pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
36732469 | null | s2 | 3,558 | {
"abstract": "OmcZ nanowires produced by Geobacter species have high electron conductivity (>30 S cm"
} | 21 |
39174531 | PMC11341674 | pmc | 3,559 | {
"abstract": "While different stages of mutualism can be observed in natural communities, the dynamics and mechanisms underlying the gradual erosion of independence of the initially autonomous organisms are not yet fully understood. In this study, by conducting the laboratory evolution on an engineered microbial community, we reproduce and molecularly track the stepwise progression towards enhanced partner entanglement. We observe that the evolution of the community both strengthens the existing metabolic interactions and leads to the emergence of de novo interdependence between partners for nitrogen metabolism, which is a common feature of natural symbiotic interactions. Selection for enhanced metabolic entanglement during the community evolution repeatedly occurred indirectly, via pleiotropies and trade-offs within cellular regulatory networks, and with no evidence of group selection. The indirect positive selection of metabolic dependencies between microbial community members, which results from the direct selection of other coupled traits in the same regulatory network, may therefore be a common but underappreciated driving force guiding the evolution of natural mutualistic communities.",
"introduction": "Introduction Microorganisms are typically part of communities that display a large taxonomic diversity and in which members are often linked through obligatory metabolite exchanges 1 – 4 . These interactions likely developed through a stepwise process, resulting in a gradual erosion of independence of the initially autonomous organisms 5 , 6 . Notably, similar processes likely guided eukaryogenesis 7 , 8 and the emergence of obligate symbiotic interactions 5 , 6 . However, whilst communities composed of partners linked by various degrees of entanglement can be found in nature, the investigation of these evolutionary snapshots allows drawing only limited conclusions about the dynamics, molecular mechanisms and selection forces behind transitions towards increased cooperation 5 , 9 . Artificial synthetic communities may hence represent valuable models for the controlled observation of evolutionary processes in a relatively short time 10 . This approach was previously used to assemble mutualistic communities through the co-cultivation of microbial partners carrying auxotrophies 11 – 17 , some of which could be evolved towards reinforced metabolite exchanges 18 – 24 . Nevertheless, the next phase of the community evolution predicted by ecological models 25 – 27 - an increase in the interdependence mediated by the loss of traits - has not been experimentally reproduced so far. Moreover, there is still only limited validation for theoretical frameworks proposed to explain how enhanced cooperation may be evolutionarily favoured over selfish behaviours 5 , 6 , 10 , 11 , 15 , 25 , 28 , 29 . In this study, we aim to experimentally reproduce the transition towards increased cooperation by evolving an interkingdom m utualistic co nsortium between auxotrophs of E scherichia coli and S accharomyces cerevisiae (MESCo). One advantage offered by MESCo is the ability to control the aggregation status between its two partners, since E. coli is naturally able to co-aggregate with yeast via type I fimbriae 17 . This allows us to assess the possible influence of group selection, the major theoretical mechanism proposed to drive the emergence of cooperation 5 , 6 , 24 , 28 . Moreover, natural symbiotic interactions, including the most pronounced examples of metabolism reduction, often involve eukaryotic and prokaryotic organisms 6 , 30 – 33 . Compared to the previously investigated monospecies consortia between either bacteria 13 , 18 or yeast 14 , 20 , the two MESCo partners display larger differences in their metabolism and exometabolome profiles 34 , which is expected to favour cross-feeding interactions 35 and may allow for a greater degree of mutual de novo adaptability and metabolic specialisation of the partners. Indeed, the evolution of MESCo in our experiments leads not only to an extensive reinforcement of the initial metabolic dependencies as observed previously for purely bacterial 18 , 19 , 22 or eukaryotic 20 communities, but also to the emergence of a new interdependence in the consortium resulting from a strong reduction of ammonia assimilation by the yeast partner, and consequently its increased reliance on the bacterial exometabolome as source of nitrogen. Finally, through the use of model organisms as the community members, we perform an in-depth molecular investigation of the underlying genetic and metabolic changes.",
"discussion": "Discussion Ecological models predict that enhanced partner addiction should emerge from the co-evolution of interdependent organisms 27 , 41 , such as those exchanging essential metabolites in mutualistic and symbiotic communities 2 – 4 , 32 . In this study, we report experimental observation of an increase in metabolic entanglement during the experimental laboratory evolution of an engineered interkingdom mutualistic community between auxotrophs of E. coli and S. cerevisiae (MESCo). Using this synthetic interkingdom community formed by two model organisms enabled us to mechanistically describe several characteristic steps in the progression of communities towards tighter cooperation. This evolution firstly included the strengthening of pre-existing interactions, through the self-serving enhanced uptake of the exchanged metabolites by both partners, as previously observed in bacterial or yeast communities 16 , 20 , 21 . The cooperation was further promoted by the costly increase in sharing the partner-serving metabolite by E. coli , and at least in one instance also by yeast. Previous observations of the enhanced excretion of the partner-serving metabolites were primarily made in the context of laboratory evolution of spatially structured communities 18 , 19 , 24 , and the formation of multicellular clusters was even favoured by the co-evolution 23 . These instances of selection on cooperative traits could thus be interpreted as a consequence of local cooperation within small neighbourhoods 42 , favouring group selection that is normally assumed to be a prerequisite for the evolution of cooperation 5 , 6 , 10 , 11 , 15 , 28 . In contrast, although group selection was specifically enabled in our experiments in the MESCo Agg community, the cooperative metabolite sharing rather emerged as a consequence of the pleiotropic mutations in the same regulatory component that simultaneously increased production of the partner-serving metabolite and uptake of the self-serving metabolite. While it has previously been demonstrated that pleiotropy can stabilise existing cooperation against the emergence of cheaters through regulatory coupling between cooperative and private traits 43 – 45 , the relevance of pleiotropy in the evolution of social traits has been questioned 29 , 46 . Nonetheless, at least in one case pleiotropy was proposed to explain the selection of the ecm21 mutations in a mutualistic yeast community 20 , which may also apply to our experiments where the same yeast gene was mutated. Repeated instances of the apparently indirect selection on increased interdependency between partners during the short evolution of our MESCo communities suggest that pleiotropy may be a generally important, and previously underappreciated, factor in the evolution of sociality, promoting the emergence of social traits. Our results also indicate a mechanism that could favour selection of such pleiotropic over purely self-serving mutations, because of the observed negative impact of the latter on community growth. Besides reinforcements of the pre-existing interactions, the evolved MESCo communities showed repeated emergence of a new level of dependency, with the yeast partner becoming increasingly reliant on E. coli for assimilation of ammonium, the primary nitrogen source in the medium during the co-culture evolution. This increased entanglement evolved through sequential inactivation of the major pathways of ammonium assimilation in yeast. Importantly, fixation of the most prominent of these mutations, gdh1 , was apparently contingent on the prior fixation of the argR mutation in E. coli , indicative of co-evolution between partners. Although the underlying selection pressure remains to be fully elucidated, the reduced ability of yeast to use ammonium may cause a rewiring of the nitrogen assimilatory pathways to enhance the uptake and consumption of arginine, thereby providing mutants with an increased scavenging ability for this metabolite, and thus with a competitive fitness advantage in a cross-feeding community. Additionally, the assimilation of ammonium under conditions of cross-feeding may cause some metabolic imbalance 47 , for example in the redox potential 48 , which could select for its inactivation. In either case, we conclude that this erosion of autonomy is again selected indirectly, as a consequence of regulatory trade-offs within the yeast metabolic network. In summary, we observed that high regulatory connectivity of the cellular genetic and metabolic networks leads to the fixation of mutations that enhance existing and create novel metabolic interdependencies in the evolving synthetic microbial community, even in the absence of group selection. We speculate that similar mechanisms may drive the emergent division of labour during the evolution of natural communities, including other dependencies based on shared nitrogen-containing compounds that are common in symbiotic interactions 32 , 49 , 50 ."
} | 2,395 |
35464318 | PMC9022003 | pmc | 3,560 | {
"abstract": "The leaky integrate-and-fire (LIF) spiking model can successively mimic the firing patterns and information propagation of a biological neuron. It has been applied in neural networks, cognitive computing, and brain-inspired computing. Due to the resistance variability and the natural storage capacity of the memristor, the LIF spiking model with a memristor (MLIF) is presented in this article to simulate the function and working mode of neurons in biological systems. First, the comparison between the MLIF spiking model and the LIF spiking model is conducted. Second, it is experimentally shown that a single memristor could mimic the function of the integration and filtering of the dendrite and emulate the function of the integration and firing of the soma. Finally, the feasibility of the proposed MLIF spiking model is verified by the generation and recognition of Morse code. The experimental results indicate that the presented MLIF model efficiently performs good biological frequency adaptation, high firing frequency, and rich spiking patterns. A memristor can be used as the dendrite and the soma, and the MLIF spiking model can emulate the axon. The constructed single neuron can efficiently complete the generation and propagation of firing patterns.",
"conclusion": "7. Conclusion In this work, the LIF model with the non-volatile memristor is proposed successfully, and we aim to develop the application of memristor in neuroscience. We choose the flux-controlled memristor to combine with the LIF spiking model and get the MLIF spiking model. To demonstrate the superiority of the MLIF model over the LIF model, we compared the firing patterns of the two models. The simulation results show that the MLIF model has good biological spiking frequency adaptation, higher firing frequency, and rich firing patterns. The MLIF model can reproduce the firing behavior of biological neurons very well. Due to the intrinsic characteristics of a memristor, it can potentially promote the analysis and application of biological neural models. This work has experimentally proved that a single memristor can be used as a synapse, performs the function of integration and filtering of the dendrite, and realizes the function of integration and firing of soma. An individual neuron constructed entirely by memristors can emulate passive and active propagations over time; finally, it efficiently transmits the information. Our MLIF model converts the current pulses to potential spikes, corresponds to the Morse code sequence. The simulation results indicate that the MLIF model can successfully generate and recognize the Morse code. Therefore, the proposed MLIF model can be a potential building block for reproducing the behaviors of a biological neuron and constructing the spiking neural networks.",
"introduction": "1. Introduction Neurons, as the dual-role of the function units of the perceiving-conducting stimulus and the information processing, can carry out particular tasks of sensory, motor, neural responses, and cognition (Hirokawa et al., 2019 ), and so on. Many neuron models emerged (Hodgkin and Huxley, 1952a ; FitzHugh, 1961 ; Morris and Lecar, 1981 ; Bernander et al., 1994 ; Izhikevich, 2003 ) to mimic the functions of a biological neuron, especially the LIF spiking model. It is a simpLIFied and much easier model for hardware implementation and large-scale integration (Slepova and Zhilenkov, 2018 ). The primary purpose of an artificial neuron is to mimic the functions of biological neurons in an energy effectiveness and scalability way. The typical LIF model consists of a capacitor and a resistor. The external stimulus is applied to the LIF model until a threshold is reached, and then the action potential is produced (Han and Meyyappan, 2018 ). It is widely applied in bioinspired and brain-inspired neuromorphic information processing systems (Belkaid and Krichmar, 2020 ; Neves and Timme, 2020 ; Yang and Kim, 2020 ; Doutsi et al., 2021 ). Although the LIF model can reproduce the firing behaviors of neurons after each activation, the previous pulse cannot be retained, and the biological spiking frequency adaptability does not perform very well. To solve these deficiencies, we need to find a new device to promote the LIF neuron model. A memristor is a potential element to emulate the function and behavior of a biological synapse or neuron (Hu et al., 2016 ; Choi et al., 2018 ; Chen et al., 2019 ; Greenberg-Toledo et al., 2019 ; Xia and Yang, 2019 ; Wang et al., 2020 ; Shi and Zeng, 2021 ) gets a lot of attention. The non-volatile memristor modulates its conductance due to ion motion, similar to the phenomena in biological neurons and synapses. Therefore, these advantages enable the memristor to become an inevitable choice as a building block between artificial neural networks and biological neural networks. The LIF oscillatory neuron with a memristor is used to perform threshold and firing functions (Jiang and Hu, 2021 ), and the LIF neuron with a threshold switching memristor realizes the firing behavior is driven by the threshold (Dev et al., 2020 ). A flexible memristor is integrated into the LIF neuron, generates four firing patterns, and implements the transformation between analog signals and spiking signals (Zhu et al., 2021 ). The TSM (threshold switching memristor) LIF neuron circuit experimentally performs the integrate and fire behaviors (Xu et al., 2022 ). The diffusive memristor LIF neuron model mimics neuron integration, leakage, spatiotemporal, and firing activities (Yang et al., 2020 ). The LIF neuron combined with a TSM can show the “leaky-integrate-fire” function and low power consumption (Lu et al., 2020 ). Even though the LIF neuron model with a memristor had achieved lots of progress in emulating biological neurons, the implementation of retaining the previous pulse and performing the biological spiking frequency adaptability has not been yet explored in the MLIF neuron model. In our work, we first experimentally implement the MLIF neuron model. The memristor exhibits non-volatile behavior to “remember” the previous pulses by applying a series of pulses. In addition, the biological spiking frequency adaptability performs very well by combining the LIF neuron with a memristor. Furthermore, an individual neuron model formed by memristors is presented, and the distortionless transmission of the action potential is realized. The primary work is to construct the memristive leaky integrate-and-fire spiking model after integrating a memristor to the LIF spiking model. In section 2, the LIF spiking model will be introduced to analyze it. In section 3, the MLIF spiking model is constructed. When the distinct stimuli act on the LIF and MLIF spiking models, the MLIF spiking model performs good biological adaptation, high firing frequency, and rich firing patterns in section 4. The memristor experimentally simulates the functions of synapse, dendrite, and soma, and an individual neuron is entirely constructed by memristors. It successfully reproduces the firing patterns and vividly emulates the information transmission of a biological neuron in section 5. Finally, the proposed model is further verified by the generation and recognition of Morse code in section 6. Section 7 is the conclusion of the article."
} | 1,825 |
30567498 | PMC6299973 | pmc | 3,561 | {
"abstract": "Background Genome-scale metabolic modeling is a cornerstone of systems biology analysis of microbial organisms and communities, yet these genome-scale modeling efforts are invariably based on incomplete functional annotations. Annotated genomes typically contain 30–50% of genes without functional annotation, severely limiting our knowledge of the “parts lists” that the organisms have at their disposal. These incomplete annotations may be sufficient to derive a model of a core set of well-studied metabolic pathways that support growth in pure culture. However, pathways important for growth on unusual metabolites exchanged in complex microbial communities are often less understood, resulting in missing functional annotations in newly sequenced genomes. Results Here, we present results on a comprehensive reannotation of 27 bacterial reference genomes, focusing on enzymes with EC numbers annotated by KEGG, RAST, EFICAz, and the BRENDA enzyme database, and on membrane transport annotations by TransportDB, KEGG and RAST. Our analysis shows that annotation using multiple tools can result in a drastically larger metabolic network reconstruction, adding on average 40% more EC numbers, 3–8 times more substrate-specific transporters, and 37% more metabolic genes. These results are even more pronounced for bacterial species that are phylogenetically distant from well-studied model organisms such as E. coli . Conclusions Metabolic annotations are often incomplete and inconsistent. Combining multiple functional annotation tools can greatly improve genome coverage and metabolic network size, especially for non-model organisms and non-core pathways. Electronic supplementary material The online version of this article (10.1186/s12864-018-5221-9) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusions This analysis has led us to make recommendations for providing a more comprehensive metabolic genome annotation. We found that a single annotation tool is often insufficient unless one is only interested in core metabolism where different tools often agree. Organisms that are phylogenetically far removed from well-studied model organisms are particularly susceptible, in which case annotation tools will tend to diverge far more. In addition, one can trade off confidence in predictions versus greater coverage by using the intersection or union of multiple annotation tools. BLASTing against a database of reference sequences is generally an inefficient method for annotating enzymes but may be useful to cover more recently assigned EC numbers not yet included by other tools. Still, all these efforts require manual curation to bring together annotation from multiple sources. More tool development is needed to merge annotations beyond simple EC numbers, and a universal reference database for well-balanced reactions and metabolites would be a very valuable resource to merge annotations that use different reaction nomenclatures [ 38 – 40 ]. Likewise, now that annotation tools such as TransportDB are producing significant numbers of transporter annotations with substrate predictions that are precise enough to be included in metabolic modeling, more tool development may be needed to fully take advantage of these substrate predictions in Flux Balance Analysis methods, and move beyond the current implicit assumption used by most algorithms that all metabolites can be transported when needed.",
"discussion": "Results and discussion Discrepancies in metabolic annotations between different tools In total, the RAST, KAAS, EFICAz and BRENDA tools produced 47,447 Gene-EC annotations (“gene X codes for an enzyme with EC number Y”) across the 27 reference genomes, for an average of 1757 annotations per genome. The metabolic gene-EC annotations produced by these automated genome-wide annotation tools differed drastically (Fig. 1 ). Each tool produced on average between 23% (EFICAz) and 48% (BRENDA) unique gene-EC annotations that were not predicted by any of the other tools. Overall, fewer than a quarter of all gene-EC annotations were agreed on by at least 3 tools. Fig. 1 Large differences exist between the sets of Gene-EC annotations generated by the four annotation tools across the 27 reference genomes When two annotation tools both assigned a particular gene an EC annotation, the two tools assigned at least one identical EC annotation in more than 50% of cases (Table 2 ). BRENDA on average had the lowest agreement with other tools (56.0–69.7%). Note also that BRENDA had a larger fraction (47.5%) of gene-EC annotations not shared by any other tools (Fig. 1 ). In contrast, EFiCAz showed the highest agreement with other tools (69.7–86.4%) and had the lowest number of gene-EC annotations not shared by other tools (23.4%). Table 2 Percentage of gene-EC annotation agreements that exist between pairs of tools Tool Combination Gene-EC Agreements KEGG-RAST 16,697/20,915 (79.8%) KEGG-EFICAz 14,413/16,677 (86.4%) KEGG-BRENDA 3777/6748 (56.0%) RAST-EFICAz 12,977/15,694 (82.7%) RAST-BRENDA 3907/6288 (62.1%) EFICAz-BRENDA 3902/5601 (69.7%) The denominator is the number of genes across the 27 reference genomes that are covered by both tools. The numerator counts the number of such genes for which both tools provide at least one identical gene-EC number annotation. Comparing annotation tools against each other can give a sense of which tools are closest to a consensus annotation, or which tools seem to be outliers, however assessing the integrity of these predictions is difficult without experimental validation. Therefore, to determine which tool provides the best ratio of true/false annotation predictions we compared their predictions to the EcoCyc database [ 37 ]. EcoCyc is a gold-standard continuously updated database of experimentally determined and extensively hand-curated enzymatic functions in Escherichia coli K-12 substr. MG1655, the most-studied model organism in modern biology. We used the gene-EC numbers annotated in EcoCyc as a set of true positives to evaluate how well the two most commonly used automated annotation tools, RAST and KEGG, are able to assign function to the enzymes in E. coli K-12. Overall, there was a high degree of overlap between the RAST and KEGG predictions (Fig. 2 ) with EcoCyc, however neither tool covered all of EcoCyc, and both tools predicted a small number of reactions not experimentally validated. Fig. 2 Gene-EC annotations produced by KEGG and RAST for E. coli K-12, compared to the EcoCyc gold standard. The sets and intersections are drawn proportionally to the number of annotations in each One major caveat of using E. coli to evaluate the quality of annotation tools is that so much of our knowledge of microbial metabolism is based on E. coli , and therefore annotation tools can be expected to be trained or optimized on E. coli to some extent, so performance on E. coli is not necessarily indicative of results on other organisms. For example, the KEGG annotation provided by KAAS is done by calculating bidirectional best BLAST hits against annotated reference genomes including E. coli , essentially providing a direct lookup of E. coli annotations in the KEGG database. Coverage of the metabolic network reconstruction While the individual gene-EC annotations examined in the previous section reflected the quality and agreement between annotation tools, the total set of EC reactions annotated for a genome by each tool reflects the size and coverage of its metabolic network reconstruction. In this case, we simply counted the total number of different EC numbers, regardless of whether multiple genes are annotated with the same EC number (isozymes), or whether genes were annotated with multiple EC numbers (multifunctional enzymes). On average, the four tools combined produced 868 EC reactions per genome, with the largest agreement between RAST and KEGG (Fig. 3 ). In general, KEGG produced a larger number of unique EC numbers, which could indicate more over-prediction, or more comprehensive pathway coverage. Note that both RAST and KEGG also generate many reactions without official EC numbers, so in some cases these annotation tools may produce annotations that are minor variants or subsets of the canonical EC number reaction in EcoCyc. Fig. 3 Reaction overlap between the annotation tools (average number of EC numbers per genome) EFICAz produced the least number of unique EC numbers but had high agreement with RAST and KEGG (86% of EFICAz EC numbers were also generated by both RAST and KEGG – see Fig. 3 ), even though it uses very different annotation methods, which suggests it can be helpful in validating those annotation tools. Note that EFICAz also produces incomplete “three-digit” EC number annotations (e.g. 1.2.3.-) which may be useful for hole filling but were not considered in this analysis. BLASTing against the BRENDA database of reference enzymes produced the smallest number of annotations, but a high fraction of unique EC numbers. Interestingly, of the top 10 unique EC numbers produced by this method, only one is also covered by RAST and KEGG, two of the EC numbers have been deprecated by the Enzyme Commission, and six are EC numbers that have been assigned in 2000 or later and may not have been incorporated into the predictions by the other annotation tools yet. So even though a simple BLAST against a reference database such as BRENDA proves to be one of the less effective means for assigning metabolic functions (compared to the more sophisticated HMM or protein family based methods used by the other tools), it may still have some value to capture recently described enzymes not already covered by the other tools. While counting the number of EC numbers reflects the size of the metabolic network, counting the numbers of genes that have received any metabolic annotation reflects the genome annotation coverage. Additional file 1 : Figure S1 shows the number of genes in all of the reference genomes annotated with one or more EC numbers by each of the tools. On average, 1361 genes per genome were assigned a function with at least one tool, and more than 65% of these genes were assigned a metabolic function by more than one tool. The results show that just as each tool adds a significant number of reactions to the metabolic network model, each tool also significantly contributes to the number of genes covered with metabolic annotations. The EC numbers on which the different tools most often agree across the 27 reference genomes tended to belong to well-studied core metabolic pathways. Out of the 79 EC numbers on which all four tools agreed in at least half of the genomes (Additional file 2 ), more than three quarters (61/79) were involved in biosynthesis or biodegradation of amino acids, nucleotides, carbohydrates, and cofactors; or in the processing of RNA, DNA and proteins. In contrast, almost none of these EC numbers were involved in biosynthesis or degradation of fatty acids, lipids, aromatics compounds, or secondary metabolites. The differing sets of annotations produced by each tool can enable a user to trade off confidence for coverage, with higher confidence obtained when accepting only annotations that were agreed upon by multiple tools (the intersection), or higher coverage obtained by using the combined set of annotations from multiple tools (the union). To examine the effect of taking the intersection (higher confidence) or union (higher coverage) of the annotation tools, we compared combinations of the four annotation tools against the E. coli K-12 “gold standard” metabolic reactions in EcoCyc (Fig. 4 ). These combinations included the single tool annotations, as well as the union and intersection of all four tools combined in pairs, triplets and quartets. The resulting EC annotations from these combinations were then compared to the 1064 EC numbers from EcoCyc, and the number of true positives, false positives, and false negatives were calculated for each combination, as well as the resulting Precision and Recall (Table 3 ). Fig. 4 Precision vs Recall of EC numbers for different combinations of tools on EcoCyc. Individual tools are denoted by B, E, K, or R for BRENDA, EFICAz, KEGG, and RAST, respectively. For each combination of tools, we calculated precision and recall for both the union and intersection of the sets of EC numbers generated by each tool. The union corresponds to the set of EC numbers generated by at least one of the tools in the combination, while the intersection corresponds to those EC numbers generated by every single tool in the combination Table 3 Definitions of Precision, Recall and associated terms Term Formula Definition True Positive TP EC numbers predicted by tools and found in EcoCyc. False Positive FP EC numbers predicted by tools and not found in EcoCyc. False Negative FN EC numbers in EcoCyc but not predicted by tools. Precision TP/(TP + FP) Fraction of predicted EC numbers that are in EcoCyc. Recall TP/(TP + FN) Fraction of EC numbers in EcoCyc correctly predicted by tools. A more lenient annotation policy (e.g. merging annotations from all tools) will tend to generate fewer false negatives but more false positives, achieving a higher Recall at the expense of lower Precision. Conversely, a more restrictive annotation policy (e.g. only including EC numbers if all tools agree on them) can increase Precision, but at the expense of a lower Recall. Figure 4 shows a plot of precision versus recall for all the different combinations of tools. Out of the individual tools, KEGG performed best in terms of both precision and recall on this dataset (although as mentioned before, performance on E. coli K12 may not reflect performance on other genomes), and a simple Blast against the BRENDA database performed worst. Combinations that contain some union of the tools have a higher recall than each of the individual tools in the combination, but a somewhat lower precision (> 80%). In contrast, intersections of annotations from two or more tools show very high (> 90%) precision but much lower recall (< 65%). A consensus annotation that produces both higher recall and higher precision might be achieved by means of a weighted sum of all the annotation sources, similar to the approach taken by EnzymeDetector [ 33 ]. Even though we expected the E. coli K-12 genome to be a best-case annotation candidate, there were still significant differences in the annotations produced by the different tools, with each tool only covering a subset of the known enzymes in EcoCyc. The four annotation tools annotated a significantly larger fraction of the genome, and showed much more agreement on E. coli than on more remote lineages such as Actinomycetes, Bacteroidetes, or Clostridia (Fig. 5 a). For E. coli K-12, 60% of EC numbers were agreed on by 3 or more tools, while 28% EC numbers come from only a single tool. In contrast, for C. difficile 630, only 33% of EC numbers were agreed on by 3 or more tools, and 48% of EC numbers come from only a single tool. Note that this also seems to justify the use of multiple annotation tools in the recent work on modeling of Clostridium beijerinckii NCIMB 8052 [ 28 ]. Compared to the five E. coli strains in our dataset, the annotation tools also cover on average around 30% fewer genes for the 13 genomes in the bottom half of Fig. 5 b. We see a similar effect when we compare the annotation coverage for B. subtilis – arguably the best studied Gram-positive model organism – with all 8 other Gram-positive genomes in our dataset. These results suggest that genome coverage for each tool, and agreement in annotations across tools are significantly worse for organisms that are more phylogenetically distant from well-studied model organisms, making it all the more important to combine multiple tools when annotating these genomes. Fig. 5 Genome coverage and overlap in annotations varies across genomes. a Horizontal bars represent the fraction of the total number of EC numbers for each genome produced by only a single tool, or by two, three or all four tools. The 27 reference genomes were sorted with respect to the fraction of EC numbers that were predicted by 3 or more tools (blue bars). The top of the list is dominated by model organisms such as E. coli , B. subtilis , and closely related organisms. As we move farther away from such well-studied model organisms, the fraction of unique EC numbers predicted only by a single tool (red bars) increases, at the expense of those predicted by multiple tools. b The fraction of genes annotated as enzymes by each tool likewise decreases as we move farther away from model organisms such as E. coli . Note that two of the organisms with a drastically reduced genome content, Candidatus Portiera aleyrodidarum BT-QVLC and Candidatus Evansia muelleri, also have a relatively higher fraction of core metabolic enzymes Transporter annotations Knowledge of the molecules and substrates an organism can transport and exchange with the environment can help to build a more accurate metabolic model. Both RAST and KEGG include membrane transport annotations, yet both tools yielded on average only 114 and 204 transporter predictions per genome, respectively (Fig. 6 a and Additional file 1 : Figure S2). Many of these annotated transporters lack substrate predictions (52% of transporter annotations in RAST, 25% in KEGG) or have ambiguous substrate predictions (ranks 3–4 (Table 4 , Additional file 3 ); 20% in RAST, 28% in KEGG), while less than half have substrate predictions that are sufficiently detailed to be incorporated in a metabolic model (ranks 1–2; 28% in RAST, 48% in KEGG; Fig. 6 b). In contrast, TransportDB produces an average of 426 transport annotations per genome, and most of those have specific substrate predictions (59% rank 1–2; 32% rank 3–4, 10% rank 5; Fig. 6 b). Fig. 6 a Total number of genes annotated as transporters, regardless of substrate. b Transporter annotations with substrates predictions specific enough to be included in metabolic models (rank 1 or 2) Table 4 Examples of substrate annotation ranking Rank Substrate Examples 1 Metabolite that can be incorporated as a transport reaction in a metabolic model • Fe • lysine 2 Substrate(s) that map to a small number of possible transport reactions • Mg/Co/Ni • aromatic amino acid 3 Broader substrate classes not directly usable to construct a metabolic network • dipeptide • sugar 4 Very broad class of substrates • multidrug efflux • protein 5 No substrate annotated Transporter substrates were ranked from most specific (rank 1) to least specific (no substrate, rank 5). See Additional file 3 for the full table Transporter annotations by RAST, KEGG and TransportDB showed surprisingly little overlap. Out of the more than 15,000 genes annotated as transporters (regardless of substrate prediction), the three tools only agree on 2.8% (423/15,161). Out of those, only 130 genes are annotated by all three tools with a specific substrate prediction (ranks 1–2). When two or more tools provide a sufficiently specific substrate annotation, the substrate annotations tend to agree 85% of the time, even if they may not be identical (for example, one transporter was annotated as “leucine/valine”, “leucine”, and “branched-chain amino acid” by TransportDB, RAST and KEGG respectively). Overall, the detailed transporter annotations by TransportDB’s Transporter Automatic Annotation Pipeline provide a significant advance over more general metabolic annotation tools such as RAST and KEGG."
} | 4,906 |
38026874 | PMC10666632 | pmc | 3,562 | {
"abstract": "Microalgae have emerged as a promising, next-generation sustainable resource with versatile applications, particularly as expression platforms and green cell factories. They possess the ability to overcome the limitations of terrestrial plants, such non-arable land, water scarcity, time-intensive growth, and seasonal changes. However, the heterologous expression of interested genes in microalgae under heterotrophic cultivation (dark mode) remains a niche area within the field of engineering technologies. In this study, the green microalga, Chlorella sorokiniana AARL G015 was chosen as a potential candidate due to its remarkable capacity for rapid growth in complete darkness, its ability to utilize diverse carbon sources, and its potential for wastewater treatment in a circular bioeconomy model. The aims of this study were to advance microalgal genetic engineering via dark cultivation, thereby positioning the strain as promising dark-host for expressing heterologous genes to produce high-value phytochemicals and ingredients for food and feed. To facilitate comprehensive screening based on resistance, eleven common antibiotics were tested under heterotrophic condition. As the most effective selectable markers for this strain, G418, hygromycin, and streptomycin exhibited growth inhibition rates of 98%, 93%, and 92%, respectively, ensuring robust long-term transgenic growth. Successful transformation was achieved through microalgal cell cocultivation with Agrobacterium under complete darkness verified through the expression of green fluorescence protein and β-glucuronidase. In summary, this study pioneers an alternative dark-host microalgal platform, using, Chlorella , under dark mode, presenting an easy protocol for heterologous gene transformation for microalgal host, devoid of the need for expensive equipment and light for industrial production. Furthermore, the developed genetic transformation methodology presents a sustainable way for production of high-value nutrients, dietary supplements, nutraceuticals, proteins and pharmaceuticals using heterotrophic microalgae as an innovative host system.",
"conclusion": "4 Conclusion Our work presented the first report on the successful transformation of the green microalga, C. sorokiniana AARL G015, using Agrobacterium -mediated transformation under heterotrophic cultivation. The transformation was confirmed through GUS and GFP expression analysis. To enhance the transformation efficiency, further additional optimization of specific strains and cultivation modes should be evaluated. The successful transformation of heterologous genes through a cost-effective technique under completely dark mode marked a significant advancement in the field of engineered microalgae. This finding provides a molecular toolkit for the high-value production of phytochemicals for further industrial applications, offering a cost-effective, less labor-intensive, and time-saving alternative platform. Furthermore, the development of new molecular tools and techniques is crucial for fully harnessing the economic potential of microalgae as circular model organisms. Genetic engineering improvement of specific strains and exploration of promising hosts for downstream applications are valuable areas for future research. This research marks a prominent step and fill the gap towards unlocking the potential of heterotrophic microalgae for genetic engineering applications. Further development of methods and strategies for transgene expression in non-model microalgae are critically required.",
"introduction": "1 Introduction Is traditional production both sustainable and sufficient? The current need for sustainable platforms, resources, and cell factories, particularly for the production of valuable phytochemicals has prominently directed attention towards microalgal production due to its short doubling time. This focus is driven by its potential alignment with the concept of a bio-circular economy model, which aims to make the most of resources. This microalgal production and applications sector have undergone a significant shift over the past 5 years ( Maria et al., 2023 ). With the global population projected to surge to 9.9 billion by 2050 ( United Nations, 2023 ), coupled with severely environmental shifts, the advancement of biotechnologies has become important to solve these issues. Utilizing alternative green cell factories—specifically microalgae—to express foreign genes can enhance foods, feeds, recombinant proteins, biopharmaceuticals, and high-value compounds. A wide range of biological chemicals is derived from fruits, vegetables, whole grains, and other part of plants. These compounds produced by plants are known as phytochemicals. However, concerns regarding plant cultivation limitations persist, such as arable land, seasonal changes, climatic conditions, time-consuming processes, and high production costs ( Jareonsin and Pumas, 2021 ). In order to address these challenges, a sustainable approach to develop eco-friendly industries has been required. Most biopharmaceutical products are currently manufactured in animal cells, but each host has its limitations, including low yield, high cost, virus contamination, and expensive medium costs, among others. As a result, alternative hosts for phytochemical production have been continually investigated. In this sense, eukaryotic microalgae hold high metabolic potential within the context of the circular bioeconomy concept, and offer tremendous metabolic potential to serve as a suitable platform for plant chemicals production, as mentioned earlier. Genetic engineering of microalgae provides cutting-edge tools to expand the platform for food and feed production ( Kusmayadi et al., 2021 ). Furthermore, eukaryotic algae not only share evolutionary ancestry with land plants but also fulfill most of the necessary criteria, particularly for plant chemicals ( Novoveska et al., 2019 ; Saini et al., 2019 ). They exhibit metabolic potential and post-translational modification pathways appropriate for phytochemical production ( Weiner et al., 2018 ). Additionally, microalgal metabolism precursors are more closely associated with phytochemicals compared to those in prokaryotic hosts ( Lauersen et al., 2018 ). Therefore, further developments are needed to establish microalgae as a regular resource for interested compounds. Microalgal genera such as Arthrospira , Dunnaliella , and Chlorella have emerged as an ideal platform for large-scale production due to their Generally Recognized as Safe (GRAS) status recognized by the Food and Drug Administration (FDA) ( Yaakob et al., 2014 ). Among these, Chlorella holds significant industrial potential due to several key factors, including high growth rate, ability to grow in mass culture reactors over extended periods, robustness in coping with various harsh conditions. With an expected value of USD 412.3 million by 2028, Chlorella sp. holds one of the highest market values, with a targeted yield of 5,000 tons dry matter per year ( Levasseur et al., 2020 ). Additionally, Chlorella is considered to be safe with lower risk of viral, prion, or bacterial endotoxin contamination ( Yaakob et al., 2014 ). Moreover, Chlorella exhibits high nutritional value and high lipid content exceeding that of most terrestrial plants and is recognized as a source for biofuel production. Thus, recombinant products derived from microalgae hold significant advantages and are closely related to secondary metabolites produced by desired plant genes. Currently, plant-derived ingredients have shown positive correlations in numerous plant ingredients aspects. These include the utilization of phytochemical ingredients produced from microalgae as a manufacturing, such as polyunsaturated fatty acid (PUFAs) for nutritional purposes, cannabinoids for medical use, terpenoids for pigments and supplements, Cytochrome P450s for plant metabolites, astaxanthin for food coloring, hydrocarbon for high-quality fuel applications, and plant hormone ( Laban, 2019 ; Jareonsin and Pumas, 2021 ). With the advancements in microalgal biotechnologies, including genomics, bioinformatics, analyses, and genetic and metabolic engineering, further studies should focus on developing heterotrophic hosts for novel products synthesis through gene insertion. Microalgae offer sustainability advantages as an alternative expression host for recombinant production due to their ability to perform correct post-transcriptional and post-translational modifications, cost-effectiveness, and shorter expression times without significant land or water usage ( Yang et al., 2016 ). However, the field of microalgal engineering is still relatively new, and the engineering technologies for microalgae are not as well-developed as those for heterotrophic microbes ( Lu et al., 2021 ). Microalgal biotechnology currently faces several challenges as follow: i) limited model microalgal species, ii) limited reports on expression and transformation under heterotrophic mode, iii) low transformation efficiencies, and v) despite the greater advantages offered by heterotrophic microalgae compared to autotrophic microalgae, there is still a lack of knowledge regarding the biocircular economic model for utilizing sustainable organisms as a dark host for future industrial development. The benefits of dark microalgal hosts include more economical nutrient options, lower instrument costs, and ease of operation and maintenance ( Jareonsin and Pumas, 2021 ). Their rapid growth can adapt to large-scale production within a few weeks without concerns about light penetration, leading to reduce costs related to employee hiring time, and electricity when using artificial light sources, and so on ( Yang et al., 2016 ). In heterotrophic microalgal cultivation, 80% of the production cost depends on organic carbons sources ( Park and Moon, 2018 ), while bacterial cultivation can reach up to 60% ( Omoregie et al., 2019 ). Certain heterotrophic microalgae, particularly Chlorella , can be cultivated using wastewater or by-products as nutrient sources, contributing to both economic and environmental sustainability ( EI-Sheekh et al., 2014 ). For example, in our previous studies ( Chen et al., 2019 ; Jareonsin et al., 2023 ), we found that the most cost-effective option is to use poultry effluent with molasses as carbon sources, saving up to 80% compared to mBG11 medium (reducing the cost from $0.55 to $0.11) and 52% compared to the wastewater with 10 g/L glucose (reducing the cost from $0.23 to $0.11), in term of cost, as shown in Figure 1 . While bacteria cultivation with commonly used media such as yeast extract, nutrient broth, cooked meat medium, and lactose broth can cost $3.56, $1.59, $134.06, and $2.19 per liter ( Omoregie et al., 2019 ). Utilizing wastewater or by-product strategies for heterotrophic microalgal hosts offers an alternative method compared to employing other heterotrophic host organisms like bacteria and yeast. Even though bacteria can use inexpensive media, it is crucial to note that certain low-cost media may lack rigorous quality control and reproducibility, especially in ureolytic bacteria ( Cuzman et al., 2015 ). The price to be paid not only the production costs, but also the following effects from many dimensional disadvantages in terms of applicability, environmental impacts, and cost-effectiveness. In another perspective, microalgae hosts can indirectly contribute to the reduction of their entire production chain. For instance, in certain countries, like Spain, cost can exceed $1.72/kg for various chemicals and $0.13/kWh for energy use in cleaning industrial wastewater ( Ruiz et al., 2022 ). FIGURE 1 Cost analysis of Chorella sorokiniana AARL G015 using commonly used medium (mBG11) and wastewater with different organic carbon sources (glucose and molasses). Additionally, yeast , Saccharomyces cerevisiae , cultivated on laboratory medium costs $0.85 per liter, while using batch production medium can cost $0.035-$0.045 per liter ( Malairuang et al., 2020 ). In this sense, coculture with yeast and microalgae could reduce the cost and enhance the interested product ( Qin et al., 2019 ). For example, the co-cultivation of Chlorella sp. and Rhodotorula glutinis in starch processing effluent has the capability to more efficiently convert nutrients into single-cell compared to the monoculture approach. Consequently, the co-fermentation field of microalgae and yeast emerges as a promising and practical strategy for cost-effective and sustainable production ( Lu et al., 2023 ). Furthermore, an effective and industrially scalable heterotrophic Chlorella sorokiniana GT-1 could reduce costs to $1,601 per ton of biomass if the biomass concentration reached 200 g/L, spending annual costs of equipment depreciation and power consumption lower than Chlorella protothecoids , reducing 24% of the overall production cost ( Jin et al., 2021 ). In a particular gene aimed to produce plant substances for future phytochemical applications, microalgae may offer more favorable conditions for genetic engineering compared to yeast. Moreover, microalgae metabolism involves the production of precursors more closely associated with phytochemicals than that of prokaryotic hosts ( Lauersen et al., 2018 ). Despite yeast is an excellent eukaryotic host due to its low cost a scalability, the common occurrence of hypermannosylation in yeast can result in misfolded proteins and activity malfunction ( Yusibov and Mamedov, 2010 ). This occurrence may result in extra costs depending on the genes of interest. However, a limited number of large companies, such as Terravia Holding, Inc., Nutrinova, DSM, and Corbion are engaged in heterotrophic cultivation ( Santin et al., 2022 ). While autotrophic cultivation is the main strategy in microalgal production, there are some limitations ( Ruiz et al., 2022 ). New insights are needed for this field to enhance transformation efficiencies, establish specific transformation systems for individual strains and products, develop target editing methods, explore alternative advance technologies for microalgae, and address the remaining challenges, particularly by using microalgae that can easily and completely grow under dark cultivation as a heterotrophic host. Is it possible to grow microalgae without using light? Most microalgae are capable of growing in wide range of environments. The three cultivation modes for microalgae—autotrophic, mixotrophic, and heterotrophic cultivation—are categorized by the source of energy and carbon ( Dreesen et al., 2010 ). As the economic and industrial scales increase to meet human demands, heterotrophic cultivation (dark mode) holds a significant advantage in biomass production and economic value in larger scale, especially for high-volume and cost-effectiveness. Heterotrophic cultivation enables the utilization of wastewater, offers easy control of factors, supports bioremediation, and operates within a closed system that requires less environmental factors compared to autotrophic and mixotrophic cultivation, which rely on light for being the source of energy ( Chen et al., 2023 ). In the search of a suitable microalga strain to develop innovative biotechnology, the green microalga, Chlorella sorokiniana has intrigued as a candidate due to its enriched fast-growing nature and its potential for multiple applications across various industries under different cultivation modes ( Park et al., 2019 ; Yun et al., 2020 ; Politaeva et al., 2021 ). Our previous studies have revealed that Chlorella sorokiniana AARL G015 exhibits robust growth, reaching full nutrient utilization within a short period of 5–14 days, and achieving higher biomass production (3–6 g/L) in dark mode ( Jareonsin et al., 2023 ). This cultivation mode reduces the cost of production as low as 0.02 $/g. As recombinant microalgae strains are being developed, it is crucial to evaluate the expression of heterologous genes under various cultivation types, particularly in dark cultivation as mentioned above. Fortunately, some of the most common microalgae are capable of heterotrophic growth. However, there is a lack of researches in the engineering field and further investigations are required. Heterotrophic microalgal host, referred to as a ‘dark host’ in this context, can be independently managed without the risk of environmental contamination, as they can be cultivated in a fermenter or closed system ( Khan et al., 2016 ). However, information on transgene expression under dark mode with C. sorokiniana for strain development is very limited and needed for studies and developments. \n Agrobacterium tumefaciens -mediated transformation technique is one of the methods used to deliver genes into host cells. This technique has shown success in a broad spectrum of cells, especially from both plant and microalgae, such as Chlamydomonas reinhardtii , Chlorella vulgaris , Dunaliella salina , Haematococcus pluvialis , and others ( Dehghani et al., 2018 ). Furthermore, a new assay for the screening of transgenic strains has been developed which employs quantitative analysis of β-glucuronidase (GUS) histochemical assay with X-gluc ( Yedahalli et al., 2018 ). Based on other studies, the transformation efficiency of three microalgal species was found to be as follows: Chlorella sp (12.25%), Scenedesmus bajacalifornicus (2.92%) and Ankistrodesmus sp. (3.5%) using Agrobacterium transformation method ( Sanitha et al., 2014 ). This transformation method is flexible and easy to perform with less cost; however, there are limited research studies providing specific results for microalgal strains. Therefore, there is a gap to fill for more accessible knowledge and information regarding heterologous expression in microalgae. Therefore, there is a need to identify a new eukaryotic microalgal dark host that can effectively produce heterologous proteins and phytochemicals for various biotechnological applications to improve quality of nutrition. In this study, the genetic modification of a specific dark microalgal host was evaluated as a promising alternative host. Our aim was to produce phytochemicals, and develop an efficient transformation system for Chlorella as a dark host, thereby enhancing the potential of transgenic microalgae for the production of desirable substances to meet human needs. Up to our knowledge, this is the first report of a genetic engineering system in C. sorokiniana under completely dark cultivation via Agrobacterium -mediated transformation. The increasing interest in finding new and innovative algal host systems and platforms is motivated by the potential to create sustainable organisms that can serve as hosts or cell factories. This approach aims to address the challenges related to the requirement for light, cost, and energy at an industrial scale.",
"discussion": "3 Results and discussion 3.1 Biomass growth under dark cultivation \n Chlorella sorokiniana AARL G015 was successfully cultivated in mBG11, which our previous study demonstrated to be the most suitable medium for this strain under heterotrophic cultivation ( Jareonsin et al., 2023 ). Glucose was selected as a carbon source due to its common use in industrial fermentation processes ( Shu, 2007 ). The cell growth revealed the maximum specific growth rate (µ max ) of 8.52 days -1 and doubling time (dt) of 0.14 days. In this medium, the microalga rapidly grow within 7 days after a 4-day lag phase and then last approximately 10–14 days before entering the stationary growth phase. The exponential phase was chosen to perform genetic transformation procedures. The cells obtained through spread and streak plating were inoculated to liquid cultures for transformation. As illustrated in Figure 3 , the biomass dry weight was determined by harvesting microalgal cells for a week. The maximum biomass production of C. sorokiniana AARL G015 occurred on the fourth day, reaching 5.5 g/L. This production is 8.6 times higher than the maximum cell yield of C. vulgaris –a common model for microalgal host–under heterotrophic condition ( Cai et al., 2021 ). It is also 3.1 times higher than that of C. sorokiniana FC6 IITG cultivated under mixotrophic cultivation ( Kumar et al., 2014 ) ( Table 2 ). Subsequently, dry biomass production dropped after 4 days. The cells slightly decreased from fifth to seventh day, after which the cells entered the stationary phase. This result indicated that this strain is suitable for cultivation under dark mode. Moreover, these results implied that this strain holds promise for large-scale industrial applications due to its fast growth and high yield production. Furthermore, once the targeted gene is expressed, this microalga has the potential to produce a substantial of crucial amino acids such as arginine, lysine and cysteines as reported by Chen et al. (2023) , which could serve as an alternative protein source for reducing or replacing fish meal in aquafeeds, meeting the urgent demand. Recently, other research has focused on upscaling C. sorokiniana FZU60 to achieve ultra-high lutein production under dark mode. The novel fed-batch strategy, as demonstrated by Xie et al. (2022) , has shown to significantly enhance the performance and commercial viability of lutein derived from microalgae. After the transformation, the microalgal host can continually be employed under dark cultivation with expression, making this upcoming strain viable for multipurpose. Additionally, understanding the specific growth under specific conditions is crucial before assessing other criteria of host platform. This growth evaluation will contribute to a successful cultivation process, potentially enabling high-density cultures that facilitate cheaper and easier downstream processing steps on a scaling-up ( Geada et al., 2023 ). FIGURE 3 \n Chlorella sorokiniana AARL G015 growth characteristic under heterotrophic cultivation in mBG11. TABLE 2 Growth characteristics of Chlorella sp. Microalgae \n C. sorokiniana AARL G015 \n C. sorokiniana FC6 IITG \n C. vulgaris \n \n C. vulgaris \n \n Medium \n mBG 11 \n mBG 11 \n Modified Bristol mBG 11 \n \n Cultivation mode \n Heterotroph Mixotroph Heterotroph Heterotroph \n Carbon source \n Glucose (10 g/L) Sodium acetate Glucose (20 g/L) Glucose (10 g/L) \n Scale \n 50–100 mL 100 mL 200–500 mL 150 mL \n Specific growth (µ, d −1 ) \n 5.13 - 1.371 1.18 \n Biomass (g/L) \n 5.5 1.75 - 0.64 \n Biomass productivity (g/L/D) \n 2.48 0.11 0.687 3.2 \n Lipid content (%) \n 11.69 39.2 - - \n References \n This study \n Kumar et al. (2014) \n \n Cao et al. (2023) \n \n Cai et al. (2021) \n Other studies support the idea that Chlorella can utilize waste more efficiently than Chlamydomonas due to its ability to digest and use waste nutrients while simultaneously removing pollutants from wastewater. This implies that in the near future, the use of Chlorella in genetic field can also meet the bio-circular model by using waste for culturing. However, the genetic toolbox of Chlamydomonas is more advanced than that of Chlorella . Therefore, the need to develop approaches across various strains and cultivation modes is crucial. 3.2 Selectable marker for dark Chlorella host We investigated the growth response of C. sorokiniana AARL G015 to different concentrations of commonly used antibiotics and determined the optimal screening conditions for liquid culture transformed cells. To ensure accurate results and long-term stability, a range of antibiotic concentrations from low to high was tested to verify the percentage of growth inhibition to prevent false positive transformants. Furthermore, for Agrobacterium transformation, the elimination of this bacterium after infection is crucial. This is typically achieved using CTX or RIF. Therefore, the antibiotic sensitivity of CTX and RIF was tested at high concentrations (250, 500, 1,000 μg/mL) to confirm that these two antibiotics have less effects on microalgal growth. This evaluation ensures that in Agrobacterium -mediated transformation, which often involves the use of these two antibiotics in protocols for bacteria elimination, there is no significant interference with the growth of transgenic microalgal cells. The result from low antibiotic concentrations (25, 50, 100 μg/mL) showed that certain antibiotics could only temporarily inhibit the growth of C. sorokiniana AARL G015 during a short exposure period (day 2–4) ( Supplementary Figure S1 ). After this phase, the microalga exhibited a gradual resurgence in growth under dark conditions for STR, CLO, and G418, and NEO ( Figure 4A ). Intriguingly, the strain became slightly less sensitive to certain antibiotics when exposed to lower concentrations, leading to just slower growth. In particular, SPEC and AMP were unable to inhibit growth entirely. Therefore, the inhibition percentage appeared as negative in Figure 4A , located within the inverted column. However, by the 14-day of exposure, the inhibitory effect of antibiotics dramatically dropped, indicating that this strain is capable of tolerating low antibiotic concentrations, particularly at a concentration of 100 μg/mL of CLO, SPEC, and ZEO, which were unable to inhibit microalgal growth on day 14. Notably, this strain exhibited complete tolerance to AMP at all lower concentrations without decreasing its growth. The physiological changes indicated by the green color in response to low antibiotic concentrations (as shown in Figure 4B ) suggested the microalga can survive and gradually grow. Conversely, higher antibiotic concentrations (250, 500, 1,000 μg/mL) caused significant inhibition of the biomass concentration in this strain. C. sorokiniana AARL G015 proved to be sensitive to certain antibiotics under dark mode, particularly G418, hygromycin, and streptomycin, which resulted in the highest growth inhibition: 98% at 6 days (500 μg/mL), 93% at 4 days (250 μg/mL), 92% at 4 days (250 μg/mL), respectively ( Figure 5A ). In contrast, the strain exhibited slightly less sensitivity to AMP and SPEC with growth inhibition of less than 40% at higher concentration (1,000 μg/mL). While CLO demonstrated minimal inhibition at low concentrations, higher concentrations led to the highest growth inhibition (89%) on day 12 ( Supplementary Figure S2 ). FIGURE 4 Antibiotics sensitivity of antibiotics under dark mode at 100 μg/mL: (A) growth inhibition (%) (B) physical characteristic of C. sorokiniana AARL G015 of some antibiotics on day 0, day 4, day 8 and day 14. FIGURE 5 Antibiotics sensitivity at high antibiotic concentrations (250, 500, 1,000 μg/mL) under dark mode of C. sorokiniana AARL G015 (A) growth inhibition (%) of the most potent antibiotics affecting the microalgal growth after 14 days exposure (B) culture color at various antibiotic concentrations on the last day of exposure. Cefotaxime and rifampicin, commonly used as selection markers for Agrobacterium , did not have a strong effect on the microalgal growth even high concentration. As shown in Figure 5B , the physical color of the microalgal culture with CTX and RIF appeared green compared to control culture. Hence, cefotaxime was selected as the suitable marker for Agrobacterium in further transformation. The evaluation and screening of suitable antibiotics become crucial, especially given the lack of research under dark mode for this strain. These results provide essential data for advanced genetic engineering of microalgae and expanding the pool of selectable markers, particularly for heterotrophic cultivation. Currently, for Chlamydomonas , a model microalga, six antibiotic resistances are commonly used as selectable markers: zeocin, hygromycin, kanamycin, paromomycin, sulgadiazine, and spectinomycin ( De Carpentier et al., 2020 ). In this study, we found another six antibiotics, namely, G418, HYG, STR, NEO, KAN and CLO, as alternative markers for C. sorokiniana AARL G015 under dark cultivation. The suggested antibiotics containing gene cassettes in the plasmid that provide resistance to STR (200 μg/mL), HYG (150 μg/mL), and NEO (100 μg/mL). 3.3 Agrobacterium -mediated transformation under dark mode The pCAMBIA1304 vector containing GUS gene and mgfp5 reporter was transformed into A. tumefaciens vir helper strain 4404. A single clone of transgenic A. tumefaciens was selected for transformation of C. sorokiniana AARL G015. Subsequently, the transgenic microalga was selected under agar plate containing STR (200 μg/mL), HYG (150 μg/mL), and NEO (100 μg/mL) as a selectable marker ( Figure 6A ). FIGURE 6 The growth of transformed microalgal cells on solid medium using Agrobacterium -mediated transformation method: (A) the growth of transformants carrying pCAMBIA 1304 which was contained GFP gene, GUS gene, and HYG gene under selective medium (B) the growth of transformants after cocultivation with Agrobacterium on day 3, and (C) the growth of transformants on selective medium on day 13 at the first round. This marks the first study of an efficient genetic transformation system in heterotrophic microalgae, Chlorella sorokiniana AARL G015, as a dark host using Agrobacterium -mediated transformation to express heterologous genes. C. sorokiniana cells were cultured on mBG11 agar medium supplemented with 100 µM of acetosyringone, for 3 days ( Figure 6B ) at 25°C in complete darkness. The cocultivation process involved the introduction of Agrobacterium containing the vector at a cell density of OD 600 = 0.6. After transformation, the transgenic microalga was fully developed on selective media within 2 weeks ( Figure 6C ). However, further investigation is required to improve and optimize the transformation efficiency to this heterotrophic strain. The ability to transform larger-sized genes into the host genome of A. tumefaciens has significant advantages on C. sorokiniana AARL G015. Previous studies have shown that this transformation method has been successful in a few freshwater microalgal strains, including Chlamydomonas rienhardii , H. pluvialis , C. vulgaris , Dunaliella tertiolecta and C. sorokiniana ( Cha et al., 2012 ; Sanitha et al., 2014 ; Norzagaray-Valenzuela et al., 2017 ; Sharma et al., 2021 ). Nevertheless, it is important to note that Agrobacterium -mediated transformation method has been limited only to auto/mixotrophic cultivation modes. Additionally, various parameters can influence the transformation efficiency using this approach, such as the microalgal species, type of cocultivation media, acetosyringone concentration, mode of microalgal cultivation, and cocultivation duration. Based on the report from Sharma et al. (2021) , BG11 cocultivation medium was found to produce the greatest number of transformed colonies for autotrophic C. sorokiniana compared to TAP medium, which is commonly used in model microalgae. In this study, mBG11 medium was used for heterotrophic C. sorokiniana AARL G015 due to the dark cultivation condition where the microalga needs to use organic carbon sources from medium instead of utilizing energy from light. Moreover, our cocultivation duration was similar to other studies, with resistant colonies of this strain appearing after 3 days of cocultivation ( Cha et al., 2012 ; Sharma et al., 2021 ). Another report showed that higher concentrations of acetosynringone than 100 µM can reduce the transformation efficiency ( Zho et al., 2009 ). Therefore, in this study, we used 100 µM as the acetosynringone concentration. There have been research papers discussing genetic approaches to understanding the genes involved in carbon source utilization for enhanced growth, as well as gene regulation to control glucose metabolism. However, there is a lack of information on heterologous genes expression on heterotrophic cultivation. This research demonstrated that certain strains, particularly C. sorokiniana AARL G015, which can grow better in heterotrophic mode, should be studied for the development of their genetic transformation under dark cultivation conditions . \n 3.4 GUS biochemical expression GUS histochemical assay was performed to check GUS expression as a reporter in transformed cells, following the antibiotic screening for resistant cells. This assay allowed transformants to react with the substrate (X-gluc), resulting in blue color appearance. Figure 7 shows the transformed C. sorokiniana AARL G015 in weakly blue-colored cells, while the non-transformant shows no blue color. However, the screening for transient GUS expression exhibited weak coloration. There are many variables that affect the quality of the histochemical localization. Several factors are involved in the GUS gene expression pattern, such as promoter efficiency, random positioning effects in the nuclear genome, the rigid cell wall, which may possibly lead to improper penetration of GUS substrate, and a lack of GUS gene expression, as well as epigenetic-derived transgene silencing ( Kathiresan et al., 2009 ; Doron et al., 2016 ; Norzagaray-Valenzuela et al., 2017 ; Miamioh, 2019 ). As a consequence, the transgenic cells were visualized using GFP gene as a reporter. FIGURE 7 GUS histochemical analysis. Microscopic view of transformation of heterotrophic microalga with pCAMBIA1304 vector harboring GUS gene reporter (A) WT; untransformed control Chlorella sorokiniana \n (B) T; transgenic C. sorokiniana showing GUS expression. Hence, there is a need to study and improve this technique for many other microalgal species. Using the GUS gene as a reporter makes transgenic monitoring easier and less time-consuming ( Su et al., 2016 ). Additionally, further studies are necessary to explore factors that may influence the expression of GUS gene in specific strains. The goal is to enhance GUS gene expression and address gaps in the genetic toolbox for non-model microalgae. 3.5 Detection of GFP in transformed algal cells Visualization of GFP using a laser confocal microscopy (Leica stellaris 5, Germany) confirmed the expression of GFP in transgenic cells as green, while non-transformed cells exhibited only red chlorophyll autofluorescence. The putative C. sorokiniana transformants carrying the pCAMBIA1304 vector were recovered for the third time in liquid-selection medium. The result demonstrated that transgenic C. sorokiniana containing pCAMBIA1304 could be easily distinguished from non-transformant cells ( Figure 8 ). Strong fluorescence emissions were observed in C. sorokiniana with pCAMBIA1304 and were localized in the cells where the bright green signal appeared, whereas non-transformants showed no green. FIGURE 8 Confocal microscopy analysis (A) GFP expression in transgenic cell culture (B) wild-type cell cultures. From left to right: (i) cells detection in chloroplast auto-fluorescence channel, (ii) cells detection in the GFP fluorescence channel, (iii) merged images, and (iv) phase contrast image cells. 3.6 Stability of transformants After undergoing 5-7 rounds of subcultures on a non-selective medium, followed by transferring to the specific selective medium, Chlorella transformants were still able to grow on the selective medium. This outcome implies the stability of transgenes within C. sorokiniana AAGL G015. In another study, the use of Agrobacterium transformation for delivery of exogenous materials to Euglena gracilis demonstrated the stability of the mutants for over a year, through 12 rounds of cultivation. In contrast, other transformation methods such as biolistic bombardment, and electroporation techniques lacked stable integration of the transforming DNA into the host genome ( Chen et al., 2022 ). Have hypes and hopes by using this advancement technologies in dark-microalgal host? Tremendous breakthroughs in the new discovery of novel expression platforms for producing biopharmaceuticals or phytochemicals are needed. Heterotrophic microalgae, as a sustainable and scalable host for recombinant technology, hold promise. Certain microalgal species naturally possess pathways to synthesize vital substances for nutrition and food. Recent progress in microalgae-based product development has yielded various industrial applications, including enhanced textural and color properties, improved sensory quality, antioxidant capabilities, and elevated protein content. Additionally, microalgae represent a third-generation biofuel and energy source, benefitting from their short life cycle, environmental adaptability, and widespread distribution that align well with economic systems. However, for reasons of high cost and unavailability of genetic information for commercially suitable strains, they have not yet reached industrial maturity and commercial success. So far, a considerable effort has been given to tackle the bottleneck of various methods, including various nutritional-, environmental-, and physiological alteration of cultivation, metabolic and genetic engineering ( Pierobon et al., 2018 ; Chen and Lee, 2019 ). To satisfy the large market demand, a high level of technology and mechanized harvesting techniques are required. Future studies must explore the integration of new efficient technology of downstream processes including extraction, concentration, conversion, and purification of recombinant product from microalgae. It is also essential to establish microalgal host using genetically modified microalga for both endogenous and heterologous expression. To achieve economically efficient large-scale production and utilization of microalgae for production, sustainable and eco-innovative processing techniques are necessary. These techniques should efficiently transform raw microalgal biomass into value-added products or isolated ingredients without compromising on their nutritional and environmental benefits. Using microalgal hosts to produce plant substances offers advantages such as consistent synthesis without seasonal limitations and serves as a valuable tool for studying the involved precursors and enzymes of desired phytochemicals ( Koo et al., 2013 ; Eisohly et al., 2014 ). Choosing a suitable host for plant production requires careful consideration, especially with the projected increase in global demand for production along with zero waste. This marks a new phase for the alternative resources and tools, fostering innovation in this sector. Additionally, the ability to manipulate microalgae’s excretion system could open up possibilities for direct release of fuels or other metabolites into the medium, enhancing their potential as versatile bioproduction platforms. Microalgae also show promise in producing recombinant proteins of high industrial relevance due to their rapid growth cycle and cost-effective cultivation compared to other expression hosts ( Gramegna et al., 2020 ). These advancements shed light on the molecular aspects of algal phytochemical production and pave the way for optimizing microalgae as a platform for therapeutic and industrially significant recombinant protein production. As a ‘green’ alternative to existing mammalian, yeast, or bacterial systems, microalgae are poised to play a prominent role in the future of biotechnology and functional product industries."
} | 9,826 |
26888174 | PMC4758071 | pmc | 3,565 | {
"abstract": "A recurrent spiking neural network is proposed that implements planning as probabilistic inference for finite and infinite horizon tasks. The architecture splits this problem into two parts: The stochastic transient firing of the network embodies the dynamics of the planning task. With appropriate injected input this dynamics is shaped to generate high-reward state trajectories. A general class of reward-modulated plasticity rules for these afferent synapses is presented. The updates optimize the likelihood of getting a reward through a variant of an Expectation Maximization algorithm and learning is guaranteed to convergence to a local maximum. We find that the network dynamics are qualitatively similar to transient firing patterns during planning and foraging in the hippocampus of awake behaving rats. The model extends classical attractor models and provides a testable prediction on identifying modulating contextual information. In a real robot arm reaching and obstacle avoidance task the ability to represent multiple task solutions is investigated. The neural planning method with its local update rules provides the basis for future neuromorphic hardware implementations with promising potentials like large data processing abilities and early initiation of strategies to avoid dangerous situations in robot co-worker scenarios.",
"discussion": "Discussion The brain efficiently processes and predicts sequential activity patterns in the context of planning tasks 10 11 12 13 . Understanding these processes and how information is possibly encoded in stochastic neurons are major goals of theoretical neuroscience. In this paper, we demonstrated how recurrent spiking neural networks can solve planning problems and provide a solid theory based on the framework of probabilistic inference. The model reproduces observed neural dynamics, predicts that contextual information is one of the key modulating factors and is a promising low-energy control strategy for mobile robot applications. Theoretical contributions Spiking neural networks are a reasonable neuroscience model as verified by data 35 36 37 38 . Their capabilities in solving planning tasks however have been underexplored to date. It was shown that spiking networks can encode and draw samples from arbitrary complex distributions 39 40 , models of temporal sequences can be learned 7 8 and Bayesian filtering was studied 38 41 42 43 . Our model builds on these probabilistic sampling results and a solution to planning problems is suggested that approximates a stochastic process for planning through forward sampling from a parameterized model distribution. The parameters of the model distribution denote the synaptic weights of a population of afferent neurons for which local Hebbian update rules were derived from the principle of probabilistic inference. The derivations include arbitrary differentiable activation functions and postsynaptic potential shapes for the neuron model (details are provided in the supplement ). Links to expectation maximization 27 32 and policy gradient methods 28 29 30 were established, where the resulting offline learning rules are similar to Monte Carlo policy evaluation 25 and the network parameters subject to these updates converge to the globally optimal solution for WTA network dynamics. The online learning rules resemble the online Monte Carlo policy iteration algorithm and converge to local lower bounds of the optimum. The correctness of the neural planning method was validated in two toy tasks, a finite horizon planning task with a known optimal solution and an infinite horizon task, where the neural network achieved the same performance level as Monte-Carlo policy evaluation 25 . Implications for neuroscience Previous neural models that implement path planning have focused on attractor networks or potential fields 15 44 45 46 47 and their activity was related to hippocampal firing 17 18 . A deficit of these models is however that the path which was taken to reach a desired state cannot be modeled with attractors. To overcome this limitation a sequence of successive metastable states in attractor networks was proposed 19 but it is left unclear how these networks can be trained from rewards. We followed a different approach where attractors emerge through reinforcement of rewarding trajectories. As a result, different input neurons with its synaptic weights can model different routes to multiple attractors. Thus, the proposed model extends the modulation abilities of attractor networks and can be validated, e.g., in a study on contrasting planning of safe versus straight-line paths. We demonstrated in simulation results that the proposed recurrent neural network can reproduce the dynamically changing firing rates observed during hippocampal sweeps 10 . The input modulated activity in our network hypothesizes that a cognitive map representation (the recurrently connected state neurons) receives contextual input from other brain regions. Potential sites for these contextual inputs are projections from the entorhinal and the prefrontal cortex 48 49 . It is worth mentioning that the network is not limited to model hippocampal sweeps. It may be used to model frequently observed dynamically changing firing rates from other brain regions 50 51 52 53 . Embedded in the framework of probabilistic inference, the proposed network can be naturally extended in multiple ways, e.g., the place cells encoding the state transition model might be learned 17 , actions might be encoded additionally 34 , forward and backward replays 11 12 can be simulated, or multiple cognitive maps can be installed 18 . Furthermore, Poisson neurons were chosen for simplicity and the model generalizes to noisy integrate and fire neurons 54 . Implications for robotics State-of-the-art planning algorithms in robotics generate movement plans within seconds and scale to many degrees of freedom 55 56 57 . Spiking neural networks could not compete with these methods due to the encoding of continuous variables, e.g., in our robot experiment, we used population codes 58 59 to encode a two-dimensional continuous state variable and more than a few hundred neurons could not be simulated without risking to run out of memory in a standard computer. A potential solution that is currently under investigation are factorized population codes which scale but can not capture correlations (which are needed to avoid obstacles). Another promising alternative are neuromorphic chips which were already used to learn 62-dimensional joint angle sequences of human jump motions in recurrent networks 60 . In addition, related spiking network models were proposed which also build on winner-take-all circuits and local plasticity rules 20 21 22 . Therefore, it is reasonable to assume that the presented theory provides the basis for future neural controller implementations in neuromorphic hardware for robot control. In contrast to previous work on spiking neurons in a reinforcement learning framework 34 , we followed here a model-based approach where the recurrent dynamics of the network can be reused to learn multiple related tasks with different sets of weights from the context neurons (e.g., representing different goal positions or obstacles). The state transition model does not need to be re-learned when switching between environments. Furthermore, our model has the advantage that multi-modal solutions to planning tasks can be learned. This feature was exploited in an obstacle avoidance task in a real robot, where the network randomly sampled one out of two paths. This ability to encode non-linear mappings is in particular beneficial for learning forward and inverse kinematic models in robotics."
} | 1,934 |
35438241 | PMC9322558 | pmc | 3,566 | {
"abstract": "Abstract 5‐hydroxymethylfurfural (HMF) is produced upon dehydration of C6 sugars in biorefineries. As the product, it remains either in aqueous solutions, or is in situ extracted to an organic medium (biphasic system). For the subsequent oxidation of HMF to 2,5‐furandicarboxylic acid (FDCA), ‘media‐agnostic’ catalysts that can be efficiently used in different conditions, from aqueous to biphasic, and to organic (microaqueous) media, are of interest. Here, the concept of a one‐pot biocatalytic cascade for production of FDCA from HMF was reported, using galactose oxidase (GalOx) for the formation of 2,5‐diformylfuran (DFF), followed by the lipase‐mediated peracid oxidation of DFF to FDCA. GalOx maintained its catalytic activity upon exposure to a range of organic solvents with only 1 % (v/v) of water. The oxidation of HMF to 2,5‐diformylfuran (DFF) was successfully established in ethyl acetate‐based biphasic or microaqueous systems. To validate the concept, the reaction was conducted at 5 % (v/v) water, and integrated in a cascade where DFF was subsequently oxidized to FDCA in a reaction catalyzed by Candida antarctica lipase B.",
"conclusion": "Conclusion Galactose oxidase showed an excellent stability in a myriad of organic solvents, retaining its activity upon exposure to the solvents in biphasic systems or in media with reducing water contents (microaqueous). Therefore, the oxidation of HMF to DFF could be successfully established in ethyl acetate‐based biphasic/microaqueous systems with water contents as low as 1 % (v/v). The reaction was successfully scaled up to a 100 mL scale, with only a minor decrease in the conversion of HMF to DFF. Following this, the reaction was coupled to a second step, namely the CalB‐catalyzed synthesis of peracetic acid, which in situ oxidizes DFF to FDCA. Due to the low solubility of FDCA both in ethyl acetate and water, the products precipitated, simplifying the downstream processing. The herein presented approach showed relevance for biorefineries, because the use of ‘media‐agnostic’ biocatalysts enables the synthesis of FDCA in a broad range of conditions, from pure aqueous media to other systems without bulk water added. Given the many types of (crude) effluents expected in biorefineries, having these catalysts in hand may be of high importance, as they can be straightforwardly adapted to on‐demand conditions. Once the proof‐of‐concept has been shown, further optimization steps need to be considered to reduce reaction times and improve overall yields and productivities. Apart from fine‐tunning the different enzyme proportions and loadings, it must be noted that the concept has been demonstrated with a commercial wild type enzyme preparation of GalOx, and not with other variants (e.g., the M 3‐5 mutant), which has been reported to display a higher affinity towards HMF as the substrate. \n [12] \n Likewise, from the standpoint of reaction engineering, improving the mixing conditions, in particular, providing the optimal amount of oxygen, are aspects to be considered. When it comes to the overall cascade, the tendency of FDCA to precipitate due to its insolubility in the reaction medium opens new frontiers towards easier downstream processing. Overall, the versatility of the reaction media, together with straightforward FDCA recovery may be important assets for the implementation of the synthetic route in future biorefineries.",
"introduction": "Introduction 5‐hydroxymethylfurfural (HMF) is considered to be one of the future key building blocks from biorefineries. HMF can be produced from C6 sugars originating from lignocellulose (mostly glucose), upon a triple acidic dehydration. In some cases, glucose is first isomerized to fructose, as the dehydration of fructose proceeds in a more straightforward manner. Glucose from lignocellulose is typically found in water as a product of cellulose hydrolysis and thus, the produced HMF may remain in aqueous media or may be in situ extracted in an organic solvent (through a biphasic medium). Therefore, for the subsequent valorization of HMF, robust catalysts that show versatility and can perform reactions both in aqueous solutions and in (microaqueous) organic systems are highly desirable. Biocatalysis is intuitively considered as a technology to be implemented in aqueous media, since the majority of biochemical processes occur in water. However, with the rise of biocatalysis in the 1980s, and the necessity to explore the versatility of the application of enzymes in different industrial processes, the initiative to use enzymes in non‐conventional (non‐aqueous) media arose, and proved to be successful.[ \n 1 \n , \n 2 \n ] Advantages of the use of water‐free media in biocatalysis are the greater stability of enzymes, and the straightforward recovery of the catalyst. \n [3] \n In particular, the advantages of using water as a reaction medium are easily outweighed in the case of industrial processes, where enzymes typically catalyze non‐natural organic substrates which have a higher solubility in non‐aqueous media. \n [4] \n Although a minimum amount of water is necessary for the hydration of the enzyme molecule, this is relatively low and no bulk quantities of water are needed.[ \n 5 \n , \n 6 \n ] Non‐conventional media encompass a wide array of potential reaction systems, ranging from ionic liquids (ILs), deep eutectic solvents (DES), neat solvents, microaqueous reaction systems (MARS), and biphasic systems consisting of water and organic media. \n [7] \n Recent developments showed great potential of using enzymes in low‐water media, which became not limited to lipases. \n [8] \n When it comes to biorefineries, the possibility that enzymes can efficiently catalyze reactions in media containing different water proportions is clearly an asset, as it provides adaptation to varied (crude) effluent types. Galactose oxidase (GalOx, EC 1.1.3.9) gained momentum in recent years, mostly due to extensive research in the field of enzyme engineering. \n [9] \n GalOx catalyzes the oxidation of primary alcohols to their corresponding aldehydes, using molecular oxygen as a cosubstrate. \n [9] \n As in all other enzymes that need gaseous (co)substrates, the cosubstrate solubility in water (e.g., the solubility of O 2 is 0.25 m m at 20 °C) renders to reach high product titers. However, the solubility of gaseous substrates is higher in organic solvents than water, which opens up new process windows for the use enzymes in organic media. \n [10] \n Despite the potential, only a handful of examples showcasing GalOx in non‐conventional media have been reported so far.[ \n 11 \n , \n 12 \n ] GalOx immobilized on an epoxy resin showed exceptional tolerance to several neat organic solvents, when used for oxidation of 3‐fluorobenzyl alcohol. \n [11] \n Additionally, engineered variants of GalOx have been used in their free form in the presence of various different organic cosolvents in aqueous media. \n [12] \n \n The substrate scope of GalOx is broad, and includes sugar‐derived compounds, primary and benzyl alcohols, as well as HMF. Namely, HMF poses a great platform for the synthesis of different biobased and biomass‐derived chemicals, most prominently 2,5‐furandicarboxylic acid (FDCA). FDCA has been named one of the top twelve value‐added chemicals from biomass. \n [13] \n FDCA serves as a backbone of polyethylene furanoate (PEF), a material that poses a 100 % biobased alternative to conventional polyethylene terephthalate (PET) plastic. \n [14] \n \n Since HMF contains an alcohol group and an aldehyde group, the biosynthesis of FDCA from HMF requires three consecutive oxidation steps, which means the reaction can take place through a myriad of intermediate products. \n [15] \n A variety of processes for the oxidation of HMF to FDCA have been explored, including the use of electrocatalysis, photocatalysis, or catalysis using metal oxides. \n [16] \n \n However, biocatalysis remains the golden standard of sustainable green production. One of the first cascades, albeit not completely biocatalytic, was proposed in 2013, in which HMF was first oxidized by 2,2,6,6‐tetramethylpiperidine‐1‐oxyl (TEMPO) to obtain DFF, which was then oxidized to FDCA using peracetic acid formed in situ with Candida antarctica lipase B (CalB). \n [17] \n A combination of magnetic lipase and TEMPO as the mediator was successfully used to oxidize HMF to FDCA via 5‐formyl‐2‐furoic acid (FFA) at ambient temperature and pressure. \n [18] \n Most recently, the aforementioned chemoenzymatic approach was also combined with whole‐cell biocatalysis using Trichoderma reesei filamentous fungi in order to obtain FDCA using 5‐hydroxymethyl‐2‐furan carboxylic acid (HMFCA) as the intermediate product. \n [19] \n In another study, whole‐cell biocatalysis was employed in a cascade reaction to transform HMF to FDCA using vanillin dehydrogenase (VDH1) and HMF/furfural oxidoreductase (HmfH) co‐expressed in Escherichia coli . \n [20] \n The possibility of using a single enzyme to catalyze the oxidation of HMF to FDCA was explored using the relatively recently discovered 5‐hydroxymethylfurfural oxidase (HMFO). \n [21] \n The same enzyme was used to convert HMF to FFA, which was then transformed into FDCA using CalB. \n [22] \n \n Pseudomonas putida strain S12 expressing HMFO was used for efficient whole‐cell biotransformation of HMF to FDCA. \n [23] \n A tandem reaction combining Escherichia coli and wild‐type Pseudomonas putida KT2440 resulted in a full conversion of HMF with high DFF yields. \n [24] \n In a different approach, as HMF poses an inhibitor in the production of biofuels from lignocellulose waste, immobilized cells of Burkholderia cepacia H‐2 were used to remove HMF from the process by transforming it to FDCA. \n [25] \n Another proposed strategy was to use an aryl‐alcohol oxidase (AAO) to transform HMF to FFA, subsequently transforming it to FDCA using an unspecific peroxygenase. \n [26] \n Likewise, a one‐pot cascade reaction using GalOx, horseradish peroxidase (HRP) and periplasmic aldehyde oxidase (PaoABC), with the same intermediate product was also established. \n [27] \n Overall, the biocatalysis community has been witnessing great progress in the conversion of HMF to FDCA in different reaction systems via enzymatic or chemoenzymatic methods using isolated enzymes or whole‐cells. Although numerous biocatalytic synthesis routes to FDCA have been proposed, for their scalability further optimizations related to productivity and cost‐effectiveness are typically needed. \n [16] \n The tandem reaction utilizing GalOx and CalB for FDCA formation has been reported in the past in aqueous solutions. While being promising, the use of pure water‐based media restricts its use to only aqueous effluents from biorefineries, and furthermore, it obliges to the separation of intermediate products, due to the incompatibility between the two reaction steps (oxidase catalysis and lipase‐mediated peracid oxidations). \n [28] \n \n Given the needs of providing highly integrated systems for biorefinery, in which crude media – containing more or less water and impurities – can be used, herein, we propose a three‐step bio‐bio‐chemocatalytic reaction cascade for the synthesis of FDCA (Scheme 1 ). The first step of the reaction is based on the oxidation of 5‐hydroxymethylfurfural (HMF) to 2,5‐diformylfuran (DFF) using GalOx. The enzyme uses molecular oxygen (O 2 ) as the cosubstrate, yielding hydrogen peroxide (H 2 O 2 ) as the by‐product. The H 2 O 2 will be taken up by the second enzyme, CalB, which will utilize it to convert ethyl acetate into peracetic acid. \n [17] \n The in situ formed peracetic acid will then oxidize DFF to FDCA. Scheme 1 Conceptual enzymatic cascade for the synthesis of 2,5‐furandicarboxylic acid (FDCA) from 5‐hydroxymethylfurfural (HMF) through 2,5‐diformylfuran (DFF) using galactose oxidase (GalOx) and Candida antarctica lipase B (CalB). In the case of chemoenzymatic cascades, the use of a biphasic or microaqueous system brings the best of both worlds: (i) water for the preference of enzymes, as well as (ii) an organic solvent for the preference of the hydrophobic substrates.[ \n 29 \n , \n 30 \n , \n 31 \n ]",
"discussion": "Results and Discussion The primary focus of this study was to explore the ‘medium engineering’ scope for GalOx catalysis, with the aim of facilitating a smooth transition between the two steps of the reaction cascade, as well as potentially opening the possibility of establishing a one‐pot cascade. The solvents used were chosen due to their different chemical structures, as well as physicochemical characteristics such as boiling point, water solubility, and hydrophobicity (log P) (Table S1). \n [32] \n Cyclopentyl methyl ether (CPME) was chosen as the environmentally friendly and green(er) alternative to traditional solvents, which greatly contributes to the broader picture of establishing the production of biobased plastics.[ \n 33 \n , \n 34 \n ] The motivation behind using ethyl acetate originated from the goal of using it as both a substrate and solvent component in potential following steps of the cascade. \n [17] \n Additionally, lipases, catalyzing the second reaction step, exhibit the phenomenon of interfacial activation, meaning that they show higher activity in biphasic systems in comparison to pure aqueous conditions. \n [35] \n \n Effect of biphasic systems on galactose oxidase activity Medium engineering can eliminate the need for extensive downstream processing for the isolation of intermediates. For this purpose, it was first necessary to assess the potential effects of organic solvents on the catalytic performance of GalOx. As pure water‐based media have proven successful (although with downstream units), biphasic systems with 50 % (v/v) of phosphate buffer were set as a starting point, and the range was consequently expanded with biphasic systems of lower water contents down to 1 % (v/v), falling into the category of microaqueous systems (without bulk water quantities). \n [19] \n The setup was analyzed with the preselected water‐immiscible organic solvents of different physicochemical characteristics, with three of the most representative shown in Figure 1 .\n Figure 1 Residual activities of GalOx in biphasic systems with 50, 20, 10 and 5 % (v/v) of aqueous phase and cyclopentyl methyl ether (CPME, a), dodecane (b), and ethyl acetate (EtOAc, c) at 1200 rpm and 25 °C measured over the course of 72 h using 100 m m D‐galactose as the substrate. The residual activity is expressed relative to the initial activity of GalOx in 100 m m NaPi buffer at pH 7.4 and time zero. The experiments were performed in biological triplicates. The presence of organic solvents in the system did not display a significant effect on the residual activity of GalOx over the course of 72 h. CPME, dodecane, and EtOAc are solvents with vastly different chemical structures, water miscibility and log P value. However, regardless of the values of the aforementioned parameters, the residual activities remained above the 80 % mark in most cases. No visible correlation was found between the obtained results and the properties of the organic solvents, considering the physicochemical properties and their chemical structures. Moreover, the results corroborate that the log P value cannot be used as a sole criterion to predict biocompatibility of the solvent, given the complexity of establishing enzymatic reactions in biphasic media. \n [36] \n The ten‐fold reduction of the aqueous content in the biphasic system from 50 % (v/v) to 5 % (v/v) did not result in further loss of activity, as was the case for all investigated solvents. Effect of microaqueous systems on galactose oxidase activity Encouraged by the obtained positive results even at water content as low as 5 % (v/v), the subsequent reduction of the aqueous phase led the study into the area of microaqueous systems with 1 % (v/v). The residual activity was monitored for 48 h, and the obtained results were highly comparable to those obtained at the same timepoint with the other biphasic systems. The enzyme kept the highest activity in toluene, and the lowest in CPME, although residual activities higher than 70 % were recorded in all cases (Figure 2 ).\n Figure 2 Residual activity of GalOx in microaqueous (two‐liquid‐phase) systems with various organic solvents and 1 % (v/v) of aqueous phase, at 1200 rpm and 25 °C, measured after 48 h using 100 m m D‐galactose as the substrate. Effect of neat solvents on galactose oxidase activity Inspired by the results, the GalOx stability was also evaluated in neat as well as in water‐saturated solvents, but no catalytic activity could be observed. It must be noted, however, that the exposure to solvents does not affect the activity of the enzyme, as shown in previous works where both free‐ and immobilized GalOx exhibited significant tolerance towards organic solvents.[ \n 11 \n , \n 12 \n ] To corroborate this, upon 20 h of exposure to the aforementioned solvents, buffer was added to the reaction vessels. Thus, the systems were reconstituted to biphasic systems with 50 % (v/v) aqueous phase, and the activity was measured after two hours of equilibration (Figure 3 ).\n Figure 3 Residual activity of GalOx upon exposure to neat organic solvents for 20 h (monophasic system), and 2 h after subsequent reconstitution to biphasic systems with 50 % (v/v) aqueous phase at 1200 rpm and 25 °C using 100 m m D‐galactose as the substrate. The control experiment represents the enzyme without exposure to organic solvents. In some cases, the exposure to neat solvents and regeneration of GalOx in 50 % (v/v) H 2 O biphasic system led to a remarkable increase in activity. A similar observation was made for decarboxylation of ferulic acid, where the enzyme exhibited higher stability (3.5‐fold increase in half‐life time) in biphasic systems compared to aqueous media. \n [37] \n Importantly, the results confirm that the exposure to organic solvents does not have a detrimental effect on the enzyme, and that the enzyme retains sufficient activity to be used for subsequent catalysis. Oxidation of HMF to DFF in EtOAc‐based biphasic and microaqueous media As previously underlined, the actual motivation for exploring the stability of GalOx in the presence of organic media is to use the enzyme in a multistep chemo‐enzymatic reaction for the synthesis of FDCA (Scheme 1 ). Therefore, the oxidation of 1 m m HMF to DFF was systematically studied in EtOAc based biphasic systems of various water contents (at 1 mL scale), using GalOx, with addition of auxiliary enzymes: horseradish peroxidase and/or catalase (Table S2). HRP was added as an activator, whereas catalase was added to remove H 2 O 2 . The highest conversion of HMF was 58.6±3.6 %, achieved when all three enzymes were used in pure buffer. The conversions in biphasic and microaqueous systems with ethyl acetate were lower, however, the reduction in the aqueous phase content from 50 % (v/v) to 1 % (v/v) did not result in a significant reduction in conversion, providing operational options for the water‐free biocatalytic cascade to FDCA. Catalysis using GalOx requires a careful interplay of different process parameters. The active site of GalOx contains a tyrosine radical bound to a copper(II) ion, which is reduced to a nonradical tyrosine copper(I) complex during oxidation of the substrate. \n [38] \n This tyrosine radical can also undergo a reduction to a nonradical copper(II) complex, which is inactive. \n [39] \n Therefore, the enzyme requires a single‐electron oxidation in order to regenerate the active site. \n [39] \n Although extensive studies have been made to find alternative activator molecules, such as potassium ferricyanide (K 3 [Fe(CN) 6 ]), the most commonly used activator remains horseradish peroxidase (HRP). \n [40] \n Based on the results shown in Table 1 , the addition of HRP indeed had a positive effect on the reaction yield, especially noticed in the reaction performed in pure buffer, both when added to GalOx alone, or in presence of catalase as well.\n Table 1 Conversion of HMF to DFF in EtOAc‐based biphasic systems with 50, 20, 10, 5, and 1 % (v/v) of 0.1 m NaPi buffer at pH 7.4, as well as pure buffer using GalOx and HRP and/or catalase as auxiliary enzymes. \n Added aqueous phase [% (v/v)] \n \n Conv. [a] [%] \n \n Conv. [b] [%] \n \n Conv. [c] [%] \n \n Conv. [d] [%] \n \n 1 \n \n 2.6±0.1 \n \n 4.6±0.4 \n \n 3.7±1.0 \n \n 1.9±0.2 \n \n 5 \n \n 6.7±0.4 \n \n 11.5±1.7 \n \n 8.0±0.7 \n \n 6.9±1.6 \n \n 10 \n \n 7.6±0.6 \n \n 11.5±0.7 \n \n 7.6±0.2 \n \n 5.6±1.8 \n \n 20 \n \n 8.8±1.5 \n \n 15.6±0.4 \n \n 10.0±0.5 \n \n 6.6±1.0 \n \n 50 \n \n 15.0±0.8 \n \n 26.1±0.2 \n \n 16.5±1.6 \n \n 9.2±0.4 \n \n 100 \n \n 24.9±1.4 \n \n 36.7±0.6 \n \n 33.0±1.8 \n \n 25.6±1.3 \n [a] 2 mg mL −1 GalOx, 1 mg mL −1 HRP, 1 mg mL −1 catalase. [b] 4 mg mL −1 GalOx, 1 mg mL −1 HRP, 1 mg mL −1 catalase. [c] 4 mg mL −1 GalOx, 2 mg mL −1 HRP, 1 mg mL −1 catalase. [d] 4 mg mL −1 GalOx, 1 mg mL −1 HRP, 2 mg mL −1 catalase. All reactions were performed at 25 °C, 200 rpm for 72 h and subsequently analyzed using HPLC. All reactions were performed in biological triplicates. Wiley‐VCH GmbH Hydrogen peroxide (H 2 O 2 ) is the by‐product of GalOx‐catalyzed oxidation, however, it has been shown to both inhibit and deactivate the enzyme. \n [41] \n Namely, hydrogen peroxide is not soluble in organic media, which means that its full amount is confined to the aqueous phase. Hence, the concentration of H 2 O 2 increases with the decrease in water content in the biphasic system. In order to alleviate this, catalase was added to the reaction to dismutate H 2 O 2 into H 2 O and O 2 . \n [38] \n The addition of catalase alone resulted in much lower DFF yields, in the case of pure buffer as a reaction medium. Interestingly, in the case of biphasic systems the obtained yields with the addition of catalase alone, were lower than those achieved in catalase‐free systems, or where it was present along with HRP. Oxygen serves as a cosubstrate for GalOx, and it is therefore necessary to facilitate enough O 2 in the system. Assuming the saturation of the enzyme with the substrate, as well as atmospheric pressure conditions, the reaction rate will be almost directly proportional to the oxygen concentration. \n [38] \n Although some oxygen is provided, or more precisely, can be recycled from the breakdown of H 2 O 2 by catalase, it is necessary to provide additional oxygen for the process. In higher scale reactors this can easily be achieved by bubbling.[ \n 38 \n , \n 42 \n ] In low volume reactions, one possible solution is to maximize the headspace volume to reaction volume ratio within the reaction vessel. \n [32] \n \n Taking all these aspects into account, a combination of galactose oxidase, horseradish peroxidase, and catalase resulted in the highest HMF conversions when comparing the results obtained in pure buffer (Table 1 ). Therefore, based on the screening experiments, further optimization of the reaction ensued. In experiments that followed, all three enzymes were used in all trials, however, at different concentrations, in order to assess the optimal ratio of the enzymes. A comparison of the results obtained with 2 mg mL −1 GalOx, 1 mg mL −1 HRP, and 1 mg mL −1 catalase revealed that the increase in headspace, and therefore, oxygen availability in the system, resulted in a higher observed conversion of HMF to DFF in the case of all biphasic systems. An interesting result can be seen in the case of the reaction performed in 100 % buffer, where the introduction of the headspace volume resulted in a nearly halved yield to the previously recorded result. When comparing the results obtained with different enzyme combinations, the achieved results are significantly better in all four cases, in comparison with the initial results. The combination of 4 mg mL −1 GalOx, 1 mg mL −1 HRP, and 1 mg mL −1 catalase faired best, resulting in the highest DFF yields in all biphasic systems, as well as pure buffer. Scale‐up towards FDCA synthesis Once the operational conditions were set, the subsequent step was performing the reaction at preparative conditions. For this, the reaction was performed using the enzyme concentrations that showed the most promising results (the combination of 4 mg mL −1 GalOx, 1 mg mL −1 HRP, and 1 mg mL −1 catalase). With the goal of minimizing the water content, but at the same time maintaining sufficient productivity of the reaction, the biphasic system of choice was ethyl acetate containing 5 % (v/v) of buffer. The total reaction volume was 0.1 L and it was performed in 2 L glass bottles in order to keep roughly the same liquid‐to‐headspace ratio at 150 rpm. Close examination of the scaled‐up reaction revealed that, contrary to extensive literature search, the oxidation of HMF to DFF is much faster than anticipated (∼20 % yield in 3 h), with the majority of DFF produced in the initial few hours (Figure S2). A decrease in the residual activity was observed with the increase of the concentration of DFF in a biphasic system (Figure S3). Dialdehydes, such as DFF, can interact with the surface amino acids of the enzyme. The most prominent example is glutaraldehyde, a crosslinking molecule commonly used for enzyme immobilization.[ \n 43 \n , \n 44 \n ] The DFF partitioning to the aqueous phase was also noticed, and it was especially visible at higher concentrations. At 30 °C its saturation concentration (maximum solubility) in water is approximately 15 g L −1 (∼120 m m ), whereas at temperatures below 20 °C DFF tends to precipitate, and therefore can easily be separated from HMF and removed from the aqueous solution. \n [45] \n \n Upon 72 h, CalB was added directly into the reaction, along with H 2 O 2 (stepwise) and the reaction temperature was elevated to 40 °C. Remarkably, upon 24 h of reaction, the two phases disappeared, and a uniform liquid phase was formed, together with a white precipitate, which was confirmed to be FDCA by HPLC analysis. FDCA has a relatively low solubility in water and ethyl acetate in comparison to other solvents. \n [46] \n Moreover, the acidic conditions (pH 3) facilitated by the presence of (per)acetic acid caused FDCA to precipitate as also documented in the literature. \n [47] \n \n Thus, the herein established bio‐bio‐chemocatalytic concept may provide useful operational options for straightforward downstream processing when using the biphasic system or the microaqueous approach (Scheme 2 ). Scheme 2 Conceptual approach for the biphasic (or microaqueous) system for the synthesis of FDCA from HMF."
