of the robot task sequencing problem and trajectory planning problem, briefly called the robot TSTP problem. The objective is the simultaneous minimization of the robot’s moving time and energy consumption. The robot TSTP problem is an NP-hard problem; it involves task point sequencing, IK solution calculation and selection, trajectory parameter determination, and trajectory planning. The solution space and computational complexity are much larger than those of any single optimization problem. To solve the robot TSTP problem, a hybrid method that combines the DQN algorithm and the QPI algorithm is proposed. First, the TSTP problem is modeled as a Markov decision process; then, we use the DQN algorithm to select the task points, IK solutions, and trajectory parameters at each decision step. Finally, the QPI algorithm is used to plan the moving trajectories between task points. Through detailed experiments, we evaluated the competitiveness of our proposed DQN algorithm in solving the robot TSTP problem in terms of computation speed and solution quality. In addition, the experimental results confirm the superiority of the TSTP model. The TSTP model extends the optimization space and enhances the DQN algorithm, enabling it to find more excellent solutions than any of the sequential optimization models. Furthermore, the TSTP model also increases the production flexibility of the robotic cell because it can easily adjust the scheduling and controlling parameters to respond to changes in production targets or conditions. Finally, when solving the TSTP problem, setting the time–energy optimization objective is more reasonable than setting a single objective, as it can explore a solution that can improve production efficiency while maximizing energy use. Currently, our research work only considers one robot’s production time–energy optimization, which limits the application of the proposed approach. The future direction is to apply these approaches to (a) multi-robot production environments, (b) stochastic dynamic production environments, and (c) a combination of the above.",
"introduction": "1. Introduction An increasing number of robots are being deployed on production floors to achieve intelligent manufacturing. This has motivated the development of production optimization technology in robotics. Robot task sequencing [ 1 ] and trajectory planning [ 2 ] are two traditionally separate optimization problems in robotics. The task sequencing problem focuses on finding the optimal task execution sequence for the robot in order to achieve certain objectives. It is analogous to the traveling salesman problem (TSP), but it is more complex because of the robot’s kinematic redundancy (also known as the IK solution). The robot trajectory planning problem involves determining the timing of the motion law that the robot follows along a predefined geometric path while satisfying specific requirements, such as trajectory smoothness and accuracy, and achieving desired objectives, such as those related to the execution time, energy consumption, vibration, and their combinations; a general overview is given in [ 3 ]. In the robotic work cell, there exists a unique group of application scenarios that involve task sequencing and trajectory planning problems simultaneously, such as spot welding, freeform surface inspection, or spray painting [ 4 ]. In these applications, the robot is required to visit a set of task points with no predefined sequence to perform corresponding tasks, and it must finally return to its initial state. Clearly, the visiting sequence of the task points and the trajectories followed for the robot to reach each task point have strong effects on the robot’s production efficiency. Additionally, different task-point visiting sequences can lead to alterations in the robot’s moving path, consequently impacting the robot’s movement trajectories. Therefore, it is necessary to address the task sequencing and trajectory planning problems simultaneously. However, the robot task sequencing and trajectory planning problems are traditionally treated as two separate issues. The robot motion trajectories are often predefined at the production control level, and then, the production scheduling level conducts the task sequence planning. This sequential optimization model diminishes the optimization space of the production problems and can improve productivity to a certain extent, but the optimization solutions are often suboptimal because of the neglect of the underlying synergies associated with the optimization objectives among the task sequencing and trajectory planning problems. In addition, the solution space for the TSTP problem grows exponentially with the number of manufacturing points. For an exhaustive search, the computational complexity is O ( m ! g m ) , where m is the number of task points and g is the number of IK solutions for each point. Traditional methods adopted to solve task sequencing and trajectory planning problems, such as metaheuristics and dispatching rules, face challenges in terms of time efficiency and ensuring high-quality solutions. These methods either suffer from slow processing times or struggle to consistently achieve optimal solutions that meet various objectives. Over the past few years, deep reinforcement learning algorithms (DRL) [ 5 ] have enabled significant advancements; their excellent self-learning and self-optimization qualities enable them to solve complex decision-making problems quickly and accurately, which has resulted in their extensive utilization across various robot work cell optimization problems [ 6 ]. Motivated by the aforementioned co-optimization requirements and DRL algorithms, in this paper, we take the optimization problems in task sequencing and trajectory planning as a monolithic problem and model it as a Markov decision process. Furthermore, a DRL-based optimization method is developed to address the above co-optimization problem. The major contributions are as follows. To the best of our knowledge, we are the first to combine the traditionally separate optimization problems in task sequencing and trajectory planning into a monolithic problem, called the robot TSTP problem, providing an integrated view in the discrete manufacturing domain. To solve the above TSTP problem, we employ a DRL-based policy for decision optimization. During the learning process, a specific state representation, action space, and reward function are carefully designed. Typically, in the action space, each action step considers the selection of task points, IK solutions, and trajectory parameters concurrently. Given the absence of a benchmark test for scheduling and control co-optimization, the feasibility of the proposed TSTP model and the effectiveness of the DRL are validated through a demonstration abstracted from a specific real-world case: a spot-welding task in an automation plant with a UR5 robot. The rest of this article is organized as follows. Section 2 provides a summary of the related work. Section 3 formulates the mathematical model of the robot TSTP problem under study. Section 4 presents the methodology. Section 5 presents the experiments and results. Section 6 concludes this study."
} | 2,148 |
32238866 | PMC7113308 | pmc | 8,777 | {
"abstract": "Microbial cells in the seabed are thought to persist by slow population turnover rates and extremely low energy requirements. External stimulations such as seafloor hydrocarbon seeps have been demonstrated to significantly boost microbial growth; however, the microbial community response has not been fully understood. Here we report a comparative metagenomic study of microbial response to natural hydrocarbon seeps in the Gulf of Mexico. Subsurface sediments (10–15 cm below seafloor) were collected from five natural seep sites and two reference sites. The resulting metagenome sequencing datasets were analyzed with both gene-based and genome-based approaches. 16S rRNA gene-based analyses suggest that the seep samples are distinct from the references by both 16S rRNA fractional content and phylogeny, with the former dominated by ANME-1 archaea (~50% of total) and Desulfobacterales, and the latter dominated by the Deltaproteobacteria, Planctomycetes, and Chloroflexi phyla. Sulfate-reducing bacteria (SRB) are present in both types of samples, with higher relative abundances in seep samples than the references. Genes for nitrogen fixation were predominantly found in the seep sites, whereas the reference sites showed a dominant signal for anaerobic ammonium oxidation (anammox). We recovered 49 metagenome-assembled genomes and assessed the microbial functional potentials in both types of samples. By this genome-based analysis, the seep samples were dominated by ANME-1 archaea and SRB, with the capacity for methane oxidation coupled to sulfate reduction, which is consistent with the 16S rRNA-gene based characterization. Although ANME-1 archaea and SRB are present in low relative abundances, genome bins from the reference sites are dominated by uncultured members of NC10 and anammox Scalindua , suggesting a prevalence of nitrogen transformations for energy in non-seep pelagic sediments. This study suggests that hydrocarbon seeps can greatly change the microbial community structure by stimulating nitrogen fixation, inherently shifting the nitrogen metabolism compared to those of the reference sediments.",
"conclusion": "Conclusion Our study revealed the responses of subseafloor microbial communities to hydrocarbon seepage in deepwater GOM, by employing both gene- and genome-based analysis of metagenomes. Particularly, based on 16S rRNA gene analysis, community structure was revealed to be significantly altered, with higher fractions of 16S rRNA gene and the notable dominance of ANME archaea in the seep sites. Further, functional gene analysis suggested that anaerobic methane oxidation and nitrogen fixation, both presumed to be performed by ANME archaea, were intensified by hydrocarbon seepage, while nitrogen transformation such denitrification and anammox are dominant in the reference sediments. These metabolic features predicted in gene-based analysis were confirmed by genomic analysis, with high relative abundances of ANME archaea in the seep sites and the dominance of denitrifying Woeseia and anammox Scalindua in the reference sites. Collectively, this study revealed that natural seafloor hydrocarbon seepages in the deep ocean can alter the microbial community structure and change the dominant redox regime from nitrogen loss for energy production to sulfate-based methane oxidation and nitrogen fixation.",
"introduction": "Introduction Abundant microbial cells are found in marine sediments beneath the ocean, accounting for 1/3 rd to 1/20 th of total microbial biomass on the planet 1 . Microbial cells in sediments beyond the bioturbation zone are thought to experience extremely low energy availability 2 , and persist in a maintenance state without proliferation for at least tens of years 3 . At a sediment site with a near-constant sedimentation rate (i.e. under a steady state), microbial biomass follows a power-law decreasing trend with depth 1 , 4 , presumably due to the decrease of energy availability 2 . The responses of sedimentary microbial communities to environmental changes was documented in Peru Margin sediments, where cell increases were seen in relation to the sulfate-methane transition zones 5 , 6 . Subseafloor microbes were also observed to uptake amended substrates and proliferate in laboratory incubations 7 , 8 . However, their responses to other environmental disturbances and energy pulses have not been well studied. Natural seafloor hydrocarbon fluid and gas seepage is a prevalent process on the continental margin 9 and pelagic area 10 of the Gulf of Mexico (GOM), and could serve as an important environmental influence on the local sedimentary microbial community. These seep fluids are enriched in carbon and can markedly alter the microbial community structure and function 11 , 12 . Metagenome sequencing has been previously performed to better assess the microbial response to petroleum seeps in GOM 13 . However, most of the analyses performed focused on functional genes 14 , 15 rather than genomes. Microbial genomes in hydrocarbon-impacted sediments were reported in the Guaymas Basin 16 and compared to the nearby reference sites 17 , to elucidate the changes in metabolic functionality and dependencies caused by hydrocarbon seepages. In Guaymas Basin sediments, despite contrasting community compositions, similar overall community functionality was observed between hydrothermal and non-hydrothermal sediments 17 . Functional redundancy of the microbial communities was invoked to explain this observation 17 . In this study, we collected sediment cores from multiple sites with and without natural hydrocarbon seepages in the GOM to investigate the impacts of this environmental disturbance to microbial communities. We focus on the subseafloor sediments to test the hypothesis that the seep-related communities persist at low metabolic rates in the deep sediments and are insensitive to hydrocarbon seepages, as had been previously suggested for GOM sediments 11 . Here we employ both gene- and genome-based approaches to characterize the microbial community composition and structure, overall metabolic functionality, and inter-dependencies by comparing the metagenomes from seep and reference environments. With our genomic and functional insights into the microbial communities inhabiting the two contrasting geochemical conditions, we see that in addition to microbial cycling of sulfate and methane, nitrogen cycling dynamics must be severely shifted between seep and reference sediments.",
"discussion": "Discussion We performed a metagenomic study focusing on subseafloor sediments from two contrasting habitats in the Gulf of Mexico to examine the microbial responses to seafloor hydrocarbon seepage. We used five samples from sites bearing natural hydrocarbon seepage and two from nearby reference sites without detectable seepage. Based on gene- and genome-centric analyses, this study reveals that hydrocarbon seepage can significantly alter the community structure. Although metagenome-assembled genomes are helpful to provide information of microbes in natural environments beyond their identities 16 , 17 , 54 , 55 , complete microbial genomes are rarely reconstructed from natural environments 56 . Importantly, reconstructed genomes often lack 16S rRNA 57 , which is a robust phylogenetic marker and is critical to constrain the taxonomy of novel genomes on the basis of existing extensive 16S rRNA gene sequences 23 . Also, the assembly-binning approaches typically only recover members of a community that have sufficient coverage, minimal population heterogeneity, few sequence repeats and distinct genome nucleotide composition 58 . As a result, current genome-centric approaches do not capture the full phylogenetic and functional diversity of a microbial community, particularly in complex environments. These deficiencies have led to the emergence of gene-centric analysis pipelines such as phyloFlash 23 and graftM 24 , which bypass the bottleneck of assembly of complex communities and can provide a comprehensive view of diversity of the 16S rRNA gene and other functional genes based on well-curated reference databases. Combining our gene and genomic analysis results, we offer detailed insights into the subsurface community. Microbial community at site D24 showed mixed features of both seep and reference sites, although this sample was collected within Seep-1. Most of the analyzed microbial features such as the DNA yield, 16S rRNA gene and mcrA abundances assessed by qPCR (Fig. 2 ), ratio of 16S rRNA gene in the metagenome (Fig. 3a ), abundance of ANME archaea (Fig. 3b ), and functional gene ratios of hydrogenase and nitrate reductase (Fig. 4a ), was lower than the seep sites but higher than the two reference sites. Considering that hydrocarbon seepages in GOM are ephemeral and can migrate over time 59 , it is very likely that the duration time of seepage at this site was shorter than the rest seep sites. Therefore, we excluded this site for the MAG comparison analysis. High fractions of ribosomal RNA in the seep sites suggesting active growth We detected significantly higher fractions of the 16S rRNA gene in the metagenome at the seep sites than the reference sites. Microbes with different growth rates tend to have different copies of ribosomal RNA operons ( rrn ), and microbes exhibiting higher reproductive rates were revealed to have higher rrn copy numbers 60 . Therefore, the copy number of genes encoding the ribosome has been suggested as a predictor of both growth rate and carbon use efficiency because of a proteome allocation trade-off 60 . The different overall microbial community structures, especially the markedly different proportions of archaea observed between the two types of sites, could explain the different fractions of 16S rRNA gene in the total communities. However, archaea are known to have fewer copies of rrn per genome than bacteria 61 and therefore are not likely to cause the higher fractions of 16S rRNA genes detected in the metagenomes of the seep sites. The faster proliferation rates of microbes stimulated by energy-rich substances emitting from the subsurface seepages is the most plausible explanation of the higher fractions of rrn in the metagenomes of the seep sites. Faster growth rates at the seep sites could also explain the higher microbial abundances detected by qPCR measurements (Fig. 1 ). Microorganisms in marine sediments beyond the bioturbation zone are always assumed to persistent in a maintenance state with turnover time of hundreds of years; 62 , 63 our data, however, reveal that subseafloor microbes can respond to and experience significant growth due to the impact of hydrocarbon seepages. Whether the subseafloor microbes are still undergoing growth or not is unclear. Experimental approaches such as H 2 18 O-labeling 64 , 65 are necessary to apply to marine sediments to confirm the proposed microbial in situ growth here and elsewhere 19 . The overall community structure and functionality differences between the two types of sites The overall microbial community structure was significantly changed by the presence of active hydrocarbon seepage (Table 2 ; Fig. 3 ). The notable dominance of ANME archaea detected at the seep sites (~50% of total) can be explained by the growth stimulated by methane in hydrocarbon seepage. It is worth noting that such archaeal dominances were not detected in the seep sites by qPCR of 16S rRNA genes, which is likely attributed to the known primer biases against archaea 66 , 67 . In the reference sites Planctomycetes (mainly anammox Scalindua ), NC10, Acidobacteria, Alphaproteobacteria and Gammaproteobacteria were more frequently detected, suggesting that these groups are well adapted to the typical pelagic sediment environment found in the reference sites 68 . The relative abundances of Deltaproteobacteria, Latescibacteria, and Zixibacteria are similar between the two types of sediments, indicating these are part of the core microbiome in the sediments and are not affected by the hydrocarbon seepage. We also assessed the potential functional differences between the two types of environments based on the functional gene contents in the metagenomic datasets. We detected notably high fractions of mcrA and nifH genes in the seep sediment metagenomes (Fig. 4c ), both of which are largely derived from ANME archaea 69 , 70 . ANME archaea were >45% of the taxa detected in the seep sites while virtually absent in the reference sites (Fig. 3 ). ANME archaea at the seep sites could have exhibited significant growth under the stimulation of hydrocarbon seepage, supported by the 20-fold higher mcrA gene abundances relative to the reference sediments (Fig. 2c ). The nitrogen fixation capacity of ANME archaea could provide ammonia as the nitrogen source for biomass synthesis during the microbial proliferation in the seep sites. The bloom of ANME archaea have resulted in higher methane oxidation rates previously noted in GOM cold seep sediments compared to reference sediments 71 . The increase in the relative abundance of archaea was also observed in the hydrothermal sediments in Guaymas Basin 17 . We also detected significantly higher fractions of mtrA gene in the metagenome datasets at the seep sites, the function of which was suggested as iron oxidation 72 , suggesting that iron oxidation could be more strongly affected at the seep sites than the reference ones. We detected notably higher fractions of the functional genes for denitrification ( narG , nirK and nosZ ), anammox ( hzo gene), and aerobic methane oxidation (mxaF/xoxF gene) in the reference sites, suggesting that these processes are more prevalent in typical pelagic sediments. This is consistent with the fact that the genomes of Woeseia and NC10, both capable of nitrate/nitrite reduction, and anammox Scalindua were mainly or exclusively recovered in the reference sites by metagenome assembly and binning (Fig. 6 ). Interestingly, in the reference sites we also detected higher abundances of the mxaF/xoxF genes, which are involved in aerobic methane oxidation 53 . This suggests that methane consumption in the reference sites is largely coupled to oxygen respiration, and that oxygen may be more steadily available than in the seep sites. This increase in oxygen availability in the sediments is likely due to lower overall microbial activities than those in the seep sediments under the stimulation of reduced substance fluxes. Functional genes of carbon fixation (RuBisCo), sulfate reduction ( sat gene), and hydrogen oxidation (NiFe- and FeFe-hydrogenase) are not statistically different between the two types of sediments. This observation suggests that microorganisms responsible for these processes are core members of the microbial communities in these deep-sea sediments and are not affected by hydrocarbon seepages. This is consistent with the similar amounts of Deltaproteobacteria (mainly Desulfobacterales) detected in both environments by 16S rRNA gene-based analysis (Fig. 3b ) since Desulfobacterales have been suggested to play an important role in H 2 consumption and sulfate reduction in marine sediments 73 . Metabolic dependency between different groups in the two contrasting environments To better understand the biogeochemical impacts of hydrocarbon seepage on sedimentary microbial communities in the Gulf of Mexico, we examined the major ecological roles of the dominant genomes in the two contrasting types of sediments (Fig. 7 ). In the seep sites, Chloroflexi (mainly Anaerolineae) and Zixibacteria are the major organic matter degraders that are capable of degrading organic matter to low-molecular weight (LMW) carbon compounds such as acetate and ethanol 16 , 17 , 74 . These LMW compounds can be used as the electron donor by sulfate-reducing Desulfubacterales. In addition to using sulfate as the electron acceptor, Desulfubacterales may also form consortia with ANME archaea and accept electrons from ANME archaea when performing sulfate-dependent methane oxidation under the stimulation of the presumed upward flux of methane from the subsurface. In addition, the higher relative abundances of Bathyarchaeota in the seep sites (Fig. 6 ) may be due to their capability of aromatic compound degradation 75 and could contribute to the hydrocarbon degradation in GOM seep sediments. The co-enrichment of ANME and Bathyarchaeota has also been reported in the hydrothermal-influenced biofilm on the chimney wall of the Soria Moria vent field 76 , supporting that these two archaeal groups may have similar responses to environmental impacts. Figure 7 Major geochemical processes and the associated microbial genomes in the seep (upper) and reference (lower) sites of GOM. Only metabolic pathways of the dominant microbial genomes in the two contrasting sediment sites are shown. Arrows represent metabolic capabilities that were inferred in the MAGs reconstructed from GOM deep-sea sediments by genome annotation. Taxa enriched in the seep and reference sites are shown in orange and blue, respectively, while the common taxa in both sites are shown in black. OM: organic matter. The initial hydrolysis and fermentation of organic matter in the reference sediments is likely performed by the abundant taxa Chloroflexi and Zixibacteria (Fig. S2 ). The further degradation of organic matter could be carried out by putative denitrifying bacteria capable of using nitrate/nitrite as the electron acceptors, such as Woeseia 77 and NC10 bacteria 78 . Nitrite produced from nitrate reduction and ammonium produced from the degradation of organic matter could be consumed by microbes of the genus of Scalindua (Fig. 7 ), a specific taxon capable of anaerobic ammonium oxidation 79 . The dominance of these taxa in the recovered MAGs is consistent with higher fractions of denitrification and anammox functional genes detected in the reference sites, suggesting that nitrogen transformation for energy generation is prevalent in the typical pelagic sediments in GOM. Our results run counter to conclusions from the previously studied seeps in the Gulf of Mexico, where deeper communities were not seen to be greatly impacted compared to the surface-expression of the seeps 11 . Perhaps the removal of the surface comparison allows a more even comparison between deep sediments. However, our study also runs counter to the data from Guaymas Basin sedimentary seeps, where communities in-seep and outside-of-seep were found to be functionally redundant 17 . Guaymas Basin sediments may be under different nutrient limitations compared to Gulf of Mexico sediments, so nutrient influences on microbial communities may not be as readily seen there. Our work shows that seepage influences on the microbial community structure should be complexed with additional measurements to fully understand what limits or supports community development and geochemical cycling. Overall, the finding that nitrogen fixation is largely present at the seeps versus nitrogen loss (via denitrification and annamox) present at the reference sites (Fig. 4b ), suggests that the carbon use efficiency varies drastically between the two settings. While carbon and nitrogen dynamics are not often focused in the seafloor, the study of soils has shown that C and N balances greatly impact microbial processing 80 and nitrogen fixation has been documented in carbon enriched sites 12 . This contrasting nitrogen metabolism pattern in GOM sediments - the prevalence of nitrogen fixation at seep sites and nitrogen loss at non-seep sites - was also supported by 15 N isotope compositions of previous studies 71 . Microbial selection toward nitrogen-fixing bacteria in marine sediments has also been reported in laboratory microcosm containing hydrocarbon (crude-oil) 81 . Considering GOM seep sediments are considered organic-rich (3–6% weight percent 71 ) and nitrogen-poor (C/N ratio in the range of 20–40 in sediments <10 cm 71 ), our results indicate that the newly fixed nitrogen by diazotrophs is a prerequisite of the rampant microbial growth in seep sediment since N limitation may limit microbial growth. Meanwhile, our results show that outside seep environments, the deep pelagic sediments (>1,400 m water depth) may use nitrogenous compounds as an energy source. This investigation of seep and reference sediments from the Gulf of Mexico deep biosphere shows that sedimentary microbes may be limited by different factors in adjacent settings due to the bottom-up environmental perturbation of natural hydrocarbon seepage."
} | 5,148 |
38906896 | PMC11192741 | pmc | 8,779 | {
"abstract": "Lacustrine methane emissions are strongly mitigated by aerobic methane-oxidizing bacteria (MOB) that are typically most active at the oxic-anoxic interface. Although oxygen is required by the MOB for the first step of methane oxidation, their occurrence in anoxic lake waters has raised the possibility that they are capable of oxidizing methane further anaerobically. Here, we investigate the activity and growth of MOB in Lake Zug, a permanently stratified freshwater lake. The rates of anaerobic methane oxidation in the anoxic hypolimnion reached up to 0.2 µM d −1 . Single-cell nanoSIMS measurements, together with metagenomic and metatranscriptomic analyses, linked the measured rates to MOB of the order Methylococcales. Interestingly, their methane assimilation activity was similar under hypoxic and anoxic conditions. Our data suggest that these MOB use fermentation-based methanotrophy as well as denitrification under anoxic conditions, thus offering an explanation for their widespread presence in anoxic habitats such as stratified water columns. Thus, the methane sink capacity of anoxic basins may have been underestimated by not accounting for the anaerobic MOB activity.",
"introduction": "Introduction Methane (CH 4 ) is a powerful greenhouse gas, abundantly produced in marine and freshwater environments. Fortunately, its emissions are strongly mitigated by the activity of microorganisms that have the capacity to oxidize methane to CO 2 1 – 3 . These microorganisms, termed methanotrophs, are therefore considered a major biological methane filter. Both archaea and bacteria possess the capacity to activate and oxidize methane, and their relative contribution to methane removal varies in different aquatic habitats 4 – 6 . In anoxic marine sediments, the majority of methane is consumed by consortia of anaerobic methanotrophic archaea (ANME) and Deltaproteobacteria that couple anaerobic oxidation of methane (AOM) to sulfate reduction 5 , 7 . Additionally, anaerobic methane oxidation in marine and freshwater habitats can be linked to nitrate or nitrite reduction and is performed either by nitrate/nitrite-dependent ANME archaea 8 , 9 or by NC10 bacteria related to Candidatus Methylomirabilis oxyfera 10 – 14 . Aerobic methanotrophic activity is typically attributed to methane-oxidizing bacteria (MOB) belonging to the Alpha- or Gammaproteobacteria 15 . These bacterial methanotrophs activate methane via a soluble or particulate methane monooxygenase that catalyzes the initial oxidation of methane to methanol using molecular oxygen. Methane oxidation by NC10 bacteria is also mediated by a methane monooxygenase and, therefore, oxygen-dependent. However, these bacteria couple methane oxidation to denitrification via a unique nitrite-dependent methane oxidation pathway that produces oxygen intracellularly via nitric oxide dismutation 11 , which enables NC10 bacteria to thrive in anoxic environments 12 – 14 , 16 , 17 . With the exception of NC10 bacteria, bacterial methane oxidation is considered an obligate oxygen-dependent process constrained to oxic environments. However, it has been shown recently that various taxa of MOB differ in their oxygen requirements and inhabit distinct niches in the water column 18 , 19 . Gammaproteobacterial MOB (gamma-MOB) typically exhibit the highest activity under hypoxic conditions 20 , 21 and, therefore, thrive primarily at the oxic–anoxic interface in the water column or sediment, where they experience high fluxes of both methane and oxygen 22 . Interestingly, gamma-MOB belonging to the ubiquitous gammaproteobacterial order Methylococcales are also commonly detected in anoxic waters or sediments of stratified lakes and marine basins 18 , 23 – 32 . The abundance of gamma-MOB in anoxic waters is poorly understood because there is so far no mechanistic explanation for methane oxidation by these microorganisms in the complete absence of oxygen. It has therefore been proposed that even in anoxic environments, the activity of MOB might be linked to the periodic occurrence of trace amounts of oxygen (e.g. refs. 24 , 33 ), which are undetectable with current techniques 34 . For example, in shallow lakes, where light penetrates into anoxic water depths, aerobic methane oxidation can be sustained through in situ oxygen production by photosynthetic algae 27 , 35 . Lateral and/or vertical intrusions of oxic waters into anoxic depths were also proposed to sustain methanotrophic activity in situ 24 . Additionally, some gamma-MOB evolved specific mechanisms to minimize the overall oxygen demand of the cell by either effectively directing oxygen towards particulate methane monooxygenase (pMMO) activity 36 – 39 or by performing anaerobic respiration 31 , 33 , 40 – 42 . Denitrification, the stepwise reduction of nitrate or nitrite to N 2 or N 2 O, has been commonly invoked as a mechanism allowing for the survival of bacterial MOB under oxygen limitation. This metabolism has so far only been demonstrated experimentally for Methylomonas denitrificans cultures 41 , but many cultured and environmental gamma-MOB possess the genes for respiratory nitrate reduction 19 , 33 , 42 , 43 . Alternatively, cultured methanotrophs of the genus Methylomicrobium have been shown to conserve energy through fermentation-based methanotrophy rather than aerobic respiration or denitrification, thus representing another strategy to overcome oxygen limitation 38 . In this process, MOB switch to a novel fermentation mode at low oxygen tensions and convert methane to fatty acids (including formate, acetate, succinate, lactate, and hydroxybutyrate) and hydrogen, via primary oxidation of methane to formaldehyde and pyruvate, followed by fermentation of these compounds. This process would have important implications for lacustrine methane cycling in that it would allow methane carbon to be retained in anoxic waters in the form of microbial biomass 44 . Recent studies suggested a broad prevalence of this metabolism among gammaproteobacterial MOB in lake water columns based on metagenomic analyses 45 , 46 . Irrespective of the utilized energy-conserving pathway (anaerobic respiration or fermentation), in all cases, molecular oxygen is still thought to be required for the initial activation of methane via the pMMO. A widespread capacity of gamma-MOB for thriving anaerobically in anoxic environments would bear great significance for the in situ activity of this important microbial group. Yet, at this point, the occurrence and prevalence of gammaproteobacterial ‘anaerobic’ methane oxidation in the environment remains elusive. Here, we investigated the activity and growth of different aerobic gammaproteobacterial and Methylomirabilis -like methanotrophs in the water column of Lake Zug, a deep, permanently stratified freshwater lake in central Switzerland, using a combination of stable isotope incubations, single-cell imaging mass spectrometry, as well as metagenomic and metatranscriptomic analyses.",
"discussion": "Discussion In stratified lacustrine and marine water columns, aerobic methane oxidation at the oxic–anoxic interface is typically a major sink for upwards-diffusing methane and, thus, an important biological methane filter 18 , 28 , 30 , 32 , 47 . During our sampling campaign in September 2017, hypoxic incubations from the oxic–anoxic interface (123 m depth) displayed high potential rates of methane oxidation (up to 1.8 µM d −1 ). However, the rates were preceded by a ca. 2-day long lag phase, indicating a necessity for an adaptation period for the methanotrophic community, potentially due to low in situ activity at the time of sampling. Indeed, the in situ methane concentration profiles showed that very little methane diffused all the way up to the oxic–anoxic interface; instead, most methane was consumed already 10–15 m below, suggesting a likely sink for methane in the anoxic hypolimnion at and below 135 m water depth (Fig. 1a ). Correspondingly, high and linear rates of anaerobic methane oxidation (up to 0.2 µM d −1 ) with immediate onset could be detected in the anoxic incubations from both deep water depths (135 and 160 m; Fig. 1c ). For comparison, the anaerobic rates were 5 to 20-fold lower than rates from parallel hypoxic incubations with water from these same water depths, to which low concentrations of oxygen (ca. 10 µM O 2 ) were added (Fig. 1b ). Despite the evidence for ongoing, apparently anaerobic, methane oxidation, anaerobic methane-oxidizing archaea were not detected in the deep anoxic water depths. Instead, the methane-oxidizing community in the anoxic waters was dominated by aerobic gammaproteobacterial methanotrophs (gamma-MOB) (Fig. 1e ) as well as denitrifying methanotrophs of the NC10 phylum. This is in agreement with previous analyses of the methanotrophic community in this lake 12 , 28 . Morphologically, four distinct cell types could be distinguished among the gamma-MOB targeted with a specific FISH probe (Fig. 2a ), of these Crenothrix -like filaments 42 and small cocci (presumably Methylovulum -like MOB) were numerically least abundant and their cell numbers increased ( Crenothrix ) or remained constant (small cocci) throughout the water column (Figs. 1e and S 2 ; Supplementary Note 2 ). Most abundant were small rods, whose abundance decreased with increasing depth. In contrast, large rods—which, due to their large size, constituted a large proportion of the MOB biomass (Fig. S2 )—were more abundant below the oxic–anoxic interface (Fig. 1e ). The conspicuous morphology of the large rods detected by a gamma-MOB-specific FISH probe is strongly reminiscent of methanotrophs from the Methylobacter genus of the Methylococcales order. Cultured representatives of Methylobacter tundripaludum are typically large, rod-shaped and free-living 48 , 49 , although other morphologies have been observed for other species 46 . Environmental Methylobacter -related methanotrophs are frequently detected in hypoxic and anoxic freshwater environments 24 , 26 , 31 , 50 . Indeed, our molecular analyses confirmed the presence and abundance of Methylobacter spp. in all three incubation depths (Fig. 5b ). Multiple genomic bins assigned or related to Methylobacter spp. were retrieved from our samples, and their relative abundance increased with depth, supporting the prevailing opinion that members of this genus thrive in low-oxygen environments and might play a key role in methane removal there 29 , 31 . In our anoxic incubations that were setup to mimic the conditions in the anoxic hypolimnion (no added O 2 , added nitrate, excess methane), we consistently detected growth and activity of the Methylobacter -like MOB; in fact, these were the only MOB exhibiting substantial growth on methane in the absence of oxygen (Fig. 3 ). Surprisingly, the activity of these MOB—with regards to cell-specific 13 C methane assimilation rates (9.8 vs. 8.6 fmol 13 C cell −1 d −1 ) or 13 C-based growth rate (0.39 vs. 0.34 d −1 )—did not significantly differ with or without oxygen (Table S2 ), suggesting that their activity is not suppressed under oxygen-limiting conditions. This is in striking contrast to the other three MOB morphotypes, which exhibited comparable rates of growth and activity only when oxygen was added to the incubations. The putative Crenothrix -like filaments occasionally also exhibited high 13 C enrichment under anoxic conditions, although less consistently than the large Methylobacter-like MOB. However, Crenothrix -like MOBs were solely identified based on their filamentous cell shapes; it is therefore possible that not all of the measured filaments were indeed Crenothrix (Supplementary Note 3 ). These results suggest that while high diversity exists among methane-oxidizing bacteria under hypoxic conditions, distinct Methylobacter -like MOB are largely responsible for methane oxidation in the anoxic waters of Lake Zug. Importantly, this implies that the lower rates of anaerobic methane oxidation measured in the hypolimnion compared to the oxic–anoxic interface were a consequence of a lower number of MOB that remained active under anoxic conditions rather than a general decrease in cell-specific methanotrophic activity of MOB in the anoxic hypolimnion. It should be noted that using 13 C-methane assimilation as a proxy for activity precludes us from evaluating the potential contribution of Ca . Methylomirabilis-like bacteria in our samples. This is because these methanotrophs preferentially grow on inorganic carbon 51 and thus did not get enriched in 13 C in our incubations (Fig. 2b ). However, based on their uptake of 15 N-nitrate over time (Fig. S7 ) and their moderate abundance in the deep water (Fig. 1e ) we assume that they may also contribute to methane oxidation and denitrification in the Lake Zug hypolimnion, even under non-bloom conditions, similar to other lakes 14 . It remains to be elucidated why the Methylobacter -like MOBs are so notably more successful in anoxic environments than other MOB, as well as how their activity can be sustained in the absence of oxygen. Importantly, we do not rule out the presence of trace amounts of oxygen in our ex-situ bottle incubations, either originating from microbial production or oxygen contamination during sample handling (Supplementary Note 5 ). In fact, based on our current understanding of methane oxidation by bacteria, this trace oxygen is likely indispensable for the initial oxidation of methane to methanol by methane monooxygenase. In this regard it is interesting to note that also aerobic ammonium oxidizers, which use a related enzyme, ammonia monooxygenase, appear capable of utilizing traces of oxygen to sustain their activity in oxygen-deficient environments 52 . In the environment, oxygen is most likely introduced at low concentrations periodically through e.g. intrusions and mixing, which is to some extent a possible scenario also for Lake Zug 53 . Alternative biological mechanisms have been proposed, such as dark oxygen production by NC10 bacteria 11 , 54 and ammonium-oxidizing archaea (AOA) 55 , as well as oxygen production via water splitting 56 . However, the relevance of these processes as a potential oxygen source in anoxic environments remains to be determined. Alternatively, some methanotrophs appear to be capable of fermentation-based methanotrophy under low oxygen conditions, a process, in which methane carbon is converted into fatty acids, with little biomass synthesis 38 . We indeed found that all MOB bins detected in our lake metagenomes, except bin 7, encoded as well as transcribed several fermentation genes, including those of the mixed acid fermentation pathway (Fig. 5b ). This is in agreement with the observations that many bacterial methanotrophs possess the genetic repertoire for fermentation, including lake taxa 19 . Our transcriptome data support the notion that fermentation-based methanotrophy may indeed be employed in situ, preferentially in anoxic waters. The employment of the fermentation-based methanotrophy would minimize the oxygen requirements of the MOB while simultaneously leading to methane carbon being retained in the low-oxygen waters in the form of microbial biomass instead of being oxidized to CO 2 . Indeed, our analyses show that methane carbon was efficiently assimilated into microbial biomass, likely via direct assimilation of methane by proteobacterial methanotrophs as well as uptake of excreted fermentation products by non-methanotrophs. Interestingly, in contrast to the proteobacterial MOB, we did not detect substantial uptake of 13 CH 4 -derived carbon by the anaerobic Ca . Methylomirabilis methanotrophs (Fig. 2b ), in agreement with their proposed growth on inorganic carbon 51 . These data show that despite their high abundance and proposed activity in anoxic waters, Ca . Methylomirabilis-like bacteria do not contribute to the retention of methane carbon in the lake hypolimnion by direct assimilation. Methane carbon retention seemed to be particularly pronounced in the anoxic depths, where rates of methane carbon assimilation matched or even exceeded methane oxidation to CO 2 (Fig. 1b, c ; Table S1 ). These data strongly suggest that ‘fermentation-based methanotrophy’ occurred under the investigated conditions, thus enhancing the retention of methane carbon in the hypolimnion. Additionally, the production of volatile fatty acids may, in turn, support the growth and activity of mixo- and heterotrophic microbial communities. It should be noted that the partitioning of methane-derived carbon between biomass assimilation and complete oxidation to carbon dioxide may depend on the environmental conditions as well as the composition of the methanotrophic community. Indeed, even in Lake Zug, the average proportion of assimilated methane carbon varied between years, with methane carbon assimilation contributing less to total methane turnover in October 2018 and May 2019 than in September 2017. In any case, our data provide first indications for the environmental implications of the proposed fermentation process that bears great relevance to our understanding of the cycling of methane-derived carbon in anoxic environments. Additionally, our results suggest that a high proportion of methane carbon can be assimilated in addition to being oxidized to carbon dioxide under hypoxic and anoxic conditions (Fig. 6 ). Therefore, the methane sink capacity of anoxic basins may be underestimated, with up to 60% of methane carbon being retained in the anoxic hypolimnion in the form of biomass. Fig. 6 Schematic overview of microbial methane oxidation in the anoxic hypolimnion of Lake Zug. The left panel depicts representative water column profiles of oxygen and methane. The right panel illustrates the MOB groups, comprising gammaproteobacterial MOB (filaments, large rods, small rods, cocci) as well as anaerobic NC10 bacteria and their respective contribution to methane oxidation and assimilation at and below the oxic–anoxic interface. Note that the abundance, activity, and cell size of the different gamma-MOB groups are reflected by their respective numbers, color shading, and size in the figure. In the presence of oxygen, all four distinct Methylococcales-related gamma-MOB groups showed methane-dependent growth at comparable rates, whereas under apparent anoxia, only one group, the large rod-shaped gamma-MOB, was persistently active. Based on this, we conclude that under hypoxic conditions, all identified MOB groups contribute to methane oxidation, while large rod-shaped MOBs, as well as presumably NC10 bacteria, are responsible for the observed methane oxidation under anoxic conditions. Under anoxia, the assimilation of methane carbon into biomass exceeded the oxidation to carbon dioxide in this sampling campaign (see also Fig. S1 ). We envision that aerobic methane oxidation by gamma-MOB under apparent anoxia can be sustained by traces of oxygen that are periodically mixed into the anoxic waters (green arrows), in combination with anaerobic processes such as denitrification and/or fermentation that allow for energy conservation independently of oxygen. This image was created with Adobe Illustrator. In addition to fermentation, some aerobic MOBs have been proposed to be capable of conserving energy through nitrate or nitrite respiration 41 . Most cultured gamma-MOB encode some genes for respiratory nitrate/nitrite reduction in their genomes 43 , which enable them to switch to anaerobic respiration when oxygen becomes limiting 42 , 57 , 58 . Environmental data also support a link between methane oxidation and denitrification, e.g. nitrogen loss in an Indian freshwater reservoir was enhanced by high methane concentrations 59 . We have compelling genomic evidence for denitrification potential within the Lake Zug MOB community, as denitrification genes ( narGHJI , napAB , nirKS , and norBC ) were found in the majority of pmoA -containing bins affiliated with the Methylococcales order (Fig. 5b ). Of these, bin 4 (here classified as UBA4132 clade) closely affiliates with Crenothrix sp., which were shown to be capable of methane-dependent growth in anoxic, denitrifying conditions previously 42 . Notably, bins 5, 8, and 10 that belonged to the Methylobacter A clade also contained a combination of three different denitrifying enzymes. Members of the Methylobacter genus have been shown to encode genes for partial denitrification previously 31 , 57 , 60 . Interestingly, even though many gamma-MOB possessed and expressed genes related to denitrification (as well as fermentation-based methanotrophy), methane-dependent growth under apparent anoxia could only be observed for one discrete MOB morphotype (i.e. large rods). More cultivation-based as well as cultivation-independent studies are still needed to better understand the prevalence of anaerobic metabolisms of this environmentally relevant group of microorganisms. The measured denitrification rates (to 30 N 2 ) in our anoxic, nitrate-supplied incubations (Fig. 1d ) were high enough to support methane oxidation at both anoxic depths, assuming a stoichiometric ratio of 5:8 for methane (oxidation to CO 2 ) and nitrate (reduction to N 2 ) 61 . However, no direct link between denitrification and methane oxidation rates could be shown experimentally, likely due to the large contribution of heterotrophic organisms to denitrification. It should be noted that to date, Ca . Methylomirabilis spp. are the only methanotrophs, which have been shown to denitrify all the way to N 2 62 . Methane-oxidizing gammaproteobacteria, on the other hand, notoriously lack the capacity to reduce N 2 O further to N 2 , and the nosZ gene encoding for N 2 O reductase was consistently absent from our MOB bins. As such, the production of labeled N 2 O (both as 45 N 2 O and 46 N 2 O) in our incubations (Fig. S4 ) might include a contribution from methane-oxidizing bacteria. The non-linear nature of the N 2 O accumulation throughout the experiment is indicative of rapid turnover of the produced N 2 O to N 2 by N 2 O-reducing (non-methanotrophic) bacteria. The sustained activity of Methylobacter -like MOB under anoxia indicates that these microorganisms efficiently remove methane not only at oxic-anoxic interfaces but also at anoxic water depths that contain nitrate and might periodically receive oxygen. Interestingly, the calculated per-cell methane oxidation rates of Methylobacter -like MOB under anoxic conditions (ca. 9.1 fmol CH 4 cell −1 day −1 ) (Supplementary Note 6 ) exceed the reported cellular rates of ‘true’ anaerobic archaeal and bacterial methanotrophs (ANME groups and NC10) making the ‘aerobic’ MOB very efficient ‘anaerobic’ methane oxidizers. This ecological success may be more pronounced in perturbation-prone anoxic water columns compared to e g. diffusion-driven anoxic and sulfidic sediments, as facultative aerobic bacterial methane oxidizers, may better cope with dynamic environments. Collectively, our findings highlight the need for a deeper understanding of the ecological role of gamma-MOB in anoxic environments. Recently, the importance of aerobic methanotrophy for lacustrine carbon cycling was documented 63 and our results now accentuate the importance of this process also under anoxic conditions. With the increasing occurrence of seasonal or permanent anoxia in lakes, the importance of methanotrophy in methane removal within lacustrine systems can be expected to intensify, thus making a significant contribution to greenhouse gas mitigation and carbon storage in future scenarios."
} | 5,919 |
34969854 | PMC8740591 | pmc | 8,780 | {
"abstract": "Significance Landscapes are spatially heterogeneous, and the measurement of spatial pattern is dependent upon observation scale. Understanding plant populations requires assessing their extrinsic interactions with the environment as well as intrinsic biological processes and has been difficult because of the inability to track both plant abundance and health on appropriate scales. We introduce remote sensing observations that assess the abundance and health of giant kelp, an important ecosystem-structuring species, over regional and local scales. We find that both extrinsic nutrient availability and intrinsic senescence processes regulate population dynamics but on regional and local scales, respectively. This suggests that future satellite missions will be able to assess plant abundance and health and their interactions with the environment on local to global scales.",
"discussion": "Discussion The regional scale relationship between frond Chl:C and seawater temperature ( Fig. 1 ) has implications for the growth and persistence of giant kelp forests. Field studies show that Chl:C is positively related to seawater nitrate concentrations up to 1 μM and is positively correlated with changes in canopy biomass in southern California ( 34 ). This study also shows that declines in canopy Chl:C led to large reductions in net primary production in the following months ( 34 ). The observed decline in regional scale Chl:C begins between 14 and 15 °C, the temperature range associated with 1 μM seawater nitrate concentration, and reaches a minimum at 19 °C, when seawater nitrate concentrations are typically below detection ( SI Appendix , Fig. S1 ). Enhanced physiological condition related to nutrient supply is linked to increased growth rate and production ( 33 , 34 ) and is likely related to stable recruitment, as reproductive structures are proportionally largest during cool, and thus nutrient-rich, seawater conditions ( 37 ). Mean regional scale Chl:C retrievals during the spring upwelling season were positively related to multidecadal kelp forest persistence ( SI Appendix , Fig. S4 ). This link between upwelling season Chl:C and decadal kelp persistence implies that nutrient availability underlies regional scale kelp persistence patterns. The local scale variations in Chl:C at times exceeded the range of regional scale Chl:C observed across the entire study area ( Fig. 2 D ). Nutrient-rich waters in the western SBC are rapidly exchanged through large kelp forests, as nutrient concentration observations made within and outside giant kelp forests are similar (refs. 38 and 39 ). Previous work in other regions has suggested that giant kelp nutrient uptake rates may be affected by ambient currents and mixing ( 40 , 41 ). In the SBC, current speeds along the mainland coast typically increase in the late spring and summer months ( 42 ) and coincide with a reduction in seawater nutrient concentrations ( Fig. 2 B ). Thus, it is unlikely that fine scale hydrodynamics have a large influence on the physiological condition patterns observed. Together, this implies that the local scale variability in canopy Chl:C observed across a single image ( Fig. 2 D and SI Appendix , Fig. S5 ) was too large to be driven by local scale nutrient limitation, suggesting an alternate driver for local scale variations in Chl:C. Local scale Chl:C was negatively associated with age across both single and multiple image dates with lower values found in older canopies ( Fig. 2 C – H ). The strong negative relationship between Chl:C and canopy age implies that intrinsic biotic processes, such as senescence, provide an explanation of local scale kelp population dynamics ( Fig. 3 A ). A synthesis of kelp frond cohort survivorship observations ( 9 ) showed that frond loss rate reaches a maximum at an age of ∼100 d simultaneous with Chl:C reaching its minimum ( Fig. 3 A ). This pattern is consistent with our field observations of multiple frond cohorts across the spring and summer in which chlorophyll concentration decreased dramatically at ∼50 d irrespective of ambient nitrate concentrations ( Fig. 3 B and C ). The sudden nonlinear decline in Chl:C with increasing age displayed the expected survivorship pattern of a senescing population ( 8 ) indicating that local scale reductions in physiological condition are associated with frond senescence and not nutrient availability. The local scale dynamics of Chl:C reveal frond demographic patterns that presage canopy decline. Seasonally increased seawater nitrate concentrations promote recruitment and frond elongation to form a surface canopy, which presents a uniformly elevated Chl:C as the forest is dominated by young, vigorously growing fronds ( Fig. 2 C , F , and I ; refs. 22 , 43 , and 44 ). Over the following months, new fronds are initiated, and the canopy develops into a mixture of growing and senescing fronds, depressing the average Chl:C in older areas, while new individuals bearing younger fronds emerge along the edges of the canopy ( Fig. 3 D , G , and J ). The microscopic stages of giant kelp have high light requirements and are vulnerable to intraspecific competition through shading ( 22 , 45 ). In a fully developed kelp canopy, the edges have higher benthic light levels enabling the growth of new recruits ( 45 , 46 ). As warm water is advected into the SBC during the summer and kelp becomes nitrogen limited, the regional scale physiological condition of the canopy decreases, and frond initiation is reduced ( 34 ). The lack of new fronds replacing fronds lost due to senescence further decreases the mean Chl:C of the canopy ( Fig. 2 E , H , and K ). Canopy fronds make the largest contribution to production ( 47 ), simultaneously translocating photosynthate to promote the growth of subsurface juvenile fronds ( 48 ). The reduction of photosynthetic performance in aging fronds inhibits frond initiation, preventing new frond cohorts from reaching the surface and ultimately leading to a loss of canopy ( 35 , 48 ). Previous studies of giant kelp canopy dynamics have reported increasing stochasticity at scales less than ∼1 km, indicating that other ecological phenomena, such as mortality, dispersal, or recruitment, may be responsible for local abundance patterns ( 49 ). Cavanaugh et al. ( 25 ) described an exponential decrease in southern California kelp population synchrony within distances of 200 m that corresponded to the spatial patterns of recruitment. Giant kelp exhibits episodic reproductive synchrony ( 50 ), and the high light requirements of microscopic stages ( 22 ) favor recruitment at the edges and away from established canopy. Deeper areas of the reef may be provided with elevated summertime seawater nutrients from internal waves, fueling the growth and eventual canopy formation of the late season cohorts ( 39 , 51 ). Spatially synchronous recruitment of kelp sporophytes may result in the synchronous decline in physiological condition and loss of canopy fronds later in the year, exposing the scales of these benthic processes to detection by aerial imagery. In kelp systems in which recruitment of sporophytes is spatially asynchronous, these patterns in senescence would be difficult to observe at the scales used in this study. Although local scale giant kelp dynamics are primarily governed by senescence, frond growth is fundamentally linked to extrinsic environmental drivers. The physiological condition of newly matured fronds is a function of the external nutrient and light environment and is positively related to frond initiation, net primary production, and biomass accumulation ( 34 ). Elevated seawater nitrate concentrations lead to vigorous growth in the spring and early summer, producing dense kelp canopies that reach maximum biomass shortly after seawater nitrate concentrations fall ( Fig. 4 A – D ; ref. 29 ). Since the life span of kelp fronds is generally unaffected by seawater nitrate conditions ( 9 ), reef areas that produce canopy earlier in the year will possess a greater proportion of older fronds and lose canopy at a faster rate in the subsequent months, especially in the absence of wave disturbance ( Fig. 4 B , E , and F ). Variable spatial patterns in frond demographics can lead to high standing biomass months after optimal growing conditions have passed, making it difficult to directly link regional nitrate conditions to local scale kelp dynamics ( 52 ). This is of particular importance, as marine heat waves and associated low-nutrient conditions have been implicated in the decline of kelp forests globally ( 53 ) with spatially diverse resilience patterns ( 54 ). While temperatures known to be deleterious to giant kelp were not observed during the hyperspectral time series at our focal site, significantly elevated temperatures were observed after the data shown in Fig. 2 and were associated with a complete loss of canopy. While the use of satellite imagery for monitoring kelp canopy dynamics is extensive, pixels are often binned at the regional scale, overlooking local scale dynamics captured by the imagery ( 55 ). Future studies should take advantage of this fine scale resolution to explore local patterns of resistance and resilience of these systems to high-temperature events. Teasing apart the roles of the external environment and intrinsic biotic factors will aid in understanding plant population dynamics across ecosystems. Spatiotemporal estimates of physiological condition may benefit studies of lowland tropical forests and semiarid drylands due to the presence of many perennial evergreen species ( 15 ). Observations of plant physiology across scales will likely reveal spatial patterns of enhanced growth, stress, impending declines, and reproductive processes, such as inflorescence, missed by existing satellite sensors. Nascent repeat, global-scale imaging spectrometers open the possibility of simultaneously studying the physiology and abundance of populations, uncovering processes and patterns at a variety of scales, and giving ecologists a suite of powerful tools for disentangling the intrinsic and extrinsic factors regulating plant population dynamics."
} | 2,550 |
22936780 | null | s2 | 8,782 | {
"abstract": "Relative to the atmosphere, much of the aerobic ocean is supersaturated with methane; however, the source of this important greenhouse gas remains enigmatic. Catabolism of methylphosphonic acid by phosphorus-starved marine microbes, with concomitant release of methane, has been suggested to explain this phenomenon, yet methylphosphonate is not a known natural product, nor has it been detected in natural systems. Further, its synthesis from known natural products would require unknown biochemistry. Here we show that the marine archaeon Nitrosopumilus maritimus encodes a pathway for methylphosphonate biosynthesis and that it produces cell-associated methylphosphonate esters. The abundance of a key gene in this pathway in metagenomic data sets suggests that methylphosphonate biosynthesis is relatively common in marine microbes, providing a plausible explanation for the methane paradox."
} | 223 |
26663423 | PMC5021209 | pmc | 8,783 | {
"abstract": "Summary Warm fluids emanating from hydrothermal vents can be used as windows into the rocky subseafloor habitat and its resident microbial community. Two new vent systems on the M id‐ C ayman R ise each exhibits novel geologic settings and distinctively hydrogen‐rich vent fluid compositions. We have determined and compared the chemistry, potential energy yielding reactions, abundance, community composition, diversity, and function of microbes in venting fluids from both sites: P iccard, the world's deepest vent site, hosted in mafic rocks; and V on D amm, an adjacent, ultramafic‐influenced system. V on D amm hosted a wider diversity of lineages and metabolisms in comparison to P iccard, consistent with thermodynamic models that predict more numerous energy sources at ultramafic systems. There was little overlap in the phylotypes found at each site, although similar and dominant hydrogen‐utilizing genera were present at both. Despite the differences in community structure, depth, geology, and fluid chemistry, energetic modelling and metagenomic analysis indicate near functional equivalence between V on D amm and P iccard, likely driven by the high hydrogen concentrations and elevated temperatures at both sites. Results are compared with hydrothermal sites worldwide to provide a global perspective on the distinctiveness of these newly discovered sites and the interplay among rocks, fluid composition and life in the subseafloor.",
"introduction": "Introduction Deep‐sea hydrothermal vents are energy‐rich environments where phylogenetically and physiologically diverse microbial communities thrive both above and below the seafloor. Microbial communities are especially enriched in regimes where oxygenated seawater mixes with reduced hydrothermal fluids, thus providing abundant sources of oxidants and reductants for microbial metabolism in a thermal regime that can support life (Kelley et al ., 2002 ). These mixed fluids, termed diffuse vents, offer a window into microbial communities beneath the seafloor (Huber and Holden, 2008 ), and have previously been found to host diverse bacterial and archaeal populations (Takai and Horikoshi, 2000 ; Huber et al ., 2007 ; Perner et al ., 2007 ; Nunoura and Takai, 2009 ; Akerman et al ., 2013 ; Anderson et al ., 2013 ). The population structure and functional repertoire of these microbial communities are intimately linked to the physical and chemical environment, both of which, in turn, are directly influenced by the geologic setting. Most previous microbial characterizations of subseafloor communities have focused on mafic (basalt)‐hosted systems. The discoveries of diverse ultramafic (peridotite)‐influenced hydrothermal systems, as well as back‐arc and other types of deep‐sea volcanism, have expanded the known range of geologic, physical and geochemical conditions that support microbial life (Takai et al ., 2004 ; 2008 ; Kelley et al ., 2005 ; Brazelton et al ., 2006 ; Perner et al ., 2007 ; 2010 ; Nunoura and Takai, 2009 ; Huber et al ., 2010 ; Takai and Nakamura, 2010 ). Microbial characterization of the few known ultramafic sites shows that microbial assemblages are compositionally and functionally distinct when compared with nearby and distant mafic‐hosted counterparts (Schrenk et al ., 2004 ; Kelley et al ., 2005 ; Nercessian et al ., 2005 ; Brazelton et al ., 2006 ; 2011 ; 2012 ; Perner et al ., 2007 ; 2010 ; Flores et al ., 2011 ; Roussel et al ., 2011 ). Commonly detected organisms in these ultramafic sites include abundant methane‐ and hydrogen‐metabolizing chemolithoautotrophs, which are often present in relatively lower abundance in mafic‐hosted systems (Flores et al ., 2011 ; Ver Eecke et al ., 2012 ). It is believed that the difference in microbial population structure and function between mafic and ultramafic sites reflects the high concentrations of dissolved hydrogen and carbon in hydrothermal fluids at ultramafic‐hosted systems (McCollom and Shock, 1997 ; Takai et al ., 2004 ; McCollom, 2007 ; Amend et al ., 2011 ; Ver Eecke et al ., 2012 ). Hydrogen is particularly important because it can be used to reduce ferric iron, sulfur species, oxygen and nitrate, for example, and hydrogen‐driven microbial ecosystems have been described both in terrestrial habitats (Stevens et al ., 1995 ; Chapelle et al ., 2002 ; Brazelton et al ., 2013 ) and the marine environment at deep‐sea hydrothermal vents (Takai et al ., 2004 ; 2006 ; Kelley et al ., 2005 ; Nealson et al ., 2005 ; Brazelton et al ., 2012 ; Ver Eecke et al ., 2012 ; Schrenk et al ., 2013 ). Thermodynamic studies of hydrothermal fluids at both mafic and ultramafic sites also suggest a critical role for hydrogen and methane in structuring microbial communities at deep‐sea hydrothermal vents. In particular, thermodynamic models predict that ultramafic‐hosted systems are capable of providing almost twice as much energy as mafic‐hosted systems due to enrichment of fluids in hydrogen and methane (McCollom, 2007 ; Takai and Nakamura, 2010 ; Amend et al ., 2011 ). Located on Earth's deepest and slowest spreading mid‐ocean ridge, the study site is situated in the western Caribbean, where the axial rift valley floor occurs at a depth of ∼ 4200–6500 m ( Fig. S1 ; Kinsey and German, 2013 ). In 2009–2010, two novel hydrogen‐rich vent sites were discovered along the Mid‐Cayman Rise (MCR): Piccard , the world's deepest hydrothermal vent field, which is basalt‐hosted and situated at a depth of 4960 m; and Von Damm , an ultramafic‐influenced system that is located at a lateral distance of just ∼ 20 km from Piccard , but at a much shallower depth of 2350 m, near the summit of an oceanic core complex (German et al ., 2010 ; Connelly et al ., 2012 ; Kinsey and German, 2013 ). The Piccard hydrothermal field is hosted in a classically neovolcanic setting atop a volcanic spur, which comprised exclusively mounded pillow basalts. There is no geologic evidence for ultramafic rocks in the vicinity of the Piccard hydrothermal field, nor in any other comparable neovolcanic ridge‐axis settings, worldwide (Beaulieu et al ., 2013 ; Kinsey and German, 2013 ). What is novel about the Piccard hydrothermal field, compared with all other neovolcanically hosted hydrothermal fields, is its extreme depth: at 4960 m, it is more than 800 m deeper (hence, at higher pressure) than all previously reported mid‐ocean ridge hydrothermal fields (Beaulieu et al ., 2013 ). The end‐member vent fluids at Piccard are acidic (pH 3.2) with high sulfide (12.3 mM) and hydrogen (20.7 mM), but very little methane is present (0.13 mM), which is consistent with a basalt‐dominated site (McDermott, 2014 ; Reeves et al ., 2014 ). In contrast, the Von Damm site, located at 2350 m on an oceanic core complex, is thought to be influenced by reactions with ultramafic rocks (German et al ., 2010 ; Connelly et al ., 2012 ). No basalt has been found at this site (Connelly et al ., 2012 ; Bennett et al ., 2013 ); end‐member fluids show chemical compositions that are only slightly acidic (pH 5.6), with up to 19.2 mM hydrogen and 2.8 mM methane, consistent with the presence of ultramafic host rocks (Reeves et al ., 2014 ; McDermott et al ., 2015 ). Here, we report on the first microbiological characterization of the subseafloor microbial communities in fluids venting at each site using a combination of chemical measurements, energetic modelling, total cell and domain‐specific enumeration, targeted stable isotope tracing experiments, and 16S rRNA gene amplicon and metagenomic sequencing. With these data, we compare the potential energy available for microbial metabolism and the subseafloor microbial community diversity, function and activity at the two newly discovered vent fields to hydrothermal sites worldwide to provide new insights into subseafloor microbial communities at deep‐sea hydrothermal vents.",
"discussion": "Discussion The discovery of the Piccard and Von Damm hydrothermal systems not only adds to the range of physical and chemical conditions known to support subseafloor microbial life, but also allows us to test thermodynamic predictions and compare them with both phylogenetic and functional microbial data from the same mixed fluids used in our expanded modelling effort. Given the fact that we cannot directly sample the subseafloor environment in active, unsedimented hydrothermal systems, but instead are sampling a mixture of habitats and fluids at the seafloor, we used an integrative approach including chemical measurements, energetic modelling, stable isotope tracing and ‐omics data to infer the microbially mediated processes occurring beneath the seafloor. Previous thermodynamic modelling efforts have suggested that ultramafic systems will host more diverse microbial communities compared with their mafic counterparts due to higher concentrations of hydrogen and reduced carbon species (McCollom, 2007 ; Amend et al ., 2011 ). Our modelling results show that despite the large differences in host‐rock, depth, and end‐member fluid temperature and geochemistry between Von Damm and Piccard, the catabolic energies available for key redox reactions are very similar, with the only difference being more energy yield for methane‐based metabolisms at Von Damm. While mafic‐hosted systems with extremely high hydrogen concentrations were not included in previous modelling nor yet identified in field efforts, the present work supports the critical role of hydrogen in determining microbial community composition at deep‐sea hydrothermal vents in both mafic and ultramafic‐influenced hydrothermal sites. Moreover, it emphasizes the importance of temperature and pressure in regulating the composition of hydrothermal fluids that regulate chemical environments inhabited by microbial ecosystems and the inability to predict microbial community composition from rock host type alone. Our study also substantiates conclusions from previous modelling efforts (McCollom, 2007 ; Takai and Nakamura, 2010 ; Amend et al ., 2011 ; Nakamura and Takai, 2014 ) that ultramafic systems will host more diverse lineages and metabolic functions compared with mafic counterparts due to elevated concentrations of methane. However, despite these differences in energy yields and diversity, the microbial communities at both sites show a high level of similarity, both functionally and phylogenetically, and are distinct compared with other hydrothermal systems on Earth. At Von Damm and Piccard, both V6 amplicon and metagenomic analyses indicate that lineages within highly conserved genera of microbes, e.g. Methanothermococcus , Archaeoglobus and Sulfurovum , are important in the subseafloor at MCR. These three predominant genera have representative cultivars that are all capable of using hydrogen, and our metagenomic analysis shows the presence of diverse hydrogenases as well. The fact that these organisms dominate in both settings is consistent with the high concentrations of hydrogen in vent fluids at Von Damm and Piccard and the important role of hydrogen in microbial metabolism. Despite these similarities, however, at the fine scale of analysis allowed by Illumina sequencing of the V6 region, we showed that while a few OTUs were found at both sites (OTU0 in the archaea, Archaeoglobaceae , and OTU1 in the bacteria, Sulfurovum ), most of the other phylotypes were found almost exclusively at only one site, with individual orifices at each vent field being more similar to each other than to orifices at the other vent field. These results show there are distinct communities of subseafloor microorganisms at both Piccard and Von Damm, and either environmental selection (fluid chemistry, pressure, temperature) or physical separation (depth, geographic distance, limits to dispersal, genetic drift), or a combination of both, has allowed for fine‐scale phylogenetic differentiation of these communities. Differences in methanogenic activity were also detected between the two sites. The stable isotope tracing experiments showed methanogenesis with formate occurring at Von Damm, but not at Piccard. This result is consistent with thermodynamic models for carbon speciation that predicts higher formate concentrations at Von Damm mixed fluids relative to Piccard (Seewald et al ., 2006 ; McDermott, 2014 ). The thermodynamic predictions have been confirmed by formate measurements in mixed fluids that indicate concentrations up to 669 μmol kg −1 at Von Damm and substantially lower concentration at Piccard (max 58.1 μmol kg −1 at Hot Chimlet; McDermott et al ., 2015 ). Formate concentrations up to 158 μmol kg −1 have also been measured at the ultramafic site Lost City, although whether or not microorganisms are using formate at Lost City has not been demonstrated (Lang et al ., 2010 ). The high methane concentrations (up to 2.8 mM in end‐member fluids; Bennett et al ., 2013 ; Reeves et al ., 2014 ; McDermott et al ., 2015 ) and potentially other reduced carbon sources at the ultramafic‐influenced Von Damm vent field (Bennett et al ., 2015 ; McDermott et al ., 2015 ) also likely account for the greater diversity in both domains and functions relative to Piccard. While microbial communities at all Piccard vents were highly similar to one another, much more variability and diversity were observed among the vents at Von Damm. In addition, within Von Damm, there were a number of vents that looked extremely different from Piccard and other vents at Von Damm, such as Shrimp Hole, where the microbial composition was similar to those observed in cold seep sediments. Shrimp Hole is the only location at Von Damm where live tube worms were seen (Plouviez et al ., 2015 ), and both genera of tube worms found ( Escarpia and Lamellibrachia ) have been reported in sedimented hot vents and cold methane seeps (Black et al ., 1997 ). At Shrimp Hole, the archaeal population was dominated by sequences related to the unknown Methanosarcinales GOM Arc I group, previously described from sediments overlying a hypersaline methane seep in the Gulf of Mexico (Lloyd et al ., 2006 ). In the bacterial community at Shrimp Hole, the most abundant sequence found was related to the propionate‐oxidizing family Desulfobulbaceae within the Deltaproteobacteria , a frequently encountered group at methane‐rich sediments and in consortia with ANME (Niemann et al ., 2006 ). The high relative abundance of putative sulfate‐reducing bacteria at Shrimp Hole is consistent with sulfur isotope data from this site, indicating that microbial sulfate reduction is occurring in the shallow subseafloor (Bennett et al ., 2015 ). Anaerobic methane‐oxidizing archaea were also frequently detected at Von Damm, including Shrimp Hole, but very rarely at Piccard. This is consistent with geochemical modelling results, where the anaerobic oxidation of methane is predicted to be moderately exergonic in ultramafic systems, but not in mafic systems (Amend et al ., 2011 , this study). The ANME lineages at Von Damm are not the same as those seen at Lost City, where a single phylotype of archaea known as the Lost City Methanosarcinales dominates (Schrenk et al ., 2004 ). The presence of ANME and sulfate‐reducing bacteria in the venting fluids at Shrimp Hole suggests seep‐like characteristics (i.e. warm fluids with high methane concentrations and high activity of anaerobic methane oxidizers) on the outer perimeter of the Von Damm vent field, similar to ‘hydrothermal seeps’ seen along the Costa Rica Margin, for example (Levin et al ., 2012 ). While our metagenomic work was consistent with conclusions from the V6 data with respect to both the dominant lineages and the diversity of lineages at each site, it also allowed further insight into the potential metabolisms of these communities. Our analysis revealed near functional equivalence among dominant lineages at Von Damm and Piccard based on major subsystems, as well as key hydrogen, sulfur, methane and carbon fixation genes. However, direct comparison of the thermodynamic modelling and functional profiles indicates a number of discrepancies, where the prediction from modelling does not appear consistent with the relative abundance of a particular metabolic gene. For example, specific genes involved in methanogenesis and sulfate reduction, mcr and dsr, respectively, were found in both metagenomes at relatively low abundance, although our thermodynamic modelling shows anaerobic hydrogenotrophic metabolisms should be favoured. These discrepancies can be explained by a number of important points in comparing modelling and metagenomics. First, metagenomics only highlight the functional potential, not necessarily genes that are transcribed, which is in contrast to the thermodynamic modelling, which predicts what metabolisms are most likely to be active given the energy yields under different mixing regimes, and thus temperatures. Second, our thermodynamic modelling shows these anaerobic metabolisms are favoured at 80°C under anoxic conditions, which is not the condition at the point of sampling due to mixing with seawater. Beneath the seafloor at some uncharacterized depth, these organisms are likely thriving in anoxic warm niches, where methanogenesis and sulfate reduction might be favoured. Finally, it is important to note that archaea only make up 1–10% of the microbial community in these mixed fluids, as shown by qPCR data and the extraction of 16S rRNA genes from the metagenomes, which might also explain the low abundance of these archaeal genes in the studied metagenomes. Similarly, sqr and nap, key genes involved in sulfide oxidation and nitrate reduction, respectively, are the most frequently detected reads in the metagenomes, with most mapping to Epsilonproteobacteria . However, thermodynamic modelling suggests that the catabolic energies available from sulfide oxidation and hydrogen denitrification are relatively low. A number of Epsilonproteobacteria are known to carry genes for both sulfide and hydrogen oxidation, together with nitrate reduction, such as in the genome of Sulfurovum NBC37‐1 where six copies of the sqr gene, one copy of napA and three different [Ni‐Fe] hydrogenases were found (Nakagawa et al ., 2007 ; Meyer and Huber, 2014 ). Our analysis of taxonomic annotation for each gene highlights that Sulfurovum at MCR have the genomic potential to carry out both sulfide and hydrogen oxidation, as well as denitrification, confirming their metabolic flexibility, even though we cannot say which metabolisms they are using at MCR. Despite these discrepancies between the thermodynamic modelling and metagenomics analysis, overall, they are consistent with both the amplicon data and modelling efforts. Genes for all the modelled metabolisms were identified in the metagenomes, including a wide range of genes involved in hydrogen, sulfur and methane metabolism. Importantly, metabolisms that were not predicted based on modelling efforts were not seen at the level of sequencing carried out. For example, thermodynamic modelling predicted very little methane oxidation at Piccard, and this was also seen in the metagenome, with the methanol dehydrogenase and methane monooxygenases genes only found at Von Damm, not Piccard. The combined modelling and ‐omics effort of this study will provide the groundwork for future work focusing on in situ activity of organisms that will improve our understanding of which organisms and metabolisms are ‘turned on’ under subseafloor‐relevant conditions. In addition to comparing thermodynamic predictions to real field data from venting fluids at MCR, we also compared the newly discovered vents to sites across the globe. With the exception of Shrimp Hole, the global taxonomic analysis showed that the MCR vents are more closely related to one another than any other vent site, highlighting their distinct community structure, and the inability of only rock‐host type or vent fluid chemistry to predict microbial community composition. While there are certainly weaknesses to using only taxonomic profiles rather than more sensitive techniques such as OTUs or oligotypes to compare communities, results clearly show which dominant taxa are driving differentiation among sites. For example, the MCR vents were more similar to eruption‐associated ephemeral diffuse vents than to any other environment surveyed due to high relative abundance of Methanothermococcus and Sulfurovum . Meyer and colleagues ( 2013 ) suggest these anaerobic methanogens and other microaerobic taxa in Axial snowblower come from deeper communities that are seeded from the volatile‐rich eruptive fluids, which could explain the high level of microbial community similarity between Axial snowblower vents and the hot and reduced MCR vent fluids. Similarly, the high abundance of hyperthermophilic sulfate reducers Archeaglobaceae , commonly found in sulfide chimneys, where sulfate and either organics or hydrogen are provided by seawater and hot fluids, respectively, supports the high similarity level between sulfide chimneys of Lau Basin and the MCR samples. Interestingly, the two ultramafic sites with similarly high concentrations of hydrogen to Von Damm and Piccard – Rainbow with 12–16 mM end‐member hydrogen (Charlou et al ., 2002 ) and Lost City with ∼ 10 mM hydrogen (Kelley et al ., 2002 ) – showed distinct lineages not found at the MCR vents, including Methanocaldococcus at Rainbow, novel ANME lineages at Lost City, and Thiomicrospira at both. In line with these findings, the analysis also showed that sequences of the Epsilonproteobacterial group, Sulfurovum , the key bacteria in the MCR, have not been detected in such abundance in any other hydrothermal systems. Bacterial communities in fluids from volcanoes of the Mariana Arc and Axial Seamount are dominated by the aerobic or microaerobic, mesophilic group Sulfurimonas . These frequently detected lineages in diffuse fluids are also found in hydrothermal sediments (Inagaki, 2003 ), associated with vent animals (Takai et al ., 2006 ), in coastal marine sediments (Sievert et al ., 2008 ), and in redox clines (Grote et al ., 2012 ), but were strikingly absent in MCR fluids. Similarly, the Gammaproteobacteria lineage SUP05, an ubiquitous deep ocean aerobic sulfur‐oxidizing bacterium that is abundant in other vent environments (Anantharaman et al ., 2012 ; Akerman et al ., 2013 ; Anderson et al ., 2013 ; Dick et al ., 2013 ) were not detected at the MCR. The absence of both of these lineages is likely attributed to the highly reducing nature of the MCR vent sites, as well as the elevated temperatures of mixed fluids. Consistent with thermodynamic predictions, at temperatures above ∼ 40°C, the (micro)aerobic metabolisms of these psychrophilic or moderately mesophilic groups are not expected. This comparison of microbial taxonomic profiles across different deep‐sea hydrothermal vent environments, spanning a wide diversity of systems from ephemeral eruptive hydrothermal events to sulfide chimneys at the global scale, confirms the distinct composition and structure of the MCR subseafloor microbial communities. Altogether, these findings from the newly discovered MCR vent fields represent an important advance in the field of hydrothermal and subseafloor research, shedding light on new aspects of diversity for venting types, fluid chemistry and subseafloor microbial communities along the spectrum of known volcanism on our planet."
} | 5,907 |
33791444 | null | s2 | 8,784 | {
"abstract": "To protect against diverse challenges, the immune system must continuously generate an arsenal of specialized cell types, each of which can mount a myriad of effector responses upon detection of potential threats. To do so, it must generate multiple differentiated cell populations with defined sizes and proportions, often from rare starting precursor cells. Here, we discuss the emerging view that inherently probabilistic mechanisms, involving rare, rate-limiting regulatory events in single cells, control fate decisions and population sizes and fractions during immune development and function. We first review growing evidence that key fate control points are gated by stochastic signaling and gene regulatory events that occur infrequently over decision-making timescales, such that initially homogeneous cells can adopt variable outcomes in response to uniform signals. We next discuss how such stochastic control can provide functional capabilities that are harder to achieve with deterministic control strategies, and may be central to robust immune system function."
} | 269 |
36557674 | PMC9784785 | pmc | 8,785 | {
"abstract": "Microbial fuel cells are a promising technology for future wastewater treatment, as it allows cleaning and power generation simultaneously. The bottleneck of microbial fuel cells is often its cathodes because they determine the power output. Gas diffusion electrodes might overcome this bottleneck due to their low production costs and high oxygen reduction rates. However, biofilm formation on the gas diffusion electrodes reduces their performance over time. In this work, a new reactor design of the microbial fuel cell using rotating gas diffusion electrodes is presented. The biofilm growth on the electrode during operation was observed and its effect on the performance of the microbial fuel cell was examined. In addition, different antifouling strategies were investigated over a period of 80 days. It was found that already after 7 days of operation a complete biofilm had grown on an untreated gas diffusion electrode. However, this does not seem to affect the performance of the cells in the beginning. Differences in the performance of the reactors with and without an antifouling strategy only become apparent from day 15 onwards. The use of UV radiation and antibacterial membranes leads to the best results with maximum power densities of approx. 200 mW m −2 while the untreated microbial fuel cell only achieves a maximum power density of approx. 20 mW m −2 .",
"conclusion": "4. Conclusions and Outlook In this contribution, a new MFC reactor concept was investigated and successfully operated. Five different anti-fouling strategies to prevent biofilm growing on GDEs were integrated in this new reactor design and compared to a GDE without any anti-fouling strategy. All strategies have in common that they work in situ, which means that the GDE did not have to be removed from the reactor. Comparing the power generation and the biofilm growth on the electrode, UV and membrane were the most promising anti-fouling strategies. Although the power generation of the reactor with UV-LED was much higher in the beginning, membrane as anti-fouling strategy is recommended. This is mostly due to the high operational costs. The UV-LED works at approx. 6.1 V and 0.25 A. The consumption will be 36.6 Wh d −1 , if the UV-LED is operated continuously. This means the UV-LED consumes more power than the MFC produces. For future application, the duration of the UV irradiation should be reduced to enable net power generation by MFC. Additionally, the use of UV-LEDs in stack systems is another challenge, as there must be a UV-LED on each side of the rotating cathode. Coating the GDE with a membrane causes some higher production costs, but no additional energy consumption during operation. The membrane also causes a better separation of the cathode chamber from the anode chamber and prevents oxygen crossover from the GDE backside to the wastewater. For future application, the membrane should be prepared by coating the GDE with a sprayable solution of the membrane material. This approach avoids the gluing of an additional membrane and simplifies the production of MEAs. The in situ prevention of biofilm growth on chemical catalytic GDEs is appreciated as a pre-requisite towards technical scale wastewater treatment applications of single-chamber MFC using GDEs, because a long-term, low-cost operation is necessary for commercial breakthrough of MFCs. Costs for antifouling strategies, e.g., for energy (polarity reversal, UV-LED) as well as for maintenance (e.g., cathode replacement) need to be minimized. Antifouling membrane coating seems to be a promising technology for the desired application, as they may prevent biofilm growth and guarantee long-term operation of MFC with sufficient power output.",
"introduction": "1. Introduction Microbial fuel cells (MFCs) are a promising technology for simultaneous wastewater treatment and power generation [ 1 ]. From the early 20th century [ 2 ] until today, uninterrupted research activities can be registered. Despite several thousands (>18,000) of publications even in the year 2020 (Google scholar, keyword microbial fuel cell), there is no commercial breakthrough, yet [ 3 ]. MFCs are still in the stadium of basic research. The reasons for that are manifold, starting from high investment costs, high operational costs [ 4 ], still low power output [ 5 ], and the use of toxic materials such as potassium ferricyanide [ 6 ]. Several research groups identified especially the cathode as the bottleneck, which means that the cathode determines the power generation and the operation (time) of the MFC [ 7 ]. For that reason, new cathodes have to be developed, which deliver high and stable electrical power outputs [ 8 ]. Gas diffusion electrodes (GDEs) are an encouraging technology to overcome that bottleneck as they can be made of unharmful materials such as carbon and additional catalysts for oxygen reduction reaction (ORR) and they are also able to deliver a high electric output [ 9 ]. However, GDEs are prone to fouling, which affects long-term stability and leads to a decrease in cell performance over time. Therefore, more and more antifouling methods are being researched. Table 1 shows an overview of selected MFCs with antifouling investigations and their power densities. Anti-fouling strategies can be categorized into two groups: in situ strategies and non-in situ strategies. In situ means that the anti-fouling strategy can be operated directly in the reactor without any GDE-disassembling. This minimizes operational costs and prevents operational outages, especially regarding a future technical operation. For that reason, five possible in situ strategies were selected and investigated: UV radiation (1), coating with copper compound (2), polarity reversal (3), rotation speed adjustment (4) and membrane coating (5). All approaches were carried out once to examine their feasibility and effectiveness. The details of each strategy are as follows: (1) UV light is known for its antibacterial effect. The wavelength of approx. 240–290 nm affects the RNA and DNA of organisms and this leads to inactivation [ 25 ]. UV light disinfection is state of the art in drinking water disinfection [ 26 ]. When using UV rays against fouling, it has been shown that higher intensity improves the protection against fouling [ 27 ]. The required effective dose varies depending on the type of bacteria, mold spores and algae [ 28 ]. In addition, a higher exposure time also improves the antifouling effect [ 29 ]. (2) Copper is known for its biocidal effect on a wide range of microbials [ 30 ] and application spectrums, e.g., hospitals [ 31 ]. Therefore, copper or copper compounds are a promising method to prevent bacterial growth on the surface of the GDE. (3) Polarity reversal is known in electrochemical systems to prevent the cathode from scaling. During polarity reversal, the cathode becomes the anode and oxidation process occurs on the former cathode [ 32 ]. This technique has to be adapted for use in MFCs. (4) Mechanical ablation by increasing the shearing force is another option for removing parts of the biofilm. Conventional ablation by using blades or wipers was not possible, as determined in prior experiments. Any direct contact with blades or wipers destroys the surface of the GDE, whereby gaps and consequently electrolyte breakthroughs occur. For that reason, shear forces should remove the biofilm of the GDE. It is known that it is possible to control the structure and thickness of the electroactive biofilm on the anode by adjusting the shear force (e.g., flow rate of the pump) [ 33 ]. This technique is applied for the GDE to remove the unwanted bacteria from the surface of the GDE. (5) The functional amino groups are known for their antibacterial effect. The negatively charged cell membranes adsorb to the positively charged amino groups, resulting in damage to the cell. Therefore, a coating with these groups might prevent bacterial growth of the GDE [ 34 ]. For the first time, five possible anti-fouling strategies were investigated and compared with a new reactor concept based on rotating cathodes.",
"discussion": "3. Results and Discussion Figure 3 shows the biofilm growth during 11 days in an MFC without anti-fouling strategy, taken by the Raspberry Pi camera. Within 24 h after filling the reactor with wastewater, a biofilm on the surface is visible. After seven days, the complete surface is covered with a biofilm. In the following days, the biofilm grows to a thickness of approx. 3 to 5 mm, manually measured after disassembling the GDE. Then the biofilm was removed from the GDE for organic dry matter analysis and qualitative bacteria screening by Amodia Bioservice GmbH. The average organic content of the samples, based on the dry matter, is about 80.3%. DNA sequence analysis confirmed the presence of acidaminococcus fermentans, victivallis lenta, acidaminococcus intestini and parabacteroides chartae. These are all anaerobic and mesophilic bacteria. The reduction of the power generation of the cells does not correlate with the biofilm growth in the beginning (see Figure 4 ). The biofilm seems not to affect the power generation initially, as differences in the performance of the reactors with and without antifouling strategy become evident only from day 15 onwards. This is likely because the biofilm at the beginning is thin enough so that the ions can still migrate through the biofilm in a sufficient number. The electrical current between the anode and cathode, and thus the ion exchange, is very low compared to chemical fuel cells. It is also likely that the biofilm grows inside the pores of the GDE, so that the bacteria may use the oxygen of the GDE. Within 80 days, the power densities of the five reactors show different developments. The lowest power density (PD) was obtained with Reactor 2 (copper resinate). This low PD could be explained by the coating itself. The resinate formed a compact layer on the GDE surface, which might have closed all the GDE’s pores, which reduced the performance of the GDE drastically. A visual inspection of the GDE shows complete biofilm growth ( Figure 5 b). It is possible that the copper in the resinate is not accessible and could not develop its antibacterial effect. The second lowest PD was achieved with Reactor 4 (varying rotation speed). It is obvious that the highest rotation speed of the GDE element was not high enough to prevent biofilm growing on the GDE surface. A higher rotation speed might be more sufficient in biofilm preventing, but a higher rotation speed could not be adjusted due to the motor parameters. The achieved PD with that reactor stayed almost constant over the time with a maximum of approx. 20 mW m −2 at the beginning. An examination of the GDE confirms the assumption. The entire electrode surface is affected by a biofilm ( Figure 5 d). Reactor 3 (polarity reversal) obtained quite better PDs, rising to a maximum of approx. 45 mW m −2 at day 70. The polarity reversal seems to be a potential strategy for biofilm inhibition on the GDE surface, as it has a positive effect on the PD. This can also be verified by the surface of the GDE, which shows less biofilm growth than an untreated electrode ( Figure 5 f). The thickness of the microbial film is quite thin so that the black electrode surface underneath is still visible ( Figure 5 c). The highest PDs were achieved with Reactors 1 (UV) and 5 (membrane). Both reactors reached a maximum PD of about 200 mW m −2 , but after 75 days both reactors reached a stable PD of approx. 150 mW m −2 . Both reactors showed an increasing power generation, but Reactor 1 benefitted from a faster increasing PD. The PD of the reactor with the UV-LED decreased at day 58, because the feed was interrupted. Accordingly, the use of the antibacterial membrane and the UV irradiation seems to reduce the biofilm infestation. This can be confirmed by the photos of the removed cathodes. The irradiated electrode does not show any biofilm on its surface. Only scaling effects can be seen ( Figure 5 a). The GDE coated with a membrane shows slight biofilm growth but the black surface of the electrode can still predominantly be seen ( Figure 5 e). In addition, the biofilm does not appear to be cross-linked to the membrane, allowing an easy removement."
} | 3,068 |
25343514 | PMC4351914 | pmc | 8,786 | {
"abstract": "The energy metabolism of essential microbial guilds in the biogeochemical sulfur cycle is based on a DsrAB-type dissimilatory (bi)sulfite reductase that either catalyzes the reduction of sulfite to sulfide during anaerobic respiration of sulfate, sulfite and organosulfonates, or acts in reverse during sulfur oxidation. Common use of dsrAB as a functional marker showed that dsrAB richness in many environments is dominated by novel sequence variants and collectively represents an extensive, largely uncharted sequence assemblage. Here, we established a comprehensive, manually curated dsrAB /DsrAB database and used it to categorize the known dsrAB diversity, reanalyze the evolutionary history of dsrAB and evaluate the coverage of published dsrAB -targeted primers. Based on a DsrAB consensus phylogeny, we introduce an operational classification system for environmental dsrAB sequences that integrates established taxonomic groups with operational taxonomic units (OTUs) at multiple phylogenetic levels, ranging from DsrAB enzyme families that reflect reductive or oxidative DsrAB types of bacterial or archaeal origin, superclusters, uncultured family-level lineages to species-level OTUs. Environmental dsrAB sequences constituted at least 13 stable family-level lineages without any cultivated representatives, suggesting that major taxa of sulfite/sulfate-reducing microorganisms have not yet been identified. Three of these uncultured lineages occur mainly in marine environments, while specific habitat preferences are not evident for members of the other 10 uncultured lineages. In summary, our publically available dsrAB /DsrAB database, the phylogenetic framework, the multilevel classification system and a set of recommended primers provide a necessary foundation for large-scale dsrAB ecology studies with next-generation sequencing methods.",
"introduction": "Introduction The DsrAB-type dissimilatory (bi)sulfite reductase is a key microbial enzyme in both the reductive and the oxidative steps of the biogeochemical sulfur cycle. Utilized by microorganisms that catalyze redox reactions involving sulfur-containing compounds as components of energy metabolism, it catalyzes the reduction of sulfite to sulfide during anaerobic respiration with sulfate, sulfite or organosulfonates as terminal electron acceptor, and functions in reverse during sulfide oxidation. DsrAB enzymes are heterotetramer proteins with an α 2 β 2 structure and possess iron-sulfur clusters and siroheme prosthetic groups ( Dahl et al. , 1993 ). The α and β subunits are encoded by the paralogous genes dsrA and dsrB , respectively, which are organized in a single copy operon with dsrA preceding dsrB ( Dahl et al. , 1993 ; Karkhoff-Schweizer et al. , 1995 ; Wagner et al. , 1998 ; Larsen et al. , 1999 ; Pereira et al. , 2011 ). Given the presumed antiquity of siroheme and the proposed existence of DsrAB before the separation of the domains Bacteria and Archaea ( Wagner et al. , 1998 ; Dhillon et al. , 2005 ; Loy et al. , 2009 ), DsrAB enzymes are considered very ancient and might have had a fundamental role in mediating biological conversions of sulfur compounds by some of the first microorganisms in the anoxic, reduced atmosphere environments of the primordial Earth ( Wagner et al. , 1998 ; Canfield and Raiswell, 1999 ; Huston and Logan, 2004 ). It is now recognized that the distribution of dsrAB among extant microorganisms was driven by a combination of divergence through speciation, functional diversification and lateral gene transfer (LGT) between unrelated taxa ( Loy et al. , 2008b ). DsrAB enzymes are best known from sulfate-reducing microorganisms (SRMs) because of their global relevance for biogeochemical cycling of sulfur and carbon ( Pester et al. , 2012 ; Bowles et al. , 2014 ). DsrAB catalyzes the last and main energy-conserving step in the dissimilatory sulfate reduction pathway that is conserved in all cultivated SRM, which are distributed in four bacterial ( Proteobacteria —class Deltaproteobacteria , Nitrospirae , Firmicutes , Thermodesulfobacteria ) and two archaeal phyla ( Euryarchaeota , Crenarchaeota ). The canonical pathway essentially consists of the enzymes ATP sulfurylase (Sat), adenosine 5′-phosphosulfate reductase (Apr) and dissimilatory (bi)sulfite reductase (Dsr). However, a new, yet unresolved pathway for sulfate reduction was suggested to operate in a syntrophic microbial consortium that mediated the anaerobic oxidation of methane coupled to sulfate reduction and polysulfide disproportionation ( Milucka et al. , 2012 ). Surprisingly, sulfate was reduced by an unknown mechanism in the archaeal partner resulting in the formation of disulfide and not by the deltaproteobacterial partner that harbors the canonical DsrAB-based pathway. DsrAB genes are also present in some microorganisms that are unable to use sulfate as a terminal electron acceptor including sulfite-reducing microorganisms (e.g., Desulfitobacterium , Desulfitibacter and Pyrobaculum ) ( Simon and Kroneck, 2013 ), sulfur-disproportionating bacteria (e.g., Desulfocapsa sulfexigens ) ( Finster, 2008 ) and in organisms that metabolize organosulfonates to internally produce sulfite for respiration (e.g., the taurine-consuming gut bacterium Bilophila wadsworthia ) ( Devkota et al. , 2012 ). The physiological role of DsrAB in anaerobic syntrophs of the spore-forming Firmicutes genera Pelotomaculum and Sporotomaculum , which possess and transcribe dsrAB but are incapable of reducing sulfite, sulfate or organosulfonates ( Brauman et al. , 1998 ; Imachi et al. , 2006 ), is unknown. Some but not all sulfur-oxidizing bacteria (SOB) carry a reversely operating DsrAB that is homologous, yet phylogenetically clearly distinct from DsrAB enzymes that function in sulfite reduction ( Schedel and Trüper, 1979 ; Loy et al. , 2009 ). Unlike most SRM, SOB do not share a common sulfur metabolism pathway, but exploit various, partially redundant enzyme systems for the oxidation of a range of reduced sulfur compounds with intermediate oxidation states ( Kelly et al. , 1997 ; Kletzin et al. , 2004 ; Friedrich et al. , 2005 ). DsrAB is essential for the oxidation of sulfur globule repositories ( Pott and Dahl, 1998 ; Dahl et al. , 2005 ) and might thus provide these SOB with an advantageous backup sulfur metabolism in environments with varying concentrations of reduced sulfur compounds. Thus far, dsrAB have been detected in free-living and symbiotic sulfur-storing SOB of the phyla Proteobacteria (classes Alpha -, Beta -, Gamma- and Deltaproteobacteria ) and Chlorobi ( Loy et al. , 2009 ; Swan et al. , 2011 ; Sheik et al. , 2014 ). With a few significant exceptions that are indicative of LGT of dsrAB among major SRM taxa, DsrAB and 16S rRNA phylogenies are largely congruent ( Klein et al. , 2001 ; Zverlov et al. , 2005 ; Loy et al. , 2009 ). Consequently, dsrAB have been frequently exploited as phylogenetic marker genes in amplicon sequencing-based environmental studies ( Dhillon et al. , 2003 ; Nakagawa et al. , 2004 ; Leloup et al. , 2006 ; Loy et al. , 2009 ; Moreau et al. , 2010 ; Mori et al. , 2010 ; Pester et al. , 2010 ; Lenk et al. , 2011 ). Application of the dsrAB approach ( Wagner et al. , 2005 ) in diverse environments has uncovered an extensive hidden diversity of dsrAB sequences that are not closely related to dsrAB from any recognized organisms. New sequencing techniques have opened up opportunities for large-scale α- and β-diversity studies of dsrAB . However, the currently available dsrAB sequence set is largely uncharacterized, which poses considerable problems in identifying and classifying newly obtained environmental sequences. A comprehensive classification framework for streamlined computational analyses of large dsrAB sequence data sets is thus urgently needed. A first step toward a dsrAB classification system has been made by a meta-analysis of dsrAB diversity that focused on freshwater wetland SRM ( Pester et al. , 2012 ). This study highlighted the existence of at least 10 major monophyletic lineages that were only composed of environmental sequences and similar in intralineage diversity to known SRM families. Furthermore, several primers targeting reductive and oxidative dsrAB types have been published and applied for PCR-based environmental monitoring of the diversity and abundance of sulfur-cycling microorganisms ( Wagner et al. , 1998 ; Kondo et al. , 2004 ; Geets et al. , 2005 ; Loy et al. , 2009 ; Mori et al. , 2010 ; Lenk et al. , 2011 ; Steger et al. , 2011 ; Lever et al. , 2013 ), but it is unclear how well these primers cover the currently known dsrAB diversity and thus how suitable they are for such purposes. In the present study, we established a comprehensive, manually aligned and curated database of nucleic acid and inferred amino-acid sequences of dsrAB that are available in public sequence repositories, and provided a robust, taxonomically and phylogenetically informed classification system for the entire environmental dsrAB diversity. This allowed us to classify and systematically quantify the uncharted dimensions of dsrAB diversity and to reveal its distribution across various environments. We further used the database to determine the in silico coverage of all published dsrA- or dsrB -targeted primers to provide guidance for future PCR-based studies.",
"discussion": "Results and discussion The DsrAB consensus tree provides a robust phylogenetic framework for environmental studies For a dsrAB census, we created a comprehensive database of 7695 sequences with ⩾300 nucleotides length and sufficient quality that derived from 530 amplicon sequencing, metagenome or genome studies. For more reliable phylogenetic inferences, we constructed a DsrAB consensus tree using a core data set of 1292 sequences with ∼1.9 kb length ( Figure 1 and Supplementary Figures S1–S3 ). The DsrAB tree has four main basal branches that delineate three major DsrAB protein families, namely the reductive bacterial type, the oxidative bacterial type and the reductive archaeal type. The fourth branch is so far only represented by the second dsrAB copy of Moorella thermoacetica ( Pierce et al. , 2008 ; Loy et al. , 2009 ). Through paralogous rooting analysis, we show that this copy and the reductive archaeal-type DsrAB family represent the deepest branches in the DsrAB tree and add support to the previously proposed early evolution of DsrAB as a reductive enzyme ( Wagner et al. , 1998 ) ( Supplementary Results and discussion and Supplementary Figure S4 ). To assess the minimum number of dsrAB -containing species that are currently represented in the 1292 sequence core data set, we initially inferred a species-level sequence identity cutoff from the linear regression in a plot of corresponding pairs of 16S rRNA gene and non-laterally acquired reductive- and oxidative bacterial-type dsrAB identities ( Supplementary Figure S5 ) ( Kjeldsen et al. , 2007 ). A dsrAB nucleic acid sequence identity of 92% over the ∼1.9 kb fragment is equivalent to a 16S rRNA sequence identity of 99%, which is a frequently used threshold for delineating species-level OTUs ( Stackebrandt and Ebers, 2006 ). We recommend using a more conservative threshold of 90% dsrAB sequence identity, because two organisms with <90% dsrAB identity generally have <99% 16S rRNA identity ( Supplementary Figure S5 ) and will likely represent two different species, given that dsrAB is usually present as a single copy per genome. Application of the 90% threshold showed that already the core data set represents a minimum of 779 species-level OTUs, of which 647 are of the reductive and 118 of the oxidative bacterial DsrAB type. For comparison, ∼240 species of SRM are currently present in the List of Bacterial Names with Standing in Nomenclature ( Euzéby, 1997 ). The reductive bacterial-type DsrAB family cluster mostly contains bacteria that use sulfate, sulfite or organosulfonates as terminal electron acceptors ( Loy et al. , 2008b ), and also from sulfate/sulfite-reducing archaea that received dsrAB via LGT from ancestral bacterial donors (see section below). Two hundred and ninety-nine environmental sequences of the core data set were not affiliated with members of described taxonomic families and clustered in 13 stable, monophyletic lineages, which were designated ‘uncultured DsrAB lineages 1 to 13' (note that lineages 1 to 10 were defined previously; Pester et al. , 2012 ) using a combination of dsrAB sequence identity-based and phylogenetic criteria ( Figure 1 , Supplementary Figure S1 and Supplementary Materials and methods ). Each of these 13 lineages could represent a taxonomic family whose members are yet uncultured or not known to possess dsrAB , illustrating the enormous unexplored diversity of dsrAB -harboring microorganisms in the environment. Our phylogenetic analysis even provided indications for further lineages of environmental dsrAB sequences ( Figure 1 ), but these did not meet our conservative criteria to label them as an ‘uncultured family-level DsrAB lineage'. Importantly, only very few sequences ( n =4) of the uncultured family-level lineages contain internal stop codons, which are not confirmed and might result from sequencing errors. Furthermore, nonsynonymous/synonymous substitution rate ratios of the branches that lead to the 13 uncultured family-level lineages are clearly below one ( ω =0.05–0.37), which highlights strong purifying selection and suggests that these dsrAB variants are being expressed as functionally active proteins ( Yang, 1997 ; Yang et al. , 2000 ). Although a very recent loss of function will not be evident in the DsrAB sequence record, it is nevertheless unlikely that this vast environmental dsrAB diversity is primarily caused by uncontrolled mutation rates owing to the lack of or reduced selective pressure, for example, in viruses ( Anantharaman et al. , 2014 ) or microorganisms that received dsrAB via LGT yet do not make use of them. At a higher phylogenetic level, we could reproduce three previously described ‘superclusters' ( Pester et al. , 2012 ), namely the Deltaproteobacteria supercluster, the Nitrospirae supercluster, which was previously named Thermodesulfovibrio supercluster ( Supplementary Results and discussion ), and the environmental supercluster 1, which each comprise at least two uncultured DsrAB family-level lineages and/or known SRM families ( Figure 1 ). DsrAB of the euryarchaeal genus Archaeoglobus and related sequences from thermophilic environments form a separate branch in the reductive bacterial-type DsrAB family tree. All remaining sequences, namely those that are not affiliated with the three superclusters and the Archaeoglobus cluster, did not group consistently at a higher phylogenetic level ( Steger et al. , 2011 ; Pester et al. , 2012 ), and we have thus not designated them as a supercluster but as the Firmicutes group sensu lato . These high-order groups/superclusters are named after the main phylum/class that they affiliate with but do not necessarily imply a taxonomic affiliation. Similar to the Deltaproteobacteria supercluster, the highly diverse Firmicutes group contains dsrAB from cultivated members of different phyla and many environmental sequences ( Supplementary Results and discussion ). Oxidative-type DsrAB sequences from SOB form a monophyletic enzyme family that is phylogenetically distinct from all other DsrAB sequences ( Figure 1 and Supplementary Figure S4 ). The branching pattern of the tree suggests that oxidative DsrAB evolved by an ancient functional adaptation from an ancestral reductive DsrAB before the diversification into extant DsrAB-carrying phyla. Sequences from known SOB of the classes Alpha -, Beta - and Gammaproteobacteria and the phylum Chlorobi form separate clusters in the DsrAB tree that are generally in accordance with the organismal taxonomy ( Figure 1 , Supplementary Figure S2 and Supplementary Materials and methods ). Only Thioalkalivibrio nitratireducens branches outside the Gammaproteobacteria cluster. Its DsrAB sequence is remarkably different (67–71% amino-acid identity) from the DsrAB of three other Thioalkalivibrio species (as opposed to 87–95% DsrAB identity among these three species). Metagenomic ( Sheik et al. , 2014 ) and single-cell genome ( Swan et al. , 2011 ) analyses have recently identified reverse dsrAB and accessory genes for sulfur oxidation in members of the deltaproteobacterial SAR324 clade. These deltaproteobacterial reverse DsrAB branch with DsrAB of Chlorobi and Magnetococcus marinus , a species that has been provisionally included in the Alphaproteobacteria ( Bazylinski et al. , 2013 ). Interestingly, the root of the oxidative-type DsrAB branch is not located between the Proteobacteria and the Chlorobi . Instead, Chlorobi form a monophyletic cluster with M. marinus and the putative sulfur-oxidizing deltaproteobacterium ( Figure 1 ), which provides phylogenetic support for the acquisition of dsrAB by Chlorobi via LGT from a sulfide-oxidizing proteobacterial donor. Such a scenario has been previously postulated based on the absence of dsrAB in the deep-branching Chlorobi member Chloroherpeton thalassium ( Frigaard and Bryant, 2008 ). Archaeal-type dsrAB sequence diversity is mainly represented by sequenced genomes and metagenomes because PCR primers commonly used for amplification of dsrAB do not bind to archaeal-type dsrAB . So far, three genera within the hyperthermophilic family Thermoproteaceae (order Thermoproteales ) of the phylum Crenarchaeota , namely species of Pyrobaculum ( n =7), Vulcanisaeta ( n =2) and Caldivirga ( n =1), are known to harbor this type of dsrAB , and each genus is represented by a distinct monophyletic group in the archaeal DsrAB tree ( Figure 1 and Supplementary Figure S3 ). Members of all three genera of archaeal-type DsrAB-carrying organisms are able to reduce thiosulfate and elemental sulfur ( Molitor et al. , 1998 ; Itoh et al. , 1999 , 2002 ). So far, sulfate reduction has been shown only for Caldivirga maquilingensis ( Itoh et al. , 1999 ); however, Vulcanisaeta species might also be capable of sulfate reduction ( Itoh et al. , 2002 ), as genes for the complete canonical sulfate reduction pathway are present in the genomes of Vulcanisaeta distributa ( Mavromatis et al. , 2010 ) and V. moutnovskia ( Gumerov et al. , 2011 ). dsrAB are robust phylogenetic marker genes for sulfur compound-dissimilating microorganisms Phylogenetic signal is blurred in genes that are subject to (i) frequent LGT between unrelated organisms and (ii) duplication and subsequent functional diversification ( Koonin et al. , 2001 ; Gogarten et al. , 2002 ). The identification of dsrAB in members of bacterial ( Actinobacteria , Caldicerica , oxidative-type dsrAB in Deltaproteobacteria ) and archaeal ( Aigarchaeota ; formerly known as pSL4 or hot water crenarchaeotic group I candidate division ( Nunoura et al. , 2011 ); note that it is under debate whether Aigarchaeota members represent their own phylum or belong to the Thaumarchaeota ( Brochier-Armanet et al. , 2011 )) phyla previously not known to harbor these genes necessitates the re-evaluation of dsrAB as phylogenetic markers. Using an established phylogenetic approach ( Zverlov et al. , 2005 ; Loy et al. , 2009 ), we directly compared consensus trees of DsrAB and 16S rRNA sequences originating from 254 pure cultures and genome sequences for topological incongruences that are indicative of LGT. Owing to the apparent functional adaptation of DsrAB, this analysis was carried out separately for organisms using the reductive ( Figure 2 ) and the oxidative sulfur energy metabolism ( Figure 3 ). Our analysis confirms that reductive-type DsrAB and 16S rRNA branching patterns are generally similar and reproduces known topological inconsistencies regarding (i) a group of Firmicutes that most likely acquired dsrAB from deltaproteobacterial ancestors of the Desulfatiglans anilini (formerly Desulfobacterium anilini ; Suzuki et al. , 2014 ) lineage ( Figure 2 ) ( Klein et al. , 2001 ; Zverlov et al. , 2005 ), (ii) members of the phylum Thermodesulfobacteria and (iii) members of the euryarchaeotal genus Archaeoglobus that possess bacterial-type dsrAB . Besides these documented cases, we have obtained evidence for further possible dsrAB LGT events ( Supplementary Results and discussion and Supplementary Figure S6 ). The Aigarchaeota member clearly has a reductive-type dsrAB that was received either directly by LGT from a bacterial donor, possibly a member of the phylum Firmicutes , or indirectly from a yet unknown, bacterial dsrAB -containing archaeon ( Figures 1 and 2 ). The presence of bacterial dsrAB in the Aigarchaeota member and members of the genus Archaeoglobus seems to be the result of at least two independent LGT events. The stable monophyletic grouping of the actinobacterium Gordonibacter pamelaeae with the Firmicutes genera Desulfosporosinus and Desulfitobacterium in the DsrAB tree ( Figures 1 and 2 ) suggests LGT from an unknown donor. Although the deep, independent position of the Caldiserica phylum member in both the DsrAB tree and the 16S rRNA tree is inconclusive regarding LGT, complementary analyses also indicate a foreign origin of its dsrAB ( Supplementary Results and discussion and Supplementary Figure S6 ). These results provide a first view into the possible evolutionary paths that led to the presence of a reductive bacterial-type dsrAB in the bacterial phyla Actinobacteria and Caldiserica , and the archaeal candidate phylum Aigarchaeota , but in-depth insights can only be obtained when more dsrAB sequences from members of these phyla are available. Based on larger sequence data sets, we confirm that branching patterns of DsrAB and 16S rRNA trees of SOB are largely congruent ( Loy et al. , 2009 ), with one exception ( Figure 3 ). In the DsrAB tree, T. nitratireducens is not related to three other species of the genus Thioalkalivibrio , but branches independently from other Proteobacteria . One possible explanation for the phylogenetic position of DsrAB of T. nitratireducens is xenologous replacement of its orthologous dsrAB with dsrAB from an unknown and unrelated proteobacterial donor. A robust DsrAB consensus tree and knowing the discrepancies in 16S rRNA and DsrAB-based phylogenies of described taxa are important for a phylogenetically well-informed interpretation of dsrAB diversity in environmental samples. The detection of reverse dsrAB in a metagenome bin ( Sheik et al. , 2014 ) and single-cell genomes ( Swan et al. , 2011 ) of the deltaproteobacterial SAR324 clade, whose sulfide-oxidizing members are related to deltaproteobacterial SRM in the 16S rRNA tree, illustrates that inferring general physiological traits such as sulfate/sulfite reduction or sulfur oxidation from 16S rRNA phylogeny can be problematic. In contrast, DsrAB phylogeny clearly distinguishes oxidative versus reductive sulfur metabolism. Despite some limitations, dsrAB also remain useful phylogenetic markers because an environmental dsrAB sequence is identified with high certainty as a member of a recognized taxon if it clusters unambiguously within this taxon. In contrast, the taxonomic identity of an organism represented by an environmental dsrAB sequence that branches outside a recognized taxon, such as members of the 13 uncultured family-level DsrAB lineages, is uncertain. Environmental distribution of dsrAB -carrying organisms We further examined the environmental distribution of the 1292 core dsrAB sequences and of 6403 partial dsrA or dsrB sequences with a minimum length of 300 nucleotides. Owing to the many non-overlapping sequences, partial dsrA and dsrB sequences could not be jointly clustered into sequence identity-based species-level OTUs. Instead, they were phylogenetically placed into the consensus tree ( Figure 1 ) without changing its topology ( Figure 4 ). The majority of the 6403 shorter sequences is affiliated with described families and uncultured family-level lineages ( n =5893; 92%) or unclassified environmental sequences of the core data set ( n =409; 6%) ( Figure 1 ). Only few sequences ( n =101; 2%) do not branch within sequence clusters defined by the core data set. A large proportion ( n =2349; 35%) of the 6686 sequences on the reductive bacterial DsrAB branch are not affiliated with known taxa (i.e., families, genera) that are represented by cultured organisms. For example, uncultured family-level lineage 9 ( n =559) contains a similar number of sequences as the family Desulfovibrionaceae ( n =531) that harbors many, taxonomically well-described Desulfovibrio species ( Loy et al. , 2002 ; Muyzer and Stams, 2008 ). We additionally grouped the 7695 dsrAB sequences into eight broad environmental categories (i.e., marine, estuarine, freshwater, soil, industrial, thermophilic, alkali-/halophilic and symbiotic) ( Supplementary Materials and methods ) to gain insights into the environmental distribution patterns of members of major phylogenetic DsrAB lineages. Most sequences derive from marine environments (31%), followed by freshwater (24%), industrial (16%) and soil environments (11%). Members of most major DsrAB lineages are widely distributed among various environments with starkly contrasting biogeochemical properties, which provides limited indications of the possible ecological factors that gave rise to evolution of the many, phylogenetically distinct lineages at the approximate taxonomic rank of families ( Figure 4 ). However, there are notable exceptions that are indicative of environmental preference. Members of the uncultured family-level lineages 2, 3 and 4 are almost exclusively found in marine environments. Not surprisingly, sequences affiliated with the deltaproteobacterial families Desulfohalobiaceae and Desulfonatronumaceae , which include known halophilic and alkaliphilic SRM ( Ollivier et al. , 1991 ; Pikuta et al. , 2003 ; Jakobsen et al. , 2006 ; Sorokin et al. , 2008 ), derive predominately from high-salt and/or high-pH environments. Oxidative bacterial-type dsrAB sequences of Chlorobi are most often detected in freshwater habitats. This is, however, possibly a biased representation, as two studies of freshwater environments have provided 94% of all available Chlorobi dsrAB sequences. Microorganisms with archaeal-type dsrAB seem to be restricted to hot environments, but this also needs to be interpreted with caution, because of the low number of available environmental sequences from this DsrAB enzyme family. Analogous to marker genes for other functional guilds ( Mussmann et al. , 2011 ), detection of reductive and oxidative dsrAB or their transcripts in an environmental sample is not to be mistaken with the actual physiological capability for dissimilatory sulfate/sulfite reduction and sulfur oxidation, respectively ( Pester et al. , 2012 ). In addition to the presence of dsrAB in syntrophic bacteria, which are apparently incapable of using sulfate, sulfite or organosulfonates, environmental fragments of dsrAB might derive from viruses or virus-like particles that infect SRM ( Rapp and Wall, 1987 ; Walker et al. , 2006 ; Stanton, 2007 ) or SOB ( Anantharaman et al. , 2014 ), and thus possibly serve as vectors for LGT or supplement the sulfur metabolism of their microbial hosts. Although DsrAB has thus far been shown to function exclusively in dissimilation, it is conceivable that some organisms use it for detoxification of sulfite ( Johnson and Mukhopadhyay, 2005 ; Lukat et al. , 2008 ). A list of in silico -evaluated primers allows selection of best primer combinations for future environmental dsrAB surveys We evaluated the in silico coverage (i.e., the fraction of sequences in the target group that is matched by the primer) of 103 published, individual dsrAB -targeted primers and primer mixtures and 28 primer pairs against the updated dsrAB sequence database ( Supplementary Results and discussion and Supplementary Tables S1–S4 ). Although most primers are highly specific for dsrAB sequences, only few primers or primer mixes target ⩾50% (coverage of perfectly matched sequences) and/or ⩾90% (coverage of sequences with up to one weighted mismatch) of reductive- or oxidative bacterial-type dsrAB sequences ( Supplementary Tables S1 and S2 and Supplementary Figure S7 ). The forward primers DSR1Fmix a–h, DSR1728Fmix A–E, DSR67F, dsrB F2a–i and reverse primers DSR4Rmix a–g, DSR698R, dsrB 4RSI1a–f have highest coverage values for reductive bacterial-type dsrAB and are recommended for future use. Reverse dsrAB sequences are best covered by forward primers rDSR1Fmix a–c, rDSRA240F, DSR1728Fmix A–E, DSR874F, dsrB F1a–h and reverse primers rDSR4Rmix a–b, rDSRB808R, PGdsrAR and dsrB 4RSI2a–h. It is noteworthy that DSR1728Fmix A–E and dsrB F1a–h have relatively high coverage for both reductive- and oxidative bacterial-type dsrAB . Of the 28 previously published primer pair combinations ( Supplementary Tables S3 and S4 ), only nine have a good coverage of >75% when hits with up to one weighted mismatch are considered ( Table 1 ) and are recommended for further use. Primer pairs DSR1Fmix a–h/DSR4Rmix a–g (∼1.9 kb dsrAB PCR product) and DSR1728Fmix A–E/DSR4Rmix a–g (∼0.4 kB dsrB PCR product) have highest perfect-match coverage of 47% and 70%, respectively, for reductive bacterial-type dsrAB . Primer pairs rDSR1Fmix a–c/rDSR4Rmix a–b and DSR1728Fmix A–E/rDSR4Rmix a–b, which amplify the homologous regions in SOB, have even higher coverage of 97% and 90%, respectively. Separate coverage values obtained for the five major groups within the reductive bacterial-type DsrAB tree indicate that the DSR1F/DSR4R primer pair mix is biased against sequences of the Firmicutes group sensu lato and the Nitrospirae supercluster. While new primer variants should be designed to improve in silico coverage, already many environmental sequences belonging to these two groups were obtained by using the DSR1F/DSR4R primer variants ( Kaneko et al. , 2007 ; Martinez et al. , 2007 ; Leloup et al. , 2009 ; Wu et al. , 2009 ). For an improved coverage of the environmentally occurring dsrAB diversity by amplicon sequencing, it is therefore recommended to apply the aforementioned primer pairs at low stringency (e.g., low annealing temperature) to allow for binding of non-perfectly matching target sequences. This also promotes more uniform amplification by the different primers in a degenerate primer mixture ( Higuchi et al. , 1993 ). Amplification of complex environmental DNA extracts with highly degenerate primers ( Supplementary Tables S1–S4 ) at low stringency unfortunately increases the likelihood for unspecific PCR products ( Wagner et al. , 2005 ). This is particularly a problem if degenerate primers are applied for denaturing gradient gel electrophoresis, terminal restriction fragment length polymorphism or real-time PCR analyses. Hence, PCR performance/biases must be carefully evaluated for each primer combination individually, and specificity of amplification should additionally be confirmed by sequencing of the environmental PCR product or the extracted denaturing gradient gel electrophoresis bands. To assist researchers during the evaluation of existing and development of new dsrAB -targeted oligonucleotides, we have incorporated our database into the probeCheck webserver for straightforward in silico testing of primer specificity and coverage ( Loy et al. , 2008a )."
} | 8,003 |
36092273 | PMC9428804 | pmc | 8,787 | {
"abstract": "Microbial cell factories (MCFs) and cell-free systems (CFSs) are generally considered as two unrelated approaches for the biosynthesis of biomolecules. In the current study, two systems were combined together for the overproduction of agroclavine (AC), a structurally complex ergot alkaloid. The whole biosynthetic pathway for AC was split into the early pathway and the late pathway at the point of the FAD-linked oxidoreductase EasE, which was reconstituted in an MCF ( Aspergillus nidulans ) and a four-enzyme CFS, respectively. The final titer of AC of this combined system is 1209 mg/L, which is the highest one that has been reported so far, to the best of our knowledge. The development of such a combined route could potentially avoid the limitations of both MCF and CFS systems, and boost the production of complex ergot alkaloids with polycyclic ring systems.",
"introduction": "1 Introduction Ergot alkaloids display diverse biological activities and are often found in fungus-infected grains or fungi-associated plants. There are about over 100 ergot-alkaloid type natural products, and currently, six FDA-approved ones are on the market, including treatments for parkinsonian syndrome and vasoconstrictors [ 1 ]. Ergot alkaloid agroclavine (AC, Fig. 1 A) antagonizes depression that is induced by noradrenaline in the cerebral cortex of the rat [ 2 ] and was once used for ergot-based drug synthesis via the oxidation of the top methyl group [ 3 ]. Ergot alkaloids generally possess complex conjugated rings with several stereocenters ( Fig. 1 A), which led to their total synthesis being very difficult and not at all profitable [ 1 ]. The biosynthesis of ergot alkaloids requires tryptophan, Dimethylallyl pyrophosphate (DMAPP), and S -Adenosyl methionine (SAM) [ 5 , 16 ]. Tryptophan in fungus is synthesized by a shikimic acid pathway containing 6 enzymes and DMAPP is generated by the mevalonate pathway containing 7 enzymes, whereas the SAM regeneration cycle requires 3 enzymes ( Fig. 1 A). Starting from those precursors, three prenylated-tryptophan derivatives, dimethylallyltryptophan (DMAT), N -methyldimethylallyltryptophan ( N –Me-DMAT), and prechanoclavine (PCC) are generated under the successive catalysis of prenyltransferase DmaW, methyltransferase EasF, and FAD-linked oxidoreductase EasE. Two additional enzymes chanoclavine synthase EasC and FAD-linked dehydrogenase EasD catalyze the biosynthesis of chanoclavine (CC) and chanoclavine-aldehyde (CCA), the two common tricyclic precursors for all ergot alkaloids. Finally, aldehyde dehydrogenase EasA and agroclavine dehydrogenase EasG complete the biosynthesis of AC. Overall, 22 enzymes are involved in the biosynthesis of AC in the native host ( Fig. 1 A). In conclusion, ergot alkaloids employ a large number of enzymes for the biosynthesis of precursors as well as tailoring modifications [ 4 , 5 ], which led to their construction of microbial cell factories (MCFs) and cell-free systems (CFSs) difficult. Fig. 1 The biosynthetic pathway of AC. (A) The biosynthetic pathway of AC in the native hosts. (B) the combined system designed for overproduction of AC in the current study. Fig. 1 MCFs and CFSs both have been developed to overproduce natural biomolecules, which are two alternative approaches of chemical synthesis as well as natural extractions. The titers of products like vitamins, amino acids, and a few natural products have reached the industrial standard by using environmentally friendly and sustainable MCFs [ 6 , 7 ]. Despite this, it remains problematic to obtain high titers for many complex natural products due to i) limited precursor supply, ii) uncontrollable native metabolic network, iii) inefficient product transportation caused by the cell membrane, and iv) low expression and low activity of foreign enzymes. Indeed, accompanied by the complexity of the biosynthetic pathways, the difficulties of the construction and engineering of MCFs for natural biomolecules become exponential growth. A recent work, which takes over 20-year works and invests millions of dollars, reached the industrial-scale production of artemisinic acid [ 8 ]. Whereas for many other biomolecules, including morphine [ 9 ], cannabidiols [ 10 ], and tropane alkaloids [ 11 ], their titer is still below 10 mg/L in the yeast, and further engineering is essential for their industrial application. CFSs separate the enzyme synthesis from the following catalysis, and thus could effectively avoid the limitations caused by cell membranes, fermentations, and uncontrolled native metabolic networks in MCFs [ 12 ]. Benefitted from the absence of cell membranes, CFSs allow for easy manipulation and optimization [ 12 ]. In addition, CFSs can achieve higher product titers, faster reaction rates, and better tolerance of toxic precursors. There are two types of CFSs based on the methods of preparations: cell extract-based and purified enzyme-based [ 13 , 14 ]. However, commercially CFS for complex biomolecules is rare, until Merck recently reported the commercial development of a nine-enzyme cascade for the synthesis of islatravir [ 15 ]. CFSs generally require many enzymes, which could be logistically challenging for scaling. This combined with the expensive cofactors and starting materials, also makes CFSs difficult to bring costs down. To avoid the limitations that arise from both MCFs and CFSs, and to take full advantage of both systems, we developed a combined system to synthesize clinically important and structurally complex AC in this study ( Fig. 1 B). We split the enzymes of its biosynthetic machinery into two sections, which were transformed into an MCF ( Aspergillus nidulans ) and a CFS system, respectively ( Fig. 1 B). The key to this designed MCF-CFS combined system is determining its splitting point for the construction of two systems. The combined system was demonstrated to efficiently synthesize AC, with the titer up to 1209 mg/L and an engineering period of up to 3 months. Thus, the combined MCF-CFS system represents a fast, robust as well as practical engineering methodology for laboratory use as well as industrial application.",
"discussion": "4 Discussion The combined system established in this study seems to be superior compared to either the MCF system or the CFS system. The non-accessible soluble protein of EasE as well as the expensive cofactors, substrates, and a large number of enzymes required make the overproduction of ergot alkaloids via the CFS system difficult to achieve. In accord with this, no such studies have been reported for ergot alkaloids or AC ( Fig. 5 ). On the other hand, the highest production of AC using the MCF systems is up to 79 mg/L ( Fig. 5 ), which is not profitable for industrial production. To overcome those dilemmas, a combined MCF ( A. nidulans )-CFS system was developed for the overproduction of AC in the current study, which led to the highest titer (1209 mg/L, Fig. 5 ) in the literature to the best of our knowledge. Fig. 5 The overall titers (mg/L) of AC generated from different strategies. Fig. 5 Currently, over half of the ergot alkaloids were still produced by a field-production mode, while the ergot alkaloids isolated from ergot fermentation can be impure because of the abundance of similar metabolites. These difficulties are exacerbated for ones with a low natural abundance, such as chanoclavine and AC. Thus, developing a new mode to synthesize ergot alkaloids could be beneficial for the production of ergot-based pharmaceuticals. We apply a A. nidulans -based MCF to overproduce PCC at the titer of 2050 mg/L in the bioreactor, followed by the CFS system based on a four-enzyme cascade to overproduce the tetracyclic ergot alkaloid AC at the titer of 1209 mg/L. Future optimization of this combined system would further improve the AC production and accelerate the industrial production of this clinically important natural product. The concept of this new combined MCF-CFS system, with picking a suitable splitting point as the key step, could be further applied for the biosynthesis of more structurally complex natural products and pharmaceutical molecules, which are currently inaccessible through either the MCF or CFS systems."
} | 2,049 |
35756008 | PMC9221998 | pmc | 8,789 | {
"abstract": "We investigated the fungus Aspergillus fumigatus PD-18 responses when subjected to the multimetal combination (Total Cr, Cd 2+ , Cu 2+ , Ni 2+ , Pb 2+ , and Zn 2+ ) in synthetic composite media. To understand how multimetal stress impacts fungal cells at the molecular level, the cellular response of A. fumigatus PD-18 to 30 mg/L multimetal stress (5 mg/L of each heavy metal) was determined by proteomics. The comparative fungal proteomics displayed the remarkable inherent intracellular and extracellular mechanism of metal resistance and tolerance potential of A. fumigatus PD-18. This study reported 2,238 proteins of which 434 proteins were exclusively expressed in multimetal extracts. The most predominant functional class expressed was for cellular processing and signaling. The type of proteins and the number of proteins that were upregulated due to various stress tolerance mechanisms were post-translational modification, protein turnover, and chaperones (42); translation, ribosomal structure, and biogenesis (60); and intracellular trafficking, secretion, and vesicular transport (18). In addition, free radical scavenging antioxidant proteins, such as superoxide dismutase, were upregulated upto 3.45-fold and transporter systems, such as protein transport (SEC31), upto 3.31-fold to combat the oxidative stress caused by the multiple metals. Also, protein–protein interaction network analysis revealed that cytochrome c oxidase and 60S ribosomal protein played key roles to detoxify the multimetal. To the best of our knowledge, this study of A. fumigatus PD-18 provides valuable insights toward the growing research in comprehending the metal microbe interactions in the presence of multimetal. This will facilitate in development of novel molecular markers for contaminant bioremediation.",
"conclusion": "Conclusion In this study, we found the fungus A. fumigatus PD-18 developed stress coping strategies by secreting a suite of proteins that were either unique or upregulated/overexpressed when compared to the control. The proteomics study revealed that maximum proteins that were upregulated belonged to KOG class translation, ribosomal structure, and biogenesis. This study also highlighted the enhanced expression of antioxidants superoxide dismutase, molecular chaperone heat shock proteins, and involvement of proton transporter, such as ATPase. These proteins are involved in the tolerance and detoxification of multimetals by the fungus. A. fumigatus PD-18. Therefore, this investigation on the response of cellular proteomes to multimetal stress enabled us to better understand the cellular mechanism regarding the cumulative effect of the inorganic heavy metals stress on microbes. Further, it will be conducive to screening the key genes coding for enzymes that have higher resistance to these inorganic pollutants with enhanced capability to transform pollutants.",
"introduction": "Introduction Heavy metals, such as copper (Cu), chromium (Cr), cadmium (Cd), zinc (Zn), lead (Pb), and nickel (Ni), occur in the water bodies, such as river water, and drains, in developing countries that are above the permissible mandates as prescribed by the Food and Agriculture Organization ( Bhattacharya et al., 2015 ). These heavy metals are released from numerous small- and medium-scale enterprises (pesticide, textile, electroplating, fertilizer, batteries, etc.) due to lack of or improper sewage treatment plant systems ( Bhattacharya et al., 2015 ; Bhardwaj et al., 2017 ; Gola et al., 2020 ). These hazardous contaminants are discharged into various water bodies via irrigation during agricultural activities. This results in bioaccumulation of these contaminants that enter the food chain and cause detrimental health ailments, such as cancer in human beings ( Kou et al., 2018 ; Li et al., 2019 ; Abed et al., 2020 ; Chauhan et al., 2021 ; Irawati et al., 2021 ). Thus, these heavy metals are the environmental priority contaminants threatening the environment and therefore must be remediated before discharge into the environment. The conventional physico-chemical methods, such as chemical precipitation, ion exchange, adsorption, membrane filtration, coagulation-flocculation, and flotation, are usually utilized to remediate these harmful contaminants. These techniques have the disadvantages of being expensive, having low selectivity, production of additional sludge, and further treatment is required for better results ( Dey et al., 2021 ; Zamora-Ledezma et al., 2021 ). On the other hand, living and actively growing microbial cells can be a lucrative option for bioremediation ( Malik, 2004 ). Of the bioremediation techniques, mycoremediation has shown to be a promising technology having the potential to ameliorate these hazardous chemicals ( Dey et al., 2020 ; Sabuda et al., 2020 ). The superiority of fungi over single-celled microbes, such as bacteria, to remediate these recalcitrant heavy metals is well-documented ( Deshmukh et al., 2016 ). In addition, the fungi are omnipresent, multifarious, and have a wider arsenal to acclimatize to environmental limitations, such as immoderations of temperature, extremes of pH, higher metal concentrations, and low nutrient accessibility, due to their morphological diversity ( Anand et al., 2006 ). Besides, fungal mycelia have enhanced enzymatic and mechanical contact with the pollutant due to a greater cell to the surface ratio ( Sagar and Singh, 2011 ). Further, in microbial cells, such as fungus, heavy metals are key components in the number of catalytic and structural proteins that are integral to biochemical processes. The outcome of these processes differs depending on the type of metal involved and its concentration inside the cell ( Gadd, 1994 ). Particularly, the species of Aspergillus have a high metal uptake capacity for metals, such as Cu ( Dusengemungu et al., 2020 ). Also, filamentous fungi develop signature metabolic pathways which are species-specific to survival in the harsh environment of heavy metals and other contaminants that are utilized as nutrients and energy sources ( Kuhn and Käufer, 2003 ). Accordingly, proteins expressed in cells under such diverse conditions and at different times are dissimilar ( Boopathy, 2000 ; Keller, 2015 ; Wisecaver and Rokas, 2015 ). Also, there is evidence that fungal resistance toward one element does not necessarily infer resistance to another element even though the elements possess similar valency charges ( Høiland, 1995 ). For example, the toxicity of Pb toward the microbe is less compared to other toxic metals, such as Cd, As, and Hg ( Jiang et al., 2020 ). Further, in fungi, the metallothioneins formation is predominantly induced by the heavy metal Cu ( Jaeckel et al., 2005 ). Thus, understanding the altered heavy metal uptake in the fungus in presence of diverse environmental conditions needs to be evaluated. To gain such mechanistic insight, high throughput techniques, such as proteomic analysis by LC-MS/MS, can be used to identify and characterize proteins involved in the multimetal resistance mechanism in a filamentous fungus ( Pandey and Mann, 2000 ; Kraut et al., 2009 ; Zhang et al., 2010 ). Moreover, proteomics facilitates the development of new and important protein biomarkers that specify and monitor metal contamination in the environments as protein type and also their copy numbers are estimated by the translational regulation ( Ohno et al., 2014 ). Several researchers have studied the change in proteomes under various metal conditions. Selamoglu et al. (2020) have assessed the change in the differential proteome expression of the prokaryotic organism Rhodobacter capsulatus in presence of 5 μM Cu by nano-LC-MS/MS. About 75 proteins were significantly regulated. Most of the proteins present were responsible for maintaining Cu homeostasis. Further, Lotlikar et al. (2020) enumerated the variable expression of proteins in a eukaryotic fungus Penicillium chrysogenum in the presence of 100 mg/L and 500 mg/L Cu. Several key proteins related to genetic information, carbohydrate metabolism, glycan biosynthesis and metabolism, amino acid metabolism, and energy metabolism were expressed. Cherrad et al. (2012) highlighted the overaccumulation of proteins of the oxidoreductase family when exposed to Cd, Cu, and Ni but not when exposed to Zn. Thus, the secretion of proteins in fungus is extremely dynamic and its production depends on different environmental triggers. However, despite these growing proteomic studies dealing with the bioremediation of single heavy metal, there is a dearth of information on the modulated proteins triggered by the cumulative toxicity of hexametals, namely, Cd, Cr, Cu, Ni, Pb, and Zn, specifically in a filamentous fungus. Interestingly, Aspergillus fumigatus has high adaptability when subjected to an altered environment ( Bakti et al., 2018 ). Also, as established in our previous works ( Dey et al., 2016 , 2020 ; Bhattacharya et al., 2020 ), A. fumigatus PD-18 is a filamentous fungus capable of removal of 30 mg/L multimetal exceptionally well. Furthermore, there are structural and functional similarities between the numerous genes of lower eukaryotes, such as fungi and mammals ( Bae and Chen, 2004 ). Thus, this fungus A. fumigatus PD-18 would be a good eukaryotic model for helping us understand how cells adopt various cellular strategies and lay a foundation study to decipher the enzymes produced in the presence of simultaneous effects of multimetal cocktail on interaction with a fungus for scale-up process.",
"discussion": "Discussion In this study, we identified the chief classes of proteins that are crucial for the resistance and tolerance of A. fumigatus PD-18 for multimetals, including intracellular and extracellular mechanisms of metal uptake. These two processes are further elaborated. Intracellular and Extracellular Mechanism of Metal Uptake by Fungus The mechanism of the fungal resistance and tolerance could be attributed to its occurrence at the metal-contaminated site ( Baker, 1987 ; Gadd and White, 1993 ). The fungal detoxification process involves strategies, such as intracellular bioaccumulation, extracellular precipitation, biotransformation, biomineralization, and biosorption, that involve several signaling pathways ( Wang et al., 2019 ). In addition, the reduction of heavy metals from the cell comprises of activation of different metabolic processes in the cell that are stimulated by heavy metals ( Goyal et al., 2003 ). Fungi largely counter heavy metals in two ways. The first way involves averting the metal uptake and its passage inside the fungal cell. This happens chiefly by a decrease in metal uptake or increased efflux of metals, metal biosorption to the impermeable cell walls by metal binding polysaccharides, peptides, extracellular formation of complexes, and the release of organic acids that chelates heavy metals outside the cell. Thus, the extracellular mechanism operates in the cell to counter heavy metals by circumventing the entry of heavy metals. Particularly, the secondary metabolites, such as citric acid, oxalic acids, succinate, and fumarate, which are low molecular weight compounds (<900 daltons) are secreted by fungal species namely A. niger specifically in response to heavy metals exposure and they bind the heavy metals extracellularly ( Kolen, 2013 ). Also, the secondary metabolite oxalic acid is produced as an intermediate compound in the biochemical tricarboxylic acid cycle (TCA) ( Munir et al., 2001 ). Here, we found the upregulation of the key enzymes that are involved in the TCA cycle viz. phosphoglycerate kinase (2.91-fold), glyceraldehyde-3-phosphate dehydrogenase (2.38-fold), enolase (1.8-fold), and pyruvate kinase (1.86-fold) which mediates the enhanced secretion of oxalic acids. In the second way, the fungus subsists the high concentration of metals inside the cell by tolerance after the process of detoxification via metal chelation by synthesizing ligands, such as metallothioneins and phytochelatins, that bind heavy metals intracellularly or by compartmentalization of heavy metals within the cell organelles of vacuoles by polyphosphates. The three main classes of intracellular peptides binding metal ions are phytochelatins (PCs), metallothioneins (MTs), and glutathione (GSH). MTs are low molecular weight cysteine-rich metal-binding proteins that have high affinity toward both the essential metal ions, such as Cu and Zn, non-essential metal ions, such as Cd, Hg, and Ag, and also have large metal-binding capacities ( Reddy et al., 2014 ). Further, MTs chelate heavy metals by forming thiolate bonds with the heavy metals. Glutathione S transferases (GSTs) are enzymes that metabolize heavy metals and other contaminants by catalyzing the binding of glutathione to non-polar compounds comprising of electrophilic nitrogen, carbon, and sulfur atom ( Morel et al., 2009 ). Usually, the metals Cd, Cu, Pb, and Zn are removed via glutathione (GSH)-mediated sequestration. However, in this study, there was no evidence of the production of proteins glutathione, metallothioneins, and phytochelatins despite the presence of these heavy metals Cd, Cu, Pb, and Zn in the multimetal mixture. The reason for this phenomenon could be attributed to the dynamics of individual heavy metals when present in a mixture. As the expression of GST is related to the type of heavy metal, its concentration, and the extent of treatment time of the heavy metal ( Shen et al., 2015 ). Further, it is reported that heavy metals induce oxidative damage to the cell membranes of fungi by the generation of reactive oxygen species (ROS). These ROS are detoxified by the production of antioxidants that are components from the thioredoxin system, such as peroxiredoxins, NADPH dehydrogenases, catalase, superoxide dismutase, and peroxidase, that enables the fungus to confront the reactive-oxygen species that accumulate in the cell on exposure to the metals ( Zhang et al., 2015 ). Thus, in principle, intracellular mechanisms decrease the metal load in the cytosol ( Sandau et al., 1996 ; Gadd, 2000 , 2007 , 2010 ). In this study, we found 3.45-fold upregulation of the antioxidant protein Cu-Zn superoxide dismutase. Other important functional groups were detected that expressed amino acid metabolism, lipid metabolism, energy metabolism, and also the proteins involved in signal transduction, transcription, translation, or DNA repair. In general, the upregulated proteins are stimulated to display the fungal resistance against the contaminants’ stress, while the downregulated proteins are suppressed by the action of the pollutants’ toxicity. The highly upregulated protein hydroxymethylglutaryl-CoA synthase with 5.19-fold upregulation is responsible for the production of secondary metabolite carotenoid from the precursor molecule of acetyl-CoA when stimulated by the heavy metal stress ( Bhosale, 2004 ). The protein serine/threonine phosphatase with 3.42-fold upregulation is responsible for maintaining the conformation of cell organelles and proteasomes ( Dias et al., 2019 ). The other proteins, such as G-protein beta subunit SfaD , with 2.98-fold upregulation depicted the role of G-protein-coupled receptors (GPCRs) in heavy metal bioremediation. These are the largest transmembrane receptors that aid in communicating the extracellular signals, such as stresses of heavy metals into the intracellular sites. G-protein-coupled receptors (GPCRs) regulate the important effector molecules, such as adenylate cyclase and phospholipase C, and regulate the function of kinase and ion channel by producing secondary messengers, such as cAMP, thereby inducing signaling cascades ( El-Defrawy and Hesham, 2020 ). From Figures 2A , discussion of the highly regulated selected KOG classes are as follows: Post-translational Modification, Protein Turnover, Chaperones Protein homeostasis is crucial for the cell proliferation and viability of all organisms. Further, cellular signaling is greatly affected by the protein homeostasis under different physiological conditions and environmental stresses, such as heavy metals, and therefore they can be suitable biomarkers. Molecular chaperones aid in the delivery of metal ions to the cell organelles and metalloproteins ( Lotlikar and Damare, 2018 ). The two vital regulators of molecular chaperones in the proteostasis network are heat shock transcription factor Hsf1 and heat shock protein Hsp90. Under a stressful environment, heat shock protein enables the folding of newly synthesized proteins and helps in the degradation of damaged/misfolded proteins with the help of the ubiquitin-proteasome system ( Hossain et al., 2020 ). Heat shock proteins bind to the denatured proteins, compelling them to refold into their native conformation and regain their original structure ( Feng et al., 2018 ). In this study, we observed 42 proteins upregulated that belonged to post-translational modification, protein turnover, and chaperones. There was upto 4-fold upregulation of proteasome regulatory particle subunit Rpt3 (KOG0727), upto 2-fold upregulation of Hsp70 chaperone ( HscA ) (KOG0101), and upto 2.5-fold upregulation of protein geranylgeranyltransferase (KOG1439). The enzyme geranylgeranyltransferase I (GGTase I) aids in the catalysis of the post-translational transfer of lipophilic diterpenoid geranylgeranyl molecule to the cysteine residue of proteins with the termination at CaaX motif ( Rho1p and Cdc42p ). This alteration helps in the membrane localization of the protein and thereby rendering it biologically active. Rho1p is a regulatory subunit of 1,3-β-D-glucan synthesis and contributes to the cell wall synthesis in fungi which is vital for cell viability under stressful condition of excess metals ( Singh et al., 2005 ). Translation, Ribosomal Structure, and Biogenesis Different proteins related to protein translation under multimetal stress were overexpressed. Here, we found 60 proteins of translation, ribosomal structure and biogenesis upregulated. There was upto 3.0-fold increase in glutamyl-tRNA synthetase (KOG1147), upto 2.5-fold increase in eukaryotic translation initiation factor 3 subunit B (KOG2314), upto 2.0-fold increase in 60S ribosomal protein L23 (KOG1751), upto 2.3-fold increase in mitochondrial translation initiation factor IF 2 (KOG1144), and 2.0-fold increase 40S ribosomal protein S10b (KOG3344). Similar elements of protein synthesis, such as translation initiation factor 5A, elongation factor 2, 40S and 60S ribosomal proteins, ATP-dependent RNA helicase, and aspartyl-tRNA synthetase, were overexpressed in Phanerochaete chrysosporium under Cu stress as a result of the need for production of new proteins or renewal of the damaged proteins ( Okay et al., 2020 ). Intracellular Trafficking and Secretion and Vesicular Transport Proteins such as ion transporters and other solutes are crucial for processes such as detoxification, cell nutrition, cell signaling, cellular homeostasis, and resistance toward metal stress. These polytopic transmembrane proteins are translated altogether and folded in the endoplasmic reticulum (ER) of the eukaryotic cells that are later ultimately arranged to their respective membrane location through vesicular secretion. During any physiological or stressful environment, transporters undertake several regulated turnovers. Thus, in the process, transporters briefly interact dynamically with multiple proteins ( Dimou et al., 2021 ). In this study, we found 18 proteins of intracellular trafficking and secretion and vesicular transport upregulated. The levels of the proteins were expressed in higher amounts (SEC31) (KOG0307) by 3.3-fold, endosomal cargo receptor (P24) (KOG1692) by 2.9-fold, and mitochondrial inner membrane translocase (KOG2580) by 2.4-fold in A. fumigatus under the effect of multimetal stress. The vesicle (Ves) are tissues composed of a lipid bilayer whose size varies ∼nanometers to micrometers. The Ves structures fuse with the plasma membrane of the cell and eject the trapped materials either inside or outside of the cytoplasm. There are three types of intracellular Ves viz. protein complex I (COPI)-coated Ves, protein complex II (COPII)-coated Ves, and BAR-domain protein Ves. These Ves proteins aid in physiological processes, such as the exchange of proteins and RNA intercellularly ( Jiang et al., 2020 ). Energy Production and Conversion Heavy metals, such as Cd, Cu, Ni, and Zn, function as cofactors in bacteria and fungi. However, excess amounts of these metals are toxic to these cells and also produce reactive oxygen species ( Liu et al., 2017 ). The need for metabolic energy in the fungal increases during abiotic stress, such as exposure to excess heavy metals. Thus, ATPases are responsible for the biochemical and physiological processes by the production of energy. Heavy metal ATPases (HMAs) or P-type ATPases can be categorized into three groups namely, A, B, and C. Further, P-type ATPases are utilized by numerous organisms to facilitate the transport of cations viz. Na + , K + , and Ca 2+ . To eliminate these excess metals, fungal HMA Saccharomyces cerevisiae CCC2 (Group A) localizes metals to metal-containing proteins, for example, in the case of copper metal, copper-containing protein FET3 in trans-golgi compartment transports metals to the cell membrane via efflux pumps, such as cadmium efflux pumps, encoded by fungal HMA Saccharomyces cerevisiae PCA1 (Group B and Group C) ( Adle et al., 2007 ). Saitoh et al. (2009) studied the CCC2-type HMA gene that targets copper-containing proteins from the fungus Cochliobolus heterostrophus by cloning. There was upto 0.78-fold upregulation in the production of V-type ATPase and upto 1-fold upregulation in the production of mitochondrial ATPase subunit ATP4. This gene has other multifarious roles, such as in the formation of dark brown colored melanin pigment located in fungal cell walls, that also sequester metals ( Chang et al., 2019 ). Amino Acid Transport and Metabolism The nitrogen cycle is essential for nitrogen assimilation and transformation and also for stress tolerance. Heavy metals impact the enzymes that play important role in nitrogen metabolism ( Khouja et al., 2014 ). There was upto 4-fold upregulation in the production of aminotransferase (KOG1549) in response to multimetal. A similar response was observed by Okay et al. (2020) , where the production of enzyme aspartate aminotransferase was enhanced in the fungus Phanerochaete chrysosporium to tackle the Cu stress. This enzyme has a possible role in the renewal of the mitochondrial NAD/NADH imbalance. Analysis of Biological Pathways and Protein–Protein Interactions In addition to these mechanisms, there are contributions from other regulatory systems, such as cross-talks in various pathways, interconnection amongst these different pathways, and regulation of different genes. These regulatory networks of the microbial proteins are intricate and play crucial roles in resistance to metal contaminants by modifying the series of specific functional proteins/non-proteins and altering the different metabolic enzymes at the cellular level. The metabolic processes related to the detoxification of contaminants are usually regulated by the complete set of proteins and their networks instead of a single enzyme ( Zhao and Poh, 2008 ; Feng et al., 2018 ). The STRING analysis displayed protein–protein interaction networks that are related to the resistance and tolerance mechanism A. fumigatus PD-18 for multimetals. Protein interactions in the first cluster (green) are involved in ribosome and preinitiation factors (60S and 40S ribosomal proteins). Ribosomal proteins form the protein part of ribosomes and participate in protein synthesis in cells in conjunction with rRNA. Thus, the increased expression of the large subunit of ribosome renders resistance against abiotic stresses, such as heavy metals, radiation, cold, and salt ( Liu et al., 2014 ). The important hub proteins expressed were cytochrome-c oxidase that is mitochondrial proteins and catalyst in the electron transport chain and is responsible for the transport of heavy metals particularly copper ( Dias et al., 2019 ). The second cluster (blue) had hub protein ATP synthase. The important interconnecting protein in the network is the septin protein of KOG class cell cycle control, cell division, and chromosome partitioning, and is involved in vesicular trafficking and countering the apoptotic cell death initiated by the toxicity of heavy metals. This energy-intensive process is mediated by ATP synthase to maintain the cellular structure and function under the lethal environment of heavy metals. Excess to or above the permissible limit of heavy metals exposure can substantially delimit the growth of organisms. Here, more growth is initiated in the fungal cell as a result of the enhanced activity of ATP synthase ( Yıldırım et al., 2011 ). This was evident in our study by the upregulation of tubulin by 1.1-fold and actin proteins by 2.2-fold which are responsible for cellular division and growth ( Marks et al., 1986 ). This enhanced growth in the fungus also corroborated with our previous study where there was an increased dry weight of the fungal biomass under 30 mg/L multimetal ( Dey et al., 2016 ). These results showed that the proteins in this network played important functions in cell functioning under heavy metals stress. The mechanism of hexametal uptake by A. fumigatus PD-18 has been summarized in Figure 4 ."
} | 6,405 |
20033048 | null | s2 | 8,792 | {
"abstract": "Sequencing of bacterial and archaeal genomes has revolutionized our understanding of the many roles played by microorganisms. There are now nearly 1,000 completed bacterial and archaeal genomes available, most of which were chosen for sequencing on the basis of their physiology. As a result, the perspective provided by the currently available genomes is limited by a highly biased phylogenetic distribution. To explore the value added by choosing microbial genomes for sequencing on the basis of their evolutionary relationships, we have sequenced and analysed the genomes of 56 culturable species of Bacteria and Archaea selected to maximize phylogenetic coverage. Analysis of these genomes demonstrated pronounced benefits (compared to an equivalent set of genomes randomly selected from the existing database) in diverse areas including the reconstruction of phylogenetic history, the discovery of new protein families and biological properties, and the prediction of functions for known genes from other organisms. Our results strongly support the need for systematic 'phylogenomic' efforts to compile a phylogeny-driven 'Genomic Encyclopedia of Bacteria and Archaea' in order to derive maximum knowledge from existing microbial genome data as well as from genome sequences to come."