} | 6,652 |
36802247 | null | s2 | 3,567 | {
"abstract": "Caldicellulosiruptor species are proficient at solubilizing carbohydrates in lignocellulosic biomass through surface (S)-layer bound and secretomic glycoside hydrolases. Tāpirins, surface-associated, non-catalytic binding proteins in Caldicellulosiruptor species, bind tightly to microcrystalline cellulose, and likely play a key role in natural environments for scavenging scarce carbohydrates in hot springs. However, the question arises: If tāpirin concentration on Caldicellulosiruptor cell walls increased above native levels, would this offer any benefit to lignocellulose carbohydrate hydrolysis and, hence, biomass solubilization? This question was addressed by engineering the genes for tight-binding, non-native tāpirins into C. bescii. The engineered C. bescii strains bound more tightly to microcrystalline cellulose (Avicel) and biomass compared to the parent. However, tāpirin overexpression did not significantly improve solubilization or conversion for wheat straw or sugarcane bagasse. When incubated with poplar, the tāpirin-engineered strains increased solubilization by 10% compared to the parent, and corresponding acetate production, a measure of carbohydrate fermentation intensity, was 28% higher for the Calkr_0826 expression strain and 18.5% higher for the Calhy_0908 expression strain. These results show that enhanced binding to the substrate, beyond the native capability, did not improve C. bescii solubilization of plant biomass, but in some cases may improve conversion of released lignocellulose carbohydrates to fermentation products."
} | 392 |
33428722 | PMC7878175 | pmc | 3,570 | {
"abstract": "ABSTRACT Microbial community engineering aims for enrichment of a specific microbial trait by imposing specific cultivation conditions. This work demonstrates that things may be more complicated than typically presumed and that microbial competition can be affected by seemingly insignificant variables, like in this case the type of acid used for pH control. Aerobic bioreactors pulse fed with acetate operated with hydrochloric acid resulted in the enrichment of Plasticicumulans acidivorans , and changing the pH controlling agent to sulfuric acid shifted the community towards Zoogloea sp . Further research demonstrated that the change in community structure was not directly caused by the change in acid used for pH control, but resulted from the difference in corrosive strength of both acids and the related iron leaching from the bioreactor piping. Neither system was iron deficient, suggesting that the biological availability of iron is affected by the leaching process. Our results demonstrate that microbial competition and process development can be affected dramatically by secondary factors related to nutrient supply and bioavailability, and is way more complex than generally assumed in a single carbon substrate limited process.",
"conclusion": "CONCLUSIONS Aerobic, pulse fed sequencing batch bioreactors allow efficient enrichment of microbial communities with superior PHA storing capacity. This work has demonstrated that minute differences in medium composition may strongly affect microbial competition and therewith affect the PHA producing capacity significantly. Here we elucidated that the type of acid used for pH-control affected the bioavailability of iron and therewith determined the microbial community structure and the PHA producing capacity: Even though no iron limitation was observed in any of the systems, the titration of additional iron through corrosion of reactor inlets facilitated the enrichment of the well-known PHA producer Plasticicumulans acidivorans . By changing the hydrochloric acid for sulfuric acid as pH controlling agent, the corrosion of acid inlet points of the reactor halted and an immediate change in functional response was observed, followed by a change in microbial community towards a Zoogloea sp . dominated culture. The results described in this work demonstrate that apparently insignificant variations in medium composition can induce secondary nutrient limitations and have a major impact on the functional and structural development of microbial enrichments.",
"introduction": "INTRODUCTION Microbial community engineering (MCE) utilizes ecological selection principles to enrich microbial communities with specific functional properties, e.g. the production of chemicals and bioenergy (Kleerebezem and van Loosdrecht 2007 ). MCE finds its roots in the work of Baas Becking: ‘Everything is everywhere, but the environment selects’ (Baas-Becking 1934 ). By using selective environments, we aim to enrich and maintain microbial communities with desired functionalities under non-axenic, i.e. open, conditions. MCE can contribute to the circular economy and valorize non-sterilized mixed substrate streams, thereby unlocking the tremendous carbon and energy resources originating from heterogeneous feedstocks currently regarded as waste streams. In general, laboratory enrichment studies are designed with the idea in mind that one substrate is present in the influent in a rate determining concentration. Typically, in selective conditions favoring heterotrophic growth, the limiting substrate is the carbon and energy source. All other essential growth nutrients are supplied in excess with the objective to characterize the process as a function of a single substrate limitation. In this way, microbial competition is assumed to be investigated as a function of a single variable, and conclusions can be drawn in terms of the dependency of system development on this variable. For example, the competition in chemostat enrichment experiments generally is assumed to be determined by the affinity for one limiting substrate. To which extent this assumption holds true is rarely verified due to the large number of medium constituents that would need to be tested. Nevertheless, it remains largely unclear if microbial ecosystem development depends on the concentration and bioavailability of secondary substrates such as trace elements. In this work, we describe our analysis of an unanticipated secondary limitation encountered in experiments aiming for enrichment of a polyhydroxyalkanoates (PHA) producing microbial community. PHA is a polymer with chemical properties that make it an interesting bioplastic that is fully biodegradable (Chen 2009 ; Tamis et al . 2018 ). Enrichment of PHA producing microorganisms can be established by aerobic cultivation in alternating presence and absence of the carbon substrate. Over the past 10 years, this strategy was shown repeatedly to enable the effective enrichment of the superior PHA-producer Plasticicumulans acidivorans from sewage sludge, and it has been elemental to mixed culture PHA research (Johnson, Kleerebezem and van Loosdrecht 2009 ; Jiang et al . 2011 ; Tamis et al . 2014 ; Marang, van Loosdrecht, and Kleerebezem 2016 ; Stouten et al . 2019 ). Typically, the enrichment of P. acidivorans from activated sludge can be established within 30 generations (Stouten et al . 2019 ). In a new attempt to enrich for P. acidivorans we operated a Sequencing Batch Bioreactor (SBR) for more than 60 generations. Although conditions known to enable effective enrichment of P. acidivorans were applied, no enrichment of P. acidivorans was established and the functional performance in terms of substrate conversion rates was markedly different. The sole operational difference with previous systems was the choice of acid used for pH control; Previous enrichments were conducted with HCl, while in this enrichment H 2 SO 4 was used. Based on this observation we decided to investigate in more detail the role of the type of acid used for pH control in the process, with the objective to identify the secondary factors that determine enrichment and functional process development.",
"discussion": "DISCUSSION Microbial competition is affected by the type of acid used for pH control In this work we have demonstrated that in a pulse fed aerobic bioreactor fed with acetate as sole carbon and energy source, the type of acid ( \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{{H_2}S{O_4}}}$\\end{document} and \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} ) used as pH controlling agent has a paramount impact on the functional properties of the process and the microbial community established in steady state (Fig. 1 ). Switching the pH controlling agent resulted in the functional and microbial transition to the alternating state: \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{{H_2}S{O_4} \\to HCl}}$\\end{document} became equivalent to \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} and vice versa. The reproducibility of these remarkable transitions was verified over twenty times with biomass from both \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} and \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{{H_2}S{O_4}}}$\\end{document} . The changes in functional performance of the microbial communities throughout all transition experiments were highly comparable, which further emphasizes the dependency of the process on the acid used for pH control. Microbial competition is directly affected by iron leaching due to corroding reactor inlets Additional experiments demonstrated that the type of anion supplied with the acid used for pH control only had a minor impact on the enrichment. Apparently, the differences observed were an indirect effect of the type of acid used for pH control. Abiotic experiments demonstrated that steel from the bioreactor corroded in HCl, and not in H 2 SO 4 resulting in increased concentrations of iron, chromium and nickel (Fig. S2, Supporting Information). Independent titration experiments with these main constituents of 316L steel, demonstrated that supplementing \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{{H_2}S{O_4}}}$\\end{document} with chromium and nickel did not enable the establishment of a process equivalent to \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} , nor the functional enrichment of P. acidivorans (Fig. 5 ). The presence of chromium was correlated with minor shifts in functionality and morphology and could therefore be a contributing factor to the microbial competition as observed in the cultivations (Lemire, Harrison and Turner 2013 ). The titration of iron did result in a steady state operational performance of \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{{H_2}S{O_4}}}$\\end{document} fully equivalent to \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} and enrichment of P. acidivorans . This led to the remarkable conclusion that some form of iron limitation was preventing the enrichment of P. acidivorans even though less than 50% of iron in the original medium was consumed. Iron bioavailability and its role in microbial competition From the results in this research it was not apparent which factors affect the biological availability of iron for P. acidivorans . Iron is the most important micro nutrient for almost all microorganisms, and due to iron's complex speciation and precipitation properties microorganisms have evolved to scavenge iron at very low concentrations (Lankford and Byers 1973 ). Natural ecosystems show dissolved iron concentrations below 1 nM (Falkowski 1998 ). Neilands ( 1981 ) describes how microbes with high iron affinity due to siderophores and cognate transport apparatus still grow optimal at iron concentrations below 0.1 µM. The genome of P. acidivorans contains several high-affinity iron transporters (e.g.: catecholate siderophore receptor: PWV65540.1, iron ABC transporter: PWV65654.1, ferrous iron transporter: PWV63195.1), making it unlikely that the measured iron Fe(III) concentrations (>9 µM) are limiting its growth rate (Göker 2017 ). Although iron was added to the medium as Fe(II), it oxidizes to Fe(III) in aerobic conditions, resulting in complex speciation (Davison and Seed 1983 ). EDTA was added as chelating agent to the medium to prevent precipitation of iron salts and oxides. The high binding strength of the EDTA-Fe(III) complex (k f_Fe(III) \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}$ \\approx $\\end{document} 25) reduces the free iron concentration to the order of 10 −29 M Fe(III) in the bioreactors, possibly affecting the iron uptake rate (Anderegg 1977 ). Additionally, the method of trace metal dosing was shown to influence microbial functionality in anaerobic digestion, where continuous titration at low concentrations achieved higher rates than pulse dosing excessive amounts (Gonzalez-Gil, Kleerebezem and Lettinga 1999 ). The oxidation state of iron, its complexation with chelating agents, and the manner in which iron is added to the medium, i.e. through pulse or titration, was demonstrated to play a key role in microbial competition and enrichment in this work. Transition state—altered functional performance of the microbial community. \n \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${\\textit{SBR}_{HCl}}$\\end{document} enrichments dominated by P. acidivorans showed an almost instantaneous change in functional performance when the acid agent was switched to H 2 SO 4 (Fig. 3 ). Despite the change in functional performance within 1 to 2 generations, the change in abundance of the dominant microbial community members seemed to take at least 10 generations as shown through 16S rRNA gene sequencing and microscopy (Fig. 2 ). The perceived change in function is related to microbial abundance and activity, here it is likely that the activity of the dominant community members changed. During the first cycles after the acid switch, the functional performance of P. acidivorans dominated enrichments showed high similarity to enrichments dominated by Zoogloea sp . The apparent decrease in the biomass specific substrate uptake rate upon the transition from HCl to H 2 SO 4 as pH controlling agent diminishes the competitive advantage of P. acidivorans in the experiments described in this work, allowing the minor community members to increase in abundance during the consecutive cycles. Pure culture cultivations with P. acidivorans and Zoogloea sp . might aid in unraveling the specific mechanisms behind the current observation by looking at transcriptomics and proteomic changes under different iron limitations. These observations evidently raise the question to what extent we may overlook other limitations and corresponding differences in functional distinct behavior due to choices in bioreactor operation and medium composition. Secondary limitation may have an important impact on microbial community structure and functioning—The environment selects From these experiments it becomes apparent that seemingly negligible operational differences may impact enrichment results. The results as observed in this study raise the question to which extent our assumption of single limiting factors in enrichment studies can be supported. Starting from the hypothesis of Baas Becking ‘everything is everywhere; but the environment selects’ (Baas-Becking 1934 ), the question arises how much diversity and alternative functionalities can be unlocked and are currently overlooked by the nuances of enrichment studies. Many microbiologists have anecdotal evidence where their tricks of the trade allowed them to cultivate and isolate specific microbial species. In most microbiology literature, minute differences often go unnoticed or undocumented, possibly hampering the elucidation of novel biological mechanisms. Some noteworthy exceptions include the works of Zeikus and Thauer in their respective labs on Methanobacterium thermoautotrophicum (Zeikus and Wolfe 1972 ). Eventually the difference in growth rate measured in both laboratories was related to the nickel in the needles in Zeikus’ lab (Schönheit, Moll, and Thauer 1979 ). And a more recent publication from the group of Op den Camp explains the growth dependencies of Methylacidiphilum fumariolicum on rare earth metals, which were present in medium supplemented with mudpot water from Solfatara (Pol et al . 2014 ). The above-mentioned findings resulted from close observation and critical analysis of operational practices. In order to facilitate the discovery and understanding of secondary factors affecting enrichment studies, more systematic and comparative research is required."
} | 4,297 |
28246634 | PMC5298852 | pmc | 3,572 | {
"abstract": "Both habitat diversity and species diversity are important factors in ecosystem productivity and nutrient cycling.",
"introduction": "INTRODUCTION Ecosystems around the globe are facing habitat homogenization due to human activities ( 1 – 3 ). Homogenized ecosystems will reduce the diversity of species and consequently diminish valuable ecosystem functions and services ( 4 ). In addition to affecting species diversity, reduced habitat diversity may also directly impair ecosystem functioning because of reduced cross-habitat exchanges of material and energy. Understanding the functional consequences of biodiversity loss across multiple scales of organization, that is, at both habitats and species levels, is thus critical. However, studies on the functional consequences of changes in biodiversity have been solely designed to investigate the effects of diversity among and within species (taxonomic or functional) ( 4 ). The effects of changes in habitat diversity on ecosystem functioning remain unexplored. Here, we scale up the framework for consideration of biodiversity and ecosystem functioning by investigating whether ecosystems comprised of a diversity of habitats have higher levels of functionality than ecosystems with low habitat diversity ( Fig. 1 ). Our experiment directly addresses the issue of potential effects of habitat homogenization on ecosystem functioning. It merges theory on biodiversity and ecosystem functioning ( 4 ) with landscape ecology ( 5 ) and thus incorporates the concept of meta-ecosystems ( 6 ). We propose that, like interspecies interactions, habitats can facilitate each other, and that ecosystem-wide functioning is promoted by habitat complementarity. Complementarity among habitats includes the exchange of material and energy, such as various forms of oxidants and reductants (for example, oxygen and organic matter). The effect of habitat diversity can be direct but also indirect via changes in species diversity ( Fig. 1 ). One example of interactions among habitats is the interplay between mangroves and coral reefs in tropical ecosystems ( 7 ). Mangroves serve as nursery habitats that affect the community structure and biomass of fish on neighboring coral reefs ( 8 ), which, in turn, may protect mangrove by functioning as a breakwater ( 9 ). Similar positive interactions may be common in ecosystems that are composed of a mosaic of different habitats. Fig. 1 Conceptual diagram of our framework. Ecosystem homogenization (caused by, for example, human disturbance) results in a change in habitat diversity ( A ). Because habitats have different physical and chemical characteristics, they are likely associated with different sets of species. Loss of habitat diversity thus potentially leads to loss in species diversity (the union of the species in all habitats, indicated by different symbols) ( B ). Changes in habitat diversity can affect ecosystem functioning not only directly through changes in structural complexity and the cross-habitat exchange of nutrients and other resources ( C ) but also indirectly via changes in species diversity. Natural ecosystems perform many ecosystem functions simultaneously; all of these functions have the potential to be positively or negatively affected by biodiversity ( 10 ). To quantify the overall effect of habitat diversity on ecosystem functioning across a range of functions, we therefore used an index of multifunctionality ( 10 ), which summarizes the focal ecosystem functions that are examined in our study. Because habitat diversity generally favors species diversity ( 11 ), the net effect of habitat diversity on ecosystem multifunctionality was partitioned into direct and indirect effects using structural equation modeling (SEM). Moreover, the relationships between habitat diversity, species diversity, and ecosystem multifunctionality may differ across seasons. Therefore, we also tested how changes in habitat diversity affect ecosystem multifunctionality during different seasons. As model systems, we used experimental ecosystems consisting of four different natural habitats common in coastal marine ecosystems: cyanobacterial mats, plant meadows (the seagrass Ruppia maritima ), silty mud, and sandy beach. Cores with sediment and overlying water from each habitat were sampled in the field and arranged randomly into ecosystems to form a diversity gradient, including one, two, three, or four habitat types. This setup allowed for interactions between and within the habitats via a common water column. Given the importance of shallow-water coastal systems in terms of productivity and nutrient cycling ( 12 ), we measured four key biogeochemical processes that describe the productivity and nitrogen cycling in these ecosystems: gross primary production (GPP), nitrogen fixation, denitrification, and uptake of dissolved inorganic nitrogen (DIN). We hypothesized that higher habitat diversity within an ecosystem (i) directly enhances ecosystem multifunctionality and (ii) increases bacterial and microalgal diversity, and that (iii) an increase in bacterial and microalgal diversity also increases multifunctionality. To examine whether these hypotheses are valid across seasons, we repeated the experiment in spring, summer, and autumn, representing different combinations of light, temperature, and concentrations of organic matter and inorganic nutrients.",
"discussion": "DISCUSSION In landscape ecology, spatial heterogeneity is a central and causal factor of ecological systems. The spatial arrangement of different elements in a landscape matrix affects fluxes of energy and material ( 5 ), but the field of biodiversity–ecosystem functioning has not yet incorporated this organizational aspect into experimental investigations. The importance of scale and spatial heterogeneity has been discussed in verbal ( 14 , 15 ) and theoretical ( 6 ) frameworks, but a concept for experimental tests has not been described. Our experiment demonstrates that habitat diversity directly and indirectly drives ecosystem multifunctionality and that habitat homogenization can threaten ecosystem multifunctionality beyond the loss of species diversity. We propose that our general framework ( Fig. 1 ), in which we scale up from the level of species to the level of the entire habitats, provides a profitable way forward if net ecosystem consequences of environmental change are to be modeled and managed. In our experiment, multifunctionality during summer was larger than the expected sum of all individual habitats, with the highest habitat diversity showing an observed level of multifunctionality that was 66% higher than expected from single habitat treatments ( Fig. 6E ). This is the equivalent to overyielding at the level of species; that is, polycultures are more productive than monocultures due to positive interactions, such as complementarity ( 13 ). Habitats can be complementary in terms of both their species composition and their physical structure ( 5 – 8 ). Biogeochemical and structural characteristics of shallow-water ecosystems can differ significantly from each other, a circumstance that allows for potentially positive interactions ( 16 ). For example, if nitrogen fixation is favored in one habitat, organismal growth may be supported in adjacent habitats with less available nitrogen. Moreover, Ruppia meadows affect oxygen availability in the surrounding sediment and overlying water ( 17 ), which connected the habitats in our experiment. Oxygen originally produced by Ruppia , for example, may therefore stimulate mineralization of organic nitrogen and further oxidation to nitrate in habitats that are typically less oxygen-rich, such as silty mud. However, in contrast to species diversity, diversity effects of habitats can only be attributed to positive interactions between habitats and not to niche complementarity or selection effects [sensu Loreau and Hector ( 13 )]. While species in habitat mixtures can increase in relative abundance and thereby drive diversity effects, the proportion of each habitat in this experiment was constant over time. Each type of habitat constituted 25% of the total surface in the four-habitat mixture over the course of the experiment. Therefore, the observed diversity effect cannot be explained by high-performing single habitats. Furthermore, there cannot be any niche complementarity among habitats, as there can be for species, because there are no “habitats” for habitats. Although a habitat can harbor different numbers of species, total habitat cover cannot deviate from 100%. Therefore, we attribute ecosystem-level overyielding in our study to positive interactions among habitats. As shown in fig. S5, GPP is considerably higher in mixtures containing three and four habitat types compared to habitats containing a single habitat type. GPP must therefore be influenced by strong positive interactions. The flux of inorganic nitrogen and nitrogen fixation shows similar, although weaker, patterns. In contrast, denitrification in mixtures is the average of what it is in habitat “monocultures,” signaling no positive interactions. The fact that multifunctionality and single functions in the habitat mixtures deviate from what would be expected on the basis of the functioning in the single habitat treatments ( Fig. 6 and fig. S5) provides evidence that the four habitats did exchange material and energy via the common water column. The structure of the four habitat types in our experiment was physically, chemically, and biologically different during the three seasons. Temperature and light (fig. S6) and concentrations of inorganic nutrients (fig. S7) and organic material (fig. S8, A and B) all displayed seasonal differences, which affected the habitat-defining properties. In spring, the lack of effects on multifunctionality was probably a consequence of the high within- and low between-habitat dissimilarity ( Fig. 2A ) in combination with low water temperature and bacterial diversity due to the hyperdominance of a single operational taxonomic unit (OTU) most closely affiliated with Pseudomonas sp. The structural dissimilarity of the different habitats in spring indicates physical sediment disruption, causing habitat homogenization. This is a direct consequence of the winter season conditions, such as erosion by freezing, ice, storms, low light, and low availability of organic nutrients. All these environmental factors induce variation in the biogeochemical properties in sediments of the individual habitats ( 18 ). By contrast, the four habitats were clearly separated in summer due to increased growth of autotrophic components, such as Ruppia meadows, cyanobacterial mats and well-developed diatom mats, higher temperature, and more stable weather conditions. Thus, structural differences could underlie the observed direct relationship between habitat diversity and multifunctionality. In summer, habitat diversity directly affected multifunctionality, while in autumn multifunctionality was mediated via bacterial diversity. Hence, our study also adds to a growing body of evidence showing the importance of indirect effects in mediating ecosystem processes ( 19 – 21 ). Indirect effects have previously been found to be at least as important as direct effects in structuring communities ( 20 , 21 ). Comparing the results of the linear models (estimating net effects) and SEM (partitioning net effects into direct and indirect effects) provided insights into the importance of the indirect effects in our experiment. For example, both bacterial community diversity and habitat diversity were significantly correlated to ecosystem multifunctionality in summer, suggesting that both aspects of diversity were driving the relationship. However, when bacterial diversity was analyzed simultaneously with habitat diversity in a SEM framework, only habitat diversity directly drove multifunctionality. This is in accordance with recent literature, which shows that biogeochemistry can be the single most important driver for the functioning of bacteria-dominated systems ( 22 ) and that changes in bacterial diversity generally have weak effects on ecosystem functioning; only around 25% of the dilution-to-extinction experiments reviewed by Roger et al. ( 23 ) found positive relationships [although there are also recent studies showing strong effects of bacterial diversity; for example, Delgado-Baquerizo et al. ( 24 )]. However, in autumn, the effect of habitat diversity on ecosystem multifunctionality was indirect and mediated through changes in bacterial diversity. Thus, both structural complementarity and bacterial diversity were linked to ecosystem multifunctionality, but seasonal changes in ecosystem components, such as biogeochemistry and physical structure ( 18 , 25 ), may have affected their relative importance. This finding illustrates the value of using statistical methods (for example, SEM) that partition between direct and indirect effects if the consequences of changes in biodiversity within ecosystems are to be explained and predicted. Benthic microalgal diversity was unrelated to ecosystem multifunctionality during all three seasons. Because it was unfeasible to identify all benthic diatoms visually to species level, identification methods might have been insufficiently precise to detect potential differences in species diversity. A general positive effect of species richness of primary producers on production has previously been shown ( 4 , 26 ), but the consequences of changes in benthic microalgal diversity are not well understood. An observational study that related benthic microalgal diversity to functioning found both positive and negative relationships ( 27 ). The only study so far in which the species richness of benthic diatoms was manipulated reported positive effects of diversity on production ( 28 ). However, the maximum richness was only eight species, and natural communities of benthic diatoms are far more diverse ( 27 ). Diversity and community composition of other types of microbiota, such as meiofauna and protozoa, might play a role in ecosystem multifunctionality. However, a full mechanistic elucidation of all relationships between the biotic and abiotic components of habitat diversity on ecosystem-scale multifunctionality was beyond the scope of our study and warrants further research. The present study demonstrates a direct link between habitat diversity and ecosystem-scale multifunctionality, an association that was partly mediated by effects associated with bacterial diversity. From a management perspective, our results support the concept ( Fig. 1 ) that habitat homogenization can have negative consequences for ecosystem functioning and the ecosystem services these functions underpin. Environmental management often focuses on the ecosystem and habitat level ( 29 ). However, the continuing focus in biodiversity-functioning research on species, at the expense of other levels of diversity, is unlikely to provide managers with the full range of information that they need ( 30 ). In managing complex ecosystems efficiently, simplification of ecosystem knowledge is often necessary. Our paper thus provides a framework for scaling up biodiversity–ecosystem functioning research to simplify ecological information. Studying the diversity and composition of habitats instead of, or in concert with, species diversity has the potential to provide more relevant data for management decisions. In conclusion, our findings advance the core finding of biodiversity–ecosystem functioning research that species diversity often plays a central role in influencing the magnitude and stability of ecosystem functioning. We demonstrate that the diversity of habitats within an ecosystem can complement species diversity and even independently influence multifunctionality. An aspect of habitat diversity that is not considered in our study is the dispersal and migration of organisms between the habitats. Examples of these phenomena include the observed large-scale synergistic effects between coral reefs and mangrove ecosystems ( 8 ) and the small-scale dispersal that affects the biodiversity-functioning relationship within metacommunities ( 31 ). Nonetheless, to the extent that our study can be extrapolated to other systems, our results suggest that functions may be negatively affected if habitat diversity declines, even though species diversity is sustained."
} | 4,113 |
35425297 | PMC8979244 | pmc | 3,573 | {
"abstract": "The integration of superior mechanical properties and fast healing efficiency for self-healing polyurethane supramolecular elastomers is challenging due to the confliction between high chain mobility for healing and high chain rigidity for mechanical properties. Herein, a strategy to design a “hard–soft” hard domain by the cooperation of quadruple hydrogen bonds (HBs) in the mainchain as restriction units and single HBs in the side chain as diffusion units is reported. The resulting transparent supramolecular elastomer exhibited fast self-recoverability, good puncture resistance and superior mechanical properties with a tensile strength of 20.5 MPa, an extensibility of 2043.7%, a toughness of 146.1 MJ m −3 and a tear resistance of 13.8 kJ m −2 . Moreover, the fast self-healing capacity (healing efficiency > 82% within 3 h under moderate condition) was realized due to the soft effects of weak HBs in the side chain on the strong HBs in the mainchain. Taking advantage of the merits of the supramolecular elastomer, a flexible sensor was simply fabricated, which showed good self-repairable and stable sensing properties. Thus, the elastomer has great potential in the field of flexible electronics and wearable devices.",
"conclusion": "Conclusion In this study, a design concept based on “hard–soft” hard domain by the cooperation of quadruple HBs in the mainchain as restriction units and single HBs in the side chain as diffusion units is proposed to synthesize a novel polyurethane supramolecular elastomer. The elastomer exhibited colorless transparency, high strength, extensibility and toughness, superior puncture and tear resistance, fast self-recoverability and self-healing capacity. The synchronous improvement of the mechanical and self-healing properties was based on the corporation of the two different hard domains and related strong and weak HB networks. Thereinto, the strength of the weak HBs at side chain is the key factor, which not only had direct effects on the exchange and diffusion performance, but also had decisive effects on the self-healing capacity at room temperature. The regulation of the strength between strong HBs in the mainchain and weak HBs in the side chain provides a promising construction strategy for tuning the conflict between mechanical property and self-healing property, which facilitates to find use in the field of flexible electronics and wearable devices.",
"introduction": "Introduction Self-healing polymeric materials with the ability to repair cracks and restore original functions, have been widely developed to improve service reliability, extend service lifetime and reduce material consumption and environmental pollution. 1–4 As such, these promising materials show great value in numerous fields, including flexible electronics, 5–9 soft robotics, 10,11 protective coatings 12,13 and healthcare devices. 14–17 To put self-healing polymeric materials into use in the above fields, the key factor lies in the contradiction between fast healing efficiency and good mechanical properties of the materials. 18–20 The high diffusibility and exchange of the healing agent for fast healing conflict with the good chain rigidity, entanglement and crystallinity of materials for good mechanical properties. 21–23 The integration of good mechanical properties and fast healing efficiency for self-healing polymeric materials is a challenging task, which attracts great interest in both academia and industry. Among numerous self-healing polymeric materials, elastomer especially polyurethane (PU) is the most suitable candidate to meet the aforementioned requirements. 24 PU usually exhibits two phases: soft matrix and hard domains. The soft matrix has segments with low glass transition temperature ( T g ), which endows the material with deformability. The hard domains have segments with strong interactions, which provides the material with good shape fixity. 25–27 At the same time, lots of macromolecular diols and vast amounts of diisocyanates and chain extenders can be used as soft and hard segments, respectively. Thus, the desired mechanical and healing properties may be obtained by adjusting the composition or ratio of soft and hard segments. Basically, two kinds of interactions are often introduced into the hard segments of PU to regulate the self-healing property. One is the non-covalent interaction, including hydrogen bonds (HBs), 18,19,22,24,28–32 metal–ligand coordination, 33,34 ionic interactions, 35 π–π stacking, 36 host–guest interactions 37 and van der Waals forces. 38 The other is the dynamic covalent interaction such as Diels–Alder reaction, 39 disulfide bond, 40 transesterification, 41 siloxane exchange, 42 olefin metathesis, 43 hindered urea bond, 44 boronic ester 45 and so on. Among these, HBs have attracted significant attention in terms of binding energy and self-healing condition. Numerous efforts have been done to fabricate fascinating PU with good mechanical and self-healing property through HBs. 19,32,46–54 For example, Kang et al. introduced multistrength (weak and strong) HBs into poly(urethane urea), which resulted in the elastomer with high stretchability, high toughness, notch-insensitive and water-insensitive self-healing. 32 Xu et al. proposed a “multiphase active HBs” strategy to create a loosely packed hard segment of poly(urethane urea), which also resulted in a ultra-stretchable, notch-insensitive and fast self-healing elastomer. 46 Based on this, Wang et al. further putted forward the “dynamic hard domains” concept. A poly(urethane urea) with decent mechanical property, outstanding self-recoverability and self-healing property was fabricated. 47 However, the tensile strength of the above elastomers was somewhat not good enough because of the relatively low binding energy of the hard segments. By adopting quadruple HBs with high binding energy as side chain in the hard segments, Song et al. reported a healable PU with remarkable tensile strength. However, the healing efficiency of the elastomer was relatively low. 19 To the best of our knowledge, the simultaneous improvement of the tensile strength, stretchability and self-healing property is difficult to achieve by the reported hard segments regulation, because of the natural limitation of the single type of the hard segments. Herein, based on the different demand of the mechanical and self-healing properties to the segments, we propose a strategy to improve the two properties simultaneously through the “hard–soft” hard domain design. The quadruple HBs in the mainchain with high binding energy form “hard” hard domain and act as strong physical crosslinking points, which provide not only high tensile strength, but also good shape fixity. The single HBs in the side chain with low binding energy form “soft” hard domain, which diffuse and exchange easily and result in the fast healing efficiency at a moderate condition. 55,56 Moreover, by the cooperation of the two different hard segments, the elastomer with high toughness, puncture and tear resistance is fabricated. Besides, the self-healing property of the elastomer in the applied flexible electronic is also simply examined.",
"discussion": "Results and discussion To synchronously regulate the mechanical and self-healing properties, HMIC and hindered phenol (HP) units were introduced into the mainchain and side chain of PU-HM-HP elastomer, respectively ( Fig. 1a ). The chemical structures of the elastomer were confirmed by the 1 H-NMR and FT-IR analysis (Fig. S15 and Table S1 † ). HMIC can form quadruple HB interactions with an AADD–DDAA (A-proton acceptor, D-proton donor) type, which exhibits a high dimerization constant (>10 7 L mol −1 ) and excellent stability. 57,58 Thus, HMIC acted as strong physical crosslinking points in the elastomer with the combination of rigid HMDI ( Fig. 1b ). HP units can form single HB interactions, which shows a weak binding energy and strong exchangeability. 59 By attaching HP units to the end of a long side chain, the easy diffusion of HP units was achieved in the elastomer. Besides, the excessive entanglement of chains in the elastomer was also weaken due to the long side chain ( Fig. 1b ). Further, the influence of the two components on the microstructure of the elastomer was studied by FT-IR. The stretching vibration region of 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 group of PU-HM-HP elastomers can be divided into four bands by using Gauss Lorzne peak fitting (Fig. S16 † ). The bands appeared at about 1720 cm −1 , 1695 cm −1 , 1665 cm −1 and 1635 cm −1 are attributed to free urethane C O groups, H-bonded urethane C O groups (and free urea C O groups), disordered H-bonded urea C O groups and ordered H-bonded urea C O groups, respectively. 19 PU-HM-HP samples with high HP content showed a high intensity of the free and H-bonded urethane C O groups and an extremely low intensity of the ordered and disordered H-bonded urea C O groups (Fig. S17 † ), which indicated a “soft and weak” hard domain. The situation of samples with high HMIC content was opposite, which indicated a “hard and strong” hard domain. By the cooperation of the “soft and weak” hard domain and the “hard and strong” hard domain, PU-HM-HP elastomers with desired properties were achieved. For example, the average transmittance of the colorless PU-HM-HP films under visible range from 400 to 800 nm was about 90% ( Fig. 1c ). Moreover, PU-HM-HP exhibited excellent repeatable deformation recovery and superior puncture resistance. The shape of the 0.7 mm thick sample was almost completely recovered within 30 s from an elongation of 650% ( Fig. 1d and S18 † ) and the 0.7 mm thick sample was not pierced even punctured up to an elongation of 500% ( Fig. 1e and S19 † ). Fig. 1 Schematic and performance of PU-HM-HP elastomer. (a) Chemical structure of PU-HM-HP. (b) Schematic illustration of microphase-separated structure of PU-HM-HP. (c) Transmittance spectrum of PU-HM-HP film with a thickness of 180 μm. Inset photograph of the film is transparent and colorless. (d) Optical images of PU-HM4-HP6 that can recover its original length after being stretched to 650% strain and (e) optical images of PU-HM4-HP6 that can tolerate puncture. In addition, basic characterization experiments of PU-HM-HP were shown in the ESI. † The number-average molecular weight of the elastomers ranged from 61 to 91 kDa with polydispersity indexes of about 1.98 (Fig. S20 and Table S2 † ). The T g of the elastomers was detected at about −45 °C and neither exothermic nor endothermic peaks were observed in the DSC curve, which indicated the noncrystalline hard domains in PU-HM-HP (Fig. S21 † ). The XRD results further confirmed the amorphous structure of PU-HM-HP (Fig. S22 † ). The initial decomposition temperature was about 300 °C, which indicated excellent thermal stability of PU-HM-HP (Fig. S23 † ). The mechanical properties of PU-HM-HP elastomers with different HM/HP molar ratio were quantitatively examined via uniaxial tensile testing. Typical strain–stress curves of the elastomers were shown in Fig. 2a and all PU-HM-HP samples showed rubber-like properties. PU-HM0-HP10 exhibited a young's modulus of 0.283 ± 0.016 MPa, a tensile stress of 1.383 ± 0.134 MPa and an elongation of 1536.246 ± 114.104% ( Fig. 2b and c ), which was attributed to the weak restraint of the “soft” HP-type side chain dominated hard domains to the elastomer. With HM/HP molar ratio increasing, the young's modulus and tensile stress gradually increased, which was attributed to the strong crosslinking effects of the “hard” HMIC dominated hard domains. In particular, the young's modulus showed a significant improvement from 1.896 ± 0.082 MPa of PU-HM4-HP6 to 2.670 ± 0.143 MPa of PU-HM5-HP5 ( Fig. 2b ). Associated with the strain-induced crystallization phenomenon only observed in PU-HM5-HP5 system during tensile testing (not presented in this paper) and the significant increase of the intensity of the ordered H-bonded urea C O groups from PU-HM4-HP6 to PU-HM5-HP5 as well as the highest intensity of disordered H-bonded urea C O groups of PU-HM4-HP6 (Fig. S17 † ), the relatively regular alinement of chains induced by the strong restriction of HMIC hard domains was generated in PU-HM5-HP5 system, which resulted a tensile stress of 27.237 ± 1.051 MPa. 60 Although the elongation at break showed a fluctuation change along with HM/HP molar ratio variation, the elongation of all PU-HM-HP samples was higher than 1450%, which was mainly attributed to the weakened entanglement of the chains and the “sacrificial bond” effects of the weak HBs. Typically, PU-HM4-HP6 exhibited the highest elongation of 2043.749 ± 136.098% with a relatively high tensile strength of 20.531 ± 0.528 MPa, which resulted in a toughness of 146.097 ± 8.241 MJ m −3 . PU-HM5-HP5 exhibited the elongation of 1831.124 ± 95.457% and resulted in the highest toughness of 211.935 ± 12.957 MJ m −3 ( Fig. 2d ). It is noteworthy that the toughness of our elastomers was higher than that of man-made fibers such as Kevlar 49 (50 MJ M −3 ), Nylon 66 (80 MJ M −3 ) and spider dragline silk (180 MJ m −3 ). Moreover, the toughness was comparable or even exceeded the record of some materials reported in the literature. 24,28,31,37,55,61–66 Such superordinary toughness is beneficial for avoiding crack formation and prolonging service lifetime. Fig. 2 The mechanical properties of PU-HM-HP elastomers. (a) Strain–stress curves of PU-HM-HP with a loading velocity of 100 mm min −1 . (b) Young's modulus variation of PU-HM-HP. (c) Tensile strength and elongation at break variation of PU-HM-HP. (d) Toughness variation of PU-HM-HP. Inset graph illustrates the calculation method of toughness. (e) Optical images of the single-edge-notched PU-HM4-HP6 sample from 0% to 800% strain and (f) fracture energy variation of PU-HM-HP. Inset graph illustrates the calculation method of fracture energy. Tear resistance of elastomers is also important for the protection of the relevant devices. As Fig. 2e and S24 † shows, PU-HM4-HP6 with a cut notch could be stretched up to 8 times without fracture, which was also attributed to the weakened entanglement of the chains and the strong–weak HB network effects. 19,32 With HM/HP molar ratio increasing, the fracture energy of PU-HM-HP system gradually increased ( Fig. 2f ). PU-HM4-HP6 and PU-HM5-HP5 showed a fracture energy of 13.798 ± 0.184 kJ m −2 and 16.698 ± 0.641 kJ m −2 , respectively, which revealed the good tear resistance of the elastomers by comparing with natural rubber (fracture energy is about 10 kJ m −2 ). The “sacrificial bond” effects of the weak HBs at small deformation and the crosslinking effects of the strong HBs at large deformation of PU-HM-HP allows energy dissipation at multiple scales. To reveal the role of different HBs in energy dissipation, five successive loading-unloading process with a strain of 500% in every cycle were performed (Fig. S25 † ). All PU-HM-HP exhibited large hysteresis loops in the first loading–unloading process and sharply decreased in the second loading–unloading process, which was attributed to the massive breaking of the multiple HBs. Then the hysteresis showed slight decrease in the following three loading–unloading process, which was mainly assigned to the recombination of the weak HBs. In each cycle, the higher HM/HP molar ratio, the larger the hysteresis area ( Fig. 3a ). The hysteresis difference between the first and second successive cyclic tensile testing (Δ W ) represents the ability of chain rearrangement, a larger value represents the harder rearrangement because of the molecular interactions. The ratio between the integrated area in the hysteresis loop and that under the loading curve of the first cyclic tensile testing ( W 1 / W ) represents the efficiency of energy dissipation, a higher value represents the better efficiency. With HM/HP molar ratio increasing, the Δ W gradually increased due to the increasing restriction of the strong HBs. The W 1 / W first decreased then increased and PU-HM0-HP10 showed the highest value among the elastomers, which indicated that the efficiency of energy dissipation of the weak HBs was better than that of the strong HBs and the high density of the strong HBs was benefit for the efficiency ( Fig. 3b ). Besides, with more strong HBs, the stress difference between the first and second successive cyclic tensile testing at 500% strain and residual strain of the first cyclic tensile testing gradually increased and decreased, respectively ( Fig. 3c ). Dissipated energy during loading-unloading process under 100% to 500% strains was also examined (Fig. S26 † and 3d ). At low strain (100%), the hysteresis areas of all PU-HM-HP showed little difference, which was attributed to the preferentially breaking of weak HBs. With strain increasing, the difference of the hysteresis areas of PU-HM-HP became larger and larger, which was attributed to the different amounts of strong HBs dissociation at high strain. Thus, the effective energy dissipation process was achieved due to the combination of the weak sacrificial HBs and the strong restrictive HBs, which resulted in a tough PU-HM-HP elastomer. Fig. 3 The energy dissipation of PU-HM-HP elastomers. (a) Dissipated energy during five successive loading-unloading process with a strain of 500% in every cycle. (b) Variation of the hysteresis difference between the first and second successive cyclic tensile testing (Δ W ) and the hysteresis ratio of the first cyclic tensile testing ( W 1 / W ). (c) Variation of the stress difference between the first and second successive cyclic tensile testing at 500% strain and residual strain of the first cyclic tensile testing. (d) Dissipated energy during loading–unloading process under 100% to 500% strains. (e) Cyclic tensile testing curves in successive cycles of PU-HM4-HP6 with a strain of 500% in each cycle and a delay time of 0 h and 1 h, and (f) the hysteresis ratio between the 1 h cyclic tensile testing and the initial cyclic tensile testing. To elucidate the recovery property of PU-HM-HP elastomers, cyclic tensile testing curves in successive cycles were conducted after 1 h relaxation ( Fig. 3e ). As Fig. 3f shows, the hysteresis ratio between the 1 h cyclic tensile testing and the initial cyclic tensile testing gradually decreased with HM/HP molar ratio increasing. PU-HM0-HP10 had a hysteresis ratio of about 88% after 1 h relaxation while that of PU-HM5-HP5 was about 42%. The reconnection process arising from weak HBs dominates the faster elastic recovery, while the strong HBs showed a weak dynamic, which needed more time or more energy to completely reformed. The introduction of single HBs at the end of a long side chain makes the easy exchange and diffusion of the HBs, which is expected to improve the self-healing efficiency. To verify our assumption, PU-HM-HP elastomers healed at different temperature for different time were conducted. For quadruple HBs contained elastomers such as PU-HM4-HP6 ( Fig. 4a ), the tensile strength healed at room temperature (25 °C) for 6 h could reach about 1.3 MPa due to the easily exchange and diffusion of single HBs at side chain. However, the tensile strength showed little change when the healing time was extended, as the quadruple HBs were stable at room temperature. When healed at 80 °C, the tensile strength exhibited a noticeable increase within 3 h, while the elongation at break possessed a large value even healed for only 1 h. Therefore, the healing efficiency in this paper is defined as the ratio between the healed and virgin tensile strength for accurate assessment. As Fig. 4b shows, only PU-HM0-HP10 system had an excellent healing efficiency at room temperature (about 100% for 6 h healing). For other PU-HM-HP elastomers, though the healed tensile strength could reach about 1.3 MPa, the increased virgin tensile strength with HM/HP molar ratio increasing resulted in the rapid decrease of the healing efficiency at room temperature. For high temperature self-healing (80 °C), PU-HM0-HP10 to PU-HM3-HP7 samples exhibited a high healing efficiency after 2 h healing (>92%). PU-HM4-HP6 showed a healing efficiency of above 82% after 3 h healing. However, the healing efficiency for PU-HM5-HP5 was only about 36% even after 3 h healing. As is well known, the healing process goes through five steps: (a) surface rearrangement, (b) surface approach, (c) wetting, (d) diffusion and (e) randomization, during which the diffusion is the most critical step for healing. 76,77 For PU-HM-HP system, the introduction of weak HBs in the side chain not only weakened chain entanglement of soft domain, but also had soft effects on the strong restraint hard domain, which promoted the chain diffusion and resulted in the fast self-healing of the elastomer. As the scratch recovery of PU-HM1-HP9 and PU-HM4-HP6 showed ( Fig. 4c ), the scratches of about 100 μm shrink to about 17 μm and 32 μm for PU-HM1-HP9 and PU-HM4-HP6 within 5 min, which was attributed to the fast diffusion of chains. Then the scratches showed little change in the following 1 h. The size of the scratch of PU-HM4-HP6 after 1 h healing was about 2 times than that of PU-HM1-HP9, which was attributed to the restriction of the quadruple HBs. The non-desirable healing efficiency of PU-HM5-HP5 was mainly attributed to the regular arrangement of chains caused by the restriction of the quadruple HBs as mentioned above. Therefore, PU-HM4-HP6 exhibited the balanced mechanical and self-healing properties. As Fig. 4d shows, though the incision of PU-HM4-HP6 system was not fully repaired, the sample could lift a 5 kg object without changes of incision after healing at 80 °C for only 1 h. Such fast healing efficiency had obvious advantages over other reported UPy-type self-healing elastomers ( Fig. 4e ). Fig. 4 The self-healing properties of PU-HM-HP elastomers. (a) Representative stress–strain curves of virgin and self-healed PU-HM4-HP6 sample. (b) The healing efficiency of PU-HM-HP elastomers at different temperature for different time. (c) Optical microscopy images of scratch recovery of PU-HM1-HP9 and PU-HM4-HP6 films. (d) Load-bearing test of the self-healed PU-HM4-HP6 sample and (e) graphic comparison of self-healed tensile strength, healing time and healing efficiency at 80 °C for UPy-type self-healing elastomers. 19,67–75 Taking advantage of PU-HM-HP elastomer, a flexible sensor composed of MWCNTs and PU-HM4-HP6 was fabricated. The conductivity and sensing property in response to vocalization, finger and wrist bending were simply examined. The healed sensor film, which was cut in half with a razor blade and healed at 80 °C for 1 h after splicing, was used for the conductivity test ( Fig. 5a ). When power on, the LED lamp connected in series with the healed film was lighted up, which indicated the conductive network of the film was reconstructed after healing. When the film was stretched to 65%, the brightness of LED lamp remained unchanged and no crackle was observed, which indicated that the conductive film exhibited good flexibility and stability. When the film was stretched further to 130%, the LED lamp dimed, which was attributed to the obvious crackle of the MWCNTs layer on the surface of PU-HM-HP layer. With this conductive and flexible sensor film, the sensing property of the virgin and healed sample (cut in half with a razor blade and healed at 80 °C for 1 h after splicing) was compared. The current signal of the virgin and healed sensors in response to vocalization could both be detected, which indicated the high sensitivity of the sensors ( Fig. 5b ). Both of the virgin and healed sensors showed a repeatable and wide response peak, while the sensing sensitivity of the healed sensor showed a little decline compared with that of the virgin sensor. For the current signal in response to finger and wrist bending, both of the virgin and healed sensors showed a sharp double splitting peak ( Fig. 5c and d ). Despite fluctuations, both the peaks of the virgin and healed sensors showed good stability and repeatability. Though the sensing sensitivity of the healed sensors showed a certain degree of decline for both actions, PU-HM-HP elastomer still showed great potential in the field of flexible electronics and wearable devices associate to the superior puncture and tear resistance of the elastomer. Fig. 5 The conductivity and sensing properties of PU-HM-HP elastomers. (a) Optical images of conductive behavior of the healed sensor under stretching. (b) Optical images and the current signal of the virgin and healed sensors in response to vocalization. (c) Optical images and the current signal of the virgin and healed sensors in response to finger bending and (d) optical images and the current signal of the virgin and healed sensors in response to wrist bending."
} | 6,343 |
35557509 | PMC9090350 | pmc | 3,574 | {
"abstract": "ABSTRACT The rice leaf, combining the surface properties of lotus leaves and shark skin, presents outstanding superhydrophobic properties motivating its biomimesis. We created a novel biomimetic rice-leaf superhydrophobic surface by a three-level hierarchical structure, using for a first time stereolithographic (SLA) 3D printed channels (100µm width) with an intrinsic roughness from the printing filaments (10µm), and coated with TiO 2 nanoparticles (22 and 100nm). This structure presents a maximum advancing contact angle of 165° characterized by lower both anisotropy and hysteresis contact angles than other 3D printed surfaces, due to the presence of air pockets at the surface/water interface (Cassie-Baxter state). Dynamic water-drop tests show that the biomimetic surface presents self-cleaning, which is reduced under UV-A irradiation. The biomimetic surface further renders an increased floatability to 3D printed objects meaning a drag-reduction due to reduced water/solid contact area. Numerical simulations of a channel with a biomimetic wall confirm that the presence of air is essential to understand our results since it increases the average velocity and decreases the friction factor due to the presence of a wall-slip velocity. Our findings show that SLA 3D printing is an appropriate approach to develop biomimetic superhydrophobic surfaces for future applications in anti-fouling and drag-reduction devices.",
"conclusion": "4. Conclusion The combination of SLA 3D printed microchannels with a TiO 2 -HTMS coating presented in this study can mimic the superhydrophobic behavior of the rice leaf. Our results show that apart from the designed channels and the presence of nanoparticles, the roughness generated by the SLA 3D printing filaments is relevant to adding a hierarchical structure to the material. On the flat printed surfaces, the presence of these filaments provides 35° of anisotropy and allows a hydrophobic wettability (contact angles measured >95°) in the perpendicular direction, and a hydrophilic behavior in the parallel direction (contact angles measured <74°). The TiO 2 -HTMS coating allowed to reach superhydrophobic states for all microchannel surfaces, showing contact angles >160°. The best biomimicry is obtained with 100µm coated microchannels with an advancing contact angle value of 165°, a contact angle hysteresis <9°, and a contact angle anisotropy of 5°, which meets the requirements of a rice leaf biomimetic superhydrophobic surface. This wettability was reached by a hierarchical three-level structure: printing filaments (around 10µm scale), microchannels (around 100µm scale), and nanometric coating (22–100nm scale). Air pockets under the drops placed over the biomimetic surface confirm the Cassie-Baxter state of the material. Dynamic tests show the self-cleaning property of the surface and its tunable wettability under UV-A irradiation. Floatability experiments show that the biomimetic surface can modify the contact area of an object placed over water. Numerical simulations confirm that the presence of air in a channel having a biomimetic bottom-wall increases the average velocity and decreases the friction factor, as a slip velocity exists on the biomimetic boundary layer. Our findings show that the combination of SLA 3D printing with a TiO 2 -HTMS coating is an excellent strategy to develop biomimetic superhydrophobic surfaces for future applications in fog-harvesting, anti-fouling, and drag reduction devices.",
"introduction": "1. Introduction Biomimesis aims to solve human problems by mimicking the solutions that nature has already found through the collaboration of different scientific fields, such as chemistry, biology, physics, nanotechnology, tribology, material sciences, and engineering. An example of this is the biomimicry of the topography of natural surfaces, which aims to reproduce the unique arrangement of hierarchical micro and nano-structures found in the surfaces of the lotus leaf [ 1 ], the rice leaf [ 2 ], and insect wings [ 3–5 ], among others. It has applications in the development of new materials with topographies that allow anti-biofouling, drag reduction, detection of analytes, and improved catalysis. Recent investigations report the use of nanoarchitectonics for the development of new biomimetic surfaces, which construct micro and nano-metric topographies through the manipulation of atoms and molecules [ 6 ]. One of the challenges solved by nature is the control of the interaction between water and a surface through superhydrophobic structures [ 7 ], allowing for a drag reduction in fluids [ 8 ]. Viscous drag and pressure resistance are the two main mechanisms by which objects resist the movement of fluids, or vice versa, playing a crucial role in the development of hydrodynamic structures. From the different contributions to the total drag force, viscous drag is the most relevant requiring the largest amount of energy to overcome [ 9 ]. Nature faces this problem using surfaces that either repel water (for laminar flows) or stabilize the flow behavior or vortices (for turbulent flows). For instance, the shark skin experiences drag reduction for turbulent flows due to the presence of denticles with riblets on their surfaces [ 10 ]. The specific surface morphology and superhydrophobicity found in nature can add other functionalities such as self-cleaning, which is found for the lotus leaf [ 1 ], the rice leaf, and butterfly wings [ 3 ]. These naturally occurring surfaces repel droplets of water without opposing any resistance to their movement due to the presence of air pockets at the interface, thus motivating a biomimetic design of materials to solve relevant industrial problems such as pollution, corrosion, icing, and biofilm formation [ 7 , 11–14 ]. Superhydrophobicity is a type of wettability tending to repel a drop of water with contact angles above 150°, a contact angle hysteresis below 10° (difference between the advancing and receding angle), a sliding angle of less than 5°, and high stability in the Cassie-Baxter state (i.e. presence of air trapped between the surface and the water droplet) [ 15 , 16 ]. In nature, superhydrophobic surfaces have a hierarchical structure at the micro-and nanometric scale [ 17 ]. Regarding different natural materials featuring this behavior, the rice leaf is highlighted because it combines the properties of the lotus leaf and the shark skin [ 8 ], thus having a three-level hierarchical structure as depicted in Figure 1a . This structure is based on a micrometric arrangement of channels or riblets (around 50µm in height with a peak separation separated of 200µm) with rough papillae on top of them (around 2µm) covered by wax (changed surface chemistry) [ 18 ]. This hierarchical structure provides a superhydrophobic behavior with a contact angle of 164° and a sliding angle anisotropy of approximately 6° [ 2 , 8 ].\n Figure 1. (A) Scheme of the hierarchical structure of the superhydrophobic rice leaf surface composed of micrometric riblets and hydrophobic papillae with nanometric features on top of them. This structure forms air pockets between the surface and the water droplet. (b) Geometry and coordinate axis of the biomimetic surface studied and simulated. Here, the flow moves through the x axis with a velocity u , while h represents the height of the microchannel, H stands for the gap between microchannels, W is their width ( d ), and h is the distance between the bottom and the top of the channel. (c) Silanization process of the TiO 2 nanoparticles, where the HTMS chains attach via the hydroxyl functional groups and provide a hydrophobic behavior. (d) Perpendicular and parallel views of the surface for contact angle measurements orientation. Hierarchical superhydrophobic surfaces can be manufactured using either physical modification (for instance by plasma and molding processes) [ 19 , 20 ], or hydrophobic coatings (for instance by spin-coating, aerosol, and electrochemical deposition) [ 21 , 22 ]. Apart from those processes, 3D printing has been gaining interest in the last years since it allows for the creation of objects with many possibilities in terms of precision, sizes, shapes, mechanical and surface properties, including from delicate prototypes to robust engineering products. By adding layers of a polymeric material, 3D printing can create micropatterns needed in the hierarchical structure of a superhydrophobic surface through techniques such as fused deposition modeling (FDM) [ 23–25 ], inkjet printing [ 26 ], one-step 3D printing [ 27 ], immersed surface accumulation-based 3D (ISA-3D) printing [ 28 ], and mold imprinting [ 2 , 3 , 8 ], among others. However, irrespective of the method used, the printed structure cannot reach a superhydrophobic wettability by itself as polymeric resins are mostly hydrophilic [ 29 ]. Moreover, they do not have a hierarchical structure, which directly asks for a surface modification such as the application of a nanostructured, hydrophobic coating. Titanium dioxide (TiO 2 ) based coatings are studied since they can provide a superhydrophobic character to different materials such as cellulose sponges [ 30 ], polydimethylsiloxane (PDMS) [ 31 ], and flat glass pieces [ 32 ]. They can be applied by dip-coating, spraying, and spin-coating, among others [ 33 ]. Additionally, coatings consisting of TiO 2 nanoparticles have been combined with other substances due to their photocatalytic, bactericidal, and self-cleaning properties [ 32 , 34–37 ]. Rice leaf biomimetic surfaces have also been developed using both nanostructured silica coatings and 3D printing including fused deposition modeling (FDM) and mold imprinting [ 8 , 23 ], although other coatings can also be used for this purpose (for example polydivinylbenzene powder) [ 2 ]. Regarding the printing process, stereolithographic (SLA) 3D printing has advantages over other 3D printing techniques regarding its deposition speed (10 5 mm 3 h −1 ) and high precision (25µm) [ 38 ], besides being the oldest and, therefore, the most mature 3D printing techniques [ 39 ]. However, despite all these advantages, it has been barely used to manufacture biomimetic surfaces. Few examples can be found dealing with lotus leaves with a contact angle of 139° without any additional coating and shark skin with a drag reduction of 10% in pipe experiments [ 40 ], as well as a cactus surface coated with TiO 2 nanoparticles having a contact angle below 160° [ 41 ]. A biomimetic surface based on a rice leaf is yet to be explored by SLA. Moreover, the use of the microstructure provided by the SLA 3D printing filaments for the construction of a biomimetic hierarchical structure has not been reported yet. Consequently, this study aims at mimicking the three-level hierarchical structure of the rice leaf using SLA 3D printed microchannels modified by the incorporation of TiO 2 nanoparticles functionalized by hexadeciltrimethylsiloxane (HTMS). Our results show that besides the morphology of the designed 3D printed structure generated by SLA (of around 100µm), the imperfections from the presence of SLA 3D printed filaments (of around 10µm) render a hierarchical structure providing a further roughness to the printed surface, to which the nanoparticles (22 and 100nm) were added. The combination of SLA 3D printing and TiO 2 coatings can be used in many applications, as it allows for the printing of infinite shapes with a biomimetic surface. Numerical simulations were further carried out to confirm the underlying mechanism of the dynamic superhydrophobic properties arising from air pockets that the biomimetic material provides.",
"discussion": "3. Results and discussion 3.1. Numerical results As discussed above, superhydrophobic surfaces enable the existence of air pockets between a drop of water and the surface arising from their hierarchical structure [ 17 , 55 ]. To confirm the effects that a rice leaf biomimetic surface with trapped air could have on the dynamics of water transport, we carried out numerical simulations of a simplified system consisting of water moving inside a rectangular channel under laminar flow conditions (that favor the existence of the air pockets) [ 56 ]. The superhydrophobic surface was simulated through a simplified structure consisting of a flat hydrophilic surface on the top wall and a single superhydrophilic microchannel surface at the bottom formed by the presence of a solid-riblet (right-side) and air pocket (left-side). The boundary conditions at the bottom surface are zero velocity at the solid-water interface as well as shear force and velocity continuity at the air–water interface. The lateral walls of the rectangular channels presented periodic boundary conditions to replicate the periodical arrangement of the experimental biomimetic surface. Both channels have the same height of H=9.5 · 10 −4 m (not including the microchannel), the channel width is 2 · 10 −4 m, the microchannel width is 1 · 10 −4 m and its height is 5 · 10 −5 m. We aim at comparing the velocity profiles obtained from this superhydrophobic surface having microchannels with the results from a flat surface. The simulations were conducted for pressure gradients of 100, 500, 1000, 5000 and 10,000Pa m −1 to allow a laminar flow under the specific channel dimensions. An example of the velocity profiles obtained for each configuration of the channel (flat and with microchannel) is displayed in Figure 2(a,b) for a pressure gradient of 5000Pa m −1 . In this context, the effect of the microchannel at the bottom wall can be approximated as a change in the shape of the velocity profile ( Figure 2b ), as a non-zero value at the biomimetic surface (H=0) increasing the maximum velocity at the center of the channel by 9%, and the average velocity by 13%. This phenomenon is known as ‘slip velocity’ [ 57 ] and occurs when the shear force exhorted by the surface (air/water interface) over the fluid is not enough to stop the layer that advances at the interface (boundary layer), which can be theoretically modeled as a penetration of the boundary layer into the material [ 58 ]. For superhydrophobic surfaces, this slip velocity is generated through the decrease in the shear force exerted by the air pockets compared to the solid surface due to its small viscosity compared to water [ 59 ]. The values obtained for the average velocity, slip velocity, shear stress, and friction factor for each case are presented in Tables 1 and 2 (for the flat and biomimetic bottom walls, respectively).\n Figure 2. (A) Velocity profiles obtained via numerical simulations for a pressure gradient of 5000 Pa m −1 on a flat channel (left) and a surface with one microchannel at the bottom wall (right). Maximum velocities are at the center of the channel (red): .58m s −1 and .63m s −1 , respectively. No-slip boundary conditions can be observed at the solid–fluid interface (yellow), and velocity continuity (slip velocity) can be observed at the air-water interface (microchannel, red line). (b) Velocity profiles were obtained via numerical simulations for the flat channel and the surface with one microchannel for a pressure gradient of 5000 Pa m −1 . the plotted velocity value corresponds to the average velocity for each channel height. Table 1. Numerical simulations results for the flat bottom wall Flat bottom wall Pressure gradient [Pa/m] Re Average velocity [m/s] Slip velocity [m/s] τ bottom wall [N/m 2 ] Friction factor f [-] 100 14 7.53×10 −3 0 4.36×10 −5 6.16×10 −3 500 72 3.77×10 −2 0 2.18×10 −4 1.23×10 −3 1000 143 7.53×10 −2 0 4.36×10 −4 6.16×10 −4 5000 715 3.77×10 −1 0 2.18×10 −3 1.23×10 −4 10000 1431 7.53×10 −1 0 4.36×10 −3 6.16×10 −5 \n Table 2. Numerical simulations results for the biomimetic bottom wall Biomimetic bottom wall Pressure gradient [Pa/m] Re Average velocity [m/s] Slip velocity [m/s] τ bottom wall [N/m 2 ] Friction factor f [-] 100 22 8.53×10 −3 2.01×10 −3 4.12×10 −5 4.54×10 −3 500 108 4.26×10 −2 1.00×10 −2 2.06×10 −4 9.09×10 −4 1000 216 8.53×10 −2 2.01×10 −2 4.12×10 −4 4.54×10 −4 5000 1080 4.26×10 −1 1.00×10 −1 2.06×10 −3 9.09×10 −5 10000 2160 8.53×10 −1 2.01×10 −1 4.12×10 −3 4.54×10 −5 The microchannel surface allows for a 26% reduction of the friction factor compared to the flat channel. Under these conditions, the Re values also increased proportional to the velocity increase, although showing values lower than 2200 meaning laminar flow in all the cases (controlled by the pressure gradients). The slip velocity has a direct proportionality with the pressure gradient, and its value is about 24% of the average velocity on the experiments with microchannels. Flat channels do not provide a slip velocity, as a no-slip boundary condition was imposed for this type of surface. At higher pressure gradients (displaying turbulent flow), the air pockets formed by the biomimetic surface are not present [ 56 ]. The simulated slip length provided by the biomimetic configuration can be estimated by extending the velocity profile slope obtained at the surface/water interface until reaching the position where the boundary layer would theoretically end [ 60 ], obtaining a value of b=8.96×10 −5 m. This corresponds to 10% of the total height of the channel, which could allow for 23% of drag reduction in a Poiseuille flow configuration according to Choi et al. [ 61 ]. Considering the effect of a biomimetic wall in the simulations carried out, it is expected that the experimental results show a relevant change in the interaction of water droplets with the different surfaces given the presence of air gaps, decreasing the adhesion of the material with the fluid. This effect will be reflected in the advancing and receding contact angle measurements, where less friction of the material with the water will deform the droplets interacting with the surface less. The speed of the droplets, when deposited on inclined surfaces, should increase considerably. 3.2. Surface characterization 3.2.1. Analysis of TiO 2 -HTMS nanoparticles and coated surface The modified nanoparticles (TiO 2 -HTMS) were characterized through XRD and TEM techniques. Figure 3a shows the powder XRD diffractogram of TiO 2 -HTMS nanoparticles presenting diffractions peaks at 2θ=25.3 °, 37.0 °, 37.9 °, 38.5 °, 48.0 °, 54.0 °, 55.1 °, 62.7 °, 68.9 °, 70.3 °, 75.2 °, and 2θ=27.5 °, 36.1 °, 41.3 °, 53.8 °, and 56.6 °, corresponding to mixed phases of anatase and rutile, respectively [ 62 , 63 ]. TiO 2 -HTMS nanoparticles exhibit compact self-organization, as shown in the TEM image of Figure 3b arising from the attractive interaction of the alkyl chains on the surface of the nanoparticles. The size of the spherical nanoparticles is shown in Figure 3c as a distribution histogram. The mean size of TiO 2 -HTMS is 35nm, exhibiting a high frequency of nanoparticle sizes between 10–20nm and 35–40nm. The chemical modification of nanoparticles does not change the commercial product’s size. Figure 3b otherwise shows well disperse nanoparticles meaning a posterior efficient and well-packed coverage.\n Figure 3. Microstructural and chemical characterization of TiO 2 -HTMS nanoparticles and coated surface. (a) Powder XRD diffractogram of TiO 2 -HTMS nanoparticles (blue and red indices correspond to anatase and rutile phases, respectively). (b) Representative TEM image of TiO 2 -HTMS nanoparticles. (c) Histogram of TiO 2 -HTMS nanoparticle’s size distribution. (d) TGA curve of TiO 2 -HTMS nanoparticles under a nitrogen flow (dotted line corresponds to the first derivative of the TGA curve). (e) ATR-FTIR spectra of TiO 2 (blue line, a), TiO 2 -HTMS (green line, b), pure 3D printed resin (magenta line, c), and TiO 2 -HTMS coated 3D printed resin (red line, d). (f) SEM-EDS elemental analysis for the surface of the TiO 2 -HTMS coated polymeric resin (top left) showing the presence of carbon (green), oxygen (red), and titanium (blue). TGA estimated the degree of functionalization of TiO2-HTMS nanoparticles. Thermal degradation of silanized metal oxide nanoparticles provides information of the organic content [ 64 ]. Figure 3d shows the thermogram of TiO 2 -HTMS under a nitrogen atmosphere. The curve exhibits a 5.25% mass loss between 100°C and 650°C corresponding to the organic HTMS moiety. The mass loss in two steps could be due to the different sizes of nanoparticles, each one having a different modification degree due to the different surface area. The chemical modification of TiO 2 nanoparticles and the coating of the 3D printed structures were studied via ATR-FTIR spectroscopy ( Figure 3 ). Figure 3e shows the FTIR spectra of the nanoparticles before and after silanization with HTMS as well as the 3D printed resin before and after the application of the hydrophobic coating. The spectrum of TiO 2 (blue line in Figure 3a ) shows a broadband at 3350cm −1 , corresponding to the stretching vibration mode of the O-H bonds of hydroxyl groups, a weak band at 1630cm −1 , corresponding to the bending vibration mode of the O-H bonds of moisture water, and an intense band at 600cm −1 associated to stretching vibration mode of the Ti-O-Ti bonds [ 65 , 66 ]. The spectrum of TiO 2 -HTMS (green line in Figure 3a displays two new sharp and medium intensity bands at 2920cm −1 and 2850cm −1 , corresponding to the asymmetrical and symmetrical stretching of the C sp3 -H bonds of alkyl chains [ 67 ]. It also shows bands around 1465cm −1 , 1170cm −1, and 1030cm −1 , corresponding to scissoring of C sp3 -H bonds, and symmetrical as well as asymmetrical stretching of Si-O-Si bond (formed by condensation), respectively [ 68 ]. The shoulder of the band observed in the TiO 2 -HTMS spectrum at 950cm −1 corresponds to vibrational modes of Ti-O-Si bonds [ 69 ], which implies that the nanoparticles were successfully modified by HTMS. The commercial photo-cross-linkable resin used for the 3D printing is based on urethane dimethacrylate and other methacrylate monomers [ 52 ]. This resin presents bands at 3350cm −1 , 2950cm −1 , 2860cm −1 , 1700cm −1 , 1530cm −1 , 1452cm −1 , 1240cm −1 and 1150cm −1 , corresponding to the respective vibrational mode of N -H (stretching), C sp3 -H (asymmetrical stretching), C sp3 -H (symmetrical stretching), C=O (stretching), C-N (stretching) combined with N -H (bending), C sp3 -H (bending), C-O (ester, stretching) and C-O-C (stretching) bonds (magenta line in Figure 3a ), respectively [ 70 ]. After coating the 3D printed structure with TiO 2 -HTMS (red line in Figure 3e ), the characteristic bands of TiO 2 -HTMS and the polymer substrate were observed, confirming the physical bonding of the modified nanoparticles on the 3D printed structure. FE-SEM image with elemental analysis results ( Figure 3f ) shows two of the elemental components of the commercial resin (C and O), and further confirms the effective coating of the 3D printed structures by dip coating (Ti). 3.2.2. Surface topography FE-SEM images of the flat and the micropatterned printed surfaces are displayed in Figure 4 . The flat surface ( Figure 4a ) displays a roughness of around 10µm that is characteristic of the printing filament used for SLA 3D printing. This roughness implies that the printed surface is anisotropic depending on the printing direction. The surface topography of the printed microchannels shows rounded edges ( Figure 4b ) instead of the designed perfect rectangles due to the spot-size of the laser beam (around 140µm) used [ 71 ], as the microchannels are printed perpendicular to the printing layer orientation (the printing layers are added in the plane zy, adding height on the axis x of Figure 1b ). However, this mismatch between the designed and obtained morphologies even amplifies the biomimetic aspect of the surface as rice leaves have wavy riblets. The printed channels further mismatch the designed dimensions, especially their height (from 50 to 30µm). To overcome this limitation, three structures with different heights (50, 75, and 100µm) were 3D printed allowing for further analysis of the effect of this variable. Noteworthy, the roughness arising from the 3D printed filaments (observed in the flat surface) is also displayed in this case providing a hierarchical roughness to the microchannels thus improving the similarity to the papillae observed on the rice leaf surface ( Figure 1a ). After applying TiO2–HTMS coating with two particle sizes (22 and 100nm) ( Figure 4c ), the nanoparticles are distributed over the printing filaments ( Figure 4(d,e) ). Figure 3b shows a TEM image of the nanoparticles used to coat the surface, where the different nanoparticle sizes are arranged randomly and close-packed over the treated surface ( Figure 4e ).\n Figure 4. FE-SEM images of the 3D SLA printed surfaces (perpendicular view, plane zy): (a) flat surface showing printing layers, (b) printed microchannels (100µm designed height) with printing layers on top, (c) biomimetic coated microchannels, (d) TiO 2 -HTMS nanoparticles deposited on the microchannels, and (e) top view of the TiO 2 -HTMS nanoparticles placed over the printing filaments. These results show that the surface coating of an SLA 3D printing material mismatches the original rectangular design ( Figure 5a ) and allows for the generation of a three-level hierarchical structure biomimicking the morphology of the rice leaf surface associated with: 3D printed microchannels (100µm scale), roughness from the printing filaments (10µm scale), and nanoparticles (22 and 100nm scale), as displayed in Figure 5b .\n Figure 5. (A) Designed two-level hierarchical surface for the biomimetic material: designed microchannel and nanoparticles. (b) Actual three-level hierarchical surface for the biomimetic material: printed microchannel, printing filament, and nanoparticles. 3D printing has been previously used for developing biomimetic structures and characterized regarding the roughness. For instance, the characteristic roughness of the SLA 3D printing technique is also reported in the work of Wang et al. [ 72 ] for developing hydrophobic surfaces, where structures of micrometric posts (diameter of 300µm) are manufactured and roughness of the order of 30µm is obtained on their sides. However, by using this configuration the printing filaments do not interact with a drop of water, since they are not present on the upper face of the posts and do not necessarily contribute to the wettability of the surface unless the drop penetrates the air spaces (where roughness could even increase the adhesion of the fluid). Li et al. [ 40 ] used the multiscale SLA 3D printing technique to produce a lotus flower surface and shark denticles but did not report a characteristic roughness for the printing filaments. 3D printing by TPP (two-photon polymerization) is another technique that has been used to produce biomimetic microstructures [ 73 ] inspired by the surface of the Salvinia molesta plant. Although they managed to print eggbeater-like structures of the order of 10µm, their surfaces are smooth and without roughness on a smaller scale, which is observed in the natural sample of the plant. The Salvinia molesta surface has also been mimicked using immersed surface accumulation 3D printing [ 28 ]. Likewise, 3D printing by FDM (fused deposition modeling) has been used to manufacture biomimetic structures of the lotus flower [ 74 ] and others similar to the rice leaf [ 23 ]; however, none of these works report a hierarchical roughness attributable to the 3D printing technique. 3.2.3. Surface wettability Examples of the contact angles measured for the flat and the coated biomimetic surfaces are shown in Figure 6 for the advancing ( Figure 6(a,e) ) and receding ( Figure 6(b,f) contact angle in the perpendicular direction as well as the advancing ( Figure 6(c,g) ) and receding ( Figure 6(d,h) ) contact angles in the parallel direction, respectively. By measuring these values, two relevant parameters were measured: anisotropy and hysteresis.\n Figure 6. Optical images of the wettability of the flat uncoated surface: (a) perpendicular advancing, (b) perpendicular receding, (c) parallel advancing, and (d) parallel receding; and biomimetic surface (100µm coated microchannels): (e) perpendicular advancing, (f) perpendicular receding, (g) parallel advancing, and (h) parallel receding. Trapped air in the biomimetic surface can be observed in images (e) and (f) in the blue circle. the printed microchannel direction (view) is indicated by a red arrow ( ↑ perpendicular and → parallel). \n Uncoated surfaces The contact angles obtained for the perpendicular orientation are summarized in Figure 7a . The flat surfaces display values of 106±7° and 96±8° for the advancing and receding contact angle, respectively, with an average hysteresis value of 10°. The hydrophobic behavior observed for the flat uncoated surfaces can be explained through the chemical composition of the polymeric resin main component (urethane dimethacrylate), which structure is mainly apolar.\n Figure 7. (A) Contact angle measurements (advancing and receding) perpendicular to the printing filaments, (b) contact angle hysteresis measurements for the perpendicular and parallel orientations, (c) parallel measurements of the contact angle (advancing and receding), and (d) contact angle anisotropy (difference between the perpendicular and parallel direction measurements for the advancing and receding contact angle) for the flat surfaces and the ones with microchannels (50, 75, and 100µm), with and without TiO 2 -HTMS coating. Surfaces having microchannels with a depth of 50 and 75µm maintain an advancing contact angle of around 106±5°, but these samples decrease the receding contact angle to 85±11° and 82±8°, respectively, thus increasing the hysteresis to around 21° ( Figure 7b ). Surfaces with 100µm microchannels exhibit an increased advancing and receding contact angle of about 121±7° and 102±4°, respectively, which implies a hysteresis of 18°. Hysteresis results from the higher adhesion of a drop during its volume reduction or receding process ( Figure 6b ) and contact angle measurements show that the topography increases this adhesion thus duplicating the hysteresis (~20°) as compared with the flat samples (10°). All uncoated samples (including the flat surface) show advancing contact angles associated with a hydrophobic wettability (>90°), with microchannels of 100µm presenting the highest hydrophobicity (15° higher values compared to the flat sample) thus confirming the relevance of the topography. Receding contact angles behave hydrophobically for the flat surface and microchannels of 100µm (96 ° and 102 °, respectively), while the surfaces with microchannels of 50 and 75µm presented hydrophilic values <90°. The contact angles for the parallel orientation are presented in Figure 7c shows lower values (less hydrophobic) compared to the perpendicular measurements for all uncoated samples. In this case, the drop of water does not have to move through changing topography during the dynamic tests, meanwhile, for the perpendicular tests, the drop must move alternately between the peaks and valleys of the riblets. For instance, for all uncoated surfaces, the advancing contact angles have values lower than 74±7°, while the receding contact angles decreased to values lower than 60±5°. Despite these changes, hysteresis values for both orientations are similar in the case of the uncoated surfaces ( Figure 7b ). This implies that the pinning force between the material and the fluid does not depend on the orientation of observation. The results of the anisotropy measurements are presented in Figure 7d . The flat surface shows a difference of 32±4° and 35±4° between perpendicular and parallel measurements for the advancing and receding contact angle, respectively. This confirms that the printing filaments create a slightly anisotropic surface, as can be seen in Figure 6(a–d) . This behavior is amplified by the presence of microchannels, which increase the anisotropy to values around 45±4° for the advancing and receding angles (50 and 75µm microchannels), and 55±4° (100µm microchannels), respectively. This tendency results from the higher perpendicular contact angles obtained for increased structural depths while keeping the contact angle constant in the parallel direction across all depths fabricated.\n Coated surfaces Perpendicular measurements show that all coated surfaces become more hydrophobic as compared to the uncoated surfaces ( Figure 7a ). The flat-coated surface increases its advancing contact angle to 130±21° (25° higher than uncoated surface), and all microchannels presented angles higher than 161±4°, which implies a superhydrophobic behavior, surpassing the values obtained by previous studies [ 2 , 8 , 23 ]. In the case of the microchannels with a depth of 100µm, an average value of 165±4° was obtained, which coincides with the characteristic value of 164° of the actual rice leaves. The receding contact angle presented a similar tendency increasing to 128±20° for the flat-coated surface and to values higher than 155±7° for all microchannels. The hydrophobic coating decreased the hysteresis values ( Figure 7b ) to 2° and lower than 9° for the flat and microchannel surfaces, respectively, which is a further requirement for superhydrophobic surfaces (<10°) [ 15 ]. Pictures from these tests show that air pockets exist at the interface between the drop and the surface explaining this superhydrophobic behavior through a Cassie-Baxter state ( Figure 6e ). The advancing contact angle in parallel orientation ( Figure 7c ) also increased significantly after coating to 124±20° for the flat surface (50° higher) and to 160±8° (100° higher) for all microchannels. The receding contact angle values increased to 117±21° for the flat surface (57° higher) and to >153±4° for all microchannels (at least 113° higher). These higher differences compared to the perpendicular values mean that the coating significantly decreases the adhesion of the fluid to the surface in the parallel direction. Hysteresis values ( Figure 7b ) are lower than 7° for all coated surfaces and do not show significant differences compared to the perpendicular measurements. Anisotropy measurements ( Figure 7d ) show that the coating has a pronounced effect on all surfaces. Flat samples reduce their anisotropy values to 7±5° and 16±4° for the advancing and receding contact angle, respectively (over 20° difference compared to the uncoated case). Microchannels having a depth of 50 and 75µm decrease their anisotropy values to <4±2° (compared to 45° for the uncoated case), while 100µm microchannel with a depth of 100µm show values of 5±4° and 2±1° for the advancing and receding contact angle, respectively (compared to 55° for the uncoated case). This tendency coincides with the behavior of both a natural rice leaf with an advancing contact angle anisotropy of 6° and previous biomimetic rice leaf surfaces developed via mold imprinting coated with silica nanoparticles (5°) [ 2 ]. The decreased anisotropy after coating confirms the presence of air pockets at the interface that avoid the direct interaction between the fluid and the surface. Regarding previous investigations, Figure 8 shows the comparison between the advancing contact angle measurements obtained for Bixler et al. [ 8 ], Lee et al. [ 23 ], Wu et al. [ 2 ], Lee et al. [ 18 ], and this research. Bixler et al. [ 8 ] report a contact angle of 155 ° for a polymeric urethane surface molded with a natural sample of the rice leaf and coated with silica nanoparticles, which implies an increase of 37 ° compared to the uncoated sample. This increase is less than the one obtained for the material developed in this work, which corresponds to 44 ° for the surface with 100µm microchannels. Regarding the hysteresis of the contact angle, the authors report around 8 ° for the coated biomimetic surface, a value very similar to the 9 ° obtained for the perpendicular orientation of the newly developed material. This indicates that both materials have a very similar adhesion to a drop of water, reflecting the great biomimetic aspect of the material developed in the present investigation compared to a surface developed with a mold of a natural rice leaf. However, the perpendicular contact angle obtained in this work exceeds the reference by 10 °.\n Figure 8. Comparison between the advancing contact angle (perpendicular direction) obtained for previous investigations and this research. Lee et al. [ 23 ] reported a contact angle of 114 ° for a smooth PLA material printed in 3D by FDM and coated with silica nanoparticles, which is less than the advancing contact angle measured on the flat-coated surface developed in this work both in the perpendicular and parallel direction (130 and 124 °, respectively). In the case of the surfaces presented by the authors with 100µm microchannels, they reported a perpendicular contact angle of 157 ° and parallel contact angle of 150 °, lower values than those reported in this work (165 and 160 °, respectively). Although the authors do not report hysteresis values for the contact angle, they do report the sliding angle, which corresponds to 12 ° for the perpendicular orientation and 10 ° for the orientation parallel to the microchannels. This does not coincide with the <5 ° limit required in the slip angle for superhydrophobic surfaces that do not adhere to water, therefore this material does not meet the requirements to be non-adherent and possible reducer of drag forces. Wu et al. [ 2 ] developed biomimetic structures of the rice leaf using PDMS molds and silica nanoparticles, using different dimensions (10–60µm deep and 20–200µm wide). They report a contact angle of 160 ° (5 ° less than in this work) and measure the adhesion of the surface to a drop of water by the sliding angle. These surfaces, as they are not printed, do not show roughness due to the presence of printing filaments, and only show a hierarchical two-level structure (microchannel – nanoparticle). The authors indicate that the anisotropy of the sliding angle increases with the depth of the microchannels and the spacing between them, the latter being the most relevant parameter. All reported sliding angle values are <8 °, which implies that all configurations show low adhesion to a drop of water. Lee et al. [ 18 ] reported micro-grooves 40µm wide (PDMS) coated with nano-silica, which depending on the number of coating layers can have a contact angle of 170 ° (RMS roughness 80nm), a value that can even be exceeded when the coating generates an isotropic surface (completely covers the microchannels). For the best anisotropic case (roughness RMS 80nm, 9 coating layers), the hysteresis of the contact angle has a value close to 10 °, thus fulfilling the requirement for a superhydrophobic material that does not adhere to a drop of water. However, for the case with 2 coating layers (comparable with the present work), which corresponds to an RMS between 20–40nm as indicated by the authors, the static contact angle barely reaches 150 ° for the perpendicular orientation (<150 ° for the parallel orientation, at least 15 ° less than in this work), and the authors indicate that the droplets adhere to the surface. According to this, the material presented in this work does not exceed the wettability properties granted by the hierarchical structure developed in the present investigation. Therefore, superhydrophobic biomimetic materials can be designed through an SLA 3D printer and a nano-coating that is different than previous investigations that mimicked a rice leaf hierarchical structure surface and wettability using silica nanoparticles [ 2 , 8 , 23 ]. The contact angle results confirm that only the printing filaments, the printed microstructure, or the nanoparticle coating, cannot reach a superhydrophobic wettability by themselves or combine only two of them. Therefore, the three-level hierarchical structure must be present to obtain the desired superhydrophobic behavior. Hereafter, complementary experiments were carried out on the 3D printed biomimetic surface with the highest depth of 100µm resulting in the greatest superhydrophobicity. 3.3. Dynamic experiments 3.3.1. Dynamics of water drops Dynamic experiments were conducted on an inclined surface (25°) to observe the movement of a water droplet (having methylene blue for visualization purposes) using both the uncoated 3D printed flat surface and the 3D printed biomimetic rice leaf material (Supporting Video 1). On the uncoated hydrophilic 3D printed flat surface, the water droplets spread and attach to the material ( Figure 9a ), while on the biomimetic surface the water droplets do not experience resistance to motion thus flowing over the material ( Figure 9b ).\n Figure 9. Dynamic adhesion experiments: (a) flat uncoated surface with inclination, water drops (blue) are placed over the material with a pipette (blue) and they adhere to it. (b) inclined biomimetic surface (100µm designed coated microchannels), a water drop (blue) is placed over the surface, and it rolls off without any resistance to motion induced by the material. This behavior agrees with the results from the contact angle measurements ( Figure 7 ) as the presence of a three-level hierarchical structure formed by the printing filaments, 3D printed microchannels, and the nanoparticle coating renders superhydrophobic characteristics to the material due to the presence of air pockets at the interface. The anisotropy of the biomimetic surface was studied measuring the velocity of drops that roll over inclined surfaces, in the perpendicular and parallel direction to the printed microchannels ( Figure 10 , Supporting Video 2). Three different drops with an average volume of 14µL were placed over the inclined surface (25 °) in each direction, and then a video was recorded using a high-speed camera (2873 fps). Results show that drops have an average velocity of 239mm/s in the parallel direction, and 154mm/s in the perpendicular direction (55% higher in the parallel direction). This represents the natural behavior of water drops that roll over rice leaves, which prefer the parallel direction to the microchannels, as riblets act as a guide for the droplets to move over the surface.\n Figure 10. Drop velocity comparison for the biomimetic inclined surface (25°) between the parallel and perpendicular direction to the printed microchannels. the drops have a volume of approximately 14µl, the surface has a width of 20mm and a length of 40mm. 3.3.2. Self-Cleaning properties One of the main properties arising from natural superhydrophobic surfaces is self-cleaning, which can be observed in leaves that remove pollution by a simple rain shower as water drops roll over the surface thus removing the contaminants [ 75 ]. In plants, hydrophobic self-cleaning surfaces may both protect against harmful microorganisms, which are growth-inhibited by dry plant surfaces and ensure efficient gas exchange by keeping a thin film of air clinging to the surface when the leaves are submerged [ 76 ]. This effect has been drawing attention in the last few years for different applications in materials science, such as fabrics, photovoltaic panels surfaces, anti-corrosion, and fog-harvesting surfaces, among others [ 77–79 ]. To test the self-cleaning behavior in our case, activated carbon powder was scattered over inclined surfaces (both flat and biomimetic 3D printed materials) mimicking organic contamination, and then water drops were added (Supporting Video 3) [ 32 ]. In the case of the uncoated inclined flat 3D reference surface ( Figure 11a ), the water drops adhered to the material. Although they adsorbed the carbon powder, the contaminant was not removed from the surface (even after the addition of 30 drops). In contrast, when water drops were located on the top of the inclined 3D printed biomimetic surfaces ( Figure 11b ), they rolled off removing the contaminant. The surface stayed with a very low amount of carbon after the addition of just 30 water drops, mimicking the self-cleaning characteristic of the rice leaf. This is due to the decreased shear force that the biomimetic surface exerts over the water drops allowing them to roll over it easily due to the air pockets existing between the material and the fluid ( Figure 6e ). This self-cleaning property has been previously reported on flat plates coated with TiO 2 nanowires [ 32 ], PDMS biomimetic surfaces coated with silica nanoparticles [ 8 ], perfluorosilane-coated titanium dioxide nanoparticles sprays [ 33 ], and perfluorodecyltriethoxysilane-modified silica/PDMS sprays [ 80 ], all without the involvement of 3D printing, and on 3D multijet printed biomimetic feathers of the bird Anser cygnoides domesticus [ 81 ].\n Figure 11. (A) Self-cleaning experiment for the flat printed surface inclined and covered with activated carbon and washed with distilled water. as the drops adhere to the surface, the contaminant cannot be removed. (b) Self-cleaning experiment for the biomimetic surface (100µm designed coated microchannels) inclined and covered with activated carbon. as the water drops do not adhere to the surface and roll off easily, they carry the contaminant away and clean the surface. (c) Dynamic adhesion experiment for an inclined biomimetic surface after 30minutes of UV light radiation, drops of water (blue) are placed over the surface and they adhere to the material. 3.3.3. Effect of UV-A irradiation Previous investigations have shown that TiO 2 coated surfaces can change their wettability when exposed to UV light irradiation going from superhydrophobic to hydrophilic behavior, because of the photocatalytic properties of TiO 2 nanoparticles [ 32 ]. Our rice leaf biomimetic surface was placed under UV-A light radiation (365nm, 48 W) for 30minutes and then, the dynamic adhesion experiment was repeated ( Figure 11c ) showing a dramatic change in the surface behavior since the drops adhered to the surface after irradiation instead of freely rolling off (see Figure 9b ). This change from superhydrophobic to hydrophilic behavior can be explained through the photocatalytic properties of TiO 2 molecules, which increase the -OH groups after UV-A irradiation [ 82 ]. These groups can react and easily attach water molecules, which explains the affinity of the surface for water droplets. The UV irradiation effect provides the material a tunable wettability and can increase the application possibilities for the new biomimetic surface. An example of this is mist collecting surfaces, which are characterized by combining superhydrophobic and hydrophilic areas to direct dew drops that settle on the surface, for which they stand out as a low-cost, accessible, and energy-saving mechanism, that provides a feasible alternative to obtain freshwater [ 83 , 84 ]. 3.4. Floatability experiments Floatability is observed in nature in the insect known as water strider ( Gerris remigis ), which can float and walk over water due to its superhydrophobic legs that trap air on its surface [ 85 ]. The trapped air enabling the superhydrophobic behavior also influences the dynamics of a submerged object. This was verified in a study presented by McHale et al. [ 86 ] in which the air plastron (i.e. trapped air) formed over the surface of submerged spheres allowed up to 10% of drag reduction, observed as an increment on their terminal velocity. Dong et al. tested the effect of a superhydrophobic surface on the velocity of model ships, obtaining 39% drag reduction compared to uncoated flat surfaces [ 87 ]. A small boat ( Figure 12a ) was 3D printed to confirm the drag reduction effect in our biomimetic superhydrophobic surface by floatability experiments as compared to a standard 3D printed surface. Three types of 3D printed surfaces were tested: flat uncoated, and both uncoated and coated 100µm microchannels. The boats, having a weight of .8 grams and measuring 10mm of length, 4mm of height, and 2mm thick, were gently placed over water (with methylene blue to improve the visual analysis) and a picture was taken when their movement was stabilized ( Figure 12(b–d) . Both uncoated boats reached an equilibrium characterized by a partially submerged state without surpassing the water/air interface ( Figure 12(b,c) ), meaning a floating state, for which the bottom and the lateral parts of the boats were in contact with water. In contrast, the biomimetic boat ( Figure 12d ) perfectly floated over the water surface, and only the bottom surface was in contact with it (approximately 15% of the boat volume), where the keel of the boat is partially submerged with trapped air. This means that the biomimetic surface decreases the contact area of the boat with the fluid, therefore decreasing the drag force that the fluid can exhort over the structure [ 88 ], which implies a decrease in the thrust force exerted by the water proportional to the decrease in the submerged volume (~85%). This means that the biomimetic boat could carry on its surface a mass of water equivalent to 85% of its volume before being submerged in a position equivalent to the flat boat. This effect combined with the fact that air pockets exist inside the biomimetic microchannels ( Figures 6e and 12d ) reduces the resistance to motion of an object in the water.\n Figure 12. Floatability experiments for a small boat (a) on water (colorized with methylene blue) for an uncoated flat surface(b), 100µm microchannel uncoated surface (c), and a rice leaf biomimetic surface (100µm microchannel coated surface) (d). the position of the boat is indicated by a white circle. Trapped air can be observed at the keel of the biomimetic boat (d). 3.5. Recyclability test The capability of the biomimetic material to maintain its wettability after being in contact with a drop of water was tested. A drop of water (4µL) was placed on the same spot over the biomimetic surface, then the perpendicular contact angle was measured, and the drop was removed being absorbed with paper. The process was repeated every 2minutes until 30 cycles were reached. This experiment was repeated on three different spots over the surface. The averaged results, presented in Figure 13 , show that the surface goes from superhydrophobic to hydrophobic (contact angle >140°), as the surface decreases its contact angle 19°. As the printed structure remains the same, the change in the contact angle value is due to the detachment of the nanoparticles, which can be solved by a simple reapplication of the coating. Also, the wettability of the surface can be affected by the photocatalytic activity of the TiO2 nanoparticles ( Figure 11c ), so the material must be kept away from sunlight.\n Figure 13. Recyclability experiment for the perpendicular contact angle of the biomimetic surface. a drop of water (4µl) was placed over the material 30 times (cycles). 3.6. Drop bouncing dynamics Drop bouncing dynamics defines the capability of a surface to repel water droplets during their impact as observed in practical applications such as self-cleaning, anti-icing, anti-corrosion, and water repellent surfaces in general [ 89 ]. A water drop, having a volume of 6µL, was released over the four different surfaces using the goniometer dispensation system, and then the bouncing dynamics were recorded at 2873 fps (see Supporting Video 4). The releasing height ( h ) of this experiment was varied from 20 to 100mm (distance from the tip of the needle to the surface), which translates to a range of Weber number ( W e = ρ ν 2 D σ ) between 11 and 57. Depending on the interactions between the water drop and the surface, two main processes can be observed: adhesion on the surface or a conventional bouncing. In the case of the flat-uncoated and microchannel-uncoated surfaces for all heights, and for the flat-coated surface for h >20, the drop adheres to the material. This behavior means that the material does not repel falling water drops. These results coincide with the observations made in the inclined surface experiments, where the drops do not roll over. Otherwise, in the case of the flat-coated surface with h =20, the drop bounces having a contact time of 14ms and a bounce dynamic, measured through the pancake quality, of Q = D m i n D m a x =.5 (where D min is the minimum jumping diameter and D max is the maximum spreading diameter), which is in the range of conventional bouncing [ 90 ]). In the case of the microchannel-coated surface (biomimetic) for all the heights, the drop bounces off the material showing a conventional bouncing. The material therefore repels the fluid, where the drop spreads over the surface showing a pancake shape ( Q <.6 ) during impact but regroups and jumps recovering the shape of a droplet. Figure 14 displays photograms showing the highest position reached by the drop during the bouncing dynamic in the case of h =60mm for the different surfaces, where the drop only bounces off the biomimetic surface ( Figure 14d ).\n Figure 14. Drop bouncing dynamics comparison (first contact and highest position): (a) flat uncoated, (b) microchannels uncoated, (c) flat coated, and (d) microchannels coated (biomimetic). the drop, having a volume of 6µl (scale bar showing 10mm), falls from a height of 60mm ( We =34), only bouncing off the biomimetic surface (d). Figure 15 presents the contact time and pancake quality for each Weber number found during the test on the biomimetic material, which values coincide with previous studies [ 89–91 ]. When the velocity of the falling drop increases (higher heights), the contact time also increases, and the pancake quality decreases, due to the penetration of the fluid inside the air gaps that the biomimetic surface allows. An increase in the contact area implies an increment in the shear force over the drop’s surface, but the kinetic energy of the drop allows for a higher bounce. Because of this, the biomimetic surface repels falling drops at a range of Weber numbers. This observation, together with the previous dynamic experiments, shows that the new material has a promising prospect in dynamic water-repelling applications.\n Figure 15. Contact time and pancake quality ( Q ) for the biomimetic surface (100µm microchannels coated), for drop releasing heights from 20 to 100mm (11 < We <57). Conventional bouncing is observed in all the experiments ( Q <.6), where contact time increases with the impact velocity of the drop (6µl)."
} | 13,653 |
31438681 | PMC6746187 | pmc | 3,575 | {
"abstract": "DNA nanotechnology\nallows for the realization of complex nanoarchitectures\nin which the spatial arrangements of different constituents and most\nfunctions can be enabled by DNA. When optically active components\nare integrated in such systems, the resulting nanoarchitectures not\nonly provide great insights into the self-assembly of nanoscale elements\nin a systematic way but also impart tailored optical functionality\nto DNA origami. In this Letter, we demonstrate DNA-assembled multilayer\nnanosystems, which can carry out coordinated and reversible sliding\nmotion powered by DNA fuels. Gold nanoparticles cross-link DNA origami\nfilaments to define the configurations of the multilayer nanoarchitectures\nas well as to mediate relative sliding between the neighboring origami\nfilaments. Meanwhile, the gold nanoparticles serve as optical probes\nto dynamically interact with the fluorophores tethered on the filaments,\nrendering in situ detection of the stepwise sliding processes possible.\nThis work seeds the basis to implement DNA-assembled complex optical\nnanoarchitectures with programmability and addressability, advancing\nthe field with new momentum."
} | 288 |
29170661 | PMC5684177 | pmc | 3,576 | {
"abstract": "Rhizobia form symbiotic nitrogen-fixing nodules on leguminous plants, which provides an important source of fixed nitrogen input into the soil ecosystem. The improvement of symbiotic nitrogen fixation is one of the main challenges facing agriculture research. Doing so will reduce the usage of chemical nitrogen fertilizer, contributing to the development of sustainable agriculture practices to deal with the increasing global human population. Sociomicrobiological studies of rhizobia have become a model for the study of the evolution of mutualistic interactions. The exploitation of the wide range of social interactions rhizobia establish among themselves, with the soil and root microbiota, and with the host plant, could constitute a great advantage in the development of a new generation of highly effective rhizobia inoculants. Here, we provide a brief overview of the current knowledge on three main aspects of rhizobia interaction: trade of fixed nitrogen with the plant; diplomacy in terms of communication and possible synergistic effects; and warfare, as antagonism and plant control over symbiosis. Then, we propose new areas of investigation and the selection of strains based on the combination of the genetic determinants for the relevant rhizobia symbiotic behavioral phenotypes."
} | 324 |
40009242 | PMC11865136 | pmc | 3,577 | {
"abstract": "Arbuscular mycorrhizal (AM) fungi (phylum Glomeromycota) are obligate symbionts with plants influencing plant health, soil a(biotic) processes, and ecosystem functioning. Despite advancements in molecular techniques, understanding the role of AM fungal communities on a(biotic) processes based on AM fungal taxonomy remains challenging. This review advocates for a standardized trait-based framework to elucidate the life-history traits of AM fungi, focusing on their roles in three dimensions: host plants, soil, and AM fungal ecology. We define morphological, physiological, and genetic key traits, explore their functional roles and propose methodologies for their consistent measurement, enabling cross-study comparisons towards improved predictability of ecological function. We aim for this review to lay the groundwork for establishing a baseline of AM fungal trait responses under varying environmental conditions. Furthermore, we emphasize the need to include underrepresented taxa in research and utilize advances in machine learning and microphotography for data standardization."
} | 272 |
34357223 | PMC8307917 | pmc | 3,578 | {
"abstract": "Sensors are an important part of the organization required for robots to perceive the external environment. Self-powered sensors can be used to implement energy-saving strategies in robots and reduce their power consumption, owing to their low-power consumption characteristics. The triboelectric nanogenerator (TENG) and piezoelectric transducer (PE) are important implementations of self-powered sensors. Hybrid sensors combine the advantages of the PE and TENG to achieve higher sensitivity, wider measurement range, and better output characteristics. This paper summarizes the principles and research status of pressure sensors, displacement sensors, and three-dimensional (3D) acceleration sensors based on the self-powered TENG, PE, and hybrid sensors. Additionally, the basic working principles of the PE and TENG are introduced, and the challenges and problems in the development of PE, TENG, and hybrid sensors in the robotics field are discussed with regard to the principles of the self-powered pressure sensors, displacement sensors, and 3D acceleration sensors applied to robots.",
"conclusion": "6. Concluding Remarks This paper summarizes the pressure sensors, motion sensors, and 3D acceleration sensors that have been used in robots in recent years, and briefly describes the working principles, structures, and characteristics of these sensors. As a research hotspot in recent years, self-powered sensors have developed rapidly. Among them, the PE and TENG are important self-powered sensors and have been extensively investigated. Research on PE sensors started relatively early, and relatively mature technologies exist with regard to the materials, structure, and signal processing of self-powered sensors. Particularly, acceleration sensors have been widely used. In recent years, PENG research has greatly improved the sensitivity of PE sensors, particularly with regard to pressure sensors. Owing to the structural characteristics of its material, the PE is mainly used in high-precision displacement sensors and small-range displacement detection applications, and has certain frequency requirements. Owing to their structural characteristics, TENG sensors have great development potential as self-powered pressure sensors, self-powered displacement sensors, and self-powered acceleration sensors. Among them, the TENG self-powered displacement sensor has unique advantages. The TENG displacement sensor has a large range, even in high-precision displacement measurement, and good output characteristics. The PE–TENG hybrid sensor combines the advantages of the PE and TENG, and has higher sensitivity, wider measurement range, and better output characteristics. It is important that different structural designs of the PE–TENG hybrid can detect more information regarding various robot movements. Self-powered sensors also have unique advantages in low-power electronic devices. It is very important for low power electronic devices to improve energy conversion efficiency through the optimization of self-powered sensors. In recent years, there have been many achievements in the research on improving the interface of self-powered sensors [ 114 , 115 ] for enhancing the energy conversion efficiency [ 116 , 117 , 118 ]. Because TENG sensors will be used in practical or even commercial applications, many aspects should be investigated in future work. The PE–TENG hybrid acceleration sensor still requires extensive research with regard to miniaturization. The size of the hybrid sensor is still relatively large, and reducing the volume of the self-powered sensor will inevitably reduce the sensor’s output. Maintaining higher sensor output with smaller sensor volume is also a topic that should be investigated. In sensor research, self-powered hybrid sensors use the advantages of different sensors to greatly improve their measurement range and output characteristics. Hence, self-powered hybrid sensors are also a promising field for future sensor research.",
"introduction": "1. Introduction With the development of science and technology, intelligent robots have been developed through the fusion of multiple technologies, including machine vision technology, communication technology, sensor technology, and even biotechnology. Intelligent robots have been introduced into various fields, in addition to the traditional industrial manufacturing fields [ 1 ], such as underwater detection, medical treatment, service, and other fields [ 2 , 3 , 4 , 5 ]. The working environment of intelligent robots is also becoming increasingly complex. To improve the intelligence of robots, human control interventions must be reduced when the environment changes, the robot’s perception of the working environment and its own state must be improved, and corresponding feedback actions and posture adjustment must be carried out according to the situation. The robots’ detection of the environment mainly relies on vision and various sensor types. Machine vision allows robots to approximately model the environment, but the perception accuracy is low. For example, when a robot is grabbing objects, it cannot directly understand the grasping force through vision. Although machine learning is used to replace touch sensors for object perception [ 6 , 7 , 8 ], machine vision is far from sufficient for achieving precise control and coping with complex and changing environments. In recent years, research on soft robots, and particularly research on soft touch sensors applied to robots [ 9 , 10 , 11 ], has intensified. Compared with traditional touch sensors, soft touch sensors have higher sensitivity and can perceive more subtle tactile information from the external environment. Moreover, soft sensors have a lower Young’s modulus and can better adapt to the environment through deformation [ 12 ]. In particular, the interaction between robots and humans has become safer and more comfortable [ 1 , 13 , 14 ]. Robots need increasingly more sensors [ 14 ], which requires smaller sensors, higher integration, and lower power consumption. Because self-powered sensors can use their own energy to provide information, they have lower power consumption and only need two wires for signal transmission, and can even can realize signal transmission with only one wire. Therefore, self-powered sensors have been widely investigated [ 15 , 16 , 17 ]. Piezoelectric and triboelectric energy are important implementation methods for self-powered sensors. The piezoelectric transducer (PE) uses piezoelectric materials. When the piezoelectric materials deform, they are internally polarized. Two types of piezoelectric materials are typically used, namely crystal materials and ceramic materials. Crystal materials have limitations such as high production cost, sensitivity to moisture, and difficulty in obtaining consistent orientations. However, polycrystalline ceramic materials are easy to process [ 17 ], and have relatively low manufacturing cost, higher sensitivity than crystals [ 18 ], and a wide range of applications. In recent years, structures such as nanofibers [ 19 ], nanowires [ 20 ], nanoparticles [ 21 ], nanocubes [ 22 ], and nanorods [ 23 ] have been used in piezoelectric material structures. Among them, nanofibers and nanowires can effectively convert the strain caused by low-strength mechanical external forces into electrical energy [ 19 ]. This piezoelectric transducer with nanostructures is known as the piezoelectric nanogenerator (PENG). Piezoelectric materials combined with self-sensing technology can be used as a sensor while acting as an actuator [ 24 , 25 ]. The main advantages of self-sensing are the embeddability of measurement technology, low cost and lack of a need for additional sensors [ 24 ]. Through piezoelectric actuator self-sensing technology, precise positioning applications such as micromanipulation and microassembly can be realized [ 25 , 26 , 27 , 28 , 29 , 30 ]. The main source of triboelectricity is the principle of triboelectric nanogenerators (TENGs). The triboelectric nanogenerator invented by Wang et al. in 2012 uses the triboelectric effect and electrostatic induction effect to convert mechanical energy into electrical energy [ 31 ]. The TENG structure is simple, easy to manufacture, and has low cost. Owing to these characteristics, it is more convenient to use a TENG to convert wind energy [ 32 ], water wave energy [ 33 ], and the mechanical energy of human movement into electrical energy [ 34 ]. As a self-powered sensor, the TENG has good output characteristics, as has been demonstrated by studies considering pressure sensors, displacement sensors, and inertial sensors. Additionally, because the TENG can flexibly select materials, and the materials can be soft materials [ 35 ], it can be effectively applied to soft robots [ 36 ]. Owing to their particular structure and material characteristics, the PE sensor and TENG have their particular advantages and disadvantages in different application scenarios. Sensors made by combining the PE and TENG have higher sensitivity, wider measurement range and better output characteristics [ 37 ]. Owing to their outstanding characteristics, the PENG and TENG have been extensively investigated in recent years, and it has been found that these sensors have good development potential [ 38 ]. The artificial intelligence design model provides a new method for designing, predicting and optimizing the structure and materials of PENG and TENG [ 39 ]. The sensors implemented in robots mainly include pressure sensors, displacement sensors, and three-dimensional (3D) acceleration sensors. The motion perception and environment perception of a robot are inseparable from these three types of sensors. Based on research pertaining to PE, TENG, and PE–TENG hybrid (PTENG) sensors, this paper summarizes the principles used by these three sensors to realize pressure sensing, displacement sensing, and three-dimensional (3D) acceleration sensing, and their corresponding applications in the field of robotics ( Figure 1 ). In the first part, the principles of the PE and TENG are briefly introduced. The next few sections introduce the research results obtained for PE, TENG, and hybrid self-generation sensors in the field of pressure sensors, displacement sensors, and spatial acceleration sensors over the last year. Additionally, the sensors’ potential applications in the field of robotics are introduced. Finally, the opportunities, challenges, and future expectations for PE sensors, TENG sensors, and PE–TENG hybrid sensors in the practical application of robot perception are introduced."
} | 2,650 |
38444203 | PMC10950044 | pmc | 3,579 | {
"abstract": "Abstract Microbial communities are not only shaped by the diversity of microorganisms and their individual metabolic potential, but also by the vast amount of intra- and interspecies interactions that can occur pairwise interactions among microorganisms, we suggest that more attention should be drawn towards the effects on the entire microbiome that emerge from individual interactions between community members. The production of certain metabolites that can be tied to a specific microbe-microbe interaction might subsequently influence the physicochemical parameters of the habitat, stimulate a change in the trophic network of the community or create new micro-habitats through the formation of biofilms, similar to the production of antimicrobial substances which might negatively affect only one microorganism but cause a ripple effect on the abundance of other community members. Here, we argue that combining established as well as innovative laboratory and computational methods is needed to predict novel interactions and assess their secondary effects. Such efforts will enable future microbiome studies to expand our knowledge on the dynamics of complex microbial communities.",
"introduction": "Introduction Microorganisms seldom live as isolated species. Rather they co-exist in a community with other microorganisms, within small consortia or as part of complex cohorts that can span all domains of life (Faust and Raes 2012 ). Within these consortia, different types of intra- and interspecies interactions occur that can have beneficial, negative or neutral effects for one or multiple interacting participants. With that, interactions can range from being advantageous for all interacting organisms (mutualism) to having an overall negative impact on all microorganisms involved (competition) (Faust and Raes 2012 ). Ongoing discussions revolve around the prevalence of distinct microbial interactions. While some studies contend that negative interactions such as competition dominate (Foster and Bell 2012 , Palmer and Foster 2022 ), others suggest this perspective might be biased. The challenge lies in distinguishing mutualistic and commensalistic interactions (Kehe et al. 2021 , Kost et al. 2023 ). These studies propose that both positive and negative impact interactions are widespread within microbiomes, challenging the notion of negative interactions' overwhelming abundance. A multitude of conceptually different interaction mechanisms, such as the excretion of antimicrobial compounds or cross-feeding of metabolites and the formation of biofilms, are underlying each type of interaction. Traditionally, these interactions are defined at the molecular level, meaning that they include a direct exchange or conversion of molecules, such as primary or secondary metabolites, toxins, siderophores or signaling compounds between two or more organisms (Braga et al. 2016 ). The direct correlation between the exchange of molecules and a positive or negative impact on the exchanging organisms make these interactions relatively easy to quantify and classify. However, microbial interactions at the molecule level are solely centered around observing effects on the interacting players. The “emerging effects” of microbial interactions, which are less direct consequences of microbial interactions that can impact the entire community, are often overlooked, or not considered. While these emerging effects of microbial interactions are more challenging to qualify and quantify, their influence on complex communities and even their host environment can potentially be more impactful than the individual interaction itself. Here, we suggest that future studies that aim to improve our understanding of the functioning of complex microbial communities should (i) combine innovative methods to uncover novel microbial interactions that are of relevance within an environment and at the same time also (ii) consider the emerging effects that these interactions have on other community members, and thereby the community itself. The gained information in positive and negative impacts of individual microbial interactions on the entire population will drastically increase our understanding of the dynamics in complex microbial communities and enable us to better protect and utilize the ecosystem functions and services that these communities provide."
} | 1,090 |
36145836 | PMC9502053 | pmc | 3,580 | {
"abstract": "Drought generates a complex scenario worldwide in which agriculture should urgently be reframed from an integrative point of view. It includes the search for new water resources and the use of tolerant crops and genotypes, improved irrigation systems, and other less explored alternatives that are very important, such as biotechnological tools that may increase the water use efficiency. Currently, a large body of evidence highlights the role of specific strains in the main microbial rhizosphere groups (arbuscular mycorrhizal fungi, yeasts, and bacteria) on increasing the drought tolerance of their host plants through diverse plant growth-promoting (PGP) characteristics. With this background, it is possible to suggest that the joint use of distinct PGP microbes could produce positive interactions or additive beneficial effects on their host plants if their co-inoculation does not generate antagonistic responses. To date, such effects have only been partially analyzed by using single omics tools, such as genomics, metabolomics, or proteomics. However, there is a gap of information in the use of multi-omics approaches to detect interactions between PGP and host plants. This approach must be the next scale-jump in the study of the interaction of soil–plant–microorganism. In this review, we analyzed the constraints posed by drought in the framework of an increasing global demand for plant production, integrating the important role played by the rhizosphere biota as a PGP agent. Using multi-omics approaches to understand in depth the processes that occur in plants in the presence of microorganisms can allow us to modulate their combined use and drive it to increase crop yields, improving production processes to attend the growing global demand for food.",
"conclusion": "5. Perspectives and Conclusions As stated above, a large body of evidence highlights the role of specific microbial strains within the main rhizosphere microbial groups in conferring drought tolerance to plants. However, much less known are the physiological, molecular, and biochemical mechanisms and responses displayed by plants as a consequence of the presence and action of these PGP microorganisms. In the case of yeasts, these microorganisms have very interesting PGP characteristics that have not been widely reported, especially regarding water stress. Nevertheless, recent studies have demonstrated their great potential as coadjutants in plant growth to cope with other abiotic stresses. Such evidence makes us presume that yeasts can be a key element with a great biotechnological value that needs to be explored as a tool to face the food shortage that promises to be accentuated with the advance of GCC. Therefore, the next steps for the designing of biofertilizers supported by the use of multi-omics approaches could represent a significant leap in the research regarding the role of rhizosphere microbial communities in plant production under drought conditions. In this sense, the integration of omics platforms will strongly support the mechanistic understanding that underlies the use of beneficial microorganisms in plant production through rhizosphere engineering. While it is necessary to realistically recognize that the tangible results could primarily be framed at the local level, considering particular soil conditions, rhizospheres, and crop plants, the research at the pilot scale will establish the basis for the generation and massification of optimized biofertilizers. However, some considerations must be addressed, as the development of local biological collections for the ex situ maintenance of PGP microbes able to enhance drought tolerance represents a valuable resource for testing their applicability in different crops, in line with the global advice for the maintenance of genetic resources oriented to agriculture and food production [ 156 ]. Moreover, the projection of the fundamental and mechanistic bases studied in the plant, considering different efficient rhizosphere modifications, can also be focused on different types of environmental stress, such as the low availability of nutrients, salinity, heavy metals, extreme pH values, and many other environmental and soil constraints that currently affect enormous areas of arable land surfaces worldwide.",
"introduction": "1. Introduction Projections of agroclimatic models indicate a strong impact of global climate change (GCC), represented by both temperature increases of 0.5 to 2 °C by 2100 and a significant decrease in rainfall [ 1 , 2 ]. This change will undoubtedly promote a reconversion of agronomic practices to produce the necessary food for a growing world population, both in volume and requirements of high-quality products [ 3 ]. Regardless of the plant species cropped, agricultural production depends on a significant proportion of water being supplied mainly by rainfall, which has had a strong impact on the production of the last few seasons due to a marked mega-drought in many places around the world [ 1 ]. Under this complex scenario, the availability of water sources for irrigation is also a major structural problem, since the availability of adequate infrastructure to allow water storage is insufficient to ensure access to water during periods of higher demand [ 4 ]. Such constraints require the implementation of innovative agricultural approaches to increase resilience to climate variability, such as the incorporation of new species to be cropped. Additionally, lands where new crop species are chosen to be incorporated in must be evaluated in terms of alternatives to increase the “water use efficiency”, together with an increased plant tolerance to drought [ 5 , 6 , 7 , 8 ]. The above is a major challenge for global agricultural activity given the extent of the predicted climate change effects on agriculture [ 9 ]. Consequently, one of the major questions to be answered under this complex paradigm is not how to access more water in the short–medium term (that could only be answered at the infrastructural level and in the long term) but how it would be possible to keep the current yields of plant production with the scarce water availability, or even how to increase them based on the projected demand. An alternative that has generated much interest in recent years is the use of plant growth-promoting (PGP) microorganisms as inoculants (biofertilizers) in agricultural plants [ 6 , 7 , 10 , 11 ]. The above is based on the numerous microbial strains that have developed tolerance to the environmental stresses in which they commonly are exposed. This, together with the microbial fast growth and multiplication rate, can generate a high abundance of some desirable traits in a very short time. Among the multiple PGP microbial groups, the arbuscular mycorrhizal (AM) fungi stand out because they establish strict mutualist symbiosis with most of the agricultural plant species, mainly characterized by providing mineral nutrition (as P) and water transport to the host plants [ 5 , 12 , 13 ]. Moreover, the development of biotechnological tools using AM fungi constitutes one of the most environmentally friendly alternatives to address the above-described constraints in a context of resilient agriculture [ 12 , 14 , 15 , 16 , 17 ]. Additionally, other free-living microorganisms such as yeast and bacteria present diverse PGP traits, which also support their use as biofertilizers. While their beneficial effects on plant growth have been extensively described, they also produce effects in plants that can be based on (or can promote) molecular, biochemical, and physiological changes that are lesser known. Despite the increasing use of microbial inoculants from both monospecific isolates and consortia of yeast and bacteria (for instance, [ 7 , 18 ]), some points need to be addressed, such as: (a) the basis by which such microorganisms generate beneficial responses in plants; and (b) the multiple microbial or “ecological” interactions that may occur in the rhizosphere [ 19 , 20 ]. In the last case, it may not necessarily produce positive effects, being in some cases even negative [ 21 ]. Therefore, as a basis for the design of optimized bioinoculants, it is necessary to understand their degree of compatibility. This can be achieved by avoiding negative interactions such as competition or predation and promoting commensalism and cooperation [ 22 ]. Finally, the microbial functionality in generating desirable responses in the host plants will be registered as yield and food quality increases. There are previous experiences that link, at the molecular level, the responses of plants to inoculation with certain PGP microorganisms [ 11 , 23 , 24 ]. Meanwhile, to the best of our knowledge, there are no systematized works to develop microbial consortia engineered to different hosts, including microbe–microbe interactions + PGP traits + shifts in plant mechanisms, to develop optimized biofertilizers for a specific plant species. In contrast, it is commonly concluded that the effects observed by the inoculation with a PGP microorganism in a model plant could easily be transferred to other species without a deep understanding of the physiological, biochemical, or molecular changes in the host plant (regarding the uninoculated plant) [ 7 , 10 ]. The above scenario generates double uncertainty: (i) whether the inoculant can be beneficial to more than one host plant species; (ii) whether the plant responses can be mechanistically equivalent between different host plants using the same inoculant. Currently, the multiple soil–microorganism–plant interactions represent an unexplored “black box”. Much less effort has been put into the validation at the field level of an inoculant’s effectiveness beyond the argued socially friendly decrease in the use of chemical inputs, as commonly advertised by marketing agricultural companies. Against this scenario, the current “omics” tools emerge as an attainable alternative to clarify “what is happening in this black box”, allowing us to describe and predict behaviors in plants based on powerful genomic, transcriptomic, proteomic and metabolomic tools. Therefore, with this focused review, we hoped to summarize the current and updated knowledge regarding the role of microbial tools used as bioinoculants to improve the plant production. In addition, we highlight the projections based on the use of single-omics platforms and multi-omics approaches. With this, our aim was to elucidate the mechanisms that explain the improvements in yield and food quality under drought stress, as one the main sensible effects of climate change in agriculture."
} | 2,657 |
24957022 | PMC4101502 | pmc | 3,581 | {
"abstract": "The generation of efficient production strains is essential for the use of eukaryotic microalgae for biofuel production. Systems biology approaches including metabolite profiling on promising microalgal strains, will provide a better understanding of their metabolic networks, which is crucial for metabolic engineering efforts. Chlamydomonas reinhardtii represents a suited model system for this purpose. We give an overview to genetically amenable microalgal strains with the potential for biofuel production and provide a critical review of currently used protocols for metabolite profiling on Chlamydomonas . We provide our own experimental data to underpin the validity of the conclusions drawn.",
"conclusion": "8. Conclusions In this article, we have comprehensively reviewed the published protocols for GC-MS-based metabolite profiling on Chlamydomonas . Thereby, we focused, as far as accessible from the publications, on the employed harvesting methods of cells and extraction methods. In addition, we have presented data that encourage proper method establishment, and we have provided possibilities for normalizing GC-MS metabolite profiles from Chlamydomonas for subsequent reliable statistical analyses. Up to now and despite a dozen publications on this topic, the harvesting of Chlamydomonas cells for metabolic profiling has not yet reached a universally applied standard. The quenching method in cold methanol-water represents the most frequently used technique ( Table 2 ). While, apparently, quenching seems to be applicable for some relatively leakage-resistant strains, it is not recommended for the frequently used wall-less strains. In this respect, fast filtration represents the best alternative to quenching for GC-MS-based approaches. With harvesting times in the lower seconds range [ 124 , 126 ], fast filtration is superior to centrifugation, but has not yet been frequently used for metabolomics studies on Chlamydomonas ( Table 2 ). There still seems to be room for harvest method improvements, such as combinations of quenching and filtration [ 168 ], different quenching solutions [ 169 ] or switching experimental setups, such as experimentation with cells on filters, which, so far, has only been applied to yeast and E. coli [ 98 , 132 ]. As suggested by our data, measurements of quenched whole culture samples (including growth medium) are only recommended for targeted approaches with adapted growth media, where the analyzed intracellular metabolites are not masked by extracellular excess and where matrix effects can be largely excluded [ 96 , 104 ]. The buffer systems used for Chlamydomonas metabolite extraction were found to be basically very similar and consisted of methanol, chloroform and water in different combinations ( Table 2 ). The analysis of different numbers of Chlamydomonas cells processed and measured by GC-MS revealed a limited linear range for biomass. This notion seems to be of importance for quantitative comparisons, because matrix effects in complex samples may strongly and differently affect the response of individual analytes. Therefore, biomass should be balanced between samples for crucial processing steps, like extraction, derivatization and GC-MS measurement. Hence, the amount of cells to be used is a trade-off: sufficient material to be above the detection limit for most metabolites, but as little material as possible to avoid inhibitory effects by the sample matrix. Our results suggest that extracts from 1–5 × 10 5 cells finally analyzed in the GC-MS seem to meet these requirements. A thorough assessment of cellular biomass is necessary to reliably compare cellular metabolite levels between different samples, experiments or laboratories. The OD750 is an easy estimate for the biomass of Chlamydomonas ; alternatively, the total cellular volume may be less influenced by the cellular composition. Other intracellular parameters, like total protein, starch or chlorophyll contents, could be used, but bear the risk of being affected by the biological treatment. Other methods to decrease inter-sample variance rely on the properties of each chromatogram, like TIC normalization and probabilistic quotient normalization (PQN). TIC normalization and PQN, in addition to methods derived from the standardization of microarrays, may thus be more robust against biologically-induced variation of culture-specific factors, like e.g. cell count. Yet, irrespective of the applied normalization technique, the suitable scaling or transformation represents an additional prerequisite to allow for the proper statistical analysis of GC-MS metabolomic data. The use of systems biology to identify targets for the metabolic engineering of organisms is emerging [ 170 , 171 ] and attractive for the application to microalgae. Metabolomics and other high-throughput data can be used to screen microalgal strains and to refine or test predictions from genome-scale metabolic models [ 65 ]. Accordingly, GC-MS constitutes a valuable method for algal biofuel research. Nevertheless, to ensure that results are truly transferrable between experiments, organisms and laboratories, metabolomics methods, like GC-MS, should be reliably developed and should reach standards that go beyond the presently discussed issues [ 73 ].",
"introduction": "1. Introduction Fossil fuels are getting increasingly costly as their world-wide supply is declining, while global energy demands are steadily rising. Thus, extensive effort is currently being put into the development of renewable energy sources. In this context, microalgae receive significant attention as potential “biofuel crops” [ 1 ]. The intense interest in microalgae for renewable energy production has several reasons: algae are photosynthetic organisms that turn solar energy into chemical energy and other valuable products, thereby capturing CO 2 . Microalgae can be grown in open ponds or closed systems situated on non-arable land, and some species can even grow on waste or salt water, thus reducing fresh water consumption. The photosynthetic yield of microalgae is considered to be higher than that of conventional land-grown biofuel crops [ 1 , 2 ]. Microalgal diversity provides species able to produce molecular hydrogen, ethanol or triacylglycerides, which are easily accessible as biofuels [ 3 , 4 ]. However, the large diversity of microalgae has not yet been exploited for the production of biofuels, nor has the possibility of improving potential production strains by means of molecular breeding [ 1 , 2 ]. In this article, we give a short overview of microalgal strains amenable to metabolic engineering and of promise for biofuel production. We discuss the importance of systems biology approaches for the identification of targets for metabolic engineering and, in this context, focus on GC-MS-based metabolite profiling for the analysis of algal metabolism. With a focus on the model alga, Chlamydomonas reinhardtii , we review published methods for cell harvest, metabolite extraction and data normalization and support conclusions by our own experimental data."
} | 1,771 |
38994271 | PMC11238178 | pmc | 3,582 | {
"abstract": "Data-driven spiking neuronal network (SNN) models enable in-silico analysis of the nervous system at the cellular and synaptic level. Therefore, they are a key tool for elucidating the information processing principles of the brain. While extensive research has focused on developing data-driven SNN models for mammalian brains, their complexity poses challenges in achieving precision. Network topology often relies on statistical inference, and the functions of specific brain regions and supporting neuronal activities remain unclear. Additionally, these models demand huge computing facilities and their simulation speed is considerably slower than real-time. Here, we propose a lightweight data-driven SNN model that strikes a balance between simplicity and reproducibility. The model is built using a qualitative modeling approach that can reproduce key dynamics of neuronal activity. We target the Drosophila olfactory nervous system, extracting its network topology from connectome data. The model was successfully implemented on a small entry-level field-programmable gate array and simulated the activity of a network in real-time. In addition, the model reproduced olfactory associative learning, the primary function of the olfactory system, and characteristic spiking activities of different neuron types. In sum, this paper propose a method for building data-driven SNN models from biological data. Our approach reproduces the function and neuronal activities of the nervous system and is lightweight, acceleratable with dedicated hardware, making it scalable to large-scale networks. Therefore, our approach is expected to play an important role in elucidating the brain's information processing at the cellular and synaptic level through an analysis-by-construction approach. In addition, it may be applicable to edge artificial intelligence systems in the future.",
"introduction": "1 Introduction Elucidating the mechanisms underlying information processing in the brain represents a great challenge in neuroscience. In parallel to collecting data with experiments, building brain models has proven to be a powerful approach to enable in-silico analysis and provide a framework for understanding information processing in the brain. Macroscopic models (Kawato, 1999 ; Frank et al., 2001 ; Norman and O'Reilly, 2003 ; Walther and Koch, 2006 ) describe information flow at the functional level and present an overview of neural processing. In contrast, spiking neuronal network (SNN) models emulate the brain at the cellular and synaptic level and provide their in-silico counterparts, which are more tractable and easier to manipulate. From an engineering perspective, properly built SNN models are expected to be capable of intelligent information processing equivalent to the brain. Silicon neuronal network (SiNN) chips, which are highly power-efficient neuromorphic hardware optimized for SNN models, have already been developed (Merolla et al., 2014 ; Qiao et al., 2015 ; Davies et al., 2018 ). Therefore, they have great potential for next-generation artificial intelligence (AI) applications. The structure of the brain is highly diverse, which makes it demanding to capture the comprehensible rules about the network topology. In addition, a wide variety of neuronal and synaptic properties has been reported. The data-driven approach intends to replicate the brain by semi-automatically incorporating vast amounts of anatomical and physiological data. Several large-scale data-driven SNN models (Markram et al., 2015 ; Bezaire et al., 2016 ; Ecker et al., 2020 ) that reproduce a part of the mammalian cortex and hippocampus have been built. They were designed to replicate the network topology, neuronal anatomy and electrophysiology, and synaptic properties, and they successfully reproduced the characteristic spiking activities seen in the target regions. However, in mammalian brains, the considerable number of neurons makes it challenging to measure the exact connection topology between the neurons. Hence, the network topology was inferred based on statistical data. In addition, because each brain region closely interacts with various other brain regions, it is not trivial to understand the specific function of the target region. Generally, data-driven models employ the ionic-conductance-based neuronal models, which can reproduce arbitrary electrophysiological properties but incur enormous computational costs. For example, the model in Bezaire et al. ( 2016 ) runs on a supercomputer consisting of 3,488 processors, and its simulation speed is 1,600 times slower than real-time. Moreover, these models are not suitable for implementation on SiNNs because they involve complex calculation processes that require enormous circuit resources. In this study, we built a data-driven SNN model for the olfactory nervous system of Drosophila melanogaster (fruit fly). The system is a relatively small (~2,200 neurons) network having a known function, whose complete network topology, or connectome, is available. The electrophysiological activity of neurons was reproduced by using the piecewise quadratic neuron (PQN) model, which is a lightweight neuron model suitable for digital arithmetic circuit implementations (Nanami and Kohno, 2016a , b , 2023 ; Nanami et al., 2016 , 2017 , 2018 ). The PQN model was adopted to reduce the computational cost and enable the SNN model to be run on a SiNN chip. It focuses on reproducing the key dynamics behind neuronal activities with lightweight calculations. The model is designed using the dimension reduction techniques of nonlinear dynamics such that the dynamical structure of the activity of the target neuron is preserved. Unlike integrate-and-fire (I&F) based models, such as the leaky I&F model, Izhikevich (IZH) model (Izhikevich, 2003 ), and adaptive exponential I&F model (Brette and Gerstner, 2005 ), the dynamics in the neuronal spike are not replaced by a resetting of the membrane potential. I&F-based models are generally more lightweight than the PQN model. However, they have been reported to have limitations in the reproducibility of neuronal activities. For example, because their spike amplitudes are fixed, they cannot reproduce the propagation of spike intensity observed in some brain regions including the hippocampus (Alle and Geiger, 2006 ). In addition, the IZH model can only reproduce spiking within a limited range of stimulus intensities (Nanami and Kohno, 2016b ). Furthermore, the phase-resetting curve of the Class II mode in Hodgkin's classification (Hodgkin, 1948 ) of the IZH and AdEx models differs from the typical shape (Nanami and Kohno, 2023 ). In addition to the aforementioned advantages, the PQN model supports the efficient implementation on digital arithmetic circuits. Thus, the SNN model can be executed efficiently (power and speed) with a SiNN on field-programmable gate arrays (FPGAs) and application-specific integrated circuits (ASICs). The results in this study were obtained using a SiNN on an entry-level low-cost FPGA chip to demonstrate its potential for low-power brain-morphic artificial intelligence (AI) applications. In recent years, brain-inspired AI has become popular, where spike-based machine learning (Yang and Chen, 2023a , b ; Yang et al., 2023a , b ) is studied mainly using I&F-based models. These studies built massively parallel information processing systems inspired by the brain's structure to enable advanced and robust information processing with low power consumption. In contrast, here we aim to provide an in silico platform that more faithfully reproduces neuronal connectivity and information processing in brain microcircuits, which is distinct from the objective of brain-inspired AI. The fruit fly brain comprises 100,000 neurons. Moreover, its connectome was recently revealed (Scheffer et al., 2020 ). It is compact compared to the mammalian brain but capable of complex information processing. Its olfactory nervous system consists of brain regions including the antennal lobe and the mushroom body, the anatomy and physiology of which have been widely studied (Wilson, 2013 ; Modi et al., 2020 ). The function and activity of each type of neuron in these regions are better characterized in the context of sensory input and behavioral output than those of the mammalian cortex and hippocampus, enabling us to adequately verify the reproducibility of the model. However, previous modeling studies (Wessnitzer et al., 2012 ; Faghihi et al., 2017 ; Kennedy, 2019 ) (not data-driven) of the olfactory nervous system used simplified I&F-based neuron models, which did not fully reproduce the electrophysiological properties of each type of neurons. Specifically, they did not reproduce the characteristic spiking activities seen in the olfactory nervous system including (1) odor-evoked oscillatory firing in the projection neurons (PNs) and local neurons (LNs) (Tanaka et al., 2009 ), (2) absence of oscillations in Kenyon cells (KCs) (Turner et al., 2008 ), (3) different contributions of LN subclasses to the formation of oscillations (Tanaka et al., 2009 ), and (4) temporal dynamics of firing in mushroom body output neurons (MBONs) (Hige et al., 2015 ). Thus, it is uncertain whether they accurately capture information processing mechanisms in the olfactory nervous system. More sophisticated, ionic-conductance-based SNN models of the insect brain (Bazhenov et al., 2001a , b ) had been built for the antennal lobe of locust. However, they were not data-driven and did not reproduce most of the aforementioned characteristics of spiking activities. This is likely because they modeled only PNs and LNs, and also lacked electrophysiological data on identified neurons. Here we built a model of a fly olfactory system incorporating the connectome data as well as neuronal and synaptic electrophysiological properties of neurons. Our model successfully reproduced not only the aforementioned characteristic spiking activities (1)–(4) of the constituent cells, but also olfactory associative plasticity, the primary function of the olfactory system. Although we did not intend to implement every single known neuron or connection in our model, this study lays a foundation for building lightweight data-driven SNN models and is expected to aid in understanding the brain and developing brain-morphic AI systems.",
"discussion": "4 Discussion In this study, we built the first data-driven SNN model of the olfactory nervous system of Drosophila melanogaster . Our modeling approach proposed a way to overcome the trade-off between replicating the detailed biological data (the connectome and electrophysiological activities of neurons) and the computational cost, such that the model can run in real-time on a low-power SiNN chip while reproducing the characteristic neuronal activities in the brain. Features of previous data-driven models (Markram et al., 2015 ; Bezaire et al., 2016 ; Ecker et al., 2020 ) that reproduced parts of the mammalian cortex and hippocampus as well as this work are compared in Table 1 . Specifically, our model went beyond the preceding models in the following four aspects: the higher reproducibility of (1) synaptic connectivity, (2) characteristic spiking activities, (3) neuronal functions, and (4) the lower computational cost. Whereas the preceding models reproduced the electrophysiological and morphological properties of each type of neuron using multicompartmental ionic-conductance-based models, our model reproduced electrophysiological properties using the PQN model, which requires a lower computational cost. In Markram et al. ( 2015 ) and Ecker et al. ( 2020 ), the Tsodyks–Markram (TM) synapse model (Tsodyks and Markram, 1997 ) with a stochastic mechanism was used to accurately reproduce synaptic physiology, whereas in Bezaire et al. ( 2016 ), the double exponential synapse model reproduced the rising and decaying time constants of the synaptic current for each type of synaptic connection. In this study, the decay time constant of the double exponential synapse model was fitted to electrophysiological data for the corresponding type of neurotransmitter. In the preceding models, synaptic connections were randomly determined based on the position and morphology of individual neurons and statistical information for each neuron type. However, in this model, they were based on the connectome (HEM, 2020 ; Scheffer et al., 2020 ) identified from comprehensive electron microscopy images. In the preceding models, the vast number of neurons and complex structures of the mammalian brain limited the validation of the models. In Markram et al. ( 2015 ) and Bezaire et al. ( 2016 ), synchronous oscillations at the network level were validated, but not for each type of neuron. Spiking activities were not examined in Ecker et al. ( 2020 ). Additionally, the preceding models did not reproduce the function of the network, as mammalian cortical and hippocampal functions at the circuit level have not yet been elucidated. In contrast, because the olfactory nervous system has a smaller network size and its function is clearer, we were able to demonstrate that our model successfully reproduces olfactory associative learning, characteristic spiking activities of each type of neuron, such as odor-evoked oscillatory firing in PNs and LNs, absence of oscillations in KCs, different contributions of LN subclasses to the formation of oscillations, and temporal dynamics of firing in MBON-α1. Whereas the preceding models required supercomputers owing to their enormous computational cost, our model was light enough to be simulated on an entry-level low-cost FPGA chip at 0.37 watts, which may be acceptable for small robots and portable AI devices. In addition, whereas the simulation speed in Bezaire et al. ( 2016 ) was approximately 1,600 times slower than real time, our model performs real-time simulations. Table 1 Comparison of the data-driven SNN models. \n This paper \n \n (Markram et al., 2015 ) \n \n (Bezaire et al., 2016 ) \n \n (Ecker et al., 2020 ) \n Target Drosophila olfactory nervous system Microcircuit of rat neocortex CA1 of rat hippocampus CA1 of rat hippocampus Model PQN model Ionic-conductance -based model Ionic-conductance -based model Ionic-conductance -based model Scale 2,200 neurons 31,000 neurons 340,000 neurons 400,000 neurons Reproducibility of neuronal electrophysiology High High High High Reproducibility of neuronal morphology No High High High Reproducibility of synaptic properties Low (PQN synapse) High (TM with stochastic) Medium (double exponential) High (TM with stochastic) Reproducibility of synaptic connectivities High (connectome-based) Medium ( morphology-based statistical method ) Medium (neuronal distance -based statistical method ) Medium ( morphology-based statistical method ) Reproducibility of characteristic spiking activities High Medium Medium Low Reproducibility of functions High No No No Computing environment FPGA(0.3W) Supercomputer (4-rack IBM Blue Gene/Q) Supercomputer (3,488 processors) Supercomputer Simulation speed Real time Not available 1,642 times slower Not available There also are differences between our model and the latest preceding model (Kennedy, 2019 ) of the Drosophila olfactory system. Unlike our model, the preceding model is not data-driven. The preceding model used the leaky I&F model, and did not reproduce the electrophysiological properties of each class of neurons. As for the structure of the network, our model employs a slightly extended version of the preceding model. Whereas the preceding model consists of PNs, LNs, KCs, APL, and MBON, our model has another MBON and SMP354 neuron in addition, reproducing the valence-balance model (Heisenberg, 2003 ; Aso et al., 2023 ), where learning-induced plasticity in the KC>MBON synapses tips the balance of valence signals of MBONs. This competitive memory circuitry is important because it is the basis for the interactions among MBONs that are responsible for flexible and complex behavioral decisions associated with memory. As for the learning rule, both models employ reward-induced depression of KC>MBON synapses to implement olfactory associative learning. As for the synaptic connections, whereas the preceding model stochastically determines the connections between layers such as ORN>PN and PN>KC, our model precisely reproduces the connections based on the connectome database. As for the spiking dynamics, the characteristic spiking activities of each neuron are not considered in the preceding model. For example, the spiking activities of PNs and LNs are not calculated by spiking neuron models but are generated by the Poisson process. The activities of MBONs are represented using nonlinear activation functions. KCs are described by the LIF model, and their firing properties are not fitted to the in vivo data. The peak frequency of PN oscillation in this model was approximately 24 Hz, whereas experimentally observed peak frequency in the antennal lobe was 10–15 Hz (Tanaka et al., 2009 ). In the antennal lobe, PNs and LNs are connected via glomeruli, which are neuropils comprising the dendrites and axons of PNs, LNs, and ORNs. However, the model does not consider the dynamics of the glomeruli, which may cause a gap in peak frequencies. In addition, the proportion and detailed connections of the four subclasses of LNs are not known; therefore, they were not incorporated into the model and may have affected the peak frequency. A more detailed model awaits to be built to clarify the mechanism and function of oscillations in the antennal lobe. To examine the oscillations ( Figure 6 ), we only applied 3-octanol to the network. This is because the magnitude of PN's oscillations greatly depends on the identity of odors both in vivo (Tanaka et al., 2009 ) and in silico ( Supplementary Figure S8 ). Since our intention was to measure the effect of inactivation of the LN subclass on PN's oscillations, we used only one type of odor. In the future, we will comprehensively examine the relationship between oscillations and odors and clarify why the magnitude of the oscillations differs between odors. In honeybees, oscillations in the antennal lobe are necessary for distinguishing between similar odors (Stopfer et al., 1997 ). In locusts, oscillations appear not only in the antennal lobe but also in KCs, and they are believed to contribute to the sparse representation of odors in the KC population (Perez-Orive et al., 2002 ). Although the role of oscillations in Drosophila remains unclear, oscillations likely contribute to the processing of odor information given the similarity of olfactory network structure between different insects. One possible candidate is the generation of the sparse representation of odors in the antennal lobe. In this study, the PQN model employs function m ( I ), which was not incorporated into the original PQN model (Nanami and Kohno, 2023 ). This function performs a nonlinear transformation of the stimulus input so that the membrane potential behaves as expected in response to a wide range of stimulus inputs. However, this function does slightly complicate the model and has no biological counterpart. By changing the parameters and adjusting the dynamics, we expect to be able to remove this function in future works. As shown in Figure 7 , after olfactory associative learning, MBON-α1 fires for approximately 250 ms, immediately after the arrival of the odor signal, and subsequently enters a resting period, successfully reproducing the temporal firing observed in Hige et al. ( 2015 ) in MBON-γ1pedc. To our knowledge, there has been no report on the mechanism underlying this firing dynamics characteristic for the post-learning response. The result of our simulation suggests that the delayed activation of APL contributes to shaping this activity pattern. Thus, our modeling not only reproduces observed physiological data but also provides mechanistic insight by proposing an experimentally testable hypothesis. The SiNN implemented in this study operates at the same speed as the olfactory nervous system with a 100 MHz clock signal. However, if we use a higher clock, the model can provide accelerated simulations, albeit with increased power consumption. For example, we confirmed that the model can simulate four times faster than real-time using a 400 MHz clock with a Xilinx Virtex UltraScale+ xcvu37p-fsvh2892-3-e FPGA. In this implementation, the estimated power consumption was about 4W. The power efficiency and simulation speed can be further improved by using Application Specific Integrated Circuits (ASICs). As shown in Figure 9B , most of the power is consumed by the MMCM, BRAMs, and the steady-state leakage (Static). Except for a few BRAMs that are used to store the neuronal state variables, these resources are not directly used to compute the neuronal dynamics. Ignoring the reproducibility of the spiking properties and using I&F-based models instead of PQN might reduce the power of clocks, signals, logic, and DSPs. However, these resources consume only 18.5% of the total power and their impact on the overall system is expected to be small. Ionic-conductance-based models can reproduce the dynamics of the spiking process as accurately as or better than the PQN model. However, they have many exponential terms that consume a large number of DSPs in FPGA implementations (Akbarzadeh-Sherbaf et al., 2018 ; Khoyratee et al., 2019 ). Even in the most well-optimized implementation (Khoyratee et al., 2019 ), it requires more than 20,000 LUT units and more than 100 DSPs to build a network of 2,000 neurons, which would lead to significantly higher power consumption. Our modeling approach is applicable to not only FPGAs but also ASICs. Conversion from FPGA to ASIC improves power efficiency by a factor of 14 to 20 (Amara et al., 2006 ; Kuon and Rose, 2007 ). The network reproduced in this study accounts for approximately 2% of the entire brain. Thus, our approach enables the construction of an ASIC chip that simulates the entire Drosophila brain while consuming approximately 1 watt. Such chips have considerable potential in the engineering and scientific fields. Because of its low power consumption, the chip can be mounted on small insect-like robots. The resulting system is expected to move around autonomously, solve unknown tasks, and adapt to changes in the environment, similar to insects. In addition, owing to its intrinsic power efficiency, the chip can serve as a sufficiently fast simulator of the whole brain within the constraints of the power supply typically available in laboratories. It can facilitate long-term measurement of neuronal activities and is expected to contribute to the analysis of phenomena with long timescales, such as continuous learning and forgetting. To evaluate the robustness of our approach, we measured how the success rate of the olfactory associative learning varied while changing one of the empirically determined parameters ( Figure 10A ). We varied p PN_KC which scales the strength of synaptic connections from PN to KC. Increasing or decreasing from the original value ( p PN_KC = 1.03125) decreased the success rate. This is because at the lower value, the inputs from PNs to KCs are weakened, and KCs rarely fire ( Figure 10B ). As a result, KC>MBON synaptic depression, which is the basis of learning, does not occur sufficiently. When p PN_KC is large, too many KCs fire, preventing the sparse representation of odors in KCs and reducing the success rate. At present, these parameters have to be carefully tuned manually, which hinders the easy application of this approach to other nervous systems. In future research, we plan to develop a method to automatically determine these parameters to achieve the functionality of the network. Metaheuristics will be applied, just as we determined the parameters of neurons by the differential evolution algorithm. Figure 10 (A) Success rate of olfactory associative learning while changing p PN_KC . To calculate the success rate, 50 trials were performed for each odor. (B) Average number of firing KCs per trial."
} | 6,079 |
20157346 | null | s2 | 3,583 | {
"abstract": "There are hundreds of Biological Resource Centers (BRCs) around the world, holding many little-studied microorganism. The proportion of bacterial strains that is well represented in the sequence and literature databases may be as low as 1%. This body of unexplored diversity represents an untapped source of useful strains and derived products. However, a modicum of phenotypic data is available for almost all the bacterial strains held by BRCs around the world. It is at the phenotypic level that our knowledge of the well-studied strains of bacteria and the many yet-to-be studied strains intersects. This suggests we might leverage the phenotypic data from the data-poor bacteria with the omics data from the data-rich bacteria, using our knowledge of their evolutionary relationships, to map the metabolic networks of the little-known bacteria. This systems biology-based approach is a new way to explore the diversity harbored in BRCs."
} | 235 |
31641460 | PMC6802068 | pmc | 3,584 | {
"abstract": "Abstract The speed at which species adapt depends partly on the rates of beneficial adaptation generation and how quickly they spread within and among populations. Natural rates of adaptation of corals may not be able to keep pace with climate warming. Several interventions have been proposed to fast‐track thermal adaptation, including the intentional translocation of warm‐adapted adults or their offspring (assisted gene flow, AGF) and the ex situ crossing of warm‐adapted corals with conspecifics from cooler reefs (hybridization or selective breeding) and field deployment of those offspring. The introgression of temperature tolerance loci into the genomic background of cooler‐environment corals aims to facilitate adaptation to warming while maintaining fitness under local conditions. Here we use research on selective sweeps and connectivity to understand the spread of adaptive variants as it applies to AGF on the Great Barrier Reef (GBR), focusing on the genus Acropora . Using larval biophysical dispersal modeling, we estimate levels of natural connectivity in warm‐adapted northern corals. We then model the spread of adaptive variants from single and multiple reefs and assess if the natural and assisted spread of adaptive variants will occur fast enough to prepare receiving central and southern populations given current rates of warming. We also estimate fixation rates and spatial extent of fixation under multiple release scenarios to inform intervention design. Our results suggest that thermal tolerance is unlikely to spread beyond northern reefs to the central and southern GBR without intervention, and if it does, 30+ generations are needed for adaptive gene variants to reach fixation even under multiple release scenarios. We argue that if translocation, breeding, and reseeding risks are managed, AGF using multiple release reefs can be beneficial for the restoration of coral populations. These interventions should be considered in addition to conventional management and accompanied by strong mitigation of CO 2 emissions.",
"conclusion": "6 CONCLUSION This preliminary biophysical modeling suggests that without active intervention, heat tolerance is unlikely to spread beyond the very far northern reefs of the GBR before summer heat waves become annual events. Our biophysical dispersal predictions generally support findings from previous models exploring southern reefs repopulating northern reefs against prevailing currents; however, our model uniquely explores the possibility for potentially bleaching resistant corals to re‐seed central and southern reefs under scenarios of southwards gene flow. Encouragingly, limited connectivity to the central and southern sectors of the GBR is confirmed by our southward biophysical models and by the strong genetic evidence presented in previous studies (Ayre & Hughes, 2004 ; Lukoschek et al., 2016 ; van Oppen, Peplow, Peplow, Kininmonth, & Berkelmans, 2011 ). Our results show that it may take 30+ generations for thermal tolerance PALs to extend beyond single reefs in the far north through natural dispersal, which is too slow to reach central reefs of the GBR in time given IPCC warming estimates. If variants are rare throughout coral populations and only abundant in the far north, 30+ generations to leave one reef may be beyond the scope of time available given that would extend beyond the IPCC 2050 scenario. Modeling at release sites in the central and southern reefs demonstrates that single reef release sites will result in fixation rates that are also too slow. Although rates of adaptive variant spread presented here are based on estimates of selection coefficients and migration alone, our results provide a starting point for further modeling that may include a wider range of parameters, including multilocus, multiallele systems, pleiotropic, and epistatic effects. Finally, it is currently unclear how variable the genetic architecture of thermal tolerance in corals is. Thus, this single‐locus model provides a reasonable start at modeling changes in allele frequency and fixation of “adaptive” alleles in corals, but additional work is needed to more include multiallele interactions. AGF and selective breeding also involve risk, but they can be managed through the considered selection of source individuals and populations. For example, PAL spread could be limited if a single release site is used or if the site is suboptimal in terms of its physical characteristics that determine larval dispersal away from the reef. Given that the translocation of tens of thousands of individuals may be needed to reach fixation across hundreds to thousands of reefs within ~30 to over 1,000 generations, the translocation of adult colonies without further propagation is not feasible. Instead, ex situ selective breeding of corals for larval or juvenile deployment onto receiving reefs is more practical and achievable approaches to prepare central and southern reefs on the GBR for continued warming.",
"introduction": "1 INTRODUCTION The increasing pace and severity of environmental change has degraded ecosystems around the world and heightened the need for management interventions to support adaptation or restoration of species (Lindenmayer, Piggott, & Wintle, 2013 ; van Oppen, Oliver, Putnam, & Gates, 2015 ; Rinkevich, 2014 ). Coral reefs are sensitive to the effects of climate change and have declined globally, with even well‐managed reefs in the northern and central Great Barrier Reef (GBR) suffering unprecedented mortality due to heat stress in 2016 and 2017 (Hughes et al., 2018 , 2017 ; Stuart‐Smith, Brown, Ceccarelli, & Edgar, 2018 ). On coral reefs, restoration efforts have so far primarily aimed to increase local abundance of individual species with stock from wild populations or nurseries but active interventions focussed around adaptation are increasingly being proposed (Anthony et al., 2017 ). These interventions have inherent risks and should therefore only be considered after due processes that incorporate a design tree framework for planning for success and failures (IUCN, 2013 ). The genetics of source material is considered mostly in relation to the maintenance of genetic diversity (Baums, 2008 ) and has thus far rarely been explicitly addressed in the management of adaptive processes (Edwards, 2010 ). Genetic diversity underpins the potential scope for adaptation of natural populations, and its maintenance is a key objective for biodiversity management (Frankham et al., 2017 ). Genetic diversity describes the level of variation in genetic regions (polymorphic loci) and among individuals within populations. Beneficial adaptations are generated through mutation and recombination, while changes in their frequencies occur through natural selection and drift. Adaptive diversity is a subset of total genetic diversity and describes variants of genes that are directly associated with tolerance and survival (adaptive loci; Ahrens et al., 2018 ). Standing genetic variation is typically high in corals and has been demonstrated in resilient populations from highly variable environments or in survivors of bleaching (Palumbi, Barshis, Traylor‐Knowles, & Bay, 2014 ). While there are a number of candidate adaptive loci (Kenkel et al., 2014 ), and particularly those loci correlated with thermal tolerance (Dixon et al., 2015 ; Dziedzic, Elder, Tavalire, & Meyer, 2019 ; Jin et al., 2016 ), more work is needed to ground truth these putatively adaptive loci (PALs). Ecological genomic and quantitative genetic theory has been proposed as a way to develop restoration methods that go beyond maintaining genetic diversity and aims to increase resilience to future environmental pressures (Baums, 2008 ; van Oppen et al., 2015 ). Such methods include selective breeding, conditioning (e.g., stress hardening via exposure to sublethal elevated temperatures), and microbial manipulations, and are collectively termed assisted evolution (van Oppen et al., 2015 ). Variation in thermal tolerance of scleractinian corals is underpinned by both acclimatory (i.e., phenotypic plasticity) and genetic mechanisms. Partial acclimation to heat stress has been shown to be possible within the life span of a coral colony and over relatively short periods of time (Bay & Palumbi, 2015 ), mediated by rapid changes in gene expression (Barshis et al., 2013 ; Kenkel & Matz, 2016 ) and microRNAs (Gajigan & Conaco, 2017 ). While this provides promise for conditioning as an intervention tool, the extent to which acclimation is heritable from one generation to the next remains to be shown. The potential for and rate at which corals can respond to ocean warming depend on the genetic architecture of thermal tolerance—a characteristic that appears to be variable among coral species. Genetic thermal tolerance can manifest as a heritable, polygenic trait involving over 100 loci of small effect sizes each (Bay & Palumbi, 2014 ; Dixon et al., 2015 ) or a smaller number of loci with larger effect sizes (Jin et al., 2016 ). Although the genetic architecture may be complex, efforts targeted at influencing coral host genetics provide avenues for assisted evolution (van Oppen et al., 2015 ). Translocation describes the human‐mediated movement of individuals within established ranges or extant habitat, and in the more extreme case, movement beyond current ranges (assisted colonization/migration; Aitken & Whitlock, 2013 ; Kelly & Phillips, 2016 ; Seddon, 2010 ; Weeks et al., 2011 ). Assisted gene flow involves the intentional translocation of individuals with PALs into populations with an absence or low prevalence of these genetic variants to facilitate adaptation to new or anticipated local conditions (Aitken & Whitlock, 2013 ). As a variation on AGF sensu stricto, warm‐adapted corals harboring PALs may be selectively bred or crossed with cooler‐adapted conspecifics ex situ, then deployed at the cooler but warming locations. This intervention strategy may result in the introgression of heat tolerance alleles into the genomic background of the receiving population, thereby preparing those populations for further climate warming. Given that central corals transplanted southwards may bleach and potentially die when exposed to cooler winter temperatures in the south (Howells, Berkelmans, van Oppen, Willis, & Bay, 2013 ; van Oppen, Puill‐Stephan, Lundgren, De'ath, & Bay, 2014 ), selective breeding between northern corals with central and southern corals may alleviate this trade‐off. Phenotypic variation in heat tolerance exists within and among populations and this is likely underpinned by genotypic variation, leading some individuals to survive a heat wave while others do not. Despite the comparatively high thermal tolerance of northern GBR corals (Dixon et al., 2015 ), the extreme heat waves in 2016 and 2017 caused extensive coral mortality (Hughes et al., 2018 , 2017 ; Stuart‐Smith et al., 2018 ). However, surviving corals in northern populations could serve as a source of adaptive genetic variation for thermal tolerance and used in active interventions (Pardo‐Diaz et al., 2012 ; Tigano & Friesen, 2016 ). Indeed, AGF and the selection of thermally tolerant genetic stock were proposed for the management of corals reefs over ten years ago (Hoegh‐Guldberg et al., 2008 ). While it is likely that wild populations are already responding to the strong natural selection of bleaching events, AGF has the potential to accelerate the rate at which this occurs and therefore reduce the time needed for adaptive change. Given the rapid rate of coral loss on the Great Barrier Reef, feasibility assessments of proposed interventions are needed. The models presented here provide preliminary quantitative estimates as to the feasibility (scale and time frame) of applying AGF while also outlining model limitations, and thereby provide a clear path and scope for future research. The effect of AGF and selective breeding may be persistent in longer living and annually reproducing species long after the initial deployment of stock. In this scenario, translocated individuals may continue to add PALs into receiving populations throughout their reproductive life to support natural reproductive and recovery processes even after initial deployment of coral stock. By producing coral larvae or juveniles with PALs that eventually reach a size/age of reproduction, these corals become part of the reef community and therefore the natural recovery process through continual supplementation of PALs to local populations (Vallee, 2004 ). Spatial scales for many existing restoration efforts are generally square kilometers, however, restoration of many species is needed on a much larger scale given the large extent of coral bleaching and mortality (Hughes et al., 2017 ). AGF differs from natural gene flow in that it is a directed pulse of individuals with PALs (single or multiple) and can thus encompass much larger distribution areas compared to natural gene flow (Aitken & Whitlock, 2013 ; Kelly & Phillips, 2016 ). The selective breeding of individuals affords the opportunity to increase the prevalence of PALs as well as increase the production scale of individuals with those adaptive variants. Hence, such restoration and adaptation efforts may scale‐up to spatial scales better able to influence regional reef scale degradation, although this may require several generations to achieve.",
"discussion": "5 DISCUSSION 5.1 Is AGF feasible given natural recruitment rates? The addition of hundreds of thousands to millions of corals to the reef is feasible with existing breeding capacity, but high early life mortality rates need to be factored into deployment calculations. Based on previous GBR wide recruitment estimates derived from Hughes et al. ( 2000 ), fixation within ~1–2.4 generations may be possible with the previously proposed 1,000 migrants given average estimated recruitment per reef in the north (2,510 ± 774) and central (3,400 ± 1,000) GBR, but may take slightly longer in the south where recruitment appears to be less per reef (774 ± 547; Figure S2 a). Given these per reef values, the estimated 10,000 migrants needed to induce fixation through reef restoration and adaptation interventions should be feasible across at least 100 reefs. This number of migrants is also well within the limits of regional recruitment averages, suggesting ample potential for natural recovery given the estimated spread rates only if the number of migrants is high (Figure S2 b). However, the maintenance of this natural recovery potential through changes in allele frequencies may only be possible if adult populations are kept at levels similar to when these censuses were taken (1995–1997; Hughes et al., 2000 ). Given that mass scale mortality has severely decreased the number of adults in these regions (Hughes et al., 2017 ), it is likely that natural recovery potential has been severely compromised. Regardless of the life history stage used, the relatively slow fixation rates, which more generally occur over longer evolutionary timescales, strongly suggest that interventions aimed at increasing the spread of thermal tolerance PALs can have a positive effect on fixation rates for corals of the GBR. Moreover, the biophysical models of larval dispersal using multiple release sites show that some larvae have some capacity to reach further south. If given enough time (hundreds of generations), our stepping‐stone wave models suggest that only four migrants would be needed for fixation to occur if selection is strong, suggesting that even limited larval dispersal may be enough to elicit shifts in allele frequencies, consistent with genetic theory and current modeling (Matz et al., 2018 ). Furthermore, in order to potentially meet conditions for Hardy–Weinberg equilibrium, stepping‐stone wave models assume non‐overlapping generation times in which each individual is assumed to have only a single generation time. Given that many corals likely have overlapping generations, our values may underestimate fixation rates given non‐overlapping generations may lead to slower fixation as beneficial genes in the incoming population would more quickly be diluted under this assumption. Alternatively, “priority effects” that manifest in species with overlapping generations may slow fixation (Atkins & Travis, 2010 ; Gilbert et al., 2017 ). Additional modeling and empirical measurements are required to assess the impact of this assumption as well as determine what allele frequencies (i.e., soft or hard sweeps) are required to illicit increased population survival under thermal stress. 5.2 Risks and trade‐offs associated with assisted gene flow Relatively little is known about the ecological risks associated with AGF and release of ex situ selective bred corals. On a spectrum of translocation actions, AGF within species ranges has been classified as a medium level of intervention action based on the species historical distributions (Seddon, 2010 ). The formation of outbreeding depression, the breakdown of local adaptation of source populations, and lineage swamping (outcompeting/loss of local sink populations) are the three main risks associated with AGF (reviewed in Aitken & Whitlock, 2013 ). Other risks may include the depletion of local populations on source reefs or unintentional transfer of pathogens, parasites or members of the microbiome to central or southern reefs. The use of ex situ bred offspring may help partially ameliorate these risks although this approach may introduce other risks associated with the captive propagation approach. Risks associated with AGF have been quantified as “proportional to the fraction of the population replaced (5%–20%)” and well as the provenance of the coral stock used (Aitken & Whitlock, 2013 ). Given that little is known about the magnitude of effects in corals, knowledge from other species could be used to predict risk (Kelly & Phillips, 2016 ). Risk can also be mitigated by limiting the initial numbers of translocated individuals (20% of gene flow in the first pulse, and 2%–4% thereafter; Aitken & Whitlock, 2013 ). Although estimates of gene flow in corals are limited, the number of migrants per generation may range between 4 and 100 (see above and Matz et al., 2018 ). Using these estimates, the initial number of translocated individuals needed to mitigate risk associated with assisted PAL spread may equate to roughly 1–20 individuals per reef (20% of 4–100 individuals), which is well below the number needed to achieve fixation within a short time window. Other measures may therefore be needed to help mitigate risk besides limiting the number of individuals used in AGF and deployment of selectively bred corals. No organism can be perfectly adapted to every environmental pressure, resulting in ecological trade‐offs (Ferenci, 2008 ; MacArthur & Wilson, 1967 ). AGF and selective breeding may introduce genotypes that are not adapted to local environmental conditions other than temperature at the receiving location (“local is best”, Rehfeldt et al., 2017 ) and introduce trade‐offs between different traits. One concern is that the single or repeated pulse of PALs will erode or alter local adaptation among coral populations, as other loci adapted to the source reef conditions will also enter the receiving population (Drury, Manzello, et al., 2017 ; Drury, Schopmeyer, et al., 2017 ; Kenkel et al., 2013 ; Pavlova et al., 2017 ; Polato et al., 2010 ; Webber & Scott, 2012 ). Meta‐analyses across highly divergent taxa including plants, animals, fungi, and protists revealed that local populations gain ~50% fitness advantage in their native environment compared to migrants (Hereford, 2009 ). Balancing selection may function to decrease this conflict and could be implemented during these intervention practices by selecting populations that are intermediate in their phenotypic ranges (i.e., within ± 1 standard deviation of the phenotypic mean) of those traits of interest (Rehfeldt et al., 2017 ). The effectiveness of this technique will require the quantification of reaction norms, especially in southern reefs that may be particularly cold‐adapted, likely making the pairing of reefs within ± 1 standard deviation challenging between the most extreme range edges (i.e., northern and southern). Further considerations involve the ecological outcompeting by source population individuals (genetic swapping/foreign advantage), breakdown of local adaptation, outbreeding depression, and expansion load. When foreigners were placed in different environments compared to their native environment, the presence of foreign advantage was rare among taxa, especially in the circumstances where selection was strong, and where genetic variation was high (Hereford, 2009 ; Schluter, 2000 ). Selectively bred organisms (e.g., the offspring of foreign and native individuals) do not always have a competitive advantage over natives (i.e., locals can prevent the establishment of foreigners with higher thermal tolerance; van Oppen et al., 2014 ; Quigley, Willis, & Bay, 2016 ); and if present, modeling suggests that this advantage may only be transitory (<2 generations; Aitken & Whitlock, 2013 ). It has also recently been shown that the addition of foreigners does not lead to as high rates of erosion of local adaptation as originally thought (i.e., decreased rate of sweeps; Tigano & Friesen, 2016 ). A survey of translocation studies of plants also demonstrate that occurrences of maladaptation caused by outbreeding depression was likely to be low (<3.3%; Leimu & Fischer, 2008 ). Hence, the risks associated with outbreeding depression and foreign advantage may dissipate rapidly (Aitken & Whitlock, 2013 ; Harrisson et al., 2016 ; Ralls et al., 2017 ; Roitman et al., 2017 ; Tigano & Friesen, 2016 ). Finally, expansion load (i.e., the accumulation of deleterious mutations) may slow the spread of genetic variants out of northern populations into the central and southern GBR at the expansion front by removing maladapted populations (Gilbert et al., 2017 ). 5.3 Factors that may increase the efficacy of assisted PAL spread in corals A number of factors suggest the likelihood of achieving positive fitness benefits for corals through the application of AGF and selective breeding methods is considerable.\n \n Soft sweeps may be sufficient to illicit successful putative adaptive loci penetrance and therefore fitness benefits in corals . Previous studies have shown that complete fixation may not be necessary to elicit beneficial phenotypic shifts in populations, and allele frequencies as low as 0.2 may be sufficient and a more achievable outcome (Creech et al., 2017 ; Ferenci, 2008 ). Therefore, although it is unclear what exact allele frequencies below complete fixation would be sufficient to elicit these shifts, modeling complete fixation provides conservative preliminary estimates. Soft sweeps would also have the secondary benefits of safeguarding against the erosion of genetic diversity and complete elimination of native genetic variants. \n Adaptive loci are confirmed in some reefs in the northern Great Barrier Reef and may be present further south . For AGF or selective breeding to be feasible within the short term, the presence of key standing genetic variation within coral metapopulations is mandatory (pre‐existing PALs). Some adaptive variants involved with thermal tolerance in corals have already been identified in some populations of some coral species and may exist at low frequencies at cooler but warming reefs (Bay & Palumbi, 2014 ; Dixon et al., 2015 ; Jin et al., 2016 ; Louis, Bhagooli, Kenkel, Baker, & Dyall, 2017 ). Pre‐existing PALs will therefore fast track the spread of adaptive variants throughout populations. Moreover, if PALs already exist, albeit at lower frequencies, in more southern reefs, the combined effects of natural dispersal and with AGF PALs will increase the rate of spread (Hermisson & Pennings, 2017 ) and potentially decrease the risks of the introduction of maladapted variants. To estimate coefficients of selection, the difficult task of defining “beneficial loci” may be needed (Ferenci, 2008 ), although it may be initially feasible to use phenotypes to implement AGF and selective breeding. Although substantial progress has been made to characterize the genetic architecture of heat tolerance in corals, it is unclear how many loci are involved, their relative effect sizes, and level of interaction. Furthermore, it is likely that PALs will vary across many reefs in the north which may compete to establish themselves across central and southern reefs and hence slow AGF success. \n Naturally high genetic diversity (Devlin‐Durante & Baums, 2017 ). High diversity reduces the risk of outbreeding depression, one of the main risks associated with AGF (Aitken & Whitlock, 2013 ) and leads to faster rates of adaptation (Whiteley, Fitzpatrick, Funk, & Tallmon, 2015 ). The naturally high genetic diversity in many coral species may thus help to accelerate rates of beneficial fixation. Greater genome diversity (heterozygosity) has also been linked to fitter individuals that exhibit comparatively high resilience (Bay & Palumbi, 2014 ; Drury, Manzello, et al., 2017 ; Drury, Schopmeyer, et al., 2017 ; Ellegren & Galtier, 2016 ), thereby providing better source material targets for AGF. \n Large donor population sizes . Species with large donor population sizes and wide distributions make good candidates for interventions aimed at increasing PAL spread (Aitken & Whitlock, 2013 ). Large donor sizes decrease the risk of genetic drift and outbreeding, which can all impede the probability of PALs from reaching high frequencies and can erode adaptive potential (Aitken & Whitlock, 2013 ; Hereford, 2009 ). Large donor population sizes also tend to have higher rates of local adaptation (Aitken & Whitlock, 2013 ; Creech et al., 2017 ), leading to more efficient selection due to increased genetic diversity (Ellegren & Galtier, 2016 ; Savolainen, Lascoux, & Merilä, 2013 ). The potential harnessing of spawning slicks to take advantage of these large population donor sizes may also be feasible (Heyward & Negri, 1999 )."