} | 322 |
37336878 | PMC10279756 | pmc | 8,793 | {
"abstract": "The performance of any engineering material is naturally limited by its structure, and while each material suffers from one or multiple shortcomings when considered for a particular application, these can be potentially circumvented by hybridization with other materials. By combining organic crystals with MXenes as thermal absorbers and charged polymers as adhesive counter-ionic components, we propose a simple access to flexible hybrid organic crystal materials that have the ability to mechanically respond to infrared light. The ensuing hybrid organic crystals are durable, respond fast, and can be cycled between straight and deformed state repeatedly without fatigue. The point of flexure and the curvature of the crystals can be precisely controlled by modulating the position, duration, and power of thermal excitation, and this control can be extended from individual hybrid crystals to motion of ordered two-dimensional arrays of such crystals. We also demonstrate that excitation can be achieved over very long distances (>3 m). The ability to control the shape with infrared light adds to the versatility in the anticipated applications of organic crystals, most immediately in their application as thermally controllable flexible optical waveguides for signal transmission in flexible organic electronics.",
"introduction": "Introduction The speed and amount of transfer of information by optical means is one of the cornerstones of the current societal development and will very likely guide the progress of humanity in the near future. Among the variety of materials that are being considered as optical signal transducers, dynamic organic crystals 1 – 7 that have evolved into ‘crystal adaptronics’, an emerging field in materials science research 8 , have only recently garnered attention as an alternative to the traditional silica-based fiber-optics with low density, reduced scattering, flexibility 9 – 12 , and polarizing ability 13 , 14 as some of their most favorable assets. The realization that organic crystals can be used as transmitters of light has quickly led to attempts to fabricate crystal-based replicas of various commonly used optical devices and circuits 15 , 16 , an effort that capitalizes on the mechanical adaptability of some organic crystals 17 . Yet another direction of the currently very active research pursuit in this field is the attempt to diversify the range of stimuli that can be used to achieve spatial control over the optical output, which can be delivered both in the visible (see below) and near-infrared (telecom) range 18 , 19 . The available approaches to reshape crystalline optical waveguides include the application of mechanical force 20 , light 21 , 22 , humidity 23 , and magnetic field 24 . Specifically, UV light can be used to drive a photochemical reaction in the crystal, generating internal strain that translates into a bending moment 25 , 26 . Although these photochemically driven deformations can be quite significant in the extent of bending or curling, the processes of amplification of the strain caused by the photochemical reaction at a molecular scale to macroscopic bending are known to be kinetically inefficient, and such processes generally are not thought to qualify for actuating applications 27 . Moreover, the high energy of the UV light may generate chemically reactive species within the structure that could undergo side reactions, and over a prolonged period, they inevitably result in operational fatigue. Much faster deformations have been recently demonstrated by using photothermal effects caused by local heating with light; however, they result in minuscule deformations, typically on the order of less than 1° 28 . Considering that organic crystals are generally poor thermal conductors, in an attempt to achieve thermal control over the optical output, we resorted to exploring hybrid materials where they are coupled with an efficient thermal absorbent. Herein, we report a simple and efficient method for preparing hybrid organic crystals capable of responding to infrared light. The crystals were coated with a 2D compound from the MXene family, Ti 3 C 2 T x (T stands for surface terminal groups, including OH, O or F), which is known for its high efficiency in photothermal conversion 29 and has been widely applied for biological, medical, and other purposes 30 – 35 . Similar to the MXene, the hybrid organic crystals were found to be thermally absorbing and to undergo thermomechanical motion when exposed to infrared radiation. The photothermal actuation of the hybrid crystals reported here comes with multiple advantages over other methods that have been used for crystal deformation. First, since the absorber is the MXene, this approach does not require the absorption of radiation by the crystal itself, which circumvents the necessity for the crystal to absorb light (i.e., be photoreactive). Second, both high speeds and strong deformations can be achieved, and the crystal can be actuated by localized thermal excitation at a predetermined position. Finally, we demonstrate that hybrid crystals can be precisely actuated by excitation over very long distances (>3 m), which brings an added value to the potential of dynamic responsiveness of organic crystals for real-world applications, where they could be used as receivers for remote sensing, triggering, or actuation. Within a broader context, our study demonstrates that the thermal excitation can be applied to control optoelectronic elements made of these hybrid crystals, such as, for example, optical waveguides, and this concept can be expanded to control ordered arrays of optically waveguiding dynamic crystals.",
"discussion": "Results and discussion Preparation of the hybrid organic crystal The MXene used here has a typical two-dimensional lamellar structure 36 , which is often used to prepare transparent MXene multilayers 37 . The diluted aqueous suspension of the MXene appears black in color, corresponding to the strong absorption in its UV–vis spectrum between 300 and 900 nm, with a characteristic absorption peak around 760 nm (Supplementary Fig. 1 ). MXene multilayers were fabricated by using the layer-by-layer (LbL) assembly technique, by alternative deposition of positively charged polydiallyldimethylammonium (PDDA) and negatively charged MXene nanosheets onto a solid substrate 38 . This resulted in multilayered structures (PDDA/MXene) n , which can have an arbitrary number of deposited bilayers ( n ) and can be prepared on practically any type of water-insoluble substrate, including but not limited to glass, silicon, ceramics, metals, and plastics. In a typical case, a crystal of the organic compound 2,2′-((1 E ,1′ E )–1,4-phenylenebis(ethene-2,1-diyl))dibenzonitrile (for convenience, hereafter referred to as 1 ; Fig. 1a ) coated with 5 bilayers, 1 @(PDDA/MXene) 5 , had a thickness of about 200 nm and low roughness (average roughness, 20.1 nm) (Fig. 1b, c ; Supplementary Fig. 2 ) 39 . The mechanical properties of crystal 1 were tested by a three-point bending test (Supplementary Fig. 3 ). The (PDDA/MXene) n -coated crystalline hybrids are very transmissive to visible light, although expectedly, by increasing the number of deposited bilayers, the transmittance gradually decreases in the visible spectral region (≈400–780 nm, Fig. 1d, e ) 40 due to increased thickness. For instance, 1 @(PDDA/MXene) 5 has very high transmittance of 71% at 550 nm (Fig. 1e ). As shown in Fig. 1f , when exposed to infrared light, the temperature of the surface of 1 @(PDDA/MXene) 5 (Δ T ) rises with the increasing power of the infrared light up to Δ T = 61.2 °C under 744 mW infrared radiation, reaching a constant value within 10 s (Fig. 1g ). 1 @(PDDA/MXene) 5 @PDDA/poly(styrene sulfonate) (PSS) was selected to investigate the relationship between the thermal effect caused by the MXene and the surface temperature when the hybrid crystal is exposed to infrared light. As shown in Supplementary Figs. 4 and 5 and Supplementary Movie 1 , heat is transferred from the irradiated area to the surroundings due to the favorable thermal conductivity of the MXene. This raises the temperature of the surrounding crystal surface until it becomes constant after a few seconds. Furthermore, 1 @(PDDA/MXene) 1 @PDDA/PSS and 1 @(PDDA/MXene) 5 @PDDA/PSS were selected to demonstrate the effect of MXene thickness on the photothermal effect. As shown in Supplementary Fig. 6 , the temperature change of 1 @(PDDA/MXene) 5 @PDDA/PSS was about five-fold that of 1 @(PDDA/MXene) 1 @PDDA/PSS at the same power of the infrared light. These results indicate that 1 @(PDDA/MXene) 5 has excellent infrared light absorption properties. Fig. 1 Preparation and thermal properties of the MXene-polymer crystal hybrids. a Chemical structure of 1 . b Scanning electron microscopy (SEM) images of the surfaces of 1 @(PDDA/PSS) 5 and the 1 @(PDDA/MXene) 5 . The scale bar is 300 μm. c Atomic force microscopy (AFM) images of the surfaces of 1 @(PDDA/PSS) 5 and 1 @(PDDA/MXene) 5 . R q is the root mean square value of the profile deviation from the mean over the sampling length. d Photographs of different MXene layers on the 1 @(PDDA/PSS) n surface. ( n represents the number of layers of MXene). e UV–vis transmission spectra of 1 @(PDDA/PSS) 5 and 1 @(PDDA/MXene) 5 . f Increase of the temperature of the coated crystal (PDDA/MXene) 5 with increasing power of the infrared radiation (Δ T , 25 °C). g A time-dependent temperature increase of the surface of 1 @(PDDA/PSS) 5 and 1 @(PDDA/MXene) 5 upon illumination with infrared light (744 mW) at 25 °C. The dashed blue line separates the light regime (left) and the dark regime, i.e., absence of light (right). The source data is provided as a Source Data file. Encouraged by these initial results with 1 , the preparation method was slightly modified and applied to centimeter-long slender elastic crystals of three other organic compounds: 9,10-dibromoanthracene, ( Z )–2-([1,1′-biphenyl]–4-yl)–3-(anthracen-9-yl)acrylonitrile, and ( Z )–3-(furan-2-yl)–2-(4-((( E )–2-hydroxy-5-methylbenzylidene)amino)phenyl)acrylonitrile (for convenience, hereafter referred to as 2, \n 3 , and 4 , respectively, Fig. 2a ), which were obtained by using literature methods 23 , 41 , 42 . All these crystals are elastic and can be bent repeatedly into a U-shape without breaking. As shown schematically in Fig. 2b and Supplementary Fig. 7 , nascent (as-crystallized) crystals of 2 – 4 were first uniformly coated with a mixture of PDDA and a negatively charged PSS layer of ca. 650 nm thickness. The surfaces of the resulting hybrid crystals, 2 – 4 @PDDA/PSS (hereafter, 2 – 4 @P), were then coated with a mixture of (positively charged) PDDA and (negatively charged) MXene layer of ca. 200 nm thickness and described as 2 – 4 @PDDA/MXene@PDDA/PSS (for convenience, hereafter, 2 – 4 @P 2 ). Lastly, by using a needle tip, a 2 μm-thick layer of polyvinyl alcohol (PVA) with PSS, PVA/PSS, was deposited uniformly and rapidly along only one of the bendable faces and left to dry, a step that afforded a hybrid described as 2 – 4 @PVA/PSS@PDDA/MXene@PDDA/PSS (hereafter, 2 – 4 @P 3 ). Since only one of the two wide faces of the crystal is coated with the polymer, when the polymer shrinks, it gives rise to a differential strain that translates into a bending moment. This is observed as the macroscopic bending of the hybrid crystal. In the hybrid crystal, the MXene functions as a photothermal converter, while the PVA/PSS layer functions as one of the two components of a bilayer strip that generates bending moment by expansion or contraction. PVA is a common hygroscopic polymer that has a low critical solubility temperature and is well known to undergo reversible swelling via hydrogen bond formation (Supplementary Fig. 8 ) 43 . This brings about a response of the hybrid element to water vapor, and the curvature could change with variation in humidity (Supplementary Fig. 9 ). Due to the presence of the MXene all these hybrid crystals can be bent by infrared radiation. As shown with the example of 4 @P 3 in Supplementary Fig. 10 and Supplementary Movie 2 , while the crystals of 4 @PVA/PSS@PDDA/PSS practically do not show a response to infrared radiation at 184 mW, 4 @P 3 clearly bends under the same conditions. To further confirm that the bending of the hybrid crystals was a result of the thermally induced mechanical process and not a physical phase transition or a chemical reaction, NMR analysis of 2 ‒ 4 was performed before and after heating at 100 °C for 1 h, and the compounds were also analyzed by differential scanning calorimetry (DSC) (Supplementary Figs. 11 ‒ 14 ). The results ruled out phase transitions below 100 °C (note that one of the compounds undergoes a phase transition at 197 °C) or permanent chemical changes. In order to examine the effect of the coating with polymer and application of MXene on the mechanical properties, the stress‒strain profiles of the crystals 2, \n 3, \n 2 @P 3 , and 3 @P 3 were compared (Supplementary Fig. 15 ). The results confirmed that the mechanical properties of the nascent crystals were essentially retained in the hybrid crystals, with a very small change in the Young’s modulus. It is natural to expect that the crystal structure of the crystal determines its deformation. In line with this, as shown in Supplementary Fig. 16 , the bending degrees of the hybrid crystals 2 ‒ 4 @P 3 are different under identical conditions of excitation. An additional, and perhaps less obvious factor that could affect the bending, is the crystal quality. To that end, a high-quality crystal of 4 @P 3 and a crystal of 3 @P 3 of much poorer quality were selected and compared, as shown in Supplementary Fig. 17 . Both crystals bent under infrared light, which indicates that the crystal quality has a comparatively smaller effect, although the quantification of this effect is not straightforward. Fig. 2 Preparation of hybrid organic crystal arrays. a The chemical structures of crystals 2 ‒ 4 and photos (recorded under weak UV light and against a black background for clarity, green is crystal 2 , yellow is crystal 3 , and red is crystal 4 ) showing their mechanically induced bending. b Process for preparation of hybrid organic crystal arrays (kept at relative humidity RH = 62%, 64%) and their collective bending induced by illumination with infrared light (250 W). Photographs of the arrays are provided next to the illustrations. The blue and red hues are due to reflection of light from the background. All scale bars are 2 mm. Preparation of hybrid organic crystal arrays Some of the future applications of flexible crystals are based on sensing in two dimensions, which requires an ordered array of individual standing bendable crystals. In order to explore this direction of possible application, we have prepared two-dimensional arrays of the hybrid organic crystals that were bent collectively with infrared light. As shown in Fig. 2b , three different arrays were prepared by gluing crystals to capillaries embedded in a styrofoam base, including 3 × 3 arrays of 2 ‒ 4 @P 3 , 4 × 4 arrays of 4 @P 3 , and 4 × 4 arrays of 2, 4 @P 3 . These crystalline arrays, kept at relative humidity RH = 62%, 64%, were collectively bent in the same direction when exposed to infrared light (250 W). The bending is a result of the absorption of the infrared light by the MXene, which leads to the thermal contraction of the PVA/PSS layer on the crystal’s surface. When the irradiation with infrared light is terminated, the hybrid crystals return to their initial, straight shape. To realize a multidimensional bending of these photosensitive arrays, the bending direction of the hybrid organic crystals was controlled. As shown in Fig. 3a , the position of the infrared lamp (184 mW) was fixed, and the capillary glass tube below the styrofoam base was manually rotated (Fig. 2b ). As the crystal glued at the tip of the capillary glass tube rotated, it could be bent in different directions. As shown in Fig. 3b‒d (Supplementary Movie 3 ), the 3 × 3 array of 2 ‒ 4 @P 3 shows three types of collective bending under infrared light (250 W). Similarly, Fig. 3e‒j shows the optomechanical motion of 4 × 4 arrays of 4 @P 3 and 2, 4 @P 3 . Hybrid crystals with organic crystals having a different size, as well as with organic crystals of the same size but having different thicknesses of the MXene layer were also prepared to examine the effect of crystal size and MXene thickness on the performance of the 2D array. As shown in Supplementary Fig. 18 , four samples of 4 @P 3 with different sizes were bent to different angles under identical experimental conditions. Similarly, two hybrid crystals having one or five layers of the MXene, 2 @PVA/PSS@(PDDA/MXene) 1 @PDDA/PSS and 2 @PVA/PSS@(PDDA/MXene) 5 @PDDA/PSS, shown in Supplementary Fig. 19 , bent to a different degree. These results confirmed that both the size of the organic crystal and the thickness of the MXene layer determine the performance of the 2D arrays. The capability of multidirectional motion of the crystal arrays expands the range of their possible applications in two-dimensional optical transmission and detection. Fig. 3 Construction of different types of hybrid organic crystal arrays. a Photograph of hybrid organic crystal array achieving 360° bending under infrared illumination by rotating a capillary glass tube. b ‒ j Schematic diagram and photographs of different types of hybrid organic crystal arrays: b ‒ d 3 × 3 arrays of 2 ‒ 4 @P 3 , e ‒ f 4 × 4 arrays of 4 @P 3 , and g ‒ j 4 × 4 arrays of 2, 4 @P 3 . All scale bars are 2 mm. Durability and sensitivity of hybrid organic crystals The mechanical robustness, cyclability in operation over prolonged usage without fatigue, and response sensitivity are some of the main prerequisites for future applications of dynamic crystals in devices such as flexible electronic devices. The bending using light provides spatial control over the excitation, which could, in turn, be used to control the shape. A demonstration of this concept is illustrated in Fig. 4a , where four different positions on the same hybrid crystal of 3, 4 @P 3 were selected for irradiation (184 mW). As shown in Supplementary Fig. 20 , the hybrid organic crystals can be bent in both directions by changing the deposition method that had been used to prepare the PVA/PSS film. The shape of the crystal is determined by the location of bending. Moreover, the degree of bending can also be varied by the degree of heating due to light absorption. As shown in Fig. 4b , the bending of the crystals 3 @P 3 and 4 @P 3 increases with the increasing power of the infrared radiation due to an increase in temperature, mainly of PDDA/MXene (Fig. 1f ). 4 @P 3 has an inflection point in the bending angle with power at 576 mW, probably due to the fact that the horizontally incident infrared light has access to surfaces of the crystal after it bends over 90°. Fig. 4 Assessment of the performance of the hybrid organic crystals. a Bending of 3, 4 @P 3 induced by irradiation at different positions (1‒4) of irradiation with infrared light (184 mW). The red dashed line represents the position excited with infrared light. b Photographs showing the degree of bending of 3, 4 @P 3 (defined in the inset) at different powers of infrared light. The red dashed line represents the position excited with infrared light. c Testing of reproducibility of deformation of 3, 4 @P 3 in cycling mode. d Dependence of the bending angle of 3, 4 @P 3 as a function of time. The dashed blue line separates the light regime (left) and the dark regime, i.e., absence of light (right). All scale bars are 2 mm. Source data are provided as a Source Data file. Additional tests of the robustness of the hybrid actuators were carried out to characterize the reproducibility of the degree of deformation and sensitivity of 3, 4 @P 3 under irradiation with infrared light. A light-sensitive element of 3, 4 @P 3 was exposed to infrared radiation (184 mW) and sunlight (Supplementary Movies 4 and 5 ). After 100 cycles, the maximum curvature of the bent state remained nearly constant (Fig. 4c ). We tentatively attribute the slight decrease to gradual aging of the polymer coating caused by heating. The response rates of the hybrid crystal were estimated during the first cycle and the 100th cycle (Fig. 4d , Supplementary Movies 6 ‒ 9 ). The flexing and recovery time of 3, 4 @P 3 remained nearly constant, at about 1 s and 1 s for 3 @P 3 , and 2 s and 2.5 s for 4 @P 3 , respectively. Moreover, fatigue and cycling tests were performed on 3 @P 3 , and the results confirmed that the performance of the hybrid crystals was retained even after 1000 cycles (Supplementary Fig. 21 , Supplementary Movie 10 ). However, both the rate of bending and the rate of recovery decreased slightly due to a decrease in maximum curvature and hence a decrease in the reverse bending moment. Overall, we conclude that the performance of these simple light-sensitive devices is sufficiently stable, and this result favors this material as a candidate for practical applications such as optical signal transmission. In addition, as shown in Supplementary Movie 11 , we established that the actuation of the hybrid organic crystal bending can be precisely controlled by infrared light at distances of over 3 m (Supplementary Fig. 22 ). This result further highlights the prospects for long-range sensing, triggering, actuation, or other remote real-world applications of these materials. Hybrid organic crystals as optical waveguides As mentioned above, organic crystals have the advantages of relatively low optical losses and long-range ordered structures. One of the properties that are now actively being explored with slender organic crystals is their optical waveguiding capability, occasionally combined with mechanical flexibility observed with some of these crystals. Recently, similar flexible organic crystals have been shown to hold potential for optical transduction at low temperatures 43 . Flexible organic crystalline waveguides for both visible and near-infrared (telecom) ranges of the spectrum have been reported 44 – 46 . The light output of hybrid organic crystals can be precisely controlled by using a magnetic field or changes in aerial humidity 23 , 24 . Light-driven bending has also been extensively studied, especially in view of advantages such as remote, long-distance control 47 – 49 . Organic crystals have been already established as active optical waveguides 9 , 21 , 50 . The principles of this property are rooted in their optical transparency and difference in refractive index with air; when one end of the crystal is excited by ultraviolet light, the emitted light is subsequently reflected and refracted continuously inside the crystal and is eventually transmitted to its other end 51 . Figure 5a shows an infrared light-driven array of hybrid organic crystals for optical signal transmission. In Fig. 5b , one end of the hybrid organic crystal was affixed and excited by a 365 nm laser, and the optical output point at its other end was controlled by infrared radiation. This is accomplished by controlling the degree of bending by changing the point of excitation. As shown in Fig. 5 c, d , 3 @P 3 was excited by a 365 nm laser at its fixed end. The light output at the other end changes with the position of the point of exposure to an 808 nm light (184 mW). 3 @P 3 was selected to investigate the variation in the bending angle of the hybrid crystal over time under long-term irradiation with infrared light. As shown in Supplementary Movie 12 , the bending angle of the hybrid crystal remains almost unchanged up to at least 20 min of exposure to infrared light. Supplementary Movie 14 demonstrates that 3 @P 3 transmits optical signals for 20 min without visible changes in the direction of the transmission, confirming the capability of the hybrid crystal to transmit optical signals in a specific direction over a prolonged period. Fig. 5 Optical waveguiding properties of the hybrid organic crystals. a Diagram of infrared light-driven hybrid organic crystal array for optical signal transmission. The top left image is a zoomed-in representation of the crystal tip. b Photographs of a hybrid organic crystal array for optical signal transmission. c A schematic showing the dependence of the optical output point of a hybrid organic crystal on the excitation position. d Photographs showing the change in output of the optical signal of 3 @P 3 with the position of excitation with infrared light. The insets show 10-fold magnified images of the crystal tip. The broken line circles indicate the position of the optical signal output. e Photographs of a crystal of 5 before (left) and after mechanically induced bending (middle) and bending of crystals under UV irradiation (right). f Isomerization of 5 exposed to UV light. g A schematic showing the concept of an optical waveguide of crystal of 5 controlled by infrared and ultraviolet light. The red dashed arrow indicates the direction of the infrared light, and the blue arrows indicate the direction of the ultraviolet light. Scale bar: 2 mm. Having the infrared-control over the shape of the crystals at hand, hybrid organic crystals that act as optical waveguides and can be controlled by two stimuli, infrared and ultraviolet light, were also prepared. A centimeter-long slender elastic crystal of the compound ( E )–2-(4-fluorophenyl)–3-(naphthalen-1-yl)acrylonitrile (hereafter referred to as 5 ; Fig. 5e, f ) that has been reported earlier 24 was selected for the purpose. The crystal is elastic and has Young’s modulus of 2.49 GPa. It can be bent repeatedly without breaking (Fig. 5e ) (Supplementary Fig. 23 ). Crystal of 5 can also be bent by irradiation with UV light due to configurational isomerization, and this process was confirmed by NMR spectroscopy (Fig. 5f ; Supplementary Fig. 24 ). In the experiment shown in Fig. 5g and Supplementary Movie 13 , 5 @P 3 is excited by a 355 nm laser at its fixed end, and the optical output point at the other end is controlled by infrared or ultraviolet light. As shown in Supplementary Movie 14 , 3 @P 3 transmits the optical signal for 20 min without visible changes in the direction of optical transmission, confirming the capability of the hybrid crystal to transmit optical signals in a specific direction over prolonged periods of time. To confirm that the hybrid organic crystal was conductive to light, the optical waveguiding capability of 3, 5 @P 3 in the straight and bent states was tested by using a standard method 52 . The distance-dependent emission spectra were obtained by irradiating different positions of 3, 5 @P 3 by using a 355 nm laser (10 Hz, 10 ns), collecting the emission spectra at the other end of the crystal and fitting the data (Supplementary Fig. 25a‒d ). Expectedly, the emission intensity of the tip gradually decreased as the distance from the irradiation position to the tip increased due to the increased loss of emitted light with distance (Supplementary Fig. 25e‒h ). The optical loss coefficients were found to be 0.15461 dB mm − 1 and 0.15562 dB mm − 1 for the straight and bent state of 3 @P 3 , and 0.12091 dB mm − 1 and 0.12689 dB mm − 1 for 5 @P 3 (Supplementary Fig. 25i‒l ). Moreover, the optical loss of the hybrid crystal was measured after 100 bending cycles and long exposure time. As shown in Supplementary Fig. 26 , the optical loss of 3 @P 3 at 0, 50, and 100-fold bending was 0.16373, 0.16988, and 0.17649 dB mm − 1 , respectively. The optical loss after irradiation of 0, 30, and 60 min was 0.15600, 0.16238, and 0.17458 dB mm − 1 , respectively (Supplementary Fig. 27 ). These results clearly confirmed that the light transmission through the hybrid organic crystal can be controlled by both infrared and ultraviolet light. In summary, here we report a family of hybrid organic crystalline materials whose deformation is driven and can be precisely controlled by infrared light. The hybrid materials display favorable properties which combine mechanical flexibility brought about by the mechanical compliance of the organic crystals, on the one hand, and thermal sensitivity, governed by a layer of MXene on their surface, on the other. By simply adjusting the infrared light, these hybrid organic crystals can be bent to an arbitrary extent and at a desired position along their length. This deformation can be induced not only with individual crystals but also with a collection of hybrid crystals arranged in a regular two-dimensional array. The hybrid materials described here have the advantages of high sensitivity, high and controllable degree of deformation, and durability over prolonged actuation, while the bending point of each crystal can also be controlled by changing the position of excitation with infrared light. As a proof-of-concept of the immense opportunities that this approach opens for dynamic materials, we demonstrate that infrared light-driven flexible organic crystal optical waveguides can be constructed from these materials, including ordered arrays of such optical waveguides. Since photothermal bending has some advantages over other modes of excitation, such as the possibility to control the point of bending by choosing the position of excitation and to deform the crystal remotely by irradiation over long distances, adding the infrared light to the palette of available excitation stimuli such as UV or visible light, humidity, and magnetic field expands significantly the prospects for construction of flexible optical and electronic devices based on organic crystals."
} | 7,495 |
39691433 | PMC11650751 | pmc | 8,795 | {
"abstract": "ABSTRACT Ecosystem size and spatial resource flows are key factors driving species diversity and ecosystem function. However, the question of whether and how these drivers interact has been largely overlooked. Here, we investigated how ecosystem size asymmetry affects species diversity and function of two‐patch meta‐ecosystems connected through flows of nonliving resources. We conducted a microcosm experiment, mimicking resource flows between ecosystems of different sizes yet otherwise identical properties or between ecosystems of the same size. Meta‐ecosystems with asymmetric ecosystem sizes displayed higher α‐diversity but lower β‐diversity and ecosystem function (total biomass) than their unconnected counterparts. At the same time, such an effect was not found for meta‐ecosystems of identical patch sizes. Our work demonstrates how the size of ecosystems, interconnected via resource flows, can modulate cross‐ecosystem dynamics, having implications for species diversity and function across scales.",
"introduction": "1 Introduction Ecosystem size is a key factor driving species diversity. Ecologists have long known that larger ecosystems harbour more species diversity than smaller ecosystems (species–area relationship; MacArthur and Wilson 1963 ). The concept dates back to the late 1700s during the second Pacific voyage of James Cook. There, naturalists Johann Reinhold Forster and Georg Forster noted that ‘Islands only produce a greater or less number of species, as their circumference is more or less extensive’ (Forster 1778 ), which has been empirically and experimentally corroborated many times since (e.g., Fukami 2004 ; Losos and Ricklefs 2009 ; Wilson 1961 ). The various reasons why larger ecosystems harbour more species diversity remains an ongoing area of research (Losos and Ricklefs 2009 ), and the individual roles of different processes (e.g., speciation and dispersion) contributing to this pattern are still debated (e.g., Valente et al. 2020 ). The main explanation for this phenomenon has been that species go extinct at lower rates in larger ecosystems (MacArthur and Wilson 1963 , 1967 ) as they have more habitat types (Kallimanis et al. 2008 ; Williams 1943 ), more niche diversity (e.g., Ren et al. 2022 ) and experience less ecological drift (e.g., Gilbert and Levine 2017 ). The reason why larger ecosystems commonly house more species diversity has been extensively investigated through theoretical, comparative and experimental studies (e.g., Hanski and Ovaskainen 2000 ; Luo et al. 2022 ; Wang and Altermatt 2019 ). Furthermore, ecosystem size can also influence ecosystem function (LeCraw, Romero, and Srivastava 2017 ; McIntosh et al. 2024 ; Yang et al. 2021 ). For example, larger ecosystems can be more productive because communities in larger ecosystems can use resources efficiently thanks to complementary traits (complementarity effects; Delong and Gibert 2019 ). Additionally, larger ecosystems can support longer food chains (Post 2002 ), increasing or decreasing ecosystem function according to the trophic level considered (Loreau 2010 ). Furthermore, irrespective of their size, ecosystems are rarely isolated in space. Spatial flows and nonliving subsidies among ecosystems (e.g., leaf litter, carcasses, and inorganic nutrients; herein, ‘resource flows’; see Gounand et al. ( 2018 ) for a review) are—next to ecosystem size—a key abiotic factor affecting species diversity and ecosystem function. For example, salmon carcasses transported from rivers to land by wolves and bears bring abundant nutrients, which can decrease riparian plant species diversity as they promote the dominance of some species (Hocking and Reynolds 2011 ). Subsidies from marine algal wrack can either increase plant species diversity on sand dunes (Del Vecchio et al. 2017 ) or decrease plant species diversity in rainforests on tiny islands (Obrist et al. 2022 ). As another example, aquatic insects can make up a great part of the diet of riparian birds which feed on them, potentially allowing them to maintain their function (production) (Nakano and Murakami 2001 ). Supporting such empirical evidence, meta‐ecosystem theory predicts that resource flows can affect species diversity by modifying species interactions and persistence (Gravel et al. 2010 ; Marleau, Guichard, and Loreau 2014 ; Peller, Marleau, and Guichard 2022 ). For example, resource flows can delay competitive exclusion by increasing locally available resources (Gounand et al. 2017 ) or instead prevent the local establishment of dispersing species by increasing the abundance of the resident competitors (Gravel et al. 2010 ). Furthermore, resource flows should increase meta‐ecosystem production if they transport resources from ecosystems good at producing biomass through photosynthesis to ecosystems good at transforming nonliving resources into consumers (Harvey et al. 2023 ). However, despite widespread recognition that ecosystem size and resource flows can affect species diversity and ecosystem function individually, their interactive effect has largely been overlooked. These two drivers—ecosystem size and resource flows—likely interact since ecosystem size influences the amount and the effect of resource flows. For example, the size of a body of water regulates the amount of resources it exports: The larger a lake or a river, the more insects emerge from it per metre of reach (Gratton and Vander Zanden 2009 ). Furthermore, the size of the receiving watershed would determine the effects of aquatic resource import: For instance, the larger a watershed, the more diluted its fertilisation from salmon carcasses (Hocking and Reimchen 2009 ). Also, larger islands that receive algal wrack and carrions from the ocean experience a more diluted positive effect on their secondary production (Polis and Hurd 1995 ). As resource flows can influence species diversity and ecosystem function, and ecosystem size can influence resource flows, the hypothesis that ecosystem size can influence species diversity and ecosystem function through resource flows emerges naturally as a general concept. Here, we tested if and how the size of interconnected ecosystems mediates the influence of resource flows on species diversity and ecosystem function using a protist microcosm experiment (Altermatt et al. 2015 ; Benton et al. 2007 ; Cadotte and Fukami 2005 ). We constructed two‐patch meta‐ecosystems connected by resource flows between ecosystems (for clarity we use ‘patch’ as a synonym for ‘ecosystem’). We manipulated (i) the relative size of the two patches within the meta‐ecosystem (symmetric vs. asymmetric sizes) while keeping the total size of the meta‐ecosystem constant, and (ii) the connection between the two ecosystems (connected vs. unconnected). Our results showed as a proof of concept that ecosystem size asymmetry significantly influences species diversity and ecosystem function through resource flows. Specifically, we observed resource flows increasing α‐diversity and a decreasing β‐diversity and ecosystem function (total biomass) in asymmetric meta‐ecosystems when comparing their connected to their unconnected treatment. Contrastingly, resource flows did not affect α‐diversity, β‐diversity or ecosystem function in symmetric meta‐ecosystems, as shown by comparing symmetric connected and unconnected meta‐ecosystems.",
"discussion": "4 Discussion We experimentally demonstrate in a proof‐of‐concept study that ecosystem size asymmetry can mediate the effects of bidirectional resource flows on species diversity and ecosystem function. Meta‐ecosystems with asymmetric patch sizes (S L L S ) had more similar communities (lower β‐diversity) and lower function (lower total biomass) but maintained higher species diversity across the two local patches (higher mean α‐diversity) than asymmetric yet unconnected meta‐ecosystems (SL). These effects were not observed in meta‐ecosystems with symmetric ecosystem sizes, indicating a mediating role of ecosystem patch size. These results could be explained by the connection of a small ecosystem to a larger ecosystem resulting in an increase in species diversity in the small ecosystem (S L had greater species diversity than S S and S) while leaving the species diversity of the large ecosystem unchanged (species diversity was similar between L s , L L and L) and decreasing the biomass of the large ecosystem (L S had lower biomass than L L and L). As small and large ecosystems at the beginning of the experiment were identical aside from their size (resource concentration, community composition, etc.), the effects of the connection can be attributed to ecosystem size. Ultimately, our findings suggest that considering the size of interconnected ecosystems can help us understand how bidirectional resource flows shape species diversity and ecosystem function. Notably, we found that resources flowing between ecosystems of different sizes impacted both α‐ and β‐diversity by increasing the species diversity of the smaller patch within the meta‐ecosystem. Furthermore, they decreased total meta‐ecosystem biomass by decreasing the biomass of large patches, which was not compensated by the congruent increase in biomass in small patches. We suggest three mutually nonexclusive mechanisms by which small patches may have gained species diversity and biomass while large patches lost biomass compared to the unconnected control patches. All mechanisms should be the result of the effects of a change in resources coming from a difference between the resources gained by the inflow of resources and resources lost by the outflow of resources, as we precluded dispersal. First, resource quantity: Small ecosystems may have had a net import of resources. Although the volume exchanged between ecosystems was identical, larger ecosystems had a greater dominance of photosynthetic species than small ecosystems (Figure 5 ), which might have increased carbon availability more in large versus small ecosystems. Consequently, small ecosystems may have imported a greater quantity of sequestered carbon from large ecosystems relative to what they exported, creating an emergent source–sink dynamics of resources (sensu Gravel et al 2010 and Loreau et al. 2013 ). This net import of resources could have allowed small ecosystems to sustain more species diversity as more resources allow more individuals to persist, promoting a greater abundance of rare species and preventing their extinction (species energy theory, see Wright 1983 ). The net import of C‐rich necromass to small patches and net export from large patches could have increased basal resources for bacteria—primary resources of protists—in small patches at the detriment of large patches, which cascaded up on biomass production: The decrease in meta‐ecosystem biomass caused by a larger decrease in biomass in large patches than the increase in biomass in small patches could be explained by a net movement of resources to the small patch, which could have had lower recycling rates. Indeed, higher recycling rates in larger patches are reasonable to expect, as they can be found in nature (Donghao et al. 2021 ; LeCraw, Romero, and Srivastava 2017 ; Yang et al. 2021 ). In natural ecosystems, we would expect differences in ecosystem sizes to lead to differences in the quantity of resources exchanged as well, potentially through different mechanisms. For instance, the trophic island biogeography theory (Gravel et al. 2011 ; Holt 2009 ) predicts variation in the ratio between autotrophs and consumers between ecosystems of different sizes. Gravel et al. ( 2011 ) supported this prediction by parameterising a trophic metacommunity model using 50 pelagic food webs (Havens 1992 ) and showing that larger ecosystems contained more consumers relative to autotrophs. The explanation for this result is that consumers are more likely to find one of their prey in larger ecosystems and, therefore, establish. Second, resource quality: Small ecosystems may have had a net import of detritus (protist detritus) of better/different quality. If the detritus of protists was of higher quality as a resource for the local community compared to other resource forms (e.g., bacterial detritus and inorganic nutrients), it would have sustained a higher growth of individuals and, therefore, higher species diversity in the small ecosystem. Consequently, the movement of resources of higher quality to the small ecosystem and of lower quality to the large ecosystem would have increased the function of small ecosystems and decreased the function of the large ecosystem, as a meta‐ecosystem model showed that good quality subsidies should increase the function of the receiving ecosystem and bad‐quality subsidies should decrease it (Osakpolor et al. 2023 ). We would also expect this mechanism, where size differences between connected ecosystems create differences in the quality of resources exchanged and cascade to influence species diversity and function, to occur in natural ecosystems, potentially through different mechanisms. For example, ecosystems of different sizes can have different biomass distributions across trophic levels (Petermann et al. 2015 ), with often higher maximal trophic levels in larger ecosystems (Guo et al. 2023 ; Post, Pace, and Hairston 2000 ; Ward and McCann 2017 ). Moreover, different trophic levels might produce detritus of different qualities as consumers often have higher nitrogen content than producers (Elser et al. 2000 ). As a consequence, the relative quantities of biomass in trophic levels determine the overall quality of the resources exchanged with other ecosystems, which depends on ecosystem size. Third, resource heterogeneity: Small ecosystems might have imported resources that were more heterogeneous than their own. As there was greater protist species diversity in large than in small ecosystems, the corresponding exported detritus might have been more diverse with respect to carbon compounds and biomolecules, potentially creating more niches for protists to coexist in small ecosystems (resource diversity hypothesis, Lawton 1983 ). The positive correlation between detritus heterogeneity and consumer feeding on it has been observed in nature (Moore and William Hunt 1988 ; Yodzis 1988 ). We expect that also in nature differences in ecosystem size would cause differences in resource heterogeneity and, therefore, cause resource flows to influence species diversity and ecosystem function. Larger ecosystems generally have higher species diversity within trophic levels (horizontal diversity, MacArthur and Wilson 1963 , 1967 ) and higher number of trophic levels (vertical diversity or maximum food chain length, Guo et al. 2023 ; McHugh et al. 2015 ; Post, Pace, and Hairston 2000 ; Ward and McCann 2017 ). Such higher species diversity should translate into a change in biomass composition (e.g., species diversity can be related to stoichiometry; Striebel, Behl, and Stibor 2009 ) and higher resource heterogeneity, which would constitute more heterogeneous resources that would determine the effects of resource flow on species diversity and function. Our study highlights that the size of the donor ecosystem, where resource flows originate, can shape the effect of resource flows on a recipient ecosystem's species diversity. In particular, in our experiment, species diversity increased in ecosystems with small patch sizes when connected to ecosystems of large patch sizes more than when connected to ecosystems of small patch sizes. The subsidised island biogeography theory (Anderson and Wait 2001 ) states that resources flowing into an ecosystem can influence its species diversity, making its species diversity deviate from what we would expect from species–area relationships, especially in small ecosystems. There is some comparative evidence by field studies, for instance, with resource flows increasing the species diversity of bird species more on smaller than on larger islands (Obrist et al. 2020 ). However, experimental evidence of this phenomenon is largely lacking. Here, we give a formal experimental corroboration that resources exchanged between differently sized ecosystems affect species diversity and ecosystem functions (e.g., biomass) and are modulated by the differential patch size. In particular, we highlight and discuss how the size of the exporter ecosystem may mediate the quantity, quality and heterogeneity of resource flows through various mechanisms that would modulate the effect on the diversity and functioning of the recipient ecosystem. Decades of research on spatial subsidies have documented that donor ecosystems commonly vary in size. For example, islands which export nitrogen to coral reefs (Lorrain et al. 2017 ), kelp forests which exchange nonliving resources with their adjacent intertidal zone (Tallis 2009 ) or forests that export leaf litter to streams (Larsen, Muehlbauer, and Marti 2016 ). Moreover, evidence from natural systems supports our finding that donor ecosystems' size can influence recipient ecosystems' species diversity and function. Such evidence is found in lakes and rivers embedded in terrestrial watersheds of different sizes. Notably, studies found that larger watersheds can (i) increase lake primary production, as they export more phosphorus (Knoll, Vanni, and Renwick 2003 ), (ii) sustain fewer lake consumers that rely on sediments, as they export lower quantities of sediments (lower water flow, gentler slopes and increased sedimentation in terrestrial ecosystems) (Babler, Pilati, and Vanni 2011 ), and (iii) sustain longer river food chains, as they have more water flow, hence less hydrological variation and therefore a more stable environment (Sabo et al. 2010 ). This, in conjunction with our findings, suggests that subsidised island biogeography (Anderson and Wait 2001 ) would gain in integrating how the size of the connected ecosystems mediates the effects of resources on the shape of species–area relationships and possibly changes this relationship. According to our results, we expect, for example, that the species diversity of macroinvertebrates in a lake might be higher than expected by their area only (according to subsidised island biogeography) when the lake is connected to a larger rather than a small forest. In conclusion, our experiment provides experimental proof of concept that asymmetry in ecosystem size can indirectly affect species diversity and function in meta‐ecosystems through its effects on a ubiquitous connection among ecosystems–spatial flows of resources. Consequently, this implies a need to consider how ecosystem size changes resource flow between ecosystems when aiming to generally understand the drivers of species diversity and ecosystem function in spatially structured systems. Future research should focus on how ecosystem size impacts meta‐ecosystems through resource flows, testing our proposed mechanisms on resource quality and heterogeneity in relation to species diversity and incorporating other properties of resource flows, such as asynchronous flows (Nakano and Murakami 2001 ), as well as the magnitude of resource flow in relation to ecosystem size (e.g., Gratton and Vander Zanden 2009 )."
} | 4,827 |
35814646 | PMC9260433 | pmc | 8,796 | {
"abstract": "Biofilm formation by photosynthetic organisms is a complex behavior that serves multiple functions in the environment. Biofilm formation in the unicellular cyanobacterium Synechococcus elongatus PCC 7942 is regulated in part by a set of small secreted proteins that promotes biofilm formation and a self-suppression mechanism that prevents their expression. Little is known about the regulatory and structural components of the biofilms in PCC 7942, or response to the suppressor signal(s). We performed transcriptomics (RNA-Seq) and phenomics (RB-TnSeq) screens that identified four genes involved in biofilm formation and regulation, more than 25 additional candidates that may impact biofilm formation, and revealed the transcriptomic adaptation to the biofilm state. In so doing, we compared the effectiveness of these two approaches for gene discovery.",
"introduction": "Introduction Biofilm formation, the adhesion of bacteria to a surface typically involving the production of an extracellular polymeric substance (EPS), is a lifestyle that allows microorganisms to survive in the face of stresses and threats from their environment, including nutrient depletion ( Donlan and Costerton, 2002 ; Harrison et al., 2019 ), predation ( DePas et al., 2014 ; Chan et al., 2021 ), and dessication ( Lebre et al., 2017 ). Biofilms can be detrimental to human health and industrial activities, by enabling persistence and antibiotic resistance in patients and on medical surfaces ( Flemming et al., 2016 ), or by causing biofouling as on ship surfaces ( de Carvalho, 2018 ) or in corrosion of industrial piping ( Lenhart et al., 2014 ). Studies aimed at disrupting deleterious biofilms have led to a wealth of knowledge on the molecular mechanisms regulating and forming biofilms in heterotrophic or pathogenic bacteria. This knowledge includes the requirement for pili and adhesins in finding and attaching to surfaces, intracellular cyclic di-GMP secondary messenger systems for controlling activation of biofilm formation, and production and secretion of exopolymeric matrix materials that can include exopolysaccharides, amyloidgenic proteins, and extracellular DNA ( Flemming et al., 2016 ). Photosynthetic cyanobacteria, which are among the planet’s dominant producers of oxygen and fixed carbon, can also form biofilms, often in microbial mats where a combination of gas, nutrient, signaling, and light gradients determine the composition, density, and organization of the diverse community of organisms in the assemblage ( Prieto-Barajas et al., 2018 ). Cyanobacteria are a promising agricultural crop for sustainable conversion of solar energy and atmospheric carbon into replacements for petroleum products, including biofuels, industrial chemicals, nutraceuticals, and plastics ( Nozzi et al., 2013 ; Farrokh et al., 2019 ; Sitther et al., 2020 ). In this context, biofilm formation can be both disadvantageous, due to biofouling of growth vessels and piping, or beneficial for protection against predation, efficiency in harvesting the crop, or minimizing resources for growth and production in a biofilm as opposed to in a planktonic culture ( Bruno et al., 2012 ; Barros et al., 2020 ). In contrast to health applications that seek to prevent biofilm formation, the goal for agricultural scale industrial production of cyanobacterial biomass is to be able to reliably activate or deactivate biofilm formation to minimize costs and damage to equipment while maximizing productivity. Only in recent years have studies on the molecular mechanisms of biofilm formation extended to cyanobacteria, where light limitation through self-shading is an inherent consequence of biofilm structure ( Barros et al., 2020 ). For many cyanobacteria, biofilm formation and regulation appears to have a number of features in common with heterotrophic bacteria: type IV pili and S-layer proteins are required for film formation in Synechocystis sp. PCC 6803 ( Allen et al., 2019 ; Conradi et al., 2019 ), cyclic-di-GMP promotes aggregation in Thermosynechococcus vulcanus ( Enomoto et al., 2015 ) and biofilm formation in PCC 6803 ( Agostoni et al., 2016 ), a two-component regulatory system comprising a response regulator and a split histidine kinase controls cell aggregation in PCC 6803 ( Kera et al., 2020 ), and extracellular polysaccharides, such as cellulose in T. vulcanus , contribute to the extracellular matrix and cellular aggregation ( Kawano et al., 2011 ). In contrast, our studies have demonstrated that biofilm formation in Synechococcus elongatus PCC 7942, which is by default constitutively suppressed in the laboratory ( Schatz et al., 2013 ), does not require pili ( Nagar et al., 2017 ). In fact, most of the mutations discovered to date that relieve suppression and enable biofilm formation ( Supplementary Table S1 ) block formation of the Type IV pilus or its assembly apparatus, including the pilin-encoding pilA1 and pilA2 genes; the pilus assembly and Type II secretion proteins encoded by pilB (formerly referred to as t2sE ), pilC , and pilN ; and two proteins encoded by the genes hfq and ebsA that form a tripartite complex with PilB ( Schatz et al., 2013 ; Nagar et al., 2017 ; Yegorov et al., 2021 ). For the majority of these mutants, a biofilm containing 90–95% of the cell population forms 3–4 days following inoculation in bubbling tubes. Electron microscopy has confirmed that the biofilm cells lack pili. Furthermore, they have markedly diminished or altered exoproteomes, suggesting that a single Type IV/II pilus and secretion system fills these dual roles in S. elongatus ( Schatz et al., 2013 ; Nagar et al., 2017 ; Yegorov et al., 2021 ). Removal of this complex is pleiotropic—impeding other ecologically relevant behaviors such as natural competence and phototaxis-directed twitching motility ( Yang et al., 2018 ; Taton et al., 2020 ). Wild-type S. elongatus PCC 7942 (WT) does not produce biofilms under laboratory conditions due to the production and secretion of a self-suppressor ( Schatz et al., 2013 ). WT growth medium (conditioned medium, CM) contains accumulated suppressor that prevents biofilm formation by the pilB mutant, PilB ::Tn 5 (formerly named T2SEΩ), but CM from PilB ::Tn 5 does not, indicating that the mutant still responds to the suppressor it does not secrete ( Schatz et al., 2013 ). The derepression of a set of four small secreted proteins (EbfG1-4), each characterized by a secretion motif shared with microcins, enables biofilm formation in PilB ::Tn 5 ( Schatz et al., 2013 ; Parnasa et al., 2016 , 2019 ). The processing and secretion of these small proteins requires the cysteine peptidase gene pteB , which is cistronic with the ebfG genes, and the putative processing peptidase gene ebfE. Inactivation of ebfE significantly alters the exoproteome in a manner that is distinct from changes observed with inactivation of pilB ( Parnasa et al., 2019 ). The robust formation of biofilms by the recent wild isolate S. elongatus UTEX 3055 demonstrates that biofilm formation of PCC 7942 mutants uncovers a process that is reversible in the natural environment ( Yang et al., 2018 ). Decades of laboratory culturing techniques, such as decanting, that subtly select for planktonic cells ( Golden, 2019 ) has likely domesticated the PCC 7942 WT strain into a biofilm-suppressed state through genetic drift or epigenetic changes, which now requires biofilm-inducing mutations to activate the biofilm program ( Adomako et al., 2022 ). The known gene products involved in suppressing or enabling biofilm formation in S. elongatus PCC 7942 ( Supplementary Table S1 ) represent only pieces of the structural and regulatory system of biofilm formation, which must include the circuitry to synthesize, detect, and respond to the secreted repressor, as well as the repressed components that contribute to biofilm formation once self-suppression is alleviated. To obtain a more comprehensive understanding of genes involved in biofilm formation we leveraged the natural competence ( Shestakov and Khyen, 1970 ) of S. elongatus and genetic tools that include a library of transposon mutagenesis vectors for the rapid individual knockout of nearly any nonessential gene in the genome ( Holtman et al., 2005 ; Chen et al., 2012 ) and a randomly barcoded transposon sequencing (RB-TnSeq) library for quantitatively assessing the impact of gene mutations across the genome on the fitness of the strain under specific conditions, thus directly linking genotypes with phenotypes ( Rubin et al., 2015 ; Price et al., 2018 ). Analogous TnSeq libraries in other organisms have been used to study complex behaviors, as in biofilm formation in Bacillus cereus ( Yan et al., 2017 ; Okshevsky et al., 2018 ) or in community interactions of bacteria and fungi on the surface of maturing cheese ( Morin et al., 2018 ; Pierce et al., 2021 ). In addition to phenomics using RB-TnSeq, which identifies genes whose loss either enhances or reduces a cell’s fitness in a biofilm, we employed transcriptomics using RNA-Seq, which identifies genes whose transcription differs between planktonic and sessile cells. These two data sets suggest that RB-TnSeq is more powerful for identifying genes that contribute to a phenotype, whereas transcriptomics describes the underlying biology of a phenotypic state.",
"discussion": "Results and Discussion RNA-Seq Characterization of WT and Biofilm-Forming Strains Our first approach aimed to identify genes that are differentially regulated between planktonic and biofilm states. To compare the transcript profiles of cells destined to form or actively forming biofilms with those of planktonic cells, as well as the impact that conditioned (biofilm-suppressing) medium has upon these processes, total RNA was purified from WT or PilB ::Tn 5 cultures inoculated in either fresh or WT-derived conditioned medium (CM) and sampled 1 and 4 days following inoculation. Only the 4-day PilB ::Tn 5 cultures grown in fresh BG-11 produced biofilms, so transcript profiles were processed from both planktonic and biofilm fractions from these cultures ( Figure 1A ). Twenty-one samples representing triplicate experiments of seven experimental conditions were depleted of ribosomal RNA and sequenced using an Illumina platform, yielding 6–8 million reads per sample. Data were aligned to the S. elongatus PCC 7942 genome and analyzed for differentially expressed genes (DEG) using two different software pipelines: (1) Rockhopper 2, a prokaryotic-specific RNA-seq analysis platform ( Tjaden, 2015 ) and (2) a pipeline of published RNA-Seq software including Bowtie2 ( Langmead and Salzberg, 2012 ) for alignment, htseq-count ( Anders et al., 2015 ) for gene counts, and DESeq2 ( Love et al., 2014 ) for DEG analysis, with each analysis pipeline aligning approximately 99% of trimmed reads for nearly all samples, with less than 1% of these reads mapping to ribosomal RNAs (See Supplementary Files S1 and S2 for further details). Both the WT and PilB ::Tn 5 strains used in these experiments lack the smaller plasmid, pANS, and thus, its potential impact on biofilm formation was not assessed. Figure 1 RNA-Seq of WT and PilB ::Tn 5 and Day 1 DEG analysis. (A) Diagram of the time course for triplicate samples processed for Wild Type (WT) and PilB ::Tn 5 (B∇) when placed under biofilm forming conditions in either fresh BG-11 (BG) or WT-derived conditioned media (CM). Samples (circles) include planktonic samples from all test tubes and a sample of the biofilm formed by PilB ::Tn 5 on Day 4 in BG-11. (B,C) Volcano plots comparing the log 2 fold change of pairwise comparisons of samples against their −log 10 false discovery rate (FDR) as calculated by DESeq2. Data points are colored based on their DEG designation (“Hit”). Labeled shapes indicate known biofilm and pili ORFs. Dashed lines indicate the thresholds used for hit designation. Note that each plot is scaled individually. (D) Comparisons of sDEG designations (colored as in B,C ) of ORFs ( y -axis) from different relevant pairwise comparisons, as labeled above. The locations in the graph of known pili and biofilm ORFs are indicated by the colored ticks next to each graph (purple = biofilm and pili ORF, orange = pili ORF, green = biofilm ORF) and clusters of genes with common sDEG patterns across the x -axis of interest are designated with brackets and labeled with their identifiers and the number of sDEGs in the group. The number of ORFs considered in each graph is provided below the graph. On the left, the Day 1 BG-11 PilB ::Tn 5 vs. WT pairwise comparison is compared against the WT BG-11 Day 4 vs. Day 1 comparison to identify ORFs related to the stress or density state of the culture. Those ORFs that are similarly up- or downregulated in these two pairwise comparisons (groups A1 and A2) are connected to their location in the middle graph comparing all relevant Day 1 pairwise comparisons by the diagonal lines. These ORFs were then removed from the middle graph to generate the graph on the right. (E) Enrichment analyses of pili genes and known biofilm genes in the O-A and A groups of interest labeled in (D) . Significant enrichment (FDR ≤ 0.05) is designated by solid bars and stars. Groups that do not have any of the genes analyzed are designated with a negative infinity and arrow, as appropriate for the log 2 scale of the x -axis. ANOVA analysis using DESeq across all samples demonstrated that over 65% (1,776 out of 2,729) of the ORFs encoded by the genome, including all known biofilm-related ( Supplementary Table S1 ) and pili ( Supplementary Table S2 ) genes except for tsaP , are differentially expressed between at least two experimental conditions ( Supplementary File S2 Sheets 4 and 8). Additionally, Rockhopper predicted 118 noncoding RNAs (ncRNAs), of which 42 are antisense to annotated ORFs, 55 overlap with or are enclosed by previously reported potential ncRNAs ( Vijayan et al., 2011 ; Billis et al., 2014 ), and one overlaps an essential intergenic region ( Rubin et al., 2015 ; Supplementary Table S4 , Supplementary File S1 Sheet 9). All but three of these predicted RNAs were differentially expressed between at least two experimental conditions. Although RNA-binding by cyanobacterial Hfq homologs so far has not been demonstrated ( Schuergers et al., 2014 ), we hypothesize that these novel differentially expressed RNAs may represent binding targets. Both Rockhopper and DESeq analyses were performed for all 21 pairwise comparisons of the seven experimental groups ( Supplementary Files S1 Sheet 6 and S2 Sheet 5). The union of the lists of DEGs from the two analyses generated a final list of DEGs for all pairwise comparisons ( Supplementary File S3 Sheets 1 and 2); the intersection of the lists proved too stringent and excluded known biofilm forming genes that the union did not. Ten pairwise comparisons ( Figures 1 , 2 , Supplementary Figure S2 , and Supplementary File S3 Sheet 4) are relevant to the experimental variables tested: the pilB genetic alteration, biofilm state, time, and media conditions. While recognizing that more modest changes in expression may be biologically relevant, focusing on genes with fold changes >2, designated as strong DEGs (sDEGs), limited the total set considered among the 10 pairwise comparisons nearly four-fold from 1,867 DEGs to 508 sDEGs and resulted in a more limited range of sDEGS per pairwise comparison of 10 and 18 for those involving CM to 344 for pilB ::Tn 5 biofilm formation over time ( Supplementary Figures S1A,C and Supplementary File S3 Sheets 2 and 4), in contrast to 15, 67, and 1,332 DEGs, respectively. A summary of all sDEGs resulting from this analysis, as well as hits from the RB-TnSeq experiments discussed below, is provided in Supplementary Table S5 . Figure 2 Day 4 and time course RNA-Seq analysis. Volcano plots, presented as described for Figures 1B , C , of (A) Day 4 pairwise comparison of PilB ::Tn 5 biofilms vs. planktonic cells, (B) \n PilB ::Tn 5 biofilms on Day 4 vs. PilB ::Tn 5 on Day 1, and (C) \n PilB ::Tn 5 planktonic cells on Day 4 vs. PilB ::Tn 5 on Day 1. Comparison of sDEGs associated with (D) Day 4 and (E) over time are presented as described for Figure 1D . Tracking the 29 known biofilm and pili genes ( Supplementary Tables S1 and S2 , and Supplementary File S3 Sheet 6) enabled evaluation of the biological significance of these results ( Schatz et al., 2013 ; Parnasa et al., 2016 ; Nagar et al., 2017 ; Taton et al., 2020 ; Yegorov et al., 2021 ). All except 7 ( tsaP , pilD , hfq , ebfE , ebsA , pilA2 and the second pilT gene) were identified as sDEGs ( Figures 1B – D , Supplementary Figure S2 , and Supplementary File S3 Sheet 6). While it is reasonable that tsaP , pilD , and the second pilT genes do not have notable impacts on biofilm formation due to degeneracy or their encoded proteins’ roles as accessories to the pilus, the genes of the Hfq-EbsA-PilB tripartite complex ( Yegorov et al., 2021 ) or the EbfE-PteB ( Parnasa et al., 2016 ) maturation-secretion system play important roles that likely act at post-transcriptional levels ( Zhang et al., 2020 ; Yegorov et al., 2021 ). These results demonstrate the limitations of using changes in expression level as a proxy for contribution to a particular phenotype. Transcriptome Comparisons of Biofilm and Planktonic Cells Reflect the Biofilm-Proficient pilB ::Tn 5 Genotype We initially hypothesized that the Day 4 comparisons between biofilm and planktonic samples would identify novel genes required to regulate and create the biofilm. However, the sDEG patterns instead point to transcriptional consequences of the pilB ::Tn 5 mutation independent of its biofilm state or to the change in environment associated with existing in the biofilm ( Figure 2 , Supplementary Figures S2 , S3 ). None of the known biofilm and pili genes differed sufficiently in expression between pilB ::Tn 5 biofilms and planktonic cells to be considered as sDEGs ( Figure 2A , Supplementary File S3 Sheet 6), but the majority of them were either upregulated (e.g., ebfG genes, named group DF2) or downregulated (e.g., pilB operon genes, group DF5) when comparing either of the PilB ::Tn 5 samples against WT ( Figure 2D ). The sets of similarly up- or downregulated sDEGs in DF2 and DF5 are enriched for genes related to defense mechanism, motility, and proteins with predicted secretion signals ( Supplementary Figures S6B,E and Supplementary File S3 Sheets 10 and 18), consistent with known alterations in the exoproteome of PilB ::Tn 5 ( Nagar et al., 2017 ; Parnasa et al., 2019 ; Yegorov et al., 2021 ). Sets of 15 upregulated and 16 downregulated sDEGs correlate with the biofilm state (groups DF1 and DF6, respectively, in Figure 2D , Supplementary File S3 Sheet 10). The set of downregulated sDEGs (DF6) is enriched for Cyanobase categories of energy and inorganic ion metabolism ( Supplementary Figures S5A , S6D , Supplementary File S3 Sheet 18), suggesting that biofilm cells damp down core photosynthetic, ATP synthesis, nitrate, and iron nutrient metabolisms. Upregulated sDEGs (DF1) may reflect the resource limitations and self-shaded state of the biofilm. DF1 is dominated by genes encoding hypothetical proteins ( Supplementary File S3 Sheets 10 and 18) and include: sigD/rpoD3 (sigma factor), which is typically induced under high light in PCC 7942 ( Seki et al., 2007 ) and is also associated with survival under nitrogen deficiency ( Antal et al., 2016 ) and resistance to oxidative stress in PCC 6803 ( Li et al., 2004 ); and the phycobilisome degradation protein nblA . Other DF1 biofilm-specific upregulated sDEGs relate to stress responses including metal limitations, such as the high light-inducible hliA ( Ruffing, 2013 ) and genes encoding a putative ferric uptake regulator, a cyclic-AMP binding protein, and a cupredoxin. Day 1 and Day 4 samples were compared to identify genes that might be involved in the process of biofilm formation. Genes in the ebfG -cluster and pteB are the most upregulated sDEGs in both PilB ::Tn 5 biofilm and planktonic samples ( Figures 2B , C ) and are the only non-pili biofilm sDEGs over time. Although the distributed pattern of pili sDEGs over time does not highlight any specific cluster of candidates ( Figure 2E ), the set of genes that are upregulated similarly to the ebfG -cluster in both PilB ::Tn 5 samples over time (group OT1) is enriched with nine genes of unknown function ( Supplementary Figure S7B , Supplementary File S3 Sheet 12) that may play a role in biofilm formation. Biofilm PilB ::Tn 5 cells result in a much larger number of sDEGs (148 up; 196 down) than planktonic cells (61 up; 24 down) when compared to Day 1 samples ( Figures 2B , C ). Assessment of all over-time pairwise comparisons ( Figure 2E , Supplementary File S3 Sheet 12) highlights 86 up (group OT3) and 142 down (group OT4) biofilm-specific sDEGs, whose overlapping composition with the sDEGs of the Day 4 comparisons and an OT4-specific enrichment for energy metabolism, motility, and unknowns ( Supplementary Figure S5B , Supplementary File S3 Sheet 18) reinforce that the major changes in transcript abundance over time relate to adaptation to the self-shaded, nutrient-limited biofilm condition, rather than induction of components that control or form the biofilm phenotype. Day 1 Transcriptomics Highlight Regulatory and CM-Responsive Genes of Interest The four pairwise comparisons of all Day 1 samples identified nearly all known biofilm and pili genes among the highest magnitude up- and downregulated sDEGs in the comparison of PilB ::Tn 5 with WT ( Figure 1B ). Whether in fresh ( Figure 1B ) or conditioned media ( Supplementary Figure S2 ), the pilB ::Tn 5 mutation strongly decreased structural pilin gene expression, including pilB and downstream genes. The Day 1 data reliably capture the known genotype–phenotype associations related to biofilm regulation, as the ebfG genes were the most upregulated genes in PilB ::Tn 5 compared to WT in fresh media ebfG genes ( Figure 1B and Supplementary Figure S2 ), and were strongly downregulated ( Figure 1C ) in CM. The 54 upregulated (group O1) and 38 downregulated (group O4) sDEGs that differ between pilB ::Tn 5 and WT on Day 1 ( Figure 1D , Supplementary File S3 Sheet 8), independent of the media, are comprised of numerous Day 4 sDEGs that are associated with responses to the environment. These include upregulation of heat shock proteins Hsp20 and DnaK and an iron-stress chlorophyll-binding protein, and downregulation of heat shock protein GrpE. Some of these genes are sDEGs for WT when examined over time, suggesting that a pleiotropic effect of the pilB ::Tn 5 mutation is early establishment of a condition that WT assumes in a more advanced culture. These sDEGs suggest that PilB ::Tn 5 statically resembles a stationary culture, as these genes are not sDEGs when this mutant is examined over time. To filter out the sDEGs related to this apparent constitutive-stress state, we compared the sDEGs set for PilB ::Tn 5 vs. WT in BG-11 on Day 1, which reflects the genetic difference between the strains, against the sDEGs set for WT in BG-11 on Day 4 vs. Day 1, which presumably reflects response of the cells to aging of the culture and diminishing resources in low light-penetration conditions ( Figure 1D , left). This comparison identified 39 (group A1) up- and 5 (group A2) downregulated sDEGs in both pairwise comparisons, which are enriched in proteins of unknown function and Phobius-predicted secretion signals, respectively ( Supplementary Figures S4A,B , Supplementary File S3 Sheets 14 and 18). Group A1 includes stress-related genes, such as hsp20 and the iron-stress chlorophyll-binding protein previously noted, as well as transcription regulators such as sigF2 , which regulates dark-induced competence ( Taton et al., 2020 ) and is associated with pili formation and desiccation tolerance in other cyanobacteria ( Srivastava et al., 2021 ), and a Crp/Fnr transcriptional family regulator. These sDEGs support the premise that stress management genes associated with alterations in pili and protein maintenance that occur in a dense, low-light culture are regulated in WT, but are constitutively expressed at stress-response levels at all times in PilB ::Tn 5 . Based on this analysis, genes in groups A1 and A2 were removed from the Day 1 comparison and four primary groups of interest were identified as being up- or downregulated in PilB ::Tn 5 independent of the media (O-A1 and O-A4, respectively) or in fresh BG-11 media only (O-A2 and O-A3, respectively; Figure 1D , Supplementary File S3 Sheet 16). While O-A1 is not further enriched for any particular functionality beyond that of pili genes ( Figure 1E , Supplementary Figure S4C , Supplementary File S3 Sheet 18), this group of 24 sDEGs includes genes encoding two response regulator proteins, a histidine kinase, a glycosyltransferase, and 13 genes of unknown function that may contribute to the regulation of biofilm formation. O-A2 is enriched in known biofilm genes, predicted secreted proteins, and includes a response regulator protein, a lipoyltransferase, and 11 genes of unknown function that may participate in biofilm production along with the EbfG proteins ( Figure 1E , Supplementary Figure S4D , Supplementary File S3 Sheet 16). These upregulated clusters of sDEGs highlight potential novel genes involved in biofilm formation. Group O-A4 is enriched in known pili and biofilm genes and genes encoding proteins of unknown function, proteins involved in motility, and predicted secreted proteins ( Figure 1E , Supplementary Figure S4F , Supplementary File S3 Sheets 16 and 18); these properties are consistent with potential novel biofilm-repressing genes. Two genes of interest that encode proteins with cyclic di-GMP-binding domains and GAF domains, Synpcc7942_1148 and Synpcc7942_2096, may contribute to regulation of biofilm formation through cyclic di-GMP signaling, as occurs in Synechocystis PCC 6803 ( Agostoni et al., 2016 ). The pairwise comparisons pertaining to the impact of conditioned media, Day 1 in BG-11 vs. CM for both PilB ::Tn 5 and WT, resulted in small sets of 10 sDEGs each ( Supplementary Table S6 groups IG1 and IG2), with the most prominent impact of CM being the downregulation of the ebfG operon genes in PilB ::Tn 5 ( Figure 1C ), consistent with previous RT-PCR data ( Parnasa et al., 2016 ). Most of these genes encode proteins of unknown function ( Supplementary File S3 Sheet 16). Overall, the RNA-Seq experiment provides an overview of the pleiotropic phenotypes associated with the PilB ::Tn 5 mutation and a large set of genes that likely reflect the impact of the environment of the biofilm, which is self-shaded and has limited gas and nutrient exchange, rather than the active process of biofilm formation. Extensive comparative analysis does identify sDEG clusters of interest that include candidate genes for biofilm regulation, formation, and responses to conditioned media, though the majority of these potential novel biofilm genes encode proteins of unknown function and must be investigated through other means. RB-TnSeq Screen for Loci That Affect Biofilm Formation A complementary approach used an RB-TnSeq assay that measures the contribution of any given gene in the genome to influence the ability of a cell to proliferate in biofilms vs. planktonic growth. We screened a pooled RB-TnSeq library of approximately 150,000 individual barcoded insertions mapped to the S. elongatus PCC 7,942 genome that have previously shown negligible impacts from polar effects ( Rubin et al., 2015 ). PCR amplification and high-throughput sequencing of the barcodes quantifies changes in abundance of individual insertion mutants in the population under varying conditions, directly associating phenotypes with genotypes as has been shown for diurnal growth ( Welkie et al., 2018 ), c-di-AMP utilization ( Rubin et al., 2018 ), and natural competence ( Taton et al., 2020 ). We grew replicate biological samples of the RB-TnSeq library under a variety of biofilm formation conditions, including in bubbling tubes and stationary flasks in fresh BG-11 or CM, in two separate experiments ( Figure 3A ). Conditions were similar to those used for RNA-Seq experiments except that biofilm formation was allowed to proceed for two-weeks to ensure sufficient biomass to perform genome extraction and barcode counting. Because the majority of the mutant cells in the library should still secrete and respond to the biofilm-repressing signal that is characteristic of the WT and its CM ( Schatz et al., 2013 ), it was not expected a priori that the library could form biofilms. Therefore, we grew samples of a characterized biofilm forming mutant and a 1:1 mixture of WT and the biofilm-forming mutant in parallel with the library as controls. The transposon-mutant library samples, but not the WT, generated visible biofilms, although never to the extent observed for the biofilm mutant ( Figure 3B ). The 1:1 mixture of the biofilm-forming mutant and WT also produced an intermediate level of visible biofilm. Figure 3 RB-TnSeq analysis of biofilm forming cultures of the pooled TnSeq library. (A) Diagram of the RB-TnSeq library biofilm formation experiments and collection of samples. Triplicate samples were pooled prior to DNA extraction to ensure enough biomass and material for barcode sequencing. (B) Representative biofilms, or lack thereof, left attached to the glass after decanting and washing steps for the RB-TnSeq library, WT, a homogeneous biofilm formation strain, and a 1:1 mixture of WT and the biofilmation strain. Volcano plots of experiment 1 (E1) test tube (C) biofilm (BF) fractions and (D) settled fractions (S), and experiment 2 (E2) test tube fractions in BG-11 (E) biofilm, (F) settler, and (G) planktonic (P) fractions, presented as described for Figures 1B , C except that the x -axis is the fitness value and the y -axis is the T-value statistic. (H) Comparison of hit data for ORFs with significant fitness value changes, presented as described for Figure 1D , for all BG-11 test tube fractions. (I) Enrichment analyses for pili and known biofilm genes in the clusters of hits identified in (H) are presented as described for Figure 1E . Biofilm-forming library cultures were fractionated into three samples for barcode counting: ( Donlan and Costerton, 2002 ) planktonic cells that were removed by decanting the media, ( Harrison et al., 2019 ) settled cells that were removed from the emptied test tubes with gentle water washes, and ( DePas et al., 2014 ) sessile cells that required scraping to be removed from the glass culture vessels ( Figure 3A ). The planktonic sample from one replicate experiment did not produce usable sequence data. For all other experiments, the abundance of barcodes of individual transposon insertions were counted and used to calculate fitness values and statistical significance T -values for individual genes that reflect the abundance of those mutant genes in each fraction relative to their abundance in the starting culture ( Wetmore et al., 2015 ). The number of genes identified with statistically significant fitness value changes, called hits, in each fraction ranged from 16 to 103 ( Supplementary Figure S1 , Supplementary File S4 Sheet 3), much smaller than the initial DEG ranges observed for RNA-Seq, but similar to the ranges of sDEGs acquired after filtering for strong fold changes ( Supplementary Figures S1A,C ). In total, 195 genes were classified as hits in at least one of the examined samples, with 108 having absolute fitness scores greater than 1. These hit counts are both substantially reduced compared to, and have little overlap with, the sDEGs in the RNA-Seq experiment ( Supplementary Figure S10 , Supplementary Table S5 ), indicating that the RB-TnSeq data set provides a distinctly different perspective on biofilm formation than does RNA-Seq. Although the majority of genes, 1,725 or 89.8% of the 1,920 ORFs examined, show no significant fitness changes in any fraction examined ( Figures 3 , 4 , Supplementary File S4 Sheet 3), insertions in these genes were readily detected in all fractions ( Supplementary File S4 Sheet 1). These results indicate that the biofilm is composed not only of the mutants specially enriched in that fraction, but also of recruited or captured mutants that are also present in the planktonic fraction. It follows that a mutant that is under-represented in the biofilm may either be less fit to grow in the matrix or less likely to be recruited. Volcano plots demonstrate that more gene knockouts result in increased prevalence in the biofilm (positive hits) than decreased prevalence (negative hits; Figures 3C , E , 4A,C,D ). Positive hits in the fresh media test-tube experiments are dominated by known pili and biofilm genes whose loss enable film formation, such as pilB , while no pili or biofilm genes are present as negative hits in the biofilm fractions. Notably, the ebfG operon and pteB genes that prevent film formation when knocked out in a PilB ::Tn 5 background show no significant changes in fitness in any of the experiments, indicating that these mutants can be incorporated into a biofilm even when they are not producing the critical elements for biofilm formation themselves ( Supplementary File S4 Sheet 5). Analysis of the settled fractions from both test tube experiments show inconsistent results ( Figures 3D , F ). This inconsistency may be due to differences in the strength of the biofilms’ attachment to the glass vessel, so that wash steps differed in dislodging some of the biofilm as part of the settler sample. Nonetheless, the depletion of pili-related gene mutants and the ebsA mutant, which is known to lack pili ( Yegorov et al., 2021 ), from the planktonic fraction is consistent with the lack of pili resulting in faster rates of sedimentation ( Nagar et al., 2017 ; Figure 3G , Supplementary File S4 Sheet 5). Of the remaining 13 genes identified as strong negative hits (fitness < −1) in the planktonic fraction that are not known pili or biofilm associated genes ( Supplementary Table S6 group IG3, Supplementary File S4 Sheet 8), 11 are predicted to encode proteins with secretion motifs ( Supplementary Table S6 group IG4) but were not detected previously in the exoproteomes of WT, PilB ::Tn 5 , 1,127 Ω, or PilB ::Tn 5 /1127Ω ( Nagar et al., 2017 ; Parnasa et al., 2019 ). These genes either play some role in maintaining buoyancy or affect growth in the context of a biofilm. Figure 4 Conditioned media and flask RB-TnSeq analysis Volcano plots of CM test tube (A) biofilm and (B) settled fractions, and biofilm fractions from flask experiments with (C) BG-11 and (D) CM, are presented as described for Figures 1B , C . Comparison of fractions associated with (E) all test tube experiments and (F) flask experiments presented as described for Figure 1D . Comparative analysis of all fresh media test-tube experiments revealed three clusters of genes that represent consistent biofilm formers (TTBG1), and biofilm formers (TTBG2) or biofilm-depleted mutants (TTBG3) identified in one or more but not all of the experiments ( Figure 3H , Supplementary File S4 Sheet 9). TTBG2 is enriched only in known biofilm genes ( Figure 3I , Supplementary Figure S8 , Supplementary File S4 Sheet 14), and TTBG3 lacks any known pili or biofilm genes. TTBG1 includes six new loci in addition to 14 known pili and biofilm mutants. Significantly, known biofilm genes that were never identified as sDEGs in the RNA-Seq analysis are present in TTBG1 ( pilD, hfq, and ebsA ) and TTBG2 ( ebfE and pilA2 ). Thus, the remaining six genes in TTBG1, which include genes encoding a sigma factor, a response regulator, and a two-component sensor histidine kinase, represent likely candidates ( Supplementary Table S6 group IG5) for genes that participate in the self-suppression of biofilm formation, of which only Synpcc7942_0464, encoding a hypothetical protein, was a sDEG in the RNA-Seq analysis. The 59 novel hits in TTBG2 ( Supplementary Table S6 group IG6, Supplementary File S4 Sheet 9) may also participate in the self-suppression of biofilm formation, prevent sedimentation, or prevent attachment to a growing biofilm because mutations that affect any of these processes would increase propensity for inclusion in the biofilm. TTBG2 is enriched for proteins conserved among cyanobacteria or involved in translation and ribosomal biogenesis ( Supplementary File S4 Sheet 14), and it is unclear how they affect biofilm formation. The 36 genes in TTBG3 ( Supplementary File S4 Sheet 9) whose mutants are absent from the biofilm are enriched for those encoding amino acid transport and metabolism, cell motility and chemotaxis, and protein maintenance genes ( Supplementary File S4 Sheet 14), consistent with phenotypes that may be needed for either recruitment into or survival in the biofilm. In CM in test tubes, the settled fraction gave similar results to the biofilm fraction, where known biofilm and pili genes are enriched in the positive hits ( Figures 4A , B , and Supplementary Figure S8 ). Pattern analysis across all test tube experiments identified those that are unresponsive (TT1) and responsive (TT2) to CM ( Figure 4E , Supplementary File S4 Sheet 13). The majority of the known biofilm and pili genes are found in TT1, including pilB whose homogenous mutant culture does not produce biofilms in CM ( Schatz et al., 2013 ). We propose two hypotheses to explain the existence of transposon insertion mutants that are capable of producing biofilms, even in the presence of CM: (1) the self-suppressor generated by the WT-like cells in the transposon library experiments and in the mixed WT and mutant cultures was not at a high enough concentration to prevent biofilm formation at early time points, thus indicating a concentration-dependent window for enacting biofilm suppression, or (2) planktonic cells that would not initiate biofilm formation can be recruited into a mutant-seeded biofilm, which in CM would mean recruitment of CM-responsive mutant cells into a biofilm seeded by CM-unresponsive mutants. Although both hypotheses may contribute to the reduced impact of CM on heterogeneous cultures in this experiment, the latter is supported here as this analysis highlights nine novel genes identified as media-independent biofilm formers ( Supplementary Table S6 group IG7), a phenotype that was not previously encountered with S. elongatus biofilm mutants. This group includes a glycosyltransferase (Synpcc7942_0388) predicted to function in lipopolysaccharide production ( Simkovsky et al., 2012 ). The identified genes are typified by the potential to impact the properties of the cell surface and thus impact biofilm formation or incorporation. Additionally, the analysis identified three consistent biofilm forming mutants that are responsive to CM, which include sigF1 , Synpcc7942_0051, and Synpcc7942_B2646, and five mutants that biofilm only in CM, which includes a plasmid-encoded OmpR response regulator (Synpcc7942_B2647; Supplementary Table S6 groups IG 8 and IG9, Supplementary File S4 Sheet 13). Many known loci that were identified in test-tube assays did not display any significant fitness value changes in flasks ( Figures 4C , D , Supplementary File S4 Sheets 5, 11, and 12). Subtle differences between the assays, including shearing stresses and increased gas exchange associated with bubbling, may alter division rates in the biofilm or the overall strength of the biofilm and affect fitness scores. Fresh BG-11 in flasks resulted in 16 novel biofilm formation mutants that did not form biofilms in BG-11 test tubes ( Supplementary Table S6 group IG10, Supplementary File S4 Sheet 11), including tsaP and three mutants that had reduced prevalence in one of the test tube biofilms ( Supplementary Table S6 group IG11). Of these, five are genes involved in O-antigen production and susceptibility to grazing by amoeba, and were previously reported to result in cell aggregation, but not biofilm formation ( Piechura et al., 2017 ; Okshevsky et al., 2018 ; Supplementary Table S6 group IG12). We hypothesize that under the flask biofilm formation conditions, surface alterations that enhance aggregation or settling can more frequently enable incorporation into the heterogeneous biofilm. Of the 13 flask biofilm forming mutants that were responsive to CM (group F2, Figure 4F , Supplementary File S4 Sheet 12), only pilM , rntA , and tsaP are known pili genes and nine are specific to flask biofilm formation ( Supplementary Table S6 group IG13). Eight mutants are specifically enriched in the flask biofilm in the presence of CM that were not otherwise enriched in test tube biofilms or in fresh BG-11 in flasks ( Supplementary Table S6 group IG14), including three co-transcribed chemotaxis-like genes of the tax2 operon ( Supplementary Table S6 group IG15), whose function is unknown except that they are not involved in phototaxis in PCC 3055 ( Yang et al., 2018 ). Validation of Novel Genes Enabling or Repressing Biofilm Formation Together, the RNA-Seq and RB-TnSeq analyses identify hundreds of potential genes of interest ( Supplementary Table S5 ). To evaluate the efficacy of these two approaches toward novel target identification, we selected a subset of previously untested genes to knock out in WT or in PilB ::Tn 5 for biofilm evaluation, prioritizing those loci for which Unigene Set (UGS) transposon insertion vectors are available ( Holtman et al., 2005 ; Chen et al., 2012 ). Of the 91 mutants attempted in WT, 79 produced segregated knockout mutants of the intended gene of interest ( Supplementary Tables S7 and S8 ). Of these, we assayed candidates that were identified by: RNA-Seq only (30, 38%), RB-TnSeq only (38, 48%), or by both RNA-Seq and RB-TnSeq (11, 14%; Supplementary Table S7 ). Biological replicates of mutants were tested in bubbling test tubes, 96-well plates, and sometimes in flasks to increase the throughput of the biofilm assays and because different vessels impact the inclusion of a mutant in the biofilm. Most of the mutants created in a WT background did not form biofilms (49 of the 79, or 62%). Candidates based on RNA-Seq alone accounted for a higher false hit rate with 23 of the 30 assayed (77%) compared to 21 of the 38 assayed (55%) based on RB-TnSeq data alone. The remaining 30 mutants formed biofilms to some degree and were classified based on the strength, reproducibility, and diversity of vessels in which biofilms were formed. Sixteen resulted in more replicates without any biofilms than with and were deemed “mostly WT-like.” Of the remaining 14 mutants, most produced biofilms only sporadically or in a vessel-dependent manner. Only four mutants consistently produced strong biofilms in multiple vessel assays—the predicted TPR-repeat lipoprotein with an O-linked N-acetylglucosamine transferase domain (Synpcc7942_0051) that was identified in the TTBG1 RB-TnSeq group; a gene predicted without experimental confirmation by Taton et al. (2020) to encode the pilus assembly protein PilP (Synpcc7942_0168), whose impact on biofilm formation was not previously known; the gene encoding the RNA polymerase sigma factor SigF (Synpcc7942_1510); and the gene encoding the pilus assembly protein PilQ (Synpcc7942_2450). Notably, all genes identified from the TTBG1 RB-TnSeq group of interest whose validated mutations produced strong biofilm phenotypes were either never identified as sDEGs in the RNA-Seq analysis or had patterns of expression that were inconsistent with previous biofilm mutants ( Supplementary Tables S6 group IG16, and S7 ). Similarly, 13 mutations were chosen for investigation in a PilB ::Tn 5 background and assayed for mutants that no longer produce biofilms. Although none of the mutations completely abolished biofilm formation, knockout of a deacetylase (Synpcc7942_1393; DEG in RNA-Seq group O-A4 full, Supplementary Table S9 , Supplementary File S3 Sheet 15) or a pimeloyl-ACP methyl ester carboxylesterase (Synpcc7942_0774) decreased the reliability or strength of biofilm formation. Transcriptomics and Phenomics Reveal Novel Information on the Biofilm State and the Quality of the Genome Model The RNA-Seq time course data and enrichment analysis results highlight phenotypic changes associated with adjusting to the self-shaded, nutrient and gas exchange-reduced biofilm lifestyle, including general stress responses, changes in iron, sulfur, and nutrient transportation and metabolism, protein maintenance, and energy generation through photosynthesis ( Figure 5 ). The absence of mutants affected in many of these same genes in the biofilm fractions of the RB-TnSeq experiments support the hypothesis that these functions are essential for survival in biofilms. Given the substantial change in the cells’ niche as it forms biofilms, it is not surprising that a large portion of the transcriptome changes to properly adapt the cell for survival in the biofilm state, as has been observed in other systems ( Nagar and Schwarz, 2015 ; Yan et al., 2017 ; Okshevsky et al., 2018 ). Figure 5 Diagram of biofilm formation for Synechococcus elongatus PCC 7942. Self-suppression of biofilm formation in WT S. elongatus PCC 7942 is mediated by secretion through the type IV pilus (T4P) apparatus and its association with pili and a tripartite complex of the PilB, EbsA, and Hfq proteins. Suppressor present in conditioned media (CM) represses biofilm formation in some biofilm forming mutants. Removal of suppressing elements enables biofilm formation through the expression of the EbfG proteins and their processing and secretion, which involves PteB and EbfE. RNA-Seq and RB-TnSeq findings that add to this model are highlighted in yellow boxes. Strand-specific transcriptional data provided here and in previous publications ( Vijayan et al., 2011 ; Billis et al., 2014 ) can improve the genome models of PCC 7942 and related strains through resolving discrepancies between the observed transcriptome and the current genome annotation. The data confirmed that pilB is the first of four genes expressed as a single transcriptional unit ( Supplementary Figure S9A ; Vijayan et al., 2011 ). PilB::Tn5 transcript levels were reduced over 27 fold for the portion of pilB following the insertion site and the downstream genes including pilT and pilC , suggesting a polar effect. Phenotypic complementation, however, of PilB::Tn5 by introduction of pilB in trans ( Yegorov et al., 2021 ) indicates sufficient expression of the downstream genes in the mutant. Transcriptional coverage of the hfq gene does not exceed one read count until positions annotated as codons 25 or 27, depending upon the strain analyzed, suggesting that an alternative GTG start codon previously designated as amino acid position 28 begins the open reading frame ( Supplementary Figure S9B ). This shorter annotation is consistent with the NCBI conserved domain database’ identification of an Sm-like RNA binding domain from codons 35 to 95 ( Marchler-Bauer et al., 2015 ). RNA-Seq Compared With RB-TnSeq As the results for hfq and ebsA highlight, some key genes may not be identified by transcriptomic approaches because protein levels in cyanobacteria are poorly correlated with transcript levels ( Xiong et al., 2015 ) or protein functions are regulated post-transcriptionally. In contrast, RB-TnSeq directly probes the connection between genotype and phenotype in a high-throughput manner ( Wetmore et al., 2015 ; Price et al., 2018 ). In this project, we leveraged the self-fractionation of the sample that occurs through biofilm formation as an application for an RB-TnSeq library. In comparison to experiments where changes in fitness score reflect mutants’ rate of division relative to the rest of the library’s population in the presence of a selective condition ( Wetmore et al., 2015 ; Price et al., 2018 ; Rubin et al., 2018 ; Welkie et al., 2018 ; Taton et al., 2020 ), this application’s fitness scores are also affected by the capacity to reproducibly fractionate the cultures as well as mutation- and niche-dependent variations in rates of growth, survival, recruitment of WT-like cells into the biofilm, and settling, which is more reflective of the ability to remain planktonic rather than the capacity to generate biofilm structures. Nonetheless, the RB-TnSeq strategy provides a more direct identification of both known and novel gene products that participate in a complex behavior such as biofilm formation, even with fewer experimental samples, than does RNA-Seq. This conclusion is based on the higher rate of validation of hits for the RB-TnSeq data than the RNA-Seq data and the lack of known biofilm regulatory genes in the transcriptomics data ( Supplementary Figures S1 , S10 ). Although the differences in incubation times (4 days vs. 2 weeks), between homogenous and heterogenous mutant cultures, and the inherent questions that these two methods address means that this comparison is a complicated one, the data supports a greater success rate for target identification using RB-TnSeq than for RNA-Seq. Because RB-TnSeq experiments require less sequence coverage than RNA-Seq samples, this strategy is also a less expensive methodology for identifying novel candidates involved in nuanced and complex behavior. Because the WT S. elongatus PCC 7942 background constitutively represses biofilm formation, the experimental design would not identify mutants that are defective in components that are required for biofilm formation, such as the ebfG - pteB genes. These genes show the highest fold changes in transcript expression related to biofilm formation, but their inactivation in the RB-TnSeq library showed no significant changes in fitness values, likely due to the secretion of their encoded proteins by other biofilm forming mutants in the library. Future applications of Interaction RB-TnSeq (IRB-Seq), in which a second mutation (e.g., pilB ::Tn 5 ) is introduced into the RB-TnSeq library ( Rubin et al., 2018 ) to stimulate biofilm formation, or an RB-TnSeq library constructed in a native biofilm forming strain such as UTEX 3055, is more likely to identify genes that are involved in biofilm formation process rather than the self-suppression mechanism. Novel Genes of Biofilm Regulation in Synechococcus elongatus Of the large number of potential candidates for participating in regulation or formation of biofilms ( Supplementary Table S5 ), only four resulted in strong and consistent biofilm phenotypes: Synpcc7942_0051, Synpcc7942_0168, sigF (Synpcc7942_1510), and pilQ (Synpcc7942_2450). Synpcc7942_0051 shares a domain with proteins in heterotrophic bacteria that are transcriptionally regulated by a sigma-54-interacting response regulator and a histidine kinase and participate in exopolysaccharide production and protein export in biofilm forming bacteria ( Haft et al., 2006 ). Experiments based on the data presented here have recently demonstrated that this protein is in fact a glycosyltransferase that glycosylates PilA ( Suban et al., 2022 ). If regulation of this protein occurs in a similar manner in PCC 7942 biofilms, then histidine kinases and response regulators that were identified as potential genes of interest ( Supplementary Table S6 group IG17) are candidates for participating in the regulatory circuit for controlling biofilm self-suppression. These validated hits, combined with the large set of genes with unknown function highlighted by the transcriptome and phenome data sets, serve as a starting point for future investigations of novel mechanisms of biofilm regulation in S. elongatus PCC 7942."
} | 13,054 |
32546625 | PMC7298715 | pmc | 8,798 | {
"abstract": "Bacteria within a swarm move characteristically in packs, displaying an intricate swirling motion in which hundreds of dynamic rafts continuously form and dissociate as the swarm colonizes an increasing expanse of territory. The demonstrated property of E. coli to reduce its tumble bias and hence increase its run duration during swarming is expected to maintain and promote side-by-side alignment and cohesion within the bacterial packs. In this study, we observed a similar low tumble bias in five different bacterial species, both Gram positive and Gram negative, each inhabiting a unique habitat and posing unique problems to our health. The unanimous display of an altered run-tumble bias in swarms of all species examined in this investigation suggests that this behavioral adaptation is crucial for swarming."
} | 204 |
28782022 | PMC5533546 | pmc | 8,799 | {
"abstract": "High thermal conductivity in amorphous polymer films via ionization-induced chain extension and stiffening, and dense packing.",
"introduction": "INTRODUCTION Effective thermal management in applications such as batteries, automobile cooling systems, and high–power density electronic devices, where heat accumulation can have deleterious effects, is critically important to ensure system performance and reliability and to enhance lifetime. Despite their poor thermal conductivity (κ), various advantages, including light weight, low cost, and easy processability, make polymers the material of choice for several heat-intensive applications such as electronic chip encapsulation, cell phone casing, and LED (light-emitting diode) housing. These existing applications, along with emerging technologies such as flexible electronics, for which the requirements on flexibility and light weight cannot be met by most conventional thermal management materials (metals and ceramics), put greater technological incentives on developing thermally conductive polymers. Blending with high-κ fillers such as metal or ceramic particles, carbon nanotubes (CNTs), or graphene flakes is the most commonly used method to enhance polymers’ thermal conductivity ( 1 ). However, the large volume fraction of fillers required to achieve appreciable enhancement in κ often leads to undesired optical or electrical properties, increased weight, high cost [for example, ~$1000/kg for CNT versus ~$2/kg for poly(methyl methacrylate)], or loss of the easy processability generally associated with polymers. In contrast to low κ in bulk samples, constituent individual polymer chains are believed to have large κ. The thermal conductivity of a single polymer chain, in which the elastic disorder ( 2 ) between intrachain covalent and interchain van der Waals bonds is absent, was calculated to be as large as few hundreds of watts per meter per kelvin ( 3 ). Ultradrawn crystalline nanofibers with aligned polymer chains were measured to have κ of more than 100 W m −1 K −1 in the alignment direction ( 4 ). The large thermal conductivities of single- or few-chain fibers can be retained in amorphous polymers in the direction of chain orientation ( 5 , 6 ), along which heat propagation occurs predominantly through intrachain transport. Singh et al . ( 7 ) reported a significant increase in κ in amorphous polythiophene fabricated via a nanotemplate-assisted electrochemical method that allows polythiophene chains to be oriented in the vertical direction. Thermal conductivity greater than 2 W m −1 K −1 has been similarly reported for covalently grafted poly(3-methyl thiophene) brushes. Covalent grafting led to enhanced chain alignment as well as reduced energetic and positional disorder in these surface-grafted films ( 8 ). These high thermal conductivities reported in polymers with extended chain conformation stand in contrast to surface-grown polymer brushes ( 9 ) and polymer films under high pressure ( 10 ), in which the coiled conformation of polymer chains likely remained and enhancement in κ was found to be relatively moderate. However, these approaches either limit the orientation of chain extension to a certain direction or pose challenges in terms of scaling up the nanoscale films for practical applications. Therefore, it is desirable to achieve high κ in both in- and out-of-plane directions in bulk amorphous polymers using common fabrication processes. Although the mechanisms of thermal transport in amorphous materials continue to be studied ( 10 – 13 ), it is generally believed that the thermal conductivity in bulk amorphous polymers (a class of disordered solids) is inhibited by the following: (i) highly coiled and entangled intrachain structure, (ii) loose chain packing with voids that dampen the speed at which vibrations propagate, and (iii) weak nonbonding interchain interactions (for example, van der Waals and dipole-dipole) ( 14 ). Here, we demonstrate an unexplored molecular engineering route that simultaneously attacks these three bottlenecks. By using the coulombic repulsive forces between ionized pendant groups on the backbone of polyelectrolytes to “stretch” the main chain at the molecular level, we achieve significant enhancements of thermal conductivity in amorphous polymer with randomly oriented yet superiorly packed extended polymer chains and strong ionic interchain interactions ( Fig. 1A ). Fig. 1 High thermal conductivity in polyelectrolyte thin films via controlled ionization. ( A ) Illustrations of chain conformation and packing in spin-cast polymer films: coiled unionized polyelectrolyte (left) and extended ionized polyelectrolyte (right). The zoomed-in images show chain conformation at the molecular level. ( B ) Cross-plane thermal conductivity of a weak polyelectrolyte, PAA [molecular weight (MW), 100,000], and a nonionizable water-soluble polymer, PVP (MW, 40,000), thin films spin-cast from polymer solutions of different pH. Error bars were calculated on the basis of uncertainties in film thickness, temperature coefficient of electrical resistance for the heater, and heater width. Chemical structures of the polymers and ionization reaction for PAA are also shown. ( C ) Fourier transform infrared (FTIR) spectra of PAA films spin-cast from solutions of different pH. ( D ) Fraction of ionized carboxylic acid groups (α) as a function of solution pH: calculated from the FTIR spectra and by applying charge balance on PAA solutions.",
"discussion": "DISCUSSION In summary, we have used electrostatic repulsive forces to stretch the polyelectrolyte backbone at the molecular level, resulting in extended conformations, better packed chains, and enhanced modulus, all of which contribute to significantly enhanced thermal conductivities. For the spin-cast thin films, it is to be noted that centrifugal forces during spin casting may cause polymer chains to be more expanded in the in-plane direction, possibly making in-plane thermal conductivity even greater than the measured cross-plane κ ( 6 ). This unexplored route for molecular engineering of polymer thermal conductivity is also extended to making micrometer-thick blade-coated films, with thermal conductivity reaching more than 0.6 W m −1 K −1 ."
} | 1,558 |
36475793 | PMC9728961 | pmc | 8,800 | {
"abstract": "We introduce damage intelligent soft-bodied systems via a network of self-healing light guides for dynamic sensing (SHeaLDS). Exploiting the intrinsic damage resilience of light propagation in an optical waveguide, in combination with a tough, transparent, and autonomously self-healing polyurethane urea elastomer, SHeaLDS enables damage resilient and intelligent robots by self-healing cuts as well as detecting this damage and controlling the robot’s actions accordingly. With optimized material and structural design for hyperelastic deformation of the robot and autonomous self-healing capacity, SHeaLDS provides reliable dynamic sensing at large strains (ε = 140%) with no drift or hysteresis, is resistant to punctures, and self-heals from cuts at room temperature with no external intervention. As a demonstration of utility, a soft quadruped protected by SHeaLDS detects and self-heals from extreme damage (e.g., six cuts on one leg) in 1 min and monitors and adapts its gait based on the damage condition autonomously through feedback control.",
"introduction": "INTRODUCTION The past two decades have seen proliferating growth in the maturing field of stretchable sensors. Using intrinsically soft materials or structurally compliant designs, stretchable sensors have demonstrated mechanically invisible monitoring of soft-bodied systems that are otherwise challenging for the silicon-based rigid counterparts ( 1 – 4 ). For sensing applications in damage-prone environments, intrinsically stretchable sensors constituted with elastomers are inherently damage resistant to blunt impacts (e.g., running over by a car), owing to the good toughness and resilience of elastomers ( 5 – 8 ); on the other hand, they are vulnerable to damage modes such as cuts and punctures. Incorporating self-healing properties into these intrinsically stretchable sensors promises not only damage-resistant but also damage-resilient sensing capabilities. Other than the immediately followed advantage of device longevity, self-healing stretchable sensors present a unique opportunity to establish damage intelligence into soft-bodied systems, similarly vulnerable to sharp damages. Examples range from damage detection in spacesuits to alert suit air leakage by space debris during astronauts’ extravehicular activities ( 9 ) to robots that are not only resilient to damage ( 10 – 12 ) but also intelligent toward damage (i.e., able to detect the onset of damage and adapt through feedback control) when working in hazardous environments. To enable damage intelligence, stretchable strain sensors need to (i) be tough and autonomously self-healable for both resistance and resilience toward damage and (ii) able to detect damage and provide reliable dynamic sensing performance (i.e., sensitivity, range, and bandwidth) for the intended strain sensing application (e.g., motion monitoring as robotic skins or wearables). The major challenge in building such a system is achieving a self-healing sensor that satisfies both requirements at the same time, and the challenge partially originates from the complexity of the self-healing material itself. Self-healing function for strain sensors comprises the recovery of material’s mechanical properties and the restoration of sensing signals. Self-healing of polymers’ mechanical properties has been realized through (i) extrinsic self-healing with healing reagents encapsulated in microcapsules ( 13 , 14 ) or fed through vascular networks ( 15 ) (ii) intrinsic self-healing by the reorganization of polymer chains and bonds at the damaged site with stimuli or autonomously without external energy input through dynamic bonding of supramolecular interactions or dynamic covalent chemistry ( 16 – 19 ). Among these mechanisms, intrinsic self-healing through dynamic bonding promises a simple embodiment as the basis of autonomous self-healing sensors. Incorporating conductive particulates in self-healing materials, self-healing of the sensing signals has been achieved through electrical sensing mechanisms [e.g., piezoresistive ( 20 ), ionic ( 21 , 22 ), or capacitive ( 23 )] and has demonstrated quasi-static measurements (e.g., strain/flexion/pressure) ( 24 ). Few sensors, however, have been able to provide satisfactory sensing performance for dynamic strain measurements required in motion monitoring. Dynamic sensing with self-healing sensors usually exhibits large drift under continuous cyclic loading, hysteresis, heavy strain rate dependency, and limited strain range due to the intrinsic viscoelasticity from the dynamic bonding that enables self-healing. Sensors that are both autonomously self-healable and reliable in dynamic sensing have yet to be achieved. We introduce an autonomous self-healing optical sensing mechanism, networks of self-healing light guides for dynamic sensing (SHeaLDS), which exploit the damage-resilient properties intrinsic to light propagation, in combination with an intrinsic self-healing material to achieve an optomechanical sensor that both autonomously self-heals and provides reliable dynamic sensing performance. Compared to electrical signals that require physical contact for transmission, light propagation in a waveguide through total internal reflection does not require mechanical contact. Therefore, the sensing signal (i.e., output light intensity) of a self-healing optical waveguide sensor is intrinsically resilient to damage and can be autonomously recovered not only when the damaged interfaces connect but also when small gaps remain (e.g., material removal through punctures). The self-healing function of the sensor is then simplified to self-healing the mechanical properties of the sensor material. To overcome the contradiction between autonomous self-healing and material viscoelasticity-driven sensor drift and hysteresis, we design SHeaLDS into a wavy shape that achieves good dynamic sensing through structural compliance. We demonstrate SHeaLDS’ fast and autonomous self-healing, damage detection, and dynamic sensing capabilities by integration onto self-sealing actuators to construct a soft quadruped that is intelligent about damage.",
"discussion": "DISCUSSION We have introduced SHeaLDS, an autonomously self-healing and damage-resilient stretchable optomechanical sensor that achieves hyperelastic strains (140%), resilient to large works of extension (60.6 MJ m −3 ), and room temperature self-healing abilities, as well as reliable dynamic sensing performance to these large strains. Although ε ~ 100% self-healable strain of the material at room temperature is shown for our robotic demonstration, over ε > 1000% self-healable strain can be achieved with sPUU1000 at room temperature or sPUU2000 with heating for applications involving extreme deformations. Expediated healing could be achieved by heating ( Fig. 1F ). Distributed heating elements, such as resistive heating or photothermal heating through embedded nanoparticles, could be incorporated to achieve faster self-healing rates with onboard control. We demonstrate a damage intelligent soft robot, enabled by SHeaLDS, that senses injury and self-heals from severe damage on the molecular level within minutes. For robotic applications, a limitation of the intrinsically self-healing scheme is the requirement for material contact to allow self-healing of mechanical strength. The environment robots operate in could introduce impurities, such as debris or chemicals, and when damage is incurred, surface fouling could prevent clean surfaces to contact and self-heal. Proper encapsulation, design, and extrinsic or a combination of extrinsic and intrinsic self-healing mechanisms may be necessary to address the demanding self-healing conditions in these applications. Damage intelligence realized through SHeaLDS demonstrates autonomy in damage sensing, healing, and adaptation resembling physical intelligence in animals. Combined with advances in artificial intelligence, SHeaLDS presents a route toward more enduring and adaptive robots. Beyond robotics, we envision SHeaLDS to find utility in the broader area where damage intelligence is essential in damage-prone environments, such spacesuits and supersonic parachute monitoring in space ( 44 ), as well as applications where device longevity is preferred, such as wearables for human machine interaction and digital health."
} | 2,082 |
39288179 | PMC11441501 | pmc | 8,801 | {
"abstract": "Significance Microbes commonly reside within complex, multispecies communities in nature. Recent advances in high-throughput metagenomic sequencing have offered unprecedented systems-level insight into the diversity of microbial communities. However, the dissection of chemical and biological interactions is often limited by tools optimized for studying individual microbes or single compounds. Here, we report a modular and scalable microbial coculture platform designated the microbial community interaction (µCI) device for systematically measuring populations of a target strain when treated with three-member combinations of compounds or microbes. The µCI device is a simple platform that could become a routine tool in combinatorial screening of chemical and biological factors influencing microbial growth or gene expression.",
"discussion": "Discussion Here, we describe the µCI platform for systematic study of microbial community interactions. The µCI device provides measurements of chemical and microbial interactions within three-member subsets of a community and was designed to be easy to operate with only pipettes and a fluorescent microscope, making it accessible to nonengineering labs. The µCI device demonstrated its utility by rapidly identifying a rhizosphere strain, Pseudomonas sp. CI14, as highly inhibitory to other community members. The same conclusion was reached previously with laborious, pairwise experiments on agar plates and a hierarchy model ( 51 ), whereas the present experiment required only ~2 h for setup and 2 d for coculture and readout. The µCI device also makes it easy for finding positive interactions such as growth stimulators. These are underrepresented interactions in the literature since there is a lack of simple-to-use, high-throughput platforms for identifying these interactions, so their prevalence and significance in nature is not known. Also, as evidenced in our rhizosphere microbe coculture data, the µCI device enables the identification of emergent properties between members of the same community. A current limitation of the µCI device is the requirement for a fluorescently labeled target strain, which reduces its utility in studying microbes for whom a fluorescent protein reporter is not available. This also limits the platform’s ability to assay microbial communities that require anaerobic conditions (such as gut communities), as most traditional protein fluorophore reporters do not function in anaerobic conditions ( 57 ). It will be of interest to develop alternative microbial growth quantification strategies such as chemical reporter dyes or nonfluorescent imaging and image processing methods to assay microbes without a fluorescent reporter. We also envision that the µCI device will be a good platform for screening combinations of candidate microorganisms for probiotics for the treatment of diseases that are influenced or caused by microorganisms or the study of the resistance or susceptibility of communities to invaders. Perhaps the greatest power of the µCI device lies in its potential for identifying regulatory networks in communities. By replacing the constitutive GFP construct with one containing a promoter of interest, hundreds of chemicals or organisms could be screened simultaneously to find regulatory factors. Since genes that are uniquely important for survival in a community are largely undiscovered or poorly characterized, this capability could revolutionize the process of assigning functions for community-specific genes and identifying regulatory cascades. In addition, gene clusters responsible for the synthesis of small molecules are often silent. That capacity for rapid screening of potential inducers of expression could facilitate the finding of new bioactive molecules with potential for applications in agriculture and medicine."
} | 958 |
26236397 | PMC4522091 | pmc | 8,804 | {
"abstract": "Background Few strains have been found to produce isobutanol naturally. For building a high performance isobutanol-producing strain, rebalancing redox status of the cell was very crucial through systematic investigation of redox cofactors metabolism. Then, the metabolic model provided a powerful tool for the rational modulation of the redox status. Results Firstly, a starting isobutanol-producing E. coli strain LA02 was engineered with only 2.7 g/L isobutanol produced. Then, the genome-scale metabolic modeling was specially carried out for the redox cofactor metabolism of the strain LA02 by combining flux balance analysis and minimization of metabolic adjustment, and the GAPD reaction catalyzed by the glyceraldehyde-3-phosphate dehydrogenase was predicted as the key target for redox status improvement. Under guidance of the metabolic model prediction, a gapN -encoding NADP + dependent glyceraldehyde-3-phosphate dehydrogenase pathway was constructed and then fine-tuned using five constitutive promoters. The best strain LA09 was obtained with the strongest promoter BBa_J23100. The NADPH/NADP + ratios of strain LA09 reached 0.67 at exponential phase and 0.64 at stationary phase. The redox modulations resulted in the decrease production of ethanol and lactate by 17.5 and 51.7% to 1.32 and 6.08 g/L, respectively. Therefore, the isobutanol titer was increased by 221% to 8.68 g/L. Conclusions This research has achieved rational redox status improvement of isobutanol-producing strain under guidance of the prediction and modeling of the genome-scale metabolic model of isobutanol-producing E. coli strain with the aid of synthetic promoters. Therefore, the production of isobutanol was dramatically increased by 2.21-fold from 2.7 to 8.68 g/L. Moreover, the developed model-driven method special for redox cofactor metabolism was of very helpful to the redox status modulation of other bio-products. Electronic supplementary material The online version of this article (doi:10.1186/s13068-015-0291-2) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusions In this work, a rational strategy combining GSMM and synthetic promoters was employed to rebalance redox status for higher isobutanol production. A strain LA02 was firstly engineered with NADPH-dependent isobutanol synthesis pathway including heterologous Ehrlich pathway (consisting of kivd and yqhD genes) and biosynthetic 2-ketoisovalerate precursor pathway (consisting of alsS, ilvC and ilvD genes). But the isobutanol production of strain LA 02, only 2.7 g/L with 36 g/L glucose as substrate, was limited by the imbalance of redox status. The metabolism of NADH and NADPH in the GSMM was deeply investigated with the combination of FBA and MOMA algorithms, and the key target for rebalancing the redox status was predicted to be GAPDH. Consequently, a GAPDN pathway was designed and implemented experimentally to simultaneously modulate the NADH and NADPH generation. Abilities of two GAPDNs encoded by gapC and gapN from Clostridium acetobutylicum to improve redox status were compared, and the NADPH/NADP + ratios in the strain LA04 (gapN expressed) were 0.56 and 0.60 at exponential phase and stationary phase, respectively, 1.697- and 1.714-fold of that in strain LA03 (gapC expressed). The higher NADPH/NADP+ ratio indicated gapN-encoding GAPDN was more efficient to modulate redox status. Hence, gapN was the better choice. Furthermore, a suitable flux through GAPDN pathway was necessary to get a best redox status for maximizing the isobutanol production, thus, a fine-tuning of gapN expression was progressed using five different artificial constitutive promoters. As the strengths of promoters increased, the NADPH/NADP + ratios were shifted from 0.40 to 0.67 at exponential phase and from 0.43 to 0.64 at stationary phase. The best strain LA09 was finally obtained with the gapN expressed under the control of BBa_J23100 promoter. The isobutanol production of strain LA09 reached to 8.68 g/L, 2.21-fold higher than that of strain LA02, while the ethanol and lactate productions were respectively decreased by 17.5 and 48.5% to 1.32 and 6.08 g/L. The present study highlights the power and promise of the rational method combining GSMM and synthetic promoters in redox status modulation.",
"discussion": "Discussion In this work, the strain E. coli LA02 was engineered for isobutanol production by combining NADP + dependent Ehrlich pathway and biosynthetic KIV precursor pathway, but, as the key limitation for isobutanol production, the imbalance redox status with excess NADH and less NADPH supply should be solved to improve the strain performances. Actually, several studies had focused on the redox balance for increasing isobutanol production, including construct a isobutanol pathway using the redundant NADH as cofactor [ 17 ], or forcing high flux through the NADPH generating pathway, such as transhydrogenase [ 13 , 17 ] and PP pathway [ 8 ]. Especially, Qi et al. [ 8 ] used a compacted metabolic model to predict the targets for enhance isobutanol by B. subtilis using the Elementary mode analysis, and a strategy of pgi deletion and zwf overexpression was carried out with the EMP pathway disrupted, and resulted in excess supply NADPH to inhibit cell growth, which was overcome by expressing the UdhA from E. coli . However, the GAPDH involving in the EMP pathway was rationally predicted as a new key target for redox status improvement based on the GSMM in this study. Unlike the overexpression of NADPH generating pathway, the cofactor-swap of GAPDH could not only introduce a new pathway to efficiently generate NADPH depending on the high-flux dominate EMP pathway, but also eliminate the major NADH generation pathway carrying 83.6% of total NADH generation flux according model analysis (Additional file 1 : Figure S1). In practice, the cofactor-swap of GAPDH could be achieved by two approaches, one of which was to change the cofactor preference of the E. coli native glycerate-3-phosphate dehydrogenase by enzyme engineering, but it was seemingly difficult to be applied for fine-tuning the redox status because the engineered enzymes usually had catalytic efficiency both towards NAD + and NADP + , rather than used NADP + as the sole cofactor according to description in previous [ 18 ]. Fortunately, as another approach, heterologous GAPDN that catalyzed the irreversible reaction from G3P to 3PG had been found in some gram-positive bacteria such as Streptococcus and Clostridium species [ 34 ]. In this work, the gapN gene was obtained from C. acetobutylicum to encode the GAPDN pathway. Furthermore, two methods should be considered for engineering the target reaction, including replacing the native GAPDH with the GAPDN or directly heterologous overexpressing the GAPDN. For the first one, it may result in overregulation of redox status with redundant NADPH generation to cause new burden on the cell metabolism [ 28 ]. Takeno et al. replaced the NAD + dependent GAPDH of C. glutamicum strain with GAPDN from S. mutans , and the mutant strain showed significant growth deficiency [ 19 ]. In this case, the result was very consistent with the prediction of metabolic network modeling that the cofactor-swap of GAPDH decreased the specific growth rate of strain by 75.0% (Table 2 ). For the second method, it could provide a moderate improvement of redox status rather than overregulation. It has been proved feasible and practical to overexpress the gapN in the strain LA04, of which the intracellular NADPH/NADP + ratio increased by 60.0% (EP) with the isobutanol production and yield respectively increased by 95.6 and 58.3% compared to those of the stain LA02. Centeno-Leija et al. [ 28 ] also demonstrated that overexpression of GAPDN showed a better promising performance than the replacement of native GAPDH with GAPDN in modulating NADPH supply to increase polyhydroxyalkanoates production. However, it was seemingly unable to get an optimal redox status for maximizing the isobutanol production only by expression of GAPDN under the gapA promoter, and the experiment results also demonstrated the intracellular NADPH/NADP + ratio in strain LA04 (0.56 at EP) was still lower than that in the wild-type (0.65–0.99). Thus, a further modulation was necessary for increasing NADPH generation to satisfy the demand of isobutanol biosynthesis. In this work, a recovery redox status to wild-type state was achieved by a step by step increase of NADPH generation based on the modulation GAPDN pathway under the regulation of five promoters with gradual-increasing strength, and a strain LA09 of the best performance was obtained with the strongest promoter BBa_J23100. It might be doubtful that if more isobutanol would be produced with a promoter stronger than BBa_J23100 placed in front of GAPDN. But gapN has been expressed under the Trc promoter, and the cell growth of E. coli strain was decreased [ 28 ]. Thus, it did not have to get more isobutanol production with a stronger promoter. Finally, the isobutanol production and yield of the best strain LA09 were respectively increased by 2.21- and 1.50-fold compared to that of strain LA02, while the productions of byproducts (ethanol and lactate) respectively decreased by 17.5 and 51.7%. Nevertheless, as branch pathways, the biosynthesis of ethanol and lactate should be completely eliminated to reroute more carbon flux into isobutanol synthesis in future work, and then, for the new strain, it would be also need to further fine-tune the fluxes of GAPDH and GAPDN pathway to get optimal redox status for high efficient isobutanol production. Above all, as a key limitation of isobutanol production, the redox status imbalance was solved by a rational method in this work. The GSMM was used as a powerful platform to gain insight into the redox cofactors metabolism, and the most important target was predicted to be the GAPDH. Based on the model analysis and prediction, the GAPDN pathway encoded by gapN was consequently rationally designed, and then modulated with the help of artificial, constitutive promoters to improve the redox status. Finally, an efficient stain LA09 was obtained. Collectively, it indicated the GSMM combination with synthetic promoters was a powerful and promising strategy for the rational improvement of redox status, and this strategy could also be utilized to efficiently construct or optimize other cell factories for various chemicals and fuels production."
} | 2,632 |
39806455 | PMC11730135 | pmc | 8,805 | {
"abstract": "Background The widespread selective pressure of antibiotics in the environment has led to the propagation of antibiotic resistance genes (ARGs). However, the mechanisms by which microbes balance population growth with the enrichment of ARGs remain poorly understood. To address this, we employed microcosm cultivation at different antibiotic (i.e., Oxytetracycline, OTC) stresses across the concentrations from the environmental to the clinical. Paired with shot-gun metagenomics analysis and quantification of bacterial growth, trait-based assessment of soil microbiota was applied to reveal the association between key ARG subtypes, representative bacterial taxa, and functional-gene features that drive the growth of ARGs. Results Our results illuminate that resistome variation is closely associated with bacterial growth. A non-monotonic change in ARG abundance and richness was observed over a concentration gradient from none to 10 mg/l. Soil microbiota exposed to intermediate OTC concentrations (i.e., 0.1 and 0.5 mg/l) showed greater increases in the total abundance of ARGs. Community compositionally, the growth of representative taxa, i.e., Pseudomonadaceae was considered to boost the increase of ARGs. It has chromosomally carried kinds of multidrug resistance genes such as mexAB - oprM and mexCD-oprJ could mediate the intrinsic resistance to OTC. Streptomycetaceae has shown a better adaptive ability than other microbes at the clinical OTC concentrations. However, it contributed less to the ARGs growth as it represents a stress-tolerant lifestyle that grows slowly and carries fewer ARGs. In terms of community genetic features, the community aggregated traits analysis further indicates the enhancement in traits of resource acquisition and growth yield is driving the increase of ARGs abundance. Moreover, optimizations in energy production and conversion, alongside a streamlining of bypass metabolic pathways, further boost the growth of ARGs in sub-inhibitory antibiotic conditions. Conclusion The results of this study suggest that microbes with competitive lifestyles are selected under the stress of environmental sub-inhibitory concentrations of antibiotics and nutrient scarcity. They possess greater substrate utilization capacity and carry more ARGs, due to this they were faster growing and leading to a greater increase in the abundance of ARGs. This study has expanded the application of trait-based assessments in understanding the ecology of ARGs propagation. And the finding illustrated changes in soil resistome are accompanied by the lifestyle switching of the microbiome, which theoretically supports the ARGs control approach based on the principle of species competitive exclusion. \n Video Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s40168-024-02005-6.",
"conclusion": "Conclusions This study uncovers the capacity of soil microbial communities to grow ARGs across gradients of antibiotic concentrations, shedding light on the correlating bacterial taxa and their functional genetic features. Results of the analysis of the community aggregated traits demonstrated that a competitive lifestyle facilitating the growth in ARGs by enhanced energy production and resource utilization, particularly under conditions of nutrient scarcity met the sub-inhibitory of antibiotics.",
"introduction": "Introduction Antibiotic resistance is one of the greatest threats to public health today [ 1 ], and the environment is thought to play a critical role in its development and spread [ 2 ]. Investigating how microbes balance population growth with the per-microbe enrichment of antibiotic resistance genes (ARGs) is vital, as nutrient scarcity and sub-inhibitory antibiotic stress are the main constraints that influence the increased abundance of ARGs [ 3 , 4 ]. Microbes navigate trade-offs between reproduction, survival, and competition in conditions with resource limitations and survival stress [ 5 , 6 ]. These trade-offs could be framed within the trait-based life history strategy (LHS) framework, which was originally developed to elucidate the mechanisms by which organisms adapt to specific environments through trait selection [ 7 , 8 ]. Advancements in trait-based research methods have currently progressed by microbial community ecologists, as the community aggregate trait (CAT) can be directly captured through high-throughput methods based on genetic abundances [ 5 , 9 ]. This marks an advantage that facilitates the comparison of disparate communities and the formation of universal ecological hypotheses, from individual-based analysis to CATs analysis [ 10 , 11 ]. These findings present valuable methodology and hypotheses for identifying core features that soil microbes may take to adapt to antibiotic pressure. Firstly, resistance to antibiotics is not equated with stress tolerance, although the ARG expression serves as a bacterial defense against antibiotic stress. ARGs predominantly encode specialized functions against antibiotics, featuring mechanisms like efflux pumps, antibiotic-inactive enzymes, and protection of drug-binding targets, while stress tolerance traits cover a broader range of defense mechanisms, including the synthesis of the molecular chaperone [ 12 ] and sigma factors [ 13 ], enhancement of surveillance and damage repairing. Moreover, a nuanced interplay between specific antibiotic resistance and universal stress tolerance strategies is illustrated by the overlap in mechanisms such as reducing permeability, forming resting structures [ 11 ], stress sensing, signaling transduction, and the expression of global transcription regulators [ 14 – 16 ]. Second, the rising abundance of ARGs is closely tied to the growth yield in terms of bacterial proliferation. Due to this, it is considered that capacities for central carbohydrate metabolism, synthesis of the ribosome, and other precursor biomolecules (e.g., amino acids, fatty acids, nucleotide) are essential for the proliferation of ARGs. Additionally, ribosomal RNA operon copy number ( rrn ) in bacteria is also reported as a principal trait capturing the rate-yield trade-off [ 17 ]. Third, ARGs are essentially DNA fragments whose synthesis depends on the acquisition and conversion of deoxyribose, purines, pyrimidines, and phosphate. Thus, the degradation of complex substrates and the uptake of precursor molecules is crucial for the building of ARGs. Additionally, foraging traits including motility and chemotaxis processes also play an important role in perceiving and utilizing resources, especially in environments with great heterogeneity in resource accessibility [ 18 , 19 ]. Fourth, as a complement to the Y-A-S frame, the energy-producing traits were emphasized [ 9 , 20 ]. Indeed, the production and allocation of energy are crucial for the proliferation of ARGs, as it is recognized that bearing ARGs requires additional reproductive costs [ 21 – 23 ]. Since antibiotic resistance is deeply intertwined with a range of physiochemical activities in microbes, it cannot be fully captured by a single aspect of features, the core microbial traits that promote ARGs growth were still poorly understood. To this end, the main objective of this study is to reveal the taxonomic and functional-genetic features that support the growth of ARGs, especially in the case of nutritional scarcity. To achieve this objective, changes in complex microbial communities and their resulting impact on ARGs were investigated along a gradient of oxytetracycline (OTC), ranging from no antibiotic presence to a clinically relevant concentration. Subsequently, the key ARGs subtypes that drive the growth of ARGs will be identified and correlated with taxonomic features at the community level and metagenomic assembly genome level. Ultimately, applying CATs analysis to reveal genetic features under conditions of ARGs growth, if present. This study contributes to a broader application of trait-based assessments to understand the causes of antibiotic resistance, leading to a more coherent theory to block the propagation of ARGs in the environmental microbiota.",
"discussion": "Discussion Collectively, the main results obtained from this study are summarized in Table 2 , providing a comprehensive view.\n Table 2 Summary of all observations in this study CK 0.1 mg/l 0.5 mg/l 1 mg/l 5 mg/l 10 mg/l Effect on bacterial growth Control Sub-inhibition Promotion Inhibition Heavy inhibition Rel. ARG growth (normalized to CK) Control + + + − − − Abs. ARGs growth (normalized to CK) Control + + + − − − Composition similarity G1 G2 G3 Representative bacteria families Bacillaceae Micrococcaceae Rhizobiaceae Pseudomonadaceae Burkholderiaceae Microbacteriaceae Streptomycetaceae CSR/CSO lifestyle [ 11 ] ruderal/ opportunist competitor stress-tolerant YAS traits classification [ 10 ] Y&A Y&A&S S LHS behind ARG growth Promoted by the opportunists in antibiotic-free Growth of intrinsic ARGs Minor impacts on OTC resistance Further promoted by the competitor Enhanced energy production and streamlined bypass metabolism Inducing OTC resistance Stress tolerators dominating Slow proliferation with high inputs in defense mechanisms Significantly inhibited the growth of bacteria and ARGs Our findings illustrate the amplification of ARGs by microbial communities is a nuanced response to the antibiotic pressure, promoted by both the relative enrichment of ARGs and bacterial growth. At this point, both the relative and absolute abundances of ARGs display a non-monotonic growth trend across the antibiotic gradient; however, due to variations in bacterial quantities, their peak values manifest at different concentrations (Fig. 1 ). Unlike the clinical focus, which primarily concentrates on the antibiotic resistance capabilities of specific pathogens [ 52 , 53 ], the environmental viewpoint, guided by the One Health approach, emphasizes understanding the overall presence of ARGs within a region and the associated risks [ 4 , 54 ]. To this end, concentrating exclusively on the changes in relative enrichment would result in biased perceptions of the true situation. Quantification through the use of tagged gene spiking is acknowledged as a more accurate approach, though it is more demanding [ 55 ]. The method of weighting by 16S rRNA gene copies to characterize changes in the absolute abundance of ARGs is also deemed appropriate in this study, given that DNA recoveries from samples were expected to be essentially identical. It is also important to clarify that the abundance of ARGs does not equate to antibiotic resistance capacity. The mere presence of ARGs does not necessarily indicate active expression. Intrinsic functional ARGs, chromosomally carried by the host microbes may remain inactivated or unexpressed due to insufficient external stimuli, which contrasts with the expression efficiency of acquired ARGs carried on plasmids or integrons [ 4 ]. To serve different research purposes, various ARG databases have been developed, such as ResFinder, which primarily focuses on acquired resistance [ 46 ], ResFinderFG for functional ARGs [ 45 ], and the CARD database combines both [ 56 ], however, no single database is suitable for all purposes. In this study, we employed the SARG database, which encompasses both intrinsic and acquired resistance genes, as its automated pipeline for short-read analysis is more aligned with the need for absolute quantification of ARG abundance. With the absolute quantification of ARGs, an interesting phenomenon was observed under intermediate antibiotic stress (i.e., 0.5 mg/l), there was a simultaneous increase in bacterial quantity and the relative abundance of ARGs compared with CK (Table 2 ). The explanation for a similar phenomenon is that low dose antibiotics induce a hormesis effect, meaning bacteria overcompensate for the inhibition of antibiotics [ 57 ]. In addition, antibiotics at sub-lethal levels are known for accelerating the emergence and spread of antibiotic-resistant bacteria, which act as signaling molecules in promoting biofilm formation and bacterial virulence [ 3 ]. These perspectives have been preliminarily validated using bacterial populations and simplified microbial communities [ 58 , 59 ]. However, their applicability to complex microbial communities remains uncertain, as fluctuations in the abundance of ARGs may result from taxonomic shifts rather than subinhibitory antibiotic pressure [ 4 ]. By synthesizing findings from representative bacteria and CATs classification (Table 2 ), it could be proposed that this further increase in both bacteria and ARGs is the consequence of shifts in the LHS driven by microbial communities. The simultaneous growth is not due to opportunist or passive stress tolerance; instead, it signifies a proactive succession of microbes to antibiotic selection pressure. From the perspective of representative bacteria, the transition from Bacillaceae in G1 to Pseudomonadaceae in G2; indicates shifts from the ruderal to a competitive lifestyle [ 60 ]. Generally, bacteria from Bacillaceae were considered to have small genome size, high maximum growth rate, and great ability to germinate from spores within a short period [ 61 ], ensuring its opportunistic growth under antibiotic-free conditions; in contrast, Pseudomonadaceae is featured by their high catabolic diversity, siderophore production and capacity to form biofilms, enhancing their survival competitiveness in stressed environments. The Pseudomonadaceae family is also known for its substantial intrinsic resistance to tetracyclines and toxins [ 14 , 62 ]. This resistance is primarily mediated by the expression of a diverse array of functional ARGs that encode multidrug efflux pumps, which actively expel antibiotics from the bacterial cell [ 51 ], including mexAB - oprM and mexEF , as detected on MAG1. These characteristics are consistent with our findings obtained from MAGs (Fig. 3 ; Tables S11, S13–14, S16). Additionally, enrichment of genes in the oxidative phosphorylation pathway of MAG1 (belonging Pseudomonadaceae ) indicates an advantage in energy production trait (Fig. 3 C). From the CATs analysis, the end with higher quantities of bacteria and ARGs demonstrates stronger substrate acquisition capabilities and rapid proliferation traits; build on this, enhanced energy production and substrate utilization efficiency emerge as primary genetic traits that support the further increase in the abundance of ARGs, which is reflected by the strengthening of central carbon metabolism and the streamlining of low competitiveness function genes. Moreover, key ARGs encoding tetracycline resistance (Table 1 ), linking with the representative bacteria in G2, become prominent in the positive correlation networks (Fig. 2 C), which demonstrates the active adaptation of the community to increased antibiotic selection pressure. In this study, a genetic trait-based assessment approach was utilized to reveal altered LHSs under various antibiotic selection pressures. It further extends the application of this approach in the development of universal ecological theories. We mainly followed the Y-A-S theoretical framework [ 10 ], while taking into account the effects of energy production [ 20 ], motility, and chemotaxis on opportunists [ 11 ], and using MCOA as an integrating statistic analytical tool to describe aggregated genetic traits at the community level [ 5 ]. Although the development of these theories and methods was not specifically aimed at antibiotic resistance, they greatly enhance our understanding of how microbial communities balance proliferation with responses to environmental stresses. To obtain a more coherent theoretical framework for the antibiotic resistance issue, the focus could be placed on traits in terms of accessibility and mobility of ARGs [ 4 ], and the distinction between resistance and tolerance [ 63 ], in subsequent studies. Which could be complemented by the use of a genome-informed trait-based dynamic energy budget model [ 9 ]. Further, the limits of the experimental design of this study should be stated here, which focuses more on the effects of antibiotics in resource-limited conditions to mimic the typical nature. However, under conditions influenced by human activities the dominant LHSs may vary with the substantial increase in available resources [ 64 , 65 ]. In comparison with a previous study with the same soil source and experimental conditions, the maximum abundance of ARGs was observed at a higher concentration of 10 mg/l OTC with extra LB medium addition (even only 2%) [ 25 ]. This is mainly due to the increased abundance of specific efflux pumps for tetracyclines ( tet39 ) and inactivated enzymes such as tetX . However, the extra fitness cost of carrying these two ARGs might be unsustainable for microbial communities under resource-scarce conditions. Moreover, the relatively short experimental duration imposes some limitations on the generalizability of the conclusions. While differences in microbial community composition and ARG abundance across OTC treatments were observed within 24 h, it remains uncertain whether these differences would intensify or diminish with extended reaction times. This uncertainty is particularly relevant as the selective pressure exerted by OTC may gradually decrease over time due to its slow degradation. Additionally, exploring multiple classes of antibiotics represents a crucial avenue for future research, despite the added complexity introduced by potential antagonistic or synergistic interactions between antibiotics. Notably, investigating the combined effects of tetracycline and sulfonamide antibiotics could yield meaningful insights, especially when aiming to replicate conditions commonly found in farming environments [ 66 ]. The scarcity of resources prompted another consideration in the process of controlling the spread of ARGs. That is blocking the propagation of ARGs based on the principle of competitive exclusion [ 67 ], in particular, artificially enhancing competition between alternative less-ARG populations and more-ARG populations under limited resource conditions. At the population level, it was well predicted based on the coexistence theory that a fluctuating resource environment can effectively affect the competitiveness of resistant bacterial strains [ 23 ]. Gut symbionts protect against pathogens through nutrient blocking is another experimental evidence with a more complex microbial context [ 68 ]. Returning to this study, the replacement of Bacillaceae by Pseudomonadaceae with increasing concentrations of antibiotics and ARGs caught our attention. Species from these two families both ubiquitously occur in natural environments and are often co-isolated from soil samples [ 60 , 61 , 69 ], indicating comparable adaptive capacity. Indeed, interactions between members of these two families range from mutualism to competition depending on cultural conditions, however, were mainly reported as either competition or amensalism [ 61 ]. Most of the Bacillaceae members have no pathogenic potential and are relatively sensitive to antibiotics [ 70 , 71 ], and Bacillus subtilis , in particular, is commonly used as an intestinal probiotic and as a biocontrol bacterium for the protection of human, animal, and plant health [ 72 – 74 ]. Thus, the use of Bacillus subtilis spore preparations to deplete limited growth resources might be a viable method of controlling ARGs spread under sub-inhibitory antibiotic stress. Meanwhile, it would be an attempt to use the traits-based life history theory to guide microbial practice."
} | 4,930 |
31718625 | PMC6849255 | pmc | 8,806 | {
"abstract": "Background Corynebacterium ammoniagenes is an important industrial organism that is widely used to produce nucleotides and the potential for industrial production of coenzyme A by C. ammoniagenes ATCC 6871 has been shown. However, the yield of coenzyme A needs to be improved, and the available constitutive promoters are rather limited in this strain. Results In this study, 20 putative DNA promoters derived from genes with high transcription levels and 6 promoters from molecular chaperone genes were identified. To evaluate the activity of each promoter, red fluorescence protein (RFP) was used as a reporter. We successfully isolated a range of promoters with different activity levels, and among these a fragment derived from the upstream sequence of the 50S ribosomal protein L21 (P rpl21 ) exhibited the strongest activity among the 26 identified promoters. Furthermore, type III pantothenate kinase from Pseudomonas putida ( Pp coaA) was overexpressed in C. ammoniagenes under the control of P rpl21 , CoA yield increased approximately 4.4 times. Conclusions This study provides a paradigm for rational isolation of promoters with different activities and their application in metabolic engineering. These promoters will enrich the available promoter toolkit for C. ammoniagenes and should be valuable in current platforms for metabolic engineering and synthetic biology for the optimization of pathways to extend the product spectrum or improve the productivity in C. ammoniagenes ATCC 6871 for industrial applications.",
"conclusion": "Conclusions In summary, this study provides a rational strategy to isolate endogenous promoters from C. ammoniagenes ATCC 6871, which may be helpful in other similar scientific research. Through this strategy, we successfully isolated a range of promoters with different transcriptional activities and the strongest one was applied to improve the CoA production in C. ammoniagenes ATCC 6871, raising hope for further improving the industrial production level of CoA.",
"discussion": "Discussion C. ammoniagenes ATCC 6871 is a producer of CoA, however, its genetic operation has been rarely reported, and its industrial application potential still needs to be improved. In order to further increase the production of CoA, endogenous promoters from C. ammoniagenes ATCC 6871 were screened in this study. Firstly, 20 putative promoters based on genes with high transcription levels and 6 promoters upstream of molecular chaperones were identified and characterized for transcriptional activity by using RFP as the reporter gene. Among the 26 putative promoters, 20 with different transcriptional activities were isolated including 15 from the high transcription level genes and 5 from molecular chaperone genes. The output efficiency demonstrates that isolating promoters from molecular chaperones (5/6) might be a more efficient strategy than using RNA-seq (15/20). Therefore, isolating promoters from molecular chaperone genes is more effortless and cost-effective than RNA-seq. These results may ascribe to that the RPKM values of the genes only represent the abundance of RNA and do not equivalent to its promoter activities, which might be caused by multiple factors, such as the codon usage of genes, copy numbers of genes, the half-lives of mRNA and limitations inherent in transcriptomics [ 22 – 24 ]. It thus, isolating promoters based on RNA-Seq requires further verification by experiments. Nevertheless, we should note that the red fluorescence intensity of the highest promoter (P rpl21 ) identified from RNA-seq was almost 2.3 times than the highest promoter (P dnaK ) derived from molecular chaperone genes, so isolating promoters based on the genes transcriptional data may be more likely to yield promoters with highest activity. To our best knowledge, P rpl21 from the upstream sequence of the 50S ribosomal protein L21 is the strongest C. ammoniagenes ATCC 6871 promoter reported so far. The promoters obtained in this work will enrich the available promoter toolkit for C. ammoniagenes and should be valuable in current platforms for metabolic engineering and synthetic biology for the optimization of pathways to extend the product spectrum or improve the productivity in C. ammoniagenes ATCC 6871. To further verify the capacity of P rpl21 , the key gene Pp coaA in the biosynthetic pathway of CoA was overexpressed in C. ammoniagenes ATCC 6871 with the aim of reducing the feedback inhibition and increasing the production of CoA. The CoA production of manipulated C. ammoniagenes was approximate 4.4 times to the control, which confirmed that Pp coaA overexpressed and functionalized successfully in C. ammoniagene . These results prove that the selected promoter for the overexpression of foreign genes in C. ammoniagenes could be used as an efficient tool for improving the yield of major products in C. ammoniagenes . However, eliminating the burden of co-expressing RFP only slightly increase the CoA production, possibly due to the undiscovered limitations exist in the CoA synthetic pathway. Thus, more efforts should be focus on the resolving of rate-limiting steps of CoA synthesis in C. ammoniagenes ."
} | 1,294 |
39516204 | PMC11549454 | pmc | 8,807 | {
"abstract": "Evidence from homogeneous liquid or flat-plate cultures indicates that biochemical cues are the primary modes of bacterial interaction with their microenvironment. However, these systems fail to capture the effect of physical confinement on bacteria in their natural habitats. Bacterial niches like the pores of soil, mucus, and infected tissues are disordered microenvironments with material properties defined by their internal pore sizes and shear moduli. Here, we use three-dimensional matrices that match the viscoelastic properties of gut mucus to test how altering the physical properties of their microenvironment influences the growth of bacteria under confinement. We find that low aspect ratio (spherical) bacteria form compact, spherical colonies under confinement while high aspect ratio (rod-shaped) bacteria push their progenies further outwards to create elongated colonies with a higher surface area, enabling increased access to nutrients. As a result, the population growth of high aspect ratio bacteria is, under the tested conditions, more robust to increased physical confinement compared to that of low aspect ratio bacteria. Thus, our experimental evidence supports that environmental physical constraints can play a selective role in bacterial growth based on cell shape.",
"introduction": "Introduction Most natural habitats host diverse bacterial communities, which actively respond to extrinsic environmental cues that reflect the dynamic properties of their microenvironments 1 , 2 . Identifying the basis of these feedback-response cycles is critical for describing the evolution of natural microbial populations, characterizing host-symbiont crosstalk and interdependency, as well as understanding pathogenesis to design effective antibiotics. A major class of such interactions is understood within the framework of biochemical signalling, where small molecules (such as metabolites, growth factors, antibiotics, and chemoattractants) drive cellular perception and consequent responses 3 – 6 . These processes also directly impact the composition of natural microbial communities, wherein co-existing species exhibit symbiotic, mutualistic, or predatory interactions with each other 7 , 8 . However, current experimental approaches designed to explore such interactions rely on bacteria cultured in homogeneous liquid broths or on surfaces of 2D flat plates, which do not capture the structural complexity of many natural niches. Bacteria often reside in complex and disordered 3D microenvironments like soil, inter-tissue pores, and biological hydrogels such as mucus, which feature a wide range of material properties 9 – 11 . For example, the porosity and stiffness of soil is affected by its moisture content and granularity 12 . The spatial architecture of tissues also undergoes significant structural alterations as a result of ECM (extracellular matrix) deposition and enzymatic remodelling by fibroblasts and immune cells 13 . Similarly, the viscoelastic properties of mucus vary with diet, inflammatory responses, microbial enzymatic activity, and pathologies like cystic fibrosis 14 – 16 (Fig. 1a ). Furthermore, bacterial colonies physically constrained within these complex 3D matrices experience diffusion-limited growth 17 , 18 . In contrast, bacteria growing in homogeneously well-mixed liquid cultures have unrestricted access to nutrients, while colonies on a flat agar surface receive oxygen supply from the top, as well as direct nutrient access from the underlying substratum. Although diffusion-limited zones can be established during growth over time, these usually affect the colony’s core once multilayer 3D growth occurs, with cells piling on top of each other 19 . However, growth under 3D confinement could introduce such consumption-diffusion limited zones locally around small single cell-derived colonies, despite there being a global abundance of the nutrients in the bulk medium. Hence, growth under 3D confinement becomes fundamentally different from that on standard 2D interfaces. Previous work has investigated colony growth under nutrient-replete 2D conditions, extending growth-generated stresses as a putative mechanism for achieving quasi-3D conformations 20 , 21 . However, these studies do not fully capture bacterial growth in disordered, confined microenvironments. Growing evidence indicates that bacterial motility is significantly altered by confinement within a granular and porous matrix 22 and the growth of initially smooth, densely-packed bacterial populations under 3D confinement shows generic surface roughening morphodynamics driven by differential nutrient accessibility 18 . Despite this, bacterial population level growth and colony organization under complex and disordered confinement remains insufficiently described. Consequently, whether and how the physical constraints imposed by a 3D growth microenvironment influence microbial populations remains unclear. Fig. 1 An in vitro 3D growth matrix with tunable viscoelastic properties. a The physical properties of the mucosal layer are diverse and are altered by factors such as a change in diet, infection, inflammation, and enzymatic degradation of fibres by microbes. b Different gut-derived microbial strains from red flour beetles, were imaged in brightfield (pseudo-coloured for enhanced contrast). Schematic of the red flour beetle is created in BioRender (M. S. (2024) BioRender.com/y37r627). c By packing highly swollen polymeric hydrogel granules beyond jamming concentrations, we design an in vitro 3D growth medium, which provides a granular and internally porous microenvironment for bacterial growth under confinement. d – f The physical properties of the microgel growth medium are highly tunable based on the mass percentage of hydrogel granules relative to liquid LB (wt%/volume). Here, we show both soft and low confinement (0.50%), as well as stiff and high confinement (0.85%) matrices, the viscoelastic properties of which approximately match mucosal samples from natural sources 24 . Rheological measurements shown here either apply d a unidirectional shear at different rates to measure the viscosity, or, e a small amplitude (1%) of oscillatory strain at different frequencies to measure the shear moduli of the 3D growth media. The storage modulus ( G’ ) is a measure of the elastic solid-like nature, while the loss modulus ( G” ) signifies the viscous properties of the material. f The porosity of the medium, shown here as a complementary cumulative distribution function (1-CDF) of all the inter-particle pore spaces, determines the degree of confinement, and can be tuned by altering the packing fraction of the hydrogel granules. In this work, we present experimental evidence for confinement-dependent growth dynamics that selectively favour specific bacterial morphologies. We prepare transparent, granular 3D growth media—a special type of porous matrix for direct measurement of bacterial growth, as well as visualization of colony organisation under different biomimetic degrees of confinement—that broadly mimic the structural and viscoelastic properties of mucus, a biological hydrogel. Using several different bacterial strains isolated from the gut mucus of red flour beetles we show how an increase in confinement confers a growth advantage to bacteria with a higher aspect ratio. Our study combines quantitative measurements of bacterial growth and colony morphology, agent-based modelling, numerical calculations, as well as in vitro co-culture experiments, to establish a generalized principle for growth under confinement. By experimentally manipulating cellular morphology, we also demonstrate a remarkable interconvertible behaviour between high and low aspect ratio forms of the same bacterial species, which strongly implicates single-cell morphology as a broad determinant of colony architecture and growth success under physical confinement. Importantly, we show that population-level variation in growth dynamics can arise without invoking mechanisms such as genetic mutations, behavioural differences, or cellular responses to biochemical cues. Rather, our work suggests that the composition of heterogeneous microbial populations may be strongly dependent on efficient colony organization under increased confinement that selectively favours bacterial species with a high aspect ratio morphology. These principles are valuable for the experimental elucidation and theoretical modelling of microbial dynamics in complex, spatiotemporally varying natural niches.",
"discussion": "Discussion Our work provides experimental evidence implying that physical confinement can play a selective role in deciding bacterial growth fitness within their natural niches. High aspect ratio bacteria leverage growth anisotropy to form elongated colonies with higher surface area, allowing better nutrient accessibility as opposed to the spherical colonies formed by low aspect ratio bacteria. Increased confinement strongly selects high aspect ratio cells for more efficient growth—a trend that holds even for high and low aspect ratio forms of the same bacterial strain, indicating a remarkable interconvertible behaviour governed solely by cellular shape. Importantly, we demonstrate the potential to use such constraints to alter bacterial community composition under biomimetic mechanical regimes. Thus, our results represent an important conceptual advancement towards experimentally modelling bacterial growth within complex natural habitats. In microbial ecology, the present mechanistic understanding of predator-prey interactions, niche partitioning, and community organization largely relies on a framework comprising genetic mutations, behavioural patterns, and biochemical signalling 45 – 49 . Notably, our platform does not preclude such chemical interactions. Instead, the porous and permeable nature of our 3D matrix presents an opportunity to capture the dynamics of a chemical signalling landscape coupled with different mechanical regimes, enabling a more comprehensive experimental recapitulation of natural niches. While the permeable nature of our 3D matrix allows unimpeded small molecule diffusion, several biological niches such as soil and aquifers resemble semi-permeable porous media with non-deformable, rigid physical barriers. We expect that the reduced nutrient availability and severe spatial restrictions in such contexts significantly restrict colony growth and organization. Recreating such regimes in vitro will require the development of semi-permeable 3D scaffolds. Further, within the temporal window of our assays, we do not observe significant contributions from either aberrant phenotypes or perturbed physiologies. However, environmental pressures in the form of nutrient deprivation, high mechanical stiffness, and antagonistic interspecies interactions are known to induce the expression of stress genes, alter metabolism, trigger extreme phenotypes (such as filamentous growth), and favour increased mutational rate 50 – 55 . Such factors are likely to affect long-term bacterial growth— hence, future work that investigates bacteria under similar conditions over extended durations could provide interesting insights into the long-term effects of confinement on bacterial communities. Presumably, our system can be engineered to recreate such effects by altering the nutrient media composition and tuning the mechanical stiffness of the matrix. We speculate that such environmental constraints will selectively favour a subset of adaptive mutations with the potential to alter evolutionary trajectories. Furthermore, it has been reported that spatial patterning within mixed microbial communities can be strongly driven by single-cell morphology 56 . This suggests an interesting direction for future exploration. Given the selective pressure enforced by elevated physical confinement favouring rod-shaped bacteria over spherical ones, how do existing spatial patterns get altered when the microenvironmental mechanics change? Conversely, does physical confinement alter the evolution of spatial architecture within mixed microbial communities such as biofilms? Finally, our work highlights the need to explore the role of physical confinement in diverse niches that may be selected for varied life history strategies. For instance, short-term rapid growth and expansion are not universally preferred traits. In adverse conditions such as in the presence of toxins or predators when long-term cell survival is prioritized, a transient shift towards quiescence or dense collective organisations with minimally exposed surface areas may be more advantageous 57 – 61 . Such conditions and strategies may override the selection of cell shape imposed by physical confinement and should be explored further. Current descriptions of microbial growth invoke either genetic mutations or biochemical signalling as the primary mechanisms via which bacteria respond to their microenvironment. By contrast, we argue that the spatial constraints imposed by their microenvironment exert a more general effect, irrespective of specific organismal biology. Our conceptual framework can be employed to generate coarse-grained predictions of population dynamics under different biophysical regimes. Understanding how bacteria grow in complex, disordered 3D environments such as tissues, mucus, and soil, is also central for combating antibiotic resistance, improving agricultural practices, and bioremediation—underscoring the high-value practical applications of our work."
} | 3,390 |
30995811 | PMC6631277 | pmc | 8,809 | {
"abstract": "Haloarchaea, the extremely halophilic branch of the Archaea domain, encompass a steadily increasing number of genera and associated species which accumulate polyhydroxyalkanoate biopolyesters in their cytoplasm. Such ancient organisms, which thrive in highly challenging, often hostile habitats characterized by salinities between 100 and 300 g/L NaCl, have the potential to outperform established polyhydroxyalkanoate production strains. As detailed in the review, this optimization presents due to multifarious reasons, including: cultivation setups at extreme salinities can be performed at minimized sterility precautions by excluding the growth of microbial contaminants; the high inner-osmotic pressure in haloarchaea cells facilitates the recovery of intracellular biopolyester granules by cell disintegration in hypo-osmotic media; many haloarchaea utilize carbon-rich waste streams as main substrates for growth and polyhydroxyalkanoate biosynthesis, which allows coupling polyhydroxyalkanoate production with bio-economic waste management; finally, in many cases, haloarchaea are reported to produce copolyesters from structurally unrelated inexpensive substrates, and polyhydroxyalkanoate biosynthesis often occurs in parallel to the production of additional marketable bio-products like pigments or polysaccharides. This review summarizes the current knowledge about polyhydroxyalkanoate production by diverse haloarchaea; this covers the detection of new haloarchaea producing polyhydroxyalkanoates, understanding the genetic and enzymatic particularities of such organisms, kinetic aspects, material characterization, upscaling and techno-economic and life cycle assessment.",
"conclusion": "8. Conclusions As detailed in the present review, a two-digit number of different haloarchaeal species were already described as potential PHA producers. However, most of these studies were restricted to modest cultivation scales, often merely reporting on microscopic observation and fluorescence staining of PHA granules. To the best of the author´s knowledge gained from the open literature and discussions with other scientists active in this field, there are not more than four haloarchaeal species ( Hfx. mediterranei , Hpg. aswanesnsis , Hgn. amylolyticum , and Htg. hispanica ), for which PHA accumulation was studied in cultivations performed under controlled conditions in bioreactors. However, such bioreactor cultivation setups are the conditio sine qua non to get reliable kinetic data, and reasonable amounts of product for in-depth characterization. Most of all, sufficient amounts of product are needed for processing it to marketable prototype specimens; such processing is completely lacking in the case of haloarchael PHA. Moreover, techno-economic assessment of PHA production by haloarchaea, based on solid experimental data and holistic consideration of the entire production cycle, is only available for Hfx. mediterranei , for which economic and life cycle considerations were carried out based on the surplus substrates whey and waste stillage. Nevertheless, exactly these early techno-economic assessments already indicate the high potential of the extremely halophilic members of the Archaea domain for bio-economic biopolyester production of the future. Taking advantage of the broad substrate spectrum, the formation of PHA heteropolyesters of tunable composition and microstructure in dependence on the cultivation strategy, the accessibility of haloarchaea towards inexpensive and convenient product recovery from biomass, the recyclability of process side-streams (spent fermentation broth and cell debris), the detailed knowledge about the complete genome of an increasing number of haloarchaea, and the expedient robustness of such cultivation batches sets haloarchaea at the forefront of efforts dedicated to finally make PHA economically competitive polymers with plastic-like properties, which also match the end-consumer´s expectations. What is needed now is upscaling those processes at a promising lab-scale, and to tap the wealth of haloarchaea reported to produce PHA merely on a qualitative basis, or which have not yet been studied for PHA biosynthesis. In addition, one should be aware of parallel R&D activities with halophilic eubacteria as PHA production strains; here, especially the seminal works with Halomonas bluephagenensis TD01 should be mentioned, a proficient PHA production strain which can be cultivated in open bioreactor facilities [ 96 ], and which is well studied in terms of genetic manipulation [ 97 , 98 ]. Other examples for promising halophilic eubacteria as PHA producers encompass Halomonas halophila [ 99 ], or Halomonas campaniensis [ 100 ]. However, these organisms thrive best under salinities of about 60–70 g/L, which is drastically below the optimum salinity of haloarchaea, which makes the long-term stability of fermentation batches with Halomonas sp. uncertain compared with their “competitors” from the realm of haloarchaea. To summarize the aforementioned, Table 1 and Table 2 provide an overview of the PHA production processes by the individual haloarchaea discussed in the review, indicating the productivities, type of biopolyester produced, and studied production scale. While Table 1 collects the setups on smaller scale, Table 2 refers to the rather scarce number of setups carried out under controlled conditions in laboratory and pilot scale bioreactors.",
"introduction": "1. Introduction The first description of a biological polymer with plastic-like properties was published in the 1920s, when Maurice Lemoigne detected light-refractive intracellular inclusion bodies [ 1 ], today referred to as “granules”—or, more recently “carbonosomes” [ 2 ]—n resting cultures of the Gram-positive bacterium Bacillus megaterium . Based on the acidic degradation product of these inclusions, 3-hydroxybutyrate (3HB), Lemoigne correctly assumed the microscopically observed intracellular product to be the polymer of 3HB, namely poly(3-hydroxybutyrate) (PHB). In the meantime, PHB and its related homo- and heteropolyesters, as a group labelled as polyhydroxyalkanoates (PHA), have attracted global attention as biological, bio-based, biocompatible and biodegradable alternatives to established plastics of petrochemical origin in many sectors of the rocketing plastic market [ 3 , 4 ]. PHA consist of a variety of diverse building blocks, which make their material properties highly versatile [ 5 ], and can be produced biotechnologically by different continuous or discontinuous fermentation approaches and feeding strategies [ 6 ]. In principle, short chain length PHA ( scl -PHA) are distinguished from medium chain length PHA ( mcl -PHA). While scl -PHA typically constitute thermoplastic materials, mcl -PHA are known as materials with elastomeric and latex-like properties and are often of a sticky nature. Among scl -PHA, the homopolyester PHB and the copolyester poly(3-hydroxybutyrate- co -3-hydroxyvalerate) (PHBHV) are best described; in this context, increasing 3-hydroxyvalerate (3HV) fractions in the copolyester decreases melting temperature and crystallinity, which makes such PHBHV copolyesters easier to process than the rather crystalline and brittle PHB, a material of restricted applicability [ 3 , 4 ]. Apart from wild-type and genetically engineered eubacteria and recombinant yeasts, plants, and microalgae, PHA biosynthesis takes place also in the cytoplasm of various extremely halophilic species from the Archaea domain, the so called “haloarchaea”. Exclusively scl -PHA production is reported for haloarchaea, while for eubacteria, both scl - and mcl -PHA production is reported [ 3 , 4 ]. Extremely challenging habitats include environments where such highly adaptive survivalists are typically isolated; illustrative examples are the Great Salt Lake, the Dead Sea, hypersaline anoxic deep-sea basins, solar saltern crystallizers, hypersaline soil samples, salt mine boreholes, salt production pans, or even alpine dry salt rocks. The taxonomic classification of these extremely salt-demanding, typically aerobic organisms is by no means a trivial task and is based on steadily refined knowledge about the genomics, proteomics, metabolomics, and lipidomics of these organisms. Traditionally, haloarchaea are members of the family Halobacteriaceae, which belongs to the order Halobacteriales, which in turn is part of class III ( Halobacteria ) consisting of two major clades A and B, of the phylum and (sub)kingdom of Euryarchaeota, which belongs to the domain of Archaea (according to International Committee on Systematics of Prokaryotes, Subcommittee on the taxonomy of Halobacteriaceae; cited by [ 7 ]). Later, members of the class Halobacteria were re-grouped into three orders: a revised order Halobacteriales and two new orders, Haloferacales and Natrialbales, which encompass the novel families Haloferacaceae and Natrialbaceae [ 8 ]. More recently, based on phylogenetic analyses and conserved molecular characteristics, it was suggested to divide the order “Halobacteriales” into the families Halobacteriaceae, Haloarculaceae, and Halococcaceae, and the order “Haloferacales” into the families Haloferacaceae and Halorubraceae [ 9 ]. These are the currently valid designations of the families where haloarchaea demonstrated to produce PHA are grouped. Figure 1 provides a schematic overview about the phylogenetic classification of the haloarchaeal species discussed in the present review. Talking about PHA biosynthesis by haloarchaea, it took the scientific community until 1972, when Kirk and Ginzburg carried out morphological characterizations of a Dead Sea isolate, which was labeled “ Halomonas sp.” by these authors. This organism was cultivated on a highly saline medium containing 200 g/L NaCl. By using freeze-fracture and freeze-etch techniques, the authors revealed plastic-like cytoplasmic inclusion bodies, which were extracted from microbial biomass and investigated by X-ray diffractometry. Grounded solely on these examinations, the authors correctly recognized this material as the biopolyester PHB, the material already known at the time as a carbon and energy storage product for many eubacteria, as reported by Lemoigne [ 1 ] and succeeding generations of scholars. In any case, this study by Kirk and Ginzburg was the very first unambiguous description of PHA production by an archaeon [ 10 ]. Regarding the production strain “ Halomonas sp.”, it took nearly three decades until this isolate was classified as Haloarcula ( Har. ) marismortui , its currently valid species name, in a report published by Nicolaus et al. [ 11 ]."
} | 2,668 |
37468308 | PMC10629481 | pmc | 8,814 | {
"abstract": "Comparative analysis of genome-scale metabolic networks (GSMNs) may yield important information on the biology, evolution, and adaptation of species. However, it is impeded by the high heterogeneity of the quality and completeness of structural and functional genome annotations, which may bias the results of such comparisons. To address this issue, we developed AuCoMe, a pipeline to automatically reconstruct homogeneous GSMNs from a heterogeneous set of annotated genomes without discarding available manual annotations. We tested AuCoMe with three data sets, one bacterial, one fungal, and one algal, and showed that it successfully reduces technical biases while capturing the metabolic specificities of each organism. Our results also point out shared and divergent metabolic traits among evolutionarily distant algae, underlining the potential of AuCoMe to accelerate the broad exploration of metabolic evolution across the tree of life.",
"discussion": "Discussion Numerous sequencing projects and available annotation approaches generate heterogeneously annotated data. There is currently a need to homogenize annotations to make them comparable for wider-scale studies. In this work, we introduced a method to automatically homogenize functional predictions across heterogeneously annotated genomes for large-scale metabolism comparisons between species across the tree of life. We illustrated how the tool can be applied both to prokaryotes and eukaryotes, even with high levels of annotation degradation. Accounting for existing annotations in the inference of homogenized GSMNs Automatic inference of single-species GSMNs is now routinely achieved, especially for prokaryotic species, and is often systematically performed for multiple genomes. With such data at hand, one may compare the predicted metabolism among related species from a given clade and subsequently identify metabolic specificities or putative functional interactions in microbial communities ( Frioux et al. 2018 ; Machado et al. 2018 ). Such applications require consistent genome quality and similar data treatment (genome annotation, metabolic network reconstruction) to minimize biases in predictions. However, ensuring the latter is complex for eukaryotic genomes, as their enzymatic functions are difficult to characterize automatically and as they often need expert annotation. Moreover, annotation efforts can greatly vary between genomes, resulting in heterogeneous annotation and metabolic prediction quality. As the automatization of both (meta)genome reconstruction and annotation is now routinely applied, it is likely that efforts toward manual annotation will decline. However, we believe the need to manually curate annotations will remain ( Karimi et al. 2021 ). In addition, AuCoMe could also be used to homogenize annotations in several genome versions of the same species or to reconcile several annotations performed on the same genome. We have shown above that the performance of AuCoMe is superior to or on par with other commonly used reconstruction pipelines, notably gapseq, ModelSEED, and CarveMe. The originality of our metabolic inference method resides in the possibility to account for, and preserve, available expert genome annotations. Not considering the genome annotations performed by specialists may lead to the omission of unique metabolic functions that are not well described in reference databases. On the other hand, comparing metabolic networks built from well-curated annotations to those built from poorly or automatically annotated genomes will result in biases. In such cases, real metabolic differences between species cannot be distinguished from missing annotations in some genomes. AuCoMe constitutes a solution to such challenges through the propagation of expert annotations to less-characterized genomes in the process of metabolic network reconstruction. By accounting for possibly missing functional, but also structural, annotations in the input genomes, the resulting metabolic networks are homogeneous and can therefore be directly compared in both prokaryotes and eukaryotes. Method limitations and improvements AuCoMe incorporates several strategies to optimize the method's selectivity and sensitivity. Together, these strategies collectively achieve comparable GSMN reconstruction with two objectives: having comparisons as homogeneous as possible given the initial heterogeneity and incompleteness of databases and, thus, identifying errors that can be corrected during further analysis. A first limitation is illustrated by the comparison of AuCoMe reconstructions to the EcoCyc database considered as ground truth in our experiment. We observed that the GSMN automatically reconstructed from the reference genome substantially differs from the database. Extensive and systematic manual curation has been performed on this database since its creation in 1998, and we hypothesize that these efforts have not been all translated in the E. coli K–12 MG1655 annotations. As a result, several reactions were systematically missing from the automatic inferences provided by AuCoMe. This example illustrates the role of curation in producing high-quality models. The homogenization of metabolic inference proposed by AuCoMe does not aim at replacing this step but rather enabling an unbiased metabolic comparison between species. Running AuCoMe on the bacterial data set highlighted the impact of a single highly annotated genome on metabolic inference. This data set included a single well-annotated reference genome of the E. coli K–12 MG1655 strain, which caused a number of reactions initially propagated by orthology from the E. coli K–12 MG1655 genome to others to be discarded by the AuCoMe filter. Reasoning on ortholog clusters, the filter implies that several congruent genome sources are mandatory to confidently achieve an annotation propagation. Although the relevance of the filter was shown on the algal data set by avoiding the propagation of annotations related to photosynthesis to nonphotosynthetic organisms, it may be too stringent in some applications. Several improvements of the filtering approach could be devised. For example, the structural annotation step could be improved: The annotation of pseudogenes in Shigella species would have been avoided by considering the annotations as pseudogenes available for the identified loci. More generally, in addition to the difficulties of automatically estimating protein homology, the link between orthology and conservation of function is still a matter of active investigation and methodological debate ( Stamboulian et al. 2020 ; Begum et al. 2021 ). Finally, we want to emphasize that our attempts to limit the inference of false-positive reactions also directed the choice of method for the initial draft metabolic inference. We used Pathway Tools because of its several advantages such as the capacity to work with eukaryotic genomes, the suitability for parallel computing ( Belcour et al. 2020a ), and the possibility to limit gap-filling of metabolic networks. However, metabolic pathway completion performed by Pathway Tools does not systematically extend to ensuring the production of biomass. Pathway Tools was therefore adapted to our objective of avoiding to go beyond the strict interpretation of genome annotations. This goal was fulfilled, as attested by the benchmark shown in Supplemental Figure S4 , which confirms that AuCoMe GSMNs have, by design, no reaction lacking gene association. A typical use for GSMNs is their simulation, generally with flux-based approaches. As AuCoMe performs a homogenization step on GSMNs but does not provide de novo annotation, using AuCoMe without further curation might lead to missing reactions in organisms. In addition, the complexity of eukaryotes and their strong dependency on their environment make it difficult to provide a flux-based simulation-ready gap-filled model that would minimize the risk of adding false positives. For further simulation studies, GSMNs built with AuCoMe therefore still need to be gap-filled and curated ( Karp et al. 2018b ; Latendresse and Karp 2018 ). However, regarding the reactions that are present in at least one GSMN reconstructed by AuCoMe, the tool ensures that their absence in other organisms is true. In that sense, AuCoMe reduces the need for curation. Biological insights from comparison of metabolic networks across species Evolution Our examples of the Calvin cycle and phycobiliprotein synthesis show that, once all steps of the AuCoMe pipeline have been executed, the predicted metabolic capacities of the analyzed genomes reflect the biological knowledge we have of the corresponding organisms. Our approach, therefore, enables GSMNs to be compared in the light of evolutionary biology. The metabolic dendrogram calculated from the final AuCoMe reconstruction is mostly consistent with reference-species phylogeny. Indeed, numerous studies have shown that comparing GSMNs by computing a metabolic distance and arranging them into a dendrogram allows clustering organisms into groups close to the ones known by phylogenetic analysis. However, the position of species inside these groups is often different from the one of the phylogenetic groups ( Vieira et al. 2011 ; Bauer et al. 2015 ; Prigent et al. 2018 ; Schulz and Almaas 2020 ). It furthermore gives support to the hypothesis of a metabolic clock based on the congruence between molecular and metabolomic divergence in phytoplankton ( Marcellin-Gros et al. 2020 ). The difference observed in the tanglegram ( Fig. 5 B) between phylogeny and metabolic distances could be further explored. One possibility could be to look at different similarity measures for the clustering. In this work, the Jaccard distance has been used but other measures could be used. For example, if we consider an absence of a reaction in two organisms as a similarity (to represent the loss of a function) then other measures could be envisaged such as the Simple Matching Coefficient. This also opens the perspective of inferring ancestral metabolic networks to better understand the dynamics of character evolution across time ( Psomopoulos et al. 2020 ). Adaptation The second aim of reconstructing comparable GSMNs is to determine to what extent metabolic changes are the result of or the prerequisite for adaptation. In our study, we made a first attempt at this question regarding the cryptophyte G. theta . This species has several potentially plesiomorphic metabolic traits in common with other marine lineages, which may constitute adaptations to their shared marine environment. Glycine betaine, for instance, is known to be an osmoregulator or osmoprotectant in green plants ( Di Martino et al. 2003 ), and carnosine has been proposed to function as an antioxidant in red algae ( Tamura et al. 1998 ). Regarding carnitine, its physiological significance in photosynthetic organisms is still largely unknown, but antioxidant and osmolyte properties along with signaling functions have also been suggested ( Jacques et al. 2018 ). However, for now, all of this remains purely hypothetical. To dig deeper into such questions in the future, we need to be able to distinguish changes that simply result from random processes, such as metabolic drift ( Belcour et al. 2020b ), from changes that have an adaptive value. Currently, we envision two approaches that will help with this distinction. The first approach will be to further increase the number of species and lineages included in order to identify adaptive patterns, for example, among organisms occupying similar ecological niches. In phylogenomics, wide taxon sampling is recognized as one of the key features for reliable comparisons ( Young and Gillung 2020 ), whereas pairwise genomic comparisons across species are generally viewed as problematic ( Dunn et al. 2018 ). Given that, as shown above, phylogenetic signals in metabolism are stronger than the adaptive signals we can expect, this approach would also benefit from the development or adaptation of statistical models that could help detect signals of adaptation in an overall noisy data set. Such models exist, for instance, to detect selective signatures in the evolution of the protein-coding gene ( Shapiro and Alm 2008 ) but, to our knowledge, have not been developed for metabolic networks or presence/absence signatures of genes. The second related strategy consists in focusing on phylogenetically closely related species that have only recently diverged and adapted to different environments. In such cases, we anticipate that the relative importance of drift along with the noise from the phylogenetic signal will be reduced owing to the short evolutionary time since the separation. With such data sets, we may be able to reduce the level of replication required to find biologically relevant metabolic adaptations. The range of questions that could be addressed with the appropriate data set is long and includes metabolic adaptations to different environments ( Xu et al. 2020 ), food sources and domestication ( Giannakou et al. 2020 ), multicellularity ( Cock et al. 2010 ), or even life-history transitions to endophytism ( Bernard et al. 2019 ). Interactions Lastly, we anticipate that AuCoMe will provide new opportunities to study metabolic interactions between symbiotic organisms. For example, the tentative o -aminophenol oxidase activities pointed out by AuCoMe in brown algae could be involved in the protection against pathogen attacks at the cell surface. Indeed, a molecular oxygen-scavenging function in the chloroplast ( Constabel et al. 1995 ) and a defense role ( Gandía-Herrero et al. 2005 ) have been suggested for these enzymes in terrestrial plants. An o -aminophenol oxidase Streptomyces griseus is known to be involved in the grixazone biosynthesis, that is, an antibiotic ( Suzuki et al. 2006 ). Similarly, brown algal o -aminophenol oxidases or tyrosinases might be involved in the production of specific antibiotics. The o -aminophenol oxidase enzymes resemble laccases or tyrosinases. They can be involved in catechol or pigment production by oxidation ( Le Roes-Hill et al. 2009 ). Numerous references have also shown that tyrosinases are efficiently inhibited by some phlorotannins, antioxidant compounds specific to the brown algae ( Kang et al. 2004 ; Manandhar et al. 2019 ), suggesting there might be a regulation of polyphenol oxidation in certain conditions. In the same vein, metabolic complementarity has previously been used to predict potentially beneficial metabolic interaction between a host and its associated microbiome ( Frioux et al. 2018 ) and to successfully predict metabolic traits of the communities ( Burgunter-Delamare et al. 2020 ). These studies have, so far, examined large numbers of symbionts (all sequenced and annotated with identical pipelines), but usually they consider one specific host whose metabolic network was manually curated. With AuCoMe, these previous efforts could be expanded to incorporate a range of different hosts with their associated microbiota, thus facilitating the identification of common patterns in host–symbiont metabolic complementarity as well as their differences in these complementarities across different species and lineages. Just as for the question of adaptation, we believe this new scale of comparisons enabled by tools such as AuCoMe will enable researchers to move from the study of specific examples to the identification of general trends, thus approaching the biologically most relevant evolutionary constraints."
} | 3,904 |
28659873 | PMC5468387 | pmc | 8,815 | {
"abstract": "Southwest Indian Ridge (SWIR) is a typical oceanic ultraslow spreading ridge with intensive hydrothermal activities. The microbial communities in hydrothermal fields including primary producers to support the entire ecosystem by utilizing geochemical energy generated from rock-seawater interactions. Here we have examined the microbial community structures on four hydrothermal vents from SWIR, representing distinct characteristics in terms of temperature, pH and metal compositions, by using Illumina sequencing of the 16S small subunit ribosomal RNA (rRNA) genes, to correlate bacterial and archaeal populations with the nature of the vents influenced by ultraslow spreading features. Epsilon-, Gamma-, Alpha -, and Deltaproteobacteria and members of the phylum Bacteroidetes and Planctomycetes , as well as Thaumarchaeota, Woesearchaeota , and Euryarchaeota were dominant in all the samples. Both bacterial and archaeal community structures showed distinguished patterns compared to those in the fast-spreading East Pacific Ridge or the slow-spreading Mid-Atlantic Ridge as previously reported. Furthermore, within SWIR, the microbial communities are highly correlated with the local temperatures. For example, the sulfur-oxidizing bacteria were dominant within bacteria from low-temperature vents, but were not represented as the dominating group recovered from high temperature (over 300°C) venting chimneys in SWIR. Meanwhile, Thaumarchaeota , the ammonium oxidizing archaea, only showed high relative abundance of amplicons in the vents with high-temperature in SWIR. These findings provide insights on the microbial community in ultraslow spreading hydrothermal fields, and therefore assist us in the understanding of geochemical cycling therein.",
"conclusion": "Conclusions This study reported the distribution and diversity of the prokaryotic communities on the surface of chimneys collected from Longqi field the SWIR. The 16S rRNA gene analysis suggested that bacterial communities were highly diversified among all the detected samples. Compared to bacteria, the lower diversity of archaeal phylotypes agreed with other molecular surveys indicating that marine hydrothermal vent archaeal diversity is relatively limited (Huber et al., 2007 ). Phylotypes, belonging to Gammaproteobacteria and Epsilonproteobacteria , appeared to be diverse and abundant in most of samples. In contrast to the broad taxonomic coverage of bacterial community, archeal 16S rRNA gene sequences were predominated by the Thaumarchaeota and Woesearchaeota . Based on the functional analysis of bacteria and archaeal communities, sulfur-oxidation and reduction may be important energy metabolism pathways in low- and high-temperature vent chimneys with high abundance of SOB within Gammaproteobacteria and Epsilonproteobacteria . Meanwhile, ammonia oxidation may be another major pathway to provide energy for microbial ecology system on high temperature active chimneys. This paper described the results of a molecular phylogenetic analysis of chimneys collected from the Longqi field at ultra-spreading South West Indian ridge by using high through sequencing method. Our results provided more details to characterize the microbial roles in ecologic and minerogenic processes at the SWIR, especially in the S and N cycling.",
"introduction": "Introduction Hydrothermal venting is one of the fundamental processes by which heat and chemical species are transferred from the lithosphere to the ocean, and venting occurs along divergent plate boundaries in every ocean, at all spreading rates, and in a diversity of geological settings (Baker and German, 2004 ). More than 300 seafloor vent fields have been investigated in diverse settings spanning oceanic ridges, volcanic arcs, and hot spots (Corliss et al., 1979 ; Hannington et al., 2011 ). Considering that hydrothermal fluids emanate from the subsurface, these environments are considered “windows into the subseafloor” (Reveillaud et al., 2016 ). According to the disquisitive descriptions on the spreading rates, the ocean ridges have been divided into fast- (~80–180 mm year −1 full rate), intermediate- (~55–70 mm year −1 ), slow-(less than 55 mm year −1 ), and ultraslow- (less than 20 mm year −1 ) spreading ridges (Dick et al., 2003 ; Ehlers and Jokat, 2009 ). Most interest in the mid-oceanic ridges has been focused on hydrothermal activities in the fast-, slow-, and intermediate-spreading ridges (Hannington et al., 2005 ). In January–March 2007, the Chinese research cruise DY115-19 discovered an active hydrothermal field, the 49°39′E field (6 mm year −1 ) on SWIR during the Chinese research cruise DY115-19 (Zhu et al., 2010 ; Tao et al., 2012 ). Before this discovery, only the Gakkel Ridge was discovered on global ultra-slow spreading ridges (Connelly et al., 2007 ). The Gakkel Ridge ranged from 7°W to 86°E (6–11 mm year −1 ) and held numerous anomalies of the hydrothermal activity (Edmonds et al., 2003 ). The SWIR separates the African and Antarctic plates, extends from the east Rodriguez triple junction (RTJ) to the west Bouvet triple junction (BTJ), and spreads at a full rate of 14 mm year −1 (Sauter and Cannat, 2010 ). The Longqi vent field at 49°39′E, 37°47′S on SWIR was discovered and recognized as the first active field found on this ultraslow spreading ridge (Zhu et al., 2010 ). The surrounding area is basaltic-hosted environments and lacking sediments. According to the previous study of proposed modes for Longqi field, the significantly thinned crust was observed. This suggested that the tectonics were probably characterized by the early stage of the detachment fault in the area, which provided pathways for hydrothermal circulation period. Within Longqi hydrothermal field, three venting areas, the Vent S, M, and N, have been confirmed (Tao et al., 2012 ). This site offered new and exciting prospects for expanding the known ranges of minerals, fluids, biodiversity, and hydrothermal deposits at ultraslow-spreading ridge (Peng et al., 2011 ). Deep-sea hydrothermal vents are some of the most biologically productive ecosystems on the Earth, yet receive little to no input of organic matter derived photosynthetically (Rutherford, 2014 ). The ecosystems at hydrothermal vents host complex, dynamic habitats characterized by steep gradients in temperature and geochemistry (Jannasch and Mottl, 1985 ). In the ridge habitats, the permeable mineral structures, and the continued mixing of chemically-reduced, vent-derived fluids with oxidized seawater provides favorable conditions that support the growth of microbial communities (Frank et al., 2013 ). Chemoautotrophs inhabiting these areas act as important primary producers, transferring the energy from the geothermal source to the higher trophic levels through several important microbial chemosynthetic pathways such as sulfur-oxidation, nitrification, etc. (Sievert and Vetriani, 2012 ). However, most of our knowledge on the microbial communities in hydrothermal vents has come from fast-spreading ridges, such as East Pacific Rise (EPR) (Gaill et al., 1987 ; Sylvan et al., 2012 ) and the slow-spreading Mid-Atlantic Ridge (MAR) (Flores et al., 2011 ). The distributional patterns of the microorganisms that colonize deep-sea hydrothermal vent chimneys at ultraslow-spreading ridge and their link to the geologic setting remain poorly understood, partly because of sampling limits. In this study, we had a chance to obtain environmental samples on/off chimney in SWIR by Jiao Long manned submersible and applied high-through sequencing on 16S rRNA genes. In this case, we could perform detailed analysis on the microbial communities and link them with their local habitat.",
"discussion": "Discussion Methodological considerations In this study, high-throughput, DNA-based analysis of environmental samples has been applied to investigate the microbial communities of chimney samples collected from Longqi hydrothermal field. Recent studies have found archaeal DNA was poorly recovered from lower temperature, diffuse flow vents or inactive chimneys (Bourbonnais et al., 2012 ; Sylvan et al., 2012 ; Gulmann et al., 2015 ) and speculated that it was due to low low abundance or primer mismatch. In this study, we presented the archaeal community analysis of chimney samples including low-temperature deep-sea hydrothermal vent from Longqi field at SWIR using high-throughput Miseq sequencing. A total of 110 473 and 200 630 amplicons of bacterial and archaeal 16S rRNA genes, respectively, have been obtained, suggesting that the methods appeared to be efficient. Deep sequencing of archaea and bacteria from those chimneys revealed thousands of bacterial and archaeal lineages, the majority of which appeared in very low abundance just as presented in the previous studies (Sogin et al., 2006 ; Huber et al., 2007 ). Microbial diversity and connection with regional geochemical parameters According to the Venn diagrams showing (Figure S6 ), JL94D and JL94H shared the same sampling location and also the highest percent of bacterial OTUs, 9.6%, followed by the overlap between JL90 and JL94D, 7.7%, JL90, and JL94H, 7.6%. Besides, the bacterial OTUs overlap between JL95 and JL94 or JL90 about or below 6.2%. The species composition overlap reflected the microbial resemblance among the four chimneys at Longqi field on SWIR. Microbial community structures were clearly correlated to the environmental parameters, and among all the considered parameters the in situ temperature was the most influential one. Previous studies demonstrated that the biogeographical patterns of microbial communities were shaped in part by local fluid geochemistry in active hydrothermal vent chimneys (Flores et al., 2011 ), mineralogy on inactive seafloor sulfide deposits (Toner et al., 2013 ) and geological processes, such as eruption, on diffuse-flow vents (Gulmann et al., 2015 ). To evaluate the effects of linear correlation between environmental factors and microbial communities, redundancy analysis (RDA) is used in this study (Table S1 ). Compared to the other high-temperature vents and low-temperature vent of JL94D, the RDA analysis also confirmed the separation of JL95 from other chimneys, correlated with a significant decrease of sulfur, Fe, and Zn content, with the increase of temperature of fluid and manganese content of chimney (Figure S5 ). But there was no genus which was found to have a significant linear correspondence with any environmental factor. Therefore, we assumed that regional geochemical condition might have affected the structure and function of the microecosystem in the SWIR region. Inferred microbial metabolic potentials in energy metabolism Potential energy sources for deep-sea vent chemoautotrophy include reduced sulfur compounds, molecular hydrogen, reduced metals and ammonium. Classic sulfur-oxidizing bacteria have been detected as the dominant families, suggesting a strong sulfur-metabolizing potential in all tested chimney samples (Table 3 ). Many members within Epsilon - and Gammaproteobacteria are known to be chemoautotrophs utilizing inorganic sulfur as election donor to gain energy (Nakagawa et al., 2005 ; Yamamoto and Takai, 2011 ; Anderson et al., 2015 ). The deep-sea chemoautotrophic Gammaproteobacteria , possess two different sulfur-oxidization pathways including the reverse sulfate reduction and the Sox multienzyme system without SoxCD, and strictly require co-existence of reduced sulfur compounds and O 2 (Yamamoto and Takai, 2011 ). In this study, Gammaproteobacteria dominated in all detected chimney samples with the sulfide-oxidizing bacterium within the genera of Thiotrichaceae, Ectothiorhodospiraceae, Thiohalophilus , and Piscirickettsiaceae (Figure 4 ). It seemed to indicate that both the reduced sulfur compounds and O 2 were steadily supplied into the chimney habitats. Epsilonproteobacteria were known to play a significant role in carbon, nitrogen and sulfur cycling and had consistently shown to be the most numerically abundant bacteria in sediment (López-García et al., 2003 ), hydrothermal fluids (Huber et al., 2010 ), hydrothermal plumes (Nakagawa et al., 2005 ), and vent chimneys (Campbell et al., 2001 ; Opatkiewicz et al., 2009 ; Dahle et al., 2013 ). Based on our data, Epsilonproteobacterial amplicons ranged from 1.3% at high-temperature vent chimneys to 37.6% for diffusive vent chimneys. Certain sequences dominated and closely related to the known chemosynthetic, sulfur-oxidizing genera Sulfurovum and Sulfurimonas . Similar communities were also found in cool, diffusive flow at Axial Seamount on the Juan de Fuca Ridge (Akerman et al., 2013 ) and the biofilms growing on the chimney walls at the Loki's Castle vent field (Dahle et al., 2013 ). The most abundant genera of Sulfurovum and Sulfurimonas within Epsilonproteobacteria were recovered from all active chimneys in our study, and also recovered from inactive sulfides in the EPR (Sylvan et al., 2012 ). It is possible that these groups represent the widely distributed species at active sulfides and the survivable relict populations at inactive chimneys by oxidizing sulfide minerals. Table 3 Potential ecological function of tag sequences for which obvious metabolisms can be inferred. * Function Bacterial/Archaeal taxonomy Relative abundance (%) JL90 JL94D JL94H JL95 BACTERIA S oxidation Aquificae; Aquificales; Aquificaceae; Hydrogenivirga 0.016 0.105 1.780 0.000 Epsilonproteobacteria; Campylobacterales; Helicobacteraceae 36.326 24.724 8.633 1.246 Gammaproteobacteria; unknown; unknown; Thiohalophilus 0.193 0.040 0.855 1.065 Gammaproteobacteria; Thiotrichales; Thiotrichaceae 5.360 5.955 4.160 1.950 Gammaproteobacteria; Thiotrichales; Piscirickettsiaceae 2.908 0.642 0.282 0.295 Gammaproteobacteria; Chromatiales; Ectothiorhodospiraceae 0.971 0.191 2.728 2.663 Sulfate reduction Nitrospira; Nitrospirales; Nitrospiraceae; Thermodesulfovibrio 0.429 0.657 22.053 0.246 Deltaproteobacteria; Desulfarculales; Desulfarculaceae; Desulfatiglans 0.006 0.557 0.154 0.008 Deltaproteobacteria; Desulfobacterales; Desulfobacteraceae 0.034 1.786 0.604 0.016 Deltaproteobacteria; Desulfobacterales; Desulfobulbaceae 3.028 5.398 3.790 0.131 Epsilonproteobacteria; Campylobacterales; Campylobacteraceae; Sulfurospirillum 0.360 0.627 0.194 0.008 Thermodesulfobacteria; Thermodesulfobacteriales; Thermodesulfobacteriaceae; Thermosulfurimonas 0.022 0.060 1.044 0.008 Sum of S oxidation and sulfate reduction 49.653 40.742 46.277 7.636 Ammonia oxidation Gammaproteobacteria; Chromatiales; Chromatiaceae; Nitrosococcus 1.196 0.115 0.375 1.868 Nitrite oxidation Nitrospira; Nitrospirales; Nitrospiraceae; Nitrospira 0.002 0.005 0.026 0.992 Nitrate reduction Epsilonproteobacteria; Nautiliales; Nautiliaceae; Nitratifractor 0.489 0.105 0.062 0.057 Nitrification Betaproteobacteria; Nitrosomonadales; Nitrosomonadaceae; Nitrosomonas 0.002 0.010 0.004 1.418 N fixation Alphaproteobacteria; Rhizobiales 5.739 0.562 0.450 2.622 Sum of ammonia, nitrite oxidation, nitrification and N fixation 7.428 0.797 0.917 6.957 H oxidation Aquificae; Aquificales; Aquificaceae; Hydrogenobacter 0.028 0.080 1.908 0.016 Aquificae; Aquificales; Hydrogenothermaceae; Persephonell 0.072 0.261 4.958 0.033 Epsilonproteobacteria; Campylobacterales; Hydrogenimonaceae; Hydrogenimonas 0.290 0.025 0.198 0.008 Sum of H Oxidation 0.39 0.366 7.064 0.057 CH4 oxidation Gammaproteobacteria; Methylococcales; Methylococcaceae; Methylothermus 1.122 0.161 2.318 0.057 Fe(II) oxidation Zetaproteobacteria; Mariprofundales; Mariprofundaceae; Mariprofundus 1.295 2.468 1.084 0.705 Fe(III) reduction Deltaproteobacteria; Desulfuromonadales; Desulfuromonadaceae; Desulfuromusa 0.238 1.801 0.591 0.041 Sum of Fe(II) oxidation and Fe(III) reduction 2.655 4.43 3.993 0.803 Mn oxidation Alphaproteobacteria; Rhodobacterales; Rhodobacteraceae; Roseobacter 0.068 1.234 0.353 0.049 Total bacteria 60.194 47.569 58.604 15.502 ACHAEA Sulfate reduction Crenarchaeota; Thermoprotei; Desulfurococcales 0.538 0.003 7.460 0.014 Crenarchaeota; Thermoprotei; Thermoproteales 0.326 0.003 1.111 0.004 Ammonia oxidation Thaumarchaeota 45.667 1.062 24.636 42.745 Total archaea 46.531 1.068 33.207 42.763 * The relative abundance in sequencing library is for each sample's Miseq data set. Taxa are designated by class (phylum for Crenarchaeota and Thaumarchaeota), order, family, and genus . Ammonia oxidation is the first step of nitrification, in which ammonia is first oxidized to nitrite by ammonia-oxidizing bacteria and/or archaea (AOB or AOA), then subsequently to nitrate by nitrite-oxidizing bacteria (NOB). The Thaumarchaeota had rapidly gained much attention after the discovery that some of them have been able to oxidize ammonia aerobically, providing the first example of nitrification in the Archaea and therefore extending the range of microorganisms capable of this important metabolism, which was previously thought to be restricted to a few proteobacterial lineages (Könneke et al., 2005 ). Therefore, we deduced that Thaumarchaeota might be the major Ammonia Oxidizing Archaea (AOA) among the recovered microorganisms at high-temperature vent chimneys JL94H and JL95 with relative abundance over 40%, also for JL90 with ~24% (Figure 2 ). Besides, the Ammonia Oxidizing bacteria (AOB) within Nitrosococcus were recovered and represented the abundant genus (over 1%) in JL90 and JL95. Related genus of nitrifier, the Nitrospira , and Nitrosomonas , were found in all samples and with highest abundance in JL95 (Table 3 ). Recently, the completely nitrifying bacterium from the genus Nitrospira was reported (Daims et al., 2015 ), indicating that the globally distributed nitrite oxidizers fundamentally changed the picture of nitrification and might act as key microbial communities involved in nitrogen-cycling on the high-temperature chimney JL95 and other samples at Longqi hydrothermal field. The reduced metals (Fe, Mn, Cu, etc.,) are the endmembers in vent fluids and potential energy sources for deep-sea vent chemoautotrophs. Fe(II) is a common and often the most dominant metal. Microaerophilic Fe-oxidizing microorganisms (FeOM) colonize gradients of Fe(II) and oxygen, taking advantage of the available chemical energy. Vast communities of FeOM proliferate at deep sea hydrothermal vents, forming mineralized mats (Chan et al., 2013 ). The “zetaproteobacterium” Mariprofundus ferooxydans (Emerson and Moyer, 2002 ) from Loihi Seamount, and several Alpha- and Gammaproteobacteria strains are reported as the chemoautotrophic Fe(II) oxidizers described from deep-sea vents (Edwards et al., 2003 ). Mariprofundus is the sole member of the class Zetaproteobacteria in the phylum Proteobacteria . Several cultured members (JV-1, PV-1) of Fe-oxidizing Mariprofundus were isolated from deep-sea hydrothermal fields (Singer et al., 2011 ; Makita et al., 2017 ). The genus of Mariprofundus were observed as only abundant FeOM group (relative abundance over 1%) in the chimney samples JL90, JL94D, and JL94H, except in JL95 with 0.705 (Table 3 ), indicating that Fe-oxidizing bacteria within Mariprofundus were common at the Longqi field and probably played a major role in Fe oxidation. Mn(II) oxidation mediated by heterotrophic Bacillus species in Guaymas Basin hydrothermal plumes as reported previously (Dick et al., 2006 ). Those species were not recovered in any of the chimney samples collected from Longqi field at SWIR, but the genus of Roseobacter in Alphaproteobacteria were observed and inferred as the Mn(II) oxidizer. Hydrogen has also been shown to be an important energy source in vent fluids at the Logatchev and Rainbow fields on the Mid-Atlantic Ridge (Takai et al., 2006 ). Chemoautotrophs with the ability to derive energy from H 2 oxidation have been isolated from various deep-sea hydrothermal fields, including Aquificales, Epsilonproteobacteria, Desulfurococcales, Methanococcales, Thermodesulfobacteriales , and Deferribacterales (Nakagawa and Takai, 2008 ). Analysis of tag sequences revealed members of the genera Hydrogenobactera and Persephonell within the Aquificae and Hydrogenimonas in the Epsilonproteobacteria (Table 3 ). These groups are likely H 2 -oxidizing bacteria. Trans-regional distribution pattern in microbial communities of hydrothermal vents To evaluate the effects of geological and geochemical characteristic on microbial communities on the surface of active chimneys in Longqi field at SWIR, we compared the bacterial and archaeal distribution pattern with habitats of active chimney both from slow-spreading ridge of MAR and fast-spreading ridge of EPR. Overall, the microbial composition on active chimneys recovered by tag sequencing at SWIR in our libraries was different from hydrothermal vents at EPR and MAR by using cluster analysis (Figures 7 , 8 ). The results showed that all archaeal communities from chimneys at SWIR and EPR were clustered into different branches from the high-temperature vent chimney in MAR. On the other hand, the high-temperature vent chimney LS7 from MAR was highly dominated by Epsilonproterobacteria, which quite different from the bacterial composition of other active chimneys from SWIR and EPR, might lead to be clustered into a separate branch. The bacterial communities of JL95 and CH7 from EPR were surprisingly clustered in to a group with limited Epsilonproteobacteria , but with dominant Gammaproteobacterial sulfur oxidizers. The order Xanthomonadales within Gammaproteobacteria was only dominantin JL95, which was also observed in the microbial community of deep-sea sediment (Li et al., 2015 ). Within archaeal communities, the culturable genus of Pyrococcus and Thermococcus were recovered frequently in molecular environmental surveys at hydrothermal vent (Edgcomb et al., 2007 ; Flores et al., 2011 ; Li et al., 2014 ), but were absent in our archaeal tag sequences, which might be because of the distinct geographic locations and geochemical conditions related. Figure 7 Clustering analysis tree of the archaeal community structure of chimneys from SWIR, EPR and MAR. LS7 = Lucky Strike. Lucky Strike vent field located at MAR (Flores et al., 2011 ); CH7 = chimney sample from EPR, this study. Figure 8 Clustering analysis tree of the bacterial community structure of chimneys from SWIR, EPR and MAR. LS7 = Lucky Strike. Lucky Strike vent field located at MAR (Flores et al., 2011 ); CH7 = Chimney sample from EPR, this study."
} | 5,646 |
34854377 | PMC8735865 | pmc | 8,816 | {
"abstract": "From decision making to perception to language, predicting what is coming next is crucial. It is also challenging in stochastic, changing, and structured environments; yet the brain makes accurate predictions in many situations. What computational architecture could enable this feat? Bayesian inference makes optimal predictions but is prohibitively difficult to compute. Here, we show that a specific recurrent neural network architecture enables simple and accurate solutions in several environments. This architecture relies on three mechanisms: gating, lateral connections, and recurrent weight training. Like the optimal solution and the human brain, such networks develop internal representations of their changing environment (including estimates of the environment’s latent variables and the precision of these estimates), leverage multiple levels of latent structure, and adapt their effective learning rate to changes without changing their connection weights. Being ubiquitous in the brain, gated recurrence could therefore serve as a generic building block to predict in real-life environments.",
"introduction": "Introduction Being able to correctly predict what is coming next is advantageous: it enables better decisions ( Dolan and Dayan, 2013 ; Sutton and Barto, 1998 ), a more accurate perception of our world, and faster reactions ( de Lange et al., 2018 ; Dehaene et al., 2015 ; Saffran et al., 1996 ; Sherman et al., 2020 ; Summerfield and de Lange, 2014 ). In many situations, predictions are informed by a sequence of past observations. In that case, the prediction process formally corresponds to a statistical inference that uses past observations to estimate latent variables of the environment (e.g. the probability of a stimulus) that then serve to predict what is likely to be observed next. Specific features of real-life environments make this inference a challenge: they are often partly random, changing, and structured in different ways. Yet, in many situations, the brain is able to overcome these challenges and shows several aspects of the optimal solution ( Dehaene et al., 2015 ; Dolan and Dayan, 2013 ; Gallistel et al., 2014 ; Summerfield and de Lange, 2014 ). Here, we aim to identify the computational mechanisms that could enable the brain to exhibit these aspects of optimality in these environments. We start by unpacking two specific challenges which arise in real-life environments. First, the joint presence of randomness and changes (i.e. the non-stationarity of the stochastic process generating the observations) poses a well-known tension between stability and flexibility ( Behrens et al., 2007 ; Soltani and Izquierdo, 2019 ; Sutton, 1992 ). Randomness in observations requires integrating information over time to derive a stable estimate. However, when a change in the estimated variable is suspected, it is better to limit the integration of past observations to update the estimate more quickly. The prediction should thus be adaptive, that is, dynamically adjusted to promote flexibility in the face of changes and stability otherwise. Past studies have shown that the brain does so in many contexts: perception ( Fairhall et al., 2001 ; Wark et al., 2009 ), homeostatic regulation ( Pezzulo et al., 2015 ; Sterling, 2004 ), sensorimotor control ( Berniker and Kording, 2008 ; Wolpert et al., 1995 ), and reinforcement learning ( Behrens et al., 2007 ; Iglesias et al., 2013 ; Soltani and Izquierdo, 2019 ; Sutton and Barto, 1998 ). Second, the structure of our environment can involve complex relationships. For instance, the sentence beginnings \"what science can do for you is...\" and \"what you can do for science is...\" call for different endings even though they contain the same words, illustrating that prediction takes into account the ordering of observations. Such structures appear not only in human language but also in animal communication ( Dehaene et al., 2015 ; Hauser et al., 2001 ; Robinson, 1979 ; Rose et al., 2004 ), and all kinds of stimulus-stimulus and stimulus-action associations in the world ( Saffran et al., 1996 ; Schapiro et al., 2013 ; Soltani and Izquierdo, 2019 ; Sutton and Barto, 1998 ). Such a structure is often latent (i.e. not directly observable) and it governs the relationship between observations (e.g. words forming a sentence, stimulus-action associations). These relationships must be leveraged by the prediction, making it more difficult to compute. In sum, the randomness, changes, and latent structure of real-life environments pose two major challenges: that of adapting to changes and that of leveraging the latent structure. Two commonly used approaches offer different solutions to these challenges. The Bayesian approach allows to derive statistically optimal predictions for a given environment knowing its underlying generative model. This optimal solution is a useful benchmark and has some descriptive validity since, in some contexts, organisms behave close to optimally ( Ma and Jazayeri, 2014 ; Tauber et al., 2017 ) or exhibit several qualitative aspects of the optimal solution ( Behrens et al., 2007 ; Heilbron and Meyniel, 2019 ; Meyniel et al., 2015 ). However, a specific Bayes-optimal solution only applies to a specific generative model (or class of models [ Tenenbaum et al., 2011 ]). This mathematical solution also does not in general lead to an algorithm of reasonable complexity ( Cooper, 1990 ; Dagum and Luby, 1993 ). Bayesian inference therefore says little about the algorithms that the brain could use, and the biological basis of those computations remains mostly unknown with only a few proposals highly debated ( Fiser et al., 2010 ; Ma et al., 2006 ; Sahani and Dayan, 2003 ). Opposite to the Bayes-optimal approach is the heuristics approach: solutions that are easy to compute and accurate in specific environments ( Todd and Gigerenzer, 2000 ). However, heuristics lack generality: their performance can be quite poor outside the environment that suits them. In addition, although simple, their biological implementation often remains unknown (besides the delta-rule [ Eshel et al., 2013 ; Rescorla and Wagner, 1972 ; Schultz et al., 1997 ]). Those two approaches leave open the following questions: Is there a general, biologically feasible architecture that enables, in different environments, solutions that are simple, effective, and that reproduce the qualitative aspects of optimal prediction observed in organisms? If so, what are its essential mechanistic elements? Our approach stands in contrast with the elegant closed-form but intractable mathematical solutions offered by Bayesian inference, and the simple but specialized algorithms offered by heuristics. Instead, we look for general mechanisms under the constraints of feasibility and simplicity. We used recurrent neural networks because they can offer a generic, biologically feasible architecture able to realize different prediction algorithms (see LeCun et al., 2015 ; Saxe et al., 2021 and Discussion). We used small network sizes in order to produce simple (i.e. low-complexity, memory-bounded) solutions. We tested their generality using different environments. To determine the simplest architecture sufficient for effective solutions and derive mechanistic insights, we considered different architectures that varied in size and mechanisms. For each one, we instantiated several networks and trained them to approach their best possible prediction algorithm in a given environment. We treated the training procedure as a methodological step without claiming it to be biologically plausible. To provide interpretability, we inspected the networks’ internal model and representations, and tested specific optimal aspects of their behavior—previously reported in humans ( Heilbron and Meyniel, 2019 ; Meyniel et al., 2015 ; Nassar et al., 2010 ; Nassar et al., 2012 )—which demonstrate the ability to adapt to changes and leverage the latent structure of the environment.",
"discussion": "Discussion We have shown that the gated recurrent architecture enables simple and effective solutions: with only 11 units, the networks perform quasi-optimally in environments fraught with randomness, changes, and different levels of latent structure. Moreover, these solutions reproduce several aspects of optimality observed in organisms, including the adaptation of their effective learning rate, the ability to represent the precision of their estimation and to use it to weight their updates, and the ability to represent and leverage the latent structure of the environment. By depriving the architecture of one of its mechanisms, we have shown that three of them are important to achieve such solutions: gating, lateral connections, and the training of recurrent weights. Can small neural networks behave like Bayesian agents? A central and much-debated question in the scientific community is whether the brain can perform Bayesian inference ( Knill and Pouget, 2004 ; Bowers and Davis, 2012 ; Griffiths et al., 2012 ; Rahnev and Denison, 2018 ; Lee and Mumford, 2003 ; Rao and Ballard, 1999 ; Sanborn and Chater, 2016 ; Chater et al., 2006 ; Findling et al., 2019 ; Wyart and Koechlin, 2016 ; Soltani and Izquierdo, 2019 ; Findling et al., 2021 ). From a computational viewpoint, there exists no tractable solution (even approximate) for Bayesian inference in an arbitrary environment, since it is NP-hard ( Cooper, 1990 ; Dagum and Luby, 1993 ). Being a bounded agent ( Simon, 1955 ; Simon, 1972 ), the brain cannot solve Bayesian inference in its most general form. The interesting question is whether the brain can perform Bayesian inference in some environments that occur in real life. More precisely, by ‘perform Bayesian inference’ one usually means that it performs computations that satisfy certain desirable properties of Bayesian inference, such as taking into account a certain type of uncertainty and a certain type of latent structure ( Courville et al., 2006 ; Deroy et al., 2016 ; Griffiths et al., 2012 ; Knill and Pouget, 2004 ; Ma, 2010 ; Ma and Jazayeri, 2014 ; Tauber et al., 2017 ). In this study, we selected specific properties and showed that they can indeed be satisfied when using specific (not all) neural architectures. In the changing unigram and changing bigram environments, our results provide an existence proof: there exist plausible solutions that are almost indistinguishable from Bayesian inference (i.e. the optimal solution). They exhibit qualitative properties of Bayesian inference that have been demonstrated in humans but are lacking in heuristic solutions, such as the dynamic adjustment of the effective learning rate ( Behrens et al., 2007 ; Nassar et al., 2010 ; Nassar et al., 2012 ), the internal representation of latent variables and the precision of their estimates ( Boldt et al., 2019 ; Meyniel et al., 2015 ), the precision-weighting of updates ( McGuire et al., 2014 ; Nassar et al., 2010 ; Nassar et al., 2012 ), and the ability for higher-level inference ( Bill et al., 2020 ; Heilbron and Meyniel, 2019 ; Purcell and Kiani, 2016 ). The performance we obtained with the gated recurrent architecture is consistent with the numerous other successes it produced in other cognitive neuroscience tasks ( Wang et al., 2018 ; Yang et al., 2019 ; Zhang et al., 2020 ). Our detailed study reveals that it offers quasi-optimal low-complexity solutions to new and difficult challenges, including those posed by bigram and higher-level structures and latent probabilities that change unpredictably anywhere in the unit interval. We acknowledge that further generalization to additional challenges remains to be investigated, including the use of more than two categories of observations or continuous observations, and latent structures with longer range dependencies (beyond bigram probabilities). Minimal set of mechanisms What are the essential mechanistic elements that enable such solutions? We show that it suffices to have recurrent units of computation equipped with three mechanisms: (1) input, self, and lateral connections which enable each unit to sum up the input with their own and other units’ prior value before a non-linear transformation is applied; (2) gating, which enables multiplicative interactions between activities at the summation step; (3) the training of connection weights. One of the advantages of such mechanisms is their generic character: they do not include any components specifically designed to perform certain probabilistic operations or estimate certain types of latent variables, as often done in neuroscience ( Echeveste et al., 2020 ; Fusi et al., 2007 ; Jazayeri and Movshon, 2006 ; Ma et al., 2006 ; Pecevski et al., 2011 ; Soltani and Wang, 2010 ). In addition, they allow adaptive behavior only through recurrent activity dynamics, without involving synaptic plasticity as in other models ( Farashahi et al., 2017 ; Fusi et al., 2005 ; Iigaya, 2016 ; Schultz et al., 1997 ). This distinction has implications for the timescale of adaptation: in the brain, recurrent dynamics and synaptic plasticity often involve short and long timescales, respectively. Our study supports this view: recurrent dynamics allow the networks to quickly adapt to a given change in the environment ( Figure 3 ), while synaptic plasticity allows the training process to tune the speed of this adaptation to the frequency of change of the environment ( Figure 3—figure supplement 1 ). Our findings suggest that these mechanisms are particularly advantageous to enable solutions with low computational complexity. Without one of them, it seems that a very large number of units (i.e. a large amount of memory) would be needed to achieve comparable performance ( Figure 8 ) (note that universal approximation bounds in vanilla RNNs can be very large in terms of number of units [ Barron, 1993 ; Cybenko, 1989 ; Schäfer and Zimmermann, 2006 ]). These mechanisms thus seem to be key computational building blocks to build simple and effective solutions. This efficiency can be formalized as the minimum number of units sufficient for near-optimal performance (as in Orhan and Ma, 2017 who made a similar argument), and it is important for the brain since the brain has limited computational resources (often quantified by the Shannon capacity, i.e. the number of bits that can be transmitted per unit of time, which here amounts to the number of units) ( Bhui et al., 2021 ; Lieder and Griffiths, 2019 ). Moreover, simplicity promotes our understanding, and it is with the same goal of understanding that others have used model reduction in large networks ( Dubreuil et al., 2020 ; Jazayeri and Ostojic, 2021 ; Schaeffer et al., 2020 ). Since we cannot exhaustively test all possible parameter values, it might be possible that better solutions exist that were not discovered during training. However, to maximize the chances that the best possible performance is achieved after training, we conducted an extensive hyperparameter optimization, repeated for each environment, architecture, and several number of units, until there is no more improvement according to the Bayesian optimization (see Materials and methods). Biological implementations of the mechanisms What biological elements could implement the mechanisms of the gated recurrence? Recurrent connections are ubiquitous in the brain ( Douglas and Martin, 2007 ; Hunt and Hayden, 2017 ); the lesser-known aspect is that of gating. In the next paragraph, we speculate on the possible biological implementations of gating, broadly defined as a mechanism that modulates the effective weight of a connection as a function of the network state (and not limited to the very specific form of gating of the GRU). In neuroscience, many forms of gating have been observed, and they can generally be grouped into three categories according to the neural process that supports them: neural circuits, neural oscillations, and neuromodulation. In neural circuits, a specific pathway can be gated through inhibition/disinhibition by inhibitory (GABAergic) neurons. This has been observed in microscopic circuits, e.g. in pyramidal neurons a dendritic pathway can be gated by interneurons ( Costa et al., 2017 ; Yang et al., 2016 ), or macroscopic circuits, for example in basal ganglia-thalamo-cortical circuits a cortico-cortical pathway can be gated by the basal ganglia and the mediodorsal nucleus of thalamus ( O’Reilly, 2006 ; O’Reilly and Frank, 2006 ; Rikhye et al., 2018 ; Wang and Halassa, 2021 ; Yamakawa, 2020 ). In addition to inhibition/disinhibition, an effective gating can also be achieved by a large population of interacting neurons taking advantage of their nonlinearity ( Beiran et al., 2021 ; Dubreuil et al., 2020 ). Regarding neural oscillations, experiments have shown that activity in certain frequency bands (typically, alpha and beta) can gate behavioral and neuronal responses to the same stimulus ( Baumgarten et al., 2016 ; Busch et al., 2009 ; Hipp et al., 2011 ; Iemi et al., 2019 ; Klimesch, 1999 ; Mathewson et al., 2009 ). One of the most influential accounts is known as ‘pulsed inhibition’ ( Hahn et al., 2019 ; Jensen and Mazaheri, 2010 ; Klimesch et al., 2007 ): a low-frequency signal periodically inhibits a high-frequency signal, effectively silencing the high-frequency signal when the low-frequency signal exceeds a certain threshold. Finally, the binding of certain neuromodulators to the certain receptors of a synapse changes the gain of its input-output transfer function, thus changing its effective weight. This has been demonstrated in neurophysiological studies implicating noradrenaline ( Aston-Jones and Cohen, 2005 ; Salgado et al., 2016 ; Servan-Schreiber et al., 1990 ), dopamine ( Moyer et al., 2007 ; Servan-Schreiber et al., 1990 ; Stalter et al., 2020 ; Thurley et al., 2008 ), and acetylcholine ( Gil et al., 1997 ; Herrero et al., 2008 ) (see review in Thiele and Bellgrove, 2018 ). We claim that gated recurrence provides plausible solutions for the brain because its mechanisms can all be biologically implemented and lead to efficient solutions. However, given their multiple biological realizability, the mapping between artificial units and biological neurons is not straightforward: one unit may map to a large population of neurons (e.g. a brain area), or even to a microscopic, subneuronal component (e.g. the dendritic level). Training: Its role and possible biological counterpart Regarding the training, our results highlight that it is important to adjust the recurrent weights and thus the network dynamics to the environment (and not fix them as in reservoir computing [ Tanaka et al., 2019 ]), but we make no claims about the biological process that leads to such adjustment in brains. It could occur during development ( Sherman et al., 2020 ), the life span ( Lillicrap et al., 2020 ), or the evolution process ( Zador, 2019 ) (these possibilities are not mutually exclusive). Although our training procedure may not be accurate for biology as a whole, two aspects of it may be informative for future research. First, it relies only on the observation sequence (no supervision or reinforcement), leveraging prediction error signals, which have been found in the brain in many studies ( den Ouden et al., 2012 ; Eshel et al., 2013 ; Maheu et al., 2019 ). Importantly, in predictive coding ( Rao and Ballard, 1999 ), the computation of prediction errors is part of the prediction process; here we are suggesting that it may also be part of the training process (as argued in O’Reilly et al., 2021 ). Second, relatively few iterations of training suffice ( Figure 8—figure supplement 1 , in the order of 10–100; for comparison, Wang et al., 2018 reported training for 40,000 episodes in an environment similar to ours). Suboptimalities in human behavior In this study we have focused on some aspects of optimality that humans exhibit in the three environments we explored, but several aspects of their behavior are also suboptimal. In the laboratory, their behavior is often at best qualitatively Bayesian but quantitatively suboptimal. For example, although they adjust their effective learning rate to changes, the base value of their learning rate and their dynamic adjustments may depart from the optimal values ( Nassar et al., 2010 ; Nassar et al., 2012 ; Prat-Carrabin et al., 2021 ). They may also not update their prediction on every trial, unlike the optimal solution ( Gallistel et al., 2014 ; Khaw et al., 2017 ). Finally, there is substantial interindividual variability which does not exist in the optimal solution ( Khaw et al., 2021 ; Nassar et al., 2010 ; Nassar et al., 2012 ; Prat-Carrabin et al., 2021 ). In the future, these suboptimalities could be explored using our networks by making them suboptimal in three ways (among others): by stopping training before quasi-optimal performance is reached ( Caucheteux and King, 2021 ; Orhan and Ma, 2017 ), by constraining the size of the network or its weights (with hard constraints or with regularization penalties) ( Mastrogiuseppe and Ostojic, 2017 ; Sussillo et al., 2015 ), or by altering the network in a certain way, such as pruning some of the units or some of the connections ( Blalock et al., 2020 ; Chechik et al., 1999 ; LeCun et al., 1990 ; Srivastava et al., 2014 ), or introducing random noise into the activity ( Findling et al., 2021 ; Findling and Wyart, 2020 ; Legenstein and Maass, 2014 ). In this way, one could perhaps reproduce the quantitative deviations from optimality while preserving the qualitative aspects of optimality observed in the laboratory. Implications for experimentalists If already trained gated recurrent networks exist in the brain, then one can be used in a new but similar enough environment without further training. This is an interesting possibility because, in laboratory experiments mirroring our study, humans perform reasonably well with almost no training but explicit task instructions given in natural language, along with a baggage of prior experience ( Gallistel et al., 2014 ; Heilbron and Meyniel, 2019 ; Khaw et al., 2021 ; Meyniel et al., 2015 ; Peterson and Beach, 1967 ). In favor of the possibility to reuse an existing solution, we found that a gated recurrent network can still perform well in conditions different from those it was trained in: across probabilities of change points ( Figure 3—figure supplement 1 ) and latent structures ( Figure 6—figure supplement 1 , from bigram to unigram). In this study, we adopted a self-supervised training paradigm to see if the networks could in principle discover the latent structure from the sequences of observations alone. However, in laboratory experiments, humans often do not have to discover the structure since they are explicitly told what structure they will face and the experiment starts only after ensuring that they have understood it, which makes the comparison to our networks impossible in this setting in terms of training (see similar argument in Orhan and Ma, 2017 ). In the future, it could be interesting to study the ability of gated recurrent networks to switch from one structure to another after having been informed of the current structure as humans do in these experiments. One possible way would be to give a label that indicates the current structure as additional input to our networks, as in Yang et al., 2019 . One of our findings may be particularly interesting to experimentalists: in a gated recurrent network, the representations of latent probabilities and the precision of these probability estimates (sometimes referred to as confidence [ Boldt et al., 2019 ; Meyniel et al., 2015 ], estimation uncertainty [ McGuire et al., 2014 ; Payzan-LeNestour et al., 2013 ], or epistemic uncertainty [ Amini et al., 2020 ; Friston et al., 2015 ; Pezzulo et al., 2015 ]) are linearly readable from recurrent activity, the form of decoding most frequently used in neuroscience ( Haxby et al., 2014 ; Kriegeskorte and Diedrichsen, 2019 ). These representations arise spontaneously, and their emergence seems to come from the computational properties of gated recurrence together with the need to perform well in a stochastic and changing environment. This yields an empirical prediction: if such networks can be found in the brain, then latent probability estimates and their precision should also be decodable in brain signals, as already found in some studies ( Bach et al., 2011 ; McGuire et al., 2014 ; Meyniel, 2020 ; Meyniel and Dehaene, 2017 ; Payzan-LeNestour et al., 2013 ; Tomov et al., 2020 )."
} | 6,178 |
33438829 | PMC8085960 | pmc | 8,818 | {
"abstract": "Summary Microbial production of bulk chemicals and biofuels from carbohydrates competes with low‐cost fossil‐based production. To limit production costs, high titres, productivities and especially high yields are required. This necessitates metabolic networks involved in product formation to be redox‐neutral and conserve metabolic energy to sustain growth and maintenance. Here, we review the mechanisms available to conserve energy and to prevent unnecessary energy expenditure. First, an overview of ATP production in existing sugar‐based fermentation processes is presented. Substrate‐level phosphorylation (SLP) and the involved kinase reactions are described. Based on the thermodynamics of these reactions, we explore whether other kinase‐catalysed reactions can be applied for SLP. Generation of ion‐motive force is another means to conserve metabolic energy. We provide examples how its generation is supported by carbon‐carbon double bond reduction, decarboxylation and electron transfer between redox cofactors. In a wider perspective, the relationship between redox potential and energy conservation is discussed. We describe how the energy input required for coenzyme A (CoA) and CO 2 binding can be reduced by applying CoA‐transferases and transcarboxylases. The transport of sugars and fermentation products may require metabolic energy input, but alternative transport systems can be used to minimize this. Finally, we show that energy contained in glycosidic bonds and the phosphate‐phosphate bond of pyrophosphate can be conserved. This review can be used as a reference to design energetically efficient microbial cell factories and enhance product yield.",
"conclusion": "Concluding remarks and future perspectives Conservation of metabolic energy is one of the limiting factors during fermentative product formation. Here, we reviewed the mechanisms available in microorganisms to conserve energy as well as to reduce unnecessary energy expenditure. We distinguished general and product‐specific methods to conserve metabolic energy. General methods like energy‐independent transporters, mechanisms that use the difference in redox potential between redox couples or using disaccharides are the most interesting because they can be applied in many fermentation processes regardless the product of interest. Substrate‐level phosphorylation is one of the main sources of energy under fermentative conditions. It is however usually a product‐specific method (with the exception of the reaction catalysed by phosphorylating glyceraldehyde‐3‐phosphate dehydrogenase) and therefore cannot be applied in general. Some enzymes contributing to SLP have a broad substrate specificity in vitro and appear to be a good option. However, the thermodynamics of such reactions may not allow SLP to function under physiological conditions and should therefore be considered carefully. The options to increase energy conservation described in this review should not be applied to maximize energy conservation per se but be used to optimize product formation. Just as too little energy conservation has detrimental effect on product formation, the same holds for too much energy conservation, as it moves the overall reaction to a thermodynamical equilibrium and may result in decreased product yield caused by excess biomass formation. Nowadays, a lot of research is directed towards improving product titre, productivity and yield by either overexpressing enzymes involved in product pathways, knocking out competing pathways or using redox‐neutral metabolic networks. However, little research is done on using metabolic energy in a more efficient way. Energy conservation is a means to achieve high titre, productivity and yield during microbial processes. Understanding and applying the various mechanisms available in microorganisms to conserve energy is therefore a key step towards improving fermentative product formation.",
"introduction": "Introduction Metabolic engineering has been extensively used in the past decades to improve the production of chemicals by microorganisms (Atsumi et al ., 2008 ; Keasling, 2010 ; Singh et al ., 2011 ; Zhao et al ., 2013 ; Vuoristo et al ., 2015 ). Recent advances in omics and genetic techniques have allowed fast and efficient modifications of microorganisms (Datsenko and Wanner, 2000 ; Mans et al ., 2015 ), broadening the spectrum of both substrates and products (Zhang et al ., 2008 ; Jung et al ., 2010 ; Lindberg et al ., 2010 ; Yim et al ., 2011 ). Fermentation is a well‐studied metabolic concept in which a substrate is oxidized to an intermediate – resulting in the reduction of redox cofactors – after which the intermediate is reduced – regenerating the oxidized cofactors. Microbial fermentation has been used to produc biofuels and bulk chemicals (Bennett and San, 2001 ; Bechthold et al ., 2008 ; Abdel‐Rahman et al ., 2013 ; Wang et al ., 2016 ). These chemicals compete with petrochemical‐derived compounds; therefore, their manufacture requires high targets for productivity, titre and – most importantly – substrate efficiency. In microbial processes, the carbon and energy source are generally used for maintenance, growth and product formation. To maximize product yield, growth must be minimized, and product formation should ideally conserve sufficient metabolic energy to fulfil the energy requirements of the cells (Fig. 1 ). If not, a part of the substrate is dissimilated to CO 2 and H 2 O by respiration to fulfil the energy requirement of the cell. In addition, the yield of the metabolic pathway (Y P ) designed to convert substrate into product should be equal or very close to the maximum theoretical yield (Y E ). The Y E can be determined based on the ratio of the degree of reduction of substrate and product (Cueto‐Rojas et al ., 2015 ; Vuoristo et al ., 2016 ). The degree of reduction represents the number of electrons in a molecule available for chemical reactions. To reach Y E , metabolic pathways should be designed such that all electrons present in the substrate end up in the product and therefore be redox‐neutral. The use of external electron acceptors, like in respiration, deviates electrons away from the product and is therefore a less preferred option because it decreases the network yield (Weusthuis et al ., 2020 ). Fig. 1 Conservation of additional metabolic energy in the product pathway to improve product yield. On the left, classical aerobic bioconversion where part of the substrate is diverted away from product formation by dissimilation to fulfil the cells energy requirement. On the right, improved product formation by capturing metabolic energy in the product‐forming pathways. Ethanol and lactic acid are synthesized from glucose by redox‐neutral pathways that generate ATP, explaining why their practical yields approach Y E . It is not straightforward to find such pathways for other substrate/product combinations. Obtaining redox balance often requires a metabolic network consisting of at least two pathways – one resulting in cofactor reduction, the other resulting in cofactor oxidation – that together act in a redox‐balanced way. For instance, two parallel pathways combining oxidative and non‐oxidative glycolysis have been implemented in Corynebacterium glutamicum to create a near redox‐neutral metabolic network for the production of L‐glutamate (Chinen et al ., 2007 ). This approach allowed to reach a practical yield of 90% of the maximum theoretical yield on glucose. Another example is the redox‐neutral combination of oxidative and reductive branches of TCA cycle for the production of succinate, citrate, itaconate (Sánchez et al ., 2005 ; Vuoristo et al ., 2016 ) and 1,4‐butanediol (Yim et al ., 2011 ). These metabolic networks ideally should also provide energy for maintenance and growth. Table 1 shows the chemical conversion equations and thermodynamics of the previously described metabolic networks. The Gibbs free energy ΔG 0 \n ’ of these reactions is comparable to the ones of ethanol and lactic acid production from glucose (Table 1 ). These negative ΔG 0 \n ’ values show that – in principle – sufficient free energy is liberated to be conserved as metabolic energy (Cueto‐Rojas et al ., 2015 ). Metabolic reactions, however, in which energy can be conserved, are not common. Consequently, it is not straightforward to realize net ATP formation in these metabolic networks. In these cases, respiration is used to conserve metabolic energy, which has negative consequences for yield and productivity (Weusthuis et al ., 2011 ). Table 1 Thermodynamics of microbial processes using redox‐neutral pathways. The ΔG 0 ’ were calculated using eQuilibrator 2.2 with CO 2 as gas (g) and aqueous (aq) for all other compounds (Flamholz et al ., 2012 ; Noor et al ., 2012 ; Noor et al ., 2013 ; Noor et al ., 2014 ). For ATP formation, the EMP pathway and energy‐neutral substrate uptake and product efflux were used. Overall conversions ATP \n a \n \n ΔG 0 ' (kJ mol glucose ‐1 ) Redox‐neutral fermentation processes Glucose(aq) + 2 CO 2 (g) = 4/3 Citrate(aq) + 2/3 H 2 O ‒2/3 to 0 \n b \n \n ‒175 ± 12 Glucose(aq) = 12/11 1,4‐Butanediol(aq) + 18/11 CO 2 (g) + 6/11 H 2 O ‒8/11 \n c \n \n ‒216 ± 10 \n d \n \n Glucose(aq) + 6/7 CO 2 (g) = 12/7 succinate(ag) + 6/7 H 2 O 4/7 \n e \n \n ‒257 ± 8 ATP generation in existing fermentation processes Glucose(aq) = 2 Ethanol(aq) + 2 CO 2 (g) 2 ‒230 ± 13 Glucose(aq) = 2 Lactate(aq) 2 ‒187 ± 4 Glucose(aq) + H 2 O(l) = Acetate(aq) + Ethanol(aq) + 2 Formate(aq) 3 ‒211 ± 6 Glucose(aq) = Butyrate(aq) + 2 CO 2 (g) + 2 H 2 (g) 3 ‒266 ± 18 Glucose(aq) + CO 2 (g) = Succinate(aq) + Formate(aq) + Acetate(aq) 3 ‒249 ± 8 Glucose(aq) + 2 NH 3 (aq) = 2 Alanine(aq) + 2 H 2 O 2 ‒209 ± 4 a Amount of ATP produced (positive sign) or consumed (negative sign). b Calculated as described by Vuoristo et al . ( 2016 ). The –2/3 ATP was obtained by using the reversed glyoxylate cycle, the 0 ATP was obtained by combining the reductive and oxidative TCA shunts to reach redox‐neutral conversion. c Calculated using the pathway described by Yim et al . ( 2011 ) assuming that both reductive and oxidative TCA shunts were applied to obtain redox‐neutral conversion and that the acetate formed was recovered to acetyl‐CoA by means of an acetyl‐CoA synthase at the expense of two ATP equivalents. d The eQuilibrator database does not contain data on 1,4‐butanediol. The Gibbs free energy was estimated by using the value for ( S,S )‐butane‐2,3‐diol instead. e Calculated with a combined reductive and oxidative TCA shunts to reach redox‐neutral conversion. The same result was obtained with a combined reductive TCA cycle and glyoxylate cycle. John Wiley & Sons, Ltd This review focuses on the available mechanisms for the conservation of metabolic energy, – excluding those involved in respiration – and how unnecessary metabolic energy expenditure can be avoided, as well as how to implement them into product‐forming metabolic networks based on carbohydrates."
} | 2,761 |
39274679 | PMC11396315 | pmc | 8,819 | {
"abstract": "Biofouling poses a significant challenge to the marine industry, and silicone anti-biofouling coatings have garnered extensive attention owing to their environmental friendliness and low surface energy. However, their widespread application is hindered by their low substrate adhesion and weak static antifouling capabilities. In this study, a novel silicone polymer polydimethylsiloxane (PDMS)-based poly(urea-thiourea-imine) (PDMS-PUTI) was synthesized via stepwise reactions of aminopropyl-terminated polydimethylsiloxane (APT-PDMS) with isophorone diisocyanate (IPDI), isophthalaldehyde (IPAL), and carbon disulfide (CS 2 ). Subsequently, a nanocomposite coating (AgNPs-x/PDMS-PUTI) was prepared by adding silver nanoparticles (AgNPs) to the polymer PDMS-PUTI. The dynamic multiple hydrogen bonds formed between urea and thiourea linkages, along with dynamic imine bonds in the polymer network, endowed the coating with outstanding self-healing properties, enabling complete scratch healing within 10 min at room temperature. Moreover, uniformly dispersed AgNPs not only reduced the surface energy of the coating but also significantly enhanced its antifouling performance. The antibacterial efficiency against common marine bacteria Pseudomonas aeruginosa ( P .sp) and Staphylococcus aureus ( S .sp) was reduced by 97.08% and 96.71%, respectively, whilst the diatom settlement density on the coating surface was as low as approximately 59 ± 3 diatom cells/mm 2 . This study presents a novel approach to developing high-performance silicone antifouling coatings.",
"conclusion": "4. Conclusions In this study, a nanocomposite polymer coating with exceptional self-healing and antifouling properties was prepared by incorporating AgNPs into a composite PDMS-PUTI matrix. Among the formulations, AgNPs-9/PDMS-PUTI exhibited the best antibacterial and anti-algal performance. Moreover, the synergistic effect of multiple hydrogen bonds formed by the urea and thiourea groups, as well as the dynamic covalent imine bonds, conferred the composite coating with rapid self-healing capabilities. Furthermore, the incorporation of AgNPs as nanofillers into the polymer, combined with the low surface energy characteristics of silicone and the antibacterial properties of AgNPs, further optimized the fouling release and antifouling performance of AgNPs-x/PDMS-PUTI coatings. However, with the increase of AgNP content, the adhesion of the coating decreased, the roughness increased, and the self-healing ability weakened. In conclusion, AgNPS-9/PDMS-PUTI had better comprehensive properties. The cost aspect of AgNPs is significant, but by combining them with high-performance polymer matrices to create composite coatings, the amount of AgNPs can be minimized while maximizing their excellent antibacterial properties. This approach holds promise for achieving high antibacterial rates with a lower amount of AgNPs, leading to cost-effective nanocomposite polymer coatings in the future. As a low surface energy antifouling coating, this coating can be applied to various ships, effectively reducing the adhesion of marine organisms, thereby decreasing energy consumption and minimizing the risk of biological invasion. This work provides a promising pathway towards the development of high-performance silicone-based coatings for marine anti-biofouling.",
"introduction": "1. Introduction Marine biofouling refers to the adherence and accumulation of microorganisms such as bacteria, diatoms, and larger organisms like barnacles, mussels, oysters, and tubeworms on the surfaces of submerged artificial structures. It is associated with detrimental effects, including microbial corrosion, ship drag, and biological invasions, thus posing substantial safety hazards to various underwater facilities and equipment. For over a century, antifouling coatings have been an effective solution to fouling. While tributyltin (TBT) antifouling coatings exhibit excellent antifouling performance, they release organotin biocides into seawater, severely affecting the survival of other marine organisms, which led to a complete ban on their use in 2008 [ 1 , 2 ]. Self-polishing coatings (SPC) have emerged as an effective type of antifouling coating in recent years [ 3 ]. Their antifouling mechanism involves the hydrolysis of resin in water, which then slowly releases low-toxicity antifouling agents such as cuprous oxide. However, microplastics generated during resin hydrolysis may pose potential adverse effects on the marine environment [ 4 ]. In contrast, PDMS-based fouling-release coatings have recently developed rapidly owing to their environmental friendliness. The low surface free energy and elastic modulus make it difficult for fouling organisms to adhere to the surface or allow them to be easily removed by the sheer force of water flow. However, PDMS surfaces cannot remove the mucus secreted by bacteria and diatoms at low water flow rates, leading to poor static antifouling capability. Moreover, PDMS is an elastomer that generally suffers from low adhesion and susceptibility to damage, limiting its widespread application [ 5 ]. To address the aforementioned shortcomings, studies have focused on the addition of antifouling agents, the incorporation of nanofillers, and the endowment of self-healing properties to coatings. Xiong et al. [ 6 ] fabricated various LAP/PDMS composite antifouling coatings by introducing long afterglow phosphors (LAP) with different emission intensities of blue-green (BG), yellow-green (YG), and sky blue (SB) into PDMS. These coatings emit light in dark conditions, disrupting the physiological activities of marine fouling organisms surrounding the coatings, thereby improving their antifouling performance. Selim et al. [ 7 ] manufactured polydimethylsiloxane (PDMS)/ZnO nanorod (NR) composites using an in-situ method to explore the effect of ZnO NR on the superhydrophobicity and antifouling properties of silicone coatings. At a ZnO NR content of 0.5 wt%, it exhibited favorable dispersion abilities, thus enhancing the hydrophobicity, self-cleaning, and antifouling properties of the composite. To enhance coating durability, researchers have also developed self-healing antifouling coatings by constructing dynamic multiple hydrogen bond interactions between hydrogen donors and highly electronegative atoms (hydrogen acceptors) on various polymer chains, forming reversible cross-links. In our previous works [ 8 , 9 ], long-lasting antifouling coatings with efficient self-healing capabilities were prepared through multiple dynamic hydrogen bond interactions between urea and thiourea linkages, as well as disulfide bonds. In addition, the introduction of lipoic acid-benzothiazole antibacterial groups optimized the antifouling performance. Marine field tests demonstrated that this coating possesses long-term static antifouling capabilities. Additionally, imine bonds, also termed Schiff base bonds, are another type of dynamic covalent bond formed by the reversible condensation of primary amines and aldehydes. Once disrupted, dynamic imine bonds can spontaneously re-form, endowing the polymer with self-healing properties [ 10 ]. Silver nanoparticles (AgNPs) have been universally recognized for their efficient broad-spectrum antibacterial properties [ 11 , 12 , 13 ]. The positively charged silver ions are attracted to the negatively charged microbial cell membranes, enabling AgNPs to penetrate cells, disrupt cellular molecules, and cause intracellular damage. Subsequently, AgNPs induce the formation of reactive oxygen species (ROS) in bacterial cells, leading to the dephosphorylation of tyrosine, which hinders cell signal transduction pathways [ 14 , 15 , 16 , 17 ]. Liu et al. [ 18 ] constructed self-healing and long-lasting antifouling AgNP hybrid silicone coatings by synthesizing coordination polymers with Cu 2 + as the coordinating metal and polydimethylsiloxane (PDMS) and polytetramethylene glycol (PTMG) as the main chains and introducing AgNPs. The excellent antibacterial activity of AgNPs imparted the coating with superior antibacterial properties and the ability to prevent bacterial biofilm formation. Cui et al. [ 19 ] chemically bonded silver nanoparticles with silver, copper, and zinc ternary ion-exchanged zeolites through α-lipoic acid and encapsulated them in a hydrophilic polymer to generate a highly antibacterial, highly active, and durable silver ion nanocomposite antibacterial powder coating additive. The formed silver nanoparticle thin layer and hydrophilic film extended the release of active Ag + from the zeolite, whilst Ag + promoted the activation of AgNPs, thereby exerting satisfactory antibacterial effects. In this study, aminopropyl-terminated polydimethylsiloxane (APT-PDMS) was utilized as the main chain and reacted with isophorone diisocyanate (IPDI), isophthalaldehyde (IPAL), and carbon disulfide (CS 2 ) to synthesize a PDMS-based polyurea-thiourea-imine polymer (PDMS-PUTI). The multiple dynamic hydrogen bonds and imine bonds in PDMS-PUTI synergistically conferred excellent self-healing ability on the coating, allowing self-healing at room temperature. Subsequently, PDMS-PUTI was combined with AgNPs to yield the nanocomposite polymer coating (AgNPs-x/PDMS-PUTI). The broad-spectrum and highly efficient antibacterial and anti-algal capabilities of AgNPs endowed the coating with superior antifouling properties. Overall, this study aimed to develop an environmentally friendly composite coating for marine anti-biofouling."
} | 2,378 |
37902338 | PMC10734449 | pmc | 8,820 | {
"abstract": "ABSTRACT As the most common approach for the restoration of degraded grasslands, the use of grass-legume mixtures has long been recognized for its role in increasing aboveground biomass and resisting grassland degradation. However, whether the legumes in these mixtures can help neighboring plants resist the decline in biomass caused by the loss of soil microbial diversity remains a question worthy of investigation. To address this, we employed a dilution method to create a gradient of decreasing microbial diversity in soil and utilized full-factorial combinations of legumes and two grasses to investigate the crucial role of legumes in the mixture. The results showed that compared to monoculture, the mixture of Medicago sativa L. and Elymus dahuricus Turcz. enhanced the biomass of grass species under conditions of soil microbial diversity loss. We then discovered that a significantly enriched Pseudomonas (ASV53), in the grass-legume mixtures under conditions of microbial diversity loss, was positively correlated with plant biomass and nitrogen-fixing ( nifH ) gene abundance, implying that it could be a keystone species. In addition, the grass-legume mixture increased the deterministic processes of microbial community enrichment in the root zone soil by enhancing the process of homogeneous selection. Functional predictions revealed that grass-legume mixtures increased the potential abundance of N-related and phototrophy-related microbial communities in the root zone soil. This study provides an important insight into the mechanism underlying the role of legumes in increasing and maintaining grass biomass despite soil microbial diversity loss. IMPORTANCE Grass-legume mixtures are a common practice for establishing artificial grasslands, directly or indirectly contributing to the improvement of yield. In addition, this method helps maintain soil and plant health by reducing the use of chemical fertilizers. The impact of grass-legume mixtures on yield and its underlying microbial mechanisms have been a focus of scientific investigation. However, the benefits of mixtures in the context of soil microbial diversity loss remain a problem worthy of exploration. In this study, we examined different aboveground and belowground diversity combinations to elucidate the mechanisms by which grass-legume mixtures help maintain stable yields in the face of diversity loss. We identified the significantly enriched Pseudomonas genus microbial ASV53, which was gathered through homogeneous selection and served as a keystone in the co-occurrence network. ASV53 showed a strong positive correlation with biomass and the abundance of nitrogen-fixing genes. These findings provide a new theoretical foundation for utilizing grass-legume mixtures to enhance grass yields and address the challenges posed by diversity loss.",
"conclusion": "Conclusion In conclusion, there was a significant increase in the biomass of grass under grass-legume mixtures, and the legume maintained biomass stability when microbial diversity in the soil was reduced. However, grass-legume mixtures showed no significant impact on soil physicochemical properties. To elucidate the underlying mechanism of intercropping on grasses, we conducted differential analysis and network construction and identified that the pseudomonad ASV53 showed a significant increase in the mixture over the monoculture. We found that ASV53 was closely related to the nitrogen-fixing P. stutzeri A1501 and its abundance was significantly correlated with nitrogen-fixing gene copy number and grass biomass under a mixture planting pattern, and it was enriched in grass root zone soil by homogeneous selection. This study reveals new support for the restoration of artificially degraded grasslands, providing a new perspective on the soil microbial interactions in the root zone during grass-legume mixtures.",
"introduction": "INTRODUCTION Grasslands are one of the most important biomes on Earth, covering nearly one-third of the planet’s land surface ( 1 ). They provide a wide range of ecological, economic, and social benefits, including carbon sequestration, livestock grazing, and biodiversity conservation ( 2 ). However, grasslands are facing numerous threats, with degradation being one of the most significant challenges. Grassland degradation is the loss of grassland productivity and biodiversity due to various factors such as overgrazing, land-use change, climate change, and invasive species ( 3 ). This phenomenon has become a global concern, with significant impacts on the environment and human societies; consequently, the restoration of degraded grasslands has emerged as a pressing concern among scientists and local governments ( 4 ). Restoration management of degraded grasslands typically involves a range of techniques, including fencing to exclude large herbivores, raking to remove dead vegetation and reseeding with appropriate plants, fertilization to enhance soil fertility and productivity, turf transplantation to establish new vegetation patches, as well as controlling rodent and weed populations to minimize competition for resources ( 5 ). Among these practices, the establishment of artificial or semi-artificial grasslands is currently the most widely used method, as it can select suitable grass species, improve grassland productivity, and relieve the pressure on natural grassland ( 6 ). Currently, grass-legume mixtures are the most commonly used planting pattern for artificial grasslands compared with traditional monoculture ( 7 , 8 ). The key reason for selecting legumes as the main crop for a mixture in agriculture is their ability to fix nitrogen, which is highly valued for enhancing soil fertility and reducing dependence on synthetic nitrogen fertilizers ( 9 ). However, not all the nitrogen fixed by legumes necessarily flows to neighboring plants. Studies have found that when the soil environment changes, legumes will prioritize their use of fixed nitrogen, rather than exhibiting the “good neighbor” role as previously believed ( 10 ). Although there have been numerous studies exploring the benefits and mechanisms of grass-legume mixtures on grass yield, most of these studies have primarily focused on investigating the impact of different aboveground plant combinations ( 11 , 12 ). The enrichment of underground microbial communities in response to these combinations and their potential role in helping plants resist the adverse effects of diversity loss has been largely overlooked ( 13 ). Plants and microbial communities have a long co-evolutionary history, and their interactions play crucial roles in shaping soil microbial structures and ecosystem functioning ( 14 ). One important way that plants influence soil microbial communities is via the release of root exudates, which can attract or repel different microbial taxa ( 15 ). As a result, plants can selectively enrich specific microbial groups in the root zone, leading to a plant-specific microbial community structure. Due to their beneficial effects on plant growth and health, plant growth-promoting bacteria (PGPR) have been extensively studied as potential biofertilizers and biocontrol agents for various crops ( 16 – 18 ). Within the aforementioned PGPR, the most notable are the nitrogen-fixing bacteria that assist plants in acquiring nitrogen from the environment. Apart from the symbiotic nitrogen-fixing bacteria that trigger differentiated structures on the host plant (root nodules of legumes and actinorhizal plants), there is also a diverse range of free-living nitrogen-fixing bacteria that can associate with the root system of graminaceous plants, such as Klebsiella pneumoniae , Azotobacter vinelandii , and Azospirillum brasilense ( 19 ). Pseudomonas stutzeri A1501 can assist with nitrogen fixation in the Poaceae family, further expanding the range of nitrogen-fixing bacteria. P. stutzeri A1501 can colonize the root surface and invade the superficial layers of the root cortex ( 20 ). However, it is not clear whether the root zone of grasses will become enriched with specific microbial communities due to the influence of adjacent legumes within mixed grass-legume plantings. Nonetheless, plant species and cultivation methods do affect the community structure of root zone microbes, which has theoretical guidance significance in agricultural production ( 21 ), as such, it is important to clarify the role of legumes in these mixed plantings. Members of Poaceae and Fabaceae are the representative species types in grasslands, with Medicago sativa L. (legume), Elymus dahuricus Turcz. (grass), and Festuca elata Keng. (grass) being three commonly used forage species for the establishment of artificial pastures ( 22 – 24 ). In this study, we used these three plants for pot cultivation under different plant combinations and soil microbial diversities. The use of pot experiments enables the identification of microbial communities that have a significant impact on plant growth, as they provide a more sensitive reflection of microbial changes within the root zone ( 25 ). We systematically investigated the enriched microbial communities of the root zone soil in the grass-legume mixture. Our first objective was to determine whether the grass-legume mixture could help enhance the yield of grass in the context of soil microbial diversity loss. Subsequently, we intended to decipher the underlying mechanisms behind this phenomenon by focusing on the significantly enriched microbial taxa. Our findings revealed that the mixture exhibited deterministic enrichment of the Pseudomonas genus member ASV53 in the face of diversity loss. This organism showed a significant positive correlation with biomass and N-related functions. Through the assistance of these communities, the grass was able to mitigate the negative impacts of soil microbial diversity loss and maintain biomass stability.",
"discussion": "DISCUSSION In this study, we first compared the biomass trends of grass under different plant combinations and found that the MP mixture significantly increased the biomass of grass, and could maintain its production stability under our simulated soil microbial diversity loss. However, this increase in yield was not significantly correlated with soil physicochemical properties, so, we then focused on the diversity and structural differences of root zone microbial communities among the different plant combinations. The results showed no significant differences in α- and β-diversity between monocultures and mixture, but we found that the MP mixture significantly enriched Pseudomonas under soil diversity loss. We then identified ASV53 which had a very close phylogenetic relationship with the previously reported nitrogen-fixing P. stutzeri A1501. ASV53 not only showed an important role in the microbial co-occurrence network but also had a positive correlation with biomass and nitrogen fixation gene abundance in the MP mixture. Moreover, the root zone soil of grass enriched ASV53 through homogeneous selection, and functional prediction results showed that the mixture soil exhibited a stronger N-cycling potential while reducing the risk of pathogenic bacteria occurrence. The yield increase from the grass-legume mixture The use of grass-legume mixtures is a common practice in both agricultural production and the restoration of degraded grasslands ( 11 , 26 ). Besides the nitrogen fixation ability of legumes that can increase nitrogen input to neighboring plants, the mechanism underlying the yield increase in grass species through mixture has long been a focus of attention for ecologists and agricultural experts ( 27 ). Combining previous long-term field experiments ( 28 ) with the conclusions of our study, we found that there were no significant changes in soil nitrate nitrogen, ammonium nitrogen content, total nitrogen, total carbon, moisture content, and soil conductivity among different plant combinations. This indicates that the grass-legume mixture may not have a direct impact on the physical and chemical properties of soil, and the main reason for the increase in biomass of grass species under mixed planting is not related to soil properties. This could be because the nitrogen fixed by legumes is directed to meet their own needs before being provided to surrounding plants, as confirmed by related research in tropical forests ( 29 ). The “good neighbor” role of legumes requires a certain level of soil nitrogen content, and only under conditions of high soil available nitrogen, can legumes reduce competition with neighboring plants, ultimately achieving mutualism and harmonious coexistence ( 10 ). In addition, the loss of soil microbial diversity is also a significant problem faced by grassland degradation ( 30 ). Many studies have shown that there is an important relationship between microbial diversity and productivity in the soil ( 31 , 32 ). Different types of microbial communities play distinct roles, some can decompose organic matter and release nutrients ( 33 ), while others may convert these nutrients into forms that can easily be taken up by plants ( 34 ). Thus, a more diverse microbial community may promote better nutrient cycling and increase plant nutrient use efficiency, ultimately leading to higher crop yields. Some fungi can form symbiotic relationships with plants, such as mycorrhizal fungi that associate with plant roots and help them absorb nutrients ( 35 , 36 ). We first compared the variations in biomass across different plant combinations. We observed that the M and P mixture increased the biomass of P, while mixtures had no significant effect on the biomass of G. Similar observations were made with plant combinations GP and MGP. This is the reason why we focused our subsequent analysis exclusively on the individual growth of M and P and their mixed cultivation as MP. The occurrence of these patterns is not only attributed to the crucial role of the soil microbial community, which we have been investigating but also to the inherent characteristics of the plants themselves ( 37 ). As commonly used grass species in establishing artificial grasslands, G and P exhibit preferences for specific soil environments and tolerances. Empirical knowledge from livestock farming indicates that P thrives in nutrient-rich soils ( 38 ). When P is mixed with M, the nitrogen-fixing capacity of M provides a more abundant nitrogen source, creating a nutrient-rich environment that favors the growth of P. On the contrary, G responds sensitively to soil fertilization, and excessively high nitrogen content in the soil can be detrimental to its growth ( 39 ). Thus, a mixture with M does not increase G biomass. In addition, P demonstrates strong adaptability to temperature, while G prefers cooler environments ( 40 ). The controlled temperature conditions in our study may not have reached the optimal growth temperature for G. Consequently, there were no significant differences in biomass, whether in monoculture or mixture. In our artificially simulated soil microbial diversity loss experiment, both M and P monocultures showed a decrease in biomass. However, the MP mixtures helped to maintain the stable biomass of the grass, indicating that the mixtures must have formed a protective mechanism to help the root zone microbes resist the adverse effects of microbial diversity loss. Grass-legume mixtures changed the composition of the root zone bacterial community from grass In the restoration processes of degraded grasslands, the selection of plant species and planting methods can vary depending on local soil properties or climatic conditions, and both can be important factors influencing soil microbial communities ( 41 ). In our study, after a 2-month pot experiment, we found no significant differences in the α- and β-diversity of soil microbial communities among different plant combinations. This phenomenon could be attributed to multiple factors. First, functional redundancy among soil microbial communities ( 42 ) can result in a similar overall functional diversity regardless of changes in plant combinations. Second, the experimental scale, such as pot experiments, may capture changes in microbial taxa with more sensitivity, but the short duration of the experiment may not have reached the threshold to significantly impact overall α- and β-diversity. Lastly, other environmental factors, including soil type, moisture, temperature, and nutrient availability, apart from plant combinations, can also influence soil microbial communities and their diversity. If these environmental factors are similar across different plant combinations, they may mask any potential differences in the α- and β-diversity of soil microbial communities. Bacteria are key components of important functions in soil and can have beneficial effects on plant growth through direct or indirect interactions ( 43 ). Among them, genera such as Pseudomonas , which are known to promote plant growth, can help plants through mechanisms such as nodulation ( 44 ), antagonism ( 45 ), production of IAA (indole-3-acetic acid) ( 46 ), and siderophore ( 47 ). We found that the MP mixture significantly increased the abundance of Pseudomonas in the root zone soil. In addition, the abundance of Acinetobacter , a group of potential pathogens found in a variety of environments, including fresh and marine water, soil, and marine sediments ( 48 ), was decreased in the mixture, compared to P monoculture. Altogether, the significant changes in abundance of these two key microbial genera resulted in the increased yield of the grass-legume mixture. After identifying significant changes at the genus level, we aimed to further investigate which ASV played a dominant role in the process of the MP mixture. We employed microbial co-occurrence network analysis to identify key nodes in different plant combinations and found that ASV53 had the most connections with other nodes in the network for this mixture. As the ASV53 sequence was annotated as belonging to the Pseudomonas genus in the database, we constructed a phylogenetic tree using the 16S rRNA sequences of all Pseudomonas ASVs in our study and the 16S rRNA sequence of P. stutzeri A1501 , a nitrogen-fixing Pseudomonas strain reported in the literature ( 20 ). The results showed that ASV53 was closely related to A1501. P. stutzeri A1501 is a nitrogen-fixing bacterium isolated from paddy fields, which exhibits significant nitrogen-fixing activity under microaerobic conditions, and its nitrogen fixation products can be rapidly absorbed and utilized by rice plants ( 49 ). To investigate whether ASV53 possessed nitrogen fixation and plant growth-promoting abilities similar to A1501 , we incorporated nitrogen cycling-related functional gene data into the network analysis and found a strong positive correlation between ASV53 and the nitrogenase gene ( nifH ). Moreover, we found that only in the MP mixture was the abundance of ASV53 significantly positively correlated with the copy number of nifH and biomass. We also found that homogeneous selection was an important process through which grass-enriched ASV53 in the root zone soil. Although our study did not investigate the specific mechanisms of ASV53’s interactions with plants in depth, the analysis results suggest that ASV53 , similar to A1501 , plays an important role in nitrogen fixation, helping grasses and leading to increased biomass of grasses and resistance to the adverse effects of decreased biodiversity during mixture planting patterns. This finding could be further validated in subsequent experiments by isolating ASV53 from the soil of grass-legume mixtures and investigating its genetic and phenotypic characteristics. Conclusion In conclusion, there was a significant increase in the biomass of grass under grass-legume mixtures, and the legume maintained biomass stability when microbial diversity in the soil was reduced. However, grass-legume mixtures showed no significant impact on soil physicochemical properties. To elucidate the underlying mechanism of intercropping on grasses, we conducted differential analysis and network construction and identified that the pseudomonad ASV53 showed a significant increase in the mixture over the monoculture. We found that ASV53 was closely related to the nitrogen-fixing P. stutzeri A1501 and its abundance was significantly correlated with nitrogen-fixing gene copy number and grass biomass under a mixture planting pattern, and it was enriched in grass root zone soil by homogeneous selection. This study reveals new support for the restoration of artificially degraded grasslands, providing a new perspective on the soil microbial interactions in the root zone during grass-legume mixtures."
} | 5,206 |
18421347 | null | s2 | 8,822 | {
"abstract": "Sequencing DNA from several organisms has revealed that duplication and drift of existing genes have primarily moulded the contents of a given genome. Though the effect of knocking out or overexpressing a particular gene has been studied in many organisms, no study has systematically explored the effect of adding new links in a biological network. To explore network evolvability, we constructed 598 recombinations of promoters (including regulatory regions) with different transcription or sigma-factor genes in Escherichia coli, added over a wild-type genetic background. Here we show that approximately 95% of new networks are tolerated by the bacteria, that very few alter growth, and that expression level correlates with factor position in the wild-type network hierarchy. Most importantly, we find that certain networks consistently survive over the wild type under various selection pressures. Therefore new links in the network are rarely a barrier for evolution and can even confer a fitness advantage."
} | 253 |
37637129 | PMC10448053 | pmc | 8,825 | {
"abstract": "Microbes play central roles in ocean food webs and global biogeochemical processes. Yet, the information available regarding the highly diverse bacterial communities in these systems is not comprehensive. Here we investigated the diversity, assembly process, and species coexistence frequency of bacterial communities in seawater and sediment across ∼600 km of the eastern Chinese marginal seas using 16S rRNA gene amplicon sequencing. Our analyses showed that compared with seawater, bacterial communities in sediment possessed higher diversity and experienced tight phylogenetic distribution. Neutral model analysis showed that the relative contribution of stochastic processes to the assembly process of bacterial communities in sediment was lower than that in seawater. Functional prediction results showed that sulfate-reducing bacteria (SRB) were enriched in the core bacterial sub-communities. The bacterial diversities of both sediment and seawater were positively associated with the relative abundance of SRB. Co-occurrence analysis showed that bacteria in seawater exhibited a more complex interaction network and closer co-occurrence relationships than those in sediment. The SRB of seawater were centrally located in the network and played an essential role in sustaining the complex network. In addition, further analysis indicated that the SRB of seawater helped maintain the high stability of the bacterial network. Overall, this study provided further comprehensive information regarding the characteristics of bacterial communities in the ocean, and provides new insights into keystone taxa and their roles in sustaining microbial diversity and stability in ocean.",
"conclusion": "5. Conclusion In summary, our study suggests that bacterial communities in seawater have high complexity, and that stochastic processes have a more crucial role in shaping seawater bacterial communities. Additionally, the diversity of both sediment and seawater bacterial communities is positively associated with the relative abundance of sulfate-reducing bacteria (SRB). Moreover, SRB in seawater play a central role and possess complex connections with other taxa in the co-occurrence network. Loss of SRB strongly decreases network complexity and stability. Our work will provide valuable insights into understanding the keystone taxa and their roles in sustaining microbial diversity and stability in ocean.",
"introduction": "1. Introduction Oceans cover approximately 70% of the Earth’s surface and contain 97% of all water on our planet ( Sunagawa et al., 2020 ). Plankton is the dominant life form in the oceans and is comprised of highly dynamic and interacting populations of bacteria, archaea, viruses, protists, and animals that drift with the currents ( Barton et al., 2013 ; Brum et al., 2015 ). Together, these organisms play a significant role in the Earth’s biogeochemical system by contributing to almost half of the net primary production on the planet ( Guidi et al., 2016 ; Carroll et al., 2022 ). Additionally, sediment is another essential component of the ocean ecosystem and is estimated to contain approximately half (>10 29 ) of the microbial cells in the oceans ( Kallmeyer et al., 2012 ). The global ocean sampling expedition substantially promoted the development of ocean ecosystem biology, for example the Tara Oceans project ( Sunagawa et al., 2020 ). The sequencing data obtained from this expedition has provided unparalleled insights into the composition, diversity, function, and distribution patterns of bacterial, archaeal, and viral communities on a global scale. This knowledge has enabled us to better understand ocean biosphere. The interactions between microbes are also crucial for maintaining a diverse microbial community. Due to their small size, high abundance, wide distribution, and short generation time, microbes interact in complex ways ranging from mutualism to competition, via exchanges of materials and energy ( Faust and Raes, 2012 ). Such intricate ecological relationships can be represented as networks, with species as nodes and their relationships as links ( Pržulj and Malod-Dognin, 2016 ). The development of high-throughput sequencing technologies opened a new era in microbiome studies, allowing the opportunity to systematically study the interaction between microbes from various environments using co-occurrence networks. For example, researchers have explored the influence of climate warming on grassland soil microbial network complexity based on co-occurrence networks, and found that climate warming significantly increases network complexity ( Yuan et al., 2021 ). In soil, species coexistence within microbial communities is regulated by community assembly processes, and microbial co-occurrence associations tend to be higher when communities are primarily driven by dispersal limitation relative to species sorting ( Jiao et al., 2020 ). In marine, researchers explored spatiotemporal dynamics of the archaeal community in coastal sediments, and found that seasonality in archaeal co-occurrence patterns, archaea were more connected in winter than in summer ( Liu et al., 2020 ). Additionally, by linking assembly process and species co-occurrence, researchers found that microbial co-occurrence associations in the eastern Indian ocean tended to be higher when deterministic processes were weaker ( Li et al., 2021 ). In addition, networks are also used to investigate the dynamics of ecosystems ( Montesinos-Navarro et al., 2017 ; Toju et al., 2017 ; Ullah et al., 2018 ). One fundamental yet hotly debated question is whether and how the complexity of ecological network affects ecosystem stability. According to Robert May’s Complexity-stability theory, large ecosystems can maintain stability up to a certain critical complexity, This means that the complexity of ecological networks constraints their stability ( May, 1972 ). This critical complexity is determined by the number of species in the ecosystem and the intensity of interactions among them. May’s theory has been confirmed through mathematical modeling ( Thébault and Fontaine, 2010 ; Mougi and Kondoh, 2012 ; Gregorini et al., 2018 ; Qian and Akçay, 2020 ). Furthermore, over the past decade, May’s theory has been examined in real-world ecosystems, where species richness is high and species interactions are complex. After understanding interaction relationships in advance through numerous cultivation experiments. Researchers meticulously constructed ecological interactions networks among species, and investigated the relationships between species diversity, interactions, and community stability. They discovered that May’s theory is not universally applicable in natural ecosystems ( Yodzis, 1981 ; Pimm et al., 1991 ; de Ruiter et al., 1995 ). Recently, some researchers have introduced a new computational framework for estimating the complexity of ecosystems. This framework does not rely on a priori knowledge of the underlying interaction network. Their findings indicate that in natural communities, the relationship between complexity and stability aligns with May’s theory. Specifically, there is a pronounced trade-off between the number of species and their interactions. Natural communities maintain stability by decreasing complexity ( Yonatan et al., 2022 ). While it is well documented that interaction between species have a significant impact on community structure and ecosystem stability, there is limited understanding regarding the effects of SRB on the structure and stability of bacterial networks in the ocean. The aims of this study were to (I) investigate microbial variations among various habitats in eastern Chinese marginal seas; (II) gain insight into the ecological role and significance of SRB in bacterial networks; (III) assess the influence of SRB on the stability of bacterial communities in ocean ecosystems. Given the crucial contributions of SRB in ocean ecosystems, understanding the influence of SRB on the community stability can promote the overall stability and functionality of ocean ecosystems.",
"discussion": "4. Discussion Microorganisms form intricate webs of interactions within an ecological niche. In this study, we investigated the diversity, assembly process, and co-occurrence network of the bacterial community, while also exploring the functional taxa that have an impact on the diversity and complexity of the bacterial community. Our findings demonstrate that: (i) there are significant differences between the bacterial communities in sediment and seawater, with sediment bacterial communities possessing higher diversity and tighter phylogenetic distribution compared to seawater; (ii) stochastic processes are more pivotal in shaping the bacterial community of seawater compared to sediment. Furthermore, the network of bacterial communities in seawater is considerably more complex than that of sediment; (iii) SRB were enriched in core bacterial sub-communities, and the loss of SRB lead to a decrease in network complexity in seawater. Microbial interactions play a central role in maintaining the diversity of microbial communities ( Faust and Raes, 2012 ), and correlation-based network analysis has proven to be effective in exploring co-occurrence patterns and understanding microbial community structure and assembly patterns. For example, this analysis has been successfully applied to discern the direct or indirect ecological linkages among microorganisms in marine water ( Beman et al., 2011 ; Steele et al., 2011 ), soil ( Jiao et al., 2020 ), coastal sediment ( Liu et al., 2020 ), lake ( Zhang et al., 2018 ), and wastewater treatment plants ( Ju et al., 2014 ). Accordingly, in this study, we utilized network analyses to explore significant taxon co-occurrence patterns, and investigated the ecological role of the SRB in the co-occurrence networks of sediment and seawater. Our findings suggest that community composition strongly differs between sediment and seawater, with seawater possessing more complex inter-relationships among bacterial taxa than sediment. Because different groups can complement each other in a complex interaction network, the bacterial community in seawater may be more adaptable to environmental stress and possess higher functional redundancy and resiliency to environmental disturbances than sediment. Indeed, neutral model results support this notion, bacterial communities in seawater are less affected by deterministic processes than sediment. Previous studies suggested that neutral processes can modulate the occurrence frequency of species due to random fluctuations of the microbial community ( Chen et al., 2017 ), and microbial co-occurrence associations tended to be higher when communities were primarily governed by stochastic processes ( Jiao et al., 2020 ; Gao et al., 2021 ). Thus, the difference in network complexity between sediment and seawater may be due to the balance between deterministic processes and stochastic processes. In addition, the result of natural connectivity also aligns with the central ecological belief that complexity begets stability ( Yuan et al., 2021 ), the natural connectivity of low-complexity sediment was lower than that of seawater, which had a higher community complexity. Functional prediction results indicated that sulfate-reducing bacteria (SRB) were enriched in the core sub-communities. SRB are key participants in the ubiquitous process of sulfate reduction, which plays a critical role in the ocean ecosystem. Sulfate reduction encompasses assimilatory sulfate reduction and dissimilatory sulfate reduction, assimilatory sulfate reduction is present in all living organisms, during which sulfate is reduced to hydrogen sulfide and incorporated into cysteine and methionine, which are structural blocks for proteins and polypeptides ( Schiff, 1979 ). Dissimilatory sulfate reduction, which is mediated by SRB, is considered the primary process for the biomineralization of organic matter in marine sediments ( Jørgensen, 1982 ). It is estimated that 11.3 teramoles of sulfate are reduced by SRB annually. This process could account for 12 to 29% of the organic carbon being oxidized to the sea floor ( Bowles et al., 2014 ). Therefore, dissimilatory sulfate reduction is also a crucial driver of the carbon cycle. Given the vital role of SRB in ocean ecosystems, a detailed investigation of their impact on the bacterial community is necessary to reveal their significance in the biogeochemical cycles of carbon and sulfur, and provide insight into the biological factors driving the marine sulfur cycle. In the present study, we found that the SRB sub-community in sediment had higher species richness, evenness, and diversity when compared with seawater, suggesting that the sediment SRB sub-community was more diverse and evenly distributed than seawater. Variations in species diversity are a key factor for ecological stability, and positive effects of diversity on resistance are common. Therefore, SRB in sediment and seawater may be important for the bacterial community of the ocean to resist environmental change. Additionally, in sediment, 67.09% of SRB belonged to rare taxa, for which the mean relative abundance is lower than 0.001%. While in seawater, SRB mainly belonged to abundant taxa with high abundance, and only 12.09% of SRB were rare taxa. These results indicate that SRB groups exist widely in both sediment and seawater, and that rare species contribute to the greater richness and diversity of the SRB sub-communities in sediment. Moreover, recent studies also suggest that abundant taxa have closer relationships with other taxa than rare taxa in oil-contaminated soils ( Jiao et al., 2017 ). Abundant microbes constitute the majority of microbial biomass, and play a crucial role in carbon and nutrient cycling ( Pedrós-Alió, 2012 ). Therefore, these sulfur cycle-related bacterial communities have considerable potential for further exploration. Furthermore, variations in species diversity are a key factor for ecological stability, and positive effects of diversity on resistance are common. Therefore, SRB in sediment and seawater may be important for the bacterial community of the ocean to resist environmental change. In this study, we noted that seawater SRB were centrally located in their co-occurrence network, and exhibited extensive connectivity with other nodes through both direct and indirect interactions. Thus, the removal of SRB in seawater could potentially have a profound impact on the community structure, significantly reducing network complexity and ultimately leading to a decline in community stability. In order to evaluate the ecological role of SRB in the bacterial communities, we compared the node-level topological features of networks with and without SRB, and the change of natural connectivity after removing SRB or other OTUs. Our findings demonstrated that SRB in seawater possess significantly higher degree, betweenness, closeness, and eigenvector centrality than other OTUs. Recent research has shown that species with high degree centrality play a crucial role in network stability ( Banerjee et al., 2018 ), supporting our conclusion that the presence of SRB enhances network complexity and stability in seawater. Notably, our results show that the removal of SRB significantly decreases natural connectivity compared to a network with SRB, underscoring the importance of SRB for the maintenance of network complexity and stability in seawater. There are a few potential limitations that need to be considered when interpreting our findings. Firstly, our study relied on a topology-based approach to predict the ecological role of SRB in sediment and seawater communities. While this approach highlights the importance of SRB within the community, it does not fully capture the true inter-taxon correlations and metabolic connections between them. Therefore, further efforts are required to identify how SRB influence other taxa in the community, and to uncover the mechanisms behind the impact of SRB on community stability. For example, synthetic microbial communities and mathematical simulations could potentially be used to further explain the mechanisms behind this rule. Secondly, ecosystem stability is a multifaceted concept that is comprised of both resistance and resilience when facing disruptions ( Yonatan et al., 2022 ). Resistance refers to the ability of the community to remain unchanged after disturbance, while resilience refers to the rate at which a community recovers to its original status after disturbance ( Allison and Martiny, 2008 ). In our study, natural connectivity mainly refers to resistance. Due to resilience also being an essential index for evaluating community stability, further research should focus on the contribution of the SRB subcommunity to ecological resilience."
} | 4,216 |
36247100 | PMC9543674 | pmc | 8,826 | {
"abstract": "Abstract \n The increased release of dissolved organic matter (DOM) by algae has been associated with the fast but inefficient growth of opportunistic microbial pathogens and the ongoing degradation of coral reefs. Turf algae (consortia of microalgae and macroalgae commonly including cyanobacteria) dominate benthic communities on many reefs worldwide. Opposite to other reef algae that predominantly release DOM during the day, turf algae containing cyanobacteria may additionally release large amounts of DOM at night. However, this night‐DOM release and its potential contribution to the microbialization of reefs remains to be investigated. We first tested the occurrence of hypoxic conditions at the turf algae–water interface, as a lack of oxygen will facilitate the production and release of fermentation intermediates as night‐time DOM. Second, the dissolved organic carbon (DOC) release by turf algae was quantified during day time and nighttime, and the quality of day and night exudates as food for bacterioplankton was tested. Finally, DOC release rates of turf algae were combined with estimates of DOC release based on benthic community composition in 1973 and 2013 to explore how changes in benthic community composition affected the contribution of night‐DOC to the reef‐wide DOC production. A rapid shift from supersaturated to hypoxic conditions at the turf algae–water interface occurred immediately after the onset of darkness, resulting in night‐DOC release rates similar to those during daytime. Bioassays revealed major differences in the quality between day and night exudates: Night‐DOC was utilized by bacterioplankton two times faster than day‐DOC, but yielded a four times lower growth efficiency. Changes in benthic community composition were estimated to have resulted in a doubling of DOC release since 1973, due to an increasing abundance of benthic cyanobacterial mats (BCMs), with night‐DOC release by BCMs and turf algae accounting for >50% of the total release over a diurnal cycle. Night‐DOC released by BCMs and turf algae is likely an important driver in the microbialization of reefs by stimulating microbial respiration at the expense of energy and nutrient transfer to higher trophic levels via the microbial loop, thereby threatening the productivity and biodiversity of these unique ecosystems. \n Read the free Plain Language Summary for this article on the Journal blog.",
"conclusion": "5 CONCLUSIONS This study shows that some turf‐algal communities do not only release large quantities of DOM during the day, but can also do so at night. Further evidence is provided that incomplete organic matter degradation and fermentation may underlie this night‐time DOM release. As night‐time DOM is consumed by bacterioplankton two times faster compared to day‐time DOM, yet results in four times lower BGE, night‐time DOM is likely to significantly contribute to the microbialization of reefs: (a) By increasing the competitiveness of turf algae, which solidifies or even accelerates the underlying benthic community shift from calcifying to non‐calcifying taxa and (b) by further stimulating heterotrophy (i.e. respiration) in microbial communities at the expense of transfer of energy and nutrients through the microbial loop to higher trophic levels. Similar diurnal variation in the quantity and quality of DOM production may also play a role in other ecosystems. Our results may therefore encourage further investigation of diurnal variation in DOM production and its potential effects on community structure and ecosystem functioning.",
"introduction": "1 INTRODUCTION Dissolved organic matter (DOM) is a key component in the biogeochemistry and overall functioning of terrestrial, freshwater and marine ecosystems (e.g. Baines & Pace, 1991 ; Friedlingstein et al., 2020 ; Hansell & Carlson, 2001 ; Kalbitz et al., 2000 ; Thomas, 1997 ). Fixation of atmospheric CO 2 and subsequent release of DOM by photosynthetic organisms provide a major source of organic carbon into lakes, oceans and soils (Hansell & Carlson, 2001 ; Thornton, 2014 ). Rapid consumption and modification of photosynthetically derived DOM by heterotrophic microbial communities decomposes DOM to CO 2 that can be released back into the atmosphere, converts DOM into biomass (Ducklow & Carlson, 1992 ; Hansell et al., 2009 ) or sequesters it as recalcitrant DOM in the deep ocean (Hansell, 2013 ; Jiao et al., 2010 ). Shifts in the community composition of primary producers or decomposers that modify DOM production, alter DOM transformations, or affect DOM consumption may have major implications for the local, regional or even global carbon cycle. In coastal ecosystems, the DOM pool is often primarily fuelled by abundant benthic primary producers (Ziegler & Benner, 1999 ; Wada & Hama, 2013 ; Barrón et al., 2014 ; Reed et al., 2015 ). An example is provided by tropical coral reefs, which represent some of the most productive and diverse ecosystems in the world's oceans (Hatcher, 1988 ; Odum & Odum, 1955 ). Up to 50% of the photosynthates produced on coral reefs is released as DOM into the surrounding water (Davies, 1984 ; Haas et al., 2011 ). This complex mixture of organic molecules, including polysaccharides, proteins and lipids, is not directly available to most heterotrophic organisms as food source (Carlson, 2002 ; Dittmar & Stubbins, 2014 ; Hansell & Carlson, 2001 ). Processing of the DOM released by corals and algae into organic particles and the subsequent transfer to higher trophic levels via the microbial loop (Azam et al., 1983 ) and the sponge loop (de Goeij et al., 2013 ) can reduce the loss of energy and nutrients stored in this locally produced DOM to the open ocean, and is therefore considered pivotal to sustain the high productivity of coral reefs under oligotrophic conditions. A combination of anthropogenic disturbances (e.g. eutrophication, overfishing, ocean acidification, global warming) has led to a devastating decline in the abundance of scleractinian corals, and a concomitant increase in fleshy macroalgae, benthic cyanobacterial mats (BCMs) and turf algae on many reefs around the world (Gardner et al., 2003 ; Hoegh‐Guldberg, 1999 ; Hughes et al., 2017 ; McCook et al., 2001 ). Since these non‐calcifying taxa release more DOM per surface area than scleractinian corals (Haas et al., 2011 ; Mueller et al., 2014 ), this shift in benthic community composition results in an increase in benthic DOM production with potentially major implications for coral reef functioning (de Goeij et al., 2017 ; Haas et al., 2016 ; Pawlik et al., 2016 ). The relative abundance of turf algae has increased dramatically on many reefs, making them often the most abundant functional group within exposed benthic reef communities (Barott et al., 2012 ; Kramer, 2003 ; Vermeij et al., 2010 ). Turf algae are heterogeneous consortia of Chlorophyta, Phaeophyta and Rhodophyta commonly including filamentous cyanobacteria and a distinct community of other associated microbes (Barott et al., 2011 ; Connell et al., 2014 ; Fricke et al., 2011 ; Steneck & Dethier, 1994 ). Fast growth (Littler et al., 2006 ), rapid nutrient uptake (den Haan et al., 2016 ) and the ability to fix dinitrogen (Charpy et al., 2010 ; den Haan et al., 2014 ) allow turf algae to outcompete other reef organisms by rapidly occupying new space. Furthermore, their net areal primary production rates and dissolved organic carbon (DOC) release rates (i.e. during daylight) are among the highest reported for benthic primary producers on coral reefs (Adey & Goertemiller, 1987 ; Haas et al., 2010 ). In general, the release of photosynthetically fixed carbon as DOC is directly linked to primary production with a positive relationship between DOC release and light availability (e.g. Baines & Pace, 1991 ; Cherrier et al., 2014 ; Zlotnik & Dubinsky, 1989 ). Reported DOC release rates in the dark are therefore typically much lower than DOC release rates in the light (Barrón et al., 2014 ; Haas et al., 2010 ; Mueller et al., 2014 ; Zlotnik & Dubinsky, 1989 ). In contrast, BCMs dominated by the cyanobacterium Oscillatoria bonnemaisonii not only release large quantities of DOC during the day, but also release two times more DOC at night (Brocke, Wenzhoefer, et al., 2015 ). Due to the absence of light, nightime DOC release is likely caused by incomplete degradation and anaerobic fermentation of carbohydrates that have accumulated in cyanobacterial cells by their photosynthetic activity during daytime, as has previously been reported for several Oscillatoria species (Heyer et al., 1989 ; Heyer & Krumbein, 1991 ; Stal & Moezelaar, 1997 ). Cyanobacteria typically contribute 20%–50% of the total biomass of turf algae in the Southern Caribbean and are often dominated by Oscillatoria spp. (Fricke et al., 2011 ). Consequently, these turf algae may also release large amounts of DOC at night (Müller, 2015 ), but these rates as well as their contribution to carbon cycling within reef communities have not been studied. It is becoming increasingly evident that not only the quantity, but also the quality (i.e. bio‐availability and utilization) of DOM for key DOM consumers (i.e. microbes and sponges) differs considerably among primary producers. Microbes and sponges metabolize algal‐DOM faster than coral‐DOM (Campana et al., 2021 ; Nelson et al., 2013 ; Rix et al., 2017 ; Silva et al., 2021 ). Algal‐DOM further increases microbial respiration and fuels the fast but inefficient growth of opportunistic microbes (i.e. large amounts of DOM required to support growth; Nelson et al., 2013 ). This shifts the microbial community metabolism from net autotrophy to net heterotrophy (Haas et al., 2013 ) and depletes the local DOM pool to sustain a greater microbial biomass, instead of transferring energy and nutrients stored in DOM to higher trophic levels. This process is commonly referred to as the ‘microbialization of reefs’ (Haas et al., 2016 ; McDole et al., 2012 ). To date, the contribution of fermentation‐derived DOM released at night to the total reef‐wide DOM production, its bio‐availability to and utilization by microbial communities, and its potential role in the microbialization of reefs remains virtually unknown. This study therefore aims to quantify DOC released by turf algae during the day and at night (from here on referred to as ‘day‐DOC’ and ‘night‐DOC’) and to estimate how the contribution of this day‐DOC and night‐DOC release to the local DOC pool may have changed over the past 40 years on a Caribbean reef. Furthermore, we investigated the quality (i.e. C:N ratio, bio‐availability and bacterial growth efficiency) of this turf‐algal DOM for a natural bacterioplankton community to assess its potential contribution to the microbialization of reefs. Specifically, we (a) tested the occurrence of hypoxia at the water–turf algae interface; a prerequisite for fermentation‐derived night‐DOC release, (b) quantified day‐ and night‐DOC release and net primary production by turf algae and (c) assessed the quality of these day‐ and night‐exudates for bacterioplankton in bioassays. Lastly, we (d) extrapolated current DOC release fluxes on a coral reef site on Curaçao and (e) compared those to estimates of historic DOC release fluxes based on benthic community composition data from the 1970s at the same location.",
"discussion": "4 DISCUSSION In this study, we show that turf‐algal communities release large quantities of DOM during the day as well as at night. The quality of this DOM for bacterioplankton differs substantially: Night exudates are utilized more rapidly, but result in a lower growth efficiency compared to day exudates. Estimations of historic and current DOC release further suggest that reef‐wide DOC release over a 24‐hr diurnal cycle doubled over the past 40 years, due to a strong increase in the abundance of BCMs, with night‐DOC release by BCMs and turf algae accounting for >50% of the total DOC release. 4.1 DOM release and diurnal carbon budget of turf algae The DOC release rate of turf algae observed in this study during daytime (4.6 ± 2.4 mmol C m −2 hr −1 ) is among the highest reported for coral reef benthic primary producers in general, including turf algae (Brocke, Wenzhoefer, et al., 2015 and references therein). Day‐DOC release rates of turf algae vary widely from 0.1 (Mueller et al., 2016 ) to 5.5 mmol m −2 hr −1 (Haas et al., 2010 ), which can be attributed to a multitude of potential factors, including community composition and density (biomass per surface area) of turf algae (Fricke et al., 2011 ; Mueller et al., 2016 ) as well as environmental conditions (e.g. light, temperature, nutrient levels; Haas et al., 2010 ; Mueller et al., 2016 ). Highest DOC release rates are typically reported during the summer months coinciding with maximum light intensities and seawater temperatures (e.g. Haas et al., 2010 ; Naumann et al., 2010 ; Roth et al., 2021 ). Thus, here reported turf‐algal DOC release rates measured in August might be interpreted as the summer maximum. In comparison to coral reefs with pronounced seasonality, reefs of the tropical island of Curaçao experience only mild seasonal fluctuations in environmental parameters (Haas et al., 2010 ; Roth et al., 2021 ; Vermeij & Bak, 2002 ). Hence, only minor seasonal variation in DOC release rates should be expected. Similar to DOC release rates, also the percentage of NPP released as DOC varies considerably for coral reef benthic primary producers, with values ranging from 4 to 51% (Brylinsky, 1977 ; Davies, 1984 ). The here reported 36% for turf algae is in the upper range of these percentages, yet considerably higher than found in earlier studies of turf algae (6%–22%; Haas et al., 2011 ; Haas et al., 2013 ; Mueller et al., 2016 ). The major loss terms within the diurnal carbon budget of turf algae are day‐DOC release (36% of NPP daytime ) and dark respiration (33% of NPP daytime ), but a considerably amount of C is also released as night‐DOC (21% of NPP daytime ; Figure 3 ). This relative night‐DOC release is merely half as high as reported for BCMs on Curaçao (Brocke, Wenzhoefer, et al., 2015 ). After subtracting all loss terms (i.e. dark respiration, day‐ and night‐DOC release), 10% of turf algae's NPP daytime remains available for other physiological processes, such as growth, maintenance and reproduction. This can be considered low to exert anabolic processes, since primary producers from coral reefs and other benthic ecosystems typically show a surplus of >60% of their daily fixed C (e.g. Abdullah & Fredriksen, 2004 ; Haas et al., 2013 ; Hatcher et al., 1977 ). This is in line with Brocke, Wenzhoefer, et al. ( 2015 ), the only study to date where a substantial night‐DOC release was reported, who found a mere surplus of 7% of NPP daytime after subtracting night respiration and day‐ and night‐DOC release. This raises the question how most turf‐algal communities and BCMs can be successful competitors, despite their low surplus carbon budgets? 4.2 Night‐DOM release by turf algae Under dark and anaerobic conditions several cyanobacterial taxa, including Oscillatoria spp. and other benthic taxa, are known to be capable of fermentation of stored carbohydrates, resulting in the release of small organic molecules such as lactate, ethanol and acetate (Heyer et al., 1989 ; Heyer & Krumbein, 1991 ; Moezelaar et al., 1996 ; Stal & Moezelaar, 1997 ). Oscillatoria spp. are widespread both in BCMs and turf‐algal communities of the Southern Caribbean (Brocke, Wenzhoefer, et al., 2015 ; Fricke et al., 2011 ). Indeed, Brocke, Wenzhoefer, et al. ( 2015 ) showed that BCMs dominated by Oscillatoria spp. can release substantial amounts of DOC at night. Analogously, cyanobacteria in turf‐algal communities may have released substantial amounts of DOC stemming from dark fermentation of carbohydrates stored in their cells. One might speculate that cyanobacterial members of the turf‐algal community may even ferment organic molecules extracted from decaying organic material (e.g. from dead corals or detritus being trapped within turf‐algal thalli and filaments; Figure 1b ) as suggested for BCMs by Brocke, Polerecky, et al. ( 2015 ). Indeed, several cyanobacteria are capable of heterotrophic growth on external organic carbon sources (Chojnacka & Marquez‐Rocha, 2004 ; Stal & Moezelaar, 1997 ; Yu et al., 2008 ), although anaerobic heterotrophic growth on external carbon sources seems limited to only very few cyanobacterial taxa (Richardson & Castenholz, 1987 ; Stal & Moezelaar, 1997 ). Furthermore, some chlorophytes are also capable of heterotrophic growth and dark fermentation (Atteia et al., 2013 ; Chojnacka & Marquez‐Rocha, 2004 ; Kim et al., 2013 ; Ueno et al., 1998 ) and may have also contributed to the DOM release by turf‐algal communities during nighttime. Our hypothesis that the DOC release at night is at least partly driven by fermentation is supported by a rapid shift from supersaturated to hypoxic conditions following the switch from light to dark conditions in the turf algae–water interface, creating the essential conditions for fermentation processes to occur (Figure 2 ). Great caution was taken to exclude the possibility that night‐DOM release was merely a stress‐induced artefact or the result of turf algae decaying. First, the photosynthetic efficiency of incubated turf algae measured with a PAM fluorometer did not change during night incubations. This suggests no occurrence of stress or deterioration of their physiological status during the incubation period (Enríquez & Borowitzka, 2010 ; Maxwell & Johnson, 2000 ). Moreover, if the increase in DOC concentration would have been cause by decaying turf algae, this could have caused a peak in DIN and PO 4 3 − . However, none of these concentrations increased during the night incubations (Table 2 ). And lastly, while pH slightly decreased during the night incubations, this change was expected as a result of respiration. The observed change in pH was less than pH changes typically occurring on coral reefs over a diurnal cycle (Anthony et al., 2011 ; Kleypas et al., 2011 ) and therefore does not indicate decomposition of turf‐algal tissue. 4.3 Quality of day and night turf‐algal exudates for bacterioplankton Given the presumed differences in the origin of turf‐algal DOM produced during the day and at night (i.e. the release of photosynthates during the day and fermentation products at night), it can be expected that both types of DOM differ in composition and therefore in nutritional quality to DOM‐feeding organisms. Indeed, bacterioplankton utilized night turf‐algal DOC at twice the rate compared to day‐DOC (Figure 5a ), but day‐DOM supported a two times higher bacterial growth rate compared to night‐DOM (Figure 5b ). In other words, bacterioplankton utilizes more turf‐algal DOC produced at night, but grows less on it compared to DOC released during the day. The bacterial growth efficiencies (BGE; Figure 5c ) infer that approximately 30% of the day turf‐algal DOC taken up by bacterioplankton was assimilated and used to synthesize bacterial biomass, whereas the remaining 70% was respired. In contrast, night turf‐algal DOC was almost entirely respired, leaving less than 10% to support new bacterial production. Bacterial growth efficiencies of bacterioplankton communities vary widely, ranging between 1% and up to 60% (Giorgio & Cole, 1998 and references therein). For exudates of coral reef benthic primary producers, BGEs appear to be at the lower end of this range, with fleshy macroalgae ranging between 6% and 2% and calcifying taxa higher with 16%–20% (Haas et al., 2011 ; Nelson et al., 2013 ; Silva et al., 2021 ). Night turf‐algal exudates in this study compare very well to previously reported BGEs of exudates of fleshy macroalgae. BGEs of day turf‐algal exudates, however, are even higher than those of calcifying taxa and are more similar to the reef water controls of these studies (27%). High BGEs are commonly associated with a high nitrogen content of the available substrates, to maintain a stable stoichiometry within bacterial cells (Fenchel & Blackburn, 1979 ; Goldman et al., 1987 ; Vallino et al., 1996 ). Accordingly, day turf‐algal DOM—showing a higher BGE than night‐DOM—was relatively enriched in N compared to the initial C:N ratio of DOM in the incubation water (Figure 4a ). In contrast, night turf‐algal DOM tended to be relatively depleted in N, even though this difference was not significant. Furthermore, during the course of the bioassays, DOC and DON were utilized in a balanced molar ratio for day turf‐algal exudates. For night exudates on the other hand, it appeared that relatively more DON was taken up (Figure 4b ), possibly indicating N‐limitation of the growing bacterioplankton communities and/or a higher bio‐availability of night‐DON. Observed differences in BGEs of day and night exudates could additionally be explained from a bioenergetic point of view. Assuming that night‐DOM originates from incomplete organic matter degradation and fermentation, resulting fermentation products (e.g. acetate, formate, lactate) are more oxidized than photosynthates (e.g. glucose, galactose, fucose) released during the day. Hence, the biologically available energy in night exudates might not be sufficient to reduce the available carbon in these substrates to the level of bacterial cell carbon (Linton & Stephenson, 1978 ). Consequently, those more oxidized substrates are incorporated into bacterial biomass at a lower efficiency (Figure 4c ), while carbon uptake is maximized to meet maintenance energy costs from increased respiration (Figure 4a ; Giorgio & Cole, 1998 ). 4.4 Historic increase in reef‐wide DOC release and possible consequences for ecosystem functioning The dramatic shifts in benthic community composition that occurred on many Caribbean reefs over the past five decades have likely caused changes in the quantity and quality of the produced DOM. For our study site, we estimated that the reef‐wide DOC release of the benthic community over a diurnal 24‐hr cycle has doubled between 1973 and 2013 (Figure 6 ). However, as turf‐algal abundance appears to be comparable at both time points and the appearance of macroalgae only added marginally to the DOC release, the here reported increase in reef‐wide DOC release was foremost caused by a tremendous increase in BCMs at the study site. In fact, when excluding the contribution of BCMs from our estimates, the reef‐wide DOC release appears to have decreased from 27.5 mmol m −2 day −1 in 1973 to 23.9 mmol m −2 day −1 in 2013. While BCMs have certainly become a substantial component of benthic communities on Curaçaoan reefs and throughout the Caribbean (de Bakker et al., 2017 ; Ford et al., 2018 ), their actual abundance can vary enormously on short temporal and spatial scales (Brocke, Polerecky, et al., 2015 ; Kornder et al., 2021 ). Here presented reef‐wide DOC fluxes from 2013 should therefore be considered as an example under very high prevalence of BCMs. Depending on the current abundance of BCMs, reef‐wide DOC release can range from rates slightly lower compared to 1973, to twice this rate. Concomitantly, the contribution of night‐DOC to the total diurnal 24‐hr DOC release can range between 30% and >50%. The general direction of change in benthic community composition from calcifying to non‐calcifying taxa at our study site and throughout the Caribbean (Gardner et al., 2003 ; Jackson et al., 2014 ) has likely led to a substantial change in the quality of the released DOM over the past decades. During the day, the DOM released by these non‐calcifying taxa stimulates the inefficient growth of opportunistic microbes (Haas et al., 2016 ). Large amounts of DOM need to be respired to support this growth, thereby shifting the microbial community metabolism from net autotrophy to net heterotrophy (Haas et al., 2013 ; Silveira et al., 2019 ). The lower BGE of algal‐DOM compared to coral‐DOM reduces the relative amount of energy and nutrients shunted into microbial biomass and thereby its transfer to higher trophic levels through the microbial loop (Haas et al., 2016 ; McDole et al., 2012 ). As turf‐algal DOM released at night has a four times lower BGE than turf‐algal DOM released during the day (Figure 5c ), night‐DOM is likely to considerably accelerate the microbialization in the water column. This will be of particular importance during periods of high prevalence of BCMs when night‐DOC released by turf algae and BCMs can contribute >50% to the reef‐wide DOC release (Figure 6 ). Night‐DOM release may also explain the high competitiveness reported for turf algae in coral–turf algae interactions (Barott et al., 2012 ; O'Brien & Scheibling, 2018 ; Vermeij et al., 2010 ). In general, algal‐DOM increases microbial respiration in coral–algae interfaces, creating prolonged hypoxia which can harm or even kill corals (Kline et al., 2006 ; Roach et al., 2020 ; Smith et al., 2006 ). Deceased corals can then be overgrown by their competitor (Barott & Rohwer, 2012 ; Dinsdale & Rohwer, 2011 ). As macroalgae primarily release DOM during the day (Haas et al., 2010 ; Maher & Eyre, 2011 ; Mueller et al., 2014 ), microbial respiration in coral–algae interfaces is supposed to be reduced dramatically, once the day‐DOM is respired and/or diluted in the course of the night. In contrast, turf algae containing cyanobacteria and/or chlorophytes continue to release DOM during the entire night, thereby prolonging the duration of hypoxia and its detrimental effect on coral health. Moreover, the aforementioned four times lower BGE of night‐ versus day‐DOM is likely to further amplify this effect. 4.5 Night‐time DOM release beyond coral reefs Turf‐algal communities containing cyanobacteria and/or chlorophytes capable of fermenting organic material and thereby potentially releasing DOM at night are not restricted to coral reefs but are major components in a variety of intertidal and subtidal rocky habitats ranging from tropical, temperate and even polar to latitudes (Connell et al., 2014 ). In addition, Oscillatoria mats and benthic chlorophytes are also found in marine and freshwater soft bottom habitats (Guiry & Guiry, 2018 ; Vadeboncoeur et al., 2020 ). To date, ecological research into dark fermentation by cyanobacteria and chlorophytes was largely focused on the energy‐providing mechanism enabling these organisms to survive and grow under dark and anaerobic conditions (Atteia et al., 2013 ; Stal & Krumbein, 1984 ; Ueno et al., 1998 ). Yet, the potential effect of the subsequent release of fermentation products as night‐DOM on community dynamics and ecosystem processes remains to be investigated. Given its high bio‐availability (Figure 5a ) and low BGE (Figure 5c ), night‐DOM release is likely to stimulate microbial community respiration at the expense of biomass generation, which may enhance the release of CO 2 from the system (Haas et al., 2013 ; Silveira et al., 2019 ). The relative importance of night‐time DOM release is likely to increase in the future as climate change has been predicted to promote the abundance of BCMs and turf‐algal communities (Christie et al., 2019 ; Pessarrodona et al., 2022 ) as well as the prevalence of cyanobacteria within turf‐algal communities (Bender et al., 2014 ; Ford et al., 2018 ; Paul, 2008 )."
} | 6,893 |
28389641 | PMC5429656 | pmc | 8,827 | {
"abstract": "Antibiotics are typically regarded as microbial weapons, but whereas their function at concentrations lethal for bacteria is often well characterized, the role of antibiotics at much lower concentrations as possibly found under natural conditions remains poorly understood. By using whole-transcriptome analyses and phenotypic screenings of the marine bacterium Phaeobacter inhibens we found that the broad-spectrum antibiotic tropodithietic acid (TDA) causes the same regulatory effects in quorum sensing (QS) as the common signaling molecule N-acyl-homoserine lactone (AHL) at concentrations 100-fold lower than the minimal inhibitory concentration against bacteria. Our results show that TDA has a significant impact on the expression of ~10% of the total genes of P. inhibens , in the same manner as the AHL. Furthermore, TDA needs the AHL associated LuxR-type transcriptional regulator, just as the AHL molecule. Low concentrations of antibiotics can obviously have a strong influence on the global gene expression of the bacterium that produces it and drastically change the metabolism and behaviour of the bacterium. For P. inhibens this includes motility, biofilm formation and antibiotic production, all important for settlement on new host-associated surfaces. Our results demonstrate that bacteria can produce antibiotics not only to antagonise other bacteria, but also to mediate QS like endogenous AHL molecules.",
"introduction": "Introduction Since the discovery of penicillin in 1929, more than 200 antibiotic drugs that are direct natural products have been discovered and their inhibitory bioactivity has been extensively studied 1 , 2 . Whereas the function of these antibiotics at concentrations lethal for bacteria is often well characterized, their role in natural environments at sub-inhibitory concentrations remains poorly understood 3 . In recent years it was proposed that antibiotics may not only serve as killing or growth-inhibitory compounds, but also function as inter-microbial signals even below the minimal inhibitory concentration 3 – 7 . Antibiotics at sub-inhibitory concentrations have been found to interfere with bacterial cell regulation systems, altering global transcription patterns and affecting up to 5–10% of all transcripts 4 , 8 – 10 , including regulators involved in quorum sensing (QS). QS is a cell-to-cell communication system mediated by signaling molecules inducing expression of defined sets of genes, e.g., encoding for energy-intensive processes like biofilm formation, cell motility and antibiotic production, and hypothesized to facilitate adaptations to environmental stimuli 11 – 13 . The addition of antibiotics to Chromobacterium violaceum cultures enhanced production of the QS signaling molecules N -acyl-homoserine lactones (AHLs). Because of the diverse molecular structures of antibiotics, however, the possibility that they act as AHL analogues was previously ruled out 6 . Our data demonstrate, instead, that the antibiotic tropodithietic acid (TDA) can act like the endogenous AHL in the marine bacterium Phaeobacter inhibens in triggering QS. \n Phaeobacter inhibens is a strong and competitive biofilm producer 14 , found to be associated with various eukaryotic organisms, and is a member of the Roseobacter group (family Rhodobacteraceae , Alphaproteobacteria ), which is highly abundant in marine habitats 15 , 16 . Phaeobacter inhibens as well as members of the genera Ruegeria and Pseudovibrio can produce TDA 17 , however, to this day, only a few TDA-tolerant strains have been found in the natural environment, co-occurring with TDA producers 18 . TDA has anticancer activities and the mode of action of TDA is similar to that of large polyether antibiotics, which work by disrupting the proton motive force 19 . Thus, there is an increasing interest in TDA as an antibiotic and in Phaeobacter as a probiotic organism in aquacultures and other applications 20 – 22 . Production of TDA has been described for various genera, able to increase TDA gene expression by cross-feeding, suggesting that TDA functions as an autoinducer of its own synthesis 23 . Presence of genes coding for diguanylyl cyclases and c-di-GMP (cyclic dimeric guanosinmonophosphate)-specific phosphodiesterases in P. inhibens and other roseobacters suggests that c-di-GMP signalling is a universal feature of this group, and TDA production in Ruegeria mobilis was found to be influenced by intracellular c-di-GMP concentrations 24 . It was suggested that in the Roseobacter clade intra and extracellular cues are integrated via a c-di-GMP second messenger system and that expression of phenotypic traits specific for either planktonic or attached life is regulated in response to c-di-GMP concentrations 24 . Biosynthesis of TDA in P. inhibens is regulated by AHL-mediated QS based on a luxIR homologous system with pgaI encoding for the N-3-hydroxydecanoylhomoserine lactone (3OHC(10)-HSL) synthase, and pgaR for the corresponding AHL response regulator 17 . Mutants of these genes ( pgaI \n − and pgaR \n − ) were deficient in TDA production. TDA at sub-inhibitory concentrations restored TDA gene expression in pgaI \n − , but not in pgaR \n − , confirming the role of TDA as an autoinducer 17 . This prompted us to consider whether TDA regulates more than its own biosynthesis. Using whole-transcriptome analyses and phenotypic screenings, we show that gene expression and gene functions are mediated by both TDA and AHL, and demonstrate that production of TDA in P. inhibens can substitute production of its AHL.",
"discussion": "Discussion As pharmaceutical drugs, antibiotics are typically used in high concentrations to kill pathogenic bacteria, and research has focused primarily on the killing and growth-inhibitory function of antibiotics. Only in recent years antibiotics were shown to interfere with bacterial cell-to-cell communication, yet, how antibiotics affect differential bacterial responses has not been clarified 3 – 6 . Pseudomonas aeruginosa and related bacteria produce quinolones, some of which exhibit antimicrobial activity, however, some quinolones act as quorum-sensing signal molecules, controlling the expression of many virulence genes as a function of cell population density 30 . Antibiotic activity was also found for (7Z)-C14:1-AHL already in the nanomolar range 31 , but no antagonism of the producing bacteria was observed in inhibition assays and this has so far not been studied further. The data presented here provide to the best of our knowledge the first evidence of an antibiotic that regulates QS like an AHL molecule and demonstrate that bacteria can produce antibiotics that substitute their AHL molecules as global gene regulators. Our finding that a sub-inhibitory concentration of exogenous TDA is able to mediate QS and to restore WT gene expression only in conjunction with the AHL regulator confirmed that TDA interacts specifically with the LuxR-type transcriptional regulator. How exactly TDA and the AHL regulator interact and co-regulate gene expression remains to be determined. The AHL-based QS is obviously located at a high level in the regulation system of P. inhibens , controlling the expression of other signaling molecules. For example, five of seven genes associated with c-di-GMP signaling 24 , including three diguanylate cyclases, were negatively regulated by AHL and TDA (Table S1 ). In our experimental set-up, results based on WT cells represent the active QS system, as cells were harvested from cultures with high cell numbers, reaching the AHL-threshold for QS. In contrast, the AHL mutants simulate an inactive QS system equivalent to gene expression of single cells or low cell numbers 17 . The AHL- and AHL regulator-deficient strains were highly motile and produced strong biofilms, whereas motility and attachment were reduced in the WT (Fig. 3 , Fig. S3 ). This finding reflects the fact that genes involved in motility, chemotaxis and biofilm formation are more highly expressed in single cells with no activated QS to enhance their ability to find new surfaces, attach and form biofilms (“swim-and-attach”). The importance of motility for the surface attachment was reported for another roseobacter, Silicibacter sp. TM1040, by Miller and Belas 32 . This strategy likely provides an advantage for Phaeobacter spp. to settle on abiotic and host-associated surfaces, explaining its association with diverse marine eukaryotes, including seahorses, scallop larvae and algae 15 , 16 , 33 . Down-regulation of motility genes at high population densities by QS was also observed in another alpha-proteobacterium, Sinorhizobium meliloti , and it was hypothesized that attached cells do not need high motility and can thereby save energy 34 . Based on our data, we assume that P. inhibens grows on a surface until QS is activated (by the AHL and TDA), whereupon biofilm-associated genes are down-regulated, resulting in considerably reduced attachment. In Ruegeria sp. KLH11, another member of the Roseobacter group, biofilm dispersion is also induced by AHLs. It has been proposed that this prevents aggregation on crowded surfaces and promotes a balanced colonization of host organisms 35 . Dispersal rates play a key role in determining diversity and function over evolutionary timescales 36 . Therefore, our results point to another explanation, suggesting that Phaeobacter is “hitchhiking” on migrating hosts like phytoplankton cells, e.g., the cosmopolitan diatom Thalassiosira rotula , on which Phaeobacter was previously detected 37 , to access new resources. This can be an important dispersal mechanism, which could help to explain the broad geographic distribution of P. inhibens \n 15 , 36 . Another reason why QS reduces the attachment of a surface colonizer relates to the “Jekyll-and-Hyde” chemistry of P. inhibens suggested by Seyedsayamdost et al . 38 , describing a shift from mutualism to pathogenesis through the production of algaecides against its algal host. Killing of the aging host provides rapid access to resources released from the algal cell during its lysis, and QS might induce dispersion and re-association with a new, healthy host. The algaecidal precursors phenylacetyl-CoA and tropolone are also precursors and intermediate products for the biosynthesis of TDA 15 . Based on our finding that TDA mediates QS in P. inhibens , we propose that QS can mediate the switch from mutualism to pathogenesis and from attachment to dispersion. That the TDA negative mutant produced more biofilm than the WT (Fig. S3 ) is probably based on the lacking TDA as QS signaling molecule and demonstrates a role of TDA also in biofilm dispersion. The capability to produce the antibiotic TDA regulated by QS on the one hand, which mediates QS like AHLs on the other hand, implies that in P. inhibens AHL and TDA act simultaneously, probably accelerating QS and adaptation processes, for example those relevant in symbiosis. Furthermore, an organism which produces an antibiotic acting as weapon and as a signaling molecule saves on biosynthesis costs 39 . Since TDA autoinduces its own synthesis in different Roseobacters by cross-feeding 23 , it is likely to act in these and possibly other organisms as a global signal for QS. The probiotic activity of Phaeobacter spp. was attributed to TDA production 40 , however, future studies now have to show whether the probiosis is based on the antibiotic activity of TDA only or also its signaling function. The results presented here, together with prior findings that other antibiotics can interfere with cell signaling 4 , 8 – 10 , indicate that antibiotics at sub-inhibitory concentrations may have considerable impact on the microbiome in natural environments. The widespread use of antibiotics in human medicine, agriculture and aquaculture 2 , 7 , 41 , 42 might have already changed the microbial world through gene expression regulation by antibiotic signaling. The bulk of the actual impact of antibiotics in the environment has not yet been recognized, because of a lack of understanding and consideration for the role of antibiotics at sub-inhibitory concentrations in QS and, potentially more generally, in inter-microbial signaling."