} | 6,582 |
40169576 | PMC11961630 | pmc | 3,585 | {
"abstract": "The alpine grasslands of the Qinghai-Tibetan Plateau (QTP), the world’s highest plateau, have been severely degraded. To address this degradation, human-involved restoration efforts, including grassland cultivation, have been implemented. However, the impact of these practices on soil microbial community stability and its relationship with plant-soil system resilience has not been explored. In this study, we evaluate the effects of grassland restoration on microbial communities. We show that bacteria demonstrate higher composition resistance and resilience during the restoration process, when compared to fungi. The changes we observe in microbial community interactions support the stress gradient hypothesis. Our results emphasize the synergistic role of network resilience and the restoration of the plant-soil system. Importantly, we find that core microbial species significantly influence the resilience of the plant-soil system by sustaining the co-occurrence networks. These insights underscore the critical roles of microbial communities in grassland restoration and suggest new strategies for boosting grassland resilience by safeguarding core microbes.",
"introduction": "Introduction The grasslands of the Qinghai-Tibetan Plateau (QTP), the highest plateau in the world, are crucial for food supply, environmental conservation, and social development 1 . However, these alpine grassland ecosystems have undergone significant degradation in recent decades due to various drivers, including climate change, overgrazing, and anthropogenic interventions 2 . Soil microorganisms, which encompass a diverse and intricate array of biological communities that play a crucial role in organic matter decomposition, nutrient cycling and maintaining soil functionality in grassland ecosystems 3 . Notably, soil bacteria and fungi respond differently to environmental filtering, thereby influencing the distribution and diversity of the soil microbial community 4 , 5 . Mean annual temperature and aboveground net primary productivity are determining factors for changes in fungal diversity, soil pH, and N:P ratio are determining factors for changes in bacterial diversity 5 . While the significance of microorganisms in restoring degraded grasslands is well-recognized, there remains a notable gap in research regarding the disparities in soil microbial diversity, the relationship between bacteria and fungi, as well as the complexity and stability of the microbiome. Addressing these knowledge gaps is crucial for developing effective strategies to restore and preserve the fragile alpine grassland ecosystems of the QTP. The stability of an ecosystem in the face of disruption hinges on its resistance and resilience 6 . Referring to the concept of Griffiths 7 and Gao 8 et al., stability of microbial community can be defined as the ability of a microbial community’s composition and network to resist environmental disturbances (resistance) and to recover to its original state after the disturbance is eliminated (resilience). Such stability is influenced not only by the composition of its community, but also by the myriad interactions among the co-existing microbial communities. These interactions span a spectrum from negative to positive and weak to strong, encompassing both significant to non-significant relationships among bacteria, fungi, or both 9 . Investigating these correlations sheds light on the dynamics of resistance and resilience within ecosystems. The stress gradient hypothesis (SGH) suggests that the occurrence of cooperation and competitive interactions will vary inversely across abiotic stress gradients. Cooperation is more common in high-stress conditions and competition is more favorable in low-stress environments 10 – 12 . With ecosystems facing increasing anthropogenic disturbances, there is a potential for microbial communities to become destabilized. Gao et al. 9 bolstered the stress gradient hypothesis by demonstrating an increase in positive associations among microbial communities in response to drought conditions. This underscores the significance of investigating how microbial communities adapt to stressful environments. Despite these insights, there remains a notable gap in research specifically addressing how the stress gradient hypothesis applies to microbial communities during the process of grassland restoration. This area of study presents a valuable opportunity for further exploration to understand the resilience and adaptation mechanisms of microbial ecosystems under restoration actions. In a recent study, Qiao et al. utilized core amplicon sequence variants (ASVs), pinpoint key microbial species, distinguishing core species on the basis of shared ASVs, specialist ASVs, and keystones 13 . A notable outcome of their study was the discovery that the relative abundance of each core species exhibited a positive correlation with network properties (number of links, relative modularity and positive cohesion). These pivotal findings underscore the essential influence of these species in augmenting network complexity and bolstering plant health. Despite the progress made, researchers are still struggling to fully understand the specific functions of microbial core taxa in promoting microbial stability. This area remains ripe for further exploration to fully unravel the mechanisms through which core microbial species contribute to the resilience and functionality of their ecosystems. This study was carried out in the alpine grasslands of Maqin county, located in Qinghai Province, China. We utilized a space-for-time substitution approach to delve into the impacts of grassland restoration efforts. Our restoration efforts for degraded grasslands involved reseeding natural grasslands and establishing cultivated grasslands to boost grassland productivity. The study objectives are: (1) to unravel the pattern of soil microbial diversity and the intricacies of network complexity throughout the restoration of the alpine grassland, (2) to assess the resistance and resilience of fungal and bacterial communities during the period of ecological recovery, (3) to elucidate the pivotal role played by soil microbes in mediating the interactions within the plant-soil system over the course of the long-term restoration process. This comprehensive investigation was designed to shed light on the critical contributions of microbial communities to the ecological restoration and sustainability of alpine grasslands. Here, we show that core microbial species significantly contribute to the resilience of the plant-soil system by supporting co-occurrence networks.",
"discussion": "Discussion The pivotal role of environmental factors in shaping microbial communities and their subsequent consequent impact on ecosystem functions is well-recognized 14 . The extensive positive correlations between soil organisms underscore their vital contribution to maintaining global soil biodiversity and the integrity of ecological network 15 . Previous study has established a strong connection between the stability of plant and microbial community in degraded grasslands, highlighting the mutual dependence essential for community stability 16 . The space-for-time substitution method has been widely used in ecological studies to generate a chronosequence 17 . The space-for-time substitution method was used to study the steady-state transition of ecosystems in the temperate steppe of Inner Mongolia, China 18 . A previous study in North America also supported the use of the space-for-time substitution method in predicting community responses to climate change 19 . Similarly, we used the space-for-time substitution for examining the complexity and stability of microbial community during long-term restoration of alpine grasslands. We recognize the limitations associated with using space-for-time substitution methods 18 , 20 , as they involve assuming that plant, soil, and microbial processes will not be affected by a series of environmental covariates (e.g., precipitation, vegetation type). The approach was also compromised when the effects of variables such as soil type and landform were not minimized 21 . Our results supported that space-for-time substitution in this study was accountable. Previous studies have demonstrated a trade-off between resistance and resilience 22 , 23 . However, our findings suggested that bacteria exhibited greater composition resistance and resilience during degradation compared to fungi, without displaying a trade-off relationship between the two. Likewise, two studies conducted in Europe observed no trade-off between microbial resistance and resilience 24 , 25 . Due to human-involved grassland restoration efforts, the complexity of microbial networks had increased, and the interactions between microorganisms had become more intricate and complex, underscoring the importance of core species in connecting network restoration with the plant-soil system. The characteristics of bacterial-fungal networks during the restoration process showed improvement, reflecting a non-linear trend in the restoration capacity of grassland vegetation, soil, and vegetation-soil system. During the restoration of grasslands, the complexity of microbial network decreased after 4 years of restoration, possibly due to competition for nutrients among early plant growth. After 6 years of restoration, complexity had increased as more stable plants grew. However, after 8–9 years, there was a “secondary degradation” and a drop in complexity observed. Interestingly, after 12–16 years, began to rise again, indicating a trend towards recovery. Previous studies have demonstrated the intertwined nature of plant communities’ diversity and soil microorganisms, both of which are crucial for providing multiple ecosystem functions 6 , 26 – 28 . Microbes mainly enhance plant health through improved nutrient acquisition and disease resistance 29 . The intricate web of positive and negative correlations among microbial species plays a significant role in this dynamic, as demonstrated by Gao et al. 9 , highlighting the complexities of microbial co-occurrence networks. The network complexity and the negative connection interactions between bacteria and fungi both played a role in regulating soil multifunctionality, with the former promoting and the latter potentially hindering this multifunctionality 30 . Although our results clearly link microbial diversity and composition resilience, the connection to the recovery of plant soil system is less pronounced. Nevertheless, we observed a significant correlation between the degree of network recovery and plant-soil system restoration, suggesting that microbial interactions may play a pivotal role in the functioning of microorganisms. These interactions facilitate soil colonization by microorganisms 31 , indicating that the trend in microbial network resilience aligns with previous findings on the resilience of plant-soil systems, which peaked at the 16th year of recovery 32 . The growth of microbial communities is influenced by the soil environment and plants. We have directly evaluated soil nutrient factors related to abiotic stress (Supplementary Table 2 ). The average values of these factors provide a clear indication of nutrient conditions and stress levels to categorize severely degraded samples as high stress samples, non-degraded samples as low stress samples, and the after restoration (including early stage and late stage) as a stress situation falling between high and low stress. The abundance of N-NH 4 , TP, and TS increases with the increase of stress gradient, while the other nutrients show no consistent pattern. Hernandez et al. found that total cohesion decreases with increasing stress 10 , attributed to a reduction in both positive and negative cohesion along the pressure gradient. Our results, where the overall cohesion of severely degraded samples showed no significant difference compared to non-degraded samples. The relationship between network complexity and stability is complex 33 . This suggests that increased network recovery may not always correspond to microbial community stability. Hernandez et al. also found that the ratio of negative to positive cohesion decreases with increasing stress, indicating a dominance of cooperative behaviors in high stress environments 10 . Our findings, supporting the stress gradient hypothesis, suggest that microbial species exhibit more cooperative rather than competitive behavior during recovery stages and non-degraded environments. The longer the restoration period of the interconnection between fungi and bacteria in the network, the greater it grows. The resilience between fungi and bacteria best represents the network resilience. Fungi demonstrate greater resilience compared to bacteria, with a more significant increase in connections, suggesting that fungi are more volatile while bacteria maintain greater stability. Bacteria nodes demonstrated higher constancy and persistence than fungal node, indicating greater stability and longevity of bacteria within the community 34 . The higher niche width of bacteria indicates stronger environmental adaptability 35 . Fungal taxa being dominated by dispersal limitation and stochastic processes, consistent with a weaker phylogenetic signal. This indicates a more stochastic assembly process for fungi, contributing to their flexibility in metabolic functions and ability to maintain community structure under fluctuating environmental conditions 35 . In contrast, bacterial taxa are governed by heterogeneous selection, explained by a stronger phylogenetic signal, which aids their adaptation to extreme environments and favors deterministic heterogeneous selection mechanisms, leading to greater structural stability. This comprehensive analysis may explain why bacteria demonstrated higher resistance and resilience stability, and underscores the complex and distinct interactions between bacteria and fungi within soil microbial communities, highlighting the importance of considering microbial dynamics in efforts to restore degraded grasslands. The restoration of connections between fungi and bacteria closely correlates with the trend of PSQIR. The interplay between these microbial groups is crucial for the overall network recovery process. Investigation into specialized taxa have further emphasized their roles in maintaining community stability 4 , 13 , 36 . In previous research, dominant bacterial and fungal taxonomic groups exhibit strategies for drought resistanc 37 . Wang et al. have shown that elevated precipitation leads to soil C loss through interactions between microbes-plants-soil 38 . Ma et al. highlight the pivotal role of core microorganisms in enhancing network resilience during restoration and, consequently, in the restoration of plant soil health 39 . Our research builds on this by emphasizing the critical role of microorganisms in restoring plant-soil systems and exploring the impact of core species on the microorganism-plant-soil relationship. Generalist core species, which exhibit higher adaptability and play a crucial role in networks resilience, helped to rebuild the relationship between bacteria and fungi, thereby aiding in the restoration of the overall network and further promoting the restoration of the ecosystem. The significance of core species in maintaining network resilience throughout grassland restoration was a key finding of this study, revealing synergistic effect among plants, soil, and microorganisms. These findings highlight the intricate balance between specialist and generalist microorganisms in shaping the resilience and cohesion of soil microbial communities during ecological recovery processes. Understanding the roles and dynamics of these core OTUs offers valuable insights into ecosystem restoration and management strategies. Microorganisms play a crucial role in soil ecological processes, with their resistance and resilience determining the sustainability of ecosystem functionalities. Microbes can optimize nutrient cycles, improve soil structure, and increase resistance to disturbances. By introducing key microorganisms during restoration, or by managing the balance between bacteria and fungi, we can implement effective strategies for soil and plant restoration. Monitoring the health of soil microbes can serve as a valuable indicator of the progress and success of restoration efforts. These insights into the resistance and resilience of soil microbial communities are essential for assessing soil health and predicting the effects of perturbations on ecosystem functions in the face of future global change scenarios. This has profound implications in developing new strategies to enhance ecosystem resilience, particularly through the protection of core microbes in alpine grassland and other similar ecosystems globally."
} | 4,225 |
27917230 | PMC5131848 | pmc | 3,587 | {
"abstract": "The correlated motion of flocks is an instance of global order emerging\nfrom local interactions. An essential difference with analogous ferromagnetic\nsystems is that flocks are active: animals move relative to each other,\ndynamically rearranging their interaction network. The effect of this\noff-equilibrium element is well studied theoretically, but its impact on actual\nbiological groups deserves more experimental attention. Here, we introduce a\nnovel dynamical inference technique, based on the principle of maximum entropy,\nwhich accodomates network rearrangements and overcomes the problem of slow\nexperimental sampling rates. We use this method to infer the strength and range\nof alignment forces from data of starling flocks. We find that local bird\nalignment happens on a much faster timescale than neighbour rearrangement.\nAccordingly, equilibrium inference, which assumes a fixed interaction network,\ngives results consistent with dynamical inference. We conclude that bird\norientations are in a state of local quasi-equilibrium over the interaction\nlength scale, providing firm ground for the applicability of statistical physics\nin certain active systems."
} | 292 |
35922483 | null | s2 | 3,589 | {
"abstract": "Bacterial biofilms are often defined as communities of surface-attached bacteria and are typically depicted with a classic mushroom-shaped structure characteristic of Pseudomonas aeruginosa. However, it has become evident that this is not how all biofilms develop, especially in vivo, in clinical and industrial settings, and in the environment, where biofilms often are observed as non-surface-attached aggregates. In this Review, we describe the origin of the current five-step biofilm development model and why it fails to capture many aspects of bacterial biofilm physiology. We aim to present a simplistic developmental model for biofilm formation that is flexible enough to include all the diverse scenarios and microenvironments where biofilms are formed. With this new expanded, inclusive model, we hereby introduce a common platform for developing an understanding of biofilms and anti-biofilm strategies that can be tailored to the microenvironment under investigation."
} | 244 |
33268782 | PMC7710750 | pmc | 3,591 | {
"abstract": "Planktonic cultures, of a rationally designed consortium, demonstrated emergent properties that exceeded the sums of monoculture properties, including a >200% increase in cellobiose catabolism, a >100% increase in glycerol catabolism, a >800% increase in ethanol production, and a >120% increase in biomass productivity. The consortium was designed to have a primary and secondary-resource specialist that used crossfeeding with a positive feedback mechanism, division of labor, and nutrient and energy transfer via necromass catabolism. The primary resource specialist was Clostridium phytofermentans ( a.k.a. Lachnoclostridium phytofermentans ), a cellulolytic, obligate anaerobe. The secondary-resource specialist was Escherichia coli , a versatile, facultative anaerobe, which can ferment glycerol and byproducts of cellobiose catabolism. The consortium also demonstrated emergent properties of enhanced biomass accumulation when grown as biofilms, which created high cell density communities with gradients of species along the vertical axis. Consortium biofilms were robust to oxic perturbations with E. coli consuming O 2 , creating an anoxic environment for C. phytofermentans . Anoxic/oxic cycling further enhanced biomass productivity of the biofilm consortium, increasing biomass accumulation ~250% over the sum of the monoculture biofilms. Consortium emergent properties were credited to several synergistic mechanisms. E. coli consumed inhibitory byproducts from cellobiose catabolism, driving higher C. phytofermentans growth and higher cellulolytic enzyme production, which in turn provided more substrate for E. coli . E. coli necromass enhanced C. phytofermentans growth while C. phytofermentans necromass aided E. coli growth via the release of peptides and amino acids, respectively. In aggregate, temporal cycling of necromass constituents increased flux of cellulose-derived resources through the consortium. The study establishes a consortia-based, bioprocessing strategy built on naturally occurring interactions for improved conversion of cellulose-derived sugars into bioproducts.",
"introduction": "Introduction Sustainable, cost-effective production of fuels and chemicals is a major societal challenge. Lignocellulosic biomass is a promising feedstock for bioprocesses because of the large global supply, low cost, and the flexibility of the monomers to be converted into value-added products, including fuels, chemicals, and materials 1 , 2 . Consolidated, one pot, bioprocessing where lignocellulose depolymerization and product formation occur in a single vessel, is proposed to be a cost-effective strategy for producing fuels and chemicals due to process simplicity 3 , 4 . Biological routes for lignocellulose depolymerization are environmentally and economically attractive due to high substrate conversion and mild operating conditions as compared to the high energy and harsh chemical requirements of thermochemical processes 5 . Traditional bioprocessing efforts have focused primarily on using a single “superbug” to achieve all desired chemistries. However, using single organisms for consolidated bioprocess often leads to low product titers, yields, and productivities 6 – 9 . It is difficult to optimize all necessary traits simultaneously due to tradeoffs in resource allocation 10 . Resources allocated to one function are not available to optimize additional functions; this concept forms the basis of the “Darwinian Demon” ecological thought experiment 11 , 12 . Evolution and natural selection have addressed the challenge of complex, multistep processes, like lignocellulose deconstruction via consortia using division of labor 8 , 13 – 18 . Natural and assembled consortia have been used for degrading lignocellulosic substrates 8 , 19 – 24 . The assembled consortia have used combinations of fungi or fungi and bacteria. For example, Minty et al. 25 have used Escherichia coli and Trichoderma reesei to produce isobutanol from cellulose while Jin et al. 23 and Zuroff et al. 21 have assembled consortia comprised of Clostridium phytofermentans ( a.k.a. Lachnoclostridium phytofermentans ) and Saccharomyces cerevisiae to produce ethanol from cellulosic feedstocks. Biofilms are microbial aggregates encapusulated in self-produced polymers and are typically associated with an interface like a solid surface; in nature, most microorganisms reside in biofilms 26 . The biofilm phenotype is distinct from the planktonic phenotype. Rate imbalances between biotic reactions and abiotic diffusion create gradients in chemicals and metabolic activity. These gradients are largely responsible for the structure and physiology of biofilms and can be viewed as control parameters for bioprocess applications 27 – 29 . Biofilms have competitive properties for bioprocessing including high cell densities (200–300 g cell dry weight L −1 ), high volumetric productivities, reduced requirements for water, no need for energy intensive agitation, facilitated separation of biomass from supernatant, and high tolerance to stresses like pH or inhibitors 10 , 30 , 31 . There is considerable scientific interest in improving the catalytic efficiency of natural processes like nutrient cycling and applied processes like biofuel synthesis. Harnessing the emergent properties of microbial interactions has the potential to achieve this catalytic goal 9 . However, the biological compentents and interactions necessary to achieve emergent properties are not well understood. Natural systems are often extremely complex in terms of the number of species and the number of interactions, confounding the basis of emergent properties. Synthetic and artifical ecology have ability to decode the requirements of nonlinear, emergent properties 15 . In this work, an artificial consortium comprised of C. phytofermentans and E. coli was constructed. Here, the term artifical consortium is used to describe a consortium comprised of wild-type organisms that are not thought to cooccur in nature; alternatively, a synthetic consortium is defined as a consortium with at least one genetically modified population 32 . C. phytofermentans is a mesophilic, obligate anaerobe that grows on both soluble and insoluble components of lignocellulosic feedstocks 33 . C. phytofermentans is remarkable among the Clostridium genus due to its ability to catabolize a broad range of substrates. Its genome encodes over 169 carbohydrate-active enzymes, the largest number among sequenced clostridia, and its efficient ethanol production makes it a model system for cellulosic biofuel production 21 , 23 , 34 – 37 . E. coli is a well studied, facultative anaerobe capable of fermenting a broad range of substrates including glucose and glycerol which is a widely available waste product from biodiesel production 21 , 23 , 38 . E. coli is also a convenient host for metabolic engineering and can be modified to produce a wide range of biochemical products 39 , 40 . The C. phytofermentans and E. coli consortium was assembled to leverage common ecological motifs including cooccurrence of primary and secondary-resource specialists, metabolite exchange with positive feedback, and the flux of nutrients and energy between trophic levels through the catabolism of lysed biomass known as necromass 27 , 41 – 46 . Additionally, when grown as a biofilm, E. coli consumes O 2 creating an anoxic environment for C. phytofermentans . The role of each consortium member, the mechanisms of interaction, and the spatial and temporal analysis of system function were considered in this study quantifying the enhanced consortium productivity. The metrics used to quantify the emergent properties of the consortium were (1) enhanced depletion of cellulosic sugar, (2) enhanced production of ethanol as a proxy biofuel and bioproduct molecule, and (3) enhanced production of microbial biomass.",
"discussion": "Discussion An artificial consortium was assembled using principles identified in naturally occurring consortia including division of labor between primary- and secondary-resource specialists, metabolite exchange with positive feedback, and enhanced resource extraction based on necromass catabolism 27 , 41 . The cellobiose-degrading consortium comprised of C. phytofermentans , the primary resource specialist, and E. coli , the secondary-resource specialist, demonstrated the emergent properties of enhanced substrate depletion, enhanced ethanol secretion, and enhanced biomass productivity relative to the sum of monoculture properties. For example, the synergistic interactions improved planktonic and biofilm biomass productivity approximately 121% and 153%, respectively, on a mass basis (Table 1 and Fig. 2e ). A proposed model of the monoculture and consortium substrate preferences and interactions is illustrated in Fig. 9a–c . Consortial interactions also produced substantial, experimental changes in byproduct distributions after 72 h of cultivation (Table 1 and Fig. 9d–f ). The consortium used wild-type microorganisms to achieve the enhanced properties. Use of traditional metabolic engineering approaches such as deleting inefficient metabolic routes could further optimize the system as well as be used to synthesis other valuable bioproducts 39 , 40 , 61 – 63 . Fig. 9 Proposed model of monoculture and consortia interactions and experimental distribution of reduced carbon products. a \n C. phytofermentans monoculture, b \n E. coli monoculture, c \n C. phytofermentans , and E. coli binary consortia with necromass catabolism. d Experimental distribution of carbon products for C. phytofermentans monoculture after 72 h of cultivation. Areas represent percent of measured carbon moles. e Experimental distribution of carbon products for anoxic E. coli monoculture after 72 h of cultivation. Areas represent percent of measured carbon moles. d Experimental distribution of carbon products for anoxic consortium after 72 h of cultivation. Areas represent percent of measured carbon moles. Enhanced biomass productivity was proposed to be the result of a few major mechanisms. First, C. phytofermentans released cellobiase enzyme which hydrolyzed cellobiose into glucose extracellularly (Fig. 1g ). The presence of free glucose inhibited the production of additional cellulolytic enzymes (Fig. 6a, b ); when E. coli was present, it catabolized the glucose relieving inhibition of cellulolytic enzyme synthesis and created a positive feed forward loop enhancing the degradation of cellulose-derived sugar (Fig. 1g ). In the presence of O 2 , E. coli likely catabolized fermentation byproducts removing the inhibitory metabolites, creating a positive feedback loop enhancing substrate catabolism. C. phytofermentans readily formed spores, lysing the vegetative cells, and releasing necromass which was partially bioavailable for E. coli catabolism (Figs. 7 and 8b ). Additionally, the anoxic to oxic switch (AOS) cultivation would have lysed C. phytofermentans cells in the oxic zone of the biofilm, releasing necromass (Figs. 2 and 7 ). The spore-forming and O 2 -lysed cells would also release cellobiase which remained active in the presence of O 2 (Fig. 6a, b ), producing additional glucose for E. coli catabolism. Moreover, E. coli grew readily on simple substrates including free amino acids, upgrading those resources into proteins and oligomers; this upgrading combined with E. coli cell lysis would make the otherwise inaccessible resources available for the fastidious C. phytofermentans , enhancing its growth and production of cellulolytic enzymes (Fig. 8a, c ). The turnover of biomass from the primary resource population and the release of necromass is a common mechanism in natural consortia and can drive flux of material and energy between trophic levels 64 – 68 . Biomass turnover, through mechanisms like senescence, inhibitor-based cell lysis, or viral predation, can result in increased energy acquisition rates in the systems. This is a predictor of competitive consortium function based on a theory known as the “Maximum Power” principle 41 , 69 , 70 . A substantial increase in biomass productivity occurred when the consortium was transferred from anoxic to oxic conditions. One hundred and forty-seven percent more consortia mass was produced during AOS cultivation as compared to anoxic cultivation (Fig. 2e ). The increase was substantially larger (153%) than the sum of the monoculture AOS productivities, quantifying the outcome of the synergistic interactions between the two species and the oxic environment. The use of agar plates for biofilm cultivation prevented direct measurement of cellobiose utilization and ethanol production, but they are proposed to scale with biomass productivity suggesting >2-fold increase in cellobiose catabolism and ethanol production compared to monocultures. The AOS cultivation is a relatively simple strategy with a large impact and can be integrated into cultivation systems via the introduction of O 2 after the initial anoxic phase. This strategy could be applied readily to either solid phase or heterogenous (liquid + flocs) bioreactors. The timing of the anoxic to oxic transition would need to account for the system growth rates, biomass concentration, and the length scales for O 2 diffusion 71 . The simultaneous use of both anaerobic and aerobic chemistries within the biofilm provides opportunity for bioprocessing. The anoxic zone would favor the capture of sugar-derived electrons on reduced products like ethanol, while the oxic zone enables high energetic yields on byproducts like acetate and high metabolic rates which consume O 2 maintaining the anoxic zone. E. coli is a convenient biotechnological host and provides opportunities for producing a wide range of biochemicals in the anoxic, oxic, or both zones of the biofilm. Obligate aerobic or facultative E. coli strains could be cultivated in biofilms to control vertical localization, generating laminated catalytic potential 13 , 27 . This study constructed an artifical C. phytofermentans and E. coli consortium based on biomimicry of naturally occuring, microorganism interactions. The consortium demonstrated the emergent properties of enhanced substrate depletion, enhanced ethanol production, and enhanced biomass productivity. The assembled consortium had enhanced functioning during both planktonic and biofilm cultivation based on crossfeeding, positive feedback mechanisms, and the catabolism of necromass. These design features are powerful tools for improving bioprocesses and can likely be incorporated within existing bioprocesses."
} | 3,670 |
28262696 | PMC5338011 | pmc | 3,592 | {
"abstract": "Mounting evidence suggests that natural microbial communities exhibit a high level of spatial organization at the micrometric scale that facilitate ecological interactions and support biogeochemical cycles. Microbial patterns are difficult to study definitively in natural environments due to complex biodiversity, observability and variable physicochemical factors. Here, we examine how trophic dependencies give rise to self-organized spatial patterns of a well-defined bacterial consortium grown on hydrated surfaces. The model consortium consisted of two Pseudomonas putida mutant strains that can fully degrade the aromatic hydrocarbon toluene. We demonstrated that obligate cooperation in toluene degradation (cooperative mutualism) favored convergence of 1:1 partner ratio and strong intermixing at the microscale (10–100 μm). In contrast, competition for benzoate, a compound degraded independently by both strains, led to distinct segregation patterns. Emergence of a persistent spatial pattern has been predicted for surface attached microbial activity in liquid films that mediate diffusive exchanges while permitting limited cell movement (colony expansion). This study of a simple microbial consortium offers mechanistic glimpses into the rules governing the assembly and functioning of complex sessile communities, and points to general principles of spatial organization with potential applications for natural and engineered microbial systems.",
"discussion": "Discussion In this study, we examined microbial spatial self-organization as function of trophic dependency using a two-member, toluene-degrading bacterial consortium. Two interacting bacterial Pseudomonas putida strains, Pp F4 and Pp F107, grew jointly on toluene as a consortium but not as single-strain cultures, both in liquid and on solid media, thus manifesting obligate trophic mutualism ( Fig. 1 , Supplementary Fig. S1 ). The ratio of Pp F4 and Pp F107 cells, as measured by colony-forming units, converged to 1:1 in liquid cultures and on agar media independent of the initial proportions ( Fig. 4 ). Such ratio convergence is expected under conditions of obligate mutualistic growth 38 , and suggests system stability 39 . On agar surfaces, mutualistic growth with toluene led to specific expansion patterns with high degree strain intermixing, i.e., a structural organization that facilitates exchange of soluble metabolites ( Figs 2 , 3 ). It is known that Pp F107 oxidizes toluene and releases the byproduct 3-methylcatechol 32 34 , which can be used as carbon source by Pp F4. Since the oxidation of toluene into 3-methylcatechol yields no carbon, Pp F107 likely benefits from the interaction through Pp F4 byproducts (e.g., metabolic products such as acetate, which Pp F107 can use as carbon source). Growth was reduced and spatial patterns differed markedly when toluene was dissolved directly in the medium rather than provided via the gas phase ( Fig. 2 ). This is not surprising because aromatic hydrocarbons like toluene can have negative effects on bacterial growth through disruption of the structure and function of cell plasma membranes 40 41 . Normal bacterial growth can be sustained by providing volatile hydrocarbons via the gas phase, which reduces toluene toxicity to the cells 40 42 . Mutualistic expansion patterns were radically modified by replacing toluene with benzoate (while keeping all other conditions constant). Benzoate could be used independently by both members of the consortium and therefore led to a form of exploitation competition 35 . This change induced a total segregation between Pp F4 and Pp F107 populations during radial expansion on surfaces, and led progressively to the competitive exclusion of the Pp F107 genotype at the edge of the colony ( Fig. 3 , Supplementary Fig. S2 ). It was not entirely clear whether Pp F107’s reduced fitness was due only to a lower growth rate ( Supplementary Fig. S3 ), or if Pp F4 also had a direct negative effect on Pp F107 (interference competition 35 ). Both strains grew faster with benzoate than with toluene, which is also consistent with microscopic observations of cell growth and morphology ( Supplementary Fig. S8 ). Genotypic demixing and the reduction of diversity observed at the growing radial front follow general rules associated with dispersal of microbial communities in space 43 . However, strong mutualistic interactions mitigate demixing and promote coexistence 38 44 . Spatial contingency was evident through observations that i) the initial distribution of consortium members could impact growth on surfaces, but not in liquid, and ii) some configurations of consortium members perform as well as the wildtype on surfaces (as measured by CFU counts), but not in liquid ( Fig. 4 ). This is consistent with the view that spatial arrangement is required for the optimization of specific microbial processes, which is particularly relevant for the design of synthetic communities 30 45 46 . Related to this is the notion that the existence of a spatially structured habitat is a prerequisite for the evolution of mutualism from competitive ancestors 44 47 48 , or for the stabilization of exploitative interactions 11 . In our experiments with toluene we have not observed emergence of colony sectors exhibiting altered spatial growth patterns, which suggested that the conditions and time span did not select for new mutant genotypes that would enhance cooperative mutualism in the consortium. In the experiments with benzoate, we sometimes observed subpopulations of Pp F107 cells that appeared to outcompete Pp F4 cells at the edge of the growing colony (see for example in Fig. 3a ). However, we further showed that Pp F107 isolated from the edge of the colony did not compete better with Pp F4 than the original Pp F107 ( Supplementary Fig. S9 ), and thus that formation of Pp F107 sectors at the edge was likely due to chance only. It is possible that longer experimental time spans (weeks to months) could lead to emergence and enrichment of better adapted mutants and thus to onset of genetic drift in the consortium. Although such evolutionary questions are of interest, they are beyond the scope of the present study. Our results demonstrated that cooperative mutualism in toluene degradation favored the emergence of strongly intermixed patterns, with width of clonal strands decreasing with radial expansion (in the range of 10 to 100 μm) suggesting persistence of the mutualistic interactions ( Figs 2 , 3 , Supplementary Fig. S5 ). Similar values have been reported in a yeast two-member consortium that cross-feed on amino acids 38 39 , which supports the view that spatial proximity is a conserved ecological trait for metabolic cooperation between genotypes that rely on diffusion 44 49 . This intermixed pattern stood in stark contrast with strongly segregated patterns emerging under growth on benzoate that removed the obligate mutualistic dependency and resulted in alternating single strain bands >100 μm wide ( Fig. 3 ). Under mutualistic growth conditions, Pp F4 cells rely for their maintenance and growth on the aqueous diffusion of 3-methylcatechol (derived from toluene) from neighboring strands of Pp F107 cells. Hence, minimization of diffusion distances between consumer ( Pp F4) and producer ( Pp F107) favor trophic interactions, while larger physical separation can lead to starvation. Such considerations have utmost importance for soil bacteria like P. putida , because of the non-mixed, unsaturated state of their habitat. Indeed, soil contains complex pore spaces, fragmented aquatic domains and limited nutrient diffusion fields 19 50 . To better understand consortium assembly on rough porous soil surfaces where the aqueous phase dynamics and connectedness could affect spatial self-organization, we employed a porous surface model (PSM) that mimics unsaturated soil surfaces, with control over surface hydration status while allowing direct microscopic observations 36 51 . Toluene-degrading consortium grew on the PSM ( Fig. 5 ), but, as on agar surfaces ( Fig. 3 ), range expansion was reduced compared to growth on benzoate. Importantly, drier conditions (i.e., lower water matric potential and hence reduced aqueous connectivity on the PSM) reduced growth and expansion for both carbon sources, which had been previously observed with Pseudomonas putida growing on PSM 51 . We recently showed that such relatively mild suction conditions (−2 kPa) resulted in thinner and disconnected liquid films on the PSM, constraining bacterial motility and restraining dispersal radii to ≈10 μm 22 . Therefore, consortium self-assembly is likely to be limited by both low cell density and low water matric potential 12 . Importantly, results on PSM showed how template ‘intermixed’ and ‘segregated’ patterns such as observed on agar surfaces could be modified by additional non-biological factors, in this case geometry (surface roughness) and hydration conditions. Although not explored in this study, many other physicochemical factors (temperature, salinity, pH, etc.) could possibly influence spatial patterns of organization. The importance of trophic preferences and nutrient diffusion illustrated in this study contribute to the understanding of scales of microbial community organization and functioning in natural habitats such as soil. In particular, biophysical considerations are needed to interpret observations of microbial community distribution and activity in soil aggregates 52 . For instance, studies have revealed the importance of oxygen gradients 25 , of carbon source distribution and concentration 52 53 and of aggregate pore size 54 for the maintenance of stable microbial patterns and biochemical activity. These observations are best explained within a theoretical framework of microbial life in soil that encompasses physical factors (nutrient diffusion, pore geometries, water availability) and biological factors (cell motility and growth, trophic dependencies), and that is prerequisite for mechanistic understanding of microbial diversity and activity in porous media like soils 55 . Moreover, such a theoretical framework paves the way for hypothesis testing and predictions of microbial community structure and function in soils whose characteristics are (partially) known. For example, we would predict that steep gradients of one limiting carbon source would promote microbial self-organization and interactions in porous media under conditions that support cell motility, while the presence of abundant and complex mixtures of nutrient compounds would obviate the need for specific spatial positioning and thus suppress self-organization. In conclusion, our study demonstrated how simple trophic dependencies directly shape spatial patterning and self-organization of bacterial populations inhabiting surfaces. Cooperative mutualism in degrading a carbon source (toluene) imposed proximity at the micrometric range, and determined the ultimate relative abundances of bacterial partners in the consortium. Such intermixing or segregated patterns represent ecological templates that potentially drive myriad metabolic functions in natural consortia (typically surface-attached in terrestrial systems). It does not follow that patterns identical to those that we see on model surfaces would necessarily be observed in natural habitats: many confounding factors such as habitat geometry or nutrient availability could constrain pattern formation. However, our results demonstrate how trophic dependencies are a determining component of microbial self-organizing processes. In addition to trophic dependencies, we pointed out the roles of cell distribution in space and hydration conditions in controlling self-assembly of mutualistic consortium confirming theoretical predictions 12 . This is of high relevance in unsaturated soils, where average distances between microbial foci can be high and liquid connectivity relatively low. We hypothesize that the imposed proximity between microorganisms (be it driven by trophic dependency or constrained by physical factors) is likely to be an important factor in maintaining species coexistence, horizontal genetic exchanges and in supporting biogeochemical functions. Beyond soils, elucidating links between microbial spatial arrangement and function is beneficial in the context of consortia engineering for industrial, environmental, or health purposes, where the formulation of ‘natural laws’ of microbial spatial organization would prove of great value."
} | 3,144 |
38528784 | PMC10963910 | pmc | 3,593 | {
"abstract": "Abstract Medium‐chain‐length α,ω‐diols (mcl‐diols) play an important role in polymer production, traditionally depending on energy‐intensive chemical processes. Microbial cell factories offer an alternative, but conventional strains like Escherichia coli and Saccharomyces cerevisiae face challenges in mcl‐diol production due to the toxicity of intermediates such as alcohols and acids. Metabolic engineering and synthetic biology enable the engineering of non‐model strains for such purposes with P. putida emerging as a promising microbial platform. This study reviews the advancement in diol production using P. putida and proposes a four‐module approach for the sustainable production of diols. Despite progress, challenges persist, and this study discusses current obstacles and future opportunities for leveraging P. putida as a microbial cell factory for mcl‐diol production. Furthermore, this study highlights the potential of using P. putida as an efficient chassis for diol synthesis.",
"conclusion": "CONCLUSION \n P. putida KT2440 has demonstrated its potential as a non‐model microbial cell factory for mcl‐diol production. The construction and evaluation of a four‐module diol synthesis pathway have been successful, removing some barriers to mcl‐diol synthesis. However, the full exploitation of its robust flux and capacity remains untapped, with a need to address remaining hurdles in future research. Furthermore, the ability to utilize diverse and economical resources (e.g. lignocellulosic biomass, wasted plastics) could position P. putida to produce cost‐effective diols, giving it a competitive edge over other microbial cell factories. A deeper understanding of pseudomonal metabolism and regulations will be instrumental in transforming P. putida into a more efficient microbial cell factory for the sustainable production of mcl‐diols.",
"introduction": "INTRODUCTION Biomanufacturing contributes to the transition from a fossil fuel‐based economy to a sustainable and circular one. This innovative approach harnesses the metabolic capabilities of microorganisms, cells, and enzymes to successfully synthesize many valuable products, such as biofuels, bioplastics, pharmaceuticals, food and beverages, and other specialty chemicals (Garces Daza et al., 2023 ; Lu et al., 2017 ; Orsi, Beekwilder, et al., 2021 ; Werner et al., 2021 ; Zhang et al., 2017 , 2019 ; Zhao et al., 2019 ). Biomanufacturing has shown its potential to address the crises humankind faces including climate change, population growth, food security, and plastic accumulation. However, the current product portfolio derived from renewable sources remains narrow. For example, medium‐chain‐length aliphatic α,ω‐diols (mcl‐diols, C 5 ‐C 12 ) are important building blocks for polymer production and used as surfactants, lubricants, and cosmetics ingredients (Figure 1 ) (Bramson et al., 2020 ; Erle et al., 2012 ; Gross et al., 2010 ; Kluge et al., 2020 ; Llevot et al., 2015 ; Youngquist et al., 2013 ). Their production heavily relies on chemical processes that require intensive energy and expensive catalysts, while leading to greenhouse gas emissions. Meanwhile, the need for mcl‐diols is growing at a rate of 8% per year. The market volume of 1,6‐hexanediol, the most commonly used monomer, is expected to reach $1401 million by 2025 (Kim et al., 2021 ). While microbial synthesis of short‐chain diols (scl‐diols) like 1,4‐butanediol has seen significant success in engineered E. coli strains with well‐established and concise biosynthetic pathways (Burgard et al., 2016 ), progress in mcl‐diols production has been relatively constrained. The synthetic pathway of scl‐diols is rooted in intermediates of the tricarboxylic acid cycle, a well‐studied process. On the other hand, for mcl‐diols, the core synthetic module involves either the fatty acid synthetic pathway or the reversed beta‐oxidation pathway. Both pathways for mcl‐diol production encounter challenges including insufficient availability of medium‐chain precursors and inefficiencies in microbial cell factories. FIGURE 1 Applications of medium‐chain‐length aliphatic α,ω‐diols. The model microorganisms such as E. coli and Saccharomyces cerevisiae are commonly used chassis for industrial bioproduction. However popular, they are not always the ideal chassis. Medium‐chain fatty acids and alcohols that are the important intermediates of mcl‐diols synthesis were produced from diverse feedstocks via either a fatty acid synthesis pathway or reversed beta‐oxidation pathway. Due to their toxicity, low product titres were usually reported in model microorganisms. Non‐model microorganisms, therefore, need to be explored and investigated for such applications. Furthermore, the use of cheap feedstocks would make the process more economically feasible, given the low price of diols and their inefficient biosynthesis. Considering the properties of mcl‐diols and their precursors, and the features of the synthetic pathway, an ideal chassis should possess (1) tolerance to high concentrations of organic solvents, alcohols, and acids; (2) robust flux of fatty acid synthesis; (3) available genome sequence and genetic toolboxes; (4) flexible metabolism, utilizing cheap raw materials; (5) metabolic models. Metabolic engineering and synthetic biology have emerged as enabling technologies for the microbial synthesis of numerous desired compounds from sustainable feedstocks. Using these techniques, various non‐model strains are constructed and subsequently optimized through the iterative ‘Design‐Build‐Test‐Learn’ (DBTL) cycle to improve titres (g/L), rates (g/L/h), and yields (g product /g substrate ) (Nielsen & Keasling, 2016 ; Orsi, Claassens, et al., 2021 ). \n Pseudomonas putida KT2440 as a non‐model strain has attracted more and more interest in academia and industry. This gram‐negative soil bacterium is viewed as a promising microbial cell factory owing to its high metabolic versatility, low maintenance demands, tolerance to harsh conditions (including high concentrations of organic solvents and oxidative stress), efficient genetic toolbox, available genome‐scale metabolic model, and excellent capacity to produce fatty acid‐based storage materials, such as polyhydroxyalkanoate (PHA) and rhamnolipids (Batianis et al., 2020 ; Belda et al., 2016 ; Volke et al., 2020 ; Wirth et al., 2019 ). The commercial production of PHA using engineered P. putida has proven its robust flux channelling substrates via fatty acid synthesis or reversed beta‐oxidations pathway (Bluepha® and PhaBuilder®, China). In the past decade, an increasing number of tools have been developed for genome modification and gene regulation in P. putida (Martin‐Pascual et al., 2021 ). Overall, the biosynthesis of mcl‐diols in P. putida is seen as a viable alternative. In this study, we summarize the characteristics of P. putida as a non‐model microbial cell factory and the progress of mcl‐diol production in P. putida . Furthermore, we point out the opportunities and challenges of using P. putida as a platform for diol production. Finally, we proposed future prospects that can further improve its performance."
} | 1,797 |
35108266 | PMC8853641 | pmc | 3,594 | {
"abstract": "The ammonia-oxidizing bacterium Nitrosomonas europaea has been widely recognized as an important player in the nitrogen cycle as well as one of the most abundant members in microbial communities for the treatment of industrial or sewage wastewater. Its natural metabolic versatility and extraordinary ability to degrade environmental pollutants (e.g., aromatic hydrocarbons such as benzene and toluene) enable it to thrive under various harsh environmental conditions. Constraint-based metabolic models constructed from genome sequences enable quantitative insight into the central and specialized metabolism within a target organism. These genome-scale models have been utilized to understand, optimize, and design new strategies for improved bioprocesses. Reduced modeling approaches have been used to elucidate Nitrosomonas europaea metabolism at a pathway level. However, genome-scale knowledge about the simultaneous oxidation of ammonia and pollutant metabolism of N . europaea remains limited. Here, we describe the reconstruction, manual curation, and validation of the genome-scale metabolic model for N . europaea , i GC535. This reconstruction is the most accurate metabolic model for a nitrifying organism to date, reaching an average prediction accuracy of over 90% under several growth conditions. The manually curated model can predict phenotypes under chemolithotrophic and chemolithoorganotrophic conditions while oxidating methane and wastewater pollutants. Calculated flux distributions under different trophic conditions show that several key pathways are affected by the type of carbon source available, including central carbon metabolism and energy production.",
"introduction": "1. Introduction Ammonia (NH 3 ) as soluble ammonium (NH 4 + ) is one of the main pollutants in industrial wastewater effluents, reaching concentration values from 5 to 1,000 ppm [ 1 ]. Biological nitrification is the primary method to remove NH 4 + from wastewaters. This process involves the oxidation of NH 4 + to nitrate (NO 3 - ) via nitrite (NO 2 - ). Nitrification is catalyzed by ammonia-oxidizing and nitrite-oxidizing bacteria in a two-step autotrophic process [ 2 ]. Several studies have used nitrifying consortia as mechanism to remove NH 4 + and toxic pollutants (e.g., benzene, toluene, and phenol) simultaneously [ 3 – 6 ]. Nitrosomonas europaea ( Ne ) is a well-studied ammonia-oxidizing bacterium highly present in nitrifying consortia (usually from 1% to 5%) as well as an important microorganism in the nitrogen cycle [ 3 , 7 ]. Ne is typically grown with bicarbonate (HCO 3 - ) as the major inorganic carbon substrate [ 8 ]. The HCO 3 - assimilated is transformed to CO 2 through the activity of the anhydrase carbonic enzyme [ 9 ]. The CO 2 is further fixed by the Calvin-Benson-Bassham (CBB) cycle [ 10 ], producing energy by converting NH 3 to hydroxylamine (NH 2 OH), and then to NO 2 - (chemolithotrophic metabolism). Reactions catalyzed by the ammonia monooxygenase (AMO) and hydroxylamine oxidoreductase (HAO) [ 11 ]. AMO can metabolize various toxic pollutants present in wastewater, such as aromatic hydrocarbons and halogenated aliphatic compounds [ 12 – 15 ]. Physiological data highlighted have shown Ne’s versatility to utilize various substrates (pyruvate and fructose as sole organic carbon sources) under aerobic conditions [ 16 ]. However, the internal metabolic processes as the simultaneous NH 4 + assimilation and toxic compounds oxidation, or the capability of an organic carbon mineralization (chemolithoorganotrophic metabolism) by Ne are not well characterized to date. Reduced modeling approaches have been used to elucidate Ne metabolism at a pathway level, resulting in two metabolic models: a metabolic network model with 51 reactions and a genome-scale metabolic model (M-model) with 1,102 total reactions [ 17 , 18 ]. Nevertheless, there are no M-models that enable elucidating flux distributions showing the mineralization of an organic carbon or the pollutant oxidation process under ammonia assimilation conditions. Here, we reconstructed a genome-scale model for N . europaea ATCC19718 using semi-automated methods [ 19 , 20 ]. The resulting model was manually curated to improve the quality of the phenotypic predictions. The model contains 1,149 reactions, and it is capable to accurately simulate growth under chemolithoorganotrophic and chemolithotrophic conditions while oxidating pollutants and methane.",
"discussion": "3. Discussion 3.1 Model reconstruction and refinement The model i GC535 was highly curated and successfully validated under various growth conditions, achieving the highest growth phenotype accuracy ( Table 2 ) compared with all the Ne available models to date. Out of all the Ne available models (5 total), only i GC535 and i FC578 can simulate growth and nitrite production under chemolithotrophic conditions. But only i GC535 incorporates the phosphorylate oxidation chain while being fully mass and charge balanced, enabling the simulation of pollutant transformation pathways and growth under both chemolithoorganotrophic and chemolithotrophic conditions. To our knowledge, i GC535 is the first genome-scale model of Ne able to predict flux distributions simultaneous oxidation of pollutant and ammonium. Predicted growth phenotypes and fluxes across the network of i GC535 were validated using experimental data (>90% accuracy). Higher prediction accuracies were calculated for chemolithoorganotrophic growth with fructose as the sole carbon source, for oxygen uptake rates under chemolithotrophic conditions, and during methane and pollutant oxidation. The model also provides advanced quantitative insights at the metabolic level about chlorobenzene and phenol oxidation. The manual curation performed to i GC535 resulted in a metabolic network of Ne with high certainty about reaction addition and cofactors usage. For example, using the glycolytic isoenzyme NADH-dependent GAPD instead of NADPH-dependent GAPD. Chain et al. [ 10 ] suggested that Ne saves a significant amount of energy by reducing 3-phosphoglycerate through the NADH-dependent GAPD. Our model simulations showed that Ne carries flux through this reaction (-2.9597 mmol/gDW/h), thus optimizing energy utilization. 3.2 Model-driven insights into Ne metabolism Currently, the complete TCA cycle has been characterized in Ne [ 10 ]. Experimental studies have suggested that all TCA cycle enzymes are active when Ne is growing anaerobically (nitrite as the electron acceptor) while using organic carbon as the carbon source [ 44 , 45 ]. Moreover, it has been also shown that other chemolithotrophs can oxidize organic carbon sources to survive [ 46 ]. Model simulations predicted that NADH16pp oxidizes NADH to NAD + (2.05 mmol/gDW/h) when fructose is present at high concentrations. This flux is reversed when HCO 3 - is the carbon source (-10.19 mmol/gDW/h). Adessi and De Philippis, 2013 [ 47 ] suggested that NADH production instead of NAD + is due to the excess of reduced ubiquinol. Model simulations showed that Ne maintains healthy levels of ubiquinol by activating the reaction NADH16pp. Flux distributions also showed that under chemolithotrophic growth, all TCA cycle enzymes were active except for AKGDH. This prediction agrees with the study done by Beyer et al. and Hooper et al. [ 44 , 48 ], who observed that all TCA cycle enzymes were active except for AKGDH in Ne . However, SUCDi was active at a low flux (8.0939x10 -5 mmol/gDW/h). Experimental observations revealed low activity or no significant expression amount of SUCDi under chemolithotrophy growth. Even more, some attempts to measure SUCDi activity by Deutch, 2013 [ 49 ] were unsuccessful. We believe that i GC535 will provide insights into the experimental design to better understand SUCDi activity in Ne . i GC535 provides high resolution at metabolic and electron transfer levels. For example, model simulations showed that 50% of the electron flux from HAO returns to AMO under chemolithotrophic conditions, 30.8% passes to the terminal oxidase, and 19.25% goes towards NADH16pp. Wood 1986 [ 50 ], proposed that four electrons are removed from hydroxylamine oxidation by HAO, and two electrons return to AMO, 1.65 passes to the terminal oxidase, and the rest goes to NADH16pp. For this to happen, experimental evidence showed that out of all the electron flux from HAO, 50% goes to AMO, 41% to the terminal oxidase, and 8.7% to NADH16pp. Although the prediction of the flux proportion that returns to AMO coincided with that reported in the literature, the other two percentages diverged. However, the simulation correctly predicted a higher electron flux through AMO than C552oxi, which is subsequently higher than the electron flux through NADH16pp. 3.3 Metabolism change at low fructose concentrations The model predicted the increase in RBPC activity when the fructose concentration drops. Thus, CO 2 is fixed. Other organisms, such as Pseudomonas oxalaticus OX1, have also shown a progressive increase in RBPC activity and CO 2 fixation when acetate concentration (organic carbon source) decreases in formate-limited culture [ 51 ]. Overall, i GC535 is a reliable systems biology tool that will be the base to understand and generate new hypothesis about Ne metabolism under a great variety of growth conditions."
} | 2,346 |
33521845 | PMC7880967 | pmc | 3,596 | {
"abstract": "Graphical abstract Chitin is an abundant waste product from shrimp and mushroom industries and as such, an appropriate secondary feedstock for biotechnological processes. However, chitin is a crystalline substrate embedded in complex biological matrices, and, therefore, difficult to utilize, requiring an equally complex chitinolytic machinery. Following a bottom-up approach, we here describe the step-wise development of a mutualistic, non-competitive consortium in which a lysine-auxotrophic Escherichia coli substrate converter cleaves the chitin monomer N -acetylglucosamine (GlcNAc) into glucosamine (GlcN) and acetate, but uses only acetate while leaving GlcN for growth of the lysine-secreting Corynebacterium glutamicum producer strain. We first engineered the substrate converter strain for growth on acetate but not GlcN, and the producer strain for growth on GlcN but not acetate. Growth of the two strains in co-culture in the presence of a mixture of GlcN and acetate was stabilized through lysine cross-feeding. Addition of recombinant chitinase to cleave chitin into GlcNAc 2 , chitin deacetylase to convert GlcNAc 2 into GlcN 2 and acetate, and glucosaminidase to cleave GlcN 2 into GlcN supported growth of the two strains in co-culture in the presence of colloidal chitin as sole carbon source. Substrate converter strains secreting a chitinase or a β-1,4-glucosaminidase degraded chitin to GlcNAc 2 or GlcN 2 to GlcN, respectively, but required glucose for growth. In contrast, by cleaving GlcNAc into GlcN and acetate, a chitin deacetylase-expressing substrate converter enabled growth of the producer strain in co-culture with GlcNAc as sole carbon source, providing proof-of-principle for a fully integrated co-culture for the biotechnological utilization of chitin. Key Points • A bacterial consortium was developed to use chitin as feedstock for the bioeconomy. • Substrate converter and producer strain use different chitin hydrolysis products. • Substrate converter and producer strain are mutually dependent on each other. Supplementary Information The online version contains supplementary material available at 10.1007/s00253-021-11112-5.",
"introduction": "Introduction In biotechnology, it is desirable to replace food-grade feedstocks by secondary feedstocks derived from organic waste for both economic and environmental reasons (Wendisch et al. 2016 ; Abu Yazid et al. 2017 ). Chitin, one of the most abundant biopolymers on earth, is a polysaccharide that occurs in large amounts in the waste streams of different industries, e.g. in crustacean shells originating from marine fisheries, or fungal mycelium wastes arising from mushroom farming and fungal fermentations for the production of enzymes (Teng et al. 2001 ; Nisticò 2017 ). Among these sources, fungal fermentation waste is the most reproducibly available and the least contaminated one (Cai et al. 2006 ). The volume of this waste stream is difficult to quantify (Ghormade et al. 2017 ). Acetic acid production by Aspergillus niger alone probably yields about 0.1–0.2 Mt of dry mycelium annually, and this may have to be multiplied by a factor of 2–3 when fungal fermentations for other fine chemicals or enzymes are considered as well. Today, this potentially precious resource, in spite of its constant high quality, is most often either burned or transported to landfills instead of being upcycled in the interest of sustainable resource management. Most typical biotechnological producer strains, such as Corynebacterium glutamicum or Escherichia coli, do not grow on chitin as they lack a functional chitinolytic machinery (Keyhani and Roseman 1997 ; Verma and Mahadevan 2012 ). However, expressing chitinolytic enzymes in producer strains has limitations as these strains are usually genetically engineered for maximum product formation so that the additional expression of genes for chitin degradation may lead to decreased productivity (Jagmann and Philipp 2014 ; Cavaliere et al. 2017 ). A strategy to overcome this limitation and to uncouple chitin degradation from product formation is the establishment of synthetic microbial consortia (Sgobba and Wendisch 2020 ). In such a consortium, a substrate converter would generate the carbon and energy source from the substrate, e.g. chitin, for itself and for a producer strain. To avoid competition, the substrate converter should produce two different substrates that are mutually exclusively accessible to the two members of the consortium. The inevitable dependency of the producer strain on the substrate converter can stabilize the consortium, and its robustness can be further increased by implementing an auxotrophy into the substrate converter that is complemented by the producer strain. This concept of co-cultures using alternative feedstocks has mainly been applied for cellulosic substrates (Minty et al. 2013 ; Wang et al. 2015 ; Wen et al. 2017 ) but to our knowledge has only recently been applied for using chitin as a fermentation substrate (Ma et al. 2020 ). Figure 1 shows a hypothetical chitin-based consortium of a lysine-auxotrophic substrate converter strain which degrades the GlcNAc-polymer chitin to yield GlcN and acetate, growing on the latter product while making the former available for growth and lysine production by a producer strain. In principle, two different approaches could be chosen, either a top-down approach in which an existing consortium would be chosen and optimized to perform as wished, or a bottom-up approach in which the two cooperating strains are build from scratch by adding the required traits one by one (Shin et al. 2010 ; Gumulya et al. 2018 ; Gao et al. 2019 ). We opted for a bottom-up approach, using E. coli as a converter and C. glutamicum as a producer. In this scenario, the substrate converter heterologously expresses three chitinolytic enzymes: a chitinase, a chitin deacetylase, and a glucosaminidase, but is unable to take up chitobiose, GlcNAc, or GlcN. In contrast, acetate catabolism is disabled in the producer strain, while it can take up GlcN. Following the step-wise bottom-up strategy towards this eventual goal, we initially focused on the carbon sharing of the final degradation products GlcN and acetate between the two members of the consortium. Next, the enzymatic cleavage of GlcNAc into GlcN and acetate by the substrate converter was established by heterologous expression of a suitable chitin deacetylase. The third step is the generation of monomeric GlcNAc from chitin oligomers, and the final fourth step of the bottom-up approach will be the breakdown of the crystalline chitin polymer to GlcNAc oligomers by secretion of a suitable mixture of chitin degrading enzymes. Fig. 1 Design of the synthetic mutualistic consortium with E. coli and C. glutamicum for l -lysine production with chitin as sole source of carbon and energy. (1) The substrate converter EcLPPLYSA ( E. coli W3110 ΔnagE ΔmanXYZ ΔchbBCA ΔlysA Δlpp::CM ) expresses heterologous enzymes for the degradation of chitin to glucosamine (GlcN) and acetate. (2) EcLPPLYSA can only use acetate as growth substrate because of deletions in uptake systems for the other chitin degradation products. (3) CgLYS4 ( C. glutamicum DM1729 Δpta-ackA Δcat ΔldhA ΔaceAB ΔnanR ) can only use GlcN as growth substrate and for the production of l -lysine because of deletions in acetate metabolism. (4) The consortium is co-stabilized by the lysine auxotrophy of EcLPPLYSA. GlcNAc: N -acetylglucosamine, Fru-6-P: fructose-6-phosphate, GlcN-6-P: glucosamine-6-phosphate, LysA: diaminopimelate decarboxylase, ChbBCA: PTS-system chitobiose-specific, NagE: PTS-system N-acetylglucosamine-specific EIICBA component, ManXYZ: mannose-specific PTS-system\n\nIntroducing metabolic deficiencies into substrate converter strain and producer strain To establish sharing of the carbon sources resulting from chitin degradation, we first had to introduce metabolic deficiencies into the substrate converter and the producer strain. In the E. coli substrate converter strain, genes encoding transporters responsible for the uptake of the amino sugars GlcNAc ( nagE ) and GlcN ( manXYZ ) as well as for the intermediate chitin degradation product chitobiose (Plumbridge and Pellegrini 2004 ; Verma and Mahadevan 2012 ) ( chbBCA ) were deleted. In addition, the lpp gene, encoding Braun’s lipoprotein, which is located in the outer membrane of E. coli , was deleted to improve secretion of the recombinantly expressed enzymes (Shin et al. 2010 ; Chen et al. 2014 ; Müller et al. 2016 ) (Supplemental Fig. S1 ) which were later introduced for chitin degradation. For simplicity, the E. coli substrate converter with the genotype ∆ nagE ∆ manXYZ ∆c hbBCA ∆ lpp will be referred to as EcLPP or EcLPP* (‘*’ indicating that the chloramphenicol resistance gene was removed). Furthermore, lysine-auxotrophic variants of EcLPP/EcLPP* were created by deleting the lysA gene, and these strains were named EcLPPLYSA/EcLPPLYSA*. EcLPPLYSA* was tested for its ability to grow on acetate, GlcN, GlcNAc, and GlcNAc 2 compared to the E. coli W3110 wildtype strain (EcWT; Fig. 2 ). Fig. 2 Growth of the E. coli wild type strain EcWT (red circles) and its mutant strain EcLPPLYSA* (blue squares) on 20 mM (a) sodium acetate, (b) glucosamine, (c) N -acetylglucosamine, or (d) chitobiose. Error bars (mostly smaller than symbols) indicate standard deviation of three independent experiments ( n = 3) While the wild type strain EcWT was able to grow on all four substrates, the mutant strain EcLPPLYSA* had lost its ability to grow on GlcN, GlcNAc, and GlcNAc 2 (Fig. 2b–d ) while still growing on acetate in the presence of Lys (Fig. 2a ). In the C. glutamicum producer strain, the genes encoding acetate kinase ( ackA ), phosphotransacetylase ( pta ), acetyl-CoA:CoA transferase ( cat ), isocitrate lyase ( aceA ), and malate synthase ( aceB ) were deleted to prevent the strain from using acetate (Veit et al. 2009 ). The nanR gene, encoding a repressor of the genes nagA ( N -acetyl- d -glucosamine-6-phosphate deacetylase) and nagB (glucosamine-6-phosphate deaminase) was deleted to allow growth on glucosamine (Uhde et al. 2013 ). Moreover, potential cross-feeding of lactate to the substrate converter was prevented by deletion of ldhA (NAD + -dependent- l -lactate-dehydrogenase; (Okino et al. 2008 )). The resulting strain named CgLYS4 was then tested for growth on acetate, GlcN, and GlcNAc compared to the wildtype C. glutamicum (DM1729; Fig. 3 ). Fig. 3 Growth of C. glutamicum wild type strain DM1729 (red circles), and its mutant strain CgLYS4 (blue squares) on 20 mM (a) sodium acetate, (b) glucosamine, or (c) N -acetyl-glucosamine. Error bars indicate standard deviation of three independent experiments (n = 3) While the wild type strain DM1729 grew well only on acetate and poorly on glucosamine, the mutant strain CgLYS4 grew well on GlcN, but not on acetate. As expected, both wild type and mutant were unable to grow on GlcNAc, since they lacked the GlcNAc PTS system (NagE) (Matano et al. 2014 ).",
"discussion": "Discussion Establishing a bacterial co-culture based on chitin as a substrate is at the same time highly promising and highly demanding. It is promising not only because chitin is abundantly available from different waste streams, but also because it allows to set up a system in which the substrate converter and producer strains grow on different substrates produced from it, namely GlcN and acetate. It is demanding because chitin is a recalcitrant polymer that forms crystalline fibers embedded in complex matrices such as fungal cell walls, insect cuticles, or crustacean shells, requiring a complex set of enzymes for its degradation (Arnold et al. 2020 ). To establish a bacterial co-culture converting chitin to a target product, we first had to introduce different metabolic deficiencies into the substrate converter and the producer strain. Because as an attractive proof of principle, we wanted to establish lysine production from chitin , we decided to offer acetate as an energy and carbon source to the substrate converter E. coli, and GlcN as an energy, carbon and nitrogen source to the amino acid producer C. glutamicum . As a consequence, we had to delete uptake mechanisms for amino sugars in E. coli, namely for the monomers GlcNAc ( nagE ) and GlcN ( manXYZ ) as well as for the dimer chitobiose ( chbBCA ) (Plumbridge and Pellegrini 2004 ; Verma and Mahadevan 2012 ). We also had to disable C. glutamicum from using acetate by deleting genes encoding acetate kinase ( ackA ), phosphotransacetylase ( pta ), CoA-transferase ( cat ), isocitrate lyase ( aceA ), and malate synthase ( aceB ). Additionally, we had to delete the repressor-encoding gene nanR to enable C. glutamicum to grow on GlcN (Uhde et al. 2013 ). The producer strain was further improved for performance in the consortium by deleting ldhA to prevent cross-feeding of lactate to the substrate converter. Both strategies were successful even though the substrate converter strain EcLPPLYSA* was apparently still able to use GlcN after a longer incubation period. While this was observed in single culture, it is most likely of no concern in co-culture with the producer strain which utilizes GlcN much more efficiently so that it will not be available long enough for E. coli to grow on it. Not unexpectedly, establishing chitin utilization in E. coli as the substrate converter proved a lot more demanding than establishing the metabolic deficiencies. Chitin degradation to GlcN and acetate requires the introduction of a whole enzymatic cascade comprised of at least three enzymes as used in this study (Tanaka et al. 2004 ; Mekasha et al. 2017 ). In principle, two alternative approaches are feasible. The first step needs to be chitinase-catalyzed depolymerisation of chitin into small chitin oligomers. These can either be deacetylated by a chitin deacetylase yielding acetate and chitosan oligomers which can then be degraded by glucosaminidase to yield GlcN. Alternatively, these last two steps could occur in reverse order, first degrading chitin oligomers using N -acetylglucosaminidase into GlcNAc which can then be deacetylated by chitin deacetylase to yield GlcN and acetate. Given that chitin occurs in nature as crystalline fibers embedded into complex biological matrices such as fungal cell walls, insect cuticles, or crustacean shells, even more enzymes such as β-glucanases, proteases and lytic chitin monooxygenases will eventually be required for an efficient utilization of these biomaterials available on large scale from different waste streams. To develop such a system, a bottom-up approach is best suited, step-by-step establishing substrate degradation ‘in reverse’, starting with the final step (Shin et al. 2010 ; Jia et al. 2016 ; Gumulya et al. 2018 ). Depending on which of the alternative scenarios described above is chosen, the final step would be glucosaminidase-catalyzed degradation of chitosan oligomers to GlcN, or chitin deacetylase-catalyzed degradation of GlcNAc to GlcN and acetate. We opted for the latter scenario as only this one concomitantly produces both substrates required for the growth of the substrate converter and producer strain, allowing to achieve bottom-up proof-of-principle by establishing the co-culture on GlcNAc as a substrate. Consequently, we had to select a suitable chitin deacetylase able to act on the monomer GlcNAc. The only enzyme known with this ability is TkCDA from T. kodakarensis, an enzyme naturally involved in chitin utilization by this bacterium (Tanaka et al. 2004 ). As T. kodakarensis is a hyperthermophile, we tested the temperature dependency of the enzyme and found maximum enzyme activity at 54 °C, and about one third lower activity at 37 °C. Interestingly, and unexpectedly given the above scenarios, TkCDA is known, in T. kodakarensis , to act in concert with a glucosaminidase, not with a N -acetylglucosaminidase. TkCDA can act not only on GlcNAc, but also on chitin oligomers, deacetylating only the GlcNAc unit at the non-reducing end of the oligomers. The thus produced GlcN unit is then cleaved off by glucosaminidase, and the resulting smaller chitin oligomer can again be used as a substrate by TkCDA. For the purpose of establishing the co-culture-based utilization of chitin, this allowed us to start the bottom-up approach using TkCDA-catalyzed deacetylation, leaving both options, i.e. the addition of a glucosaminidase or of an N -acetylglucosaminidase, as the next step. We believe that the chitin deacetylase/glucosaminidase pathway may have evolved in T. kodakarensis to avoid the need for a N -acetylglucosaminidase which might be toxic for a bacterium with a GlcNAc containing murein-based cell wall. Therefore, an analogous approach was followed when attempting to establish the utilization of chitin polymer by the substrate converter, adding the glucosaminidase TK from the same organism and the chitinase ChiB from S. marcescens (Brurberg et al. 1996 ; Horn et al. 2006 ). When these enzymes were added separately into the substrate converter EcLPP*, they conveyed the expected abilities, i.e. to cleave chitin into chitobiose and chitobiose into GlcN. However, when the two enzymes were combined in one strain with a lysine auxotrophy and already expressing TkCDA to generate a substrate converter strain that can be tested in co-culture with lysine producing C. glutamicum strain CgLYS4, no growth of the producer strain was observed on colloidal chitin as a substrate. This was also the case when the three enzymes were produced in three separate substrate converter strains and grown in multiple co-culture with the producer strain. Most likely, enzyme production and secretion or enzyme efficiency on colloidal chitin were not sufficient to provide enough substrates for growth of the converter and/or producer strain. Unfortunately, quantification of enzyme activities and of their products as well as of lysine production was not possible due to the complexity of the M9extra medium which interfered with the HPLC analysis of the supernatants. To open up this bottleneck, it will be required to improve secretion even more than already achieved by deleting the lpp gene (Shin et al. 2010 ; Chen et al. 2014 ; Müller et al. 2016 ), and to develop a more efficient chitinolytic enzyme machinery, making use of what nature has to offer. Improving secretion might be achieved by testing different signal peptides for all three enzymes, since it has been shown in literature that the secretion efficiency not only depends on the signal peptide alone, but that it varies depending on the combination of signal peptide and enzyme (Brockmeier et al. 2006 ; Hemmerich et al. 2016 ). In addition, other secretion systems could be tested, including heterologous expression of translocation systems (Albiniak et al. 2013 ) or fusion of the recombinant protein to carrier proteins (Zhang et al. 2006 ). Well-known chitin degrading bacteria in soil, marine systems, and freshwater habitats are S. marcescens (Vaaje-Kolstad et al. 2013 ), Vibrio spec. (Meibom et al. 2004 ), and Aeromonas hydrophila (Zhang et al. 2017 ), respectively. All of them possess complex chitinolytic machineries consisting of several chitinases and a chitin-degrading lytic polysaccharide monooxygenase (LPMO). As the co-culture required a freshwater salinity, we investigated the chitinolytic enzymes of A. hydrophila . In fact, an E. coli strain secreting a chitinase from A. hydrophila as well as an LPMO from S. marcescens proved that simultaneous production and secretion of these two types of enzymes by E. coli is possible (Yang et al. 2017 ). We have identified the A. hydrophila strain AH-1 N based on an enrichment approach with chitin as substrate (Stumpf et al. 2019 ), and characterized its chitinase AH-1NChi as being rather efficient on crystalline chitin, and as acting synergistically with the LPMO AhLPMO10A from the same strain (Vortmann et al. unpublished). These may in future be used to improve the performance of the substrate converter. Alternatively, additional chitinases with activities complementing that of ChiB such as ChiA and ChiC of S. marcescens , alongside its LPMO, might be used (Purushotham et al. 2012 ; Vaaje-Kolstad et al. 2013 ; Manjeet et al. 2013 ). For all E. coli strains that heterologously expressed enzymes, the CFUs declined. A decrease in the number of dividing cells for an E. coli strain expressing a recombinant enzyme, measured by their CFU ability, has previously been reported by Andersson et al. ( 1996 ). They suggested that cells, whose heterologous gene expression was induced by IPTG, segregate and some cells enter the viable but non-culturable state (VBNC), meaning that they are incapable of division but still retain their metabolic activity. This might be explained by nutrient limitation, as a high amount of energy and carbon is needed for production of the recombinant enzymes and is therefore not available for growth. This may lead to a stress situation for the cells which has been described to lead to the VBNC-status (Oliver 2010 ; Ramamurthy et al. 2014 ). As cells which have entered the VBNC-status cannot be detected in CFU-assays, the observed decrease in CFUs of EcLPPLYSA* [TkCDA] does not necessarily imply a decrease in total number of cells in the culture, though this of course can also not be excluded. Clearly though, the amount of secreted TkCDA must have been high enough to provide enough GlcN for growth of CgLYS4 and production of sufficient l -lysine to complement lysine auxotrophy of the substrate converter, even though the CFUs of strain EcLPPLYSA* [TkCDA] decreased. Moreover, growth of the cells in co-culture has only been monitored via CFU counts on solid media, not considering that cells might behave differently in liquid media, possibly even showing growth in liquid medium. Our study provides proof-of-principle for the bottom-up development of a synthetic bacterial consortium eventually able to utilize the recalcitrant biopolymer chitin from food and biotechnology waste streams for the production of fine chemicals such as amino acids. A number of synthetic consortia have previously been described that divide labour with respect to substrate conversion and product formation, representing steps towards a fully integrated, interdependent mutualistic and non-competitive consortium as demonstrated here for the first time (Sgobba and Wendisch 2020 ). Conversion of cellulose to isobutanol has been demonstrated by co-culturing the cellulase-secreting fungus Trichoderma reesei with an isobutanol producing E. coli strain (Xin et al. 2019 ) and similarly, conversion of sugarcane bagasse slurry to ethanol has been achieved by co-culturing Saccharomyces cerevisiae that ferments glucose to ethanol and glucose negative ethanologenic E. coli that ferments xylose to ethanol (Wang et al. 2019 ). A co-culture of an l -lysine auxotrophic, naturally sucrose-negative E. coli strain and a C. glutamicum strain producing l -lysine and fructose from sucrose established commensalism in which the E. coli strain benefitted from the C. glutamicum strain that, however, was not dependent on the E. coli strain so that no mutualistic interdependence was established (Sgobba et al. 2018 ). An extension of this consortium comprised an α-amylase secreting l -lysine auxotrophic E. coli strain, allowing it to mutualistically grow on starch with a naturally amylase-negative lysine producing C. glutamicum strain (Sgobba et al. 2018 ). Growth of this mutualistic consortium required lysine cross-feeding and hydrolysis of starch to glucose, for which both strains competed as carbon and energy source for growth. The GlcNAc-converting consortium described here is equally mutualistic and depending on lysine cross-feeding, but it extends the concept significantly by avoiding competition: here, the carbon source is divided between the partners such that E. coli grows with acetate and C. glutamicum with GlcN. Stepwise extension of the concept of division of labour regarding access to substrates will likely develop further as has been seen with respect to division of labour between different steps of product formation from shorter to longer linear cascades to converging designs (Sgobba and Wendisch 2020 ). Eventually, this concept of labour division within a fully integrated, interdependent, mutualistic, non-competitive synthetic microbial consortium can be developed into a versatile platform for modular synthetic biotechnology where substrate converter strains using different substrates will be combined with producer strains yielding different products. Clearly, the benefit of using hexosamines or aminosugar containing polymers such as chitin as a substrate is that in addition to providing a carbon source, these carbohydrates also serve as a source of nitrogen which is required in high amounts for the production of many interesting organic compounds, such as amino acids. Moreover, unlike widely used carbon sources such as glucose or starch, chitin cannot be used as food, feed, or fuel, avoiding competition with these fields."