} | 3,069 |
23900844 | PMC3772215 | pmc | 8,828 | {
"abstract": "Background: Photosystem II is an essential component of oxygenic photosynthesis. Results: Photosystem II is specifically decreased in rubredoxin mutants of the green alga Chlamydomonas reinhardtii , the cyanobacterium Synechocystis sp. PCC 6803, and the plant Arabidopsis thaliana . Conclusion: Rubredoxin is required for photosystem II, and not photosystem I, accumulation in these organisms. Significance: Rubredoxin was likely important in the evolution of oxygenic photosynthesis.",
"introduction": "Introduction Photosystem II (PSII) 3 uses absorbed light energy to catalyze the splitting of water into molecular oxygen and protons and is the only enzyme complex known to be capable of this energetically unfavorable process. Working together with photosystem I (PSI), the cytochrome b 6 f complex and the ATP synthase, PSII drives oxygenic photosynthesis and is responsible for producing most, if not all, of the oxygen present in the Earth's atmosphere ( 1 , 2 ). Recent work has further refined our knowledge of the structure ( 3 ) and biogenesis ( 4 – 11 ) of PSII, but a complete understanding of how the 20–30 protein subunits and ∼70 cofactors are correctly assembled into a functional light-driven water:plastoquinone oxidoreductase remains elusive ( 12 – 14 ). The structural and functional homology between PSII and the pheophytin-quinone (type II) reaction center of anoxygenic photosynthetic bacteria, along with many geological and biological lines of evidence, indicates that these two complexes share a common evolutionary origin and that this ancestral complex was almost certainly anoxygenic ( 15 ). Despite the availability of crystal structures for both reaction centers ( 3 , 16 ), knowledge of their common ancestor is remarkably limited and therefore the evolutionary inventions and innovations that enabled a proto-PSII to use water as an electron donor (and to concomitantly release molecular oxygen) are poorly understood. Rubredoxins are [1Fe-0S] proteins in which one iron atom is coordinated by four cysteine residues ( 17 , 18 ). Most studied rubredoxins exist as small soluble proteins that act as electron carriers in a variety of biochemical processes including carbon fixation ( 19 ), detoxification of reactive oxygen species ( 20 – 22 ), and fatty acid metabolism ( 23 , 24 ). The distribution of rubredoxins within the tree of life is distinct in that these proteins appear to be limited to divergent groups within the archaea and bacteria, whereas they are found in only one group of eukaryotes: plants and photosynthetic algae that contain a plastid derived ultimately from primary endosymbiosis of a cyanobacterium. Interestingly, this eukaryotic “photosynthetic” rubredoxin is thylakoid membrane-associated and, together with the homologous protein found in cyanobacteria, appears to represent a class of rubredoxins unique to the oxygenic photoautotrophs that is distinct from all other known rubredoxins ( 25 , 26 ). Previous studies of thylakoid membrane rubredoxins have yielded conflicting results. Analysis of a mutant strain of the cyanobacterium Synechococcus sp. PCC 7002 lacking rubA , the membrane-bound rubredoxin, led to the conclusion that the function of this protein was to aid in iron-sulfur cluster assembly of PSI and that it had little or no effect on PSII ( 26 , 27 ). Surprisingly, however, an antibody raised against the membrane-bound rubredoxin from the photosynthetic cryptophyte Guillardia theta was found to react with a homologous protein in PSII-enriched particles from spinach ( 28 ). Here we report the characterization of mutants lacking the thylakoid membrane-associated rubredoxin in the green alga Chlamydomonas reinhardtii, the cyanobacterium Synechocystis sp. PCC 6803 and the flowering plant Arabidopsis thaliana . We show that these mutants exhibit a PSII-specific defect and that the role of this rubredoxin in contributing to the assembly or stability of PSII is likely conserved in oxygenic photoautotrophs.",
"discussion": "DISCUSSION Several lines of evidence show that the thylakoid-associated rubredoxin encoded by RBD1 is necessary specifically for PSII activity in Chlamydomonas . Without RBD1, the 2pac mutant does not accumulate PSII as assayed by chlorophyll a fluorescence ( Fig. 1 A ), immunodetection of PSII subunits ( Fig. 1 B ), and measurements of ΔA 520 nm ( Fig. 1 D ). PSII accumulation is clearly restored via transformation of the 2pac mutant with a small fragment of DNA containing only RBD1 , its putative promoter and its 3′ UTR ( Fig. 1 ). The mature RBD1 protein is small (∼16 kDa) and is a membrane-associated rubredoxin ( Fig. 1 B ). Membrane-bound rubredoxins are found exclusively in PSII-containing organisms and are distinct from the well-studied soluble rubredoxins found in many archaea and bacteria ( Fig. 3 A ). The PSII-specific defect of Synechocystis ( Fig. 4 ) and Arabidopsis ( Fig. 5 ) mutants lacking a RBD1 homolog indicates that the role of RBD1 in PSII assembly or stability is broadly conserved in a diverse group of oxygenic photoautotrophs. The residual amount of PSII in the cyanobacterial mutant compared with the complete absence in eukaryotic mutants is consistent with previously observed mutants in PSII assembly factors ( 4 , 7 , 9 ) and is thought to be due, at least in part, to more efficient quality control mechanisms in plastids than in their cyanobacterial relatives ( 4 , 57 ). Interestingly, PSII activity as assayed by variable fluorescence is lower than in the wild-type but apparently sufficient to support photoautotrophic growth ( Fig. 4 B ), providing an opportunity for future studies into the precise functional role of RBD1 in the proper biogenesis of PSII. Our results are in agreement with previous work showing that the membrane-bound rubredoxin is present in spinach PSII preparations ( 28 ). However, our findings with Chlamydomonas , Synechocystis , and Arabidopsis differ from reports on a rubA mutant of the cyanobacterium, Synechococcus sp. PCC 7002, which lacked PSI but not PSII activity due to a defect in iron-sulfur cluster assembly ( 26 , 27 ). This discrepancy could simply be due to a difference in RubA function in different cyanobacterial species, despite the conserved genomic location of the rubA gene next to five other genes involved in PSII function ( Fig. 3 B ). In Chlamydomonas and Arabidopsis , there appear to be multiple chloroplast-localized rubredoxins, so it is possible that one of these other homologs functions in PSI assembly. However, Chlamydomonas RBD1 (and its Arabidopsis and Synechocystis homologs encoded by At1g54500 and rubA, respectively) does not seem to play a major role in PSI assembly as evidenced by measurements of Δ A 520 nm ( Fig. 1 D ), P700 redox changes ( Figs. 1 E , 4 D , and 5 C ), as well as PsaA, PsaC and PsaD protein accumulation ( Figs. 1 B , 4 C , and 5 B ). Because PSII activity is fully restored in the complemented lines gRBD1–1 and gRBD1–2 despite relatively low levels of RBD1 protein accumulation ( Fig. 1 ), RBD1 is unlikely to be a subunit of PSII, rather it may have a catalytic, substoichiometric role in promoting PSII assembly or stability. This is further supported by the detection of wild-type levels of RBD1 in the Fud7 mutant lacking the D1 protein ( Fig. 1 B ). It is known that rubredoxins participate in electron transport reactions in some Archaea and bacteria. Indeed, the midpoint redox potential of the RBD1 homolog from the cryptophyte alga Guillardia theta was found to be ∼ +125 mV ( 25 ), a value that could enable it to participate in electron transport with, or perhaps bypassing, plastoquinone (E m ∼ +100 mV, ( 1 )). In fact, a similar role has been suggested for a pair of flavodiiron proteins that function in the photoprotection of PSII in Synechocystis ( 58 ). Rubredoxins have also been previously described as aiding in oxygen tolerance, either by reacting with reactive oxygen species directly or by helping to maintain the appropriate redox state of Fe-containing active sites in some enzymes ( 20 – 22 ). PSII, which generates oxygen and contains both a redox-active non-heme Fe as well as a redox-active heme (cytochrome b 559 ), would certainly fit the profile of a protein complex that might benefit from the presence of such an enzyme. In many microbes, the active site of superoxide reductase (SOR) contains a non-heme iron that, after catalyzing the reduction of superoxide to hydrogen peroxide, is thought to be reactivated via rubredoxin-mediated re-reduction ( 59 , 60 ). Cytochrome b 559 has recently been shown to have superoxide reductase and oxidase activity ( 61 ) and the presence of RBD1/rubA in a highly conserved gene cluster with the two subunits of this transmembrane cytochrome might also be taken to suggest an as yet unidentified interaction between these proteins. Another possible clue regarding the function of RBD1 may be found by examination of the synthesis of the membrane-bound [NiFe] uptake hydrogenase. The gene cluster required for synthesis of this hydrogenase is well conserved among anaerobic and aerobic species, but aerobic species with this enzyme contain an additional operon in which one of the genes ( hupI or hoxR ) encodes a rubredoxin ( 20 , 22 ). These rubredoxins appear to contribute to oxygen tolerance of the hydrogenase, and it is of note that some of these species are photosynthetic purple bacteria that contain type II reaction centers, which are believed to be similar to the ancestral state of PSII ( 1 ). Although they are not required for assembly of anoxygenic type II reaction centers in bacteria, it is tempting to speculate that these types of soluble rubredoxins might be representative of the ancestral state of the rubredoxin motif and that the membrane-bound rubredoxin found in the oxygenic phototrophs might be an evolutionary innovation co-opted from a soluble ancestral protein during the transition from anoxygenic to oxygenic photosynthesis."
} | 2,508 |
35445473 | PMC9325437 | pmc | 8,829 | {
"abstract": "Abstract Diseases in marine eukaryotic organisms caused by opportunistic pathogens represent a serious threat to our oceans with potential downstream consequences for ecosystem functioning. Disease outbreaks affecting macroalgae are of particular concern due to their critical role as habitat‐forming organisms. However, there is limited understanding of the molecular strategies used by macroalgae to respond to opportunistic pathogens. In this study, we used mRNA‐sequencing analysis to investigate the early antipathogen response of the model macroalga Delisea pulchra (Rhodophyta) under the environmental conditions that are known to promote the onset of disease. Using de novo assembly methods, 27,586 unique transcripts belonging to D . pulchra were identified that were mostly affiliated with stress response and signal transduction processes. Differential gene expression analysis between a treatment with the known opportunistic pathogen, Aquimarina sp. AD1 (Bacteroidota), and a closely related benign strain ( Aquimarina sp. AD10) revealed a downregulation of genes coding for predicted protein metabolism, stress response, energy generation and photosynthesis functions. The rapid repression of genes coding for core cellular processes is likely to interfere with the macroalgal antipathogen response, later leading to infection, tissue damage and bleaching symptoms. Overall, this study provides valuable insight into the genetic features of D . pulchra , highlighting potential antipathogen response mechanisms of macroalgae and contributing to an improved understanding of host–pathogen interactions in a changing environment.",
"introduction": "1 INTRODUCTION Eukaryotic organisms in the marine environment are constantly exposed to a diverse set of microorganisms ranging from beneficial symbionts to harmful pathogens. Bacterial pathogens are widely regarded as a significant threat to the health of marine organisms, having downstream consequences on ecosystem health. In recent years, there has been a rise in the number of reports of disease in marine eukaryotic hosts and there is evidence to suggest the involvement of opportunistic bacterial pathogens as aetiological agents (Burge et al., 2013 ; Egan & Gardiner, 2016 ). Opportunistic bacterial pathogens can exist as commensal organisms in the holobiont, but under certain host or environmental conditions may rapidly colonize and exploit host resources, resulting in tissue damage and disease symptoms (Burge et al., 2013 ; Casadevall & Pirofski, 2000 ). However, there is limited understanding of how marine eukaryotic hosts respond to these opportunistic bacteria under the conditions that promote the onset of disease. Of particular concern are disease outbreaks affecting macroalgae, which are being increasingly reported across both natural and farmed populations (Gachon et al., 2010 ; Ward et al., 2019 ). Macroalgae are critical for biodiversity where they provide food and habitat and contribute to primary production in temperate coastal ecosystems. Disease outbreaks can therefore have devastating and far‐reaching consequences for community functioning, leading to the long‐term decline of coastal ecosystems (Christie et al., 2009 ; Schiel & Lilley, 2011 ). To date, research surrounding diseases in macroalgae has predominantly focused on identifying pathogens and their virulence traits (Egan et al., 2014 ; Fernandes et al., 2011 ; Gardiner et al., 2015 ; Hudson et al., 2018 , 2019 ; Schroeder et al., 2003 ). Some studies have also investigated the antipathogen response strategies of the host in relation to viruses, oomycetes and algal endophytes (Im et al., 2019 ; Strittmatter et al., 2016 ; Tang et al., 2019 ; Xing et al., 2021 ), but there remains a paucity of knowledge around the response strategies of the host towards bacterial pathogens. Where research has explored the molecular response of algae to bacterial pathogens, evidence for the role of pattern recognition receptors and intracellular signalling cascades has been found (Cosse et al., 2007 ; de Oliveira et al., 2017 ; Weinberger, 2007 ). These immune processes, which are similar to those in higher plants, may assist in pathogen detection and the activation of downstream immune signalling pathways. While these studies have contributed to our understanding of algal defences, we still know comparatively very little about host–pathogen interactions on the molecular level. Bleaching disease in the red macroalga, Delisea pulchra (Bonnemaisonales, Rhodophyta), is one of the few well‐characterized examples of disease in macroalgae. This disease is caused by a bacterial infection (Case et al., 2011 ; Kumar et al., 2016 ) leading to the loss of photosynthetic pigments along the midthallus, resulting in tissue necrosis, reduced fecundity and increased herbivory (Campbell et al., 2014 ). Furthermore, the occurrence of bleaching disease is highly correlated with increased seawater temperatures in the summer months, which is thought to reduce the natural chemical defences of the alga and therefore render it more susceptible to microbial infections (Campbell et al., 2011 ). Previous investigations have found that members of the Flavobacteriaceae are enriched on bleached D . pulchra (Zozaya‐Valdes et al., 2017 ), and the bacterium Aquimarina sp. AD1 (Flavobacteriaceae, Bacteroidota) has been identified as a causative agent of bleaching disease in D . pulchra . Aquimarina sp. AD1 is considered an opportunistic pathogen as it can be found on healthy individuals and only appears to cause disease symptoms on D . pulchra under conditions of induced thermal stress (Kumar et al., 2016 ). Genome analysis of Aquimarina sp. AD1 also identified genes involved in the nutritional acquisition of algal cell‐wall components that may function as virulence traits, supporting its role as an opportunistic pathogen (Hudson et al., 2019 ). The emergence of opportunistic pathogens represents a threat to marine ecosystems which is expected to intensify in the future with growing anthropogenic pressures. Thus, D . pulchra and Aquimarina sp. AD1 are an ideal model for investigating host–pathogen dynamics in macroalgae. Here we aim to characterize the response of D . pulchra using mRNA‐sequencing (mRNA‐seq) analysis after treatment with the opportunistic pathogen, Aquimarina sp. AD1, under the environmental conditions that are known to promote the onset of disease. We further compared this gene expression response to that of a closely related bacterial strain, Aquimarina sp. AD10, that does not cause disease in D . pulchra (Kumar et al., 2016 ) in order to assess the specific antipathogen response of the macroalga.",
"discussion": "4 DISCUSSION Infectious disease outbreaks in marine organisms are being increasingly reported (Tracy et al., 2019 ), but few studies have assessed the antipathogen response of marine hosts against opportunistic pathogens, limiting our ability to protect important species in the future. Here, we first described the transcriptomic profile of the model macroalga Delisea pulchra under the environmental conditions known to promote the onset of bleaching disease. We then compared the transcriptional response of D . pulchra to an opportunistic pathogen ( Aquimarina sp. AD1) relative to a benign nonpathogen (AD10) to characterize the antipathogen response mechanisms of this macroalga. 4.1 The transcriptome of Delisea pulchra reveals its potential to detect and respond to pathogens As sessile organisms macroalgae must rapidly activate cellular stress response mechanisms when subjected to pathogen stress to mitigate damage and maintain cellular functioning. We found that genes encoding cellular stress response pathways featured prominently in the D . pulchra transcriptome (Figure 2 ), including those annotated as heat shock proteins, antioxidant enzymes and intracellular signal transduction proteins (Table S2 ). Heat shock proteins and antioxidant enzymes have a critical role in the algal immune response by preserving the integrity of cellular structures exposed to pathogen effector molecules and the algal oxidative defences (Cosse et al., 2009 ; Khan et al., 2018 ; Strittmatter et al., 2016 ). Oxidative defence in algae has also been linked to halogen metabolism, and 55 genes encoding haloperoxidase enzymes were identified in the D . pulchra transcriptome. Haloperoxidase enzymes remove H 2 O 2 via the oxidation of halide ions to form halogenated compounds and can aid in the oxidative stress response by directly removing H 2 O 2 (Cosse et al., 2007 ; La Barre et al., 2010 ). Halogenated compounds are also known to function as chemical defence molecules in macroalgae (Cosse et al., 2007 ; Ohsawa et al., 2001 ; Paul et al., 2006 ), including the brominated furanones produced by D . pulchra (de Nys et al., 1993 ). The observed number of transcripts for haloperoxidases suggests future studies will benefit from a closer evaluation of the role of halogenated compounds in D . pulchra's defence against bacterial pathogens, including Aquimarina sp. AD1. The transcriptome of D . pulchra was also rich in transcripts associated with intracellular signal transduction pathways known to coordinate the cellular response to external stress (Figure 2 ). In particular, inositol phosphate metabolism, which is linked to phosphatidylinositol signalling, plays a role in heat stress in plants (Hou et al., 2016 ; Liu et al., 2006 ) and algae (Zhang et al., 2020 ) and has recently been shown to be positively associated with the antipathogen response of the alga Pyropia yezoensis (Khan et al., 2018 ). Although signalling pathways remain relatively uncharacterized in macroalgae, in D . pulchra inositol phosphate metabolism and signalling may similarly play a role in conferring tolerance to external stressors. Genes coding for other signalling pathways, such as calcium signalling proteins and mitogen activated protein kinase (MAPK) cascades, were also expressed (Table S2 ). MAPK cascades are key pathways regulating the antipathogen response via pathogen receptor proteins (R‐proteins) in plant systems (Pitzschke et al., 2009 ; Zhang et al., 2018 ), including red algae (de Oliveira et al., 2017 ; Khan et al., 2018 ). Although further work is required to verify their role in pathogen recognition in algae, genes encoding domains associated with non‐NBS‐LRR algal R‐proteins (Collén et al., 2013 ; de Oliveira et al., 2017 ; Tang et al., 2019 ) were identified in D . pulchra (Table S2 ). Overall, the detection of transcripts involved in stress tolerance and signalling in the transcriptome of D . pulchra (Figure 2 ) suggests that it has the capacity to detect and respond to pathogen stress under the conditions of elevated temperature in which these experiments were performed. 4.2 Pathogen exposure leads to a repression of genes involved in core cellular functions known to dampen the immune response We observed the highest number of DEGs occurred in response to the AD1 treatment, relative to both the control and AD10 treatments after 24 h of exposure (Figure 3a ). This early transcriptional response of D . pulchra to a pathogen is in agreement with other studies of Rhodophyta which observed the antipathogen response to be maximal after 24 h, followed by a decrease at 48 h (de Oliveira et al., 2017 ). At 48 h of exposure only one DEG was detected between the pathogen (AD1) and non‐pathogen (AD10) treatments. These results highlight the dynamic nature of gene expression regulation in D . pulchra and suggest that differences in host gene expression during the initial exposure to an opportunistic pathogen are key to determining disease outcomes for the host (Kumar et al., 2016 ). To understand the pathogen‐response mechanisms of D . pulchra , we further explored the pathogen‐specific differential gene expression occurring between the AD1 and AD10 treatments after 24 h. Here, seven genes were upregulated in response to AD1 when compared to AD10, including a gene coding for a voltage‐gated calcium channel (Table 3 ) which may be important for eliciting downstream immune responses via calcium signalling pathways and MAPK cascades (Wurzinger et al., 2011 ). Likewise, a gene coding for a voltage‐gated calcium channel was upregulated in the AD1 treatment compared to the control, suggesting that the upregulation of this gene may be in response to pathogen‐specific traits of Aquimarina sp. AD1. The upregulation of a gene involved in halogen metabolism (Table 3 ) was also observed in D . pulchra after exposure to AD1 and may act to degrade brominated furanones (Kunka et al., 2018 ). Halogen defence compounds are considered a key defence mechanism in algae (Case et al., 2011 ; Nylund et al., 2009 ; Paul et al., 2006 ; Strittmatter et al., 2016 ) and the results here suggest that Aquimarina sp. AD1 may target these to evade the algal immune defences. Interestingly, brominated furanones are also known to antagonize bacterial quorum sensing (cell–cell communication) systems (Manefield et al., 1999 ), which are known to regulate the virulence phenotype in some bacteria. Genome analysis of Aquimarina sp. AD1 identified genes coding for putative quorum sensing functions that were absent in Aquimarina sp. AD10 (Hudson et al., 2019 ). Therefore, suppression of the chemical‐based defence responses of D . pulchra would not only result in an increased susceptibility of the alga to infection but may also enhance the virulence of Aquimarina sp. AD1. In contrast to previous studies of plant and algal antipathogen responses, the majority of DEGs between both the AD1 and AD10 treatments and the AD1 and control treatments were downregulated at 24 h (Table S4 ). A large proportion of these DEGs were related to protein synthesis and turnover functions (Figure 3c ) including ubiquitin‐mediated protein degradation, heat shock proteins and translation (Table 3 ). The ubiquitination system is widely recognized as a key component of the plant immune response (Craig et al., 2009 ; Delauré et al., 2008 ; Dielen et al., 2010 ; Trujillo & Shirasu, 2010 ) where it plays a role in tagging damaged proteins for removal (Vierstra, 2009 ). The repression of this system in D . pulchra is therefore predicted to lead to an excess of denatured proteins, impairing homeostasis and harming metabolic functioning. Moreover, in addition to their role in ubiquitination, some E3 ligases can also play a regulatory role in either initiating or repressing the plant immune response (Furniss et al., 2018 ; Marino et al., 2013 ; Ramonell et al., 2005 ). Together with the repression of the ubiquitination pathway, it is possible that the repression of two E3 ligases in D . pulchra results in an overall weakened immune response. The expression of several heat shock protein genes was also reduced in pathogen‐treated D . pulchra (Table 3 ). This observation is in contrast to studies in algae that show the upregulation of heat shock proteins in response to pathogen exposure (Khan et al., 2018 ; Strittmatter et al., 2016 ). Moreover, loss‐of‐function hsp90 mutants in Arabidopsis thaliana have been shown to inhibit the hypersensitive response and reduce R‐protein levels (Hubert et al., 2003 ), highlighting a specific role of Hsp90 in the maintenance of the plant immune response. In D . pulchra , reduced levels of hsp90 transcripts may therefore hinder the correct initiation of immune responses, resulting in an increased susceptibility to infection by opportunistic bacteria. Pathogen‐induced changes in the expression of D . pulchra genes involved in protein turnover also corresponded to decreased expression of the plastid translation elongation factor Tu (EF‐Tu) and three elongation factor subunit alpha (eEF1A) homologues, signifying an overall decrease in protein synthesis. The role of eEF1A has not been broadly characterized in plants or algae, but in yeast models the depletion of eEF1A resulted in the repression of protein synthesis (Kim & Coulombe, 2010 ). In mammalian systems, eEF1A also functions as an activator of hsp70 gene expression (Vera et al., 2014 ) and may therefore account for the transcriptional repression of hsp70 observed in this study. Moreover, the activity of eEF1A is dependent on an intact actin cytoskeleton (Gross & Kinzy, 2005 ), and thus the downregulation of two cytoskeleton components, actin and alpha‐tubulin (Table 3 ), may similarly affect translation efficiency in D . pulchra . Interestingly, EF‐Tu also acts as a chaperone and in plants it was shown to aid cellular acclimation to heat stress (Caldas et al., 2000 ; Fu et al., 2012 ; Li et al., 2018 ). Reduced expression of these key translation factors may result in a reduced capacity for protein turnover in D . pulchra and further increase its susceptibility to opportunistic pathogens such as Aquimarina sp. AD1. Further investigations should focus on evaluating the broader impact of translation factors and protein synthesis in the algal immune response. A reduced capacity for D . pulchra to mount an effective immune response in the presence of AD1 is further supported by the reduced expression of genes encoding regulatory functions. This included the downregulation of the transcription factor MYB and the small GTPase Rac1, which are known to modulate the expression of pathogen resistance genes and co‐ordinate defence following the detection of a pathogen in plants (Ambawat et al., 2013 ; Ono et al., 2001 ; Raffaele & Rivas, 2013 ). Previous studies in red algae have also reported the upregulation of MYB and Rac1 following pathogen exposure (de Oliveira et al., 2017 ), Therefore, suppression of these genes in D . pulchra may directly interfere with the initiation of an effective immune response. These contrasting findings between the current study and de Oliveira et al. ( 2017 ) may be due to the different environmental conditions under which these experiments were performed. Thus, investigating the effect of environmental conditions on the activity of regulatory proteins represents an avenue for future research. In response to the pathogen, genes coding for components of the photosynthetic pathway were downregulated in D . pulchra including parts of the Calvin cycle (e.g., Transketolase‐1) and multiple components of the photosynthetic apparatus (Table 3 ). These findings are consistent with what has been observed in plants and other algae, whereby photosynthesis is often repressed in response to pathogen exposure to allow for cellular energy to be re‐allocated towards defence (Bilgin et al., 2010 ; Cohen & Leach, 2019 ; Khan et al., 2018 ). However, genes important for cellular respiration (e.g., cytochrome C oxidase subunit [Cox2] and an ATP synthase subunit beta protein) were also downregulated, potentially restricting the availability of ATP in the cell. The downregulation of both photosynthesis and cellular respiration may therefore have a global effect on the cellular energy balance in D . pulchra , impacting protein turnover, transcription and translation, and further hindering the algal defence mechanisms. Collectively, the analysis of DEGs suggests that the antipathogen response of D . pulchra involves the repression of genes associated with a broad range of cellular functions (Figure 4 ). The overall downregulation in the expression of genes coding for protein and energy metabolism functions is hypothesized to compromise host cellular functioning, ultimately rendering the macroalga vulnerable to infection by opportunistic pathogens. Whilst such interactions have not been previously described in red algae, similar negative health outcomes are commonly reported in plants simultaneously exposed to heat and pathogen stress (Huot et al., 2017 ; Pandey et al., 2015 ; Prasch & Sonnewald, 2013 ; Wang et al., 2009 ). Although the effect of temperature was not specifically investigated here, this observation leads us to postulate that elevated temperature conditions may also influence host pathogen dynamics in macroalgae, promoting the onset of disease. The current study was conducted at a peak summer temperature for Sydney coastal waters (Australian Bureau of Meteorology), and such peak temperatures are predicted to become more frequent due to climate change, placing macroalgae at an increased risk of disease. Therefore, our understanding of macroalgal disease would greatly benefit from further studies assessing the underlying effects of environmental conditions on the interaction between D . pulchra and Aquimarina sp. AD1. FIGURE 4 Schematic diagram of the cellular response of Delisea pulchra following exposure to the pathogen Aquimarina sp. AD1. Proteins and functions encoded by differentially expressed genes are represented in light blue, with upregulated genes shown in dark blue. Here we propose that cell‐wall‐degrading enzymes produced by AD1 degrade the cell wall of D . pulchra (see Hudson et al., 2019 ). Cellular damage (red squares) is detected by D . pulchra R‐proteins, triggering an influx of calcium ions into the cell, and eliciting an antipathogen response via calcium signalling and MAPK cascades. However, reduced inositol triphosphate levels (InP3) are hypothesized to interfere with calcium signalling and downstream antipathogen responses. Downregulation of the transcription factor MYB and the small GTPase Rac1 would also probably interfere with the expression of genes involved in defence. Suppression of inositol phosphate signalling may also interfere with thermal stress resistance. Elevated temperatures would probably promote an increase in the concentration of denatured proteins (grey lines), which accumulate in the cell due to downregulation of the ubiquitin (Ub)‐mediated protein degradation pathway. Likewise, suppression of heat shock proteins Hsp70 and Hsp90 would also probably contribute to the accumulation of damaged proteins and interfere with other cellular processes, including correct protein folding of newly synthesized proteins and pathogen defence. Downregulation of the translation factor eEF1A, caused by a downregulation of cytoskeleton expression, may interfere with the functioning of Hsp70 as well as prevent the synthesis of new proteins. Downregulation of photosynthesis‐related proteins components would probably conserve ATP, but in the mitochondria downregulation of energy‐generating functions would restrict the level of ATP available for defence. Upregulation of a gene involved in halogen metabolism is hypothesized to degrade the halogenated furanone defence molecules of D . pulchra , leading to an increased susceptibility to infection We found the nonpathogen Aquimarina sp. AD10 did not elicit any DEGs at 24 h when compared to the control treatment, indicative of its role as a benign symbiont (Kumar et al., 2016 ). However, at 48 h an upregulation in the expression of genes encoding heat shock proteins (indicative of a stress response) was observed. Together with the reduced expression of genes coding for signal transduction and translation functions (Table S4 ), these results suggest that AD10 has some capacity to similarly impair the algal immune response albeit after a longer exposure time. Therefore, while Aquimarina sp. AD10 does not cause obvious signs of disease under the conditions tested here (Kumar et al., 2016 ), future work will be required to assess if prolonged exposure to AD10 results in asymptomatic impacts. In conclusion, we hypothesize that the downregulation of genes associated with immune and defence functions in D . pulchra after pathogen exposure will hinder the macroalga's ability to modulate the immune response, contributing to its vulnerability to infection and leading to tissue damage and bleaching symptoms (Figure 4 ). Overall, this study provides valuable insight into the genetic features of the model macroalga D . pulchra , highlighting potential antipathogen response mechanisms of macroalgae and has future implications for how we view host–pathogen interactions in the context of environmental change."
} | 6,076 |
35528403 | PMC9071804 | pmc | 8,830 | {
"abstract": "5,5′-Dihydroxymethyl furoin (DHMF) is a novel biobased difuranic polyol scaffold, achievable from the benzoin condensation of 5-hydroxymethylfurfural (HMF), which has recently been employed as a monomer for the preparation of cross-linked polyesters and polyurethane. Its upgrading by means of enzymatic reactions has not yet been reported. Here we demonstrated that Candida antarctica lipase B (CALB) is a suitable biocatalyst for the selective esterification of the primary hydroxyl groups of DHMF. Exploiting this enzymatic activity, DHMF has been reacted with the diethyl esters of succinic and sebacic acids obtaining fully biobased linear oligoesters with number-average molecular weight around 1000 g mol −1 and free hydroxyl groups on the polymer backbone. The structures of the DHMF-diacid ethyl ester dimers and of the oligomers were elucidated by NMR and MS analyses.",
"conclusion": "Conclusions DHMF is an emerging HMF derivative whose synthetic upgrading is receiving increasing attention. In spite of the recently reported enzymatic synthesis, further biocatalyzed elaborations of its polyfunctional structure have not yet documented. In the present study we demonstrated that the primary hydroxyl groups of DHMF can be selectively esterified using commercial CALB as the catalyst. The enzymatic reaction with diethyl succinate or diethyl sebacate as acylating agents, performed under appropriate conditions, afforded oligoesters with number-average molecular weight of about 1000 Da. In view of the growing interest for green technologies devoted to the processing of biobased chemicals, we think that the present study could be an important starting point for further researches addressed to improve the polymerization procedure by exploring the activities of other lipases and/or modifying the reaction conditions. Furthermore, the herein reported results contribute to increase the synthetic tools available to enhance the value of DHMF as multifunctional synthetic scaffold for applications in life and material sciences.",
"introduction": "Introduction The gradual replacement of fossil resources with renewable biomass-derived feedstocks is imperative for the development of a more sustainable chemical industry. From this perspective, plant oils and sugars (mono- and polysaccharides) are the main carbon sources for the production of primary building blocks, either through chemical or biotechnological processes. In 2004, the US Department of Energy listed the top value-added building blocks achievable starting from sugars. 1 In this list, as in its revision published several years later, 2 5-hydroxymethylfurfural (HMF), a furan-based building block achievable by acid-catalyzed dehydration of C 6 sugars, appeared as one of the most promising platform-chemicals. Indeed, in the last decade, this molecule has received an increasing attention by either academic and industrial researchers. 3 A multitude of catalytic approaches has been implemented to produce HMF derivatives with potential uses as fuel additives, solvents or monomers. 4–6 Within this last category, 2,5-furandicarboxylic acid (FDCA) and its diester derivatives, have been successfully employed for the production of furanic–aliphatic polyester, which are promising biobased analogues of polyethylene terephthalate 7–12 with an increased biodegradability. 13,14 Again in the field of monomers for polyester synthesis, HMF is also a valuable precursor for polyols like, 2,5-bis(hydroxymethyl)furan, 2,5-bis(hydroxymethyl)tetrahydrofuran, 1,2,6-hexanetriol, 1,6-hexanediol, achievable through selective catalytic hydrogenation. 15–17 Among them, 2,5-bis(hydroxymethyl)furan (BHMF, Fig. 1 ), is particularly attractive as biobased diol monomer because of the interesting physical and mechanical properties that its rigid heteroaromatic structure could confers to the derived polymers. In spite of this, few studies have been reported on its polycondensation with diacid or diesters 17–20 and only one group used an enzyme as catalyst. 12,21 Another furan-based polyol scaffold, the 5,5′-dihydroxymethyl furoin (DHMF, Fig. 1 ), has been recently produced through benzoin-type condensation of HMF promoted by either N-heterocycle carbene (NHC) catalysts 22–25 or by the thiamine diphosphate dependent-enzyme BAL. 26 This novel HMF derivative has been in principle proposed as a precursor of oxygenated diesel fuels or C 12 linear alkane 22 which can be obtained through its catalytic hydrogenation or hydrodeoxygenation, respectively. Lately, however, by virtue of the polyhydroxylated rigid aromatic structure, DHMF has attracted the attention of polymer chemists who adopted it for the synthesis of bio-based polyesters 27 and polyurethane. 28 In particular, Chen and coworkers obtained cross-linked polyesters by reacting DHMF with various linear diacylchlorides while, in order to synthetize linear polyesters, they employed the diol 5,5′-bishydroxymethyl furil ( Fig. 1 ) obtained by selective oxidation of DHMF. 27 Until today, a polycondensation procedure selective for the primary alcoholic groups of the triol DHMF has not yet been reported. Such a strategy would allow to prepare linear polyesters with inner secondary hydroxyl groups available for further derivatization. We hypothesized that, by analogy with the above mentioned enzymatic synthesis of BHMF-based polyesters, 21 the desired DHMF-based linear polyesters could be obtained by exploiting lipases selectivity although, until now, the enzymatic esterification of DHMF has never yet described. 29 Herein we reported the unprecedented enzymatic esterification of DHMF. Furthermore, thanks to the selectivity shown by Candida anctartica lipase B (CALB) for the primary alcoholic groups of DHMF, the reaction conducted under appropriate conditions, has been exploited to synthetize fully bio-based aromatic–aliphatic linear oligoesters, bearing free secondary hydroxyl groups on the polymer backbone. Fig. 1 Aromatic polyol monomers achievable from HMF.",
"discussion": "Results and discussion The first issue we addressed in order to evaluate the exploitability of DHMF as polyol monomer for enzymatic polymerization, has been to exploring the affinity of CALB for DHMF. We were encouraged in this attempt by a recent study which reported high yield in the CALB catalyzed esterification of HMF with various acyl donors. 30 Since the potential substrate DHMF is not commercially available, we needed to synthetize it. Apart from the enzymatic route recently reported by Chuck and Domínguez de María, 26 all the others syntheses of DHMF reported in literature, exploited the umpolung benzoin condensation of HMF catalyzed by NHC catalysts generated from imidazolium, 24 benzimidazolium 23 or thiazolium 25 salts in the presence of a suited base or, in the only case of 1,3,4-triphenyl-4,5-dihydro-l H -1,2,4-triazol-5-ylidene, 31 by the thermal decomposition of the methoxylated precursor. 22,28 Following this last methodology, we obtained the pure DHMF in 95% yield after simple filtration and washing of the solid product. With the difuran triol in hand, we moved to test the activity of the commercial immobilized CALB Novozym 435, choosing diethyl succinate as the acylating agent. The reaction was performed in dry THF, at 40 °C, with equimolar amounts of the reactants (80 mM) and 50% of enzyme (w/w with respect to DHMF). After four hours, the 1 H NMR analysis of the reaction mixture showed the presence of a set of signals attributable to the dimeric products 1 and 2 depicted in Scheme 1 . In particular, the two singlets at 5.10 and 5.00 ppm where assigned to the protons of the two kinds of esterified hydroxymethylene groups ( Scheme 1 , group A of product 1 and group B of product 2, respectively) while, the others two singlets at 4.55 and 4.68 ppm were due to the protons of the corresponding unesterfied hydroxymethylene groups ( Scheme 1 , group b of product 1 and group a of product 2, respectively). The enzyme was removed from the reaction mixture and after evaporation of the solvent, the residue was chromatographed on silica gel, giving a mixture of products (80% yield) which surprisingly contained, together with 1 and 2, a new product that, on the basis of the additional set of signal appeared in the 1 H NMR spectrum ( Fig. 2 ), we identified as the monoesterified diketone 3 ( Scheme 1 ). The results of this preliminary study demonstrated that CALB catalyzes the selective esterification of both the primary hydroxyl groups of DHMF, leaving the secondary one free. Worthy to note is the fact that, in spite of the chiral nature of DHMF, both the enantiomers of the racemic mixture used for the transesterification reactions were accepted as substrate from CALB, as demonstrated by the almost complete conversion of the reactant. In light of this, we moved to investigate the possibility to synthetize linear polyesters using DHMF and diethyl succinate as comonomers by exploiting the catalytic activity of CALB. In doing that, we followed the approach recently reported by Loos and coworkers for the enzymatic synthesis of polyesters with BHMF ( Fig. 1 ) as the diol monomer. 21 They described a three-stage method, where CALB catalyzed the formation of oligomers in the first two stages conducted at atmospheric pressure and under a weak vacuum (350 mmHg), respectively. In the third stage, performed under high vacuum (3 mmHg) the above prepolymers undergoes to polycondensation, giving oligomers (MW ranging from 1.300 to 1.500, depending on the number of carbon atoms of the diacid monomer). Using the same conditions, namely DHMF to diethyl succinate molar ratio 1 : 1, CALB 20% (w/w, with respect to DHMF) and diphenyl ether as the solvent, after 1.5 hours at atmospheric pressure and 80 °C, the 1 H NMR spectrum of the reaction mixture (spectrum a, Fig. 3 ), showed the presence of the two singlets at 5.10 and 5.00 ppm (A and B) typical of DHMF esters while the signals of the free hydroxymethylene groups (a and b) were negligible. This indicated that the main product was the trimer 4 ( Fig. 3 ) derived from the esterification of both the primary alcoholic groups of DHMF. Furthermore, the ratio between the sum of the integrals A and B and that of the multiplet at 4.15 ppm (d in Fig. 3 ) which is due to the resonance of the ethyl groups of 4 and of the unreacted diethyl succinate, suggested a low conversion. The reaction was then left at the same temperature and after 3 hours the pressure was reduced to 300 mmHg. Scheme 1 CALB catalyzed transesterification of diethyl succinate with DHMF; reaction conditions: anhydrous THF 1.0 mL, DHMF 0.08 M, diethyl succinate 0.08 M, CALB 50% (w/w with respect to DHMF), 4 h, 40 °C. Fig. 2 \n 1 H NMR spectrum (in CDCl 3 ) of the mixture of products 1, 2 and 3 ( Scheme 1 ) obtained from the CALB catalyzed transesterification of diethyl succinate with DHMF after column chromatography. Fig. 3 \n 1 H NMR spectra of the reaction mixture (10 μL in 1 mL of CDCl 3 ) acquired after 1.5 (spectrum a) and 6 hours (spectrum b) of the enzymatic polymerization of DHMF and diethyl succinate conducted at 80 °C. Reaction conditions: diphenyl ether 0.18 mL, DHMF 0.7 M, diethyl succinate 0.7 M, CALB 20% (w/w with respect to DHMF), 80 °C, 3 h under atmospheric pressure followed by 3 h at 300 mmHg. After overall 6 hours, the 1 H NMR spectrum of the reaction mixture (spectrum b, Fig. 3 ) revealed a complete conversion of the diethylsuccinate. The main product of the mixture was the trimer 4 ( Fig. 3 ) accompanied by lower amounts of the esters 1 and 2, as revealed by the presence of the small signals of the free hydroxymethylene groups (a and b). In addition, the spectrum showed the appearance of a byproduct (signals x, y and z, spectrum b, Fig. 3 ) which formation is probably due to the tendency of the hydroxymethylene groups of DHMF to dehydrate and react with alcoholic functions forming ether linkages. This behavior has been also documented for the BHMF which, under the same conditions, dehydrated and reacted with ethanol. 21 On the basis of 2D 1 H NMR COSY experiment (ESI S3 † ) and of HPLC-MS analyses (ESI S11 † ), we deduced that, in the case of DHMF, after dehydration of one of the hydroxymethylene groups, the cyclic ether 5 ( Fig. 3 ) was formed through an intramolecular mechanism. In order to reduce the formation of such byproduct, we decided to repeat the reaction starting with the lower temperature and the lower pressure of 50 °C and 18 mmHg, respectively. After six hours under these conditions, the conversion determined by 1 H NMR was about 28% and the signals of the byproduct 5 were negligible. At this point, the temperature was increased to 60 °C and the reaction was continued for additional four hours. After this period, about 80% of diethyl succinate was converted and the byproduct 5 was present in trace amount (ESI S4 † ). In order to favor the condensation of the oligomers between them and with the unreacted DHMF the pressure was reduced to 3 mmHg, in this following the procedure described by Loos and coworkers. 21 The reaction was maintained at 60 °C under vacuum for additional ten hours and then diluted with chloroform in order to precipitate the residual DHMF. After filtration, the chloroform was evaporated from the organic solution and a yellow solid was precipitated from the diphenyl ether solution by adding dropwise cyclohexane. The solid was recovered by filtration, washed with cyclohexane and dried. The mass of the solid was about 70% of the starting mass of the reactants. The 1 H NMR analysis of the solid showed, in addition to signals of the esterified hydroxymethylene groups of the DHMF units (A and B in Fig. 4 ), also the signal of the corresponding groups of the diketone units (A′ in Fig. 4 ). From the spectrum it can be deduced that about 30% of the DHMF units incorporated in the oligomers were oxidized to the corresponding diketone form [ratio of integrals A′/(A + B)]. The free hydroxymethylene groups of the DHMF and of the diketone terminal units (signals a, b and a′ in Fig. 4 ) together with the ethyl groups (signal d in Fig. 4 ), represented all the possible ending groups. Finally, the spectrum showed an intense multiplet centered at 2.60 ppm generated by the proton's resonance of the succinyl units (signal E in Fig. 4 ). The 13 C NMR spectrum of the solid, allowed to identify all the types of carbon atom present in the oligomer structures (ESI S6 † ). The number-average molecular weight ( M n ) of the oligomers calculated from the 1 H NMR was about 1.150 (see Experimental section). Several of the oligomers were recognized also by HPLC-MS analyses (ESI S12 † ). Fig. 4 \n 1 H NMR spectrum (in CDCl 3 ) and products structures of the oligoesters mixture obtained after the three-step enzymatic polymerization of DHMF with diethyl succinate. Reaction conditions: diphenyl ether 0.18 mL, DHMF 0.7 M, diethyl succinate 0.7 M, CALB 20% (w/w with respect to DHMF), 50 °C, 18 mmHg for the first step of six hours, followed by a second step of four hours at 60 °C under the same pressure and finally by ten hours at 60 °C and 3 mmHg. The CALB recovered after the polymerization was washed with THF and dried under vacuum at 40 °C. After this treatment the enzyme was used for further three reaction cycles without significant loss of activity. The number-average molecular weight of the oligomers obtained, is in the same range of that reported for the products obtained from the enzymatic polycondesation of diethyl succinate with BHMF. 21 On the other hand, recent studies suggested a beneficial effect of an increase of the diester length on the results of enzymatic polycondensation. 12,32,33 For these reasons we moved to verify the extendibility of the approach to longer diacid esters. The sebacic acid diethyl ester was chosen as the model substrate and its transesterification with DHMF was first proven at atmospheric pressure in dry THF, as described for diethyl succinate. Also in this case, after chromatography, we obtained a mixture of the three products 6, 7 and 8 ( Fig. 5 ) analogues to the compounds 1, 2 and 3 produced with diethyl succinate. After that, we moved to study the enzymatic polymerization of DHMF with diethyl sebacate following the same three-step procedure adopted for the synthesis of the DHMF-succinate oligoesters (50 °C, 18 mmHg, 6 h; 60 °C, 18 mmHg, 4 h; 60 °C, 3 mmHg, 10 h). Fig. 5 \n 1 H NMR spectrum (in CDCl 3 ) of the products mixture obtained from the CALB catalyzed transesterification of diethyl sebacate with DHMF after column chromatography; reaction conditions: anhydrous THF 1.0 mL, DHMF 0.08 M, diethyl sebacate 0.08 M, CALB 50% (w/w with respect to DHMF), 4 h, 40 °C. In this case, the 1 H NMR analysis of the reaction mixture revealed the formation of oligomers including about 30% of diketone units already after the first stage which increase up to 40% at the end of the second stage. Being the aim of the work the synthesis of polyhydroxylated linear polyesters, we decided to stop the reaction at this stage. The attempt to accelerate the rate of the polymerization by increasing the temperature was not considered because it would favor the formation of the byproduct 5 as well. After the same workup described for the polymerization with diethyl succinate a yellow solid was recovered (about 75% of the initial mass of reactants). This products mixture was analyzed by 1 H NMR ( Fig. 6 ) and HPLC-MS analyses (ESI S14 † ) and a M n of about 1.000 was calculated from the 1 H NMR. The lacking increase of molecular weight by switching from the diethyl succinate to diethyl sebacate is in agree with the results of a precedent study where the enzymatic polymerization of BHMF with both the two diethyl ester furnished oligomers of comparable M n (about 1500 g mol −1 ). 21 Fig. 6 \n 1 H NMR spectrum (in CDCl 3 ) and structures of the oligoesters mixture obtained after the two-step enzymatic polymerization of DHMF with sebacic acid diethyl ester. Reaction conditions: diphenyl ether 0.18 mL, DHMF 0.7 M, diethyl succinate 0.7 M, CALB 20% (w/w with respect to DHMF), 50 °C, 18 mmHg for the first step of six hours, followed by a second step of four hours at 60 °C under the same pressure."
} | 4,568 |
35492519 | PMC9047062 | pmc | 8,831 | {
"abstract": "The design of functionalized polymer surfaces using bioactive compounds has grown rapidly over the past decade within many industries including biomedical, textile, microelectronics, bioprocessing and food packaging sectors. Polymer surfaces such as polystyrene (PS) must be treated using surface activation processes prior to the attachment of bioactive compounds. In this study, a new peptide immobilization strategy onto hydrocarbonaceus polymer surfaces is presented. A bio-interfactant layer made up of a tailored combination of laccase from trametes versicolor enzyme and maltodextrin is applied to immobilize peptides. Using this strategy, immobilization of the bio-inspired peptide KLWWMIRRWG-bromophenylalanine-3,4-dihydroxyphenylalanine-G and KLWWMIRRWG-bromophenylalanine-G on polystyrene (PS) was achieved. The interacting laccase layers allows to immobilize antimicrobial peptides avoiding the chemical modification of the peptide with a spacer and providing some freedom that facilitates different orientations. These are not strongly dominated by the substrate as it is the case on hydrophobic surfaces; maintaining the antimicrobial activity. Films exhibited depletion efficiency with respect to the growth of Escherichia coli bacteria and did not show cytotoxicity for fibroblast L929. This environmentally friendly antimicrobial surface treatment is both simple and fast, and employs aqueous solutions. Furthermore, the method can be extended to three-dimensional scaffolds as well as rough and patterned substrates.",
"conclusion": "Conclusions We reported a new bio-functionalization method to increase surface energy and add functional groups that provide binding sites and potential active sites for hydrophobic polymer surfaces. Layers exemplarily formed on polystyrene from aqueous laccase/maltodextrin formulations serve as a bio-interfactant layers. They allow for immobilization of antimicrobial peptides based on Tet-124 resp. KLWWMIRRW with and without DOPA C-terminus modification and for formation of layers that are stable and bioactive at physiological conditions. The antibacterial properties against Escherichia coli and the fibroblast L-929 growth viability were tested. Fibroblasts survived when in contact with both, the PS and PS/bio-functionalized surface. Antibacterial properties of the PS modified with laccase/maltodextrin/Tet-124-G-BrPh-DOPA-G surface were tested against E. coli . The surface functionalization does not increase bacteria adhesion and it caused partial depletion of bacteria growth.",
"introduction": "Introduction Biomedical devices have become an essential component of the human healthcare system. Due to their low cost, widespread availability, and versatile mechanical properties, industrial polymers such as polystyrene (PS), polypropylene and polyethylene are often used in the manufacture of medical products. 1 However, these materials do not present significant antibacterial properties. In addition, their surfaces must be treated using surface activation processes prior to the attachment of bioactive compounds. 2 Despite considerable research and development into biomedical devices and implants significant risk of infections related to biomedical devices and implants persists. The development of an implant coating having broad-spectrum antimicrobial activity is challenging due to the fast rate at which bacteria evolve resistance to antibiotics. An important goal within the biomaterials field is the development of surface functionalization methods that inhibit bacterial surface attachment and growth which can postpone or prevent infections and device failure. 3 Past work has made progressing antibacterial surface design utilizing strategies such as surfaces functionalized with antimicrobial agents, low-adhesion surfaces ( e.g. superhydrophobic interfaces), and polymer matrices that release antibiotics. 4,5 A promising approach to overcome this challenge is to utilize naturally occurring antimicrobial peptides or synthetic analogues against bacterial infections. One strategy that has shown significant targeting potential is the use of antimicrobial peptides (AMPs), which exhibits a number of advantageous properties: effectiveness against a broad range of bacteria including antibiotic-resistant strains, low-toxicity to mammalian cells, small molecular size and high stability. 6,7 Furthermore, advantages of immobilized peptides comprise long-term stability. AMPs are produced by the immune systems of animals, insects and plants and are secreted when infections occur. 8–10 Consisting of approximately 12 to 100 amino acids, AMPs are amphiphilic molecules having a high content of cationic and hydrophobic amino acids providing side chains responsible for the antibacterial activity. 11 Arising from electrostatic interactions between positively charged amino acids and the negatively charged bacterial cell membrane, AMPs work by compromising membrane integrity and transport permeability. 12–15 Physical adsorption and covalent bonding strategies have been studied for AMPs immobilization on surfaces. Layer-by-layer deposition is a common physical immobilization technique applied to hydrophilic materials such as metal oxides, mainly due to electrostatic interactions 16,17 or the adding of a specific surface binding recognition sequence to the peptide. 18–20 This method may be applied on hydrophobic polymer surfaces after applying a process to increase the concentration of functional groups such as hydroxyl and carbonyl groups, and next covalent attachment using click chemistry. 21 Another viable method is solid phase peptide synthesis on polymer resins in which the protected amino acids are incorporated by assembling the peptide sequence from its C- to its N-terminus; however, these methods are time consuming and costly. 21–23 Although peptides with hydrophobic amino acids may adsorb strongly enough on hydrophobic surfaces to form a stable film under controlled conditions. The peptides when are amphiphilic likely adsorb with their hydrophobic segments facing hydrophobic surfaces and the restricted orientation of the peptide might cause a reduction in its antimicrobial properties. 24 For instance, Hilpert et al. reported and improved peptide antimicrobial properties when adsorbed at an orientation such that their hydrophobic segments are away from the interface; facing the bacteria membrane. 22 Formation of stable laccase/maltodextrin layers on highly oriented pyrolytic graphite (HOPG) was reported in Corrales et al. 25 Laccase molecules adsorb irreversible on this hydrophobic surface in short contact times; forming layers that can serve as binding sites for the interaction and attachment of other biomolecules. In this study, we present a new surface immobilization strategy for AMPs with or without −3,4-dihydroxyphenylalanine (DOPA) in their sequence which profits from using laccase as a bio-interfactant. 26 The antimicrobial peptide used for immobilization is based on the peptide Tet-124 sequence KLWWMIRRW, previously synthesized by Hilpert et al. 22 This peptide showed partial inhibition of Pseudomonas aeruginosa . Applying the adhesion mechanism of mussel foot proteins, we have modified the AMP sequence adding glycine-(4-bromophenylalanine)-(3,4-dihydroxyphenylalanine)-glycine to the original sequence reported, thus obtaining KLWWMIRRW-G-BrPh-DOPA-G. The amino acid DOPA was added in the sequence to improve the peptide adhesion, comparing the stability of both peptides on the surface after long rinsing times, and to favorably be oriented at the interface of these films. 27 We have changed the peptide C-terminal such that two glycines flank the DOPA in order to minimize steric hindrance, allowing for more flexibility, 28 and to prevent intramolecular crosslinking to neighbouring amino groups due to the potential laccase catalyzed oxidation of the DOPA. 4-Bromophenylalanine (BrPh) was added to the sequence in order to quantitatively determine the peptide adsorption fraction and to differentiate the peptide and the enzyme adsorbates within the layers. The formation of thin films in situ was characterized using quartz crystal microbalance with dissipation (QCM-D), and the resulting adsorbates were studied using X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), and water contact angle. The peptide laccase catalyzed oxidation on the surface was studied carrying on a selective derivatization reaction for DOPA containing peptides. Antimicrobial properties against Escherichia coli and, moreover, the fibroblast L-929 biocompatibility of the bio-films were tested.",
"discussion": "Results and discussion Polystyrene (PS) surface modification followed a similar protocol as used to functionalize HOPG surfaces. 25 Robust layers suitable for functionalization were formed using laccase since the protein (enzyme) irreversibly adsorb quickly to hydrophobic surfaces. There was a change in surface hydrophobicity to more hydrophilic domains, directly correlated with the decrease in the surface water contact angle, Fig. 1A . This layer could withstand rinsing with solutions at various pH levels and ultrasound over long periods of time, Fig. 1B (Table S1 † ). Similar layer LM to the obtained on HOPG was formed. 25 However, protein adsorption rate and coverage depends on time in contact with the surface, enzyme concentration, ionic strength, presence of other molecules in solution, surface affinity, among other factors. 30 Surface elemental composition as a function of time in contact with laccase/maltodextrin suspension (LMS) was investigated using XPS for durations up to 24 hours. Both nitrogen and oxygen concentrations increase monotonically in time and saturated after ≈10 min ( Table 2 , Fig. 1C and D ). Moreover, the enzyme was still active after being irreversible adsorbed on PS as confirmed by ABTS oxidation (Table S2 † ). Protein adsorption on PS surfaces could be driven by π–π interactions between protein and PS aromatic groups. 31 The laccase is widely distributed in fungus that decomposes lignocellulosic materials and its sequence is composed of 49% of hydrophobic amino acids, suggesting that its function in nature is maybe set to have affinity for hydrophobic surfaces. 32,33 Fig. 1 (A) Water contact angle as a function of time for PS after exposure to LMS. Inset: sessile drop images for the unmodified PS control sample (left) and for the surface treated PS after 1 hour (right). XPS findings indicating (B) [N] after various rinsing and desorption procedures. Surface concentration as a function of time (C) nitrogen (D) oxygen; values of two independent samples are shown as first and second point. PS elemental surface concentration analyzed after contact with LMS, LMS and Tet-124-G-BrPh-DOPA-G. Average surface roughness ( R a ) was measured using AFM topographic images Sample PS Tet124-G-BrPh-DOPA-G PS/LM LMS 10 min followed by the addition of the peptide solution for 50 minutes O 2.2 12.7 8.1 11.3 N 0.1 12.6 3.2 6 C 97.5 72.1 85.1 82.6 S — — 0.2 0.2 Br — 0.6 0.1 Na — — 0.1 Si — — 0.4 0.2 Ra 2.4 ± 0.9 nm — 3.1 ± 0.7 nm 3.8 ± 1.1 nm QCM-D was utilized to study the adsorption dynamics of peptide layer formation. Fig. 2A shows the resonant frequency shift arising from peptide adsorption on the PS surface and LM layer surface. Frequency shifts due to peptide adsorption on the LM layer (two-step functionalization process), and neat PS are detected and quantified. During the two-step functionalization process LM surface concentration increases until adsorption and desorption reach dynamic equilibrium. In the second step, the peptide solution was in contact with LM layers and a stable peptide layer was formed. Therefore, peptide layers on both PS and LM were stable against buffer rinsing. However, a thicker peptide layer was formed on LM layers. Peptide mass adsorbed on PS and PS/LM was calculated using Sauerbrey resulting in 92 and 167 ng cm −2 , respectively, Fig. 2A . To enhance the antimicrobial properties, hydrophobic residues should face away from the surface in order to increase interactions with bacteria membrane. Differences in mass adsorption could be related to surface affinity as well as to differences in the orientation of molecules building up the film: e.g. molecular long-axes lying in the surface plane for neat PS or long-axes oriented perpendicular to the surface for LM layers. Laccase and peptides could interact not only by π–π stacking but also, by electrostatic interactions between the charged functional groups and formed hydrogen bonds. 34 ITC titration curves of the Tet 124-G-BrPh-DOPA-G into laccase/maltodextrin suspension shows a characteristic curve for an enzymatic reaction on contrary to Tet-124-G-BrPh (Fig. S1 † ). Consequently, an intermolecular crosslinking could occur between laccase oxidized Tet-124-G-BrPh-DOPA-G peptide and the surface. 35 The emission fluorescence spectra obtained from the Tet 124-G-BrPh-DOPA-G solution shows an increment of intensity as expected with peptide concentration. The peptide reacts with the benzylamine to form a benzoxazole derivate molecule that has a fluorescence emission at 467 nm. Moreover, the fluorescence emission peak of the solution removed from the PS/LM/Tet 124-G-BrPh-DOPA-G modified surface suggests the presence of catechol-free-moieties on the surface. In addition, the elemental concentration on the surface after the derivatization reaction indicates that the peptide could be partially desorbed after contact with this highly alkaline solution, [N] and [O] of the PS/LM/Tet-124-BrPh-DOPA-G after the derivatization reaction was lower than the surface before the derivatization reaction and higher than the PS/LM modified surface. The results suggest a peptide layer with good stability and a possible covalent attachment ( Fig. 2B and S2 † ). Fig. 2 (A) Resonance frequency shift Δ f as a function of time for a PS coated crystal surface initially exposed to 0.1 mg mL −1 LMS and later to 0.05 mg mL −1 Tet-124-G-BrPh-DOPA-G (purple curve) and a PS coated crystal only exposed to 0.05 mg mL −1 Tet-124-G-BrPh-DOPA-G (red curve). (B) Fluorescence spectra of Tet-124-G-BrPh-DOPA-G solutions at concentration of: (a) 5 μg mL −1 , (b) 10 μg mL −1 , (c) 20 μg mL −1 , (d) 50 μg mL −1 after reaction with 10-fold benzylamine; (e) only reagents (dotted line) and supernatant solution after contact with the peptide (Tet-124 or Tet-124-G-BrPh-DOPA-G) PS/LM modified surface (continuous line). Inset: comparison of [O] and [N] on PS/LM/before and after the surface derivatization reaction. XPS findings indicate peptide partial desorption from the surface. Peptide adsorption on the LM layers resulted in increased surface concentrations of [N] and [O] as detected by XPS. Moreover, the XPS results confirmed the presence of Br, as shown in Table 2 . To determine the bulk composition of the peptide and to estimate the maximum expected Br concentration detected by XPS in a more than 10 nm thickness peptide film, a drop of aqueous peptide solution was dried, and the area probed using XPS. As shown in Table 2 , the [Br] was determined to be 0.6 at%. From this characterization, it becomes clear that for the 0.1 at% Br concentration detected in the adsorbate, film obtained after 50 min of contact between PS/LM and the peptide solution, the peptide stoichiometry will predetermine a of [N] = 2.1 at% for the brominated peptide. On the other hand, any nitrogen species that contribute to a concentration [N] > 2.1 at% after such peptide adsorption will arise from the previous enzyme adsorption contributing to the formation of a hybrid enzyme-peptide layer. Fig. 3A and B show the fitting results for the C1s signal of neat PS and PS/LM/Tet-124-G-BrPh-DOPA-G surface. The C1s spectra for the PS and substrate modified with LM and the Tet-124-G-BrPh-DOPA-G is characterized by a major peak at 284.8 eV assigned to predominantly aromatic hydrocarbonaceous species in case of PS and predominantly aliphatic hydrocarbon species in case of the peptide and laccase adsorbates, with their prominent peaks situated at 285.1 ± 0.1 eV. Four additional peaks were allocated at higher binding energies for the PS/LM/Tet-124-G-BrPh-DOPA-G sample. The peak at 286.3 ± 0.1 eV is associated with amine-like C*–N; the signal related to C*–O from alcohol/alkoxy groups is characterized by a 0.5 eV higher binding energy; the peak centered at 288.2 ± 0.1 eV is characteristic of amide groups (–NH–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); and, finally, the peak at 289.0 ± 0.1 eV is correlated with carboxyl groups. 25 The layer thickness calculated following the protocol described by Corrales et al. is approximately 1 nm and 2.7 nm for LM and LM/Tet-124-G-BrPh-DOPA-G layers on PS, respectively. 25 Fig. 3 C1s XPS findings for (A) PS control, (B) PS modified with LM/Tet-124-G-BrPh-DOPA-G when fitting the measured signal with contributions from the PS substrate, laccase and peptide bulk sample spectra. Surface topography and roughness of layers were studied using AFM. Initially, molecular dimensions of the laccase and peptide measured upon adsorption on HOPG as a model hydrophobic flat surface. Fig. 4A shows adsorbed laccase globular molecules having 3.5 ± 0.4 nm heights and larger 13 nm agglomerates appear to be randomly distributed. The peptide adsorbed on the surface forms a close thin monolayer of approximately 0.8 ± 0.2 nm on HOPG, Fig. 4D . The phase image shows contrast between the laccase molecules and substrate but not between the laccase molecules and the tip; suggesting similar adhesion forces, Fig. 4B . The peptide layer shows similar results, Fig. 4E . The layer thickness calculated between the bottom area a hole and upper LM/Tet-124-G-BrPh-DOPA-G adsorbed on HOPG is approximately 4.2 nm. The layer thickness correlates to one laccase and one peptide molecule (Fig. S3 † ). Fig. 4 AFM topography images of (A) laccase, (D) Tet-124-G-BrPh-DOPA-G. Phase image corresponding to (B) A, (E) D. Line height profile highlighted in (C) A, (F) D. \n Fig. 5A shows a PS smooth, without globular nanostructures, surface. Fig. 5B shows a surface homogeneously covered with laccase enzyme molecules. Comparison between the phase images of the LM layers and LM/peptide allowed determining the peptide distribution. The darker areas could be associated to the peptide layer adsorption on top and surrounding of laccase molecules, Fig. 5C . The average roughness, a factor that influences bacterial adhesion, 36,37 changed by less than 2 nm (see Table 2 ). The influence of the layers on a subsequent bacteria adhesion will be discussed further in this study. In order to maximize peptide concentration on LM interfactant layers, peptide adsorption time-dependence was investigated for time periods ranging from 1 min to 24 hours. Fig. 6A and B show an increase in surface concentrations of [N] and [Br] with time. After 24 hours, maximum concentrations of 7.2 at% and 0.22 at% were obtained for [N] and [Br], respectively. After comparing with the bulk composition of the peptide corresponding to a maximum surface concentration [Br] around 0.6 at%, we suggest that peptide layers characterized by [Br] = 0.1 at% are a few nanometers thin and may partially covering the surface. The [Br] concentration was higher for the peptide layers formed on LM layers than on PS even after 24 hours, Fig. 6C . Given that the isoelectric point pI value of the peptide Tet-124-G-BrPh-DOPA-G is ≈11.5, we expect that at a pH of 4.75, lysine and arginine side chains were positively charged and experience electrostatic repulsion that may prevent multilayer formation. Such supposition is in reasonable agreement with the presented AFM findings, and a more precise estimate of the peptide layer thickness may be provided following some XPS signal fitting that facilitates establishing the PS substrate signal attenuation. Another DOPA containing peptide that was previously published by the authors with charged +2 at pH 7, Ala-Lys-Pro-Ser-Tyr-Hyp-Hyp-Thr-DOPA-Lys, 38 formed a thicker and stable layer on the laccase surface (Fig. S4 † ). Consequently, we suggest that the thin peptide layers formed could be result of electrostatic repulsion between peptide molecules; not favoring multilayer formation. The influence of the amino acid DOPA on the adsorption stability of antimicrobial peptides onto PS/LM was studied by comparing the surface concentrations [Br] after exposure to either peptide solution ( Fig. 6D ), in this way considering peptides with and without DOPA modification. Similar values suggest that catechol crosslinking at the surface does not play a significant role with respect to the achieved layer thickness and stability, even if the enzyme is active after immobilization on the PS surface and the laccase enzyme is able to oxidize and to polymerize the DOPA containing peptide in solution (Fig. S1 † ). Fig. 5 AFM 3D topography images of (A) neat PS, (B) PS/LM (10 min) and then, exposed to a Tet-124-G-BrPh-DOPA-G (50 min). (C) B phase image. Fig. 6 XPS-based elemental surface concentration as a function of time (A) [N], (B) [Br]; LMS contact time 10 min and then, Tet-124-G-BrPh-DOPA-G solution up to 24 h. Comparison of adsorption of (C) Tet-124-G-BrPh-DOPA-G on PS and PS/LM surface, assessing the stability after long rinsing time (0.1 M aqueous NaCl, 24 hours exposure under shaking). (D) Comparison of [N] and [Br] between Tet-124-G-BrPh-DOPA-G and Tet-124-G-BrPh layers formed on PS/LM. We suggest that the adhesion promoting interfactant properties of the laccase layer are dominant and that the electrostatic interactions and hydrogen bonding between the functional groups facing the laccase layer surface drive the adsorption and stabilization of the peptide layer and consequently, the transport of the respective catechol moieties inwards or outwards to the catalytically active sites of the laccase. We infer that the molecular environment of such lock-and-key system may affect the reaction probabilities between laccase and catechol moieties. We attribute such hindrance to the limited diffusive mobility and the restricted conformational freedom of the adsorbed peptide on top of the bio-interfactant layer, which is due to non-covalent interactions. Further studies are necessary to determine differences in the adsorption coverage density and stability of a peptide with the DOPA modification at the N terminus modification or a comparison of hydrophilic peptides with and without DOPA. Antimicrobial properties and biocompatibility Given that it is a common bacterium colonizing medical device surfaces and potentially may cause infections, E. coli was used for antimicrobial studies. 39 MIC in solution was calculated to determine whether the changes on the C-terminus caused antibacterial activity depletion. As it can be seen in the data shown in Table 3 , the antimicrobial activity was decreased by G-BrPh and G-BrPh-DOPA-G C-terminus sequence modification. For instance, significant changes in antimicrobial activity has been reported due to the exchange of an l -amino acid to a d -amino acid in other peptides sequences. 40 Circular dichroism spectra shown in ESI Fig. S5 † indicates that both peptides have α-helical conformation in solution. 41 Further studies are necessary to determine if changes in the factors such as hydrophobicity and steric hindrance influenced the antimicrobial activity. 42–46 The antimicrobial activity of the Tet-124-G-BrPh-DOPA-G and Tet-124-G-BrPh were lower than the Tet-124; however, the peptide is still active. They show higher antimicrobial activity than the Tet-000 (peptide without reported antimicrobial activity), Table 3 . The peptide Tet-124-G-BrPh-DOPA-G was chosen for further antimicrobial tests due to the lower MIC in comparison with the Tet-124-G-BrPh. Fibroblast viability in solution was not affected at MIC of the peptides under studied conditions, Fig. 7A . There are no significant differences between cells growth negative control and cell in contact with the different peptides tested in solution at concentration of 250 μg mL −1 for all the peptides according to one-way ANOVA statistical analysis. Fibroblast cell morphology was compared between the cell's growth on PS, PS/LM and PS/LM/Tet-124-G-BrPh-DOPA-G substrates, Fig. 7B and S6. † For both PS surface modifications, cells adhered and survived in similar way that the control PS sample. Image processing (Image J) was used to measure the average cell surface density. Cell densities were determined to be 95 ± 5, 100 ± 10 and 130 ± 20 fibroblast cells/mm 2 for the PS, PS/LM and PS/LM/Tet-124-G-BrPh-DOPA-G, respectively. We did not find a significant difference applying ANOVA test ( p < 0.05) between the samples (Fig. S6 † ). To determine the influence of roughness and chemical functional groups, bacterial adhesion was tested. There was no increase in E. coli adhesion in between the neat control surface (Fig. S7 † ) and PS/LM/Tet-124-G-BrPh-DOPA-G surface. E. coli grown on LM/Tet-124-G-BrPh-DOPA-G modified surfaces displayed a log phase delay of 10 hours indicating enhanced antimicrobial surface activity. Adsorbed peptides exhibit low mobility, and therefore, antimicrobial effects can be attributed to cell membrane disruption. 47 Bacterial cell membrane disruption must arise from electrostatic destabilization rather than peptide penetration since the membrane thickness of E. coli (45–55 nm) is approximately 10 times larger than peptide length (≈5.9 nm). The membrane of Gram-negative-bacteria consist of a peptidoglycan layer between a lipid bilayer in which a high density of negatively charged lipopolysaccharides acids are in the outer lipid membrane. Cationic peptides interact with the outer cell membrane allowing for the entry of exterior molecules to diffuse into the cell. 47,48 This disrupts the electrochemical gradient across the membranes, increasing water and ion diffusion through the membrane which leads to cell swelling and osmolysis. 49,50 The sample surface modified with the peptide showed dead and alive bacteria after 18 hours of contact with E. coli . This result suggests that this surface modification caused a partial depletion of the bacteria, which may be related to a non-homogeneous peptide surface distribution as can be seen in Fig. 7D . Overall, E. coli grown on PS/LM/Tet-124-G-BrPh-DOPA-G surfaces displayed a log phase delay of 10 hours indicating enhanced antimicrobial activity. Our results demonstrated that the immobilization process was successful and preserved (at least partially) antimicrobial peptide activity. For better understanding the effect of AMP immobilization technique on the antimicrobial activity further analysis is necessary because the antimicrobial mode of action could change according to the immobilization strategy. 51 Therefore, this methodology could be used to immobilize mixtures of peptides that could act in synergy against biofilm forming bacteria. MIC of peptides against E. coli , 18 h incubation in minimal medium Peptide Sequence MIC peptide, μg mL −1 Tet000 GATPEDLNQKLS 500< Tet-124 KLWWMIRRW 14 Tet-124-G-BrPh-G KLWWMIRRWG-BrPh-G 225 Tet-124-G-BrPh-DOPA-G KLWWMIRRWG-BrPh-DOPA-G 100 Fig. 7 (A) WST-1 fibroblast viability test (NC: negative control, PC: positive control, Tet-124-G-BrPh-DOPA-G at different concentrations). (B) Fibroblast L-929 cell growth on PS/LM/Tet-124-G-BrPh-DOPA-G modified surface. Cells were stained using rhodamine phalloidin (actin filaments) and DAPI (nucleus). (C) E. coli growth as a function of time following the detachment protocol after 18 h in contact with the surface. (D) Fluorescence image of live (green) and dead (red) bacteria (LIVE/DEAD staining) on a PS/LM/Tet-124-G-BrPh-DOPA-G modified surfaces, surface rinsed with buffer."
} | 7,090 |
35072365 | PMC8922109 | pmc | 8,833 | {
"abstract": "Abstract Liquid‐repellent technology is an efficient means of energy‐saving and biofouling avoidance. However, liquid‐repellent surfaces suffer from inefficient lubricant retention under shear flow and fouling problem in marine environment. Here, the authors demonstrate a fatty acid amide (FAA)‐based oleogel for stable and sustainable lubrication in marine environment. The lubrication management of marine creatures is emulated in synthetic oleogels by incorporating solid (FAA) and liquid lubricants into the molecular meshes of polymeric networks, with the nature‐derived solid lubricant providing multifunctional synergistic effects with liquid oil molecules for slippery property and remarkable anti‐biofouling. The lubricant‐confining gel achieves shear‐stable lubricity with efficient oil management. The oleogel provides continued lubrication without biofouling for approximately 4 months in marine field tests. The gel design provides a new paradigm for sustainable and shear‐stable lubrication in marine environment.",
"conclusion": "3 Conclusion Designing liquid‐repellent surfaces from FAA‐incorporating oleogels allows shear‐stable lubrication to be engineered at the molecular level for practical applications under various flow conditions. Inspired by non‐wetting slippery marine creatures, the lubrication attributes inherent to amphiphilic FAAs are translated to the properties of oleogel networks with synergistic effects of oil molecules. The shear‐stable slippery surface is created by incorporating solid and liquid lubricant into the molecular meshes of the gel, which provides high lubricant retention under shear flow. This unique gel design enabled persistent lubrication under shear flow conditions. A dual lubricant‐based gel system generates multifunctional properties, including sustainable slippery property and anti‐biofouling properties. As demonstrated in marine field tests using a ship, FAA‐incorporating oleogels were able to sustain their multifunctional properties in an almost perfect manner under marine environments for approximately 4 months. Thus, FAA‐based oleogel design can pave the way for a shear‐stable multifunctional liquid‐repellent surface for many practical applications, including drag reduction, anti‐biofouling, anti‐icing, and self‐cleaning technology.",
"introduction": "1 Introduction Liquid‐repellent technologies have been a popular subject in both academic research and industrial applications over the past decades, including drag reduction, anti‐biofouling, self‐cleaning, anti‐icing, and medical applications. [ \n \n 1 \n , \n 2 \n , \n 3 \n , \n 4 \n , \n 5 \n \n ] Inspired by the natural non‐wetting structures of lotus leaves and Nepenthes pitcher plants, synthetic liquid‐repellent surfaces have been developed as micro/nano‐structured surfaces on a composite solid–air interface and solid–liquid interface. [ \n \n 6 \n , \n 7 \n , \n 8 \n \n ] Fundamental to these systems, micro/nano‐cavities of such surfaces are infused with non‐wetting lubricating materials such as air for superhydrophobic (SHP) surfaces and liquid lubricant for liquid‐infused surfaces (LISs). [ \n \n 9 \n , \n 10 \n \n ] Despite their promising liquid‐repellent and pressure‐stable properties, the lubricants trapped within the cavity cannot stand up to physical shear, so that the trapped lubricants, particularly those with low viscosity, can easily drain from the cavity under even low‐speed shear flow. [ \n \n 11 \n , \n 12 \n \n ] Moreover, flow‐induced shear at the flow/lubricant interface leads to an increase in surface roughness arising from curved lubricant meniscus and lubricant depletion, resulting in loss of frictional lubrication or even adverse effects on friction. [ \n \n 13 \n \n ] Recent progress in extending the lubrication lifetime, by optimizing the rheological properties, patterned surface chemistry, surface geometry, and lubricant replenishment, remains an inherent limitation in lubrication under high‐speed turbulent flow, which severely restricts its practical applications such as in marine environments. [ \n \n 11 \n , \n 12 \n , \n 14 \n , \n 15 \n , \n 16 \n , \n 17 \n , \n 18 \n , \n 19 \n , \n 20 \n , \n 21 \n \n ] In addition, the fouling resistance of slippery surfaces enables to sustain their lubrication property without biofilm‐induced drag and corrosion in marine environments, indicating the requirement of multifunctionality in liquid‐repellent surfaces. Different from previous microtexture‐based repellent technologies, marine creatures, such as coral, seaweed, and eel, exhibit slippery and anti‐biofouling properties utilizing a volumetric lubricant‐impregnated tissue system in which fatty acid‐incorporated tissues contain mucus. [ \n \n 22 \n , \n 23 \n , \n 24 \n \n ] In addition, mussels have a waterproof outermost matrix based on fatty acid amide (FAA), primarily composed of oleamide, which is a commercial slip‐induced solid‐type lubricant. [ \n \n 25 \n , \n 26 \n \n ] \n Inspired by this idea, we report a shear‐stable lubrication strategy based on FAA‐incorporating oleogels with dual solid/liquid lubricants to achieve sustainable lubrication in marine environment ( Figure \n 1 a,b ). We present a lubricant confinement strategy based on oil management through the synergetic effects of the solid and liquid lubricants in the molecular meshes of polymeric networks. The oil molecules are densely packed with high diffusive kinetics in the molecular mesh of the FAA‐incorporating oleogel. Gel‐confined solid/liquid lubricants form a molecularly continuous lubricant interface with little projected solid fraction and meniscus curvature. Furthermore, the FAA‐incorporating gel network eradicates lubricant loss under shear flow. Additionally, though most LISs rely on liquid lubricant to achieve high mobility, the dual solid/liquid lubricant system provides slippery properties with anti‐biofouling properties, indicating the multifunctional surfaces appropriate for various practical applications. In particular, the dual lubricant‐based gel system effectively resisted adhesion against various marine biological systems, compared to the conventional liquid lubricant‐based gel system. The epoch‐making multifunctional nature of FAA‐incorporating oleogels was stably sustained for approximately 4 months in long‐term marine field tests, indicating their strong potential for practical applications in marine environments. Figure 1 Design and characterization of FAA‐incorporating oleogels. a) Schematic illustration of FAA‐incorporating oleogels with shear‐stable lubrication in a marine environment. b) i) Histological cross‐section of skin tissue (blue, mucus; red, skin tissue of hagfish) of marine creatures showing mucus‐infused tissue, and ii) image of bioinspired FAA‐incorporating oleogel. c) Schematic structure of FAA‐incorporating oleogel with dual penetration of FAA and oil molecules in crosslinked PDMS polymeric network. ToF‐SIMS image showing the spatial distribution of erucamide (C 22 H 44 NO + , m / z 338.34) on a FAA‐free surface (d) and EPC5.0 surface (e). The color scale bar represents the range of pixel intensity [0, MC]. MC: maximum ion count per pixel, TC: total count in the entire image. f) Storage (G′, filled symbols) and loss moduli (G″, open symbols) of FAA‐free oleogel and EPC5.0 oleogel."
} | 1,816 |
27780257 | PMC5079562 | pmc | 8,834 | {
"abstract": "Introduction Biological communities present in soil are essential to sustainable and productive agricultural practices; however, an accurate determination of the ecological status of agricultural soils remains to date an elusive task. An ideal indicator should be pervasive, play a relevant role in the ecosystem, show a rapid and proportional answer to external perturbations and be easily and economically measurable. Rhizobacteria play a major role in determining soil properties, becoming an attractive candidate for the detection of ecological indicators. The application of massive sequencing technologies to metagenomic analysis is providing an increasingly more precise view of the structure and composition of soil communities. In this work, we analyse soil rhizobacterial composition under various stress levels to search for potential ecological indicators. General Biodiversity Indicators Our results suggest that the Shannon index requires observation of a relatively large number of individuals to be representative of the true population diversity, and that the Simpson index may underestimate rare taxa in rhizobacterial environments. Taxonomical Classification Methods Detection of indicator taxa requires comparison of taxonomical classification of sequences. We have compared RDP classifier, RTAX and similarity-based taxonomical classification and selected the latter for taxonomical assignment because it provides larger detail. Taxonomy-Based Ecological Indicators The study of significant variations in common, clearly identified, taxa , using paired datasets allows minimization of non-treatment effects and avoidance of false positives. We have identified taxa associated to specific perturbations as well as taxa generally affected in treated soils. Changes in these taxa , or combinations of them, may be used as ecological indicators of soil health. The overall number and magnitude of changes detected in taxonomic groups does also increase with stress. These changes constitute an alternative indicator to measuring specific taxa , although their determination requires large sample sizes, better obtained by massive sequencing. Summary The main ecological indicators available are the Shannon index, OTU counts and estimators, overall detection of the number and proportion of changes, and changes of specific indicator taxa . Massive sequencing remains the most accurate tool to measure rhizobacterial ecological indicators. When massive sequencing is not an option, various cultivable taxonomic groups, such as specific groups in the Actinobacteria tree, are attractive as potential indicators of large disruptions to the rhizobiome.",
"conclusion": "Conclusion We have considered several potential ecological indexes based on the current availability of metagenomic data. Our results show that the Shannon index requires large sample sizes (~5000 individuals) and the Simpson index likely underestimates rare species. OTU-based methods to measure diversity require even larger sample sizes, although they may be reduced by the use of conscientious selection of the CHAO1 or ACE estimators. Taxonomy-based indicators may provide a sensible alternative. This approach requires taxonomic imputation of individual sequences, which -according to our results- is best achieved using similarity-based methods, although other approaches such as RDP or RTAX may provide a coarser but faster short-cut for huge datasets. Choosing one or more groups to use as ecological indicator is limited by the relatively reduced NGS data available. The Indicator Value approach can provide useful information in controlled experiments or when enough data is available, and should become a routine procedure in future metagenomic analyses. When tit is not applicable, a more specific case-control approach such as the one described here should be more sensitive. As a rule, NGS technology is currently the most accurate tool to explore rhizobacterial communities, allowing identification of ecological indicators as changes in specific taxa or in the overall number and proportions of changes. When NGS is not an option, various taxa in the Actinobacteria , Verrucomicrobia and Proteobacteria trees may be selected as indicators at various taxonomic levels. In particular, relative changes in the proportions of Actinobacteria are easy to detect and commonly associated to major external perturbations. Most of the data currently available describes the rhizobacterial communities under normal agricultural practices and there is a scarcity of contrasted data on the effects of other major external perturbations. We have considered mainly soils treated with pesticides having a well-known, proven impact, and our results agree with other observations. However, all the results reported in this work should be revised as additional data collected from ecologically affected soils becomes available.",
"introduction": "Introduction Soil is the substrate that sustains terrestrial life on Earth, constituting an essential resource for the maintenance of most life processes, not only via agricultural production, but also by means of organic matter decomposition and nutrient cycling. Soil composition includes minerals, organic matter and a rich symbiotic community of micro- and meso-organisms that play a crucial role in the life-sustaining ability of soil. Viewing soil as a living ecosystem allows definition of soil health (also referred to as soil quality) in terms of its capacity to sustain plant and animal productivity under changing conditions. The diversity and evenness of the plant growth promoting bacteria (PGPB) and rhizobacteria (PGPR) communities has been reported to increase agricultural yield via a variety of mechanisms. Nitrogen fixing bacteria include symbiotic species and non-symbiotic species such as Proteobacteria , and Actinobacteria . Phosphate solubilizing microorganisms provide a biological rescue system for inorganic P and include bacteria and fungi. The most efficient belong to the genera Bacillus and Pseudomonas among bacteria, and Aspergillus and Penicillium among fungi. Other bacteria improve mineral and water uptake (NO 3 - , PO 4 3- and K + ) [ 1 , 2 ]. In addition to providing ready access to nutrients, rhizobacteria produce a broad array of antibiosis products and functions, some of which target plant pathogens and therefore protect the crop. On the other hand, some soil micro-organisms are naturally resistant to a broad range of antibiotics (the \"antibiotic resistome\"). Depending on this balance, some soils have long been known to restrict the activity of plant pathogens and are known as \"suppressive soils\". The onset of suppressiveness in a soil is usually associated with a disease outbreak triggering activation of disease-suppressor bacteria. Proteobacteria , Firmicutes and Actinobacteria have been consistently associated with disease suppression, with Actinobacteria being the most dynamic taxa when pathogenic fungi were added [ 1 , 3 , 4 ]. Soil health, depends largely on the composition and diversity of the rhizobacterial community [ 5 , 6 ]. The rhizobacterial community symbiotically adapts to specific substances secreted by the crop being cultivated. Bacterial diversity and soil health can be adversely affected by stresses throughout the cultivation periods [ 5 ], increased salinity [ 7 ], acidity [ 8 , 9 ], soil composition and climate changes [ 10 , 11 ], tillage [ 12 , 13 ], cultivation methods [ 14 ], pesticides and heavy metals[ 15 ].Proper soil management strategies can reduce negative effects and restore the balance in the soil bacterial ecosystem, increasing soil health. Organic farming may overcome or reduce these effects [ 5 ], although it may reduce yield as much as 20% [ 14 ]. Balanced fertilization has also been found to increase soil biomass and activity [ 16 ]. The physical properties of soil also affect the bacterial community. Bacteria depend on microhydration for cell transport and nutrient diffusion. Sandy soils increase the number of isolated micro-habitats without crossed competition [ 17 ]. To guide management decisions, soil health is assessed using measurable quantities known as environmental indicators. Among these, ecological indicators reflect the balance of biological processes in the ecosystem. As such, soil ecological indicators are essential to assess soil health, suppressiveness potential, contamination and potentially deleterious effects [ 18 ]. In this study, we explore potential ecological indicators related to the rhizobacterial community. Such indicators should directly reflect the role of major biological determinants of soil health. Additionally, an ideal ecological indicator should be pervasive, display a rapid and proportional answer to external perturbations and be easily and economically measurable. In the absence of extensive rhizobacterial composition data, several indexes have been used as overall indicators of soil diversity and richness. Frequently used indicators are the Shannon and Simpson indexes, operational taxonomic unit (OTU) counts and richness estimators, such as ACE or CHAO1. Both the Shannon and Simpson diversity indexes reflect the uniformity of species abundance. OTU counts provide a direct estimation of genetic richness, but they require very large sample sizes. Partial OTU counts suffer from a strong dependence on sample size and may be compared only with similar sample sizes; they are usually reported together with their corresponding rarefaction curves. Richness estimators (CHAO1 and ACE) predict diversity from partial measures of OTU counts. They also depend on sample size, although their values can stabilize earlier, reducing sample size requirements [ 19 ]. Next generation sequencing (NGS) facilitates the analysis of rhizobacterial populations in greater detail, up to, in a few cases, practically saturating the existing diversity in a given soil [ 20 ]. NGS studies have shown that soil diversity is far greater than originally expected, in the order of several thousand species [ 14 , 15 ]. Sequences can be assigned to known taxa using a variety of methods, although typically, a large fraction (~30%) of the sequences collected cannot be ascribed to any known taxa . The availability of this new NGS data facilitates the identification of novel and more specific ecological indicators. In this study, we have considered the minimum sample size required by general complexity and diversity indicators, such as the Shannon and Simpson indexes, to produce realistically representative estimates. We have also considered the existence of taxonomic groups that may act as ecological indicators. We have compared rhizobacterial communities subject to different perturbations to identify associated taxonomic ecological indicators. Aiming for maximal generality, we paid special attention to groups that are commonly present in soil and that respond consistently to external influences. If such a group can be found, then small samples or relatively simple, specific tests might be designed to assert soil health. Finally, any indicator should be economic and easy to determine, so that corrective actions may be taken swiftly. Since NGS technology is not pervasive yet, we also paid attention to cultivable rhizobacteria, as they may be used in resource-starved environments.",
"discussion": "Discussion Given the lack of a powerful direct ecological indicator, soil ecological studies have relied on indirect indicators of overall diversity and richness computed from limited experimental data. Experiments have used phospholipid fatty acid analysis (PLFA), terminal restriction fragment length polymorphisms (T-RFLP), single-strand conformation polymorphism (SSCP), denaturing/temperature gradient gel electrophoresis (DGGE/TGGE) and sequencing of a small number of 16S metagenomic sequences. The main soil health indicators currently in use are the soil chemical and physical properties, OTU content and the Shannon and Simpson indexes. Soil physico-chemical properties will undergo little changes over time and may be considered as relatively invariant in most soils. Hence, while it may be used to compare different edaphic backgrounds, it is unlikely to change in response to external stimuli making a poor indicator of ecological changes. Actually, micro-habitat fragmentation in sandy soils results in greater richness without associated changes in the Shannon and Simpson indexes [ 17 ]. The sample-size dependence of direct diversity measures (OTU and OTU-predictors such as Chao1 or ACE) has already been studied. There is a large dependence of OTU counts on sample size, requiring large and uniform samples (on the order of 50,000 to 500,000 individual reads) to estimate the actual diversity of an agricultural soil. Diversity estimators, such as Chao1 or ACE, can reduce this dependence depending on the characteristics of the underlying population [ 19 ]. However, they still require relatively large sample sizes (usually in the order of at least 10,000 to 25,000 individual sequences) to reach asymptotic levels. To further reduce the sampling size and facilitate experimental detection of external influences, the next obvious choice is to use coarser indicators. The Shannon and Simpson indexes are generally used to provide an estimation of richness and diversity in current studies. We have analysed the dependence of both indexes on the number of individuals examined. Our results show that the Simpson index may provide a reasonable estimate of its actual population value with relatively small sample sizes (about 150 individual sequences) such as those collected in many non-NGS studies. However, recent NGS studies have shown that agricultural soils may host thousands of species. This suggests that Simpson's index, with its small sample size requirements, is actually responding only to major taxa , and may be less sensitive to changes in less-abundant taxa . Consequently, the Shannon index might reflect better the actual landscape of the soil rhizobacterial diversity. The Shannon index requires at least an order of magnitude more information to be reliable. More than 5,000 individual sequences are required to obtain a value that might be within 95% of the actual population value (or more than 18,000 for 99%). According to our observations, when using smaller sample sizes, the Shannon index should be used cautiously and, preferably, to compare results only between experiments with similar sample size. Experiments with a much larger sample size may be comparable even with unequal sample sizes. It should be noted that we have used a relatively coarse precision to build the rarefaction curves, and that the samples used do not represent the full spectrum of natural soils. The values reported for the mean should be considered only a rough approximation despite the small values of their confidence intervals. These C.I. values should be considered a measure of the relative dispersion observed, not as limits for the values of the average sample size. Since the Simpson index stabilizes too early, the Shannon index seems more suitable to consider the rich information obtained from the analysis of modern experiments that yield larger sample sizes. On the other hand, once we have access to this rich information, it makes sense to search for other potential indicators. A case in point would be identification of an indicator taxonomic group. Such a group might permit elaboration of specific tests for the detection of major environmental changes with less effort and at a lower cost. Previous work has shown the relative dependence of different bacterial taxa on environmental conditions. We have looked for taxonomic groups that may act as sensitive ecological indicators. A desirable group should be present in most, if not all, soils, behave consistently under external influences, should be easy to detect and, possibly, cultivable for cheaper detection in budget-constrained environments. The analysis of various approaches to taxonomical classification confirms that similarity based assignment provides the most fine-grained results. This increased detail requires a higher computational cost. While RDP and RTAX only require a few seconds or minutes to classify a large dataset in a modern computer, running blast against the VAMPS, Silva or RDP databases may require an overnight, or longer, calculation. On the other hand, until more data is available, many individual sequences will remain unclassifiable introducing a bias in the analysis. In addition, most studies do not exhaust species diversity as this requires extensive sequencing (in the order of hundreds of thousands of reads). As NGS technology progresses and data accumulates, we may expect these shortcomings to be less relevant. Despite current limitations, NGS remains the most accurate approach for metagenomic analysis. When using NGS data, similarity based taxonomy should be preferred whenever maximal detail is desired. However, RDP Classifier and RTAX provide a practical way to speed up analyses if practical concerns require a compromise for speed and if the sacrifice of the detail from the lowest taxonomic levels is acceptable. The presence of unspecific groups used to annotate environmentally-collected metagenomic sequences further affects similarity based classification. The wealth of new data demands novel approaches to taxonomical classification. In this work we decided to use only well-characterized groups, clearly reducing the information considered. A better approach might be to classify novel reads into new, provisionally labelled, taxonomic groups to enhance taxonomic imputation of newly collected sequences. Such an approach would parallel OTU clustering followed by taxonomic imputation and suffers similar drawbacks: there is a strong dependence of similarity clustering on sample characteristics; 16S rRNA hypervariable regions may constitute a genetic continuum blurring cluster limits; the correlation of the genetic diversity of 16S rRNA sequences to overall genome diversity is unclear; and the relationship of percent similarity with the definition of taxonomic levels is still open to debate. Furthermore, OTU clustering followed by taxonomy assignment comports an initial step where several individual reads are grouped and all of them are arbritrarily assigned the taxonomy of the cluster centroid (the group's chosen representative). We have chosen to classify taxonomically each read independently to increase precision and avoid biases associated to OTU clustering. Identification of common groups is hampered at lower taxonomic levels. This is most likely because the number of expected groups increases as we proceed towards the species level and, conversely, the number of individuals in any group will decrease. The proportion of extant, clearly defined taxonomic groups also decreases as we proceed towards the species level and, with current, reduced sample sizes, it is easy to miss and be unable to classify many lower level groups. The CCA analysis of all the samples included in the analysis revealed two common trends irrespective of the taxonomic level and classification method used. Time is often associated to the first canonical coordinate, and control/conventional cultures tend to cluster separately on the second component. These results support the feasibility of identifying rhizobacterial ecological indicators of external stress. The Indicator Value method has been successfully applied to test for indicator species using metagenomic data [ 30 ] of tilled vs. non-tilled crop production. Its application to our datasets is questionable as the classification of pesticide impact is normally based, among other parameters, on the effect on taxonomic diversity at higher (usually class or phylum) levels. Still, since CCA shows a clear separation of control and conventionally cultivated samples, it might be useful to pool these samples into a common group and look for indicators comparing it to the other samples separated by treatment. The analysis identified many potential indicators, usually grouping samples in accordance to previous classifications. These indicators are, however, subject to another criticism: current metagenomic studies explore only a minor fraction of the existing species (as deduced from OTU and estimator values). It is impossible to rule out that any absent species have not been missed by chance. If a species has been missed by chance, it may be erroneously identified as a false positive indicator. When all non-common species were removed from the study, no indicators were found at any level. This does not invalidate the indicators found, but as all of them might potentially be a false positive, we cannot rely on them without additional confirmation. The Indicator value approach groups many samples by treatment and relies on group size for the effects of non-treatment differences to cancel out. In this case, variation between experiments might plausibly obscure variation within independent experiments. We can overcome this by fixing all non-treatment variables. We have selected larger samples and compared each treated sample with its respective control from the same location, soil and, time/climate cycle. An interpretation of comparative taxonomic data is difficult due to various reasons. The treatments considered have varying degrees of impact on the soil ecosystem, thereby challenging the interpretation of the changes observed. They also have different time dependences, with some treatments displaying an initial impact that can quickly be recovered, while others may display small initial effects that accumulate and increase with time. Lastly, the information available pertains to a reduced number of soils that do not reflect the overall variability potentially existing in agricultural soils. With the taxonomic information currently available, the analyses show that there are many potential indicators associated to specific treatments. Many of these correspond to taxa catalogued generically as environmental samples, suggesting that many additional indicators are possibly present, although they have not been catalogued yet. Our results indicate that our current knowledge is unlikely to identify any rhizobacterial group at the various levels that is always associated to all aggressive conditions. Nevertheless, we have been able to identify at each level the groups most commonly associated with early or long-lasting aggression and these provide a useful starting point to diagnose deleterious effects in the soil environment. Until more data is collected, it is sensible to use changes in the proportion of any of these groups, or their combinations, as indicators of alterations in soil health. NGS is nowadays the most accurate method to explore the soil bacterial community, but depending on resources, it may compensate to compromise for less accurate detection methods. In the most restricted environments, detection of cultivable groups at higher levels may be an optional approximation. Choosing an appropriate group that may function as ecological indicator in the absence of NGS data is not easy. From the major taxa present in all soils, groups at various levels in the trees of Verrucomicrobia and Actinobacteria show a response to external influence, pervasiveness (within our samples) and low variability (as measured by NSD) so that their proportional variations in response to external influences should be easier to detect. Groups in the Proteobacteria , and more specifically, δ-proteobacteria tree are also commonly affected. At the species level, E . coli , Enterobacter cloacae and Pseudomonas are frequently affected in agreement with previously published work and, combined, provide a good indicator of long-lasting damage to the soil ecosystem. There are easy and efficient methods to determine the presence of Actinobacteria in soil [ 31 ], making them attractive as potential indicators. In addition, the proportion of Actinobacteria in any given soil tends to stabilize under normal conditions [ 32 ] being resilient to pH changes (which strongly affects other major phyla ) [ 9 , 13 ]. They are producers of antimicrobial compounds [ 6 ] and they are normally present in greater proportions in agricultural soil than Verrucomicrobia . Hence, the determination of Actinobacteria should be easier and require smaller samples. We have observed early and late changes in Actinobacteria , showing an increasing effect with more aggressive treatments, in the rhizobacterial communities of maize and cotton fields treated with various herbicides or herbicide combinations [ 10 , 15 , 23 , 33 ]. Other authors have reported similar changes associated to tillage and crop rotation/succession [ 12 ]; larger abundance of Actinobacteria in suppressive soils with greater changes after the addition of fungal pathogens [ 4 ]; long term fertilization and intensive herbivory plus mowing in grasslands [ 34 ]; intensive coffee farming [ 14 ]; desert versus cultivated soil [ 35 ]; and different depths of permafrost soil samples [ 36 ]. This suggests that relative changes in the proportion of Actinobacteria groups (for instance, actynomycetes) may act as potential ecological indicators when NGS technology is not available. However, given the relative scarcity of data available, these observations should be taken with caution until more information is collected. Our knowledge of the rhizobiome is still too limited and, as data accrues, new information will need to be considered, possibly demanding a reconsideration of these observations. An alternative approach is to consider the total number of changes observed: generally speaking, more aggressive treatments affect a larger number of taxa . This suggests that the total number of significantly altered taxonomic groups and the magnitude of the change in their proportional presence may also be used to define the impact of external influences. A statistical test of the number of changes weighted by their relative magnitude should provide a useful measure of the impact of external influences on soil health. Additional metagenomic data is needed to parametrize weights and measure correlations. The main drawback is that, to be useful, many sequences are needed to properly sample the rare microbiome. Comparison of complete rhizobiomes will yield more information and, with the extension of sequencing technologies, may become the option of choice. With the characterization of the complete rhizobiomes from sufficient diverse samples, the identification of a better taxon or group of taxa that may act as indicator will also become easier."
} | 6,703 |
39057343 | PMC11278419 | pmc | 8,835 | {
"abstract": "The symbiotic relationship between arbuscular mycorrhizal fungi (AMF) and plants is well known for its benefits in enhancing plant growth and stress resistance. Research on whether key components of the AMF colonization process, such as MyC factors, can be directly utilized to activate plant symbiotic pathways and key functional gene expression is still lacking. In this paper, we found that, using a hydroponics system with Lotus japonicus , MyC factor analogue chitin oligomer 5 (CO5) had a more pronounced growth-promoting effect compared to symbiosis with AMF at the optimal concentration. Additionally, CO5 significantly enhanced the resistance of Lotus japonicus to various environmental stresses. The addition of CO5 activated symbiosis, nutrient absorption, and stress-related signaling pathways, like AMF symbiosis, and CO5 also activated a higher and more extensive gene expression profile compared to AMF colonization. Overall, the study demonstrated that the addition of MyC factor analogue CO5, by activating relevant pathways, had a superior effect on promoting plant growth and enhancing stress resistance compared to colonization by AMF. These findings suggest that utilizing MyC factor analogues like CO5 could be a promising alternative to traditional AMF colonization methods in enhancing plant growth and stress tolerance in agriculture.",
"conclusion": "5. Conclusions In this study, we found that, compared to colonization by AMF, the addition of MyC factor analogue CO5, through the activation of relevant pathways, has a superior effect on promoting plant growth and enhancing stress resistance. It can regulate different stages of plant growth and development, promoting lateral root branching and activating signaling pathways related to nitrogen and phosphorus absorption, as well as plant stress resistance. In the future, it may be possible to produce seed coating agents containing MyC factor analogue CO5. Different crop seeds can be enveloped in CO5 at the optimal concentration. This seed coating, applied during the early stages of crop growth and development, could directly stimulate plants to activate symbiotic signaling pathways, thereby achieving effects comparable to those of AMF.",
"introduction": "1. Introduction The symbiotic relationship between plants and arbuscular mycorrhizal fungi is a common and beneficial interaction that exists in more than two-thirds of terrestrial plant species [ 1 ]. AMF obtain essential carbohydrates from their host plants and, in exchange, provide the plants with mineral nutrients and water absorbed from the soil [ 2 ]. Due to the lack of strict host specificity, AMF can establish connections between the root systems of different plants. Through the extraradical hyphal network, they facilitate the redistribution of resources among various plants, thereby maintaining the stability of plant community structure and function [ 3 ]. At the ecosystem level, AMF play a significant role in nutrient cycling and regulate the ecosystem’s response to environmental fluctuations, underscoring their irreplaceable ecological significance [ 4 , 5 ]. This symbiosis plays a crucial role in enhancing plant growth and stress tolerance. The expanded mycelium network of AMF is a key factor in enhancing the stress tolerance of host plants while also aiding in the absorption of water and nutrients from the soil [ 6 ]. Once a symbiotic relationship is established, plants can rely on AMF to enhance phosphorus uptake through phosphate transporters [ 7 ]. Additionally, increasing evidence suggests that AMF can absorb both NO 3 − and NH 4 + , thereby helping plants acquire the essential nutrient nitrogen for growth and development [ 8 , 9 ]. Furthermore, numerous studies have shown that under abiotic stress, the expression of stress-related genes in host plants is influenced by AMF through different response mechanisms, such as those for drought, heavy metals, and high salinity [ 10 , 11 , 12 ]. The establishment of AMF symbiosis involves a molecular dialogue between the plant and fungal partners. Our understanding of this molecular interaction has been greatly informed by studies on rhizobium–legume symbiosis [ 13 ]. It is believed that the symbiosis involving fungal associations predates bacterial symbiosis, leading to the hypothesis that rhizobia utilize ancient mycorrhizal signaling pathways by mimicking signals used by symbiotic fungi [ 14 ]. In this context, it has been proposed that AMF release signaling molecules known as mycorrhiza (MyC) factors that are essential for recognition by plant partners. MyC factors are a mixture of lipochitooligosaccharides (LCOs) and chitin oligomers (COs). AM fungi produce diffusible symbiotic signals containing a combination of sulphated and non-sulphated simple LCOs [ 15 ]. COs, like in colonization by arbuscular mycorrhizal fungi, can trigger calcium spikes in the epidermal cells of plant roots and serve as a well-characterized elicitor of plant responses. They can activate the plant’s innate immune response by inducing the expression of genes associated with biotic stress, responses to phytopathogenic fungi [ 16 ]. Studies have shown that COs exhibit higher activity compared to both sulfated and non-sulfated MyC LCOs in inducing AM-dependent calcium spiking. This calcium spiking response is dependent on genes within the common SYM signaling pathway, particularly DMI1 and DMI2 [ 17 ]. This molecular dialogue between plant and fungal partners, mediated by MyC factors and COs, plays a crucial role in the establishment and functioning of AMF symbiosis. In practical production applications, the establishment of a symbiotic relationship requires the extensive propagation of fungal spores. Additionally, the intricate conditions for spore propagation contribute to increased costs in preparing durable AM microbial fertilizer products. Therefore, it is crucial to investigate whether MyC factor analogues demonstrate symbiotic and growth-promoting effects, as well as stress resistance, by activating gene expressions associated with plants’ symbiotic pathways. Utilizing MyC factor analogues as regulators for plant growth in AMF symbiosis shows significant potential for enhancing plant growth and stress resistance in agricultural production. To achieve this goal, we selected a representative chitin oligomer, CO5, as the material for investigating its potential growth-promoting effects on Lotus japonicus , like symbiosis with AMF. This investigation seeks to uncover the influence of externally applied MyC factor analogues on the growth, stress resistance, and signaling pathways of Lotus japonicus roots. By doing so, we aim to establish both experimental and theoretical frameworks for the potential application of MyC factor analogues as symbiotic regulators.",
"discussion": "4. Discussion AMF are important symbiotic microorganisms in soil that form arbuscular mycorrhizal associations with host plants. Once a stable symbiotic relationship is established with plants, AMF engage in nutrient exchange through arbuscular branches and hyphae, benefiting both the fungi and the plants [ 24 ]. Studies have shown that AMF produce diffusible fungal signals known as MyC factors, which can be derived from AM hyphae, germinating spores, or mycorrhizal roots [ 25 ]. These MyC factors are primarily composed of lipochitooligosaccharides and short-chain chitin oligomers [ 15 , 17 ]. Previous research has indicated that MyC factors can induce intense Ca 2 + spiking and the associated intensity [ 15 , 17 , 26 ], highlighting their potential role in signal transduction, cell communication, and environmental stimuli responses [ 27 ]. The main objective of this study was to investigate whether a representative MyC factor analogue CO5 could effectively replace the inoculation of AM fungi and achieve comparable or even superior effects compared to AMF symbiosis. This research aimed to explore the potential of CO5 in promoting plant growth and enhancing stress resistance, ultimately contributing to the development of innovative strategies for sustainable agriculture. In the experiments, the addition of the MyC factor analogue CO5 could simulate the symbiotic effect of AM fungi with Lotus japonicus . And this simulation promotes lateral root branching, enhances primary root length, and increases underground fresh weight, thereby achieving a growth-promoting effect on Lotus japonicus ( Figure 1 ). This is consistent with the work of Sonja Kosuta et al. in 2003, which facilitated lateral root branching to promote the growth of Medicago truncatula [ 28 ]. Low-molecular-weight forms of chitin, such as CO5, may have a role in nature as biostimulators of plant growth. Additionally, they are recognized as a direct source of carbon and nitrogen for biomass [ 16 ]. Furthermore, our experiments indicate that the addition of CO5 can assist Lotus japonicus in better adapting to adverse stress conditions ( Figure 2 ). The stress resistance effects resulting from the symbiosis between arbuscular mycorrhizal fungi and plants have been extensively studied [ 29 , 30 , 31 , 32 , 33 ]. Our research demonstrates that the addition of the isolated MyC factor analogue CO5 also exhibits stress resistance effects. This is consistent with the findings of Devanshi Khokhani et al., who summarized that substances like chitin oligomers can serve as elicitors of plant immune responses, activating defense signaling pathways [ 34 ]. Through RNA-seq, we observed that, compared to colonization by AMF, the sole addition of CO5 resulted in a greater number of differentially expressed genes, involving variations in enrichment across various GO and KEGG pathways. The prominent pathway identified was the integral component of the membrane (GO:0016021) and linoleic acid metabolism (map00591) ( Figure 4 ). Integral components of the membrane refer to membrane proteins that play a crucial role in cellular signal transduction, cellular recognition, and transport [ 35 ]. Linoleic acid functions in plant defense signaling and stress tolerance [ 36 ]. For instance, the Arabidopsis FAD2 omega-6 desaturase localizes to the endoplasmic reticulum and has been demonstrated to be crucial for salt tolerance during germination [ 37 ]. In the RAM1 gene, which is specifically required for AMF symbiosis [ 38 ], the gene expression in the AMF group is significantly higher than that in the CO5 group at 60 days. During the early stages of symbiosis, the main signal communication between AMF and host plants involves the mutual recognition of small molecular substances [ 39 ]. Upon the mycelium reaching the endodermis, an essential interaction for the subsequent symbiosis, namely the formation of arbuscular structures, involves the indispensable interaction with CYCLOPS, where DELLA plays a crucial role. Simultaneously, DELLA also plays a significant role in maintenance degeneration [ 40 , 41 ]. RAD17 is also a member of the GRAS family protein, and it has been reported to possess functions in meiosis [ 42 ], being required for responses to DNA damage, replication stress, and double-strand break (DSB) repair [ 43 ]. As a member of the same family as RAD1, RAD17 may also interact with RAM1. The establishment of AMF symbiotic relationships requires the involvement of plant root-derived strigolactones (SLs) [ 44 ]. And CCD7 encodes a key enzyme in the synthesis of SLs [ 45 ]. In the maintenance degeneration stage, apart from DELLA, MYB and NSP are also two crucial transcription factors. MYB, on one hand, could act as a regulator of some mycorrhizal-responsive genes in arbusculated cells, and on the other, it could be involved in the mechanisms that regulate root growth [ 27 ]. NSP1 and NSP2 coordinate fungal and plant processes associated with LCO production and perception, facilitating AMF symbiosis [ 46 ]. In the relative expression levels obtained from the RNA-seq, except for RAM1, the expression of other key genes in the CO5 group is either higher or not significantly different from that of the AMF group. It has been demonstrated that at the symbiotic level, the addition of CO5 can, to some extent, substitute the role of AMF. Ammonium is a preferred source of nitrogen for plants, and AMTs play an essential role in NH 4 + uptake [ 47 ]. Many members of the nitrate transporter 1/peptide transporter family (NPF), including NPF and NRT, are also involved in the translocation of nitrogenous compounds such as nitrate, amino acids, peptides, and plant hormones [ 48 , 49 ]. A transcriptome analysis revealed that the down-regulation of NRT3.1 nitrate transporter genes plays a significant role in the rhizobium symbiosis-induced tolerance of Medicago truncatula to arsenate [ 50 ]. After the formation of arbuscular structures, there is a substantial exchange of nutrients between AMF and host plants. For example, the phosphate transport protein PT, induced by AM fungi, begins to acquire phosphate from the membranes surrounding the arbuscules [ 51 , 52 ]. SWEET transporters represent a unique class of sugar transporters that play crucial roles in various developmental and physiological processes in plants [ 53 ]. Simultaneously, they are closely associated with pest resistance [ 54 ]. Plasma membrane intrinsic proteins (PIPs) in root cortical cells are crucial for water uptake in plants and are believed to be directly involved in cell growth [ 55 ]. Aquaporins, to which PIPs belong, are membrane channels that play an essential role in maintaining cellular water and osmotic homeostasis in plants under both normal and water-deficit conditions [ 56 ]. In these genes related to the regulation of nitrogen and phosphorus transport as well as stress resistance, the gene expression levels in the CO5 group are significantly higher than in the AMF group at 60 days. This demonstrates that the addition of CO5 surpasses AMF in promoting growth and stress resistance levels. However, the specific mechanisms by which isolated CO5 can substitute for the symbiotic effects of AMF still require further investigation. As we described before, in previous studies, MyC factors, including CO5, have been identified as the primary signaling molecules activating symbiotic responses, such as calcium oscillations, during the symbiosis between plants and AMF [ 57 ]. Although there are no related reports, we believe that CO5 plays a crucial role in symbiosis and can function independently of it. For example, strigolactones, which are other plant signaling molecules of AMF, can also function as plant hormones, independently promoting plant growth outside of the symbiotic process [ 58 , 59 ]."
} | 3,657 |
28150713 | PMC5288723 | pmc | 8,836 | {
"abstract": "Flux coupling analysis is a computational method which is able to explain co-expression of metabolic genes by analyzing the topological structure of a metabolic network. It has been suggested that if genes in two seemingly fully-coupled reactions are not highly co-expressed, then these two reactions are not fully coupled in reality, and hence, there is a gap or missing reaction in the network. Here, we present GAUGE as a novel approach for gap filling of metabolic networks, which is a two-step algorithm based on a mixed integer linear programming formulation. In GAUGE, the discrepancies between experimental co-expression data and predicted flux coupling relations is minimized by adding a minimum number of reactions to the network. We show that GAUGE is able to predict missing reactions of E. coli metabolism that are not detectable by other popular gap filling approaches. We propose that our algorithm may be used as a complementary strategy for the gap filling problem of metabolic networks. Since GAUGE relies only on gene expression data, it can be potentially useful for exploring missing reactions in the metabolism of non-model organisms, which are often poorly characterized, cannot grow in the laboratory, and lack genetic tools for generating knockouts.",
"conclusion": "Conclusion In the present work we have developed a gap analysis method, GAUGE, to resolve the cases where in silico flux coupling relationships is not in agreement with experimental gene co-expression patterns. GAUGE resolves the inconsistencies by adding reactions from KEGG database, changing the reversibility type of reactions or allowing exchange of metabolites between cytoplasm and extracellular space. We tested GAUGE on i JR904 metabolic network model of E. coli as a model that we know contains a large number of gaps. We were able to find out missing reactions that may not be recognizable by other gap filling methods. Therefore, GAUGE can be used as an alternative and complementary strategy for gap filling of metabolic networks. Usually, those methods that use topological flaws of the network such as dead-end metabolites are preferred for gap filling, since these methods can be applied without the need for an experimental dataset. For instance, obtaining gene essentiality data for every gene in an organism is not a simple task and such data is not available for many organisms. A benefit of GAUGE is that it uses a type of experimental data which is readily available for many organisms. Another beneficial feature of GAUGE is that there is the possibility to find globally optimal solutions, instead of finding solutions to solve the inconsistencies case by case. This is the approach that is also considered in very recent study of Hartleb et al . 38 . In this study, the authors present G lobal F it , a bi-leveloptimization method, which identifies minimal set of model changes to achieve a model that can correctly predict all of the experimental data of growth and non-growth cases. Here, we have validated our results by searching in the literature and databases and also by performing BLASTP to find genomic evidence of genes. It should be noted that there is a new version of E. coli metabolic network model, i JO1366, which is reconstructed in 2011 39 . We have also looked for the predicted reactions in this version of the model. Interestingly, some of these reactions are included in i JO1366. These reactions are presented in Supplementary Tables S2, S3, S4 and S5 . We should note that in the universal dataset of reactions used in this study, all of the reactions are included without considering their directionalities. It is definitely a valuable analysis to compute the Gibbs free energy change for each reaction and see in which direction it will carry flux. However, this is not a necessary step in validation of our gap filling results, since addition of reactions in each of the two directions will resolve the identified gaps. More clearly, as it is shown in Fig. 2 , addition of reactions 5 to 11 (in forward or reverse directions) will change the coupling type of reactions 1 and 2 and reactions 3 and 4 from fully coupled to directionally coupled. If one needs to add the predicted reactions of GAUGE to a metabolic network, the Gibbs free energy changes should necessarily be computed to know in which direction the reactions should be added. One should note that there are not a large number of inconsistent reaction pairs in the model. As shown in Supplementary Figure S3 , the majority of genes are involved in a low number of full coupling relations, while a large number of genes are not fully coupled to any other genes (not shown in the graph). In addition, Supplementary Figure S4 shows that those genes which are associated with larger number of reactions are generally involved in lower number of full coupling relations. Therefore, fully coupled gene pairs which are associated to fully coupled reaction pairs are not frequent in metabolic networks. However, the results presented here show that even in these situations, GAUGE can successfully predict the novel reactions for being added to the model. Another point is that, other inconsistencies may exist between experimental gene co-expressions and theoretical flux coupling relations. One such inconsistency is when a highly co-expressed gene pair is not associated to fully coupled reactions. However, in this case one cannot draw any conclusion about the incorrectness of the model. The high co-expression may exist, for example, for functionally related genes, while these genes should not necessarily be fully coupled. Another point is that if some specific biochemical pathways are activated in the cell, some genes may not be highly co-expressed anymore. The environmental conditions which activate these pathways may not be captured in the experimental gene expression data. Therefore, having highly co-expressed gene pairs with no fully coupled reactions do not mean that the model should be modified, e.g., reactions should be deleted from the model. Using more comprehensive gene expression data may decrease the number of such inconsistencies. Furthermore, as our results suggest, only certain gaps are found, and can be filled, based on gene expression data. Moreover, regulation of protein expression may occur at the post-transcriptional level, which again means that gene expression data might not be sufficient for a comprehensive gap finding. Despite these shortcomings, we show that GAUGE can be used in practice to find and fill the metabolic gaps, and its performance is comparable to the other well-known widely used gap filling tools. Therefore, it is relevant to use transcriptional level gene expression data for gap filling. GAUGE is presented here as a potential strategy for gap analysis of metabolic networks that predict different sets of reactions for addition to the model. Several parameters can be adjusted for improving the predictive power of GAUGE. Setting different thresholds for low and high correlation coefficients are one such parameter. Additionally, instead of computing Pearson correlation coefficients of gene expressions, the Boolean version of expression values may be considered, i.e., expression values higher and lower than a certain cut-off are considered as expressed or not expressed, respectively. Then, the genes which are always expressed together can be identified and labeled as fully coupled gene pairs. Another possibility is to use other measures of correlation like mutual information, instead of Pearson correlation coefficients. The gene expression data can also have an important effect on the results predicted by GAUGE. As mentioned above, completeness of dataset can affect the correlation of gene expressions, which in turn may affect the inconsistencies found between experimental observations and model predictions. For obtaining an optimized version of GAUGE with more reliable results all of the above mentioned points should be taken into account. One may also think of using protein abundance data, e.g., from proteomic databases, instead of gene co-expression data. However, we should note that protein abundance data may contain more noise, including more false negatives, compared to gene expression data, which in turn may result in more unreliable predictions. On the other hand, in the study of Notebaart et al . 21 it is shown that there is a good correlation between gene co-expressions and flux coupling relations. We should also note that, one way to improve the GAUGE predictions is to use BLAST-weighted dataset of reactions like strategies used recently 18 19 . This way, the presence of unrelated or orphan reactions may be reduced in possible solutions of GAUGE.",
"discussion": "Results and Discussion Application of GAUGE: E. coli as a case study Here, we describe the use of GAUGE to resolve the inconsistencies between experimental gene co-expression data and in silico flux coupling relationships of the i JR904 metabolic network of E. coli 23 . Characterizing a pair of metabolic genes as “correlated” or “uncorrelated” pair requires a cutoff for computed values of co-expressions. In this study we define L as the set of reaction pairs with absolute correlation coefficients of less than 0.2 and H as a set of reaction pairs with absolute correlation coefficients of greater than 0.8. The goal of GAUGE is to change the coupling type of the reaction pairs in L while keeping the coupling type of the reaction pairs in H unchanged. For the i JR904 metabolic model, L and H include 134 and 41 reaction pairs, respectively. The existence of each inconsistent reaction pair in L implies that there are some missing reaction(s) in the network. Addition of such reactions to the model will change the coupling type of the reaction pair in L from full coupling to other types of flux (un) coupling relations. We used GAUGE to resolve the inconsistencies by adding reactions from KEGG and changing the reversibility type of reactions or by adding exchange reactions to the model. Computing globally optimal solutions for resolving the inconsistencies Resolving the inconsistencies can be done by two different approaches: all inconsistencies can be resolved at once or they can be resolved once at a time. Clearly, these two approaches may result in different set of predicted gap-filling reactions. Figure 2 is a simple example that shows this difference. In this figure, suppose that (1 and 2) and (3 and 4) are inconsistent reaction pairs. Trying to resolve these inconsistencies one by one, results in the addition of reactions 5 and 6 for pair (1 and 2) and reactions 10 and 11 for pair (3 and 4). However, if these cases are resolved together, reactions 7, 8 and 9 are the minimal set of reactions that are needed for addition to the network. In order to have a globally minimal solution, we input the inconsistent reaction pairs all at once to the first step of the algorithm to calculate the maximum number of these cases that could be resolved. GAUGE identified consistency-returning suggestions for 132/134 pairs of L . Out of the 132 inconsistency cases, 54 cases were resolved by adding reactions from KEGG, 2 cases were resolved by forcing irreversible reactions to have flux in the backward direction, and the others by allowing the exchange of metabolites between extracellular space and cytoplasm. At minimum, addition of 31 KEGG reactions and 18 exchange reactions and changing the reversibility type of 1 reaction are needed to resolve the inconsistencies of these 132 cases. The detailed information about these results and the procedure of computing alternative solutions are described in the Supplementary file . Here we discuss a few examples of resolved inconsistencies by GAUGE predictions that evidence from databases or literature exist for their presence in E. coli . Methylglyoxal metabolism Figure 3a shows a part of Methylglyoxal metabolism. Reactions GLYOX and MGSA are two reactions in this pathway which are fully coupled in i JR904 with Pearson correlation coefficients of less than 0.2. GAUGE predicts 5 separate reactions to resolve the inconsistency in this case. Interestingly, for two of these reactions (R02260 and R09796) evidence can be found for their presence in E. coli 32 33 34 . In addition, one other reaction, R00203, is catalyzed by an enzyme which is known to be encoded in the E. coli K12 genome. More precisely, R00203 is catalyzed by lactaldehyde dehydrogenase (E.C. number 1.2.1.22). This enzyme is encoded in E. coli genome and catalyzes the conversion of l -lactaldehyde to l -lactate. It is also shown that this enzyme catalyzes the conversion of methylglyoxal to pyruvate (reaction R00203) in E. coli 35 . However, the K m for this conversion is higher compared to the conversion of l -lactaldehyde to l -lactate. It should be noted that R00203 and R00205 in the Figure, differ in the cofactor used by their catalyzing enzymes. Folate metabolism Figure 3b shows part of the Folate metabolism pathway. In this Figure, ADCS and DHPS2 are inconsistent reaction pairs. GAUGE predicts R03066 to be added to the model. This reaction is catalyzed by dihydropteroate synthase (E.C. number 2.5.1.15) which is encoded by a gene present in E. coli genome (b3177). Additionally, based on the KEGG database this reaction is present in folate metabolism pathway of E. coli K12. Tartrate metabolism TARTD and TARTRt7, as another inconsistent pair found by GAUGE, are shown in Fig. 3c . GAUGE predicts the addition of R01751, which is catalyzed by tartrate decarboxylase (E.C. number 4.1.1.73). d -malate oxidase is an enzyme with the same E.C. number which is encoded by b1800 gene in E. coli model, and interestingly, this enzyme is also annotated as “putative tartrate dehydrogenase” in E. coli . Purine and pyrimidine biosynthesis DHORTS and ORPT are inconsistent reaction pairs in Fig. 3d . GAUGE could not identify any reactions in KEGG to resolve the inconsistency of this case. In addition, no change in reversibility types can resolve it. However, GAUGE predicts the addition of exchange reactions for orotate or s -dihydroorotate. The gene for transporting orotate to the cell is also known to be present in E. coli 36 . Comparison of GAUGE results with other gap filling methods We have run GapFind/GapFill 29 , Smiley 12 and GrowMatch 13 algorithms on the same metabolic network, to compare their results with GAUGE. We should emphasize here that there are no standard benchmark for comparing gap analysis methods. Each method uses different kind of inputs and searches for different types of gaps. In addition, false negativity, true negativity, or even false positivity cannot be defined for the results of gap analysis methods, since a comprehensive and perfect knowledge about the metabolism of organisms does not exist. Therefore, we can only try to compare the results of different gap analysis methods by searching for evidence for the reactions predicted by each method and calculating the frequency of supported predictions for each method. GrowMatch solves two MILPs to add and remove reactions for resolving the NGG ( in silico no growth vs. in vivo growth) and GNG ( in silico growth vs. in vivo no growth) cases respectively. Since GAUGE only predicts reactions for being added to the model, only the MILP for resolving NGG cases was run to obtain comparable results. Altogether, 37 NGG cases were identified. Every NGG case was used separately and all of the alternative optimal solutions were calculated for each case. From these cases, 18 cases could be resolved using one of the three possible strategies, namely, addition of reactions from KEGG, changing irreversible reactions to reversible ones, and addition of exchange reactions. The total of 69 reactions were predicted for being added to the model or changing their reversibility type. Smiley is a method that resolves the inconsistency between observed in vivo growth phenotypes and predicted in silico growth patterns. This algorithm uses information of growth profiles on different carbon and nitrogen sources as inputs and solves an MILP formulation to add minimum number of reactions to the model to resolve false negative model predictions. Reactions were selected from KEGG dataset or dataset of exchange reactions. Using Smiley, 34 false negatives were identified and 17 out of these 34 cases could be resolved. By calculating all alternative solutions, the algorithm predicted a total number of 55 reactions for gap filling. GapFind/GapFill finds no-production metabolites in the model and add minimal set of reactions to restore the connectivity of these metabolites to the rest of the network. Using GapFind, 64 inconsistent metabolites were found in i JR904. From these cases, 63 cases could be resolved using one of the three possible strategies, namely, addition of reactions from KEGG, changing irreversible reactions to reversible ones, and addition of exchange reactions. This method predicts 84 reactions for addition to the model or changing the reversibility type. Since all these algorithms resolve the inconsistency cases one by one, to obtain comparable results, we input each inconsistent reaction pair separately to GAUGE and identified all of the possible alternative optimal solutions for each case. GAUGE predicts 89 reactions as the candidates for being added to the model or being made reversible. In the next step, the correctness of the predictions of each algorithm was validated by: looking for the presence of a link between these reactions and a gene in E. coli genome annotations in KEGG database. In other words, if, according to the KEGG database, a gene from E. coli genome can code for the catalyzing enzyme of the predicted reaction, we suppose that this reaction can occur in this organism. performing BLASTP against the E. coli K12 genome. More precisely, the best hits in the E. coli genome which have the BLASTP E value of less than 10 −20 are considered as potential coding genes for the predicted enzyme activities in E. coli . We also searched the literature to find evidence regarding the presence of enzyme activities in E. coli which are predicted by GAUGE. The detailed information about the predicted reactions by each method is presented in the Supplementary Tables S2, S3, S4 and S5 . Figure 4 shows the percentage of correct predictions of each algorithm. As shown, GapFind/GapFill and Smiley have the most successful predictions. This observation is presumably due to the logic behind these algorithms. Smiley tries to correct the false negative predictions of the model grown on different media. When the cell can grow in vivo on a media, it must have the capability to convert the available nutrients to biomass precursors. Therefore, the failure of in silico prediction definitely implies missing reactions which leads to the precise predictions by Smiley. The same is true for GapFind/GapFill. It is not reasonable to have a metabolite in the cell with no production route and GapFind searches for these metabolites. Therefore, there is a high probability that its predictions are correct. The reason for less correctness of GrowMatch predictions may be that the presence of NGG cases are not necessarily because of missing reactions. For example, the reason may be that another isozyme which is missing from the model catalyzes the same reaction. In case of GAUGE, the algorithm was tested based on gene co-expression data obtained by Pearson correlation. Using more accurate and complete gene expression datasets, choosing different thresholds for the co-expression of genes, and the application of better correlation measures 37 can potentially improve the predictions of GAUGE. “Orthogonality” of gap analysis methods Figure 5 shows the Venn diagram of the number of gap-filling reactions predicted by each method. As shown, among the reactions predicted by GAUGE only three reactions are in common with the results of Smiley, GrowMatch or GapFind/GapFill. If only positively validated reactions are considered, there will be no common reactions between predictions of GAUGE with other methods (See Supplementary Figure S2 ). These results show that GAUGE can predict different sets of reactions for being added to the model during gap filling. This finding was expected, as the logic behind our method is principally different from other gap filling approaches. In other words, GAUGE can be used as a complementary strategy to the existing strategies for filling the gaps of metabolic networks. We also investigated in which biological pathways, the predicted gap-filling reactions are involved. Only the reactions which were positively validated, are considered. The pathways for each method are shown in Supplementary Figure S3 . As it is shown, there are some biological pathways that are only captured by one single method. For example, lipopolysaccharide metabolism, riboflavin metabolism, nitrogen metabolism, valine, luecine and isoleucine degradation, lysine biosynthesis, d -alanine metabolism, and synthesis and degradation of ketone bodies are pathways that are only identified by GAUGE. In other words, Smiley, GrowMatch and GapFind/GapFill are found to be unable to explore the missing reactions of these pathways. Identification of biological pathways that are unique to only one method shows that each gap analysis method examines specific parts of the metabolism that are not considered by other methods. This may be the result of the fact that each method looks for model errors from a particular point of view, and application of popular methods like Smiley, GrowMatch or GapFind/GapFill does not eliminate the necessity of application of GAUGE. As a final note, it should be mentioned that although GAUGE is based on MILPs, in practice it works acceptably fast. For example, for i JR904 network, when inconsistencies are resolved one by one, the mean computation time of GAUGE is ~13 seconds on a PC. The computation time of resolving inconsistencies all at once is ~30 minutes. Robustness analysis of GAUGE In order to investigate how sensitive GAUGE is to the lack of GPRs, we randomly removed some of the GPRs from the model and performed GAUGE on them. Two groups of 100 random networks were generated in which 10 and 40 percent of the reaction GPRs were removed respectively. Then for each network, gene coupling relations were calculated and inconsistent reaction pairs were identified. All of the alternative solutions were computed for resolving each inconsistency. The results are shown in Fig. 6 . When we remove 40 percent of the reaction GPRs from the E. coli network, the accuracy of predictions decreases from 36 percent to about 30 percent. Therefore, GAUGE predictions is not significantly affected by the varying degrees of coverage of the GPRs. This 6 percent reduction in accuracy is probably due to the fact that by deletion of some GPRs some of the genes will become fully coupled to each other. GAUGE will mistakenly predict some reactions to be added to the model for resolving inconsistency of these cases. Another method for robustness analysis of GAUGE, is to randomly remove reactions from the model and analyze what percentage of them could be returned back using GAUGE. In the Supplementary file , we explain that this analysis is not suitable for evaluating GAUGE, since there is not a high probability that removed reactions are associated to a fully coupled gene with low co-expression."
} | 5,901 |
21076396 | null | s2 | 8,837 | {
"abstract": "The chemotaxis signalling network in Escherichia coli that controls the locomotion of bacteria is a classic model system for signal transduction. This pathway modulates the behaviour of flagellar motors to propel bacteria towards sources of chemical attractants. Although this system relaxes to a steady state in response to environmental changes, the signalling events within the chemotaxis network are noisy and cause large temporal variations of the motor behaviour even in the absence of stimulus. That the same signalling network governs both behavioural variability and cellular response raises the question of whether these two traits are independent. Here, we experimentally establish a fluctuation-response relationship in the chemotaxis system of living bacteria. Using this relationship, we demonstrate the possibility of inferring the cellular response from the behavioural variability measured before stimulus. In monitoring the pre- and post-stimulus switching behaviour of individual bacterial motors, we found that variability scales linearly with the response time for different functioning states of the cell. This study highlights that the fundamental relationship between fluctuation and response is not constrained to physical systems at thermodynamic equilibrium but is extensible to living cells. Such a relationship not only implies that behavioural variability and cellular response can be coupled traits, but it also provides a general framework within which we can examine how the selection of a network design shapes this interdependence."
} | 391 |
25759637 | PMC4338664 | pmc | 8,838 | {
"abstract": "Bio-inspired asynchronous event-based vision sensors are currently introducing a paradigm shift in visual information processing. These new sensors rely on a stimulus-driven principle of light acquisition similar to biological retinas. They are event-driven and fully asynchronous, thereby reducing redundancy and encoding exact times of input signal changes, leading to a very precise temporal resolution. Approaches for higher-level computer vision often rely on the reliable detection of features in visual frames, but similar definitions of features for the novel dynamic and event-based visual input representation of silicon retinas have so far been lacking. This article addresses the problem of learning and recognizing features for event-based vision sensors, which capture properties of truly spatiotemporal volumes of sparse visual event information. A novel computational architecture for learning and encoding spatiotemporal features is introduced based on a set of predictive recurrent reservoir networks, competing via winner-take-all selection. Features are learned in an unsupervised manner from real-world input recorded with event-based vision sensors. It is shown that the networks in the architecture learn distinct and task-specific dynamic visual features, and can predict their trajectories over time.",
"introduction": "1. Introduction Humans learn efficient strategies for visual perception tasks by adapting to their environment through interaction, and recognizing salient features. In contrast, most current computer vision systems have no such learning capabilities. Despite the accumulated evidence of visual feature learning in humans, little is known about the mechanisms of visual learning (Wallis and Bülthoff, 1999 ). A fundamental question in the study of visual processing is the problem of feature selection: which features of a scene are extracted and represented by the visual cortex? Classical studies of feature selectivity of cortical neurons have linked neural responses to properties of local patches within still images (Hubel and Wiesel, 1962 ; Olshausen and Field, 1997 ). Conventional artificial vision systems rely on sampled acquisition that acquires static snapshots of the scene at fixed time intervals. This regular sampling of visual information imposes an artificial timing for events detected in a natural scene. One of the main drawbacks of representing a natural visual scene through a collection of snapshot images is the complete lack of dynamics and the high amount of redundancy in the acquired data. Every pixel is sampled continuously, even if its output value remains unchanged. The output of a pixel is then unnecessarily digitized, transmitted, stored, and processed, even if it does not provide any new information that was not available in preceding frames. This highly inefficient use of resources introduces severe limitations in computer vision applications, since the largely redundant acquired information lead to a waste of energy for acquisition, compression, decompression and processing (Lichtsteiner et al., 2008 ). Biological observations confirm that still images are largely unknown to the visual system. Instead, biological sensory systems are massively parallel and data-driven (Gollisch and Meister, 2008 ). Biological retinas encode visual data asynchronously through sparse firing spike trains, rather than as frames of pixel values (Roska and Werblin, 2003 ). Current studies show that the visual system effortlessly combines the various features of visual stimuli to form coherent perceptual categories relying on a surprisingly high temporal resolution: the temporal offsets of on-bistratified retina cells responses show an average standard deviation of 3.5 ms (Berry et al., 1997 ; Uzzell and Chichilnisky, 2004 ). Neurons in the visual cortex also precisely follow the temporal dynamics of the stimuli up to a precision of 10 ms. In order to bridge the gap between artificial machine vision and biological visual perception, computational vision has taken inspiration from fundamental studies of visual mechanisms in animals (Hubel and Wiesel, 1962 ; Wallis and Rolls, 1997 ). One main focus of these approaches have been various computational models of simple and complex cells in the primary visual cortex (V1) Hubel and Wiesel ( 1962 ); Fukushima ( 1980 ); Riesenhuber and Poggio ( 1999 ), which are characterized by their preferred response to localized oriented bars. Typically, this orientation-tuned response of V1 cells has been modeled with Gabor Filters (Gabor, 1946 ), which have been used as the first layer of feature extraction for visual recognition tasks (Huang et al., 2004 ; Ilonen et al., 2007 ). The most well-known example of biologically inspired, although still frame-based model of object recognition is the HMAX model (Riesenhuber and Poggio, 1999 ; Serre et al., 2006 ; Mutch et al., 2010 ). It implements a feedforward neural network based on a first layer of Gabor filters followed by different layers realizing linear and non-linear operations modeled on primate cortex cells. However, HMAX like other approaches implementing neural networks to perform visual tasks (Lin and Huang, 2005 ) are still based on processing still images and therefore cannot capture key visual information mediated by time. This paper introduces an unsupervised system that allows to extract visual spatiotemporal features from natural scenes. It does not rely on still images, but on the precise timing of spikes acquired by an asynchronous spike-based silicon retina (Lichtsteiner et al., 2008 ). The development of asynchronous event-based retinas has been initiated by the work of Mahowald and Mead (Mead and Mahowald, 1988 ). Neuromorphic asynchronous event-based retinas allow new insights into the capabilities of perceptual models to use time as a source of information. Currently available event-based vision sensors (Delbruck et al., 2010 ; Posch et al., 2011 ) produce compressed digital data in the form of time-stamped, localized events, thereby reducing latency and increasing temporal dynamic range compared to conventional imagers. Because pixel operation is now asynchronous and pixel circuits can be designed to have extremely high temporal resolution, silicon retinas accomplish both the reduction of over-sampling of highly redundant static information, as well as eliminating under-sampling of very fast scene dynamics, which in conventional cameras is caused by a fixed frame rate. Pixel acquisition and readout times of milliseconds to microseconds are achieved, resulting in temporal resolutions equivalent to conventional sensors running at tens to hundreds of thousands of frames per second, without the data overhead of conventional high-speed imaging. The implications of this approach for machine vision can hardly be overstated. Now, for the first time, the strict temporal resolution vs. data rate tradeoff that limits all frame-based vision acquisition can be overcome. Visual data acquisition simultaneously becomes fast and efficient. A recent review of these sensors can be found in Delbruck et al. ( 2010 ) and Posch et al. ( 2014 ). Despite the efficiency of the sensor representation, it is far from straightforward to port methods that have proven successful in computer vision to the event-based vision domain. Much of the recent success of computer vision comes from the definition of robust and invariant feature or interest point extractors and descriptors (Lowe, 1999 , 2004 ; Bay et al., 2008 ). Although such methods have proven to be very useful for static image classification, they require processing of the whole image, and do not take temporal information into account. Dynamical features for event data should instead recognize features only from novel visual input, and recognize them as they appear in the sparse input stream. This requires a model that can continuously process spiking inputs, and maintain a representation of the feature dynamics over time, even in the absence of input. Here we present an architecture for feature learning and extraction based on reservoir computing with recurrent neural networks (Schrauwen et al., 2007 ), which integrate event input from neuromorphic sensors, and compete via a Winner-Take-All (WTA) technique to specialize on distinct features by predicting their temporal evolution. A proof of concept for the performance of the architecture is demonstrated in three experiments using natural recordings with event-based vision sensors. In the first experiment, we present a set of oriented bars to the camera in order to the show the capacity of the model to extract simple features in an unsupervised manner, using a big spatial receptive field to emphasize the graphical visualization of the learnt features. In the second experiment, the full capacity of the method is demonstrated by mapping the field of view to several small receptive fields, and showing that the model is still capable of reliably extracting features from the scene. The last experiment applies the architecture to complex object features. All experiments were conducted with real-world recordings from DVS cameras (Lichtsteiner et al., 2008 ), and thus are subject to the standard noise distribution of such sensors.",
"discussion": "4. Discussion This article presents a new architecture for extracting spatiotemporal visual features from the signal of an asynchronous event-based silicon retina. The spatiotemporal signal feeds into the system through a layer of ESN, which compute predictions of future inputs. An unsupervised learning process leads to specialization of ESNs to different features via WTA competition, which selects only the best predictors of the present input pattern for training. Whenever an already learned pattern is presented again, the system can efficiently and reliably detect it. Experimental results confirm the suitability of the feature extraction method for a variety of input patterns. The spatiotemporal feature extraction leads to robust and reproducible detection, which is a key requirement for its use in higher-level visual recognition and classification. A central characteristic of the presented technique, in contrast to conventional computer vision methods, is that it does not depend on the concept of representing visual inputs as whole image frames. Instead, the method works efficiently on event-based sparse and asynchronous input streams, which maintain the temporal dynamics of the scene due to the highly precise asynchronous time sampling ability of the silicon retina. Thus, also the extracted spatiotemporal features contain richer dynamic information, in addition to recognizing spatial characteristics. Central to the definition of spatiotemporal features in our architecture is the presence of multiple models for prediction, which compete already during learning, such that specialization can occur. Similar concepts are used by various well-known machine learning frameworks, most notably the mixture-of-experts architecture (Jacobs et al., 1991 ; Jordan and Jacobs, 1994 ; Yuksel et al., 2012 ), in which a gating function creates a soft division of the input space for multiple local “expert” models. The output of the whole network is then a combination of the expert predictions, weighted according to their responsibility for the present input. These architectures have been extended in brain-inspired architectures for reinforcement learning and control (Haruno et al., 2001 ; Doya et al., 2002 ; Uchibe and Doya, 2004 ), where multiple forward models and controllers are learned simultaneously, and the prediction performance of the forward model determines the selection of the most appropriate local controller. Mixture-of-experts architectures are closely related to learning mixture models with the EM algorithm (Dempster et al., 1977 ; Jordan and Jacobs, 1994 ), where the E-step computes a soft assignment of data points to models. Nessler et al. ( 2009 ) and Nessler et al. ( 2013 ) have proven that this can be implemented in spiking neural networks, using a soft WTA circuit to compute the E-step, and an STDP learning rule to implement the M-step. Compared to these related architectures, our new model advances in three important aspects: Firstly, whereas EM and mixture-of-experts address static input distributions, we here extend this to multiple feature predictors for spatiotemporal sequences. Secondly, our architecture allows online learning of independent features, which contrasts with batch methods like PCA or ICA that operate on the full dataset after its collection. Thirdly, our neural network architecture is specifically designed to work with spiking inputs and for implementation with spiking neurons, thus maintaining the precise dynamics of event-based vision sensors. Other spiking neural network architectures for processing DVS inputs such as spiking ConvNets (Farabet et al., 2012 ; Camuñas-Mesa et al., 2014 ), and spiking Deep-belief networks (O'Connor et al., 2013 ) do not explicitly model the dynamics of the features extracted within the networks, but instead rely on different conversion mechanisms from analog to spiking neural networks, without taking sensor dynamics into account. The features they extract are thus characterizing a current snapshot of the input, and do not take its future trajectory into account like the ESN predictors of the presented model, but nevertheless are very useful for fast recognition. This is also true for approaches that directly classify spatiotemporal spike patterns, see e.g., (Sheik et al., 2013 ; Tapson et al., 2013 ). Spiking network models that represent spatiotemporal dynamics by emulating Hidden Markov Models have recently been introduced (Corneil et al., 2014 ; Kappel et al., 2014 ). Compared to our approach, these networks do not directly learn dynamic input features, but rather identify hidden states to determine the position within longer sequences. The combination of visual sensing with bio-inspired artificial retinas and event-based visual feature extraction, as presented in this article, opens new perspectives for apprehending the mechanisms of visual information encoding in the brain. It is clear that the traditional views of visually selective neurons as static image filters for receptive fields, e.g., as Gabor-like orientation filters, which are central to many classical vision models like HMAX or Neocognitron (Fukushima, 1980 ; Serre et al., 2002 ), fails to explain how these neurons deal with the highly dynamic and sparse spike inputs from biological retinas. In the presented approach, features are naturally learned and adapted to the task. In Figure 9 it was shown that if the number of available ESNs exceeds the number of features necessary to describe a scene, only the minimum necessary number of networks are trained. This has the desirable effect that whenever a new scene with new features is encountered, the previously unused ESNs can be trained to predict novel stimulus features. This behavior has several benefits: firstly, the number of ESNs does not have to be precisely tuned, but can be set to the highest acceptable number, and only the minimum number of networks is actually recruited and trained as feature detectors by the system. Alternatively, one could employ a different strategy in which new networks are recruited to the pool, whenever all current ESNs have specialized on features. Secondly, training of feature detectors works completely unsupervised, so no higher-level controller is needed to identify what the elementary features for a scene should be. Although the precesence of a supervisor is not necessary, having such information available would still be beneficial. For instance, another processing layer could use the outputs of the WTA to control the survival of each network. If such processing layer determines that a particular network does not provide enough interesting information, the supervisor could decide to reset and release the associated ESN, so that it can detect more relevant features. The presented method has great potential for use in event-based vision applications, such as fluid and high-speed recognition of objects and sequences, e.g., in object and gesture recognition (O'Connor et al., 2013 ; Lee et al., 2014 ), or for high-speed robotics (Conradt et al., 2009 ; Mueggler et al., 2014 ). The presented architecture is almost entirely based on computation with spikes. Inputs come in the form of AER events from DVS silicon retinas, providing an event-based representation of the visual scene. The WTA circuit for choosing between feature extractors is also working with spikes, and produces spike outputs, which indicate the identity of the detected feature. The only component of the system which does not entirely use spikes is the layer of ESNs that predict the visual input, but this restriction could be lifted by replacing ESNs with their spiking counterparts, called Liquid State Machines (LSMs) (Maass et al., 2002 ), which are computationally at least equivalent to ESNs (Maass and Markram, 2004 ; Büsing et al., 2010 ). The reasons why we have chosen to use ESNs for this proof-of-principle study are the added difficulty of tuning LSMs, due to the larger number of free parameters for spiking neuron models, delays, or time constants, in addition to the higher computational complexity involved in the simulation of spiking neural networks on conventional machines, which makes it hard to simulate multiple LSMs in real-time. Overall, we expect the improvement due to using fully spike-based feature detectors and predictors to be rather minor, since the ESNs can be efficiently simulated at time steps of 1 ms, which is also the time interval at which the silicon retina is sending events through the USB bus. However, a fully spike-based architecture does have great advantages in terms of efficiency and real-time executing if it can be implemented entirely on configurable neuromorphic platforms with online learning capabilities (Indiveri et al., 2006 ; Galluppi et al., 2014 ; Rahimi Azghadi et al., 2014 ), which is the topic of ongoing research. Conflict of interest statement The Reviewer Federico Corradi declares that, despite being affiliated to the same institution as author Michael Pfeiffer, the review process was handled objectively and no conflict of interest exists. 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."