} | 6,346 |
38201679 | PMC10780632 | pmc | 3,598 | {
"abstract": "Super-wetting interface materials have shown great potential for applications in oil–water separation. Hydrogel-based materials, in particular, have been extensively studied for separating water from oily wastewater due to their unique hydrophilicity and excellent anti-oil effect. In this study, a superhydrophilic and underwater superoleophobic bamboo cellulose hydrogel-coated mesh was fabricated using a feasible and eco-friendly dip-coating method. The process involved dissolving bamboo cellulose in a green alkaline/urea aqueous solvent system, followed by regeneration in ethanol solvent, without the addition of surface modifiers. The resulting membrane exhibited excellent special wettability, with superhydrophilicity and underwater superoleophobicity, enabling oil–water separation through a gravity-driven “water-removing” mode. The super-wetting composite membrane demonstrated a high separation efficiency of higher than 98% and a permeate flux of up to 9168 L·m −2 ·h −1 for numerous oil/water mixtures. It also maintained a separation efficiency of >95% even after 10 cycles of separation, indicating its long-term stability. This study presents a green, simple, cost-effective, and environmentally friendly approach for fabricating superhydrophilic surfaces to achieve oil–water separation. It also highlights the potential of bamboo-based materials in the field of oil–water separation.",
"conclusion": "4. Conclusions In summary, the BC-coated membrane was fabricated using a green and feasible method that involved green dissolution, dip-coating, and regeneration processes. The resulting composite membrane exhibits unique properties of superhydrophilicity (water contact angle, WCA, ~0°) and underwater superoleophobicity (oil contact angle, OCA, ~151°). This super-wetting membrane has the capability to effectively separate oil and water through a process known as “water-removing”, where it can retain oil while allowing water to pass through under gravity. The separation efficiency of the prepared super-wetting membrane is above 98% for various oil/water mixtures, reaching up to 99.96%, with a permeate flux of up to 9168 L·m −2 ·h −1 . These results demonstrate the efficient oil/water separation performance of the membrane. Additionally, the membrane maintains a separation efficiency of over 95% even after 10 cycles of separation, indicating its excellent long-term stability. This study highlights the potential of using bamboo cellulose, which is cost-effective, biodegradable, and readily available, in the treatment of industrial oily wastewater.",
"introduction": "1. Introduction The increasing pollution of oily wastewater poses a significant threat to human health and the ecological environment. Separating oil and water remains a global challenge [ 1 , 2 , 3 ]. Oily wastewater pollution originates from various sources, including oil spill accidents, and industrial and domestic oily wastewater. The composition of oily wastewater is highly complex [ 4 , 5 , 6 ]. Consequently, the preparation of efficient oil/water separation materials has become a critical focus in treating oily wastewater pollution. Recently, numerous methods have been reported to address this issue, such as sedimentation separation [ 7 ], filtration [ 8 , 9 , 10 ], membrane separation [ 11 , 12 , 13 , 14 ], the flotation method [ 15 ], adsorption [ 16 , 17 , 18 ], coagulation [ 19 , 20 ], and chemical solidification [ 21 , 22 , 23 ]. However, these methods still have drawbacks, including high cost, poor separation efficiency, high energy consumption, and a lack of environmental friendliness, which limit their widespread application. Most reported oil–water separation materials have been modified using low-surface-energy substances containing fluorine (fluoroalkyl silane, FAS) [ 24 , 25 ]. Additionally, various chemical polymers/materials with non-biodegradability, such as polyurethane foam [ 26 , 27 ], polytetrafluoroethylene [ 28 , 29 ], and polydimethylsiloxane [ 30 , 31 , 32 ], have been selected as substrate materials. Therefore, there is an urgent need to design and develop cost-effective and environmentally friendly super-wetting separation materials, as well as simple and efficient separation methods, to address the serious problem of oily wastewater pollution. Cellulose is a linear polymer made of β (1–4)-d-glucose unit, which is primarily produced by plants (such as cotton, wood, straw, and bamboo), algae, tunicates, and some bacteria, serving as a structural polymer [ 33 ]. Due to its unique properties, such as availability in large quantities, low costs, biocompatibility, biodegradability and good chemical stability, cellulose also finds applications in fields such as biomedicine, tissue engineering and biomimetic materials [ 34 , 35 , 36 , 37 ]. Furthermore, a large number of hydroxyl groups are contained in cellulose, resulting in easy modification and the creation of active sites for fabricating super-wettability surfaces in oil/water separation materials. Therefore, cellulose stands out as the most favorable material for designing and preparing eco-friendly super-wetting oil/water separation materials. With the rapid development of design ideas and fabrication technologies, a significant number of cellulose-based oil/water separation materials have emerged [ 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 ]. For example, Ma et al. reported the successful development of cellulose-coated cotton fabric with hydrophilic and underwater oleophobic properties. This fabrication process involved a simple step of cellulose dissolution and regeneration, resulting in a remarkable capability to separate a wide range of highly emulsified oil–water mixtures with an excellent separation efficiency (>93.2%) and reasonable flux (>4000 L·m −2 ·h −1 ) [ 50 ]. The CNF-PDMS (cellulose nanofiber-polydimethylsiloxane) aerogel sheets with superhydrophobic, elastic, and anisotropic characteristics were prepared using a facile method that combined directional freeze-drying, soaking in a polydimethylsiloxane (PDMS) solution, and heat treatment. This method achieved continuous oil/water separation through filtration, with a flux of up to 2800 L·m −2 ·h −1 and a separation efficiency of 99.9% [ 51 ]. Li and colleagues constructed a novel carbon aerogel composed of natural microfibrils/regenerated cellulose by dissolving economical hardwood pulps in a solvent consisting of a deep eutectic solvent (DES) and NMMO·H 2 O. This carbon aerogel exhibited high compressibility, outstanding elasticity, excellent fatigue resistance, good separation performance, and recyclability [ 52 ]. Su and co-authors reported an anisotropic cellulose nanocrystalline sponge sheet using a facile and straightforward fabrication method. This sponge sheet exhibited ultrahigh water fluxes (100,000 L·m −2 ·h −1 ) and oil/water selectivity (up to 99.99%) [ 53 ]. While previous reports focused on cellulose derived from hardwood pulps, cotton, and cellulose nanocrystalline, there is relatively little research on bamboo cellulose. Therefore, the development and utilization of super-wetting materials made from bamboo cellulose hold great significance. Bamboo cellulose (BC) originates from natural bamboo and possesses characteristics such as high strength, excellent breathability, good mechanical properties, and biodegradability [ 54 , 55 , 56 , 57 , 58 , 59 ]. Peng and co-authors developed a method to prepare bamboo cellulose with different levels of Mη through one-step oxidation based on natural BC. This development has the potential to contribute to the advancement and utilization of functional materials made from bamboo cellulose [ 60 ]. Additionally, they successfully fabricated a super-hydrophobic/superoleophilic BC foam using a versatile method of surface hydrophobic modification for oil/water separation [ 41 ]. However, the oil/water separation process of cellulose-based foam with super-wetting is complex and discontinuous, which hinders its practical application in the field of oil/water separation. Therefore, it is important to develop BC-based materials for “water removing” in oil/water separation to address the issue of oily wastewater. In this study, a super-wetting membrane for \"efficient oil/water separation was fabricated using a feasible and environmentally friendly dip-coating method, based on renewable bamboo cellulose. The resulting membrane exhibited superhydrophilicity and underwater superoleophobicity, making it a promising candidate for the treatment of oily wastewater. The surface hierarchical structures and chemical composition of the super-wetting membrane were analyzed using modern techniques. Moreover, the oil/water separation performance of the super-wetting membrane was thoroughly studied through a series of experiments involving the separation of oil/water mixtures. Additionally, the membrane’s high recyclability was demonstrated through a 10-cycle oil/water separation experiment. Furthermore, the mechanism behind the oil/water separation, which involves a gravity-driven “water-removing” mode facilitated by the BC hydrogel-coated super-wetting membrane, was investigated. This research is significant as it offers a green, simple, cost-effective, and environmentally friendly approach to producing a range of super-wetting surface materials to address the issue of oily wastewater pollution. Moreover, it has the potential to enhance the practical application of BC-based materials with super-wetting in the field of oily wastewater treatment.",
"discussion": "3. Results and Discussion 3.1. Micromorphologies, Chemical Compositions and Wettability of the Super-Wetting Membrane An optical image of the BC hydrogel-coated copper mesh is shown in Figure 2 a. The transparent coating of BC hydrogel on the surface of the super-wetting membrane was clearly observed, which was attached to the surface of the copper mesh. As illustrated in Figure 2 b, an SEM image of the super-wetting membrane, the BC hydrogel coated on the copper wire surface ( Figure S2 ), especially the cross-connection position of the copper mesh, was covered with a large amount of cellulose hydrogel, and the surface roughness of the copper mesh was enhanced. The EDS data contained three elemental components: C, N, and O, indicating that a bamboo cellulose coating had been formed on the surface of the copper mesh ( Figure S3 ). Moreover, the chemical compositions on the surface of the super-wetting membrane were analyzed by XRD and FTIR spectra. As shown in Figure 2 c, the XRD spectrum exhibited raw bamboo cellulose (raw BC) and regenerated bamboo cellulose (regenerated BC). The XRD patterns of the raw BC showed three characteristic peaks at 2 θ = 14.9°, 16.2° and 22.6° corresponding to the (1 1 ¯ 0), (110), and (200) planes of cellulose I crystalline forms, respectively. However, The XRD patterns of regenerated BC show three characteristic peaks at 2 θ = 12.1°, 19.7° and 22.0° corresponding to the (1 1 ¯ 0), (110), and (200) planes. These three crystal planes belong to cellulose II. The presence of XRD peaks characteristic of cellulose II for cellulose dissolved in the LiOH·H 2 O/urea/water system indicates the fact that the cellulose structure was completely transformed from cellulose I to cellulose II after the regeneration process. The formation of the BC hydrogel was realized through a physical process by establishing hydrogen bonding (H-bonding) between the cellulose chains. Similarly, the FTIR spectra of the raw BC and regenerated BC certified that the characteristic peaks of cellulose functional groups have not changed either ( Figure 2 d), demonstrating that the formation of regenerated BC from raw BC is a physical change process. The characteristic peaks at 3400 cm −1 correspond to the O-H stretching vibration, which is the characteristic peak of cellulose. The characteristic peaks at 2870 cm −1 , 1630 cm −1 , 1370 cm −1, and 1060 cm −1 correspond to the C–H stretching vibration, C=O stretching vibration, –C–H stretching vibration and –C–O– stretching vibration, respectively. Furthermore, the surface wettability of the super-wetting membrane coated with the BC hydrogel coating was investigated. The prepared BC hydrogel-coated copper mesh exhibited a special wettability; it is superhydrophilic in the air and superoleophobic in water. An oil droplet (1,2-dichloroethane) remained spherical on the surface of the super-wetting membrane ( Figure S1 ), and the static contact angle of a water droplet (4 µL) in the air is approximately 0° ( Figure 2 e), while the contact angle of an underwater oil (4 µL, hexane) contact angle is 151° ( Figure 2 f). The BC hydrogel coating of the prepared super-wetting membrane not only increased the roughness of the copper mesh substrate surface but the hydrophilic hydroxyl groups contained in the BC hydrogel coating also enhanced the hydrophilicity of the copper mesh surface. Moreover, the super-wetting membrane has excellent underwater superoleophobicity and low adhesion—a 4 µL oil droplet (1,2-dichloroethane) quickly rolled off the surface with a tilting angle of 5.3° within 0.84 s ( Figure 2 g and Movie S3 ), resulting from the presence of micro/nanoscale structures of the hydrogel-coated super-wetting membrane that are filled with water in aquatic environments. 3.2. Performances of Oil/Water Separation The comprehensive performance of oil/water separation using a “water-removing”-type super-wetting membrane (1 wt% BC hydrogel) was systematically investigated. Optical images before and after the separation process are shown in Figure 3 a. In the mixture solution of water and dyed red toluene, the water component was quickly separated by gravity within 3.5 s, while the oil component was blocked on the surface of the super-wetting membrane. This blocking occurred because the BC hydrogel coating on the membrane surface was filled with water, creating a water/oil interface that hindered the penetration of oil components. In Figure 3 b, it can be observed that the maximum height ( h max) of the oil column (hexane, Movie S2 ) supported by the super-wetting membrane is about 19 cm, resulting in an oil intrusion pressure of up to 1.23 kPa. This high intrusion pressure is conducive to efficient oil–water separation under gravity-driven conditions. Furthermore, various oil–water mixtures, including toluene, petroleum ether, benzene, cyclohexane, hexane, paraxylene, paraffin liquid, pump oil, and peanut oil, could be successfully separated using the BC hydrogel coating super-wetting membrane driven by gravity ( Movie S1 ), as shown in Figure 3 c. The oil/water separation efficiencies for these various oils (organic solvents) were all greater than 98%. Specifically, the oil/water separation efficiencies were 99.89%, 99.96%, 99.32%, 98.61%, 98.28%, 98.12%, 98.16%, 98.61%, and 98.53% for toluene, petroleum ether, benzene, cyclohexane, hexane, paraxylene, paraffin liquid, pump oil, and peanut oil, respectively. The results showed that the super-wetting membrane modified with BC hydrogel had excellent oil–water separation efficiency. Furthermore, the water flux of six kinds of mixed solutions during the oil/water separation process was researched using the super-wetting membrane (200 N copper mesh), as shown in Figure 3 d. The water flux was 4326 L/m 2 ·h, 4528 L/m 2 ·h, 7570 L/m 2 ·h, 8485 L/m 2 ·h, 8827 L/m 2 ·h and 9168 L/m 2 ·h for toluene, petroleum ether, benzene, paraxylene, cyclohexane and hexane. The results demonstrated that the super-wetting membrane modified with BC hydrogel could achieve rapid oil/water separation. Moreover, the oil/water separation efficiency of the super-wetting membrane coated with 1 wt% BC hydrogel on a copper mesh substrate with different mesh numbers has been thoroughly explored. The results indicated that the mesh number of the copper mesh substrate had little effect on the oil/water separation efficiency of the super-wetting membrane, as shown in Figure 3 e. However, as the number of copper meshes increased, the water flux decreased, which was due to the increase in mass transfer resistance caused by the decrease in the pore size of the copper mesh. Furthermore, the stability performance of the super-wetting membranes was studied by separating a mixed solution (hexane and water). The super-wetting membranes still maintained a high oil–water separation efficiency of up to 95% after 10 cycles, indicating that the BC hydrogel-coated super-wetting membrane had outstanding stability during oil/water separation. 3.3. Oil/Water Separation Mechanism of the Super-Wetting Membrane The oil/water separation mechanism of the BC hydrogel-coated super-wetting membrane has been studied, and the liquid-wetting processes are illustrated in Figure 4 . The intrusion pressure ( Δ P ) of the super-wetting membrane could be described by Equation (4): [ 61 ]\n (4) Δ P = 2 γ R = − l γ cos θ A \nwhere γ is the interfacial tension, R is the meniscus’s radius, l is the mesh pore’s perimeter, A is the pore’s area, and θ is the contact angle on the film. According to Equation (4), the membrane surface was subjected to additional pressure (Δ P > 0) as the contact angle was θ > 90°. In contrast, the contact angle was θ < 90°, resulting in the Δ P < 0 and the liquid passing through the super-wetting membrane. Likewise, in the aquatic environment, the micro-/nanoscale structures and the BC coatings of the super-wetting membrane were filled with water, resulting in the formation of a layer of water film on the surface of the super-wetting membrane and strengthening the repulsive force between polar (water) and non-polar (oil) molecules simultaneously. Therefore, the water component was able to pass through the super-wetting membrane, while the water contacted the BC hydrogel coating, owing to the contact angle θ being nearly 0° and Δ P < 0 ( Figure 4 a). On the contrary, the oil component was blocked on the super-wetting membrane as the oil contacted the BC hydrogel coating, because of the presence of the water layer, oil was prevented from permeating the super-wetting membrane surface, and θ is obviously larger than 90° and Δ P > 0 ( Figure 4 b). According to the above analysis, the prepared super-wetting membrane may be able to achieve efficient oil–water separation."
} | 4,582 |
36743004 | PMC9893744 | pmc | 3,599 | {
"abstract": "As a new energy harvesting\ntechnology, triboelectric nanogenerators\nare widely used for vibration mechanical energy harvesting. However,\nthe current schemes ignore the composite characteristics of vibration,\nwith problems such as utilization and low collection efficiency. In\nthis paper, a random resonance cantilever beam triboelectric nanogenerator\n(RCB-TENG) with dual-mode coupled is presented, the working mode is\na coupling form of in-plane sliding and vertical contact-separation\nthat can effectively collect complex vibration energy in transverse\nand longitudinal directions. The influences of the structural parameters\nof the RCB-TENG and different dielectric materials on the output performance\nare systematically investigated. The single vibration module achieved\na power density of 463.56 mW/m 2 and a transfer charge of\n10.7 μC at a vibration frequency of 46 Hz, an increase in power\ndensity, and a transfer charge of 4.94 and 3.82 times, respectively,\ncompared to the conventional contact-separation mode. Finally, the\nRCB-TENG was tested in practice, and it was observed that nine 1 W\ncommercial LED bulbs and 500 5 mm diameter LED lamps were successfully\nlit. This work offers new ideas for distributed energy harvesting\ntechnologies and holds broad promise in the field of energy harvesting\nfrom wind, water, wave, and random vibrations caused by mechanical\nenergy.",
"conclusion": "3 Conclusions In summary, the RCB-TENG\ncoupled\nwith two types of working modes\nis proposed for harvesting random vibration energy. Compared with\nthe traditional CS-Mode, the C-Mode has a higher electrical output\nperformance. The RCB-TENG has nine vibration modules, and the single\nvibration module achieved a power density of 463.56 mW/m 2 and a transfer charge of 10.7 μC at a vibration frequency\nof 46 Hz. Compared with the conventional CS-Mode, the power density\nand transfer charge have, respectively, increased by 4.94 and 3.82\ntimes. Moreover, the short-circuit current and transfer charge of\nthe RCB-TENG can be up to 454 μA and 46 μC, respectively,\nunder conventional triggering conditions. Finally, the RCB-TENG’s\narray of vibration modules was used in and lit commercial lamps without\nan external power supply, demonstrating its potential application\nfor vibration energy harvesting and remote environmental information\nmonitoring.",
"introduction": "1 Introduction As the fourth information\nrevolution—the Internet of Things—continues\nto develop, a significant number of IP addresses, sensors, and wearable\ndevices have emerged in everyday life. Currently, batteries are the\ncommon powering choices for widely distributed sensors. However, the\nlimited energy storage capacity of conventional batteries is insufficient\nfor the continuous operation of sensors. In addition, the disposal\nof used batteries places a huge strain on the ecology. Energy harvesting\ntechnology utilizes small amounts of energy in the environment to\nprovide a continuous power supply for device power systems. Thus,\nthe investment and time spent on battery systems are saved, and a\nnew way of thinking about powering large-scale sensors is provided.\nThe triboelectric nanogenerators (TENGs), a new energy technology\nbased on frictional initiation and electrostatic induction, 1 − 4 which can convert the energy in the environment into usable electric\nenergy, 5 − 8 have received attention in energy harvesting technologies due to\ntheir small size and potential to convert a wide range of environmental\nenergy sources (e.g., wind energy, 9 − 11 vibration energy, 12 − 14 and water wave energy 15 − 21 ) into electrical output. Underpinned by Maxwell’s theory\nof displacement currents, TENGs can be classified depending on their\nmode of working into contact-separation mode, 22 in-plane sliding mode, 23 single-electrode\nmode, 24 and freestanding triboelectric-layer\nmode. 25 Due to their unique working mechanism,\nTENGs have been proven to be a simple, reliable, and cost-effective\nmeans of high-performance environmental energy harvesting. 26 , 27 As a widely distributed, renewable, and clean energy source,\nvibration\nhas always been an attractive and hot target for energy harvesting\nprocesses. TENGs have been shown to harvest vibration energy from\nthe environment, such as wearable sensors and self-powered acoustic\nsensors 28 , 29 where contact-separation mode is currently\nthe main mode of TENG working for collecting the vibration energy. 30 − 32 When collecting vibration energy, the interaction between vibration\ndistance and frequency results in a contact-separation mode with limited\nelectrical output performance. Even though higher output performance\ncan be achieved using the in-plane sliding mode, 33 the electrical stability and mechanical durability of TENGs\ncan be seriously compromised by frictional heat and material wear.\nIt is also difficult to design structures that simultaneously generate\nrelatively large displacements and high velocities. Therefore, it\nis not easy to efficiently collect vibration energy in the in-plane\nsliding mode. In general, single-direction vibrations (transverse\nor longitudinal)\nare caused by the inherent properties of the vibrating body itself.\nOn the other hand, vibrations in the orthogonal direction are caused\nby external factors. 31 Hence, most vibrations\ncan be characterized as multidirectional complex motions in which\ntransverse and longitudinal vibrations are coupled. Furthermore, the\nexisting vibration energy harvesters rarely consider composite characteristics\nof vibration and only employ a single working mechanism for harvesting\nin the TENG, with problems such as insufficient energy utilization.\nTherefore, the design of an energy harvesting device that can effectively\nharvest vibration energy is of great importance to expand practical\napplications of TENG. In this work, contact-separation and in-plane\nsliding modes are\ncoupled in the form of vibration energy harvesting, solving the problems\nassociated with the underutilization of the TENG in vibration energy\nharvesting. First, a prototype random resonance cantilever beam triboelectric\nnanogenerator (RCB-TENG) is designed and tested, the coupling mode\nallows the collection of lateral and longitudinal composite vibrations.\nConsequently, a higher electrical output can be generated while efficiently\ncollecting energy compared to the contact-separation mode. Second,\nthe effect of system parameters on the output performance is investigated\nin depth. A single vibration module achieved a power density of 463.56\nmW/m 2 and 10.7 μC at a 3 mm separation gap, 30 mm\nsliding distance, and 46 Hz vibration frequency, which are 4.94 and\n3.82 times higher than that in the conventional contact-separation\nmode, respectively. Lastly, the power management strategies for collecting\nvibration energy are also established for practical applications.\nIt is demonstrated that RCB-TENG can be used to power commercial electronics,\nwhich provides a new idea for vibration energy harvesting.",
"discussion": "2 Results and Discussion 2.1 Structure Design and Working\nPrinciple of\nthe RCB-TENG Most conventional vibration energy harvesting\nTENGs are focused on harvesting the vibrations of single mode, such\nas wind energy harvesters. 34 − 36 In such harvesters, the harvesting\nmechanism is approximately homogeneous, limiting the efficiency of\nvibration energy harvesting. In this paper, the vibration is decomposed\ninto transverse and longitudinal in-plane motions to address the issues\nmentioned above. Meanwhile, a RCB-TENG is designed, which uses a rigid\nspring plate as the vibration source and coupling mode that can significantly\nimprove the energy utilization efficiency within a single vibration\ncycle. A 3D structure of the RCB-TENG is shown in Figure 1 a. The RCB-TENG consists of\na stator, a rotor, and a vibration module. The rotor is driven by\na servo motor, and the stator is a rigid platform used to fix the\ncantilever beam and the servo motor. The specific compositions of\nthe RCB-TENG and the vibration module are shown in Figure 1 a,b, respectively. The main\nvibration module body consists of a cantilever beam, electrodes, dielectric\nmaterial, vibration housing, and a vibration source. The dielectric\nmaterial consists of a polytetrafluoroethylene (PTFE) film, paper,\nand copper that serves as the electrode. The copper electrode is bonded\nto the PTFE film and paper via the conductive adhesive to form two\ncomposite films (the first one is the copper and PTFE composite film,\nand the second is the paper and copper electrode composite film).\nThe two laminates are then bonded to the surfaces of the vibration\nhousing and the vibration source via a double-sided adhesive. The\ncopper electrode and PTFE laminate are connected to the outer surface\nof the comb finger vibration source unit, whereas the copper electrode\nand paper laminate are connected to the inner surface of the vibration\nhousing (Figure S1, Supporting Information ). The cantilever beam is a 1 mm spring steel sheet, and a damping\nblock with micro bearing on the outer edge is hung in the middle of\nthe cantilever beam to reduce friction and torque. The rotor converts\nits rotating motion into high-frequency vibration of the cantilever\nbeam by pulling the damping block with the bearing. The vibration\nmodules are evenly distributed along the circumference of the stator\n( Figure 1 c). Figure 1 Structural\ndesign of a RCB-TENG: (a) schematic diagram of the RCB-TENG\napplication scenario and structure, (b) structural design of the RCB-TENG\nvibration module, (c) assembled RCB-TENG (scale bar: 11 cm), (d) exploded\nview of the vibration module structure, (e) simplified working mechanism,\n(f) comparison of the coupling mode (C-Mode) and the contact-separation\nmode (CS-Mode) electrical output, (g) comparison of the total output\ncharge of the C-Mode with the CS-Mode, and (h) comparison of the average\npower density of the C-Mode and the CS-Mode. The RCB-TENG’s working mode is designed\nas spatially resonant\nto better employ the vibration energy, as shown in Figures 1 e and S2 in Supporting Information . The spatially resonant mode can be\nbroken down into a contact-separation mode (vertical direction) and\nan in-plane sliding mode (horizontal direction). Consequently, the\nRCB-TENG can collect complex vibrations in space, which significantly\nincreases the efficiency of energy utilization. A higher electrical\noutput performance than the traditional contact-separation mode can\nbe achieved due to the combined contact-separation mode and the in-plane\nsliding mode. As shown in Figure 1 f, a vibration module can achieve a peak short-circuit\ncurrent of 112 μA in a coupling mode at a vibration frequency\nof 46 Hz, whereas the traditional contact-separation mode can only\nachieve 37 μA under the same conditions (Figure S6, Supporting Information ). In addition, the electrical\noutput performance of different power generation modes under the same\ntest conditions has been tested. The electrical output performance\nof the single contact-separation mode is shown in Figure S6 , the single in-plane sliding mode is shown in Figure S12 , and the electrical output performance\nof the composite mode is much higher than that of the single mode\n(Figure S5, Supporting Information ). More\nimportantly, compared with the traditional contact-separation mode,\nthe coupling mode greatly improves the total transferred charge amount\nand average power density. The power density is 4.94 times that of\nthe traditional contact-separation mode, and the total transferred\ncharge is 3.82 times that of the traditional contact-separation mode\n( Figure 1 f–h). The working principle of the RCB-TENG is shown in Figure 2 . When the cantilever beam\ndamping block is continuously excited, it produces a high-frequency\nleft–right oscillation with the axis as the fixed point ( Figure 2 b), at which point\nthe RCB-TENG operates as an in-plane sliding mode in the horizontal\ndirection and as a C-Mode in the vertical direction (Note S2, Supporting Information ). For simplicity, when\nexplaining the generator theory, the RCB-TENG module motion is simplified\nto two motion states: in-plane sliding and vertical contact-separation. Figure 2 Working\nprinciple diagram of the RCB-TENG: (a) installation position\nof the RCB-TENG with the vibration module, (b) schematic diagram of\nthe vibration triggering the vibration module, (c) working principle\nof the in-plane sliding mode (PS-Mode), (d) working principle of the\nCS-Mode, (e) potential distribution in the PS-Mode, and (f) potential\ndistribution in the CS-Mode. When the cantilever beam damping block is excited,\nrelative motion\nbetween the vibration module and its source is simplified to plane\nsliding. Moreover, the sliding of the vibration source in the vibration\nshell is divided into four steps. During the initial state [ Figure 2 c(i)], the PTFE laminate\non one side overlaps the paper laminate on the other, meanwhile, positive\nand negative charges on the PTFE and paper surfaces are in equilibrium.\nAs the PTFE moves relative to the paper, the original electrostatic\nequilibrium is broken and gradually shifts to a second state [ Figure 2 c(ii)]. Due to the\ndifference in electronegativity, a potential difference is generated\nbetween the paper composite film and the PTFE composite film. Therefore,\ndriven by the potential difference, the positive charge on the copper\nelectrode in the vibration source PTFE composite film is gradually\ntransferred to the copper electrode of the paper composite film of\nthe vibration housing, and transient current is formed on the external\ncircuit, generating an electric signal. When the separation\ndistance between the two composite membranes\nis maximal [ Figure 2 c(iii)], the current in the external circuit is also the highest.\nAs the vibration source continues to move [ Figure 2 c(iv)], the positive charge on the copper\nelectrode of the vibration housing paper composite membrane flows\nback toward the copper electrode of the vibration source PTFE. This,\nin turn, generates a reverse current in the external circuit. As the\nPTFE composite film continues to move and reunites with the paper\ncomposite film, the positive charge on the paper composite film electrode\nis transferred back to the PTFE composite film electrode, and the\nelectrostatic equilibrium is restored. As the PTFE periodically moves\nbetween two vibration housings, a periodic output signal is generated. When subjected to continuous excitation, the cantilever beam oscillates\naround the center of the lower axis due to the gap between the vibration\nshell and the vibration source. Hence, the vibration source will slide\nand contact-separation periodically with the vibration shell will\nbe driven by the cantilever beam. The contact-separation process is\nshown in Figure 2 d.\nDuring the first stage [state I, Figure 2 d(i)], the paper on the surface of the comb\nfinger unit on the inside of the vibration housing fully contacts\nthe PTFE membrane of the vibration source. Positive charges accumulate\non the paper surface due to the frictional charging effect, while\nnegative charges are transferred to the PTFE surface. As the\nvibration housing moves upward under friction, the paper\nseparates from the PTFE membrane, and a potential difference is established\nbetween the two surfaces due to the difference in electronegativity.\nThis drives the transfer of electrons from the copper electrode on\nthe back of the PTFE composite membrane to the copper electrode of\nthe paper composite membrane and generates a transient current in\nthe external circuit [ Figure 2 d(ii)]. As the vibration housing continues to move upward,\nthe separation distance between the paper and the PTFE surface continues\nto increase. When the distance between the paper and the PTFE surface\nis maximal, the amount of charge transfer in the external circuit\nalso reaches a maximum value [ Figure 2 d(iii)]. It should be noted that the distance\nbetween the vibration source\nPTFE and the vibration shell paper decreases as the vibration shell\ncontinues to move upward [ Figure 2 d(iv)]. At this time, the electric potential between\nthe two composite membranes gradually disappears, and the current\nand the voltage in the external circuit gradually decrease from the\nmaximum value until the secondary contact is made ( Figure 2 d(v)). Then, the charge is\ncompletely neutralized, and the external circuit voltage and current\nare nullified. As the vibration housing begins to fall under gravity,\nthe distance between the paper and the PTFE surface increases, and\na potential difference is re-established. At a certain position, the\nexternal circuit current and voltage return to their maximum values,\nthus forming a complete external circuit signal. To illustrate\nthe working mechanism more clearly, representative\nstates in two independent modes are chosen to simulate the corresponding\npotential distribution under open circuit conditions via the finite\nelement method. The results are shown in Figure 2 e,f. Under ideal conditions, the potential\nincreases with the separation distance between the dielectric material. 2.2 Structural Optimization A comb-finger\nRCB-TENG is designed in this paper. In this section, the selection\nand optimization of the structural parameters of the vibration module\nin the RCB-TENG are clarified. Paper is uniformly used as a positive\nmaterial in the vibration module. The detailed manufacturing process\nof these devices is described in the experimental section. Within\nthe experiment, the rotor, a 3D-printed non-standard cam, is driven\nby a servo motor. The cam strikes a damping block with a miniature\nbearing on the cantilever beam, causing the beam to vibrate at frequencies\nbetween 5 and 46 Hz (Note, Figures S2 and S3, Supporting Information ). In addition, the sliding distance\nand vertical separation gap between the vibration source and the vibration\nmodule are set within the range of 10–50 and 1–5 mm,\nand the cantilever beam is selected from 0.5 to 1 mm spring steel\nsheets. The influence of system parameters (horizontal sliding\ndistance, vertical separation gap, and cantilever beam thickness)\nand the dielectric material on the output performance of the vibration\nmodule is investigated in depth to optimize the output performance\nof the vibration module. As shown in Figure 3 a,b, the effect of different sliding distances\nand vertical separation gaps on the output performance is first tested\nat 38 Hz. Here, a negative material of PTFE and a constant steel thickness\nof 0.8 mm are employed. The results show that the output performance\ntends to increase first and then decrease with an increase in the\nsliding distance and separation gap. The electrical output performance\nreaches maximum values for a short-circuit current and open-circuit\nvoltage at a sliding distance of 30 mm and a separation gap of 3 mm,\nrespectively. Too large or too small a sliding distance and separation\ngap will affect its output performance. If the sliding distance and\nseparation gap are too small, the friction resistance between dielectric\nmaterials will increase and the actual separation distance will decrease.\nOn the contrary, too large sliding distance and separation gap will\nlead to insufficient material contact and lower output performance.\nSubsequently, a study was carried out for different negative materials\n( Figure 3 c) with other\nconditions being constant (sliding distance of 30 mm, separation gap\nof 3 mm, vibration frequency of 38 Hz, and steel thickness of 0.8\nmm). Experiments were carried out using 0.1 mm thick Kapton and polydimethylsiloxane\nand PTFE films, which were attached to the electrodes and tested for\nopen-circuit voltage and short-circuit current. The highest electrical\noutput performance is achieved when PTFE is used for the negative\ndielectric layer. The corresponding peak open-circuit voltage and\nshort-circuit current were 248.53 V and 104.9 μA for a single\nvibration module, respectively. Finally, the corresponding performance\nof spring steel sheets with different thicknesses was tested at various\nvibration frequencies when the PTFE was employed as a negative dielectric\nlayer ( Figure 3 d–f).\nThe results show that the open-circuit voltage of the vibration module\nis insensitive to the vibration frequency. However, the short-circuit\ncurrent and transferred charge increase significantly with the thickness\nand vibration frequency. This is determined by the inherent characteristics\nof the cantilever beam system. The thicker the cantilever beam and\nthe higher the vibration frequency, the more signal peaks per unit\ntime and the greater the electrical output performance. Figure 3 Output performance\nof a vibration module: (a) different sliding\ndistances, (b) different separation gaps, (c) different negative materials,\n(d) open-circuit voltage, (e) short-circuit current, and (f) transfer\ncharge for different thicknesses of spring steel plates at various\nfrequencies. 2.3 Output\nCharacteristics System parameters\nfor the individual vibration modules in the RCB-TENG are determined\nbased on the conducted optimization of the vibration module structure.\nNext, the experiments were conducted for a sliding distance of 30\nmm, a vertical separation gap of 3 mm, and dielectric materials of\npaper and PTFE. The performance of the RCB-TENG was tested at 38 Hz\nwith 1 mm steel sheets for different vibration modules ( Figure 4 a–c). Figure 4 utput performance of\nthe RCB-TENG: (a) open-circuit voltage for\ndifferent numbers of vibration modules (V-M), (b) short-circuit current,\n(c) transfer charge, (d) open-circuit voltage, (e) short-circuit current,\n(f) transfer charge at different vibration frequencies, (g) root mean\nsquare value of short-circuit current for different numbers of vibration\nmodules, (h) short-circuit current, open-circuit voltage, and peak\npower, and (i) peak power of the RCB-TENG at different vibration frequencies. The current values of the nine vibration modules\nare significantly\nhigher than the amplitude of a single vibration module, with a maximum\nshort-circuit current close to 410 μA ( Figure 4 b). A single vibration module transfers almost\n10 μC. As the number of vibration modules increases, nine vibration\nmodules transfer a total of 40 μC ( Figure 4 c). However, for different numbers of vibration\nmodules, the open-circuit voltage shows minor variation due to the\nparallel connection between the vibration modules. Theoretically,\nit can be assumed that the open-circuit voltage and the total open-circuit\nvoltage are the same in the parallel circuit. Therefore, the open-circuit\nvoltage value under nine vibration modules is almost the same as the\nopen-circuit voltage for a single vibration module, that is, it is\nstabilized at approximately 250 V ( Figure 4 a). In addition, the output performance\nof the RCB-TENG was measured\nunder nine vibration module conditions and at different vibration\nfrequencies. According to Figure 4 d, as the vibration frequency increases, the open-circuit\nvoltage of the RCB-TENG first increases and then stabilizes. When\nthe vibration frequency increases from 5 to 29 Hz, the maximum value\nof the open-circuit voltage increases from 200 to about 250 V. When\nthe vibration frequency increases to 46 Hz, the voltage value remains\nalmost constant. This can be attributed to the low vibration frequency\nof the cantilever beam. Because the dielectric material between the\ncontact is insufficient, the output voltage is low. As the vibration\nfrequency increases, the contact area becomes sufficient and is no\nlonger characterized by a relatively large change. Lastly, the voltage\nvalue also gradually stabilizes. As the vibration frequency increases,\nthe short-circuit current gradually increases and can reach up to\n454 μA (Figure S7, Supporting Information ), while the total transferred charge can reach 46 μC ( Figure 4 d–f). The root mean square (rms) values and output power of this RCB-TENG\nwere also tested ( Figure 4 g–i). Figure 4 g depicts the rms current values at 38 Hz for different numbers\nof vibration modules. The rms output increases with the number of\nvibration modules, that is, up to 153.42 μA for nine vibration\nmodules. The RCB-TENG based on paper and PTFE dielectric material\ncomposition has good output performance. The power diagram for nine\nvibration modules and a frequency of 38 Hz is shown in Figure 4 h. It can be seen that the\nmaximum power can reach 22.617 mW at 500 kΩ. Figure 4 i depicts a graph of the peak\npower of the RCB-TENG for nine vibration modules at different vibration\nfrequencies. When the vibration frequency increases from 5 to 46 Hz,\nthe peak power of the RCB-TENG increases from 15.58 to 31.892 mW,\nwhile the matching resistance decreases from 3 MΩ to 500 kΩ\n(the reason for the matching resistance decrease can be found in Note\nS3, Supporting Information ). The\ndurability and relative humidity tests for the RCB-TENG have\nbeen systematically carried out. As shown in Figure S9 , the output performance of the RCB-TENG decreases slightly\nwith the continuous increase of relative humidity when the measured\naverage humidity changes from 27 to 83% (Note S6, Supporting Information ). In addition, the durability test\nof the RCB-TENG demonstrates that the output performance of the RCB-TENG\nis almost stable after running about 1260000 cycles. The electrical\noutput characteristic curve is shown in Figure S11 (Note S5, Supporting Information ). 2.4 Application\nof the RCB-TENG for Collecting\nthe Vibration Energy To further improve the performance of\nthe RCB-TENG, its current is improved using a circuit management system\nthat converts AC to DC for direct drive of electronic components.\nHere, the half-wave rectifier circuit is chosen for the power management\ncircuit. The half-wave rectifier circuit has a higher output performance\nthan the full-wave rectifier circuit in the CS and PS modes. 37 Moreover, it consists mainly of Schottky diodes,\nas shown in Figure 5 a. The DC output is obtained according to the forward bias of the\nSchottky diodes, that is, the unidirectional conductivity. The transferred\ncharge of the DC output is shown in Figure 5 c. Here, the performance and application\nof the RCB-TENG to two different modes of operation are investigated\n( Figure 5 b). The first\nmode of operation is a direct connection to a capacitor, intending\nto test the charging performance of the RCB-TENG with nine vibration\nmodules and a vibration frequency of 38 Hz ( Figure 5 b(i)). The RCB-TENG can charge a commercial\ncapacitor to over 10 V in a short time with the support of the power\nmanagement circuit ( Figure 5 d). Similarly, the RCB-TENG can charge a 10 μF capacitor\nto approximately 40 V in 20 s. Figure 5 RCB-TENG in vibration energy harvesting:\n(a) circuit diagram of\nRCB-TENG power management, (b) two typical operation modes of RCB-TENG-based\nvibration energy harvesting systems, (c) charging curve of RCB-TENG\nwith circuit management for charging four different sizes of capacitors,\n(d) RCB-TENG directly driving commercial LED bulbs (scale bar: 8 cm),\n(e) RCB-TENG brightly drives the NEFU logo (scale bar: 5 cm), and\n(f) RCB-TENG directly lights up 500 5 mm LEDs (scale bar: 8 cm). The second mode of RCB-TENG operation is directly\nconnected to\nthe load [ Figure 5 b(ii)],\nthat is, the electronic components are powered directly by the RCB-TENG\n( Figure 5 d–f).\nThe RCB-TENG can directly drive nine 1 W commercial LED bulbs ( Figure 5 d), light up to 5\nmm LED series parallel hybrid NEFU logo with a bright effect ( Figure 5 e), and directly\ndrive around 500 small 5 mm LED bulbs ( Figure 5 f). The experiments are recorded in Movies S1 , S2 , and S3 (Supporting Information). Similarly, the RCB-TENG\ncan harvest wind, water, wave, and vibration energy to power the sensors,\ncontinuously monitor the weather or vibrations generated by the machine,\nand provide alerts in case of bad weather or mechanical failure."
} | 6,988 |
20637598 | null | s2 | 3,600 | {
"abstract": "Life is a dynamic process driven by the complex interplay between physical constraints and selection pressures, ranging from nutrient limitation to inhibitory substances to predators. These stressors are not mutually exclusive; microbes have faced concurrent challenges for eons. Genome-enabled systems biology approaches are adapting economic and ecological concepts like tradeoff curves and strategic resource allocation theory to analyze metabolic adaptations to simultaneous stressors. These methodologies can accurately describe and predict metabolic adaptations to concurrent stresses by considering the tradeoff between investment of limiting resources into enzymatic machinery and the resulting cellular function. The approaches represent promising links between computational biology and well-established economic and ecological methodologies for analyzing the interplay between physical constraints and microbial fitness."
} | 232 |
21829719 | PMC3150396 | pmc | 3,602 | {
"abstract": "Recovery of an ecosystem following disturbance can be severely hampered or even shift altogether when a point disturbance exceeds a certain spatial threshold. Such scale-dependent dynamics may be caused by preemptive competition, but may also result from diminished self-facilitation due to weakened ecosystem engineering. Moreover, disturbance can facilitate colonization by engineering species that alter abiotic conditions in ways that exacerbate stress on the original species. Consequently, establishment of such counteracting engineers might reduce the spatial threshold for the disturbance, by effectively slowing recovery and increasing the risk for ecosystem shifts to alternative states. We tested these predictions in an intertidal mudflat characterized by a two-state mosaic of hummocks (humps exposed during low tide) dominated by the sediment-stabilizing seagrass Zostera noltii ) and hollows (low-tide waterlogged depressions dominated by the bioturbating lugworm Arenicola marina ). In contrast to expectations, seagrass recolonized both natural and experimental clearings via lateral expansion and seemed unaffected by both clearing size and lugworm addition. Near the end of the growth season, however, an additional disturbance (most likely waterfowl grazing and/or strong hydrodynamics) selectively impacted recolonizing seagrass in the largest (1 m 2 ) clearings (regardless of lugworm addition), and in those medium (0.25 m 2 ) clearings where lugworms had been added nearly five months earlier. Further analyses showed that the risk for the disturbance increased with hollow size, with a threshold of 0.24 m 2 . Hollows of that size were caused by seagrass removal alone in the largest clearings, and by a weaker seagrass removal effect exacerbated by lugworm bioturbation in the medium clearings. Consequently, a sufficiently large disturbance increased the vulnerability of recolonizing seagrass to additional disturbance by weakening seagrass engineering effects (sediment stabilization). Meanwhile, the counteracting ecosystem engineering (lugworm bioturbation) reduced that threshold size. Therefore, scale-dependent interactions between habitat-mediated facilitation, competition and disturbance seem to maintain the spatial two-state mosaic in this ecosystem.",
"introduction": "Introduction One of the most studied but also debated issues in ecology is the relative importance of factors affecting how organisms and ecosystems respond to disturbance [1] , [2] , [3] , [4] . One factor which may have a fundamental impact is the size of point disturbances; following a disturbance that exceeds a threshold size, local processes often change, recovery slows down, and communities may even develop into alternative stable states [5] . Such scale-dependent responses have typically been explained by weakened competitive exclusion from surrounding individuals, which increases the chance that previously inferior competitors can recruit into and dominate the center of disturbed areas [5] , [6] , [7] . For instance, ice-scour in hard bottom rockweed communities can trigger shifts to domination by mussels or fucoid macroalgae, if the disturbance is so large that the “whiplash” from surrounding rockweed cannot exclude competitors [8] . Importantly, such shifts are mediated by large disturbances, but ultimately depend on competition and space preemption. Theory and observation suggests scale-dependent ecosystem shifts can also be caused by increased abiotic stress on recolonizing individuals, when the removal of an “ecosystem engineering” species simultaneously removes the self-facilitation required for recovery. Such disturbance responses should be most common in abiotically stressed environments, where communities depend on facilitation from ecosystem engineers. These are species that via presence and/or function change abiotic conditions, concentrate resources and/or alleviate local stress, which induces positive organism-environment feedback [9] , [10] , [11] . For instance, attenuation of waves and currents by submerged vegetation creates sheltered and calm microenvironments necessary for their own recruitment [12] . Importantly, such effects typically exceed engineers both spatially and temporally, i.e. extends beyond engineer patch edges and outlive engineers [13] . As a consequence, engineers may facilitate recolonization of conspecifics into areas impacted by small disturbances; a phenomenon known as short-range facilitation [14] , [15] , [16] . If, however, the disturbed area spatially exceeds the range of the facilitation, increased abiotic stress will prevent recovery; so called long-range inhibition [17] . One example is increased hydrodynamic stress within large gaps in canopy-forming submerged vegetation [18] , [19] . Due to lack of recovery, the impacted area can in theory be colonized by other species that tolerate or even benefit from the altered abiotic conditions. However, the ecosystem shift (from a state with the engineer to a state without the engineer) will occur even if competitors do not colonize, because increased stress is the factor preventing recovery. Neither of these mechanism takes into account that those inferior competitors that may recruit into a disturbed area are often ecosystem engineers that modify abiotic conditions in other ways than the original community. Establishment of such engineers may actually counteract the spatial facilitation from un-impacted individuals in surrounding areas, and increase stress on recolonizing members of the original community. There are many examples of how contrasting engineering effects speeds up competitive exclusion of “counteracting” engineers via “habitat-mediated competition”, which contributes to the formation of spatial mosaics of engineered patches [11] , [20] . For example, the “trophic group amensalism” hypothesis suggests that bioturbators locally suppress suspension-feeders by smothering them [21] , and the “biomechanical warfare” hypothesis that antagonistic, counteracting engineers – e.g. sediment-destabilizing bioturbators and sediment-stabilizing plants – locally exclude each other partly by influencing abiotic conditions in ways that benefits conspecifics but impacts antagonists [22] . Because engineering effects (including habitat-mediated competition) exceed engineers spatially and temporally, we suggest establishment of counteracting engineers may in fact reduce the threshold disturbance size associated with increased risk for changes in recovery and shifts to alternative states increases. This is because counteracting engineering should exacerbates stress on recolonizing individuals, and thereby slow down recovery. We tested these predictions using a field survey and a removal-addition experiment in the eastern Dutch Wadden Sea (Netherlands). The study area was a near-shore intertidal mud flat characterized by a striking mosaic of two alternating habitat types ( Fig. 1A ): low tide exposed humps (from here on hummocks ) dominated by the sediment-stabilizing seagrass Zostera noltii Hornem (from here on seagrass ), alternating with low tide water-logged pools or depressions ( hollows ) dominated by the sediment-destabilizing lugworm Arenicola marina L. (from here on lugworms ). Seagrasses can facilitate themselves by stabilizing sediments and reducing erosion and turbidity through attenuation of water flow [12] . In contrast, lugworms are bioturbators that de-stabilize sediments through burrowing and feeding, which in combination with high hydrodynamic activity increases erosion of fine particles [23] . This typically leads to self-facilitation since lugworms prefer sandy substrates [11] , but also to competitive exclusion of marine plants which prefer more muddy sediments [22] , [24] . 10.1371/journal.pone.0023229.g001 Figure 1 Field survey at the study sites. A) Photograph of site East in June, showing seagrass ( Zostera noltii L.) patches growing on hummocks next to unvegetated hollows (Photo: Johan S Eklöf), (B) Seagrass ground cover (%) on hummocks and in hollows (n = 40 [20 per site]), (C) average gap width (in m, n = 6 [3 per site]), and (D) lugworm ( Arenicola marina L.) fecal cast density (0.25 m −2 ) on hummocks and in hollows (n = 40 [20 per site]), in June, August and October. Means ± 1S.E. After experimentally clearing seagrass and/or adding lugworms from areas ranging in size from 0 to 1 m 2 , we predicted that (i) the strength of seagrass ecosystem engineering and rate of recovery decreases with increasing size of disturbance, (ii) lugworm establishment success increases with size of disturbance on seagrass due to weakened spatial influence from surrounding seagrass, and (iii) lugworm engineering in disturbed areas reduces the threshold disturbance size needed to slow down seagrass recovery by exacerbating erosion caused by the removal of seagrass.",
"discussion": "Discussion Spatial mosaics of patches dominated by ecosystem engineers are common in ecosystems in general, and in inter- and sub-tidal marine systems in particular [22] , [37] , [38] , [39] , [40] , [41] . Such spatial patterning may be explained by local self-facilitation or by simple space preemption [24] , but also by habitat-mediated competition where competitors exclude each other by exacerbating local abiotic stress on their competitors [17] . Here, we show that habitat-mediated self-facilitation and habitat-mediated competition may interactively result in scale-dependent ecosystem responses to disturbance. In the largest seagrass clearings (1 m 2 ), local environmental conditions shifted from sediment stabilizing to de-stabilizing over the course of the experiment regardless of lugworm addition, most likely because of weakened (i) water flow attenuation and (ii) sediment stabilization (the latter caused by a slower relative recovery rate). This essentially caused the formerly seagrass-dominated low-tide hummocks to switch into low-tide hollows during the autumn. However, this scale-dependent response was not due to simple slowing down of recovery due to lack of facilitation [e.g. 42] or space preemption by competitors [e.g. 8] , but greatly increased risk for additional disturbance on recolonizing seagrass. Likewise, lugworm addition did not slow down seagrass recolonization, but lugworm surface sediment destabilization increased the size of hollows surrounding gaps. This was, in turn, the sole factor influencing the risk that recovering seagrass were disturbed again during autumn. As a consequence, the counteracting lugworm engineering decreased the threshold experimental clearing size associated with the additional disturbance from 1 to 0.25 m 2 . Consequently, counteracting engineering reduced the threshold size of disturbance necessary to cause shifts in ecosystem as predicted, but via a mechanism that was largely unexpected. The scale-dependent seagrass recolonization was – in contrast to our expectations – not caused by slower recolonization, but by increased risk for disturbance on recolonizing seagrasses during autumn. Since the study was not designed to identify the additional disturbance or the underlying cause(s) to it, we can only speculate on which was the disturbance. Visual observations in the field strongly suggested that waterfowl grazing – a well-known seasonal disturbance on Z. noltii in the Wadden Sea [26] , [43] , [44] , [45] – was the most likely candidate. Flocks of 150–200 Brent geese ( Branta bernicla ) and Wigeon ( Anas penelope ) were observed feeding daily on Zostera noltii from mid September to November (van der Heide et al., in preparation). Importantly, the grazing was spatially restricted to seagrasses in hollows, while seagrass-dominated hummocks covered 50% of the area (van der Heide et al. in preparation). Such hollow-specific waterfowl grazing has been observed elsewhere, and is explained by the low-tide standing water in hollows, which facilitates feeding while reducing ingestion of silt [46] . Moreover, geese typically cease feeding if food presence and/or accessibility is too low (“giving-up density”), which may happen when seagrass still remain [47] . Combined with our results, this suggests the disturbance on the recolonizing seagrass was waterfowl grazing in hollows of a size that harbored enough accessible food. Similar spatial thresholds in grazing have been demonstrated in terrestrial grasslands, and can sustain repeatedly grazed patches of short vegetation next to patches of high, ungrazed vegetation [48] . Until these suspected effects of waterfowl grazing have been thoroughly tested with experiments [see e.g. 43] , we do not exclude the possibility that other disturbances (like uprooting due to increased hydrodynamics, see Fig. 2 ) contributed to or caused the seagrass loss. Regardless of which disturbance caused the seagrass loss, hollow size alone predicted the risk for sudden seagrass disturbance ( Fig. 4A ). Meanwhile, large enough hollows were caused by the combined effect of large disturbance and lugworm addition ( Fig. 4B ). Combined with the fact that disturbances on seagrass during autumn are already known to reduce local sediment stabilization in the end of the following growth season [43] , this indicates the existence of a positive feedback between disturbance on seagrass and low sediment stability. If strong enough, such a feedback could trigger the formation of two alternative and potentially stable states; (i) undisturbed hummocks with high sediment stability, and (ii) repeatedly disturbed hollows with low sediment stability ( Fig. 5 ). Shifts between the states should – based on our results – be mediated by factors influencing sediment stability; the size of disturbance and the bioturbation by lugworms. However, they may also be influenced by the overall risk for additional disturbance (e.g. geese density or hydrodynamic conditions), the erodability of sediments (e.g. silt content), and the strength of seagrass engineering (e.g. related to shoot density and height). The hypothesis of two alternative states is, moreover, supported by the pervasive spatial hummock-hollow mosaic in our study area ( Fig. 1A–D ), which is very similar to the bimodal state distribution typical for feedback-driven systems [see e.g. 17] , [22] . However, to prove that the states are truly persistent \n [49] , it must be demonstrated that the risk for additional disturbance not only depends on hollow size, but that disturbed hollows are much more prone to additional disturbance the following year. 10.1371/journal.pone.0023229.g005 Figure 5 Conceptual model of how seagrass-dominated hummocks and lugworm-dominated hollows could constitute two alternative and potentially stable states in intertidal soft-bottom areas. Local seagrass cover and lugworm density interactively affects sediment net stability, which in turn determines the risk for disturbance on seagrass. The effects of disturbance feeds back positively on local sediment stabilization the following season, so that hollows disturbed one year has an increased risk of being disturbed again the following year. Shifts between the states should be mediated primarily by factors influencing local sediment stability; the size of disturbance on seagrass and lugworm bioturbation intensity, but also by risk for disturbance (e.g. geese grazing intensity or hydrodynamic stress), sediment relative erodability, and seagrass engineering strength. Habitat-mediated competition can alone maintain spatial mosaics if counteracting engineering is strong enough to exclude competitors [11] , [20] , [22] , [50] . In our study area, lugworms were too weak of a competitor to exclude seagrasses. This unexpected result contrasts with documented seagrass exclusion by lugworms in nearby areas [24] , and suggests the relative engineering strength of lugworms – and thereby the outcome of habitat-mediated competition – is conditional i.e. depends on other factors (e.g. sediment conditions and hydrodynamic conditions). This may explain the sometimes conflicting evidence for competition between bioturbators and intertidal plants around the world [22] , [37] , [38] , [39] , [40] , [41] . However, even though lugworms were more or less outcompeted by the end of the summer in our study, their destabilization of sediments increased the size of hollows, which in medium-sized clearings exacerbated the risk for additional disturbance. This demonstrates that bioturbation can exacerbate scale-dependent responses, in the same way as loss of sediment-stabilizing engineers has been shown to slow recovery [42] . It also emphasizes that such engineering effects may locally outlive engineers [13] . Moreover, lugworm densities in fact increased during the last month in plots where the recolonizing seagrasses were disturbed (see Fig. 3B , Medium and Large hollows during October). This suggests that removal of a superior competitor (seagrass) facilitated the starting point of population turnover; another prerequisite for persistence of an alternative state [49] . In conclusion, our study supports the notion of biota as an important factor for disturbance responses and as drivers of ecosystem heterogeneity [5] , [17] , [51] . The results illustrate that habitat-mediated self-facilitation, counteracting ecosystem engineering and physical disturbances should not be viewed in isolation, because they may interactively explain ecosystem trajectories following disturbance, and, over time, ecosystem spatial heterogeneity. Even though this study appears to be the first demonstrating the potential importance of such interactions, we believe they are a common in engineered systems [see e.g. 22] . Finally, from a local Wadden Sea perspective, the results strengthen the hypothesis [51] that recent shifts from a historical dominance of sediment-stabilizing engineering species like seagrass and reef-forming mussels, to current dominance by bioturbators such as lugworms is caused by species-specific disturbances (e.g. disease, eutrophication, overgrazing, harvest, etc.), but maintained by sediment-mediated positive feedback interactions."
} | 4,556 |
28054641 | PMC5215187 | pmc | 3,603 | {
"abstract": "Some bacteria produce and perceive quorum-sensing (QS) signals that coordinate several behaviours, including the costly processes that are exoenzyme production and plasmid transfer. In the case of plasmid transfer, the emergence of QS signal-altered invaders and their policing are poorly documented. In Agrobacterium tumefaciens , the virulence Ti-plasmid encodes both synthesis and sensing of QS-signals, which promote its transfer from a donor to a recipient cell. Here, we reported that QS-altered A. tumefaciens mutants arose during experimental evolution. All showed improved growth compared to their ancestor. Genome sequencing revealed that, though some had lost the Ti-plasmid, most were defective for QS-signal synthesis and Ti-plasmid conjugation ( traR mutations) and one exhibited a QS-signal exploitation behaviour, using signal produced by other cells to enhance its own Ti-plasmid transfer. We explored mechanisms that can limit this QS-hijacking. We showed that the A. tumefaciens capacity to inactivate QS-signals by expressing QS-degrading enzyme could attenuate dissemination of the QS signal-negative Ti-plasmids. This work shows that enzymatic QS-disruption whether encoded by the QS-producing Ti-plasmid itself, by a companion plasmid in the same donor cells, or by one in the recipient cells, in all cases can serve as a mechanism for controlling QS exploitation by QS signal-negative mutants.",
"discussion": "Discussion This study shows that the QS regulon is costly for A. tumefaciens . Cost may arise from signal production and from synthesis, assembly and functioning of the dedicated type-IV secretion system for plasmid transfer 34 . The metabolic costs associated with these processes create conditions that are favourable for the selection of QS-impaired mutants. Among the characterized mutants, we observed, most commonly, the emergence of QS-defective mutants ( traR mutations) that neither synthesised nor perceived QS-signal and therefore did not transfer Ti-plasmid, but also loss of the Ti-plasmid and emergence of a QS-hijacking mutant with strongly reduced production of QS-signal but maintaining the ability to transfer the Ti-plasmid under the influence of QS-signals emitted by neighbouring cells. The master regulator TraR thus appears the preferential target for reducing the cost associated with the QS-activity, but one could not exclude that mutations in other loci, such as traI coding for QS-signal production, may occur. In Sinorhizobium(Ensifer) meliloti and Pseudomonas aeruginosa , mutations in the genes coding the master QS-regulators ExpR and LasR are also preferentially selected under costly QS-conditions 10 35 36 . When infecting plants, another process is costly for A. tumefaciens : the transfer of the T-DNA from the Ti-plasmid to the plant cells. The costly activation of this process also contributes to the selection of lower virulence in vitro and in planta via either Ti-plasmid loss or rearrangements in the Ti and At plasmids or the chromosome 14 15 16 17 18 . Altogether, these data suggest that two costly processes, T-DNA transfer and QS regulon occur when A. tumefaciens infects and colonizes the plant host. Selection should favour bacterial mutants that minimise these costs, for example, by losing or modifying integrity and conjugation capacity of the Ti-plasmid. On the other hand, the two processes contribute to maintain the Ti-plasmid in the infective population. Firstly, the bacteria harbouring a functional Ti-plasmid have a growth advantage in the opine-rich environment of the plant tumour, because they can consume these opines 19 . Secondly, some opines such as agrocinopines activate QS-activity and Ti-plasmid conjugation, which will lead to Ti-plasmid reintroduction into cells that have lost their Ti-plasmids. This should generate a complex dynamic of repeated cycles of loss and gain, summarized in Fig. 4 . Importantly, since cells containing a Ti-plasmid are several orders of magnitude less competent for the acquisition of another one 37 , QS-hijacking limitation is important because the maintenance of QS-producing Ti-plasmids depends largely on their ability to transfer to new host cells that contain no Ti-plasmid. We explored how the emergence and proliferation of QS-altered individuals in A. tumefaciens populations might be limited. A first mechanism may consist of limiting availability of QS-signal to non-producing cells. The main molecular actor of this process would be the regulatory protein TraM, which binds to the QS-master regulator TraR, preventing QS-signal synthesis 33 . Here we propose a complementary mechanism involving the QS-signal degrading lactonase. We show that a QS-signal degrading enzyme can impede the spread of QS signal-negative plasmids that do not encode QS-signal but make use of the QS-signal produced by other individuals to initiate their own transfer to recipient cells. The QS-signal degrading enzyme reduced the spread of QS signal-negative plasmids whether it was produced in cis by the QS-signal producing plasmid itself ( Fig. 3 ), in trans by a companion plasmid in the same cells ( Figure S2 ) or even in an independent recipient cells ( Figure S3 ), while not impeding the transfer of the Ti-plasmids from QS-signal-producing cells ( Fig. 3 ), consistent with previous observations 28 30 . The fact that QS-degrading enzymes affect only the transfer of QS signal-negative Ti-plasmids and not that of QS-producing plasmids can be explained at a mechanistic level. Indeed when QS-signal is bound to the TraR sensor, it is protected from lactonase-mediated degradation 30 . Cells hosting the QS signal-negative plasmid had only access to extracellular QS-signal (unbound to TraR) that diffused from the producing cells and could therefore be degraded by lactonase ( Fig. 3 ). In other QS-bacteria such as the opportunistic pathogen P. aeruginosa , QS-degrading enzymes 25 26 27 may also contribute to policing QS-cheating or QS-hijacking behaviours. This work reveals a novel paradigm involving QS-degrading enzymes for limiting QS exploitation by selfish invaders in cooperative bacterial populations. Given the broad distribution of QS-degrading enzymes in bacteria we suspect that this phenomenon might be of special importance to limit crosstalk between different QS system using the same QS signals in complex microbiomes. The QS-exploitation protection we describe could be complementary with the expected frequency-dependent dynamics of QS signal-negative and QS-emitting plasmids among bacterial populations 9 38 and the possibility of horizontal transfer of the QS-genes for converting hijackers into QS-cooperating partners 39 . We consider that conversion by horizontal transfer is unlikely in the case of A. tumefaciens because cells containing a Ti-plasmid are far less competent for the acquisition of a novel Ti-plasmid 37 . More generally additional factors may influence plasmid transfer and susceptibility to QS exploitation by signal negative plasmids. We found that different host genotypes differ in how QS signal-negative Ti-plasmids perceive and use the QS-signal for their own transfer. Finally, environmental heterogeneity may also affect the dynamics of QS signal-negative plasmids and cells by exerting variable selection pressures on collective behaviours 40 . In natural environments, animal and plant hosts, as well as several microorganisms in the associated microbiota produce a variety of compounds that may interact with QS-signal and its degradation 41 42 43 , thereby modifying the conditions of QS-plasmid transfer."
} | 1,903 |
35648853 | PMC9159573 | pmc | 3,604 | {
"abstract": "Integration of rigid components in soft polymer matrix is considered as the most feasible architecture to enable stretchable electronics. However, a method of suppressing cracks at the interface between soft and rigid materials due to excessive and repetitive deformations of various types remains a formidable challenge. Here, we geometrically engineered Ferris wheel–shaped islands (FWIs) capable of effectively suppressing crack propagation at the interface under various deformation modes (stretching, twisting, poking, and crumpling). The optimized FWIs have notable increased strain at failure and fatigue life compared with conventional circle- and square-shaped islands. Stretchable electronics composed of various rigid components (LED and coin cell) were demonstrated using intrinsically stretchable printed electrodes. Furthermore, electronic skin capable of differentiating various tactile stimuli without interference was demonstrated. Our method enables stretchable electronics that can be used under various geometrical forms with notable enhanced durability, enabling stretchable electronics to withstand potentially harsh conditions of everyday usage.",
"introduction": "INTRODUCTION Stretchable electronics enable a wide variety of previously unknown functions by offering various form factors not possible with rigid electronics ( 1 – 7 ). Recently, stretchable display ( 8 – 10 ), battery pack ( 11 – 13 ), sensor array ( 9 , 14 , 15 ), heater ( 16 ), and logic circuit ( 17 ) have been demonstrated. In stretchable electronics, components (rigid devices, electrodes, and sensors) that are vulnerable to lateral strain have been designed to be protected against mechanical deformation through various geometric engineering (e.g., serpentine, rigid island, and wrinkling). Among them, the strategy of placing nonstretchable devices on the rigid island arrays minimizes the lateral strain applied to the devices because the polymer matrix surrounding islands has a relatively low elastic modulus than the island and is therefore predominantly stretched ( 4 , 10 , 11 ). However, the interfacial cracks between rigid islands and polymer materials due to mechanical mismatch (difference in modulus, stretchability) are likely to occur under excessive or repetitive stretching ( 18 , 19 ), ultimately leading to crack propagation and device failure. To solve this challenge, several approaches to improving interfacial bonding strength between islands and the polymer matrix have been reported, mainly through chemical treatments ( 16 , 20 ). However, the usable soft polymer materials are limited, and the organic chemicals are not biocompatible. Furthermore, the fabrication process is not appropriate to produce the devices on the island array because organic chemicals must be coated to specific areas. Therefore, a universal approach to enhance the mechanical stability at the interface is essential to fabricate multifunctional and stretchable electronics capable of withstanding various deformation modes. Here, we present geometrically engineered rigid islands showing excellent mechanical stability at the interface ( Fig. 1A ). The interlocking structure in the proposed Ferris wheel–shaped island (FWI) effectively suppresses crack propagation at the interface. The optimized geometrical shapes of FWIs depend on the mechanical properties (e.g., toughness and stretchability) of the polymer materials (e.g., Ecoflex, Dragon Skin, and Ecoflex Gel). The repetitive interlocking structure prolongs the fatigue life against various three-dimensional (3D) deformation modes such as twisting, poking, and crumpling as well as stretching in the 1D direction. Furthermore, we demonstrate several applications taking advantage of the FWI array in Ecoflex and intrinsically stretchable electrode: stretchable electronics operating under various deformations (left image in Fig. 1B ) and electronic skin (e-skin) detecting tactile stimuli (right image in Fig. 1B ). The proposed approach will greatly enhance the durability of stretchable electronics under practical usage, thus strengthening their commercial viability. Fig. 1. FWIs embedded in Ecoflex substrate for highly durable stretchable electronics. ( A ) Schematic illustration of stretchable electronics with the Ferris wheel–shaped island (FWI) array in Ecoflex. ( B ) Left: schematic illustration of stretchable electronics operating under various deformations; right: schematic illustration of electronic skin (e-skin) detecting tactile stimuli. ( C ) Left: photographs of PLA islands embedded in Ecoflex; right: photographs comparing the maximum stretchability of circle-shaped island (CI) and FWI in Ecoflex substrate. The CI and FWI in Ecoflex are stretched to 75 and 175%, respectively. ( D ) Digital image correlation (DIC) images showing the progress of crack propagation for the CI and FWI in Ecoflex under stretching. ( E ) Stress versus strain for the CI (red trace) and FWI (blue trace) in Ecoflex under stretching. ( F ) The strain at failure according to the angle. The islands are rotated at specific angles, embedded in Ecoflex matrix, and stretched vertically. Scale bars, 1 cm (C) and 5 mm (D). Photo credit: J. C. Yang, Korea Advanced Institute of Science and Technology (KAIST).",
"discussion": "DISCUSSION For practical use of stretchable electronics in the future, it is of high importance to ensure high durability under potentially harsh conditions that the devices can be exposed to under everyday usage. In this sense, simple lateral strain testing is insufficient to qualify stretchable electronics toward practical applications. To bridge this gap, we developed FWIs with strong mechanical stability at the interface with soft polymer. Because of suppression of interfacial crack propagation by the interlocking structure, the FWI improved strain at failure under stretching and prolonged fatigue life under various deformation modes (stretching, twisting, poking, and crumpling). Various design parameters of FWI have a great influence on the mechanical stability of the substrate and depend on the mechanical properties of polymer materials. For practical demonstrations of stretchable electronics, we printed intrinsically stretchable electrodes and placed rigid components (LED and coin cell) on the FWI arrays. Furthermore, we fabricated e-skin capable of differentiating various physical stimuli. Our technique can be generally applied to a wide variety of stretchable electronics to impart high durability under various deformation modes, thus bringing stretchable electronics closer to commercialization in the near future."
} | 1,650 |
39031305 | PMC11425271 | pmc | 3,605 | {
"abstract": "Abstract Hydrogels present attractive opportunities as flexible sensors due to their soft nature and tunable physicochemical properties. Despite significant advances, practical application of hydrogel‐based sensor is limited by the lack of general routes to fabricate materials with combination of mechanical, conductive, and biological properties. Here, a multi‐functional hydrogel sensor is reported by in situ polymerizing of acrylamide (AM) with N,N′‐bis(acryloyl)cystamine (BA) dynamic crosslinked silver‐modified polydopamine (PDA) nanoparticles, namely PAM/BA‐Ag@PDA. Compared with traditional polyacrylamide (PAM) hydrogel, the BA‐Ag@PDA nanoparticles provide both high‐functionality crosslinks and multiple interactions within PAM networks, thereby endowing the optimized PAM/BA‐Ag@PDA hydrogel with significantly enhanced tensile/compressive strength (349.80 kPa at 383.57% tensile strain, 263.08 kPa at 90% compressive strain), lower hysteresis (5.2%), improved conductivity (2.51 S m −1 ) and excellent near‐infrared (NIR) light‐triggered self‐healing ability. As a strain sensor, the PAM/BA‐Ag@PDA hydrogel shows a good sensitivity (gauge factor of 1.86), rapid response time (138 ms), and high stability. Owing to abundant reactive groups in PDA, the PAM/BA‐Ag@PDA hydrogel exhibits inherent tissue adhesiveness and antioxidant, along with a synergistic antibacterial effect by PDA and Ag. Toward practical applications, the PAM/BA‐Ag@PDA hydrogel can conformally adhere to skin and monitor subtle activities and large‐scale movements with excellent reliability, demonstrating its promising applications as wearable sensors for healthcare.",
"conclusion": "3 Conclusion In summary, PAM/BA‐Ag@PDA hydrogel was developed as biofunctional sensor for health monitoring. Various ratios of PAM, BA, Ag@PDA were measured, and the optimal combination was chosen as the strain sensor with excellent tensile properties and resilience. In vitro adhesion tests showed the hydrogel sensor had good adhesion to tissue. Abundant H‐bonds and Ag‐thiolate coordination within the hydrogel matrix also provided the self‐healing property. The PAM/BA‐Ag@PDA sensor was antioxidative, antibacterial, and biocompatible by evaluation of DPPH free radical scavenging assay, bacterial killing test, and Live/Dead staining. All of the above are important for the hydrogel sensor in wearable applications. Characterization of the hydrogel sensor was conducted, which revealed the sensor had desired sensitivity, high linearity, low hysteresis, and good repeatability. Corresponding applications of human motion detection and human‐machine interfaces were conducted to show the potential of PAM/BA‐Ag@PDA hydrogel for a variety of applications in wearable sensors and electronic skins.",
"introduction": "1 Introduction Flexible sensors that can adhere to arbitrary surfaces and transduce mechanical deformations into electrical output signals have attracted tremendous attention and have been widely used for the applications of health monitoring, soft robots and human–machine interaction system. [ \n \n 1 \n , \n 2 \n , \n 3 \n \n ] Compared with traditional flexible sensors composed of elastomeric matrix (e.g., polydimethylsiloxane (PDMS), polyurethane (PU), and Ecoflex) and conductive components (e.g., liquid metals, metal nanoparticles/nanowires, MXene), [ \n \n 4 \n , \n 5 \n \n ] hydrogels have emerged as a promising alternative for wearable electronics owing to their similarities to natural tissues and versatility in electrical, mechanical, and biofunctional engineering. [ \n \n 6 \n , \n 7 \n \n ] By rationally designing the hydrogel networks, hydrogel can be endowed with multiple combined conductive, mechanical, and biological properties. [ \n \n 8 \n , \n 9 \n \n ] Despite significant advances and prospects, hydrogel‐based electronics for personalized health monitoring that can synchronously meet the demands of good mechanical strength, high stretchability, desirable sensitivity, excellent reliability, and biocompatibility still remain numerous challenges. [ \n \n 10 \n , \n 11 \n \n ] For example, using conductive polymers, such as polypyrrole (PPy), polyaniline (PANI), poly(3,4‐ethylenedioxythiophene) (PEDOT), as the frameworks of hydrogel sensors can obtain high conductivity, [ \n \n 12 \n \n ] but their biocompatibility, stretchability (intrinsic stiff and hydrophobic characteristics of the conductive polymers) and biodegradability need to be improved. [ \n \n 13 \n \n ] Doping conductive agents into hydrogels as fillers would suffer phase separated instability, limited stretchability, or leakage issues. [ \n \n 14 \n \n ] Moreover, most existing hydrogel sensors have complex components or complicated manufacturing process, restricting their wide applications. [ \n \n 15 \n , \n 16 \n \n ] Thus, developing advanced hydrogel sensor with combinations of stretchability, strength, sensitivity, biocompatibility, and ease of scalable fabrication is important for the next generation of wearable electronics. [ \n \n 17 \n \n ] \n Meanwhile, to achieve long‐term conformal attachment on skin with longevity, hydrogel sensors with multi‐functional properties, including self‐adhesion, self‐healing as well as anti‐inflammation/anti‐bacterial capability have received considerable attention for biomedical applications. [ \n \n 18 \n , \n 19 \n , \n 20 \n \n ] A variety of self‐adhesive and self‐healable hydrogel sensors were explored through synergistic reversible interactions of dynamic covalent (imine, acylhydrazone, disulfide bonds and borate ester bonds, etc.), non‐covalent (hydrogen bonding, ionic interaction, metal coordination, and host‐guest interaction, etc.) and irreversible covalent bonding (amide bond and carbon‐sulfide bond, etc.). [ \n \n 21 \n , \n 22 \n , \n 23 \n \n ] Recently, Yu et al. developed a self‐adhesive, self‐healing, biocompatible strain/pressure hydrogel sensor by in situ polymerization of acrylamide (PAM) in the presence of polydopamine‐modified carbon nanotubes (PDA@CNTs). The abundant hydrogen bonding and π‐π stacking derived from the PDA ensure the PDA@CNT/PAM hydrogel sensor with an autonomous self‐healing behavior (97.3% healing efficiency after 6 h), and the CNTs enhance the electronic transmission capacity and mechanical properties (17.5 kPa tensile strength at >700% strain) of the hydrogel sensor. [ \n \n 24 \n \n ] Furthermore, Zhang et al. developed a super stretchable, self‐healing, adhesive ionic conductive stress sensors based on cellulose nanofibrils (CNFs) modified phenylboronic acid‐ionic liquid (PBA‐IL)/acrylamide hydrogel. Owing to the multiple combined H‐bonds, electrostatic interactions, and borate ester bonds, the hydrogel sensor exhibited a tensile stress of 369.5 kPa ± 9.1 kPa with a self‐healing efficiency about 92% after 150 min. [ \n \n 25 \n \n ] To further improve the healing speed and controllability, stimulus (e.g., heat, pH, near‐infrared (NIR) light) triggered self‐healing hydrogel sensors were investigated. [ \n \n 26 \n \n ] Although tremendous progress, several issues remain unsolved. For instance, the mechanical strength of autonomous self‐healing hydrogel sensors is relatively weak. [ \n \n 27 \n , \n 28 \n \n ] Multiple synergistic interactions can improve the mechanical strength and self‐healing capability of hydrogel sensors, but their biocompatibility or stretchability might be compromised. [ \n \n 29 \n \n ] Achieving optimal properties for a desired hydrogel sensors with excellent mechanical properties, high healing efficiency, and desirable bioactivity are generally difficult (Table S1 , Supporting Information). [ \n \n 29 \n , \n 30 \n , \n 31 \n \n ] \n To address these, we propose a multi‐functional nanocomposite hydrogel (named PAM/BA‐Ag@PDA) with appealing characteristics as a strain sensor for human health monitoring by free radical polymerization of acrylamide (AM) with N,N′‐bis(acryloyl)cystamine (BA) modified silver‐doped polydopamine nanoparticles (BA‐Ag@PDA) ( Scheme \n \n 1 \n ). By introducing the crosslinkable BA‐Ag@PDA nanoparticles in the PAM hydrogel, improved mechanical strength with low hysteresis and good fatigue resistance can be achieved through the high‐functionality nanoparticle crosslinks and multiple interactions within the hydrogel networks. Moreover, the PAM/BA‐Ag@PDA hydrogel can be endowed with electrical sensitivity and NIR stimuli‐responsive ability due to the intrinsic conductive and photothermal ability of both Ag and PDA. Further, owing to the reversible crosslinks of BA‐Ag@PDA within the hydrogel network, by simultaneously involving abundant H‐bonds and Ag‐thiolate coordination, the PAM/BA‐Ag@PDA hydrogel can restore its original function after mechanical failure. Benefiting from the presence of catechol groups, the PAM/BA‐Ag@PDA hydrogel exhibits repeatable adhesiveness for conformal skin contact. Besides, due to the synergistic antibacterial effect of Ag with PDA and an antioxidant effect from PDA, the PAM/BA‐Ag@PDA is able to meet the increasing biologic requirements in future health monitoring. Scheme 1 Schematic illustration of the PAM/BA‐Ag@PDA nanocomposite hydrogel for wearable sensors.",
"discussion": "2 Results and Discussion 2.1 Characterization of the PAM/BA‐Ag@PDA Hydrogel The Ag@PDA nanoparticles were synthesized through in situ redox reaction between AgNO 3 and PDA nanoparticles due to the presence of dihydroxyphenyl groups on PDA. [ \n \n 32 \n \n ] Before Ag‐modification, the PDA nanoparticles displayed a uniform spherical shape with an average hydrodynamic diameter of 196 nm (Figures S1a and S2 , Supporting Information), and the C, N, and O elements were homogeneously distributed in the nanoparticles. By contrast, the surface of Ag@PDA nanoparticles was deposited with large amounts of Ag dots. The elemental distribution map of Ag indicated the homogeneous presence of Ag on PDA (Figure S1b , Supporting Information). To further obtain highly functionalized networks, the Ag@PDA nanoparticles were modified by BA through the dynamic silver‐thiolate coordination interactions (named BA‐Ag@PDA), and then introduced as the crosslinkers within the PAM network. TEM image showed that the BA‐Ag@PDA nanoparticles had a lichee‐like morphology with an average diameter of 238 nm ( Figure \n \n 1 a ; Figure S2 , Supporting Information). The corresponding SEM‐EDS mapping images of BA‐Ag@PDA indicated the C, N, O, Ag, and S elements were homogeneously distributed in the BA‐Ag@PDA (Figure 1b ). Figure 1 Characterization of the BA‐Ag@PDA nanoparticles. a) TEM image of the BA‐Ag@PDA nanoparticles. b) SEM and EDS mapping images of the BA‐Ag@PDA nanoparticles. c) XRD patterns of the BA, PDA, Ag@PDA, and BA‐Ag@PDA. d) XPS analysis of the PDA, Ag@PDA, and BA‐Ag@PDA nanoparticles. e) UV–vis–NIR absorption spectra of the BA, PDA, Ag@PDA, and BA‐Ag@PDA. f) FTIR spectra of the BA, dopamine, PDA, Ag@PDA, and BA‐Ag@PDA. g) SEM and EDS mapping images of the PAM/BA‐Ag@PDA network in the dry state. h) Tensile stress–strain curves of the PAM/BA‐Ag@PDA hydrogel with different contents of AM and Ag@PDA. Insert: Image of the 15AM/1.4BA‐1.0Ag@PDA group under twist and stretch condition. i) Compressive stress–strain curves and j) lap shear stress–strain curves of the PAM/BA‐Ag@PDA hydrogel with different contents ofAg@PDA. X‐ray diffraction (XRD) results indicated that the PDA had amorphous features with a broad peak around 2𝜃 = 24.2°. [ \n \n 33 \n \n ] By contrast, obvious (111), (200), (220), and (311) diffraction peaks of metallic Ag at 38.2°, 44.1°, 64.3°, and 77.3° were observed in the XRD pattern of Ag@PDA and BA‐Ag@PDA. [ \n \n 34 \n \n ] (Figure 1c ). Noteworthy, the diffraction peaks of crystalline Ag phase in BA‐Ag@PDA were weakened than that of Ag@PDA after the incorporation of BA, whereas four new diffraction peaks (2θ) at 27.9°, 46.3°, 54.88°, and 57.54°, which could be indexed to the (021), (−123), (212), and (140) planes of Ag 2 S (JCPDS Card no. 01‐089‐3840) were observed in the BA‐Ag@PDA. Moreover, a specific diffraction peak located at 31.76° assigned to the diffraction peak of BA was found, in which the slight shift from 31.76° to higher angle of 32.32° might be ascribed to Ag‐thiolate interaction and the formation of Ag 2 S. Consistent with these findings, the Raman spectra at 1345 and 1582 cm −1 in PDA, Ag@PDA and BA‐Ag@PDA corresponded to the stretching and deformation of aromatic rings in PDA backbone (Figure S3 , Supporting Information). [ \n \n 35 \n \n ] Moreover, X‐ray photoelectron spectroscopy (XPS) results of Ag@PDA and BA‐Ag@PDA at 367.8 eV (Ag 3d 5/2 ) and 373.8 eV (Ag 3d 3/2 ). [ \n \n 36 \n , \n 37 \n \n ] clearly indicated the existence of Ag on PDA (Figure 1d ; Figure S4 , Supporting Information). Meanwhile, a typical S 2p peak was detected in the XPS spectrum of BA‐Ag@PDA. [ \n \n 38 \n , \n 39 \n \n ] (Figure 1d ). Further, in the UV–vis spectra, BA‐Ag@PDA showed a redshifted absorption peak from 440 to 462 nm arising from the surface plasmon resonance effect of Ag nanoparticles (Figure 1e ; Figure S5 , Supporting Information). [ \n \n 40 \n , \n 41 \n \n ] Next, we reconfirmed the existence of alkene on the Ag@PDA nanoparticles by Fourier transform infrared spectroscopy (FTIR). Compared with Ag@PDA, a new absorption peak at 1634 cm −1 ascribed to the alkene vibration was found in the BA‐Ag@PDA, further indicating the successful preparation of BA‐Ag@PDA (Figure 1f ). To demonstrate the advantages of BA‐Ag@PDA in constructing high‐performance PAM‐based hydrogel, we first investigated the gelation behavior, mechanical property, and self‐healing ability of the PAM hydrogel using BA as the crosslinker. We found that the precursor solution of 15% AM containing 1.4 mg mL −1 of BA could not be crosslinked within 12 min via thermally initiated polymerization (Figure S6 , Supporting Information). Even though the gelation time was prolonged to 3 h, it was still unable to form fully crosslinked hydrogels. After demolding, a soft and partially crosslinked gel was observed for the PAM hydrogel. Moreover, the PAM hydrogel was fragile (9.66 kPa, Figure S7 , Supporting Information), and no self‐healing phenomenon occurred in the PAM hydrogel under NIR irradiation (Figure S8 , Supporting Information). SEM images indicated that the PAM hydrogel using BA as the crosslinker instead of the BA‐Ag@PDA nanoparticles displayed a collapsed and irregular network (Figure S9 , Supporting Information). While the networks of PAM/BA‐Ag@PDA exhibited branched architectures, in which the elements of Ag and S belonged to BA‐Ag@PDA were mainly distributed in the linear regions (Figure 1g ). Moreover, after swelling, a more compact and denser microporous network with pore size of about 10–50 µm was observed for the PAM/BA‐Ag@PDA than that of PAM. Since nano‐/microparticles can be introduced into polymer networks as high‐functionality crosslinks, providing tunable mechanical properties for hydrogels, [ \n \n 42 \n , \n 43 \n \n ] we systematically analyzed and optimized the properties of PAM/BA‐Ag@PDA hydrogel for their applications as wearable sensor. Detailed preparation information of the PAM/BA‐Ag@PDA hydrogels with different contents of AM, BA, and Ag@PDA is summarized in Table S2 (Supporting Information). Tensile stress–strain curves of the PAM/BA‐Ag@PDA hydrogels showed that a higher amount of AM resulted in a greater stretchability of hydrogel. Increasing the amount of BA‐Ag@PDA, the tensile stress of hydrogel first increased and then decreased. According to the definition of hydrogel networks, the PAM/BA‐Ag@PDA hydrogel networks belong to the category of unconventional polymer networks, in which both the high‐functionality architectures (multiple polymer chains can be interconnected at the crosslink of BA‐Ag@PDA nanoparticles) and multiple interactions (physical and reversible interactions) among polymer chains exist in the network. [ \n \n 42 \n , \n 43 \n , \n 44 \n \n ] (Scheme 1 ). In the dry state of a conventional polymer network, the end‐to‐end distances of a polymer chain at the relaxed and fully stretched states are defined as N b and Nb, respectively, where N is the number of Kuhn monomers in each polymer chain, b is the length of each Kuhn monomer. Therefore, the stretch limit of the polymer chains (λ lim ) in a dry polymer network can be calculated as λ lim = Nb ÷ N b = N . Although the PAM/BA‐Ag@PDA hydrogel networks are a kind of multimodal polymer networks, the longest polymer chains in PAM/BA‐Ag@PDA hydrogel can maintain its integrity up to the stretch limit, which can be calculated as: λ lim = N m a x ÷ λ s , where N max is the number of Kuhn monomers on the longest polymer chain, and λs is the effect of swelling on the λ lim . Following this principle, the stretchability of PAM/BA‐Ag@PDA hydrogel is closely related with the number of Kuhn monomers on the longest polymer chain. Correspondingly, with a constant amount of BA‐Ag@PDA, the strain of PAM/BA‐Ag@PDA hydrogel increases with the increase of AM monomer amount. Moreover, at a constant amount of AM monomer, a higher amount of BA‐Ag@PDA provides higher density of functionality crosslinks per unit volume in the hydrogel network, causing decreased number of Kuhn monomers in each polymer chain. Therefore, the tensile strain of PAM/BA‐Ag@PDA hydrogel at a high amount of BA‐Ag@PDA was significantly decreased (Figures S10–S14 , Supporting Information). Our experimental observations are in good agreement with the theoretical analyses. Tensile strength is commonly defined as the stresses at which the ultimate tensile failure occurs in the uniaxial tensile test, which can be evaluated as σ f = m f f f , where f f is the force required to fracture a single polymer chain, m f are the numbers of simultaneously fractured polymer chains per unit area of the polymer network at the deformed states. The design principle of strong hydrogels is to form substantial number of polymer chains in per unit area of polymer network to fracture simultaneously. [ \n \n 42 \n \n ] A greater number of polymer chains fracture simultaneously gives a higher tensile strength of the hydrogels under larger deformation. In the network of PAM/BA‐Ag@PDA hydrogel, there were multiple polymer chains bridging between two adjacent BA‐Ag@PDA crosslinks. At different ratios of AM, BA‐Ag@PDA, the number and lengths of polymer chains between adjacent crosslinks as well as the density of dynamic Ag‐thiolate coordination between polymer chains and crosslinks were different, thereby endowing the PAM/BA‐Ag@PDA hydrogel with different mechanical strength. Based on the strength principle and our experimental data, we found that without enough BA‐Ag@PDA crosslinks in the hydrogel networks, the AM monomers were mainly polymerized into long PAM polymer chains, resulting in limited number of polymer chains but prolonged polymer chains per unit volume. Therefore, a compromised tensile strength but excellent tensile strain was observed for the PAM/BA‐Ag@PDA hydrogels containing a low concentration of BA‐Ag@PDA (Figures S10 and S11 , Supporting Information). While, increasing the concentration of BA‐Ag@PDA, the tensile strength of PAM/BA‐Ag@PDA hydrogels first increased and then decreased. Notably, at the highest concentration of BA‐Ag@PDA, the tensile strength of PAM/BA‐Ag@PDA hydrogels was significantly decreased at either low or high concentration of AM monomers (Figures S10 and S14 , Supporting Information), indicating that a high crosslinks density might result in nonuniform lengths of polymer chains in the PAM/BA‐Ag@PDA networks. As the hydrogels undergo deforms, the short polymer chains and long polymer chains are successively fractured but not simultaneously fractured. Among them, the PAM/BA‐Ag@PDA hydrogel containing 20% AM, 1.0 mg mL −1 of BA and 2.0 mg mL −1 of Ag@PDA (named 20AM/1.0BA‐2.0Ag@PDA) achieved highest tensile strength (up to 789.89 kPa) with a maximal stretching ratio of 1154.37% (Figures S10 and S12 , Supporting Information). Despite attractive tensile properties, the 20AM/1.0BA‐2.0Ag@PDA showed unsatisfactory reversibility with obvious permanent deformation after the stress was removed (Figure S15 , Supporting Information). A hysteresis (19.3%) was observed in the loading‐unloading curve (Figure S16 , Supporting Information). The significant hysteresis would cause unreliable output electrical signal, impeding the practical use of hydrogel sensor. Comparatively, the PAM/BA‐Ag@PDA hydrogels containing 15% AM, 1.4 mg mL −1 of BA and 1.0 mg mL −1 of Ag@PDA (named 15AM/1.4BA‐1.0Ag@PDA) displayed good tensile stress/stretchability (349.80 kPa at 383.57% strain, Figure 1h and Figure S13 , Supporting Information), low hysteresis (5.2%, Figure S16 , Supporting Information) and desirable stability (Figure S17 , Supporting Information), guaranteeing its potential as strain sensor. Further, the 15PAM/1.4BA‐1.0Ag@PDA hydrogel remained undamaged when compressed to a large strain of 90% with a maximal compressive stress of 263.08 kPa (Figure 1i ). When applying 30%, 50%, and 70% compression strains, the 15AM/1.4BA‐1.0Ag@PDA hydrogel also showed neglectable recovery loss, revealing its excellent compression resilience under various compression strains (Figure S18 , Supporting Information). In addition, the 15AM/1.4BA‐1.0Ag@PDA hydrogel could directly adhere to organs or tissues without assistance of tapes (Figure S19 , Supporting Information). As determined by lap shear test using pork skin as adherends, the 15AM/1.4BA‐1.0Ag@PDA hydrogel showed an adhesion strength of 5.00 kPa (Figure 1j ; Figure S20 , Supporting Information). Collectively, the PAM/BA‐Ag@PDA hydrogel containing 15% AM, 1.4 mg mL −1 of BA and 1.0 mg mL −1 of Ag@PDA with optimal mechanical properties and good resilience was adopted for the application of strain sensor. 2.2 Sensing Properties of the PAM/BA‐Ag@PDA Hydrogel Based Resistive Strain Sensor Next, we measured the conductivity of PAM/BA‐Ag@PDA hydrogel. Notably, the conductivity of PAM/BA‐Ag@PDA hydrogel was 2.51 S m −1 , which was 4.5‐fold higher than that of PAM (Figure S21 , Supporting Information). To quantitatively characterize the flexible strain sensor based on PAM/BA‐Ag@PDA hydrogel, the hydrogel sensor was directly connected to an impedance analyzer to record the real‐time change in resistance under various stimuli from a universal testing machine. As depicted in Figure \n \n 2 a , the dumbbell‐shaped hydrogel was stretched from 0% to 200% without breakage, displaying excellent stretchability. The relative change in resistance of the hydrogel‐based resistive strain sensor was generated at loading/unloading speeds of 50% strain per min to 150% strain per min with the same tensile strain (200%), as shown in Figure 2b . The output curves were stable and monotonic, showing high linearity and positive correlation with tensile strain. The sensitivity of the strain sensor was evaluated by gauge factor (GF), which is defined as GF = Δ R / R 0 ε × 100 % , where ΔR and R 0 represent the change in resistance and the initial resistance of the hydrogel strain sensor respectively. [ \n \n 45 \n \n ] The GF value for the conductive hydrogel across a 200% strain range was approximately 1.86, indicating its good potential sensitivity as a strain sensor. The response time was another key factor for evaluating the performance of a sensor when detecting rapid vibrations. Figure 2c shows that the sensor provided a fast response to both stretching and relaxing processes. The response time and recovery time were 138 and 161 ms, respectively, enabling the device to sense real‐life human motions timely, such as human–machine interfacing, [ \n \n 46 \n \n ] subtle throat movement, [ \n \n 47 \n \n ] and muscle contracting, [ \n \n 48 \n \n ] etc. The slight drift in peak plateau data is mainly associated with the viscoelastic effect of hydrogels. [ \n \n 25 \n , \n 46 \n \n ] When the velocity of loading is high, residual deformation occurs, leading to a gradual upward drift in resistance. This effect is particularly pronounced during response time testing, where the loading speed reaches as high as 1100 mm min −1 . Figure 2 Strain sensing properties of the PAM/BA‐Ag@PDA hydrogel‐based resistive strain sensors. a) Photographs of the PAM/BA‐Ag@PDA hydrogel sensor to strains during the uniaxial stretching test. b) Relative resistance changes (ΔR/R 0 signal) of the PAM/BA‐Ag@PDA hydrogel sensor at different stretching speeds (50% min −1 , 100% min −1 , 150% min −1 ) for one cycle. c) Response and recovery time of the PAM/BA‐Ag@PDA hydrogel sensor under small tensile strain. d) Dynamic response of the PAM/BA‐Ag@PDA hydrogel sensor under a series of unloading step‐down strains from 100% to the initial state. e,f) Relative resistance changes of the PAM/BA‐Ag@PDA hydrogel sensor under increasing strain increments of 5% and 40%. g) Relative resistance changes of the PAM/BA‐Ag@PDA hydrogel sensor during cyclic tensile test at 10%, 20%, 30%, and 40% strain (strain frequency: 0.04 Hz). h) Relative resistance changes of the PAM/BA‐Ag@PDA hydrogel sensor at 50%, 100%, and 200% strain for 100 cycles (stretching–releasing speed: 120 mm min −1 ). Step and hold tests were conducted to evaluate the strain sensor stability and limit of detection (LOD) with different magnitudes of loading/unloading processes. As shown in Figure 2d , a 10% per step release from 100% strain to the initial state resulted in the sensor's resistance change behavior responding remarkably fast and robustly to the applied strain. Additionally, similar to the unloading process, the sensor also exhibited outstanding stability and excellent reliability whether under small (5%) or large (40%) stretching process step by step. Figure 2e,f shows the response signal trend of the PAM/BA‐Ag@PDA hydrogel sensor as a regular step‐like change with negligible overshoot. The dynamic responsive behaviors of the sensor under cyclic loading were displayed in Figure 2g , which exhibited a uniform and consistent output to the cyclic loading applied. The responses to different strains were almost identical except for the peak magnitude of the output curves with a noise‐free and stable strain response observed from 10% to 40%. For practical applications, strain sensors also need to be highly stretchable to achieve optimal sensing performance, such as accommodating up to 100% strain in human joint motion. [ \n \n 49 \n \n ] Furthermore, the ability of strain sensors to meet the demand of large deformations enhances their wearability, preventing discomfort from constriction, especially when applied to complex biological surfaces (>50% strain). [ \n \n 50 \n \n ] Accordingly, the durability and fatigue resistance of the hydrogel‐based strain sensor were also experimentally tested by continuously stretching/releasing strains of 50%, 100%, and 200%. As shown in Figure 2h , ΔR/R 0 increased as the strain increased, and the sensor also adapted well to continuous different degrees of strain. The real‐time resistance variation signals maintained nearly identical values with only a slight attenuation observed after 100 cycles for different degrees of strain, revealing its high durability in long‐term usage. 2.3 Sensing Properties of the Self‐Healed PAM/BA‐Ag@PDA Hydrogel The self‐healing ability of material to recover from physical damage is another attractive feature for wearable sensor with improved durability and longevity. Owing to the reversible Ag‐thiolate coordination, [ \n \n 51 \n , \n 52 \n \n ] the PAM/BA‐Ag@PDA exhibited great NIR‐responsive self‐healing performance during the cutting‐healing process ( Figure \n \n 3 a ). We measured the tensile property of the self‐healed PAM/BA‐Ag@PDA hydrogel. After 10 min of NIR irradiation, the healed hydrogel could withstand 364.2% strain and a tensile stress of 282.53 kPa (Figure 3a ; Figure S22a , Supporting Information). The tensile stress–strain data indicated that the healing efficiency of the PAM/BA‐Ag@PDA hydrogel was related to the NIR exposure time (Figure S22b , Supporting Information). Moreover, the Ag@PDA nanoparticles could also provide dynamic interactions with the BA‐PAM polymer chain via the thiolate‐metal coordination, endowing a distinctive self‐healing function between the PAM and PAM/BA‐Ag@PDA hydrogel (Figure S8 , Supporting Information). Figure 3 a) Self‐healing behavior of the PAM/BA‐Ag@PDA hydrogel after NIR irradiation. b) The LED electrical performance of the self‐healed PAM/BA‐Ag@PDA hydrogel. c) Relative resistance changes of the self‐healed PAM/BA‐Ag@PDA hydrogel at 50% and 100% strain for 200 consecutive cycles (stretching–releasing speed: 120 mm min −1 ). Relative resistance changes of the d) original and e) self‐healed PAM/BA‐Ag@PDA hydrogel at 30th, 60th, 90th, 120th, 150th, and 180th cycles in the strain range of 0–100%. In the light‐emitting diode (LED) circuit experiment, the brightness of LED bulb immediately went off when the PAM/BA‐Ag@PDA sensor was cut off. After NIR irradiation for 8 min, the red LED was lighted up again due to the self‐healing ability of the PAM/BA‐Ag@PDA hydrogel (Figure 3b ). To further confirm its reliability and durability, the sensing performances of the healed PAM/BA‐Ag@PDA hydrogel during cyclic tensile test were investigated at 50% and 100% strain. As shown in Figure 3c , the healed PAM/BA‐Ag@PDA hydrogel still maintained a good stability with negligible decay of resistance after 200 consecutive stretching–releasing cycles. Correspondingly, the output ΔR/R 0 signals of the original and self‐healed PAM/BA‐Ag@PDA hydrogel at 100% strain of different cycles (30th, 60th, 90th, 120th, 150th, and 180th) were shown in Figure 3d,e , in which only a slight upward drift was observed between each cycle. The main reason for the drift in various cycles was that the plastic deformation of hydrogel existed under prolonged large deformation. [ \n \n 53 \n , \n 54 \n \n ] In terms of the cyclic tensile test at 200% strain, the relative resistance amplitude grew larger after 100 stretch cycles, which was mainly due to the cracks appeared in the healed hydrogel under large deformation (Figure S23 , Supporting Information). 2.4 Antioxidant, Antibacteria, and Biocompatibility of the PAM/BA‐Ag@PDA Hydrogel As a skin‐attachable electronic for health detection, there is a clinical need for improved the bioactivity and biocompatibility of sensors to prevent inflammation, microbial infections, and potential toxicity. Dopamine, the natural polyphenol existing in human, fruits, and vegetables, contains abundant phenolic hydroxyl groups that can react with free radicals to terminate the oxidation via hydrogen transfer and single electron transfer reactions, possessing outstanding antioxidant properties. [ \n \n 55 \n , \n 56 \n , \n 57 \n \n ] To verify these functions, the antioxidant effect, antibacterial activity, and cell cytotoxicity of the PAM/BA‐Ag@PDA hydrogel were evaluated by 2,2‐diphenyl‐1‐picrylhydrazyl (DPPH) free radical scavenging assay, bacterial killing test, and Live/Dead staining, respectively. As indicated in Figure \n \n 4 a , the UV–vis absorbance of DPPH radical at 517 nm gradually deceased in response to increasing concentration of PAM/BA‐Ag@PDA hydrogel. At a concentration of 20 mg mL −1 , 80.53% of the DPPH radical was eliminated by the PAM/BA‐Ag@PDA hydrogel (Figure 4b ). Moreover, due to the conjugated π–π structure in PDA that electron leap can occur under 808 nm NIR irradiation, the PAM/BA‐Ag@PDA exhibited excellent photothermal conversion ability under NIR irradiation. Increasing the power density of NIR, an increased photothermal conversion performance was observed. With a NIR power density of 1.5 W cm −2 , the temperature of PAM/BA‐Ag@PDA could reach ≈65.9 °C in 5 min, and the photothermal conversion capability maintained stable after 4 heating‐cooling cycles (Figure 4c ). Figure 4 a) UV–vis absorption spectra of the DPPH free radicals incubated with different concentrations of PAM/BA‐Ag@PDA hydrogel. b) DPPH radical scavenging percentage of the PAM/BA‐Ag@PDA hydrogel at different concentrations. c) Photothermal efficiency of the PAM/BA‐Ag@PDA hydrogel at different NIR power densities. Insert image: Photostability of the PAM/BA‐Ag@PDA hydrogel after four on/off NIR irradiation (1.5 W cm −2 ) cycles. d) Antibacterial activity of the PAM/BA‐Ag@PDA hydrogels with or without NIR irradiation (808 nm, 1.5 W cm −2 , 3 min). e) Biocompatibility of the PAM/BA‐Ag@PDA hydrogel determined by CCK‐8 and Live/Dead assay. The antibacterial abilities of PAM/BA‐Ag@PDA hydrogels were investigated by bacterial colony‐forming units (CFUs) counting test on agar plate (Figure 4d ). In vitro antibacterial tests proved a significant antibacterial activity of PAM/BA‐Ag@PDA to both Escherichia coli ( E. coli , Gram‐negative bacteria) and Staphylococcus aureus ( S. aureus , Gram‐positive bacteria). Compared with PBS, 48.54% of E. coli and 50.03% of S. aureus were killed by PAM/BA‐Ag@PDA. Moreover, owing to the synergistic effects of bacteria capture, Ag ion release and photothermal conversion from Ag@PDA, the NIR‐treated PAM/BA‐Ag@PDA hydrogel exhibited a further enhanced antibacterial effect. The bacteria‐killing efficiency of NIR‐treated PAM/BA‐Ag@PDA for E. coli and S. aureus were calculated to be 95.74% and 98.27%, respectively. In accordance with the in vitro observation, the PAM/BA‐Ag@PDA also exhibited a potent antibacterial activity in a S. aureus ‐infected rat full‐thickness skin defect model. The levels of viable bacteria for control, PAM/BA‐Ag@PDA and NIR‐treated PAM/BA‐Ag@PDA were 103.56%, 42.22%, and 4.21%, respectively, reconfirming the excellent and broad‐spectrum antibacterial activity of PAM/BA‐Ag@PDA. Additionally, CCK‐8 results indicated that the PAM/BA‐Ag@PDA hydrogel could support the normal proliferation of cells (Figure 4e ). There was no significant difference in cellular proliferation between the PAM/BA‐Ag@PDA and control. Consistently, Live/Dead staining of the cells co‐incubated with PAM/BA‐Ag@PDA hydrogel showed a similar green fluorescence to the control group, reconfirming that the PAM/BA‐Ag@PDA sensor had no side effects on cytocompatibility. 2.5 Practical Applications for Human Movement Detection and Human–Machine Interface In our daily life, various human activities necessitate strain sensors to recognize distinct degrees of skin strain, joint bending, etc. The excellent strain sensing properties of the PAM/BA‐Ag@PDA hydrogel‐based resistive strain sensors enable their use in various scenarios. As depicted in Figure \n \n 5 a , the hydrogel sensor was adhered to a volunteer's neck to monitor his neck shake to left and right shoulder. The highly reversible and repeatable response signals were recorded when the volunteer tilted his head, showcasing the potential application of this technique for aiding cervical syndrome patients in rehabilitation. Similarly, Figure 5b demonstrates the application of the PAM/BA‐Ag@PDA sensor on the laryngeal node to measure the throat's movement up and down during swallowing activities. The sensor maintained a stable response during the act of drinking water, demonstrating the potential application of this sensor in addressing oropharyngeal dysphagia resulting from stroke or other diseases. As seen in the inset of Figure 5c , a single PAM/BA‐Ag@PDA sensor was adhered to the index finger joint to sense the bending angle. When the index finger was continuously bent at 30°, 60°, and 90° and held for approximately 3 s, the recorded signal showed distinct ladder‐shaped resistance changes with the gradually increased bending angle. The results indicated that the PAM/BA‐Ag@PDA sensor was promising for use in motor function evaluation with good stability. In Figure 5d , the hydrogel sensor was employed to monitor biceps contraction and relaxation according to the variation of resistance with good repeatability. Additionally, changes in contracting frequency also resulted in changes in resistance with the same frequency, highlighting the sensor's ability to real‐time and accurately record muscle movements. The most notable characteristic of the obtained sensor was its high sensitivity to various vibration stimuli. By attaching the hydrogel sensor onto the chest of a volunteer, the periodic expansion and contraction states were captured precisely. Figure 5e shows that the sensor maintained a stable response during standing, running, and recovery states. From the normal state to running, both the velocity and intensity of breath waveforms sharply increased, indicating that volunteers required increasing amounts of oxygen during exercise to support the activity and metabolize the produced carbon dioxide. As the subject gradually cooled down, the breath waveform began to converge toward the normal state. It holds significant potential for use in athlete training, sports competitions, as well as monitoring respiratory‐related diseases. Figure 5 Detecting strain changes in human movement for healthcare monitor and human–machine interface. Relative resistance responses of the PAM/BA‐Ag@PDA based resistive strain sensor to: a) neck twisting, b) subtle throat movements, c) finger bending with different amplitudes, d) biceps contracting, e) real‐time sensing of breathing during standing (normal), running, and recovery by attaching the sensor to the abdomen. f) Photograph of the robot hand integrated with five hydrogel‐based resistive strain sensors on the back of fingers. g) Electrical signals of the robot hand for different gestures. h) Photograph of the application of the PAM/BA‐Ag@PDA hydrogel sensor to monitor knee‐joint activity during the squats. i) Real‐time relative resistance changes of the PAM/BA‐Ag@PDA hydrogel sensor used for large deformation of the knee‐joint. As a further proof‐of‐concept demonstration of human–machine interaction, the hydrogel sensors were attached to the back surface of a robot fingers to measure the time‐dependent single signal of each finger, as shown in Figure 5f . When the robot fingers moved, data regarding the movements were captured, which could be aggregated into a sophisticated array of electrical signals representing various gestures. As displayed in Figure 5g , the finger data channels were independent, and each electrical signal included a set of crests and troughs, which could reflect the position of the fingers. The hydrogel sensors could accurately identify different motions when the robot hand grasped two different objects: the grip of table tennis required bending all five fingers, resulting in peak signals in all five channels of relative resistance change. Simultaneously, the grip of a key involved only the thumb and index finger, thus corresponding strain hydrogel sensors registered changes in their output signals. Sign language such as “ok” and “yeah” could be distinguished by the difference in the fast‐changing signal details; signal changes in all channels were clearly observed when the robot hand gestured. The exceptional ability to enable the robot hand sense various gestures endows this hydrogel strain sensor with promising potential for human–machine interface and sign language communication. [ \n \n 58 \n \n ] Moreover, to ensure accurate measurement under large deformation, we tested the PAM/BA‐Ag@PDA hydrogel sensor's capability to monitor large‐scale bending‐related human knee‐joint movements. As shown in Figure 5h,i , when the volunteer changed his transitions from a standing position to a full squat, the knee joint's bending angle progressively increased from 0° to over 90°. Synchronously, the strain sensor converted the knee‐joint movements into resistance output signals with excellent repeatability and reliability."