} | 4,653 |
34064870 | PMC8150720 | pmc | 8,840 | {
"abstract": "A quickly tunable wettability pattern plays an important role in regulating the surface behavior of liquids. Light irradiation can effectively control the pattern to achieve a specific wettability pattern on the photoresponsive material. However, metal oxide materials based on light adjustable wettability have a low regulation efficiency. In this paper, zinc (Zn) superhydrophobic surfaces can be obtained by femtosecond-laser-ablated microholes. Owing to ultraviolet (UV) irradiation increasing the surface energy of Zn and heating water temperature decreasing the surface energy of water, the wettability of Zn can be quickly tuned photothermally. Then, the Zn superhydrophobic surfaces can be restored by heating in the dark. Moreover, by tuning the pattern of UV irradiation, a specific wettability pattern can be transferred by the Zn microholes, which has a potential application value in the field of new location-controlled micro-/nanofluidic devices, such as microreactors and lab-on-chip devices.",
"conclusion": "4. Conclusions In summary, a Zn superhydrophobic surface can be achieved by femtosecond-laser-ablated microholes. The Zn superhydrophobic surface can be quickly transformed into a hydrophilic surface by adjusting both the UV irradiation and water temperature. UV irradiation can control the wettability pattern of the Zn surface, while the hot water can decrease the surface tension of water to obtain a low surface energy of water. Therefore, the specific wettability pattern can be transferred through microholes in Zn foil. Compared with the adjustment of the UV radiation or temperature alone, the proposed method can effectively improve the wettability-transformation efficiency and achieve specific wettability patterns. Therefore, the method has good potential application prospects in new location-controlled micro-/nanofluidic devices, such as microreactors and lab-on-chip devices, among others.",
"introduction": "1. Introduction In nature, the lotus has excellent superhydrophobic surfaces [ 1 ]. After long-term research, it has been determined that in the characterization of superhydrophobic surfaces, they must have a certain rough structure and low surface energy [ 2 ]. Therefore, there are two methods for preparing superhydrophobic surfaces: one is fabricating a rough structure with low surface energy; the other is fabricating a rough surface and then spraying a low-surface-energy modifier onto it. However, the fixed wettability surface cannot satisfy the needs of a functional surface. Tunable wettability surfaces play an important role in regulating the surface behavior of liquids, which can be used for printing [ 3 ], droplet transfer [ 4 ], microfluidic systems [ 5 ], oil-water separation [ 6 , 7 ], and underwater gas collection [ 8 , 9 , 10 ]. In recent years, the tunable wettability of some metal oxide materials under light stimulation has attracted much attention [ 11 , 12 , 13 ]. The metal oxide ZnO, which is an important light-sensitive material, has the advantages of good stability and low cost, but its regulation efficiency is low. For example, Yong et al. [ 14 ] used a femtosecond laser to ablate cross micro-/nanostructures on the surface of a zinc (Zn) plate, and realized the regulation between superhydrophobic and superhydrophilic surfaces in ultraviolet (UV) and dark environments. However, the superhydrophobic Zn surface must be irradiated by UV light for 24 h to become a quasi-superhydrophilic surface, and then the quasi-superhydrophilic surface must be placed in a dark environment for 7 d to become a superhydrophobic surface. Tian et al. [ 15 ] used UV light to irradiate a stainless-steel mesh covered with ZnO nanorods to achieve regulation between superhydrophobic and superhydrophilic surfaces. However, the regulation efficiency was also low. To improve the efficiency of wettability regulation, Bai et al. [ 16 ] used stearic acid ethanol solution and sodium hydroxide to quickly adjust the wettability of stainless steel coated with nano Zn oxide. The regulation can be quickly switched in 15 min, but this method requires two kinds of chemical reagents to tune the wettability and does not have the ability to form a wettability pattern. At the same time, the temperature can also be used to adjust the wettability. For example, Liu et al. [ 17 ] realized the rapid regulation between superhydrophobic and superhydrophilic aluminum (Al) surfaces by adjusting the temperature of the Al surface and the pressure between the water and Al, and still achieved excellent recoverability, stability, and repeatability after 10 cycles. Moreover, Ngo et al. [ 18 ] achieved the adjustment of wettability by heating copper, Al, and titanium at different temperatures. The higher the temperature, the shorter the conversion time from superhydrophilic to superhydrophobic. However, this method does not have the ability to form a wettability pattern. Furthermore, tunable wettability can be achieved by changing the water temperature [ 19 ]. Liu et al. [ 20 ] investigated the repellent hot water of superhydrophobic surfaces and found that the superhydrophobic surfaces usually exhibited a high repellency to cool water. However, such surfaces show a remarkably decreased repellency to hot water, which can be attributed to a decrease in the surface tension. Moreover, the surface structures can be destroyed by elevated temperatures [ 21 ]. Meanwhile, the magnetic field [ 22 ], electric field [ 23 ], and mechanical force [ 24 ] have been commonly used to achieve tunable wettability. For example, Tian et al. [ 25 ] used ferromagnetic micro-nanomaterials to prepare a magnetic fluid/ZnO nanoarray. A water droplet could follow the motion of the gradient composite interface structure as it responded to the gradient magnetic field motion, which achieved tunable wettability. In addition, Tian et al. [ 26 ] further proposed a method for photoelectric coordination to achieve tunable wettability. The photosensitizer material titanium phthalocyanine was coated on the ZnO nanorods, and the structure was modified by heptafluorododecyltrimethoxysilane. The capillary pressure between the micro/nanostructures is tuned by a voltage combined with the light-modified photosensitizer in order to tune the wettability. Yang et al. [ 27 ] used a femtosecond laser to prepare the superhydrophobic surface of silicone rubber. The superhydrophobic surface of silicone rubber was transformed from the “petal” state to the “lotus” state by stretching. Compared with light and thermal tune wettability, the advantage of the magnetic field is that it can quickly achieve tunable wettability, but the disadvantage is that magnetic materials must be added to soft materials, which limits the scope of application. The voltage can quickly tune wettability, but the material must be conductive. The shortcomings of the mechanical force are also due to the limitations of the material. This method can only use materials with flexibility and high ductility to achieve tunable wettability. The above methods mainly serve to control the properties of materials in order to achieve tunable wettability, which is easily affected by the defects of material properties. At present, optical, magnetic, electrical, mechanical forces and other factors are seldom combined with liquid temperature to achieve tunable wettability. Because the temperature of the liquid is not affected by the material properties, it can not only improve the efficiency of tunable wettability but also provide a new idea for achieving tunable wettability. Therefore, to achieve a quickly tunable wettability pattern, in the work described in this paper, a femtosecond laser was used to fabricate microholes on Zn foil placed in the dark and heated at 100 °C for 12 h to obtain a superhydrophobic surface. The wettability pattern was then able to be obtained by controlling the pattern of UV irradiation. Moreover, with an increasing water temperature, the surface tension of water decreases. Therefore, the wettability pattern of the Zn-foil surface can be quickly transferred to paper through microholes. This method has potential applications in new location-controlled micro-/nanofluidic devices, such as microreactors and lab-on-chip devices."
} | 2,061 |
33195112 | PMC7606986 | pmc | 8,841 | {
"abstract": "The comprehension of the underlying mechanisms of the interactions within microbial communities represents a major challenge to be faced to control their outcome. Joint efforts of in vitro , in vivo and ecological models are crucial to controlling human health, including chronic infections. In a broader perspective, considering that polymicrobial communities are ubiquitous in nature, the understanding of these mechanisms is the groundwork to control and modulate bacterial response to any environmental condition. The reduction of the complex nature of communities of microorganisms to a single bacterial strain could not suffice to recapitulate the in vivo situation observed in mammals. Furthermore, some bacteria can adapt to various physiological or arduous environments embedding themselves in three-dimensional matrices, secluding from the external environment. Considering the increasing awareness that dynamic complex and dynamic population of microorganisms (microbiota), inhabiting different apparatuses, regulate different health states and protect against pathogen infections in a fragile and dynamic equilibrium, we underline the need to produce models to mimic the three-dimensional niches in which bacteria, and microorganisms in general, self-organize within a microbial consortium, strive and compete. This review mainly focuses, as a case study, to lung pathology-related dysbiosis and life-threatening diseases such as cystic fibrosis and bronchiectasis, where the co-presence of different bacteria and the altered 3D-environment, can be considered as worst-cases for chronic polymicrobial infections. We illustrate the state-of-art strategies used to study biofilms and bacterial niches in chronic infections, and multispecies ecological competition. Although far from the rendering of the 3D-environments and the polymicrobial nature of the infections, they represent the starting point to face their complexity. The increase of knowledge respect to the above aspects could positively affect the actual healthcare scenario. Indeed, infections are becoming a serious threat, due to the increasing bacterial resistance and the slow release of novel antibiotics on the market."
} | 550 |
37745781 | PMC10512918 | pmc | 8,842 | {
"abstract": "ABSTRACT Phase-change memory (PCM), recently developed as the storage-class memory in a computer system, is a new non-volatile memory technology. In addition, the applications of PCM in a non-von Neumann computing, such as neuromorphic computing and in-memory computing, are being investigated. Although PCM-based devices have been extensively studied, several concerns regarding the electrical, thermal, and structural dynamics of phase-change devices remain. In this article, aiming at PCM devices, a comprehensive review of PCM materials is provided, including the primary PCM device mechanics that underpin read and write operations, physics-based modeling initiatives and experimental characterization of the many features examined in nanoscale PCM devices. Finally, this review will propose a prognosis on a few unsolved challenges and highlight research areas of further investigation.",
"introduction": "Introduction Recently, the demand for data storage has exponentially increased and attracted considerable attention compared with other memory technologies. Since the development of data storage technologies, several memory technologies have emerged, providing diverse choices for the next generation of storage [ 1–5 ]. To improve data-computing efficiency and energy utilization, a series of universal memory data storage devices with advantages, such as low power consumption and moderate price of dynamic random-access memory (DRAM) and high operating speeds of static random-access memory (SRAM) was proposed [ 6–12 ]. However, SRAM and DRAM are volatile, and data is lost after a power outage. Conversely, non-volatile memory can store data for tens of years in the power-off state and thus has been attracting increasing research attention. The fast and non-volatile properties can bridge the memory wall between the volatile DRAM and the non-volatile solid-state disk. The resistance random-access memory is a promising candidate for high-density storage owing to its simple crossbar structure [ 9 , 13 ]. Commercialized magnetic random-access memory (MRAM) is advantageous for its low power and high-speed consumption. As non-volatile memory, the high-performance and high-scalability phase-change random-access memory (PCRAM) can operate in the order of nanoseconds and has been a leading candidate among the emerging nonvolatile memory technologies for next-generation electronic devices [ 14–16 ]. In PCRAM, phase change materials are the core components, so in-depth research on the microstructure characterization and device applications of phase-change materials is of great significance. Phase change materials are the core materials of phase change memory, and the research and development of phase change materials directly affect the performance and application of phase change memory. Therefore, the research and development of phase change materials is an important foundation for the development of phase change memory. Data storage in PCRAM is realized by the large electrical and optical properties of its crystalline and amorphous states in chalcogenide materials. In general, the flagship phase change material is ternary Ge 2 Sb 2 Te 5 , which is a pseudo-binary chalcogenide material typically comprising Sb 2 Te 3 and GeTe [ 17–22 ]. In amorphous Ge – Sb–Te (GST) compounds, the atoms are randomly distributed without long-range order and can be sequentially crystallized into an equilibrium hexagonal structure, as illustrated in Figure 1(a) [ 23 ]. The crystalline states generally exhibit low electrical resistivity and low transmittance, whereas the amorphous states exhibit high electrical resistivity and high transmittance. The crystalline state exhibits electricity two to six orders of magnitude lower than that of the amorphous state, and the two states can be reversibly switched by current pulses or light pulses at different durations and intensities. The PCM can be crystallized from the amorphous under the application of a long low current pulse (above crystalline temperature Tc, below melting point Tm) for the SET/write operation at a temperature of ∼500 K; and recrystallized from the crystalline state under the application of a short high current pulse (above Tc) for the RESET/erase operation at a temperature > 1000 K, as shown in Figure 1(b) [ 24 ]. The crystallization speed is tens to hundreds of nanoseconds, while the amorphization process can be as fast as hundreds of picoseconds.\n Figure 1. (a) Crystal structures of amorphous and crystalline Ge2Sb2Te5 (GST). Reproduced by permission from [ 23 ], copyright [2015, Springer] (b) Diagram of operation principle in PCM. Reproduced by permission from [ 24 ], copyright [2020, IOP]. The PCM acts as a core layer in commercialized PCRAM and plays an important role. The first chalcogenide PCM, GeSiAsTe, was originally proposed by Ovshinsky et al. in 1968, and the GST compounds were subsequently discovered. Following these pioneering reports, in the GST alloys, Ag 5 In 5 Sb 60 Te 30 and Ag 4 In 2 Sb 67 Te 26 were found to exhibit fast crystallization and favorable optical reflectivity contrast, facilitating their wide application in rewritable optical storage products, such as digital versatile disks (DVD)-RAM, DVD-rewritable disks, and Blu-ray discs. With the development of material science, PCMs are increasingly being discovered and more detailed phase diagrams are being developed. Until now, Te-containing chalcogenide PCMs have been significantly studied, but few non-chalcogenide alloys and elementary substances have. The crystallization mechanisms of PCMs can be broadly divided into nucleation-driven and growth-driven types, according to the crystallization kinetics. For the nucleation-driven type PCMs, such as Sb 2 Te 3 , crystallization occurs via the formation of critical nuclei and subsequent growth, exhibiting fast nucleation. In contrast, the crystallization of those growth-driven PCMs occurs in the amorphous region surrounded by a crystalline matrix; it rapidly proceeds, and no sizeable nor robust crystalline seeds form during the short timescale involved in the growth process. Moreover, doping engineering, including W-doped Ge 2 Sb 2 Te 5 , Sc-doped Sb 2 Te 3 , and Se-doped GeTe, is frequently used to improve the intrinsic properties of PCM, providing different material characteristics to consequently meet the requirements in different scenarios. In this review, we outline the recent advancements in the characterization of PCM and device performances, including physical characteristics, crystallization dynamics, structural design, and storage performance. Moreover, we anticipate further advancements in PCM technology."
} | 1,667 |
39063819 | PMC11278828 | pmc | 8,843 | {
"abstract": "The growing energy consumption and the need for a circular economy have driven considerable interest in the anaerobic digestion (AD) of organic waste, offering potential solutions through biogas and digestate production. AD processes not only have the capability to reduce greenhouse gas emissions but also contribute to the production of renewable methane. This comprehensive review aims to consolidate prior research on AD involving different feedstocks. The principles of AD are explored and discussed, including both chemical and biological pathways and the microorganisms involved at each stage. Additionally, key variables influencing system performance, such as temperature, pH, and C/N ratio are also discussed. Various pretreatment strategies applied to enhance biogas generation from organic waste in AD are also reviewed. Furthermore, this review examines the conversion of generated digestate into biochar through pyrolysis and its utilization to improve AD performance. The addition of biochar has demonstrated its efficacy in enhancing metabolic processes, microorganisms (activity and community), and buffering capacity, facilitating Direct Interspecies Electron Transfer (DIET), and boosting CH 4 production. Biochar also exhibits the ability to capture undesirable components, including CO 2 , H 2 S, NH 3 , and siloxanes. The integration of digestate-derived biochar into the circular economy framework emerges as a vital role in closing the material flow loop. Additionally, the review discusses the environmental benefits derived from coupling AD with pyrolysis processes, drawing on life cycle assessment investigations. Techno-economic assessment (TEA) studies of the integrated processes are also discussed, with an acknowledgment of the need for further TEA to validate the viability of integrating the biochar industry. Furthermore, this survey examines the techno-economic and environmental impacts of biochar production itself and its potential application in AD for biogas generation, aiming to establish a more cost-effective and sustainable integrated system.",
"conclusion": "7. Conclusions The AD process stands out as a straightforward technology for treating organic waste, offering significant environmental, energy, and economic potential. Various pathways facilitated by different types of microorganisms enable the conversion of organic waste into biogas. Key operational parameters such as temperature, moisture content, pH, organic loading rate, and C ratio influence process phases, affecting overall efficiency. A substantial amount of digestate is generated annually from AD, necessitating careful management for agronomic purposes. Factors governing digestate storage duration include local regulations, weather conditions, soil type, crop cycles, and operational procedures. Improper handling can lead to methane and ammonia emissions, impacting fertilizer quality, greenhouse gas emissions, public perception, and economic viability. Considering ecological impacts, organic content quality, disposal costs, and nutrient availability, land application alone is insufficient. Therefore, exploring alternative valorization methods for digestate, particularly through pyrolysis for biochar production, is essential. Pyrolysis technology optimizes AD by converting digestate into biochar and enhancing overall sustainability through additional oil and syngas production and higher carbon stability. While conventional biochar has been extensively researched for soil improvement and contaminant remediation, studies on digestate-derived biochar remain limited but promising. Biochar derived from digestate holds potential as a soil conditioner, contributing to carbon sequestration and environmental benefits such as contaminant remediation. It improves microbial growth, system buffering capacity, and Direct Interspecies Electron Transfer (DIET), leading to increased methane production. Future research should focus on fully characterizing digestate-derived biochar and its application in AD systems. Understanding how biochar characteristics impact AD efficiency, including surface area, pore structure, functional groups, elemental composition, and ash content, is crucial. Mechanisms such as ammonium sorption and microbial pathways need further exploration. Additionally, biochar shows promise in upgrading biogas by adsorbing impurities like CO 2 , H 2 S, NH 3 , and siloxanes. Integrating digestion with pyrolysis presents an effective solution to challenges associated with digestate management. In conclusion, utilizing digestate-derived biochar in AD systems or as a biogas purification material can advance circular-economy-driven processes. Further research is necessary to assess the technical, financial, and environmental efficiencies of alternative process designs integrating digestion with pyrolysis.",
"introduction": "1. Introduction The circular economy model has received considerable attention on the policy agenda, especially in the context of climate change and the decarbonization of the economy. Returning waste to supply chains and meeting the energy demands of local communities and businesses is crucial for rationalizing waste management, increasing resource efficiency, and effectively implementing the circular model. Human demands exert persistent pressure on Earth, and global society continues to grow. Resources are however both critical and limited, and their use must be maximized [ 1 ]. Efficient management is essential to promote circularity, combat climate change, and achieve sustainable development [ 2 ]. Integrating a circular economy with biotechnology can support local businesses’ growth, while simultaneously preserving the environment from uncontrolled waste disposal [ 3 , 4 ]. An example of this potential is the substantial prospect of reusing approximately 1 million tons of used cooking oil generated in the European Union (EU), often discharged into public sewage systems, elevating wastewater treatment costs [ 5 ]. The accumulation of municipal and residential waste is increasing as the world population grows, particularly in densely populated areas and in tourist destinations. In addition, agriculture faces challenges, such as significant amounts of agricultural leftovers, livestock residues, and biodegradable waste that are unsuitable for human and animal consumption. The annual global waste generation is estimated between 7 and 9 billion tons, with over 2 billion tons representing municipal solid waste (MSW) [ 6 , 7 ], projected to reach 3.4 billion tons by 2050. Since waste management is one of the urgent and crucial issues facing contemporary civilization, efforts are being made to develop technologies to limit waste accumulation in landfills. This includes waste separation in industrialized nations, which makes waste disposal and recycling easier and less expensive. Additionally, organic waste, being predominately biodegradable, can be subjected to anaerobic digestion or incineration. Furthermore, sewage sludge, a byproduct of municipal and industrial wastewater treatment, must be eliminated. Despite the successful reduction of waste that ends up in landfills, the above-mentioned traditional methods can pose various negative environmental implications, including greenhouse gas emissions as well as groundwater, soil, and air contamination. The European Commission, Council, and Parliament have reached a preliminary agreement on the Circular Economy Package of waste-reduction policies. This agreement leads the EU to a good level of waste management sustainability. Waste-to-gas technology, particularly through anaerobic digestion (AD) in which biodegradable matter is converted into biogas, is recognized as an ecologically sustainable waste management technique, AD can address concerns related to waste management, renewable energy production, sustainable food production, and nutrient recycling in a circular and sustainable manner. In numerous European countries, AD is supplanting emission-intensive waste management alternatives, including landfilling in the agro-industrial sector. Beyond enhancing resource efficiency and reducing CO 2 emissions, biogas plants bring about a positive economic impact and provide ecologically friendly energy [ 8 ]. As part of the development of a circular economy, new approaches could improve the AD sector. AD of organic waste is a well-established process that generates biogas, a biofuel used for heat and electricity generation or injection into the natural gas grid after upgrading [ 9 , 10 ]. At the same time, a huge quantity of digestate is produced as residual waste [ 11 ]. Digestate contains concentrated organic and inorganic compounds with a high moisture content and agronomic value [ 12 ]. Direct use in agriculture may face challenges, such as the quality of the digestate and the limited availability of suitable land near the biogas plant. Given the highly contentious nature of this topic and the extensive research being conducted on the valorization of digestate in agriculture, it is worthwhile to explore alternatives for its valorization. It is worth noting that the unprocessed release of digestate without appropriate treatment can have negative implications for the quality of the surrounding environment [ 13 , 14 ]. Therefore, there is a recognized imperative for comprehensive digestate treatment before its final disposal, even though this process may incur substantial costs [ 15 ]. In response to the call to embrace a circular economy in modern societies, substantial efforts have been dedicated to recovering valuable elements from digestate, transforming them into renewable resources. As a result, using digestate to generate biofuels and value-added products by combining AD with other processes emerges as a promising alternative [ 11 ]. Consequently, digestate can be utilized as a feedstock for pyrolysis-based biochar synthesis. Given the possible applications of biochar, it represents a viable and beneficial option for managing this AD by-product [ 16 , 17 ]. Additionally, the circular economy model for digestate handling is attractive for closing the material loop, since digestate could be utilized for biochar synthesis before reintroduction into anaerobic digesters or for various other purposes [ 18 ]. The key factors enhancing the efficiency of AD when biochar is added are the physicochemical qualities of the biochar (including its porosity, electrical conductivity, pH, redox properties, etc.) [ 18 ]. The present review provides an overview of the AD process, critical operational factors, and potential approaches utilized to improve the recovery of organic waste. It also details the valorization of digestate produced by AD, in particular exploring the potential for biochar production from this feedstock and discussing the impacts of biochar on the AD process. As a result, this review aims to elucidate the ability to generate biochar from digestate, examine its impacts on AD when used as additives, and demonstrate its potential benefits for bio-CH 4 upgrading, thereby illustrating a circular system. The results of this review can guide policy-makers, waste management professionals, and renewable energy developers in implementing more efficient and sustainable waste-to-energy systems. By detailing the potential for integrating AD and biochar production, we provide a roadmap for developing circular waste management solutions that maximize resource recovery and minimize environmental impact. From an academic perspective, this study synthesizes current knowledge in fields such as waste management, renewable energy, and materials science. By linking these generally distinct areas, we aim to identify new research directions and interdisciplinary opportunities in circular bioeconomy systems. This integrated approach can significantly contribute to climate change mitigation, resource conservation, and sustainable development goals by transforming waste management from a linear to a circular model. Finally, beyond the environmental and economic analysis of combining AD and pyrolysis, the life cycle analysis of the processes under consideration is briefly reviewed."
} | 3,047 |
23603788 | PMC3631768 | pmc | 8,845 | {
"abstract": "Over the past decades, our understanding of nacre's toughening origin has long stayed at the level of crack deflection along the biopolymer interface between aragonite platelets. It has been widely thought that the ceramic aragonite platelets in nacre invariably remain shielded from the propagating crack. Here we report an unexpected experimental observation that the propagating crack, surprisingly, invades the aragonite platelet following a zigzag crack propagation trajectory. The toughening origin of previously-thought brittle aragonite platelet is ascribed to its unique nanoparticle-architecture, which tunes crack propagation inside the aragonite platelet in an intergranular manner. For comparison, we also investigated the crack behavior in geologic aragonite mineral (pure monocrystal) and found that the crack propagates in a cleavage fashion, in sharp contrast with the intergranular cracking in the aragonite platelet of nacre. These two fundamentally different cracking mechanisms uncover a new toughening strategy in nacre's hierarchical flaw-tolerance design.",
"discussion": "Discussion A cube-corner nanoindenter was used to probe the mechanical response difference between the geologic single-crystal aragonite and nacre's aragonite platelets. Figure 4 depicts the load-displacement curves of nanoindentations made on the (002) planes of the respective geologic single-crystal aragonite and nacre's aragonite platelet with the peak indentation load (300 μN) (Section 5, Supplementary materials ). For the geologic single-crystal aragonite, at least two pronounced displacement pop-ins (strain avalanche) can be detected in the loading curve and the first one occurs at the displacement of approximately 6.6 nm. In this aspect, the indentation depth is sufficiently small and its corresponding contact zone can be considered as the nominally defect-free volume 29 30 . The sudden displacement excursion is a result of plasticity onset, immediately after perfect elasticity process. Triggering such plastic deformation requires the stress level to be on the order of theoretical strength (~E/10) 29 30 . Obviously, the stress concentration at the first pop-in undoubtedly reached the theoretical strength, initiating the plastic deformation onset. In comparison, no discernible pop-in can be detected in the nanoindentation loading curve for nacre's individual aragonite platelets, implying the stress concentration underneath the nanoindenter did not achieve the theoretical strength. Gao et al. 16 proposed that for a brittle material, there exists a critical length scale based on Griffith criterion as follows where h* is the critical length scale, parameter α is equal to , γ is the surface energy, E m and σ th are elastic modulus and theoretical strength, respectively. For the brittle geologic monocrystal aragonite 16 , the critical length scale h* is estimated to be approximately 30 nanometers. When the aragonite particle size is below or equivalent to this critical length scale, the stress concentration can not be accumulated anymore, as reflected by the smooth loading curve for nacre's aragonite platelets. Prior to achieving the theoretical strength, the weak nanoparticle boundary can serve as the relieving outlet for potential stress accumulation. Whereas, once the grain size far exceeds the critical length scale, i.e. the geologic monocrystal aragonite, the stress upon deformation can be significantly concentrated, as validated by the onset of first pop-in event under nanoindentation. Here, we can reasonably argue when the aragonite particle size falls into coarse-grained regime (micrometer-scale), the stress concentration upon deformation accordingly emerges and triggers catastrophic brittle fracture (Section 6, Supplementary materials ). The association between length scale and deformation mechanism, i.e. dislocation in metallic material, shear band in metallic glass, crack in ceramic, is a critically important guideline of how to acquire fracture toughness from microstructure design. The fracture toughness of metallic materials strongly depends on the spacing between dislocation nucleation sites and obstacles 3 . Owing to its micrometer-sized affected volume surrounding crack tip, the attempt to keep such a small length scale via refining grain size, will hence impose potential microstructure obstacles on dislocation activities at the crack tip. Likewise, the characteristic microstructure length scale associated to the maximum spatial extension for shear band has also been found in metallic glasses 31 32 . Through introducing crystalline imhomogeneities in metallic glasses, the traveling distance of shear band will be shortened prior to evolving into an unstable crack, thereby avoiding catastrophic failure. Similarly, it is conceivable that for ceramic materials tailoring the length scale such as embedding second-phase in matrix to intercept crack propagation will enhance the crack extension resistance, which will in turn contributes to fracture toughness improvement. However, this toughening approach typically suffers from stress concentration at the crack tip prior to being intercepted. Whereas, the finding of crack manifestation in individual aragonite platelets works out a new critical length scale, referring to a criterion whether there exists stress concentration at the crack tip. Of course, the role of nanoparticle boundary in biominerals, relative to other pure nanocrystalline ceramics, is probably unique in releasing the stress concentration 33 . In this regard, its exceptional role can be validated by two characteristic traits from the load-displacement response. Through comparing with geologic monocrystal aragonite, the load-displacement curve for nacre's aragonite platelets exhibits a much lower loading slope, and more remarkable viscoelastic effect in the nanoindentation holding segment, implying the existence of biopolymer matrix (soft phase) inside nacre's individual aragonite platelets. It is worthy pointing out that several slight fluctuations in the loading curve of nacre's aragonite platelets can not be regarded as displacement pop-ins, but related to the deformation response from aragonite nanoparticle rotation (Section 7, Supplementary materials ). The direct evidence that individual aragonite platelets defeat crack catastrophic propagation with its nanoparticle-architecture overturns the previous understanding that considers nacre's aragonite platelet as simply brittle monocrystal. Such a cracking mechanism favors local buffering in case a crack rigidly invades the aragonite platelet. Certainly, it is not expected for aragonite platelets to witness a catastrophic failure arising from confined stress concentration. This manifestation is reminiscent to the mechanical protection role of individual threads of spider webs, where local damage-control promotes the integral mechanical optimization 34 35 . The nanoscale localized damage-tolerance design principle in nacre is probably ubiquitous for other biominerals, i.e. tooth 36 , bone 37 , sea urchin spine 38 , ancient armored fish 39 ."
} | 1,777 |
26132866 | null | s2 | 8,846 | {
"abstract": "Increasing rare earth element (REE) supplies by recycling and expanded ore processing will result in generation of new wastewaters. In some cases, disposal to a sewage treatment plant may be favored, but plant performance must be maintained. To assess the potential effects of such wastewaters on biological treatment, model nitrifying organisms Nitrosomonas europaea and Nitrobacter winogradskyi were exposed to simulated wastewaters containing varying levels of yttrium or europium (10, 50, and 100 ppm), and the extractant tributyl phosphate (TBP, at 0.1 g/L). Y and Eu additions at 50 and 100 ppm inhibited N. europaea, even when virtually all of the REE was insoluble. Provision of TBP with Eu increased N. europaea inhibition, although TBP alone did not substantially alter activity. For N. winogradskyi cultures, Eu or Y additions at all tested levels induced significant inhibition, and nitrification shut down completely with TBP addition. REE solubility was calculated using the previously developed MSE (Mixed-Solvent Electrolyte) thermodynamic model. The model calculations reveal a strong pH dependence of solubility, typically controlled by the precipitation of REE hydroxides but also likely affected by the formation of unknown phosphate phases, which determined aqueous concentrations experienced by the microorganisms."
} | 334 |
38049399 | PMC10696030 | pmc | 8,850 | {
"abstract": "Microbial rhodopsin, a significant contributor to sustaining life through light harvesting, holds untapped potential for carbon fixation. Here, we construct an artificial photosynthesis system which combines the proton-pumping ability of rhodopsin with an extracellular electron uptake mechanism, establishing a pathway to drive photoelectrosynthetic CO 2 fixation by Ralstonia eutropha (also known as Cupriavidus necator ) H16, a facultatively chemolithoautotrophic soil bacterium. R. eutropha is engineered to heterologously express an extracellular electron transfer pathway of Shewanella oneidensis MR-1 and Gloeobacter rhodopsin (GR). Employing GR and the outer-membrane conduit MtrCAB from S. oneidensis , extracellular electrons and GR-driven proton motive force are integrated into R. eutropha ’s native electron transport chain (ETC). Inspired by natural photosynthesis, the photoelectrochemical system splits water to supply electrons to R. eutropha via the Mtr outer-membrane route. The light-activated proton pump - GR, supported by canthaxanthin as an antenna, powers ATP synthesis and reverses the ETC to regenerate NADH/NADPH, facilitating R. eutropha ’s biomass synthesis from CO 2 . Overexpression of a carbonic anhydrase further enhances CO 2 fixation. This artificial photosynthesis system has the potential to advance the development of efficient photosynthesis, redefining our understanding of the ecological role of microbial rhodopsins in nature.",
"introduction": "Introduction Microbial carbon dioxide (CO 2 ) fixation through photosynthesis is one of the foundations of the global carbon cycle 1 , 2 . Photosynthetic microbes harvest solar radiation to convert CO 2 and water into organic compounds, contributing around 50% of primary productivity on earth 3 . In nature, almost all known light-harvesting mechanisms in microorganisms are either chlorophyll- or rhodopsin-based systems 4 . Chlorophyll-based photosystems are multi-component pigment-protein complexes (i.e., photosystem I and II) coupled with a series of redox protein complexes to form a photosynthetic electron transport chain, which use light as the energy source to split water, accompanied by the production of ATP and reductants (e.g., NADPH). In comparison, rhodopsin-based photosystems are much simpler with only a light-activated proton pump generating proton motive force 5 . There was a lot of evidence suggesting that rhodopsin can support ATP synthesis to enhance inorganic carbon assimilation in the presence of organic electron donors 6 , 7 . However, rhodopsin cannot solely drive the CO 2 fixation pathway with its proton pumping, due to the lack of electron generation. Microbial rhodopsins as the most widespread proteins in the microbial world are hypothesised to have a significant role in microbial carbon fixation 8 . Recently, rhodopsin phototrophy has been shown to generate sufficient energy for reversing electron transfer which can drive NADH synthesis via NADH dehydrogenase 9 and then power CO 2 fixation 10 . These findings indicated the proton gradient generated by rhodopsin can support the regeneration of NAD(P)H via an electron transport chain, providing a driving force for CO 2 reduction 10 . Microbial rhodopsins are widely used as optogenetic tools in synthetic biology and have been successfully expressed in model bacteria such as Escherichia coli 11 , Ralstonia eutropha 10 and Shewanella oneidensis 9 . Numerous studies 12 – 14 demonstrated that the engineered bacteria with rhodopsin can convert light energy into intracellular chemical energy. For example, the engineered E. coli with rhodopsin has been shown to improve bioproduction in the light 12 . Additionally, the expression of rhodopsin in the electroactive bacterium S. oneidensis can power the electrosynthesis of reduced products in cathodic conditions 9 . Integrating a rhodopsin-based photosystem with chemoautotrophic bacteria represents a promising approach to increase biosynthesis from CO 2 . Chemoautotrophic bacteria can utilise inorganic chemicals or an electrode as the electron donor to assimilate CO 2 into valuable compounds. R. eutropha H16, a chemolithoautotroph, is a model microbial chassis, owing to its CO 2 fixation pathway and metabolic versatility. We recently engineered the autotrophic bacterium R. eutropha H16 with a Gloeobacter rhodopsin (GR) and created a redox loop by integrating it with an external electrode 10 . Electrons supplied by the electrode were transferred into the engineered R. eutropha to drive CO 2 fixation. The process was mediated by an electron-shuttling molecule such as flavin, and powered by rhodopsins 10 . While this system can effectively incorporate CO 2 , it displayed a relatively low electron transfer rate and efficiency. We hypothesised that engineering an efficient electron transfer interface on the cell membrane could be a key strategy for improving the electron transport from the electrode to the cells. Natural electroactive bacteria, such as S. oneidensis MR-1, use the outer-membrane-bound Mtr pathway, facilitating bidirectional electron transfer between extracellular substrates and the quinone pool 15 , 16 . In the model photosynthetic bacterium Rhodopseudomonas palustris TIE-1, there are also cytochromes presented on the outer membrane to enable the electron transfer between intracellular and extracellular electron acceptors and electrons 17 . It is hypothesised that artificially combining extracellular electron uptake with intracellular reverse electron transfer could lead to the construction of a synthetic photosynthetic electron transport chain in R. eutropha H16. In addition, the proton-pumping capacity could be another limiting factor to rhodopsin-based photoelectrosynthesis, which is associated with ATP and NADH synthesis 18 . As with R. palustris , the proton gradient not only drives ATP production via ATP synthase but also plays a key role in reversing the function of proton-translocating NADH dehydrogenase for regeneration of reducing equivalents 17 . Therefore, increasing the proton-pumping activity of rhodopsin may also increase the efficiency of the artificial system. In this work, R. eutropha H16 is engineered to establish a photon and electron harvesting system by the heterologous expression of the outer-membrane conduit MtrCAB complex from S. oneidensis MR-1 and the GR protein from Gloeobacter violaceus . This synthetic biology design combining the Mtr pathway and rhodopsin-based phototrophy enables the construction of an electrochemically driven photosynthetic electron transport chain. The engineered R. eutropha can directly take electrons from a cathode through the Mtr pathway and use the proton motive force generated by light-activated rhodopsin to power ATP and NAD(P)H synthesis. We integrate the photoelectrosynthetic R. eutropha with a solar-driven electrochemical system to achieve artificial photosynthesis (Fig. 1 ). To enhance CO 2 fixation efficiency, flavin and canthaxanthin are introduced to combine with Mtr and GR, respectively, improving the electron transfer rate and proton-pumping capacity. Additionally, a β-carbonic anhydrase ( can ) is overexpressed to enhance CO 2 fixation, as increased can levels facilitate the interconversion between bicarbonate and CO 2 . Overall, artificial photosynthesis could help design alternative methods for CO 2 fixation, providing a valuable reference for engineering non-photoelectrosynthetic bacteria into photoelectrotrophs, and offering insights into future investigations of the potential interplay between the rhodopsin-based metabolism and extracellular electron transfer, along with their joint effect on microbial CO 2 fixation. Fig. 1 Construction of an artificial photosynthesis system by integrating a photoelectrochemical system with genetically engineered cells expressing rhodopsin and an outer-membrane conduit MtrCAB. In the artificial photosynthesis system, light energy is absorbed by a solar panel and rhodopsin to generate electricity and drive the metabolism, respectively. The engineered bacteria obtain electrons from the electrode, mediated by the Mtr complex and flavins, to synthesise reducing power (i.e., NADH and NADPH). In the presence of ATP and NADPH, the Calvin–Benson–Bassham (CBB) cycle is activated to drive CO 2 fixation for biosynthesis.",
"discussion": "Discussion This research demonstrates that an artificial photosynthetic system can be established by redesigning a non-phototrophic bacterium R. eutropha . This system can be potentially extended to transform other non-photosynthetic bacteria into photosynthetic forms. The artificial photosynthetic system was inspired by natural photosynthesis. We have developed an electrochemical system powered by light energy, to split water, mimicking the function of photosystem II. To achieve the function of photosystem I for the regeneration of NADH/NADPH, rhodopsin was incorporated to generate proton motive force, facilitating the process of electron transfer (Fig. 5 ). Unlike common H 2 -mediated microbial electrosynthesis 30 , we engineered a transmembrane conduit Mtr complex as an interface between an extracellular cathode and the intracellular electron transport chain. Such electron transfer efficiency, in theory, could be comparable to an H 2 -mediated system 31 . To rule out hydrogen generation and ensure that the Mtr-mediated electron transfer was the major mechanism, a relatively low potential at ~1.7 V was applied with no evolution of hydrogen detected in the headspace (Supplementary Fig. 4c ). The control condition in the absence of light showed no cell growth (Fig. 4f ), consistent with the previous reports that low potential conditions would be difficult to support cell growth 30 , 32 . The formation of the photosynthetic electron transport chain designed in this study is fundamentally similar to natural photosynthesis, in which electrons are transferred via a series of redox enzymes accompanied by proton movement, whilst avoiding the generation of intermediates such as H 2 30 and formate 33 . Such interactions between electrons and protons drive the photochemical reactions. Fig. 5 Schematic illustration of the light reaction pathways of natural photosynthesis and designed artificial photosynthesis. Potentials for hydrogen evolution reaction (E Hydrogen evolution ) and NAD(P)H (E NAD(P)H ) synthesis are highlighted by dash lines. Part of redox enzymes are omitted. Oxygen evolution with O 2 bubbles was observed on the anode, and biofilms were observed on the cathode with a brightfield microscopic image and FITC fluorescence image of biofilms attached to the cathode using SYTO TM 9 staining. The scale bars represent 20 μm. The engineered photoelectrotrophic R. eutropha in this study is similar to the anoxygenic phototroph Rhodopseudomonas palustris which is a model bacterium for studying phototrophic extracellular electron uptake 17 . Both the genetically engineered R. eutropha and the native R. palustris are phototrophic; however, their photosystems are unable to solely drive CO 2 fixation due to the absence of photosystem II which is essential for water splitting and electron generation. In the presence of an electrode, both microorganisms are able to harvest electrode-supplied electrons to drive CO 2 fixation pathway in the light. However, the underlying mechanisms of these microbial systems are different. The engineered R. eutropha employs rhodopsin to generate proton motive force. We used the synthetic design to couple proton movement with electron transfer to generate the energy required for CO 2 fixation. In contrast, R. palustris ’s photosystem P 870 naturally involves electron transfer and can excite electrons in the light 17 . Although the synthetic design showed comparable performance in supporting photoelectrotroautotrophic growth, further investigations are needed to deepen our understanding of the photosynthetic electron transport chain. In the process of electron transfer in the engineered R. eutropha , the native biomolecule plays an important role in obtaining electrons from the Mtr pathway. In the model electroactive bacterium S. oneidensis MR-1, the outer-membrane MtrCAB pathway is usually coupled with an inner-membrane protein CymA to involve electron transfer 34 . However, a recent study on the engineered E. coli with an Mtr pathway showed that CymA might not be necessary for inward electron transport 35 . In addition, several other inner-membrane enzymes such as NapC have been suggested to perform the same function as CymA 19 . Furthermore, a recent study of S. oneidensis reported that hydrogenase could obtain electrons from the Mtr pathway 36 . Inspired by the finding in S. oneidensis MR-1, we propose that, in the native system of R. eutropha , the periplasm-facing membrane-bound hydrogenase could also be involved in Mtr-mediated inward electron transfer. This study demonstrates that the expression of MtrCAB complex enhances extracellular electron transfer in R. eutropha . Nevertheless, the enhancement in electron transfer remains limited compared to other heterologous microorganisms, such as E. coli which achieved an approximate 2 to 6-fold increase in the electron transfer efficiency through the Mtr pathway expression 19 , 37 . This observed limitation could be attributed to relatively low MtrCAB expression in R. eutropha . The expression of the Mtr pathway can potentially be improved by optimisation of codon and ribosome binding site, as well as the addition of precursors like 5-aminolevulinic acid 38 for heme synthesis. Additionally, the introduction of additional proteins such as periplasmic small tetraheme cytochrome (STC) may further enhance electron transfer within the periplasmic space 38 . The investigation of heterologous Mtr pathway expression in non-native bacteria has received considerable attention in recent years 19 , 37 , 38 . The introduction of Mtr pathway has improved the extracellular electron transfer of R. eutropha in this study (Fig. 4h ). However, the exact functionality of the Mtr expression in R. eutropha and other heterologous bacteria requires more rigorous verification. For example, there is a remote possibility that MtrCAB may not be directly involved in electron transfer but, instead, might increase membrane permeability, enhancing electron transfer via endogenous electron mediators such as flavins and quinones. The reversing function of NADH dehydrogenase is key to regenerating NADH 9 . The reverse electron transfer from the quinol pool (with a midpoint potential, E m , approximately –0.08 V) to NAD + ( E m = –0.32 V) is thermodynamically unfavourable (Fig. 5 ). The reversed process requires a highly reduced quinol pool to supply electrons and a high proton motive force to drive protons back through the NADH dehydrogenase in the reverse direction to forward catalysis. In this study, electrons flow inwardly from the external electrode to the Mtr pathway then leading to the reduction of the quinol pool. The effective potential of the highly reduced quinol pool should be more negative than its electrochemical midpoint potential and approached NADH potential. In the meantime, the proton motive force generated by rhodopsin facilitated the NADH dehydrogenase-mediated reverse electron flow for NADH regeneration. In R. eutropha , there are two types of NADH dehydrogenase, namely nuo and ndh 21 . nuo is a proton-translocating protein and ndh is a non-proton-pumping membrane-bound protein. According to a previous study of S. oneidensis , nuo is critical to inward electron transfer from the electrode for NADH generation 39 . In this system, the NADH generation was impeded by the addition of cyanide m-chlorophenyl hydrazone (CCCP), a protonophore (Supplementary Fig. 11 ). Therefore, it is logical to reason the significant role of the proton-translocating nuo in R. eutropha in the process of inward electron transfer for NADH generation. In this system, the proton motive force primarily relies on the photoreaction of the proton-pumping rhodopsin. There are two strategies for enhancing the proton pumping of rhodopsin. One approach is increasing light intensities, thereby elevating the proton-pumping rate of the rhodopsin. We found that increasing light intensities led to enhanced biomass increases in R. eutropha RHM5-GR-Mtr (Supplementary Fig. 12 ). The other approach is binding an antenna molecule to rhodopsins. In nature, microorganisms cannot control the light intensities of natural light, but they synthesise the antenna molecules to bind with rhodopsin for enhanced energy absorption and energy generation 40 . In the system presented here, the exogenous antenna molecule canthaxanthin was introduced which can be potentially biosynthesised. The addition of canthaxanthin did not adversely affect the viability of cells (Supplementary Method 3 and Supplementary Fig. 13 ). Furthermore, the GR-canthaxanthin complex was reported to have higher thermal stability compared to GR 27 . Reactive oxygen species (ROS) is a potential concern in photoelectrosynthesis because of their toxicity to cells 30 , 32 , 33 . In this study, a high proton motive force and reverse electron transfer could be possible factors contributing to ROS generation. To avoid the risk of ROS generation, we employed a dual-chamber electrochemical reactor separated by a Nafion proton exchange membrane (PEM) to reduce the effect of anodic oxygen generation on microbial carbon fixation in the cathodic chamber. This design ensures that oxygen produced from water splitting in the anodic chamber cannot permeate the PEM, thus preventing its entry into the cathode chamber to generate ROS. In addition, the N 2 /CO 2 syngas was continuously pumped into the cathode chamber to maintain an anaerobic condition. Although the anaerobic cathode chamber can effectively avoid ROS generation, it could sacrifice the cellular energy generated via the aerobic respiration pathway. Therefore, a strategic approach might involve controlling oxygen levels while simultaneously enhancing microbial resistance to ROS. Achieving a balance between reducing ROS effects and improving energy generation would be important for the future scaling-up and long-term operation of the system. There are some strategies to reduce ROS when bacteria are exposed to the risk of ROS. For example, engineering bacteria to express lycopene has been shown to improve microbial resistance toward ROS 41 . The addition of antioxidants, such as glutathione, has been shown effectively to reduce ROS in the system 42 . This artificial photosynthesis system is relatively simple, in contrast to natural chlorophyll-based photosynthesis which contains complex membrane structures after millions of years of evolution. Calculations can be made with respect to the overall photosynthesis efficiency of this system. But as yet, such calculations will have to be based on so many assumptions that a meaningful number is not yet attainable. Chlorophyll does not absorb photons in the photosynthetically available radiant (PAR) waveband evenly, with only a low degree of absorbance occurring in the green region. In contrast, rhodopsin strongly absorbs green-blue light (~500 nm). The complementary light absorption offers the possibility of integrating chlorophyll and rhodopsin photosystems in either a pure culture 43 or a mixed community 4 to maximise the utilisation of solar energy. The artificial photosynthesis outlined in this research is modularised, allowing for further optimisation. In this study, the addition of flavin and canthaxanthin enhanced the electron transfer rate and the proton-pumping rate, respectively. Material engineering such as electrode modification can provide effective approaches to further increasing energy efficiency 44 . It is even possible to achieve maximum utilisation of light energy by extending rhodopsin’s light absorption to non-visible wavelengths, which has already been shown feasible for the GR photosystem 45 . Many studies have demonstrated that microbial rhodopsins are major contributors to solar energy harvesting in the ocean, compared to chlorophyll 4 , 46 – 48 . Antennas such as canthaxanthin are found binding to rhodopsin and may have an important effect on rhodopsin phototrophy 40 . Rhodopsin phototrophy is recognised as independent of electron transfer and does not involve redox processes. In recent years, the proton motive force generated by rhodopsin has been demonstrated to drive NADH dehydrogenases in reverse, which couples the rhodopsin-based photosystem with the electron transport chain 9 , 10 . Interestingly, there is evidence showing transmission of Mtr-mediated extracellular electron transfer among oceanic bacteria 49 . In addition to the solid-phase electrode, syntrophic metabolisms 50 and inorganic minerals 51 could be used as natural electron sources for bacteria growth if they contain both rhodopsin and Mtr complexes. In nature, the ubiquity of rhodopsins might allow microbes with Mtr pathways to use various electron sources besides just water splitting. Therefore, our demonstration of artificial photosynthesis in R. eutropha could guide future research into connecting rhodopsin photosystems to extracellular electron transfer systems at the gene level in oceanic environments. It is plausible to hypothesise that autotrophic microorganisms, equipped with Mtr-like proteins and rhodopsin in nature, could potentially fix CO 2 powered by solar energy."
} | 5,411 |
30783143 | PMC6381156 | pmc | 8,851 | {
"abstract": "Recent explorations of scientific ocean drilling have revealed the presence of microbial communities persisting in sediments down to ~2.5 km below the ocean floor. However, our knowledge of these microbial populations in the deep subseafloor sedimentary biosphere remains limited. Here, we present a cultivation experiment of 2-km-deep subseafloor microbial communities in 20-million-year-old lignite coalbeds using a continuous-flow bioreactor operating at 40 °C for 1029 days with lignite particles as the major energy source. Chemical monitoring of effluent samples via fluorescence emission-excitation matrices spectroscopy and stable isotope analyses traced the transformation of coalbed-derived organic matter in the dissolved phase. Hereby, the production of acetate and 13 C-depleted methane together with the increase and transformation of high molecular weight humics point to an active lignite-degrading methanogenic community present within the bioreactor. Electron microscopy revealed abundant microbial cells growing on the surface of lignite particles. Small subunit rRNA gene sequence analysis revealed that diverse microorganisms grew in the bioreactor (e.g., phyla Proteobacteria , Firmicutes , Chloroflexi , Actinobacteria , Bacteroidetes , Spirochaetes , Tenericutes , Ignavibacteriae , and SBR1093). These results indicate that activation and adaptive growth of 2-km-deep microbes was successfully accomplished using a continuous-flow bioreactor, which lays the groundwork to explore networks of microbial communities of the deep biosphere and their physiologies.",
"introduction": "Introduction Over the past two decades, scientific ocean drilling has demonstrated that numerous microbes exist in the global deep subseafloor sediment, from the continental margins to open ocean gyres, comprising approximately 10 29 microbial cells and 4 Pg of biomass carbon on our planet 1 , 2 . Porewater geochemistry suggests that organic matter-fueled microbial energy respiratory activity is extraordinary low, ranging from 2.8 × 10 −18 to 1.1 × 10 −14 moles/e − /cell/year between the anoxic eastern equatorial Pacific and the oxic South Pacific Gyre sediments, respectively 3 – 5 . Culture-independent molecular ecological studies (e.g., PCR-mediated 16S rRNA and functional gene analysis, or metagenomics) of the above-mentioned subseafloor settings showed that they harbor diverse microbial communities, most of which are phylogenetically distinct from those living in the Earth’s surface environments 6 – 9 ; hence, their physiology and metabolic functions still remain largely unknown 10 , 11 . To gain insight into deep subseafloor microbial life, cultivation is crucial. Previous cultivation efforts on sediment core samples, however, indicated a high resistance of deeply buried microbial communities to conventional batch-type cultivation techniques. Consequently, only a small fraction of indigenous deep microbes could be isolated thus far from ≥10 m below seafloor (mbsf) sediment samples, whose members are primarily affiliated with the phyla Proteobacteria , Firmicutes , Actinobacteria , and Bacteroidetes or Euryarchaeota genera Methanoculleus , Methanococcus , and Methanosarcina 8 , 12 – 14 . Nevertheless, stable isotope tracer incubation experiments combined with nanometer-scale secondary ion mass spectrometry (NanoSIMS) analysis confirmed that more than 70% of the total microbial cells are viable, despite having very slow biomass turnover rates 15 , 16 . Thus, cultivation of deep subseafloor microbes through batch-type techniques may be impeded by their extraordinarily low metabolic activity under energy-limited conditions 5 and/or the “substrate-accelerated death” phenomenon, wherein microbial cells are damaged when suddenly exposed to high substrate concentrations in rich laboratory media 17 . Given the limited success of previous efforts to cultivate deep subseafloor microbes, new cultivation approaches are needed. Parkes et al . 18 applied a high-pressure anaerobic enrichment system (i.e., DeepIsoBUG) for gas hydrate-bearing sediments and successfully obtained some anaerobic bacteria (e.g., genera Acetobacterium and Clostridium ). Imachi and co-workers (2011, 2014, 2017) 19 – 21 applied a continuous-flow bioreactor cultivation technique to overcome the limitation of batch-type cultivation and successfully enriched previously uncultured lineages from deep subseafloor sediments. These studies employed a down-flow hanging sponge (DHS) reactor system, which was originally developed for treating municipal sewage in developing countries at a low cost 22 . Specifically, a polyurethane sponge used in the DHS reactor ensures medium pore space to provide a larger surface area for microbial colonization and extended cell residence time. Such continuous-flow bioreactor cultivation can maintain the low concentrations of substrates found in the natural environments and outflow the accumulated metabolic products that may inhibit microbial growth. These continuous-flow reactors thereby might increase the culturability of subseafloor microorganisms in a controlled manner and serve as better sources (incubators) for the isolation of microorganisms than the original samples. Recently, using a DHS reactor, Inagaki et al . 23 established a methanogenic enrichment culture from ~2-km-deep subseafloor coalbed samples obtained using the riser-drilling technology of the deep-sea drilling vessel Chikyu during the Integrated Ocean Drilling Program (IODP) Expedition 337. In this study, we report the extensive microbiological and biogeochemical investigations over 1000 days of DHS reactor operation, including the detailed cultivation procedure, microbial community structure, and microbial metabolism during the course of the bioreactor operation. We observed that phylogenetically diverse indigenous microbial populations were cultivated in the bioreactor. The cultivars seemingly grow on and transform coalbed-derived organic matter. Additionally, three anaerobic microorganisms, including a methanogenic archaeon, were obtained in pure culture from the bioreactor enrichment culture.",
"discussion": "Discussion In this study, we successfully obtained a methanogenic microbial community from 2-km-deep subseafloor coalbed layer samples. Increased sequence frequency of some indigenous populations that were present in the inoculum samples (Figs 5 and 6 , Supplementary Table S5 ) clearly indicated maintenance of cellular growth and population in the DHS reactor, suggesting that the continuous-flow DHS reactor system could be a powerful tool to cultivate indigenous microbial communities from the deep subseafloor sedimentary biosphere. The unique feature of the polyurethane sponge carrier contributed to the successful cultivation of a microbial community from lignite samples. Based on a previous study using subseafloor sediment 21 , we first assumed that the microorganisms enriched from the lignite samples would grow on the sponge surface and within its pore space. However, as shown in Fig. 4 and Supplementary Fig. S2 , the microbes colonized entirely on the lignite particle surfaces. The polyurethane sponge carriers functioned to retain lignite particles in the reactor column for long-term continuous-flow cultivation, whereas the lignite particles themselves served as microbial habitats. Consequently, the DHS reactor system is also applicable to cultivate fastidious microorganisms on solid substances such as coals, rocks, and minerals. Together, our chemical data is consistent with the existence of a microbial ecosystem that degrades lignite-derived organic material to diverse intermediate organic substances (e.g., acetate and high molecular weight, humic-like aromatic compounds) and methane as metabolic end products (Figs 2 and 3 , Supplementary Table S1 ). For example, the remarkable increase in methane concentration at around 600–700 days of the reactor operation shows a clear link between chemical and microbial data (Fig. 1a ). The δ 13 C values of the methane decreased sharply during this period (Fig. 2b ), indicating the occurrence of active CO 2 reducing methanogenesis. Shift of effluent pH values from approximately 7.3 to 7.8 (Fig. S1b) and an increase of the detection frequency of Methanobacterium sequences in the SSU rRNA gene-tag sequencing libraries (Fig. 5b ) also support the presence of active CO 2 reducing methanogenesis during this time. This active methanogenesis should be supported by the supply of methanogenic substrates (i.e., hydrogen and formate) that would be generated by active lignite-degradation. Microbial lignite-degradation in the reactor indeed was demonstrated by the increase of the AC/M ratio (Fig. 3c ). We interpret acetate to be a fermentation product of lignite-derived organic matters as acetate δ 13 C values were similar to those of higher-plant derived saturated hydrocarbons detected within natural lignite samples of Site C0020 33 . No known homoacetogen-related sequences in the SSU rRNA gene-tag sequence libraries (Supplementary Table S5 ) and absence of homoacetogen growth in the batch-type cultivation using the reactor enrichments (Supplementary Table S8 ) also suggests that acetate is primarily produced via fermentation rather than by homoacetogenesis 25 . EEM fluorescence spectroscopy analysis also indicated humic-like compounds as precursors for fermentation and that an ongoing turnover of these compounds is driven by microbial activity (Fig. 3 ). The majority of the methane produced in the DHS reactor is most likely generated via hydrogenotrophic (CO 2 reduction) methanogenesis. Carbon isotope fractionation between lignite-derived methane (minimum δ 13 C of −94.0‰; mean of −68.1‰) and organic matter (δ 13 C of −28‰ to −36.5‰) 33 was approximately about 70‰ at maximum, which is equivalent to that between methane and CO 2 in hydrogenotrophic methanogenesis under mesophilic, H 2 -limited conditions 34 . However, the δD values of the methane (−232.1 to -393.0‰) point to the potential contribution of aceticlastic/methylotrophic methanogenesis 24 , 35 , 36 . The methylotrophic signals are consistent with recent findings of position-specific 13 C enrichments of methoxy groups within the Shimokita lignites from Site C0020 37 , consistent with their microbial utilization by methanogens. Moreover, as the δ 13 C values of methane fluctuated and sometimes became isotopically heavy (from 210 to 455 days, from 693 to 721 days, and after 805 days, Fig. 1b ), aceticlastic/methylotrophic methanogenesis likely co-occurs next to hydrogenotrophic methanogenesis resulting in smaller stable carbon isotope fractionation 24 . The activity of both pathways was supported by SSU rRNA gene tag sequencing analysis, in which minor but archaeal OTUs were affiliated with aceticlastic/methylotrophic methanogen Methanosarcina and aceticlastic methanogen Methanosaeta together with the predominant OTUs of hydrogenenotrophic methanogen Methanobacterium . The enriched microbial community comprised phylogenetically diverse microorganisms (Figs 5 , 6 and Supplementary Fig. S3 , and Tables S4 – S6 ). During the cultivation period, major OTUs (>1% in any of the sponge carrier libraries) were affiliated with the phyla Proteobacteria , Firmicutes , Chloroflexi , Actinobacteria , Bacteroidetes , Spirochaetes , Tenericutes , Ignavibacteriae , and SBR1093. The phylum-level community composition was similar to those reported in the previous molecular studies of methanogenic communities in coal environments 38 – 42 and enriched communities from coal and its associated water 43 – 47 . However, the detailed composition of microbial community members enriched in our reactor at below the phylum level differed from the previously reported compositions, potentially due to the different sources of the inoculum and enrichment conditions (e.g., marine vs. terrestrial). When comparing detection frequency of the dominant OTUs in the sponge carrier and effluent samples, several OTUs were exclusively detected from either the sponge carrier or effluent samples (Supplementary Table S5 ). This result suggests that two types of microbial communities exist in the DHS reactor: one is the sessile community with members attached to the lignite surface and, depending on the substrates, inhered in lignite; the other one is the planktonic community with members relying on the substrates released from the degradation of lignite. This observation also suggests that stepwise biodegradation of organic substances in lignite were carried out in the DHS reactor, but inferring the detailed metabolic functions of each host microorganisms for the OTUs is difficult only from the SSU rRNA gene tag sequencing data (Supplementary Table S5 ). Therefore, to address this, further polyphasic investigation (e.g., subsequent isolation, metagenome/transcriptome analyses and stable isotope labeling experiments) using the DHS reactor enrichment culture is needed. Although the metabolic functions are not yet clarified for each microbial species at this moment, the host microorganisms for most of the predominant OTU probably play an important role in the degradation of coal-derived organic substances. For example, OTUs of the class Clostridia were frequently detected from the reactor samples at different times (e.g., OTUs 433, 285 and 198, Fig. 6 and Supplementary Tables S5 ). In general, Clostridia members are strictly anaerobic bacteria that degrade a wide variety of heterotrophic organic compounds 48 , 49 ; therefore, the reactor-enriched Clostridia bacteria may be relevant to the lignite-degradation. A Clostridia strain capable of decomposing polycatechol and humic substances was isolated from groundwater samples, including groundwater associated with a coal bed recovery site 50 . Likewise, other predominant OTUs belonging to Gammaproteobacteria , Alphaproteobacteria , Actionobacteria , and Bacteroidetes may be associated with the degradation of lignite-derived organic matter 40 , 51 , as the bacterial groups contain metabolically versatile bacteria. Notably, two predominant OTUs are closely related to hydrocarbon-degrading bacterial groups: n-alkanes degrader Alcanivorax (OTU164) and polycyclic aromatic hydrocarbon degrader Nitratireductor (OTU291). Those genera are generally recognized as aerobic microorganisms, but some strains can grow under anaerobic conditions 52 – 54 . Oxygen contamination could not be excluded from long-term reactor operation. Therefore, the presence of microbes degrading hydrocarbons using contaminant oxygen cannot be completely ruled out. On the other hand, a previous metagenomic study has identified high gene proportions for aerobic hydrocarbon metabolism enzymes in Canadian subsurface coalbed samples 55 . Therefore, there is a possibility that those microorganisms perform hydrocarbon-degradation under anaerobic conditions. We successfully obtained three isolates including a methanogen from the DHS reactor enrichment (Fig. 7 ). Identical 16S rRNA gene sequences with indigenous population-derived OTUs 37, 313, and 433 (Fig. 6 , Supplementary Table S5 ) support that these isolates originated from the 2-km-deep coalbed layers. Notably, the reactor isolates, Methanobacterium sp. strain MZ-A1 and Tenericutes sp. strain MZ-XQ share (almost) identical 16S rRNA gene sequences with Methanobacterium sp. strain MO-MB1 (100% sequence identity) and Tenericutes sp. strain MO-XQ (99.4%), respectively, both of which were isolated from shallow subseafloor sediments of the same drilling site in our previous study 19 . To examine if the evolutionary genome diversification is potentially affected by the geological time (approximately 20 million years of difference is estimated between the deposition ages of 2-km-deep coalbed layer and shallow subseafloor sediments), we are currently performing comparative genome analysis and physiological characterization for both “deep isolates” (i.e., strains MZ-A1 and MZ-XQ) and “shallow isolates” (i.e., strains MO-MB1 and MO-XQ). It is possible that deep subseafloor coalbed environments harbor more extant and diverse microorganisms than estimated herein. We performed alpha-diversity analysis of SSU rRNA tag sequencing data to estimate cultured microbial community diversity and richness in the DHS reactor (Supplementary Table S7 ). Notably, Shannon diversity index and Simpson’s evenness scores increased along with the reactor operation time, although enrichment cultivation generally exhibits a strong selective bias for microbial populations that can adapt to cultivation conditions. Conversely, species richness scores (i.e., Chao 1 and ACE) decreased immediately upon initial bioreactor cultivation, then increased and became stable. Generally, species richness scores decrease as enrichment cultivation proceeds. These results suggest that some predominant microbial populations in the inoculum sample could not grow, but many minor microbial populations could grow in the bioreactor. Specifically, we identified 97 and 149 OTUs, apparently derived from indigenous populations, from the samples before and after cultivation, respectively (excluding singleton OTUs, Supplementary Tables S4 and S5 ). Thus, microbial diversity in the inoculum samples was estimated to be low, although various indigenous populations were originally present in the inoculum samples. The low diversity estimation in the inoculum sample might be derived from (i) insufficient amount of DNA for PCR amplification obtained from most microorganisms owing to low abundance 23 ; (ii) insufficient sequence read numbers to cover the entire microbial community in the inoculum samples; and/or (iii) possible DNA extraction and PCR amplification biases affecting the tag sequencing results. Specifically, DNA extraction bias may have a significant impact on the estimation of microbial diversity, specifically in the inoculum sample. DNA extraction from deeply buried microbial cells is known to be difficult as compared to from other habitats, possibly because of rigid cell forms 56 . Using a standard DNA extraction kit, Morono et al . 57 demonstrated that at least 70% of subseafloor sedimentary microbial cells remained intact in DNA extraction residue. This suggests that many subseafloor sedimentary microbes possess a rigid cell envelope to provide resistance against common DNA extraction chemicals. Moreover, Lomstein et al . 58 reported abundance of bacterial endospores in deep subseafloor sediments. Endospores are dormant and mechanically tough structures produced by certain members of bacteria within phylum Firmicutes 59 . We also detected members of the phylum Firmicutes that were closely related to endospore-forming members in the inoculum and enriched samples (Supplementary Tables S5 ). Endospores are unlikely to be detected by either nucleic acid fluorescence staining 60 , 61 or rRNA-targeted fluorescence in situ hybridization techniques 62 , 63 . They are also resistant to physical-chemical cell lysis procedures of common DNA extraction methods 64 . Therefore, microbial components from endospores and rigid cell envelope forms may withstand DNA extraction and result in the underestimation of microbial diversity estimation in the pre-cultivation samples (Supplementary Tables S7 ). However, in the continuous-flow bioreactor, microbial cell membrane permeability and fluidity likely increased to incorporate energy substrates for active growth, potentially enhancing DNA yield, Shannon diversity index, and Simpson’s evenness scores in the enriched samples (i.e., Supplementary Tables S2 and S7 ). Thus, performing molecular-based analyses of microbial diversity not only on natural samples, but also on cultivated or enriched samples under the similar conditions to their natural habitat would provide important clues to clarify the microbial diversity in the deep subseafloor sediments. In summary, our >1000-day-long DHS bioreactor operation for 2-km-deep, 20-million-year-old lignite core sample demonstrates that a good fraction of the deeply buried subseafloor microbial community is cultivable, including key players in the anaerobic heterotrophic microbial ecosystem, such as various fermenters and methanogens. The enriched community consists of phylogenetically diverse microorganisms and possibly contains a concert of microbes that can convert complex coaly organic matters to methane. The enriched microbial community may be applicable to biological techniques that stimulate methane-production from low-rank coals and coalbed methane layers 40 , 65 – 69 . The cultured methanogenic community thus represents an attractive microbial entity for both science and engineering."
} | 5,207 |
33022255 | null | s2 | 8,852 | {
"abstract": "The role of microbes in sustaining agricultural plant growth has great potential consequences for human prosperity. Yet we have an incomplete understanding of the basic function of rhizosphere microbial communities and how they may change under future stresses, let alone how these processes might be harnessed to sustain or improve crop yields. A reductionist approach may aid the generation and testing of hypotheses that can ultimately be translated to agricultural practices. With this in mind, we ask whether some rhizosphere microbial communities might be governed by 'keystone metabolites', envisioned here as microbially produced molecules that, through antibiotic and/or growth-promoting properties, may play an outsized role in shaping the development of the community spatiotemporally. To illustrate this point, we use the example of redox-active metabolites, and in particular phenazines, which are produced by many bacteria found in agricultural soils and have well-understood catalytic properties. Phenazines can act as potent antibiotics against a variety of cell types, yet they also can promote the acquisition of essential inorganic nutrients. In this essay, we suggest the ways these metabolites might affect microbial communities and ultimately agricultural productivity in two specific scenarios: firstly, in the biocontrol of beneficial and pathogenic fungi in increasingly arid crop soils and, secondly, through promotion of phosphorus bioavailability and sustainable fertilizer use. We conclude with specific proposals for future research."
} | 390 |
29226470 | null | s2 | 8,853 | {
"abstract": "Hydrogel particles are versatile materials that provide exquisite, tunable control over the sequestration and delivery of materials in pharmaceutics, tissue engineering, and photonics. The favorable properties of hydrogel particles depend largely on their size, and particles ranging from nanometers to micrometers are used in different applications. Previous studies have only successfully fabricated these particles in one specific size regime and required a variety of materials and fabrication methods. A simple yet powerful system is developed to easily tune the size of polypeptide-based, thermoresponsive hydrogel particles, from the nano- to microscale, using a single starting material. Particle size is controlled by the self-assembly and unique phase transition behavior of elastin-like polypeptides in bulk and within microfluidic-generated droplets. These particles are then stabilized through ultraviolet irradiation of a photo-crosslinkable unnatural amino acid (UAA) cotranslationally incorporated into the parent polypeptide. The thermoresponsive property of these particles provides an active mechanism for actuation and a dynamic responsive to the environment. This work represents a fundamental advance in the generation of crosslinked biomaterials, especially in the form of soft matter colloids, and is one of the first demonstrations of successful use of UAAs in generating a novel material."
} | 353 |
26779171 | PMC4705240 | pmc | 8,854 | {
"abstract": "Bacillus species present a major concern in the dairy industry as they can form biofilms in pipelines and on surfaces of equipment and machinery used in the entire line of production. These biofilms represent a continuous hygienic problem and can lead to serious economic losses due to food spoilage and equipment impairment. Biofilm formation by Bacillus subtilis is apparently dependent on LuxS quorum sensing (QS) by Autoinducer-2 (AI-2). However, the link between sensing environmental cues and AI-2 induced biofilm formation remains largely unknown. The aim of this study is to investigate the role of lactose, the primary sugar in milk, on biofilm formation by B. subtilis and its possible link to QS processes. Our phenotypic analysis shows that lactose induces formation of biofilm bundles as well as formation of colony type biofilm. Furthermore, using reporter strain assays, we observed an increase in AI-2 production by B. subtilis in response to lactose in a dose dependent manner. Moreover, we found that expression of eps and tapA operons, responsible for extracellular matrix synthesis in B . subtilis , were notably up-regulated in response to lactose. Importantly, we also observed that LuxS is essential for B. subtilis biofilm formation in the presence of lactose. Overall, our results suggest that lactose may induce biofilm formation by B. subtilis through the LuxS pathway.",
"introduction": "Introduction Bacteria often use quorum sensing (QS) as cell–cell communication mechanism to control expression of genes that affect a variety of cellular processes ( Fuqua et al., 1994 ; Miller and Bassler, 2001 ; Bai and Rai, 2011 ). QS is based on production, secretion and response to small signaling molecules, termed autoinducers (AI; Bai and Rai, 2011 ). AI-2, a furanosyl-borate-diester ( Chen et al., 2002 ) is referred as a “universal autoinducer” as it is found in numerous Gram positive and Gram negative bacteria ( Schauder and Bassler, 2001 ; Xavier and Bassler, 2003 ). AI-2 is synthesized by LuxS through steps involving conversion of ribose-homocysteine into homocysteine and 4,5-dihydroxy-2,3pentanedione (DPD), a compound that cyclizes into several furanones in the presence of water ( Schauder et al., 2001 ). QS modulates various cellular processes involved mainly in the regulation of virulence factors, sporulation, motility, toxin production ( Hammer and Bassler, 2003 ; Henke and Bassler, 2004 ; Smith et al., 2004 ; Waters and Bassler, 2006 ) and formation of a structured multicellular community of bacterial cells, also termed biofilm ( Hall-Stoodley et al., 2004 ; Kolter and Greenberg, 2006 ). It appears that biofilm formation is the most successful strategy for bacteria to survive unfavorable environmental conditions ( Stewart and Costerton, 2001 ; Hall-Stoodley et al., 2004 ). Bacteria in biofilms are highly resistant to disinfection and antibiotic treatments, therefore biofilm formation is considered as a major problem in the industrial fields and in medicine ( Simoes et al., 2010 ). Bacillus subtilis is a Gram-positive non-pathogenic bacterium, which is a facile model microorganism for biofilm research. B. subtilis possesses the ability to form different types of biofilms in different environmental conditions: colony type biofilm at solid-air interface, pellicle at liquid–air interface as well as submerged biofilm at solid-liquid interface ( Vlamakis et al., 2013 ). B. subtilis cells can sense different environmental and physiological signals, which may activate one of its histidine sensor kinases. Those kinases are responsible for phosphorylation of Spo0A, a master regulator in the cell. Phosphorylated Spo0A leads to down-regulation of the transcriptional repressors AbrB and SinR, which keeps expression of genes for production of extracellular matrix turned off when conditions are not propitious for biofilm growth ( Branda et al., 2006 ; Vlamakis et al., 2013 ). When a signal is introduced for biofilm formation, B. subtilis cells are shifted from motile bacteria to bacterial chains that stick together by producing an extracellular matrix ( Branda et al., 2001 ; Kobayashi, 2007 ). The matrix has an important role during the biofilm formation. It provides an attaching source for other bacteria in the surrounding environment and therefore plays a crucial step in biofilm progression ( Branda et al., 2001 ; Kobayashi, 2007 ). The matrix consisted of two main components, an extracellular polysaccharide (EPS) synthesized by the products of the epsA-O operon, and amyloid fibers encoded by tasA located in the tapA-sipW-tasA operon ( Branda et al., 2006 ; Vlamakis et al., 2013 ). Biofilms formed by Bacillus species are vastly found throughout dairy processing plants ( Oosthuizen et al., 2001 ). Moreover, the major source of contamination of dairy products is often associated with members of the Bacillus genus ( Sharma and Anand, 2002 ; Simoes et al., 2010 ). Recently, it was found that certain milk components enhance biofilm formation by Bacillus species ( Pasvolsky et al., 2014 ). Lactose, a β1,4-linked disaccharide, is the main carbohydrate in milk and numerous dairy products. Our previous study showed that lactose increases biofilm formation by the Gram-positive bacteria Streptococcus mutans ( Assaf et al., 2015 ). Since lactose is an abundant disaccharide sugar in milk and its products, it might serve as an environmental trigger for biofilm formation by other bacteria too, for instance B. subtilis . Interestingly, it has been shown that B. subtilis might use QS to regulate motility and biofilm formation ( Lombardía et al., 2006 ). However, the link between sensing environmental cues and the QS induced biofilm formation by B. subtilis is poorly known. Therefore, the aim of this study was to investigate the role of lactose, the primary sugar in milk, on biofilm formation by B. subtilis and its possible link to QS process.",
"discussion": "Discussion Our results show that lactose triggers bundle formation as well as formation of colony type biofilm by B. subtilis. This result falls in line with our previous study which showed that lactose enhances biofilm formation by Streptococcus mutans ( Assaf et al., 2015 ). Expression of epsA-O and tapA operons, which are responsible for biofilm matrix production, were notably increased when lactose was added to the LB medium ( Figure 2 ). Interestingly, induction in expression of both operons is correlated with biofilm bundles formation by B. subtilis cells. Bundle formation is one of the first stages in biofilm development ( Branda et al., 2001 ). Moreover, investigation of the mutant strains for these operons shows absence of the bundling phenotype as a response to lactose ( Figure 3 ). This result indicates that lactose induce biofilm formation depends on tapA and epsA-O expression. In recent years, there has been an increasing interest in the quorum-sensing signaling molecules related to food spoilage. Various signaling compounds associated with QS, such as AI-2, have been detected in different food systems such as milk ( Pinto et al., 2007 ). Furthermore, studies have shown that QS is important for social behavior of B. subtilis and other bacteria ( Lombardía et al., 2006 ). Using V. harveyi as a reporter strain for bioluminescence, we were able to track the level of produced AI-2 molecules. We observed an increase in the AI-2 production as a response to lactose in dose dependent manners ( Figure 4 ). It has been shown previously that simple dietary sugars can affect QS, specifically production of AI-2 by Klebsiella pneumoniae ( Zhu et al., 2012 ). In our study, the cell density of all tested samples was the same at the sampling time, consequently, changes in the AI-2 production is apparently not related to the cell density but to the metabolic state of the bacteria. Thus, our results support previous studies that showed that AI-2-dependent signaling is a reflection of metabolic state of the cell and environmental factors and not cell density ( Bassler, 1999 ; Beeston and Surette, 2002 ). Previous studies also suggested that activation of QS through LuxS can be regulated in response to sugar metabolism by cyclic AMP receptor protein molecules ( Lyell et al., 2013 ). In B. subtilis cells, lactose may affect the energetic metabolic balance in the cell, and through second messengers such as cyclic AMP, or CCP can lead to expression of QS genes such as luxS . The main finding of this study is the apparent link between lactose induced biofilm formation and activation of QS system through increased production of AI-2 molecules in B. subtilis . Addition of synthetic precursor for AI-2, DPD, to the media resulted in enhanced bundle formation as well as up-regulation of tapA expression ( Figure 5 ). Similarly, the direct effect of AI-2 molecules on EPS biosynthesis has been observed previously in Vibrio cholera e where the AI-2 molecules up-regulated expression of the EPS biosynthesis genes ( Hammer and Bassler, 2003 ). According to our results, examination of biofilm formation in CDM of the B. subtilis ΔluxS mutant resulted in deficiency of biofilm formation (bundle, and colony types) ( Figures 6A and 7 ). These results suggested that QS via LuxS cascade plays an important role in biofilm formation in the presence of lactose. This is consistent with previous research which showed that LuxS is important for B. subtilis social behavior (motility and biofilm formation) ( Lombardía et al., 2006 ). Another study showed that blocking the AI-2 pathway, using an AI-2 analog, inhibited biofilm formation by B. subtilis ( Ren et al., 2002 ). Similar results were found for Hafnia alvei , a food-related bacterium that can be found in dairy products. QS in H. alvei is required for differentiation of individual cells into a complex multicellular structure of biofilm ( Souza Viana et al., 2009 ). Interestingly, we observed that the luxS mutant strain could form pellicle in biofilm promoting medium LBGM ( Figure 8 ). Although, a pellicle formation in LBGM appears to be not LuxS dependent, it seems that in CDM there is a slight induction in pellicle formation in response to lactose ( Figure 6B ). As it was shown recently ( Shemesh and Chai, 2013 ), glycerol and manganese activate KinD-Spo0A pathway for matrix production. In case of lactose, it seems that enhanced production of AI-2 affects not directly on the biofilm formation cascade. This may explain the differences found between phenotypes in CDM supplemented with lactose and in LBGM. Activation of biofilm formation via QS system might be an additional regulatory mechanism which enables fine tuning of the biofilm formation pathway that has been previously described ( Shemesh and Chai, 2013 ). FIGURE 8 Pellicle formation by B. subtilis in LBGM is not LuxS dependent. WT and ΔluxS cells were used for pellicle formation in LBGM. The pictures were taken using a Zeiss Stemi 2000-C microscope with an axiocamERc 5s camera. Images are representative of two biological repeats. The LuxS system possesses an inherent metabolic function in the activated methyl cycle; phenotypic defects in luxS mutants may not strictly be attributed to AI-2 signaling but possibly to metabolic disturbances. For instance, biofilm defects in a Lactobacillus rhamnosus luxS mutant are not restored by AI-2 molecules but rather by the addition of cysteine, indicating a sole metabolic role of LuxS ( Lebeer et al., 2007 ). In order to test whether the deficiency of biofilm formation in the presence of lactose in the mutant strain is due to AI-2 signal molecules or due to metabolic reason, we used DPD for complementation tests. It was shown previously that the synthetic AI-2 precursor (DPD) has the ability for specific AI-2 complementation during biofilm formation by Streptococcus intermedius ( Ahmed et al., 2008 ). In the complementation test, we observed restoration of the biofilm phenotype. The ΔluxS mutant showed ability for increased bundle formation in media supplemented with lactose and 200 μM of DPD ( Figure 7 ), indicating that the abolished biofilm formation is mostly connected to AI-2 and not to LuxS enzyme function. In overall, results of the present study suggest that QS via LuxS system plays an important role in biofilm formation induced by lactose in B. subtilis . As lactose affects activation of LuxS system, it is likely related to activation of Spo0A which leads to biofilm formation through a known pathway of up-regulation of the extracellular matrix operons. Moreover, Spo0A has been shown to be a negative regulator of LuxS system ( Lombardía et al., 2006 ). Additional research on lactose in association with QS will further elucidate the role of QS in biofilm formation of Bacilli and the effect of this dairy component on biofilm related gene expression."
} | 3,231 |
20979240 | null | s2 | 8,856 | {
"abstract": "No abstract available"
} | 5 |
35280331 | PMC8914556 | pmc | 8,857 | {
"abstract": "We report a highly stretchable hydrogel based on the crosslinking structure between calcium cations and alendronates (ALN) conjugated with poly-γ-glutamate (γ-PGA), a typical biodegradable polymer. γ-PGA with ALN (γ-PGA-ALN) forms the hydrogel in the aqueous solution containing CaCl 2 . The hydrogel shows 2000% of stretchability and reversible stretching without failure at a strain of 250%. The fracture strain and stress are tunable by varying the concentration of either γ-PGA-ALN or CaCl 2 , indicating the importance of fine-tuning of the density of the cross-linkage to control the mechanical properties of the hydrogel. We believe the biodegradable polymer based highly stretchable hydrogel has potential for use in various fields such as tissue engineering.",
"conclusion": "4 Conclusions In this work, we developed the highly stretchable hydrogel consisting of the biodegradable polymer by engineering the crosslinking structure between calcium cations and ALNs in a γ-PGA-ALN polymer network. γ-PGA-ALN formed a hydrogel in the presence of CaCl 2 and the obtained γ-PGA-ALN gel showed 2000% of stretchability and reversible stretching without failure at a strain of 250%. The mechanical properties of the γ-PGA-ALN gel could be controlled by tuning the concentration of γ-PGA-ALN or calcium cations in addition to the introduction of the permanent cross linker. Since NHDF adhered on/in γ-PGA-ALN gel by mixing gelation and subsequent culture, we believe the biodegradable polymer-based hydrogel with tunable mechanical properties reported in this study has the potential to be used as a scaffold in tissue engineering for bone regeneration.",
"introduction": "1 Introduction Biodegradable polymers are of great interest in the fields of tissue engineering and/or drug delivery [ 1 ]. Mechanical properties such as strength, stiffness, toughness or stretchability are often some of the most important functions in the application of these materials [ 2 ]. Mechanobiological studies have revealed that the mechanical properties of living constructs such as the extracellular matrix (ECM) strongly affect various biological processes from the molecular to tissue level, including tissue morphogenesis, cancer progression and tissue re-modeling [ 3 ] as well as ensuring the mechanical stability and structural integrity of tissues and organs [ 4 ]. Regulation of the mechanical properties of ECM often relies on the non-covalent cross-linkages including hydrogen bonding, electrostatic interaction and/or hydrophobic interactions in addition to the covalent crosslinked structure [ 5 , 6 ]. Non-covalent crosslinking allows reversible association/dissociation, so it can exchange the crosslinking points upon mechanical stress, enabling the high and reversible stretchability of ECM [ 7 , 8 ]. The introduction of non-covalent crosslinking in a polymer is also an effective strategy for enabling the control of the mechanical property of a synthetic hydrogel. A supramolecular host-guest recognition [ [9] , [10] , [11] ], [ [9] , [10] , [11] ] metal coordination [ 12 ], electrostatic interaction [ 13 ], hydrophobic interaction [ 15 ], and/or hydrogen bonding [ 14 ] have been introduced in synthetic hydrogels as non-covalent crosslinkers. Calcium ions are essential for the construction of inorganic-organic hybrid materials for various biological functions such as protection and/or mechanical support in living systems. These are often three-dimensional macromolecular assemblies of proteins, polysaccharides, and/or glycoproteins [ 16 ]. In the fabrication of hydrogel materials, it has been reported that the alginate forms a hydrogel by the crosslinking structure with calcium cations, known as the egg box junction [ 17 ]. Suo et al. reported that an alginate-calcium hydrogel with an interpenetrating network consisting of crosslinked polyacrylamide showed remarkable stretchability and toughness [ [18] , [19] , [20] ]. Taking inspiration from biomineralization, Cölfen et al. reported a highly stretchable hydrogel consisting of amorphous calcium carbonate and poly-acrylic acid, a material they named ‘mineral plastic’ [ 21 ]. Bisphosphonates such as alendronate (ALN) are known to have a high affinity with calcium cations, and have been used medicinally for osteoporosis [ 22 , 23 ]. In addition to its use as a small molecule drug, the conjugation of ALN with a polymer has been considered to be a promising strategy for engineering a bone-targeting carrier for drug delivery, regeneration therapy and/or tissue engineering [ [24] , [25] , [26] ]. Since ALN forms a 2:1 complex with a calcium cation [ 27 ], the interaction between ALNs and calcium cation can function as a non-covalent crosslinker in a hydrogel network. It has been reported that hydrogels consisting of polymers with ALN showed self-healing properties due to the dynamic nature of the cross-linkage. [ [28] , [29] ]. However, little has been reported on the fabrication of highly stretchable hydrogels based on the crosslinking between a biodegradable polymer functionalized with ALN and calcium cations. In this study, we demonstrate a highly stretchable hydrogel with tunable mechanical properties consisting of poly-γ-glutamate functionalized with ALN (γ-PGA) and calcium cations. γ-PGA is a naturally occurring polymer secreted by a Bacillus subtilis strain , and provides a potential resource for environmental and biodegradable materials [ 30 ]. We previously reported γ-PGA hydrogels crosslinked with disulfide bonds for the construction of three-dimensional engineered tissues by collapsing the hydrogels in response to a reduction condition [ 31 , 32 ]. ALN conjugated γ-PGA is expected to form crosslinking structure in response to calcium cations for the provision of a biodegradable hydrogel. We believe the highly stretchable hydrogel with tunable mechanical properties consisting of the biodegradable polymer would be a promising material as a scaffold in tissue engineering for bone regeneration.",
"discussion": "3 Results and discussion 3.1 Preparation of γ-PGA-ALN gels ALN was introduced in γ-PGA, a typical biodegradable polymer [ [30] , [31] , [32] , [33] ] by amide coupling reaction using (4-4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methyl-morpholinium chloride (DMT-MM) as a coupling reagent ( Fig. 1 a) [ 34 ]. The obtained γ-PGA-ALN was characterized by 1 H NMR, 31 P NMR and FT-IR ( Figs. S1 and S2 ). The incorporation amount of ALNs in γ-PGA was determined to be 11% from the elemental analysis. To validate the capability of γ-PGA-ALN to form a hydrogel responding to calcium cations, 100 μL of 10 wt% γ-PGA-ALN was added into 1.5 mL of water containing 1 M CaCl 2 and incubated at room temperature ( Fig. 1 b). A transparent hydrogel immediately appeared upon the addition of γ-PGA-ALN into the CaCl 2 solution, and then the hydrogel turned white within 10 min (γ-PGA-ALN in Fig. 1 c). Here, we confirmed that γ-PGA without ALN showed no gelation in the CaCl 2 solution (γ-PGA in Fig. 1 c). In addition, gelation was not observed when γ-PGA-ALN was added into the solution containing mono-cations such as lithium, sodium and potassium due to the weak affinity for ALN. γ-PGA-ALN formed a fragile film on the solution surface in response to MgCl 2 , whereas γ-PGA-ALN formed rigid gels in response to BaCl 2 and CuCl 2 ( Fig. 1 d). Since these divalent cations have different size [ 35 ], these results may indicate that the size of di-valent cations affects the capability to form the stretchable polymer network with γ-PGA-ALN. We concluded from these results that ALNs in γ-PGA form crosslinking structures with calcium cations by selective interaction ( Fig. 1 e) [ 28 , 29 ]. The stress-strain curve measurement of the molded γ-PGA-ALN gel revealed the high stretchability of the hydrogel: the fracture strain and stress were determined to be 1900% and 23 kPa, respectively ( Fig. 2 a and b). The loading–unloading curve of the hydrogel shows a hysteresis loop, indicating the energy dissipation of the hydrogel ( Fig. 2 c). These results may indicate an extension of polymers from collapsed globules to a stretched coil structure upon tension, that is coupled with exchanges of the crosslinking points between ALN and calcium cations. The cyclic loading–unloading curve also confirmed a reversible stretch of the hydrogel without significant loss of stress at a strain of 250%. In addition, the hysteresis loop decreased with the increase of the number of the cycle from 1 to 3, and then reached constant for further cycles. This may arise from the rearrangement of the hydrogel network occurred during the cyclic test. Fig. 1 Preparation of γ−PGA-ALN gels. (a) Synthesis of γ-PGA-ALN by amide coupling reaction. (b) Schematic illustration of the γ-PGA-ALN hydrogel preparation. (c) Photographs showing the formation of the hydrogel upon the addition of γ-PGA-ALN into an aqueous solution containing CaCl 2 . The transparent pellet (left) became a white hydrogel (middle) within 10 min, whereas γ-PGA did not form a gel (right). (d) γ-PGA-ALN in the aqueous solution containing 1 M LiCl, NaCl, KCl, CaCl 2 , MgCl 2 , BaCl 2 or CuCl 2 . (e) The proposed structure of crosslinking structure between ALNs and a calcium cation. Fig. 1 Fig. 2 Stretching property of γ-PGA-ALN gels. (a) Photograph showing the high stretchability of the γ-PGA-ALN gel. Schematic illustration depicts the proposed conformational change of γ-PGA-ALNs in the hydrogel. (b) The stress-strain curve of the γ-PGA-ALN gel prepared from 10 wt% γ-PGA-ALN and 1 M CaCl 2 . (c) The stress-strain curves for the reversible stretch of the γ-PGA-ALN hydrogel (10 wt%). The inserted figure shows the stress for each cycle. Fig. 2 3.2 Effect of calcium ion and polymer concentrations on mechanical property of γ-PGA-ALN gels Next, we prepared the γ-PGA-ALN gel with various polymer concentrations at a fixed CaCl 2 concentration (1 M) and incubation time (10 min) to demonstrate tunable mechanical properties of the hydrogel ( Fig. 3 a). The fracture stress increased with an increase in concentration of γ-PGA-ALN, whereas the fracture strain decreased due to the high density of crosslinking structure in the hydrogel. The influence of ALN incorporation amount in γ-PGA was also investigated using γ-PGA functionalized with 37% ALNs ( Figs. S1 and S2 ). γ-PGA with 37% ALNs formed aggregation at lower calcium concentration than γ-PGA with 11% ALNs, whereas, obtained aggregation was rigid and no longer stretchable hydrogel ( Fig. 3 b). When ALNs are concentrated excessively in polymer networks, polymers would form excess crosslinked networks with calcium ions both intra- and inter-polymer, that fix the γ-PGA-ALN as a rigid globule polymer. We concluded from these results that the fine tuning of the density of the cross-linkage between ALN and calcium cations in a polymer network allows the control of the mechanical properties. We further investigated the influence of the CaCl 2 concentration and incubation period during the gelation on mechanical properties of the hydrogel consisting of γ-PGA with 11% ALNs at a fixed concentration of precursor polymer solution (30 wt%). The concentration of calcium cations during the gelation significantly affected the mechanical properties of hydrogels when the incubation time was fixed at 10 min: the fracture strain decreased from 3000%–200% and the fracture stress increased from <10 kPa to 400 kPa with the increase of CaCl 2 concentration from 100 mM to 4 M ( Fig. 3 c). Although there was the trade-off relationship between stretchability and stress, a maximum stretchability of γ-PGA-ALN hydrogel (3000% of maximum strain) observed at lowest CaCl 2 concentration was higher than or comparable with that in earlier reports on highly stretchable hydrogels [ [18] , [19] , [20] , [36] , [37] , [38] ]. We concluded from this result that the tuning of the occupancy of ALNs with a calcium cation in the hydrogel allows the control of the mechanical properties of the biodegradable polymer based γ-PGA-ALN hydrogel across a wide range. On the other hand, the influence of the incubation period on the stress and strain was slight when the hydrogel was prepared in the presence of 1 M CaCl 2 ( Fig. 3 d). It may be more likely that the density of the cross-linkage between ALN and calcium cations is dominated by the equilibrium stability of the complexation. Fig. 3 Effect of calcium ion and polymer concentrations on mechanical property of γ-PGA-ALN gels. (a) The fracture stress and strain of the γ-PGA-ALN gel prepared from various concentrations of γ-PGA-ALN. The concentration of CaCl 2 was fixed at 1 M. (b) The effect of the population of ALN conjugated with γ-PGA on gelation. 30 wt% of γ-PGA-ALN was added into water containing various concentrations of CaCl 2 and incubated for 10 min. (c) The effect of CaCl 2 concentration on the stress and strain of the γ-PGA-ALN gel. The incubation time was fixed at 10 min. (d) The effect of the incubation time in the gelation process on the stress and strain of the γ-PGA-ALN gel. CaCl 2 concentration was fixed at 1 M. Fig. 3 3.3 Crosslinking mechanism of γ-PGA-ALN gels The interaction between γ-PGA-ALN and calcium cations was assessed by the evaluation of the absorption amount of calcium cations by the hydrogel ( Fig. 4 ). When the incubation time was fixed at 10 min, the absorption amount linearly increased with an increase in concentration of CaCl 2 from 0.1 to 2 M and decreased at 4 M ( Fig. 4 a). The decreased absorption at 4 M CaCl 2 may have arisen from the squeezing effect induced by the salting out. On the other hand, the effect of incubation time on the absorption was slight ( Fig. 4 b). These results are consistent with the arguments put forward earlier. Interestingly, the calcium ions absorbed by the gel were five to ten times greater than the number of ALNs in the gel. This result indicates that remaining glutamate residues in the polymer function to retain excess calcium ions in the hydrogel by weak electrostatic interaction [ 23 ], perhaps, that provide calcium ions to the alendronates effectively upon the stretch as sacrificial bonds [ 9 ]. We further assessed the effect of the introduction of the additional covalent crosslinker, ethylene glycol diglycidyl ether (EGDGE) in the γ-PGA-ALN gel on the mechanical properties of the gel ( Fig. 4 c and S3 ). The addition of EGDGE during the molding process resulted in the increased fracture stress and decreased fracture strain, respectively, due to the introduction of a permanent crosslinking structure ( Fig. 4 d). This result emphasizes the importance of the dynamic nature of the crosslinking between ALN and calcium cations for the high stretchability. This result also indicates the capability of the introduction of the permanent crosslinker to control the mechanical stability and property of γ-PGA-ALN hydrogel. Indeed, the stability of the hydrogel in PBS could be prolonged by the introduction of EGDGE ( Fig. S4 ). Lastly, we preliminary confirmed the adhesion and elongation of normal human dermal fibroblasts (NHDF) on γ-PGA-ALN hydrogel ( Fig. S5 ), indicating its suitability as a scaffold for tissue engineering, although further investigations are currently ongoing. Fig. 4 Crosslinking mechanism of γ-PGA-ALN gels. (a) The effect of CaCl 2 concentration on the absorption of calcium cations by the γ-PGA-ALN gel. The incubation time was fixed at 10 min. (b) The effect of the incubation time on the absorption of calcium cations by the γ-PGA-ALN gel. CaCl 2 concentration was fixed at 1 M. (c) The chemical structure of γ-PGA-ALN covalently crosslinked by EGDGE. (d) The effect of covalent crosslinking on the mechanical properties of the γ-PGA-ALN gel. The weight ratio of EGDGE/γ-PGA-ALN is inserted in the figure. The concentration of γ-PGA-ALN was fixed at 10 wt%. Fig. 4"
} | 3,972 |
34873881 | PMC8649616 | pmc | 8,858 | {
"abstract": "Abstract Waste plastics are non‐degradable constituents that can stay in the environment for centuries. Their large land space consumption is unsafe to humans and animals. Concomitantly, the continuous engineering of plastics, which causes depletion of petroleum, poses another problem since they are petroleum‐based materials. Therefore, energy recovering trough pyrolysis is an innovative and sustainable solution since it can be practiced without liberating toxic gases into the atmosphere. The most commonly used plastics, such as HDPE, LDPE (high‐ and low‐density polyethylene), PP (polypropylene), PS (polystyrene), and, to some extent, PC (polycarbonate), PVC (polyvinyl chloride), and PET (polyethylene terephthalate), are used for fuel oil recovery through this process. The oils which are generated from the wastes showed caloric values almost comparable with conventional fuels. The main aim of the present review is to highlight and summarize the trends of thermal and catalytic pyrolysis of waste plastic into valuable fuel products through manipulating the operational parameters that influence the quality or quantity of the recovered results. The properties and product distribution of the pyrolytic fuels and the depolymerization reaction mechanisms of each plastic and their byproduct composition are also discussed.",
"conclusion": "6 Conclusion and Future Prospects The pervasive consumption of plastics paired with their long lifetime and the huge volumes in use causes environmental challenges. Various strategies are practiced to manage these disposals, such as recycling, landfilling, and incineration. Unfortunately, these strategies are not sufficient to effectively manage the plastic waste without releasing toxic chemicals. This unveils the importance of plastic recycling and innovative technologies to deal with this problematic surplus in an environmentally friendly manner. The pyrolysis technology is considered the best method and sustainable solution that may be economically profitable on a large scale, addressing environmental concerns regarding waste minimization, carbon sequestration, soil amendment, energy/heat supply, and value‐added products. Pyrolysis technology is used to convert waste plastics into liquid oil (fuel) and other valuable byproducts such as char and gases under controlled conditions and is considered to be a relatively environmentally friendly technology when compared to uncontrolled incineration and landfilling practices. The pyrolysis products depend on several process parameters such as temperature, catalyst choice, heating rate, carrier gases, retention time, kind of plastics, reactor, and pressure. A yield of up to 80–90 % of fuel oil by weight can be recovered from thermoplastic waste. The produced liquid oils are either gasoline or diesel fractions, which have comparative characteristics with commercial diesel or gasoline fuels. The gases recovered from the pyrolysis of waste plastics are mainly C 1 −C 4 compounds for the polyolefins, and additional gases like CO, CO 2 , and HCl are obtained from PC, PET, PVC. High temperature and retention time are the main limitations of the pyrolysis of plastic wastes that need to be optimized to make the process more economical and environmentally friendly. Pyrolysis chars can also be activated for use in the absorption of heavy metals, smoke, and odor, and as an input to other material productions. Several catalysts have been studied to upgrade the quality of liquids obtained from the pyrolysis process. Specifically, exploration and utilization of naturally occurring and cheaper catalysts like natural zeolites require more research. Moreover, catalyst modification requires further attention to improve their performance to optimize the pyrolysis method. Nano‐sized zeolite crystals should be investigated for their conversion efficiency and selectivity of fuels oils obtained from waste plastics. Integrating and combining catalysts to bimetallic, trimetallic, and other systems should also be studied for use in catalytic pyrolysis. Efforts should be made to lower the pyrolysis temperature in order to lower energy consumption, in addition to using the latest technologies such as plasma, microwave irradiation, and continuous systems to provide tunable and scalable pyrolysis procedures. Computer‐based technologies and models, including such based on quantum mechanics, should be incorporated to investigate the detailed pyrolysis reaction mechanism of the polymers. PP, PE, PS, and PC waste plastics are suitable for pyrolysis with a good yield of fuels, but PVC and PET are not suitable because of hazardous chlorinated gas escaping, at already a low temperature, from PVC and toxic heteroatom‐containing gases produced from PET in addition to their low liquid yield. Therefore, more detailed studies are required to improve the products from these polymers, or other chemical or mechanical recycling methods should be applied for such types of plastics. Incorporating PET or PVC with biomasses during the pyrolysis process should be investigated for synergistic advantages. Safety is one of the primary concerns during the construction and designing of the pyrolytic reactors. Therefore, reactors should be designed using standard codes so that they perform under high temperature and pressures without rupture. Finally, as municipal waste management is challenging in most urban environments and cities, responsible government should incorporate proper waste management systems in their policies.",
"introduction": "1 Introduction Plastics are among those materials innovated by human beings for their need. They are highly inevitable materials employed in a wide range of applications making our day‐to‐day activities easy in home, shop packing, marketing, constructions, and healthcare, due to their lightweight, chemical stability (do not rust or rot), availability, and can be used repetitively. Their replacement for natural resources such as metal pipes and woody materials gains much acceptance. \n [1] \n Single‐use plastics such as masks, gloves, containers, medical packaging, and utensils of the ongoing COVID‐19 pandemic are certainly affecting waste plastic management. \n [2] \n Plastics do not have natural equivalents because they can replace many natural nonrenewable resources such as metals, woods, glasses in different sectors. \n [3] \n The ceaseless growth of the consumption of plastics is vast and has been rising steadily because of the advantage derived from their flexibility, low cost, and durability throughout the world. As shown in Figure 1 below, the USA leads the world plastic per capita consumption with 142 Kg/year. \n [4] \n Nearly 6.3 billion metric tons of plastics have been manufactured in 2015, of which 79 % of the total product was sent to landfill, 12 % incinerated, and 9 % recycled. \n [5] \n Half of the total plastics manufactured in the European Union end up as waste every year and become the third‐largest contributor to municipal solid waste (MSW) after food and paper wastes.[ \n 1 \n , \n 6 \n , \n 7 \n ] The annually generated plastic waste is expected to grow at a rate of 3.9 % per year. \n [8] \n Globally, the population growth, industrial expansion, consumer demand, and depletion of resources have contributed to the wide usage of plastics and become more serious to the environment than ever. These days, plastics are among the most environmentally devastating and challenging wastes due to their huge quantities and disposal difficulties. \n [10] \n The huge volume of waste plastics that resulted from the dramatic growth in their production and consumption give rise to serious concerns, as they do not degrade and remains in municipal refuse for decades. Plastic wastes are more voluminous than the other organic wastes and thus take up a lot of landfill space that is becoming scarce and expensive. Incineration of the accommodated waste plastics cannot be a popular solution since a huge amount of gases (HCl, dioxins, SO x , NO x , and CO 2 among others), are emitted, and if these toxic gases are inhaled for a long period of time, it can lead to respiratory problems in addition to their contribution to the global warming and acid rain.[ \n 11 \n , \n 12 \n ] Recycling requires an effort to perform the transformation sequence, which may lead to environmental, labor‐intensive, and cost impacts. Moreover, recycling faces a lot of challenges such as logistics, manpower, and financial concerns, as well as a lack of consumer awareness and education. The cost could be an issue if recycled plastic has to compete with virgin plastics that are manufactured at a comparatively low cost. Besides, not all plastic wastes are equally good for recycling. \n [13] \n \n Figure 1 Global per capita consumption of plastics (Kg year −1 ). Reproduced with permission from Ref. [9]. Copyright 2010, Elsevier. Disposal of waste plastics does not stop in one location as they can enter into oceans, lakes, rivers, and water bodies which causes the formation of garbage patches that can poison the health of the entire aquatic life. Furthermore, they can also affect the economy and food supply of the societies that have depended on fishing. \n [14] \n Waste plastics not only cause damage to the ocean but also damage groundwater sources.[ \n 15 \n , \n 16 \n ] Plastics such as high‐density polyethylene (HDPE), low‐density polyethylene (LDPE), polypropylene (PP), polystyrene (PS), polyvinyl chloride (PVC), polyethylene terephthalate (PET), and polycarbonate (PC) are now indispensable, relatively cheap, durable, and versatile synthetic materials and their application in industrial fields are continually increasing. Petroleum‐based plastics are mainly composed of hydrocarbons, but they contain different additives like antioxidants, colorants, stabilizers, and plasticizers. When the plastics are discarded, the additives are also undesirable from the environmental point of view since they might be water‐soluble. Plastics are non‐biodegradable materials that are extremely troublesome components for landfilling since their lifespan is too high (Table 1 ).[ \n 4 \n , \n 10 \n , \n 11 \n , \n 17 \n ]\n Table 1 Density, crystallinity and lifespan of thermoplastic polymers. \n [18] \n \n \n Plastic \n \n Density [23/4 °C] \n \n Crystallinity [%] \n \n Lifespan [year] \n \n PE \n \n 0.91–0.925 \n \n 50 \n \n 10–600 \n \n PP \n \n 0.94–0.97 \n \n 50 \n \n 10–600 \n \n PS \n \n 0.902–0.909 \n \n 0 \n \n – \n \n PET \n \n 1.03–1.09 \n \n 0–50 \n \n 450 \n \n PVC \n \n 1.35–1.45 \n \n 0 \n \n 50–100+ \n Wiley‐VCH GmbH According to their origin, wastes plastics are classified as industrial and municipal waste that have different compositions and properties when subjected to different management strategies. \n [19] \n These days, waste plastic managements are challenging in urban settings, since their huge quantities accommodate as a byproduct or faulty product from the commerce and agriculture sectors. Of the total waste plastics, over 78 wt % correspond to thermoplastics, and the remaining are thermosets. Thermoplastics such as HDPE, LDPE, PP, PS, and PVC are composed of polyolefins that have the possibility to be recycled easily. However, recycling of thermosets waste plastics is challenging due to epoxy resins and polyurethane origins. \n [9] \n \n HDPE, LDPE, PP, PS, PVC, and PET are reported as the most common municipal solid wastes in Europe. \n [20] \n However, PE plastics (HDPE and LDPE) are the most popular and make up over 40 % of the total content of municipal solid wastes. [21 ] The chemical processes such as thermal and catalytic methods of converting the waste into energy and value‐added fuels/chemicals are promising techniques to eliminate the plastic refuse, which otherwise is a major cause of environmental contamination. Extracting fuel oils from waste plastics can also decrease the dependence on fossil fuel since the plastic manufacturing industry uses nearly 6 % of petroleum produced worldwide. Therefore, it is like ‘killing two birds with one stone’ in terms of saving the supply of energy and alleviating environmental concerns.[ \n 22 \n , \n 23 \n , \n 24 \n ] Nowadays, tertiary recycling technologies of converting waste plastics to chemicals and value‐added fuels such as pyrolysis, gasification, and depolymerization are of recent interests for waste management. Pyrolysis (also called thermolysis), thermal cracking, catalytic cracking, and liquefaction convert waste plastics to gases, liquids, and waxes under high temperatures, either in the absence of a catalyst (thermal) or in the presence of a catalyst (catalytic pyrolysis). \n [25] \n \n Waste plastics are resources that open many opportunities like job‐creating, growth, innovation, and sustainability and have multiple effects on society and the economy. Some countries banned landfill and incineration; therefore, energy recovery from the waste resources through pyrolysis is the best choice for waste management. Having all the environmental risks of the non‐degradable waste plastics, it is timely and urgent to review their conversions to energy fuels through thermal or catalytic pyrolysis. Prior to our review, some excellent review papers had already been communicated by Miandad et al.,[ \n 26 \n , \n 27 \n ] Kasar et al., \n [28] \n Williams, \n [29] \n Dwivedi et al., \n [30] \n Lopez et al., \n [31] \n Wong et al., \n [32] \n Chen et al., \n [33] \n which basically deal with the effectd of the operational parameters on the pyrolysis, and the latest review by Hou et al. \n [25] \n summarizes the catalytic degradation of the plastics that cover the reaction mechanisms of the polymers in detail. However, the present review is presented in a comprehensive way of discussing the pyrolysis of the most frequently used polymers, including the factors that affect the pyrolysis process, physicochemical properties of the recovered fuels, proposed reaction (depolymerization) mechanism of each plastic which has not been covered in detail in the previous review articles. Therefore, this review could be used as a guide for the researchers to rationally design their experiments and to improve new approaches."
} | 3,529 |
31852953 | PMC6920442 | pmc | 8,859 | {
"abstract": "A central issue in ecology is understanding how complex and biodiverse food webs persist in the face of disturbance, and which structural properties affect disturbance propagation among species. However, our comprehension of assemblage mechanisms and disturbance propagation in food webs is limited by the multitude of stressors affecting ecosystems, impairing ecosystem management. By analysing directional food web components connecting species along food chains, we show that increasing species richness and constant feeding linkage density promote the establishment of predictable food web structures, in which the proportion of species co-present in one or more food chains is lower than what would be expected by chance. This reduces the intrinsic vulnerability of real food webs to disturbance propagation in comparison to random webs, and suggests that biodiversity conservation efforts should also increase the potential of ecological communities to buffer top-down and bottom-up disturbance in ecosystems. The food web patterns observed here have not been noticed before, and could also be explored in non-natural networks.",
"introduction": "Introduction The complexity of food webs and the multitude of stressors affecting ecosystems limit our comprehension of how disturbance propagates in ecological communities, hampering biodiversity conservation and management. To cope with such complexity, ecologists have investigated the relationship between the topology of food webs and their stability 1 , seeking to understand whether and how the number of species underlies biodiversity architecture and stability to perturbation 2 . May 3 argued that in randomly assembled multispecies systems, increasing species richness reduced system stability. This prompted several subsequent studies, which established that the arrangement of food webs with respect to the number of species they contain is not random 4 – 6 , and that such non-random structures are more stable to perturbation than what is expected by chance 7 – 10 . Recently, studies have focused on network-level properties of food webs that promote community persistence 9 , 11 , resilience to perturbations 10 and resistance to species loss 12 – 14 . Nevertheless, our understanding of disturbance propagation mechanisms in real ecosystems is still limited. In addition, while food web modelling has fundamentally advanced our understanding of their structure and dynamics, the complexity of the phenomena, the quantity of data involved and the computational skills required have limited the ability to translate such science-based knowledge into practical advice supporting policy formulation 15 , 16 . Such insight is particularly urgent in a scenario of global change, given the evidence suggesting that both anthropogenic pressure and climate change have the potential to disrupt top-down and bottom-up control mechanisms that regulate ecological communities 17 – 21 , highlighting the need for management strategies aimed at preserving stable food webs in ecosystems 22 , 23 . Within food webs, disturbance can propagate along food chains either from a basal resource towards its consumers and their predators (bottom-up pathway) or from a predator to its prey and their resources (top-down pathway). In directional terms, such pathways can be grouped into sink or source sub-webs, which include all the food chains originating from or converging to a single basal or top species respectively 24 . By propagating through feeding links, even small changes in species’ traits and single extinction or invasion events can significantly impact ecological communities 13 , 25 . Nevertheless, despite the importance of top-down and bottom-up controls on the dynamics of populations and ecological processes 19 , 20 , 24 , 26 , the mechanisms determining the distribution of species in sink and source sub-webs, as well as the relationship between species richness and the proportion of species sharing food chains in real food webs, have yet to be fully explored. We investigated the relationships between species richness, food web complexity and the distribution of species into source and sink sub-webs. We also considered a third type, the cross sub-web, which involves disturbance propagating in both directions from an intermediate species (Fig. 1 ). These represent fundamental aspects of biodiversity organisation, and underlie the potential of disturbance to propagate across trophic levels 13 , 24 , 25 , 27 , 28 . Accordingly, we assumed that the higher the proportion of species co-present in at least one food chain within a food web, the higher the potential of disturbance to propagate 13 . The intrinsic vulnerability of food webs to the propagation of disturbance along food chains was thus quantified as the proportion of species (P) included in each source (P B ), cross (P C ) and sink (P T ) sub-web, originating from each basal, intermediate or top species respectively. Here, we focused on detritus-based food webs. The detritus compartment plays a fundamental role in ecosystem structure and functioning 29 , 30 , and climate change is expected to affect detritus inputs and organic matter decomposition rates in ecosystems 31 , 32 . Nevertheless, very little information that might help to understand the structure of these donor-controlled systems is available. Figure 1 Panel (a) Comparison of three simple networks containing the same number of species (nodes) and trophic links (lines). In network 1, all species are part of one food chain, and a disturbance originating from any given species would directly propagate along the food chain to all the remaining ones. In network 2, a bottom-up disturbance spreading from species 1 would affect species 4-5-6 only. In network 3, the modification of a single link with respect to network 2 would mean that only species 4 would be directly affected by a bottom-up disturbance propagating from species 1. Panels (b–d) Food sub-webs exemplifying various propagation pathways along food chains for disturbance starting from a single species (circled): ( b ) bottom-up propagation from a basal resource (red), ( c ) cross propagation from a primary consumer (orange), and ( d ) top-down propagation from a predator (yellow). We compared food webs belonging to both aquatic and terrestrial habitats in the Mediterranean region (Table 1 ), and we quantified their vulnerability to disturbance propagating from species on various trophic levels. We then sought to verify whether the observed P B , P C and P T values were (i) dependent on species richness (S) and web connectance (Cmin) and (ii) predictable using a mechanistic food web model 6 or similar to what would be expected by chance. This was achieved by comparing field data with random food web models, as well as webs generated using the niche model 6 , a simple yet predictive food web model based on observed S and Cmin values and a few foraging rules constraining the probability of feeding interaction between species. The comparison of observed and random food web structures can provide important insight into the organisation of ecological communities. Indeed, the association of food web stability with non-random structures enhances our understanding of the ecological factors constraining the observed patterns and hence the mechanisms underlying the emergence and persistence of complex food webs 7 . Table 1 Food webs representative of various habitats in the Mediterranean region. Locality indicates the region, country and GPS coordinates for each habitat. Habitat Locality S Cmin L/S Lagoon Sardinia (IT) 39°13′N 09°03′E 37 0.075 2.76 Stream, upstream Lazio (IT) 41°51′N 13°00′E 45 0.073 3.30 Stream, downstream Lazio (IT) 41°31′N 13°24′E 32 0.094 3.00 Lake Lazio (IT) 42°05′N 12°12′E 28 0.130 3.65 Corn field Catalunya (SP) 41°38′N 00°35′E 27 0.111 3.00 River, upstream Lazio (IT) 41°58′N 12°30′E 21 0.114 2.39 River, downstream Lazio (IT) 41°49′N 12°25′E 25 0.111 2.78 Beech forest Lazio (IT) 42°19′N, 12°10′E 34 0.106 3.59 S: number of species in the food web; Cmin: food web connectance; L/S: species’ feeding linkage density.",
"discussion": "Discussion The distribution of species across source, cross and sink sub-webs differed between habitats in a predictable way in accordance with food web size (S) and connectance (Cmin). The higher the number of species, the lower the proportion of them co-present in at least one food chain. This indicates that the vulnerability of ecosystems to disturbance propagation across trophic levels falls as their species richness rises, which implies that more biodiverse communities are potentially more able to buffer disturbance. Our observations are also consistent with recent results from studies based on model food webs of similar species richness 13 , where higher connectance intensified the effect (i.e. caused a higher rate of secondary species loss) of single perturbation events (i.e. successful species invasions). As well as species richness, a high number of intermediate species relative to the total reduced the potential of disturbance to propagate along food chains. In addition, our data satisfied expectations from the link-species scaling law, which predicts constant linkage density among species. According to foraging optimisation theories, (i) a low and roughly constant linkage density between species arises from the selection by consumers of the lowest number of food items that maximises their net energy intake 9 , 35 , and (ii) prey species richness promotes predator trophic specialisation, as shown in the food webs analysed here 32 , 36 and elsewhere 9 , 37 . This suggests that energetic constrains on foraging strategies, which operate at the individual level 38 , may give rise to food webs which are intrinsically less vulnerable to the propagation of disturbance than what would be expected if consumers tended to generalise as the number of their potential food sources increased 39 . Notably, the random food web models failed to reproduce real patterns, while niche model-generated webs closely predicted the observed values. Our results thus imply that the distribution of species in source, cross and sink sub-webs significantly deviates from random, and that the patterns observed in real food webs benefit ecological communities by reducing vulnerability to disturbance propagation along food chains compared to what would be expected if distribution was determined by chance. Regardless of habitat type, source sub-webs included a higher proportion of total species than cross and sink sub-webs. This implies that disturbance propagating from a basal resource would have a significantly higher potential to spread throughout the food web than disturbance propagating from a top or an intermediate species. We speculate that the observed patterns may promote stability in detritus-based systems, where the availability of basal resources (i.e. organic detritus and colonising microfungi) is expected to be much more stable over time than invertebrate populations, often characterised by marked spatial-temporal variations and sensitivity to stress 40 , 41 (see Fig. S2 for supporting results). On the other hand, global change is expected to affect the dynamics of the detritus compartment in ecosystems 31 , 32 , 42 , 43 . Thus, our data suggest that there is a strong likelihood of future environmental changes in Mediterranean ecosystems giving rise to bottom-up effects mediated by the structure of detritus-based food webs. The coefficient of variation of Pi values increased with the number of species within each food web, and it was inversely related to food web vulnerability to disturbance. Together with the P B , P C , P T and V values, this makes the variation of species’ Pi values within a food web a useful a-priori indication of its vulnerability. We acknowledge that our observations are based on static food web structures, and that the reorganisation of trophic links and/or changes in linkage strength between species in accordance with their trophic niche plasticity could modify the effects of disturbance 25 . Nevertheless, our analysis of sub-webs enabled us to evaluate the intrinsic vulnerability of real food webs to a range of disturbance propagation pathways (i.e. bottom-up, cross, and top-down). This approach is expected to yield information useful to management strategies based on the risk of ecosystem-specific disturbance propagation, regardless of food web compilation methods. Analysis of the scaling relationships between the size of networks and that of the sub-webs of which they are composed could also be extended to non-natural systems (e.g. financial 44 , transport 45 or disease vector dispersal 46 networks). This would make it possible to verify whether (i) systemic risk increases with network size or (ii) the rules governing such systems mean that an increase in size is accompanied by increased potential to buffer disturbance, as observed in the food webs analysed in this study. In conclusion, we show that increasing species richness and foraging constraints on food web complexity limit the proportion of species that share food chains in ecosystems. This pattern has not been noticed before. It reduces the intrinsic vulnerability of biodiverse ecological communities to disturbance propagation, and promotes the emergence of species-rich food webs where species organise into effective disturbance-buffering structures in accordance with predictable rules. It is also consistent with the predictions of theoretical research, which highlight the positive effect of adaptive trophic behaviour by consumers on the stability of model food web networks 9 . Conversely, the observed results imply that species-poor ecosystems could be highly vulnerable to the propagation of disturbance along food chains, including bottom-up disturbance associated with climate change 31 , 42 , with increased risks for the stability of food webs and the ecosystem services they support 47 – 49 . While this study focused on detritus-based food webs, the observed results may be extended to more complete food webs including primary producer-based food chains. Indeed, detrital and herbivore energy pathways are closely interconnected 8 , and our observations are consistent with webs generated using the niche model, which was developed and tested against whole food web structures in both aquatic and terrestrial ecosystems 6 . Notably, our food web analysis provides mechanism-based evidence indicating that efforts devoted to biodiversity conservation also increase the potential of natural communities to buffer disturbance in ecosystems by maintaining biodiverse, relatively less complex, and ‘safer’ food web structures for their constituent species."
} | 3,694 |
31448892 | PMC6753654 | pmc | 8,860 | {
"abstract": "Cell\nlipid membranes are the site of vital biological processes,\nsuch as motility, trafficking, and sensing, many of which involve\nmechanical forces. Elucidating the interplay between such bioprocesses\nand mechanical forces requires the use of tools that apply and measure\npiconewton-level forces, e.g., optical tweezers. Here, we introduce\nthe combination of optical tweezers with free-standing lipid bilayers,\nwhich are fully accessible on both sides of the membrane. In the vicinity\nof the lipid bilayer, optical trapping would normally be impossible\ndue to optical distortions caused by pockets of the solvent trapped\nwithin the membrane. We solve this by drastically reducing the size\nof these pockets via tuning of the solvent and flow cell material.\nIn the resulting flow cells, lipid nanotubes are straightforwardly\npushed or pulled and reach lengths above half a millimeter. Moreover,\nthe controlled pushing of a lipid nanotube with an optically trapped\nbead provides an accurate and direct measurement of important mechanical\nproperties. In particular, we measure the membrane tension of a free-standing\nmembrane composed of a mixture of dioleoylphosphatidylcholine (DOPC) and dipalmitoylphosphatidylcholine\n(DPPC) to be 4.6 × 10 –6 N/m. We demonstrate\nthe potential of the platform for biophysical studies by inserting\nthe cell-penetrating trans-activator of transcription (TAT) peptide\nin the lipid membrane. The interactions between the TAT peptide and\nthe membrane are found to decrease the value of the membrane tension\nto 2.1 × 10 –6 N/m. This method is also fully\ncompatible with electrophysiological measurements and presents new\npossibilities for the study of membrane mechanics and the creation\nof artificial lipid tube networks of great importance in intra- and\nintercellular communication.",
"conclusion": "Conclusions In summary, we introduce a microfluidic-based\nplatform to interface\nfree-standing membranes with optical tweezers for nanomanipulation,\nnanotube formation, and electrophysiological measurements. We use\nour tool to directly measure the membrane tension without assuming\nany values for the bending rigidity or nanotube radius. Moreover,\nour approach offers control over the solutions on both the outside\nand inside of a nanotube, where positive and negative membrane curvatures\noccur respectively. These are physiologically relevant membrane topologies. 50 − 52 For these reasons, our approach extends the range of tools available\nto quantify forces in cell biomechanical processes, 53 for instance, to study the mechanosensitivity associated\nwith cell motility, auditory, and tactile functions. It also opens\nup new possibilities for the creation and the dynamical study of artificial\nlipid tube networks mimicking biological structures, i.e., lipid tubes\npart of cell organelles 54 and lipid tubes\nthat extend from cells for communication. 55",
"introduction": "Introduction Mechanical forces at the cell membrane\nplay an important role in\nmany vital biological processes, such as endo- and exocytosis, 1 − 3 inter- and intracellular communication, 4 cell division, 5 and cell spreading. 6 , 7 A large number of these cellular processes depend on unequal conditions\non each side of the membrane (e.g., proton-motive force-dependent\nprocesses). Direct measurements of these forces represent a major\nexperimental challenge as they require the integration of force measurement\ntechniques, such as optical tweezers, with lipid bilayers while allowing\nthe independent control and measurement of physiological conditions,\nelectric potential or pH, on both sides of the cell membrane. Previous approaches have used artificial membranes, which mimic\ncell membranes in vitro and offer more control over physicochemical\nconditions than in vivo systems. Such approaches include supported\nlipid bilayers, black lipid membranes, and lipid vesicles. 8 − 10 The combination of optical tweezers with supported lipid bilayers\nor giant unilamellar vesicles (GUVs) has contributed to our biophysical\nunderstanding of lipid nanotube formation, 11 , 12 the influence of protein crowding on membrane nanotube mechanics, 13 and the role of proteins involved in membrane\nfission 14 and fusion. 15 However, these approaches are limited because they do not\nallow equal access and control over the conditions on both sides of\nthe membrane. Here, we present a design for an experimental\nplatform ideally\nsuited to the study of biological membrane processes. A free-standing\nmembrane is formed between the two microchannels of a flow cell. Our\ndevice integrates optical tweezers with a flow cell that provides\naccess to both leaflets of the membrane independently, thereby affording\nindependent and dynamic control over physiological conditions on each\nside of the membrane. The flow cell supports electrophysiology measurements,\nwhich we demonstrate by monitoring the capacitance of the membrane\nin real time. While several approaches to form free-standing membranes\nin microdevices are reported, 16 − 19 they have all been hindered by the presence of an\nannulus 20 of solvent generally trapped\nwithin the lipid membrane, which is responsible for severe optical\naberrations 21 that prevent optical trapping\nclose to the membrane. In contrast, we here show that the optical\ntweezers in our device can trap beads and accurately measure forces\narbitrarily close to and on both sides of the membrane. We achieve\nthis by reducing the presence of organic solvent between the two leaflets.\nThe robustness of our approach and its ability to measure forces on\nboth sides of the free-standing membrane are demonstrated by pushing\noptically trapped microspheres through the free-standing lipid bilayers\nto quantify the membrane tension and form lipid membrane nanotubes,\na biologically relevant structure. This microfluidic platform is ideal\nfor biophysical studies of biomolecules interacting with membranes.\nTo demonstrate this, the cell-penetrating HIV-1 trans-activator of\ntranscription (TAT) peptide 22 is introduced\ninto the microchannel to be inserted into the membrane. We find the\npresence of TAT reduces the membrane tension.",
"discussion": "Results and Discussion Interfacing\nFree-Standing Lipid Bilayers with Optical Tweezers The free-standing\nlipid membranes are formed inside a microfluidic\ndevice consisting of two parallel microchannels connected by one or\nseveral rectangular apertures of 100 μm × 85 μm ( Figures 1 and S1 ). The lipid membranes are formed by the contact\nof two lipid monolayers at the water–solvent interface over\nthe apertures connecting the two channels ( Figure 1 b,c). Membranes formed in these devices,\nas opposed to GUVs, have both sides of the membrane readily accessible.\nThe polymer chosen for the fabrication of the device is the photopolymerized\nthiol-ene resin Norland Optical Adhesive 81 (NOA81) that allows the\nformation of rigid and transparent microdevices, compatible with optical\ntechniques. 23 − 25 NOA81 is impermeable to air and water vapor, 24 thus avoiding evaporation and being a favorable\ncandidate for the formation and long-term stability of lipid membranes.\nNOA81 is generally described to be compatible with organic solvents,\nexcept for chlorinated solvents like chloroform, showing in some cases\na swelling of ∼30%. 23 , 26 , 27 This susceptibility for chloroform makes it a good candidate for\nthe preparation of membranes with a smaller annulus. Figure 1 Design of the microdevice\ncombining free-standing membranes with\noptical tweezers. (a) Picture of a representative microfluidic device\nused for mechanical measurements. Free-standing lipid bilayers are\nformed over the apertures connecting the two microchannels. The white\nsquare indicates the position of one of the apertures. (b, c) Pictures\nof the process of membrane formation (b) before the organic solvent\nreaches the aperture and (c) after membrane formation (A, air; O,\norganic phase; and W, aqueous phase; white arrows indicate the direction\nof the flow). To determine whether chloroform\ncan indeed reduce annulus size,\ntwo different approaches using different organic solvents are followed.\nIn both approaches, 1,2-dioleoyl- sn -glycero-3-phosphocholine/1,2-dipalmitoyl- sn -glycero-3-phosphocholine (DOPC/DPPC) (2:1 molar ratio)\nis used as lipid component. In the first approach, lipid membranes\nare formed by subsequently flowing a mixture of decane/chloroform/methanol\n(7:2:1 v/v) and an aqueous solution containing lipids. In the second\napproach, membranes are prepared using the same lipid composition\nin chloroform followed by the aqueous solution. As shown in Figure 2 a, the membranes\nformed with the solvent mixture decane/chloroform/methanol exhibit\na thickened appearance at the edges, which corresponds to the annulus\nthat is easily observable with bright-field and fluorescence microscopy.\nHowever, in membranes formed with chloroform ( Figure 2 b), no apparent annulus is observed. In fact,\nthe accumulated chloroform residues are directly observed to shrink\nat the contour of the microstructures of the device, in agreement\nwith the permeation of chloroform in NOA81 ( Figure S2 ). The reduction of the annulus due to material permeability\nis in accordance with previous observations in polydimethylsiloxane\n(PDMS) and the known susceptibility of NOA81 to chloroform. 23 , 28 Figure 2 Effect\nof the lipid membrane annulus on optical imaging and optical\ntrap stiffness. (a, b) Bright-field (left) and confocal fluorescence\noptical microscopy (right) images of lipid membranes formed using\n(a) a mixture of decane/chloroform/methanol as an organic solvent\nand (b) only chloroform as an organic solvent. (c, d) Pictures of\na trapped bead near a membrane prepared using (c) decane/chloroform/methanol\nmixture and (d) chloroform. The distances between the bead and the\nmembrane are indicated at the bottom of the pictures. (e) Optical\ntrap stiffness in the x -axis, perpendicular to the\nmembrane plane, as a function of the trap position with respect to\nthe membrane. Trap stiffness measurements are all done with 1 μm\nbeads and a laser power of 1.3 W (measured before entering the microscope\nobjective) near the membranes formed using decane/chloroform/methanol\n(red circles) and chloroform (blue squares). The position represents\nthe distance between the trap center and the membrane. The bars represent\nthe standard deviation between measurements. Samples prepared following the two different approaches described\npreviously, with a large or reduced annulus, are studied in combination\nwith optical tweezers. Membranes prepared with decane/chloroform/methanol\nshow clear optical aberrations as the trapped bead is brought toward\nthe membranes to a point where the bead cannot be trapped anymore\n( Figure 2 c). We find\nthat optical trapping is hindered in the vicinity of these membranes.\nA decay in optical trapping stiffness is measured when the trapped\nbead is brought close to the membrane, as depicted by the red circles\nin Figure 2 e. For bead–membrane\ndistances above 50 μm, stiffness values remain relatively constant\n(∼0.7 pN/nm), while at a distance shorter than 18 μm\nfrom the lipid bilayer, it is not possible to successfully trap a\nparticle ( Figure 2 e). On the other hand, due to the negligible size of the annulus when\nmembranes are formed using chloroform, the trap stiffness remains\nalmost unaffected at distances of 1–200 μm from the lipid\nbilayer ( Figure 2 e,\nblue squares) and the optical appearance of the beads remains unchanged\n( Figure 2 d). This behavior\nis independent of the laser power used ( Figure S3 ) and is attributed to the reduction in the optical aberrations\ncaused by the annulus. 21 These observations\nconfirm the importance of the solvent accumulated within the membrane\nand the improvement of the trapping stiffness in the vicinity of the\nmembranes with a smaller annulus. Capacitance Measurements\nduring Membrane Formation Our microfluidic approach enables\nstraightforward electrophysiology\nmeasurements by simply adding electrodes in the microdevice. In this\nway, we investigate the membrane’s electrical capacitance,\nwhich informs us about membrane formation kinetics and about whether\norganic solvent remains within the bilayer. 17 We find that lipid membranes with the small annulus form within\na few seconds and reach a steady capacitance value 16 ± 10 s\nafter initial contact of the lipid monolayers. The average steady\ncapacitance value is C M = 49.2 ±\n2.4 pF. The membrane is estimated to cover the full cross section\nof the gap (8500 μm 2 ) because the size of the annulus\nis negligible compared to the membrane surface area. Using the gap\ncross-sectional area for the membrane surface area results in a specific\ncapacitance of 0.6 μF/cm 2 . This value is in accordance\nwith the specific capacitance reported for phospholipid bilayers composed\nof a mixture of DOPC and dioleoylphosphatidylethanolamine (DOPE). 29 A specific capacitance of 0.6 μF/cm 2 is also predicted for DPPC bilayers considering the measured\ndielectric constant (ε r = 3.2) and thickness (5 nm). 30 Substantial amounts of chloroform within the\nbilayer would result in a lower specific capacitance; therefore, there\nis no significant amount of solvent trapped within the leaflets. Lipid Nanotube Formation The combination of the free-standing\nmembranes with optical tweezers enables nanomanipulation of the lipid\nbilayers to form nanotubes. Membrane nanotubes are a ubiquitous structure\nfound in cells and used for inter- and intracellular exchange and\ntransport. 31 − 34 They are also found in different cellular organelles, such as the\nendoplasmic reticulum, 35 , 36 mitochondria, 37 and Golgi apparatus. 38 In the\ncell, lipid nanotubes are thought to be formed by spontaneous curvature 39 but also by the application of force from molecular\nmotors and the cytoskeleton. 40 In vitro,\nthey are constructed in many studies via direct micromanipulation\nusing optical tweezers interfaced with a GUV. 11 , 14 , 41 , 42 Here,\nlipid nanotubes are formed by two different ways: (1) by pulling a\nmembrane containing biotinylated lipids with a trapped streptavidin-coated\nmicrobead ( Figure 3 a) and (2) by pushing a trapped microbead across the membrane ( Figure 3 b). For both pulling\nand pushing experiments, trapped beads are displaced at 1 μm/s\nfrom or toward the membrane, respectively. The pulling approach, which\nis more conventional, 11 , 14 , 43 requires the addition of biotinylated lipids to the original lipid\nmixture of DOPC/DPPC (2:1) and the use of a streptavidin-coated microbead.\nIn the conditions tested, this approach requires several contacts\nbetween the free-standing membrane and the bead for successful bead\nattachment via biotin–streptavidin bond creation. In contrast,\nthe pushing approach results in nanotube formation in all attempts.\nIn that case, the bead is wrapped by the membrane without the use\nof functionalization. With this approach, networks of lipid nanotubes\nwith increasing complexity can be created through the use of multiple\noptical traps. To demonstrate this capability, we form two neighboring\ntubes by two optical traps and the coalescence of the tubes is observed\nin real time ( Figure 3 c). Tubes pushed from these free-standing lipid bilayers are as long\nas 550 μm ( Video S1 ), limited by\nthe width of the channels in the microdevice, suggesting that longer\nnanotubes may be achievable in wider channels. Figure 3 Lipid tube formation.\n(a) Bright-field images of a lipid tube formed\nby pulling a patch of membrane with an optically-trapped bead. The\nbead is first moved toward the membrane and then pulled away, as shown\nby the blue arrows. (b) Bright-field images of a lipid tube formed\nby pushing a bead against a free-standing lipid bilayer. (c) Bright-field\nimages of two separate lipid tubes held by two optical traps. From\ntop to bottom, the traps are brought closer to one another, as shown\nwith blue arrows, until the two tubes contact and coalesce. (d) Six\nrepresentative force–displacement curves obtained when pushing\na 2 μm bead against the same lipid membrane. Membrane Tension Measurements Figure 3 d shows typical force–displacement\ncurves for a bead pushed against the free-standing bilayers. For convenience,\nwe split the pushing process into two chronological segments: deforming\nthe free-standing membrane and extending the nanotube. As shown in Figure 3 d, the force increases\nmonotonically with displacement during the initial phase of deformation\nuntil reaching the maximum, or overshoot, force. Then, a sharp transition\noccurs when the nanotube is formed, after which the force remains\nconstant while the tube is extended. This behavior is qualitatively\nsimilar to observations reported previously for tubes pulled from\na GUV, 11 where the force also increases\nuntil a sharp drop in force is observed when the tube is formed. However,\nfor the pushing approach, the forces needed to create a tube are not\ndefined by the patch of contact between biotinylated membrane and\nbead, as is the case when pulling a tube. 11 As a result, in the pulling experiments, it is not possible to directly\nextract the membrane tension, bending rigidity, and tube radius from\nthe force–displacement curves alone, as energy conservation\nof the tube-pulling process only provides two equations for the three\nunknowns. Therefore, the pulling approach would require the use of\nadditional sensors, such as micropipettes, or the assumption of one\nof the unknown values, such as the bending rigidity. 11 On the contrary, the pushing approach allows for a straightforward\ndetermination of the mechanical properties of the membrane since the\nprocess is independent of bond formation between the bead and membrane. We hypothesize that for pushing the maximum force, or overshoot\nforce, would depend on the radius of the bead, which is invariable\nduring an experiment, unlike the patch area for pulling experiments. Figure 4 a shows the force–displacement\ncurves obtained for three different bead sizes. As shown in Figure 4 b, the maximum force\nindeed increases proportionally with the bead diameter, while the\nforce required for tube extension remains constant ( Figure 4 c) and is independent of bead\nsize. Since the maximum force exhibits a linear relationship with\nbead size ( Figure 4 b), we expect that the relevant mechanical property resisting membrane\ndeformation before nanotube formation is only tension. If bending\nrigidity contributions were significant, they would result in a nonlinear\nrelationship between the maximum force and particle diameter. Figure 4 Force measurements\nwhen pushing beads of various sizes against\na DOPC/DPPC lipid bilayer. (a) Force–displacement curves for\ntubes formed by pushing beads of 1, 2, and 5 μm diameters ( N = 10, 15, and 14 curves, respectively), with representative\ncurves shown in red, blue, and yellow, respectively, and all other\ncurves shown in gray. (b) Maximum force and (c) tube extension force\nas a function of bead diameter. To model the mechanics during the pushing approach, we first consider\nthe free-standing membrane deformation. This process is assumed to\nbe quasi-steady since pushing speeds ranging from 0.05 to 1.0 μm/s\nresult in overlapping force–displacement curves ( Figure S4 ). Therefore, a force balance is conducted\non the bead ( Figure 5 a). The two forces acting on the bead at any given time are the force F from the optical tweezers and an opposing force F σ due to the membrane tension, which is\ndependent on the angle θ of the membrane at a radial distance\nδ from the center of the bead. An expression for this force\nis given by F σ = 2πδσ cos θ,\nwhere σ is the membrane tension. The two geometrical parameters\n(θ and δ) are measured from videos taken during the force\nmeasurements. By balancing the forces, the surface tension can be\nexpressed as . As shown in Figure 5 b, we find that the surface tension, σ,\nis independent of bead size, with an average value and standard deviation\nof 4.63 ± 0.74 × 10 –6 N/m. This value\nagrees with those obtained previously using optical methods to measure\nthe thermal fluctuations of free-standing bilayers. 44 Figure 5 Membrane properties are extracted from force curves. (a) Representative\nimage from video recordings used to measure the angle θ of the\nmembrane at a radial distance δ from the center of the bead.\nThe shown force balance is used to measure the membrane tension. (b)\nMembrane tension, (c) bending rigidity, and (d) radius of lipid nanotubes\nare not statistically different for bead diameters 1, 2, and 5 μm\n(Kruskal–Wallis one-way analysis of variance, p > 0.05). From the obtained surface tension,\nthe bending rigidity, κ,\nand tube radius, R t , can be obtained using\nthe force associated with nanotube extension. 45 , 46 The free energy of the tube extension is , where R t and L t are the tube radius and length, respectively,\nand F t is the tube extension force. 47 As the energy must remain constant at equilibrium,\nthe bending rigidity, κ, and surface tension, σ, are related\nas follows 1 With eq 1 and , we find\na membrane rigidity of 3.11 ±\n0.56 × 10 –20 J and a tube radius of 58.8 ±\n10.6 nm ( Figure 5 c,d).\nThe bending rigidity values obtained are within the range of those\npreviously reported. 48 This microfluidic\nplatform enables studies of biomolecule–membrane\ninteractions. To demonstrate this, an aqueous solution of the cell-penetrating\nHIV-1 trans-activator of transcription (TAT) peptide is injected into\nthe microchannel to interact with the membrane. The TAT peptide is\nan arginine-rich peptide that has been shown to interact with lipid\nbilayers 49 and to carry cargo across cell\nmembranes. 22 To characterize the interactions\nbetween the TAT peptide and membranes, we measured the tension of\nDOPC/DPPC membranes in the presence of TAT peptide, by pushing 2 μm\nbeads with the optical tweezers ( Figure 6 ). We find that the TAT peptide lowers the\nmembrane tension to an average value and standard deviation of 2.08\n± 0.16 × 10 –6 N/m. Figure 6 Effect of TAT peptide\non membrane properties. (a) Force–displacement\ncurves for tubes formed by pushing beads of 2 μm diameter against\na DOPC/DPPC lipid bilayer without (gray) and with (green) TAT peptides, N = 15 and 16 curves, respectively. (b) Membrane tension\nextracted from the force–displacement curves."
} | 5,589 |
34947029 | PMC8708846 | pmc | 8,861 | {
"abstract": "The establishment of lignocellulosic biorefineries is dependent on microorganisms being able to cope with the stressful conditions resulting from the release of inhibitory compounds during biomass processing. The yeast Kluyveromyces marxianus has been explored as an alternative microbial factory due to its thermotolerance and ability to natively metabolize xylose. The lignocellulose-derived inhibitors furfural and 5-hydroxymethylfurfural (HMF) are considered promising building-block platforms that can be converted into a wide variety of high-value derivatives. Here, several K. marxianus strains, isolated from cocoa fermentation, were evaluated for xylose consumption and tolerance towards acetic acid, furfural, and HMF. The potential of this yeast to reduce furfural and HMF at high inhibitory loads was disclosed and characterized. Our results associated HMF reduction with NADPH while furfural-reducing activity was higher with NADH. In addition, furans’ inhibitory effect was higher when combined with xylose consumption. The furan derivatives produced by K. marxianus in different conditions were identified. Furthermore, one selected isolate was efficiently used as a whole-cell biocatalyst to convert furfural and HMF into their derivatives, furfuryl alcohol and 2,5-bis(hydroxymethyl)furan (BHMF), with high yields and productivities. These results validate K. marxianus as a promising microbial platform in lignocellulosic biorefineries.",
"conclusion": "5. Conclusions In conclusion, our results evidence the higher potential for the application of industrial isolates in lignocellulosic biorefineries compared to laboratory strains, as a result of their higher tolerance to lignocellulosic-derived inhibitors. To the extent of our knowledge, this is the first study exploiting K. marxianus as a whole-cell biocatalyst to produce furfuryl alcohol and BHMF. Furthermore, as far as we know, the furfuryl alcohol productivities presented in our work using glucose (5.74 g/L/h) or xylose as a co-substrate (6.46 g/L/h) are the highest reported for yeast, and the 99.65% BHMF yield attained is the highest reported in the literature. Moreover, given the high cell viability at the end of the bioconversion assays, further experiments with cell recycling in a fed-batch scheme and using HMF or furfural-enriched medium obtained from renewable carbohydrates (e.g., lignocellulosic biomass) should be performed to establish a sustainable process to produce furfuryl alcohol and BHMF by K. marxianus . On the other hand, the production of these high-value compounds will greatly impact the economic feasibility of lignocellulosic biorefineries.",
"introduction": "1. Introduction The yeast Kluyveromyces marxianus has been emerging as an alternative cell factory to produce ethanol, high-value chemicals, and enzymes with a wide range of applications in food, feed, and pharmaceutical industries [ 1 ]. Due to its Qualified Presumption of Safely (QPS) and Generally Regarded as Safe (GRAS) status, the features of K. marxianus have rendered it as an attractive chassis for different industrial applications. These features include: a high growth rate among other eukaryotes [ 2 ], thermotolerance (the ability to grow at temperatures up to 52 °C), tolerance to low pH [ 3 ], the ability to metabolize a broad range of sugar substrates (glucose, xylose, lactose, fructose, arabinose, galactose, among others), and the ability to produce lytic enzymes [ 4 ]. Accordingly, K. marxianus application in a biorefinery context is gaining special interest due to its ability to ferment various low-cost feedstocks, such as cheese whey, fruit peels, and the sugars derived from lignocellulosic biomass. Lignocellulosic-based biorefineries have been developed and are suggested to be environmentally sustainable alternatives to replace the use of fossil fuels to produce biofuels and value-added chemicals. One of the challenges associated with the implementation of these biorefineries relies on the inhibitory compounds present in the hydrolysates after the pretreatment or hydrolysis steps of the biomass. These compounds include acetic acid, furfural, 5-hydroxymethylfurfural (HMF), 4-hydroxybenzaldehyde, syringaldehyde, catechol, and vanillin, that cause the inhibition of cell growth and microbial fermentation [ 5 ]. Even though there has been an increased usage of K. marxianus in lignocellulose-to-ethanol processes, the knowledge of its stress physiology to multiple inhibitor resistance in lignocellulosic hydrolysates is still poor compared to Saccharomyces cerevisiae . Regarding the aldehyde compounds, K. marxianus , as S. cerevisiae , is known to reduce them into their corresponding alcohols to reduce their toxicity [ 6 ], being capable of producing ethanol in the presence of 10 mM of HMF or furfural [ 7 ]. Furthermore, K. marxianus was identified as an HMF tolerant yeast among several non-conventional yeasts [ 8 ]. In fact, a K. marxianus strain isolated from a Mezcal fermentation process was found to have a higher tolerance to HMF and furfural than the commercial S. cerevisiae Ethanol Red strain [ 9 ]. Despite being generally considered as undesirable inhibitors in lignocellulosic processes, HMF and furfural (obtained from dehydration of glucose and xylose, respectively) were identified as the top promising compounds to be obtained from biomass to reach economically viable biorefineries [ 10 ]. These furans present a versatile composition—an aromatic furan ring and reactive functional groups (aldehyde group in furfural; aldehyde and alcohol groups in HMF), which makes them promising building-block platforms that can be converted into a wide variety of compounds with applications in diverse areas, such as plastic, pharmaceutical, and textile industries [ 11 , 12 ]. Up until now, the production of these higher-value derivatives has been mainly based on chemical catalysis, with drawbacks such as harsh reaction conditions, expensive catalysts, and low selectivity [ 13 ]. More recently, biocatalysis has appeared as a more environmentally friendly alternative, with the use of whole-cell biocatalysts presenting advantages over the use of purified enzymes, e.g., the easiness of catalyst recycling and the regeneration of cofactors [ 14 ]. Following that, some microorganisms have been reported to be used in the bioconversion of HMF and furfural into their higher value derivatives, with the production of their corresponding alcohols, 2,5-bis(hydroxymethyl)furan (BHMF) and furfuryl alcohol, receiving great attention [ 11 , 12 , 15 ]. Furfuryl alcohol has been used as a precursor mainly to produce resins for the foundry industry, but also as an intermediate to produce diuretic furosemide, ranitidine, an antiulcer drug, and chloroquine, used in the treatment of malaria. Moreover, furfuryl alcohol has applications in the wood industry to produce impregnating agents, resins, or adhesives [ 16 ]. In turn, BHMF has been studied for its application in the synthesis of polymers, resins, and ethers for the replacement of phthalate plasticizers or the production of polyesters [ 17 , 18 , 19 ]. Considering this, and the necessity to reach economically viable biorefineries through the production of both bioethanol and high-value compounds [ 20 , 21 ], in this work, we aimed to (1) evaluate the profile of industrial K. marxianus isolated from cocoa fermentation processes in terms of xylose consumption and tolerance towards lignocellulosic-derived inhibitors, and (2) explore the more stress-tolerant K. marxianus isolate as a whole-cell biocatalyst to produce high-value HMF- and furfural-derivatives (BHMF and furfuryl alcohol).",
"discussion": "4. Discussion During the pretreatment and hydrolysis steps of the lignocellulosic biomass, several inhibitors are released, such as furfural, HMF, and acetic acid, which can be toxic and hinder microbial fermentation [ 5 ]. As such, the search for robust microorganisms able to cope with these inhibitory compounds is of utmost importance to achieve efficient fermentations. The non-conventional yeast K. marxianus has emerged as a promising biofactory, currently being mainly explored for ethanol production [ 25 ]. Xylose assimilation in K. marxianus is accomplished by the xylose reductase/xylitol dehydrogenase (XR/XDH) pathway, where xylitol is the first intermediate, and its accumulation is caused by the cofactor imbalance between XR and XDH enzymes [ 26 , 27 ]. In this work, we explored the growth profile of eight K. marxianus strains isolated from cocoa fermentation and the laboratory strain CBS6556 at 37 °C in a xylose-rich medium. Since strains S8, S9, and S11 presented similar growth rates in YPX medium, their stress tolerance to lignocellulosic-derived inhibitors was assessed in the same medium with 4.8 mM of HMF, 7.3 mM of furfural, and 2.8 g/L of acetic acid at pH 5.0. It is worth noting that the pH of the medium in this experiment was adjusted to 5.0 since it is reported that at pH 4.0 or 4.5 K. marxianus growth is retarded when acetate is added to the medium [ 28 ]. During the lag phase, furfural and HMF were being converted to less toxic derivatives, allowing the exponential growth to start after that period. Strains isolated from cocoa fermentations showed higher tolerance to the inhibitors tested than the laboratory strain. Moreover, the low concentration of ethanol achieved could be a result of the strains only being able to consume half of the acetic acid introduced in the medium, as previously observed by Nitiyon et al. (2016) [ 7 ]. Gathering these results, it is evident that K. marxianus strain variability is reflected in their ethanol and xylitol production and their stress tolerance, as previously reported by Nitiyon et al. (2016) [ 7 ] for K. marxianus BUNL-21 and DMKU3- 1042, and by Wilkins et al. (2008) for K. marxianus strains IMB2, IMB4, and IMB5 [ 29 ], as examples. Recently, Wang et al. (2018) performed a transcriptomic analysis by RNA-seq after K. marxianus growth at 42 °C in the presence or absence of a mixture of lignocellulosic inhibitors similar to the ones used in our study (3 g/L acetic acid, 0.7 g/L furfural, 0.7 g/L HMF, and 0.28 g/L phenols). The authors showed that most of the differentially expressed genes from glycolysis, gluconeogenesis, pyruvate metabolism, and NADPH metabolism, among other pathways, were down-regulated, while genes involved in the TCA and pathways involved in stress response were up-regulated (such as alcohol dehydrogenases encoding isoform genes, ADH3 , ADH4 , and ADH6 ). This suggests that K. marxianus boosted up the mechanisms for the detoxification of these inhibitors [ 30 ]. Considering that strains S8 and S9 seemed to present a similar detoxification capacity, strain S9’s ability to cope with a higher concentration of HMF and furfural was tested in YPX medium and compared with the laboratory strain CBS6556. As a result, strain CBS6556 did not grow under any condition, while strain S9 only grew in the condition with furfural alone. Moreover, in the xylose-rich medium, the presence of 10.4 mM of furfural is known to increase acetic acid yield from 0.09 g/g xylose to 0.13 g/g xylose compared to a control without furfural [ 31 ]. In our work, strain S9 was able to achieve a higher yield of approximately 0.47 g/g xylose. One of the possible explanations for acetic acid accumulation in YPX is the need for NADPH as a cofactor for xylose reductase, which can be supported by the up-regulation of the ALD4 gene coding for acetaldehyde dehydrogenase on YPX [ 7 ]. We hypothesized that the growth inhibition in YPX was caused by NADPH cofactor competition between xylose reductase, the first enzyme of the oxidoreductive pathway in K. marxianus , and the oxidoreductases involved in HMF detoxification [ 5 , 27 ]. To evaluate the effect of carbon sources on the detoxification of HMF, we tested the same inhibitory conditions in YPD medium. As a result, the growth of both strains was not inhibited, and strain S9 proved to be able to detoxify HMF and furfural more readily than the laboratory strain CBS6556. Temperature and ethanol tolerance are attributes that cocoa fermentation strains may display, given that the temperature can reach 40 °C and ethanol can be present at 10–12% under the operation conditions [ 32 ]. The strain S9 proved to be able to tolerate 37 °C and was capable of detoxifying a high concentration of inhibitory compounds (furfural and HMF) more readily than CBS6556, revealing its higher potential for application in fermentations with real hydrolysates. S. cerevisiae strains isolated from industrial distilleries also showed superior ethanol production than laboratory strains [ 33 ] and the importance of selecting the best chassis for the target application has been clearly demonstrated [ 23 ]. Under glucose growth, acetic acid is produced from acetaldehyde by aldehyde dehydrogenases to regenerate NADH in the cytoplasm [ 9 ]. As furfural detoxification requires NADH, this explains the higher production of acetic acid in the presence of furfural. Glycerol production by both S9 and CBS6556 strains was higher in the condition with 47.6 mM of HMF than in the presence of 20.8 mM of furfural. In fact, in the presence of 20.8 mM of furfural, a low concentration of glycerol was detected in the case of strain S9, and it was not detected in CBS6556 fermentation under this condition. Reports in the literature suggest that furfural detoxification substitutes the formation of glycerol to reoxidize the excess NADH during fermentation and to maintain the intracellular redox balance [ 34 ]. Furthermore, an HMF and furfural reducing activity assay corroborated the hypothesis of NADPH cofactor imbalance in YPX medium since NADPH was confirmed to be the preferred cofactor for HMF reduction and NADH for furfural reduction in both strains. Moreover, this assay reinforced that strain S9, isolated from an industrial environment, is more robust than the laboratory strain CBS6556 given its higher furfural and HMF-reducing activity. In K. marxianus , NADH regeneration occurs in the second step of the XR/XDH pathway [ 26 ], and NADH was proved in our study to be the preferred cofactor for furfural reduction, which explains why only furfural detoxification was possible in the xylose-rich medium. In S. cerevisiae , glucose assimilation via the pentose phosphate pathway (PPP) and through the action of glucose-6-phosphate dehydrogenase regenerates NADPH [ 35 ]. The same was demonstrated in Kluyveromyces lactis , a close yeast to K. marxianus [ 36 , 37 ]. As such, assuming that K. marxianus uses the same mechanism, this explains why strains S9 and CBS6556 were able to grow in YPD medium and detoxify HMF. Four alcohol dehydrogenases (KmAdh) were found in the K. marxianus genome [ 38 ], and later its characterization revealed that these enzymes prefer NAD + /NADH as cofactors over NADP + /NADPH. Specifically, KmAdh1 and KmAdh2 could efficiently reduce furfural using NADH as a cofactor, which corroborates our observations in the furfural reducing activity assay [ 39 ]. In another work, a broad specific NADPH-dependent aldehyde reductase, KmGRE2, with a 46% similar identity to the S. cerevisiae S288c GRE2 was identified in the K. marxianus strain DMB1 and showed 60% more relative activity towards furfural than HMF. The results presented in that study reveal that KmGRE2 could be involved in furfural detoxification [ 40 ]. We verified that furfural and HMF were mainly converted by K. marxianus strains S9 and CBS65556 to their corresponding alcohols (furfuryl alcohol and BHMF, respectively), and strain S9 was faster than CBS6556 in the production of both alcohols. The low concentrations detected of the corresponding carboxylic acids, HMFCA and furoic acid, were expected since it is reported that under anaerobic conditions, HMF and furfural are mainly converted to their corresponding alcohol while under aerobic conditions, furfural is known to be converted to furoic acid [ 6 , 41 ]. Oliva et al. (2004) evaluated the inhibitory effect of furfural on the growth and fermentation of K. marxianus CECT 10875 and further detected the production of furfuryl alcohol at 42 °C in the presence of 30 g/L glucose. Using an inoculum size of 4% ( v / v ), the authors showed that 2 g/L (21 mM) of furfural, a concentration that inhibited 25% of the strain growth at 24 h, was completely reduced to furfuryl alcohol after 8 h [ 6 ]. In our work, K. marxianus strain S9 was able to produce 21.09 mM of furfuryl alcohol in the presence of 20 g/L of glucose after 12 h using an inoculum at OD 600 nm 0.1, which is lower than the inoculum used by Oliva et al. (2004). Finally, considering the demonstrated potential of K. marxianus strain S9 to produce furan alcohol derivatives, we performed a whole-cell bioconversion assay under anaerobic conditions and with a higher inoculum to favor the production of the high-value compounds furfuryl alcohol and BHMF. Our results indicate that future experiments for the biocatalytic synthesis of furfuryl alcohol from the xylose-rich liquid fraction of lignocellulosic hydrolysates could be efficient due to the higher conversion, yield, and productivity observed in the presence of xylose (vs. glucose). It also shows the major relevance of the redox balance between xylose consumption and furfural reduction for furfuryl alcohol production. Despite slow xylose consumption during the bioconversion assay, the pre-inoculum performed in YPX probably created an NADH pool necessary for furfural conversion into furfuryl alcohol. When pulsing furfural during xylose consumption, the redox state and energy metabolism of S. cerevisiae cells was found to be more severely affected than during glucose consumption. Therefore, it can be assumed that in K. marxianus, furfural addition could cause similar effects, which could explain why cell viability was lower in the assays with xylose than with glucose [ 42 ]. There are some reports in the literature describing the production of BHMF by whole-cell biocatalysts. To the extent of our knowledge, the BHMF yield reported in our work (99.65%) is the highest reported for yeast [ 43 , 44 , 45 ], bacteria [ 46 , 47 ], and fungi [ 48 , 49 ] in batch and fed-batch modes using a synthetic medium with glucose as co-substrate. Specifically, in yeast, Li et al. (2016) reported BHMF production at 35 °C by Meyerozyma guilliermondii SC1103 from 100 mM of HMF with an inoculum of 20 g/L of wet weight using glucose as a co-substrate and obtained a yield of 86% [ 43 ]. In another work, an S. cerevisiae strain harboring an aryl alcohol dehydrogenase from M. guilliermondii was reported to produce BHMF with a yield of 94% from 250 mM of HMF with an inoculum of 60 g/L of wet weight using glucose as co-substrate [ 44 ]. Furfuryl alcohol production from furfural has also been previously described. To the extent of our knowledge, the furfuryl alcohol productivities presented in our work using glucose as a co-substrate (2.89 g/L/h from 33.5 mM of furfural, and 5.74 g/L/h from 66 mM of furfural) were the highest among the reported for yeast [ 50 , 51 , 52 ] and bacteria [ 53 , 54 , 55 , 56 , 57 ] in a batch using a synthetic medium. Moreover, the productivities obtained in our work using xylose as co-substrate (3.18 g/L/h from 33.5 mM of furfural, and 6.46 g/L/h from 66 mM of furfural) were also the highest among the reported for yeast [ 58 , 59 ] in a batch using a synthetic medium. For example, Mandalika et al. (2014) reported the production of furfuryl alcohol at 30 °C by S. cerevisiae UWOP587-2421 from 25 g/L of furfural with an inoculum of 10 g/L of wet weight using glucose as a co-substrate and achieved a productivity of 0.96 g/L/h [ 50 ]. Recently, Kılmanoğlu et al. (2021) intended to optimize the pretreatment and enzymatic hydrolysis conditions of tomato pomace, ultimately aiming to produce alcohols and esters by K. marxianus at 28 °C and using a 5% seed culture. The authors were able to produce 0.28 g/L of furfuryl alcohol with a productivity of 0.01 g/L/h. Although low productivity and a final concentration of furfuryl alcohol was obtained in the work by Kılmanoğlu et al. (2021) and also considering that the main goal of the work was not the production of furfuryl alcohol, the authors presented evidence that K. marxianus can produce furfuryl alcohol from furfural from an inexpensive real hydrolysate [ 60 ]."