} | 9,941 |
38947931 | PMC11213128 | pmc | 3,607 | {
"abstract": "Controlling stochastic temporal networks remains an open challenge in control theory. While predictable temporal networks with known future dynamics enhance controllability, real-world networks often exhibit stochasticity and unpredictability, making control harder. Here, we investigate control mechanisms for stochastic temporal networks by analyzing how biological controllers, such as shepherd dogs, manage panicked flocks of sheep. We studied a century-old shepherding competition, the sheepdog trials, where small groups of sheep unpredictably switch between fleeing and following behaviors–effectively forming stochastic temporal networks. Unlike large, cohesive flocks, these small, indecisive flocks are difficult to control, yet skilled dog-handler teams excel at both herding and splitting them (shedding) on demand. Using a stochastic choice model to describe the sheep’s behavioral shifts, we found that trained dogs exploit stochastic indecisiveness, typically seen as an obstacle, as a control tool, enabling both herding and splitting of noisy groups of sheep. Building on these insights, we developed the Indecisive Swarm Algorithm (ISA) for artificial agents and benchmarked its performance against standard approaches, including the Averaging-Based Swarm Algorithm (ASA) and the Leader-Follower Swarm Algorithm (LFSA). ISA minimizes control energy in trajectory-following tasks and outperforms alternatives under noisy conditions. By framing these results within a stochastic temporal network perspective, we demonstrate that even probabilistic knowledge of future dynamics can enhance control efficiency in specific scenarios. These findings establish a framework for managing stochastic temporal networks with applications in noisy, behavior-switching animal collectives, swarm robotics, and opinion dynamics.",
"introduction": "Introduction Emergent collective dynamics, where simple local interactions give rise to complex global behaviors, govern a wide range of systems. Examples include swarm robotics 1 , animal collectives 2 , social networks such as opinion dynamics 3 , pedestrians’ movements 4 , and vehicular traffic 5 . Controlling these systems is challenging, as their behaviors often defy traditional control methods 6 – 8 . Unlike systems with predictable, linear dynamics, emergent systems are best described as complex networks that require multiscale strategies to address both the microscopic interactions between individual agents and the macroscopic patterns that emerge at the group level 8 . Most of these networked systems introduce additional complexity when individual agents (nodes) switch between different behaviors, leading to temporal restructuring in the network. Biological collectives and social interactions in humans serve as prime examples of such behavior switching 9 – 13 (see SI Section 1 and Table S1 for a full list of behavior-switching systems from ants and locusts to seals and humans). Carrier ants transporting cargo alternate roles between lifters and pullers based on their orientation and the nest’s position 9 , sheep in small flocks randomly switch between leading and following roles 12 , and during an epidemic outbreak, humans frequently switch between different interaction partners, facilitating spread of diseases 14 . These systems highlight the need for control strategies that account for the stochastic and context-dependent nature of individual behavior transitions and their cascading effects on the evolving temporal networks, where edges dynamically reorganize over time 15 – 17 . Recently, it has been shown that temporal restructuring can improve the controllability of a network 18 . Specifically, temporal networks require less time and less energetic cost to be controlled than their static counterparts 19 , 20 . This counterintuitive observation relies on the fact that the future dynamics of the network are predictable and are exploited in designing the controls in the previous steps. However, when the switching dynamics are stochastic and unpredictable, temporality can make the control process more energetically demanding compared to a static network 21 . Therefore, despite advances in control theory and swarm robotics 7 , 8 , managing the dynamics of stochastic temporal networks remains an open challenge, particularly in systems where individual agents exhibit behavior switching. Predator-prey systems provide a natural framework for studying the challenges of controlling such noisy networks with behavior-switching dynamics. For instance, flocks of starlings confuse raptors by transitioning between complex dynamic patterns. Similarly, large herds of wildebeests intermittently shift between selfish herding and solitary flight when confronted by predators like cheetahs. In response, predators, instead of complex control mechanisms, adopt simplified strategies like focusing on a fixed point in space rather than tracking individual prey 22 , 23 . This allows them to split vulnerable individuals before leveraging speed and agility to secure their target 24 – 26 . These examples suggest that effective control of stochastic temporal networks with behavior-switching individuals does not always require precise prediction of behavioral transitions. In this work, we analyze such control mechanisms by studying shepherd dogs managing small flocks of sheep in a competition called the sheep-dog trials. Two key features of these competitions make them model systems for investigating control mechanisms in stochastic temporal networks. First, during these trials, when threatened, panicked sheep oscillate unpredictably and indecisively between fleeing from the dog and following other sheep, forming a stochastic temporal network. Trained shepherd dogs are highly effective at managing these noisy flocks under fluctuating conditions (SI Video 1). Second, unlike interactions between predators and large herds of animals in the wild, the sheepdog trials competitions provide a controlled environment where the behavior-switching dynamics of the sheep can be observed, quantified, and analyzed (see SI Section 2&3 for history and competition rules). By bridging empirical observations with quantitative modeling to analyze various tasks in the sheepdog trials competition, we find that shepherd dogs utilize the behavior switching in sheep for herding and splitting (shedding) the flocks. Behavior switching dynamics have been previously studied in the context of animal collectives and human societies (voter models) using individual-based stochastic choice models 11 , 27 – 30 . In this work, we build on the existing framework to frame sheepdog trials as a control problem for “indecisive collectives” — systems where agents stochastically alternate between different behaviors and interaction partners in the presence of an external agent. This paper is structured as follows: We begin by exploring the nuances and rules of sheepdog trials. Next, we present a stochastic framework to develop quantitative metrics such as “pressure” and “lightness” that capture the nuanced behavior of sheep. The framework is based on qualitative insights from experienced handlers, and empirical data on sheep-dog dynamics. We then present a stochastic choice model and the master equation to model the indecisive transitions in sheep movement, comparing our model’s predictions with observed dynamics. Building on this, we investigate whether sheep indecisiveness could benefit the dog. Our findings reveal that stochastic indecisiveness can aid the dog in both herding and shedding tasks. Finally, we extend our analysis to develop the Indecisive Swarm Algorithm (ISA), a swarm control strategy inspired by shepherding dynamics. By modeling ISA as a non-reciprocal stochastic temporal network and comparing it against the standard Averaging-Based Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA), we demonstrate that for specific control tasks like herding, ISA minimizes control energy requirements.",
"discussion": "Discussion Summary This study investigates control mechanisms for noisy, indecisive collectives, using sheepdog trials as a model system. These trials challenge trained shepherd dogs to herd and shed (split) small flocks of sheep ( N s ≤ 5), where the dynamics differ markedly from larger flocks. Unlike the cohesive selfish herd behavior seen in large groups, sheep in small flocks stochastically transition between fleeing (solitary behavior) and following the group (collective behavior), making them harder to control (i.e., an indecisive herd). By combining qualitative insights from expert dog handlers with a stochastic modeling framework, we analyze how trained dogs manage these indecisive sheep collectives. We find that sheep behavior depends on two key factors: the dog’s threat level and the sheep’s switching dynamics. Within the shepherding community, these factors are encapsulated by the terms “pressure” (the dog’s threat) and “lightness” (the isotropy of the sheep’s responsiveness). Light and heavy sheep exhibit distinct behaviors during herding and shedding tasks. To translate this nuanced qualitative knowledge into a quantitative framework, we developed a stochastic model to describe indecisive sheep behavior. The model reveals that trained dogs employ a two-step control strategy: first aligning stationary sheep to a desired orientation (orientation step) before increasing threat to initiate movement (movement step). Focusing on the orientation step, we modeled sheep as stationary agents that reorient stochastically. This analysis formalized the concepts of pressure and lightness, confirming their utility as core descriptors of sheep behavior. Comparing the model to data from actual sheepdog trials, we find that high isotropy aids group cohesion (for herding) but complicates splitting, while the dynamics of indecisive sheep are largely governed by the two parameters, pressure and lightness. We also investigated whether indecisiveness benefits the controller rather than solely posing a challenge. Extending our framework to simulate both orientation and movement steps in a 2D arena, we compared indecisive sheep agents with standard averaging-based Vicsek-type agents. While averaging agents outperform indecisive agents under low noise conditions, the reverse is true at higher noise levels. For shedding tasks, averaging agents consistently fail to split, while indecisive agents shed easily, irrespective of noise levels. These results highlight how trained dogs exploit the sheep indecisiveness as a tool and underscore the importance of the two-step control process. Finally, we explored whether indecisiveness could improve control strategies in artificial systems. Developing the Indecisive Swarm Algorithm (ISA), we compared it against the Averaging-based Swarm Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA) in a trajectory-following task. ISA agents successfully followed predefined trajectories at low stimulus intensities from the controller, unlike ASA and LFSA agents, which deviated significantly. Framing swarm algorithms as stochastic temporal networks, we identified two tunable timescales: the dynamics update timescale ( τ d ) and network restructuring timescale ( τ n ). By defining temporality 𝒯 = τ d / τ n , we showed that adjusting 𝒯 reproduces all three algorithms: ASA ( 𝒯 → ∞ ) , L F S A ( 𝒯 → 0 ) , and ISA ( 𝒯 = 1 ). Borrowing the concept of control energy from control theory, we quantified the stimulus intensity required to steer a swarm. ISA required the least control energy, demonstrating its effectiveness in herding noisy swarms. Our findings reveal the counterintuitive advantages of indecisiveness in controlling noisy collectives, with applications ranging from sheepdog trials to artificial swarms. By introducing deliberate indecisiveness, controllers can enhance their ability to perform complex tasks, such as herding and splitting, while also reducing effort in simpler tasks like trajectory-following. Why Sheepdog Trials are Challenging If indecisive agents require less control effort, why are sheepdog trials considered so challenging? To address this, we extended our indecisive model to large group sizes (SI Section 8). While the model was originally designed to explain the behavior of small groups ( N s ≤ 5) in response to external stimuli, its extension to larger group sizes captures dynamics consistent with known sheep behaviors. This broader application allowed us to propose a unified phase diagram for indecisive behavior (see SI Section 8 for details), offering insights into transitions between different behavioral regimes as group size and stimulus specificity change. The phase diagram ( Figure 8 ) illustrates the likelihood of individuals being influenced by controlling stimuli ( α ), intragroup interactions ( γ ), or random noise (non-specific stimulus) ( ε ) . Stimulus specificity, defined as the ratio of α / ε , measures the strength of external stimuli to noise. External stimuli, such as a dog’s pressure or the departure of an informed sheep, are key factors driving transitions between behaviors. We identify three distinct behavioral regimes: flocking (red), dominated by intra-group interaction, resulting in cohesive group behavior; fleeing (blue), dominated by specific stimuli where individuals act independently, ignoring the group; and grazing (green), dominated by random noise, with individuals disregarding both specific stimuli and the group. In small groups, increasing stimulus specificity shifts behavior from grazing to fleeing. In larger groups, flocking dominates under typical stimulus intensities. However, when stimulus specificity becomes extremely high - such as during a predator attack or an encounter with an untrained dog - the flocking phase transitions to fleeing, even in large groups ( Figure 8 ). We validated our model’s predictions by comparing them with prior empirical studies of sheep behavior. King et al. 32 (circle) observed that intermediate-sized groups (46 sheep) exhibited selfish herd behavior under high stimulus specificity, with herding dogs inducing cohesion. Toulet et al. 60 (square) found that when a trained sheep departs intermediate-sized groups (8-32 sheep), the group reaches a consensus to follow or ignore the individual, demonstrating the dominance of intra-group interactions even under mild stimuli (low specificity). Ginelli et al. 61 and Gomez-Nava et al. 12 (star and triangle) studied group dynamics without external stimuli. Ginelli focused on large groups (100 sheep), while Gomez-Nava examined small groups (4 sheep). Both identifyied intermittent grazing and flocking epochs, aligning with the grazing-flocking transition boundary in our model. These behaviors suggest an evolutionary anticipation of external threats as a defense mechanism. Our model (red line) predicts that small groups transition from grazing to uncontrolled fleeing through a narrow flocking phase as external stimulus increases. This prediction explains why managing small flocks is particularly difficult in sheep-dog trials. Since individual sheep vary in their responsiveness to stimuli, effectively herding or splitting small flocks requires the dog to balance intra-group cohesion with individual responsiveness, as excessive stimulus risks triggering chaotic fleeing. This underscores the complexity of controlling small, indecisive collectives, where behavioral transitions depend on a delicate interplay of external stimuli, noise, and group interactions. Temporality and Indecisiveness Temporal networks have been shown to require significantly less control energy than static networks 18 . This efficiency arises from their ability to leverage changing topologies to exploit favorable configurations, thereby reducing the need to counteract unfavorable system dynamics. In contrast, static networks, with their fixed structures, often force controllers to expend substantial energy to navigate energetically costly directions or to overcome inherent system dynamics. A useful analogy is sailing, where adjusting the sail to align with shifting wind directions enhances efficiency, rather than struggling against them 18 . However, this framework assumes that the controller has prior knowledge of future topology changes. Without such foresight, temporality can actually increase control energy by orders of magnitude compared to static networks 21 . This raises a key question: can temporality still offer advantages in the absence of knowledge about future changes? We demonstrate that for a specific class of control problems–herding–temporality can significantly reduce control energy, even without prior knowledge of topology changes. While traditional controllability in the context of complex networks involves the ability of the controller in steering the system from any initial state to any desired state within the state space 8 , 62 , herdability focuses on guiding all agents (nodes) to a fixed consensus state along a predefined trajectory 54 . Our analysis reveals that indecisive collective–stochastic temporal networks with restructuring timescales equal to system dynamics timescales ( 𝒯 = 1 )–are optimal for minimizing control energy. This finding offers a new perspective on leveraging temporality for efficient control of noisy living and robotic swarms, even in the absence of topology foresight. Broader Implications and Future Directions Without external stimuli (e.g., a dog or a handler), our indecisive model extends a general stochastic framework widely applied across diverse systems, including auto-catalytic biochemical reactions 63 , heterogeneous cancer cell populations 27 , collective animal movement 10 , 11 , 28 , and human opinion dynamics 64 (SI Section 1, Table S1). By introducing the concept of an external controller, or “shepherd,” our analysis establishes a foundational framework for controlling noisy groups in a variety of domains. For instance, Zajdel et al. 65 demonstrated a shepherd-dog-inspired mechanism to guide cells along specified trajectories, highlighting the potential for shepherding strategies in cellular systems. Building on these insights, our framework could guide the design of effective control mechanisms to herd and sort heterogeneous cell collectives. Such strategies hold promise for applications like promoting wound healing through coordinated cell movement or selectively isolating healthy cells from infected populations. More broadly, our approach bridges seemingly disparate fields, providing a foundation for algorithms capable of effectively controlling stochastic, indecisive swarms. While we presented a simplified model to explore the effects of sheep indecisiveness in sheep-dog-handler interactions, the real-world dynamics of this system are far more intricate. Shepherd dogs can instinctively predict sheep movements, but expertly trained dogs uniquely integrate instinct with handler commands to achieve precise coordination. In successful teams, the handler and dog operate cohesively, eliminating the need for constant monitoring. Instead, they function as a unified entity, sharing cognitive resources to analyze and anticipate the sheep’s behavior in real time 34 , 35 , 66 . Systematically studying these interactions, spanning verbal, physical, and visual modalities, could reveal more rich complexities hidden in these multi-species control dynamics, and offering insights into principles of decentralized and stochastic collective control."
} | 4,883 |
38947931 | PMC11213128 | pmc | 3,607 | {
"abstract": "Controlling stochastic temporal networks remains an open challenge in control theory. While predictable temporal networks with known future dynamics enhance controllability, real-world networks often exhibit stochasticity and unpredictability, making control harder. Here, we investigate control mechanisms for stochastic temporal networks by analyzing how biological controllers, such as shepherd dogs, manage panicked flocks of sheep. We studied a century-old shepherding competition, the sheepdog trials, where small groups of sheep unpredictably switch between fleeing and following behaviors–effectively forming stochastic temporal networks. Unlike large, cohesive flocks, these small, indecisive flocks are difficult to control, yet skilled dog-handler teams excel at both herding and splitting them (shedding) on demand. Using a stochastic choice model to describe the sheep’s behavioral shifts, we found that trained dogs exploit stochastic indecisiveness, typically seen as an obstacle, as a control tool, enabling both herding and splitting of noisy groups of sheep. Building on these insights, we developed the Indecisive Swarm Algorithm (ISA) for artificial agents and benchmarked its performance against standard approaches, including the Averaging-Based Swarm Algorithm (ASA) and the Leader-Follower Swarm Algorithm (LFSA). ISA minimizes control energy in trajectory-following tasks and outperforms alternatives under noisy conditions. By framing these results within a stochastic temporal network perspective, we demonstrate that even probabilistic knowledge of future dynamics can enhance control efficiency in specific scenarios. These findings establish a framework for managing stochastic temporal networks with applications in noisy, behavior-switching animal collectives, swarm robotics, and opinion dynamics.",
"introduction": "Introduction Emergent collective dynamics, where simple local interactions give rise to complex global behaviors, govern a wide range of systems. Examples include swarm robotics 1 , animal collectives 2 , social networks such as opinion dynamics 3 , pedestrians’ movements 4 , and vehicular traffic 5 . Controlling these systems is challenging, as their behaviors often defy traditional control methods 6 – 8 . Unlike systems with predictable, linear dynamics, emergent systems are best described as complex networks that require multiscale strategies to address both the microscopic interactions between individual agents and the macroscopic patterns that emerge at the group level 8 . Most of these networked systems introduce additional complexity when individual agents (nodes) switch between different behaviors, leading to temporal restructuring in the network. Biological collectives and social interactions in humans serve as prime examples of such behavior switching 9 – 13 (see SI Section 1 and Table S1 for a full list of behavior-switching systems from ants and locusts to seals and humans). Carrier ants transporting cargo alternate roles between lifters and pullers based on their orientation and the nest’s position 9 , sheep in small flocks randomly switch between leading and following roles 12 , and during an epidemic outbreak, humans frequently switch between different interaction partners, facilitating spread of diseases 14 . These systems highlight the need for control strategies that account for the stochastic and context-dependent nature of individual behavior transitions and their cascading effects on the evolving temporal networks, where edges dynamically reorganize over time 15 – 17 . Recently, it has been shown that temporal restructuring can improve the controllability of a network 18 . Specifically, temporal networks require less time and less energetic cost to be controlled than their static counterparts 19 , 20 . This counterintuitive observation relies on the fact that the future dynamics of the network are predictable and are exploited in designing the controls in the previous steps. However, when the switching dynamics are stochastic and unpredictable, temporality can make the control process more energetically demanding compared to a static network 21 . Therefore, despite advances in control theory and swarm robotics 7 , 8 , managing the dynamics of stochastic temporal networks remains an open challenge, particularly in systems where individual agents exhibit behavior switching. Predator-prey systems provide a natural framework for studying the challenges of controlling such noisy networks with behavior-switching dynamics. For instance, flocks of starlings confuse raptors by transitioning between complex dynamic patterns. Similarly, large herds of wildebeests intermittently shift between selfish herding and solitary flight when confronted by predators like cheetahs. In response, predators, instead of complex control mechanisms, adopt simplified strategies like focusing on a fixed point in space rather than tracking individual prey 22 , 23 . This allows them to split vulnerable individuals before leveraging speed and agility to secure their target 24 – 26 . These examples suggest that effective control of stochastic temporal networks with behavior-switching individuals does not always require precise prediction of behavioral transitions. In this work, we analyze such control mechanisms by studying shepherd dogs managing small flocks of sheep in a competition called the sheep-dog trials. Two key features of these competitions make them model systems for investigating control mechanisms in stochastic temporal networks. First, during these trials, when threatened, panicked sheep oscillate unpredictably and indecisively between fleeing from the dog and following other sheep, forming a stochastic temporal network. Trained shepherd dogs are highly effective at managing these noisy flocks under fluctuating conditions (SI Video 1). Second, unlike interactions between predators and large herds of animals in the wild, the sheepdog trials competitions provide a controlled environment where the behavior-switching dynamics of the sheep can be observed, quantified, and analyzed (see SI Section 2&3 for history and competition rules). By bridging empirical observations with quantitative modeling to analyze various tasks in the sheepdog trials competition, we find that shepherd dogs utilize the behavior switching in sheep for herding and splitting (shedding) the flocks. Behavior switching dynamics have been previously studied in the context of animal collectives and human societies (voter models) using individual-based stochastic choice models 11 , 27 – 30 . In this work, we build on the existing framework to frame sheepdog trials as a control problem for “indecisive collectives” — systems where agents stochastically alternate between different behaviors and interaction partners in the presence of an external agent. This paper is structured as follows: We begin by exploring the nuances and rules of sheepdog trials. Next, we present a stochastic framework to develop quantitative metrics such as “pressure” and “lightness” that capture the nuanced behavior of sheep. The framework is based on qualitative insights from experienced handlers, and empirical data on sheep-dog dynamics. We then present a stochastic choice model and the master equation to model the indecisive transitions in sheep movement, comparing our model’s predictions with observed dynamics. Building on this, we investigate whether sheep indecisiveness could benefit the dog. Our findings reveal that stochastic indecisiveness can aid the dog in both herding and shedding tasks. Finally, we extend our analysis to develop the Indecisive Swarm Algorithm (ISA), a swarm control strategy inspired by shepherding dynamics. By modeling ISA as a non-reciprocal stochastic temporal network and comparing it against the standard Averaging-Based Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA), we demonstrate that for specific control tasks like herding, ISA minimizes control energy requirements.",
"discussion": "Discussion Summary This study investigates control mechanisms for noisy, indecisive collectives, using sheepdog trials as a model system. These trials challenge trained shepherd dogs to herd and shed (split) small flocks of sheep ( N s ≤ 5), where the dynamics differ markedly from larger flocks. Unlike the cohesive selfish herd behavior seen in large groups, sheep in small flocks stochastically transition between fleeing (solitary behavior) and following the group (collective behavior), making them harder to control (i.e., an indecisive herd). By combining qualitative insights from expert dog handlers with a stochastic modeling framework, we analyze how trained dogs manage these indecisive sheep collectives. We find that sheep behavior depends on two key factors: the dog’s threat level and the sheep’s switching dynamics. Within the shepherding community, these factors are encapsulated by the terms “pressure” (the dog’s threat) and “lightness” (the isotropy of the sheep’s responsiveness). Light and heavy sheep exhibit distinct behaviors during herding and shedding tasks. To translate this nuanced qualitative knowledge into a quantitative framework, we developed a stochastic model to describe indecisive sheep behavior. The model reveals that trained dogs employ a two-step control strategy: first aligning stationary sheep to a desired orientation (orientation step) before increasing threat to initiate movement (movement step). Focusing on the orientation step, we modeled sheep as stationary agents that reorient stochastically. This analysis formalized the concepts of pressure and lightness, confirming their utility as core descriptors of sheep behavior. Comparing the model to data from actual sheepdog trials, we find that high isotropy aids group cohesion (for herding) but complicates splitting, while the dynamics of indecisive sheep are largely governed by the two parameters, pressure and lightness. We also investigated whether indecisiveness benefits the controller rather than solely posing a challenge. Extending our framework to simulate both orientation and movement steps in a 2D arena, we compared indecisive sheep agents with standard averaging-based Vicsek-type agents. While averaging agents outperform indecisive agents under low noise conditions, the reverse is true at higher noise levels. For shedding tasks, averaging agents consistently fail to split, while indecisive agents shed easily, irrespective of noise levels. These results highlight how trained dogs exploit the sheep indecisiveness as a tool and underscore the importance of the two-step control process. Finally, we explored whether indecisiveness could improve control strategies in artificial systems. Developing the Indecisive Swarm Algorithm (ISA), we compared it against the Averaging-based Swarm Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA) in a trajectory-following task. ISA agents successfully followed predefined trajectories at low stimulus intensities from the controller, unlike ASA and LFSA agents, which deviated significantly. Framing swarm algorithms as stochastic temporal networks, we identified two tunable timescales: the dynamics update timescale ( τ d ) and network restructuring timescale ( τ n ). By defining temporality 𝒯 = τ d / τ n , we showed that adjusting 𝒯 reproduces all three algorithms: ASA ( 𝒯 → ∞ ) , L F S A ( 𝒯 → 0 ) , and ISA ( 𝒯 = 1 ). Borrowing the concept of control energy from control theory, we quantified the stimulus intensity required to steer a swarm. ISA required the least control energy, demonstrating its effectiveness in herding noisy swarms. Our findings reveal the counterintuitive advantages of indecisiveness in controlling noisy collectives, with applications ranging from sheepdog trials to artificial swarms. By introducing deliberate indecisiveness, controllers can enhance their ability to perform complex tasks, such as herding and splitting, while also reducing effort in simpler tasks like trajectory-following. Why Sheepdog Trials are Challenging If indecisive agents require less control effort, why are sheepdog trials considered so challenging? To address this, we extended our indecisive model to large group sizes (SI Section 8). While the model was originally designed to explain the behavior of small groups ( N s ≤ 5) in response to external stimuli, its extension to larger group sizes captures dynamics consistent with known sheep behaviors. This broader application allowed us to propose a unified phase diagram for indecisive behavior (see SI Section 8 for details), offering insights into transitions between different behavioral regimes as group size and stimulus specificity change. The phase diagram ( Figure 8 ) illustrates the likelihood of individuals being influenced by controlling stimuli ( α ), intragroup interactions ( γ ), or random noise (non-specific stimulus) ( ε ) . Stimulus specificity, defined as the ratio of α / ε , measures the strength of external stimuli to noise. External stimuli, such as a dog’s pressure or the departure of an informed sheep, are key factors driving transitions between behaviors. We identify three distinct behavioral regimes: flocking (red), dominated by intra-group interaction, resulting in cohesive group behavior; fleeing (blue), dominated by specific stimuli where individuals act independently, ignoring the group; and grazing (green), dominated by random noise, with individuals disregarding both specific stimuli and the group. In small groups, increasing stimulus specificity shifts behavior from grazing to fleeing. In larger groups, flocking dominates under typical stimulus intensities. However, when stimulus specificity becomes extremely high - such as during a predator attack or an encounter with an untrained dog - the flocking phase transitions to fleeing, even in large groups ( Figure 8 ). We validated our model’s predictions by comparing them with prior empirical studies of sheep behavior. King et al. 32 (circle) observed that intermediate-sized groups (46 sheep) exhibited selfish herd behavior under high stimulus specificity, with herding dogs inducing cohesion. Toulet et al. 60 (square) found that when a trained sheep departs intermediate-sized groups (8-32 sheep), the group reaches a consensus to follow or ignore the individual, demonstrating the dominance of intra-group interactions even under mild stimuli (low specificity). Ginelli et al. 61 and Gomez-Nava et al. 12 (star and triangle) studied group dynamics without external stimuli. Ginelli focused on large groups (100 sheep), while Gomez-Nava examined small groups (4 sheep). Both identifyied intermittent grazing and flocking epochs, aligning with the grazing-flocking transition boundary in our model. These behaviors suggest an evolutionary anticipation of external threats as a defense mechanism. Our model (red line) predicts that small groups transition from grazing to uncontrolled fleeing through a narrow flocking phase as external stimulus increases. This prediction explains why managing small flocks is particularly difficult in sheep-dog trials. Since individual sheep vary in their responsiveness to stimuli, effectively herding or splitting small flocks requires the dog to balance intra-group cohesion with individual responsiveness, as excessive stimulus risks triggering chaotic fleeing. This underscores the complexity of controlling small, indecisive collectives, where behavioral transitions depend on a delicate interplay of external stimuli, noise, and group interactions. Temporality and Indecisiveness Temporal networks have been shown to require significantly less control energy than static networks 18 . This efficiency arises from their ability to leverage changing topologies to exploit favorable configurations, thereby reducing the need to counteract unfavorable system dynamics. In contrast, static networks, with their fixed structures, often force controllers to expend substantial energy to navigate energetically costly directions or to overcome inherent system dynamics. A useful analogy is sailing, where adjusting the sail to align with shifting wind directions enhances efficiency, rather than struggling against them 18 . However, this framework assumes that the controller has prior knowledge of future topology changes. Without such foresight, temporality can actually increase control energy by orders of magnitude compared to static networks 21 . This raises a key question: can temporality still offer advantages in the absence of knowledge about future changes? We demonstrate that for a specific class of control problems–herding–temporality can significantly reduce control energy, even without prior knowledge of topology changes. While traditional controllability in the context of complex networks involves the ability of the controller in steering the system from any initial state to any desired state within the state space 8 , 62 , herdability focuses on guiding all agents (nodes) to a fixed consensus state along a predefined trajectory 54 . Our analysis reveals that indecisive collective–stochastic temporal networks with restructuring timescales equal to system dynamics timescales ( 𝒯 = 1 )–are optimal for minimizing control energy. This finding offers a new perspective on leveraging temporality for efficient control of noisy living and robotic swarms, even in the absence of topology foresight. Broader Implications and Future Directions Without external stimuli (e.g., a dog or a handler), our indecisive model extends a general stochastic framework widely applied across diverse systems, including auto-catalytic biochemical reactions 63 , heterogeneous cancer cell populations 27 , collective animal movement 10 , 11 , 28 , and human opinion dynamics 64 (SI Section 1, Table S1). By introducing the concept of an external controller, or “shepherd,” our analysis establishes a foundational framework for controlling noisy groups in a variety of domains. For instance, Zajdel et al. 65 demonstrated a shepherd-dog-inspired mechanism to guide cells along specified trajectories, highlighting the potential for shepherding strategies in cellular systems. Building on these insights, our framework could guide the design of effective control mechanisms to herd and sort heterogeneous cell collectives. Such strategies hold promise for applications like promoting wound healing through coordinated cell movement or selectively isolating healthy cells from infected populations. More broadly, our approach bridges seemingly disparate fields, providing a foundation for algorithms capable of effectively controlling stochastic, indecisive swarms. While we presented a simplified model to explore the effects of sheep indecisiveness in sheep-dog-handler interactions, the real-world dynamics of this system are far more intricate. Shepherd dogs can instinctively predict sheep movements, but expertly trained dogs uniquely integrate instinct with handler commands to achieve precise coordination. In successful teams, the handler and dog operate cohesively, eliminating the need for constant monitoring. Instead, they function as a unified entity, sharing cognitive resources to analyze and anticipate the sheep’s behavior in real time 34 , 35 , 66 . Systematically studying these interactions, spanning verbal, physical, and visual modalities, could reveal more rich complexities hidden in these multi-species control dynamics, and offering insights into principles of decentralized and stochastic collective control."
} | 4,883 |
38947931 | PMC11213128 | pmc | 3,608 | {
"abstract": "Controlling stochastic temporal networks remains an open challenge in control theory. While predictable temporal networks with known future dynamics enhance controllability, real-world networks often exhibit stochasticity and unpredictability, making control harder. Here, we investigate control mechanisms for stochastic temporal networks by analyzing how biological controllers, such as shepherd dogs, manage panicked flocks of sheep. We studied a century-old shepherding competition, the sheepdog trials, where small groups of sheep unpredictably switch between fleeing and following behaviors–effectively forming stochastic temporal networks. Unlike large, cohesive flocks, these small, indecisive flocks are difficult to control, yet skilled dog-handler teams excel at both herding and splitting them (shedding) on demand. Using a stochastic choice model to describe the sheep’s behavioral shifts, we found that trained dogs exploit stochastic indecisiveness, typically seen as an obstacle, as a control tool, enabling both herding and splitting of noisy groups of sheep. Building on these insights, we developed the Indecisive Swarm Algorithm (ISA) for artificial agents and benchmarked its performance against standard approaches, including the Averaging-Based Swarm Algorithm (ASA) and the Leader-Follower Swarm Algorithm (LFSA). ISA minimizes control energy in trajectory-following tasks and outperforms alternatives under noisy conditions. By framing these results within a stochastic temporal network perspective, we demonstrate that even probabilistic knowledge of future dynamics can enhance control efficiency in specific scenarios. These findings establish a framework for managing stochastic temporal networks with applications in noisy, behavior-switching animal collectives, swarm robotics, and opinion dynamics.",
"introduction": "Introduction Emergent collective dynamics, where simple local interactions give rise to complex global behaviors, govern a wide range of systems. Examples include swarm robotics 1 , animal collectives 2 , social networks such as opinion dynamics 3 , pedestrians’ movements 4 , and vehicular traffic 5 . Controlling these systems is challenging, as their behaviors often defy traditional control methods 6 – 8 . Unlike systems with predictable, linear dynamics, emergent systems are best described as complex networks that require multiscale strategies to address both the microscopic interactions between individual agents and the macroscopic patterns that emerge at the group level 8 . Most of these networked systems introduce additional complexity when individual agents (nodes) switch between different behaviors, leading to temporal restructuring in the network. Biological collectives and social interactions in humans serve as prime examples of such behavior switching 9 – 13 (see SI Section 1 and Table S1 for a full list of behavior-switching systems from ants and locusts to seals and humans). Carrier ants transporting cargo alternate roles between lifters and pullers based on their orientation and the nest’s position 9 , sheep in small flocks randomly switch between leading and following roles 12 , and during an epidemic outbreak, humans frequently switch between different interaction partners, facilitating spread of diseases 14 . These systems highlight the need for control strategies that account for the stochastic and context-dependent nature of individual behavior transitions and their cascading effects on the evolving temporal networks, where edges dynamically reorganize over time 15 – 17 . Recently, it has been shown that temporal restructuring can improve the controllability of a network 18 . Specifically, temporal networks require less time and less energetic cost to be controlled than their static counterparts 19 , 20 . This counterintuitive observation relies on the fact that the future dynamics of the network are predictable and are exploited in designing the controls in the previous steps. However, when the switching dynamics are stochastic and unpredictable, temporality can make the control process more energetically demanding compared to a static network 21 . Therefore, despite advances in control theory and swarm robotics 7 , 8 , managing the dynamics of stochastic temporal networks remains an open challenge, particularly in systems where individual agents exhibit behavior switching. Predator-prey systems provide a natural framework for studying the challenges of controlling such noisy networks with behavior-switching dynamics. For instance, flocks of starlings confuse raptors by transitioning between complex dynamic patterns. Similarly, large herds of wildebeests intermittently shift between selfish herding and solitary flight when confronted by predators like cheetahs. In response, predators, instead of complex control mechanisms, adopt simplified strategies like focusing on a fixed point in space rather than tracking individual prey 22 , 23 . This allows them to split vulnerable individuals before leveraging speed and agility to secure their target 24 – 26 . These examples suggest that effective control of stochastic temporal networks with behavior-switching individuals does not always require precise prediction of behavioral transitions. In this work, we analyze such control mechanisms by studying shepherd dogs managing small flocks of sheep in a competition called the sheep-dog trials. Two key features of these competitions make them model systems for investigating control mechanisms in stochastic temporal networks. First, during these trials, when threatened, panicked sheep oscillate unpredictably and indecisively between fleeing from the dog and following other sheep, forming a stochastic temporal network. Trained shepherd dogs are highly effective at managing these noisy flocks under fluctuating conditions (SI Video 1). Second, unlike interactions between predators and large herds of animals in the wild, the sheepdog trials competitions provide a controlled environment where the behavior-switching dynamics of the sheep can be observed, quantified, and analyzed (see SI Section 2&3 for history and competition rules). By bridging empirical observations with quantitative modeling to analyze various tasks in the sheepdog trials competition, we find that shepherd dogs utilize the behavior switching in sheep for herding and splitting (shedding) the flocks. Behavior switching dynamics have been previously studied in the context of animal collectives and human societies (voter models) using individual-based stochastic choice models 11 , 27 – 30 . In this work, we build on the existing framework to frame sheepdog trials as a control problem for “indecisive collectives” — systems where agents stochastically alternate between different behaviors and interaction partners in the presence of an external agent. This paper is structured as follows: We begin by exploring the nuances and rules of sheepdog trials. Next, we present a stochastic framework to develop quantitative metrics such as “pressure” and “lightness” that capture the nuanced behavior of sheep. The framework is based on qualitative insights from experienced handlers, and empirical data on sheep-dog dynamics. We then present a stochastic choice model and the master equation to model the indecisive transitions in sheep movement, comparing our model’s predictions with observed dynamics. Building on this, we investigate whether sheep indecisiveness could benefit the dog. Our findings reveal that stochastic indecisiveness can aid the dog in both herding and shedding tasks. Finally, we extend our analysis to develop the Indecisive Swarm Algorithm (ISA), a swarm control strategy inspired by shepherding dynamics. By modeling ISA as a non-reciprocal stochastic temporal network and comparing it against the standard Averaging-Based Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA), we demonstrate that for specific control tasks like herding, ISA minimizes control energy requirements.",
"discussion": "Discussion Summary This study investigates control mechanisms for noisy, indecisive collectives, using sheepdog trials as a model system. These trials challenge trained shepherd dogs to herd and shed (split) small flocks of sheep ( N s ≤ 5), where the dynamics differ markedly from larger flocks. Unlike the cohesive selfish herd behavior seen in large groups, sheep in small flocks stochastically transition between fleeing (solitary behavior) and following the group (collective behavior), making them harder to control (i.e., an indecisive herd). By combining qualitative insights from expert dog handlers with a stochastic modeling framework, we analyze how trained dogs manage these indecisive sheep collectives. We find that sheep behavior depends on two key factors: the dog’s threat level and the sheep’s switching dynamics. Within the shepherding community, these factors are encapsulated by the terms “pressure” (the dog’s threat) and “lightness” (the isotropy of the sheep’s responsiveness). Light and heavy sheep exhibit distinct behaviors during herding and shedding tasks. To translate this nuanced qualitative knowledge into a quantitative framework, we developed a stochastic model to describe indecisive sheep behavior. The model reveals that trained dogs employ a two-step control strategy: first aligning stationary sheep to a desired orientation (orientation step) before increasing threat to initiate movement (movement step). Focusing on the orientation step, we modeled sheep as stationary agents that reorient stochastically. This analysis formalized the concepts of pressure and lightness, confirming their utility as core descriptors of sheep behavior. Comparing the model to data from actual sheepdog trials, we find that high isotropy aids group cohesion (for herding) but complicates splitting, while the dynamics of indecisive sheep are largely governed by the two parameters, pressure and lightness. We also investigated whether indecisiveness benefits the controller rather than solely posing a challenge. Extending our framework to simulate both orientation and movement steps in a 2D arena, we compared indecisive sheep agents with standard averaging-based Vicsek-type agents. While averaging agents outperform indecisive agents under low noise conditions, the reverse is true at higher noise levels. For shedding tasks, averaging agents consistently fail to split, while indecisive agents shed easily, irrespective of noise levels. These results highlight how trained dogs exploit the sheep indecisiveness as a tool and underscore the importance of the two-step control process. Finally, we explored whether indecisiveness could improve control strategies in artificial systems. Developing the Indecisive Swarm Algorithm (ISA), we compared it against the Averaging-based Swarm Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA) in a trajectory-following task. ISA agents successfully followed predefined trajectories at low stimulus intensities from the controller, unlike ASA and LFSA agents, which deviated significantly. Framing swarm algorithms as stochastic temporal networks, we identified two tunable timescales: the dynamics update timescale ( τ d ) and network restructuring timescale ( τ n ). By defining temporality 𝒯 = τ d / τ n , we showed that adjusting 𝒯 reproduces all three algorithms: ASA ( 𝒯 → ∞ ) , L F S A ( 𝒯 → 0 ) , and ISA ( 𝒯 = 1 ). Borrowing the concept of control energy from control theory, we quantified the stimulus intensity required to steer a swarm. ISA required the least control energy, demonstrating its effectiveness in herding noisy swarms. Our findings reveal the counterintuitive advantages of indecisiveness in controlling noisy collectives, with applications ranging from sheepdog trials to artificial swarms. By introducing deliberate indecisiveness, controllers can enhance their ability to perform complex tasks, such as herding and splitting, while also reducing effort in simpler tasks like trajectory-following. Why Sheepdog Trials are Challenging If indecisive agents require less control effort, why are sheepdog trials considered so challenging? To address this, we extended our indecisive model to large group sizes (SI Section 8). While the model was originally designed to explain the behavior of small groups ( N s ≤ 5) in response to external stimuli, its extension to larger group sizes captures dynamics consistent with known sheep behaviors. This broader application allowed us to propose a unified phase diagram for indecisive behavior (see SI Section 8 for details), offering insights into transitions between different behavioral regimes as group size and stimulus specificity change. The phase diagram ( Figure 8 ) illustrates the likelihood of individuals being influenced by controlling stimuli ( α ), intragroup interactions ( γ ), or random noise (non-specific stimulus) ( ε ) . Stimulus specificity, defined as the ratio of α / ε , measures the strength of external stimuli to noise. External stimuli, such as a dog’s pressure or the departure of an informed sheep, are key factors driving transitions between behaviors. We identify three distinct behavioral regimes: flocking (red), dominated by intra-group interaction, resulting in cohesive group behavior; fleeing (blue), dominated by specific stimuli where individuals act independently, ignoring the group; and grazing (green), dominated by random noise, with individuals disregarding both specific stimuli and the group. In small groups, increasing stimulus specificity shifts behavior from grazing to fleeing. In larger groups, flocking dominates under typical stimulus intensities. However, when stimulus specificity becomes extremely high - such as during a predator attack or an encounter with an untrained dog - the flocking phase transitions to fleeing, even in large groups ( Figure 8 ). We validated our model’s predictions by comparing them with prior empirical studies of sheep behavior. King et al. 32 (circle) observed that intermediate-sized groups (46 sheep) exhibited selfish herd behavior under high stimulus specificity, with herding dogs inducing cohesion. Toulet et al. 60 (square) found that when a trained sheep departs intermediate-sized groups (8-32 sheep), the group reaches a consensus to follow or ignore the individual, demonstrating the dominance of intra-group interactions even under mild stimuli (low specificity). Ginelli et al. 61 and Gomez-Nava et al. 12 (star and triangle) studied group dynamics without external stimuli. Ginelli focused on large groups (100 sheep), while Gomez-Nava examined small groups (4 sheep). Both identifyied intermittent grazing and flocking epochs, aligning with the grazing-flocking transition boundary in our model. These behaviors suggest an evolutionary anticipation of external threats as a defense mechanism. Our model (red line) predicts that small groups transition from grazing to uncontrolled fleeing through a narrow flocking phase as external stimulus increases. This prediction explains why managing small flocks is particularly difficult in sheep-dog trials. Since individual sheep vary in their responsiveness to stimuli, effectively herding or splitting small flocks requires the dog to balance intra-group cohesion with individual responsiveness, as excessive stimulus risks triggering chaotic fleeing. This underscores the complexity of controlling small, indecisive collectives, where behavioral transitions depend on a delicate interplay of external stimuli, noise, and group interactions. Temporality and Indecisiveness Temporal networks have been shown to require significantly less control energy than static networks 18 . This efficiency arises from their ability to leverage changing topologies to exploit favorable configurations, thereby reducing the need to counteract unfavorable system dynamics. In contrast, static networks, with their fixed structures, often force controllers to expend substantial energy to navigate energetically costly directions or to overcome inherent system dynamics. A useful analogy is sailing, where adjusting the sail to align with shifting wind directions enhances efficiency, rather than struggling against them 18 . However, this framework assumes that the controller has prior knowledge of future topology changes. Without such foresight, temporality can actually increase control energy by orders of magnitude compared to static networks 21 . This raises a key question: can temporality still offer advantages in the absence of knowledge about future changes? We demonstrate that for a specific class of control problems–herding–temporality can significantly reduce control energy, even without prior knowledge of topology changes. While traditional controllability in the context of complex networks involves the ability of the controller in steering the system from any initial state to any desired state within the state space 8 , 62 , herdability focuses on guiding all agents (nodes) to a fixed consensus state along a predefined trajectory 54 . Our analysis reveals that indecisive collective–stochastic temporal networks with restructuring timescales equal to system dynamics timescales ( 𝒯 = 1 )–are optimal for minimizing control energy. This finding offers a new perspective on leveraging temporality for efficient control of noisy living and robotic swarms, even in the absence of topology foresight. Broader Implications and Future Directions Without external stimuli (e.g., a dog or a handler), our indecisive model extends a general stochastic framework widely applied across diverse systems, including auto-catalytic biochemical reactions 63 , heterogeneous cancer cell populations 27 , collective animal movement 10 , 11 , 28 , and human opinion dynamics 64 (SI Section 1, Table S1). By introducing the concept of an external controller, or “shepherd,” our analysis establishes a foundational framework for controlling noisy groups in a variety of domains. For instance, Zajdel et al. 65 demonstrated a shepherd-dog-inspired mechanism to guide cells along specified trajectories, highlighting the potential for shepherding strategies in cellular systems. Building on these insights, our framework could guide the design of effective control mechanisms to herd and sort heterogeneous cell collectives. Such strategies hold promise for applications like promoting wound healing through coordinated cell movement or selectively isolating healthy cells from infected populations. More broadly, our approach bridges seemingly disparate fields, providing a foundation for algorithms capable of effectively controlling stochastic, indecisive swarms. While we presented a simplified model to explore the effects of sheep indecisiveness in sheep-dog-handler interactions, the real-world dynamics of this system are far more intricate. Shepherd dogs can instinctively predict sheep movements, but expertly trained dogs uniquely integrate instinct with handler commands to achieve precise coordination. In successful teams, the handler and dog operate cohesively, eliminating the need for constant monitoring. Instead, they function as a unified entity, sharing cognitive resources to analyze and anticipate the sheep’s behavior in real time 34 , 35 , 66 . Systematically studying these interactions, spanning verbal, physical, and visual modalities, could reveal more rich complexities hidden in these multi-species control dynamics, and offering insights into principles of decentralized and stochastic collective control."
} | 4,883 |
38947931 | PMC11213128 | pmc | 3,608 | {
"abstract": "Controlling stochastic temporal networks remains an open challenge in control theory. While predictable temporal networks with known future dynamics enhance controllability, real-world networks often exhibit stochasticity and unpredictability, making control harder. Here, we investigate control mechanisms for stochastic temporal networks by analyzing how biological controllers, such as shepherd dogs, manage panicked flocks of sheep. We studied a century-old shepherding competition, the sheepdog trials, where small groups of sheep unpredictably switch between fleeing and following behaviors–effectively forming stochastic temporal networks. Unlike large, cohesive flocks, these small, indecisive flocks are difficult to control, yet skilled dog-handler teams excel at both herding and splitting them (shedding) on demand. Using a stochastic choice model to describe the sheep’s behavioral shifts, we found that trained dogs exploit stochastic indecisiveness, typically seen as an obstacle, as a control tool, enabling both herding and splitting of noisy groups of sheep. Building on these insights, we developed the Indecisive Swarm Algorithm (ISA) for artificial agents and benchmarked its performance against standard approaches, including the Averaging-Based Swarm Algorithm (ASA) and the Leader-Follower Swarm Algorithm (LFSA). ISA minimizes control energy in trajectory-following tasks and outperforms alternatives under noisy conditions. By framing these results within a stochastic temporal network perspective, we demonstrate that even probabilistic knowledge of future dynamics can enhance control efficiency in specific scenarios. These findings establish a framework for managing stochastic temporal networks with applications in noisy, behavior-switching animal collectives, swarm robotics, and opinion dynamics.",
"introduction": "Introduction Emergent collective dynamics, where simple local interactions give rise to complex global behaviors, govern a wide range of systems. Examples include swarm robotics 1 , animal collectives 2 , social networks such as opinion dynamics 3 , pedestrians’ movements 4 , and vehicular traffic 5 . Controlling these systems is challenging, as their behaviors often defy traditional control methods 6 – 8 . Unlike systems with predictable, linear dynamics, emergent systems are best described as complex networks that require multiscale strategies to address both the microscopic interactions between individual agents and the macroscopic patterns that emerge at the group level 8 . Most of these networked systems introduce additional complexity when individual agents (nodes) switch between different behaviors, leading to temporal restructuring in the network. Biological collectives and social interactions in humans serve as prime examples of such behavior switching 9 – 13 (see SI Section 1 and Table S1 for a full list of behavior-switching systems from ants and locusts to seals and humans). Carrier ants transporting cargo alternate roles between lifters and pullers based on their orientation and the nest’s position 9 , sheep in small flocks randomly switch between leading and following roles 12 , and during an epidemic outbreak, humans frequently switch between different interaction partners, facilitating spread of diseases 14 . These systems highlight the need for control strategies that account for the stochastic and context-dependent nature of individual behavior transitions and their cascading effects on the evolving temporal networks, where edges dynamically reorganize over time 15 – 17 . Recently, it has been shown that temporal restructuring can improve the controllability of a network 18 . Specifically, temporal networks require less time and less energetic cost to be controlled than their static counterparts 19 , 20 . This counterintuitive observation relies on the fact that the future dynamics of the network are predictable and are exploited in designing the controls in the previous steps. However, when the switching dynamics are stochastic and unpredictable, temporality can make the control process more energetically demanding compared to a static network 21 . Therefore, despite advances in control theory and swarm robotics 7 , 8 , managing the dynamics of stochastic temporal networks remains an open challenge, particularly in systems where individual agents exhibit behavior switching. Predator-prey systems provide a natural framework for studying the challenges of controlling such noisy networks with behavior-switching dynamics. For instance, flocks of starlings confuse raptors by transitioning between complex dynamic patterns. Similarly, large herds of wildebeests intermittently shift between selfish herding and solitary flight when confronted by predators like cheetahs. In response, predators, instead of complex control mechanisms, adopt simplified strategies like focusing on a fixed point in space rather than tracking individual prey 22 , 23 . This allows them to split vulnerable individuals before leveraging speed and agility to secure their target 24 – 26 . These examples suggest that effective control of stochastic temporal networks with behavior-switching individuals does not always require precise prediction of behavioral transitions. In this work, we analyze such control mechanisms by studying shepherd dogs managing small flocks of sheep in a competition called the sheep-dog trials. Two key features of these competitions make them model systems for investigating control mechanisms in stochastic temporal networks. First, during these trials, when threatened, panicked sheep oscillate unpredictably and indecisively between fleeing from the dog and following other sheep, forming a stochastic temporal network. Trained shepherd dogs are highly effective at managing these noisy flocks under fluctuating conditions (SI Video 1). Second, unlike interactions between predators and large herds of animals in the wild, the sheepdog trials competitions provide a controlled environment where the behavior-switching dynamics of the sheep can be observed, quantified, and analyzed (see SI Section 2&3 for history and competition rules). By bridging empirical observations with quantitative modeling to analyze various tasks in the sheepdog trials competition, we find that shepherd dogs utilize the behavior switching in sheep for herding and splitting (shedding) the flocks. Behavior switching dynamics have been previously studied in the context of animal collectives and human societies (voter models) using individual-based stochastic choice models 11 , 27 – 30 . In this work, we build on the existing framework to frame sheepdog trials as a control problem for “indecisive collectives” — systems where agents stochastically alternate between different behaviors and interaction partners in the presence of an external agent. This paper is structured as follows: We begin by exploring the nuances and rules of sheepdog trials. Next, we present a stochastic framework to develop quantitative metrics such as “pressure” and “lightness” that capture the nuanced behavior of sheep. The framework is based on qualitative insights from experienced handlers, and empirical data on sheep-dog dynamics. We then present a stochastic choice model and the master equation to model the indecisive transitions in sheep movement, comparing our model’s predictions with observed dynamics. Building on this, we investigate whether sheep indecisiveness could benefit the dog. Our findings reveal that stochastic indecisiveness can aid the dog in both herding and shedding tasks. Finally, we extend our analysis to develop the Indecisive Swarm Algorithm (ISA), a swarm control strategy inspired by shepherding dynamics. By modeling ISA as a non-reciprocal stochastic temporal network and comparing it against the standard Averaging-Based Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA), we demonstrate that for specific control tasks like herding, ISA minimizes control energy requirements.",
"discussion": "Discussion Summary This study investigates control mechanisms for noisy, indecisive collectives, using sheepdog trials as a model system. These trials challenge trained shepherd dogs to herd and shed (split) small flocks of sheep ( N s ≤ 5), where the dynamics differ markedly from larger flocks. Unlike the cohesive selfish herd behavior seen in large groups, sheep in small flocks stochastically transition between fleeing (solitary behavior) and following the group (collective behavior), making them harder to control (i.e., an indecisive herd). By combining qualitative insights from expert dog handlers with a stochastic modeling framework, we analyze how trained dogs manage these indecisive sheep collectives. We find that sheep behavior depends on two key factors: the dog’s threat level and the sheep’s switching dynamics. Within the shepherding community, these factors are encapsulated by the terms “pressure” (the dog’s threat) and “lightness” (the isotropy of the sheep’s responsiveness). Light and heavy sheep exhibit distinct behaviors during herding and shedding tasks. To translate this nuanced qualitative knowledge into a quantitative framework, we developed a stochastic model to describe indecisive sheep behavior. The model reveals that trained dogs employ a two-step control strategy: first aligning stationary sheep to a desired orientation (orientation step) before increasing threat to initiate movement (movement step). Focusing on the orientation step, we modeled sheep as stationary agents that reorient stochastically. This analysis formalized the concepts of pressure and lightness, confirming their utility as core descriptors of sheep behavior. Comparing the model to data from actual sheepdog trials, we find that high isotropy aids group cohesion (for herding) but complicates splitting, while the dynamics of indecisive sheep are largely governed by the two parameters, pressure and lightness. We also investigated whether indecisiveness benefits the controller rather than solely posing a challenge. Extending our framework to simulate both orientation and movement steps in a 2D arena, we compared indecisive sheep agents with standard averaging-based Vicsek-type agents. While averaging agents outperform indecisive agents under low noise conditions, the reverse is true at higher noise levels. For shedding tasks, averaging agents consistently fail to split, while indecisive agents shed easily, irrespective of noise levels. These results highlight how trained dogs exploit the sheep indecisiveness as a tool and underscore the importance of the two-step control process. Finally, we explored whether indecisiveness could improve control strategies in artificial systems. Developing the Indecisive Swarm Algorithm (ISA), we compared it against the Averaging-based Swarm Algorithm (ASA) and Leader-Follower Swarm Algorithm (LFSA) in a trajectory-following task. ISA agents successfully followed predefined trajectories at low stimulus intensities from the controller, unlike ASA and LFSA agents, which deviated significantly. Framing swarm algorithms as stochastic temporal networks, we identified two tunable timescales: the dynamics update timescale ( τ d ) and network restructuring timescale ( τ n ). By defining temporality 𝒯 = τ d / τ n , we showed that adjusting 𝒯 reproduces all three algorithms: ASA ( 𝒯 → ∞ ) , L F S A ( 𝒯 → 0 ) , and ISA ( 𝒯 = 1 ). Borrowing the concept of control energy from control theory, we quantified the stimulus intensity required to steer a swarm. ISA required the least control energy, demonstrating its effectiveness in herding noisy swarms. Our findings reveal the counterintuitive advantages of indecisiveness in controlling noisy collectives, with applications ranging from sheepdog trials to artificial swarms. By introducing deliberate indecisiveness, controllers can enhance their ability to perform complex tasks, such as herding and splitting, while also reducing effort in simpler tasks like trajectory-following. Why Sheepdog Trials are Challenging If indecisive agents require less control effort, why are sheepdog trials considered so challenging? To address this, we extended our indecisive model to large group sizes (SI Section 8). While the model was originally designed to explain the behavior of small groups ( N s ≤ 5) in response to external stimuli, its extension to larger group sizes captures dynamics consistent with known sheep behaviors. This broader application allowed us to propose a unified phase diagram for indecisive behavior (see SI Section 8 for details), offering insights into transitions between different behavioral regimes as group size and stimulus specificity change. The phase diagram ( Figure 8 ) illustrates the likelihood of individuals being influenced by controlling stimuli ( α ), intragroup interactions ( γ ), or random noise (non-specific stimulus) ( ε ) . Stimulus specificity, defined as the ratio of α / ε , measures the strength of external stimuli to noise. External stimuli, such as a dog’s pressure or the departure of an informed sheep, are key factors driving transitions between behaviors. We identify three distinct behavioral regimes: flocking (red), dominated by intra-group interaction, resulting in cohesive group behavior; fleeing (blue), dominated by specific stimuli where individuals act independently, ignoring the group; and grazing (green), dominated by random noise, with individuals disregarding both specific stimuli and the group. In small groups, increasing stimulus specificity shifts behavior from grazing to fleeing. In larger groups, flocking dominates under typical stimulus intensities. However, when stimulus specificity becomes extremely high - such as during a predator attack or an encounter with an untrained dog - the flocking phase transitions to fleeing, even in large groups ( Figure 8 ). We validated our model’s predictions by comparing them with prior empirical studies of sheep behavior. King et al. 32 (circle) observed that intermediate-sized groups (46 sheep) exhibited selfish herd behavior under high stimulus specificity, with herding dogs inducing cohesion. Toulet et al. 60 (square) found that when a trained sheep departs intermediate-sized groups (8-32 sheep), the group reaches a consensus to follow or ignore the individual, demonstrating the dominance of intra-group interactions even under mild stimuli (low specificity). Ginelli et al. 61 and Gomez-Nava et al. 12 (star and triangle) studied group dynamics without external stimuli. Ginelli focused on large groups (100 sheep), while Gomez-Nava examined small groups (4 sheep). Both identifyied intermittent grazing and flocking epochs, aligning with the grazing-flocking transition boundary in our model. These behaviors suggest an evolutionary anticipation of external threats as a defense mechanism. Our model (red line) predicts that small groups transition from grazing to uncontrolled fleeing through a narrow flocking phase as external stimulus increases. This prediction explains why managing small flocks is particularly difficult in sheep-dog trials. Since individual sheep vary in their responsiveness to stimuli, effectively herding or splitting small flocks requires the dog to balance intra-group cohesion with individual responsiveness, as excessive stimulus risks triggering chaotic fleeing. This underscores the complexity of controlling small, indecisive collectives, where behavioral transitions depend on a delicate interplay of external stimuli, noise, and group interactions. Temporality and Indecisiveness Temporal networks have been shown to require significantly less control energy than static networks 18 . This efficiency arises from their ability to leverage changing topologies to exploit favorable configurations, thereby reducing the need to counteract unfavorable system dynamics. In contrast, static networks, with their fixed structures, often force controllers to expend substantial energy to navigate energetically costly directions or to overcome inherent system dynamics. A useful analogy is sailing, where adjusting the sail to align with shifting wind directions enhances efficiency, rather than struggling against them 18 . However, this framework assumes that the controller has prior knowledge of future topology changes. Without such foresight, temporality can actually increase control energy by orders of magnitude compared to static networks 21 . This raises a key question: can temporality still offer advantages in the absence of knowledge about future changes? We demonstrate that for a specific class of control problems–herding–temporality can significantly reduce control energy, even without prior knowledge of topology changes. While traditional controllability in the context of complex networks involves the ability of the controller in steering the system from any initial state to any desired state within the state space 8 , 62 , herdability focuses on guiding all agents (nodes) to a fixed consensus state along a predefined trajectory 54 . Our analysis reveals that indecisive collective–stochastic temporal networks with restructuring timescales equal to system dynamics timescales ( 𝒯 = 1 )–are optimal for minimizing control energy. This finding offers a new perspective on leveraging temporality for efficient control of noisy living and robotic swarms, even in the absence of topology foresight. Broader Implications and Future Directions Without external stimuli (e.g., a dog or a handler), our indecisive model extends a general stochastic framework widely applied across diverse systems, including auto-catalytic biochemical reactions 63 , heterogeneous cancer cell populations 27 , collective animal movement 10 , 11 , 28 , and human opinion dynamics 64 (SI Section 1, Table S1). By introducing the concept of an external controller, or “shepherd,” our analysis establishes a foundational framework for controlling noisy groups in a variety of domains. For instance, Zajdel et al. 65 demonstrated a shepherd-dog-inspired mechanism to guide cells along specified trajectories, highlighting the potential for shepherding strategies in cellular systems. Building on these insights, our framework could guide the design of effective control mechanisms to herd and sort heterogeneous cell collectives. Such strategies hold promise for applications like promoting wound healing through coordinated cell movement or selectively isolating healthy cells from infected populations. More broadly, our approach bridges seemingly disparate fields, providing a foundation for algorithms capable of effectively controlling stochastic, indecisive swarms. While we presented a simplified model to explore the effects of sheep indecisiveness in sheep-dog-handler interactions, the real-world dynamics of this system are far more intricate. Shepherd dogs can instinctively predict sheep movements, but expertly trained dogs uniquely integrate instinct with handler commands to achieve precise coordination. In successful teams, the handler and dog operate cohesively, eliminating the need for constant monitoring. Instead, they function as a unified entity, sharing cognitive resources to analyze and anticipate the sheep’s behavior in real time 34 , 35 , 66 . Systematically studying these interactions, spanning verbal, physical, and visual modalities, could reveal more rich complexities hidden in these multi-species control dynamics, and offering insights into principles of decentralized and stochastic collective control."
} | 4,883 |
25125642 | PMC4132618 | pmc | 3,609 | {
"abstract": "The benefits of using transgenic switchgrass with decreased levels of caffeic acid 3- O -methyltransferase (COMT) as biomass feedstock have been clearly demonstrated. However, its effect on the soil microbial community has not been assessed. Here we report metagenomic and metatranscriptomic analyses of root-associated soil from COMT switchgrass compared with nontransgenic counterparts."
} | 97 |
31459428 | PMC6648469 | pmc | 3,611 | {
"abstract": "Long-term sensing\nof dissolved oxygen in aqueous solution always suffers from adherence\nof algae, barnacles, and clams and formation of biofilms on the sensor\nsurface, which strongly influences the diffusion of oxygen into the\nsensor film. Metabolism of these adhered species consumes oxygen and\ncauses bias on sensor readout. Therefore, commercial sensors are equipped\nwith mechanical brushes to constantly clean the sensor surface, which\nsignificantly complicates the sensor design and causes damage to the\nsensor surface. In addition, extra energy storage and mechanical structures\nare required, which make an optical sensor bulky and limit its service\nlife. We have developed a robust and highly sensitive dissolved oxygen\nsensor with good mechanical stability and self-cleaning capability.\nThe sensor was fabricated by doping oxygen-sensitive probe PtTFPP\nwith superhydrophobic coating. The 3 to 5 nm micro/nanostructures\nformed from silica sol were solidified with silicone resin, which\nendowed the sensor film with excellent mechanical stability. The sensor\nfilm exhibits antifouling, antiabrasion, and self-cleaning properties.\nThere is no need of mechanical brushes to clean sensor surfaces, which\ngreatly simplifies the sensor design. Owing to the porous structure,\nthe sensor shows high quenchability, with I 0 / I 100 of 77. All these features guarantee\nthat the sensor could be used in harsh and dirty conditions for long-term\nmonitoring of dissolved oxygen concentration.",
"conclusion": "Conclusions We have developed a highly sensitive oxygen\nsensor film with good mechanical stability and antifouling and self-cleaning\nproperties. The sensor film was simply fabricated by doping the oxygen-sensitive\nprobe PtTFPP in superhydrophobic sol suspension and cast along with\nsilicone resin to form superhydrophobic micro/nanohierarchical structures.\nBecause of the unique porous structure, the sensor film shows fully\nreversible response toward oxygen quenching, and quenchability (expressed\nas I 0 / I 100 ) is as high as 77. The oxygen-sensitive probes are firmly encapsulated\nand uniformly distributed inside the hydrophobic coating so that they\ndo not leach and aggregate. Extensive wear-resistant tests (including\ncontinuous tape peeling, knife scratching, and sandpaper abrasions)\nprove that the sensor film has excellent mechanical stability. More\nimportantly, the sensor film shows excellent antifouling and self-cleaning\nproperties, on which the formation of biofilms and adherence of clams\nand other marine lives could be minimized. Because of its simple fabrication\nprocedure, this kind of superhydrophobic sensor film could be massively\nproduced. The antifouling and self-cleaning properties of the sensor\nfilm greatly simplifies sensor design. A compact sensor could be fabricated\nwithout using mechanical cleaning brushes, which significantly reduces\nsensor cost, maintenance frequency and cost, and extends sensor service\nlife. All these properties endow that the sensor film could be used\nin harsh and dirty conditions and for long-term monitoring of oxygen\nconcentration without worrying about the influences of surface contamination.",
"introduction": "Introduction Molecular oxygen is\nprobably one of the most important chemicals for life on earth. Continuously\nmeasuring dissolved oxygen concentration has been proven vital to\nmany industrial areas, including fermentation, dense fish-farming,\nlong-distance transportation of natural gases, biofuel production,\nbiomedical bioreactors, marine research and exploration, and water\nquality monitoring and pollution control, to name only the major applications. 1 − 3 In early time, dissolved oxygen is mainly measured using a Clark\nelectrode and a Mackereth-type electrode. 4 However, with decades of innovation, sensing of molecular oxygen\nbased on luminescence quenching has become one of the major methods\nof choice, simply because of its high sensitivity, ease of system\nsetup, no consumption of oxygen during measurement, and remote sensing.\nMore importantly, as a fierce competition to the classic Clark and\nMackereth-type electrodes, 5 luminescence\nquenching-based optical oxygen sensors present a unique feature that\nthey could be easily sterilized without worrying about the connecting\ncable. Such sensors could be glued inside the container, and signal\nis readout remotely outside. This becomes extremely important for\nfermentation and biomedical bioreactor applications, where complete\nsterilization is the key to prevent contamination. The concentration\nof dissolved oxygen has also been determined using Winkler titration\n(mostly in an environmental monitoring sector), which is a quantitative\nanalytical method based on chemical reaction. However, this method\nis not suitable for long-term and online monitoring of dissolved oxygen\nbecause of its irreversibility in nature. Optical sensors are\ncomposed of a luminescent sensor film, an optical fiber, and a signal\nanalysis and readout system. 1 The sensor\nfilm is the key component in optical oxygen sensors, and it always\ndirectly contacts with aqueous samples and senses variation of dissolved\noxygen concentration. This method has been successfully applied in\nlots of fields, but is mostly used in clean aqueous solutions. However,\nin real applications, there are microorganisms, algae, clams, and\nother marine lives in the testing samples. Long-term use of an optical\noxygen sensor film in these samples will always result in growth of\nalgae, formation of biofilm, and adherence of shellfish, 6 which will create a barrier for oxygen diffusion\nin the sensor film. In addition,\nmetabolism of these adhered marine species also influences the local\noxygen concentration nearby the sensor film, due to which the dissolved\noxygen concentration in the sample could not be correctly measured.\nIn order to prevent adherence of biofilms, algae, and shellfish on\nthe sensing film, commercially available oxygen sensors (especially\nsensors used in marine monitoring buoys) are equipped with brushes\nto clean the sensor surface (typical products used in buoys from YSI\nInc., www.ysi.com ). However,\nthe oxygen sensor materials are not wear-resistant. Continuous abrasion\nduring the brush-cleaning process damages the sensor film and significantly\nreduces sensor performance and service life, resulting in increases\nin maintenance frequency and costs. The addition of mechanical component\nto drive the brush also complicates sensor design, increases sensor\ncost, and makes the sensor bulky and heavy. The innovation in\nsuperhydrophobic materials provides extraordinary features that these\nmaterials could be used as coatings to create antifouling 7 and self-cleaning surfaces. 8 Superhydrophobic surfaces rely on the formation of a thin\nair layer that can decrease the attractive interactions between the\nsurface and the liquid and are inspired from natural nonwetting structures,\nparticularly the lotus leaf, 9 one of green,\nnontoxic, nonfouling surfaces. 8 , 10 , 11 However, the major challenges of artificial superhydrophobic surfaces\nfor practical applications are their poor mechanical robustness and\ndifficulties in massive production. The superhydrophobic surface can\nbe formed via the classic top-down, bottom-up, or the combination\nof both approaches, which require sophisticated fabrication and nanomaterial\nsynthesis process. 12 The assignment of\nmicro/nanostructures on the surface of interest requires precise control\nof the location at a microscopic scale, which strongly restricts the\nmassive production of the superhydrophobic surface. In addition, the\norderly arranged superhydrophobic surface based on the micro/nanostructures\nis fragile and is highly susceptible to mechanical wear. 13 − 16 These surfaces could be easily damaged during practical use. Notable\nprogress has been made over the past decade, and superhydrophobic\nsurface can now be formed via the robust sol–gel approach,\nwhich greatly simplifies the fabrication process. 12 However, majority of fabricating superhydrophobic surfaces\nstill rely on the adhesive layer to enhance its mechanical stability, 15 , 17 − 20 and such a two-layer system is undesirable during practical coating.\nChen and co-workers developed a large-scale fabrication method for\nrobust hydrophobic surface coating with high rigidity and good flexibility. 21 Superhydrophobic sol was first fabricated via\na one-pot reaction, which was then mixed with silicone resin and applied\ndirectly on the surface of interest via routine coating processes.\nThe fabricated superhydrophobic surface exhibits good flexibility,\nhigh rigidity, and good mechanical stability. Because of the\nsuperhydrophobic nature and porous structure, superhydrophobic coating\ncould be an ideal material for oxygen sensing because it exhibits\nexcellent oxygen permeability. However, only few attempts have been\nmade to use superhydrophobic coating for oxygen sensing. Mitsuishi\nand co-workers 22 covalently immobilized\nan oxygen-sensitive probe PtTPP on an amphiphilic acrylamide-based\ncopolymer and cast the copolymer with another fluoropolymer to form\na nanoparticle-based superhydrophobic surface. The fabricated oxygen\nsensor film exhibited high quenchability, with an I 0 / I 100 over 120, which could\nbe attributed that the probe is directly exposed to the surface and\nis in direct contact with molecular oxygen. However, fabrication of\nthe sensor requires tedious copolymer synthesis and dye labeling,\nand the nanoparticle-based superhydrophobic surface may easily be\ndamaged from mechanical abrasion. In this work, we have developed\na highly sensitive dissolved oxygen sensor with excellent mechanical\nstability and self-cleaning capability. The sensor is composed of\nmicro/nanohierarchical structured superhydrophobic coating that is\ndoped with an oxygen-sensitive probe. The hierarchical structure was\nthen solidified using silicone resin to endow the sensor film with\ngood mechanical stability. After tape-peeling, intensive scratching,\nand sandpaper abrasion, the sensor film still maintains good superhydrophobic\nnature with contact angle (CA) larger than 150°. Owing to the\nporous structure and high gas permeability of silicone resin, the\nsensor showed excellent quenchability toward oxygen quenching with I 0 / I 100 over 77.\nBecause of its superhydrophobic property, algae, biofilms, and clams\nare difficult to adhere on the surface of the sensing film. This feature\nkeeps the sensor surface clean over long-term usage, and there is\nno need to install cleaning brushes, which will make sensor design\nmuch simple, more energy-saving, and with largely extended service\nlife. Moreover, the fabrication of the sensor film is very simple,\nand there is no sophisticated synthesis and fabrication technique\ninvolved, which could be applied for massive production.",
"discussion": "Results and Discussion The preparation of the sensor film is very simple and straightforward,\nand there is no sophisticated preparation and nanofabrication involved.\nSilica nanoparticles with a size of 3–5 nm were synthesized\nvia a classic Stöber method, and their surface is then modified\nwith γ-aminopropyltriethoxysilane (APTES) to form aggregates.\nThese irregular shapes of aggregates form micro/nanohierarchical structures.\nThen, hydrophobic 1 H ,1 H ,2 H ,2 H -perfluorooctyltriethoxysilane (PFOTS)\nwas used to lower the surface energy. Finally, the oxygen-sensitive\ndye PtTFPP and the hydrophobic silica sol were mixed with silicone\nresin to form a viscous mixture. The mixture was then coated onto\na glass slide to form the sensor film. Figure 1 a shows three\ndrops of water on the sensor surface, which are perfect spherical\nwater droplets formed with a CA of 168°, illustrating the superhydrophobic\nnature of the sensor surface. Low surface energy and micro/nanostructures\nare two essential physical ingredients to prepare the superhydrophobic\nsurface. 16 Figure 1 c,d shows the scanning electron microscopy\n(SEM) images of the sensor film with small and large magnification,\nrespectively. It is obvious that numerous silica nanoparticles are\ngathered together to form the micro/nanohierarchical structures. The\nunique structure helps trapping a thin layer of air that decreases\nattractive interactions between the solid surface and liquid, leading\nto superhydrophobicity. Besides, the sensor also exhibits a fractal-like\nporous structure, which is also important to enhance the hydrophobicity. 23 In addition, the sensor surface shows a CA of\n145° for methylene iodide (Figure S1 in the Supporting Information ), revealing that the sensor surface\nhas low adherence to oil. The calculated surface-free energy was 0.54 ± 0.07 mN/m, indicating that the\nsensor could not only prevent contamination and adherence of marine\nlife, but also avoid adherence of organic pollutants. These properties\nwould be very useful when the sensor was applied for oxygen monitoring\nin the area where accidental oil spilling happens. Figure 1 (a) Photograph of three\nwater droplets on the sensor film. The doping of oxygen-sensitive\nprobe PtTFPP gives the sensor film a pink color. (b) Water droplets\non the sensor film exhibits a CA of 168.85°, indicating the superhydrophobic\nproperty of the sensor surface. (c,d) SEM images of the sensor surface\nwith low and high magnification; the sensor exhibits rough surface\nwith a micro/nanohierarchical structure, which is the basic framework\nfor the superhydrophobic surface. The silica sols were firmly immobilized\nby silicone resin, which guarantee the mechanical stability and preserve\nthe superhydrophobic property over long-term use at harsh conditions. Figure 2 shows the surface properties of the sensor\nfilm after continuous tape peeling, knife scratching, and sandpaper\nabrasions. The resistance to these external destructions, even in\nexceptionally harsh conditions, is important in real applications.\nThe current major challenge for practical application is its poor\nmechanical robustness. 15 , 16 To evaluate the mechanical robustness,\nwe first implemented the tape-peeling test for the sensor film. 24 As shown in Figure 2 a,b, after 10 cycles of tape peeling, ignorable\neffects in superhydrophobicity occurred. The CA of the sensor film\nslightly declines from 167.6° to 165.7°, and the sliding\nangle (SA) slightly increases from 2° to 4°. It seems that\nthe increases of SA are related to the decreases of CA. Continuing\nto 40 cycles, the sensor still reserves its superhydrophobicity. The\nSEM images of the sensor surface before and after 40 cycles of repetitive\ntape peeling indicate no obvious surface morphology change (see Figure\nS2 in the Supporting Information ). We also\nperformed a knife scratching test according to the ASTM standard.\nFigure S3 in the Supporting Information showed that the sensor was well adhered to the supporting matrix.\nEven after knife scratching with observable crack, the sensor surface\nmaintains its superhydrophobic property with a CA of 162°. We\nfurther challenged the sensor surface with a sandpaper abrasion test. 15 Figure 2 c shows the schematic of the sandpaper abrasion test; the\nsamples were placed face-down to the sandpaper (240#) under a loading\nof 100 g and moved along the ruler at a constant speed. A 10 cm abrasion\ntest was defined as one cycle. As summarized in Figure 2 d, the sensor preserves its superhydrophobic\nnature even after 40 cycles of sandpaper abrasion; the CAs were all\nabove 150°. All this information proves the good mechanical stability\nof the sensor film. Figure 2 (a) Illustration of the tape-peeling test with 4 kg of\nloading weight. (b) CAs and SAs measured after different cycles; the\nsensor surface maintains its superhydrophobic property even after\n40 cycles of tape peeling, indicating that the surface has good mechanical\nstability over tape peeling. (c) Scheme of one cycle of the sandpaper\nabrasion (240 grit) test with 100 g of loading weight and a moving\ndistance of 10 cm. (d) CAs of the sensor surface after continuous\nsandpaper abrasion. The superhydrophobic property was preserved after\nharsh sandpaper abrasion test, proving the great mechanical stability\nof the sensor surface. The superhydrophobic sensor surface also exhibits self-cleaning\ncapability. As shown in Figure 3 , the sensor film was placed with a tilting angle of 5°,\nand manganese dioxide power was put on the sensor surface as manual\ndirt. When one drop of water containing CuSO 4 (as colorant) slides\nover the MnO 2 powder, the powder dirt is removed from the\nsensor surface. There is no adherence and remains of dirt on the sensor\nsurface, demonstrating that the sensor film possesses good self-cleaning\ncapability. In comparison, the hydrophobic sensor surface without\nsuperhydrophobic property does not have self-cleaning capability,\nas shown in Figure S4 . Figure 3 Self-cleaning property\nof the sensor film. The sensor film was placed at a tilting angle\nof 5°, and MnO 2 powder was used as manual dirt. When\na drop of water containing CuSO 4 passes through the dirt,\nthe surface remains clean. We further studied the distribution of oxygen-sensitive probe\nPtTFPP in the sensor film. As shown in Figure S5 in the Supporting Information , elemental analysis and\nimaging reveal that the probes (Pt distribution) are evenly distributed\nin the superhydrophobic sensor film and there is no obvious aggregation\nobserved. More importantly, the sensor film is semitransparent, and\nthe rough surface could scatter excitation light, resulting more efficient\nexcitation of the oxygen-sensitive probe and enhanced brightness of\nthe sensor film. Figure 4 a shows the luminescence spectra of the sensor film at three different\ndissolved oxygen concentrations. It is obvious that molecular oxygen\ncould efficiently quench the luminescence of PtTFPP in the superhydrophobic\nfilm. The quenchability (expressed as I 0 / I 100 , which stands for the intensity\nmeasured in an oxygen-free solution and in a pure oxygen saturated\nsolution, respectively) of the sensor reaches 77. The time trace curve\nreveals that the sensor could measure dissolved oxygen concentration\nin the range of 0–20 mg/L, which is shown in Figure 4 b. The response time is less\nthan 6 s. (The response time is defined as the time required for the\nintensity to decrease to 90% of the equilibrium value. Note that the\nresponse time is measured in gas phase, as it is difficult to control\nthe filling time of the flow-through cell, Figure S6 in the Supporting Information .) Figure 4 (a) Emission spectra\nof oxygen sensors at three different dissolved oxygen concentrations.\nMolecular oxygen efficiently quenched the luminescence of PtTFPP in\nthe sensor matrix. (b) Response of sensor at different dissolved oxygen\nconcentrations and its reversibility. The sensors exhibit high quenchability\nand full reversibility; the luminescence intensity was monitored at\n650 nm. (c) Stern–Volmer plot of oxygen sensors as a function\nof dissolved oxygen concentration. (d) Photostability of oxygen sensors. The high quenchability and fast\nresponse could be attributed to the porous structure of the sensor\nfilm and good gas permeability of silicone. Such kind of porous silica\nstructure has been proved to be a good matrix for highly sensitive\noxygen sensors. 25 − 29 The sensor response is fully reversible, which guarantees monitoring\nof oxygen concentration variation over long term. The almost linear\nStern–Volmer plot shown in Figure 4 c further proves that the oxygen-sensitive\nprobes are uniformly distributed in the sensor film. The Stern–Volmer\nconstant is 2.65 L/mg. The linear calibration curve is beneficial\nfor two-point calibration and significantly simple sensor calibration\nprocess and data analysis. More importantly, the sensor shows\nexcellent photostability as shown in Figure 4 d. The sensor was exposed to intense illumination\nat 395 nm with a power density of 13.2 W/m 2 ; there is no\nmore than 10% of intensity loss after 1 h of exposure. In addition,\nthe sensor film is fully compatible with a commercial optical oxygen\nmeter, which uses much weaker illumination from light-emitting diodes\nand measures phase difference rather than luminescence intensity.\nThe influences of photobleaching on sensor performance could be easily\neliminated. All these features confirm that the sensor has good quenchability,\nfast response, and good stability, which is suitable for long-term\nmonitoring of dissolved oxygen. The sensor film was finally\ntested for its antifouling property. We have compared the adherence\nof Navicula pelliculosa on four different\nsurfaces, including polydimethylsiloxane (PDMS), polystyrene, polyethylene\nterephthalate (PET), and the superhydrophobic sensor film. All the\nfilms were immersed in a suspension of Navicula pelliculosa for 48 h and then their surfaces were washed with water. After drying\nat room temperature, the Navicula pelliculosa on the surface was imaged via autofluorescence of chlorophyll on\nthe fluorescence microscope. 30 The chlorophyll\nwas excited in the wavelength range of 595–645 nm, and the\nemission was recorded in the range of 665–715 nm. Figure 5 shows the overlay\nimages of bright-field image and autofluorescence image of Navicula pelliculosa . Results show that Navicula pelliculosa could adhere to all surfaces\ntested. Statistic results show that there are 40 ± 16.2, 1760\n± 165.1, 1346 ± 369.9, and 786 ± 136.8 Navicula pelliculosa per square millimeter attached\non the superhydrophobic sensor film, PDMS, polystyrene, and PET support,\nrespectively. Obviously, the superhydrophobic sensor film exhibits\nmuch less attachment of Navicula pelliculosa . Close observation could conclude that Navicula pelliculosa is mainly trapped in the hole of micro/nanostructure and could easily\nbe washed away with repetitive washing steps. However, the adhered Navicula pelliculosa on PDMS, polystyrene, and PET\nsurface attached firmly and could not be easily removed via a simple\nwashing step. These results further demonstrate that the superhydrophobic\nsensor film shows excellent antifouling property. Figure 5 Overlay images of bright-field\nimage and autofluorescence image of adherence of Navicula\npelliculosa on different surfaces after 48 h of incubation.\n(a) Superhydrophobic sensor film. (b) PDMS sensor film. (c) Polystyrene\nsensor film. (d) PET support."
} | 5,581 |
34739926 | null | s2 | 3,612 | {
"abstract": "Microbial communities are a key part to tackling global challenges in human health, environmental conservation, and sustainable agriculture in the coming decade. Recent advances in synthetic biology to study and modify microbial communities have led to important insights into their physiology and ecology. Understanding how targeted changes to microbial communities result in reproducible alterations of the community's intrinsic fluctuations and function is important for mechanistic reconstruction of microbiomes. Studies of synthetic microbial consortia and comparative analysis of communities in normal and disrupted states have revealed ecological principles that can be leveraged to engineer communities towards desired functions. Tools enabling temporal modulation and sensing of the community dynamics offer precise spatiotemporal control of functions, help to dissect microbial interaction networks, and improve predictions of population temporal dynamics. Here we discuss recent advances to manipulate microbiome dynamics through control of specific strain engraftment and abundance, modulation of cell-cell signaling for tuning population dynamics, infiltration of new functions in the existing community with in situ engineering, and in silico modeling of microbial consortia to predict community function and ecology."
} | 332 |
20715762 | null | s2 | 3,613 | {
"abstract": "Mechanical properties of tissue scaffolds have major effects on the morphology and differentiation of cells. In contrast to two-dimensional substrates, local biochemical and mechanical properties of three-dimensional hydrogels are difficult to control due to the geometrical confinement. We designed synthetic 3D hydrogels featuring complexes of four-arm poly(ethylene glycol) (PEG) and collagen mimetic peptides (CMPs) that form hydrogels via physical cross-links mediated by thermally reversible triple helical assembly of CMPs. Here we present the fabrication of various PEG-CMP 3D hydrogels and their local mechanical properties determined by particle tracking microrheology. Results show that CMP mediated physical cross-links can be disrupted by altering the temperature of the gel or by adding free CMPs that compete for triple helix formation. This allowed modulation of both bulk and local stiffness as well as the creation of stiffness gradients within the PEG-CMP hydrogel, which demonstrates its potential as a novel scaffold for encoding physicochemical signals for tissue formation."
} | 274 |
30270957 | null | s2 | 3,615 | {
"abstract": "Heterogeneous substrates with moderate and extreme wettability contrasts were fabricated by comprising of superhydrophobic/hydrophilic and superhydrophobic/extremely hydrophilic surfaces, respectively. The interactions of water droplets impinging on the surfaces with sharp wettability contrasts were investigated experimentally. The impinging droplets that slightly touch the hydrophilic or extremely hydrophilic areas on each substrate exhibit a directional rebounding towards the more wetting surfaces, i.e., hydrophilic or extremely hydrophilic surface. The trajectory and landing distance of the rebounded droplets were tailored by controlling the releasing height of the droplet, wetting contrast across the border, and portion of the droplet touching the more wetting surface of the substrates with wettability contrasts. The landing distance of the droplet increases with the increased releasing height and higher wettability contrast across the border. Increasing the portion of the impinging droplet touching the more wetting surface of the heterogeneous substrates leads to the shorter landing distance of rebounded droplets."
} | 284 |
37706333 | null | s2 | 3,616 | {
"abstract": "Drawing inspiration from natural systems, such as the highly segmented structures found in silk fibroin, is an important strategy when designing strong, yet dynamic biomaterials. Polymer-peptide hybrids aim to incorporate the benefits of hierarchical polypeptide structures into synthetic platforms that are promising materials for hydrogel systems due to aspects such as their biocompatibility and structural tunability. In this work, we demonstrated the utility of poly(ethylene glycol) (PEG) peptide-polyurea hybrids as self-assembled hydrogels. Specifically, poly(ε-carbobenzyloxy-L-lysine)-"
} | 148 |
36319743 | PMC10156819 | pmc | 3,617 | {
"abstract": "This review evaluates oilseed crop soybean endophytic bacteria, their prospects, and challenges for sustainable agriculture. Soybean is one of the most important oilseed crops with about 20–25% protein content and 20% edible oil production. The ability of soybean root-associated microbes to restore soil nutrients enhances crop yield. Naturally, the soybean root endosphere harbors root nodule bacteria, and endophytic bacteria, which help increase the nitrogen pool and reclamation of another nutrient loss in the soil for plant nutrition. Endophytic bacteria can sustain plant growth and health by exhibiting antibiosis against phytopathogens, production of enzymes, phytohormone biosynthesis, organic acids, and secondary metabolite secretions. Considerable effort in the agricultural industry is focused on multifunctional concepts and bioprospecting on the use of bioinput from endophytic microbes to ensure a stable ecosystem. Bioprospecting in the case of this review is a systemic overview of the biorational approach to harness beneficial plant-associated microbes to ensure food security in the future. Progress in this endeavor is limited by available techniques. The use of molecular techniques in unraveling the functions of soybean endophytic bacteria can explore their use in integrated organic farming. Our review brings to light the endophytic microbial dynamics of soybeans and current status of plant microbiome research for sustainable agriculture.",
"conclusion": "Conclusions This review evaluates endophytic bacteria in soybean and other food crops. The bioprospecting of these bacteria enhances their potential for sustainable yield enhancement. Soybean was discussed as a reference crop for oilseed crops due to its economic importance, high yield, and nutritional value. Soybean harbors some endophytic microbes important in agriculture. Beneficial endophytic microbes inhabiting different parts of the plants can potentially contribute to the growth of soybeans and other food crops. For instance, root nodule bacteria and endophytic bacteria enhanced nitrogen fixation in soybean, which promotes their yield and other yield parameters, enhance immunity, and boost plant defense against diseases. However, the root endophytes are emphasized because of high metabolic activities occurring below ground level due to the high quantity of metabolite secretion, which contributes to plant physiological functions. Different conventional and molecular techniques have been employed in the past to unravel endophytic microbes in some plants; nevertheless, each method comes with shortcomings. For instance, some endophytes can be difficult to culture on media despite their viability, such that culturing method can only unravel a lesser percentage (i.e., 0.1%) of endophytic populations. Hence, the advancement of endophytic microbes discovery using molecular techniques has proven more promising, although, with diverse challenges. Extracted endophytic bacteria DNA might contain traces of plant DNA, the chloroplast, and mitochondria DNA, which are identical to the targeted endophytic bacteria DNA [ 160 ]. Host depletion techniques have been employed to remove a substantial amount of plant DNA that might be present in the DNA extracted from the plant tissues. Conversely, the use of fluorescence in situ hybridization (FISH) is inefficient, because it can only be carried out in a natural habitat. Oilseed crop soybean is economically important due to their high yield, and nutritional value. The mechanism employed by the endophytes present in the seed, shoot, leaves, roots, and other microbes inhabiting the rhizosphere, bulk soil in plant growth promotion, and disease control still needs to be emphasized, although some research information are available on them. The variation in the diversity and population of microbes inhabiting different plant parts can be due to difference in the geographical locations, cropping system, developmental stage of the plants and the farming practices adopted. These key factors may affect the crop yield, microbial diversity and their ability to produce secondary metabolites. It is therefore very important to understand the mechanisms behind the production of secondary metabolites in soybean as a measure to improve their production, oil content, antioxidant content, seed quality, carbohydrates, chemical composition, and yield in different environments and also as a model to the research of other crops. More research should also be carried out to help understand the use of endophytes in the agriculture, industry, and medical industries, owing to the production of bioproducts. Better still, there is a need to develop robust bioinformatics tools and analytical techniques with the existing technologies to generate microbiome data as a guide for further experiments. Employing this approach by ecologists, environmental and computational scientists, microbiologists, agriculturists, and industrialists aims to provide insights into plant microbiome research as a reference for further studies. Hence, the authors conclude and recommend that the current approaches highlighted in this review will be of help to researchers in understanding the dynamics, prospect, and potential of endophytic microbes in soybeans and other food crops as agricultural bio-input to ensure food security and sustainable agriculture.",
"introduction": "Introduction Globally, diverse oilseed crops are cultivated for edible oil production to safeguard humans from malnutrition and related illnesses [ 1 ]. Their production rate differs from one country to another due to adaptation and growth under different weather conditions by region (e.g., temperate, tropical, and subtropical) [ 2 ]. The major type of oilseed crops are canola, groundnut, palm oil, sunflower, soybean, peanut, rapeseed, and cottonseed [ 3 ]. In 2020/2021, statistics of USDA showed an account of 362.05 soybeans, 68.87 rapeseed, 49.46 sunflower seed, 47.79 peanuts, and 41.80 of cottonseed, 19.96 palm kernel, and 5.75 copra world oilseed production (million metric tons) with soybean estimated of about a 90% production in the USA [ 4 ]. Also, in Sub-Saharan Africa, Nigeria produces and exports a larger percentage of soybean annually. Soybeans are leguminous plants in the family Fabaceae. Interest in soybean cultivation relies on their economic value, the edible oil-producing potential of about 20%, and protein content of 20–25% [ 5 ]. Notably, soybeans serve as an inexpensive and excellent source of high-quality edible oil and protein for humans as compared to other leguminous crops and animal protein [ 6 ], and can be a supplement food source for livestock. Yet, soybean’s market value and maximum utilization are less explored in many countries [ 7 ]. Soybean can be processed into composite food products, substituting animal proteins, i.e., eggs, meat, and milk. The uncertainties and challenges facing soybean cultivation may include poor and inefficient farming systems, drought, disease invasion, pest attack, lack of disease-resistant cultivars, etc. [ 8 – 10 ]. Diseases such as stem and root blight, bacterial leaf blight, downy mildew, bacterial pustule, rust, purple seed stain, frog-eye leaf spot, brown spot, charcoal rot, and soybean mosaic virus are the most common peculiar to soybean [ 11 ]. The control of disease in plants and crops under storage can be achieved by either biological, chemical, or physical means. Therefore, adopting proper control measures against phytopathogens in soybean can sustain plant health and crop productivity. From antiquity, farmers adopted diverse cropping systems (crop rotation, mixed farming, organic farming, etc.) and agricultural practices (e.g., agrochemicals, irrigation, and harrowing) to mitigate bottlenecks limiting the cultivation of soybean and other food crops. Over time, agrochemical use has been a major concern to environmentalists, ecologists, and microbiologists due to the negative impact on the ecosystem [ 12 , 13 ]. The peculiarity of these challenges is not limited to soybean cultivation alone, but other economical food crops. In recent times, research efforts are on the increase to devise a sustainable means of improving soybean, and other food crop production in order to help solve food scarcity, hunger, and malnutrition [ 14 ]. Because of the environmental threats posed by the synthetic fertilizer application and the incessant population increase, the need to employ biorational approaches and sustainable measures to enhance soybean production has become imperative. Naturally, soybean houses endophytic microbes capable of increasing the nitrogen pool in the soil to enhance plant nutrition for higher productivity [ 15 ]. The natural occurrence of these nitrogen-fixing bacteria is a promising way to reclaim lost soil nutrients for food production to meet the demand of the ever-growing population and relieve farmers of the cost and over-dependence on chemical fertilizers by farmers. Thus, harnessing endophytic bacteria as bioinoculants to oppose chemical fertilizers is critical as the best alternative. The plant root endosphere represents discreet regions occupied by diverse, endophytic microorganisms [ 16 ], where these microbes exhibit mutualistic, neutral, or antagonistic relationships with the host plants. The emphasis on the root-associated bacteria will be most considered in this review, as soybean root nodules naturally contain diverse nitrogen-fixing bacteria (NFB) [ 17 ]. The complementary effects of root-associated bacteria and root nodule NFB can positively influence plant growth and survival under nitrogen-limiting soils [ 18 ]. Here we emphasize that the nitrogen-fixing potential of endophytic bacteria in leaves, stems, seeds, flowers, ovules, etc. may be of greater importance in plant growth when compared to root nodules NFB only. Nevertheless, comparative studies of these bacteria from various plant organs upon inoculation under greenhouse and field experiments are required to ascertain this claim, requiring further studies. The molecular insights into plant–microbe interactions have unveiled important functions of some endophytic microbes, which suggests their maximum exploration as bioinoculants in sustaining plant growth and health [ 19 ]. For instance, a few beneficial nodule endophytic microbes associated with soybeans have been assessed under greenhouse and field trials to enhance soybean yield and in vitro screening for their antimicrobial properties against phytopathogens [ 20 ]. The interdependence of endophytic bacteria with the host plants confers beneficial effects in soybeans and other food crops, such that it stimulates plant growth promoters, antibiosis activity against phytopathogens for plant health, defense against oxidative stress, and yield enhancement without any pathogenic effects [ 21 , 22 ]. Limited information is available in the literature on the plant growth stimulation and biocontrol potential of endophytic bacteria inhabiting soybean, thus limiting their ecological services. Nevertheless, exploring endophytic bacteria as bioinoculants can provide several opportunities in mitigating diverse agricultural problems, such as biotic and abiotic stress, and climate change. Furthermore, addressing the challenges and uncertainties limiting plant microbiome biotechnologically will ultimately reveal the amazing realities of incorporating endophytic resources from soybean and other food crops into agricultural management. Our review brings to light the endophytic microbial dynamics of soybeans and current status of plant microbiome research for sustainable agriculture."
} | 2,912 |
21556754 | PMC3098350 | pmc | 3,619 | {
"abstract": "Next-generation sequencing provides technologies which sequence whole prokaryotic and eukaryotic genomes in days, perform genome-wide association studies, chromatin immunoprecipitation followed by sequencing and RNA sequencing for transcriptome studies. An exponentially growing volume of sequence data can be anticipated, yet functional interpretation does not keep pace with the amount of data produced. In principle, these data contain all the secrets of living systems, the genotype–phenotype relationship. Firstly, it is possible to derive the structure and connectivity of the metabolic network from the genotype of an organism in the form of the stoichiometric matrix N . This is, however, static information. Strategies for genome-scale measurement, modelling and predicting of dynamic metabolic networks need to be applied. Consequently, metabolomics science—the quantitative measurement of metabolism in conjunction with metabolic modelling—is a key discipline for the functional interpretation of whole genomes and especially for testing the numerical predictions of metabolism based on genome-scale metabolic network models. In this context, a systematic equation is derived based on metabolomics covariance data and the genome-scale stoichiometric matrix which describes the genotype–phenotype relationship.",
"conclusion": "Conclusions and perspectives NGS will enable the systematic and comparative investigation of the genotype–phenotype relationship. However, before this relationship reveals its secrets, a comprehensive strategy of metabolic modelling and metabolic measurements has to be established. Here, I have presented a systematic and conceptual equation connecting the genotype and the molecular dynamic phenotype. This equation can be exploited in future for the inverse modelling of the dynamic molecular phenotype and will be instrumental in the interpretation of the corresponding genotype. To achieve these accurate predictions of a dynamic metabolism in newly sequenced organisms, the following improvements are essential. For validating models of metabolism metabolomics will play a key role; however, metabolome coverage needs to be enhanced by combining different analytical procedures and novel technologies. Furthermore, we need to aim for improved cellular resolution of metabolite profiles. Because of the complexity of metabolism it can be helpful to dissect the system into smaller parts and analyse these discretely. This procedure coincides with targeted pathway analysis in metabolomics (see also Fig. 6 and earlier). The complete structure can be reconstructed by defining biochemical modules and assembling these modules into large-scale networks. Two recent studies demonstrated how genotyping and metabolite profiling can be combined on a robust statistical basis [ 78 , 79 ]. In the study by Gieger et al. [ 79 ] genome-wide association studies with the human metabolic phenotype were performed using a commercial metabolite profiling platform. The observation from this study was that common genetic polymorphisms induce major differentiations in the metabolism of the individuals. These results strongly support the general strategy of personalized health care and nutrition in combination with metabolite profiling and genotyping [ 79 ]. In the study of Chan et al. [ 78 ] a large panel of Arabidopsis plants were genotyped and investigated by GC-TOF-MS metabolite profiling. One of the conclusions from this study was that genotype–metabolite associations are sensitive to environmental fluctuations. This opens up a completely new avenue for environmental studies combining rapid NGS genotyping and molecular profiling using omics technologies such as metabolomics and proteomics. Finally, the integrative approach combining multilevel measurements and modelling approaches in the targeted organism (see Fig. 7 ) [ 21 ] is the conclusive goal. The combination of transcript, protein and metabolite data is especially relevant since there is no initial, as yet readable information from the genome sequence on which enzyme is active or inactive. However, active or inactive enzymes will give different biochemical states and will result in a different stoichiometric matrix N and a different Jacobian J (see earlier). Thus, we need knowledge of the activity or presence and absence of messenger RNAs and proteins. Furthermore, metabolic flux analysis is crucial to reveal active pathways and flux distributions. Techniques such as metabolic labelling with stable isotopes can be exploited in combination with genome-scale metabolite profiling to reveal the in vivo activity of whole pathways and enzymes [ 23 – 25 ]. In combination with the genome-scale investigation of the molecular network of an organism to understand its networking properties, it is as important to continue classic biochemical studies to elucidate protein functions on a much smaller scale and case by case. Integrating this knowledge into the information about the network dynamics of the molecular components might finally result in a functional understanding of the system in relation to the genotype.\n Fig. 7 Integrative approach combining genome sequencing, dynamic modelling and omics analysis to reveal a mechanistic and predictive understanding. EFM elementary flux modes, FBA flux balance analysis, MCA metabolic control analysis \n In conclusion, NGS in combination with metabolomics science will be a powerful tool for the investigation of the genotype-phenotype relationship and the ab initio prediction of metabolism in newly sequenced organisms.",
"introduction": "Introduction We have witnessed an exponential growth of public genome sequence releases ( http://www.genomesonline.org/ ). In principle, this amount of data will enable us to investigate any subtle aspect in living systems. However, the process of whole-genome assembly and functional gene annotation of de novo sequenced organisms is far behind the speed of data generation using next-generation sequencing (NGS) technologies [ 1 – 4 ]. Whole genome assembly and ab initio gene prediction is in the first instance dependent on algorithms. In recent studies, approaches have been presented for functional annotation of newly sequenced genomes combining complementary DNA [expressed sequence tag (EST), messenger RNA or RNA-sequencing data] with gene predictions [ 5 ]. More recently, proteogenomic studies have used proteomics data to reveal new gene models [ 6 – 8 ]. A truly systems biology approach is the integration of several layers of molecular information in conjunction with metabolic modelling [ 7 ]. In this study, genome annotation with metabolomics data and a structural modelling approach were combined for the first time [ 7 ]. Besides the qualitative or structural investigation of genome function and metabolic networks, the next aim is to explore the quantitative prediction. Here, dynamic modelling is the key approach. The final goal is genome-scale metabolic reconstruction and quantitative understanding and prediction of metabolism in a newly sequenced organism, the genotype–phenotype relationship. By reviewing the literature it becomes clear that this is the limiting step in the functional interpretation of whole genomes and organisms. Only an iterative cycle of improving genome annotation, structural and dynamic modelling and comparison of the predictions with experimental data will be successful, necessitating not only further development of computer-based annotation of gene functions and modelling algorithms but also the integration of whole metabolome profiling approaches. In the next sections I will explore the strategies and limitations of how to connect metabolomics data and genome-derived metabolic reconstruction and suggest a complete workflow. A systematic equation is derived for the genotype–phenotype relationship."
} | 1,961 |
35479639 | PMC9037085 | pmc | 3,620 | {
"abstract": "Revealing community assembly and their impacts on ecosystem service is a core issue in microbial ecology. However, what ecological factors play dominant roles in phyllosphere fungal community assembly and how they link to crop quality are largely unknown. Here, we applied internal transcriptional spacer high-throughput sequencing to investigate foliar fungal community assembly across three cultivars of a Solanaceae crop (tobacco) and two planting regions with different climatic conditions. Network analyses were used to reveal the pattern in foliar fungal co-occurrence, and phylogenetic null model analysis was used to elucidate the ecological assembly of foliar fungal communities. We found that the sensory quality of crop leaves and the composition of foliar fungal community varied significantly across planting regions and cultivars. In Guangcun (GC), a region with relatively high humidity and low precipitation, there was a higher diversity and more unique fungal species than the region of Wuzhishan (WZS). Further, we found that the association network of foliar fungal communities in GC was more complex than that in WZS, and the network properties were closely related to the sensory quality of crop. Finally, the results of the phylogenetic analyses show that the stochastic processes played important roles in the foliar fungal community assembly, and their relative importance was significantly correlated with the sensory quality of crop leaves, which implies that ecological assembly processes could affect crop quality. Taken together, our results highlight that climatic conditions, and plant cultivars play key roles in the assembly of foliar fungal communities and crop quality, which enhances our understanding of the connections between the phyllosphere microbiome and ecosystem services, especially in agricultural production.",
"introduction": "Introduction Assembly patterns within microbial communities are an important topic in microbial ecology and are closely related to the functioning of plant-associated ecosystems ( Konopka, 2009 ). There are many different factors that could affect the structure of the plant-associated microbial community. For example, soil microbial communities were different across various plant species ( Ma et al., 2019 ), because of the range of nutrient content available in the leaf and root litter that alters decomposer abundance ( Otsing et al., 2018 ). The foliar endophytic fungal community in Cirsium arvense could be associated with the soil nutrients and arbuscular mycorrhizal (AM) colonization ( Eschen et al., 2010 ), and the latter would be further affected by root exudates, such as methyl salicylic acid and acibenzolar- S -methyl ( Mannaa et al., 2020 ). Furthermore, climate change, such as warming, would decrease fungal species richness and change foliar fungal community composition, especially at the end of the growing season ( Faticov et al., 2021 ). In the case of elevated atmospheric carbon dioxide, the growth of trees would also lead to the changes in the composition of microbial communities that colonize the fallen leaves ( Kelly et al., 2010 ). Crop cultivar, tissue type, and climatic factors can all significantly influence phyllosphere fungal community structure; moreover, location-dependent climate conditions could contribute to the differences in abundance, diversity, and presence of genera containing pathogens, whereas the root communities were less affected by climatic factors ( Latz et al., 2021 ). In addition, drought changes the composition of the root microbiome, where changes in the relative abundance of specific bacterial groups were associated with increased drought tolerance in plants ( Fitzpatrick et al., 2018 ). Although bacterial abundance was negatively affected by O 3 stress, it was found that the fungal abundance was substantially stimulated (up to 12-fold compared with non-fumigated plants at 20°C). These changes were accompanied by modifications of the genetic structures and a relative increase in amino acids catabolism ( Changey et al., 2018 ). The above findings advanced our understanding of the drivers in shaping plant-associated communities, but the mechanisms of assembly in the foliar fungal community remain largely unknown. Exploring network assembly in microbial communities and their responses to environmental changes is fundamentally vital for the understanding of community organization ( Zhou et al., 2010 ). Microbial community networks can provide a mechanistic association between species in a specific environment and information on the dynamics of community structure as a function of time or other external variables ( Cardona et al., 2016 ). For example, climate change, such as warming, can significantly increase network complexity, including network size, connectivity, and number of keystone species ( Faticov et al., 2021 ), whereas elevated CO 2 can increase modularity and hierarchy ( Zhou et al., 2010 ). The community assembly of plant-associated microbes may have some differences, such as the rhizosphere microbial networks. In the rhizosphere, complexity of microbial ecological network increases with the growth of wild oat plants ( Shi et al., 2016 ). Artemisia annua (sweet wormwood) promoted a specific root-associated microbial community assembly process, with increased abundance of plant growth–promoting microorganisms and building of interkingdom association networks, which may be beneficial for the fitness of plants in the natural environment ( Shi et al., 2021 ). Together, these results revealed that network assembly of plant-associated microbial communities could be closely related to plant growth. However, how network assembly of the phyllosphere microbial community is affected by climatic conditions and crop cultivars have been less well studied. Phylogenetic analyses based on null model provide a conceptual background for understanding the ecological processes of community assembly that determine which, and how many, species live in a particular environment ( Campbell et al., 2011 ). Foliar fungi are of great importance to host plant growth and health and can also affect ecosystem functions. Most importantly, host environmental filtering caused by fungal infections outweighs competitive exclusion in driving foliar fungal community assembly in symptomatic leaves ( Liu X. et al., 2021 ). Community co-occurrence theories can be explained mainly by niche-based theory and the null model ( Gravel et al., 2006 ; Jiao et al., 2020 ). On the one hand, niche-based theory ( Zhou and Ning, 2017 ) posits that deterministic processes play a key role in the community assembly process. Different species occupy different niches, and ecological selection can affect the community co-occurrence. On the other hand, the neutral model demonstrates that all species are equivalent on ecological function, and the community assembly is affected by stochastic processes but not their ecological abilities. The environment can play a vital role in stochastic processes that correlate with community assembly. However, the deterministic and stochastic processes that shape phyllosphere fungal community assembly have not been extensively explored. The present study aims to reveal the assembly of phyllosphere fungal communities inhabiting a Solanaceae (tobacco) crop across climatic conditions and cultivars. We set up a large-scale field experiment with three crop cultivars and in areas with different climatic conditions in Hainan, China. We assessed the sensory quality of crop leaves and explored the phyllosphere fungal communities using internal transcriptional spacer (ITS) high-throughput sequencing technology. We hypothesized that (i) the sensory quality of crop leaves and the composition of the foliar fungal community are significantly affected by both the planting region and cultivar; (ii) the association network of foliar fungal communities and its characteristics are closely related to the sensory quality of the crop; (iii) the community assembly of all samples would be dominated by the ecological drift, which is community phylogeny structure with little effect on the sensory quality of the crop.",
"discussion": "Discussion In our study, we studied various plant cultivars and ecological regions with different climatic conditions in Hainan, China. Our results showed that the sensory quality and foliar fungal community structure showed significant differences across different planting regions and crop cultivars. There were significant associations between the sensory quality of crop leaves and foliar fungal diversity. Furthermore, ecological association networks of crop foliar fungi in GC were more complex than those in WZS, and the community assembly was dominated by ecological drift. Network assembly had significant correlations with sensory qualities, whereas ecological processes of community assembly had little effect on sensory qualities. These results highlighted how the phyllosphere fungal community is closely associated with crop quality. Effects of Foliar Fungal Community on Crop Quality Our results demonstrated that planting regions with different climate conditions and cultivars had significant effects on the sensory quality of a Solanaceae crop, which is likely to be led by the changes in foliar fungal community composition. Climate may affect the soil conditions and further alter the fungal community. Moreover, cultivation methods also will change the soil nutrients composition and the microbial network structure in the rhizosphere or foliar fungal communities. This is supported by previous studies of the cucumber rhizosphere, where AM fungi (AMF) significantly altered the nutrient composition of the branches of the host plant, with the strongest contrast observed between cucumber-irregular symbiotic plants and non-mycorrhizal cucumber plants ( Ravnskov and Larsen, 2016 ). The composition of soil fungal communities changed with continuous cucumber cultivation, which may have been caused by the combined cultivation period of cucumber and excessive application of chemical fertilizers ( Sun et al., 2021 ), such as nitrogen fertilizer and phosphate fertilizer. Fungal community diversity and microbial interaction play key roles in plant growth and metabolism. Proteobacteria can directly inhibit Firmicutes from entering into the endophytic community and consequently modify the microbial community ( Chen et al., 2020 ). Endophytes have been isolated from Coffea canephora , and the high biodiversity of fungal endophytes in coffee plants may help us understand the plant growth process ( Vega et al., 2010 ). Soybean rhizosphere may act as allelochemicals in the interactions between root and soil microbial community in a long-term monocropped soybean field ( Guo et al., 2011 ). AMF and plant growth–promoting bacteria are beneficial to horticultural crops, which could increase yield and enhance crop quality ( Emmanuel and Babalola, 2020 ). Terpenoids are a group of structurally diverse natural products that are widely used in the flavor and fragrance industry. Furthermore, it was clarified that the fungal sesquiterpene synthase’s function differs between the phyla Ascomycota and Basidiomycota ( Zhang et al., 2020 ). It was evident that there is interaction between the indigenous microbial community and grain metabolism even with good-quality, mature malting barley. In the malting ecosystem, the fungal community markedly contributed to the production of microbial β-glucanases and xylanases and was also involved in proteolysis ( Laitila et al., 2007 ). Elevated temperature also increases aphid abundance but decreases AMF colonization rates of the wheat grain, which implies that climate may affect crop quality by altering plant-associated fungal communities ( Tian et al., 2019 ). Drivers in Shaping the Structure of Foliar Fungal Community Many factors could affect the foliar fungal community structure, including plant cultivar ( Martins et al., 2011 ), soil physical and chemical characteristics, and climate ( Kauserud et al., 2013 ). Our results showed that planting regions with different climates and plant cultivars are key factors. The PCA results indicated that both cultivar and planting region climate could affect the community structure of foliar fungi, whereas through PERMANOVA using distance matrices ( Alekseyenko, 2016 ), we found that the impact of ecological environment factors ( R 2 = 0.32, p < 0.05) was more significant than crop cultivar ( R 2 = 0.42, p < 0.05; Table 3 ). TABLE 3 The impact of ecological environment factors and crop cultivar on the structure of the foliar fungal community using permutational multivariate analysis of variance using distance matrices based on Bray–Curtis similarity index. \n df \n Sum of sequence \n R \n 2 \n \n F \n \n p \n Planting region 1 72.83 0.32 2.61 0.0166 Cultivar 2 93.00 0.42 1.67 0.2333 The phyllosphere represents one of the most abundant habitats for microbiota colonization ( Chen et al., 2020 ), and the role played by interactions between phyllosphere microorganisms in modifying the fungal community composition cannot be neglected. Related metastudies have identified climate as an important driving factor in different aspects of fungal biogeography, including the global distribution of common fungi and the composition and diversity of fungal communities ( Vetrovsky et al., 2019 ). Climatic variability might modify trait selection in fungi, including spore size and dispersal characteristics. Changes in the composition and characteristics of fungal communities will have an important impact on interaction with plant communities and ecosystem functions ( Andrew et al., 2016 ). Climate change may affect ecosystem functioning due to the narrow climatic tolerances of key fungal taxa. Mycorrhizal fungi appear to have narrower climatic tolerances than pathogenic fungi ( Vetrovsky et al., 2019 ). Our results showed that the differences between tobacco cultivars play an important role in phyllosphere fungal community structure and affect the microbial co-occurrence pattern in phyllosphere fungal communities. Similarly, cucumber cultivars inoculated with different AMF had differential responses in terms of growth and branch nutrient composition, which revealed that plant cultivar could affect the microbial community functional diversity ( Ravnskov and Larsen, 2016 ). Moreover, AMF also can enhance ecosystem resilience and reduce the negative impact of increased precipitation on nutrient losses ( Martinez-Garcia et al., 2017 ). Also, mycorrhizal fungi can promote or hinder the successful spread of plants away from harsh environments ( Bennett and Classen, 2020 ). Community Assembly of Foliar Fungal Community on Leave Surface The ecological assembly process is vital for the construction of microbial communities ( Sloan et al., 2006 ). Spatial turnover in the composition of biological communities includes (ecological) drift, selection, and dispersal. Quantitatively estimating the influences of selection, dispersal, and drift is fundamental to our understanding of ecological systems ( Stegen et al., 2013 ). In our study, the community NRI of fungi on leaf surfaces in GC region was between 0.74 and 1.01, whereas in WZS, it was between 0.78 and 1.26; furthermore, the NTI and NRI significantly varied among cultivars and planting regions ( Figure 5 ), which implicates that the stochastic process plays a key role in the local species diversity and spatial turnover. Other studies have previously elucidated the importance of drift in community assembly process of the legume root nodules, including the core rhizobial communities (genus Mesorhizobium ) that were driven by dispersal limitation in concert with drift (81.1% of nodA communities, Ramoneda et al., 2020 ). During the degradation of straw, ecological drift was important across all stages of decomposition ( Bao et al., 2020 ). The βNTI was linearly related to the sensory quality of crop leaves, indicating that the ecosystem services may depend, to some extent, on the assembly process of microbial communities. This may have resulted from stochastic processes in the assembly of foliar fungal communities. Although these results have advanced our understanding of the relationships between microbial community assembly and crop quality, future work is still required to further reveal the connections between foliar fungal community and plant molecular metabolic mechanism."
} | 4,124 |
25324833 | PMC4183113 | pmc | 3,621 | {
"abstract": "A remarkable example of biological engineering is the capability of some marine animals to take advantage of photosynthesis by hosting symbiotic algae. This capacity, referred to as photosymbiosis, is based on structural and functional complexes that involve two distantly unrelated organisms. These stable photosymbiotic associations between metazoans and photosynthetic protists play fundamental roles in marine ecology as exemplified by reef communities and their vulnerability to global changes threats. Here we introduce a photosymbiotic tidal acoel flatworm, Symsagittifera roscoffensis , and its obligatory green algal photosymbiont, Tetraselmis convolutae (Lack of the algal partner invariably results in acoel lethality emphasizing the mandatory nature of the photosymbiotic algae for the animal's survival). Together they form a composite photosymbiotic unit, which can be reared in controlled conditions that provide easy access to key life-cycle events ranging from early embryogenesis through the induction of photosymbiosis in aposymbiotic juveniles to the emergence of a functional “solar-powered” mature stage. Since it is possible to grow both algae and host under precisely controlled culture conditions, it is now possible to design a range of new experimental protocols that address the mechanisms and evolution of photosymbiosis. S. roscoffensis thus represents an emerging model system with experimental advantages that complement those of other photosymbiotic species, in particular corals. The basal taxonomic position of S. roscoffensis (and acoels in general) also makes it a relevant model for evolutionary studies of development, stem cell biology and regeneration. Finally, it's autotrophic lifestyle and lack of calcification make S. roscoffensis a favorable system to study the role of symbiosis in the response of marine organisms to climate change (e.g., ocean warming and acidification). In this article we summarize the state of knowledge of the biology of S. roscoffensis and its algal partner from studies dating back over a century, and provide an overview of ongoing research efforts that take advantage of this unique system.",
"conclusion": "Concluding remarks Non-model organisms are generally prone to an uncertain fate in biological science until reliable protocols are developed for a transfer from field to laboratory. Continuous laboratory cultivation and in vitro rearing under artificial conditions often represent major challenges: it enables to have access to precise windows of the developmental program from the first cleavage to the death of the model. Several new animal models have recently emerged for exploring metazoan diversity, for investigating the origin of body plans, and for tracing metazoan evolution and other crucial life history traits. Here, we have stressed diverse fields of research that are being investigated or could potentially be investigated using S. roscoffensis . Specific features of S. roscoffensis could be useful and complementary for the analysis of more general phenomena such as the molecular basis of the mechanisms of photosymbiosis: the easy handling of S. roscoffensis and the culture of its photosymbiotic partner represents an emerging model opening new designs of experimental protocols for alternatively questioning endosymbiosis or attain a more concrete idea of reef ecology and the fragility of photosymbioses that face major environmental changes which result in phenomena such as mass bleaching of the emblematic photosymbiotic coral. We also report the potential of employing a combination of approaches for exploring the molecular and physiological basis of other processes such as brain regeneration and the potential of glial cells-like in neural development, function and health but also the superposition of circadian/circatidal rhythms in the field of chronobiology (Tessmar-Raible et al., 2011 ), these being crucial phenomena in many marine organisms which have been rarely considered in the classical literature (Martin, 1907 ). 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 Photosymbiosis represents around 50% of marine photosynthesis, as exemplified by corals and other reef animals as well as by the remarkable biomass and diversity of oceanic photosymbiotic protists (unicellular eukaryotes) from the taxonomic super-groups Rhizaria, Alveolates, and Stramenopiles (Baldauf, 2008 ). An association between a host (multi- or unicellular) and an “algal” photosymbiont represents, in principle, a “domestication” of photosynthesis that can result in a trophic independence as long as the partners are located in the euphotic zone in view of the infinite source of solar energy. As a consequence, photosymbiosis has a strong global ecological impact and, hence, the adaptations and innovations that contribute to photosymbiotic interactions require both in depth functional and molecular analysis. This analysis is also important since the evolutionary success of photosymbiosis, based on innovative combinations of hosts/photosynthetic symbionts, contributes markedly to marine diversity. Finally, a comprehensive understanding of the mechanisms underlying the recruitment of photosynthetic symbionts by heterotrophic organisms may also contribute to and understanding of the ancestral evolutionary events of acquisition of plastid-like endosymbionts by a primitive heterotrophic eukaryotic cells, i.e., primary and secondary endosymbiosis. We introduce here Symsagittifera roscoffensis (formerly Convoluta roscoffensis ), a marine flatworm-like animal that belongs to the phylum Acoela (a = non, coela = cavity), as an ideal photosymbiotic biological model system for exploring the functional biology, molecular biology and evolution of photosymbioses. S. roscoffensis , described as a “Plant-Animal” (Keebles, 1910 ) a century ago, is endemic to the intertidal zone of European coasts of the Atlantic Ocean and the English Channel, and at low tide forms huge colonies composed of millions of individuals (Figures 1 , 2A–C ). The vivid green color of the colonies is due to the abundant presence of a unicellular green algal endosymbiont, namely the prasinophyte Tetraselmis convolutae (Figure 3A ). This animal presents a wealth of biological features that are relevant for investigating key phenomena of the photosymbiotic interaction such as the development, genetics, and cell biology of symbiogenesis that are not easily addressed in other more conventional model systems. Importantly, while other experimental systems are well established for the study of some of these aspects, the combination of features exhibited by S. roscoffensis and its associated symbionts (and indeed acoels in general) is exceptional in its breadth and diversity, and this makes it possible to address a suite of biological questions that cannot be easily approached in any other single model system. Indeed “to resolve the puzzle of metazoan evolution and development, bioinformatic and experimental approaches must be applied to a wider range of species than just the standard model organisms” (Tessmar-Raible and Arendt, 2004 ). Figure 1 Life cycle of Symsagittifera roscoffensis . In the background minute and elongated adults (around 4 mm length) swimming and gliding upon and inside the sand at low tide. Figure 2 (A) Original drawings from Graff ( 1891 ) depicting structures of S. roscoffensis \n (1) together with a closely related (dwarf) species S. schultzii (ex Convoluta schultzii ) (2) and in particular the localization of algae in the epidermal area. a : frontal organ (sensing), b : statocyst c : mouth, d : ovocyte, e : female genital pore, f : bursal nozzle [the bursa nozzle leads from the seminal bursa (g) to the female genital opening (e)], g : seminal bursa, h : sagittocysts, i : seminal vesicles, j : male genital pore. (B) Original drawings from Delage ( 1886 ) a : frontal organ (sensing), b : (statocyst) c : central nervous system (gray area surrounding the statocyst). (C1) Adult symbiotic S. roscoffensis (see also magnification of the anterior pole in D ). Note the presence of four visible lines running along the body corresponding to thin longitudinal dorsal and lateral neural bundles. (C2) Dark green mat on the beach: a typical colony of S. roscoffensis in the residual flows on ebb tides close to Roscoff, Brittany, France. (D) Magnification of the anterior part. Note the reduced presence of algae surrounding the anterior pole and the occurrence of reddish structures; these rod-shaped epidermal mucus secretion bodies are called rhabdoids (Smith et al., 1982 ). Yellow arrows show the two photoreceptors flanking the central statocyst involved in gravity sensing (the blue arrow). ( E) Egg capsule (cocoons) with early developmental stage embryos (E1) 2-cell stage with two macromeres. (E2) 4-cell stage: first duet of micromeres. Figure 3 (A1) TEM cross section of a S. roscoffensis adult. On the top of the image, the ciliated (cilia: cil) epidermis is visible above a net of both circular and longitudinal muscles (cm and lm). Algal cells (alg) are in contact with the acoel cells below the muscles layer. Algal cells exhibit two specific features, the chloroplasts (Chl) and the pyrenoid with starch accumulation on the outside (the white mass). Note some algal extensions close to the epidermis, above the muscle boundary. (A2) Detailed scheme of the free-living unicellular green alga, Tetraselmis convolutae (modified from Parke and Manton, 1967 ): c, chloroplast; f, flagellum; g, golgi body; m, mitochondrion; mf, muciferous body; n, nucleus; p, pyrenoid; rb, refractive body of unknown nature; ss, starch shell; s, stigma; st, stroma starch; t, theca. (B,C) These two pictures exhibit a clear physical proximity of germinal cells, respectively sperm (Spz in B ) and ovocyte (ov in C ), and algae (alg). Such a situation could increase possible horizontal gene transfers. From the late nineteenth century onwards, this intriguing green flatworm has been the subject of investigations of the symbiosis between the metazoan host and the unicellular green alga that were focussed mainly on trophic exchanges and energy contribution. An important development for experimentation occurred in the 1960s when the phycologist L. Provasoli (1908–1992, Lehman, 1993 ), managed to establish laboratory cultures in New York of multiple generations of wild type S. roscoffensis originating from the coasts of Brittany, France (Parke and Manton, 1967 ; Provasoli et al., 1968 ). However, for various reasons, from 80's onward research on S. roscoffensis slowly declined despite the clear potential for addressing key biological issues ranging from photosymbiosis to regeneration, and including development, ecology, and genetics. This lack of experimental progress is especially unfortunate for issues in metazoan evolution. This is because S. roscoffensis as an acoel represents something of a Gordian knot in the debate concerning the features of the ancestor of the all bilaterian animals (the so-called “urbilaterian”). In this debate some scientists propose that all Bilateria are descendants of a primitively coelomate ancestor (archicoelomate hypothesis, Salvini-Plawen, 1978 ; Balavoine, 1998 ) implying that flatworms have secondarily lost their coelom and are thus not basal bilaterians, while others argue that acoel flatworms are primarily simple and that their acoelomate condition is representative of the ancestor of the Bilateria (planuloid-acoeloid hypothesis, Salvini-Plawen, 1978 ; Baguna and Riutort, 2004a ). Thus, from a historical point of view, the current renaissance in experimental investigations of S. roscoffensis represents a renewal with a rich and relevant background of over a century of studies dedicated to this acoel flatworm from the early 1870s until the 1980s. In this report, we review the current state of knowledge concerning the biology, ecology, and physiology of this animal, and we highlight the many new avenues of research made possible by the precise experimental control of the ontogeny of this marine flatworm including induced symbiogenesis through co-culture of the acoel and the algal symbiont."
} | 3,108 |
32354088 | PMC7285240 | pmc | 3,622 | {
"abstract": "The nitrogen environment and nitrogen status of reef-building coral endosymbionts is one of the important factors determining the optimal assimilation of phototrophic carbon and hence the growth of the holobiont. However, the impact of inorganic nutrient availability on the photosynthesis and physiological state of the coral holobiont is partly understood. This study aimed to determine if photosynthesis of the endosymbionts associated with the coral Stylophora pistillata and the overall growth of the holobiont were limited by the availability of dissolved inorganic carbon and nitrogen in seawater. For this purpose, colonies were incubated in absence or presence of 4 µM ammonium and/or 6 mM bicarbonate. Photosynthetic performances, pigments content, endosymbionts density and growth rate of the coral colonies were monitored for 3 weeks. Positive effects were observed on coral physiology with the supplementation of one or the other nutrient, but the most important changes were observed when both nutrients were provided. The increased availability of DIC and NH 4 + significantly improved the photosynthetic efficiency and capacity of endosymbionts, in turn enhancing the host calcification rate. Overall, these results suggest that in hospite symbionts are co-limited by nitrogen and carbon availability for an optimal photosynthesis.",
"introduction": "1. Introduction Symbiotic associations between animals and photosynthetic microorganisms have resulted in some of the best physiological adaptations that have evolved in the animal kingdom [ 1 ]. Such associations have indeed increased fitness and aptitude for food acquisition, related to the photosynthetic capacity of the autotrophic symbionts and the capture of external prey by the host. The symbiosis established between reef-building corals (Cnidaria; Anthozoa; Scleractinia) and dinoflagellates of the Symbiodiniaceae family [ 2 , 3 ], is one of the most well-known marine nutritional symbioses, because corals are the cornerstone of reef ecosystems, providing, through their calcification process, a tridimensional structural framework that support a high biodiversity and productivity [ 4 ]. Their ecological success in nutrient-poor tropical waters is due to the sharing and the tight recycling of nutrients between the partners [ 5 , 6 ]. Symbionts, through photosynthesis, transform dissolved inorganic carbon (DIC) and nitrogen (DIN), as well as other nutrients, into organic molecules which are translocated to the host for most of its metabolic needs, such as respiration, calcification and energetic reserves [ 7 , 8 , 9 ]. In turn, the animal provides its symbionts with nutrients coming from the environment and from its metabolic waste products [ 10 , 11 , 12 ]. An efficient acquisition of DIC by the symbionts is a key process for the health of the symbiotic association, because the amount of DIC fixed and assimilated by the symbionts will also determine the amount of photosynthates transferred to the host for its own needs [ 9 ]. Both partners have developed multiple ways of acquiring and concentrating carbon for photosynthesis. In addition to CO 2 derived from respiration, the animal delivers DIC from the water column (mainly HCO 3 − ) to the symbionts through a combination of external and intracellular carbonic anhydrases [ 13 , 14 , 15 ], specific transporters [ 16 ] and H + -ATPase [ 17 ]. Then the CO 2 concentrating mechanism (CCM) of the algae takes over to maximize carbon fixation [ 18 , 19 ]. Even though oceanic concentrations of DIC are high, DIC can be a source of limitation for the primary production of the symbionts [ 15 , 20 , 21 ] although not in all coral species or environmental conditions [ 15 , 22 ]. This is due to the fact that the adequacy of the supply of DIC to the symbionts (with respect to saturation of demand) depends on diverse environmental factors, but also on symbiont density or holobiont metabolic needs, and is not yet well understood. Since symbiotic carbon acquisition is a key process for coral health, conditions which lead to increased photosynthetic efficiency of the symbionts under varying environmental conditions are therefore of interest for reef biologists, especially in this era of global changes. The nitrogen environment and nitrogen status of the coral symbionts is one of the important factors determining optimal photosynthetic carbon acquisition and hence holobiont growth. In plants and free-living algae, carbon fixation and nitrogen assimilation are interdependent, because fixed carbon need to be coupled to a nitrogen source for amino acid production [ 23 , 24 , 25 , 26 , 27 ]. For instance, it was shown that CO 2 deprivation inhibits NH 4 + uptake in the unicellular alga Cyanidium caldarium [ 24 ], but also that internal nitrogen levels modulate the photosynthetic response of the organisms to DIC enrichment [ 28 ]. In the coral-dinoflagellate association, the carbon-nitrogen coupling has been less studied than in plants and algae and has led to opposite conclusions, also depending on the nitrogen source (ammonium versus nitrate) and the nitrogen to phosphorus (N:P) ratios considered [ 29 , 30 ]. Overall, the above studies showed negative impacts of increased N:P ratios and nitrate enrichment on coral photosynthesis and health and a positive effect of ammonium enrichment. Nevertheless, for the same nitrogen source such as ammonium, controversies still exist. On the one hand, overall carbon acquisition (per holobiont surface area) or per symbiont cell was enhanced by heterotrophic feeding (ammonium source) or direct ammonium supplementation, suggesting that nitrogen, not carbon, was limiting symbiont photosynthesis [ 31 , 32 , 33 ]. On the other hand, when nitrogen enrichment of the seawater increased symbiont density [ 34 , 35 , 36 , 37 ], symbionts may compete for CO 2 acquisition, and present decreased rates of photosynthesis per symbiont cell [ 34 , 36 ]. Such discrepancy in the effect of nitrogen enrichment on carbon acquisition can be due to differences in DIC limitation of the symbionts but remains to be investigated. All together, these results suggest that the role of inorganic nutrient availability on the photosynthetic capacities of the coral holobiont is partly understood. The aim of this study was to test the DIC and ammonium (NH 4 + ) limitation for the photosynthesis of the symbionts associated to the scleractinian coral Stylophora pistillata . We first assessed whether symbionts of S. pistillata , were DIC or NH 4 + limited under our oligotrophic culture conditions, by enriching seawater with either 4 µM ammonium or 6 mM bicarbonate. We also tested DIC limitation under nitrogen-enriched conditions. Such conditions will occur more and more frequently in eutrophic coastal reef waters, following urban and agricultural development [ 30 , 38 , 39 ]. In the present context of global change and anthropogenic impacts on coastal waters, experimental studies like this one will help develop adaptive management strategies for the future of coral reefs.",
"discussion": "4. Discussion This study reports the first assessment of a combined inorganic carbon (HCO 3 − ) and nitrogen (NH 4 + ) enrichment on the physiological traits of a coral-dinoflagellate symbiosis. Although positive effects on coral physiology could already be observed with the supplementation of one or the other nutrient, the most important and positive changes were observed when both nutrients were provided, due to an interactive relationship between NH 4 + and DIC acquisition. These results suggest that coral symbionts are co-limited in nitrogen and carbon for an optimal photosynthesis. 4.1. The Increased Availability of HCO 3 − and NH 4 + Improves Photosynthetic Processes in Endosymbionts The increased availability in bicarbonate or ammonium alone had significant, but time-dependent effects on symbiont physiology. Since symbionts are generally nitrogen limited in hospite [ 54 ], ammonium supplementation induced a fast response of the symbionts, which showed a significant increase in growth and pigment content as well as an increase in some of their photosynthetic processes (E O per symbiont and rETR) after only 7 days of incubation. Conversely, the rate at which iC supplementation affected symbiont physiology was dependent on the parameter considered. Nevertheless, after 21 days, the increased availability in either bicarbonate or ammonium improved significantly the photosynthetic efficiency (rETR PSII ) and/or capacity (O 2 production per symbiont cell) of the endosymbionts of S. pistillata ( Figure 3 A and Figure 2 F), partly due to a significant increase in chlorophylls content per cell. This suggests that the activity of the photosynthetic apparatus is impacted by the nitrogen availability and that the in hospite symbionts are CO 2 limited. The positive effect of bicarbonate supplementation also translated into a significant decrease in NPQ and DPS xanthophylls ( Figure 3 ). These results are indicative of lower excitation pressure on Symbiodinium ’s photosynthetic apparatus in the enriched conditions (see below for more explanations). A DIC limitation of in hospite symbionts is in agreement with previous studies, which reported an increase in photosynthesis per symbiont cell or chlorophyll content [ 15 , 20 , 21 ]. Tansik et al. [ 15 ] postulated that the cause of such DIC limitation of photosynthesis might be related to the activity of the coral external carbonic anhydrase and therefore, to the host regulation of DIC delivery to endosymbionts. The major finding of this study is the co-limitation of symbiont photosynthesis by inorganic nitrogen and carbon. Indeed, several symbiont parameters, such as the chl and carotenoid content per cell, the rates of oxygen production per cell and the rETR were significantly higher under the double iC & iN enrichment compared to all other conditions after 21 days of incubation ( Figure 2 D,F, Figure S2 and Figure 3 A). Concomitantly, symbionts showed the highest decrease in NPQ (ca. 30% Figure 3 C) and a constant low de-epoxidation state of the xanthophylls compared to control ( Figure 3 D). These results suggest that the excitation pressure on Symbiodinium ’s photosynthetic apparatus in the enriched conditions is lower and more energy is allocated to photochemistry. The 50% reduction in the O 2 -dependent rETR at the induction of photosynthesis ( Figure 3 B) comes in support to this hypothesis. Indeed, the O 2 -dependent rETR is a parameter estimating the amplitude of AEF involving O 2 as the final electron acceptor, and has been demonstrated to be related to the Mehler reaction (i.e., electrons originating from PSII that reduce O 2 at the PSI acceptor side) in Symbiodiniaceae [ 30 ]. As the Mehler-reaction acts as a photoprotective mechanism, a decrease of O 2 -dependent rETR is indicative of an increased availability of electron acceptors downstream PSI. The evolution of the relationship between rETR PSII and the O 2 net exchange rates during steady state photosynthesis ( Figure 4 ) is also consistent with the above results. Such relationship became more linear at the highest light intensities (> 300 µmol photons m −2 ·s −1 ) after 3 weeks of the double iC & iN enrichment suggesting that coral symbionts have a more efficient light phase, very likely because more electron acceptors downstream PSI are available at the induction of photosynthesis and during steady state. It also suggests that rapid variations in light intensity (e.g., underwater light flecks) and the excess energy that co-occurs, could be more easily withstood by the photosynthetic apparatus of the endosymbionts. The coupling between photosynthesis and nitrogen metabolism is well-known in microalgae [ 55 ]. Since N is involved in the composition of photosynthetic pigments like chlorophylls and phycobilins, N limitation was shown to change the ratio of chlorophyll- a to accessory pigments, to cause a reduction in thylakoid absorptivity, to negatively impact photosynthetic enzymes such as Rubisco, as well as the efficiency of energy transfer through the photosynthetic chain (reviewed in [ 56 ]). In corals, once ammonium is absorbed by the endosymbionts, it is assimilated into glutamate and glutamine thanks to the glutamine synthetase (GS)/glutamate synthase (GOGAT) cycle [ 57 ]. This metabolic pathway is driven by energy and reducing power derived from photosynthesis (reduced Ferredoxin, ATP and NAD(P)H)) and the synthesized products are the precursors for the biosynthesis of major N compounds (amino acids, nucleic acids, chlorophylls, secondary metabolites…) [ 58 ]. All together these results suggest that the increased availability of both HCO 3 − and NH 4 + improves the photosynthetic efficiency and capacity by providing more electron sinks through the Calvin-Benson cycle and the ammonium assimilation pathway. Moreover, by positive feedback, the resulting higher photosynthetic activity provides more carbon molecules onto which nitrogen can be fixed and thus impacts positively the symbiont biomasses. 4.2. The Increased Availability of HCO 3 − and NH 4 + Sustains the Growth of the Holobiont When expressed per unit of coral surface area, gross photosynthesis rates, which are now a proxy of the total amount of photosynthetic carbon acquired by the holobiont, were increased in all enriched conditions ( Figure 2 E). Such a response at the holobiont level has been previously reported on S. pistillata when supplemented with bicarbonate [ 59 ] or ammonium [ 33 , 34 , 36 ]. In our study it is related to the increased symbiont biomass that occurred within a week. While symbiont growth limitation by nitrogen availability is relatively well-known (e.g., [ 33 , 35 , 37 , 60 , 61 ]) it has been less reported for CO 2 /HCO 3 − availability, alone or in combination with iN, which is a new observation in coral studies. The stimulation of symbiont growth observed here suggests that the enhancement of photosynthesis is accompanied by a higher carbon fixation in the symbionts. A co-limitation of primary productivity by carbon and nitrogen has been well studied in terrestrial plants and macroalgae. In these models, it was demonstrated that N uptake and assimilation rates were significantly enhanced following culture at high pCO 2 (e.g., [ 62 , 63 , 64 ]). The increased demand in mineral nutrients such as nitrogen and phosphorus is generally needed to support a higher growth rates of plants and algae under high iC availability. It has to be noticed that the uptake rates of iN observed in the present study were also always higher under the control and iC enrichments ( Figure S1 ), suggesting higher needs in iN under these two conditions; however, since the uptake rates of iN were measured at the holobiont level (host + symbionts + microbiome), we cannot conclude about the occurrence of a similar effect in coral endosymbionts. In addition, a meta-analysis on plants demonstrated that high iC levels significantly decreased nitrogen concentrations in plant tissues by 2 to 57% depending on species, as a consequence of a combination of increase carbohydrate concentrations, starch accumulation, decreased investment in Rubisco, and changes in tissue chemical composition (reviewed in [ 65 ]). Although this remains to be tested, similar changes might have occurred in our coral symbiosis following iC enrichment, which have requested higher nitrogen acquisition by the symbionts, only possible, in oligotrophic waters, with an additional input of iN. In turn, the increased nitrogen input has promoted Symbiodinium proliferation within host tissues, resulting very likely in a greater competition for CO 2 /HCO 3 − between individual endosymbiotic cells [ 66 ]. Therefore, when CO 2 is provided concomitantly with nitrogen, photosynthetic rates at the symbiont and the holobiont levels are further improved. The enhancement of areal gross photosynthesis does not necessarily result in enhanced carbon translocation into the host tissue, and thus enhanced host benefit. This is due to the fact that growing symbiont populations might turn into parasitism (keeping nutrients for their own needs) under certain conditions such as thermal stress or nutrient enrichment [ 67 , 68 , 69 ]. The impact of nutrient enrichment on the coral-algae symbiosis is however dependent on the forms (organic versus inorganic, ammonium versus nitrate) and nitrogen:phosphorus ratio [ 29 , 70 ]. Although nutrient supply in the form of plankton particles or ammonium, increases both the areal rates of photosynthesis and of carbon translocation in the same S. pistillata species than as used here [ 71 , 72 ], no study has investigated the effect of a double enrichment in inorganic carbon and nitrogen on the translocation rates of photosynthates. Nevertheless, the fact that calcification rates were similarly increased in all enriched treatments (either with iC, iN or iN & iC) compared to the control conditions suggest that the improvement in the photosynthetic efficiency of the symbionts, and the overall increase in the areal rates of photosynthesis benefit to the coral host. Indeed, calcification is an energy-demanding process, which cannot be achieved under sub-optimal nutritional conditions [ 73 ]. Some previous studies actually linked increased calcification rates with high bicarbonate/carbonate concentrations [ 20 , 57 , 74 ]. They explained the effect by the establishment of a high aragonite saturation state (Ω aragonite), favorable for calcification, or by a suppression of competition between calcification and photosynthesis for iC. However, the results obtained here with iN enrichment alone suggest that the enhancement of calcification in our experiment was not linked to a higher Ω aragonite, or to a direct nitrogen or carbon effect on the calcification process, but rather to higher rates of photosynthesis, which provided more energy, as well as organic molecules for the synthesis of the organic matrix of the coral skeleton [ 75 ]. As previously observed [ 76 ], the increase in calcification rate under the iC&iN enrichment was much lower than the photosynthetic increase (22% against 133%), suggesting that coral calcification was already at a sub-optimal rate under non enriched conditions or that it was limited by other factors, independent of the photosynthetic process. Eutrophication, with excessive amount of N, P, and CO 2 , and insufficient amount of dissolved O 2 , is becoming a serious problem causing a global deterioration of reef environments [ 77 , 78 ]. Our results however suggest that there is an interactive effect of high iC and iN on the photosynthesis and general metabolism of scleractinian corals like S. pistillata which will have to be further studied. In particular, it is well-known that nitrate, which is the main by-product of eutrophication, has a different effect on coral metabolism than ammonium [ 29 , 30 , 70 , 71 ], in particular under low phosphorus availability. Future studies should therefore aim to investigate the combined effects of all these nutrients when considered under their different forms and availabilities."
} | 4,803 |
39720189 | PMC11667706 | pmc | 3,623 | {
"abstract": "Polyethylene Terephthalate (PET) is a petroleum-based plastic polymer that, by design, can last decades, if not hundreds of years, when released into the environment through plastic waste leakage. In the pursuit of sustainable solutions to plastic waste recycling and repurposing, the enzymatic depolymerization of PET has emerged as a promising green alternative. However, the metabolic potential of the resulting PET breakdown molecules, such as the two-carbon (C2) molecule ethylene glycol (EG), remains largely untapped. Here, we review and discuss the current state of research regarding existing natural and synthetic microbial pathways that enable the assimilation of EG as a carbon and energy source for Escherichia coli . Leveraging the metabolic versatility of E. coli , we explore the viability of this widely used industrial strain in harnessing EG as feedstock for the synthesis of target value-added compounds via metabolic and protein engineering strategies. Consequently, we assess the potential of EG as a versatile alternative to conventional carbon sources like glucose, facilitating the closure of the loop between the highly available PET waste and the production of valuable biochemicals. This review explores the interplay between PET biodegradation and EG metabolism, as well as the key challenges and opportunities, while offering perspectives and suggestions for propelling advancements in microbial EG assimilation for circular economy applications.",
"conclusion": "7 Conclusion PET biodegradation and assimilation represents a promising avenue with enormous potential for sustainable plastic waste management. This allows the sustainable conversion of PET into valuable NGFs such as EG, thereby alleviating the dependency on other commonly used feedstocks like glucose. Recent advancements in enzyme engineering and microbial biotechnology are paving the way towards enhancing the efficiency of PET biological depolymerization, although careful and complex optimization of PET hydrolases and metabolic pathways is still needed to reach a fully efficient bioprocess. By integrating this process into a circular economy framework, E. coli can be further engineered to utilize EG as a primary carbon source, channelling it into central metabolic pathways to produce value-added products. Continued research and development in optimizing both PET biodegradation and EG assimilation in E. coli will be critical for maximizing the efficiency and viability of this bioconversion process. In conclusion, by harnessing the power of biotechnology, synthetic biology, and systems biology, researchers will be able to unlock the full potential of EG as a renewable microbial feedstock and pave the way for a more sustainable and resource-efficient future.",
"introduction": "1 Introduction Polyethylene Terephthalate (PET) is a petroleum-based plastic polymer that is designed to endure for decades, if not centuries, which is highly disadvantageous when released into the environment via plastic waste leakage. PET is produced from a condensation reaction between ethylene glycol (EG) and terephthalic acid (TPA) ( Fig. 1 ), two chemicals commonly derived from petroleum feedstock ( Benyathiar et al., 2022 ). In 2021, it was estimated that global production reached 390.7 Mt of plastic, with 90.2% still originating from fossil-based feedstock ( Plastics Europe, 2022 ). Although PET accounted for 6.2% of the annual global plastic production, this polymer is primarily and extensively used in single-use packaging, such as beverage bottles and food containers, as well as in textile manufacturing (polyester fibres), which are highly prone to being downcycled or mismanaged (Grant et al., 2022; Plastics Europe, 2022 ). In Europe, despite PET being one of the most recyclable plastics, the majority of PET waste is not currently managed within a circular model (Grant et al., 2022). Fig. 1 Chemical reaction between terephthalic acid (TPA) and ethylene glycol (EG) to form polyethylene terephthalate (PET) . The PET polymer has the Resin Identification Code 1 and can be identified with acronym PET or PETE. Fig. 1 Consequently, plastic waste management and recycling are widely discussed topics today due to environmental concerns. The most commonly used recycling method is mechanical recycling, which entails grounding and melting the plastic waste into plastic pellets by extrusion. This process can cause undesirable changes in polymer properties in every cycle, limiting its reusability ( Achilias and Karayannidis, 2004 ; Benyathiar et al., 2022 ). Alternatively, PET can be chemically recycled, partially broken down into small oligomers and other chemical substances, or completely into its monomer units, which can be repolymerized into a new oligomer. However, chemical recycling processes such as methanolysis and glycolysis are energy-intensive, requiring high pressures and temperatures, as well as downstream separation and purification steps. Additionally, the reaction catalysts employed are often difficult to remove from the final reaction, cannot be reused, and pose environmental threats ( Achilias and Karayannidis, 2004 ; Benyathiar et al., 2022 ). Moreover, recycled PET faces challenges in competing with virgin PET in terms of availability, price, and quality, resulting in only a small fraction of the PET produced worldwide being recycled ( Grant et al., 2022 ). In 2019, 91.25% of plastic was still either buried in landfills, incinerated for energy generation, or mismanaged ( Benyathiar et al., 2022 ; OECD, 2023 ). The Great Pacific Garbage Patch serves as just one striking example of plastic waste being lost into the environment ( National Geographic, 2024 ). Plastic waste pollution has thus been recognized as one of the most significant global environmental challenges of our lifetime. In the quest for more sustainable alternatives to PET recycling and repurposing, researchers have focused on the biodegradation of PET using enzymes and engineered microorganisms. This pursuit aims to develop a biological recycling process, marking a significant area of research in recent years. Many examples include the application of enzymes in vitro that are capable of degrading the PET polymer along with derived engineered enzyme variants with catalytically improved PET breakdown capabilities ( Khairul Anuar et al., 2022 ; Sui et al., 2023 ). Some of the most prominent cases include FAST-PETase, a mutant PETase from Ideonella sakaiensis engineered using a machine learning algorithm, which can completely degrade a number of different untreated postconsumer-PET products within one week, and even as quickly as 24 h ( Lu et al., 2022 ). Another example is a variant of the leaf-branch compost cutinase (LCC), which degrades 90% of pretreated post-consumer PET in just 10 h ( Tournier et al., 2020 ). Notably, some cases are already taking their first steps in industrial application ( Carbios, 2021 ). Microorganisms have also been engineered to biodegrade PET in vivo , with some examples including E. coli ( Benavides Fernández et al., 2022 ) but also some other species such as the non-conventional yeast Yarrowia lipolytica ( Kosiorowska et al., 2022 ). Recent research has demonstrated how TPA resulting from in vitro enzymatic digestion can be repolymerized to create a new PET film with similar properties to virgin PET ( Tournier et al., 2020 ). However, production of new PET from biologically obtained monomers has only recently started to be studied and is still not well developed. This process faces several associated hurdles, particularly regarding purity requirements. For instance, purifying EG from enzymatically obtained PET hydrolysate can be difficult and costly due to its high solubility in water and elevated boiling temperature, thus requiring complex and energy-intensive separation processes and distillation steps ( Tournier et al., 2020 ; Wong et al., 2023 ). In alternative, if PET waste can be cost-effectively biodegraded, the resultant hydrolysed PET monomers could be directly used by microorganisms to fuel their metabolism and produce useful molecules and target compounds. Particularly, the two-carbon (C2) molecule EG holds great promise for fuelling microbial metabolism as a next-generation feedstock (NGF) for bioprocesses, even being considered a viable candidate for replacing commonly used microbial feedstocks such as glucose ( Pandit et al., 2017 ). Thus, PET could potentially be recycled into new added-value compounds in a circular economy manner through microbial assimilation of EG. However, for efficient assimilation and/or upcycling of PET-derived EG, it is crucial to understand the EG assimilating pathways, as well as comprehend how bacteria can be engineered for improved EG assimilation and/or conversion. This review aims to survey the known natural and synthetic pathways for microbial EG assimilation and recent advancements in metabolic engineering (ME) strategies. Specifically, our focus lies on the exploitation of synthetic pathways for consuming EG as a carbon source to support microbial growth and produce compounds of interest, focusing on the industrially well-established bacterium Escherichia coli . We further identify and discuss possible metabolic bottlenecks as well as other associated challenges with an EG assimilation bioprocess using PET as feedstock."
} | 2,344 |
31348998 | null | s2 | 3,624 | {
"abstract": "Microbial-based chemical synthesis serves as a promising approach for sustainable production of industrially important products. However, limited production performance caused by metabolic burden or genetic variations poses one of the major challenges in achieving an economically viable biomanufacturing process. To address this issue, one superior strategy is to couple the product synthesis with cellular growth, which renders production obligatory for cell survival. Here we create a pyruvate-driven metabolic scenario in engineered Escherichia coli for growth-coupled bioproduction, with which we demonstrate its application in boosting production of anthranilate and its derivatives. Deletion of a minimal set of endogenous pyruvate-releasing pathways engenders anthranilate synthesis as the salvage route for pyruvate generation to support cell growth, concomitant with simultaneous anthranilate production. Further introduction of native and non-native downstream pathways affords production enhancement of two anthranilate-derived high-value products including L-tryptophan and cis, cis-muconic acid from different carbon sources. The work reported here presents a new growth-coupled strategy with demonstrated feasibility for promoting microbial production."