} | 5,129 |
30255095 | PMC6145314 | pmc | 8,863 | {
"abstract": "Empty palm fruit bunch fiber (EPFBF) is a renewable resource in oil palm plantations that can be used for lignocellulosic bioethanol production. To enhance ethanol productivity with high-lignin-content EPFBF, the biomass was prepared with an alkali-thermal pretreatment (sodium hydroxide, 121°C, 60 min). The delignification yield was 55.4–56.9%, in proportion to the amount of sodium hydroxide, from 0.5 to 2.0 M. The lignin and hemicellulose contents of EPFBF were reduced by the pretreatment process, whereas the proportion of cellulose was increased. During enzymatic saccharification using Celluclast 1.5L and Novozyme 188 enzyme cocktails, about 62% of glucan was converted to a fermentable sugar. In simultaneous saccharification and fermentation, comparison among three ethanologenic yeast strains showed Saccharomyces cerevisiae W303-1A to be a candidate for maximum ethanol yield. In a batch fermentation with alkali-pretreated EPFBF hydrolysate, 21 g/L ethanol was obtained within 28 h, for a production yield of 0.102 g ethanol/g dry EPFBF or 0.458 g ethanol/g glucose. Moreover, a fed-batch fermentation produced 33.8±0.5 g/L ethanol with 1.57 g/L/h productivity in 20 h. These results show that the combination of alkaline pretreatment and biomass hydrolysate is useful for enhancing bioethanol productivity using delignified EPFBF.",
"conclusion": "4. Conclusions Alkali-thermal pretreatment of EPFBF with sodium hydroxide was effective in reducing hemicellulose and lignin contents and enhancing the enzymatic digestibility and fermentability of the biomass. The alkali-pretreated EPFBF showed ~55.4% delignification efficiency. In the pretreated biomass, 62% of the cellulose was hydrolyzed by the Celluclast 1.5L/Novozyme 188 enzyme cocktail and was converted to fermentable sugars during enzymatic saccharification. In a small-scale culture of simultaneous saccharification and fermentation supplemented with the pretreated biomass, S. cerevisiae , the best ethanol producer among the three yeast strains tested, produced 14.5 g/L ethanol with 0.14 g ethanol/g biomass and 82.2% fermentable sugar conversion. In a batch involving separate hydrolysis and fermentation supplemented with the alkali-pretreated EPFBF hydrolysate, 21 g/L ethanol was obtained within 28 h for a production yield of 0.102 g ethanol/g dry EPFBF or 0.458 g ethanol/g glucose. In addition, a fed-batch involving separate hydrolysis and fermentation could produce 33.8±0.5 g/L ethanol with 1.57 g/L/h productivity in only 20 h. These results confirm that alkali-pretreated EPFBF effectively reduced the hemicellulose and lignin components. Moreover, separate hydrolysis and fermentation using biomass hydrolysate may be useful for producing bioethanol with high productivity.",
"introduction": "1. Introduction Lignocellulosic biomass is a renewable bioresource for second-generation bioethanol production. Lignocellulosic biomass is composed primarily of cellulose, hemicellulose, and lignin. Cellulose and hemicellulose can be hydrolyzed or degraded enzymatically to glucose and a variety of pentose and hexose sugars, respectively, which can then be fermented to produce bioethanol [ 1 ]. However, the rigid cellulose structure, combined with the amorphous hemicellulose and lignin cross-linked structure, is chemically complex and resistant to degradation [ 2 ]. Thus, physical, chemical, and biological methods are needed to convert the complex structures of lignocellulose into fermentable sugars [ 3 , 4 ]. The production of bioethanol from lignocellulosic biomass requires four main steps: physical and chemical pretreatment of the lignocellulosic biomass, enzymatic hydrolysis of the lignocellulosic biomass, fermentation of the resulting sugars, and finally distillation of the ethanol [ 1 ]. Of these processes, pretreatment of the biomass is the most important step in saccharification efficiency, determining the ultimate bioethanol production yield [ 3 , 5 , 6 ]. The selection of the pretreatment procedure depends on the proportions of cellulose, hemicellulose, and lignin in the biomass and removing interfering materials from the biological steps that affect the overall bioprocess [ 3 , 4 ]. The palm oil industries in Indonesia and Malaysia generate ~8.2 million tons of lignocellulosic agricultural waste and byproducts per year [ 7 , 8 ]. An abundant byproduct in the palm oil industry is empty palm fruit bunch fiber (EPFBF), which consists of 27.6-32.5% lignin, 41.3-46.5% cellulose, and 25.3-33.8% hemicellulose [ 9 , 10 ]. EPFBF contains a relatively higher lignin content rather than other agriculture residual biomasses and feedstocks as a comparison of per gram of biomass (% lignin: % cellulose: % hemicellulose); soybean straw (19.2:44.2:5.9); wheat straw (23.4:38.2:21.2); corn stover (17.6:37.5:22.4); and switch grass (17.6:31.0:20.4) [ 11 – 14 ]. Lignocellulosic EPFBF is potentially a low-cost material and an alternative renewable bioresource, instead of food sources, such as corn, sugar cane, and other food stocks, for the production of bioethanol [ 7 , 8 , 11 ]. EPFBF contains relatively high levels of hemicellulose and lignin per gram biomass, compared with other lignocellulosic biomasses [ 7 – 10 ]. Thus, pretreatment is needed to reduce the hemicellulose and lignin contents before bioethanol fermentation. The pretreatment processes should enhance the proportion of cellulose in the EPFBF [ 8 – 10 ]. Acidic, alkali, and sequential acidic-alkali pretreatments, combined with high temperature or high pressure, have been applied in “conventional” chemical treatment processes [ 3 , 7 , 12 ]. Acidic pretreatments are known to be effective with lignocellulosic biomasses in reducing the hemicellulose content. Similarly, alkali pretreatments have been reported as simple processes for the delignification of biomass under mild conditions with minimal sugar degradation and without the formation of inhibitory compounds [ 12 ]. Thus, both acid and alkali pretreatments can reduce hemicellulose and lignin together to enhance the cellulose content of a biomass [ 10 , 13 – 15 ]. In this study, to increase the cellulose content of EPFBF, the alkali pretreatment was applied and the effects of this pretreatment were evaluated on reducing nonfermentable components in a lignocellulosic biomass for bioethanol production and productivity with the enzymatic biomass hydrolysate.",
"discussion": "3. Results and Discussion 3.1. Alkali Treatment of EPFBF and Its Composition To determine the composition of the EPFBF raw material prior to alkali pretreatment, the cellulose, hemicellulose, and lignin contents were analyzed using the analytical procedures of the US DOE. Dried EPFBF (100 g) consisted of 39% cellulose, 17% hemicellulose, and 28.8% lignin. Compared with previous data [ 7 – 9 ], the EPFBF used in this study contained a relatively high amount of lignin. Lignin interacts strongly with cellulose in biomass. It needs to be reduced or eliminated to enhance the enzymatic digestibility of the cellulose, to generate glucose, a fermentable sugar, and to increase bioethanol fermentation yield using yeast strains ( Figure 1 ). To prepare a high-cellulose-content biomass, alkali pretreatment was performed to reduce lignin in the EPFBF. After the biomass had dried completely, the change in the chemical composition in alkali-pretreated EPFBF was analyzed ( Figure 2 ). The concentration of sodium hydroxide affected the solubility of the biomass and the loss of cellulose, hemicellulose, and lignin contents (Supplementary Table S2 ). Alkali-thermal treatment within sodium hydroxide extracted up to 38.5–51.9% of the biomass into the soluble fraction. With 3 M sodium hydroxide pretreatment, the insoluble residue fraction contained 34.2 g cellulose, 4.9 g hemicellulose, and 9.0 g lignin per gram of the residual biomass. After alkali treatment, 10.2±0.5 % lignin and 18.9±1.2% hemicellulose were left in the residual biomass and the removal yields were ~71.1% and ~18.9%, respectively. Under high sodium hydroxide concentration from 2.5 M to 3.0 M, the half of the biomass was extracted as soluble fractions. On the other hand, the pretreated biomass did not lose much cellulose content, from 36.8 g to 34.2 g per gram of the residual EPFBF, with an increase in the sodium hydroxide concentration ranging from 0.5 to 2.0 M. The delignification yield with the alkali-thermal pretreatment was 55.4–56.9%, in proportion to the amount of sodium hydroxide. However, > 57% of the lignin could not be removed from the biomass under the alkali-thermal pretreatment conditions, even with a high concentration of sodium hydroxide. The alkali-thermal treatment of the biomass demonstrated that sodium hydroxide can extract hemicellulose and lignin effectively and increase the cellulose content per gram biomass. Sodium hydroxide treatment was effective for delignification of the biomass. The main advantage is that the alkali process condition is relatively mild [ 12 , 14 , 18 ]. These mild conditions prevent condensation of lignin, resulting in a high lignin solubility. Due to the mild conditions, degradation of sugars to furfural, HMF, and organic acids, considering potential inhibitors for enzymatic saccharification as well as ethanol fermentation, is prevented. Moreover, the sodium hydroxide solution could be reused several times. In addition, to compare with other chemical pretreatment processes using aqueous ammonia, organosolv, and ionic liquids, the alkali treatment does not need the special equipment and the reactor. The cost of sodium hydroxide and of the alkali recovery is cheaper rather than the other chemical catalysts. 3.2. Enzymatic Hydrolysis of EPFBF To assess the enzymatic hydrolysis of alkali-thermal-pretreated EPFBF with 1 M sodium hydroxide, 10% (w/v) of the biomass, which contained 35.1% cellulose, 12.1% hemicellulose, and 8.8% lignin per 100 g dry biomass, was digested with different concentrations of cellulase (Celluclast 1.5L), from 20 to 100 FPU (filter paper units) and a constant concentration of β -glucosidase (Novozyme 188; 40 CBU) for 72 h. The amount of glucose generated by enzymatic saccharification increased with increasing amounts of cellulase per unit pretreated EPFBF ( Figure 3(a) ). At 72 h, the cellulose-degrading enzyme cocktail generated 23.6–34.6 g/L glucose and 14.1–14.6 g/L xylose. At enzyme loading ratios of 0.5 to 2.5 Celluclast 1.5L per Novozyme 188, the enzymatic digestibility efficiency at 72 h ranged from 42.3% to 62% with the pretreated biomass. At a ratio of 2.5 FPU/g cellulose to 1 CBU/g cellulose, 34.6 g/L glucose and 14.6 g/L xylose were obtained from the pretreated biomass. However, the untreated biomass saccharification with equal amount of cellulase cocktail dose produced 7.1–10.4 g/L glucose and 4.2–4.5 g/L xylose under equal enzyme dose (data not shown). Compared with untreated biomass, the alkali-thermal-pretreated EPFBF produced 2-3 times higher amounts of fermentable sugars in the enzymatic hydrolysis. The chemical pretreatment could extract lignin from the biomass and reduce the strong interactions in cellulose-hemicellulose-lignin complexes in the biomass structure, enhancing the enzyme reaction efficiency [ 19 ]. Moreover, the morphological analysis of the untreated and the alkaline-treated biomass by scanning electron microscopy (SEM) showed clearly that the alkaline pretreatment made the biomass surface many cracks, porous, and rough structures. These morphological changes in the pretreated biomass might be for the cellulase to enhance accessibility as well as enzymatic hydrolysis of the polysaccharide structures (Supplementary Fig. S1 ). Nevertheless, an increase in enzyme dose did not significantly affect the amount of fermentable sugar generation. At the final reaction time point, the amount of glucose increased up to a 1.5 enzyme loading ratio of Celluclast®1.5/Novozyme 188, whereas the sugar concentrations decreased at 2.0 and 2.5 enzyme cocktail ratios (data not shown). It would be a product inhibition for the cellulase: the monosaccharides, disaccharides, and oligosaccharides generated by enzymatic hydrolysis might inhibit cellulase cocktails as a product inhibition. High concentration sugars produced by the cellulase could decrease further hydrolysis reaction [ 20 ]. Additionally, sugar production increased time-dependently, whereas enzymatic hydrolysis rates decreased exponentially due to product inhibition ( Figure 3(b) ). The generated glucose, xylose, and incompletely digested cellobiose may inhibit the hydrolysis reactions of the cellulase cocktail enzymes. The amount of xylose hydrolyzed from xylan in the biomass showed a near-constant value of ~14 g/L, even when the enzyme loading ratio was changed. The constant amount of xylose with enzyme supplementation was probably due to the β -xylosidase activity present in Novozyme188, which was loaded at a fixed concentration (40 CBU). All of the data presented for the enzymatic saccharification showed that pretreated EPFBF could be used to generate high concentrations of glucose, rather than xylose and other sugars, for bioethanol production. 3.3. Simultaneous Saccharification and Fermentation of Alkali-Pretreated EPFBF To test ethanol production with the enzymatic hydrolysis of alkali-pretreated EPFBF, three yeast strains, Saccharomyces cerevisiae W303-1A [ 10 , 13 ], Kluyveromyces marxianus CBS1555 [ 15 , 16 ], and Scheffersomyces stipitis ( Pichia stipitis ) CBS5776 [ 17 ], were used as ethanol producers. Batch cultivation of each yeast strain was performed in a 50-mL culture volume in a 250-mL Erlenmeyer flask with 5 g of alkali-pretreated EPFBF, supplemented with a 100 FPU Celluclast 1.5L and 40 CBU Novozyme 188 cocktail. Before cell inoculation, prehydrolysis was performed at 42°C for 12 h. Then, 5% (v/v) yeast inoculum (5 OD600 nm) was added and cultured further (30°C, 200 rpm, 18 h) for ethanol fermentation. The prehydrolysis step generated ~12.3±1.5 g/L glucose in each flask for yeast cell growth with no lag phase. The ethanol productivity of each yeast strain under these culture conditions is summarized in Table 1 . With simultaneous saccharification and fermentation (SSF), S. cerevisiae W303-1A was the “best” strain for maximum ethanol concentration and production yield: 10% (w/v) alkali-pretreated EPFBF was converted to 14.5 g/L ethanol with 0.14 g ethanol/g dried biomass. The yeast used 82.2% of the fermentable sugars for ethanol production, and no reduction in ethanol concentration was observed in culturing for 18 h. On the other hand, K. marxianus CBS1555 and Scheffersomyces stipitis CBS5776 produced 10.6 g/L ethanol with 0.11 g ethanol/g dried biomass and 10.1 g/L ethanol with 0.10 ethanol/g dried biomass, respectively. These yeast strains converted 60.1% and 57.2% glucose, respectively, in the biomass hydrolysate to ethanol. In SSF, among the yeast strains tested, S. cerevisiae W303-1A had maximum ethanol productivity and yield ( Table 1 ). Although inoculum size and initial glucose concentration were almost equal amounts in culture broth, the ethanol productivity of S. cerevisiae W303-1A was about 1.4-fold higher than those of K. marxianus CBS1555 and Scheffersomyces stipitis CBS5776 strains. These different productivities among the ethanologenic yeast might be due to the individual strain characteristics such as glucose uptake and metabolism, stress responsibility, and redox balance in the biomass hydrolysate [ 21 ]. 3.4. Batch Fermentation Supplemented with Alkali-Pretreated EPFBF Hydrolysate To assess the fermentability of the alkali-pretreated EPFBF hydrolysate, separate hydrolysis and fermentation reactions were performed using S. cerevisiae W303-1A in a 1 L jar fermentor. The enzymatic hydrolysate was prepared as described in Methods. The initial concentrations of glucose and xylose in the hydrolysate of alkali-pretreated EPFBF were 47.4±2.5 g/L and 19.0±1.2 g/L, respectively. After yeast inoculation, ethanol fermentation proceeded gradually for 30 h ( Figure 4(a) ). Yeast cell growth was slow until 4.5 h in a lag phase, increased exponentially, and entered a stationary phase between 16 and 24 h ( Figure 4(b) ). The concentration of glucose decreased gradually due to cell growth. All of the glucose in the culture broth was depleted at 28 h. The amount of xylose, which is a nonfermentable sugar, did not decrease. The highest ethanol concentration was 21 g/L at 28 h, giving an ethanol yield of 0.102 g ethanol/g dry EPFBF and 0.458 g ethanol/g glucose at 28 h. Ultimately, 85.4% of the fermentable sugar was used by the yeast for ethanol fermentation. No reduction in ethanol concentration was observed until glucose depletion. Cell growth and ethanol production in the separate hydrolysis and fermentation (SHF) were faster than those in the simultaneous saccharification and fermentation (SSF) (Supplementary Fig. S2 ). Ethanol fermentation of the alkali-pretreated EPFBF by S. cerevisiae W303-1A produced several metabolites, including acetic acid, lactic acid, succinic acid, and glycerol (data not shown). All metabolites produced by the yeast were at concentrations > 1 g/L, lower than that of the ethanol. In addition, these metabolites accumulated with no consumption in the fermentation broth. These metabolites did not appear to affect cell growth or ethanol productivity during the fermentation. Levels of these byproduct metabolites were probably too low to have any negative effects on cell metabolism due to the short fermentation time in the SHF process. These results indicated that SHF produced large amounts of ethanol with less byproducts and no inhibition by metabolites in the short fermentation time. 3.5. Fed-Batch Fermentation Supplemented with Alkali-Pretreated EPFBF Hydrolysate To increase ethanol production and yield, separate hydrolysis and fermentation in a fed-batch were performed using the alkali-pretreated EPFBF hydrolysate. When the residual glucose concentration fell below 0.5 g/L during the fermentation, the filtered enzymatic hydrolysate was fed into the fermentor to maintain ~10.0±1.5 g/L glucose. By 20 h, 33.8±0.5 g/L ethanol was produced ( Figure 5(a) ), and the residual glucose concentration was completely depleted. The productivity of ethanol was 1.57 g/L/h, and the production yield was 0.102 g per g EPFBF and 0.465 g per g glucose, with 91.2% sugar conversion efficiency. The yeast strain grew exponentially with no lag phase for 4 h and then entered a stationary phase at 8 h ( Figure 5(b) ). The cell density was maintained until the culture end time with no growth decrease. The initial glucose was consumed rapidly by the yeast by 8 h. In addition, the fed sugar in each 2-h period was also taken up rapidly by the viable yeast cells until the fermentation end point. Xylose could not be fermented by the yeast and thus accumulated in the fermentor ( Figure 5(a) ). Total reducing sugar decreased markedly until 8 h in the fermentation, and then the nonfermentable sugar increased gradually ( Figure 5(b) ). In the fed-batch cultivation, byproduct metabolites, such as organic acids and glycerol, were produced, but at < 0.53±0.12 g/L by the end of fermentation (data not shown). The cultivation time for ethanol production was only 20 h until the glucose was completely depleted. It may be that the byproducts did not accumulate in the fed-batch culture due to the short fermentation time. These results showed that separate hydrolysis and fermentation in a fed-batch using the alkali-pretreated EPFBF hydrolysate could produce bioethanol with high productivity in a short operation time."
} | 4,917 |
37754561 | PMC10581246 | pmc | 8,864 | {
"abstract": "ABSTRACT Petroleum-source and black carbon-source aromatic compounds are present in the cold seep environments, where ANaerobic MEthanotrophic (ANME) archaea as the dominant microbial community mediates the anaerobic oxidation of methane to produce inorganic and organic carbon. Here, by predicting the aromatics catabolic pathways in ANME metagenome-assembled genomes, we provide genomic and biochemical evidences that ANME have the potential of metabolizing aromatics via the strategy of CoA activation of the benzene ring using phenylacetic acid and benzoate as the substrates. Two ring-activating enzymes phenylacetate-CoA ligase (PaaK ANME ) and benzoate-CoA ligase (BadA ANME ) are able to convert phenylacetate to phenylacetyl-CoA and benzoate to benzoyl-CoA in vitro, respectively. They are mesophilic, alkali resistance, and with broad substrate spectra showing different affinity with various substrates. An exploration of the relative gene abundance in ANME genomes and cold seep environments indicates that about 50% of ANME genomes contain PCL genes, and various bacteria and archaea contain PCL and BCL genes. The results provide evidences for the capability of heterotrophic metabolism of aromatic compounds by ANME. This has not only enhanced our understanding of the nutrient range of ANME but also helped to explore the additional ecological and biogeochemical significance of this ubiquitous sedimentary archaea in the carbon flow in the cold seep environments. IMPORTANCE ANaerobic MEthanotrophic (ANME) archaea is the dominant microbial community mediating the anaerobic oxidation of methane in the cold seep environments, where aromatic compounds are present. Then it is hypothesized that ANME may be involved in the metabolism of aromatics. Here, we provide genomic and biochemical evidences for the heterotrophic metabolism of aromatic compounds by ANME, enhancing our understanding of their nutrient range and also shedding light on the ecological and biogeochemical significance of these ubiquitous sedimentary archaea in carbon flow within cold seep environments. Overall, this study offers valuable insights into the metabolic capabilities of ANME and their potential contributions to the global carbon cycle.",
"introduction": "INTRODUCTION In the cold-seep ecosystems, the anaerobic oxidation of methane (AOM) coupled with sulfate reduction is the primary energetic and biogeochemical process, which is mediated by a consortium of ANaerobic MEthanotrophic (ANME) archaea and sulfate-reducing bacteria ( 1 ). During the AOM process in the cold seep environments, the ANME are the key players in the methane assimilation with sulfate as electron acceptors ( 1 ). ANMEs [ANME-1 ( 2 ) and ANME-2 ( 3 )] oxidize methane using the reversed methanogenic pathway with different electron-transporting pathways. By the metabolism of methane assimilation, ANME possesses the potential ability to produce low-molecular organic acid acetate ( 4 ) and propanoate ( 1 ) as an alternative novel carbon flow supporting large autotrophic and heterotrophic bacterial populations existing in the cold-seep areas. It was also reported the presence of the petroleum-source polycyclic aromatic hydrocarbons through oil seeps ( 5 ), and black carbon-source aromatic compounds deposited from the surface, including aromatic acid, phenol, and naphthalene ( 6 ). These results in the accumulation of petroleum and sedimentary aromatic hydrocarbons in the deep-sea cold seep environments. Taken together, the presence of aromatics and the ANME as the key players in cold seep environments into consideration, we propose that ANME may be involved in the metabolism or utilization of aromatics, in addition to its chemoautotrophic lifestyle. During the microbial metabolism of aromatics, different activation strategies are employed for aromatic hydrocarbons activation, such as phosphorylation, fumarate insertion, hydroxylation, carboxylation, and methylation ( 7 ). Indeed, CoA ligation involved in ring activation is one of the major approaches for aromatic degradation ( 8 ), even for the subsequent reactions after fumarate insertion, carboxylation, decarboxylation, and methylation strategies. For the CoA ligation approach, phenylacetic acid (PAA) ( 8 ) and benzoic acid (BA) ( 9 ) are two major substrates for the aromatic acid CoA ligases forming the central aromatic intermediates ( 10 ). The two aromatic acid CoA ligases, phenylacetate-CoA ligase (PCL) and benzoate-CoA ligase (BCL) catalyze the ligation of CoA to the aromatic acids producing phenylacetyl-CoA (PA-CoA) and benzoyl-CoA (BA-CoA). Then the two CoA thioesters are transformed into nonaromatic compounds by corresponding multicomponent epoxidases, resulting in the formation of acetyl-CoA and succinyl-CoA in the aerobic pathways ( 9 , 11 ). For the anaerobic degradation pathway, PA-CoA is firstly oxidated forming BA-CoA ( 12 ) and the BA-CoA is then dearomatized by the multicomponent ATP-dependent ( 13 ) or ATP-independent ( 14 ) benzoyl-CoA reductase (BCR). Since the PCL and BCL are recruited in both aerobic ( 9 , 11 ) and anaerobic ( 10 , 11 ) aromatics degradation pathways (Fig. S1), the two aromatic acid CoA ligases activating the aromatic rings play a key role in the control of the potential precursors into the downstream degradation pathway. In contrast to the extensive research on aromatic compounds degradation in bacteria, only few cases from archaea have been reported to degrade aromatic compounds, illustrated by halophilic archaea using dioxygenases ( 15 ) and hyperthermophilic archaea using BCR ( 16 ). A study on the deep-sea archaea to degrade the aromatic compounds showed that Thermoprofundales (Marine Benthic Group D archaea) was involved in the aromatics degradation through PAA via the CoA activation strategy in the deep-sea hydrothermal sediment ( 8 ). Nevertheless, no further studies have been reported on archaea from deep-sea cold seep, hampering us from understanding the role of archaea on aromatics degradation and acting as a possible microbial carbon pump in seep area. To verify the hypothesis that archaea ANME in cold seep may be involved in the metabolism or utilization of aromatics, this study explored the presence of aromatic degradation genes in ANME genomes, the function and properties of PCL and BCL from ANME, and the relative abundance of the PCL and BCL genes in ANME genomes and cold seep environments. The potential of ANME to transform or utilize the aromatic hydrocarbons via the pathway involving PCL and BCL was explored by genomic analyses and enzymatic characterization. This study may provide evidences to elucidate that the dominant archaea ANME in cold seeps may be involved in the metabolism or utilization of aromatics apart from its chemoautotrophic lifestyle.",
"discussion": "DISCUSSION ANME was usually considered to act as chemoautotrophs to oxidize methane ( 3 ) or produce acetate ( 4 ) or propanoate ( 1 ) supporting heterotrophic bacterial populations in the cold seeps in the previous reports. In this study, the presence of enzymatically active PCL and BCL proteins in the MAGs of ANME archaea tentatively indicated that ANME may utilize the strategy of the CoA ligation to active the benzene ring and probably then transform or degrade the aromatic compounds ultimately. The results in this study suggest that, apart from the methane oxidation and production of low-molecular organic acid as an autotrophic microorganism, ANME may be involved in the heterotrophic aromatic compounds metabolism or help some heterotrophic microorganisms to transform aromatics in the cold spring environment. The functional genes are actually PCL gene. Nevertheless, expression pattern condition in ANME in vivo was to be studied further. On the other hand, because the ANME population are commonly regarded as uncultivated and to form syntrophic aggregates with sulfate-reducing bacteria ( 3 ), it is generally accepted that it is not practical to enrich or isolate the ANME population for performing in vivo assays. Thus, the enzymatic assay is virtually the best evidence to support the metabolic potential of ANME. The presence of genes involved in the PAA and BA pathway for the aromatic degradation, and the absence of monooxygenase and ring-cleavage dioxygenase genes in ANME genomes, along with the anaerobic methanotrophic character of ANME implied that several aromatic compounds may be transformed or utilized by ANME without the participation of oxygen. To take it a step further, instead of activating the benzene ring using the introduction of hydroxyl groups by oxygenation and the substrate ring-cleavage by dioxygenation reaction, the aromatics transformation by ANME may be through the route of aromatic CoA thioesters. For the strategy of the ring activation by CoA ligation, the formed PA-CoA or BA-CoA could be degraded through epoxidation pathway ( 9 , 11 ) or reductive pathway ( 9 , 12 ). In a study regarding the uncultured Thermoprofundales archaea from hydrothermal sediment, genomic and transcriptomic evidences established that aromatic degradation in Thermoprofundales was using the PCL and BCR enzymes through the reductive pathway ( 8 ). But in ANME, no BCR or epoxidases encoding genes were found in the aromatic degradation pathway in this study. The broad substrate spectra of Paak ANME and BadA ANME observed in this study could enable ANME to effectively adapt to benthic sediment environments under various carbon substrate conditions, especially in the oligotrophic pelagic ecosystem in the Gulf of Cádiz ( 18 ). Similarly, the PCLs from the uncultured archaea Thermoprofundales were also with broad substrate spectrum ( 8 ). But this was different for the PCLs from the terrestrial bacteria strain B. cenocepacia J2315 ( 8 ) and the thermophilic bacterium Thermus thermophilus strain HB27 ( 19 ), both with a much narrower substrate spectra. On the other hand, many BCLs reported did not accept monohydroxybenzoate isomers as substrates but usually use 2-aminobenzoate. And the reported K \n \n m \n values of the BCLs were less than 50 µM ranging from 0.6 µM to 45 µM ( 20 ). In contrast, the BadA ANME in this study has a similar affinity but uses monohydroxybenzoate isomers instead of 2-aminobenzoate. The PCL gene was widely distributed in ANME genomes for about half of ANME MAGs containing the PCL gene. It might be more than 50% presence in ANME MGAs of PCL gene if considering the incompleteness of many metagenome-assembled genomes. And it could get different results if we have more ANME MAGs. The PCL and BCL genes were identified in many bacteria and archaea in 77 cold seep metadates, indicating a wide distribution of aromatic metabolism in cold seep microbes. The ubiquitous distribution of PCL gene was assayed in the previous research ( 8 ). The more frequent distribution of the PCL gene than BCL in ANME genomes and metadata may indicate that the precursor aromatic compounds in cold seep environments may be more likely ring-activated via phenylacetate. The close relationship between BCLs from ANME and Methanosarcinales archaea suggested their common evolutionary origin in aromatics metabolism, further indicating the close phylogenetic relation between Methanosarcinales and ANME-2 described previously ( 3 ). It was reported that Chloroflexi ( Dehalococcoidia ) and Deltaproteobacteria in the cold seep area potentially degrade aromatic hydrocarbons using benzoyl-CoA as central metabolite but without BCL in some MAGs ( 21 ), the BCL genes in ANME may well help these bacteria to process the transformation of benzoate. On another hand, all the five ANME MAGs contain the putative MFS-type benzoate permease (with the identity of about 30% with the functionally identified benzoate transporter BenK), further enhancing the possibility of the transformation of BA by ANME in the cold seep environments. In conclusion, this study has not only enhanced our understanding of the nutrient range and metabolic repertoire of ANME but also helped to explore the additional ecological and biogeochemical significance of this ubiquitous sedimentary archaea ANME in the carbon flow in the cold seep environments."
} | 3,050 |
38600907 | PMC11004305 | pmc | 8,865 | {
"abstract": "Training neural networks to perform different tasks is relevant across various disciplines. In particular, Recurrent Neural Networks (RNNs) are of great interest in Computational Neuroscience. Open-source frameworks dedicated to Machine Learning, such as Tensorflow and Keras have produced significant changes in the development of technologies that we currently use. This work contributes by comprehensively investigating and describing the application of RNNs for temporal processing through a study of a 3-bit Flip Flop memory implementation. We delve into the entire modeling process, encompassing equations, task parametrization, and software development. The obtained networks are meticulously analyzed to elucidate dynamics, aided by an array of visualization and analysis tools. Moreover, the provided code is versatile enough to facilitate the modeling of diverse tasks and systems. Furthermore, we present how memory states can be efficiently stored in the vertices of a cube in the dimensionally reduced space, supplementing previous results with a distinct approach.",
"conclusion": "6 Conclusions In this work, all steps to build and analyze an RNN have been presented for a sample task. We started from the model description in terms of the equations, discretization, and code implementation. We discussed different options that are available for code implementation depending on the considered model and scientific questions. Then, we described the task parametrization and network training protocol. We also presented a set of tools to analyze the results using open-source scientific libraries making use of the different visualization tools that allow extracting relevant features. We used the Flip Flop task as an example, but other relevant tasks could be considered, as mentioned in Section 3. For example, “Perceptual Decision Making” (Britten et al., 1992 ), “Context-dependent Decision Making” (Mante et al., 2013 ; Zhang et al., 2021 ), working memory tasks such as “Delay match to sample with two items” (Freedman and Assad, 2006 ) or “Parametric working memory” (Roitman and Shadlen, 2002 ). In this work, motivated by Sussillo and Barak ( 2013 ), a working memory task such as a 3-bit Flip Flop, was chosen to show the entire process: from the differential equations of the RNN model, discretization, through the parameterization of the task and the methods of analysis for the activity of the network against the different stimuli on the network. The use of open-source scientific frameworks designed and maintained for large communities, such as the tools used here, allows enhancing research. This is why we are currently using tools that are more transparent in terms of code and documentation because they are open to being modified and improved by thousands of users. Regarding the limitations, the proposed method was evaluated on a single cognitive task, namely Flip Flop. It is not clear whether the proposed pipeline would generalize to other more complex types of cognitive tasks. We did not include other explicit biological constraints in this example. We could extend it to include sparsity or Dale's law, for example. Further work could address such research directions to complement generalization and biological details.",
"introduction": "1 Introduction Machine learning methods, especially Deep Learning, have achieved remarkable success across diverse tasks in various domains. These include speech processing (Ogunfunmi et al., 2019 ), bioinformatics (Min et al., 2016 ), where algorithms predict protein structures, discover drugs and analyze gene expression data, and image recognition (Litjens et al., 2017 ) where deep learning classifies images and detects objects. The emergence of open-source frameworks dedicated to Machine Learning, such as Pytorch, Tensorflow and Keras (Abadi et al., 2015 ; Chollet et al., 2015 ; Paszke et al., 2019 ) has produced huge changes in the development of technologies we use every day for different tasks in research. Due to their novelty and complexity, it can be challenging to properly learn how to utilize these frameworks in different relevant scientific domains, such as the development of models in Computational Neuroscience, which will be the aim of the present work. To bridge the gap between theoretical knowledge and practical application, clear tutorials or primers are crucial. These resources should equip researchers not only with the ability to implement the algorithms but also with the skills to solve diverse problems pertinent to their field. Recurrent Neural Networks (or RNNs) were originally invented by Paul Werbos, who also invented backpropagation, a fundamental tool for training these models (Werbos, 1990 ). This also includes the concept of latent variables. The problem of training neural networks to perform different tasks is relevant across various disciplines that go beyond Machine Learning. In particular, RNNs are of great interest in different scientific communities. These models also have great relevance concerning control systems and other areas such as electronics (Alianna J. Maren and, Auth.; Deng, 2013 ; Dinh et al., 2014 ; Mohajerin and Waslander, 2017 ). One relevant problem to address with them is how to build models for the study of dynamical systems and how to extract meaningful information from them. In general, Neural Networks are algorithms that allow us to model different systems. According to the Universal Approximation Theorem, a neural network with one hidden layer containing a sufficient but finite number of neurons can approximate any continuous function to a reasonable accuracy under certain conditions for activation functions (Hornik, 1991 ). This theorem has been extended to RNNs. It is well known that dynamical systems can be approximated by continuous-time RNNs (Funahashi and Nakamura, 1993 ). In particular, RNNs are widely used in the field of Computational Neurosciences to describe the behavior of cortical areas, which presents great recurrence in their connections (Murphy and Miller, 2009 ). They are related to the processing of temporal information and the production of time-dependent outputs. The basic premise of RNNs is that the feedforward connection weights in a Multilayer Perceptron (MLP) neural network (McCulloch and Pitts, 1943 ) can be modified using prior activation history as well as the immediately presented stimulus. This mechanism can be considered to encapsulate, in a very simple model, the much broader and more interesting task of guiding neural behavior. Factors that influence neural interactions and even growth can be included within this simple model. In this context, the broader scope of systems neuroscience relates to a detailed and careful analysis of RNNs. The realm of temporal influence within systems neuroscience has a long and substantive history. The work by Levi-Montalcini and Booker ( 1960 ); Levi-Montalcini ( 1987 ), was among the earliest to show how specific signaling proteins (nerve growth factors, or NGFs) could influence temporal evolution within an organism. More recently, Baldassarro et al. ( 2023 ) showed, in an in vitro study, that NGFs could influence the proliferation of fetal brain multipotent stem cells, pushing them into a specific oligodendrocyte cell lineage and also influencing the differentiation of oligodendrocyte precursor cells. These works are simply examples of how the complex process of influencing neural cell growth and differentiation can be influenced over time, by introducing specific signaling mechanisms. For this, the notion of RNNs encapsulates a much larger suite of neural processes. In this way, RNNs allow the incorporation of realistic characteristics at the biological level, such as Dale's law (Dale, 1935 ; Rajan and Abbott, 2006 ; Song et al., 2016 ; Jarne and Caruso, 2023 ), sparsity or different characteristics of interest in animal models. In the field of Machine learning, more sophisticated architectures such as LSTM (Long Short Term Memory units) or GRU (Gated recurrent units) are widely spread and have been used to process temporal sequences since they do not have the same limitations as RNNs to process long time dependencies (Bengio et al., 1994 ; Pascanu et al., 2013 ; Chung et al., 2014 ; SHI et al., 2015 ; Gudowska-Nowak et al., 2020 ). Other powerful models are based on spiking neural networks (SNNs). Several recent studies have made significant contributions to the field of brain-inspired intelligence. These studies demonstrate the potential of this field to achieve high-level intelligence, high accuracy, high robustness, and low power consumption (Yang et al., 2022a , b , 2023 ; Yang and Chen, 2023a , b ). The primary reason for using simple RNN models lies in their ability to comprehend neural computation through collective dynamics, a phenomenon intricately linked to motor control, brain temporal tasks, decision-making (Mante et al., 2013 ), neural oscillations and working memory (Vyas et al., 2020 ; Jarne and Caruso, 2023 ; Pals et al., 2024 ). Analyzing the dynamics inherent in these models allows us to formulate various hypotheses regarding the functioning of different brain areas and to offer an interpretation for the experimental results observed (Barak, 2017 ; Kao and Hennequin, 2019 ). An illustrative instance involves the recent utilization of RNNs to transfer learned dynamics and constraints to a spiking recurrent neural network in a one-to-one fashion (Kim et al., 2019 ). A well-established fact is that the dynamics of a network are heavily influenced by the eigenvalue spectrum of the weight matrix describing synaptic connections (Zhou et al., 2009 ). Thus, the significance of investigating this distribution lies in elucidating various aspects of the dynamic behavior of the system, which is why, in Section 5.2, such analysis will be presented and described. There are general tutorials available on artificial neural networks, such as Yang and Wang ( 2020 ). However, in this work, we will focus extensively on RNNs and their application in Computational Neuroscience because they play a relevant role in understanding complex neural processes and dynamics. Throughout this tutorial, we will delve into the architecture, training methodologies, and practical aspects of the RNN implementation. We explore also their significance and potential contributions to the field. A simple RNN was chosen and trained to perform a time-series processing task inspired by Computational Neuroscience studies (Sussillo, 2014 ). The implementation of the network, the training, and the tools are carefully described here, as well as different forms to obtain the information that allows a suitable description of the system under study. Training an RNN to perform temporal tasks has many difficulties and can be done through various paradigms. Here it is proposed to approach the problem through supervised learning. The entire procedure is described in detail. Among the different tasks, the Flip Flop was chosen as a case example. On one hand, a Flip Flop is the simplest sequential system that one can build Floyd ( 2003 ). To be precise, a 3-bit memory was studied, which is a task composed of a set of Flip Flops as the one shown in Figure 1 . This is also a working memory task considered previously in other works in Computational Neuroscience (Sussillo and Barak, 2013 ; Barak, 2017 ; Jarne, 2022 ). The parameterization of the chosen task, one fundamental key in any work related to trained RNNs, is as described in Sussillo and Barak ( 2013 ). It is also revisited here. Gradient descendant minimization was used to take advantage of different optimized implementations of the available algorithms. The code implementation is presented using Tensorflow and Keras. The reason for this choice is that such scientific libraries are open-source, their use is rapidly growing, and they are becoming increasingly popular. One can find excellent documentation for software development about them Gulli and Pal ( 2017 ); Ramsundar and Zadeh ( 2018 ); Singh and Manure ( 2019 ). Also, we have new tools such as Google Colaboratory that allow implementing and testing models directly online. Figure 1 A Flip Flop. Binary task designed to store one bit of information. It has two inputs, Set (S) and Reset (R), which in here will be represented by the temporal signal states of B 0 +1 and -1 in amplitude, and one output Q. The output represents the current state of the Flip Flop and can be either 1 or -1. The focus of this paper is on elucidating how a trained RNN operates, with code provided for detailed study. The “Flip Flop problem” is chosen to illustrate the study. Every step is thoroughly explained, from parameterizing the task to describing the dynamics of trained networks.This example is used to show how the problem of training networks can be studied using these computing tools applied in any temporal task in general, but also to discuss the limitations that networks have and the alternatives to solve them. The rest of the paper is organized as follows. In Section 2, the description of the dynamics, discretization and code examples are presented. In Section 3, the task parametrization is shown. Section 4 describes the training protocol. In Section 5, the results, different analyses of the network, tools and software are discussed in detail. Finally, Section 6 includes the final remarks."
} | 3,361 |
36465038 | PMC10107873 | pmc | 8,868 | {
"abstract": "Abstract The Pseudomonas putida group in the Gammaproteobacteria has been intensively studied for bioremediation and plant growth promotion. Members of this group have recently emerged as promising hosts to convert intermediates derived from plant biomass to biofuels and biochemicals. However, most strains of P. putida cannot metabolize pentose sugars derived from hemicellulose. Here, we describe three isolates that provide a broader view of the pentose sugar catabolism in the P. putida group. One of these isolates clusters with the well‐characterized P. alloputida KT2440 (Strain BP6); the second isolate clustered with plant growth‐promoting strain P. putida W619 (Strain M2), while the third isolate represents a new species in the group (Strain BP8). Each of these isolates possessed homologous genes for oxidative xylose catabolism ( xylDXA ) and a potential xylonate transporter. Strain M2 grew on arabinose and had genes for oxidative arabinose catabolism ( araDXA ). A CRISPR interference (CRISPRi) system was developed for strain M2 and identified conditionally essential genes for xylose growth. A glucose dehydrogenase was found to be responsible for initial oxidation of xylose and arabinose in strain M2. These isolates have illuminated inherent diversity in pentose catabolism in the P. putida group and may provide alternative hosts for biomass conversion.",
"introduction": "INTRODUCTION Lignocellulosic biomass is an abundant and sustainable global source as feedstocks for the production of biofuels and bio‐based products (Dahmen et al., 2019 ; Paul & Dutta, 2018 ). Biofuels and bio‐based chemicals have been traditionally produced from lignocellulosic hydrolysates by microorganisms such as Saccharomyces cerevisiae and Escherichia coli (Liu et al., 2021 ). However, these microbial hosts are limited in substrate range and are sensitive to toxic inhibitors that are often present in hydrolysates (Piotrowski et al., 2014 ). Therefore, other potential hosts with broader substrate ranges and higher tolerance to inhibitors have been developed to complement S. cerevisiae and E. coli (Keasling et al., 2021 ). Among the most promising is Pseudomonas putida KT2440, which has been recently reclassified as Pseudomonas alloputida KT2440 (Bentley et al., 2020 ; Dong et al., 2019 ; Shi et al., 2017 ; Shields‐Menard et al., 2018 ; Wang et al., 2018 ). Pseudomonas alloputida is a representative of the P. putida group in the Gammaproteobacteria, members of which have been intensively studied for their role in bioremediation and plant growth promotion. Pseudomonas alloputida KT2440 is of particular interest because of its ability to convert plant‐derived aromatics and has been engineered to produce a variety of biofuels and bio‐based chemicals from both glucose and aromatics (Banerjee et al., 2020 ; Dong et al., 2019 ; Johnson et al., 2019 ). Pentose sugars xylose and arabinose are the predominant constituents of hemicellulose; xylose makes up a substantial amount of the total plant sugars (10%–25% of dry biomass) followed by arabinose (usually 2%–3%, although some hydrolysates contain up to 20%) (Agrawal et al., 2015 ; Dehghanzad et al., 2020 ; Kumar, Binod, et al., 2018 ; Narisetty et al., 2022 ; Rocha et al., 2015 ; Zhang et al., 2014 ). However, P. alloputida KT2440 is not able to catabolize pentose sugars (Isikgor & Becer, 2015 ; Lim et al., 2021 ). Therefore, expanding the substrate range of P. alloptuida to include pentose sugars will improve the overall carbon conversion efficiency of lignocellulosic hydrolysates. Several approaches have been used to engineer P. alloputida to utilize pentose sugars. The xylose isomerase pathway from E. coli , which converts xylose to intermediates in the pentose phosphate pathway, has been expressed in P. alloputida and used to convert xylose to cis‐cis ‐muconic acid (Dvořák & de Lorenzo, 2018 ; Le Meur et al., 2012 ). Two oxidative pathways for xylose catabolism that proceed through xylonate as an intermediate, the Weimberg pathway, originally characterized in Caulobacter crescentus , and the Dahms pathway from E. coli , have been expressed in P. alloputida and have been used to produce rhamnolipids and indigoidine (Bator et al., 2020 ; Lim et al., 2021 ). Heterologous expression of isomerase ( E. coli ) and oxidative ( Burkholderia ambifaria ) pathways for arabinose catabolism has allowed it to grow on this pentose sugar (Elmore et al., 2020 ). However, problems such as genetic instability, long lag‐phases and low cell density have been encountered during these engineering efforts (Elmore et al., 2020 ; Jeffries, 2006 ; Kang et al., 2018 ; Meijnen et al., 2009 ). A complementary approach to genetic engineering is to obtain isolates related to P. alloputida KT2440 that can natively grow on pentose sugars. These isolates may serve as alternative hosts for the production of biofuels and bio‐based chemicals. One member of the P. putida group, Pseudomonas taiwanensis , grows on xylose (Köhler et al., 2015 ). Analysis of the P. tawianensis genome indicated that it possessed a variant of the oxidative xylose pathway found in C. crescentus , converting xylose through oxidative intermediates (xylonate and 2‐keto‐3‐deoxyxylonate) to α‐ketoglutarate. The P. taiwanensis pathway has been expressed in P. alloputida KT2440, conferring on it the ability to grow on xylose (Bator et al., 2020 ). Here, we describe a targeted isolation approach to obtaining Pseudomonas species related to P. alloputida KT2440 that grow on xylose and arabinose. These isolates illuminate the extent of pentose catabolism in the P. putida group and provide possible new hosts for metabolic engineering.",
"discussion": "DISCUSSION This study has provided a broad perspective on the ability of members of the P. putida group to metabolize pentose sugars relevant to lignocellulose bioconversion. Targeted isolation studies carried out with xylose and p ‐coumarate as the targeted substrates provided isolates that expressed an oxidative pathway for xylose catabolism similar to the Weimberg pathway in C. crescentus (Shen et al., 2020 ). Comparison of the isolates to the P. putida group genomes indicates that xylose oxidation is relatively restricted in this group. Strain BP6 and BP7 cluster with P. alloputida , the species which is the most thoroughly characterized of the P. putida strains, including strain KT2440 (Keshavarz‐Tohid et al., 2019 ). The BP6 and BP7 strains are most closely related to P. alloputida LF54, which is a representative of the most divergent clade of P. alloputida (Passarelli‐Araujo et al., 2021 ). The M2 and M5 strains are affiliated with P. putida W619, which has been characterized for its ability to promote the growth of plants (Taghavi et al., 2005 ). P putida W619 has been shown to grow on both xylose and arabinose (Davis et al., 2013 ). The survey also revealed other members of the P. putida group ( Pseudomonas sp . BP8 and P. vranovensis ) that grew on xylose along with the previously characterized P. taiwanensis . Interestingly, the ability to grow on arabinose is even more constrained and is only present in P . monteilii and P . plecoglossicida along with the clade containing P. putida M2/M5 and W619. These results are consistent with pentose oxidation being a niche activity in the P. putida group. Strain BP8 had genes for additional pathway for xylose oxidation that may divert intermediates in the oxidative pathway to pyruvate and glycolate. Proteins for both pathways were present in the xylose‐grown cells. Integrated genomic and proteomic analyses demonstrated that homologues of the Weimberg pathway (XylDXA and AraDXA) were responsible for xylose and arabinose oxidation in the P. putida group strains. These proteins are likely responsible for converting the oxidized sugar (xylonate and arabinoate) through multiple dehydrations and an oxidation to produce α‐ketoglutarate. Repression of the genes encoding these proteins using CRISPRi in P. putida M2 demonstrated that only xylD interference repressed xylose growth. This result is consistent with studies in P. alloputida KT2440 that expression of only xylD is required to confer growth on xylose by these isolates, and that the activities of XylX and XylA can be recruited using other proteins (Lim et al., 2021 ). In the putative operon for xylonate oxidation, there is a transcriptional regulator and permease found in common in all the strains. However, the newly isolated strains have an annotated transporter (MHS family metabolite: H+ symporter) that is not present in P. tawianensis . The presence of the transporter, which was identified in the proteomes of Strains M2, BP6 and BP8, may improve growth on xylose relative to P. taiwanensis VLB120 (Table 1 ). The importance of this transporter was reinforced in the P. putida M2 CRISPRi experiments, as repression of this transporter inhibited the xylose growth of Strain M2. The possible substrate for the transporter is xylonate, as both CRISPRi‐based repression of gcd in Strain M2 and deletion of gcd in Strain BP6 affected the ability of the strains to grow on xylose, indicating that Gcd oxidizes xylose to xylonate in the periplasm and then it is transported into the cytosol by the transporter present in the isolates described here. The xylose dehydrogenase activity of Gcd is consistent with previous studies on P. alloputida KT2440 (Dvořák & de Lorenzo, 2018 ), P. putida NCTC 10936 (Hardy et al., 1993 ) and P. taiwanensis VLB120 (Köhler et al., 2015 ), which demonstrated that Gcd was required for xylose oxidation. Interestingly, CRISPRi experiments also demonstrated that Gcd was required for arabinose growth, suggesting that arabinose is also oxidized to arabinoate in the periplasm and then the arabinoate is transported into the cytosol. The requirement of Gcd for pentose‐based growth in multiple P. putida group strains and for multiple substrates indicates there is a link in the metabolism of all three major lignocellulose‐derived sugars in the members of the P. putida group. The catalytic promiscuity of Gcd for lignocellulose‐derived hexose and pentose sugars has also been demonstrated in sugar catabolism in Sulfolobus solfataricus and Sulfolobus acidocaldarius (Nunn et al., 2010 ). The growth rates of the native xylose‐oxidizing strains in the P. putida group are comparable or better than strains of P. alloputida KT2440 that have been engineered to grow on xylose (Elmore et al., 2022 ). Therefore, these isolates may be suitable complements to strain KT2440 for metabolic engineering focused on lignocellulose bioconversion."
} | 2,689 |
39009031 | PMC11283204 | pmc | 8,869 | {
"abstract": "Abstract Lignocellulose (dry plant biomass) is an abundant cheap inedible residue of agriculture and wood industry with great potential as a feedstock for biotechnological processes. Lignocellulosic substrates can serve as valuable resources in fermentation processes, allowing the production of a wide array of chemicals, fuels, and food additives. The main obstacle for cost-effective conversion of lignocellulosic hydrolysates to target products is poor metabolism of the major pentoses, xylose and L-arabinose, which are the second and third most abundant sugars of lignocellulose after glucose. We study the oversynthesis of riboflavin in the flavinogenic yeast Candida famata and found that all major lignocellulosic sugars, including xylose and L-arabinose, support robust growth and riboflavin synthesis in the available strains of C. famata . To further increase riboflavin production from xylose and lignocellulose hydrolysate, genes XYL1 and XYL2 coding for xylose reductase and xylitol dehydrogenase were overexpressed. The resulting strains exhibited increased riboflavin production in both shake flasks and bioreactors using diluted hydrolysate, reaching 1.5 g L −1 .",
"conclusion": "Conclusions The engineered strains of the yeast C. famata overproduced riboflavin on lignocellulosic hydrolysate. Riboflavin accumulation increased in the recombinant strains overexpressing the XYL1 and XYL2 genes, which code for xylose reductase and xylitol dehydrogenase, respectively. The highest riboflavin accumulation during bioreactor cultivation by BRPI/XYL1, using bagasse hydrolysate as the carbon source, reached 1.5 g L −1 .",
"introduction": "Introduction Riboflavin (vitamin B 2 ) is a water-soluble vitamin, produced by all plants and most of microorganisms. It is essential for growth and reproduction of humans and animals. Riboflavin is the precursor of flavin nucleotides (FMN and FAD), crucial coenzymes involved in various oxidoreductive reactions (Abbas and Sibirny 2011 , Schwechheimer et al. 2016 ). Riboflavin is an important biotechnological product that is used mainly in agriculture as a feed additive, as well as in the food industry and medicine (Beztsinna et al. 2016 , You et al. 2021 ). Riboflavin is currently biotechnologically produced by engineered strains of the bacterium Bacillus subtilis and the filamentous fungus, Ashbya (Eremothecium) gossypii (Ruchala et al. 2022 ). To enhance microbial riboflavin production, strains were improved through metabolic engineering and classical selection (Zhao et al. 2021 ). These approaches involved random mutagenesis induced by chemical exposure and UV irradiation, as well as random and site-directed mutagenesis achieved through genetically engineered deletions, insertions, or substitutions (Zhao et al. 2021 ). Additionally, the fermentation process was optimized by selecting and adjusting medium components and their concentrations (You et al. 2021 ). However, further research is required to enhance industrial riboflavin processes, focusing on improving key steps, including fermentation conditions, purification techniques, and the utilization of recycled sources. The price of riboflavin depends significantly on the price of the carbon substrate used (Kato and Park 2012 ). It would be especially important to use waste as a carbon source. In our current work, we studied riboflavin production by the yeast Candida famata (teleomorph form is known as Debaryomyces subglobosus Nguyen et al. 2009 ). The riboflavin overproducing strain of C. famata was used for more than 10 years for industrial riboflavin production. However, its production was discontinued due to genetic instability of the used strain (Abbas and Sibirny 2011 ). We constructed stable non-reverting riboflavin-overproducing strains of C. famata using a combination of classical selection methods and metabolic engineering. These strains accumulated 1–1.45 g L −1 of riboflavin in flasks and reached 16.4 g L −1 in a 7 L bioreactor during fed-batch cultivation (Dmytruk et al. 2011 , Dmytruk et al. 2014 ). A yeast strain overexpressing the RFE1 gene, which encodes riboflavin excretase, produced 1.7 g L −1 of vitamin B 2 (Tsyrulnyk et al. 2020 ). Modulation of the purine biosynthesis pathway further enhanced riboflavin production, resulting in up to 2.85 g L −1 (Dmytruk et al. 2020 ). The engineered strains produced riboflavin in media with glucose as the carbon source. The latest strain, expressing Sef1—a positive regulator of riboflavin synthesis—under a lactose-inducible promoter, or the structural RIB6 gene, achieved riboflavin production of approximately 2.5 g L −1 using milk whey, a byproduct of cheese production, as the carbon source (Ruchala et al. 2022 ). Lignocellulosic hydrolysates can serve as valuable resources for fermentation processes, allowing the production of a wide range of chemicals, fuels, and food additives. For instance, ethanol (Robak and Balcerek 2020 ), organic acids (Jimenez-Quero et al. 2020 ), polymers (Kawaguchi et al. 2017 ), and enzymes (Namnuch et al. 2021 ) are among the diverse products that can be synthesized through this approach. Different yeast species have been considered for the production of chemicals, mainly ethanol and xylitol, using different lignocellulosic substrates, such as horticultural waste olive tree pruning, rice straw, corncob, and sugarcane bagasse (Bergmann et al. 2019 ). In this study, we demonstrated that engineered strains of C. famata exhibit riboflavin overproduction in medium containing lignocellulose hydrolysate. We investigated both, previously isolated, and newly constructed recombinant strains with overexpression of XYL1 and XYL2 genes encoding for xylose reductase and xylitol dehydrogenase, respectively, for their riboflavin production capabilities in flask and bioreactor experiments.",
"discussion": "Discussion When considering the importance of chemical feedstock, the production of high-value substances from renewable sources is becoming increasingly appealing. Lignocellulosic hydrolysates are attractive substrates in microbial fermentation processes for the production of commercially relevant compounds (Baptista et al. 2021 ). Previous reports have demonstrated riboflavin production from lignocellulosic hydrolysates by bacteria. Using corn cob hydrolysate as a carbon source, growth and riboflavin biosynthesis were optimized by the thermophilic strain Geobacillus thermoglucosidasius (Wang et al. 2022 ). The overexpression of the mannose-6-phosphate isomerase gene manA , xylose isomerase gene xylA , and xylulokinase gene xylB to enhance mannose and xylose consumption in Corynebacterium glutamicum resulted in a 56% increase in riboflavin productivity from sugars derived from lignocellulose breakdown (Pérez-García et al. 2021 ). However, the riboflavin production by the mentioned microorganisms reached only 121 mg L −1 and 27 mg L −1 , which are too low for considering them as viable industrial producers of vitamin B2 at this stage. So far, no microorganism that could efficiently produce riboflavin in lignocellulosic hydrolysates has been reported. To address this gap, we tested the ability of the C. famata riboflavin-overproducing strain BRPI, which was constructed in our previous work (Dmytruk et al. 2020 ), to grow on the main sugars found in lignocellulosic hydrolysates and to produce riboflavin. Both, hexoses (glucose, fructose, mannose, galactose) and pentoses (L-arabinose, xylose) supported growth of BRPI (Table 1 ). Even more intriguing was the result that BRPI synthesized riboflavin on all tested carbon sources (Table 1 ). We have demonstrated, for the first time, that C. famata is capable of overproducing riboflavin on L-arabinose and xylose. It has been shown that L-arabinose is a good carbon source for vitamin B 2 production by recombinant Bacillus subtilis (Oraei et al. 2018 ). BRPI was also able to grow and produce riboflavin in synthetic medium with a glucose/xylose mixture and in medium with diluted sugarcane straw hydrolysate (Fig. 2 , 3 , 5 ). The overexpression of the XYL1 gene, encoding xylose reductase, resulted in a 10–15% increase in riboflavin production in medium with diluted hydrolysate. However, additional overexpression of the XYL2 gene, which encodes xylitol dehydrogenase, did not lead to a further increase in riboflavin synthesis. The maximally achieved riboflavin titer of around 1.5 grams per liter show promise for future research and practical applications. This work represents the first communication, to the best of our knowledge, on the overproduction of riboflavin using lignocellulosic hydrolysate. Additional increases in riboflavin production could be attained through the two strategies involving strain construction: (i) enhancing riboflavin synthesis by activating pathways responsible for riboflavin precursors (GTP, ribulose-5-phosphate) synthesis, and (ii) further enhancing xylose utilization by overexpressing xylulokinase. It was observed that undiluted hydrolysate did not support the growth of C. famata strains, likely due to the inhibitory effects of furfural, HMF and acetic acid. This finding suggests the need for further work in selecting strains that are insensitive to these inhibitors. For instance, adaptive laboratory evolution could be considered as a method to achieve this goal (Ujor and Okonkwo 2022 ). Our results indicate that less diluted hydrolysate supports faster growth and higher riboflavin production (Fig. 5 ). Bearing this in mind, optimizing the fed-batch mode in bioreactors with hydrolysate as the carbon source could lead to increased riboflavin production. Finally, it is worth emphasizing that C. famata is an ideal candidate for riboflavin production from lignocellulosic hydrolysates due to its robust growth and high riboflavin yield. Notably, it can efficiently utilize not only glucose and xylose but also other major sugars present in hydrolysates. These unique attributes, uncommon among native microorganism strains, offer promising opportunities for the development of industrially viable riboflavin producers using lignocellulose, pectin, beet pulp, and other relevant residues."
} | 2,554 |
31134748 | null | s2 | 8,870 | {
"abstract": "The question whether communities should be viewed as superorganisms or loose collections of individual species has been the subject of a long-standing debate in ecology. Each view implies different spatiotemporal community patterns. Along spatial environmental gradients, the organismic view predicts that species turnover is discontinuous, with sharp boundaries between communities, while the individualistic view predicts gradual changes in species composition. Using a spatially explicit multispecies competition model, we show that organismic and individualistic forms of community organisation are two limiting cases along a continuum of outcomes. A high variance of competition strength leads to the emergence of organism-like communities due to the presence of alternative stable states, while weak and uniform interactions induce gradual changes in species composition. Dispersal can play a confounding role in these patterns. Our work highlights the critical importance of considering species interactions to understand and predict the responses of species and communities to environmental changes."
} | 276 |
33281592 | PMC7691602 | pmc | 8,871 | {
"abstract": "Engineering neural networks to perform specific tasks often represents a monumental challenge in determining network architecture and parameter values. In this work, we extend our previously-developed method for tuning networks of non-spiking neurons, the “Functional subnetwork approach” (FSA), to the tuning of networks composed of spiking neurons. This extension enables the direct assembly and tuning of networks of spiking neurons and synapses based on the network's intended function, without the use of global optimization or machine learning. To extend the FSA, we show that the dynamics of a generalized linear integrate and fire (GLIF) neuron model have fundamental similarities to those of a non-spiking leaky integrator neuron model. We derive analytical expressions that show functional parallels between: (1) A spiking neuron's steady-state spiking frequency and a non-spiking neuron's steady-state voltage in response to an applied current; (2) a spiking neuron's transient spiking frequency and a non-spiking neuron's transient voltage in response to an applied current; and (3) a spiking synapse's average conductance during steady spiking and a non-spiking synapse's conductance. The models become more similar as additional spiking neurons are added to each population “node” in the network. We apply the FSA to model a neuromuscular reflex pathway two different ways: Via non-spiking components and then via spiking components. These results provide a concrete example of how a single non-spiking neuron may model the average spiking frequency of a population of spiking neurons. The resulting model also demonstrates that by using the FSA, models can be constructed that incorporate both spiking and non-spiking units. This work facilitates the construction of large networks of spiking neurons and synapses that perform specific functions, for example, those implemented with neuromorphic computing hardware, by providing an analytical method for directly tuning their parameters without time-consuming optimization or learning.",
"introduction": "1. Introduction Neuromorphic computing hardware is becoming more widely available (Khan et al., 2008 ; Pfeil et al., 2013 ; Benjamin et al., 2014 ; Gehlhaar, 2014 ; Merolla et al., 2014 ; Ionica and Gregg, 2015 ; Davies et al., 2018 ). Such chips have non-traditional architecture, with highly-parallel processing and specialized circuits that mimic neural and synaptic dynamics. These chips mimic the communication of spiking neural networks, whose discrete communication events (i.e., spikes) reduce the communication overhead relative to continuous networks. Many canonical brain networks have been tested with these chips including decorrelation networks, winner-take-all networks, and balanced random networks (Pfeil et al., 2013 ), as well as other networks that perform complex computations, such as multi-object recognition (Merolla et al., 2014 ) and keyword-matching (Blouw et al., 2018 ), using <100 mW of power in the process. Neuromorphic hardware is advancing both computational neuroscience (Eliasmith and Anderson, 2002 ; Eliasmith et al., 2012 ) and artificial intelligence (Pfeil et al., 2013 ; Benjamin et al., 2014 ; Merolla et al., 2014 ), and soon will play a critical role in robotics. Animals' mobility shows that neuron-based control is effective, and several groups have already developed neural-inspired controllers that could benefit from the low power and parallel computing of neuromorphic hardware (Ayers et al., 2010 ; Floreano et al., 2014 ; Dasgupta et al., 2015 ; Hunt et al., 2017 ; Szczecinski and Quinn, 2017b ; Dürr et al., 2019 ). However, to apply these neuromorphic chips to robotics, these controllers must be converted into a chip's specific neural model, which may not be trivial. All chips use a variant of the integrate-and-fire model (Brunel and van Rossum, 2007 ). Toward this goal, we have developed methods for applying our functional subnetwork approach (FSA) for designing non-spiking recurrent neural networks (Szczecinski et al., 2017b ) to the specific generalized integrate-and-fire (GLIF) model used by Intel's Loihi chip (Mihalaş and Niebur, 2009 ; Davies et al., 2018 ). We will show that these models (i.e., non-spiking and GLIF) have several parallels that enable a network designer to map between them. The details of this comparison are listed at the end of the Introduction. Setting parameter values in dynamic neural networks can be extremely difficult. Even when network architecture is created directly from animal architecture, parameters cannot be practically measured across all neurons and synapses, and there may be thousands of parameters that dynamically interact. Therefore, these parameter values must be set by the modeler such that the network produces the desired behavior. Some methods and tools have been developed to assist neural designers when mapping network behavior to a desired output, for example, the Neuroengineering Framework and its browser-based design program, Nengo (Eliasmith and Anderson, 2002 ; Maass and Markram, 2004 ; Bekolay et al., 2014 ). These methods seek to build populations of neurons, whose average activity (i.e., spiking frequency) encodes a value of interest. Each population can then interact with others to perform specific operations, such as arithmetic or calculus. This technique is very powerful, enabling the construction of brain-scale networks (Eliasmith, 2013 ). However, one drawback is that within each population, the connectivity is random, and may not provide insight into how biological networks are structured at small scales. This method is also not ideal for modeling networks with relatively few neurons, such as those that have been described in the locomotion networks of animals (e.g., Bueschges et al., 1994 ; Sauer et al., 1996 ; Berg et al., 2015 ). As an alternative to this technique we have developed methods for explicitly computing neural and synaptic parameter values for non-spiking dynamical neural networks that perform arithmetic and calculus (Szczecinski et al., 2017b ). Such networks are also called “recurrent neural networks,” because each neuron's instantaneous state is a function of its own history, producing a form of self-feedback. While such recurrent dynamics can make it difficult to tune networks, such continuous dynamical neural models enable direct analysis of a network's eigenvalues, equilibrium points, and therefore, individual neuron behavior in response to specific inputs (Szczecinski et al., 2017a , b ). Such analysis can be difficult to perform on spiking networks, but is particularly important in robotics, in which engineers seek to guarantee a robot's stability and the controller's robustness to parameter changes or sensor noise. The resulting networks are sparse and based on known anatomy, similar to related robotic controllers composed of analog very large scale integration (VLSI) circuits or efficient, discrete-time neuron models (Ayers and Crisman, 1993 ; Ayers et al., 2010 ). Such non-spiking, continuous-valued models theoretically have the same activation dynamics as the average spiking frequency of a population of spiking neurons, all of whom receive the same (noisy) inputs (Wilson and Cowan, 1972 ). The current manuscript explores this assertion by identifying relationships between the parameter values in the non-spiking model used in our previous work (Szczecinski et al., 2017a , b ) and those in a GLIF model (Mihalaş and Niebur, 2009 ; Davies et al., 2018 ). Applying the functional subnetwork approach to spiking networks has three primary benefits: First, it enables rapid and direct assembly of spiking networks that have predictable performance; second, it enhances neural robot controllers with richer dynamics than recurrent neural networks; and third, it is a tool for implementing neural controllers on neuromorphic hardware for robots. In this manuscript, we demonstrate the first of these benefits by extending our previously-developed design tools for non-spiking models (Szczecinski et al., 2017b ) to spiking models. We present three “parallels” between the non-spiking and spiking models that enable the extension of our non-spiking network design techniques to spiking networks. This extension includes reducing the impact of non-linear relationships within a network. We derive three parallels between these models: P1. The steady-state spiking frequency of a spiking neuron is parallel to the steady-state depolarization of a non-spiking neuron because each is proportional to the current applied to the neuron. We refer to both of these quantities as the “activation” of the neuron. The activation of each model can be related to the other via model parameters, and specific parameter values increase the similarity between spiking frequency and non-spiking depolarization. P2. The instantaneous spiking frequency of a spiking neuron is parallel to the instantaneous depolarization of a non-spiking neuron. Each exhibits a transient response when stimulated. The decay rate of each model can be related to the other via model parameters, and specific parameter values increase the similarity between the spiking frequency time constant and the non-spiking membrane time constant. P3. The time-averaged conductance of a spiking synapse is parallel to the conductance of a non-spiking synapse because each is proportional to the activation of the pre-synaptic neuron. Both spiking and non-spiking synapses can be designed to implement a given “gain” value, i.e., the ratio between the post-synaptic (i.e., receiving) and pre-synaptic (i.e., sending) neurons' activations. Specific parameter values increase the similarity between the time-averaged spiking synapse conductance and the non-spiking synapse conductance. This manuscript is organized as follows. The methods in section 2 present the non-spiking and GLIF models and compute fundamental quantities for each, including equilibrium points and useful relationships between parameter values and variables. We use these expressions to extend our FSA for designing non-spiking networks to spiking models. The results in section 3 demonstrate parallels P1-P3 and leverage them into a sequential process for designing a spiking pathway. In section 4, the results from section 3 are applied to a neuromuscular model of a stretch reflex, and the resulting motion of the models is compared. Finally, the discussion in section 5 summarizes the work, explains how neurobiologists and roboticists can apply this work to their research, and proposes future work. To aid the reader, variable names are defined in Table 1 . Table 1 List of variables and descriptions. Variable Description NON-SPIKING U ¯ Membrane voltage, state variable C - m e m Membrane capacitance, constant parameter G mem Membrane conductance/leak conductance, constant parameter I app Applied current, input variable G ¯ max Maximum non-spiking synaptic conductance, constant parameter G s - Instantaneous non-spiking synaptic conductance, piecewise linear function of the pre-synaptic neuron's voltage E s - Non-spiking synaptic reversal potential, constant parameter SPIKING U Membrane voltage, state variable θ Spiking threshold, state variable C mem Membrane capacitance, constant parameter G mem Membrane conductance/leak conductance, constant parameter I app Applied current, input variable θ 0 Initial spiking threshold, constant parameter τ θ Spiking threshold time constant, constant parameter m Proportionality constant that determines the change in θ relative to U , constant parameter G max Maximum spiking synaptic conductance, constant parameter G s Instantaneous synaptic conductance, state variable τ s Synaptic time constant, constant parameter E s Spiking synaptic reversal potential, constant parameter\n\n3.5. Spiking Pathways May Introduce Unwanted Artifacts We have shown that we can apply our non-spiking signal pathway design rules to spiking networks by treating many values as their average over time. However, there are some unintended artifacts of this approach that reduce performance. The first is an intermittent drop in a post-synaptic neuron's spiking frequency ( Figure 7A ). The way the system is tuned, the post-synaptic neuron should fire every time that the pre-synaptic neuron fires. Occasionally, however, the post-synaptic spike time is delayed relative to the synaptic current. This manifests as a temporary drop in the instantaneous spiking frequency. The prediction of the average spiking frequency is intermittently incorrect as a result. Figure 7 (A) If a pre-synaptic spike does not elicit a spike in the post-synaptic neuron, then its spiking frequency will intermittently drop. (B) The spiking frequency of a post-synaptic neuron whose threshold hyperpolarizes in response to membrane voltage depolarization may oscillate at low frequencies. In both columns, spikes are indicated by violet asterisks (no cosmetic “spikes” are plotted). The second artifact is a periodic instantaneous spiking frequency (about the predicted spiking frequency, Figure 7B ). This occurs when m < 0, and thus the spiking threshold θ decreases when the neuron is depolarized, making it easier for the neuron to spike. Particularly at low stimulus frequencies, the neuron may exhibit a periodic spiking frequency ( Figure 7B ). However, one can see that the prediction of the average spiking frequency remains accurate. In the following section, we show that both of these unwanted artifacts can be eliminated by adding more neurons and synapses to the network.",
"discussion": "5. Discussion 5.1. Summary The analysis and numerical results in this manuscript show how continuous, non-spiking leaky-integrator neural dynamics can approximate the dynamics of a population of identical GLIF spiking neurons with randomized interconnections. The parallels in encoding and information transfer manifest through three analogs: (1) A spiking population's mean spiking frequency is analogous to the membrane voltage of a leaky integrator, (2) the transient spiking frequency of a spiking neuron can mimic a non-spiking neuron's transient voltage, and (3) a spiking synapse's average conductance is proportional to the pre-synaptic neuron's spiking frequency, analogous to how a non-spiking synapse's conductance is proportional to the pre-synaptic neuron's membrane depolarization. Since the dynamics are approximately similar, a network built from either type can be used to encode and transfer information in an equivalent manner. Therefore, networks of either type can have similar overall behavior, and either type might effectively be used to model and understand the nervous system. The parallels between non-spiking neural models and models of populations of spiking neurons have been known for quite some time (Wilson and Cowan, 1972 ). However, the process and tools needed to set parameters within networks of these models to achieve desired and/or equivalent behavior have been lacking. In an attempt to apply the classical analysis from (Wilson and Cowan, 1972 ) to the practical application of programming neuromorphic hardware, we extended our functional subnetwork approach (FSA) to tuning networks of GLIF neurons and synapses. This extension enables one to tune control systems built from either non-spiking nodes or populations of spiking neurons. We presented a step-by-step method for tuning practical networks of populations of spiking neurons and synapses. We provided examples showing how increasing the number of neurons makes data transmission more ideal (i.e., match the expected population average activity). Finally, we provided a practical example of how such a method can be used to tune a neuromechanical model for control, and how the non-spiking and spiking implementations compare and contrast. Despite the similarities between the models' activation in response to inputs (section 3.1) and how this activation maps to synaptic conductivity (section 3.3), we observe differences in the models' transient responses. A spiking neuron's smoothed (i.e., time-averaged) transient spiking frequency is a good match for a non-spiking neuron's transient membrane voltage if the spiking neuron is not initially spiking ( Figures 6 , 10 ). This is because the spiking threshold must change from the initial value θ 0 to the steady state value when a spike occurs, θ ∞ * . However, the spike-time threshold θ ∞ * is insensitive to a neuron's input current (and therefore the neuron's spiking frequency), meaning that the amplitude of the transient is very small once a neuron is spiking. This is not true for the non-spiking model, whose transient amplitude depends strongly on the input current. Therefore, the response properties of networks tuned to have long, exaggerated transient responses are a good match initially, but not once a neuron is already spiking ( Figure 12 ). In all other respects, however, these models can be tuned to produce the same responses. 5.2. Expanding These Methods The analysis in this manuscript primarily provides a framework for designing rate-coding networks, based on their steady-state spiking frequency. We have already shown how steady-state analysis contributes to designing arithmetic and dynamic (i.e., differentiation and integration over time) networks (Szczecinski et al., 2017b ). However, not all neural computation is rate-coding, meaning that additional techniques are needed to expand this work to engineer direct encoding of other signals to produce more sophisticated neural behaviors. For instance, a non-spiking model captures class II excitability, in which a neuron's spiking frequency exponentially approaches a steady-state spiking frequency when subjected to a step input (Mihalaş and Niebur, 2009 ). However, the GLIF spiking neuron model used in this work can also exhibit class I excitability in which there is no transient spiking frequency, a behavior that a non-spiking neuron cannot replicate. Such a response might be useful in a reflex pathway, in which delayed sensory feedback may destabilize the system. In addition, a spiking neuron can exhibit phasic excitability, in which its spiking frequency starts high, but decays to 0 during a step input (Mihalaş and Niebur, 2009 , see also Figure 3C ). In the future, we plan to investigate whether such phasic responses could be used to replace the differentiation network from the non-spiking FSA approach (Szczecinski et al., 2017b ) with a single neuron, a technique we have used in the past, but did not characterize thoroughly (Szczecinski et al., 2014 ). Exploiting single-neuron properties in this way could enable designers to pack more computational capability into chips with limited (albeit large) network sizes, such as Loihi (Davies et al., 2018 ). However, creating small networks that seek to exploit single-neuron properties may reduce the accuracy of encoding, decoding, and other operations within the network. Alternative systems, such as the Neuroengineering Framework (NEF) (Eliasmith and Anderson, 2002 ), rely on nodes (ensembles) with many neurons (on the order of 10–1,000) to accurately encode information into the network. What makes NEF so powerful is the relatively hands-off design process, wherein the user specifies the intended function and number of neurons per node, and then NEF learns the intra-node parameter values necessary to perform that function (Eliasmith et al., 2012 ; Bekolay et al., 2014 ). This approach is much less onerous to the designer than the FSA, which requires explicit tuning of parameters for encoding, decoding, and other functions. We anticipate that these two approaches may complement one another, wherein the direct network tuning accomplished by the FSA could be used to initialize tuning within the NEF. In our experience, the FSA can be used to select initial network parameters that aid subsequent parameter optimization (Pickard et al., 2020 ). As a next step, we plan to compare the accuracy and efficiency of networks tuned via the FSA with those tuned via the NEF. We expect that the NEF may achieve arbitrarily high accuracy, but possibly at a computational cost. The FSA could be used to initialize learning networks in a less random way, requiring fewer neurons per node and less time to train. 5.3. When to Use Spiking or Non-spiking Neurons A question that follows from this work is that if non-spiking and spiking neuron dynamics have many parallels, how does a modeler choose to use one or the other type? We believe that both types are useful in different contexts, depending on the knowledge available about the system to be modeled, the research question addressed by the model, and the real-world application of the network (e.g., in robotics). A natural choice is to model spiking neurons in the nervous system with spiking models, and non-spiking neurons with non-spiking models. However, biomechanical constraints determine whether animals use spiking or non-spiking neurons. Specifically, action potentials can be transmitted over long distances, whereas graded (i.e., non-spiking) signals tend to dissipate over even short distances. This may be why many non-spiking neurons have been identified in insects and other small animals, particularly for integrating sensory information (Burrows et al., 1988 ; Sauer et al., 1996 ); they have less room in their bodies to house networks of many spiking neurons, and their small bodies do not require them to transmit information over long distances. No matter why animals have spiking or non-spiking neurons, a computer model does not share these biochemical constraints, so it is worth deciding how to model networks based on the computational hardware available. One purely technological motivation to use spiking models is for model implementation on neuromorphic hardware. Chips, such as Loihi (Davies et al., 2018 ), SpiNNaker (Khan et al., 2008 ), TrueNorth (Merolla et al., 2014 ), and others (Pfeil et al., 2013 ; Gehlhaar, 2014 ; Ionica and Gregg, 2015 ) use non-traditional architecture to simulate hundreds of thousands of spiking neurons and hundreds of millions of synapses in real-time while using on the order of one watt of power. Neuromorphic computers tend to use spiking models because they have less communication overhead than non-spiking networks. For spiking networks, communication can be binary (1 bit per spike) and only must occur after a spike occurs, at a maximum of 200–300 Hz (Gerstner et al., 1997 ; Carter and Bean, 2010 ) but more often below 1–2 Hz (Kerr et al., 2005 ). In contrast, non-spiking synapses need to be updated during every simulation step, because they depend on the pre-synaptic neuron's continuous membrane voltage. Such a requirement significantly increases overhead relative to spiking models. Spiking neurons and synapses are also attractive because they enable the use of spike timing dependent plasticity (STDP) learning techniques (Gerstner et al., 1996 ; Markram et al., 1997 , 2012 ), a powerful class of machine learning tools. These methods have been applied to many stimulus-recognition tasks, such as hearing, speech, and vision. They measure the coincidence of incoming spikes to increase or decrease the strength of synapses in the network, and thus rely on spiking models. These networks are typically classified as “self-organizing,” meaning that they initially have no structure, but develop their own structure as connections are pruned due to disuse. However, many parts of the nervous system have exquisite structure, and lend themselves to being directly modeled, structure, and all. Thus, we believe that the methods in this manuscript may serve to produce baseline parameter values for a highly structured network, which may then use spike-based mechanisms to tune itself over time, a technique like that utilized in Nengo (Bekolay et al., 2014 ). In such cases, populations of spiking neurons would be preferable to non-spiking population models, but could be initialized using the tuning rules presented in this manuscript. Additionally, spiking neurons and synapses may be beneficial because even a simple model like that used in this work can produce a wide variety of behaviors and responses (Mihalaş and Niebur, 2009 ). For example, setting m > 0 can produce spike frequency adaptation and phasic spiking, which are known to be critical for filtering sensory feedback in locomotory systems (Mamiya et al., 2018 ; Zill et al., 2018 ). Such rate-sensitive responses can be produced by small networks of non-spiking neurons and synapses (Szczecinski et al., 2017b ), but force the modeler to use more neurons than may be necessary. Therefore, if the modeler knows that single neurons in their model system generate more complex responses than class I or II excitability, then spiking models should be utilized. However, if the neuron responses in the network are simple, then the model could be implemented as a network of non-spiking neurons instead. We believe non-spiking networks may be particularly beneficial if a model is not meant to run on specialized neuromorphic hardware. Simulating the membrane voltage of each spiking neuron in a population requires storing and updating orders of magnitude more states than simply using a non-spiking node to represent the mean activity of the population. In addition, throughout this study, we observed that the timing of spikes was sensitive to the simulation step used (i.e., spikes cannot happen between time steps), and simulations only closely matched our analytic predictions as the time step became very small. We also tested adaptive stepping algorithms (Matlab's ode45 and ode15s solvers); however, the discontinuous nature of spikes required the use of event functions that halt simulation whenever a spike occurs, complicating the code. In general, we found that it took longer to simulate the dynamics of a spiking network relative to those of a non-spiking network. These reasons motivate implementing networks on traditional computers with non-spiking models. Finally, non-spiking neurons may contribute to models by representing neuromodulatory neurons that cause large-scale changes to network behavior. In Szczecinski et al. ( 2017b ), we not only designed “signal transmission” pathways as in this work, but also “signal modulation” pathways, in which one neuron could modify the gain of another neuron's response to its synaptic inputs. When we tested signal modulation pathways built from spiking neurons and synapses, the results were poor (data not shown). Due to the discrete nature of spikes, modulation was inconsistent, leading to unpredictably varying firing frequencies from the “modulated” neuron. However, it may be advantageous to instead construct networks in which signals are transmitted via spiking neurons, but modulated via non-spiking neurons. These non-spiking neurons in effect model neuromodulatory neurons, whose voltage reflects the concentration of a neuromodulator in the network, and whose non-spiking synapses represent the activation of receptors sensitive to that particular neuromodulator. Such pathways may change the resting potential, membrane conductance, and time constant of other neurons throughout the network (for a review, see Miles and Sillar, 2011 ). Such non-spiking neurons would have long time constants, enabling them to modify network performance on much longer timescales than that of a single spike. Such neurons could receive either descending input from the brain, or ascending input from local sensory neurons. The result would be a model that can modulate its control system based on exteroception and interoception, and exhibit truly adaptive, context-dependent behavior."
} | 6,955 |
25613225 | PMC4303876 | pmc | 8,872 | {
"abstract": "The forest timberline responds quickly and markedly to climate changes, rendering it a ready indicator. Climate warming has caused an upshift of the timberline worldwide. However, the impact on belowground ecosystem and biogeochemical cycles remain elusive. To understand soil microbial ecology of the timberline, we analyzed microbial communities via 16s rRNA Illumina sequencing, a microarray-based tool named GeoChip 4.0 and a random matrix theory-based association network approach. We selected 24 sampling sites at two vegetation belts forming the timberline of Shennongjia Mountain in Hubei Province of China, a region with extraordinarily rich biodiversity. We found that temperature, among all of measured environmental parameters, showed the most significant and extensive linkages with microbial biomass, microbial diversity and composition at both taxonomic and functional gene levels, and microbial association network. Therefore, temperature was the best predictor for microbial community variations in the timberline. Furthermore, abundances of nitrogen cycle and phosphorus cycle genes were concomitant with NH 4 + -N, NO 3 − -N and total phosphorus, offering tangible clues to the underlying mechanisms of soil biogeochemical cycles. As the first glimpse at both taxonomic and functional compositions of soil microbial community of the timberline, our findings have major implications for predicting consequences of future timberline upshift.",
"discussion": "Discussion Investigating soil microbial communities is important for understanding microbe-mediated biogeochemical cycle and ecosystem functioning 29 . A main goal of this study is to identify the environmental factor(s) that best explain microbial community variation. It has been proposed that above-ground plant communities drive below-ground microbial diversity 30 . Previous studies have indicated that plants affect microbial communities via rhizodeposits and exudates and provide organic C and N to the soil microbial community 31 32 . The different vegetation types could select for distinct soil microorganisms due to formation of a variety of microhabitats that support a diverse collection of species 33 34 . Thus, we examined whether soil microbial functional diversity was affected by plant diversity. However, significantly higher microbial taxonomic and functional diversity were detected in the shrubland than the coniferous forest ( Table 1 ). Considering higher plant diversity, species number and biomass in the coniferous forest, this finding suggested that plant diversity might not be the determinant of microbial diversity, which was consistent with several previous studies showing that soil microbial community was uncorrelated with plant diversity 35 36 37 . The inconsistency between plant diversity and microbes could be ascribed to different ecophysiological traits of plant species, which would exert strong effects on soil biological properties 38 39 . Thus, the plant community composition is more likely to affect soil microbial community composition than plant diversity 35 40 41 . In addition, this could also be explained by a chance effect that a keystone species present in the plant community results in greater effects on soil processing than the contribution of the total plant diversity 42 . Notably, the dominant species were different between the shrubland and the coniferous forest. The plant litter of Abies fargesii in the coniferous forest with relatively higher C/N ratios provides poor nutrients for microbial growth 43 , and this may have a tendency to reduce microbial diversity and activity. On the contrary, Fargesia murielae , one of the dominate plants in the shrubland, would buffer temperature extremes and modify the local microenvironments and soil quality 44 , which might be helpful for microbial survival. Temperature has been shown to be the major factor in timberline formation 9 . Here we showed that soil temperature, among all of measured environmental parameters, showed the most significant and extensive linkages with microbial biomass, microbial diversity and composition at both taxonomic and functional gene levels. Correlation networking analysis also showed that temperature was an influential environmental parameter. Therefore, temperature was the best predictor for microbial community variations at the timberline. Our study showed that the temperature was significantly lower in coniferous forest, which was consistent with previous studies showing that the timberline forest typically had a colder temperature than its adjacent vegetation at higher altitudes 2 45 46 , since the closed forest canopy protects the soil from sunrays and results in a lower temperature. Temperature can affect metabolic rates via the kinetics of the biological process, famously coined as “the Red Queen runs faster when she is hot”, and consequently contributes to organism differentiation and diversification 47 48 . However, it is also possible that temperature indirectly affects microbial community by controlling soil nutrient availability, which further affects soil microbial community 49 . A comparison among the shrubland, grassland, farmland and reforested land in an adjacent region to our study sites showed that the shrubland had the highest soil quality value 50 . Similar studies on alpine ecosystems showed that treeless soil at higher altitudes had more labile C, greater microbial activity than soil of closed forest canopy 51 52 . Consistently, higher soil nutrient contents of dissolved organic carbon, available nitrogen, total phosphorus and total sulfur in the shrubland provided a suitable environment for microbial communities, as indicated by higher microbial biomass and diversity. CCA showed that microbial community compositions were correlated with δ 15 N, total phosphorus, total sulfur; Al and Fe ( Fig. 3 ), in addition to correlations between microbial biomass, functional diversity and NH 4 + -N ( Supplementary Fig. S2 ). Among them, N appeared to be a key factor affecting microbial community composition, diversity and productivity in many N-limited terrestrial ecosystems 53 . The correlation between microbial biomass and NH 4 + -N was consistent with precious study showing that N limitation was an influential factor in affecting microbial community 54 . Meanwhile, our study showed that low microbial diversities were detected in sampling sites with high C/N ratios, which was consistent with the finding that microbial community structure and activity were negatively coupled to the soil C/N ratio 43 . This indicated that an upward expansion of forest resulting from global warming would lead to changes in soil organic quality and the activities of the underground communities. A parallel 16s rRNA and shotgun sequencing study on tallgrass prairies showed a strong positive correlation between taxonomic and functional gene diversity 55 , suggesting a low degree of functional redundancy. By contrast, we did not detect strong correlation between overall taxonomic and functional gene diversity, which might be attributed to the differences in techniques or ecosystems. However, there was a positive correlation between bacterial amoA gene abundance and ammonium oxidizers ( Fig. 5C ), a functionally narrow group utilizing the conversion of ammonia into nitrite as the sole energy source 22 . It can thus be predicted that reduction in ammonium oxidizer species is associated with decreases in the functional potential of bacterial ammonium oxidization contained within these soil communities. Correlation networking analyses show important details of community assembly rules reflecting ecological processes such as cooperation, competition, habitat filtering and historical effects, and can represent mathematical interaction/coupling among different populations and/or functional genes that regulate system functions 56 . The shorter path length and higher average connectivity and transitivity of coniferous forest networks suggested that microbial interaction/coupling was high. The high interaction might be ascribed to certain deterministic processes such as habitat filtering, reducing the spread of trait values and reflecting shared ecological tolerances 57 , which was consistent with the observations of more positive interactions in the coniferous forest networks. Modularity helps control disturbances by compartmentalizing social-ecological systems 58 , the higher modularity of the shrubland network indicated that the microbial system would be more resistant to changes, both at the taxonomic and functional level, which might be ascribed to higher microbial diversity of shrubland ( Supplementary Table S1 ). Profiling microbial communities in the habitats bordering the timberline is crucial for predicting the dynamics of microbial community changes and ecosystem functioning, since the trend of timberline upshift resulting from climate warming is likely to continue. Our results showed that temperature was the best predictor for microbial community formation and there were clear linkages between microbial functional potentials and soil biogeochemistry cycles of the timberline. Based on these results, we predict that timberline upshift resulting from global warming would cause distinct changes in microbial communities and soil C and N pools. However, the general trend of timberline shift is modified by local, regional and temporal variations 59 . It would be interesting to expand the observations in this study to other timberlines or conduct time-series experiments in order to test the generality of the observations."
} | 2,405 |
37701813 | PMC10494312 | pmc | 8,875 | {
"abstract": "Summary Microbes shape their habitats by consuming resources and producing a diverse array of chemicals that can serve as public goods. Despite the risk of exploitation by cheaters, genes encoding sharable molecules like siderophores are widely found in nature, prompting investigations into the mechanisms that allow producers to resist invasion by cheaters. In this work, we presented the chemostat-typed “resource partition model” to demonstrate that dividing the iron resource between private and public siderophores can promote stable or dynamic coexistence between producers and cheaters in a well-mixed environment. Moreover, our analysis shows that when microbes not only consume but also produce resources, chemical innovation leads to stability criteria that differ from those of classical consumer resource models, resulting in more complex dynamics. Our work sheds light on the role of chemical innovations in microbial communities and the potential for resource partition to facilitate dynamical coexistence between cooperative and cheating organisms.",
"introduction": "Introduction Theoretical ecology has long been fascinated by the intricate feedback between microbes and their microhabitats. 1 Central to this ecological feedback is the “chemical environment,” which includes the substances in the local environment that directly impact and are influenced by microbes. 2 , 3 Examples of such substances include essential “resources,” such as carbon, nitrogen, and oxygen, which are supplied to the extracellular environment and consumed by microorganisms to promote their own growth. 4 , 5 , 6 In theoretical ecology, the consumer-resource model has been employed for decades in studying this feedback, where the dimension of the chemical space has been demonstrated to be of key importance. 7 For instance, MacArthur et al. validated the competitive exclusion principle (CEP), which asserts that the number of stably coexisting species cannot exceed the number of resources they are competing for. 8 , 9 Tilman et al. demonstrated that in a chemical space of dimension two, the zero-net growth isoclines and the consumption vector dictate the outcome of competition. 10 Moreover, in a multi-species resource competition system with a minimal chemical dimension of three, Huisman et al. established that oscillatory and even chaotic dynamics could arise, allowing for dynamic coexistence that surpasses the upper limits of CEP. 11 , 12 Actually, despite numerous research efforts aimed at reconciling the conflict between CEP and the apparent biodiversity in nature, most of which include additional factors such as spatial or temporal heterogeneity, 13 , 14 , 15 , 16 a definitive answer has yet to be achieved. 17 , 18 Microorganisms have developed ways to overcome the upper limit of competitive exclusion. By leaking metabolic byproducts or actively producing and secreting secondary metabolites, cells can expand the chemical diversity of their habitats. 19 , 20 Such chemical innovations, in which microorganisms generate new chemical dimensions previously nonexistent in the environment, have profound implications for microbial ecology. 21 , 22 For example, cross-feeding through metabolic byproducts has been suggested to promote stable coexistence even with a single carbon source. 23 Meanwhile, the vast range of bacteriocins and antibiotics contributes to the chemical warfare among microorganisms. 24 , 25 , 26 Overall, microbial chemical innovations increase the chemical dimensions in the environment, raising the upper bounds of competitive exclusion. However, the active production of secondary metabolites often requires microbial cooperation, such as various quorum sensing molecules that coordinate collective decision-making, 27 or niche-construction molecules such as siderophores or secretory proteases that serve as “public goods” to enhance the microenvironment for the entire population, 28 , 29 , 30 From a game theory perspective, secondary metabolites produced for cooperation bring “the tragedy of the commons” 31 : If a cheater strain can exploit public goods without contributing to their production, how can the cooperating strains that provide these extra chemical dimensions remain competitive? Big questions in biology often have specific solutions in particular systems. The siderophore system, a diverse family of microbial secondary metabolites used for iron-scavenging, provides an excellent model to investigate the game between cooperators and cheaters and the ecological effect of self-generated chemical dimension. 32 , 33 Iron is one of the most limiting resources for microbes, required for crucial processes like energy metabolism and DNA synthesis. 34 However, the bioavailable iron concentration is orders of magnitude lower than what is required for normal microbial growth in most environments. 4 , 35 Most microorganisms acquire iron via siderophores, a type of small molecule with a strong iron-binding affinity. 36 Siderophores are secreted into the extracellular environment to chelate iron, and the iron-siderophore complex forms a new chemical dimension only absorbed by cells via corresponding membrane receptors. 37 However, siderophore production carries a metabolic cost that slows microbial growth, making siderophores a costly public good that may result in the “tragedy of the commons” in game theory: cheaters that do not invest in siderophore production but still benefit from the presence of siderophores can have a selective advantage over producers, leading to a decrease in overall siderophore production and potentially reducing the availability of iron for all microbes in the community. 38 , 39 Many theories have been proposed to explain the prevalence of siderophore synthesis pathways, 40 , 41 most of which entail spatial factors that facilitate kin selection or group selection. 42 , 43 , 44 Despite this, microorganisms actively produce siderophores of various types in well-mixing environments such as the ocean. 45 , 46 Recent experiments have suggested that private siderophores, which are accessible only to their producers, may be crucial for the survival of siderophore-producing microorganisms. 47 In many organisms, certain modifications during the multi-step process of siderophore production can transform the secretory molecules into a membrane-attached form. 47 , 48 , 49 This privatization of siderophores may shift the game between cooperators and cheaters toward the “snowdrift” scenario, where producers can benefit from their membrane-attached siderophores when diffusible siderophores are scarce. According to evolutionary game theory, the snowdrift scenario allows both cooperation and defection to persist at a stable equilibrium through negative frequency-dependent selection. 50 However, membrane-attached siderophores have a considerably lower diffusion radius in iron scavenging. 29 , 37 The marginal benefits conferred by membrane-attached siderophores have not been quantitatively assessed by ecological models, and it is unclear whether they provide cooperators with a sufficient advantage over cheaters. Many questions remain to be systematically explored through mathematical formulation, such as whether and how cooperators and cheaters can coexist, how different allocation strategies between membrane-attached and publicly shareable siderophores change the system dynamics, and which strategies might be optimal considering within- and between-species interactions. Furthermore, can siderophore-mediated iron competition help address a profound question in ecology: whether self-generated resource dimensions fundamentally change the properties of mathematical models about resource competition? In this work, we employed a chemostat-type “resource partition model” to investigate the iron competition mediated by siderophores, accounting for the allocation of limited cellular resources between biomass accumulation and the production of membrane-attached (private) and publicly sharable (public) siderophores. By incorporating private siderophores into the resource allocation strategies, new classes of strategies emerge, including “partial cooperators,” who produce both types of siderophores, and “self-suppliers,” who only produce membrane-attached siderophores. Coexistence between the partial cooperators and the pure cheaters is made possible by private siderophores, by providing partial cooperators with a growth advantage over cheaters when public siderophores are scares. Notably, such coexistence between two species can occur via dynamical oscillation. Further stability analysis revealed that creating new resource dimensions by species, i.e., siderophore production in our model, modifies the stability criteria of the classical consumer-resource model, allowing for more diverse dynamics than the classical model. In summary, our model of iron competition suggests that the division of the iron resource by siderophores increases the upper bound of coexistence, and the privatization of siderophores in cellular strategies provides advantages to partial cooperators to achieve coexistence. Our analysis of the stability criteria revealed that microbial niche construction by creating new resource dimensions adds more dynamics than the classical model, which may contribute to the diversity and complexity of the microbial world.",
"discussion": "Discussion Microbial diversity has been a long-standing area of interest in microbial ecology. 3 While the range of species may be constrained by chemical parameters, microorganisms can expand this limit by generating chemical diversity in their microhabitat. In this study, we constructed a model of siderophore-mediated iron competition to address two broader questions: first, how can cooperators responsible for chemical diversity withstand extinction triggered by cheaters, and second, what are the ecological implications of microbes creating new chemical dimensions? Addressing the first question, we proposed the privatization of siderophores as a “game changer” that provides partial cooperators an advantage over pure cheaters. By avoiding diffusion losses and cheater exploitation, this privatization strategy reflects a microbial instance of the “snowdrift” scenario in game theory. 50 While cheaters reap the benefits of public goods without contributing, cooperators prioritize access to the goods to gain advantages when public siderophores are scarce. 55 This instance shows that negative frequency-dependent selection in microbial communities can be mediated by siderophore production and consumption. Analytically, we derived that the necessary condition for coexistence is that the cooperators who invest more in public siderophores also invest more in private siderophores. Answering the second question led us to develop an updated model for microbial interactions called the “resource partition model.” This model incorporates the idea that organisms not only consume but also actively produce resources. Furthermore, the public siderophore is not considered a “resource” until it forms a complex with ferric iron, which is the actual resource that microorganisms require. This adds another layer of meaning to the term “resource partition model”: externally supplied resources (iron) and microbe-generated resources (siderophore) interact to form the actual resources (iron-associated siderophores) that are taken up by the microorganisms. Chemical innovation and resource partition Microbes interact by influencing their shared environment. 56 Classical ecological models have mostly emphasized the “consumption” aspect of such influences. 57 , 58 With the rapid development of microbiology, it has become increasingly clear that bacteria have enormous potential to introduce new chemicals into their microhabitats, providing chemical innovations that shape their community interactions. 59 Siderophores are just one example of the many secondary metabolites that microbes actively produce for their own benefit. Other examples include antibiotics, bacteriocins, signaling molecules, and even bacterial vesicles, all of which can be considered as “chemical dimensions” generated by microbes. 1 , 26 A general ecological framework for such “self-generated dimensions” has yet to be established. The conventional consumer resource model provides intuitive coexistence criteria where species should preferentially consume the resource that limits their growth, and the consumption vectors clearly segregate supply conditions into zones of the same stability. 10 , 60 In contrast, in this “resource partition model,” due to the reversed direction of consumption vectors, there are additional parameter areas where the Routh-Hurwitz (RH) criterion can change signs. Consequently, in regions of steady-state equilibrium in the classical model, non-equilibrium dynamics become possible. Our analysis suggests that an ecosystem with “self-generated dimensions” tends to be more dynamic and complex. Oscillation usually bridges two distinct states and often plays critical roles in various biological systems. 61 Notably, the oscillation in our model differs from the oscillation in the work of Huisman et al., where the oscillation bridges the stable equilibrium and the chaos with higher-than-CEP dynamical coexistence. 11 , 12 In our model, the parameter region of oscillation locates between total extinction and stable coexistence between partial cooperators and cheaters. Oscillation here is more of a “danger zone”, indicating that the system is on the verge of collapse, similar to the early warning signatures of ecosystems. 62 Indeed, our continuous equations assume that organisms can recover from arbitrarily small values, but the troughs of oscillatory dynamics increase the probability of stochastic extinction in natural systems. 63 , 64 Recent studies in game theory have detected oscillations when the environment is made explicit, showing that oscillatory dynamics prevent the extinction of cooperators. 65 , 66 , 67 Given the specificity of oscillatory dynamics in ecology and the propensity of resource partition models to enter the oscillation zone, it would be intriguing to explore the facilitative or destructive functions of oscillations in diverse ecological systems with chemical innovations. Resource partitioning and cross-feeding are two mechanisms that enable microbes to increase the resource dimensions in their environment, enabling higher species diversity. While both byproducts and siderophores play crucial roles in microbial ecosystems, there are notable differences between these two mechanisms. Cross-feeding occurs primarily via passive leakage of metabolic byproducts that are no longer utilized by producers, while resource partitioning mediated by siderophores involves active competition among microbes for binding sites on the limiting iron. This competitive process results in different siderophore-iron complexes that serve as actual iron resources, which can only be absorbed by microorganisms possessing corresponding receptors. Furthermore, resource partitioning involves an additional level of complexity, as it requires the formation of distinct siderophore-iron complexes based on factors such as affinity and concentration. These characteristics distinguish resource partitioning from cross-feeding, highlighting the need for a nuanced understanding of microbial interactions and the mechanisms that underlie them. Spatial factors, private goods, and adaption Spatial factors were commonly considered in theoretical models to explain the coexistence mechanisms of cheaters and cooperators. 43 , 68 , 69 Structured habitats, such as soil and the human body, are typical in natural habitats. Even in weakly structured environments like the ocean, many microbes can construct structured environments by adhering to surfaces or assembling into groups. 70 It has been suggested that coexistence can be promoted by highly structured habitats, where the sharing of public goods is spatially limited to local cooperation. 43 However, in models with weakly structured or highly mixed habitats, the dissipation of public goods and the costs associated with their production impose losses and burdens on cooperators, still leading to the “tragedy of the commons.” 71 In siderophore-mediated iron competition, privatization enables coexistence in homogeneous habitats. Although our model was specifically developed for certain species, such as marine microbes or actinomycetes, this public-private goods framework provides generalizable insights on the necessary conditions for coexistence: the organism that shares more must also preserve more. Furthermore, our results suggested that the partial cooperator might serve as a universal strategy of microbes, as it facilitates coexistence among different species. Recently, studies have discussed the private properties of “public goods,” converting them into a continuous metric (partially privatized public goods). Niehus et al. found that the privatized siderophores evolved to be upregulated. 48 The work of Lerch et al. discovered that when two public goods exist, a species producing a completely public good can exclude non-producers by partially privatizing another public good, 72 and this ability is highly reliant on nutrient supply, underscoring the significance of environmental conditions for sustaining cooperation. Several studies have bridged the microbial habitat with their goods’ private and public properties. Kummerli et al. revealed that the more structured the habitat, the more microbes utilize siderophores with high diffusibility. 73 Similarly, Garcia et al. found that bacteria living in more structured habitats encode more extracellular proteins with higher diffusibility. 74 These findings are consistent with the adaption of siderophore privatization proposed by our model and other works, 48 , 72 where the private portion is more advantageous in highly mixed environments. Another microbial adaptation in iron competition involves the regulation of resource allocation strategies in response to the environment. For example, quorum sensing (QS) can regulate the degree of privatization of siderophores, such as in marine Vibrio harveyi , where cells produce more membrane-attached siderophores at low cell density and switch into more soluble siderophores at higher cell density. 75 By balancing the private and public benefits, QS regulation can promote cooperation and enhance the efficiency of resource utilization in microbial communities. In the future, more research is needed to identify the regulatory mechanisms underlying microbial cooperation in natural environments. Besides the membrane-attached form, other forms of siderophore privatization, such as keeping siderophores intracellularly for relieving oxidative stress, have been proposed to confer cheater resistance. 76 We hypothesize that the division of the iron resource by siderophores provides a universal mechanism for “resource privatization”: numerous siderophores exist in the natural world, each with its specific receptors. 76 , 77 For a given type of siderophore, microorganisms with corresponding receptors can share it as a public good, whereas microbes without comparable receptors perceive it as inaccessible “private goods”. 29 Theoretical models suggest that the populations of different cooperators utilizing distinct siderophores can be regulated by their cheaters, similar to how parasites impose negative frequency selection. 78 Additionally, a model with multiple types of siderophores suggests that coexistence between cooperators and cheaters is possible if a “loner” uses a less efficient siderophore. 51 In the future, it will be exciting to investigate iron interactions in a more realistic biological setting and systematically examine the ecological consequences of microbial chemical innovation. It would be interesting to explore the role of microbial chemical innovation in community dynamics and stability in different habitats. The results of such investigations could lead to the development of novel strategies for understanding, controlling, and predicting microbial community dynamics, with important applications in areas such as agriculture, medicine, and environmental management. Limitations of the study Our research highlights the role of siderophores in promoting dynamical coexistence through resource partitioning, yet we acknowledge several limitations that warrant discussion. First, our model system was a highly simplified chemostat model, with constant supply, constant dilution, and well-mixed environment. These assumptions may not fully capture the complexity of natural systems, where the environment is heterogeneous in both space and time. Second, to obtain analytical results, we only discussed the two-dimensional case and did not explore the effects of high-dimensional scenarios or multiple types of siderophores. Third, for convenience, our model made certain assumptions about species parameters, and these assumptions may not hold in all ecological contexts. While our study offers valuable insights into the principles of siderophore-mediated interactions, further expansion and refinement of our model are necessary to fully understand more complex systems. We believe that future studies could benefit from incorporating additional ecological factors, such as spatial structure, varying resource availability, and diverse microbial communities."
} | 5,382 |
26044993 | PMC4486115 | pmc | 8,877 | {
"abstract": "Many promising hydrogen technologies utilising hydrogenase enzymes have been slowed by the fact that most hydrogenases are extremely sensitive to O 2 . Within the group 1 membrane-bound NiFe hydrogenase, naturally occurring tolerant enzymes do exist, and O 2 tolerance has been largely attributed to changes in iron–sulphur clusters coordinated by different numbers of cysteine residues in the enzyme’s small subunit. Indeed, previous work has provided a robust phylogenetic signature of O 2 tolerance [ 1 ], which when combined with new sequencing technologies makes bio prospecting in nature a far more viable endeavour. However, making sense of such a vast diversity is still challenging and could be simplified if known species with O 2 -tolerant enzymes were annotated with information on metabolism and natural environments. Here, we utilised a bioinformatics approach to compare O 2 -tolerant and sensitive membrane-bound NiFe hydrogenases from 177 bacterial species with fully sequenced genomes for differences in their taxonomy, O 2 requirements, and natural environment. Following this, we interrogated a metagenome from lacustrine surface sediment for novel hydrogenases via high-throughput shotgun DNA sequencing using the Illumina™ MiSeq platform. We found 44 new NiFe group 1 membrane-bound hydrogenase sequence fragments, five of which segregated with the tolerant group on the phylogenetic tree of the enzyme’s small subunit, and four with the large subunit, indicating de novo O 2 -tolerant protein sequences that could help engineer more efficient hydrogenases. Electronic supplementary material The online version of this article (doi:10.1007/s00284-015-0846-2) contains supplementary material, which is available to authorized users.",
"introduction": "Introduction The microbial world is a rich reserve of species and metabolic capabilities, which are being exploited to tackle grand challenges in energy, biotechnology and drug discovery. However, despite our knowledge of this vast diversity, we appear to rely on a few well-characterised organisms in biotechnological applications. Thus, there is a tendency when optimizing a biotechnology process to genetically engineer these organisms rather than seek out more efficient natural organisms. While genetic engineering can be used to enhance performance, it does not always lead to a superior enzyme as has been the case with O 2 tolerance and NiFe hydrogenases [ 2 ]. Hydrogenases are enzymes of great biotechnological interest because they catalyse the H 2 half-cell reaction [2H + + 2 e′ ⇔ H 2 ] that can be manipulated to produce hydrogen from sunlight [ 3 , 4 ] or sustainably use hydrogen in fuel cells driven by biocatalysts [ 5 , 6 ]. However, while all three types of hydrogenases, [Ni–Fe], [Fe–Fe] and [Fe-only], can catalyse this reaction for the vast majority of enzymes that we know about, this reaction is severely attenuated, or even irreversibly halted, in the presence of O 2 . Given that O 2 is either present or produced in every major reaction exploited by these proposed technologies, this intolerance is a major stumbling block that must be overcome [ 4 , 5 ]. Of all the different types of hydrogenases, the membrane-bound NiFe subtype (MBH) has a few well-characterised O 2 -tolerant members. One in particular, from the bacteria Ralstonia eutropha , is currently utilised in experimental enzymatic fuel cells (EFCs) [ 5 ]. However, this aero tolerance comes with the cost of both a total bias towards H 2 oxidation [ 7 – 10 ] and a decreased efficiency [ 11 ] when compared with O 2 sensitive hydrogenases (herein referred to as standard hydrogenases) [ 12 ]. As the ideal hydrogenase for many of these technologies would be both oxygen tolerant and efficient, significant work has gone into creating such an enzyme via genetic manipulation [ 7 , 8 , 13 , 14 ]. Although efforts thus far have met with little success [ 2 ], when combined with research on the structural chemistry [ 15 – 21 ] of MBHs, these studies have delivered significant insight into the mechanisms and gene sequence that are responsible for O 2 tolerance. Tolerance has been linked to specific amino acid residues mostly in the small subunit (α) and to a lesser extent in the large subunit (β) of the NiFe MBH. The bulk of the evidence shows that O 2 tolerance is a function of the proximal Fe–S cluster coordinated by key cysteine residues in the enzyme’s small subunit [ 1 , 8 , 14 ]. Specifically, O 2 -tolerant enzymes possess six cysteine residues (6C group) instead of the typical four conserved cysteine residues (4C group) found in standard hydrogenases. There is also emerging evidence from the enzyme’s large subunit that shows a histidine (H229) residue could serve to further stabilise the proximal cluster in the presence of O 2 [ 22 , 23 ]. Taken together, this information produces a reliable phylogenetic signature that can be used to identify potential de novo O 2 -tolerant enzymes from sequence alone. Indeed, a recent study identified at least 30 additional MBH sequences with the phylogenetic signature suggestive of O 2 tolerance from publically available fully sequenced genomes of microbes that can be cultured [ 1 ]. Cultured isolates represent a small fraction (<1 %) of the diversity of bacteria on the earth [ 24 ], so with the vagaries of evolution acting over billions of years on MBH in distantly related organisms it seems reasonable to speculate that there may be an untapped diversity of O 2 -tolerant enzymes in nature. However, sequence information will be of little use if not provided within the context of the organisms’ natural habitat as these organisms have evolved to exploit these diverse environments by fine-tuning their metabolisms to these conditions. Therefore, differences in natural habitats and subtleties of O 2 metabolisms of an organism could have shaped an enzyme that is more biotechnologically suitable than the ones currently under use. In this study, we interrogated publically available MBH sequences that had the phylogenetic signature of O 2 tolerance for differences in taxonomic group, natural environments and oxygen requirements. Following this, we used whole metagenome next generation sequencing to isolate novel O 2 -tolerant NiFe MBH sequence fragments from an environmental sample.",
"discussion": "Discussion The immediate need for efficient biotechnologies to solve current problems such as sustainable sources of energy has driven us to explore the microbial landscape for unique organisms and enzymes. However on its own, sequence information is not enough and needs to be paired with contextual information about the organism and the environment it inhabits. The observation that the 4C and 6C enzymes are not randomly distributed across phyla, natural environments or oxygen requirements suggests that these factors have influenced the evolution of these enzymes. Overall, there appears to be a shift from harsh, nutrient poor environments, to more anodyne, nutrient-rich environment. The 4C enzyme grouping mostly contained organisms that could metabolise sulphur and metals anaerobically. The 6C enzyme group contained organisms with aerobic metabolisms as well as many species that could either photosynthesize or metabolise nitrogen. This supports the idea of a redox up-shift via high-potential terminal electron acceptors within the 6C group [ 1 ]. An exception is the enzyme from R. eutropha , utilised in some experimental biotechnologies [ 3 – 5 ]. This organism exploits the relatively redox poor “ Knallgas ” reaction despite being part of the 6C group. A more superior enzyme could be purified from phylogenetically related organisms that exploit high-potential redox reactions. In particular, the purple non-sulphur bacteria display an array of metabolic capabilities and are actively being investigated for their potential in H 2 technologies [ 33 – 40 ]. Having evolved different modifications, some of which may provide increased efficiency, exploring these groups might greatly benefit the engineered systems. The presence of O 2 in the organism’s natural environment could also affect enzyme efficiency. In environments with a flux of O 2 , there is still the possibility that when O 2 is present, the MBH is either not expressed or is expressed but has a lowered efficiency. A relationship between expression, aerobicity and enzyme function was demonstrated with hyd-5 in S. enterica [ 22 , 41 ]. Twenty-four percent of the enzymes from our study were detected in predominantly aerobic environments with a potentially constant exposure to O 2 . Therefore, one could speculate that aerobes would possess hydrogenases that are extremely tolerant and potentially more efficient than the ones in the versatile group and the obligate anaerobes. Indeed, in experiments, the aerobe H. marinus retained a higher percentage of activity compared to the microaerophilic R. eutropha after exposure to air [ 42 ]. The aerobes B. vietnamiensis G4 , P. naphthalenivorans CJ2 and M. petroleiphilum PM1 have phylogenetically related hydrogenases to R. eutropha (Figure S1) that might have evolved greater efficiency due to their aerobic heterotrophic lifestyles. The current study recovered MBH sequence fragments via shotgun high-throughput sequencing of metagenomic DNA from an environmental sample. Although the number of MBH sequence fragments recovered in our study was lower than expected, it is comparable to work utilising similar techniques to discover novel hydrogenases from the global ocean survey of surface waters [ 32 ]. Of the five new 6C sequences, four segregate with aerobic/facultative heterotrophic α- and β- proteobacteria on the small subunit tree, and similar to their cultured neighbours, could also make use of high-potential redox couples. The full-length small subunit sequence 20 from the Rhodocyclales family (Table S2) segregates with the aerobic β-proteobacteria Methylibium petroleiphilum and Polaromonas naphthalenivorans . Both are aquatic aerobic heterotrophs that can use methyl tert-butyl ether [ 43 ] and Naphthalene [ 44 ], respectively, as sole carbon sources. Similarly, the close-to-full-length fragment 218 segregates with Decloromonas aromatica species and is most likely from this or a closely related organism. The phylogenetic tree and analyses of taxonomic groups, natural environments and oxygen requirements enabled us to place all 44 de novo MBH sequence fragment amongst the 177 MBH identified from the database search providing a powerful tool for bio-inspired enzyme engineering. Using the database analysis, we can now make use of well-characterised biology and couple it to the uniqueness offered by an untapped reserve of natural diversity, greatly enhancing a simple BLAST search. It stands to reason that these de novo sequences contain novel combinations of amino acid residues that could be utilised to engineer the hydrogenases from organisms that lend themselves well to pure-culture. In addition, identifying related sequences can target primer design and identify genomes of organisms that can serve as a scaffold for downstream procedures to assemble and retrieve ‘missing’ gene information in order to re-create a full enzyme."
} | 2,814 |
28099471 | PMC5242510 | pmc | 8,878 | {
"abstract": "Rhizobia-legume symbiosis is the most well researched biological nitrogen fixation system. Coating legume seeds with rhizobia is now a recognized practical measure for improving the production of legume corp. However, the efficacy of some commercial rhizobia inoculants cannot be guaranteed in China due to the low rate of live rhizobia in these products. A greenhouse experiment was conducted to assess the effects of different rhizobial inoculant formulations on alfalfa productivity and nitrogen fixation. Two rhizobia strains, (ACCC17631 and ACCC17676), that are effective partners with alfalfa variety Zhongmu No. 1 were assessed with different concentrations of ammonium molybdate in seed-coat formulations with two different coating adhesives. Our study showed that the growth, nodulation, and nitrogen fixation ability of the plants inoculated with the ACCC17631 rhizobial strain were greatest when the ammonium molybdate application was0.2% of the formulation. An ammonium molybdate concentration of 0.1% was most beneficial to the growth of the plants inoculated with the ACCC17676 rhizobial strain. The sodium carboxymethyl cellulose and sodium alginate, used as coating adhesives, did not have a significant effect on alfalfa biomass and nitrogen fixation. However, the addition of skimmed milk to the adhesive improved nitrogenase activity. These results demonstrate that a new rhizobial seed-coat formulation benefitted alfalfa nodulation and yield.",
"conclusion": "Conclusion Mo-enriched rhizobial seed-coat inoculants significantly improved alfalfa plant heights, dry weight, root nodule number, nodule weight, and nitrogenase activity. The best Mo concentration for rhizobia strain ACCC17631 is 0.2% (W/V) and rhizobia strain ACCC17676 was most effective at 0.1% (W/V) Mo addition. CMC and AE were applied as adhesive agents to combine solid inoculate and seeds and had no significant effect on plant growth and nodulation. Surprisingly, nitrogenase activity was enhanced significantly by adding skimmed milk into AE adhesive. Overall, the rhizobia seed-coat formulations, which included 0.1% (0.2%) Mo addition and used alginate + skimmed milk as adhesive, were beneficial to alfalfa production and N-fixation in this greenhouse study.",
"introduction": "Introduction Alfalfa ( Medicago sativa L.) is a leading forage species with wide distribution and the largest cultivated acreage in China [ 1 , 2 ]. As perennial leguminous forage with access to fixed atmospheric nitrogen (N 2 ), alfalfa has a long history in livestock production and grassland restoration due to its high nutritional value. In recent years, the Chinese dairy industry has been shaken by product-quality scandals that are mainly due to the lack of high-quality protein in cattle feed. Therefore, alfalfa production can play a critical role in improving dairy quality as well as the success of China’s dairy industry. However, because of its small-size seed and limited seed nutrients, alfalfa can be negatively impacted by unfavorable germination conditions such as drought and cold. Seed-coat products are increasingly recognized to improve seed germination and increase alfalfa production. Rhizobia-legume symbiosis is the most well researched biological nitrogen fixation system. It has been proved that rhizobia can increase host-plant nitrogen supply by colonizing its roots and exchanging nitrogen fixed by the bacteria for plant photosynthate within root nodules [ 3 , 4 ]. Coating alfalfa seeds with rhizobia inoculation products can also increase seed size, making them uniformly shaped and protected from certain pests. This is conducive to mechanized planting, as well as improving nitrogen fixation and alfalfa yield [ 5 ]. There is typically limited success from coating seeds with rhizobia because it is difficult to maintain living and active bacterial cells. The efficacy of some commercial rhizobia inoculants cannot be guaranteed in China due to the low rate of live rhizobia in these products [ 6 , 7 ]. Factors such as temperature, humidity, and toxic substances all affect the survival of rhizobia in the seed-coating agent [ 8 ]. As an essential component of nitrate reductase and nitrogenase, molybdenum (Mo) plays a central role in nitrogen metabolism [ 9 ]. Studies have found molybdenum fertilizer can significantly increase the activities of nitrogenase, glutamine synthetase and asparagine synthetase, enhancing symbiotic N 2 fixation capacity of root nodules and the nitrogen metabolism of plants [ 10 ]. Thus, adding an appropriate amount of molybdenum to the inoculation formulas could improve seed respiration and increase the survival of the rhizobial cells [ 11 ]. However, an overabundance of Mo would change the permeability of bacteroid cell membranes and prevent the normal transportation of ammonia. This could inhibit nitrogenase activity, resulting in reduced nitrogen fixation in root nodules [ 12 ]. Moreover, it has been suggested that the responses of different strains of rhizobia to Mo fertilization are dissimilar and depend on various factors, especially Mo concentration in the seed-coating formulation [ 11 , 13 ]. Utilizing an adhesive agent in seed-coat inoculants enables the rhizobial carrier medium to attach to seeds and prevents or diminishes the direct threat of substances that could potentially contaminate the rhizobia on the seed surface [ 14 ]. Researchers report rhizobia activity can be significantly influenced by different adsorbent substances [ 15 , 16 ]. Due to low cost and environmental safety, both carboxymethyl cellulose (CMC) and alginate (AER) are the most common polymeric material, used for commercial microorganism inoculation. These polymers have been demonstrated to protect the rhizobia against adverse environmental conditions and retain a large number of viable cells even after six months of storage [ 15 , 17 , 18 ]. Therefore, it is critical to assess new seed-coat formulations to determine their potential quality and efficiency for rhizobial inoculation. The objectives of this study were to investigate the effect of different concentrations of ammonium molybdate and different adhesive agents on the activity of rhizobia in seed-coat inoculant formulations as well as potential impacts on alfalfa biomass. These assessments will provide technical support for achieving high yields of high-quality alfalfa in agricultural production.",
"discussion": "Discussion When considering the inoculation of legume crops with rhizobia, the first step is finding a highly effective strain. R . meliloti ACCC17631 and ACCC17676, isolated from Xinjiang and Shandong province in China and cultured with alfalfa, were effective partners for Zhongmu No.1 alfalfa. Our results indicate ACCC17676 was more beneficial to aboveground plant production and root nodulation, as well as nitrogenase activity than inoculation with ACCC17631, however, ACCC17631 has been shown to have more environmental adaptability and higher tolerance for adverse conditions [ 29 ]. We also tested the stress resistance of these two strains. ACCC17676 is more sensitive to salt and high pH, whereas the growth rates of ACCC17631 relatively unaffected by salt or alkali with NaCl concentration of culture media up to 7.0% (W/V) or culture pH up to 9.5. Therefore, both of these strains would be the suitable candidates for commercial rhizobial inoculation production and should be chosen for different specific soil conditions and properties. Typically, legumes have small Mo requirements but low Mo availability can inhibit nitrogenase activity in root nodules [ 30 ]. The benefits of sowing seeds enriched with Mo include increasing plant yields, as shown for common beans [ 31 – 33 ] and soybeans [ 11 ]. Our greenhouse study demonstrated that 0.1%-0.2% molybdenum supplied in the rhizobia seed-coat formulations significantly increased the plant heights and aboveground biomasses for both rhizobia strains. This agrees with Rahman et al. (2008) who reported that mungbeans exhibit superior yield when they combined rhizobia inoculant along with 1.0 kg Mo/ha in silty clay loam soils [ 34 ] and Campo et al. (2009) who demonstrated inoculation of Mo-rich seeds significantly increased soybean yield and total N compared to the non-inoculated seeds and control [ 11 ]. The majority of existing research focuses on the single effect of Mo seed enrichment of leguminous crops [ 35 ], or the root nodulation response to supplemental Mo fertilizer with peat-based powders or liquid inoculants [ 36 ]. The development and impact of applying Mo directly into seed-coat inoculations have received much less attention, particularly for alfalfa. The reason may be that plant seeds and rhizobia are susceptible to Mo overdosing, which could reduce the survival of rhizobia, damage seed respiration, and decrease plant nodulation and N 2 fixation [ 37 , 38 ]. These toxic effects of Mo depend on the plant species and strains of microbes, soil pH, and other environmental factors [ 39 – 42 ]. Therefore, the Mo tolerance threshold of different rhizobia strains and crops should be assessed carefully. Our research showed rhizobia ACCC17631 grew well with Mo content below 0.4% but ACCC17676 cells were suppressed with Mo content at or above 0.2%. Further testing revealed that the growth and nodulation of alfalfa were strongly inhibited when Mo concentration exceeded the rhizobia tolerance threshold content. Similar results have been reported when high concentrations of Mo had negative osmotic effects on Bradyrhizobium [ 42 , 43 ] and an inhibitory effect on the activity of phosphatase in pea plumules [ 44 , 45 ]. However, legume crop yield and nodulation did not always respond to the addition of Mo to seed inoculant formulas. No significant effects were observed due to the Mo content of black beans, common beans, and soybeans [ 13 , 35 , 46 ] These different observations most likely relate to differences in plant genetics with regard to Mo accumulation in seeds [ 47 ]. During the last two decades, polymeric materials were evaluated as potential bacterial carriers or seed inoculant adhesive agents due to their low cost, strong compatibility with bacteria, and ease of field application [ 48 – 50 ]. CMC and AE are the most common high molecular synthetic compounds for entrapped microbes [ 51 , 52 ], can stick closely to coated materials, sustain live rhizobial cells, are harmless to cotyledons, and are not toxic in the environment [ 53 , 54 ]. While many reports have focused on the ability of polymers to sustain rhizobia on legume seeds, especially on cowpea and soybean seeds [ 49 , 55 – 58 ], their effects on root nodulation and N-fixation have not been well quantified. CMC is a cellulose-derived ester and pre-studies indicated that CMC has proper chemistry characteristics for stable microbe storage [ 48 , 59 , 60 ]. Alginate is a naturally formed polymer and wildly used for entrapping microorganisms in two main forms, encapsulated formulations and alginate beads [ 61 ]. Alginate encapsulation and alginate beads have been verified to improve microorganism survival time, but alginate products are difficult to blend with seeds for commercial application. The bacteria released from the beads must migrate through the soil, compete with native microflora, and strive to attach to seeds [ 62 ]. Our study explored a new way to compare the effects of different polymers as the adhesive agent for seed-coat inoculations. Although no significant differences were observed in plant height, biomass production, nodule number, or nodule weight between the adhesive agents, the activity of nitrogenase improved significantly when skimmed milk was added to the AE-adhesive. Nitrogenase activity within root nodules is typically the limiting factor for N-fixation in legume crops [ 63 ]. Alginate formulations containing skimmed milk have effectively conserved the viability of many Gram-negative bacteria [ 64 , 65 ]. The highest experimental survival of Azosprillum was achieved by the addition of skim milk to Alginate beads, with greater than 88% survival even after 150 days of storage [ 66 ]. Nitrogenase activity would be linked with rhizobium effectiveness, active nodule number, and nodule weight. Vieira et al. (1998) reported that more effective rhizobium had a positive influence on nitrogenase activity [ 67 ]. In our study, nitrogenase activity improved in the alginate + skimmed milk treatment, which may be as a result of increased adhesive agent protection of live rhizobial cells. It is a slow process to develop new and effective rhizobial seed-coat inoculant formulations. Rhizobium inoculation may function ideally under precise greenhouse conditions, with scientific equipment and the management of technical personnel, but it is difficult to achieve similar results for an effective product used in field conditions by microbiologically untrained farmers [ 68 ]. The introduced bacteria need to find an empty niche in the soil and compete with the well-adapted native microflora. Generally, the industry procedures for developing microbial inoculants will include the following steps: isolation and screening of the best rhizobial strain, identification of the strains, formulation, growth chamber and greenhouse testing, field microplot testing, release to industry, mass production, testing in farmer’s fields, registration and commercialization, and finally wide field use [ 61 ]. We recognize our work is the beginning of a lengthy process and many issues will need to be addressed; such as, how we can improve the survival of rhizobia in the formulation, how we can ensure good field performance of seed-coat inoculants, how we can develop low-cost technologies for extending the shelf life of inoculation products. Formulation and field function of inoculants are a matter of sustainable agriculture and environmental development rather than merely a technical challenge. Further studies will be conducted to assess the benefits of the seed coat formulations in different soil conditions and explain the potential microbiological and material science mechanisms."
} | 3,527 |
34007059 | PMC8528912 | pmc | 8,880 | {
"abstract": "Members of the genus Acidithiobacillus , now ranked within the class Acidithiobacillia , are model bacteria for the study of chemolithotrophic energy conversion under extreme conditions. Knowledge of the genomic and taxonomic diversity of Acidithiobacillia is still limited. Here, we present a systematic analysis of nearly 100 genomes from the class sampled from a wide range of habitats. Some of these genomes are new and others have been reclassified on the basis of advanced genomic analysis, thus defining 19 Acidithiobacillia lineages ranking at different taxonomic levels. This work provides the most comprehensive classification and pangenomic analysis of this deep-branching class of Proteobacteria to date. The phylogenomic framework obtained illuminates not only the evolutionary past of this lineage, but also the molecular evolution of relevant aerobic respiratory proteins, namely the cytochrome bo 3 ubiquinol oxidases.",
"introduction": "Introduction Members of the genus Acidithiobacillus are among the most widely studied extremely acidophilic prokaryotes [ 1 ]. The genus comprises Gram-negative autotrophic bacteria that are obligate acidophiles. While all Acidithiobacillus spp. can catalyze the dissimilatory oxidation of sulfur compounds, some members of the genus grow by oxidizing ferrous iron, using oxygen as electron acceptor [ 2 ]. These combined physiological traits are used extensively in biotechnological applications, including the biomining of metal sulfide ores [ 3 ]. From a fundamental perspective, the acidithiobacilli are model bacteria for the study of chemolithotrophic energy conversion reactions and pathways under acidic conditions (reviewed in ref. [ 4 ]). More recently, they have provided key insights into the evolution of aerobic respiration [ 5 ]. First considered to be members of the Betaproteobacteria [ 6 ], the acidithiobacilli were later assigned to Gammaproteobacteria [ 7 ]. With the advance of genome sequencing and phylogenomic analysis, they were subsequently repositioned as a sister class to the Beta -, Gamma -, and Epsilonproteobacteria [ 8 ], together with Thermithiobacillus tepidarius [ 9 ]. The class Acidithiobacillia (often referred to as the acidithiobacilli) contains a single order and two families, the Acidithiobacillaceae and the Thermithiobacillaceae , the latter including non-acidophilic taxa. T. tepidarius and Thermithiobacillus plumbiphilus are the only acknowledged species of the Thermithiobacillaceae [ 10 ]. Like Acidithiobacillus spp., these bacteria are aerobic, obligately chemolithoautotrophic sulfur-oxidizers, and in common with only one Acidithiobacillus sp. ( Acidithiobacillus caldus ), moderately thermophilic [ 10 ]. Considering the high taxonomic rank of the Acidithiobacillia , little attention has been paid to their true diversity so far, thereby leaving a considerable gap in our knowledge of these extremophiles. Variability among members of this class is evident. Phenotypically, the acidithiobacilli span a wide range of growth temperatures, ranging from >50 °C in T. tepidarius [ 11 ] and A. caldus [ 12 ] to 4 °C in Acidithiobacillus ferrivorans [ 13 ], and of pH optima, ranging from 7 in T. tepidarius [ 11 ] to 1.8 in Acidithiobacillus ferrianus [ 14 ], with Acidithiobacillus sulfuriphilus growing in a pH range of 5 units [ 15 ]. While knowledge of the physiological and phylogenetic diversity of Thermithiobacillus is currently limited, bacteria confirmed as members of the Acidithiobacillus genus have been isolated from many different settings since the 1950s (reviewed in ref. [ 16 ]). Many isolates [ 17 ] and clones [ 18 ] have been assigned to distinct phylogenetic subgroups or genomovars [ 19 ], some of which are probably novel species [ 20 ]. Current consensus recognizes ten species within the Acidithiobacillus genus, five of which are able to couple the dissimilatory oxidation of elemental sulfur and reduced inorganic sulfur compounds (but not ferrous iron) to molecular oxygen: Acidithiobacillus thiooxidans [ 21 ], Acidithiobacillus albertensis [ 22 ], A. caldus [ 12 ], and Acidithiobacillus sulfuriphilus [ 15 ]. The other species are additionally capable of using ferrous iron as an electron donor (in the presence of oxygen) or ferric iron as an electron acceptor, coupled to sulfur oxidation (under anaerobic conditions): Acidithiobacillus ferrooxidans [ 23 ], A. ferrivorans [ 13 ], Acidithiobacillus ferridurans [ 24 ], Acidithiobacillus ferriphilus [ 25 ], and A. ferrianus [ 14 ]. The ability to grow with hydrogen as electron donor using either oxygen or ferric iron as electron acceptor is also widespread among the acidithiobacilli [ 26 ]. While some members of the Acidithiobacillus genus are well characterized physiologically, many isolates remain unassigned or uncultured; therefore, their metabolic characteristics remain unknown [ 27 ]. As more genomes and metagenome-assembled genomes (MAGs) assigned to acidithiobacilli are sequenced, the knowledge gap between genomic information and physiology has expanded [ 28 , 29 ]. Nonetheless, accumulating evidence confirms that most species of the Acidithiobacillus genus are highly diverse at the genomic level, exhibiting both species-to-species and strain-to-strain heterogeneity [ 30 ]. Intraspecific variations can be accounted for by diverse plasmids present [ 31 ], and by a highly diverse set of integrated mobile genetic elements [ 32 – 34 ]. Recently, pangenomic studies have provided additional insights into the core and exclusive functions of certain species of Acidithiobacillus [ 35 , 36 ]. However, studies performed so far have (i) overlooked the existence of different taxonomic ranks within the class Acidithiobacillia , (ii) involved few representative strains of each species, and/or (iii) undertaken biased sampling efforts, recovering strains from a limited range of habitats or geographical origins. Recognizing the limits of current knowledge of Acidithiobacillia , we elected to test the hypothesis that the Acidithiobacillia class encompasses a greater number of genera and species than previously acknowledged. To this end, we undertook systematic efforts to sample, culture, and sequence new representatives of the class from diverse environments. These included unconventional and little-reported niches, such as geothermal acidic watersheds and near-neutral mineral environments. In addition, we have sequenced the missing genomes of type strains of validated species of acidithiobacilli. This work has aimed to reconstruct a robust genome-based phylogeny enabling the exploration of the ecological and metabolic origin of acidithiobacilli on the basis of their extant representatives. Our work represents the state of the art for the knowledge of these extremely acidophilic bacteria and their physiology, and sheds light into their evolutionary history.",
"discussion": "Discussion Using diverse complementary approaches and an extensive set of genomes, we present here the most comprehensive study to date of the taxonomic and phylogenetic structure of the Acidithiobacillia class. Novel results supported the assignment of Acidithiobacillus sulfuriphilus strain CJ-2, A. caldus strains, and Acidithiobacillus -like spp. from the Copahue Volcano and similar environments to three different prokaryotic genera (‘ Ambacidithiobacillus ’, ‘ Fervidacidithiobacillus ’, and ‘ Igneacidithiobacillus ’), which are well separated from known Acidithiobacillus taxa, as well as Thermithiobacillus . Our analysis also revealed the existence of several novel species of acidithiobacilli, which significantly expands the genomic diversity of this class. Importantly, the phylogenomic analysis described here has provided a novel and robust framework for understanding the eco-physiology and evolution of extant species of the class. For example, it has indicated the deep-branching position of A. ferrianus among iron/sulfur-oxidizing acidithiobacilli (cf. Fig. 2 ), suggesting that the currently available genomes may only partially cover the phylogenetic space of these extremophiles. Indeed, the functional profile of iron/sulfur-oxidizing acidithiobacilli partially overlaps that of their deepest branching relative, ‘ Ambacidithiobacillus sulfuriphilus ’, especially in regard to traits of energy conservation, such as the combined presence of type 1 (ubiquinone-reducing) membrane hydrogenases and nitrogenase (cf. Fig. 4 ). Such traits are present also in Gamma - and Betaproteobacteria with sulfur-oxidizing physiology [ 93 ], which form part of the sister clades of Acidithiobacillia (Fig. 2 ). Such clades clearly evolved after separation from the Alphaproteobacteria lineage, which is recognized as the deepest branching class of Proteobacteria [ 94 ]. By inference, the class Acidithiobacillia is among the basal lineages of Proteobacteria. On the basis of the growth optima and tolerance ranges of the deepest branching taxa, it may be inferred that the last common ancestor of the Acidithiobacillia class was a thermophile. Present-day strains of T. tepidarius and ‘ F. caldus ’ can endure temperatures well >50 °C [ 11 , 12 ], while accumulating evidence suggests that ‘ Igneacidithiobacillus ’ spp. are also thermotolerant (up to 40 °C [ 27 , 95 ]). Quite probably, this ancestor was a neutrophile as is present-day T. tepidarius (optimal pH 7 [ 11 ]), which gradually developed acid tolerance much like ‘ Ambacidithiobacillus sulfuriphilus ’ (with a reported pH growth range between 1.8 and 7.0 [ 15 ]). Sulfur oxidation can be considered the most common physiology across the whole class of Acidithiobacillia , the ancestor of which were probably microaerophilic as extant members of the Thermiacidithiobacillus genus, which have terminal oxidases with high affinity for oxygen, such as the cbb 3 oxidase (Fig. 5 ). These bacteria, however, are unlikely to represent ancestral models for extant acidithiobacilli because their terminal oxidases, as well other traits of energy metabolism, are close to those of sulfur-oxidizing Betaproteobacteria and unrelated to those of other members of the class (Fig. 5 and data not shown). Conversely, the metabolic trait of iron oxidation is restricted to a specific clade of Acidithiobacillia , supporting the possibility that iron chemolithotrophy might have arrived late in the evolutionary history of the class, acquired by HGT from other primitive iron-oxidizing microorganisms. However, it remains intriguing that the same taxa possessing the aerobic traits of iron oxidation are the only facultative anaerobes in the class [ 26 ] and have the elements of nitrogen fixation, which is an ancient trait characteristic of anaerobic bacteria. The presence of several of these ancient traits in deep-branching ‘ Ambacidithiobacillus sulfuriphilus ’ (e.g., the nitrogenase) supports the possibility that some of them were present already in the common ancestor of the acidithiobacilli, after branching of the lineage of extant Thermothiobacillus . Subsequent differential loss might then explain their patchy distribution in current taxa (Fig. 6 ), as has occurred in other taxa exposed to strong selective conditions during their evolutionary course (e.g., sulfur-oxidizing symbionts of Gammaproteobacteria [ 93 ]). While initially iron reduction may have provided the ecological advantage for retaining nitrogen fixation under anaerobic conditions, active oxygen consumption achieved through iron oxidation may still provide local situations of virtual anaerobiosis favoring the function of oxygen-sensitive nitrogenases. Such bacteria were almost certainly aerobes (or facultative anaerobes) and capable of bioleaching trace metals associated with the abundant pyrite minerals that were present on the exposed continental crust. Therefore, such bacteria were the likely progenitors of extant iron/sulfur-oxidizing acidithiobacilli. Fig. 6 Emerging evolutionary scenario for heme copper oxidases (HCO) in the Acidithiobacillia class. Two variant A2-type (COX) and three variant A1-type (CYO) oxidases were identified in the genomes of sequenced acidithiobacilli. Left, characteristics of extant A1-2, A1-1a/1b, and A2-type HCO CoxB and CyoA subunits with respect to the conservation of the Cu A biding motif (number of conserved residues of the six-residues Cu A -binding site and the number of conserved cysteines in the CXXXC moiety). The A1-1a variant gene clusters are unusual in that they group together (in the same gene cluster) the deepest branching CtaA heme A synthase-encoding genes, with genes encoding late-diverging CyoAB structural subunits. Right, evolutionary transitions inferred for the HCO terminal oxidases of the Acidithiobacillia class, based on the phyletic patterns of both COX/CYO and CtaAB-encoding genes in extant lineages of the Acidithiobacillia class. Only representative species-specific patterns are depicted for simplicity. Colored circles identify the lineages by the acronyms, and colored boxes the HCO variants present in each lineage. Fully colored boxes indicate presence and opaque boxes, absence. Boxes crossed with red lines along the dendrogram, represent probable instances of gene loss in ancestral branches of the Acidithiobacillia class tree. The dendrogram is based on CP tree depicted in Fig. 2 . Conservation of the cysteins of the CXXXC Cu A -binding motif and gene context-based association with heme A biosynthesis genes is symbolized in the right panel accordingly. Detailed analysis of energy conversion traits that we present here suggests that a long missing transitional state in the evolution of extant A1-type bo 3 quinol oxidases form A2-type aa 3 HCO oxidases has been captured in the Acidithiobacillia class. Following this finding, we propose a probable evolutionary scenario for these terminal oxidases. In response to bursts of cyanobacteria-produced oxygen in emerged lands of primordial earth [ 96 ], ancestors of extant acidithiobacilli (presumably microaerophilic sulfur-oxidizers), evolved or acquired a primordial A2-type cytochrome c -oxidase (“A2-COX- aa 3 ”)—including its assembly factors CtaA and CtaB. Shortly after, the capacity to oxidize iron, in addition to sulfur, provided an initial ecological advantage for ancestral acidithiobacilli. Later on in their evolution, depletion of pyrite and other sulfidic minerals, and reduction in ambient oxygen levels [ 96 ] promoted the transition of cytochrome c -oxidase into a Cu A -lacking ubiquinol oxidase, still retaining heme A (“A1-COX- aa 3 ”). This new ubiquinol oxidase, became the dominant terminal oxidase after the loss of the cytochrome bc 1 complex, no longer necessary for driving energy production via reverse electron transfer [ 78 ]—once sulfur compounds became the major electron donors to the respiratory chain. Subsequently, loss of heme A biosynthesis and complete degeneration of the Cu A -binding motif, brought about the transition of the A1-COX- aa 3 into the A1 bo 3 quinol oxidases found today in most Proteobacteria. In conclusion, this study has significantly expanded the genomic countours of the Acidithiobacillia and highlighted important lacunae in our comprehension of the phylogenetic space occupied by the class. The data and conclusions call for further exploration of unconventional environments, including those that may be considered relics of the Earth’s past, and for additional genomic studies of the Acidithiobacillia . The phylogeny obtained thus far provides grounds to delineate the likely evolutionary scenarios of these extremophiles and further completes the emerging pamorama on the evolution of the phylum Proteobacteria."
} | 3,948 |
36200894 | PMC9603131 | pmc | 8,881 | {
"abstract": "ABSTRACT Microalgae can adapt to extreme environments with specialized metabolic mechanisms. Here, we report comparative physiological and genetic regulation analyses of Chlorella sorokiniana from different environmental regions of an arctic glacier, desert, and temperate native lake in response to N deprivation, for screening the optimal strain with high lipid accumulation. Strains from the three regions showed different growth and biochemical compositions under N deprivation. The arctic glacier and desert strains produced higher soluble sugar content than strains from the native lake. The arctic glacier strains produced the highest levels of lipid content and neutral lipids under N deprivation compared with strains from desert and native lake. At a molecular level, the arctic strain produced more differentially expressed genes related to fatty acid biosynthesis, glycolysis gluconeogenesis, and glycerolipid metabolism. The important functional genes acetyl coenzyme A (acetyl-CoA) carboxylase (ACCase), fatty acid synthase complex, pyruvate dehydrogenase component, and fatty acyl-acyl carrier protein (acyl-ACP) thioesterase were highly expressed in arctic strains. More acetyl-CoA was produced from glycolysis gluconeogenesis and glycerolipid metabolism, which then produced more fatty acid with the catalytic function of ACCase and acyl-ACP thioesterase in fatty acid biosynthesis. Our results indicated that the C. sorokiniana strains from the arctic region had the fullest potential for biodiesel production due to special genetic regulation related to fatty acid synthesis, glycolysis gluconeogenesis, and glycerolipid metabolism. IMPORTANCE It is important to reveal the physiological and genetic regulation mechanisms of microalgae for screening potential strains with high lipid production. Our results showed that Chlorella sorokiniana strains from arctic glacier, desert, and temperate native lake had different growth, biochemical composition, and genetic expression under N deprivation. The strains from an arctic glacier produced the highest lipid content (including neutral lipid), which was related to the genetic regulation of fatty acid biosynthesis, glycolysis gluconeogenesis, and glycerolipid metabolism. The functional genes for acetyl-CoA carboxylase, fatty acid synthase complex, pyruvate dehydrogenase component, and fatty acyl-ACP thioesterase in the three pathways were highly expressed in arctic strains. The revelation of physiological and genetic regulation of strains from different environmental regions will contribute to the microalgae selection for high lipid accumulation.",
"introduction": "INTRODUCTION Overconsumption of fossil fuel by industry and transportation leads to decreasing world fossil fuel reserves and serious environment problems, e.g., atmospheric pollution and global warming ( 1 ). It is important to search for alternative and ecologically friendly energy resources. Recently, biodiesel, which can be produced from a variety of sources (including plants, animals, and microbes) has been proposed as a renewable energy source ( 2 ). The search for sustainable sources of biofuels has led to renewed interest in microalgae as a potential feedstock, due to the potential to synthesize and accumulate large quantities of lipids in some species ( 3 , 4 ). The advantage of microalgae is that they can accumulate large quantities of oils in the form of triacylglycerols (TAGs), which are preferred renewable oils because they possess a high molar ratio of hydrogen to carbon ( 4 – 9 ). Microalgae grown under nutrient limitation exhibit considerable variation in their biochemical composition, depending on the limiting nutrient and the degree of limitation. The nitrogen deprivation response is perhaps the best-characterized inducer of lipid accumulation in microalgae ( 10 – 13 ). In addition to increased total lipid content, N deprivation can also induce changes in fatty acid chain length and saturation ( 14 , 15 ). Moreover, some species of microalgae can be efficiently transformed, which makes it possible to enhance the productivity of natural compounds through genetic strain engineering strategies ( 16 , 17 ). Among the oleaginous microalgae, the Chlorella genus of green algae ( Chlorophyceae ) is commonly considered a promising candidate, due to its high photosynthetic efficiency, lipid productivity, and fast growth ability ( 18 ). Chlorella microalgae are able to adapt to a variety of environmental conditions, including extreme habitats ranging from desert soil to the polar region ( 19 ). Although Chlorella has proved to be a potential industrial microalgae for biological engineering, it is unclear whether strains from some environments are more suitable for producing high levels of lipids ( 15 , 20 – 24 ). At present, screening oleaginous microalgae strains which have great potential to produce TAGs is of great importance for biodiesel production. The potentially oleaginous strains could also be used in transgenic microalgae, which could be used as industrial microalgae based on biological engineering ( 25 ). This biological engineering technology shows great promise to simplify the production process and significantly decrease biodiesel production costs. Many microalgae can adapt to extreme habitats ranging from desert soil to arctic regions ( 19 , 26 ). Their long-term adaptation shows that the microalgae can adjust their physiological responses to adapt to such environments ( 27 ). In particular, microalgae from extreme environments possess special physiological features and molecular mechanisms for their adaption ( 28 ). In this context, microalgae strains in various environments may have different physiological and genetic regulation metabolisms, including those for growth and lipid accumulation. For example, it was indicated that the increased total lipid content and fatty acid composition of Antarctic microalgae could provide its adaptability to low temperatures ( 29 ). Some desert microalgae have also been shown to have lipid productivity ( 30 ). Therefore, comparative analyses of microalgae strains from different environmental regions in response to nitrogen stress may be used to screen the potential industrial microalgae for biofuel production and also be used in transgenic microalgae for biofuel production. Presently, few studies have compared microalgae from extreme and native environmental regions, e.g., polar, desert, and temperate regions. Here, we report the growth, lipid accumulation, and transcriptome analysis of three Chlorella sorokiniana strains from different environmental regions, an arctic glacier, desert soil, and temperate native lake, in response to nitrogen stress. The responses in terms of their cell growth, biochemical compositions, lipid productivity, fatty acid profiles, and functional genetic expression to N deprivation were compared. We aimed to reveal the physiological and genetic regulation of microalgae from different environmental regions with N deprivation for selecting potential industrial microalgae and providing the basis for transgenic microalgae in biofuels.",
"discussion": "RESULTS AND DISCUSSION Isolation and identification of Chlorella strains from different environmental regions of arctic glacier, desert soil, and temperate native lake. Microalgal biomass, a promising source of biofuel, could contribute to the decrease in our dependence on fossil fuel while offering multiple environmental advantages compared with traditional biofuel land crops. Long-term evolution has made possible the development in algae of physiological adaptations to cope with different natural environments ( 31 ). It was found that nitrogen deprivation plays an important role in lipid accumulation of microalgae ( 11 – 13 ). It is possible to screen microalgae strains with high lipid production under N deprivation from different environmental regions by comparing their physiological and genetic regulation mechanisms. Chlorella , a kind of oleaginous microalgae, is one of the genera of microalgae that has achieved commercial success on a large scale and can occupy extreme habitats. In the present research, three different Chlorella sp. strains were isolated from arctic glacier (referred to here as Chorella -Arc), desert soil ( Chorella -Des), and temperate native lake ( Chorella -Nat) regions. They were first identified based on morphological characteristics. Then, both rbcL and tufA sequences were obtained from the three strains and were analyzed using BLAST against sequences in GenBank. The identification of these Chlorella sp. strains from the three environmental regions is illustrated in Fig. 1 . The corresponding GenBank accession numbers are as follows: for tufA , KR154271.1 ( Chlorella -Nat), KR154255.1 ( Chlorella -Des), and ON602074 ( Chlorella -Arc); for rbcL , KM514865.1 ( Chlorella -Nat), KM514884.1 ( Chlorella -Des), and ON602073 ( Chlorella -Arc). It was shown that the three strains clustered together with other Chlorella sorokiniana strains (from published sequences) in the tufA phylogenetic tree, which was independent from the Chlorella variabilis clade. The three strains also clustered together with other published C. sorokiniana sequences in the rbcL phylogenetic tree. Although Chlorella -Arc was not much closer to Chlorella -Des and Chlorella -Nat, it was in the same clade with the other two C. sorokiniana sequences. Thus, the results indicated that the three strains from desert, template native lake, and arctic glacier were identified as C. sorokiniana by molecular tools. FIG 1 Maximum-likelihood trees of Chlorella strains inferred from tufA and rbcL gene sequences (the three strains used in this study are underlined). Data are based on 1,000 replications; bootstrap values of >50% are shown above the internodes. Effect of nitrogen deprivation on growth of the three C. sorokiniana strains. The growth rates of the three C. sorokiniana strains from arctic glacier, desert, and template native lake in the medium deprived of different nitrogen are shown in Fig. 2 , with numbers of cells provided. Generally, a lower or higher nitrogen concentration than 2.98 mmol/liter did not promote growth for the three strains. Chlorella -Des grew well in 0.74 mmol/liter, which may have been due to the extremely scarce nutrition in a desert. In addition, the growth of the three strains seemed to be region dependent under both N deprivation and N-replete cultural conditions. By comparison, the strain from the arctic glacier was more nitrogen sensitive than was the temperate native lake algae. The reason for the difference growth rates may be that the strains in different environmental regions have distinct mechanisms of growth adaptations and oil accumulation. So far, the mechanism of oil accumulation in microalgae, including N gene expression of nitrogen starvation-induced accumulation, is still not clear ( 13 , 24 , 32 ). The distinct responses to N deprivation and N-replete conditions among strains from different environmental regions indicated that microalgae could perform special adaptation to the nutritional limits in a regional environment, which is significant when selecting high-quality strains for producing biodiesel, especially based on genetic engineering. FIG 2 Growth curves of Chlorella -Arc (a), Chlorella -Des (b), and Chlorella -Nat (c) cultured under different sodium nitrate concentrations. The points represent means ± SD of triplicate samples. Effect of nitrogen deprivation on biochemical composition. The biochemical compositions of Chlorella -Arc, Chlorella -Des, and Chlorella -Nat were all measured at the end of exponential growth. The detailed time for Chlorella -Arc was on day 12, and the detailed time for both Chlorella -Des and Chlorella -Nat was day 10. There was obvious accumulation of chlorophyll in response to the increased nitrogen content for all three strains, but the chlorophyll contents of Chlorella -Arc and Chlorella -Des were generally constant after the nitrogen content reached 2.94 mmol/liter ( Fig. 3a ). This result was consistent with a previous report that degradation of chlorophyll could be induced by nitrogen deprivation in Chlorella protothecoides ( 33 ). Actually, chlorophyll breakdown is the most conspicuous symptom of leaf senescence and fruit ripening ( 34 ), and the character of chlorophyll breakdown in microalgae is similar to the visual degreasing observed during leaf senescence and fruit ripening. Compared with Chlorella -Nat, the Chlorella -Arc and Chlorella -Des strains showed higher chlorophyll contents in lower N concentrations between 0.37 and 8.82 mmol/liter. High chlorophyll content could be an indication of high photosynthetic capacity. The Chlorella -Arc and Chlorella -Des strains possibly evolved special regulation mechanisms for photosynthesis to producing chlorophyll under nutrition limitation in adapting to the extreme environments. The Chlorella -Nat strain has a normal regulation mechanism for producing chlorophyll. Thus, with increasing nutrition concentrations, the photosynthetic capacity of Chlorella -Nat is enhanced and thus the strain produces more chlorophyll. There were no significant changes in protein contents for any strains under the N-deprivation regimen, and the protein contents were just slightly higher for Chlorella -Des and Chlorella -Nat under the N-replete regimen ( Fig. 3b ). It was indicated that Chlorella -Arc had a significantly higher protein content at all treatments among the tested strains and had an obvious increased trend of protein content in response to the increased nitrogen content; this finding was consistent with a report that protein is the main product of photosynthesis in polar algae due to the lower temperatures ( 27 ). FIG 3 The changes of chlorophyll (a), protein (b), and soluble sugar (c) concentrations of Chlorella -Arc, Chlorella -Des, and Chlorella -Nat cultured under different sodium nitrate concentrations. The points represent means ± SD of triplicate samples. Different letters above the bars indicate significant differences ( P < 0.05) between growth temperatures. The soluble sugar contents of the three strains did not show consistent trends with respect to N deprivation and N-replete conditions ( Fig. 3c ). The strains from the desert and the arctic glacier produced higher soluble sugar contents than the strains in the native lake region, indicating that this organism in desert and arctic regions tends to resist the extreme weather by increasing soluble sugar in cells, especially when sufficient nutrition is available. The seasonal and diurnal temperature fluctuations have a strong influence on the metabolic function and photosynthesis of microalgae in outdoor environments ( 35 ). For example, Chlorella strains in desert have a good capacity for sugar accumulation because of the diurnal temperature fluctuations ( 35 ). In addition, it has been demonstrated that microalgae in polar regions can resist the intense radiation by increasing the sugar content in cells ( 36 ), and this is consistent with the result here indicating that the strain from the arctic glacier had higher sugar content than the template native lake strain. Effect of nitrogen deprivation on neutral lipid and lipid content. The changes of neutral lipids for Chlorella -Arc, Chlorella -Des, and Chlorella -Nat in response to the N deprivation and N-replete conditions are shown in Fig. 4 . We found that the neutral lipid accumulated obviously in response to the N deprivation for all three strains from the different environments ( Fig. 4 ), but it was apparent that the Chlorella -Arc strain accumulated the highest content of neutral lipids among the three strains. FIG 4 Changes in neutral lipid concentrations based on fluorescence intensity of Chlorella -Arc, Chlorella -Des, and Chlorella -Nat cultured under different sodium nitrate concentrations. The bar charts represent means ± SD of triplicate samples. Different letters above the bars indicate significant differences ( P < 0.05) between growth temperatures. The lipid content and lipid productivity of the three Chlorella strains are shown in Fig. 5 . It can be clearly seen that the Chlorella strains from the three environmental regions of polar, desert, and temperate native lake had different changes in lipid contents in response to N deprivation. For all three strains, the total lipid content under the N deprivation condition was higher than that under the N-replete condition, and this was consistent with the changes of neutral lipids. It was apparent that all three strains had higher lipid contents at the lowest nitrogen concentrations of 0.37 mmol/liter and 0.74 mmol/liter, which indicated that the N deprivation had a positive effect on the lipid accumulation. Although the Chlorella -Arc strain had low lipid productivity, it had the highest lipid content among the three strains under N deprivation and N-replete culture conditions, as the lipid contents reached 43% and 54% under the nitrogen concentration of 0.74 mmol/liter and 0.37 mmol/liter, respectively. The Chlorella -Des and Chlorella -Nat strains had highest lipid contents 40% and 41%, respectively. The reason for the high lipid content in Chlorella -Arc may be that such a strain has to achieve high utilization of nitrogen and develop special lipid accumulation processes to produce lipid for storing energy for long-term survival in extremely cold regions ( 37 ). FIG 5 Biomass productivity, lipid productivity, and lipid content of Chlorella -Arc, Chlorella -Des, and Chlorella -Nat cultured under different sodium nitrate concentrations. The bar charts and points represent means ± SD of triplicate samples. It was clear that among the three strains, Chlorella -Arc from the polar region could be a promising candidate for biodiesel feedstock, since it could accumulate higher lipid content under the N deprivation condition, and in particular the highest neutral lipid contents, which are the main materials for biofuels. Although Chlorell a-Arc had a low growth rate, it could be used in transgenic microalgae for biofuel production based on the genetic strain engineering strategies that offer a means to accelerate the commercialization of algal biofuels by improving the rate and total accumulation of microalga lipids ( 38 , 39 ). The reason for the difference of growth rates and lipid contents of the three strains may be related to their adaptation to local environments in the long-term evolution on the genetic level. The different environmental factors could lead to their changes in physiology, which are controlled by gene regulation ( 40 ). At present, more studies are focusing on the molecular regulation of nitrogen starvation-induced lipid accumulation ( 11 , 13 , 24 ). For example, Goncalves and colleagues in 2016 ( 24 ) provided evidence to support the hypothesis that transcription factor ROC40 has a role in N-induced lipid accumulation, and they uncovered multiple previously unknown proteins modulated by short-term N in green algae. Effect of nitrogen deprivation on fatty acid composition. Fatty acid composition, which affects the quality of biodiesel production, is strongly strain specific ( 41 ). In our research, the fatty acid compositions of the three Chlorella strains from arctic glacier, desert, and temperate native lake regions in response to N deprivation were well compared. Different patterns of fatty acid composition were found among the three strains from the different environmental regions ( Tables 1 , 2 , and 3 ). For each strain, the fatty acid composition was also different under N deprivation and N-replete conditions. The fatty acid compositions of the three Chlorella strains were mainly composed of C 16 and C 18 fatty acids; this was similar to findings reported elsewhere ( 35 , 42 , 43 ). C 18:3 (linolenic acid) contents in Chlorella -Arc and Chlorella -Nat strains were lower than 6% under all treatments ( Table 2 and Table 3 ). Thus, the fatty acid compositions of all three Chlorella strains under N deprivation met the requirements of European Standard EN14214 for biodiesel production ( 44 ). TABLE 1 Fatty acid compositions of Chlorella -Des cultured under different sodium nitrate concentrations Fatty acid % of total fatty acids (mean ± SD) when cultured in NaNO 3 concn of: 0.37 mmol/liter 0.74 mmol/liter 1.47 mmol/liter 2.94 mmol/liter 5.88 mmol/liter 8.82 mmol/liter 11.76 mmol/liter SFAs C 16:0 21.89 ± 2.40 16.89 ± 2.12 9.64 ± 5.81 1.30 ± 0.21 11.48 ± 7.38 2.27 ± 0.09 14.16 ± 0.88 C 17:0 9.70 ± 2.02 12.95 ± 1.39 15.70 ± 0.25 14.92 ± 1.77 16.06 ± 3.78 15.66 ± 1.27 12.72 ± 0.20 C 18:0 1.33 ± 0.09 0.20 ± 0.34 0.36 ± 0.63 0.31 ± 0.54 0.64 ± 0.71 ND 0.34 ± 0.48 C 20:0 ND a 0.13 ± 0.22 0.17 ± 0.30 0.44 ± 0.38 0.28 ± 0.25 ND 0.28 ± 0.39 Total 32.92 ± 4.53 30.15 ± 2.65 29.30 ± 1.38 16.97 ± 2.29 28.46 ± 2.92 17.93 ± 1.37 27.49 ± 2.31 MUFAs C 16:1 3.03 ± 1.45 2.26 ± 3.28 5.55 ± 0.91 7.58 ± 0.84 3.86 ± 3.49 4.52 ± 0.08 5.52 ± 1.78 C 17:1 2.65 ± 0.31 2.30 ± 2.01 3.26 ± 0.30 3.52 ± 0.21 2.03 ± 1.76 3.44 ± 0.02 3.45 ± 0.05 C 20:1 27.65 ± 1.54 36.14 ± 1.70 39.71 ± 7.16 40.50 ± 0.52 37.60 ± 1.32 40.30 ± 0.11 34.76 ± 2.02 C 18:1n9t 2.57 ± 0.53 ND ND ND ND ND ND C 18:1n9c 15.66 ± 0.15 17.66 ± 2.18 17.67 ± 4.20 17.92 ± 1.07 15.92 ± 1.47 18.55 ± 0.43 16.07 ± 0.90 Total 51.57 ± 2.92 58.36 ± 2.46 59.25 ± 1.31 69.52 ± 1.73 59.40 ± 3.47 66.81 ± 0.47 59.80 ± 1.13 PUFAs C 18:2n6t 2.30 ± 0.41 2.93 ± 2.62 5.20 ± 1.96 5.20 ± 0.61 4.40 ± 0.17 5.81 ± 1.32 5.29 ± 0.15 C 18:2n6c 11.52 ± 1.18 8.56 ± 1.00 7.96 ± 1.01 8.31 ± 0.26 7.73 ± 0.61 9.45 ± 2.22 7.41 ± 0.66 Total 13.82 ± 0.78 11.49 ± 1.62 11.44 ± 0.06 13.51 ± 0.63 12.13 ± 0.77 15.26 ± 0.89 12.70 ± 0.81 UFA/SFA ratio 2.01 2.12 2.42 4.89 2.51 4.59 2.64 a ND, not detected. TABLE 2 Fatty acid compositions of Chlorella -Arc cultured under different sodium nitrate concentrations Fatty acid % of total fatty acids (mean ± SD) when cultured in NaNO 3 concn of: 0.37 mmol·L −1 0.74 mmol·L −1 1.47 mmol·L −1 2.94 mmol·L −1 5.88 mmol·L −1 8.82 mmol·L −1 11.76 mmol·L −1 SFAs C 16:0 8.12 ± 0.69 3.36 ± 1.11 5.39 ± 1.42 8.86 ± 0.49 3.95 ± 0.88 25.95 ± 1.57 24.41 ± 2.68 C 20:0 12.51 ± 1.02 17.81 ± 0.58 11.53 ± 0.66 14.54 ± 0.53 16.39 ± 0.56 2.21 ± 1.30 12.08 ± 1.27 Total 20.63 ± 2.32 21.17 ± 0.46 16.92 ± 0.28 23.40 ± 0.46 20.34 ± 0.56 28.16 ± 6.85 36.49 ± 5.69 MUFAs C 16:1 12.51 ± 0.34 6.64 ± 0.46 5.27 ± 0.03 3.3 ± 0.29 5.98 ± 0.67 7.04 ± 0.79 9.81 ± 3.27 C 18:1n9t 6.23 ± 0.28 22.47 ± 0.23 16.35 ± 1.35 15.11 ± 1.23 15.62 ± 0.86 3.50 ± 4.5 0.46 ± 0.23 C 18:1n9c 20.46 ± 0.08 18.34 ± 0.65 20.63 ± 2.56 23.89 ± 3.45 18.47 ± 1.32 11.57 ± 6.75 13.67 ± 4.20 C 20:1 16.33 ± 0.45 15.93 ± 0.22 16.43 ± 0.73 21.41 ± 0.70 12.91 ± 2.13 39.92 ± 0.06 33.05 ± 3.83 Total 55.53 ± 0.60 63.38 ± 3.46 58.68 ± 0.70 53.71 ± 4.56 52.89 ± 0.45 62.03 ± 11.25 56.99 ± 12.41 PUFAs C 18:2n6t 17.13 ± 0.8 8.65 ± 015 7.75 ± 0.56 2.67 ± 00.56 7.52 ± 0.05 9.82 ± 8.45 7.52 ± 5.79 C 18:2n6c 2.98 ± 0.23 3.25 ± 2.25 12.77 ± 1.38 15.67 ± 1.34 13.53 ± 0.57 2.21 ± 3.13 11.07 ± 9.33 C 18:3n3 3.73 ± 0.19 3.55 ± 0.41 3.88 ± 0.10 4.55 ± 0.55 5.72 ± 0.61 nd a nd Total 23.84 ± 0.22 15.45 ± 0.87 24.40 ± 3.56 22.89 ± 2.45 26.77 ± 4.23 12.02 ± 3.20 18.60 ± 8.27 UFA/SFA ratio 3.84 5.47 3.10 3.37 2.74 2.63 2.07 a ND, not detected. TABLE 3 Fatty acid compositions of Chlorella -Nat cultured under different sodium nitrate concentrations Fatty acid % of total fatty acids (mean ± SD) when cultured in NaNO 3 concn of: 0.37 mmol/liter 0.74 mmol/liter 1.47 mmol/liter 2.94 mmol/liter 5.88 mmol/liter 8.82 mmol/liter 11.76 mmol/liter SFAs C 16:0 20.65 ± 1.46 28.81 ± 0.51 20.74 ± 2.52 31.4 ± 0.43 29.34 ± 0.77 16.08 ± 0.07 17.91 ± 2.14 C 18:0 4.51 ± 0.55 2.09 ± 0.75 3.92 ± 0.79 0.63 ± 0.05 1.91 ± 0.21 10.18 ± 0.67 14.27 ± 3.21 Total 25.16 ± 0.38 30.90 ± 0.42 24.66 ± 1.67 32.03 ± 0.09 31.25 ± 0.18 26.26 ± 0.75 25.06 ± 6.81 MUFAs C 16:1 1.29 ± 0.16 0.85 ± 0.07 1.27 ± 0.07 2.50 ± 0.32 0.59 ± 0.15 6.28 ± 0.53 6.66 ± 0.87 C 18:1n9t 10.08 ± 0.90 26.70 ± 1.05 30.31 ± 0.59 18.37 ± 1.06 22.72 ± 1.28 0.45 ± 0.07 0.67 ± 0.42 C 18:1n9c 12.26 ± 0.52 2.66 ± 1.14 5.48 ± 1.08 1.07 ± 0.19 1.16 ± 0.10 19.15 ± 0.38 17.31 ± 1.56 C 20:1 3.99 ± 0.90 1.63 ± 0.25 5.52 ± 1.48 1.76 ± 0.48 2.09 ± 0.90 28.14 ± 0.75 19.64 ± 0.74 C 20:1n9 18.19 ± 1.01 12.12 ± 1.55 9.64 ± 1.03 17.52 ± 0.57 19.26 ± 1.01 3.80 ± 0.54 3.91 ± 0.63 Total 45.84 ± 1.05 43.96 ± 1.36 52.22 ± 0.70 41.22 ± 3.28 45.82 ± 1.56 57.82 ± 0.07 55.29 ± 7.56 PUFAs C 18:2n6t 5.95 ± 1.69 1.81 ± 0.88 3.73 ± 1.08 0.77 ± 0.19 2.70 ± 0.96 15.92 ± 0.73 ND C 18:2n6c 18.61 ± 1.38 21.78 ± 1.88 15.93 ± 1.60 25.38 ± 1.08 18.93 ± 1.66 ND a ND C 18:3n3 4.24 ± 0.99 1.55 ± 0.39 3.46 ± 0.38 0.60 ± 0.09 1.31 ± 0.07 ND ND Total 28.81 ± 0.65 25.58 ± 1.45 23.12 ± 1.45 26.75 ± 2.35 22.93 ± 0.78 15.92 ± 0.74 19.65 ± 0.74 UFA/SFA ratio 2.97 2.25 3.06 2.12 2.2 2.81 2.99 a ND, not detected. The lipids containing high amounts of saturated fatty acids (SFAs) and monounsaturated fatty acids (MUFAs) with low polyunsaturated fatty acid (PUFA) content, especially C 18:3 , are more suitable for biodiesel production, since PUFA can lead to low oxidative stability during storage ( 45 ). In our study, the sum of SFA and MUFA content was generally high (67.08 to 83.08%) in strains from all three environmental regions, and the sum of PUFA content in Chlorella -Arc and Chlorella -Des strain was generally lower than that in Chlorella -Nat under different sodium nitrate concentrations ( Tables 1 to 3 ). The proportion of C 18:3 was low (<6%) in the Chlorella strain from the arctic and native lake regions and was absent in the strain of the desert region ( Tables 1 to 3 ). In addition, high UFA and SFA are important for stress resistance, since unsaturated fatty acids are more easily oxidized by oxygen radicals than are saturated fatty acids ( 19 ). Chlorella -Arc possessed the highest UFA/SFA ratios under the treatment of N deprivation, which indicated that this Chlorella strain might be the most suitable strain that not only could accumulate high lipids but also could have suitable fatty acid composition for biodiesel production under the N deprivation condition. Quality of RNA sequencing data and gene functional annotation. A total of 40.37 Gb bases were generated for all strains. The clean reads quality metrics after filtering sequences that contained low-quality, adaptor-polluted, and or a high content of unknown base (N) reads, are shown in Table S1 in the supplemental material. The abundance of raw reads of the 6 strains were from 50.63 to 52.26 Mb. The clean reads ratios of all strains ranged from 84.38% to 89.74%. The number of transcripts for the six samples were 21,333, 25,771, 44,358, 51,269, 26,359, and 26,770, with a mean length ranging from 915 bp to 1,159 bp and N50 length from 1,379 bp to 1,733 bp ( Table 4 ; see also Fig. S1 to S6). The numbers of unigenes for the six samples were 18,441, 22,145, 39,597, 45,602, 22,354, and 22,789, respectively, with a mean length of 952 bp to 1,257 bp and an N50 length of 1,457 to 1,912 ( Table 4 ; see also Fig. S1 to S6). These reads and assembly qualities indicated that the transcriptome sequencing performed well for functional analysis. TABLE 4 Quality statistics of transcripts and unigenes for each strain from three environmental regions cultured under N deprivation (1/8ND) compared to the control group (ND) Sample Total no. Total length Mean length N50 N70 N90 % GC Transcripts Nat-1/8ND 21,333 24,730,273 1,159 1,721 1,157 529 66.99 Nat-ND 25,771 28,764,323 1,116 1,721 1,131 480 66.96 Arc-1/8ND 44,358 40,980,088 923 1,433 873 378 61 Arc-ND 51,269 46,925,918 915 1,379 858 381 59.56 Des-1/8ND 26,359 28,928,276 1,097 1,730 1,140 446 66.29 Des-ND 26,770 29,424,456 1,099 1,733 1,133 447 66.15 All transcripts 195,860 Unigenes Nat-1/8ND 18,441 23,193,984 1,257 1,857 1,269 605 67.04 Nat-ND 22,145 26,846,722 1,212 1,872 1,251 546 67.01 Arc-1/8ND 39,597 38,202,965 964 1,524 935 389 61.07 Arc-ND 45,602 43,438,848 952 1,457 913 392 59.65 Des-1/8ND 22,354 26,864,972 1,201 1,912 1,288 503 66.26 Des-ND 22,789 27,187,125 1,192 1,879 1,264 505 66.1 All unigenes 77,917 87,573,379 1,123 1,841 1,170 446 62.47 All the unigenes assembled were annotated to the NR, NT, GO, COG, KEGG, Swissprot, and Interpro databases, respectively. For all annotated unigenes, the NR, Swissprot, COG, and KEGG databases showed higher annotation proportions ( Fig. 6A ). Among the 7 databases, the NR database had the most unique annotation ( Fig. 6B ). There were 22,473 shared annotated genes among the 7 databases. The top annotated species included C. sorokiniana , Chlorella variabilis , Guillardia theta , Auxenochlorella protothecoides , and Coccomyxa subellipsoidea , whereas Chlorella variabilis and Chlorella sorokiniana showed higher annotation proportions ( Fig. 6C ). This was not strange, since the genomes of C. sorokiniana and C. variabilis strains are available in public databases ( 46 , 47 ). The COG functional enrichment was performed for all unigenes where the lipid and amino acid transport and metabolism were included ( Fig. 7 ). The translation, ribosomal structure and biogenesis, cell cycle control, cell division, chromosome partitioning, and general function predictions accounted for the top classification categories ( Fig. 7 ). The KEGG database annotated the unigenes into five categories, including cellular processes, environmental information processing, genetic information processing, metabolism, and organism systems, whereas the translation and global and the overview maps accounted for the top annotation. FIG 6 (A) Annotation showing proportions of unigenes in seven databases. (B) Annotation showing differences in seven databases for C. sorokiniana . (C) Species distribution for annotation of C. sorokiniana . FIG 7 Distribution of COG function classifications for C. sorokiniana . Genetic expression patterns and pathways of differentially expressed genes. First, the gene expression distributions of the six samples from the three strains were compared, including their shared and unique unigenes. From the principal-component analysis (PCA) plot of gene expression ( Fig. 8 ), we noted that the strains DesND and Des1/8ND from desert clustered together closely and the NatND and Nat1/8ND strains from the native lake clustered together loosely. However, the ArcND and Arc1/8ND strains were separated. The comparison of the shared and unique unigenes among the strains are shown in Fig. 9 , where the distributions of the unigenes were generally consistent with the PCA analysis. These indicated that the strains under different nitrogen stress in the desert region have similar gene expression patterns. The strains under different nitrogen stress in temperate regions also had similar gene expression patterns. However, the arctic strains under different nitrogen concentrations showed significant differentially expressed genes (DEGs). FIG 8 PCA plot for the differences in expression levels among the strains of the three environmental regions. FIG 9 The shared and unique unigene distributions of strains among the three environmental regions. The DEGs among the six samples of the three strains revealed upregulated and downregulated genes among them ( Fig. 10 ). It was apparent that the Nat1/8ND versus Des1/8ND, NatND versus DesND, and NatND versus ArcND comparisons showed more DEGs, but the DesND versus Des1/8ND comparison showed fewer DEGs. The ArcND versus Arc1/8ND comparison also showed more DEGs than the DesND versus Des1/8ND comparison. This was consistent with the result from the PCA plot that showed strains from the same environmental regions had different genetic expression patterns, except for strains from the arctic region, which showed significantly different genes. These results suggest that the genetic regulated mechanisms of C. sorokiniana from different environmental regions differ from each other. In particular, the strains from arctic regions under the 1/8 N deprivation condition showed significantly different expression genes from the control group, which further suggested that the C. sorokiniana strain from the arctic region possessed a special mechanism for high lipid production (discussed below). Extreme environments like the polar region possibly make the microalgae possess special regulation mechanisms for nutrient production due to adaptive evolution ( 28 ). Multiple DEGs among the six samples were selected randomly for quantitative PCR (qPCR). The 18S gene was selected as the reference gene. In general, the statistical results indicated that the relative expression levels of the unigenes by qPCR were generally consistent with the actual expression levels (fragments per kilobase of transcript per million mapped reads [FPKM]) ( Fig. 11 ). FIG 10 Statistics of upregulated and downregulated genes among six samples of Chlorella- Arc, Chlorella- Des, and Chlorella- Nat, where ND and 1/8ND indicate the concentration of NaNO 3 (with 1 set for the control, (2.94 mmol/liter) and 1/8 times lower than the control, respectively. For example, Nat1/8ND versus Arc1/8ND comparison shows that Arc1/8ND generated more upregulated genes (blue) than downregulated genes (red) in comparison with Nat1/8ND. FIG 11 Consistency verification of relative expression levels from RT-PCR and actual expression levels from RNA-Seq. Among their DEGs, the most common were those obtained among the six C. sorokiniana samples ( Fig. 12 ). The genes that were related to carbon and lipid metabolism were included in the DEGs, including acetyl coenzyme A (acetyl-CoA) carboxylase (ACCase), fatty acid synthase complex (FAS), diacylglycerol acyltransferase (DGAT), pyruvate dehydrogenase component (PDCom), phosphoenolpyruvate carboxylase (PEPC), long-chain acyl-CoA synthetase (LC-FACS), fatty acid desaturase ( 35 , 40 , 48 , 49 ). These DEGs were highly expressed in the C. sorokiniana strain from the arctic regions (especially under 1/8 N deprivation), e.g., ACCase, LC-FACS, DGAT, PEPC, PDCom, and FAS. These findings suggested that the C. sorokiniana strain from the arctic region expressed more genes related to carbon and lipid metabolism to produce a high amount of lipids, consistent with our physiological results. First of all, the accumulation of large quantities of lipids in microalgae requires a continuous supply of acetyl-CoA for fatty acid biosynthesis ( https://www.kegg.jp/kegg-bin/show_pathway?cvr00061 ) ( 40 ) ( Fig. 13 ). In the glycolysis gluconeogenesis process of Chlorella variabilis , pyruvate, which is a key precursor in central carbon metabolism and lipid synthesis, can be transferred to acetyl-CoA under the catalytic action of PDCom ( Fig. 13 ) ( https://www.genome.jp/kegg-bin/show_pathway?cvr00010 ) ( 49 ). Meanwhile, PEPC is important for the production of pyruvate. Here, we showed that the arctic strain expressed more PDCom than the desert or temperate strains, and Arc1_8ND produced more PEPC than all other strains of the three regions ( Fig. 12 ). The physiological results also indicated that the arctic strains produced more sugar, which can then produce more acetyl-CoA through the function of PDCom and PEPC in glycolysis gluconeogenesis. DGAT plays an important role in acetyl-CoA and triacylglycerol (TAG) biosynthesis, also referred to as catabolism of fatty acids ( 35 , 48 ). In glycerolipid metabolism, TAG and acetyl-CoA are finally produced with the catalysis of DGAT ( Fig. 13 ). In this study, DGAT was highly expressed in Arc1_8ND ( Fig. 12 ). ACCase initiates fatty acid biosynthesis in microalgae and plays an important role for catalyzing the reaction converting acetyl-CoA to malonyl-CoA in the initial step ( 40 ). It was indicated that ACCase was highly expressed in Arc1_8ND ( Fig. 12 ). The fatty acyl-ACP thioesterase transfers long-chain acyl-ACP to long-chain fatty acids in the last step of fatty acid biosynthesis ( Fig. 13 ). We found that Arc1_8ND expressed the highest fatty acyl-ACP thioesterase compared to other strains. FIG 12 Expression comparisons of the highest DEGs among the strains of the three environmental regions. FIG 13 KEGG pathways of fatty acid biosynthesis, glycolysis gluconeogenesis, and glycerolopid metabolism. The genes highlighted in blue are functional catalyzing enzymes. In conclusion, the high lipid production of arctic strains is related to pathways of glycolysis gluconeogenesis, glycerolipid metabolism, and fatty acid biosynthesis. PDCom and DGAT were expressed highly in arctic strains, and thus more acetyl-CoA was produced in the glycolysis gluconeogenesis and glycerolipid metabolism. Finally, more acetyl-CoA initiated fatty acid biosynthesis through the catalytic function of ACCase and acyl-ACP thioesterase, which were highly expressed in the arctic strain."
} | 9,272 |
25505896 | PMC4245776 | pmc | 8,883 | {
"abstract": "The porin-cytochrome (Pcc) protein complex is responsible for trans-outer membrane electron transfer during extracellular reduction of Fe(III) by the dissimilatory metal-reducing bacterium Geobacter sulfurreducens PCA. The identified and characterized Pcc complex of G. sulfurreducens PCA consists of a porin-like outer-membrane protein, a periplasmic 8-heme c -type cytochrome ( c -Cyt) and an outer-membrane 12-heme c -Cyt, and the genes encoding the Pcc proteins are clustered in the same regions of genome (i.e., the pcc gene clusters) of G. sulfurreducens PCA. A survey of additionally microbial genomes has identified the pcc gene clusters in all sequenced Geobacter spp. and other bacteria from six different phyla, including Anaeromyxobacter dehalogenans 2CP-1, A. dehalogenans 2CP-C, Anaeromyxobacter sp. K, Candidatus Kuenenia stuttgartiensis, Denitrovibrio acetiphilus DSM 12809, Desulfurispirillum indicum S5, Desulfurivibrio alkaliphilus AHT2, Desulfurobacterium thermolithotrophum DSM 11699, Desulfuromonas acetoxidans DSM 684, Ignavibacterium album JCM 16511, and Thermovibrio ammonificans HB-1. The numbers of genes in the pcc gene clusters vary, ranging from two to nine. Similar to the metal-reducing (Mtr) gene clusters of other Fe(III)-reducing bacteria, such as Shewanella spp., additional genes that encode putative c -Cyts with predicted cellular localizations at the cytoplasmic membrane, periplasm and outer membrane often associate with the pcc gene clusters. This suggests that the Pcc-associated c -Cyts may be part of the pathways for extracellular electron transfer reactions. The presence of pcc gene clusters in the microorganisms that do not reduce solid-phase Fe(III) and Mn(IV) oxides, such as D. alkaliphilus AHT2 and I. album JCM 16511, also suggests that some of the pcc gene clusters may be involved in extracellular electron transfer reactions with the substrates other than Fe(III) and Mn(IV) oxides.",
"conclusion": "Conclusions In addition to their shared similarities in biological functions and protein compositions, the Pcc and Mtr systems also share similar traits in their genetic organizations, such as association with additional genes encoding c -Cyts. Analyses of the amino acid sequences of the Pcc-associated c -Cyts suggest that they may be involved in redox cycling of quinone/quinol pool in the cytoplasmic membrane and electron transfer across the periplasm and outer membrane. The first two proposed functions of the Pcc-associated c -Cyts are very similar to the proposed functions of some Mtr-associated c -Cyts. Together, these shared similarities suggest that c -Cyts play critical roles in mediation of electron transfer across not only the outer membrane but also the periplasm as well as redox cycling of quinone/quinol pool in the cytoplasmic membrane in both Pcc- and Mtr-mediated electron transfer pathways. Although it still lacks experimental verification, genomic analyses of the pcc gene clusters suggest that the Pcc system may be involved in extracellular electron transfer reactions with the substrates other than Fe(III) and Mn(III, IV) oxides, such as S 0 , SeO 2− 4 , and SeO 2− 3 , which is also very similar to the Mtr system that is involved in extracellular reduction of dimethyl sulfoxide and extracellular oxidation of Fe(II) in addition to extracellular reduction of Fe(III) and Mn(III, IV) oxides. Major differences are found between the Pcc and Mtr systems. In the Mtr system, apparent co-evolution among the porin-like outer-membrane proteins, periplasmic c -Cyts and the outer-membrane c -Cyts is suggested (Shi et al., 2012b ). In the Pcc system, extremely diverse amino acid sequences of the outer-membrane c -Cyts indicate no apparent co-evolution between the outer-membrane c -Cyts and the porin-like outer-membrane proteins/the periplasmic c -Cyts. This observation supports pervious suggestions that the Pcc and Mtr systems evolve independently (Liu et al., 2014 ). Frequent detections of the atypical heme-binding motifs in the Pcc outer-membrane c -Cyts is another unique feature of the Pcc system. Compared to those with the typical heme-motifs, the outer-membrane c -Cyts with atypical heme-binding motifs are much less characterized. Thus, future research should focus on the detailed characterizations of biochemical, biophysical and electrochemical properties of these c -Cyts with atypical heme-binding motifs. 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.",
"introduction": "Introduction Geobacter spp. are a group of Gram-negative bacteria whose hallmark feature is transfer of metabolically-derived electrons to appropriate electron-accepting substrates external to the bacterial cells, such as oxidized metals, electrodes and even other microorganisms (Lovley et al., 2004 , 2011 ; Summers et al., 2010 ). Geobacter spp. are found in a wide range of habitats and are distributed world-wide. They are important in different environmental processes, including biogeochemical cycling of carbon and iron and attenuation of metal, radionuclide, and organic contaminants. Geobacter spp. have also been harnessed for a variety of biotechnology applications, such as bioremediation of contaminants in the subsurface sediments, generation of electrical current as microbial fuel cells, and electrosynthesis of organic compounds (Lovley et al., 2004 , 2011 ). In order to use extracellular substrates as terminal electron acceptors, Geobacter spp. have developed pathways to transfer electrons from the quinone/quinol pool in the cytoplasmic membrane, across the periplasm and the outer membrane to the extracellular substrates (Lovley, 2006 ; Weber et al., 2006 ; Shi et al., 2007 , 2009 ; Bird et al., 2011 ). Previously, we identified and characterized a trans-outer membrane porin-cytochrome (Pcc) protein complex for transferring electrons across the outer membrane during extracellular reduction of Fe(III) by G. sulfurreducens PCA. The identified Pcc complex of G. sulfurreducens PCA consists of a porin-like outer-membrane protein (OmbB or OmbC), a periplasmic 8-heme c -type cytochrome ( c -Cyt, OmaB or OmaC) and an outer-membrane 12-heme c -Cyt (OmcB or OmcC). The genes that encode Pcc proteins are adjacent to each other in the genome (i.e., the pcc gene cluster) of G. sulfurreducens PCA that possesses total four pcc gene clusters, two of which are involved in extracellular reduction of Fe(III)-citrate and ferrihydrite [a poorly crystalline Fe(III) oxide]. Isolated Pcc complex reconstituted in proteoliposomes transfers electrons from the reduced methyl viologen inside the liposomes across the lipid-bilayer to Fe(III)-citrate or ferrihydrite. The pcc gene clusters are present in all eight sequenced Geobacter genomes and 11 other phylogenetically diverse bacterial genomes. Widespread distribution of the pcc gene clusters in phylogenetically diverse bacteria reflects the importance of Pcc proteins in trans-outer membrane electron transfer by the Gram-negative bacteria (Liu et al., 2014 ). Furthermore, the characterized function and organization of the Pcc complex of G. sulfurreducens PCA are very similar to that of the Mtr (i.e., metal-reducing) porin-cytochrome extracellular electron transfer complex in Shewanella oneidensis MR-1, despite the fact that Pcc and Mtr proteins are phylogenetically unrelated (Liu et al., 2014 ). In S. oneidensis MR-1, the characterized Mtr porin-cytochrome protein complex also consists of a porin-like outer-membrane protein (MtrB), a periplasmic 10-heme c -Cyt (MtrA) and an outer-membrane 10-heme c -Cyt (MtrC), and is responsible for electron transfer across the outer membrane during extracellular reduction of Fe(III) oxides (Hartshorne et al., 2009 ; Richardson et al., 2012 ; White et al., 2013 ). The Pcc and Mtr complexes appear to have evolved independently to a common functional role in mediating electron transfer across the bacterial outer membrane. The observed functional and organizational similarity between the Pcc and Mtr protein complexes collectively demonstrates that porin-cytochrome protein complex is a common mechanism shared by different groups of Gram-negative bacteria for trans-outer membrane electron transfers (Liu et al., 2014 ). Despite detailed characterization of the Pcc complexes of G. sulfurreducens PCA and discovery of widespread distribution of pcc gene clusters in the Gram-negative bacteria, other features of the pcc gene clusters, such as their genetic organization, phylogenetic relationship and potential biological functions, had not been previously investigated. In this report, we further analyzed the characteristics of identified bacterial pcc gene clusters.",
"discussion": "Results and discussion Overview As shown in Figure 1 , the pcc gene clusters were found all sequenced Geobacter genomes. These include the genomes of G. sulfurreducens PCA, G. bemidjiensis Bem, G. daltonii FRC-32, G. lovleyi SZ, G. metallireducens GS-15, G. uraniireducens Rf4, Geobacter spp. M18, and Geobacter spp. M21. Numbers of the pcc gene clusters found in Geobacter genomes varied, ranging from one in G. bemidjiensis Bem and Geobacter sp. M21 to four in G. sulfurreducens PCA. Notably, for the Geobacter genomes with > one pcc gene clusters, at least two pcc gene clusters were consistently adjacent to each other. In G. sulfurreducens PCA, the orfS-ombB-omaB-omcB , and orfR-ombC-omaC-omcC gene clusters that are adjacent to each other are the result of gene duplication because at the amino acid sequence level, OmbB/OmbC and OmaB/OmaC are 100% identical, respectively, and OrfR/OrfS and OmcB/OmcC are 99 and 71% identical, respectively (Leang et al., 2003 ; Leang and Lovley, 2005 ; Aklujkar et al., 2013 ; Liu et al., 2014 ). Given that GM18_3461/GM18_3467, GM18_3462/GM18_3468, and GM18_3463/GM18_3469 are 100% identical at their amino acid sequence levels, respectively, the adjacent GM18_3461-GM18_3462-GM18_3463 and GM18_3467-GM18_3468-GM18_3469 gene clusters of Geobacter sp. M18 are also the result of gene duplication. The corresponding components of the pcc gene clusters that are adjacent in the genomes of G. daltonii FRC-32, G. metallireducens GS-15, and G. uraniireducens Rf4 are <64% identical at the amino acid sequence level, suggesting that they are unlikely to have arisen as a result of gene duplication (Figures 1A , 2 , 3 ). Figure 1 Identified pcc gene cluster in Geobacter spp. (A) and other bacteria (B) . The genes encoding different type of proteins are labeled with different colors: green/the Pcc porin-like outer-membrane proteins; red/the Pcc periplasmic c -type cytochromes ( c -Cyts); purple/the Pcc outer-membrane c -Cyts; black/the cytoplasmic membrane c -Cyts; pink/the periplasmic c -Cyts; yellow/the porin-like outer-membrane c -Cyt; light blue/transcriptional factors; dark blue/chitinase and orange/hypothetic proteins. The numbers displayed above the gene clusters are part of their locus tags whose letter parts are displayed at left side of the gene clusters with exception that PCA is displayed for the gene clusters identified from G . sulfurreducens PCA. The numbers for the pcc genes are in red, green, blue or purple, while the numbers for the gene associated with pcc genes are in black. Gbem: Geobacter bemidjiensis Bem; Geob : Geobacter sp. FRC-32; Glov: Geobacter lovleyi SZ; Gmet: Geobacter metallireducens GS-15; Gura: Geobacter uraniireducens Rf4; GM18: Geobacter sp. M18; GM21: Geobacter sp. M21; A2cp1: Anaeromyxobacter dehalogenans 2CP-1; Adeh: A. dehalogenans 2CP-C; AnaeK: Anaeromyxobacter sp. K; Kuster: Candidatus Kuenenia stuttgartiensis; Dacet: Denitrovibrio acetiphilus DSM 12809; Selin: Desulfurispirillum indicum S5; DaAHT2: Desulfurivibrio alkaliphilus AHT2; Dester: Desulfurobacterium thermolithotrophum DSM 11699; Dace: Desulfuromonas acetoxidans DSM 684; IALB: Ignavibacterium album JCM 16511; and Theam: Thermovibrio ammonificans HB-1. Figure 2 Phylogenetic analyses of the Pcc porin-like outer-membrane proteins (A) and periplasmic c -type cytochromes (B) . The phylogenetic trees were constructed with MEGA6 and confidence levels are indicated in the major nodes by the bootstrap values (%) in red. The numbers in the parenthesis next to the locus tags are the numbers of their predicted trans-outer membrane motifs of the porin-like outer-membrane proteins (A) and heme-binding motifs of the periplasmic c -type cytochromes (B) . The phylogenetic groups of the Pcc porin-like outer-membrane proteins (A) and periplasmic c -type cytochromes (B) are indicated by Roman numerals. The trees are not drawn to scale. Figure 3 Phylogenetic analyses of the Pcc outer-membrane c -type cytochromes . The phylogenetic trees were constructed with MEGA6 and confidence levels are indicated in the major nodes by the bootstrap values (%) in red. The numbers in the parenthesis next to the locus tags are the numbers of their predicted typical (CX 2 CH)/atypical (CX 3-15 CH) heme-binding motifs of the c -type cytochromes. The tree is not drawn to scale. The pcc gene clusters were also found in the genomes of a group of phylogenetically and functionally diverse bacteria. These include the dissimilatory Fe(III)-reducing bacteria Anaeromyxobacter dehalogenans 2CP-1, A. dehalogenans 2CP-C, Anaeromyxobacter sp. K, and Desulfuromonas acetoxidans DSM 684; the selenate [SeO 2− 4 ]- and selenite [SeO 2− 3 ]-respiring bacterium Desulfurispirillum indicum S5; the elemental sulfur (S 0 )-reducing bacteria Desulfurivibrio alkaliphilus AHT2, Desulfurobacterium thermolithotrophum DSM 11699, and Thermovibrio ammonificans HB-1; the anammox bacterium Candidatus Kuenenia stuttgartiensis; the dissimilatory nitrate-reducing bacterium Denitrovibrio acetiphilus DSM 12809 and the moderately thermophilic chemoheterotrophic bacterium Ignavibacterium album JCM 16511 (Figure 1B ) (Roden and Lovley, 1993 ; L'Haridon et al., 1998 ; Myhr and Torsvik, 2000 ; Narasingarao and Haggblom, 2007 ; Sorokin et al., 2008 ; Iino et al., 2010 ; Rauschenbach et al., 2011a ; Giovannelli et al., 2012 ; Nissen et al., 2012 ; Speth et al., 2012 ). Only one pcc gene cluster was found in the genomes of each of these searched microorganisms (Figure 1B ). Notably, the pcc gene clusters identified from three Anaeromyxobacter spp. and Candidatus Kuenenia stuttgartiensis have only two genes that are predicted to encode a porin-like outer-membrane protein and a periplasmic c -Cyt with 12 or 14 heme-binding motifs, respectively, which is similar to some of the mtr gene clusters (Figures 1 , 2A,B ). The mtr genes found in other metal-reducing bacteria, such as Shewanella spp., are also clustered in the bacterial genomes (Fredrickson et al., 2008 ; Liu et al., 2012 ; Shi et al., 2012b ). Some of the mtr gene clusters, especially those involved in Fe(II) oxidation, lack the genes encoding the outer membrane c -Cyts, such as mtrC gene (Jiao and Newman, 2007 ; Hartshorne et al., 2009 ; Liu et al., 2012 ; Shi et al., 2012b ; Emerson et al., 2013 ). In S. oneidensis MR-1, the MtrAB or MtoA/MtrB complex alone that possess only 10 hemes can transfer electrons across the outer membrane for extracellular reduction of Fe(III) (Coursolle and Gralnick, 2010 ; Liu et al., 2012 ). Moreover, MtrAB can be co-purified without MtrC and the purified MtrAB complex reconstituted in proteoliposomes can transfer electrons across the lipid-bilayer of the liposomes (Hartshorne et al., 2009 ; White et al., 2013 ). The 10-heme MtrA polypeptide contains 333 amino acid residues with a calculated molecular mass of 36.0 kDa. Insights into the MtrA structure, determined by small-angle X-ray scattering and analytical ultracentrifugation, suggest that this protein is rod-shaped with length of 104 Å (Firer-Sherwood et al., 2011 ). The Pcc periplasmic c -Cyts of Anaeromyxobacter spp. and the Candidatus Kuenenia stuttgartiensis possess 338–344 amino acid residues with calculated molecular masses of 36.9–39.2 kDa and are predicted to have 12–14 hemes. If they are structurally similar to MtrA, the Pcc periplasmic c -Cyts of Anaeromyxobacter spp. and the Candidatus Kuenenia stuttgartiensis along with their respective porin-like outer-membrane proteins could potentially provide a span sufficient to transfer electrons across the outer membrane even without the outer-membrane c -Cyt counterparts. In all cases, a gene encoding a porin-like outer-membrane protein is always associated with a gene encoding a periplasmic c -Cyt and in most cases a gene encoding an outer-membrane c -Cyt (Figures 1A,B ). The pcc -associated genes Most of the mtr gene clusters contain additional genes encoding c -Cyt (Fredrickson et al., 2008 ; Shi et al., 2011 , 2012b ; Liu et al., 2012 , 2013 ). These Mtr-associated c -Cyts are involved in extracellular reduction of Fe(III) oxides on the exterior side of the outer membrane, such as OmcA and UndA; quinone/quinol redox cycling in the cytoplasmic membrane, such as CymA; and probably electron transfer in the periplasm, such as MtrK/MtoD (Lower et al., 2009 ; Reardon et al., 2010 ; Shi et al., 2011 , 2012a , b ; Edwards et al., 2012 , 2014 ; Marritt et al., 2012a , b ; McMillan et al., 2012 , 2013 ). They are key components of the pathways that collectively mediate electron transfer between quinone/quinol pool in the cytoplasmic membrane and the extracellular electron donors or acceptors, whose electron transfer processes are spanning the entire width of bacterial cell envelope (Liu et al., 2012 ; Richardson et al., 2012 ; Shi et al., 2012a , b ). Similar to the mtr gene clusters, additional genes encoding c -Cyt, nearly all of which possessed >one heme-binding motifs, often associated with the pcc gene clusters (Figure 1 ). Given the lack of sequence conservation among Geobacter c -Cyts as well as among the Pcc c -Cyts (Butler et al., 2010 ; Liu et al., 2014 ), it is not surprising that these additional c -Cyts show no apparent sequence similarity to the c -Cyts that are proposed to be involved in quinol oxidation in the cytoplasmic membrane, electron transfer in the periplasm or extracellular reduction of Fe(III) in Geobacter spp. (Lovley, 2006 , 2012 ). Lack of sequence conservation among the Pcc c -Cyts suggests that different Pcc complexes may interact with different periplasmic and outer-membrane c -Cyts for intermolecular electron transfer (Liu et al., 2014 ). Thus, it is reasonable to hypothesize that some of these Pcc-associated c -Cyts are also key components of pathways that transfer electrons between the quinone/quinol pool in the cytoplasmic membrane and substrates external to the bacterial cells. Consistent with this speculation, the c -Cyts encoded by the pcc -associated genes are predicted to be localized in the cytoplasmic membrane (Dester_ 0351, Geob_1683, and Theam_0871), the periplasm (IALB_1840 and GSU_2645) and the outer membrane (GM18_3465) (Figure 1 ), similar to the c -Cyts encoded by the mtr -associated genes (Shi et al., 2012b ). Notably, the predicted cytoplasmic membrane c -Cyts Dester_0351 of D. thermolithotrophum DSM 11699 and Theam_0871 of T. ammonificans HB-1are 79% identical and each protein contains six typical heme-binding motifs (CX 2 CH), two atypical heme-binding motif (CX 3−5 CH) of the c -Cyt and 24 histidine residues in which 16 are the putative ligands for c -type hemes. Furthermore, BLAST search identifies low sequence similarity between Dester_0351/Theam_0871 and cytochrome b subunits of bacterial formate dehydrogenases that also use histidine residues as heme ligands (Gross et al., 2004 ). Thus, some of the extra histidine residues found in the amino acid sequences of Dester_0351 and Theam_0871 may be the ligands for the b -type hemes. The cytochrome b subunits of bacterial formate dehydrogenase are the cytoplasmic membrane proteins with quinone reduction activity in which the hemes are involved (Gross et al., 2004 ). Previously, we found that the genes encoding the c -Cyts with sequence similarity to cytochrome b subunits of bacterial formate dehydrogenase (MtrH/MtoC) are part of the mtr gene clusters where they are proposed to be involved in quinone/quinol cycling in the cytoplasmic membrane (Shi et al., 2012b ). Although Dester_0351/Theam_0871 and MtrH/MtoC share very low sequence identity (<17%) in which most identity is in the regions of their heme-binding motifs, Dester_0351 and Theam_0871 may also be involved in quinone/quinol cycling in the cytoplasmic membrane, similar to MtrH and MtoC. Interestingly, GM18_3465 of Geobacter sp. M18 is predicted to be a porin-like, 10-heme and outer-membrane c -Cyt with 21 trans-outer membrane motifs by the Hidden Markov Model with the posterior decoding method using a dynamic programming algorithm. The posterior decoding method using a dynamic programming algorithm is better in prediction than that of Viterbi and N-best algorithms (Bagos et al., 2004a , b ), which also predict that GM18_3465 is a porin-like outer-membrane protein with different trans-outer membrane motifs. All the heme-binding motifs are found in the long solvent-exposed loops: five heme-binding motifs in loop 5, two each in loop 9 and 11 and one in loop 10 (Figure 4 ). Although the porin-cytochrome is a common mechanism shared by different groups of Gram-negative bacteria for transferring electrons across the outer membrane, all previously identified and characterized porin-cytochrome proteins are complexes that each consists or is predicted to consist of a porin-like outer-membrane protein, a periplasmic c -Cyt and in most cases an outer-membrane c -Cyt (Hartshorne et al., 2009 ; Liu et al., 2012 , 2014 ; Richardson et al., 2012 ; Shi et al., 2012b ; White et al., 2013 ). The current porin-cytochrome model proposes that the porin-like outer-membrane proteins function as scaffolds through which the c -Cyts are inserted (Richardson et al., 2012 ; Liu et al., 2014 ). To the best of our knowledge, GM18_3465 is the first reported case of a putative porin-like outer-membrane c -Cyt with multiple hemes. Based on the current porin-cytochrome model, we propose that GM18_3465 may contain a cytochrome domain and a trans-outer membrane domain into which the cytochrome domain may also be inserted for mediating trans-outer membrane electron transfer. A key question is whether GM18_3465 alone can transfer electrons across the outer membrane. Figure 4 The amino acid sequence of GM18_3465 of Geobacter sp. M18 . The N-terminus and predicted short solvent-exposed loops are in green, the predicted trans-outer membrane motifs are in red, and predicted long solvent-exposed loops are in blue. The 21 trans-outer membrane motifs are numbered sequentially and the numbers are displayed in red and Arabic numerals and above the amino acid sequence. The long solvent-exposed loops are numbered sequentially and numbers are displayed in blue and Roman numerals and above the amino acid sequence. The heme-binding motifs are underlined and in italic. Other genes associated with the identified pcc gene clusters include those encoding hypothetical proteins and a putative chitinase, in addition to the genes encoding the transcription factors OrfR and OrfS in the omcB -associated gene clusters of G. sulfurreducens PCA, where OrfR regulates expression of ombB-omaB-omcB gene cluster (Leang and Lovley, 2005 ) and OrfS may regulate expression of ombC-omaC-omcC gene cluster (Figure 1 ). The pcc porin-like outer-membrane proteins and periplasmic c -Cyts In the Mtr system, all identified porin-like outer-membrane proteins (i.e., MtrB/MtoB) contain 28 predicted trans-outer membrane motifs, including PioB of the phototrophic Fe(II)-oxidizing bacterium Rhodopseudomonas palustris TIE-1. All identified periplasmic c -Cyts (i.e., MtrA/MtoA/PioA) possess or are predicted to possess 10 hemes (Pitts et al., 2003 ; Shi et al., 2005 , 2012b ; Jiao and Newman, 2007 ; Fredrickson et al., 2008 ; Hartshorne et al., 2009 ; Liu et al., 2012 ; White et al., 2013 ). In the Pcc system, predicted trans-outer membrane motifs found in the porin-like outer-membrane proteins varied, ranging from 18 to 22, while predicted heme-binding motifs in the periplasmic c -Cyts also varied, ranging from 5 to 14 (Figures 2A,B ) (Liu et al., 2014 ). Consequently, the Pcc porin-like outer-membrane proteins are much smaller than MtrB/MtoB/PioB and are predicted to form the pores on the outer membrane that may also be smaller than those formed by MtrB/MtoB/PioB. The amino acid sequence identity among Pcc porin-like outer-membrane proteins and periplasmic c -Cyts also vary greatly, ranging from 8 to100% and from 13 to100%, respectively (Tables S1 – S4 ). Despite the sequence differences, phylogenetic analyses revealed that both Pcc porin-like outer-membrane proteins and periplasmic c -Cyts were clustered into five different groups, except IALB_1839 and IALB_1838 of I. album JCM 16511 and Kuster_4034 and Kuster_4025 of Candidatus Kuenenia stuttgartiensis, which are distantly related to the rest of their respective counterparts (Figures 2A,B ). Within each phylogenetic group, the proteins are often more closely related to each other than to those in the different groups (Figures 2A,B and Tables S1 – S4 ). Remarkably, the porin-like outer-membrane protein and periplasmic c -Cyt from the same gene cluster are always found in similar corresponding phylogenetic groups. For instance, OmbB and OmbC of G. sulfurreducens PCA are in Group I of the porin-like outer-membrane proteins, while OmaB and OmaC of G. sulfurreducens PCA are in Group I of the periplasmic c -Cyts (Figures 1 , 2A,B ). These results suggest that within their respective phylogenetic groups, the Pcc porin-like outer-membrane proteins and periplasmic c -Cyts may be co-evolved. The Pcc porin-like outer-membrane proteins and periplasmic c -Cyts from Geobacter spp. are found in their respective phylogenetic Group I and V. The Pcc porin-like outer-membrane proteins and periplasmic c -Cyts of Desulfurivibrio alkaliphilus AHT2, Desulfuromonas acetoxidans DSM 684, and Desulfurispirillum indicum S5 are placed in their respective phylogenetic Group II, while those of Denitrovibrio acetiphilus DSM 12809, Desulfurobacterium thermolithotrophum DSM 11699, and Thermovibrio ammonificans HB-1 are in the Group III. Phylogenetic Group IVs include Pcc porin-like outer-membrane proteins and periplasmic c -Cyts of the Anaeromyxobacter spp. analyzed (Figure 2 ). As discussed in the Overview section, the pcc gene clusters of A. dehalogenans 2CP-1, A. dehalogenans 2CP-C, and Anaeromyxobacter sp. K and Candidatus Kuenenia stuttgartiensis lack the genes encoding the outer-membrane c -Cyts. Lack of the outer-membrane c -Cyts maybe one of the reasons that the Pcc periplasmic c -Cyts associated with these bacteria are larger and have more heme-binding motifs than rest of the periplasmic c -Cyts. An exception is IALB_1838 of I. album JCM 16511 that is the largest Pcc periplasmic c -Cyt identified to date, which possesses 400 amino acid residues and 14 heme-binding motifs (Figure 2B ). With extra hemes, these larger periplasmic c -Cyts could transfer electrons across the outer-membrane in the absence of outer-membrane c -Cyts. The pcc outer-membrane c -Cyts In the Mtr system, all the outer-outer membrane c -Cyts (i.e., MtrC) have or are predicted to have10 hemes (Shi et al., 2006 , 2012b ; Hartshorne et al., 2007 ; Fredrickson et al., 2008 ; Clarke et al., 2011 ). In the Pcc system, the typical heme-binding motifs (i.e., CX 2 CH) found in the outer-membrane c -Cyts varied from 1 to 15 (Figure 3 ) (Liu et al., 2014 ). We noticed that each of GM18_3461 and GM18_3467 of Geobacter sp. M18 contained only one typical heme-binding motif, while each of their corresponding periplasmic c -Cyts had five typical heme-binding motifs (Figures 2B , 3 ). The combined 6 typical hemes associated with these proposed protein complexes would not form the heme-based electron conduits that are sufficiently long to span entire width of a typical Gram-negative bacterial outer membrane. Further analyses revealed that in addition to a typical heme-binding motif, each of GM18_3461 and GM18_3467 contained five atypical binding motifs with sequences of CX 3−4 CH, which were previously confirmed to bind heme covalently (Stevens et al., 2004 ). Thus, these c -Cyts may bind up to 6 hemes covalently. Given that the 10-heme MtrA/MtoA c -Cyt alone could transfer electrons across the outer membrane and the lipid-bilayer of proteoliposomes (Hartshorne et al., 2009 ; Liu et al., 2012 ; White et al., 2013 ), the Pcc protein complex of Geobacter sp. M18 that is predicted to consist of a 5-heme periplasmic c -Cyt and a 6-heme outer-membrane c -Cyt should have enough hemes to form the conduits for the efficient transfer of electrons across the outer membrane. In addition to GM18_3461 and GM18_3467 of Geobacter sp. M18, atypical heme-binding motifs (i.e., CX 3−15 CH) are also found in the Pcc outer-membrane c -Cyts GSU_2724 of G . sulfurreducens PCA, Gura_1837 of G. uraniireducens Rf4, Geob_1686 of Geobacter sp. FRC-32, and IALB_1837 of I. album JCM 16511 (Figure 3 ). To date, no atypical heme-binding motif has been found in the Pcc periplasmic c -Cyts or the Mtr c -Cyts. It should be noted that atypical heme-binding motifs are also found in other outer-membrane c -Cyts, such as OmcZ of G. sulfurreducens PCA (Inoue et al., 2010 ). It remains to be determined whether the atypical heme-binding motifs with the sequence of CX >4 CH can also covalently bind hemes. The identity among the Pcc outer-membrane c -Cyts varies from 4 to 100% (Tables S5 , S6 ). The Pcc outer membrane c -Cyts are not, however, clustered into distinct phylogenetic groups corresponding to those found in the Pcc porin-like outer-membrane proteins and periplasmic c -Cyts (Figures 2 , 3 ). The lack of phylogenetic groups similar to those found in other Pcc components are attributed to the extreme sequence diversity among the Pcc outer-membrane c -Cyts. This demonstrates that the Pcc porin-like outer-membrane proteins/periplasmic c -Cyts and outer-membrane c -Cyts are unlikely co-evolved, which is in contrast to the apparent co-evolution of the Mtr porin-like outer-membrane proteins, periplasmic c -Cyts and outer-membrane c -Cyts (Shi et al., 2012b ). In the Mtr system, without the outer-membrane c -Cyt, the porin-like outer-membrane protein and periplasmic c -Cyt can work as a single functional unit for mediating electron transfer across the outer membrane (Hartshorne et al., 2009 ; Liu et al., 2012 ; Shi et al., 2012b ; White et al., 2013 ). Consistent with these previous findings in the Mtr system are the observations of apparent co-evolution only between Pcc porin-like outer-membrane proteins and periplasmic c -Cyts within their phylogenetic groups and the pcc gene clusters encoding only Pcc porin-like outer-membrane proteins and periplasmic c -Cyts in this study. Biological implications In addition to Geobacter spp., other Fe(III)-reducing bacteria identified with the Pcc proteins included A. dehalogenans 2CP-1, A. dehalogenans 2CP-C, and Anaeromyxobacter sp. K and D. acetoxidans DSM 684. Notably, the abundance of the Pcc periplasmic c -Cyt Adeh_3392 of A. dehalogenans 2CP-C increased under Mn(IV)-reducing conditions, compared to that when Fe(III)-citrate was provided as a terminal electron acceptor (Nissen et al., 2012 ), which is consistent with the proposed role of Adeh_3392 in extracellular electron transfer by A. dehalogenans 2CP-C. Identification of the pcc gene cluster in D. acetoxidans DSM 684 is also consistent with previous findings that D. acetoxidans DSM 684 was phylogenetically related to G. metallireducens and c -Cyts were involved in reduction of solid-phase Fe(III) or Mn(IV) oxides by D. acetoxidans DSM 684 (Roden and Lovley, 1993 ). Among other bacteria with pcc gene clusters, only D. alkaliphilus AHT2 and I. album have been tested for their ability to grow on Fe(III) or Mn(IV) oxides and were found to be unable to use either as a terminal electron acceptor (Sorokin et al., 2008 ; Iino et al., 2010 ). It remains unknown whether the remaining bacteria with pcc gene clusters can use Fe(III) and Mn(III, IV) oxides as the terminal electron acceptors. However, it should be pointed out that pcc gene clusters may not be restricted to mediation of extracellular reduction of Fe(III) and Mn(III, IV) oxides. Mtr proteins are directly involved in extracellular reduction of dimethyl sulfoxide and extracellular oxidation of Fe(II), in addition to extracellular reduction of Fe(III) and Mn(III, IV) oxides (Gralnick et al., 2006 ; Jiao and Newman, 2007 ; Liu et al., 2012 , 2013 ; Shi et al., 2012b ). Similarly, the Pcc proteins found in the bacteria that are not known to reduce Fe(III) or Mn(III, IV) oxides may also be involved in extracellular electron transfer reactions with other substrates. A common trait shared by D. alkaliphilus AHT2, D. thermolithotrophum DSM 11699, and T. ammonificans HB-1 is their utilization of insoluble S 0 as the terminal electron acceptor (L'Haridon et al., 1998 ; Sorokin et al., 2008 ; Giovannelli et al., 2012 ). The ability of these bacteria to reduce S 0 extracellularly may be attributed in part to their possession of the Pcc proteins. The SeO 2− 4 - and SeO 2− 3 -respiring bacterium D. indicum S5 reduces water soluble SeO 2− 4 and SeO 2− 3 to water insoluble selenium (Se 0 ), which forms Se 0 -containing granules outside the bacterial cells (Narasingarao and Haggblom, 2007 ; Rauschenbach et al., 2011a ). Although they were once thought to be localized in the periplasm, the terminal reductases for SeO 2− 4 and SeO 2− 3 in D. indicum S5 have not been identified (Rauschenbach et al., 2011b ). Formation of Se 0 -containing granules outside the bacterial cells after reduction of SeO 2− 4 and SeO 2− 3 by D. indicum S5 and existence of the pcc gene cluster in D. indicum S5 collectively suggest that the reduction of SeO 2− 4 and SeO 2− 3 may occur extracellularly. Hence, we suggest that Pcc proteins Selin_2480, Selin_2481, and Selin_2482 are associated with the outer membrane where they catalyze extracellular reduction of SeO 2− 4 and SeO 2− 3 . Extracellular reduction of SeO 2− 4 and SeO 2− 3 will avoid accumulation of insoluble Se 0 intracellularly, which may be detrimental to the cells of D. indicum S5. This is very similar to microbial extracellular reduction of chromium and uranium, which is considered as a detoxification mechanism (Belchik et al., 2011 ; Cologgi et al., 2011 ). It should be pointed out that unlike the mtr gene clusters that are also found in genomes of the Fe(II)-oxidizing bacteria (Jiao and Newman, 2007 ; Liu et al., 2012 , 2013 ; Shi et al., 2012b ; Emerson et al., 2013 ), to date no pcc gene cluster has been identified in any known Fe(II)-oxidizing bacterium."
} | 8,903 |
37688565 | PMC10505500 | pmc | 8,884 | {
"abstract": "Abstract Summary The chem16S package combines taxonomic classifications of 16S rRNA gene sequences with amino acid compositions of prokaryotic reference proteomes to generate community reference proteomes. Taxonomic classifications from the RDP Classifier or data objects created by the phyloseq R package are supported. Users can calculate and visualize a variety of chemical metrics in order to explore the effects of redox, salinity, and other physicochemical variables on the genomic adaptation of protein sequences at the community level. Availability and implementation Development of chem16S is hosted at https://github.com/jedick/chem16S . Version 1.0.0 is freely available from the Comprehensive R Archive Network (CRAN) at https://cran.r-project.org/package=chem16S .",
"conclusion": "4 Conclusions The merger of taxonomic and genomic information in community reference proteomes allows microbial communities to be represented in chemical terms. chem16S makes it easy to use taxonomic classifications produced by the RDP Classifier or OTU/ASV tables in “phyloseq-class” objects to calculate and visualize chemical metrics for community reference proteomes and thereby to gain new insight into the role of physicochemical factors in genomic adaptation to environments.",
"introduction": "1 Introduction Chemical analysis of protein sequences can reveal new aspects of genomic adaptation. Carbon oxidation state ( Z C ) represents the degree of oxidation of carbon atoms that results from bonds with other atoms (i.e. H, N, O, and S in the primary sequences of proteins). Protein Z C tends to be lower in the genomes of methanogens that inhabit anoxic environments compared to those that are occasionally found in oxygenated environments ( Dick et al. 2023 ). This suggests that geochemical relative stability models are applicable to genomic variation, but identifying the role of the environment is complicated by the cosmopolitan nature of many organisms. Microbial communities represent a localized collection of genomes. Recently we described the combination of taxonomic abundances from the Ribosomal Database Project (RDP) Classifier ( Wang et al. 2007 ) with reference proteomes derived from the Reference Sequence (RefSeq) database of the National Center for Biotechnology Information (NCBI) ( O’Leary et al. 2016 ) to generate community reference proteomes ( Dick and Tan 2023 ). This allows new tests of hypotheses about genomic adaptation. For instance, the thermodynamic prediction that protein Z C is positively correlated with environmental oxidation–reduction potential has been confirmed for bacterial communities at a global scale ( Dick and Meng 2023 ). Community reference proteomes are inferred by taxonomic comparisons with genomic databases rather than directly derived from protein extracts or community DNA (i.e. metaproteomes and metagenomes). Despite this, trends of Z C for community reference proteomes are mostly consistent with those for protein sequences inferred from shotgun metagenomes ( Dick and Tan 2023 ). Different natural abundances and extraction efficiencies for cytoplasmic and membrane proteins in metaproteomic experiments can explain in part the relatively weak correspondence of Z C between metaproteomes and community reference proteomes ( Dick and Meng 2023 ). For these reasons, metrics calculated for community reference proteomes are indicators of genomic variation rather than protein expression. The code that was developed in our recent studies was consolidated to create the chem16S package. An important new addition to the package are reference proteomes derived from the Genome Taxonomy Database (GTDB) ( Parks et al. 2022 ). Using the GTDB for both taxonomic classification and reference proteomes avoids the uncertain mapping between the RDP training set and the NCBI taxonomy. Moreover, functions have been added to chem16S to process data objects generated with the phyloseq package ( McMurdie and Holmes 2013 ), allowing for seamless calculation and visualization of chemical metrics for microbial communities by more researchers."
} | 1,023 |
28943999 | PMC5586202 | pmc | 8,885 | {
"abstract": "Summary \n Animals often display a marked tendency to return to previously visited locations that contain important resources, such as water, food, or developing brood that must be provisioned. A considerable body of work has demonstrated that this tendency is strongly expressed in ants, which exhibit fidelity to particular sites both inside and outside the nest. However, thus far many studies of this phenomena have taken the approach of reducing an animal's trajectory to a summary statistic, such as the area it covers. Using both simulations of biased random walks, and empirical trajectories from individual rock ants, Temnothorax albipennis , we demonstrate that this reductive approach suffers from an unacceptably high rate of false negatives. To overcome this, we describe a site‐centric approach which, in combination with a spatially‐explicit null model, allows the identification of the important sites towards which individuals exhibit statistically significant biases. Using the ant trajectories, we illustrate how the site‐centric approach can be combined with social network analysis tools to detect groups of individuals whose members display similar space‐use patterns. We also address the mechanistic origin of individual site fidelity; by examining the sequence of visits to each site, we detect a statistical signature associated with a self‐attracting walk – a non‐Markovian movement model that has been suggested as a possible mechanism for generating individual site fidelity.",
"introduction": "Introduction The phenomenon of ‘recurrence’, in which the movement of an individual is biased towards a set of previously visited locations, is widespread in the animal kingdom (Gonzalez, Hidalgo & Barabasi 2008 ; Boyer, Crofoot & Walsh 2011 ; Schreier & Grove 2014 ). Recent work comparing the mobility patterns of humans and vervet monkeys has shown that recurrence is a fundamental statistical property common to both (Boyer, Crofoot & Walsh 2011 ). In humans, examples of important locations at which recurrence is most strongly expressed include homes, workplaces, restaurants and the transit routes that connect them (Sun et al . 2013 ). In other non‐human animals these locations might take the form of watering holes, foraging patches, leks, or nesting areas where there are brood that must be provisioned regularly. Depending upon the study system and the context, preferential bias towards previously visited locations has been variously labelled recurrence (Gonzalez, Hidalgo & Barabasi 2008 ; Song et al . 2010 ; Boyer, Crofoot & Walsh 2011 ), recursion (Bar‐David et al . 2009 ; Benhamou & Riotte‐Lambert 2012 ; Fagan et al . 2013 ; Berger‐Tal & Bar‐David 2015 ), site tenacity (Hahn & Maschwitz 1985 ), site allegiance (Dejean & Turillazzi 1992 ), site recognition (Salo & Rosengren 2001 ), site fidelity (Lamb & Ollason 1994 ; Schwarzkopf & Alford 2002 ; Giuggioli & Bartumeus 2012 ), spatial fidelity (Sendova‐Franks & Franks 1995 ), ‘ortstreue’ (Rosengren & Fortelius 1986 ) and route fidelity (Rosengren 1971 ). Recursive movement has been particularly well documented in the social insects – ants, bees, wasps and termites – where the phenomenon is most often referred to as site fidelity. Social insects show strong site fidelity both outside the nest (Traniello, Fourcassié & Graham 1991 ; Fourcassié & Traniello 1994 ; Lamb & Ollason 1994 ; Schatz, Lachaud & Beugnon 1995 ; Beverly et al . 2009 ; Salo & Rosengren 2001 ), and within it (Seeley 1982 ; Sendova‐Franks & Franks 1995 ; Jandt & Dornhaus 2009 ; Frohschammer & Heinze 2009 ; Baracchi et al . 2010 ; Jeanson 2012 ). For example, wood ant workers show a strong tendency to re‐use one of the multiple foraging trunk‐trails emanating from the nest mound (Rosengren , 1971 , 1977 ), a preference that can persist over several seasons (Rosengren 1971 ). Inside the nests of ants and bees, there is a strong division of labour, in which work is divided into discrete tasks that are spatially segregated into different zones, with each zone being populated by a particular set of worker task specialists (Seeley 1982 ; Mersch, Crespi & Keller 2013 ; Baracchi & Cini 2014 ). This division of labour is thought to increase colony productivity, and has led to social insects being ecologically dominant in many ecosystems (Oster & Wilson 1978 ). Hence the origin and quantification of individual spatial fidelity have been and continue to be, of considerable interest to scientists interested in the organization of animal societies. Here we study site fidelity in colonies of the rock ant, Temnothorax albipennis . We chose this species because the workers exhibit site fidelity within the nest (Sendova‐Franks & Franks 1993 ), and because the nearly two‐dimensional geometry of natural rock ant nests – flat cavities between rock layers – makes them ideal for studies of spatial movement. A variety of methods is now available for identifying different spatio‐temporal components of site fidelity. For example, there has been a recent growth in methods for identifying routine movement patterns, such as periodic returns to previously‐visited locations (Bar‐David et al . 2009 ; Riotte‐Lambert, Benhamou & Chamaillé‐Jammes 2013 ; Péron et al . 2016 ), or repetitive sequences of visits to particular locations (Riotte‐Lambert, Benhamou & Chamaillé‐Jammes 2017 ). Similarly, there are several tools to evaluate whether there is a stable home range over which the animal typically roams, or a core area to which it frequently returns, such as comparing the degree of spatial overlap between consecutive time periods (Cooper 1978 ; Van Beest et al . 2013 ), or checking whether the time‐series of the total area that the animal covers (Van Moorter et al . 2009 ), or its net displacement (Börger, Dalziel & Fryxell 2008 ), saturate over time. Despite this plethora of techniques, many studies of within‐nest site fidelity in social insects still adopt a ‘reductive’ approach in which a complex spatio‐temporal object – an animal trajectory – is aggregated over time and space into a single summary statistic such as the area the trajectory covers (Jandt & Dornhaus 2009 ; Baracchi et al . 2010 ; Baracchi & Cini 2014 ). This preference may be derived from the nest wall severely circumscribing individual movement; as the total area covered and the net displacement of a physically bounded random walk both saturate over time, it is difficult for the above methods to distinguish between an agent that moves randomly within the nest, and one that has a preference for one (or several) parts of the nest. Hence, the primary motivation for the current study is to provide an analytical framework that can identify those individuals that exhibit site fidelity that can pinpoint the sites to which they are loyal, and that is robust to the presence of physical boundaries. The second motivation stems from the fact that existing measures of spatial fidelity are often based upon descriptions of the space use patterns of individuals (Sendova‐Franks & Franks 1993 ; Frohschammer & Heinze 2009 ; Baracchi et al . 2010 ; Benhamou & Riotte‐Lambert 2012 ; Mersch, Crespi & Keller 2013 ) or groups Baracchi & Cini ( 2014 ), rather than upon quantitative comparisons between the observed pattern and an absolute standard (but see Sendova‐Franks & Franks 1995 ; Jandt & Dornhaus 2009 ). In other words, rigorous hypothesis‐testing, involving comparisons between the observation and the expectation under the assumption of random movement, as predicted by a mathematical or statistical null model, has sometimes been lacking. In the first part of the paper, we present an extension of a recent site‐centric framework which has been developed for the analysis of human digital mobility traces (Crandall et al . 2010 ; Sun et al . 2013 ) and animal movement ecology (Boyer, Crofoot & Walsh 2011 ; Benhamou & Riotte‐Lambert 2012 ; Lyons, Turner & Getz 2013 ; Fagan et al . 2013 ; Berger‐Tal & Bar‐David 2015 ). Contrary to the traditional reductive approach in which the trajectory is reduced to a single summary statistic, in the site‐centric framework space is discretized into a regular grid, and the visitation statistics of a given individual for each site are analysed independently. In our extension, we demonstrate that sites to which individuals exhibit positive or negative biases can be identified by comparing these site‐visitation statistics with an absolute standard, provided by null model synthetic trajectories that exhibit no spatial biases. Further, using both simulations of biased random walks, and empirical analysis of ant trajectories, we show that this combined framework is more sensitive at identifying individuals that exhibit site fidelity than the traditional reductive approach. Whilst our understanding of the social organization of colonies of ants (Sendova‐Franks et al . 2010 ; Blonder & Dornhaus 2011 ; Jeanson 2012 ), bees (Naug & Smith 2007 ; Otterstatter & Thomson 2007 ), and other highly social species (Williams & Lusseau 2006 ; Drewe 2010 ), has been greatly advanced by the application of tools from network science, these tools are only just beginning to be applied to the spatial organization of these societies (see e.g. Mersch, Crespi & Keller 2013 ; Baracchi & Cini 2014 ; Richardson & Gorochowski 2015 ). Therefore, in the second part of the paper, we use the results of the site‐centric analysis of the ant trajectories to construct spatial networks in which each edge represents the spatial overlap in the site‐visitation patterns of two ants. We then show how modern network partitioning methods can be used to identify groups of ants with distinctive space use patterns. Recent theoretical modelling has shown that biologically interesting behaviours, such as the establishment of a territory, core area, or home range, can emerge when an individual's movement decisions are influenced by its historical movement patterns (Van Moorter et al . 2009 ; Foster, Grassberger & Paczuski 2009 ; Spencer 2012 ; Fagan et al . 2013 ; Boyer & Solis‐Salas 2014 ; Berger‐Tal & Bar‐David 2015 ; Merkle, Potts & Fortin 2017 ). Furthermore, there are now a range of methods for detecting such history dependence in real‐world animal movement data (Börger, Dalziel & Fryxell 2008 ; Bar‐David et al . 2009 ; Riotte‐Lambert, Benhamou & Chamaillé‐Jammes 2013 ; Merkle, Fortin & Morales 2014 ; Riotte‐Lambert, Benhamou & Chamaillé‐Jammes 2017 ; Péron et al . 2016 ). Two of the mechanisms for generating history‐dependent movement include internal (cognitive) memory, and external (chemical) signals deposited into the environment. Indeed, considering their exceedingly small (<1 mm 3 ) brains, rock ants exhibit impressive capacities for both internal (i.e. neuronal, McLeman, Pratt & Franks 2002 ; Stroeymeyt, Franks & Giurfa 2011 ; Bowens, Glatt & Pratt 2013 ), and external (i.e. pheromonal, Mallon & Franks 2000 ) memory storage formats. Therefore, the last part of this paper examines whether the trajectories of individual rock ants exhibit non‐Markovian properties, that is, whether movement decisions are history dependent.",
"discussion": "Discussion In this paper, we have presented an analytical framework that leverages the spatial and temporal information contained within an animal trajectory to identify important sites within the environment, identify groups of animals with distinctive space‐use patterns, and shed light on the mechanisms that underpin animal movement. We have also shown that this combined framework is considerably more sensitive than previous approaches which reduce a complex spatio‐temporal object – an animal trajectory – to a single summary statistic. This sensitivity is derived from the formal statistical hypothesis testing provided by comparisons between the original trajectories and the synthetic trajectories produced by the RW null model. It should be emphasized that this null model could profitably be combined with other site‐specific methods for quantifying local space‐use intensities, such as those of Benhamou & Riotte‐Lambert ( 2012 ) and Lyons, Turner & Getz ( 2013 ), to identify locations that are more intensively exploited or more frequently revisited than expected by chance alone. In addition to these methodological results, our application of the above combined framework to within‐nest ant trajectories also provided several novel biological conclusions. The first concerns the theory of ‘organisational immunity’ (Schmid‐Hempel & Schmid‐Hempel 1993 ; Stroeymeyt, Casillas‐Pérez & Cremer 2014 ), which predicts that animal societies in which there is a reproductive division of labour, should possess structural features – such as bottlenecks, or compartmentalization – that inhibit transmission of pathogens to the reproductive individuals. The spatial network analysis provides two lines of support for the presence of organizational immunity in T. albipennis . The first was that 21 of 23 colonies were segregated into two groups, with the group that contained the queen always being the group that was found closest to the biological centre of the colony, that is, the brood pile. The second was that even though queens were typically found at the centre of the nest, the sites to which they exhibited bias overlapped little with those of most workers, which led to them occupying peripheral positions on the spatial network. Thus, queens were spatially and socially isolated from the workers. In social insect colonies, it is typically the outside‐nest workers that are most likely to expose the colony to risk, for example, by bringing back a pathogen after a foraging trip outside the nest (Schmid‐Hempel & Schmid‐Hempel 1993 ). As here, the ‘other’ group overlapped little with the brood pile, and was instead concentrated around the nest entrance, it is likely that this group contained many such outside‐nest workers. Therefore, the compartmentalization of the colony into layered groups and the isolation of the queen within the innermost group could be interpreted as organizational features that reduce the exposure of the colony to pathogens. The second conclusion concerns the mechanisms responsible for generating site fidelity. Although the presence of site fidelity is well documented across a range of social insect species, to the best of our knowledge nothing is known about how individuals first establish and then maintain bias towards a set of important sites. The finding that the longer an ant dwells at a site the more quickly it will return after leaving it, indicates that ant movement is not compatible with a Markov movement model, or in other words, ‘history’ influences current behaviour. The observed statistical signatures appear consistent with a particular class of non‐Markov movement model, the so‐called self‐attracting walk, in which sites become progressively more attractive with each visit. Indeed, the self‐attracting walk has been proposed as a candidate mechanism that would allow an animal to establish and maintain fidelity towards a set of important sites (Tan et al . 2001 ; Foster, Grassberger & Paczuski 2009 ). However, it is important to note two caveats. First, this association is a correlation, so it cannot be claimed that the long site dwell times cause short returns. Second, whilst the presence of a statistical signature of a self‐attracting walk indicates that this may underly the generation of site fidelity in T. albipennis ants, this result does not say anything about the nature of the ‘memory’ that allows the reinforcement to be brought about. Nevertheless, there are at least two (potentially complimentary) candidate mechanisms, namely, chemical pheromones deposited onto the substrate, and internal place memory. As rock ants have evolved sophisticated strategies for chemical marking (Mallon & Franks 2000 ), and navigation strategies (McLeman, Pratt & Franks 2002; Bowens, Glatt & Pratt 2013), future research should concentrate on elucidating their contributions to the generation of site fidelity. In this paper, we have outlined a combined framework for identifying the sites to which individuals are attracted, and for identifying groups of individuals that share a common set of sites to which they are attracted. We hope that the clear advantages of the site‐centric framework over traditional reductive approaches, will encourage others to delve further into the mechanisms that govern animal movement."
} | 4,136 |
33955585 | PMC11468872 | pmc | 8,887 | {
"abstract": "Abstract Coatings with low sliding angles for liquid drops have a broad range of applications. However, it remains a challenge to have a fast, easy, and universal preparation method for coatings that are long‐term stable, robust, and environmentally friendly. Here, a one‐step grafting‐from approach is reported for poly(dimethylsiloxane) (PDMS) brushes on surfaces through spontaneous polymerization of dichlorodimethylsilane fulfilling all these requirements. Drops of a variety of liquids slide off at tilt angles below 5°. This non‐stick coating with autophobicity can reduce the waste of water and solvents in cleaning. The strong covalent attachment of the PDMS brush to the substrate makes them mechanically robust and UV‐tolerant. Their resistance to high temperatures and to droplet sliding erosion, combined with the low film thickness (≈8 nm) makes them ideal candidates to solve the long‐term degradation issues of coatings for heat‐transfer surfaces."
} | 240 |
34522924 | null | s2 | 8,889 | {
"abstract": "Alkenyl phenols are utilized by nature in the construction of one of the most important biopolymers, lignin. Using similar building blocks, an array of distinct structures can be formed by selective dimerization of the starting phenols to form lignans, neolignans, oxyneolignans, and norlignans. Given the multitude of possible outcomes, many methods have been reported to affect the desired bond formations and access these biologically relevant scaffolds. The most biomimetic of these methods, discussed here, involve the unprotected phenols undergoing oxidative bond formation that proceeds "
} | 148 |
26441143 | null | s2 | 8,891 | {
"abstract": "The alkyne is an important functionality widely used in material science, pharmaceutical science, and chemical biology, but the importance of this functionality is contrasted by the very limited number of enzymes known to be involved in alkyne biosynthesis. We recently reported the first known carrier protein-dependent pathway for terminal alkyne formation, and in silico analysis suggested that this mechanism could be widespread in bacteria. In this paper, we screened additional homologous gene cassettes presumed to be involved in alkyne biosynthesis using both in vitro biochemical study and an E. coli-polyketide synthase (PKS) reporting system for in vivo analysis. We discovered and characterized a new terminal alkyne biosynthetic pathway comprised of TtuA, -B, and -C from Teredinibacter turnerae T7901. While the acyl-CoA ligase homologue (TtuA) demonstrated promiscuity in the activation and loading of medium-chain fatty acids onto the carrier protein (TtuC), the desaturase homologue (TtuB) showed stringent substrate specificity toward C10 fatty acyl moieties. In addition, TtuB was demonstrated to be a bifunctional desaturase/acetylenase that efficiently catalyzed two sequential O2-dependent dehydrogenation reactions. A novel terminal-alkyne bearing polyketide was further produced upon coexpression of ttuABC and a PKS gene in E. coli. The discovery and characterization of TtuA, -B, and -C provides us with a new bifunctional desaturase/acetylenase for mechanistic and structural study and expands the scarce enzyme inventory for the biosynthesis of the alkyne functionality, which has important applications in synthetic and chemical biology."
} | 416 |
36080933 | PMC9460910 | pmc | 8,893 | {
"abstract": "The required navigation performance (RNP) procedure is one of the two basic navigation specifications for the performance-based navigation (PBN) procedure as proposed by the International Civil Aviation Organization (ICAO) through an integration of the global navigation infrastructures to improve the utilization efficiency of airspace and reduce flight delays and the dependence on ground navigation facilities. The approach stage is one of the most important and difficult stages in the whole flying. In this study, we proposed deep reinforcement learning (DRL)-based RNP procedure execution, DRL-RNP. By conducting an RNP approach procedure, the DRL algorithm was implemented, using a fixed-wing aircraft to explore a path of minimum fuel consumption with reward under windy conditions in compliance with the RNP safety specifications. The experimental results have demonstrated that the six degrees of freedom aircraft controlled by the DRL algorithm can successfully complete the RNP procedure whilst meeting the safety specifications for protection areas and obstruction clearance altitude in the whole procedure. In addition, the potential path with minimum fuel consumption can be explored effectively. Hence, the DRL method can be used not only to implement the RNP procedure with a simulated aircraft but also to help the verification and evaluation of the RNP procedure.",
"conclusion": "5. Conclusions In this study, the multi-task extended HRL (MHRL) algorithm was proposed for implementing the RNP procedure and exploring a path with minimum fuel consumption (i.e., for solving the problems of aircraft control and path planning). To solve the multi-channel coupling control problem in aircraft control, a flight controller based on multi-task DRL was raised to learn the generalities of the three dimensions of control (heading, altitude, and velocity). As for the path planning for minimizing the fuel consumption, the MHRL algorithm was suggested. In this algorithm, the bottom layer is for the aircraft control, while the top agent is for exploring a path with minimum fuel consumption. The simulation experimental results revealed that the multi-task DRL controller can realize end-to-end control of an aircraft and can be used to solve the coupling control problem in similar scenes. The MHRL algorithm can control aircraft to explore a path with the minimized fuel consumption while complying with the safety specifications, offering an idea for RNP procedure design and verification. As also shown, the MHRL algorithm can solve a complex problem effectively by having the problem divided into several sub-problems and conquering them separately. It is hopeful to apply this method to solve complex problems in the aviation field. Nevertheless, the proposed algorithms and experiments also have some limitations. In addition to the altitude and velocity, more restrictions (e.g., the descending rate and the turning rate) should be considered in implementing the RNP procedure. Moreover, the exploration of a path with minimum fuel consumption was conducted on the given RNP procedure. In the practical design of the RNP procedure, the departure and destination points are the considerations for judging the minimum fuel consumption. It is expected to solve these problems by setting a more reasonable reward function in combination with traditional path planning algorithms, and comparing the method proposed in this paper with the lowest fuel consumption path obtained from historical flight data is the future work of this paper, the acquisition of more real and reliable historical flight data is a difficult problem, using data from a simulated air traffic control system maybe is a solution.",
"introduction": "1. Introduction As proposed by the International Civil Aviation Organization (ICAO), the performance-based navigation (PBN) procedure is a new technology for the next generation of air traffic [ 1 ]. To be specific, it refers to the aircraft’s requirements for the precision, integrity, usability, continuity, function, and other performances of the system when flying along the designated path, as per the designed instrument flying procedure or within the specified route or air space under the corresponding condition of navigation infrastructures. The PBN procedure contains two basic navigation specifications: the rules for implementation of area navigation (RNAV) and the required navigation performance (RNP) [ 1 ]. Among them, the RNAV specifies that the aircraft should be able to fly along any expected path within the coverage of the navigation signal or within the scope of data calculated by the aircraft’s avionic device or within both of the two scopes, while the RNP is the RNAV added with onboard performance monitoring and alerting (OPMA) ability. Compared with the RNAV, the RNP can realize a higher navigation precision and only depends on the global radio navigation satellite system (RNSS). Currently, one of the key technologies for flying with the RNP procedure is to have an aircraft fly along the expected path in conformance with the RNP safety specification. The designed RNP procedure should be safe, economic, and convenient while considering the benefits to the air traffic control officers, the approach tower, the aircraft crews, the airport ridership, and the environment. A flight procedure should be put forward by the demander firstly, then undergo the coordination between the procedure designer and the procedure’s stakeholders, and iterated many times before being applied [ 2 ]. In general, a practical flight procedure is a flight path determined as per the safety specifications of the flight procedure and the constraints of the stakeholders. However, the economic optimization of this path is still worthy of study. Deep reinforcement learning (DRL) is a technology that combines the deep learning (DL) and the reinforcement learning (RL), where the former is to perceive the environment and the latter is to solve a problem with a decision [ 3 ]. Compared with DL, the DRL mainly obtains data by interaction with the environment. This technology has been successfully applied in many aspects such as AlphaGo [ 4 ], investment [ 5 ], UAV path planning [ 6 ], and robot control [ 7 ], which is considered as one of the closest to the artificial general intelligence (AGI) approach [ 8 , 9 ]. With excellent performance in continuous control, the DRL method has been widely used in path planning for aircrafts or robots. Compared with the A* algorithm, particle swarm, and other traditional path planning algorithms, the DRL-based path planning algorithm has two advantages: first, it can implement the path planning without the complete information about the environment; second, it considers the dynamic or kinematic performance of the controlled object. In this study, the DRL algorithm was proposed the first time for optimizing the flight path in the RNP procedure. To be specific, the DRL algorithm was used to have an aircraft implement the RNP procedure as per the safety specifications, based on which an optimal path with minimum fuel consumption was sought further. The main contributions of this paper can be highlighted as follows: (a) A flight controller based on multi-task deep reinforcement learning is proposed, which can solve the multi-channel coupling control problem existing in aircraft control and provide an effective solution for similar coupling control problems. (b) A DRL method named MHRL is proposed, which combines multi-task RL with hierarchical reinforcement learning (HRL) for integrated control and decision-making. MHRL can solve the problem of flight control in accordance with safety regulations and path planning considering fuel consumption. (c) The proposed work provides possible application prospects for the research that needs to consider both aircraft control and decision-making. The remaining of this paper is organized as follows. Section 2 briefly discusses the related work, the RNP procedure, the DRL algorithm, and the Dryden wind turbulence model are introduced. In Section 3 , the proposed DRL-RNP algorithm for aircraft control and path planning are presented. Section 4 describes the environment settings and structure of the model, and the simulation results and discussion are given. Finally, Section 5 provides some concluding remarks and future directions.",
"discussion": "4. Simulation, Experiments, and Discussion This study mainly involved two experiments. The first was conducted on the proposed multi-task RL-based flight controller. The results have validated that the multi-task RL-based flight controller had a better control effect than the flight controller trained by single-task RL and could improve the multi-dimensional flight control effect. The second mainly revealed that the proposed MHRL method could have an aircraft to implement the RNP procedure and explore a path with minimum fuel consumption in compliance with the safety specifications of the procedure. The simulation and experiments involved in this study were carried out on a PC device with the i7-8700K CPU, 16GB memory, and NVIDIA GeForce GTX 1070Ti. 4.1. Aircraft Controller Based on Multi-Task DRL 4.1.1. Environment Settings In this study, the aircraft was controlled in three dimensions: i.e., heading, altitude, and velocity. For a single-task RL controller, a new control task was set at a time interval of 150 s, which came from the default reinforcement learning control tasks of JSBSim. For heading control, a new task was designed to change ±10° from the current heading; for altitude control, a new task was designed to change ±30 ft from the current altitude; for velocity control, a new task was designed to change ±20 ft/s from the current velocity. Concerning the multi-task RL-based controller, the control task randomly selected from the above three single-tasks at a time interval of 150s. The above four controllers (three single-task controller and one multi-task controller) are trained on the A320 aircraft. The control surface and throttle ( δ e , δ a , δ r , δ T ) are the action spaces. The observations of single-task controller were the difference between the currently altitude, heading, velocity and the expected altitude, heading, velocity of the aircraft ( d e l t a a l t i t u d e , d e l t a h e a d i n g , d e l t a v e l o c i t y ), as well as the attitude (Euler angle), altitude, and velocity. The observations of the multi-task controller add i d t a s k based on the observations of single-task controller. Regarding the setting of the reward function, it was expected to grant a large reward when the aircraft’s heading, altitude, and velocity reached the expected ranges. Moreover, a special reward was set for roll angle, with an expectation that the aircraft would not have an overlarge maneuver. The settings of the reward function are expressed as below: (23) r heading = e − ∣ Δ heading scale heading r altitude = e − | Δ a l t t i t u d e | scale altitude r velocity = e − | Δ velocity | scale velocity r roll = e − roll scale roll r = r heading * r altitude * r velocity * r roll 1 4 \nwhere, r h e a d i n g , r a l t i t u d e , r v e l o c i t y , and r r o l l are the rewards corresponding to the heading, altitude, velocity, and roll angle, respectively; Δ h e a d i n g , Δ a l t t i t u d e , and Δ v e l o c i t y are the differences between the current values and expected values of the specific variables; r o l l is the roll angle; s c a l e h e a d i n g , s c a l e a l t i t u d e , s c a l e v e l o c i t y , and s c a l e r o l l are scaling indices of the specific variables to ensure that the heading, altitude, velocity, and roll angle within the allowed error range can obtain a large reward. In this study, s c a l e h e a d i n g was set at 3°, s c a l e a l t i t u d e at 10 ft, s c a l e v e l o c i t y at 5 ft/s, and s c a l e r o l l at 30°. Moreover, Equation (23) is the step reward, the episode reward is calculated by Equation (24), r is step reward, k is the step number in an episode, and the initial environment settings are shown in Table 3 .\n (24) e p i s o d e r e w a r d = ∑ t = 1 k r t 4.1.2. Model Settings The hyper-parameters in the model were set as follows: the learning rate was 0.0001 and the discount factor was 0.98; the soft update coefficient was 0.05 and replay buffer size was 20,000; the mini-batch size was 1024. This model contains nine layers of neural networks. The first two layers have 512 neurons, and the remained layers have 256 neurons. The hyper-parameters and network structures in the model are listed in Table 4 . 4.1.3. Simulation Results and Analysis Figure 8 presents the reward curves of three single-task (heading, velocity, and altitude) and the multi-task over 2.0 × 10 6 steps of simulation with the same reward function, observations, and model. The horizontal axis of Figure 8 is the step number and the vertical axis is the episode reward. Compared with using episode number as horizontal axis, using total steps as horizontal axis can better reflect the step number in an episode. As shown in Figure 8 , the single-task RL controller was not available for effective control of the aircraft’s velocity or heading separately, but the altitude. The reward for multi-task RL controller was higher than single-task RL controller. Moreover, within the same step numbers, the points on the multi-task reward curve are less than the alt-task reward curve, which indicates that the step numbers in a multi-task episode is greater than that in an alt-task episode. Both episode reward and step numbers in an episode indicate that the multi-task RL algorithm is more effective than the single-task RL algorithm for the aircraft’s heading, velocity, and altitude controls. Figure 8 is reckoned that the single-task DRL controller is not effective for the aircraft’s velocity and heading controls but the altitude control, which is consistent with the characteristics of the aircraft. For the fixed-wing aircraft, all the controls should be conducted when the aircraft is maintained at a certain flight altitude; otherwise, it is easy to cause a stall and further lose control of the aircraft. Although the reward functions of heading and velocity controls considered the altitude control (Equation (23)) (maintaining at the initial altitude), the heading or velocity change disturbed the single-task DRL controller’s perception of the altitude control so that the single-task DRL controller lost the control of the aircraft’s heading or altitude. In the first 5 × 10 5 steps of training, the reward curve of the multi-task DRL controller was almost the same as that of the single-task (altitude control) DRL controller. Hence, it is deemed that the multi-task DRL controller learned the aircraft’s altitude control first and then its heading and velocity controls, where the later learning made the controller more robust. Consequently, the multi-task RL controller’s reward curve was higher than the single-task (altitude control) DRL controller. To further verify that the multi-task DRL controller is effective for the single-task DRL controller, and has high robustness, we counted the success rates of having the aircraft reach the expected heading, altitude, and velocity at different task switching frequencies with or without the multi-task DRL algorithm, as shown in Figure 9 a,c,e. In the three figures, the horizontal axis represents the time interval of the heading, altitude, or velocity change; the vertical axis represents the success rate in 100 tests at this time interval. From Figure 9 a,c,e we can find that with the increase in time interval of heading, altitude, and velocity change, the success rate of multi-task DRL controller can arrive about 100%, 80%, and 60%, which are better than the success rate without the multi-task DRL method controller. Although the heading and altitude control arrive a high success rate, the velocity control is not good as expected, the reason we think is that the designed velocity control task is a scene with low initial speed, in which the velocity control is more difficult. The three experiments adopted the same initial conditions as the environmental settings described in Section 4.1.1 . Further, we also observed the stability of the multi-task DRL controller in simultaneous control of the heading, velocity, and altitude. With the expectation that the controller could maintain the controls in the other two channels stable when that in one channel changed, three experiments were conducted additionally to see the changes in the other two channels when one channel changed, for example, when heading change 10°, we observed changes in altitude and velocity. In the three tests, the environment settings were the same as that described in Section 4.1.1 . The simulation times were identical, 320 s; at the seconds of 80 s, 160 s, and 240 s, the controller changed the heading, altitude, or velocity +/−10°, +/−30 ft, or +/−20 ft/s from the current heading, altitude, or velocity. The results are exhibited in Figure 9 b,d,f. As demonstrated, when the heading changed, the velocity fluctuated by 10 ft/s at most and the altitude by 10 ft at most; when the altitude changed, the velocity fluctuated by 10 ft/s maximally and the heading by 5° maximally; when the velocity changed, the heading fluctuated by no more than 5° and the altitude by no more than 25 ft. These results showed that the multi-task DRL controller has perfect stability. In addition, the stability of the controller also has a certain scope of application. For example, when the aircraft altitude changes a lot, the aircraft velocity cannot always remain unchanged. Moreover, we designed five traditional controllers (PID) to control the heading, altitude, and velocity of the aircraft, and compared them with the multi-task DRL controller. The five PID controllers are pitch, roll, heading, altitude, and velocity controller. The structures of pitch, roll, and velocity controller are shown in Figure 10 , the structures of heading and altitude controller are shown in Figure 11 . From the Figure 10 , we can observe that the input value of the PID controller is the target pitch, roll or velocity, and the output is the control surface command ( δ e , δ a , δ T ). From the Figure 11 , we can observe that the input of the heading or altitude PID controller is the target heading or altitude, and the output is the roll angle or pitch angle, and then the roll or pitch PID controller outputs the control surface command ( δ e , δ a ) to realize the heading and altitude control. The results of the pitch and roll PID controller are shown in Figure 12 . Figure 12 a is pitch control and Figure 12 b is roll control. Figure 13 compared the PID controller and multi-task DRL controller. From Figure 13 , we can find that for heading control, the rise time, adjustment time, and overshoot of the multitask DRL controller are better than those of the PID controller, but the fluctuation is large after stabilization. For altitude and velocity control, the overshoot and stability of multi-task DRL controller are obviously better than PID controller. It is worth noting that coupling often occurs in the parameter adjustment of PID controllers, that is, the parameter adjustment of one PID controller affects the other PID controller. The results of the PID controller are shown in Figure 12 and Figure 13 have undergone about 80 parameter adjustments, and the final parameters are shown in Table 5 . In general, we think that the multi-task DRL controller is superior to the PID controller in terms of stability and has more advantages in dealing with control coupling problems. 4.2. Path Planning Based on MHRL In this part, two experiments were conducted: the first was mainly to test whether the MHRL method could meet the safety specifications or not when implementing the RNP approach procedure; the second was mainly to test the MHRL method’s exploration of the path with minimum fuel consumption in compliance with the safety specifications of the RNP approach procedure. 4.2.1. Environmental Settings In these two experiments, it was expected that the MHRL method could complete the initial and intermediate approach stages of the RNP procedure. According to the literature [ 27 ], two routes were designed, similar to a half of a “Y” shaped approach procedure ( Figure 14 ). The longitude, latitude, half width, altitude, and velocity limits of the waypoints were listed in Table 6 , where 5000 ¯ ¯ means that the altitude of the aircraft at this waypoint is about 5000 ft, 2300 ¯ means that the altitude of the aircraft at this waypoint is not less than 2300 ft. The initial settings of the aircraft in Table 7 , where the initial position of the aircraft was at the first waypoint of the approach routes, and the heading was the direction of the initial approach leg. The observations of top agent can be divided into four parts: RNP flight procedure safety specification, wind turbulence, fuel consumption, and airplane state. RNP flight procedure safety specification contains three-part limits: altitude limit, velocity limit, and distance to expected path limit. Moreover, the leg type and distance to next waypoint are also be observed. Wind turbulence contains along wind speed, cross wind speed and vertical wind speed. Fuel consumption is the remaining fuel at every simulation step. Airplane state contains altitude and roll angle, which is less than the airplane state variable observations in flight controller agent. The observations are as follows: o b s = Δ a l t m a x , Δ a l t m i n , Δ v m a x , Δ v m i n , Δ d i s p a t h , Δ d i s n e x t w p t , v a l o n g − w i n d , v c r o s s − w i n d , v v e r t i c a l − w i n d , f u e l , a l t , r o l l \nwhere Δ a l t max m i n is the difference between airplane altitude and the leg max (min) altitude limit, Δ v max m i n is the difference between airspeed and the leg max (min) velocity limit, Δ d i s path is the distance to expected path, Δ d i s nextwpt is the distance to next waypoint, v along corss , vertical − wind is the along (cross, vertical) wind speed, f u e l is the remaining fuel, a l t is airplane altitude, r o l l is roll angle. In addition, the observations are normalized in the implementation. The action space of top agent contains seven i d t a s k , which mean seven basic maneuvers: left turning, right turning, ascending, descending, acceleration, deceleration, and hold on. These seven basic maneuvers are combined by heading, altitude, and velocity control. We set the value of left (right) turning is 10°, the value of ascending (descending) is 30 ft, the value of acceleration (deceleration) is 20 ft/s, which are the same as the control task of flight controller in Section 4.1.1 . The action space as follows: a c t i o n s p a c e = i d l e f t t u r n i n g , i d r i g h t t u r n i n g , i d a s c e n d i n g , i d d e s c e n d i n g , i d a c c e l e r a t i o n , i d d e c e l e r a t i o n , i d h o l d o n The reward function of top agent is divided into two parts: the first part is the reward for complying the flight procedure under safety specifications, which is the main part, the second part is the reward for the remained fuel mass. The settings of the reward function are expressed as below: (25) r A W / 2 = 1 , i f Δ d i s path < 1 2 A W , e l s e 1 × 10 − 4 r a l t i t u d e = 1 , i f alt min < alt airplane < alt max , e l s e 1 × 10 − 4 r v e l o c i t y = 1 , i f v min < v a i r p l a n e < v max , e l s e 1 × 10 − 4 r d i s 2 n e x t w p t = e − Δ d i s nextwpt scale dis r f u e l = e − Δ fuel scale fuel r 1 = r A W 2 * r a l t i t u d e * r v e l o c i t y * r d i s 2 n e x t w p t 1 4 r 2 = r 1 + r f u e l \nwhere r A W 2 , r a l t i t u d e , r v e l o c i t y , and r d i s 2 n e x t w p t are the first part reward, 1 2 A W is the protection zero width, which is shown in Figure 2 . and calculated by Equation (8), a l t min m a x is the leg min (max) altitude limit, v min m a x is the leg min (max) velocity limit, the s c a l e d i s we set 5000 m, which is about half distance of two waypoints. r f u e l is the second part reward, Δ f u e l is the current fuel consumption, s c a l e f u e l is the scale index, for route 1, we set it 252 lbs, for route 2, we set it 298 lbs, which are the fuel consumption without consideration of fuel consumption reward. r 1 is the step reward without consideration of fuel consumption, r 2 is the step reward with consideration of fuel consumption. 4.2.2. Model Settings The RL algorithm of the top agent also adopted the SAC algorithm. Compared with the SAC algorithm model in the multi-task RL controller, this model had fewer layers of networks. Its hyper parameters were set as follows: the learning rate was 0.0001 and the discount factor was 0.98; the soft update coefficient was 0.05 and the replay buffer size was 20,000; the mini-batch size was 1024. This model contained seven layers of neural networks. The first two layers had 512 neurons and the remained layers had 256 neurons. The hyper-parameters and network structures of the model are provided in Table 8 . 4.2.3. Simulation Results and Analysis The reward curves for implementing the RNP procedures of route 1 and route 2 with and without MHRL method are illustrated in Figure 15 . The implementation of without MHRL is to add the rewards of safety specification and fuel consumption to the reward function in the multi-task DRL controller. As shown in Figure 15 , the rewards for implementing route 1 and route 2 with MHRL method were both significantly higher than those without MHRL. At the initial time of the training, the rewards for implementing the procedure with MHRL method were higher than that without MHRL method, the reason of this is that we think the top policy in MHRL method could output action to keep the aircraft stable and obtain a reward even though the top agent had not learned the effective policy. Figure 16 shows the details of implementing the initial and intermediate approach procedures of route 1 and route 2 with MHRL method. In the following experiment, the Dryden wind turbulence model was used to simulate the impact of wind on the aircraft. The wind turbulence had three levels: light, moderate, and severe. The wind turbulence velocity at each level is calculated according to the altitude and airspeed of the aircraft, Equations (1)–(7) and Table 1 . Refs. [ 25 , 26 ] describe in detail the equation derivation and code implementation of Dryden wind turbulence model. For route 1, the MHRL method could have the aircraft fly along the expected path with the three degrees of wind turbulence. In the whole process, the aircraft’s altitude, velocity, and distance to the expected path complied with the safety specifications of the RNP procedure. For route 2, the aircraft was also maintained flying along the expected path with MHRL method at the three degrees of wind turbulence. In the whole process, the aircraft’s altitude and the distance to the expected path satisfied the safety specification of the RNP procedure. Although an overspeed event occurred due to the rapid decline of aircraft altitude in the first 50 s, the top agent can reduce the speed by climbing altitude to meet the safety specifications. In addition, in the second half of route 2, the top agent seems to learn to slowly reduce the height and speed at the same time. It also can be seen from the Figure 16 that the influence of wind turbulence on the distance to expected path is more obvious than that of wind turbulence on velocity and altitude. The reason is that we think it is related to the optional actions of top agent. Turning left and right cannot well offset the impact of crosswind on aircraft flight. Flying in a sideslip attitude may achieve better results, which is a direction that this research can improve in the future. The above results are based on without consideration of fuel consumption ( r 1 in Equation (25)), then the MHRL method was taken to explore a path with minimum fuel consumption ( r 2 in Equation (25)) in implementing the flight procedure. The experimental results of route 1 and route 2 are provided in Figure 17 and Table 9 , Figure 18 and Table 10 . Figure 17 exhibits the flight paths of route 1 obtained by MHRL method with and without fuel consumption reward under windless conditions. The color bar shown on the right side represents the flight altitude. As observed, the two paths were less different in the initial approach stage of route 1 and largely different in the intermediate approach stage; moreover, the top agent in the MHRL method attempted to reduce fuel consumption by shorter flight path. Table 9 illustrates the fuel consumptions of MHRL method with and without fuel consumption reward at four wind levels: the first row presents the fuel consumption without fuel consumption reward, and the second row presents the fuel consumption with fuel consumption reward; the fuel consumptions shown in the table are the averages of 50 tests obtained at four wind levels; the last row presents the percentage of reduction in the fuel consumption (the fuel consumption difference between the two cases divided by the fuel consumption without fuel consumption reward). It can be seen that the MHRL method with fuel consumption reward could help reduce about 5% fuel consumption for route 1. Figure 18 displays the flight paths of route 2 obtained by MHRL method with and without fuel consumption reward under windless conditions, where the right color bar represents the flight altitude. As observed, the two paths were largely different in the initial approach stage of route 2 and less different in the intermediate approach stage; compared with the path without fuel consumption reward, the path with fuel consumption reward can reduce fuel consumption by selecting appropriate turning time and further reducing the flight path. Table 10 lists the fuel consumptions of MHRL with and without fuel consumption reward at four wind levels: the first row presents the fuel consumption without fuel consumption reward, and the second row presents the fuel consumption with fuel consumption reward; the fuel consumptions shown in the table are the averages of 50 tests obtained at four wind levels; the last row presents the percentage of reduction in the fuel consumption (the fuel-consumption difference between the two cases divided by the fuel consumption without fuel consumption reward). It can be observed that the MHRL method with fuel consumption reward contributed to reducing about 13% fuel consumption for route 2. It can be seen from Table 9 and Table 10 that whether for route 1 or route 2, the higher the wind turbulence level, the more fuel consumption. The reason for this is, we believe, related to the frequent change of maneuvers and the increase in flight path under wind turbulence. In addition, another topic worth discussing is the design of reward function considering fuel consumption ( r 2 in Equation (25)), which is also an important and difficult part in DRL. In fact, we tried several different reward functions to reduce fuel consumption, for example, taking r f u e l (in Equation (25)) as an item in r 1 (in Equation (25)), only taking r f u e l as r 2 (in Equation (25)), etc.; however, none of them obtain better results than r 2 , and some even lead to worse results than ignoring fuel consumption. This makes us have to consider that this only is the lowest fuel consumption path under r 2 , not for route 1 or route 2. In summary, the proposed MHRL method can control aircraft execute the RNP approach procedure while satisfying the safety specification. Meanwhile, the MHRL method with fuel consumption reward is effective for exploring a path with minimum fuel consumption, and the basic maneuver of top agent in MHRL method and the reward function with fuel consumption are the improvement directions of this study."
} | 8,044 |
28903461 | PMC5585692 | pmc | 8,895 | {
"abstract": "Abstract Gene retroposition is an important mechanism of genome evolution but the role it plays in dinoflagellates, a critical player in marine ecosystems, is not known. Until recently, when the genomes of two coral-symbiotic dinoflagellate genomes, Symbiodinium kawagutii and S. minutum , were released, it has not been possible to systematically study these retrogenes. Here we examine the abundant retrogenes (∼23% of the total genes) in these species. The hallmark of retrogenes in the genome is the presence of DCCGTAGCCATTTTGGCTCAAG, a spliced leader (DinoSL) constitutively trans -spliced to the 5′-end of all nucleus-encoded mRNAs. Although the retrogenes have often lost part of the 22-nt DinoSL, the putative promoter motif from the DinoSL, TTT(T/G), is consistently retained in the upstream region of these genes, providing an explanation for the high survival rate of retrogenes in dinoflagellates. Our analysis of DinoSL sequence divergence revealed two major bursts of retroposition in the evolutionary history of Symbiodinium , occurring at ∼60 and ∼6 Ma. Reconstruction of the evolutionary trajectory of the Symbiodinium genomes mapped these 2 times to the origin and rapid radiation of this dinoflagellate lineage, respectively. GO analysis revealed differential functional enrichment of the retrogenes between the two episodes, with a broad impact on transport in the first bout and more localized influence on symbiosis-related processes such as cell adhesion in the second bout. This study provides the first evidence of large-scale retroposition as a major mechanism of genome evolution for any organism and sheds light on evolution of coral symbiosis.",
"introduction": "Introduction Gene birth is a major driver of genome evolution and shapes the configuration and function of the genome. A major mechanism of gene birth is gene duplication, and can occur by whole genome duplication (WGD) or segmental duplication (SD) events ( Wolfe and Shields 1997 ; Crow etal. 2006 ). Retroposition, the integration of an RNA back into the genome (also called mRNA recycling), is also frequent and has been well documented in mammals and fruit flies ( Kaessmann etal. 2009 ). However, because the regulatory elements (i.e., promoters) that drive the expression of the parent genes are not included in the transcripts, most of the retrogenes become “dead on arrival” pseudogenes ( Kaessmann etal. 2009 ). Only those few retrogenes (90 ∼ 100 in mammals or flies) that successfully obtain new promoters can survive ( Marques etal. 2005 ; Bai etal. 2007 ). Currently known mechanisms of retrogene survival include 1) using the promoter of other genes when the retrocopy is inserted into an intron or an active region of the chromosome; 2) exploiting a distant promoter by transcribing a long 5′-UTR; 3) inheriting a promoter from an alternatively transcribed parental copy; 4) using a “proto-promoter” with promoter potential in CG islands; and 5) accumulating mutations to form a promoter de novo (reviewed in Kaessmann etal. 2009 ). Despite their infrequent survival, retrogenes play a critical role in shaping genome architecture as they are prone to sub or neofunctionalization and become a valuable source of new genes. Dinoflagellates, a group of unicellular eukaryotic protists closely related to the Apicomplexa, are important marine primary producers, major contributors to harmful algal blooms (commonly known as “red tides”), and the essential symbionts of reef corals. They display many unusual cytological characteristics ( Lin 2011 ) including generally undetectable levels of histones, permanently condensed chromosomes and widely varied and generally large (∼1–250 Gbp) genomes. Furthermore, the classical TATA box may have been replaced by TTT(T/G) as the core promoter motif. This sequence is one of the few highly conserved motifs in the region upstream of S. kawagutii coding sequences ( Shoguchi etal. 2013 ; Lin etal. 2015 ). Furthermore, a TBP-like factor with higher affinity to this motif has been found in dinoflagellates ( Guillebault 2002 ) and is the only type of a TATA-box binding protein present. Interestingly, mRNA maturation requires trans -splicing of a 22-nt dinoflagellate spliced leader [DinoSL, DCCGUAGCCAUUUUGGCUCAAG (D = U, A or G)] to their 5′ termini ( Zhang etal. 2007 ). One of the best studied dinoflagellate lineages is Symbiodinium , which harbors species living endosymbiotically in a wide range of cnidarian hosts including corals and anemone. Due to its importance for coral reef growth and its small genome size (for dinoflagellates, still 50-fold that of Plamodium falciparum ), Symbiodinium has become a valuable model for dinoflagellate genome research. No evidence for WGD or SD has been documented in Symbiodinium genomes ( Shoguchi etal. 2013 ; Lin etal. 2015 ), and thus the major forces driving the rapid expansion of the immense dinoflagellate genomes are still poorly understood. Retrogenes in dinoflagellates, discovered by the presence of relict DinoSL sequences ( Slamovits and Keeling 2008 ), have great potential to contribute to genome evolution as they provide a large source of new genes. However, the genome-wide scale, tempo, and evolutionary as well as ecological implications have not been systematically investigated due to the lack of dinoflagellate genome sequences. Furthermore, as none of the known mechanisms of promoter acquisition can account for a large-scale retention of retrogenes in dinoflagellate genomes, the mechanism facilitating survival of dinoflagellate retrogenes is also an enigma. The recent availability of two dinoflagellate genome assemblies ( Symbiodinium kawagutii and S. minutum ) ( Shoguchi etal. 2013 ; Lin etal. 2015 ) has afforded an opportunity to address these questions. In this study, we used data from the two genomes and identify retrogenes globally by searching for relict DinoSL sequences in the upstream 5′ region of gene coding sequences. We found that >20% of the genes in both genomes have relict DinoSL sequence. We suggest that the unique trans -splicing process which adds TTT(T/G), the potential equivalent of the TATA box in dinoflagellates, may act to introduce a basal promoter into the retrogenes thus facilitating their survival. Extensive analyses of sequence characteristics and functional distribution of the retrogenes reveal that these retrogenes have emerged in two major bouts, which might have promoted the emergence and radiation of Symbiodinium species.",
"discussion": "Discussion The birth of genes is important in shaping the genome and driving the evolution of species. In most of the known eukaryotic genomes, WGD and SD efficiently generate new gene copies. As chromosome fragments are duplicated, the genes on it are also doubled. The regulatory elements are always duplicated together with the coding regions, and as a result, a new copy initially has an expression pattern similar to the parental copy. The gene copy can gradually change its function or become silenced by pseudogenization or posttranscriptional regulation, such as by small RNAs. In contrast, generation of new gene copies by retroposition is a very different process. Retroposition is an mRNA-mediated gene birth process similar to retrotransposition but differing in that the gene that is being copied does not encode the endonuclease and reverse transcriptase activities required for retrotransposition. The endonuclease and reverse transcriptase activities required for retroposition are presumably provided by LINEs ( Pavlicek etal. 2007 ). In most organisms, retrogenes are very often “dead on arrival” due to the lack of regulatory elements (i.e., promoter), and are termed pseudogenes. Only the few retrogenes that integrate close enough to an existing promoter to be activated by it will survive. The retrogenes that are expressed typically play critical roles in shaping the genome and phenotypes because their expression pattern will be different from the parental gene copies, leading to dramatic changes in phenotypes. Retroposition Is Widespread in Dinoflagellate Genomes Dinoflagellate nucleus-encoded mRNAs are capped by a 22-nt leader sequence before maturation ( Zhang etal. 2007 ). This character allows easy identification of retrogenes in dinoflagellate genomes by searching for the DinoSL or its relicts in the upstream region of genes. The search results show the retrogenes account for >20% of the total genes in both the Symbiodinium genomes, with some of them (2 ∼ 3%) having been recycled multiple times ( fig. 1 A ). Besides Symbiodinium , retrogenes have also been found in ESTs from many other dinoflagellates including Alexandrium , Oxyrrhis , Heterocapsa , and Prorocentrum ( Slamovits and Keeling 2008 ; Jaeckisch etal. 2011 ; Lee etal. 2014 ). These results suggest large scale retroposition is widespread in dinoflagellates. These larger dinoflagellate genomes may have experienced more extensive retroposition than Symbiodinium because ∼12% of their cDNAs, ∼4–6 times greater than what is found in Symbiodinium genomes, have more than three DinoSLs ( Slamovits and Keeling 2008 ) which is indicative of multiple recycling events ( table 1 ). The large number of retrogenes in both the Symbiodinium genomes allows their character and evolution to be analyzed in more detail. Surprisingly, this analysis indicates that retroposition occurred in two major bouts during evolution, whose timing corresponds first to a time before the emergence of Symbiodinium and second to the radiation of Symbiodinium species. It is interesting that the times of retroposition estimated from the two different Symbiodinium genomes are virtually identical ( fig. 2 ). The first bout of retroposition occurred ∼60 Ma. This time is very close to the time (∼55–75 Ma, the Cretaceous–Paleogene boundary) when WGD events frequently occurred in plant genomes ( Van de Peer etal. 2017 ). The Cretaceous–Paleogene boundary is marked by a number of catastrophic events, which had led to major climate changes ( Petersen etal. 2016 ). A mass extinction of species (60–70% of all plant and animal) including nonavian dinosaur occurred at this time. These environmental changes whose effects are so noticeable in the terrestrial organisms may thus be mirrored in the burst of retroposition in marine dinoflagellates. The second retroposition occurred at ∼6 Ma, after the split of S. kawagutii and S. minutum , suggesting that the large-scale retroposition process is not species-specific but rather may also have been due to another major episode of environmental changes. If true, it suggests that retroposition may also have occurred at these times in other dinoflagellate genomes. As the interval between these two bouts of retroposition is rather small, and retroposition seems to be easily triggered by environmental factors, we speculate that there might be more than two bouts of retroposition during dinoflagellate evolution. Although this is certainly possible it seems likely they will be undetectable using our current data due to several limitations, including that 1) the SL is rather short, hence very old retrogenes would mutate beyond recognition; 2) the loss of parental copies would result in fewer or false pairs for calculation of Ks; and 3) the saturation of synonymous mutations may limit detection of old bouts. Despite these caveats, the frequency of retroposition identified in Symbiodinium genomes clearly indicates that large-scale retroposition events have occurred several times during dinoflagellate evolution. The Survival of Dinoflagellate Retrogenes Is Facilitated by DinoSL As mentioned earlier, most retrogenes in other genomes have become pseudogenes and their sequences have degenerated. We thus asked what mechanisms might have facilitated survival of dinoflagellate retrogenes. We noticed that DinoSL, which is constitutively trans -spliced to the 5′ end of mRNA, has a TTT(T/G) motif, proposed to be the dinoflagellate equivalent of a TATA box. This leader sequence is retroposed together with mRNA and might be able to promote the transcription of retrogene. If this hypothesis is true, we predicted that 1) the DinoSLs located close to the promoter region (probably within the first 100-bp upstream from the start codon) would be more conserved than the others regardless of their ages; 2) the TTT(T/G) motif in particular would be more conserved; 3) there would be a limited number of multiply recycled retrogenes. In a multi-recycled retrogene, it seems reasonable to assume that the DinoSL closest to the coding region represents an early retroposition event while the one further away represents a more recent retroposition event—upstream DinoSL sequences would thus be younger than the downstream ones. We find DinoSLs located at −50 to −100 bp are the most highly conserved ( fig. 1 C ), as predicted, and note that this pattern of sequence conservation is thus not related to their ages. Furthermore, we also find an increased conservation of TTT(T/G) motifs. If the TTT(G) in the recycled DinoSL served as a promoter, the primary transcripts from the retrogenes would lack a DinoSL, and secondary recycled retrogenes would thus also contain only one DinoSL upstream of its coding sequence. Only when a retrogene was transcribed from a site upstream of the recycled DinoSL would a secondary retrogene with multiple DinoSLs be observed. This might be expected to be a relatively infrequent event, and indeed, the number of retrogenes with multiple DinoSLs is very small in both Symbiodinium genomes ( table 1 ). Our result regarding the conservation of DinoSL motifs appears different from what has been described in earlier reports conducted on ESTs ( Slamovits and Keeling 2008 ; Jaeckisch etal. 2011 ), in which the authors found that the degree of conservation increases as a function of the distance between the DinoSL and the start codon. This difference is presumably attributable to the different data sets used in the two types of studies. In EST studies, the analyses were performed on 5′-UTRs, which are necessarily downstream of the promoter, and any DinoSLs upstream of the transcriptional start site are simply not present. Genomic sequence, on the other hand, allows these upstream DinoSLs to be observed. Since the sequence actually used as a promoter is likely to be more conserved independent of where it is located relative to the start codon, we suggest that the poorly conserved DinoSL sequences further upstream in the genome are not used as a promoter element and are thus less conserved ( supplementary fig. S1 , Supplementary Material online). Of the two factors affecting the degree of DinoSL sequence conservation, one the age of the sequence and the other the selective pressure exerted by its ability to function as a promoter, only the age of the sequence can be observed in the EST data set. We found the highest degree of conservation in DinoSLs located between −50 and −100 ( fig. 1 D ), a location where dinoflagellate promoters are likely to be found based on the usual length of 5′-UTR ( Zhang etal. 2007 ; Kim etal. 2011 ). Our analysis of dinoflagellate retrogenes supports the initially described model of dinoflagellate retroposition ( Slamovits and Keeling 2008 ), a model which differs from that in humans and other model organisms. In dinoflagellates, genes are transcribed from active regions and the transcripts are rapidly capped with DinoSL. Although most of these DinoSL-capped transcripts subsequently have the introns removed and are exported to the cytoplasm for translation, some appear to have been rapidly reverse transcribed and integrated into the genome. The TTT(T/G) motif in the DinoSL cap may then serve as a basal promoter to facilitate the survival of retrogenes in the genome ( fig. 4 A ).\n Fig. 4. —The retroposition process in dinoflagellate. (A) Gene retroposition process inferred from the characteristics of retrogenes. A gene transcript (step 1) is trans -spliced through which the DinoSL sequence is added to the 5′ terminus of a nascent pre-mRNA (step 2), followed by cis -splicing (step 3) before they were exported to cytoplasm for translation (step 4); Reverse transcription and genome integration (step 5) can occur at either of these steps inside nucleus (step 1, 2, 3 and 4); in dinoflagellates, the DinoSL sequence harboring core promoter motif “TTTT” and “TTTG” potentially serves as a promoter enabling the survival of retrogenes; therefore, most retrocopies stemmed from nascent transcripts (step 1) would be dead upon their birth due to the lack of promoter; the whole procedure is limited in the nucleus. (B) A scheme showing the self-enforcing model enabling the fixation of the increased gene activities. It is unknown what protein or mechanism mediated the reverse transcription. The GO category “RNA-dependent DNA replication” was enriched during both the retroposition bouts, suggesting retrotransposons might be the major driver of retroposition. Lee etal. (2014) proposed the LTR-retrotransposon, Ty1/copia, as a candidate but direct evidence is still lacking ( Lee etal. 2014 ). Retroposition in the human genome is mediated by LINE1 ( Pavlicek etal. 2007 ). Pavlicek et al. proposed that mRNA with higher stability, smaller length, and which is translated on free as opposed to bound ribosomes will have a higher chance of retroposition. However, we did not find that any of these terms affect dinoflagellate retrogenes suggesting that machinery mediating retroposition differs from that in humans. Furthermore, we found that some retrogenes appeared to have retained the introns from their parents. Although these retrogenes were filtered out in our downstream analysis in order to exclude possible paralog duplicates in the genome, the presence of these genes suggests that retroposition could happen before or during the splicing process of introns. This could also account for the fact that retrogenes frequently insert into a locus near their parents. If true, the retroposition process in dinoflagellate is likely to occur before the transcripts are exported to the cytoplasm. Ancient Gene Activities Predicted by the Number of Retrogenes Both a previous report ( Pavlicek etal. 2007 ) and our own data ( table 2 ) indicate that more highly expressed genes have a greater chance to be retroposed. As a result, more retrocopies in a gene family born, for example, during the second retroposition bout, suggest that this family was actively transcribed at that time. This, in principle, provides an estimate of the transcript profiles at the time of the first and second retroposition bout. To attempt to predict cell behavior from these transcript profiles, we categorized the retroposed genes into GO biological processes and performed an enrichment analysis. These results showed that biological processes related to retroposition, such as “RNA-dependent DNA replication” and “DNA replication,” were activated at both the bouts, as expected. There is also enrichment in the “response to stress” at both the bouts of retroposition, supporting the idea that retroposition may have been stimulated by environmental changes. In line with our conclusion that the second retroposition occurred after the emergence of Symbiodinium , processes related to photosynthesis, carbon fixation (i.e., photosystem II stabilization, ATP synthesis coupled proton transport), and symbiosis establishment were enriched at the second bout. The expansion and diversification of these families may have facilitated the radiation of Symbiodinium strains into new hosts ( fig. 2 B ). We have previously reported ( Lin etal. 2015 ) that the biosynthesis pathway of the N-Glycans essential for host recognition had diverged between S. kawagutii and S. minutum . Interestingly, several important transport processes including “transmembrane transport” and “ion transport” were enriched during the first bout. This suggests the cell may have been ready for the evolution of symbiotic capacities. A Self-Enforcing Mechanism Is Adopted by the Dinoflagellate Genome The retroposition of highly expressed genes may have had the effect of “fixing” an increased level of gene expression by increasing its gene copy in the genome. This is likely to be particularly acute for any genes stimulated by the same factors that activate the retroposition process. Our data implies that some dramatic environmental changes had led to the increased levels of retroposition, and other genes whose expression levels increased in response to these changes would thus also have a higher chance of retroposition. The increased abundance of transcripts of these genes, when fixed in the genome, could then promote the evolution of new adaptive phenotypes ( fig. 4 B ). This mechanism presumably enabled dinoflagellate genomes to accumulate genes engaged in the resistance against environmental stresses, as observed in the two sequenced dinoflagellate genomes. In summary, gene retroposition is widespread in the two Symbiodinium genomes examined here. We suggest that preservation of the retroposed genes may be facilitated by trans -splicing of a leader sequence containing a basal transcriptional promoter element to the 5′-UTR of the mRNAs. Our analysis of the time when retroposition was most active during the evolutionary history of Symbiodinium genomes reveals two major bouts of gene retroposition, one during the emergence of Symbiodinium and the second during its radiation among different hosts. The retroposition events and the retrogenes enriched during these processes likely played critical roles in driving genome evolution, especially genome expansion and shaping of the phenotype. This unusual evolutionary mechanism for dinoflagellate genome evolution may have been driven by environmental factors stimulating gene expression, followed by fixation of the increased transcript levels due to an increase in the number of gene copies per gene family ( fig. 4 B ). It appears that large-scale retroposition, instead of WGD or SD, has been the major mechanism driving the evolution of dinoflagellate genomes."
} | 5,547 |
38541495 | PMC10972417 | pmc | 8,896 | {
"abstract": "The study examined the correlation between collagen coating damage and self-healing under various tribological conditions. It confirmed that the friction coefficient and degree of damage on the collagen coating varied based on contact and sliding conditions. The friction coefficient, measured at 0.56 for a single sliding cycle under a 350 mN normal load, demonstrated a notable decrease to 0.46 for 2295 cycles under 30 mN, further reducing to 0.15 for 90 cycles under a 20 mN normal load. As the normal load increased, the friction coefficient decreased, and with repeated sliding cycles under the same load, the coefficient also decreased. Water droplets induced a self-healing effect on collagen coating, causing wear tracks to vanish as fibers absorbed water. Severe wear tracks, with broken fibers and peeled coating, showed limited self-healing. In contrast, mild wear tracks, with compressed yet connected fibers, exhibited the self-healing phenomenon, making the wear tracks disappear. Real-time observations during 90 cycles under a 20 mN normal load highlighted the formation of mild wear tracks with intact collagen fibers, providing quantitative insights into self-healing characteristics. To preserve the self-healing effect of the collagen coating, it is essential to ensure tribological conditions during contact and sliding that prevent the disconnection of collagen fibers.",
"conclusion": "4. Conclusions In this study, a thin-film collagen coating was formed using a protein solution extracted from the tail of a rat. The inside of the collagen coating had a structure in which thin collagen fibers were intricately entangled in the form of a net. The friction and wear characteristics of collagen coatings were evaluated under various normal-load and sliding conditions. The self-healing phenomenon of the wear track formed on the surface of the collagen coating was observed according to each contact and sliding motion condition, and the conditions for the self-healing effect of the collagen coating were confirmed. The average friction coefficient was measured as 0.56 for 1 cycle of sliding under a normal load of 350 mN, 0.46 for 2295 cycles of sliding under 30 mN, and 0.15 for 90 cycles of sliding under 20 mN. As the normal load decreased, the friction coefficient generally decreased. Under the same normal load, the friction coefficient and wear rate tended to decrease as the number of sliding cycles increased. In the experiment in which the sliding motion was performed while changing the length of the sliding stroke, the transition period of the friction coefficient was clearly shown for each sliding stroke section. In the case of the sliding motion being performed for 90 cycles under a normal load of 20 mN while observing the degree of wear track formation every 30 cycles in real time, the collagen fibers were not damaged and a mild wear track was formed, whereas in all other experiments, a severe wear track was formed in which the collagen fibers were completely broken and the coating peeled off. In the case of severe wear tracks, the self-healing effect of the collagen coating did not emerge, whereas the mild wear tracks self-healed in contact with water. Therefore, through this study, the conditions of contact and sliding motion required to maintain the self-healing properties of the collagen coating were identified without completely damaging the collagen fibers. This study anticipates progress in the realm of bio-inspired coatings, capitalizing on the specifically tailored self-healing attributes of collagen coatings to enhance durability across varied applications. The potential impacts span crucial areas like medical devices, wearable technologies, and micro/nano systems, ensuring prolonged functionality with minimal maintenance. A deep understanding of collagen coatings’ self-healing mechanism provides valuable insights, fostering innovative solutions in resilient biomaterial development and contributing to advancements in biomedical engineering.",
"introduction": "1. Introduction The phenomenon in which a material damaged by an external force returns to its original shape is called self-healing. Studies on self-healing properties have been conducted using various materials such as metals, ceramics, and polymers [ 1 , 2 , 3 , 4 , 5 ]. In particular, studies on polymeric materials with self-healing properties based on chemical reactions have been conducted in various fields [ 6 , 7 ]. Studies have been reported to develop composites based on microcapsules containing healing materials, microvascular structures capable of supplying healing materials, or polymer materials with self-healing abilities [ 8 , 9 , 10 , 11 ]. Since these studies induce self-healing through chemical reactions using healing agents, it is difficult to apply them to areas related to biomaterials or areas where chemical contamination is a concern. In this respect, a method using collagen or hydrogel materials, which are biocompatible and do not require reaction with other chemicals, is required. Because collagen and hydrogel materials have swelling properties and self-healing abilities, they can be usefully applied in various fields related to medical/biomaterials [ 12 , 13 , 14 ]. In recent studies, there has been a notable advancement in the development of conductive strain sensors based on collagen and hydrogel materials, showcasing self-healing characteristics. These sensors have found application in detecting human movements and signals, demonstrating their utility in various academic domains, including self-healing sensors, electronic skins, and wearable electronics [ 15 , 16 , 17 , 18 ]. Especially, they are applied as a surface coating material for micro/nano systems, and a self-healing effect that restores damaged areas can be expected. In this regard, studies on the self-healing properties of collagen coatings and hydrogel coatings formed as thin films have been reported [ 19 , 20 ]. The friction and wear characteristics of the collagen coating and the hydrogel coating were analyzed, and the self-healing property, by which wear scars formed on the surface of the coating disappear spontaneously on contact with water, was confirmed. The self-healing mechanism was explained as the wear marks formed on the surface of the collagen and hydrogel coatings disappear as the internal structure of the coating changes due to the swelling phenomenon, in which the collagen fibers and hydrogel fibers expand due to contact with water [ 21 ]. The repeatability of the self-healing ability of collagen and hydrogel coatings was also confirmed. However, it was confirmed that collagen and hydrogel materials have relatively weak mechanical strength at the level of polymer materials, and when the contact pressure is very high, the collagen and hydrogel coatings are permanently damaged and their self-healing ability disappears. When these materials are manufactured in bulk, they exhibit healing properties that enable them to reattach even after being completely cut. However, when applied as a thin film coating to protect the surface, it is judged that the presence or absence of the self-healing effect is determined according to the degree of damage. Bulk forms of collagen and hydrogel materials, even if subjected to permanent damage in certain areas, exhibit a high potential for self-healing by reacting in other regions. However, when formed as extremely thin film coatings, collagen and hydrogel coatings fail to manifest self-healing properties in the presence of localized permanent damage. Therefore, to enable the application of thin film collagen and hydrogel coatings in diverse fields, it becomes imperative to establish the conditions under which the self-healing effect can occur. It is confirmed that there is no study result that has quantitatively evaluated whether the self-healing effect appears according to the degree of damage of the collagen coating and the hydrogel coating. That is, the minimum standards for the degree of damage or contact sliding conditions for maintaining the self-healing ability of the collage coating and the hydrogel coating have not yet been established. In this study, we tried to confirm the minimum degree of surface damage to maintain the self-healing ability of the collagen coating. The friction and wear characteristics of the collagen coating were analyzed while adjusting the contact pressure applied to the coating surface and the number of sliding repetitions, and it was confirmed whether the damaged area became self-healing by contact with water. Through this, the minimum level of surface damage required for the self-healing effect to appear was confirmed, and the contact sliding tribological conditions were identified.",
"discussion": "3. Results and Discussion Figure 3 shows the surface image of the collagen coating. Figure 3 a shows the 3D surface morphology of the collagen coating captured by 3D-LSCM, and Figure 3 b,c show low- and high-magnification SEM images of the collagen coating, respectively. The collagen coating has a structure in which microscale collagen fibers formed using a protein solution extracted from rat tails are intricately entangled in a net shape [ 23 ]. From the three-dimensional surface morphology image taken through 3D-LSCM, it was confirmed that the average surface roughness of the collagen coating was approximately 3.53 μm. Through the SEM image enlarged with high magnification, it was confirmed that the inside of the collagen coating was randomly entangled with micro-scale thin collagen fibers in a network structure. The thickness of the collagen coating confirmed through SEM analysis of the cut side of the collagen-coated specimen was confirmed to be approximately 30 μm. Previous studies have reported that when the wear track part formed on the collagen coating comes into contact with water, a self-healing phenomenon occurs in which the wear track disappears as the collagen fibers absorb water and swell [ 24 ]. The repeatability of the self-healing effect of the collagen coating was verified, but it was necessary to verify whether the self-healing effect of the collagen fibers could be expressed even when the collagen fibers were completely cut off. The self-healing phenomenon of the collagen coating means that the wear track disappears because the collagen fibers, which have been compressed by repeated contact sliding movements, absorb water molecules and swell. If the collagen fibers are disconnected, the swelling effect of the collagen fibers is limited, and it is thought that the disconnected fiber space will remain empty. To confirm this, when the collagen-coated surface was picked and scratched with sharp tweezers, the collagen coating was dented and torn off, as shown in Figure 4 a. Because the collagen coating was scratched by applying artificially strong surface contact pressure, the entangled collagen fibers were completely broken. Water droplets were dropped on the damaged area to swell the collagen fibers, and the results of observing the surface after completely drying the water are shown in Figure 4 b. Although the morphology of the surface was changed due to the swelling of the collagen fibers, it was confirmed that the part that was forcibly torn off by the tweezers remained damaged without recovery. The hypothesis that the self-healing ability of the collagen coating disappears when the collagen fibers are broken was verified. In order to maintain the self-healing effect of the collagen coating, the need to understand the contact and sliding conditions in which the collagen fibers are maintained without permanent damage has emerged. In order to quantify the contact and sliding conditions, it is necessary to use a spherical ball as a counter tip and use a tribotester that can constantly adjust the normal load and sliding speed [ 25 , 26 ]. The friction and wear characteristics of the collagen coating according to normal load and sliding cycle were analyzed. After applying a vertical load of 350 mN to the collagen coating, a sliding motion of reciprocating once with a sliding stroke of 2 mm at a sliding speed of 1 mm/s was performed. Figure 5 a shows the change in friction coefficient during one cycle of sliding motion. At the beginning of sliding, the friction coefficient gradually increases, then saturates around 0.55 and shows a large fluctuation. In the latter part of the slide, the friction coefficient rose slightly and exceeded 0.6, and the overall average friction coefficient was 0.56. The fluctuation of the friction coefficient throughout the sliding motion and the increase in the friction coefficient in the second half of the sliding movement are assumed to have been caused by the wear of the collagen coating. Figure 5 b shows the wear track formed on the collagen-coated surface by 1 cycle of sliding motion. It can be seen that the collagen coating was completely peeled off to the extent that the substrate was exposed. In the middle of the wear track, partial collagen fibers were connected without being completely broken. Water droplets were dropped on the wear track formed in this way to check whether self-healing effect appeared. As shown in Figure 5 c, the wear track that was completely torn off and damaged did not recover. It can be seen that the parts where the collagen fibers were connected in the middle of the wear track maintained the connection as it was. A load of 350 mN induces a high contact pressure, and it is considered to be a severe contact and sliding condition to the extent that the collagen fibers are completely broken and separated even in a single sliding cycle [ 27 ]. In order to greatly alleviate the contact pressure condition for the collagen coating, a sliding motion was applied under a normal load of 30 mN, which was lowered by more than 10 times compared to the severe contact condition. Instead of greatly relaxing the contact pressure condition, the sliding cycle was increased, and the sliding stroke was increased from 1 mm to 2 mm and 4 mm in the middle of the sliding motion. The sliding motion was performed for 765 cycles for each sliding stroke section, and the sliding speed was set to 4 mm/s. Figure 6 a shows the change in the friction coefficient according to the total sliding cycle. First, in the case of sliding motion in the sliding stroke section of 1 mm, the friction coefficient increased from 0.4 to 0.5 at the beginning of the sliding cycle and was maintained, and then abrupt fluctuations occurred in the latter half of the sliding cycle. As the contact area between the surface area of the wear track formed on the collagen coating and the counter tip increased, the friction coefficient was thought to be increased by the repulsive force blocking the advancing direction of the tip. In addition, it is expected that serious damage to the collagen coating occurred in the section where the fluctuation of the friction coefficient occurred. After performing the sliding motion for 765 cycles with a sliding stroke of 1 mm, the sliding motion was performed with a sliding stroke of 2 mm immediately afterwards. This continuously damaged the part of the wear track corresponding to the 1 mm stroke that had been damaged to some extent, while forming a new wear track in the extended sliding stroke part. It is for the same reason that the sliding stroke was changed to 4 mm after 765 cycles of sliding motion. As the sliding stroke was changed, there was a difference in the friction coefficient for each section of the sliding stroke in the simultaneous sliding motion. In the 2 mm sliding stroke section, the friction coefficient was lower than that of the 1 mm sliding stroke, showing a level of 0.4, and as the sliding cycle increased, the friction coefficient gradually increased. The initial friction coefficient started at 0.4 in a sliding stroke of 1 mm, and as the wear track was formed, the friction coefficient increased due to the increase in the contact area. As the sliding stroke increased to 2 mm, the counter tip passed the new surface of the collagen coating, showing an initial friction coefficient of 0.4, and it is considered that the overall friction coefficient decreased. It is speculated that the reason why the friction coefficient did not increase rapidly in the newly increased 1 mm section of the 2 mm wear track is that the wear particles separated from the collagen coating by 765 cycles of sliding motion in the 1 mm sliding stroke adhered to the surface of the counter tip, causing a friction reduction effect. However, in the 1 mm sliding stroke section where the wear track was first formed in the 2 mm sliding stroke, the collagen coating was more damaged as the cumulative sliding motion exceeded 1000 cycles. It seems that the friction coefficient increased due to the increase in the contact area as the large wear track was formed in the newly contacted sliding stroke section of the remaining 1 mm. After sliding for 1530 cycles, when the sliding stroke was increased to 4 mm, the friction coefficient rapidly increased to a level of 0.5 or more, and then decreased after about 1700 cycles. In the already damaged 1 mm sliding stroke, collagen fibers were completely ripped off, which is expected to have accelerated the damage of the collagen coating. Due to the modified sliding stroke, there was a difference in surface height inside the wear track. It is believed that the newly formed wear tracks on the 2 mm and 4 mm sliding strokes were also torn off due to the shearing action caused by the frictional force acting horizontally on the surface. In this process, it seems that the friction coefficient rapidly rose and then rapidly decreased as it came into contact with the glass surface, which was the substrate. As shown in Figure 6 b, when comparing the average friction coefficient for each sliding stroke section, the friction coefficients for sliding strokes of 1 mm, 2 mm, and 4 mm were 0.50, 0.43, and 0.44, respectively. The highest value was reached at 1 mm sliding stroke and the lowest value at 2 mm sliding stroke. The average friction coefficient over the entire experiment was measured to be approximately 0.46. As shown in Figure 6 c, after a total of 2295 cycles of sliding, the wear tracks formed on the collagen-coated surface were completely torn off, and the substrate surface was also damaged. This was observed after dropping water droplets on completely damaged traces of the collagen-coated surface. As shown in Figure 6 d, it can be seen that the collagen fibers were completely broken, and the collagen coating was torn off, leaving severe wear tracks. Even if the normal load was greatly reduced and the contact pressure condition was alleviated, it can be judged that the collagen fibers were permanently damaged if the number of sliding cycles was high [ 28 ]. After applying a normal load of 20 mN to the collagen coating to further alleviate the contact pressure condition, a sliding friction motion was performed. In addition, in order to observe the formation process of the wear track in real time and perform the experiment only until the collagen fibers were completely broken, a friction test was performed with a custom-built tribotester mounted on the 3D-LSCM. The wear track formed on the surface of the collagen coating was observed every 30 cycles, and the change in friction coefficient was analyzed in real time, as shown in Figure 7 a. When a normal load of 20 mN was applied, the initial sliding motion started with a friction coefficient of approximately 0.26. After dropping rapidly below 0.2, the friction coefficient was maintained at approximately 0.18 and then gradually decreased. After observing the wear track that formed on the collagen coating at the point of 30 cycles, an additional sliding friction motion was performed again for 30 cycles, and as a result, the friction coefficient was maintained at 0.14, and this value was maintained for a total of 90 cycles thereafter. Figure 7 b–e show the average friction coefficient, 2D profile and 3D surface images of the wear track, wear volume and wear rate measured every 30 cycles of sliding motion, respectively. The average friction coefficient values measured for each 30, 60, and 90 cycles of sliding motion were 0.17, 0.14, and 0.13, respectively, and tended to decrease as the sliding cycle increased. The average friction coefficient in the entire experiment was approximately 0.15. In the case of the wear track measured every 30 cycles, the width and depth of the wear track increased as the sliding cycle increased, as shown in Figure 7 c. As a result of calculating and measuring the wear volume through 2D profile image analysis of the wear track, we found that it was 7.85 × 10 6 μm 3 , 8.28 × 10 6 μm 3 , and 9.82 × 10 6 μm 3 for 30, 60, and 90 cycles of sliding motion, respectively. As the sliding motion was repeated, it was a natural result that the amount of accumulated wear increased. However, when compared with the wear volume per cycle, that is, the wear rate, as the sliding cycle increased by 30 cycles, it was confirmed that the wear rate gradually decreased to 2.62 × 10 5 µm 3 /cycle, 1.38 × 10 5 µm 3 /cycle, and 1.09 × 10 5 µm 3 /cycle, respectively. This means that the rate of increase in the amount of wear was low compared to the increase in the number of sliding cycles. In the initial contact sliding motion, the collagen coating was compressed to form a wear track, but the collagen coating was not completely torn off within at least 90 cycles. This is to keep the amount of wear relatively small. However, when contact and sliding conditions became severe due to a further increase in the sliding cycle or an increase in the normal load, the collagen coating was completely peeled off and the amount of wear increased rapidly. After sliding for 90 cycles, it was found that the collagen fibers were much compressed from the measured width and depth of the wear track, and it was judged that further sliding would cause permanent damage to the collagen coating, so the experiment was stopped. Figure 7 f shows a 3D image of the wear tracks with the wear tracks not completely peeled off. It showed the shape of a deep and wide wear track, and it can be seen that the wear track was deeply compressed over 20 μm. When water drops were dropped on the finally formed wear track, the deep and wide wear track completely disappeared and self-healed to a flat surface, as shown in Figure 7 g. When the collagen fibers were not completely cut off, but only compressed, the collagen fibers absorbed water molecules and swelled to restore the compressed wear track area. The hypothesis was verified that when the contact pressure and sliding conditions were severe, the collagen fibers were broken and the self-healing effect disappeared as the coating was peeled off, and the self-healing effect appeared only when the collagen fibers were connected to each other. Figure 8 shows a schematic diagram explaining the mechanism of the self-healing effect according to the degree of damage of the collagen coating. In this study, when a normal load of 20 mN was applied and sliding motion was performed for 90 cycles, the collagen fibers in the collagen coating did not break and a mild wear track was formed in a compressed state, and the collagen fibers absorbed water molecules and swelled. As the collagen fibers were rearranged, a self-healing phenomenon occurred in which the wear track disappeared [ 29 ]. On the other hand, under severe contact and sliding conditions, such as when a normal load of 350 mN was applied and 2295 cycles of sliding motion were performed under a normal load of 30 mN, the collagen fibers were broken, the coating peeled off, and a severe wear track was formed. However, even in contact with water droplets, the swelling and realignment of collagen fibers were limited, so the wear tracks did not self-heal and permanent damage remained. Through this study, it was confirmed that in order to maintain the self-healing effect of the collagen coating, permanent damage resulting from breaking the collagen fibers should not occur. In addition, the normal load and sliding cycle conditions required for maintaining the connection without the breakage of the collagen fibers were identified."
} | 6,083 |
38577983 | PMC7615837 | pmc | 8,899 | {
"abstract": "Abstract The growth and success of many bacteria appear to rely on a stunning range of cooperative behaviours. But what is cooperation and how is it studied?"
} | 39 |
35108392 | PMC9118995 | pmc | 8,900 | {
"abstract": "Abstract Optimization of metabolism to maximize production of bio-based chemicals must consistently balance cellular resources for biocatalyst growth and desired compound synthesis. This mini-review discusses synthetic biology strategies for dynamically controlling expression of genes to enable dual-phase fermentations in which growth and production are separated into dedicated phases. Emphasis is placed on practical examples which can be reliably scaled to commercial production with the current state of technology. Recent case studies are presented, and recommendations are provided for environmental signals and genetic control circuits.",
"conclusion": "Concluding Remarks/Future Directions Examples given above use E. coli as production host and specific genetic control components discussed are from bacterial systems, but the concepts are more broadly applicable. Amyris has used a maltose switch-based system with multilevel control using yeast in commercial scale fermentation processes (Chua et al., 2019 ). Additionally, non-genetic metabolic control strategies providing separation of growth and production phases and strain stability have been successfully applied to improve industrial fermentation processes (Pooth et al., 2020 ; Sandoval et al., 2014 ). While the primary focus of this review is on highlighting proven genetic control strategies available as scalable engineering solutions for commercial bioprocesses, many other tools are in development and may soon find their way into industrial processes. The phosphate limitation switch-based multilevel control strategy employed by DMC in commercial application provides the first published example of its kind, but a commercial process using a well-developed system such as quorum sensing could reasonably be anticipated in the near future. Other technologies, such as optogenetics (Pouzet et al., 2020 ), may be further from commercial application but offer great potential for processes with tight, dynamic control far beyond what is currently possible. The use of contract genetic control system development and scale-down testing services to access synthetic biology expertise in development and validation of robust genetic control for a planned commercial process may enable more deployment of the technology in the near future. Large companies seeking competitive advantages for their bioprocesses but lacking the requisite expertise in these areas may benefit from these services, along with startups with ideas and funding looking to minimize research and development overhead and time to market.",
"introduction": "Introduction Commercial bioprocesses require reliable and practical control strategies to realize their tremendous potential. A recent report from the McKinsey Institute looking at applications of biological innovations stated that biology could be used to produce “60 percent of the physical inputs to the global economy” and that the applications studied could have a $4 trillion direct economic impact within the next 20 years (Chui et al., 2020 ). Many of the cell factory-dependent bioprocesses needed to produce those physical inputs will rely on tight control of host metabolism to be successful. While production of high-cost compounds such as those for the therapeutics market might permit less cost-efficient processes, production of commodity chemicals will not. A significant portion of the forecasted impact of biological innovation on the global economy will depend upon a cost advantage from biological production of commodity chemicals. Optimization of metabolism to maximize production efficiency must consistently balance cellular resources for biocatalyst and desired compound production. Toxicity of products and intermediates may need to be managed, often dynamically. Metabolic control strategies must be precisely and reliably matched to the process input and infrastructure constraints. Unlike infrastructure and process control design (Crater & Lievense, 2018 ; Hill et al., 2020 ), genetic control of production host metabolism is scientifically feasible but not currently precise or predictably engineered at industrial scale. How can the progress and promise of synthetic and systems biology for genetic control strategies be translated into practical and scalable engineering solutions for commercially successful bioprocesses? What tools are available and how ready are they for industrial application? What are the key considerations in choosing components of a metabolic control strategy for different types of processes? This review addresses these and other related questions, providing overall guidance towards developing a reliable, dynamically controlled biocatalyst that can be successfully taken to scale in industrial bioreactors. For the sake of brevity, the examples presented are limited to Escherichia coli ; however, the same overall design principles can be adapted to other host organisms."
} | 1,227 |
32047188 | PMC7012894 | pmc | 8,903 | {
"abstract": "When an ant dies within a nest, a worker ant carries its corpse away from the nest and drops it onto a pile known as an ant cemetery. These ant cemeteries form cluster patterns, and the dynamics of the corpse piles have been studied experimentally. The aim of the present study was to investigate how sensitivity to the presence of nest-mates would influence the corpse-carrying behaviour of ants, and how this would impact the dynamics of corpse pile clustering. This was achieved by developing an agent-based computational model in which simulated ‘ants’ (the agents) carry and drop ‘corpses’, resulting in the growth of the corpse pile. In the model, the probability of an ant dropping a corpse was tuned according to the presence or absence of nest-mates. The pile dynamics of the resulting model showed a partial match with the time series evolution of corpse piles observed with real ants in previous experimental studies. Although the switch of probabilities is a thought experiment, our results suggest that the corpse-carrying behaviour of worker ants might be influenced by interactions with their nest-mates because there is evidence that ant behaviour can be influenced by encounter rates.",
"introduction": "Introduction The collective behaviour of groups of ants seems to be the result of simple mechanisms at the agent level. This is an example of emergent behaviour, wherein the growth or evolution of more complex forms arises from simple rules. Studies have shown that the dynamics of ant systems at the macro-level can evolve via self-organising construction by worker ants (ants that perform tasks such as tending to the nest and looking after the larvae) 1 – 6 . Ants form various clustering patterns, which have been studied as examples of collective behaviours 4 – 8 , and they are known to aggregate many types of objects, including seeds, larvae and the corpses of their dead workers 9 – 11 . These clusters can form in the presence of environmental heterogeneities. Necrophoresis is a behaviour observed in social insects in which the corpses of dead members of the colony are carried away from the nest; this prevents infection spreading through the colony 12 . Necrophoresis by worker ants can cause the formation of piles of dead ants known as ant cemeteries. Through a combination of experiments and modelling, this phenomenon has been rigorously demonstrated to be the result of simple clustering rules. The worker ants pick up corpses, carry them away from the nest and then drop them to form piles, resulting in the formation of several clusters. Interestingly, surviving clusters have been observed to undergo sigmoidal growth, with some clusters growing while others disappearing, which suggests that cluster formation can be autocatalytic 13 . Observation has revealed that worker ants drop or pick up corpses according to the local density of corpses. An ant carrying a corpse drops that corpse on a specific cluster with a probability that increases with the cluster size; thus, clusters grow through self-enhancement 13 . This phenomenon is an example of stigmergy, a mechanism that mediates agent–environment interactions. Stigmergy has been extensively studied in social insects 14 – 17 . Stigmergy can explain indirect task coordination in which the actions of individual agents are stimulated by a trace left in the environment; examples include the construction of termite mounds and digging activity by leaf-cutting ants. This way, individual actions indirectly influence the future repetition of those actions 15 – 17 . With the ant cemeteries, the ants construct the clusters via a form of indirect communication, with the cluster size influencing the probability an ant will drop the corpse it is carrying. Stigmergy can be a form of self-organisation, and it can produce collective, complex and intelligent structures. Studies have suggested that the behaviour of ants can also be determined by earlier direct interactions with their nest-mates 18 – 21 . For example, harvest ants perform midden work by carrying objects such as a dead ant or an item of refuse to a pile, where they sort them. They seem to engage in this behaviour more often following encounters with other midden workers, with which they interact via brief antennal contacts 20 . The activities of worker ants are facilitated and regulated by interactions with nest-mates, and this can also contribute to the collective behaviours of colonies 18 , 19 . Pattern formation by worker ants can be coordinated by ant-to-ant interactions, with an individual ant’s contribution increased or reduced through interactions with nest-mates 21 . High encounter rates increase the probability of an ant performing a particular work activity. In this way, the collective building behaviours of insects can be influenced by their nest-mates; examples include the collective nest excavation of leaf-cutting ants and tunnel excavation by termites 22 , 23 . Simulations, such as agent-based computational models, have been widely used to investigate the collective behaviours and decision-making processes of insects, including ants 24 – 28 . This methodology has been applied to the formation of ant cemeteries, with the development of an activator–substrate model, a complex function describing the time evolution of a local cluster of corpses 13 . Models such as this assume that ants can detect the number of corpses within their detection areas; this seems to have a direct influence on the explosive increase in pile size. The dropping rate per corpse-carrying ant seems to increase with the number of corpses detectable by the ant, reaching an asymptotic value. However, it is unlikely that an ant is able to estimate that value accurately. Simpler mechanisms might make more sense; for example, an ant might drop a corpse it is carrying based on a threshold number of corpses, following a step function. Threshold models are often proposed to describe the emergent behaviour of social insects 29 , 30 . The aim of this study was to develop an agent-based model that describes the growth of ant corpse piles. It hypothesises two criteria for determining whether an ant drops the corpse it is carrying: the ant encounters more than a threshold number of corpses, and there is at least one nest–mate present locally. Thus, even if an ant detects the threshold number of corpses, it still tends to leave that location without dropping the corpse if its nest-mates are locally absent. This second criterion fills a gap between the dropping rate obtained directly from a threshold function and the experimentally observed dropping rate per corpse-carrying ant. For the sake of simplicity, these criteria were tested by developing a one-dimensional pile growth model in which the simulated ants could only move in two directions and corpse piles and more than one ant could occupy the same position. The results of this model were in agreement with aspects of those found in experimental studies of ant cemeteries.",
"discussion": "Discussion This study investigated the growth of ant corpse piles by developing an agent-based model. In the model, the probability for whether an ant dropped the corpse it was carrying depended on whether there was another ant at the same location. The resulting model reproduced phenomena observed in the pile growth of real ants 13 . In the model, many small clusters of corpses were produced at an early stage, but after the number of piles reached a maximum value, some piles grew and others disappeared through a process of autocatalytic cluster formation. Models in previous studies have generally assumed phenomenological results and functions directly obtained from experimental studies 2 . However, recent studies of macro-behaviours have used a simple threshold rule rather than a complicated function 29 , 30 . Previous models have hypothesised that ants can accurately detect the number of corpses within their detection areas. In our models, the agents did not need to accurately detect the number of corpses within their corpses, which is a different assumption from previous studies. However, as shown in results obtained from the Threshold-only Model, it is not enough for the agents to use only a simple threshold rule. Agent–agent interactions, namely nest–mate presence, will enable agents to exhibit a more complex pattern at the macro-level. On the other hand, in the Interaction Model, the carrying agents do not always drop the corpse just because the local cluster of corpses exceeds the threshold, which might result in facilitating the growth of specific clusters and inhibiting the growth of others. Therefore, agent–agent interactions will make a difference with respect to the probability of dropping a corpse between these two models. In fact, this probability in the Interaction Model appears to reproduce the observed probability of dropping a corpse in the process of the formation of clusters of ant corpses. Although the agents randomly walk, their spatial distribution can be unbalanced, triggering the effect of nest–mate presence on the probability of drop. Therefore, nest–mate presence might be an important piece of information for ant workers to determine whether the pile growth is stable or temporal. This is because they cannot recognise whether a corpse pile at their location will grow without pause or decrease as time goes on. Some corpses within a pile may be carried away by agents before the pile has grown in size, whereas those that are not fully grown can become fragmented. Ant workers may present complex and collective patterns by effectively using such information combined with stigmergy. In that regard, the present model may describe how to model the decision-making processes of ants involved in the growth of ant corpse piles. Although there is no direct evidence that nest–mate presence has an influence on the coordination of the probability of dropping corpses, such an effect may bridge the gap between threshold response at the micro-level and the spatial macro-pattern of ant cemeteries. In fact, the growth of ant cemeteries and their pattern formation seem to be influenced by the surrounding environment 31 , 32 ; because individual ants adjust their behaviours to the environment. Furthermore, the performance of worker ants can be increased or decreased according to interactions with their nest-mates 20 . In the present model, the contribution of each ant to the construction of clusters was influenced by whether other ants were present at the cluster location. Such a strategy combined with stigmergy may explain the construction and the formation of ant cemeteries. Nest–mate interactions are known to regulate and facilitate some building activities in insects 22 , 23 , and similar behaviours have been observed in other aggregation behaviours by ants, such as bridge formation 33 , 34 . The varying contribution of individual ants may therefore be a key effect that explains the dynamics of ant cemeteries and their activity."
} | 2,750 |
35541583 | PMC9076477 | pmc | 8,904 | {
"abstract": "A pristine soil environment supports a healthy soil biodiversity, which is often polluted with recalcitrant compounds. The bioelectrochemical remediation of the contaminated soils using bioelectrochemical systems (BESs) is gaining significant attention with respect to the restoration of the soil ecosystem. In this direction, a microbial fuel cell (MFC, an application of BES), was employed for the treatment of total petroleum hydrocarbons (TPHs) in a soil microenvironment at three ranges of pollution (loading conditions – 320, 590 and 840 mg TPH per L). TPHs degraded effectively in the soil-electrode vicinity in the range of 158 mg TPH R per L (320 mg TPH per L) and 356 mg TPH R per L (840 mg TPH per L). The study also demosntrated a maximum bioelectrogenesis of 286.7 mW m −2 (448 mV at 100 Ω) at the highest TPH loading concentration studied (840 mg TPH per L). The conditions prevailing in the soil MFC also facilitated the removal of sulfates (114 mg SO 4 2− per L; 62.64%) and the removal of total dissolved solids (910 mg TDS per L, 12.08%) at an 840 mg TPH per L loading condition. The pH of the outlet wastewater prevailing in the mild alkaline range of 7.6 and 8.4, along with improved sulfate and TPH removal in the respective conditions suggested suitable conditions for sulfate-reducing bacteria (SRB). This study also signified the sustainability of the process for the effective treatment of hydrocarbon contaminated soil that also generates green energy.",
"conclusion": "4 Conclusions The present study demonstrated that the placement of the anode with a well-developed electroactive biofilm in the soil allows for the petroleum-based hydrocarbons contaminants to assist in the bioelectricity generation. It was also observed that the bioelectricity generation (average, 276 mW m 2 ; 440 mV) increased with the increase in the hydrocarbon contaminants. The use of the soil MFCs was found to be feasible for the treatment of highly petroleum hydrocarbon contaminated soils (up to 840 mg TPH per L). The removal of the total dissolved solids (12.08% removal) and sulfates (62.64% removal) was attributed to the power generated in the soil MFC system. The improvement in the sulfate removal with an increase in the TPH loading conditions demonstrated the role of organic matter in the sulfate reduction (320 mg TPH per L, 49 mg L −1 ; 840 mg L −1 , 113 mg L −1 ). In this study, the importance of bioelectrochemistry in the treatment of complex and recalcitrant compounds has been illustrated. The pattern of the diesel range organic (DROs) degradation depicted that the high carbon molecules degraded to lower carbon molecules. The RDAP analysis also depicted the sustainability of the soil MFCs.",
"introduction": "1. Introduction The rapid expansion and development of industrial activities increased the discharge of pollutants into the environment that causes a deterioration of water resources. It consequently resulted in the deterioration of the quality of soil and sediments, which further affected the quality of the entire ecosystem. 1,2 Currently, petroleum-based resources are considered as the major source of energy required for industrial and human activities. This resulted in the release of petroleum hydrocarbon pollutants to various components of the environment such as soil and water. 3–6 Alkanes, aromatics, nitrogen–sulphur–oxygen compounds (NSO) and asphaltenes are the main fractions of petroleum hydrocarbons. 7,8 The alkanes are C 1 (methane) to C 40 compounds with straight-chain or branched-chain hydrocarbons, whereas aromatic compounds contains at least one benzene ring in the structure. NSOs and asphaltenes are complex molecules that exert complex toxic effects on the ecosystem. The release of crude oil constituents into the environment has the potential to exhibit adverse effects on ecology and human health. 9–11 Environmental contamination and ecosystem deterioration are a global concern that demands for innovative and cost-effective remediation technologies. 12,13 Total petroleum hydrocarbons (TPHs) are lethal to beneficial soil organisms and human beings, and create a serious concern among the research community and governments. 14,15 Chronic petrogenic contamination shows adverse effects on various components of the ecosystem. 16,17 This has demanded the research on remediation techniques for petroleum contamination. Based on the complexity of the pollutants and sites of contamination, and to avoid secondary pollutants from the treatment process, biological processes show merits over chemical and physico-chemical processes. 18,19 Currently, an increased number of bioremediation technologies using plants (phytoremediation), bacteria, algae and fungi are being explored for the degradation of TPH-contaminated surfaces and subsurface soils. 8,20–22 A novel method called bioelectrochemical systems (BES), such as microbial fuel cells (MFCs) and microbial electrolysis cells (MECs), which deal with the interface of a biofilm and electrode environment, has shown superior functions for the degradation of petrogenic pollutants. 23–25 The degradation or oxidation of pollutants in an anodic environment has facilitated green energy generation. Several other pollutants, such as sulfates, nitrates, and chemical oxygen demand (COD), can also be treated simultaneously by BES. 26,27 The electrochemically-active biofilm present on the anode of MFCs generates bioelectricity from the treatment of organics, whereas MECs treat specific pollutants under mild external applied potentials for different types of pollutants. 28–30 The biological and electrochemical effects combined in BES systems create hybrid mechanisms that facilitate complex bio-electrochemical treatment processes. 32–34 The bioelectrochemical energy generating from the system is green in nature and non-pollutant. It treats complex pollutants deteriorating the environment. Herein, the present study was aimed to remediate soil contaminated with recalcitrant petroleum-based hydrocarbons via a novel soil MFC configuration. Petroleum hydrocarbons present in the soil act as an electron donor and carbon electrodes as electron acceptors that develop a biopotential in the soil environment. The combination of a generated biopotential and associated microbial activity help to generate biological electricity in concurrence with bioelectro-remediation. At this stage of the study, the primary focus was towards soil bioelectro-remediation, MFC performance and its evaluation of sustainability, rather than the microbial aspects.",
"discussion": "3. Results and discussion 3.1 Degradation of petroleum hydrocarbons in the subsurface soil environment of MFC Petroleum hydrocarbons are complex, but can be degraded by specific bacteria at a relatively lower rate compared to simple wastewater of a biological origin. Bioelectrochemical systems adapt mixed consortia from environmental sources for the degradation of various types of pollutants, which also helps for the generation of green energy in the form of bioelectricity. 40 In the present study, the observed trend of TPH degradation at the three different loading conditions clearly depicted the efficiency of the soil MFC ( Fig. 1 ). During the first cycle of operation (inlet substrate concentration, 320 mg TPH per L), a substrate degradation of 142 mg TPH per L (44.38%) was recorded during the 7 days of operation. Among the 3 cycles of operation with 320 mg TPH per L, a maximum TPH degradation of 158 mg L −1 was registered, which represented a 49.38% removal efficiency (average of 320 mg TPH per L, 146.7 mg TPH R per L and 45.83%). Control experiments were performed to compare the metabolic action of soil bacteria alone and to identify the specific function of the anodic electrogenic biofilm on TPH degradation. Under the same loading condition (320 mg TPH per L), the control experiments depicted a maximum TPH degradation of 28 mg L −1 (TPH degradation efficiency, 8.75%) with an average degradation of 23.3 mg TPH per L (7.29%) ( Fig. 1b ). Fig. 1 Degradation of the total petroleum hydrocarbons during the three loading conditions of the (a) soil MFC and in the (b) control operations. Cycle numbers 1–3 represent 320 mg TPH per L, 4–6 represent 590 mg TPH per L and 7–8 represent 840 mg TPH per L. The increase in the TPH loading to 590 mg TPH per L indicated a substantial improvement in the substrate removal, registering a maximum value of 248 mg L −1 (42.03%). The average TPH degradation in the three cycles of operation with 590 mg TPH per L was 235 mg L −1 , which represented a 39.9% removal efficiency. The increase in the TPH loading resulted in an improvement in TPH degradation. The control reactor operation with the same TPH loading of 590 mg TPH per L, depicted a maximum TPH degradation of 45 mg L −1 (TPH degradation efficiency, 7.63%) and an average degradation of 38.6 mg TPH per L (6.55%). The further increase in the TPH concentration to 840 mg TPH per L also resulted in a similar trend with respect to substrate degradation. The maximum and average TPH removal were recorded as 356 mg L −1 and 340 mg L −1 , respectively. Similarly, the TPH degradation efficiency was also recorded as 42.38% (maximum) and 40.5% (average) ( Fig. 1a ). The substrate degradation in the control reactor was limited to a maximum degradation of 40 mg TPH per L (average, 35 mg TPH per L) with an efficiency of 4.76% (average, 4.17%) ( Fig. 1b ). Interestingly, the TPH degradation efficiency of the MFC operation at 590 and 840 mg TPH per L loading were comparable, which also claimed suitability of the MFC system for bioelectro-remediation of hydrocarbon contaminated soils. The continuous recirculation of the wastewater in the soil MFC was maintained, which might have facilitated the passing of the wastewater to the vicinity of the electrodes and the exposure of the wastewater to the bioelectrochemical reactions. Since the recirculation of the liquid content was designed in an up-flow direction, a uniform concentration of the pollutants in the soil environment was possible. However, the application of this method directly to the contaminated sites requires an understanding of the geological and physico-chemical nature of the respective soils. 3.2 Diesel range organics degradation The primary details of the evolution of TPH in the MFC were identified by tracing the DROs. DROs are highly regarded as an intermediate fraction of the petroleum refining process. Diesel hydrocarbon compounds (specifically long chain alkenes) are the most abundant compounds in DROs, which were considered to identify the evolution pattern in the soil MFC and control operations. Among the three TPH concentrations studied, total DRO degradation was increased with the increase in the available TPH concentrations ( Fig. 2 ). The highest DRO degradation of 34.69 mg DROs per L was registered with an 840 mg TPH per L variation, followed by 590 mg TPH per L (25.98 mg DROs per L) and 320 mg TPH per L (17.16 mg DROs per L), which clearly depicted the bioelectrochemical potential of the MFC towards DRO removal ( Fig. 2a ). On the contrary to the DRO removal, its removal efficiency decreased with an increase in the total DRO concentration. In the case of the 320 mg TPH per L loading condition, a maximum DRO removal efficiency of 38.78% was recorded, which further decreased to 34.41 and 31.52% for the 590 and 840 mg TPH per L conditions, respectively ( Fig. 2b ). Both the removal and removal efficiency of the DROs was several times higher than the control experiments, which also corroborated the capability of the MFC towards TPH treatment ( Fig. 2a and b ). Fig. 2 Pattern of the diesel range organic (DRO) degradation in the bioelectroremediation process and respective control operations. (a) Total concentration of the DROs, (b) DRO removal concentration and DRO removal efficiency and (c) concentrations of diverse DROs in the MFC and control operation. Further analysis was performed to identify the individual DRO compound degradation in the MFC and control operations. Compared to the control operation, the MFC visualized the removal of the DROs with the three loading conditions studied. All nine DRO compounds analyzed degraded significantly over the MFC operation. However, a higher affinity towards bioelectrochemical degradation was regarded with n -octadecane, n -eicosene and n -decosane (DROs). In all the TPH concentrations, these three compounds collectively contributed more than 50% of the total DROs degraded. Contrary to the decrease in the concentration over the MFC treatment, n -decane showed an increase in its concentration under the 320 mg TPH per L loading condition. n -Decane was the simplest among the analyzed DROs. The gradual breakdown of the higher TPH compounds or DROs via bioelectrochemical remediation might have resulted in the n -decane production that resulted in an increased concentration. The higher removal efficiency with the larger molecules rather than smaller molecules such as n -docane, surrogate and n -tetradecane might have been due to the production of smaller compounds from the degradation of the higher carbon compounds such as n -octadecane, n -eicosene and n -decosane. A previous study that was performed with petroleum refinery wastewater under applied potentials showed a higher efficiency towards DRO removal (>80%). 41 The limited efficiency in the present study was attributed to the soil environment, which provided relatively less contact with the electrode surface. On the other hand, the degradation in the present study was due to the bioelectrochemical potential that developed by the system rather than the external applied potential in the previous study. 3.3 Soil MFC performance 3.3.1 Power generation from TPH in soil microenvironment of MFC The electrochemically active bacteria that immobilized on the anode electrode were capable of producing electrons from the degradation of the organic compounds in the vicinity. The incorporation of the adapted bioanode to the hydrocarbon substrates generated a swift response to the bioelectrogenesis and resulted in a power density of 45.3 mW m −2 as well as a potential of 178 mV across the 100 Ω resistance under an operation of 320 mg TPH per L. A gradual improvement in the power generation was observed with time as well as a recorded 110.4 mW m −2 (278 mV) and 146.3 mW m −2 (318 mV) in the second and third day of operation, respectively. Furthermore, the MFC exhibited a power generation with marginal fluctuations. During the first cycle of operation, a maximum power density of 147.2 mW m −2 (321 mV) was recorded that gradually increased in each subsequent operating cycle ( Fig. 3a ). During the second and third operating cycle, the power density was registered as 153.7 mW m −2 (328 mV) and 165.1 mW m −2 (340 mV), respectively. The three operating cycles evaluated with 320 mg TPH per L showed a power density in the narrow range of 147.2 and 165.1 mW m −2 , which depicted a stable operation. Fig. 3 (a) Bioelectricity generation potential of the soil-based microbial fuel cell depicting the voltage generation during the three TPH loading conditions with a 100 ohms resistor. (b) Polarization behavior of the MFC under operation using the three TPH loading conditions. Starting from the fourth cycle of operation, the TPH loading of 590 mg TPH per L was adapted and operated consecutively for three cycles. Among the three operating cycles with 590 mg TPH per L in cycle 4, a maximum power density of 230.9 mW m −2 (402 mV) was developed across the resistance. Considering the average power densities for the 320 mg TPH per L loading operation (155.3 mW m −2 ) and 590 mg TPH per L loading (222.6 mW m −2 ), there was a significant power improvement (67.2 mW m −2 , about 40%). This clearly depicted the contribution of the available TPH in the soil matrix towards the bioelectricity generation. Upon the completion of the 3 cycles with 590 mg TPH per L, the soil MFC was fed with the next highest loading variation of 840 mg TPH per L and operated for 2 consecutive cycles, which also showed a positive sign of improved power generation. Upon the onset of cycle 7 operation (840 mg TPH per L), the power density improved to 265.4 mW m −2 , which was more than 20% over the previous cycle with 590 mg TPH per L (211.8 mW m −2 ) ( Fig. 3a ). Among all the TPH loading concentrations studied, a maximum power density of 286.7 mW m −2 was registered in the last cycle operated at 840 mg TPH per L. Among the three different TPH loading concentrations, 320 and 840 mg TPH per L showed a marginal increase in the power density with every feeding event. Contrary to the 320 and 840 mg TPH per L concentrations, the 590 mg TPH per L concentration exhibited a marginal drop. This suggested that the variation in the power density of the different operating cycles was due to a biological phenomenon. \n I – V curves (polarization behavior) were recorded from the three substrate (TPH) loading concentrations using a variable resistor box. The maximum performance period of the MFC with respect to the power generation was considered as an optimum to evaluate the bioelectrochemical behavior of the MFC. The shift in recording the polarization behavior from one point to another was allowed to recover the open circuit potential to the maximum. This allowed for the recording of the bioelectrochemical activity with minimum error. In the present study that used the three loading concentrations of TPH as a substrate, the maximum electrochemical activity was found to be in accordance with their respective bioelectrogenic activities. The maximum current density and power density were increased with the increase in the TPH loading concentrations ( Fig. 3b ). A maximum power density of 158.1 mW m −2 (at 400 Ω; current density, 388.6 mA m −2 ) was registered with the 320 mg TPH per L substrate concentration. In the case of the 590 mg TPH per L concentration, the maximum power density was increased to 190.6 mW m −2 (at 200 Ω, 882.5 mA m −2 ). With the increase in substrate loading from 320 to 590 mg TPH per L, the power density improved by 20%. Furthermore, after increasing the TPH loading to 840 mg TPH per L, a maximum power density of 239.0 mW m −2 (1127.5 mA m −2 ) was recorded with a 100 Ω resistor ( Fig. 3b ). The polarization study also suggested that the electron discharge capacity for the MFC under the respective operating conditions was adapted. The point at which the maximum power density was recorded was denoted as the cell design point. In the present study, the cell design point decreased with the increase in the TPH loading concentration (400 Ω, 320 mg TPH per L; 200 Ω, 590 mg TPH per L and 100 Ω, 840 mg TPH per L). Considering the lowest cell design point from all the loading TPH concentrations, i.e. 100 Ω, the power density was found to increase with an increase in the substrate concentration (320 mg TPH per L, 138.0 mW m −2 ; 590 mg TPH per L, 182.0 mW m −2 and 840 mg TPH per L, 239.0 mW m −2 ). This phenomenon suggested that the electron discharge and bioelectrochemical response were directly proportional to the substrate concentration used for bioelectrogenesis. 3.3.2 Half-cell potentials and sustainable power The half-cell potentials of the MFCs determined the efficiency of the individual performance of the anode and cathode processes. The anode and cathode potentials were recorded across the different resistances (50 Ω – 30 kΩ) using a variable resistor box (similar to I – V curve analysis). The details of the anode and cathode potentials against the external resistance are presented in Fig. 4a . The air cathode configuration used for the present study maintained a similar cathodic oxygen reduction mechanism and was not influenced by the TPH concentration in the soil column. Therefore, the cathodic potential was found to be almost stable (in the range of 338 to 357 mV) for all three variations studied. Among the three TPH concentrations studied, the anode potential was found to increase with an increase in the substrate concentration, which confirmed the contribution of the electrons generated from the TPH degradation towards potential development. A maximum anode potential of −366 mV was recorded with the 840 mg TPH per L concentration followed by the 590 (−323 mV) and 320 mg TPH per L (−264 mV) concentrations. The anodic potential trend was found to be very similar to the polarization behavior exhibited by the MFC at the respective concentrations. Fig. 4 Evaluation of the sustainable performance of the soil MFC during the three TPH concentrations. (a) Anodic and cathodic potentials as well as the (b) sustainable power and resistance value at a relative decrease in the anodic potential (RDAP) point. The anode potential was found to decrease with the decrease in the external resistance, which depicted the electron discharge phenomenon. The lower resistance used in the anodic half-cell potential circuit facilitated the relatively higher electron discharge that resulted in a drop in the anodic potential. This phenomenon helped for the derivation of the relative decrease in the anodic potential (RDAP). The RDAP facilitated the sustainable resistor point for the stable performance of the MFC where the electron discharge and production reached equilibrium. The 320 mg TPH per L concentration unveiled a RDAP at 7.0 kΩ of the external resistance with a sustainable power generation of 0.45 mW ( Fig. 4b and S2 † ). The RDAP resistance value increased with the increase in the TPH loading concentrations. In the case of the 590 and 840 mg TPH per L concentrations, the sustainable power was found to be 0.22 mW (at 15.5 kΩ) and 0.20 mW (at 18.0 kΩ), respectively. 3.4 Sulfate removal at various TPH concentrations Sulfate is one of the major pollutants present in the waste/wastewater generated from petroleum industries. High sulfate concentrations present in wastewater are evidence for the adverse effects on the ecosystem. The toxic and acidic gases generated from the sulfate rich wastewater are either carcinogenic or create serious health problems, which affects the eyes, skin, lungs, intestines and nervous system. 42 Material corrosion is also another challenge that is associated with sulfate rich wastewater. 43,44 The reduced products may volatilize into the atmosphere and result in acid rain. Sulfate contamination also has adverse effects on the soil microenvironment, which may further led to groundwater contamination. 45,46 BESs have been considered as an effective method for sulfate removal. 35,47,48 In the present study bioelectricity generation was also concomitantly investigated for sulfate removal ( Fig. 5 ). The initial concentration of sulfate was kept constant, which allowed for identifying the influence of the organic matter on the sulfate removal. Thus, it was also determined that the sulfate removal increased with an increase in the TPH concentration. At the initial TPH loading condition of 320 mg TPH per L, the highest sulfate removal of 54 mg L −1 with an efficiency of 29.67% (average, 49 mg L −1 and 26.92%) was observed ( Fig. 5a ). When the soil MFC was shifted to 590 mg TPH per L, an improvement in sulfate removal was found to be significant, which registered a maximum sulfate removal of 84 mg L −1 (46.15%). The improvement in the sulfate removal was compared with the registered higher TPH degradation at the higher loading conditions ( Fig. 5a ). Similar results were also observed when the TPH concentration was further increased to 840 mg TPH per L, which resulted in the highest sulfate removal of 114 mg L −1 (average, 113 mg L −1 ) with a 62.64% removal efficiency (average, 62.08%). Sulfate removal in the control reactor operation was also evaluated under the three TPH loading concentrations, which was in the range of 20 to 27 mg L −1 ( Fig. 5b ). Since the sulfate concentration used was constant for all of the operating conditions, a significant difference in the sulfate removal and sulfate removal efficiency was not identified with respect to the initial TPH concentration. Apart from the sulfate degradation from the mixed bacteria in the MFCs, the sulfate-reducing bacteria (SRB) also played an important role in the power generation. SRB reduced the sulfate to sulfide and then sulfide was oxidized to elemental sulfur, which was deposited on the surface of the electrode, resulting in power generation. 47,49 In this case, sulfate also acted as a substrate for power generation. However, with the available analysis in the present study, the involvement of the sulfate reducing bacteria in the power generation could not be illustrated. The SRB species such as Desulfovibrio desulfuricans , Desulfuromonas acetoxidans and Desulfobulbus propionicus were confirmed to produce electricity with a concomitant sulfate reduction. 47–49 The sulfate reduction in BES also exhibited the advantage of generating value-added elemental sulfur from toxic sulfide. 50 Fig. 5 Sulfate removal pattern registered in the (a) soil MFC and (b) control operations during the three loading conditions of the petroleum-based hydrocarbons. Cycle numbers 1–3 represent 320 mg TPH per L, 4–6 represent 590 mg TPH per L and 7–8 represent 840 mg TPH per L. 3.5 Removal of dissolved solids Petroleum-based wastewaters such as petroleum refinery wastewater and produced water were characterized to have high solid concentrations (TDS). Produced water contains TDS in the range of 65 to 220 g L −1 , whereas refinery wastewater contains 15–45 g L −1 of TDS. 51 Accidents and improper wastewater management in petroleum industries may result in soil contamination. The presence of electroactive biofilms in the MFCs facilitated the development of the bioelectrogenic conditions. The direct electric field developed an electrochemical gradient, which separated the charged ions to the oppositely-charged electrodes. This phenomenon was attributed for the removal of the salts/TDS. 51,52 The employed wastewater in the present study had a TDS concentration of 7530 mg L −1 , which depicted a significant removal during the MFC operation through the three different TPH loading conditions. However, the TDS removal efficiency was found to depend on the concentration of TPH in the wastewater, which was directly proportional to the bioelectrogenic potential evidenced at the respective TPH loading concentration. During the 320 mg TPH per L loading condition, a maximum TDS removal of 690 mg L −1 (average, 580 mg L −1 ) was achieved with a maximum removal efficiency of 9.16% (average, 7.70%; Fig. 6 ). Under the 590 mg TPH per L loading conditions, the TDS removal was improved to 730 mg L −1 (average, 683 mg L −1 ) and a 9.69% efficiency (average, 9.07%). A further increase in the substrate loading to 840 mg TPH per L also depicted an improved TDS removal and removal efficiency of 910 mg L −1 and 12.08%, respectively. In the case of the control operation, the outlet TDS concentration was estimated in the range of 7540 to 7600 mg L −1 , which exhibited a marginal improvement in the TDS concentration (improvement in TDS, 10 to 70 mg L −1 ) from an inlet value of 7530 mg L −1 ( Fig. 6b ). The increase in the TDS concentration was irregular with the different TPH loading concentrations. A maximum TDS increase of 70 mg TDS per L was registered in the second cycle of the 590 mg TPH per L loading condition, whereas a minimum TDS increase was registered in the third cycle of 320 mg TPH per L and the first cycle of the 840 mg TPH per L loading conditions. The improvement in TDS might have been due to the evaporation loss that happened during the operation. Even though the inlet and outlet reservoirs used in the operational set-up were closed, it was possible to have an evaporation loss from the soil column of the reactor. The improvement in the TDS removal with respect to the TPH concentration correlated well with the power density registered at the respective conditions. This also supported the ion mobility phenomenon for the TDS removal. The membrane-less configuration adapted in the present study also facilitated the mobility of the ions towards the oppositely-charged electrodes. The microbial desalination cells (MDCs) having three chambered configurations also followed a similar principle. However, the presence of the middle chamber helped in the separation of the ions (desalination). 53 Apart from the ion mobility, the direct and indirect anodic oxidation phenomena also helped to adsorb the pollutants on the anode surface. 31 After eight cycles of operation for about 2 months, a gradual drop in the power density was observed. A detailed observation was made on the electrodes depicted scaling (salts deposition) on the electrode surface, which is called electrodeposition. 54,55 This might have hindered the electric conductivity of the electrode due to the drop in the recorded potential. The visible difference in the TDS removal from the control and MFC operations also clearly showed the influence of the electrochemical activity in removing TDS from the wastewater used in the operation. Fig. 6 (a) Removal of the total dissolved solids (TDS) during the three loading conditions of TPH in the soil MFC. (b) Outlet TDS values evaluated in the soil MFC and control operations (the dotted line represents the value of the inlet TDS). (c) Change in pH (from inlet pH of 7.82). Cycle numbers 1–3 represent 320 mg TPH per L, 4–6 represent 590 mg TPH per L and 7–8 represent 840 mg TPH per L. 3.6 Variation in the pH with TPH loading pH is a sensitive parameter that is influenced by many biological and chemical processes in the environment. In MFCs, a pH change can be observed with the variation in the substrate degradation and the change in the components of wastewater. The wastewater that was considered in the present study was found to show an alkaline pH of 7.82. The pH was found to change by the end of each operating cycle. However, the change in the pH was found to depend on the TPH concentration adapted for the respective cycle. In the case of 320 mg TPH per L, the wastewater pH demonstrated a marginal drop towards a neutral pH (average, pH 7.6), whereas with 590 mg TPH per L, the pH was shifted to alkaline conditions ( Fig. 6c ). The maximum change in the pH at 590 mg TPH per L was in cycle 6 (pH = 8.4). For the highest substrate concentration studied (840 mg TPH per L), the outlet pH was 8.32 (maximum). A visible difference in the pH of the control and MFC operations was identified. In the case of control operations, the outlet pH was found to be increased in all of the TPH loading conditions ( Fig. 6c ). In the case of 320 mg TPH per L, the average pH was registered as 7.90, which was significantly higher than that for the MFC operation (7.6). The outlet pH in the control operations was increased with the increase in the TPH loading conditions. The 590 mg TPH per L loading conditions depicted an average outlet pH of 8.12 (maximum pH, 8.15). Upon a further increase to an 840 mg TPH per L loading condition, stability in the outlet pH (8.1) was observed. Among the bioelectrogenic conditions and control operations, a visible difference in the outlet pH was observed. This also agreed with the influence of the metabolic activities of the bioelectrochemical bacteria in TPH metabolism compared to only the soil microflora. The change in ion concentration is one factor for the change in the pH. The change in conductivity of the produced water also influenced the pH during the electrochemical sulfur production by Jain et al. 56 The conditions that prevailed in the soil based MFC system may have stimulated electrochemical oxidation and oxidative desulfurization, etc. , which might have induced a change in the pH of the produced water. 57–60 The TDS removal that was registered during the MFC operation might have been one of the reasons for the change in the pH, which was determined upon evaluating the anion and cation concentrations in the wastewater and their respective charges. On the other hand, the change in the sulfate concentrations also predominantly changed the solution pH. The activity of the SRB improved with an increase in the availability of organic matter, which in turn resulted in the increased pH. 61 For the 590 and 840 mg TPH per L loading conditions, the higher organic matter available might have increased the activity of the SRB. Under these conditions, increased sulfate removal was identified. This might have resulted in the increase in the pH towards alkaline conditions."
} | 8,170 |
32180337 | PMC7415357 | pmc | 8,905 | {
"abstract": "Summary The transition to sustainable agriculture and horticulture is a societal challenge of global importance. Fertilization with a minimum impact on the environment can facilitate this. Organic fertilizers can play an important role, given their typical release pattern and production through resource recovery. Microbial fertilizers (MFs) constitute an emerging class of organic fertilizers and consist of dried microbial biomass, for instance produced on effluents from the food and beverage industry. In this study, three groups of organisms were tested as MFs: a high‐rate consortium aerobic bacteria (CAB), the microalga Arthrospira platensis (‘Spirulina’) and a purple non‐sulfur bacterium (PNSB) Rhodobacter sp. During storage as dry products, the MFs showed light hygroscopic activity, but the mineral and organic fractions remained stable over a storage period of 91 days. For biological tests, a reference organic fertilizer (ROF) was used as positive control, and a commercial organic growing medium (GM) as substrate. The mineralization patterns without and with plants were similar for all MFs and ROF, with more than 70% of the organic nitrogen mineralized in 77 days. In a first fertilization trial with parsley, all MFs showed equal performance compared to ROF, and the plant fresh weight was even higher with CAB fertilization. CAB was subsequently used in a follow‐up trial with petunia and resulted in elevated plant height, comparable chlorophyll content and a higher amount of flowers compared to ROF. Finally, a cost estimation for packed GM with supplemented fertilizer indicated that CAB and a blend of CAB/PNSB (85%/15%) were most cost competitive, with an increase of 6% and 7% in cost compared to ROF. In conclusion, as bio‐based fertilizers, MFs have the potential to contribute to sustainable plant nutrition, performing as good as a commercially available organic fertilizer, and to a circular economy.",
"conclusion": "Conclusions The individual and blended microbial fertilizers were shown to have an equal fertilization, mineralization and storage performance as a state‐of‐the‐art organic fertilizer. This means that these microbial fertilizers can be a viable option to set up production chains for local production of organic fertilizers. Particularly, CAB and a blend of CAB/PNSB (85%/15%) appear to be cost‐effective alternatives, with only an increase of 6% and 7%, respectively, in cost.",
"introduction": "Introduction On a global scale, the agro‐industry is expanding due to the growth of the world’s population, which is expected to reach almost 10 billion by 2050 (Food and Agriculture Organization of the United Nations, 2017 ). To ensure food security and sustainable agriculture and horticulture, adequate plant nutrition and hence fertilization is indispensable. In agriculture, the fertilizers are applied directly onto the fields to produce arable crops (e.g. grains, potatoes). In horticulture, these fertilizers are applied to the soil or added to growing media (GM) for an effective production of fruits, vegetables and ornamental plants under controlled conditions (Gruda, 2009 ; Dixon and Aldous, 2014 ). In Flanders, the production value of the horticultural market is 2.2 times higher than the arable crop market in agriculture (Platteau and Van Bogaert, 2014 ). The global demand for fertilizers amounts to an estimated 110 million tons (Mt) of N, 47.0 Mt P 2 O 5 and 37.5 Mt K 2 O per year (Agriculture Production & International Trade and Market Intelligence Services, 2019 ; Food and Agriculture Organization of the United Nations, 2019 ). The vast majority of these fertilizers is supplied in the form of inorganic, ‘synthetic’ nutrients. Organic fertilizers, which are typically bio‐based, containing recovered resources from agro‐industry (byproducts, residues, and side and waste streams), only made up an estimated 5% of the total fertilizer market value in 2019 (Statistics Market Research Consulting Pvt Ltd, 2018 ; Brandessence Market Research and Consulting Pvt, 2019 ; Mordor Intelligence, 2020 ). However, the compound annual growth rate (CAGR) of the organic fertilizer market is estimated at 14.3%, which considerably exceeds the CAGR of 3.8% for the overall fertilizer market (Statistics Market Research Consulting Pvt Ltd, 2018 ; Mordor Intelligence, 2020 ). Marketed organic fertilizers are typically solid and based on animal or plant materials (Sonneveld and Voogt, 2009 ), such as blood meal, cocoa shells, animal manures and soybean meal, amongst others (Wuang et al. , 2016 ). All organic fertilizers have in common that they release the nutrients gradually since the major part of the N, and P is bounded in complex molecules such as proteins and is released by decay or through decomposition by the microbial community associated with the soil or GM (Grunert et al. , 2016a , 2016b ). Organic fertilizers may offer appealing benefits for plant and soil health, and the environmental footprint of plant production. This slow‐release characteristic results in a decrease of N leaching losses. Furthermore, the organic fertilizers can increase the organic matter content of the soil, which improves the exchange capacity of nutrients, promotes soil aggregates, increases soil water retention and buffers the soil against acidity, alkalinity, salinity, pesticides and toxic heavy metals (Chen, 2006 ; El‐Haggar, 2007 ; Paungfoo‐Lonhienne et al. , 2012 ). Microbial fertilizers (MFs) constitute an emerging type of organic fertilizers composed of dried microbial biomass. As other organic fertilizers, MF is a direct source of plant macronutrients such as nitrogen, phosphorus, and potassium (N/P/K). The MF serves as a substrate to the microbiome present in the growing medium or soil, mineralizing the supplemented biomass, thereby rendering nutrients available for plant growth. This is not to be confused with a biofertilizer which contains active cells that can act as biocatalysts and contributes to the microbiome activity of the GM or the soil. A common application of these active cells is to render nutrients in the rhizosphere available to the plant roots, thereby providing an indirect fertilizer function. Other known application relate to bio‐stimulation, bio‐fortification and nutrient fixation (Sakarika et al. , 2020 ). Microbial biomass, whether for use as MF or as microbial protein (single‐cell protein), can be efficiently produced in bioreactors, converting up to 100% of the inorganic nutrients to biomass, thus minimizing/avoiding environmental emissions (Verstraete et al. , 2016 ). The microbial biomass can also capture and concentrate dissolved nutrients from dilute side streams, and as such assist in local nutrient loop closure, through rendering an additional recovered source of nutrients available (Doucha et al. , 2005 ; Lee et al. , 2015 ; Alloul et al. , 2019 ). Effluents from the food and beverage sector (i.e. potato processing, beer brewery) are not well explored as a source for bio‐based fertilizer production, even though they often contain substantial amounts of N, P and K (Falletti et al. , 2015 ). To comply with environmental regulations, nitrogen and phosphorus must be removed, avoiding eutrophication and oxygen depletion in the receiving water bodies. However, the removal process generally leads to the loss of these nutrients either to the air (i.e. N 2 and N 2 O) or to waste sludge (i.e. P precipitated with iron or aluminium; N and P in microbial biomass). Nutrient recovery and reuse is mostly limited to applying some of the resulting biosolids on farmland, mostly as a slurry, with low nutrient release efficiencies (Kanagachandran and Jayaratne, 2006 ; DC water, 2020 ). There is tremendous improvement potential for the recovery and reuse of valuable nutrients from agro‐industrial side streams to provide the much needed fertilizers. While recovery as refined mineral compounds (e.g. struvite, ammonium sulfate) is technologically ready for implementation (Grunert et al. , 2019 ), the currently associated costs may slow down market entry (De Vrieze et al. , 2019 ). In this study, three metabolic types of MF will be discussed: a consortium of aerobic bacteria (CAB), the microalga Arthrospira platensis (Spirulina) and the purple non‐sulfur bacterium (PNSB) Rhodobacter sp. Figure 1 presents the conceptual scheme of MF production and application. The microbial biomass is ideally produced in high‐rate systems (maximum production, high N content) at a low solid retention time (young biomass), to assimilate high nutrients levels, after which it is harvested and dried (down to around 10% moisture content). Spiller et al. ( 2019 ) provides a more detailed description of the processes and mass balances used to produce these three metabolic types of MF on potato wastewater. In terms of assessing fertilizer performance, N/P/K content, mineralization rate and fertilization effects are important parameters. As with other organic fertilizers, success relies on an active GM/soil microbiome (Grunert et al. , 2016a , 2016b ). Fig. 1 Conceptual scheme of the production of three metabolic types of microbial fertilizer (MF, dried microbial biomass powder), e.g. consortium of aerobic bacteria (CAB), microalgae (MA) and the purple non‐sulfur bacterium (PNSB). Subsequently, it is shown how MFs are applied in growing media or soils to produce plants. Despite a strong research interest in producing nutrient‐rich biomass, the literature on applying dried microbial biomass as fertilizer is scarce and lacks a systematic approach. Typically, one metabolic type of MF is investigated at non‐comparable nutrient dosages, often with only an inorganic rather than an organic fertilizer as control, or without positive fertilization control. Furthermore, storage behaviour, mineralization patterns and economic aspects are not addressed. For CAB, no fertilization tests are available on biomass with a low age, i.e. expected to have high nutrient content and slow release pattern. Kanagachandran and Jayaratne ( 2006 ) performed a study on sundried CAB grown at relatively high cell age on brewery wastewater, combined with compost. Germination and plant growth of pumpkin seeds and chili seeds was stimulated when these CAB were added to the compost. For Spirulina, promising results were demonstrated in leafy vegetables, compared to an inorganic fertilizer as control (Wuang et al. , 2016 ). This Spirulina biomass was grown on aquaculture wastewater, but no additional information was available on drying procedure. For PNSB, freeze‐dried Rhodopseudomonas spp. or Rhodobacter spp. gave positive results for the production of tomatoes (Kondo et al. , 2010 ), pakchoi (Wong et al. , 2014 ), spinach (kondo et al. , 2008 ) and rice (Kobayashi and Haque, 1971 ). These PNSB studies either used an inorganic fertilizer or no fertilizer as a control. The novelty in this research is the systematic comparison of novel types of MF (CAB, Spirulina and Rhodobacter sp.), in which a reference organic fertilizer (ROF) was used as a practical benchmark (positive control). Focus was on the key aspects of investigating the feasibility of this new value chain, ranging from nutrient levels, over shelf‐life of dried fertilizer and industrial storage after blending with growing medium, to fertilization properties on two crops and a cost estimation. Furthermore, to the authors’ knowledge, a number of these aspects have not been reported before for microbial fertilizers, including its use to produce herbs and ornamental plants, the compositional dynamics of stored products and an economic analysis.",
"discussion": "Discussion Microbial fertilizer production and quality Overall, the MFs showed comparable N and P quality with ROF (Table 1 ). The higher phosphorus content of Rhodobacter sp. can be explained by the phosphate accumulating capacities of PNSB (Hiraishi et al. , 1991 ; Lai et al. , 2017 ). In terms of drying conditions, as for other organic products subjected to heat processing (e.g. corn; Odjo et al. , 2015 ), a considerable degree of expertise may still need to be developed on this topic for cost‐effective microbial biomass drying in terms of heating rate and level, and their impact on the final quality of the product. Furthermore, the microbial community of CAB was not analyzed in this study, with the application in mind of using the powder product as a direct source of plant macronutrients (N, P, K). It is known that such communities are inherently dynamic and furthermore impacted by process conditions (Meerburg et al. , 2016 ; Langer et al. , 2019 ). Future research on fertilizers based on complex communities could benefit from understanding whether and which compositional changes may alter their fertilization properties. Alternatively, microbial biomass that is not dried and mostly viable, a so‐called biofertilizer, is known to perform indirect fertilization (e.g. biological N 2 fixation), or possess bioactive compounds to promote plant growth and health (Shen et al. , 2019 ; Sakarika et al. , 2020 ). It is currently unknown whether bioactive compounds remain present and active after drying, and this research line should be further explored to include all growth‐promoting effects. Fertilization performance Parsley growth showed an equal performance for all MFs (and their blends) when compared to ROF for plant height and fresh weight. These results suggest that the developed MFs have a nutrient release pattern close to the actual needs of the plant. However, CAB performed significantly better when the fresh weight of the parsley plants was applied as an evaluation parameter. The follow‐up experiment with the ornamental plant petunia (only performed for CAB) confirmed the better performance of CAB when compared to the ROF. CAB obtained the highest fresh weight (206 g) and a superior rooting score. The results for CCI and number of flowers showed similar trends for CAB and ROF. Mineralization behaviour The MFs showed similar nitrogen mineralization patterns in the GM compared to ROF, i.e. mineralization rate as well as the ammonium and nitrate content. However, an advantage of the use of the MFs compared to ROF is the slightly lower ammonium concentrations during mineralization (without and with plants). Too high ammonium concentrations might cause ammonium toxicity, resulting in stunted growth, necrotic leaves, inhibition of the primary root growth, and, in severe cases, even plant death (da Silva et al. , 2016 ). Additionally, it was shown that there is a high correlation between the EC and nitrogen mineralization efficiencies (Fig. 4D ), which is the sum of the ammonium and nitrate concentrations. The correlation coefficients between EC and nitrogen mineralization efficiency for ROF, CAB, Spirulina and Rhodobacter sp. were 0.96, 0.95, 0.96 and 0.98 respectively. Therefore, the EC could be used as a proxy for nitrogen mineralization. All fertilizers showed lower nitrification rates during mineralization in the packed GM without plants compared to during the parsley growth (Table 2 ). The major difference in nitrification rate between the mineralization without and with plant growth (5.5–8.6 higher) can be explained by the better accessibility for oxygen and water in the pots used for the plant trial compared to the 40‐l storage bags. The mineralization of the MFs in the packed GM (industrial storage) showed that more than 70% of the organic N is converted to inorganic species after 77 days. It is therefore recommended to store GM and MF separate and only mix these when needed to align the slow‐release characteristic with the plant’s needs. Therefore, a storage test of the individual MFs was performed to determine its shelf‐life. Storage The storage test showed slight hygroscopic activity (increase of 3–5% in moisture) for CAB and Rhodobacter sp. It is known that particle size distribution influences product properties, such as caking (i.e. clumping of biomass through the uptake of water). In a study on lactose powders (also an organic material), smaller particles showed higher moisture sorption and a greater caking tendency (Carpin et al. , 2017 ). The higher hygroscopicity of the CAB biomass might therefore be attributed to the smaller particle size of this end‐product (about 0.5 mm granular size). Spirulina and Rhodobacter sp. both had larger particle sizes. Some caking was noticeable in the sample bags of CAB over the course of the storage test. Additionally, the moisture content increase might be caused by the extra dry texture at the onset. Rhodobacter sp. had an initial dry weight of 95%, which is higher than necessary for the end‐product. To avoid moisture increases in the biomass, vacuum packaging for storing can be considered. Overall, the VS content and the TKN/TS content had not decreased by 10%, and thus, the shelf‐life of the biomass is at least 91 days. Cost estimations of microbial fertilizers A cost assessment of the MF (and their blends) was made in comparison with the commercially available ROF. CAB proved to be a cost‐effective alternative for ROF, with only an increase in cost of 6%. The blends with phototrophic bacteria (i.e. Spirulina and Rhodobacter sp.) resulted in increases between 7 and 20%. However, phototrophic microorganisms (i.e. Spirulina and Rhodobacter sp.) are known to contain plant growth promoting substances such as phytohormones, vitamins, carotenoids and antifungal substances (Serdyuk et al. , 1993 ; Spolaore et al. , 2006 ). The activity and content of bioactive compounds that remains active after drying is unknown, but could have an added‐value in the MF. Future research should focus on the effect of these bioactive compounds in the MF, since this could justify the higher cost of the blends."
} | 4,463 |
19537980 | PMC2990325 | pmc | 8,908 | {
"abstract": "The colony is the functional unit of natural selection for most social insects including the Florida harvester ant, Pogonomyrmex badius . To address reproduction in the species variables were evaluated relevant to the colony-level (sociometry), and social growth (sociogenesis). Colonies become reproductively mature when the worker population reaches ~700 individuals. The production of males and gynes (reproductive females) occurs only in spring, and is highly synchronized from its onset, which in turn allows synchronization of mating in early summer. Worker production follows sexual production, and continues until colonies go dormant in winter. Once mature, colony investment into reproduction is a constant proportion of colony size (isometric), regardless of the sex ratio produced. As individual male body size increases, they become leaner, whereas the amount of fat stored by gynes is highly variable. Larger colonies produce larger males, but gyne size is a constant across the range of colony sizes. As colony size increases, investment into males increases faster than investment into gynes. Therefore, there is a trend of increasing male bias in the sex ratio with increasing colony size (although a single outlier complicates this conclusion). We interpret our results in the light of sexual and natural selection as documented in related species. We also report the first documentation of male production by workers in the genus Pogonomyrmex .",
"introduction": "Introduction In eusocial organisms, selection acts on the collective unit (colony), rather than individuals ( Bourke and Franks 1995 ; Korb and Heinze 2004 ). Ultimately, natural selection favors strategies that maximize the number of new colonies successfully established. The manner in which new colonies are founded and grow differ tremendously in the social insects, and ants run the gamut from independent (a solitary foundress) to completely dependent (i.e., budding) colony founding ( Holldobler and Wilson 1990 ). Harvester ants of the genus Pogonomyrmex have a well-described mating system that is analogous to broadcast spawning. Colonies send out males and gynes (reproductive females) in highly synchronized nuptial flights ( Strandtman 1942 ; Michener 1948 , 1960 ; Van Pelt 1953 ; Nagel and Rettenmeyer 1973 ; Clark and Hainline 1975 ; Holldobler 1976 ; Davidson 1982 ; Mintzer 1982 ; Rust 1988 ; Harmon 1993 ; Cole and Wiernasz 1997 ). The sexual individuals, once released from the parent colony, increase their chance of successful fertilization by aggregating at leks (see Holldobler 1976 ). Once mated, the queen founds a new colony (different species may found colonies solo or cooperatively: Johnson 2004 ). Parent colonies can better their odds of producing a successful founding queen or the male(s) that inseminated her using a variety of strategies, at a variety of levels. For example, by modifying the size or fat reserves of individual males and gynes, a parent colony can produce higher quality reproductives that are more likely to succeed at colony founding (examples from Pogonomyrmex occidentalis : Abell et al. 1999 ; Wiernasz et al. 2001 ; Wiernasz and Cole 2003 ). At the colony level they can alter the sex ratio and total numbers of individuals produced (see review by Mehdiabadi et al. 2003 ), as well as changing reproductive timing, and the duration of reproduction (i.e., the number of mating flights in which they participate). If colonies are able to reproduce in multiple years, they may be able to augment lifetime reproductive output by increasing the size of their worker population because larger colonies have higher reproductive potential in Pogonomyrmex ( Mackay 1981 ; Cole and Wiernasz 2000 ; Tschinkel 1993 ). Therefore, to obtain maximal fitness, colonies must balance investment into growth and reproduction, and then further modify their investment in reproduction in a way that maximizes their genes in the next generation ( Stearns 1992 ; Bourke and Franks 1995 ). In this paper the Florida harvester ant ( Pogonomyrmex badius ), was used to explore the ontogeny (sociogenesis) of reproduction at the colony level, and the strategies that may maximize fitness across colony sizes. Our knowledge of this species’ seasonal variation and sociogenesis ( Tschinkel 1998 , 1999a , Tschinkel b ) allow us to investigate in detail how selection has shaped colony investment in growth and reproduction. The genus, Pogonomyrmex , is well studied and ecologically important ( Taber 1998 ; Johnson 2001 ), but colonies are often inaccessible because they typically nest in rocky soils, limiting our knowledge of the colony-level processes underlying ecological and evolutionary questions. Because P. badius lives in deep sandy soils, we are able to excavate whole colonies to examine how they invest in both reproduction and growth. This paper addresses the following questions about allocation to reproduction: 1) what is the timing of growth and reproduction during the annual and lifetime cycle? 2) what are the patterns of resource investment in reproductive individuals, at both the individual and colony-level? and 3) how does colony ontogeny affect sexual allocation? Site description, colony excavation, and measurements All colonies were collected in the Apalachicola National Forest, about 15 km southwest of Tallahassee, Florida. Nineteen colonies were excavated in an area that was clear-cut in 1999 and planted with longleaf pine, few of which were taller than 1 m at the time of this study. Ground cover vegetation was shrubby and herbaceous, predominantly shiny blueberry ( Vaccinium myrsinites ). Due to the clear-cut and the lack of P. badius in the surrounding forest, it is likely that none of the study colonies was greater than 4 years old. Mature, reproductive, colonies were selected by mound size, a good proxy for colony size ( Smith 2004 ). Reproductive colonies were of two experimental treatment groups, with their charcoal covering removed or removed and replaced (n = 9 in each group). The experiment was designed to test whether fitness was affected by charcoal on top of the mound (large nests can have 1/3kg of charcoal, ~300,000 pieces). Harvester ants are notorious for collecting objects, including charcoal, which are deposited on top of their nest Smith and Tschinkel 2005 ). This experiment addressed the adaptive potential (affect on reproductive fitness) of objects on top of the nest. Because no effects of charcoal were detected ( Smith 2004 ), data were pooled across the treatments for the analysis presented here (17 of 19 excavated colonies produced sexual offspring). Colonies were excavated with a shovel as in Tschinkel (1998) , where an initial hole (pit) was dug adjacent to the colony, and chambers were exposed sequentially from the top down, working from the pit. For completeness, soil was excavated either 25cm deeper than ants were found or to the water table. Ants were brought back to the lab, frozen, sorted and counted. All larvae were thought to be workers, based on their size when excavated. In their final instars sexual larvae are much larger than those of workers. After counting them, ants were dried at 60°C in an oven for at least 48 hrs, and weighed to the nearest milligram. A sample of 10 males and 10 gynes were individually weighed on an electronic microbalance (to the nearest mg for gynes, and tenth of a mg for males), and their head widths estimated by measuring the inter-ocular distance. Fat was extracted using ethyl ether as the solvent in a Soxelet extractor for at least 24 hours. The energetic content of individuals was calculated based on the conversion values of Peakin (1972) for fat and lean mass (fat = 39.33 J/mg; lean = 18.87 J/mg). Reproductive schedule The relationship between the collection date of a colony and the proportion of pupae and callows in the nest was examined to evaluate reproductive synchrony across colonies. If this relationship is linear then colonies are synchronized in their production schedules, all having started producing at about the same time. To estimate the time at which colonies began production the pupation rate was extrapolated to 100% pupae, and egg, larval and pupal times estimated from the literature. This information is not known for P. badius , so data was used from other species ( Porter 1988 ). To verify the estimate of reproductive onset three colonies were excavated, noting the presence or absence of brood in each. The first was on 11-March, 2004 (C7), the second on 25-March (C25), and the third on 15-April (C56). Both March colonies were returned to the field after assessment, while C56 was retained as a lab colony because the queen was accidentally decapitated during excavation. This colony was used to assess the possibility of worker reproduction in the absence of their queen. Workers were divided into 2 sub-colonies of approximately equal size and casually observed for 3 months. Individual-level size and allometries To assess how total mass and fat content changed relative to individual size (approximated by head width) individuals were pooled from all colonies. To confirm the assumption that dry weight is a good approximation of the energetic cost of manufacturing (excluding maintenance cost – respiration – of individuals) dry weight and energetic content of both workers and reproductives was regressed. Boomsma (1989 ; Boomsma et al. 1995 ) demonstrated that using manufacturing cost overestimates the energetic cost of females in species with a sexual size dimorphism (due to respiration costs), and proposed correcting mature adult dry weights using a power conversion of 0.7. Comparisons of estimates of sex ratio (expressed as the proportion female) were made using the manufacturing cost and the transformed adult dry weight estimate (x 0.7 ) using a t-test. Whether gyne fat content was affected by the date of collection (since adult gynes accumulate large fat reserves prior to flights) was examined by regressing dark (mature) gyne proportion fat against the collection date. A positive slope would indicate a sampling bias. Colony-level allometries Colony size, the number of mature workers in the nest, was used as the independent variable for all allometric analyses. The dependent variables are estimates of sexual production, represented by their weight or the number of individuals. Weights are sensitive to when colonies, especially those sampled in spring, are excavated because they have differing proportions of ants in each developmental state, and adults weigh more than pupae. Therefore, the end-weight of workers and sexuals was estimated to account for sampling effects. Numerical sexual production was the sum of sexual pupae and adults of both sexes, and by weight was the number of males multiplied by the average male weight from that colony, plus the number of gynes multiplied by the average gyne weight across colonies (some colonies did not have mature gynes). Statistical methods and calculations Variables were transformed to satisfy assumptions of statistical tests, using a log 10 transformation on most variables, but arcsine transformation on proportions. Allometries were done using Type-I regression because it is a standard method employed for such analyses, though we recognize that Type-II regression is the more appropriate statistical tool. Isometry (slope different than 1) was tested using a t-test. Individual allometries used data pooled from all colonies, whereas colony allometries used the average or total per colony. All allometries of gynes are of dark individuals (presumed mature). All analyses were done in Statistica 6.0 ( Statsoft 2003 ).",
"discussion": "Discussion Production schedule Production in this species is highly synchronized among colonies, as would be expected, since nuptial flights (i.e., mating) are somewhat unpredictable events, usually following the first heavy summer rain. Furthermore, successful flights require the participation of multiple colonies, or colonies risk sib-mating (which occurs at high levels in P. occidentalis , Cole and Wiernasz 1997 ). Therefore, all colonies have some of both sexes ready to fly when the first rain comes, but continue to rear more for subsequent flights. This reproductive synchrony is likely achieved through a cue that initiates production at the beginning of spring. Knowing the production schedule of this species facilitates experimental work by providing an estimate of what is in the colony at any given time, and when colonies switch from their annual reproductive- to worker-producing phase. Observations made on a single queenless colony suggest that multiple workers within a nest will become reproductive when the queen is removed. Although uninseminated, the workers can still produce males and potentially gain in fitness. This is relevant to sexual allocation because it cannot be assumed that all male production is from the queen. When a queen is removed from a nest the ovaries of workers become more active (have more developed oocytes - Smith, unpublished data), making it likely that the queen inhibits male production by workers. To our knowledge, this is the first report of male production by workers in the genus Pogonomyrmex . Individual-level size and allometries There was a pronounced sexual dimorphism in this species, with females much larger and more expensive than males. The energetic cost of manufacturing individuals (excluding respiration costs) is well predicted by dry weight, justifying the use of dry weight as a proxy for energetic investment (colony-level allometries). Furthermore, accounting for respiration costs as suggested by Boomsma (1989) did not significantly change the ratio of investment between the sexes, and both are not different than 1:1. Therefore, results were interpreted using only dry weights. There are pronounced allometric relationships of male characters. As males increase in size they increase proportionately in weight, but become leaner. P. badius queens are known to mate with many males ( Rheindt et al., 2004 ) in the mating swarms (leks), so that male size may be very beneficial in obtaining a mate ( Davidson 1982 ; Abell et al. 1999 ; Wiernasz et al. 2001 ). As males increase in size they are not putting on extra fat reserves, but simply increasing in lean mass. Increased lean mass may be advantageous in male-male competition, or perhaps dispersal. Male size is often correlated with increased sperm transfer (e.g., Wiernasz et al. 2001 ), which is in turn the best paternity estimate in the honeybee, another highly polyandrous hymenopteran ( Schlüns et al. 2004 ). Therefore, it would seem likely that larger males are able to obtain more fitness for the colony. Gynes do not increase in weight or fat content as a function of size, and there is less variation in gyne size. Natural selection may work to decrease variance in gyne size by 1) putting a minimum on size and fat stores for successful colony founding (e.g., Wiernasz and Cole 2003 ), and 2) putting a maximum on size due to their high manufacturing cost, possible diminishing returns of fitness, or a physiological size limit. Colony-level allometries Interpretations at the colony-level demonstrate that patterns of investment into the sexes change as a function of colony growth. Individual male size and fat content increased more slowly than colony size, but larger colonies did produce larger males ( Table 3 ). Furthermore, as colonies grew they invested disproportionately more into males, increasing male production 20-fold for every 10-fold increase in colony size ( Table 4 , Figure 4 ). On the other hand, gyne size and fat content were independent of colony size. And though the number of gynes produced by colonies was a constant function, total investment into gynes (by weight) was independent of colony size. This pattern suggests that there is a fitness benefit to producing more and larger males. Sexual selection on male body size ( Abell et al. 1999 ), as well as assortative mating ( Davidson 1982 ), are known from other Pogonomyrmex species. This result is not too surprising given the lek mating system of most species (reviewed in Holldobler 1976 ), and although the mating biology of P. badius is not well understood, mating aggregations at least occur at nests ( Harmon 1993 ; and personal observations). The interpretation of the sex ratio changes with the inclusion of a single colony, C22. If C22 is included, the average sex ratio (proportion female, dry weight) is approximately 1:1, as expected from the genetic relatedness hypothesis ( Trivers and Hare 1976 ; Nonacs 1986 ). According to this hypothesis, as the queen mates with more and more males, the sex ratio converges on a 1:1 sex investment. As colony size increased, the variance in the sex ratio decreased, until converging at ~50% female. If C22 is excluded, there was a significant negative slope to the relationship. If this result is not an anomaly, two hypotheses are suggested that could explain this result: 1) given the genetic relatedness hypothesis, worker relatedness within the colony may decrease with colony growth due to an increase in the number of patrilines among the workers, resulting in decreased female investment, or 2) the fitness gains of males and females are different, with male reproductive value increasing faster than that of females, in both quality and quantity ( Frank 1987 ). Pogonomyrmex badius are only known to found colonies haplometrotically (by a solitary foundress), and mature colonies always have a single queen (personal observations of both authors). Therefore, if hypothesis 1 were true, sperm precedence would be the likely mechanism for changes in relatedness through time, and not multiple queens. Queens of P. badius mate with an average of 9–10 males, and up to a maximum of 29 ( Rheindt et al. 2004 ), but no information is available regarding sperm precedence in harvester ants. Little information exists on sperm precedence in most social Hymenoptera, and has been best studied in honeybees, Apis mellifera . Schluns et al. (2004) report a marginal effect (P = 0.08) of insemination sequence in predicting patrilines of honeybee workers born from artificially inseminated queens. In the same study, most variation in patriline representation among workers was explained by the volume of inseminate, not order of insemination. On the other hand, various theories make predictions about differential fitness returns between and among the sexes, which would support hypothesis 2. Frank (1987) reviews several theories to predict non-linear fitness returns within a sex. Local resource enhancement predicts a synergism between related individuals of the same sex, as in cooperatively founding queens, which is not relevant in P. badius . Local resource competition and local mate competition both predict a diminishing return in colony fitness due to antagonisms between related individuals of the same sex, due either to resource (e.g., access to nest sites or food for gynes) or mate competition, which is more likely in males. The results of this study, where male investment increases much more rapidly than gyne investment, are not consistent with any of these scenarios unless related males act synergistically (i.e., local resource enhancement, where resources are presumably mates), which would seem unlikely. Multifaceted parental investment (MFPI)( Rosenheim et al. 1996 ) makes predictions for the variation of investment within sexes, where variation in individual size is due to a trade-off between resource limitations (e.g., brood/egg production or energetic limitation) and optimal offspring size. In the context of MFPI, male fitness returns are highly variable and vary with male size, where larger males presumably yield greater fitness for the colony, but gyne fitness is constant and likely constrained (see above discussion of individual gyne size). If male fitness returns are highly variable, even when producing large males, colonies with few resources may be expected to invest proportionally more resources in gynes, such that they minimize the variance in expected fitness returns. Sex ratios are highly variable both spatially, within and between sites, and temporally, between seasons, ( Elmes and Wardlaw 1982 ; Elmes 1987 ; Herbers and Banschbach 1998 ). Furthermore, hypotheses that attempt to explain sexual allocation in the hymenoptera are numerous, and may or may not differ regarding their predictions (see review by Mehdiabadi et al. 2003 ). The results and theoretical considerations of others suggest that it is best to remain skeptical about the robustness of this pattern, and collect more data in order to test particular hypotheses about sexual allocation in this species. This same skepticism applies to our entire analysis of reproduction because it draws on information from only a single year and site. Our argument for the validity and generality of the information presented is simple: the robustness of most patterns we present at least indicates our ability to understand production at the sites where it is described, and if we assume that patterns of colony development are heritable, our analysis is not likely to stray far from the general trends about this species and even genus. In summary, colonies begin their reproductive phase of life when they reach a mature worker population of ~700. In early spring (the end of March in this population), the colony begins production of males and gynes, and switches to worker production approximately a month later. Worker production continues until the beginning of winter (December in this population) when the colony becomes reproductively inactive (December to March). As colonies increase in size they are capable of producing more reproductives, and workers. Although the production of both males and gynes increase with colony size, male production increases twice as fast. This asymmetry yields a changing sex ratio as colonies grow, whether variance decreases or the slope is negative. Bigger colonies also produce bigger males. This likely increases the colony’s probability of fitness by gaining a numerical advantage over neighboring colonies, and also by producing better males (if size does indeed matter). This is not, however, the case with gyne production. All colonies produced approximately the same size of gyne, and this size is possibly the result of selective pressures to both decrease maximum size and maintain a minimum size. The fitness of a colony is the number of new colonies in the next generation, summed over all of the breeding seasons in which the colony participated. Selection favors colony growth because growth creates the largest increase of reproductive output. Colony fitness returns are possibly non-linear with size due to larger colonies’ ability to produce not only more males and females, but also larger males. Alternately, it is possible that fitness returns are constant, but the optimal investment strategy differs as a function of the amount of energy a colony has available to invest. Differing investment strategies may be represented by the sizes of males and gynes, and the total number produced of each."
} | 5,827 |
34065964 | PMC8151836 | pmc | 8,909 | {
"abstract": "This study explicated the functional activities of microorganisms and their interrelationships under four previously reported iron reducing conditions to identify critical factors that governed the performance of these novel iron-dosed anaerobic biological wastewater treatment processes. Various iron-reducing bacteria (FeRB) and sulfate reducing bacteria (SRB) were identified as the predominant species that concurrently facilitated organics oxidation and the main contributors to removal of organics. The high organic contents of wastewater provided sufficient electron donors for active growth of both FeRB and SRB. In addition to the organic content, Fe (III) and sulfate concentrations (expressed by Fe/S ratio) were found to play a significant role in regulating the microbial abundance and functional activities. Various fermentative bacteria contributed to this FeRB-SRB synergy by fermenting larger organic compounds to smaller compounds, which were subsequently used by FeRB and SRB. Feammox (ferric reduction coupled to ammonium oxidation) bacterium was identified in the bioreactor fed with wastewater containing ammonium. Organic substrate level was a critical factor that regulated the competitive relationship between heterotrophic FeRB and Feammox bacteria. There were evidences that suggested a synergistic relationship between FeRB and nitrogen-fixing bacteria (NFB), where ferric iron and organics concentrations both promoted microbial activities of FeRB and NFB. A concept model was developed to illustrate the identified functional interrelationships and their governing factors for further development of the iron-based wastewater treatment systems.",
"conclusion": "6. Conclusions The functional interrelationships presented in this study provide insights to the various synergistic and competitive relationships in wastewater treatment under iron-reducing conditions. They are crucial for further development and design of the iron-based biological technology for optimum treatment performance. The factors governing these relationships need to be considered systematically to develop guidelines for operating the treatment process. Moreover, the interrelationships reveal that there are great opportunities to develop iron-based treatment not only for wastewater management, but also for enhanced nutrient (e.g., ammonium) and biogas production.",
"introduction": "1. Introduction Anaerobic biological treatment of wastewater has been gaining increasing attention due to its simplicity, energy efficiency, and lower sludge production, greenhouse gases emission, and capital and operational costs compared to aerobic treatment processes [ 1 , 2 , 3 , 4 ]. Using an anaerobic process instead of an aerobic process can reduce operating costs by approximately $160 per metric ton, and as high as $250 for some instances [ 5 ]. Carbon dioxide (CO 2 ), sulfate (SO 4 2− ), and nitrate (NO 3 − ) are commonly used electron acceptors in anaerobic biological processes of wastewater treatment [ 6 , 7 , 8 ]. Motivated by the benefits of comanaging acid mine drainage (AMD) and municipal wastewater (MWW), cotreatment of both wastes in natural and engineering systems has previously been evaluated and showed impressive results of removing heavy metals and organic matter [ 9 , 10 , 11 ]. These studies have led to further development of innovative iron-dosed treatment processes [ 12 , 13 , 14 ]. Iron-based anaerobic treatment has multiple energy and environmental benefits including no aeration requirement, potential use of iron containing wastes, design and operation simplicity, low sludge production and CO 2 emission, and potential resource recovery from the sludge materials [ 15 ]. As a key microbial reaction in the iron-based anaerobic biological treatment, ferric reduction is coupled to organics oxidation, in which Fe (III) is reduced to Fe (II) by receiving an electron from the organics (i.e., electron donor). As Fe(III)/Fe(II) reduction potential is comparatively higher (+0.77 V at pH 2 and +0.2 V at pH 7) than other electron acceptors (e.g., sulfate, CO 2 ), iron-reducing bacteria (FeRB) can use this energy to respire a wide range of organic compounds [ 16 ]. Geobacter and Shewanella are known FeRB which were observed in most of the research on microbial iron reduction [ 17 , 18 , 19 , 20 , 21 ]. These two iron reducers have diverse ways of interacting with the ferric mineral surfaces for ferric reduction. Geobacter sp. is a strict anaerobe and mostly rely on pili (protein nanowires) as it does not secrete enough electron shuttling or chelating compounds [ 17 , 21 ]. Shewanella sp. has both direct and indirect electron transfer mechanisms including electron shuttles, ligands and pilin filaments. Organic composition also governs the type of FeRB present in a particular environment. For example, Geobacter sp. generally uses acetate and completely oxidizes it to CO 2 while Shewanella sp. uses lactate as a carbon source and oxidize it to acetate [ 17 , 18 ]. Under substrate limiting conditions such as those found in natural environments (e.g., soil, sediments, groundwater), FeRB can outcompete sulfate reducing bacteria (SRB) for organics by diverting electron flows away from SRB [ 22 , 23 , 24 ]. In wastewater treatment applications where organic matter is abundant, both FeRB and SRB can perform carbon oxidation and concurrently contribute to the removal of organics. SRB such as Desulfovibrio sp. and Desulfobulbus sp. have been reported to facilitate incomplete oxidation of larger organic substrates (e.g., lactate) to smaller organic substrates (e.g., acetate) which could subsequently be used by FeRB [ 25 , 26 ]. Such symbiotic and/or competitive dynamics between FeRB and SRB are regulated by the availability of organic substrate and electron acceptors (e.g., ferric, sulfate), and associated environmental conditions such as pH and bioavailability of the electron acceptors [ 13 ]. Ferric reduction coupled to ammonium oxidation (Feammox) is another microbial metabolic function that could be used for wastewater treatment. Most Feammox studies have been conducted in natural environments such as groundwater, soils and sediments [ 27 , 28 , 29 , 30 , 31 ] and studies related to wastewater environment are extremely limited [ 32 , 33 ]. In strict anoxic conditions, ferric reduction has been found coupled to ammonium (NH 4 + ) oxidation to produce either nitrogen (N 2 ) (Equation (1)), nitrite (NO 2 − ) (Equation (2)), or nitrate (NO 3 − ) (Equation (3)) [ 27 , 29 , 34 , 35 , 36 ]. Feammox to N 2 is energetically more favorable than Feammox to NO 2 − and NO 3 − under a wide range of conditions [ 29 ]. Huang and Jaffé [ 35 , 36 ] studied Feammox reaction in riparian wetland soils and identified Acidimicrobiaceae bacterium A6 as the predominant bacterial species responsible for Feammox reaction.\n (1) 3 Fe ( OH ) 3 + 5 H + + NH 4 + → 3 Fe ( II ) + 9 H 2 O + 0.5 N 2 \n (2) 6 Fe ( OH ) 3 + 10 H + + NH 4 + → 6 Fe ( II ) + 16 H 2 O + NO 2 − \n (3) 8 Fe ( OH ) 3 + 14 H + + NH 4 + → 8 Fe ( II ) + 21 H 2 O + NO 3 − Similar to the relationships with SRB, FeRB can potentially have symbiotic or competitive relationships with Feammox bacteria. According to redox potentials, organic carbon is a preferred electron donor compared to NH 4 + and, as a result, heterotrophic FeRB can outcompete autotrophic Feammox bacteria for Fe (III) compounds. A previous study showed that only 2% of Fe(III) reduction was observed to be associated with Feammox reaction in a paddy soil when sufficient organic substrates were present [ 30 ]. Some studies reported that FeRB such as Geobacter can play an essential role in Feammox activities [ 27 , 37 ]. An indirect relationship was established between FeRB abundance and Feammox rate in these studies, as with increasing FeRB abundance Feammox reaction rate also increased. The diverse physiological characteristics of FeRB were hypothesized as the probable reason behind their contribution to Feammox activity. In treatment of organics-rich wastewater, the functional relationships between FeRB and Feammox bacteria are expected to be significantly different and the conditions in which ammonium oxidation occurs are currently not known. Moreover, in such anaerobic/anoxic environments, fermentative bacteria, nitrogen fixing bacteria (NFB), and Anammox bacteria may also be present and their functional interrelationships are largely unknown. The objectives of this study were (i) to explicate the functional activities of various microorganisms and their interrelationships under previously reported iron dosing conditions used for wastewater treatment, (ii) to characterize the microbial diversity, abundance, and functions and to conduct comparative analyses among the different conditions, and (iii) to develop a conceptual model to illustrate the functional interrelationships of identified bacterial species and the factors that governed the microbial functions. Implications of the learned microbial diversity, abundance, metabolic functions and their interrelationships on engineering applications of the iron-dosed wastewater treatment method were discussed.",
"discussion": "3. Results and Discussion 3.1. Microbial Diversity Various phyla were identified in bioreactors R1, R2, and R3, which depict the diverse microbial compositions in all the settings ( Figure 1 ). Deltaproteobacteria, Alphaproteobacteria, Acidobacteria, Chloroflexi, Firmicutes, Bacteroidetes, and Actinobacteria are the common phyla that were observed in all the bioreactors. The bioreactor with Fe (III) iron dosing (R3) had higher microbial diversity (ten phyla) than R1 (eight phyla) and R2 (seven phyla). This was reflected in estimated diversity index Shannon’s H which ranged from 3.26 to 3.34 for the Fe (III)-dosed bioreactor (R3) and from 1.24 to 1.68 in the cotreatment bioreactor (R1). The higher diversity in R3 can be attributed to the high ferric dosing and prevalence of FeRB whereas R1 and R2 were mostly sulfidogenic. R1 was used to treat AMD/MWW mixtures that had high sulfate and low ferric concentrations after the first stage treatment. R2 was dosed with ferrous iron and had only limited ferric iron from the recycled oxidized sludge. We did not use microbial data of R4 reactor for diversity analysis, as this reactor was designed to investigate Feammox activities. 3.2. Iron-Reducing Bacteria No FeRB were characterized for R1 as that was not the scope of that study. The only FeRB observed in the Fe (II)-dosed bioreactor (R2) was Alkaliphilus metalliredigens. This species is an alkaliphilic bacterium that uses lactate, acetate, and hydrogen as electron donors for Fe (III) reduction [ 43 ]. The synthetic wastewater used in the study contained primarily acetate and lactose, and had an alkalinity of 1.68 mM, which was conducive to the prevalence of this bacteria. This strictly anaerobic bacteria from the Firmicutes phylum has the capability to thrive under extreme alkaliphilic and salinity conditions [ 44 ]. Major putative FeRB observed in the Fe (III)-dosed bioreactor (R3) were Geobacter sp., Geothrix sp., and Ignavibacteria sp. Among the three FeRB, Geobacter sp. was predominant in abundance (83%) and others included Geothrix sp. (2%), and Ignavibacteria sp. (15%). Geobacter is heterotrophic, gram-negative, non-spore-forming, curved rod-shaped bacteria belonging to the Geobacteraceae family in the Deltaproteobacteria phylum [ 45 , 46 , 47 , 48 , 49 ]. This bacterium maintains an obligately anaerobic lifestyle, and typically performs complete oxidation of small organic substrates such as acetate to CO 2 via ferric reduction. The dominance of Geobacter sp. in the Fe (III)-dosed bioreactor (R3) is attributed to the acetate (approximately 250 mg/L) as one of the main organic compounds of the synthetic wastewater. Acetate is one of the prime volatile fatty acids (VFAs) present in the real wastewater, which comprises approximately 49% to 71% of the total influent VFAs in full-scale wastewater treatment plants [ 50 , 51 ]. There are also evidences that Geobacter sp. can use lactate and ethanol via Fe(III) reduction [ 45 , 49 ]. Similar to Geobacter, Ignavibacteria has also been observed to grow well in acetate amended incubations [ 52 ]. This strictly anaerobic, moderately thermophilic, neutrophilic and obligately heterotrophic bacterium has recently been isolated from several hot springs under iron-reducing conditions [ 53 , 54 ]. Genome analysis of Ignavibacteria revealed it as a versatile bacterium which has the capability to live under both oxic and anoxic conditions by using a variety of electron donors and acceptors [ 55 ]. With the complex composition of real wastewater containing different types of electron donors and acceptors, presence and growth of Ignavibacteria can be anticipated. As ferric compounds are typically insoluble in the bioreactor at circumneutral pH, Geobacter and Ignavibacteria can facilitate the ferric reduction either by direct contact with outer-membrane cytochromes or via conductive pili structures [ 17 ]. For Geobacter sp., direct electron transfer to Fe(III) mostly occurred at the outer cell surface through c-type cytochromes [ 47 , 56 , 57 ]. Among these outer membrane (OM) cytochromes, only four of the cytochromes (OmcB, OmcS, OmcE, OmcZ) were identified to play a role in Fe (III) reduction. Another means of electron transfer for Geobacter sp. is to utilize Type IV pilin filaments, which are also known as ‘bacterial nanowires’ or ‘protein nanowires’ [ 58 ]. These filaments are composed of multiple copies of PilA proteins. Due to the high electrical conductivity of Geobacter pili, Geobacter sp. was observed to generate the highest electrical current density among exoelectrogenic bacteria [ 59 ]. This bacterial species has the potential to be used in bioelectrochemical systems for electricity generation from wastewater and/or sewage sludge to enhance energy efficiency of the iron-dosed treatment method. Geothrix sp. is phylogenetically different than Geobacter sp., but has several physiological similarities with members of the Geobacteraceae [ 48 ]. In addition to Fe (III), Geothrix sp. can utilize other electron acceptors such as Mn (IV), nitrate, fumarate, and disulfonate for redox reactions, which is also a common trait observed in the Geobacteraceae family. However, the electron transfer mechanism of Geothrix is different from Geobacter and Ignavibacteria. Geothrix sp. has the ability to facilitate iron reduction without direct contact with the insoluble Fe(III) compounds by releasing compounds that act as electron shuttles and solubilize Fe(III) from Fe(III) oxides [ 56 ]. In addition to chemical characteristics, other environmental factors may affect the growth of FeRB and SRB. Table 1 summarizes the potential growth conditions of pH and temperature for FeRB and SRB previously reported in the literature. In particular, bacteria such as Geobacter sp. and Alkaliphilus metalliredigens can grow in a broader range of temperature, making them more resilient to temperature variations than other species and adaptable for broader waste treatment applications. 3.3. Sulfate-Reducing Bacteria Putative SRB observed in the cotreatment bioreactor (R1) were Desulfovibrio sp., Desulfovirga sp., Desulfobulbus sp. and Desulfatibacillum sp.; in Fe (II)-dosed bioreactor (R2) was Desulfomonile tiedjei ; and in Fe (III)-dosed bioreactor (R3) were Desulfovibrio sp., Desulfobulbus sp., Desulfatirhabdium sp., Desulforhabdus sp. and Desulfomonile sp. The microbial analysis of R3 bioreactor revealed that the major SRB was Desulfovibrio sp. with an abundance of 38% among the total SRB ( Figure 2 ). Other SRB such as Desulfobulbus sp., Desulfatirhabdium sp., Desulforhabdus sp., Desulfomonile sp. were present in the bioreactor with abundances of 30%, 21%, 8% and 2% of the total SRB, respectively. All these SRB belong to the Deltaproteobacteria phylum and use sulfate as an electron acceptor for redox reactions. Desulfovibrio sp., Desulfobulbus sp., Desulfovirga sp., and Desulfomonile sp. can facilitate incomplete oxidation of large organic compounds (e.g., lactate), and Desulfatirhabdium sp., Desulforhabdus sp. can oxidize smaller organic substrate such as acetate and ethanol [ 25 , 26 , 60 , 61 , 62 ]. Co-existence of these bacteria suggests a synergistic relationship among these diverse SRB where Desulfovibrio sp., Desulfobulbus sp., and Desulfovirga sp. yield smaller substrates such as acetate through lactate oxidation, that can subsequently be used by other FeRB and SRB for complete oxidation of organic substrates. As wastewater is a complex mixture of various organic compounds, a diverse composition of different SRB is anticipated in iron-reducing treatment systems. The suitable temperature ranges for the growth of SRB ( Table 1 ) indicate that most of the SRB can survive at the temperatures commonly found in wastewater treatment. Similarly, the pH conducive to the growth of these bacteria overlap the pH range typically observed with wastewater effluents (6.5–8.5) [ 63 ]. microorganisms-09-01039-t001_Table 1 Table 1 Potential growth conditions pH and temperature for different FeRB and SRB observed in the bioreactors. \n Bacteria Temperature pH Reference Iron Reducing Bacteria Geobacter sp. 4–37 °C 6.5–7.5 [ 64 ] Ignavibacteria sp. 30–55 °C 6.5–8.0 [ 53 ] Geothrix sp. 35–40 °C \n [ 48 ] \n Alkaliphilus metalliredigens \n 4–45 °C 7.5–11.0 [ 65 ] Sulfate Reducing Bacteria Desulfovibrio sp. 15–45 °C 5.0–8.0 [ 66 , 67 ] Desulfobulbus sp. 10–40 °C 6.1–7.5 [ 68 , 69 ] Desulfovirga sp. 20–36 °C 6.6–7.4 [ 62 ] Desulfatirhabdium sp. 15–37 °C 6.5–8.0 [ 60 ] Desulforhabdus sp. 25–45 °C 6.6–8.5 [ 70 ] Desulfomonile sp. 30–38 °C 6.5–7.8 [ 71 ] Desulfatibacillum sp. 15–40 °C 6.6–7.8 [ 72 ] 3.4. Synergistic Relationships between FeRB and SRB A critical research question regarding the iron-based wastewater treatment is whether FeRB and SRB can perform synergistically to oxidize organics. Previous studies suggested that FeRB could inhibit SRB by competing for electron donors when the organic level is low [ 24 , 73 , 74 ]. In wastewater treatment applications, high organic content of wastewater can provide sufficient organic substrates and support the growth of both FeRB and SRB. The microbiological analyses of R3 showed the presence of diverse FeRB and SRB, and chemical analyses also corroborated that ferric and sulfate reduction contributed concurrently to organic oxidation [ 12 ]. Fe (III) and SO 4 2− concentrations (expressed by Fe/S ratio) played a significant role in regulating the activities of FeRB and SRB in the Fe (III)-dosed treatment. The overall organic oxidation rate is dependent on the individual oxidation rates of FeRB and SRB and their populations. The average abundances of putative Geobacter sp. and Ignavibacteria sp. at different Fe/S ratios (molar ratios: 0.5, 1, and 2) were 22 ± 9%, and 4 ± 2% respectively, and those of Desulfovibrio sp., Desulfobulbus sp. and Desulfatirhabdium sp. were 5 ± 2%, 4 ± 2%, 3 ± 1%, respectively ( Figure 3 ). Desulfovibrio sp. and Desulfobulbus sp. are known to facilitate incomplete oxidation of larger organic compounds to smaller compounds, which can subsequently be utilized by FeRB ( Geobacter sp. and Ignavibacteria sp.) for complete oxidation, a synergy between FeRB and SRB that occurred in the Fe (III)-dosed bioreactor (R3). An interesting trend observed with R3 was that the abundances of FeRB and SRB both increased with increasing ferric concentration. In this comparison, Fe/S molar ratios 0.5, 1 and 2 were used by changing Fe (III) and SO 4 2− concentrations to maintain the same total equivalent of electron acceptors for all the ratios ( Figure 4 ). While sulfate concentration decreased slightly with the increasing Fe/S ratio, the abundance of putative SRB increased from 12% to 16%. This was attributed to presence of Desulfovibrio sp. and Desulfobulbus sp. which have been reported capable of facilitating both ferric and sulfate reduction under iron reducing conditions [ 75 , 76 , 77 ]. For examples, sulfate reducers were reported to produce H 2 S via sulfate reduction, which can chemically reduce Fe (III) oxyhydroxides to form iron sulfides [ 78 ]. There is also evidence that these SRB could reduce Fe (III) directly through an enzymatic Fe (III) mechanism and produce siderite concretions [ 75 ]. These synergistic relationships between FeRB and SRB under the iron-reducing conditions in R3 can be an important microbial feature that contributes to the resilience of the iron-dosed biological treatment. 3.5. Feammox and Denitrifying Bacteria R4 was designed to investigate the presence and activities of Feammox bacteria in the bioreactor when organic substrate was not limited. The microbiological analysis showed the presence of Acidimicrobiaceae bacterium at a concentration of 1.84 × 10 6 gene copies/mL. This Acidimicrobiaceae sp., (represented by band A6 by Huang and Jaffe) belongs to the Actinobacteria phylum, which is the only representative of Feammox bacteria [ 35 ]. This Acidimicrobiaceae A6 is a Gram-positive, rod-shaped bacteria with an average length of 1.5–3 μm. Approximately 18% removal of NH 4 + -N with significant presence of Acidimicrobiaceae sp. in R4 showed the evidence of the Feammox activity in the Fe (III)-dosed bioreactor. The disparity in the high COD removal (90%) and the low NH 4 + -N removal (18%) indicated the competitive advantage of heterotrophic FeRB over the Feammox bacteria for ferric iron as the common electron acceptor. However, FeRB did not entirely suppress the Feammox activity. Another important aspect of Feammox reaction is the production of N products including nitrite (NO 2 − ), nitrate (NO 3 − ) or nitrogen (N 2 ) from NH 4 + . Our results showed the insignificant presence of NO 2 − and NO 3 − in the effluent of R4. The presence of denitrifying functional genes nir S and nir K with the concentrations of 1.05 × 10 10 gene copies/mL and 6.80 × 10 7 gene copies/mL, respectively, indicated the denitrifying activities in the bioreactor. These denitrifying activities were most likely stimulated by NO 2 − generated from Feammox. Due to denitrification, NO 2 − and NO 3 − did not accumulate in the bioreactor. As no Anammox bacteria were observed in the samples, Anammox reaction that transforms NO 2 − to N 2 was considered an insignificant microbial pathway in this Fe (III)-dosed bioreactor. 3.6. Fermentative Bacteria Diverse fermentative bacteria were observed in the biomass samples of the bioreactors (R1, R2, and R3, Table 2 ). All these fermentative bacteria were capable of fermenting large organic compounds to smaller compounds [ 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 , 87 , 88 ]. These smaller organic compounds can then be utilized by FeRB and SRB for further carbon oxidation. This suggests a synergistic relationship of fermentative bacteria with FeRB and SRB for substrate utilization. The presence of such wide range of fermentative bacteria contributed to the high microbial diversity under the iron-reducing conditions in these bioreactors. Apart from FeRB, strains of fermentative Clostridium were observed that are known to perform dissimilatory iron reduction [ 89 , 90 ]. The presence of Clostridium sp. in the bioreactors suggests the possibility of their contribution to Fe (III) reduction. As fermentation takes place in the absence of exogenous electron acceptors, fermentation pathway needs to produce fermentative products that can be used as electron acceptors to dispose of the electrons produced during oxidation reactions. If additional electron acceptors such as Fe (III) are present, these excess reducing equivalents (electrons) might be delivered to them. The diversion of reducing equivalents to Fe (III) might provide an energetic advantage through utilizing the oxidation of coenzyme nicotinamide adenine dinucleotide hydrogen (NADH) coupled to Fe(III) reduction to yield ATP [ 89 ] or through change in the fermentation end products. 3.7. Nitrogen-Fixing Bacteria The major NFB observed in these bioreactors were members of the Pleomorphomonas genus, which are Gram-negative, nonmotile, and pleomorphic bacteria belonging to the Alphaproteobacteria phylum [ 91 ]. They have the ability to fix atmospheric nitrogen where bioavailable N becomes limiting [ 92 ]. A previous study on potential synergy between FeRB and NFB in flooded paddy soils showed a positive correlation between the two types of bacteria [ 93 ]. The results showed that FeRB played an important role in the microbial nitrogen-fixing process in the presence of sufficient Fe (III). With increased iron concentrations, abundance of both NFB and FeRB increased. Similar results were observed by Ahmed et al. in their batch reactors, where with increasing Fe/S ratio, abundance of Pleomorphomonas sp. also increased [ 13 ]. Addition of organic carbon (e.g., glucose) also resulted in significant changes in community structures of putative FeRB and NFB [ 93 ]. The FeRB-NFB synergy promoted nitrogen fixation was attributed to two potential reaction pathways. One is that some FeRB can reduce N 2 directly to NH 3 [ 94 , 95 ], and the other is that some FeRB can indirectly promote nitrogen fixation by utilizing H 2 as an electron donor, and preventing biological nitrogen fixation inhibition by H 2 . Fermentative bacteria may add to the complexity of the FeRB-NFB synergy. For example, Clostridium sp. are known to produce H 2 by fermenting larger organic compounds and presence of hydrogen-utilizing FeRB helps prevent H 2 inhibition of NFB. Some genera of Clostridium have been reported to be capable of biological nitrogen fixation [ 96 ], which could meet the N demand from microbial growth under N limiting conditions. Collectively, FeRB, NFB, and fermentative bacteria may synergistically promote nitrogen fixation in Fe (III)-reducing bioreactors.\n\n5. Discussion Several microbial functional interrelationships and their governing factors in Figure 5 can potentially be used to meet various needs of waste treatment, and they are explored in this section. The synergistic relationship observed between FeRB and SRB can be very useful for treatment of sulfate-rich wastewaters. Sulfate-rich wastewaters are generated by many industrial processes such as paper mills and the food processing industry where sulfuric acid or sulfate rich feedstocks are used [ 97 , 98 ]. Fe (III)-dosing to provide sufficient electron acceptor in addition to sulfate for organic removal can be an effective and energy-efficient method for managing these wastewaters. In such applications, Fe/S ratio can be used as a key operating parameter to remove organic pollutants and limit sulfide toxicity by chemical precipitation of FeS. The remaining reduced chemicals (i.e., ferrous, sulfide) in the effluent of the biological treatment can be readily oxidized in a polishing unit before environmental discharge. Presence of Feammox bacteria in the Fe (III)-dosed bioreactor (R4) suggests the prospect of concurrent organic and ammonium removal in a single Fe (III)-dosed bioreactor. Nevertheless, as organic substrate level plays a major role in governing the competitive activities of heterotrophic iron reducers and autotrophic Feammox bacteria, design considerations need to be made for Fe (III)-dosed treatment to achieve satisfactory organic and nutrient removal. For example, Feammox activities may be intensified by adopting a two-stage treatment process, where the first stage is used to remove organic carbons, and the second stage is used for ammonium oxidation via Feammox. For nutrient-rich wastewaters containing low organic content, one-stage treatment may be sufficient to use Feammox activities for N removal. A thought-provoking aspect of the identified functional interrelationships is the synergy between FeRB and NFB. This synergistic relationship is augmented by both ferric iron and organic substrate, and can potentially be employed as an energy-efficient method for ammonium production. Such an engineering application would also require techniques that can recover the produced ammonium in separate process units to maintain N limiting conditions. The limiting factors of this synergism need to be identified and technoeconomic feasibility of this approach for ammonium production warrant further studies. Microbial data of these iron-based bioreactors showed insignificant presence of methanogenic bacteria, as thermodynamically Fe(III) reduction is more favorable than the methanogenic process and generally can suppress methane production [ 74 , 99 ]. However, recent studies in paddy soil environments suggested syntrophic relationships between FeRB and methanogens [ 100 , 101 ] and postulated that bioaugmentation with iron-reducing microbial consortium can intensify methanogenic process [ 102 ]. With less bioavailable ferric, Geobacter species has been found to have syntrophic associations with methanogens through direct interspecies electron transfer (DIET) and thus can attribute to increased methane production [ 101 ]. This syntrophic relationship can potentially be used in anaerobic digestion processes to enhance biogas production by iron dosing."
} | 7,356 |
36630961 | null | s2 | 8,910 | {
"abstract": "Neural activity is often described in terms of population-level factors extracted from the responses of many neurons. Factors provide a lower-dimensional description with the aim of shedding light on network computations. Yet, mechanistically, computations are performed not by continuously valued factors but by interactions among neurons that spike discretely and variably. Models provide a means of bridging these levels of description. We developed a general method for training model networks of spiking neurons by leveraging factors extracted from either data or firing-rate-based networks. In addition to providing a useful model-building framework, this formalism illustrates how reliable and continuously valued factors can arise from seemingly stochastic spiking. Our framework establishes procedures for embedding this property in network models with different levels of realism. The relationship between spikes and factors in such networks provides a foundation for interpreting (and subtly redefining) commonly used quantities such as firing rates."
} | 265 |
30885220 | PMC6421710 | pmc | 8,911 | {
"abstract": "Background Due to their high energy density and compatible physical properties, several monoterpenes have been investigated as potential renewable transportation fuels, either as blendstocks with petroleum or as drop-in replacements for use in vehicles (both heavy and light-weight) or in aviation. Sustainable microbial production of these biofuels requires the ability to utilize cheap and readily available feedstocks such as lignocellulosic biomass, which can be depolymerized into fermentable carbon sources such as glucose and xylose. However, common microbial production platforms such as the yeast Saccharomyces cerevisiae are not naturally capable of utilizing xylose, hence requiring extensive strain engineering and optimization to efficiently utilize lignocellulosic feedstocks. In contrast, the oleaginous red yeast Rhodosporidium toruloides is capable of efficiently metabolizing both xylose and glucose, suggesting that it may be a suitable host for the production of lignocellulosic bioproducts. In addition, R. toruloides naturally produces several carotenoids (C40 terpenoids), indicating that it may have a naturally high carbon flux through its mevalonate (MVA) pathway, providing pools of intermediates for the production of a wide range of heterologous terpene-based biofuels and bioproducts from lignocellulose. Results Sixteen terpene synthases (TS) originating from plants, bacteria and fungi were evaluated for their ability to produce a total of nine different monoterpenes in R. toruloides . Eight of these TS were functional and produced several different monoterpenes, either as individual compounds or as mixtures, with 1,8-cineole, sabinene, ocimene, pinene, limonene, and carene being produced at the highest levels. The 1,8-cineole synthase HYP3 from Hypoxylon sp. E74060B produced the highest titer of 14.94 ± 1.84 mg/L 1,8-cineole in YPD medium and was selected for further optimization and fuel properties study. Production of 1,8-cineole from lignocellulose was also demonstrated in a 2L batch fermentation, and cineole production titers reached 34.6 mg/L in DMR-EH (Deacetylated, Mechanically Refined, Enzymatically Hydorlized) hydrolysate. Finally, the fuel properties of 1,8-cineole were examined, and indicate that it may be a suitable petroleum blend stock or drop-in replacement fuel for spark ignition engines. Conclusion Our results demonstrate that Rhodosporidium toruloides is a suitable microbial platform for the production of non-native monoterpenes with biofuel applications from lignocellulosic biomass. Electronic supplementary material The online version of this article (10.1186/s12934-019-1099-8) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusions In this study, we have successfully demonstrated the use of R. toruloides as a production host for the conversion of lignocellulosic biomass into monoterpenes, many of which have potential applications as biofuels. The investigation of one particular biofuel molecule, 1,8-cineole, revealed that it may be a promising SI and CI fuel.",
"discussion": "Discussion In this study, we investigated R. toruloides as a potential production host for terpene-based biofuels. Sixteen monoterpene synthases producing nine different target monoterpenes were evaluated and a total of seven monoterpenes were successfully produced in R. toruloides . The highest monoterpene titer achieved was 48 mg/L of 1,8-cineole; while this is impressive for a wild-type organism, the natural flux through the essential GPP monoterpene precursor metabolite may still be limiting. The fact that the bifunctional mono/sesquiterpene synthase used in this study to assess the GPP/FPP ratio produced only the FPP-derived sesquiterpene nerolidol, suggested that R. toruloides , like other yeast, produces low amounts of the monoterpene precursor GPP [ 40 ]. Examination of the genome indicated that R. toruloides does not have a dedicated GPP synthase, but like S. cerevisiae , it has a single GPP/FPP synthase ( ERG20 ) that also appears to preferentially produce FPP at the expense of GPP. In addition, this organism has a GGPP synthase, which may also potentially use GPP as a substrate, as has been demonstrated for other GGPP synthases [ 66 ]. Therefore, between these two enzymes, it is not surprising that GPP pool is low. One strategy to enhance GPP levels would be to knock-out the GGPP synthase and replace the native ERG20 with a mutated version that preferentially produces GPP, as demonstrated in other organisms. Beyond balancing the MVA pathway flux to favor monoterpene production, another target to optimize are the terpene enzymes. The Hypoxylon sp. Hyp3 gene used to produce the highest titer of 1,8-cineole is a fungal gene and has good enzyme kinetic parameters with K m of 2.5 ± 0.6 µM and K cat of 0.295 S −1 [ 58 ]. This enzyme is derived from fungi, which means that it does not have a N-terminal plastid signal sequence like plant monoterpene or diterpene synthases, which are often expressed in chloroplast and need to be truncated to be in an active form. Many of the other terpene synthases tested in this study were from plants and were truncated, but it is often difficult to truncate the protein at the right position. For example, a previous study examined a series of N-terminal truncations to produce a “pseudomature” form of limonene synthase lacking the plastid signal sequence, and found that a N-terminal arginine pair was important for the function of the synthase as it plays a role in substrate binding and ionization [ 67 ]. Therefore, it is possible that the truncation sites chosen for some of the monoterpene synthases in this study were suboptimal, and that alternative truncations would result in better performance. These alternative truncations will be pursued in future studies along with detailed kinetic measurements of the enzymes to gain a better understanding of the impact of signal peptides on enzyme activities. Rhodosporidium toruloides is an oleaginous yeast that can accumulate triacylglycerides (TAGs) up to 70% of its weight when starved for nitrogen [ 68 , 69 ]. Strategies that limit driving carbon from the central metabolite acetyl-CoA towards lipid biosynthesis, such as deletions of TAG biosynthetic genes, could potentially allow for more carbon to be naturally diverted into the MVA pathway, resulting in higher terpene production. This diversion of flux could be further enhanced by overexpressing MVA pathway genes, especially those upstream of the GPP intermediates ( ERG10 , ERG13 , ERG12 , ERG8 , HMGR , IDI ), or transcription factors that positively regulate the pathway. The natural production of carotenoids in this organism can potentially be used as a tool to detect enhanced flux through the MVA pathway by screening for colonies with increased red pigmentation. This tool could be used for both the aforementioned targeted engineering strategies in addition to traditional mutagenesis screening. Process optimization could also be used to divert carbon away from lipid biosynthesis toward the MVA pathway, such as optimizing the carbon/nitrogen ratio, examining different carbon sources, nutrients, temperatures, pHs, osmolite concentrations, etc. Taken together, these strategies will likely improve monoterpene production significantly, and will be the focus of future studies. In addition to the natural ability to produce carotenoids, the other major attractive characteristics of R. toruloides is its ability to consume a wide range of carbon sources, including the major components derived from lignocellulose (glucose and xylose), suggesting that lignocellulosic hydrolysates may be an ideal carbon source for this organism. To explore this concept further, R. toruloides was grown on a DMR-EH corn stover hydrolysate containing glucose, xylose and acetate and was shown to grow well and produce higher titers of 1,8-cineole than from a mock hydrolysate or rich medium. DMR-EH hydrolysate is produced in a process where lignocellulosic biomass (corn stover) was deacetylate in dilute alkali, mechanically refined, and enzymatically hydrolyzed to produce high concentrations of monomeric sugars [ 70 ]. According to the 2016 D.O.E. Billion-Ton Report, there are 1.2 to 1.5 billion tons of dry lignocellulosic biomass available in the US, which, if converted into biofuel, could be used to displace 30% of current domestic petroleum consumption [ 5 ]. Our results suggest that R. toruloides can be a good addition to the lignocellulosic biofuel portfolio. Of the monoterpenes tested, 1,8-cineole was produced at the highest titers in R. toruloides . Therefore, we examined its fuel properties in more detail and found that it may be a good SI fuel, especially for downsized, boosted SI engines. The RON/MON/Sensitivity of 1,8-cineol are within the range of premium gasoline and its other properties, like high energy density, make it an appealing biofuel candidate. 1,8-cineole has an energy density that is 5–8% higher than gasoline, while ethanol has an energy density that is 34–36% lower than gasoline. This difference would allow a vehicle to get nearly twice the mileage per gallon (MPG) on a tank of 1,8-cineole as it would on a tank of ethanol assuming no change in MPG due to the lower octane and heat of vaporization (HoV) of 1,8-cineole. While the slightly lower octane and HoV will lower the potential engine efficiency of 1,8-cineole compared to ethanol, the much higher energy density would compensate for it. The only major downside of 1,8-cineole as a fuel is its high freezing point of 1 °C, limiting its use as a drop-in fuel to warm climates. However, this property could be modified by blending 1,8-cineol into gasoline itself, or with the addition of another biofuel molecule that can decrease its freezing point. Ethanol may in fact be a good candidate, and the blend could be formulated to maximize each component’s favorable fuel properties. Along those lines, previous reports have shown that when used as an additive in ethanol-gasoline SI fuel blends, 1,8-cineole can reduce fuel volatility, prevent phase separation, and improve RON [ 12 – 14 ]. Also, 1,8-cineole has been blended into diesel up to 15% by volume for use in a four stroke single cylinder diesel compression ignition (CI) engine, and showed promising fuel emission characteristics [ 15 ]. Taken further, a recent study modified a single cylinder diesel CI engine to utilize 100% eucalyptus oil (primarily composed of 1,8-cineole) as fuel [ 16 ]. Therefore, 1,8-cineole appears to be a dual-purpose molecule that can be potentially used for multiple different engine types, including the two most common engines (SI and CI) found in vehicles in the US."
} | 2,691 |
35523988 | PMC9076648 | pmc | 8,912 | {
"abstract": "Microbial life in soil is fueled by dissolved organic matter (DOM) that leaches from the litter layer. It is well known that decomposer communities adapt to the available litter source, but it remains unclear if they functionally compete or synergistically address different litter types. Therefore, we decomposed beech, oak, pine and grass litter from two geologically distinct sites in a lab-scale decomposition experiment. We performed a correlative network analysis on the results of direct infusion HR-MS DOM analysis and cross-validated functional predictions from 16S rRNA gene amplicon sequencing and with DOM and metaproteomic analyses. Here we show that many functions are redundantly distributed within decomposer communities and that their relative expression is rapidly optimized to address litter-specific properties. However, community changes are likely forced by antagonistic mechanisms as we identified several natural antibiotics in DOM. As a consequence, the decomposer community is specializing towards the litter source and the state of decomposition (community divergence) but showing similar litter metabolomes (metabolome convergence). Our multi-omics-based results highlight that DOM not only fuels microbial life, but it additionally holds meta-metabolomic information on the functioning of ecosystems.",
"conclusion": "Conclusions In this investigation, we highlight the potential of integrating heterogeneous data from multi-omics analyses to reveal functional interactions between microorganisms and dissolved organic matter (DOM). Our study expands on recent advances in understanding how microbial communities adapt to their substrate and how this shapes the process of plant litter decomposition. We identify groups of hundreds of molecules in DOM that together are highly indicative of individual plant litter types and their stage of decomposition. In the future, refining such DOM-based indicator patterns could directly inform about the metabolization and fate of plant-derived carbon in soil and contribute to assessments of the function and health of terrestrial ecosystems. Our findings suggest that bacteria secrete a variety of natural antibiotics in an effort to compete against other bacteria or fungi within the decomposer community. Competitive pressure likely drives constant adaptation and optimization of decomposer community functioning.",
"introduction": "Introduction The importance of dissolved organic matter (DOM) leaching from the litter layer for terrestrial carbon cycling is unequivocal 1 – 4 . However, the role of DOM goes beyond its function as a readily-available substrate for decomposer communities, and its role as a microbial meta-metabolome that contains signatures of ecosystem functioning is less explored 5 . The latter function has been suggested following rapid structural adaptations of decomposers during early-stage litter decomposition, when the substrate composition changes fast as well 6 . On the local scale, this leads to litter-specific decomposer community profiles 7 – 9 . On the landscape scale, however, environmental factors like pH 10 , 11 and soil geochemistry 12 at the litterfall location are suggested as major drivers of microbial community structure. How changes in substrate composition and environmental factors enforce functional adaptation of microbial communities has yet to be determined. Recently, it has been suggested that competition could be a major facilitator of both structural change and functional adaptation within topsoil microbial communities on a global scale 13 . Mechanisms of attack and defense, that happen on a molecular level between various actors across kingdoms, are being unraveled at unprecedented levels of detail 14 , 15 . Recent studies suggest a vast potential of soil microorganisms for the biosynthesis of antibiotics and targeted toxins 15 , 16 . As these substances are secreted, DOM could be an ideal medium to trace active competition functions on a community level and their relationship to changing substrates and environments 4 , 17 – 19 . In this study, we report on a litter decomposition experiment focusing on rapid functional adaptations of microbial decomposer communities to their substrate within the first weeks of litter decomposition 20 . The potential adaptation is assessed via combined analyses of litter leachates (DOM HR-MS, metabolome LC–MS) and community profiles (16S rRNA, metaproteome). We hypothesize that at the early stage of litter decomposition, the native decomposer communities (a) are already well adapted to the litter’s properties and (b) rapidly further optimize their functions to address changes in substrate composition as decomposition progresses. Furthermore, we hypothesize that the progression and mode (synergistic or antagonistic) of community-level functional adaptation are well recorded in the molecular composition of DOM. We test our hypotheses on three major vegetation types: broadleaf forest, evergreen forest and grassland. We account for site-specific factors that could affect the leaf metabolome and decomposer communities by sampling the same vegetation on two sites with distinctly different soil properties and pH 21 . Our laboratory setup excludes the possibility for hostile colonization of the litter by soil microorganisms, allowing us to attribute observed changes in community composition and functioning to adaptations within decomposer communities that natively live on the litter.",
"discussion": "Discussion We found that the qualitative (presence/absence-based) differences in pathway coverage between the litter types, time points and sites in our decomposition experiment were marginal. This finding supports a prevalent theory in microbial ecology, which suggests high functional redundancy in terrestrial microbial communities, especially in the context of organic matter decomposition 56 , 57 . In contrast, our weighted network analysis of DOM, which is based on relative abundances, suggests the emergence of highly litter-specific molecular patterns (Fig. 2 ). This finding supports theories of metabolic specialization during decomposition 8 , 58 . To unify specialization and functional redundancy during litter decomposition, we integrated our molecular network and metabolic pathway prediction (Supplementary Table S4 ). We find that the distribution of the functional metabolites in the network differed sharply between pathways (Fig. 4 ). Molecules belonging to the pathway ‘alpha-linolenic acid metabolism’ are almost exclusively found in the green module, whose members have high relative abundances in the pine litter. The pathway ‘polycyclic aromatic hydrocarbon degradation’ however, shows an opposite pattern, being under-represented in the pine litter modules (green and red). Even though we have found previously that the presence/absence-based coverages of both pathways are similar between the litter types, their relative expression patterns are highly litter-specific. In this decomposition study the litter was coarsely cut and as a result the pine needles were structurally intact (Supplementary Fig. S1 ). Therefore, the metabolization of the needle wax covers, which have been shown to contain alpha-linolenic acid, could have required priority in the decomposition process of the pine litter 59 . The breakdown of the polymerized aromatic hydrocarbons from lignin could, as a result, be delayed during pine needle degradation. Our findings suggest that even though the metabolic potential of decomposer communities is functionally redundant from a qualitative (presence/absence) perspective, the relative expressions of decomposition functions are optimized towards the properties of the immediate substrate. Figure 4 Molecular entities, that were annotated into metabolic pathways, are highlighted within the same weighted correlation network of DOM. Pathways, that are indicative of the degradation of needle waxes (top) and lignin (middle) are highly and little expressed in the pine litter, respectively, suggesting specifically optimized degradation strategies. Molecules in the pathway ‘biosynthesis of antibiotics‘ (bottom) are ubiquitously distributed. For decomposer communities it might be less important what has to be done, because many functions have to be employed in a similar manner on various litter types and are therefore redundantly distributed. It could be more important how much of the respective function has to be performed locally at each point in time and how well the respective actors within the decomposer community are able to perform these tasks. We did not find a litter-specific association to explain the competitive aspects of the decomposition processes. The molecular entities, which were annotated into the pathway ‘biosynthesis of antibiotics’, were spread ubiquitously throughout the weighted network graph. This suggests that attack and defense could be universal mechanisms among decomposer communities to compete, adapt and optimize for their substrate, supporting proposed global patterns of microbial competition in the topsoil 13 ."
} | 2,266 |
25146142 | PMC4153492 | pmc | 8,914 | {
"abstract": "“ Ferrovum myxofaciens ” is an iron-oxidizing betaproteobacterium with widespread distribution in acidic low-temperature environments, such as acid mine drainage streams. Here, we describe the genomic features of this novel acidophile and investigate the relevant metabolic pathways that enable its survival in these environments."
} | 82 |
34950886 | PMC8672035 | pmc | 8,917 | {
"abstract": "Summary Microbial taxonomic marker gene studies using 16S rRNA gene amplicon sequencing provide an understanding of microbial community structure and diversity; however, it can be difficult to infer the functionality of microbes in the ecosystem from these data. Here, we show how to predict metabolism from phylogeny using the paprica pipeline. This approach allows resolution at the strain and species level for select regions on the prokaryotic phylogenetic tree and provides an estimate of gene and metabolic pathway abundance. For complete details on the use and execution of this protocol, please refer to Erazo and Bowman (2021) ."
} | 159 |
28152560 | null | s2 | 8,920 | {
"abstract": "The marine bacterium Vibrio fischeri is the monospecific symbiont of the Hawaiian bobtail squid, Euprymna scolopes, and the establishment of this association involves a number of signaling pathways and transcriptional responses between both partners. We report here the first full RNA-Seq dataset representing host-associated V. fischeri cells from colonized juvenile E. scolopes, as well as comparative transcriptomes under both laboratory and simulated marine planktonic conditions. These data elucidate the broad transcriptional changes that these bacteria undergo during the early stages of symbiotic colonization. We report several previously undescribed and unexpected transcriptional responses within the early stages of this symbiosis, including gene expression patterns consistent with biochemical stresses inside the host, and metabolic patterns distinct from those reported in associations with adult animals. Integration of these transcriptional data with a recently developed metabolic model of V. fischeri provides us with a clearer picture of the metabolic state of symbionts within the juvenile host, including their possible carbon sources. Taken together, these results expand our understanding of the early stages of the squid-vibrio symbiosis, and more generally inform the transcriptional responses underlying the activities of marine microbes during host colonization."
} | 347 |
27048805 | PMC4817265 | pmc | 8,921 | {
"abstract": "ABSTRACT Unraveling the drivers controlling the response and adaptation of biological communities to environmental change, especially anthropogenic activities, is a central but poorly understood issue in ecology and evolution. Comparative genomics studies suggest that lateral gene transfer (LGT) is a major force driving microbial genome evolution, but its role in the evolution of microbial communities remains elusive. To delineate the importance of LGT in mediating the response of a groundwater microbial community to heavy metal contamination, representative Rhodanobacter reference genomes were sequenced and compared to shotgun metagenome sequences. 16S rRNA gene-based amplicon sequence analysis indicated that Rhodanobacter populations were highly abundant in contaminated wells with low pHs and high levels of nitrate and heavy metals but remained rare in the uncontaminated wells. Sequence comparisons revealed that multiple geochemically important genes, including genes encoding Fe 2+ /Pb 2+ permeases, most denitrification enzymes, and cytochrome c 553 , were native to Rhodanobacter and not subjected to LGT. In contrast, the Rhodanobacter pangenome contained a recombinational hot spot in which numerous metal resistance genes were subjected to LGT and/or duplication. In particular, Co 2+ /Zn 2+ /Cd 2+ efflux and mercuric resistance operon genes appeared to be highly mobile within Rhodanobacter populations. Evidence of multiple duplications of a mercuric resistance operon common to most Rhodanobacter strains was also observed. Collectively, our analyses indicated the importance of LGT during the evolution of groundwater microbial communities in response to heavy metal contamination, and a conceptual model was developed to display such adaptive evolutionary processes for explaining the extreme dominance of Rhodanobacter populations in the contaminated groundwater microbiome.",
"introduction": "INTRODUCTION Abrupt, intense environmental stress can overwhelm the ability of microbial communities to adapt by classical selection/drift/mutation mechanisms and force communities to rapidly adapt to such a stress by large-scale genomic rearrangements, including lateral gene transfer (LGT) and gene duplication (GD) ( 1 – 12 ). These genomic rearrangement events frequently underlie the natural attenuation of environmental pollutants, leading to important ecological and economic consequences ( 13 ). However, it remains challenging to identify and quantify such events from metagenome data alone ( 1 , 6 , 12 ). Most metagenomes are too diverse and complex to allow for extensive assembly, making it difficult to distinguish between native and laterally transferred genes. To ameliorate this problem, sequencing of the genomes of isolated microbes from the environments in question can serve as references for comparison ( 14 – 17 ). This strategy has been extensively used in the Human Microbiome Project to identify pathways of gene sharing between community members and hosts ( 10 , 14 , 17 – 23 ). However, for many types of environmental samples, such as soil and subsurface groundwater, few reference genomes are available, and thus, a large proportion of such communities are unrepresented in genomic databases ( 1 ). The Oak Ridge Integrated Field Research Challenge (OR-IFRC) site at the Y-12 Federal Security Complex in Oak Ridge, TN, is a well-characterized experimental field site for studying the environmental impacts of legacy waste on soil and groundwater systems ( 1 , 24 – 27 ). The groundwater contamination plume at this site is derived from spent uranium and nitric acid waste which was stored in unlined ponds that were capped in 1983. Near-source-zone groundwater contains chronically high concentrations of radionuclides (e.g., uranium), nitric acid, organics, salts, mercury, and other heavy metals, resulting in an extremely low diversity of subsurface microbial communities ( 1 , 27 , 28 ). Repeated cultivation-independent analyses of community genomic DNA and RNA from groundwater recovered from well FW106 within the near-source zone revealed a stressed microbial community of low diversity, dominated by populations of metal-resistant, denitrifying Rhodanobacter (>80% by analysis of rRNA gene abundance) ( 1 , 24 , 29 – 33 ). The FW106 metagenome showed high relative abundances of genes encoding geochemical resistance functions required for microbial survival in the presence of known environmental contaminants at the site ( 1 ), including those associated with denitrification, heavy metal resistance, and organic compound degradation. These genes were more abundant in the FW106 metagenome than in a control groundwater community from well FW301 at the OR-IFRC background site ( 1 , 34 ). Experimental and metagenomic analyses have shown that LGT of heavy metal resistance genes may partially account for their high abundance in the FW106 metagenome ( 1 ). To complement metagenomic analyses, multiple laboratories are involved in isolating and sequencing of reference genomes from these environments, including a number of strains of dominant Rhodanobacter populations ( 24 , 32 , 35 – 39 ). These reference genomes are valuable for delineation of the importance of LGT in driving the adaptation of the groundwater microbial communities in response to extremely heavy environmental contamination. Overabundance of resistance systems, as revealed by several previous studies, can result from selection favoring naturally resistant populations ( 30 , 40 – 42 ), via LGT, and/or via GD ( 1 , 6 , 7 , 10 , 11 , 43 , 44 ). Three hypotheses can provide explanations for the overabundance of geochemical resistance genes in the contaminant-stressed community: (i) they provide no selective benefit in the contaminated environment and they are overabundant simply because the host organisms are abundant, (ii) they provide selective benefit to the hosts under contaminated conditions but are not subjected to LGT or GD, and (iii) they provide selective benefit to the host under contaminated conditions and are subjected to LGT and/or GD. To test these hypotheses, multiple Rhodanobacter species were isolated from OR-IFRC sites, and their genomes were sequenced and compared to those of other Rhodanobacter reference strains and to OR-IFRC site groundwater metagenomes. The metagenomes and several of the Rhodanobacter strains were sequenced as part of previous studies, and seven additional Rhodanobacter strains were sequenced as part of this study to complement this data set. Comparative analysis showed that many geochemically important genes were highly abundant in the uranium-contaminated groundwater community due to the dominance of Rhodanobacter in the environment. A recombinational hot spot was identified in the Rhodanobacter pangenome where multiple metal resistance genes were acquired via LGT and/or GD. These results indicated that LGT could play a critical role in driving the evolution of a groundwater microbial community in response to extreme heavy metal contamination.",
"discussion": "DISCUSSION The mechanism by which specific bacterial populations adapt to intensive stress resulting from multiple environmental contaminants and extreme conditions remains an open question. Several experimental analyses at the OF-IFRC sites suggest that nitrate concentration, pHs, and heavy metal concentrations are the key drivers of community structure and functions in contaminated sites ( 30 , 41 ), and many strains have been isolated from these sites that show resistance to these particular environmental stresses ( 28 , 32 , 37 , 39 , 52 , 53 ). Comparative genomics analyses of the Rhodanobacter genomes and the OR-IFRC groundwater metagenomes show that many of the genes identified as overabundant in the contaminated-groundwater metagenome are present in most of the Rhodanobacter strains. Some of these genes, such as the core denitrification pathway genes, are vertically inherited, with no evidence of lateral gene transfer. Many metal resistance genes, including czc and mercuric resistance genes, show evidence of extensive lateral gene transfer and/or gene duplication in Rhodanobacter strains. An area of the Rhodanobacter pangenome appears to be a recombinational hot spot that allows rapid acquisition and accumulation of these genes in Rhodanobacter populations. Thus, the overabundance of putative resistance genes in the contaminated-groundwater population can be attributed to a combination of vertical inheritance, LGT, and GD in the dominant Rhodanobacter population ( Table 2 ). TABLE 2 Lateral gene transfer and gene duplication events in Rhodanobacter species relevant to geochemical resistance in the FW106 contaminated-groundwater metagenome Hypothesis Event a Stressor Gene or product Representative CDS(s) b I. No selection in response to presence of stressors, no LGT/GD VI None Housekeeping genes IIa. Selection in response to presence of stressors, no LGT/GD VI Heavy metals High-affinity Fe 2+ /Pb 2+ permease R2APBS1_2568 VI Heavy metals Cytochrome c 553 R2APBS1_0257 IIb. Selection in response to lack of stressors, no LGT/GD GL Nitrate/nitrite Nitrate/nitrite transporter R2APBS1_2892 IIIa. Selection in response to presence of stressors, LGT/GD predating contamination LGT Heavy metals High-affinity Fe 2+ /Pb 2+ permease R2APBS1_1742 LGT Heavy metals (mercury) Mercuric resistance operon genes (original) R2APBS1_1850–1856, c R2APBS1_2076–2071, c R2APBS1_2095–2101 c IIIb. Selection in response to presence of stressors, LGT/GD following contamination GD Heavy metals (mercury) Mercuric resistance operon genes (duplicates) R2APBS1_1850–1856, c R2APBS1_2076–2071, c R2APBS1_2095–2101 c LGT Heavy metals (divalent cations) Co 2+ /Zn 2+ /Cd 2+ efflux genes R2APBS1_1735–1751 LGT Heavy metals (arsenic) Arsenate reductase operon genes R2APBS1_1729–1733 LGT Nitrate/nitrite Clade 2b nitrate/nitrite transporters RhoFW510R12_08626 IV. Abundant genes that cannot currently be characterized Unknown NiFe hydrogenase III (large and small subunits) R2APBS1_2173–2174 a VI, vertical inheritance; LGT, lateral gene transfer; GD, gene duplication; GL, gene loss. b The locus tags of representatives of the gene families in the reference genome R. denitrificans 2APBS1 are listed. c It is unclear which is original and which are duplicates. However, it should be noted that the methods used for identifying putative LGT events all use different strategies to detect regions of the genome that deviate from the genomic background. Of the 51 putative laterally transferred genes listed in Table S2 in the supplemental material, 76.5% were identified as positive LGT events by at least one of the programs used (Alien_Hunter, DarkHorse, and GOHTAM) and 39.2% were identified as positive LGT events by at least two methods. Many of these genes are located in operons, and thus, it can be assumed that if some of the genes in the cluster were identified as laterally transferred, then all genes in the cluster were likely laterally transferred as well. Additional evidence supporting lateral transfer of these genes is their phylogenetic similarity to orthologs from organisms other than Rhodanobacter and the association of many of these genes in the Rhodanobacter genomes with mobile elements. Based on the results of this work, a conceptual model was constructed to predict how Rhodanobacter populations adapted to contamination and came to dominance in many of the OR-IFRC contaminated-groundwater communities. Comparative genomics analyses of Rhodanobacter isolates suggest a selection-transfer-selection dynamic by which populations of this particular genus survive and thrive in these environments ( Fig. 5 and Table 2 ). The original community prior to contamination is highly diverse, consisting of many different species with different abundances (represented by different colors in Fig. 5A ). Following the introduction of contaminants, the community goes through an initial round of selection in which only those species naturally resistant to the contaminants survive (indicated by the increase in the number of Rhodanobacter populations, represented by blue in Fig. 5B ). In the case of Rhodanobacter , vertically inherited genes, such as cytochrome c 553 genes, denitrification genes, and native metal resistance genes like czc , may impart a selective benefit on these populations ( Fig. 5B ). Simultaneous with or immediately following contaminant introduction, the surviving populations take up exogenous DNA from the environment that may allow rapid adaptation to specific contaminant profiles ( Fig. 5C ). In the case of Rhodanobacter populations, metal resistance genes, such as czc and mercuric resistance operons, appear to be the primary targets of lateral transfer and acquisition. Much of this activity appears to occur in the identified recombinational hot spot of the Rhodanobacter pangenome. The process of lateral gene transfer is likely ongoing, but once established by subsequent rounds of selection, new populations of Rhodanobacter may emerge adapted to highly specific niches within the contaminated-groundwater environment. This is consistent with some recent suggestions that ecologically determined gene transfer networks may exist in microbial communities, allowing rapid sharing of niche-adaptive genes in the flexible (or accessory) genome, and that these networks are influenced by recombination rates within populations ( 12 ). The exact nature of the vector for LGT is unknown, but circumstantial evidence from the comparative genomics analyses and other work suggests that viral and/or plasmid vectors are the primary means of LGT. It remains to be seen if plasmids specific to Rhodanobacter exist or if Rhodanobacter species are capable of conjugation. It appears that Rhodanobacter may be a key node in such a network in contaminated-groundwater environments. A second interesting question regards whether the observed LGT/GD events preceded contamination of the site or whether these are recent changes specifically in response to contamination. Some metal resistance genes, for instance, are conserved in most of the strains, suggesting they were acquired prior to contamination. On the other hand, the OR-IFRC strains often show genomic properties distinct from those of the reference strains isolated outside the OR-IFRC sites, particularly in terms of their complement of metal resistance genes. While population structures are evident in the OR-IFRC strains, it is still difficult to determine whether these represent legacy (fixed in the population prior to contamination) or opportunistic (fixed in the population following contamination) events. While definitive evidence of lateral gene transfer is difficult to identify by using sequence data alone, the biogeography and population structures of Rhodanobacter within and outside the OR-IFRC show compelling evidence of contamination-driven lateral gene transfer and gene duplication. In particular, the apparent duplication of mercuric resistance operons in R. denitrificans species provides the best evidence to date of a genome-scale evolutionary event specifically in response to contamination. Another interesting question is whether the proposed model of Rhodanobacter may be extended to other bacterial populations at the OR-IFRC sites. For instance, do other naturally resistant OR-IFRC populations, such as Pseudomonas and Stenotrophomonas , exhibit the same selection-LGT-selection dynamics as Rhodanobacter ? If so, do these populations employ the same mechanisms for accumulating these genes? Analysis of the OR-IFRC background groundwater metagenome from the OR-IFRC background site suggests a reservoir of mobile geochemical resistance genes in groundwater communities ( 34 ). If multiple populations in the contaminated-groundwater source zone are drawing from the same pool of mobile genes, different OR-IFRC populations might show similar LGT patterns. Alternatively, individual population genome dynamics may affect the rate of gene transfer and fixation in specific populations, which in turn might affect their relative fitness in these environments. As more OR-IFRC genomes become available, these hypotheses can be tested directly. The results of this analysis help explain why Rhodanobacter species are dominant in many OR-IFRC contaminated environments, but some questions remain. It is known, for instance, that Rhodanobacter populations exist in uncontaminated background communities, albeit in extremely low abundances. It is unknown whether the introduction of contaminants allows opportunistic Rhodanobacter populations to grow to dominance or whether these environments are colonized by exogenous Rhodanobacter populations. It is also not clear why Rhodanobacter populations dominate certain contaminated environments and not others. That is, do specific environments select specific populations, or are the dominance patterns observed at the OR-IFRC the result of stochastic effects ( 54 – 57 )? Further study of this phenomenon may help explain why other, seemingly equally adapted populations are not able to outcompete Rhodanobacter at specific sites and vice versa. In conclusion, comparative genome analysis of Rhodanobacter isolates and metagenomes from contaminated-groundwater sites clarifies the role and extent of LGT and GD in the accumulation of geochemical resistance genes in the population. Metal resistance genes in particular are highly susceptible to lateral transfer and have accumulated in multiple Rhodanobacter species due to a putative recombinational hot spot in the Rhodanobacter pangenome. This ability to rapidly acquire metal resistance genes via LGT likely explains in part the ability of Rhodanobacter to dominate uranium-contaminated sites."
} | 4,477 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.