} | 316 |
34135443 | PMC8209028 | pmc | 3,625 | {
"abstract": "A polydimethylsiloxane (PDMS)/Cu superhydrophobic composite material is fabricated by wet etching, electroless plating, and polymer casting. The surface topography of the material emerges from hierarchical micro/nanoscale structures of etched aluminum, which are rigorously copied by plated copper. The resulting material is superhydrophobic (contact angle > 170°, sliding angle < 7° with 7 µL droplets), electrically conductive, elastic and wear resistant. The mechanical durability of both the superhydrophobicity and the metallic conductivity are the key advantages of this material. The material is robust against mechanical abrasion (1000 cycles): the contact angles were only marginally lowered, the sliding angles remained below 10°, and the material retained its superhydrophobicity. The resistivity varied from 0.7 × 10 –5 Ωm (virgin) to 5 × 10 –5 Ωm (1000 abrasion cycles) and 30 × 10 –5 Ωm (3000 abrasion cycles). The material also underwent 10,000 cycles of stretching and bending, which led to only minor changes in superhydrophobicity and the resistivity remained below 90 × 10 –5 Ωm.",
"conclusion": "Conclusions We introduce a simple and low-cost method to fabricate elastic, conductive, and robust superhydrophobic PDMS/Cu composite material. We studied the replication process of PDMS/Cu from structured aluminum and showed that two-step etching with native oxide removal and nanostructure formation in separate steps results in best topography. Our process displays four major benefits: (i) it copies rigorously the nanostructures of aluminum; (ii) the deposition process is facile; (iii) we have a thick and durable conductive layer; (iv) overhang structures can be made. The first aspect is missing from simple polymer copying processes: they have difficulty reproducing the finer details of nanostructures. The second aspect is important for scaling to larger areas since electroless plating of copper is a very simple and cheap process that can be scaled up. The third aspect concerns both simplicity and durability of conductivity: thick copper film provides conductivity which cannot be achieved with e.g., silver nanoparticles or CNTs. The fourth aspect is important for oleophobicity: complete sacrifice of the aluminum substrate allows overhanging (retrograde) structures to be copied and released. Robust oleophobicity, however, remains to be demonstrated. Superhydrophobic, conductive, elastic films can find applications in e.g., antibacterial surfaces. Superhydrophobicity in itself can reduce bacterial adhesion due to reduced contacts on nanostructured surface, and conductivity and elasticity can be used for electrical, electro-thermal or mechanical detachment of biofilms. Since PDMS self-adhesive and its thickness can be controlled, the material could function as conductive superhydrophobic tape. It was found that the wetting properties of the resulting material can be optimized by choosing optimal etching processes and electroless plating parameters. These determine the surface topography as well as the thickness and structure of the copper layer and therefore control both the superhydrophobicity and the conductivity of the sample. It was also uncovered that the thickness of the copper layer leads to a trade-off between conductivity (the thicker, the better) and superhydrophobicity (which starts to decrease for thicker copper films beyond a certain point). The superhydrophobicity and the conductivity of the PDMS/Cu material were resistant toward stretching, abrasion, and bending. We showed that the material remained superhydrophobic and conductive, even after subjecting to 10,000 cycles of stretching at 50% of tensile strain and 1000 cycles of abrasion. The reason for the durable superhydrophobicity is likely the millimeter-thick PDMS backing layer. Diffusion of PDMS oligomers to copper surface provides self-healing which is expressed as persistent superhydrophobicity. While there was a gradual increase in resistivity, it always remained below 90 × 10 –5 Ωm, even after 3000 abrasion cycles or 10,000 stretching or bending cycles. This replication method can be extended to other polymers (e.g., polyurethanes) and metals (e.g., nickel or silver) to fabricate novel superhydrophobic metal/polymer composite materials with different combinations of mechanical–electrical-fluidic-antibacterial properties.",
"introduction": "Introduction Superhydrophobic surfaces (SHS) have high contact angles (CAs > 150°), low sliding angles (SAs < 10°), and low contact angle hysteresis (CAH) 1 – 3 . The lure of SHS owes to their potential applications in a wide range of fields, including desalination and anti-biofouling 2 , energy devices 3 , biomedical devices 4 , and heat transfer 5 .\n Two different models describe the wetting behavior of liquid droplets on a structured surface: Wenzel (penetration) and Cassie-Baxter (CB) (suspension) states. In Wenzel state, both hysteresis and sliding angles are high due to the penetration of the droplet into the pores of the surface structure. In the Cassie-Baxter state, trapped air within the micro/nanostructures, called the plastron, decreases the contact area between a droplet and the solid surface. Consequently, the droplet can easily roll off the surface at very small tilting angles 1 – 3 . The main factors of superhydrophobicity are surface roughness (surface topography), surface chemistry, the surface tension of the test liquid, and the CB state 5 – 7 . SHS can be fabricated via two main approaches: (1) adding low surface energy coating, such as fluoropolymers, to a structured surface, and (2) creating hierarchical micro/nanoscale topography of hydrophobic materials. The former method has been extensively applied to metal topographies such as etched aluminum, anodic alumina, and boehmite nanostructures 8 – 14 . When coated by hydrophobic films, excellent superhydrophobicity and in some cases also oleophobicity has been shown. These surfaces display anticorrosion 8 and resistance to alkaline and saline immersion, UV irradiation, and icing 13 . Fuoroalkylsilane (FDTS)-coated etched aluminum survived 50 tape peel tests but was not very good in alkaline and acidic tests 12 . Hierarchical hydrophobic materials approach can lead to more robust superhydrophobic surfaces because the wear of the coating is not involved 15 . One powerful method for fabricating topographical surfaces of hydrophobic materials is polymer replication. The original structures are either natural objects like plant leaves 16 , 17 or rough metal surfaces created by various etching and deposition techniques and their combinations 18 – 22 . Replicated materials include PDMS, polyurethane (PU), high-density polyethylene (HDPE), epoxies, ultra-high-molecular-weight polyethylene (UHMWPE), Polytetrafluoroethylene (PTFE), and CYTOP. Even though the polymer replica surfaces display high contact angles, they are often in Wenzel state with large sliding angles 20 . Therefore, fluoropolymer coatings are often employed 21 , 23 , and again the problem of film durability is encountered. A few superhydrophobic replicas without coatings have been demonstrated. HDPE replicated from etched aluminum showed excellent chemical durability and self-cleaning effect 19 . PDMS, PU, UHMWPE, and PTFE surfaces replicated from aluminum and alumina masters exhibited superhydrophobicity, but these films were not subjected to environmental or mechanical tests 18 . PDMS replicas displayed > 150° contact angles and underwent 10 000 pushing and bending tests successfully; however, sliding angles were > 90° for small droplets (5–10 µL), and only 70 µL droplets showed < 10° sliding angles 22 . The limited mechanical durability of superhydrophobic surfaces is often their Achilles’ heel due to the presence of vulnerable micro/nanostructures. Elastic materials are preferred over plastic and brittle materials because they survive abrasive tests better 20 . Epoxy pillars were covered by carbon nanotubes (CNT), which survived mechanical loads due to CNT elasticity 24 . As micro/nanostructures experience damage, the contact angle hysteresis increases, resulting in an increase of the sliding angle values 15 , 25 . Zhu et al . showed that when PDMS-coated copper mesh was subjected to 100 cycles of abrasion with the silicon carbide paper and load of 1 kg, the contact angle decreased from 156° to 139° due to the destruction of nanostructures 26 . Long et al . reported that the water contact angle of a PDMS-coated hierarchical aluminum surface decreased from 158° to 125° and the sliding angle rose from 2° to 90° after the abrasion experiments 27 . Other polymer-coated superhydrophobic surfaces have been subjected to abrasive tests, but under very mild conditions: 4 N load 13 , and 100 g load 10 . Additional metrics for durability have been proposed by Malavasi et al., including resistance to heating, alcohol immersion, and hydrocarbon immersion 11 . Electrical conductivity combined with superhydrophobicity has been demonstrated before, but not with polymer replication techniques. Several of the previous works have utilized carbon nanotube-based conductive superhydrophobic surfaces 28 , 29 . Zou et al . reported a thin (1 μm) layer of CNT and conductive polymer that were mixed to form a conductive layer for volatile organic vapor sensing with minimized moisture interference 28 . Yao et al. showed conductive superhydrophobic nanocomposite coatings by spraying SiO 2 nanoparticles and fluorinated multiwalled carbon nanotubes (MWCNTs) on a polymer substrate 29 . However, the mechanical durability of superhydrophobicity or conductivity was not reported in either of these works. Other routes to conductive superhydrophobic surfaces have included a porous electrodeposited zinc oxide film coated with a fluorinated monolayer 30 , and silver nanowires/particles also coated with a fluorinated monolayer 31 . However, due to the use of the monolayer hydrophobic coating, the mechanical durability is likely very limited. Previously, we have shown superhydrophobic TiO 2 /PDMS replicas from aluminum that were robust against mechanical abrasion and other environmental stressors 25 . They were replicated from etched aluminum which was coated by atomic layer deposition (ALD) TiO 2 , and the TiO 2 coating was transferred to PDMS when the aluminum substrate was sacrificially etched away. No fluoropolymer coating was used. The fabrication process is, however, limited by the requirement of ALD film, which is a costly step that is not available everywhere. Now, we show how a similar transfer process can create superhydrophobic surfaces that are not only robust against abrasion, stretch, and bending but also highly conductive. We introduce copper electroless plating (ELP) using copper (II) sulfate pentahydrate (CuSO 4 .5H 2 O) in an acidic bath (pH ≈ 2–3), using sulfuric acid and hydrochloric acid. Previous ELP processes have been carried out under alkaline conditions 32 , 33 . Electroless plating is a simple, efficient, and low-cost method for depositing a metallic layer. It is based on a chemical reaction between a reducing agent and metal ions. The thickness of plated metal can be easily controlled by changing, e.g., plating time and/or temperature 34 – 36 . Many types of bath solutions and reducing agents have been employed for copper electroless plating, including glyoxylic acid (pH ≈ 12) 32 , formaldehyde alkaline solution bath (pH ≈ 12) 33 , dimethylamine borane (pH ≈ 7.5) 37 . It has been found that low pH (< 7) baths result in uniform and smooth Cu film with adequate adhesion and low resistivity 37 , 38 . Hypophosphite-based bath is another solution (pH 9); however, it needs nickel ions for Cu plating 38 . Our process is composed of three main steps: aluminum etching, copper electroless plating, and PDMS casting on top of the plated copper. The aluminum substrate is then completely etched away, resulting in a PDMS-backed copper surface with topography defined by etched aluminum micro and nanostructures. The resulting material exhibits multiple size scales from tens of nanometers to tens of micrometers and can tolerate stretching (up to 10,000 cycles), abrasion (up to 1500 cycles), and bending (up to 10,000 cycles) with negligible effect on superhydrophobicity and only slightly reduced conductivity. No fluorine compounds are used, and the elimination of thin coatings is essential for durability.",
"discussion": "Results and discussion Fabrication of PDMS/Cu composite material Figure 1 a shows the schematic of our fabrication process. The substrate was a 0.4 mm thick aluminum plate (Al 6061-T6, 12.0\" × 12.0\", Online Metals). The plate was cleaned by acetone and isopropanol, followed by etching in a two-step process. First, Al was etched for three minutes in a phosphoric acid-based solution to remove surface oxide. After rinsing the sample with deionized (DI) water, the second etching was done in HCl for nine minutes to obtain hierarchical micro and nanostructures. Then, electroless plating was conducted in the acidic copper bath at room temperature. Table 1 summarizes the etching and plating conditions. The chemicals were purchased from Sigma Aldrich. After copper plating, PDMS (Sylgard 184 Silicone Elastomer Kit, Dow Chemical Co.) casting with the ratio of 10:1 (monomer to the crosslinking agent) was done on top of the plated copper film. The sample was cured at 65 °C for 3 h in an oven. Finally, the aluminum was removed by sacrificial etching in 6 M HCl for 10 min, followed by rinsing with DI water, drying at room temperature, and final curing at 65 °C for 2 h in an oven. Figure 1 ( a ) Schematic illustration of the fabrication process to obtain elastic, conductive, wear-resistant superhydrophobic composite PDMS/Cu material. The SEM images of 12 min single-step HCl etched Al at two different HCl concentrations: in ( b ) 3 M HCl, microscale protrusions were obtained and ( c ) 1 M HCl showing mostly unaffected surface. ( d ) The effect of etchant types used as a pre-etching step on contact angles and plated copper thickness (1 h plating at 20 °C). The SEM images show the morphologies of the PDMS/Cu surface obtained by ( e ) one-step and ( f ) two-step etching of Al substrate and one-hour copper plating. The two-step etched sample shows richer detail of micro and nanostructures. Corresponding advancing contact angles are shown in insets. ( g ) Digital photograph of composite with different thicknesses backing layer 3.50–0.25 mm displaying water contact angle around 170°. Table 1 The conditions of pre-etching, HCl etching, and the copper electroless plating bath. Process Chemicals Volume ratios and concentrations Amount Time (min) Pre-etching H 3 PO 4 :HNO 3 :H 2 O 20:1:5 – 3 Etching HCl 3 M – 9 CuSO 4 ·5H 2 O 98% 18 g H 2 SO 4 95–97% 28.8 mL Copper HCl 37% 17.5 µL 60 ELP PEG a – 90 mg SPS b – 1 mg DI water – 270 mL a Polyethylene glycol. b Bis-(sodium sulfopropyl)-disulfide. Aluminum etching process The first step was to optimize the aluminum wet etching conditions for enabling maximal superhydrophobicity of the final PDMS/Cu composite. One-step and two-step aluminum etching processes were explored to show the effect of the etching process on surface morphologies and contact angle values. The morphology and contact angles were evaluated from the final PDMS/Cu composite material. In the one-step etching process, an aluminum substrate was etched in 1 M, 3 M, and 6 M HCl solutions for 12 min. It was found that a rough structure with microscale protrusions can be obtained by using 3 M HCl solution in the one-step etching process (Fig. 1 b), while no significant roughness was achieved in 1 M HCl solution (Fig. 1 c). Aluminum was completely detached by 6 M HCl in 12 min. In the two-step etching technique, a pre-etching step was used for three minutes before HCl etching. We tested different types of etchants to investigate an appropriate type of solution for the pre-etching step, as provided in Table 2 . The results show the mixture of phosphoric acid, nitric acid, and DI water (a typical Al wet etchant used in microfabrication) resulted in the highest contact angle values of the final material, around 170° for both advancing and receding contact angles (ACA and RCA) (Fig. 1 d). The depth and size of structures can be controlled by changing etching conditions, such as the etchant type, concentration of etchant, time, and temperature of the etching process 39 – 41 . Similar results were obtained by Fernandez et al . who showed that a combination of different concentrations of HCl in the two-step etching process led to the uniform textured surfaces 40 . This can be explained by increasing the height/depth of structures obtained from the two-step etching process. Table 2 The etchants used for the pre-etching step and their volume ratios and etching time. Etchants for pre-etching Volume ratios Time (min) HF:H 2 O 1:3 3 HNO 3 :H 2 O 1:3 3 KOH:H 2 O 1:3 3 H 3 PO 4 :HNO 3 :H 2 O 20:1:5 3 Copper electroless plating Copper electroless plating was done on etched Al samples. Figure 1 d shows the effect of the pre-etching step on the contact angle and plated copper thickness after 1 h of plating at room temperature. The contact angles were measured after a complete process that involved aluminum etching, Cu plating, PDMS curing, and sacrificial etching of aluminum. Copper plating on 2-step etched Al resulted in 30 ± 2 μm thick film after 1 h, whereas the thickness was 20 ± 2 μm for the one-step etched sample. The thicker copper film shows more pronounced nanoscale surface features and resulted in improvement of the CA values, as can be seen in Fig. 1 e,f. Therefore, we chose to utilize the phosphoric-nitric acid pre-etching in all following experiments. To investigate the effect of Cu plating time on the morphology, CA values, and conductivity of PDMS/Cu material, samples were prepared with different plating times: 30, 60, and 180 min (Fig. 2 a–c). The morphology of the plated copper changed from micro-particles to a continuous film and the thickness of plated copper increased from 16 to 46 μm as plating time increased from 30 to 180 min. After 30 min of plating, copper micro-particles are clearly distinguishable on the etched aluminum surface (Fig. 2 a). After 60 min of plating, although copper formed a mostly continuous film, some aluminum is still visible (Fig. 2 b). These non-coated areas will allow PDMS on the surface after sacrificial etching of aluminum, resulting in composite PDMS/Cu surface. Increasing plating time to 180 min leads to formation of almost fully continuous copper layer, in which copper particles have coalesced and grown on top of each other (Fig. 2 c). PDMS is not able to diffuse through the thicker copper film, resulting in a decrease of contact angle values. Figure 2 Morphologies of Cu on etched Al after ( a ) 30 min, ( b ) 60 min, and ( c ) 180 min of electroless plating. The morphology of the plated Cu changed from micro-particles to a continuous film with increasing plating time. ( d ) EDX analysis of Al/Cu samples after 30 min, 60 min, and 180 min of Cu electroless plating. Even after 180 min of Cu plating, some Al is detectable, indicating that the copper film coverage is not full 100%. ( e ) The effect of copper plating time on CAs, copper thickness, and resistivity of the resulting PDMS/Cu material. This scanning electron microscopy (SEM) analysis is supported by energy-dispersive X-ray spectroscopy (EDX) (Fig. 2 d); the content of Al decreased from 53.5 to 13.7 wt% when the plating time increased from 30 to 180 min. The content of Cu, on the other hand, increased from 41.0 to 84.4 wt%, which indicates that Al was almost covered by a continuous Cu layer. Samples of Al/Cu without PDMS were used for EDX analyses. Copper plating time and the corresponding copper thicknesses, CAs, and the resistivity of the final material can be seen in Fig. 2 e. Increasing the copper thickness from 16 to 30 μm resulted in the increase of contact angles and reduction of resistivity. The plating time of 30 min resulted in a discontinuous copper film with a rather high resistivity. Increasing the thickness of copper from 16 to 30 μm resulted in a reduction of resistivity to 0.7 × 10 –5 Ωm which is metallic conductivity, even though it is roughly 400 times higher than that resistivity of bulk copper. Based on the results in Fig. 2 , we chose to focus on the material with 60 min Cu deposition to maximize the superhydrophobicity. Some other applications might benefit from longer deposition times to achieve lower resistivity at the expense of somewhat lower superhydrophobicity. Deposition times of less than 60 min are clearly suboptimal for both parameters. Su et al . obtained resistivity of ca. 30 × 10 –5 Ωm for their 84 μm thick Ag-nanoparticle composite 42 (40 × higher than ours, for a thicker layer), and sheet resistance of CNT-embedded SHS was 3 × 10 4 Ω/sq 29 , while our samples display sheet resistances of 0.4 Ω/sq. Luo et al . reported resistivity of 1 × 10 –4 Ωm 43 , and Wang et al . a resistivity of 2 Ωm for their superhydrophobic composite materials 44 , which are roughly 15 × and 100,000 × higher than our material, respectively. A superhydrophobic and conductive nanocomposite made of fluoropolymer and various allotropes of carbon exhibited resistivity of 1 × 10 –3 Ωm 45 , but it was not subjected to any mechanical or environmental tests except water droplet impalement. Composite thickness tests We also tested the actual thickness of the complete composite material (Fig. 1 g). The superhydrophobic properties remain the same for all samples with different PDMS backing plate thicknesses, ranging from 0.25 to 3.50 mm. The copper thickness was 30 µm for all the samples. The water contact angles were in the range of 167°–172° advancing and 166°-171° receding for all samples indicating no detectable difference between the samples. We chose to perform rest of the experiments with PDMS thickness of 1.5 mm. Mechanical durability tests The samples for mechanical durability tests were obtained by the two-step etching process and one-hour Cu electroless plating. We studied surface morphology, resistivity, and CAs of samples in a static stretch, cyclic stretch, abrasion cycle, and bending cycle tests. Static stretch tests The contact angles, contact angle hysteresis, and resistivity values of non-stretched and stretched samples at 25% and 50% of tensile strain (ε) are plotted in Fig. 3 a. These results are the CAs and resistivity of the material in a static stretched state. Figure 3 ( a ) The contact angles, contact angle hysteresis, and resistivity of non-stretched and stretched samples at 25% and 50% of tensile strain in a static stretch test. Photos of stretching test setup and the SEM images of static stretched samples at ( b ) 25% and ( c ) 50% of tensile values. The results indicate a slight change in the CAs of the PDMS/Cu material after 25% stretching; the advancing contact angle decreased from 172 ± 1° to 196 ± 1° and the receding contact angle decreased from 171 ± 1° to 168 ± 1°. For the stretched sample at 50% of tensile strain, the advancing contact angle decreased to 162 ± 1° and the receding contact angle to 158 ± 1°, resulting in an increase of contact angle hysteresis from 1° to 4°. The resistivity values increased from 0.7 × 10 –5 to 1.2 × 10 –5 Ωm for both 25% and 50% stretched samples. The surface morphology of tensile stretched samples can be seen in Fig. 3 b,c. The surface structures were preserved at 25% tensile strain but experienced deformation and distortion at 50% tensile strain. Contact angle hysteresis increase at 50% tensile strain is caused by damage to the hierarchical structures, as seen in Fig. 3 c. Cyclic stretch tests The advancing and receding contact angles and resistivity after cyclic stretch test of 1000 to 10,000 cycles are shown in Fig. 4 a,b (Movie S1 in the Supplementary Material). After 10,000 cycles of stretching (25%), the advancing contact angle decreased from 172 ± 1° to 161 ± 1° and receding contact angle was reduced from 171 ± 1° to 159 ± 1° (Fig. 4 a). In the case of 50% tensile strain, both advancing and receding contact angles declined to around 155 ± 1°, but superhydrophobic properties remained (Fig. 4 b). The SA with 7 μL droplet was around 7° after 10,000 stretching-relaxing cycles at 25% tensile strain, same as the non-stretched sample. For 50% tensile strain, the sliding angle was 9°. Figure 4 The contact angles and resistivity after being subjected to 10,000 stretching-relaxing cycles at ( a ) 25% and ( b ) 50% of tensile strain. ( c ) The SEM images of non-stretched and stretched samples after 10,000 stretching-relaxing cycles at 25% and 50% of tensile strain. After 1000 stretching cycles, the resistivity values increased to 6 × 10 –5 Ωm at 25% and to 12 × 10 –5 Ωm at 50% of tensile strain. As the stretch cycles increased from 1000 to 10,000, the resistivity showed a clear upward trend and gradually increased to around 60 × 10 –5 Ωm for both 25% and 50% of tensile strain, not dissimilar from 30 × 10 –5 Ωm obtained by Su et al. 42 . A possible reason for this gradual increase can be explained by free monomers in PDMS. As the sample is stretched, these mobile monomers can increasingly penetrate into any cracks or boundaries opened up in the copper layer by stretching. Besides, SEM images show that the surface morphologies of PDMS/Cu material after 10,000 cycles at ε = 25% and non-stretched sample are almost identical to each other. This means that hierarchical structures were preserved after 10,000 cycles of 25% stretching. There is slight deformation, but no change in overall topography can be observed after being subjected to the 10,000 stretching cycles at 50% tensile strain (Fig. 4 c). These results show that no serious damage is visible on the hierarchical structures of the surface after applying the cyclic stretch test and the superhydrophobicity of the PDMS/Cu material was preserved with CA > 150° and SA < 10°. In the case of silver-nanoparticle composite 42 , the contact angle remained above 150° and electrical resistance increased from 10 to 79 Ω after 200 cycles of stretching. If we estimate the resistivity increase of our samples at 200 stretch cycles from Fig. 4 a,b, we end up with something close to the factor of ten, in accordance with silver-nanoparticle composite. Abrasion tests To investigate the wear resistance of the PDMS/Cu material abrasion cycle test was conducted using silicon carbide paper (P 1200) (Movie S2 in the supplementary material). The contact angles, sliding angles, and resistivity of the abraded surface at different cycles are given in Fig. 5 a. Figure 5 ( a ) The contact angles, sliding angles, and resistivity of abraded material from 100 to 3000 cycles. ( b ) The SEM images of surface structures of non-abraded and abraded materials after 1500 and 3000 cycles of abrasion. ( c ) EDX analysis of PDMS/Cu material for non-abraded, after 1500 and 3000 cycles of abrasion. The advancing and receding contact angles decreased from around 170° to 150° after 1500 cycles, while no significant change in contact angle hysteresis and sliding angles was observed, and the material remained superhydrophobic. As the abrasion cycles increased beyond 1500 cycles, the contact angle hysteresis rose significantly from 2° to 9°, and SA increased from 7° to 20°. Resistivity steadily increased from 0.7 × 10 –5 Ωm (non-abraded) to 5 × 10 –5 Ωm after 1000 cycles of abrasion. As the surface was subject to more abrasion cycles (> 1000), the resistivity increased to 30 × 10 –5 Ωm after 3000 cycles due to the reduction of Cu content and the destruction of hierarchical structures. Wang et al. reported that when a fluorinated PDMS composite was subjected to the sandpaper-abrasion test, conductivity and superhydrophobicity were maintained for 80 abrasion cycles 44 , while our composite material maintains its properties for 1500 abrasion cycles. Wang et al. and Luo et al. showed that the static contact angle remained above 150° after 50–60 abrasion cycles 43 , 46 . However, no information was provided on the sliding angles or contact angle hysteresis, and thus the effect of abrasion on superhydrophobicity cannot be assessed. The surface morphologies of non-abraded and abraded samples are shown in Fig. 5 b. Although some rounding of structures is visible after 1500 cycles, the main surface features were preserved after 1500 cycles of sandpaper-abrasion test. As the material was subjected to 3000 cycles, destruction of finer structures can be observed and the superhydrophobicity was gradually lost. The EDX analysis shows that the Cu content decreased slightly from 73.4 (non-abraded) to 63.0 wt% after 1500 cycles of abrasion and then significantly decreased to 31.9 wt% after 3000 cycles of abrasion due to the damage of hierarchical structures (Fig. 5 c). The EDX results also show the increase in silicon and oxygen signals, which are likely due to more PDMS being visible after some of the copper was abraded away. On the other hand, the constant supply of PDMS monomers to the surface to coat freshly exposed copper is supposedly responsible for the long-lasting superhydrophobicity. This is very different compared to superhydrophobicity achieved via polymer coating: first of all, our polymer is mostly under the surface; and second, we have millimeter thickness, which is an ample supply. Nanostructured copper surface with copper oxide (CuO) has been found to exhibit low friction and improved wear resistance because wear debris is not on the surface but inside the microstructures 47 . Only few conductive superhydrophobic surfaces have undergone mechanical durability tests. These include a fluorinated PDMS composite 44 and a rubber composite 46 . The superhydrophobicity and conductivity were maintained after 1000 stretching cycles and 80 abrasion cycles for the PDMS composite 44 and after 100 stretching cycles and 60 abrasion cycles for the rubber composite 46 . Luo et al . showed the static contact angle of polypropylene/polydopamine/Ag-nanoparticle/PDMS composite remained above 150° and conductivity decreased from 89.1 to 60.0 S/cm after 50 abrasion cycles (3X less conductive than our sample after 1000 abrasion cycles) 43 . Polymer-modified Ag-nanoparticles on natural rubber were reported to exhibit superhydrophobicity, elasticity, and conductivity without fluorination 42 . These structures were not subjected to abrasive tests, even though they survived kneading, torsion, and water impact tests. Bending tests Figure 6 a shows the advancing, receding contact angles, and resistivity after cyclic bend test from 1000 to 10,000 cycles (Movie S3 in the supplementary material). The results show a small decrease in both advancing and receding contact angles after 10,000 bending cycles; the advancing contact angle decreased to 163 ± 1° and the receding contact angle was reduced to 160 ± 1°, so the superhydrophobic property of the material was preserved after cyclic bend test. The resistivity increased to 5 × 10 –5 Ωm after 1000 bending cycles, and to 90 × 10 –5 Ωm after 10,000 cycles due to deformation of hierarchical structures. The resistivity increase was more pronounced after 10,000 bending cycles compared to 10,000 stretching cycles. This is likely due to compressive failure that occurred in the bending test (Fig. 6 b). Resistivity as a function of bending was not reported by Yao et al. and Asthana et al. 29 , 45 As can be seen in Fig. 6 b, the surface did not show any changes after 1000 cycles, while significant deformation in the surface structures can be observed after being subjected to 10 000 bending cycles. Figure 6 ( a ) The advancing, receding contact angles and resistivity after 10,000 bending cycles. ( b ) The SEM images of samples after 1000 and 10,000 cycles of bending."
} | 7,986 |
30975146 | PMC6460773 | pmc | 3,627 | {
"abstract": "Background In the last years, different biotechnologically relevant microorganisms have been engineered for the synthesis of plant polyphenols such as flavonoids and stilbenes. However, low intracellular availability of malonyl-CoA as essential precursor for most plant polyphenols of interest is regarded as the decisive bottleneck preventing high product titers. Results In this study, Corynebacterium glutamicum , which emerged as promising cell factory for plant polyphenol production, was tailored by rational metabolic engineering towards providing significantly more malonyl-CoA for product synthesis. This was achieved by improving carbon source uptake, transcriptional deregulation of accBC and accD1 encoding the two subunits of the acetyl-CoA carboxylase (ACC), reduced flux into the tricarboxylic acid (TCA) cycle, and elimination of anaplerotic carboxylation of pyruvate. The constructed strains were used for the synthesis of the pharmacologically interesting plant pentaketide noreugenin, which is produced by plants such as Aloe arborescens from five molecules of malonyl-CoA. In this context, accumulation of the C 1 /C 6 cyclized intermediate 1-(2,4,6-trihydroxyphenyl)butane-1,3-dione (TPBD) was observed, which could be fully cyclized to the bicyclic product noreugenin by acidification. Conclusion The best strain C. glutamicum Nor2 C5 mu fasO BCD1 P O6 - iolT1 ∆ pyc allowed for synthesis of 53.32 mg/L (0.278 mM) noreugenin in CGXII medium supplemented with casamino acids within 24 h. Electronic supplementary material The online version of this article (10.1186/s12934-019-1117-x) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusion In the present work, we applied the knowledge of the well-characterized central carbon metabolism to tailor C. glutamicum towards increased malonyl-CoA availability by rational metabolic engineering. Additionally, our work contributes to a better understanding of the PCS Aa reaction mechanism as we could detect the intermediate TPBD, which spontaneously cyclizes to noreugenin. Acidification accelerated TPBD conversion to the product allowing for a titer of 53.32 mg/L (0.278 mM) noreugenin. Taken together, the constructed strain C. glutamicum C5 mu fasO BCD1 P O6 - iolT1 ∆ pyc represents a promising strain for the microbial production of noreugenin and other malonyl-CoA derived products.",
"discussion": "Discussion In this study, we describe engineering of C. glutamicum for increased malonyl-CoA availability and microbial synthesis of the plant pentaketide noreugenin originally found in the medicinal plant A. arborescens . Key to success was the N-terminal truncation of the type III PKS enabling noreugenin synthesis. Originally, the ten N-terminal residues of PCS Aa were believed to contribute to the formation of an expanded surface-exposed loop [ 34 ]. However, since this short amino acid stretch is not present in closely related CHS- and STS-enzymes, which could be already functionally implemented in C. glutamicum , we deleted the first 30 nucleotides of the pcs Aa gene. Eventually, this modification drastically improved noreugenin synthesis. A common approach in the context of improving malonyl-CoA availability for product synthesis aims at increasing ACC activity in the respective microbial workhorse [ 2 ]. Previously, by abolishing Snf1-dependent posttranslational regulation of Acc1, malonyl-CoA availability in Saccharomyces cerevisiae could be increased for the synthesis of fatty acid ethyl esters and 3-hydroxypropionic acid [ 37 ]. Another strategy for increasing ACC activity is overexpression of genes encoding for the respective subunits of this enzyme complex. For example, episomal overexpression of four genes coding for a heterotetrameric ACC from Photorhabdus luminescens in E. coli increased the titer of the malonyl-CoA-derived polyphenol pinocembrin sevenfold, yielding 196 mg/L of this product [ 38 ]. Furthermore, the active AccBC-AccD1 heterodimer of C. glutamicum could be functionally introduced into E. coli [ 39 ]. By following this strategy, the intracellular malonyl-CoA availability in E. coli was increased 15-fold allowing for the synthesis of 1.3 g/L phloroglucinol [ 40 ]. In this study, we developed a different strategy aiming for an increased expression of the genomically encoded ACC genes in C. glutamicum , instead of following the more traditional approach of episomal overexpression of (heterologous) genes. By mutation of the fasO -sites upstream of accBC and accD1 , repression of gene expression mediated by the transcription regulator FasR could be repealed, which almost tripled intracellular malonyl-CoA availability and thus enabled increased product formation. Important to note, mutation of both FasR binding sites was required to increase ACC activity, probably because a functional ACC requires both subunits, AccBC and AccD1, in equimolar amounts [ 39 , 41 ]. Furthermore, the intracellular malonyl-CoA availability was positively influenced by increasing glucose uptake through deregulation of the gene for the glucose/ myo -inositol permease IolT1. However, this had only limited beneficial effects on noreugenin synthesis at shaking flask scale, but might be beneficial for large-scale applications when considering important process criteria such as space–time yield. Efficacy of this strain modification was already demonstrated for the de novo synthesis of hydroxybenzoic acids in C. glutamicum from glucose [ 30 ]. Potential other strategies for increasing the intracellular malonyl-CoA availability include establishing an ACC-independent pathway for malonyl-CoA synthesis from supplied malonate through heterologous expression of genes for malonate uptake and CoA-activation of malonate originating from the malonate assimilation pathway in Rhizobium trifolii [ 42 ]. This particular approach enabled a 15-fold increase of pinocembrin production using E. coli [ 13 , 43 ]. Furthermore, state-of-the-art techniques for gene silencing or downregulation such as CRISPR interference (CRISPRi) or methods employing synthetic small regulatory RNA (sRNA) libraries or synthetic antisense RNA (asRNA) could be used to specifically knock down genes encoding for malonyl-CoA consuming enzymes [ 23 , 44 , 45 ]. Moreover, we predicted and detected the TPBD intermediate as the actual product of PCS Aa and could show that formation of TPBD is strictly limited to the exponential growth phase in C. glutamicum in which malonyl-CoA is exclusively supplied. This is in line with our observations regarding flavonoid- and stilbene synthesis with C. glutamicum [ 10 , 15 ]. Through HCl acidification of acetonitrile extracts we achieved full conversion of TPBD to noreugenin within 3 h allowing to significantly reduce the overall cultivation time. Very recently, it could be demonstrated that composition of the defined CGXII medium might not be optimal for the expression of plant-derived genes involved in anthocyanin synthesis in C. glutamicum [ 35 ]. In this particular study, supplementation of the defined AMM medium with 2 g/L casamino acids significantly improved heterologous gene expression and thus anthocyanin synthesis. Casamino acids, obtained through acid hydrolysis of casein, represent a valuable source of all proteinogenic amino acids except tryptophan [ 46 , 47 ]. Although it could be shown here that supplementation of casamino acids also promotes noreugenin synthesis, we prefer the simplicity of defined CGXII medium, especially as the composition of casamino acids varies from supplier to supplier. Nevertheless, this supplementation strategy is helpful to meet possible future challenges connected to heterologous gene expression. Alternatively, translational fusions of the target protein with the maltose-binding protein MalE from E. coli can be generated. This strategy already proved to be beneficial for the functional expression of a plant-derived O -methyltransferase gene from Vitis vinifera in C. glutamicum [ 11 ]."
} | 2,016 |
39392487 | PMC11469988 | pmc | 3,628 | {
"abstract": "The plant mycobiome plays a crucial role in the host life cycle, influencing both healthy and diseased states, and is essential for plant tolerance to drought. In this study, we used ITS metabarcoding to investigate the fungal community of the drought-resistant plant Malva sylvestris L. in Morocco along a gradient of precipitation, encompassing subhumid and semi-arid environments. We sampled three biotopes: rhizosphere, bulk soil, and root endosphere. Our findings revealed an absence of beta-diversity differences between bulk soil and rhizosphere, indicating that the plant does not selectively influence its rhizosphere mycobiome. Additionally, ASVs belonging to the genus Alternaria represented up to 30% of reads in the plant’s roots and correlated with drought ( p = 0.006), indicating a potential role for this fungal genus in mitigating drought, possibly as part of the dark septate endophyte group. Root staining and microscopic observation revealed extensive colonization by fungal hyphae and microsclerotia-like structures. Furthermore, ASVs identified as Fusarium equiseti were also correlated with low precipitation and recognized as a hub taxon in the roots. However, it remains uncertain whether this species is pathogenic or beneficial to the plant. These insights contribute to our understanding of the plant mycobiome’s role in drought tolerance and highlight the importance of specific fungal taxa in supporting plant health under varying environmental conditions. Future research should focus on characterizing these taxa’s functional roles and their interactions with the host plant to further elucidate their contributions to drought resistance. Supplementary Information The online version contains supplementary material available at 10.1007/s00248-024-02439-3.",
"conclusion": "Conclusion M. sylvestris appears to host a few fungal taxa in its roots that may aid in coping with drought stress, with the most promising being a highly abundant ASV from the genus Alternaria . Other taxa of interest include Fusarium equiseti and taxa belonging to the Mortierella genus. In contrast, the rhizosphere mycobiome does not seem to be strongly selected by the plant, but it may respond to the physico-chemical characteristics of the soil. Further investigations, including the isolation of fungal endophytes from the roots of M. sylvestris and inoculation experiments, are needed to determine the effects of these endophytes on the growth and drought tolerance of the plant. Functional studies will also provide insights into the mechanisms employed by both partners to alleviate drought stress.",
"introduction": "Introduction Developing adaptive strategies to address climate change events is unequivocally one of the most pressing challenges of the twenty-first century and beyond. Among these climate-related issues, increasing aridity, prolonged periods of drought, and excessive and erratic rainfall are of major concern [ 1 ]. Recent revelations indicate that prolonged drought and erratic rainfall are negatively impacting global food security [ 2 ]. In Morocco, long-term drought and erratic rainfall decreased agricultural productivity by approximately 17% in 2022 [ 3 ]. Plant mycobiomes—the fungal communities inhabiting plant roots, rhizospheres, and phyllospheres—are an integral component of plant biology that can mediate plant adaptation to extreme conditions such as heat and drought. The role of symbiotic fungal associations in plants’ tolerance to abiotic stress has been the focus of many recent studies [ 4 , 5 ]. Beneficial associations involving mycorrhizal and endophytic fungal communities can enhance the phenotypic and metabolic plasticity of host plants [ 6 , 7 ], resulting in improved tolerance to stresses such as drought and salinity [ 8 ]. A group of endophytic plant root mycobiome known as dark septate endophytes (DSE) have been found to extensively colonize many plant roots including trees and herbaceous plants [ 9 ]. Despite the growing awareness about the importance of this fungal group in plant growth promotion, phosphorus acquisition, and drought tolerance in arid ecosystems [ 10 , 11 ], little is known about their selection, recruitment, and turnover along arid bioclimatic gradients. Although DSE can be identified by the presence of melanized and hyaline septate hyphae, as well as microsclerotia in the intraradical roots [ 7 ], high-throughput sequencing techniques are essential to study their diversity and distribution and to identify their role in plant mycobiome assembly and selection. Environmental perturbations, such as increased warming and drought, can result in a significant turnover of plant root mycobiomes. However, the selective recruitment of specific taxa by host plants in the roots and rhizosphere can complicate the diversity and assembly of fungal communities along specific climatic or ecological gradients. For example, alpha-diversity of the fungal community in the roots and rhizosphere of castor bean ( Ricinus communis ) increased under drought compared to wet conditions; however, the higher selectivity for certain groups in drought conditions led to a greater uniformity of fungal communities, thus reducing species turnover [ 12 ]. In another study, Gargouri et al. [ 13 ] reported that while Opuntia ficus-indica mycobiome exhibited similar alpha-diversity across an aridity gradient, there was a significant difference in the beta-diversity. The relative abundances of major phyla differed between the rhizosphere and endosphere, and distinct hub taxa were recruited at each biotope within different bioclimatic zones [ 13 ]. Therefore, further assessment of plant mycobiomes is essential to understand the ecology of plant-associated fungal communities across spatially diverse aridity zones. Malva sylvestris L., commonly known as common mallow, belongs to the Malvaceae family and is widely distributed across the globe. The plant thrives along a broad gradient of aridity and demonstrates resilience to both drought and excessive rainfall [ 14 ]. M. sylvestris can grow in marginal or resource-poor ecosystem; this property enables the plant to play a role as a pioneering plant in the ecological succession of barren or disturbed habitats. Due to its ability to grow in both wet and dry climates, M. sylvestris is also a promising candidate for studying the ecological succession of plant mycobiomes across aridity gradients [ 15 ]; however, there is limited information on the ecology of M. sylvestris mycobiome. In a recent study conducted by Legeay et al. (2024) [ 15 ], Rhizobium dominated the root bacteriome of M. sylvestris , but this observation was not related to any specific variable in the study. Interestingly, the study also found no significant differences in beta-diversity between the rhizosphere and bulk soil, suggesting that M. sylvestris might exhibit weak selective recruitment of bacteriome in its rhizosphere. Therefore, this study aims to investigate the mycobiome of the plant M. sylvestris in both agricultural and natural environments to understand how fungal community structures change, particularly in arid regions, and to identify fungal taxa associated with the plant’s drought tolerance. Based on previous bacterial studies, we hypothesized that (i) M. sylvestris plants do not select specific fungal communities in their rhizosphere, and (ii) M. sylvestris plants harbor drought-responsive taxa in their roots. To test these hypotheses, we sampled 101 individuals from 13 sites across Morocco, including five agricultural and eight natural areas along a gradient of precipitation. ITS of rRNA gene metabarcoding was employed to investigate fungal communities in the bulk soil, rhizosphere, and root endosphere of M. sylvestris .",
"discussion": "Discussion There was no significant difference in alpha-diversity between bulk soil and the rhizosphere, nor was there a difference in beta-diversity in semi-arid environments. Similarly, beta-diversity was not significantly different between bulk soil and the rhizosphere, suggesting that there is no selective recruitment of the fungal community in the rhizosphere of M. sylvestris . This observation aligns with findings in the bacteriome of M. sylvestris [ 15 ] and supports our initial hypothesis. Interestingly, in a study on sorghum under drought conditions, Gao et al. [ 23 ] found an increase in stochasticity in the rhizosphere mycobiome, which they linked to a decrease in selection by the host. Similarly, the rhizosheath mycobiome of herbaceous plants in the Namib Desert did not show a strong selection by the plant [ 24 ]. One possible explanation is that M. sylvestris may conserve resources that would otherwise be used to manage its rhizosphere environment. However, unlike the bacteriome of M. sylvestris , which showed higher alpha-diversity of bacterial communities in the rhizosphere, the rhizosphere in this study did not appear to serve as a refuge for fungal diversity, as alpha-diversities were not significantly different between the rhizosphere and bulk soil. This could be due to drought stress having a greater impact on bacterial communities than on fungal communities [ 25 ], and/or the lack of beneficial conditions for fungi in the rhizosphere. Root and rhizosphere communities were highly sensitive to K, N, and C concentrations, while bulk soil communities remained unaffected. The bulk soil and rhizosphere communities responded to precipitation levels, unlike root communities, which showed no significant beta-diversity differences linked to precipitation. Additionally, rhizosphere communities were significantly influenced by the plant phenotype, whereas root communities were not. The sensitivity of rhizosphere communities to precipitation and plant phenotype, which is not observed in root communities, suggests that the roots of M. sylvestris provide refuge for fungal communities from drought stress. Alternatively, this could indicate that the selective filter imposed by the root biotope is stronger than the one imposed by dry conditions. Since some significant taxa within the roots, including those classified as “core taxa” and “hub taxa,” still respond to arid conditions, the second mechanism seems more likely. It is somewhat surprising that the rhizosphere communities differ between the two plant phenotypes, despite the lack of beta-diversity difference between bulk soil and rhizosphere. This may indicate selective recruitment of some taxa in the M. sylvestris rhizosphere under certain circumstances. This observation is supported by the fact that most of the taxa correlated with soil chemical parameters are found in the rhizosphere. However, this selective recruitment might be passive, resulting from modifications in the niche environment around the roots rather than active selective recruitment mediated by the plant [ 26 , 27 ]. Despite the lack of precipitation effects on beta-diversity in the roots, the random forest model identified some ASVs that effectively predict aridity (25% error rate). This suggests a strong interaction between aridity levels and certain taxa within the roots, even if they are not the dominant taxa. Notably, one ASV from Alternaria sp., sometimes representing up to 30% of reads, was a predictor of drought and positively correlated with total K. Another Alternaria ASV also correlated with drought in the roots. Some members of the Alternaria genus, described as dark septate endophytes, have been shown to enhance plant resistance to salinity and drought [ 28 ]. It is plausible that Alternaria spp. play a role in M. sylvestris ’s resistance to environmental stresses. Among the top 20 predictors associated with aridity, three ASVs belonged to the Mortierellomycota phylum, including one ASV in the Mortierella genus and another identified as Mortierella exigua . Mortierella species are known for their potential benefits to plants, such as enhancing P availability in the soil and exhibiting pathogen-suppressive activity [ 29 ]. Specifically, M. exigua has been noted as a potential agent for bioremediation of trace metals [ 30 ]. Mortierella species are also strongly associated with arid soils [ 31 , 32 ]. While Mortierella spp. has not been directly linked to drought stress, our results suggest that species of this genus may have some potential in this context. The remaining ASVs associated with aridity are more related to pathogenicity, including those found in humans or animals. This suggests that these taxa may exhibit opportunistic pathogenic behavior, potentially stimulated by the drought stress experienced by plants in arid regions. Nevertheless , other members of the genus Alternaria have been described as pathogens in more studies than as dark septate endophytes [ 33 ]. Another genus correlated with drought in the roots, Olpidium , is also associated with pathogenicity and reduced drought resistance [ 34 ]. Thus, it is not possible to definitively affirm a positive or negative role for Alternaria spp. in M. sylvestris without further functional and experimental studies. Another ambiguous case is the species Fusarium equiseti , a core taxon present in both roots and soil samples, with two ASVs correlated with low precipitation. F. equiseti has been shown to benefit plant growth in high-salt soils [ 35 ], but it also causes root rot in chickpeas in Morocco [ 36 ]. Therefore, it is challenging to determine from this study alone which taxa are beneficial or pathogens. In general, the findings support hypothesis 2, suggesting that M. sylvestris plants host drought-responsive taxa in their roots. Interestingly, another species of Alternaria genus, A. subcucurbitae , was identified as a core and hub taxon in the soil but was completely absent from the roots. These findings suggest that Alternaria and its associated microbial communities play a role in structuring the mycobiome of Moroccan arid environments. Alternaria is a cosmopolitan genus known for both saprophytic and pathogenic capabilities [ 36 ], including on Malvaceae plants [ 37 ]. It has been found to be the dominant genus in the foliar fungal endophytes of American plains environment, but with considerable intra-genus variation [ 38 ]. As previously mentioned, Alternaria spp. can also behave as a dark septate endophyte, and such endophytes have been shown to significantly influence the microbial communities of European forest soils [ 9 ]. M. sylvestris appears to be particularly associated with one specific species of Alternaria , which is not the most prevalent in the soil. However, it should be noted that the phylogeny of the Alternaria genus is still not well defined [ 36 ]."
} | 3,684 |
36234461 | PMC9565720 | pmc | 3,629 | {
"abstract": "Resistive random–access memory (RRAM) for neuromorphic systems has received significant attention because of its advantages, such as low power consumption, high–density structure, and high–speed switching. However, variability occurs because of the stochastic nature of conductive filaments (CFs), producing inaccurate results in neuromorphic systems. In this article, we fabricated nitrogen–doped tantalum oxide (TaO x :N)–based resistive switching (RS) memory. The TaO x :N–based device significantly enhanced the RS characteristics compared with a TaO x –based device in terms of resistance variability. It achieved lower device–to–device variability in both low-resistance state (LRS) and high–resistance state (HRS), 8.7% and 48.3% rather than undoped device of 35% and 60.7%. Furthermore, the N–doped device showed a centralized set distribution with a 9.4% variability, while the undoped device exhibited a wider distribution with a 17.2% variability. Concerning pulse endurance, nitrogen doping prevented durability from being degraded. Finally, for synaptic properties, the potentiation and depression of the TaO x :N–based device exhibited a more stable cycle–to–cycle variability of 4.9%, compared with only 13.7% for the TaO x –based device. The proposed nitrogen–doped device is more suitable for neuromorphic systems because, unlike the undoped device, uniformity of conductance can be obtained.",
"conclusion": "4. Conclusions In this study, we investigated the role of nitrogen doping on resistive and synaptic characteristics while comparing the Ta/TaO x /Pt and Ta/TaO x :N/Pt devices. The TaO x :N–based memristor exhibited uniformity of resistance state and set voltages and a 10 5 pulse endurance. Furthermore, synaptic properties such as potentiation and depression were conducted with a pulse train. The TaO x :N–based memristor exhibited more stable conductance modulation when measuring potentiation first and depression first. In contrast, the TaO x –based memristor produces variability in the I–V curves, set voltages, and device-to-device resistance. It also degrades when pulse endurance, potentiation, and depression are measured. When we use RRAM as a neuromorphic system, the device should have low resistance variability. Therefore, a TaO x :N–based device is more suitable for an artificial synapse than an undoped–TaO x device.",
"introduction": "1. Introduction New technologies such as artificial intelligence (AI) and the Internet of Things (IoT) are gaining attention, so rapid processing of vast amounts of data and information is required. However, in traditional digital computing—the von Neumann architecture—operation and storage devices are separated. Accordingly, bottlenecks occur when transferring complex data between devices [ 1 , 2 ]. Therefore, neuromorphic computing has emerged because of its parallel data processing with low power consumption and high-density structure. A neuromorphic system mimics a biological system—the human brain [ 3 , 4 , 5 ]. Recently, oxide–based device and resistance–based random–access memory designs, such as magnetic random–access memory (MRAM), ferroelectric random–access memory (FRAM), phase–change random–access memory (PRAM), spin–torque–transfer random–access memory (STT–RAM) and resistive random–access memory (RRAM) have been studied for implementation in neuromorphic systems [ 6 , 7 , 8 ]. RRAM is a promising candidate for the next generation of neuromorphic systems because of advantages such as high endurance, high–speed switching, low–power operation, scaling down capability, and multi–level cell (MLC) capability [ 9 , 10 , 11 , 12 ]. Despite these advantages, the filament–type RRAM cannot avoid large variations in conductance because the formation and rupture of filaments occur randomly, a drawback for use in neuromorphic systems [ 13 ]. Several studies have been reported to suppress the probabilistic nature of conducting filament formation, such as combining several oxide layers [ 14 ], semiconducting oxides [ 15 , 16 ], and doping techniques [ 17 ]. Nitrogen doping has been widely studied because it can control conductance accurately and effectively. Ref. [ 18 ] reported that N–doping in Ti/TiO x /Pt enhances the reliability effect on neuromorphic systems. Refs. [ 19 , 20 , 21 , 22 , 23 ] demonstrated that nitrogen doping in insulators can reduces the leakage path. Arikado, T et al., [ 24 ] stated that nitrogen eliminates oxygen vacancy related gap states by changing the charged states of V o to V o 2+, so the leakage path can be reduced. Another important effect of nitrogen in the oxide layer is the restriction of oxygen ions’ diffusion. Y. E. Syu et al., [ 25 ] stated that, due to the higher bonding energy of N–O bond than the O–O bond, nitrogen can capture the oxygen atom to localize the oxygen ion near the conducting filament. Misha et al. [ 26 ] reported that incorporating nitrogen in TaO x improves uniformity at a low operating current. The performance of N–doped memory devices compared with undoped memory device in previous reports is summarized in Table 1 . In this study, the effect of nitrogen doping on synaptic properties and DC properties was also examined, in contrast with previous studies [ 26 ]. By incorporating nitrogen, smaller variations in DC endurance, AC endurance, device–to–device resistance, and set voltages are achieved compared with the undoped film. Furthermore, conductance is efficiently and gradually modulated by pulse–train measurement in the N–doped TaO x . Therefore, we investigated enhancing RS and synaptic properties in TaO x films with nitrogen doping.",
"discussion": "3. Results and Discussion First, the electrical properties of the Ta/TaO x /Pt and Ta/TaO x :N/Pt memristors were measured. A preliminary forming process was required to produce a low–resistance state (LRS) from the initial state [ 27 , 28 ]. Both devices conducted 300 consecutive ON/OFF cycles by positive set and negative reset for the bipolar RS (BRS) after the forming process, as depicted in Figure 2 a,d. Each memristor was measured under a compliance current of 2 mA in the set process with no current limit during reset switching to ensure a fair comparison. Furthermore, a negative voltage of −1.6 V was applied to switch from the LRS to the high–resistance state (HRS), and HRS changed to LRS under a positive voltage sweep from 0 to 1.4 V. The detailed distribution of HRS and LRS values were extracted from the I−V characteristics measured at 0.2 V (V read ) during a cyclic test, as depicted in Figure 2 b,e. The TaO x –based memristor exhibited a significant difference in resistance in the set and reset operations with cycle–to–cycle resistance variabilities (σ/μ) of 27.8% (LRS) and 23.7% (HRS). However, the TaO x :N–based memristor exhibited a stable RS characteristic with relatively small cycle–to–cycle resistance variabilities of 15.7% (LRS) and 13.2% (HRS). A retention characteristic test was performed to confirm the performance of the TaO x –based and TaO x :N–based memristors, as depicted in Figure 2 c,f. Both devices can be maintained for 10 5 s without any degradation. Eight randomly selected cells were assessed to confirm the device–to–device conductance uniformity. Uniformity is also to do with the size of the possible formed clusters, and the size of both devices is equal [ 29 ]. Each cell of both devices was applied by a voltage sweep in the range of 0 to −1.6 and 0 to 1.4 V for 20 cycles with a voltage step of 0.01 V, as presented in Figure 3 a, b. For the TaO x –based device, the HRS ranges from 11.2 to 437 μS, and the LRS ranges from 0.62 to 7.3 mS ( Figure 3 a). In contrast, HRS values from 134 to 784 μS and LRS values from 3 to 4.9 mS are observed in the TaO x :N–based device ( Figure 3 b). For the TaO x –based device, the range of conductance levels was wider than for the TaO x :N–based device. We used resistance variability to numerically identify the distribution of the HRS and LRS and accurately examine the extent to which data points differ. When the variability was calculated by considering only the average of LRS and HRS values in each cell, the device–to–device HRS variability decreased from 60.7 to 48.3%, and that of LRS decreased from 35 to 8.7% when nitrogen doping was applied, as illustrated in Figure 3 c. Furthermore, the set voltage distribution of both devices was characterized in histograms, as plotted in Figure 4 a. The uniformity of the set voltages is crucial to ensure error–free operation. Set voltage is the threshold where the resistance of the I−V curve abruptly decreases from HRS to LRS. As illustrated in Figure 4 a, a wider set voltage distribution of the TaO x –based device is observed in the range of 0.5 to 1.1 V. In contrast, the TaO x :N–based device exhibits a more concentrated distribution in the range of 0.5 to 0.8 V. The statistical distribution of set voltages distribution is summarized in Table 2 . From the results of the DC characteristics, as described previously in Figure 2 , Figure 3 and Figure 4 , we confirmed that nitrogen doping on the TaO x –based device improved resistance variability and reliability—the most critical for ReRAM device applications [ 30 ]. Furthermore, a pulse endurance test was performed for up to 10 5 cycles to compare AC characteristics. A 10-μs pulse width of both set and reset pulses and amplitudes of 1.45 and −1.7 V were applied to both devices for a reliable comparison. The TaO x –based memristor exhibited unstable RS operation. Moreover, the conductance value in LRS degraded throughout the measurement, as presented in Figure 4 b. In contrast, the TaO x :N–based memristor exhibited excellent endurance of up to 10 5 cycles. Synaptic plasticity, such as long–term potentiation (LTP) and long–term depression (LTD), are crucial aspects of the application of neuromorphic systems [ 31 ]. The LTP and LTD were measured and discussed to compare the synaptic properties between the two devices. In Figure 5 a,c, 100 potentiation and depression pulses were applied with 0.1 V read voltage pulses. The amplitude of potentiation and depression pulses were set to 0.9 and −1 V for the TaO x –based memristor and 0.94 and −1.05 V for the TaO x :N–based memristor. Moreover, the pulse width was fixed to 100 nS to ensure the pulse conditions were as similar as possible while modulating conductance gradually. Ten cycles of potentiation and depression were conducted with these consecutive pulses. Figure 5 b illustrates the degradation of the potentiation and depression during cycles of the undoped device. In contrast, Figure 5 d illustrates a more constant and stable pulse measurement without degradation. When calculating the cycle–to–cycle resistance variability between the two devices, the nitrogen–doped device exhibited less variability of 4.9% compared with the 13.7% of the TaO x –based memristor. Even for nitrogen–doped devices (4.9%), cycle–to–cycle variability occurred, as depicted in Figure 5 . We reduced this variability by applying a DC set voltage to produce LRS. We then proceeded first with depression rather than potentiation, as depicted in Figure 6 a,c. As depicted in Figure 6 , 100 depression and potentiation pulses were applied with a 0.1 V read voltage pulse; ten cycles were conducted. The amplitudes of depression and potentiation pulses were set to −1.09 and 0.86 V for the TaO x –based memristor and −1 and 0.81 V for the TaO x :N–based memristor; the pulse width was set to 100 nS. As depicted in Figure 6 b, the cycle–to–cycle variation issue still occurred in the TaO x –based device. The conductance of the last point of a cycle does not match the first point of a subsequent cycle. In contrast, Figure 6 d illustrates more stable synaptic properties, with a 2.1% cycle–to–cycle variability—a smaller value than when potentiation was performed first (4.9%). Based on these pulse measurement results ( Figure 5 and Figure 6 ), device reliability is increased by incorporating nitrogen—similar to the DC measurement results. In biological neural networks, calculation and storage of information are performed simultaneously [ 32 ], so uniformity of resistance value is critical. Consequently, the TaO x :N–based memristor, which has higher cycle–to–cycle uniformity, is suitable for neuromorphic computing applications. Key performance indicators achieved in the present study are summarized in Table 3 ."
} | 3,103 |
31618815 | PMC6843377 | pmc | 3,630 | {
"abstract": "Microbial electrochemical technology provides an inexhaustible supply of electron acceptors, allowing electroactive microorganisms to generate biocurrent and accelerate the removal of organics. The treatment of wastewater contaminated by butachlor, which is a commonly used chloroacetamide herbicide in paddy fields, is a problem in agricultural production. In this study, butachlor was found to be removed efficiently (90 ± 1%) and rapidly (one day) in constructed single-chamber microbial fuel cells (MFCs). After the addition of sodium acetate to MFCs with butachlor as the sole carbon source, electricity generation was recovered instead of increasing the degradation efficiency of butachlor. Meanwhile, the microbial community structure was changed in anodic and cathodic biofilms after the addition of butachlor, following the bioelectrochemical degradation of butachlor. High-throughput sequencing showed the proliferation of Paracoccus and Geobacter in MFCs with butachlor as the sole carbon source and of Thauera butanivorans in MFCs with butachlor and sodium acetate as concomitant carbon sources. These species possess the ability to oxidize different substituents of butachlor and have important potential use for the bioremediation of wastewater, sediments, and soils.",
"conclusion": "5. Conclusions Firstly, butachlor, a commonly used chloroacetamide herbicide in paddy fields, could be removed efficiently and rapidly in constructed single-chamber MFCs. The concomitant addition of sodium acetate recovered the electricity generation instead of increasing the degradation efficiency of butachlor. Secondly, a change in microbial community structure was induced after the addition of butachlor, in order to adapt to the bioelectrochemical degradation of butachlor. Paracoccus and Geobacter in BUT and T. butanivorans in BUT-NaAC proliferated, presumably following the oxidization of different substituents of butachlor; thus, they have important potential use for the bioremediation of wastewater, sediments, and soils.",
"introduction": "1. Introduction Chloroacetamide herbicides, e.g., alachlor, acetochlor, butachlor, and metolachlor, are commonly used herbicides in rice, corn, soybean, and many other crops for controlling annual grass and broadleaf weeds [ 1 , 2 ]. Butachlor ( N -(butoxymethyl)-2-chloro- N -(2,6-diethylphenyl) acetamide) is a widely used chloroacetamide herbicide in paddy fields. Exterior drainage and outdoor rainwater are inevitably polluted by residues of butachlor and its degradation metabolites, which pose a threat to the surrounding environment, especially ground and surface waters, since they are highly toxic to some aquatic organisms [ 3 ] and are potentially carcinogenic (e.g., butachlor caused stomach tumors in rats [ 4 ]). Therefore, this risky wastewater needs to be treated before discharge. The biodegradation of organic pollutants was demonstrated as a feasible and safe treatment technology [ 1 , 5 , 6 ]. However, the degradation efficiency is seriously limited in anoxic environments, e.g., groundwater, sediment, and subsoil [ 7 , 8 ]. Moreover, butachlor and alachlor are slower to degrade than acetochlor and metolachlor, since it is possible that the electronic effect of the long-chain alkyl substituents alters the susceptibility of the benzene ring [ 9 ]. Presently, microbial electrochemical technology that provides an inexhaustible solid anode as the electron acceptor, e.g., microbial fuel cells (MFCs), can efficiently and rapidly degrade organic pollutants in anoxic environments and simultaneously generate electricity by means of electroactive microorganisms [ 10 , 11 , 12 , 13 ]. The objective of this study was to investigate the removal efficiency of butachlor in constructed single-chamber MFCs with and without an easily assimilated carbon source to elucidate whether there was a co-metabolism effect. Moreover, the capacity of electricity generation was demonstrated in MFCs using butachlor as the sole and concomitant carbon source. Finally, the change in microbial community structure was revealed in the anodic and cathodic biofilms of MFCs, especially with regard to predominant microbes, which possess potential use for the bioremediation of wastewaters, sediments, and soils.",
"discussion": "4. Discussion This study demonstrated that butachlor could be removed efficiently (90 ± 1%) and rapidly (one day) in activated-carbon air cathode MFCs, while the observed electricity generation was unsurprising. Unexpectedly, the abundance of Geobacter increased distinctly instead of decreasing in the anodic biofilm of MFCs with butachlor as the sole carbon source (in BUT). It was expected that electrons generated by the electroactive microorganisms would be consumed by the biodegradation of butachlor, but they were not. The charge output and removal efficiency of butachlor in MFCs with butachlor and sodium acetate as the concomitant carbon sources (in BUT-NaAC) were comparable to controls with sodium acetate as the sole carbon source and BUT, respectively. These results indicated that butachlor is unsuitable as fuel for MFCs; however, Geobacter may independently and/or synergistically metabolize the butachlor. For example, species of Geobacter have the ability to anaerobically oxidize aromatic compounds to benzoyl-coenzyme A, then to acetyl-coenzyme A via fatty-acid oxidation, and finally to carbon dioxide via the tricarboxylic acid cycle [ 20 , 21 , 22 ]. The amount of electricity generation in one cycle of BUT-NaAC was more than the totals of BUT and control alone. On the one hand, this suggests that some degradates (small molecular organics) were presumably transferred into electricity via the catalysis of microbes in the MFCs. On the other hand, the microbial richness in the anodic biofilm of BUT-NaAC apparently decreased compared to the control, while the abundance of Geobacter slightly increased. This suggests that there were fewer microbes to compete for the substrates with electroactive microorganisms, thereby generating more electricity. This study shows that it is infeasible to use sodium acetate as a co-metabolized carbon source to degrade butachlor. In fact, the concomitant use of sodium acetate possibly suppressed the activity of butachlor degradation, since lower microbial richness and diversity were found in the anodic biofilm of BUT-NaAC than in that of BUT. Therefore, a slightly low removal rate of butachlor was observed with BUT-NaAC. In treatments with the addition of butachlor (BUT and BUT-NaAC), the microbial richness and diversity of biofilms showed an obvious change. In BUT, the abundances of Paracoccus and Geobacter significantly increased in the anodic biofilm compared to the control, suggesting a potential decomposition effect. Previous studies found that Paracoccus spp. could efficiently degrade chloroacetamide herbicides, e.g., butachlor, alachlor, and acetochlor [ 6 ]. Some species of Paracoccus could completely mineralize chlorpyrifos, in addition to degrading 3,5,6-trichloro-2-pyridinol, pyridine, methyl parathion, and carbonfuran [ 5 ]. Species from Geobacter mainly decomposed phenol [ 20 ], benzene [ 22 ], and benzoate [ 21 ] in anaerobic conditions. Thus, it can be inferred that species of Paracoccus possess good dechlorination/degradation ability, while species of Geobacter are able to oxidize phenyl alkyl substituents. In BUT-NaAC, the amount of T. butanivorans obviously increased in the anodic biofilm compared to the control. T. butanivorans is a C2–C9 alkane-oxidizing bacterium [ 23 ] and was, thus, deemed to degrade the butachlor. Moreover, this bacterium has the ability to secrete soluble butane monooxygenase [ 24 ] and, therefore, was able to oxidize the alkoxybutyl substituent, which is the limiting step of degradation efficiency for chloroacetamide herbicides [ 6 ]. Furthermore, the majority of species of Thauera have denitrification ability [ 25 ], which further stimulates the decomposition of butachlor, involving the amide nitrogen’s alkoxybutyl, which significantly affects the biodegradability of these herbicides [ 6 ]. Unfortunately, the degradation product analysis of butachlor was unsuccessful in this study, presumably due to the low aqueous solubility of degradates [ 9 ], which will be addressed in future work to reveal the bioelectrochemical degradation pathway and to assess the function of these special microbes."
} | 2,101 |
28083017 | PMC5186783 | pmc | 3,631 | {
"abstract": "Microbial Molecular hydrogen (H 2 ) cycling plays an important role in several ecological niches. Hydrogenases (H 2 ases), enzymes involved in H 2 metabolism, are of great interest for investigating microbial communities, and producing BioH 2 . To obtain an overall picture of the genetic ability of Cyanobacteria to produce H 2 ases, we conducted a phylum wide analysis of the distribution of the genes encoding these enzymes in 130 cyanobacterial genomes. The concomitant presence of the H 2 ase and genes involved in the maturation process, and that of well-conserved catalytic sites in the enzymes were the three minimal criteria used to classify a strain as being able to produce a functional H 2 ase. The [NiFe] H 2 ases were found to be the only enzymes present in this phylum. Fifty-five strains were found to be potentially able produce the bidirectional Hox enzyme and 33 to produce the uptake (Hup) enzyme. H 2 metabolism in Cyanobacteria has a broad ecological distribution, since only the genomes of strains collected from the open ocean do not possess hox genes. In addition, the presence of H 2 ase was found to increase in the late branching clades of the phylogenetic tree of the species. Surprisingly, five cyanobacterial genomes were found to possess homologs of oxygen tolerant H 2 ases belonging to groups 1, 3b, and 3d. Overall, these data show that H 2 ases are widely distributed, and are therefore probably of great functional importance in Cyanobacteria. The present finding that homologs to oxygen-tolerant H 2 ases are present in this phylum opens new perspectives for applying the process of photosynthesis in the field of H 2 production.",
"introduction": "Introduction Microbial hydrogen (H 2 ) metabolism is a process that occurs in many different environments. In addition to being a key metabolic factor in several biological communities, H 2 has attracted considerable interest as a candidate environmentally friendly energy carrier. The use of photosynthetic organisms such as microalgae and cyanobacteria has been tested worldwide for this purpose. In cyanobacteria, the main enzymes involved in H 2 metabolism are nitrogenases and hydrogenases (H 2 ases) (Reviewed in Bothe et al., 2010 ). Nitrogenases fix molecular nitrogen (N 2 ) and produce H 2 as a byproduct (D'Eustachio and Hardy, 1964 ). H 2 ases are metalloprotein enzymes which catalyze in several microorganisms the reversible reaction:\n 2 H + + 2 e - ↔ H 2 (for a recent Review see Peters et al., 2015 ). They are usually classified into three phylogenetically independent classes: [Fe] H 2 ases, [FeFe] H 2 ases, and [NiFe] H 2 ases (Vignais and Billoud, 2007 ). Since [Fe] H 2 ases are light-sensitive enzymes (Chen et al., 2002 ), they can be considered as for limited interest in the context of H 2 photoproduction. The [FeFe] H 2 ases present in anaerobic bacteria and some phototrophic eukaryotes preferentially catalyze the evolution of H 2 at high frequencies; these enzymes are also characterized by their high sensitivity to oxygen (O 2 ) (Melis et al., 2000 ; Florin et al., 2001 ; Winkler et al., 2002 ; Peters et al., 2015 ). The [NiFe] H 2 ases, which have been found to exist in Archaea and in several aerobic and anaerobic bacterial phyla, are mainly involved in H 2 oxidation but can also catalyze the reduction of protons to H 2 (Vignais and Billoud, 2007 ). They consist of a large subunit containing the bimetallic center [NiFe] and a small subunit containing [FeS] clusters (Volbeda et al., 1995 , 1996 ; Peters et al., 2015 ). Based on a phylogenetic analysis of the large subunit, and more specifically, on two highly conserved regions located in this subunit near the [NiFe] center (the L1 and L2 regions), the [NiFe] H 2 ases have been classified into the eight different groups presented in Table 1 (Vignais et al., 2001 ; Vignais and Billoud, 2007 ). The maturation of [NiFe] H 2 ases involves six proteins (HypFCDEAB), which synthesize the non-protein ligands (CO and CN) and assemble the active site (Dernedde et al., 1996 ; Hansel et al., 2001 ; Hoffmann et al., 2006 ). In the last step in the process of biosynthesis, the C terminal part of the large subunit is cleaved by a specific peptidase (Thiemermann et al., 1996 ; Devine et al., 2009 ). Table 1 Overview of the main features of [NiFe] H 2 ases . Group Name Function H 2 O 2 sensitive/resistant References 1 Membrane bound H 2 uptake H 2 ases H 2 uptake under aerobic and/or anaerobic conditions. Oxidation Sensitive and Resistant Higuchi et al., 1999 ; Marques et al., 2010 ; Dementin et al., 2011 2a Cyanobacterial uptake H 2 ases Uptake of H 2 produced by nitrogenase. Oxidation Sensitive Oxelfelt et al., 1998 ; Tamagnini et al., 2007 ; Zhang et al., 2014 2b H 2 -signaling H 2 ases H 2 perception and signaling. Oxidation Resistant Buhrke et al., 2004 , 2005 ; Roncaroli et al., 2015 3a F 420 -reducing H 2 ases H 2 utilization during methagenosis. Oxidation evolution Sensitive Hendrickson and Leigh, 2008 ; Vitt et al., 2014 3b Tetrameric bifunctional H 2 ases Regulation and redox balance. Oxidation evolution Sensitive and resistant Bryant and Adams, 1989 ; Jenney and Adams, 2008 ; Berney et al., 2014 3c Methyl-viologen-reducing H 2 ases H 2 uptake during methagenosis. Oxidation Sensitive Kaster et al., 2011 3d Soluble bidirectional H 2 ases Regulation and redox balance. Oxidation evolution Sensitive and resistant McIntosh et al., 2011 ; Lauterbach and Lenz, 2013 4 H 2 -evolving, energy-conserving, membrane-associated H 2 ases Coupling of formate or carbon monoxide to H 2 evolution. Evolution Sensitive Bagramyan et al., 2002 ; McDowall et al., 2014 5 Actinobacteria [NiFe]-H 2 ases H 2 uptake during starvation. Oxidation Resistant Schäfer et al., 2013 Although the activity of most of the [NiFe] H 2 ases tends to be inhibited by O 2 , some members of this class remain active in the presence of O 2 and have therefore been called O 2 -tolerant. The O 2 -tolerant H 2 ases described for the first time in the anoxigenic bacterium Rubrivivax gelatinosus (Maness et al., 2002 ) occur in the Group 1 membrane-bound H 2 ases (MBH), the H 2 -signaling group (RH, Group 2b) (Buhrke et al., 2005 ; Duché et al., 2005 ), the tetrameric bifunctional H 2 ases (group 3b) (Jenney and Adams, 2008 ; Kwan et al., 2015 ), the bidirectional H 2 ases (group 3d) (Horch et al., 2013 ; Karstens et al., 2015 ) and the recently identified Group 5 Actinobacterial-H 2 ases (Table 2 ) (Constant et al., 2010 ; Lubitz et al., 2014 ). In the case of the MBH enzymes, the main difference between the standard and tolerant members focuses on the [FeS] cluster located near the [NiFe] site. Instead of the canonical [4Fe4S] present in the standard enzymes, a [4Fe3S] cluster coordinated by six cysteine residues occurs in the tolerant enzymes (Pandelia et al., 2011 ; Shomura et al., 2011 ). This proximal [4Fe3S] is the most striking feature thought to be linked to O 2 -tolerance (Goris et al., 2011 ; Lukey et al., 2011 ). The O 2 -insensitivity of the RH-H 2 ases of Ralstonia eutropha H16 depends on the size and shape of the intramolecular hydrophobic cavity giving access to the active [NiFe] site (Buhrke et al., 2005 ). The molecular mechanism underlying the O 2 -tolerance of the Group 3 SH enzymes and that of the actinobacterial H 2 ases still remains to be elucidated. Table 2 Overview of the main features of O 2 -tolerant H 2 ases in several organisms . Group Name Cluster Fe-S small subunit Structural basis of O 2 -tolerance Example References Homolog in cyanobacteria 1 Membrane bound H 2 uptake H 2 ases (MBH) p [4Fe3S] m [3Fe4S] d[4Fe4S] Transfer electron from the proximal cluster to active site to reduce O 2 to water. Rubrivivax gelatinosus , Hyd-1 Escherichia coli Maness et al., 2002 ; Evans et al., 2013 Lyngbya confervoides BDU141951 2b H 2 -signaling H 2 ases (RH) p [4Fe4S] m [4Fe4S] d[4Fe4S] The gas channel is narrower than standard H 2 ases and the O 2 cannot interact with the active site. Rhodobacter capsulatus, Ralstonia eutropha Buhrke et al., 2005 ; Duché et al., 2005 None 3b Tetrameric bifunctional H 2 ases (PfSHI) p [4Fe4S] m [2Fe2S] d[4Fe4S] No formation of the slowly reactivating state Ni-A Pyrococcus furiosus Jenney and Adams, 2008 ; Kwan et al., 2015 Cyanothece sp. PCC 7425, Leptolyngbya boryana PCC 6306, Mastigocoleus testarum BC008 3d Soluble bidirectional H 2 ases (ReSH) [4Fe4S] Reduction of O 2 in water. Cys39 and Trp42 are demonstrated important for O 2 tolerance Ralstonia eutropha Horch et al., 2013 ; Karstens et al., 2015 Aphanocapsa montana BDHKU210001 5 Actinobacteria [NiFe]-H 2 ases (AH) p [4Fe4S] m [4Fe4S] d[4Fe4S] Unknown Streptomyces avermitilis, Ralstonia eutropha Constant et al., 2010 ; Lubitz et al., 2014 None Cyanobacteria, the only prokaryotes capable of oxygenic photosynthesis, form a large and morphologically diverse bacterial group consisting of five morphological subsections. The unicellular organisms that undergo binary fission belong to subsection I ( Chroococcales ). The unicellular strains that divide through multiple fission processes form subsection II ( Pleurocapsales ), and subsection III consists of filamentous strains which are unable to perform cell differentiation ( Oscillatoriales ). The strains in subsections IV and V are filamentous and able to differentiate specific cells called heterocysts, which are dedicated to N 2 fixation (Rippka et al., 1979 ). Cyanobacteria are widely distributed in various environments (from oceans to desert crusts), where they contribute importantly to primary production and N 2 fixation processes (Garcia-Pichel et al., 2003 ). N 2 -fixation in these organisms is mainly achieved by a molybdenum-iron ([MoFe]) nitrogenase which consists of two subunits, a Fe-protein encoded by nifH , and a Mo-Fe protein encoded by nifDK genes (Smith and Eady, 1992 ). The maturation process requires three essential ( nifBEN ) and three no essential genes ( nifUSV ) (Reviewed in: Rubio and Ludden, 2008 ). The reduction of N 2 is accompanied by the formation of H 2 (Berman-Frank et al., 2003 ). Cyanobacteria contain two different [NiFe] H 2 ases: the bidirectional [NiFe] H 2 ase (Hox, Group 3d) and the uptake H 2 ase (Hup, Group 2a) (Tamagnini et al., 2007 ). The Hup H 2 ase is a heterodimeric enzyme encoded by the hupSL genes, which consumes the H 2 produced by the nitrogenase (Houchins and Burris, 1981 ; Lindblad and Sellstedt, 1990 ). The bidirectional Hox H 2 ase, which can oxidize H 2 and reduce H + , can exist in both diazotrophic and non-diazotrophic strains, and is thought to be a heteropentameric enzyme encoded by hoxEFUYH genes (Schmitz et al., 1995 ). In the unicellular cyanobacterium Synechocystis PCC 6803, the bidirectional H 2 ase has been shown to be essential under mixotrophic and nitrate limiting conditions, which suggests that this enzyme functions as electron sink for reduced flavodoxin/ferredoxin (Gutekunst et al., 2014 ). The ability of the Hox enzymes to be quickly reactivated after being inhibited by O 2 has made them the most frequently used H 2 ase in studies on H 2 production in cyanobacteria (Serebryakova et al., 1996 ; Germer et al., 2009 ; McIntosh et al., 2011 ). The main limitations of using the cyanobacterial Hox enzymes in large scale H 2 production processes are the low levels of H 2 produced and the fast reversal of the enzymatic reaction into oxidation (Tamagnini et al., 2007 ; Rögner, 2013 ). During the last decade, genetic engineering approaches were used in several studies in order to overcome these technological barriers with a relative success (Masukawa et al., 2002 ; McNeely et al., 2010 ; Baebprasert et al., 2011 ; Ortega-Ramos et al., 2014 ; Nyberg et al., 2015 ). Cyanobacterial strains and/or genomes have also been widely explored in order to unravel the complex picture of H 2 ases (Ludwig et al., 2006 ; Barz et al., 2010 ; Kothari et al., 2012 , 2013 ). These studies have opened new perspectives, since they have shed light on the H 2 production potential of strains other than those previously used as laboratory models. Since the publication of these studies, larger numbers of cyanobacterial genomes have been sequenced, which has greatly improved the genomic coverage of all the phylum (Shih et al., 2013 ). In order to investigate cyanobacterial H 2 metabolism more closely, we performed a large-scale analysis of H 2 ases genes distribution in cyanobacteria, which consisted in searching for the genes encoding H 2 ases and the proteins required for their maturation in 130 cyanobacterial genomes. The distribution of H 2 ases in the cyanobacterial phylum inhabiting various environments is discussed.",
"discussion": "Discussion The present analyses of the distribution of genes encoding H 2 ases in cyanobacterial genomes suggest that H 2 metabolism is widely distributed among the various ecological niches that have been colonized by these organisms. H 2 ase genes and the genes encoding proteins necessary to the maturation process feature prominently in the late branching clades of the cyanobacterial tree of species, which suggests that the need for H 2 production and/or uptake has followed the phylogenic evolution of this phylum. The fact that all the structural genes in these enzymes and their maturation process genes have been largely conserved in many cyanobacterial genomes indicates, if these genes are really expressed, that they might play an important physiological role in the bacterial strains inhabiting various environments. Considerable rates of H 2 production by cyanobacteria have been reported to occur in microbial mats (Marshall et al., 2012 ), and Microcoleus spp has been found to be a predominant H 2 producer in the microbial mats formed in the Elkhorn Slough estuary, Monterey Bay (Burow et al., 2012 ). These data further indicate that functional studies on H 2 ases in environmental strains in addition to laboratory models would greatly improve our understanding of H 2 metabolism in this bacterial phylum. No bidirectional H 2 ase genes were detected in the genomes of open ocean strains ( Prochlorococcus and Synechococcus in particular), in agreement with previous results (Barz et al., 2010 ). The latter study also showed that heterotrophic bacteria inhabiting this environment also lacked bidirectional H 2 ase encoding genes. The O 2 concentration of open ocean waters measured during a period of several months was found to be above 200 μM (Emerson et al., 2002 ) which may not favor the contribution of the Hox enzyme to the process of H 2 metabolism under anaerobic conditions (Khanna and Lindblad, 2015 ). The distribution of hup, hox and nif genes is highly variable in freshwater, hot spring and terrestrial environments (Figure 1 ), possibly because of the various conditions that organisms may encounter in these ecological niches. Nineteen genomes of strains belonging to subsections I, II, III and IV contain nif genes but no hup genes (Figures 1 – 3 and Supplementary Table 2 ). In this background, one might expect the H 2 production rate of nitrogenase to play an important role in the absence of uptake H 2 ase. The deletion of the hupL gene in the filamentous diazotrophic strains Nostoc PCC 7120 and Nostoc PCC 7422 has indeed been found to improve the H 2 production (Masukawa et al., 2002 ; Yoshino et al., 2007 ). In the unicellular cyanobacterium Cyanothece PCC 7822, which fixes nitrogen under aerobiosis, HupL has been shown to be essential to activity of the nitrogenase in the presence of O 2 . The authors concluded that the main function of the HupSL complex in this bacterium is the protection of the nitrogenase from O 2 (Zhang et al., 2014 ). The present data show that most of the strains possessing nif genes and lacking the uptake H 2 ase are unicellular [ Aphanocapsa montana BDHKU210001, Chroococcidiopsis sp. PCC 6712, Nodosilinea nodulosa PCC 7104, Synechococcus sp. JA-2-3B'a(2–13), Synechococcus sp. JA-3-3Ab]. All these strains are known to undergo N 2 fixation under anaerobic conditions (Suplementary Table 2 ). In future studies, it would be interesting to investigate whether the absence of the uptake H 2 ase in these strains results in high H 2 production. The finding that genes potentially encoding O 2 -tolerant H 2 ases are present in five cyanobacterial genomes is of great interest. Since Lyngbya confervoides BDU141951 genome does not contain all the accessories hox genes important for the maturation process of the MBH-O 2 tolerant enzyme, and since the genome of Leptolyngbya boryana PCC 6306 contains only the hypAB genes, it is likely that these two strains are not able to produce an active O 2 -tolerant enzyme. Whether the other three cyanobacterial strains found here to possess genes encoding for O 2 -tolerant enzyme actually produce these enzymes needs to be analyzed. The possible input of theses enzymes to the physiology of these organisms in both marine and freshwater environments is an intriguing question. These enzymes are probably involved in the oxidation of H 2 , like most of their homologs in other organisms. However, in the aerobic soil bacterium Mycobacterium smegmatis , an O 2 -tolerant H 2 ase has been found to produce H 2 , thus enabling this organism to cope with the hypoxia occurring in its ecological niche (Berney et al., 2014 ). The possibility that O 2 -tolerant H 2 ase may play a similar role in cyanobacteria is a tempting hypothesis. Whether the cyanobacterial strains found to possess genes encoding for O 2 -tolerant H 2 ases could be for interest in the context of photosynthetic H 2 production is a perspective worth exploring in the future."
} | 4,466 |
34172441 | PMC8232913 | pmc | 3,632 | {
"abstract": "Rationally engineered tolerance enables broadly efficient lignocellulosic fermentation of diverse feedstocks and bioproducts.",
"introduction": "INTRODUCTION Meaningful displacement of greenhouse gas emissions from continued oil consumption requires a renewable feedstock that is transformable into products fungible with petrofuels and petrochemicals and is deployable on a similar scale. Despite the declining cost of carbon-free electricity and rise of emission-free vehicles, studies estimate that this segment will comprise at most 31% of the global fleet by 2040 due to nonroad modes of carriage and long average ownership in the established internal combustion fleet ( 1 ). As the transportation sector remains the largest generator of carbon dioxide, the sheer number of legacy vehicles necessitates that liquid biofuels play a dominant role in any future energy mix to minimize net emissions ( 2 ). Lignocellulosic biomass, the largest renewable terrestrial resource, provides a realistic intermediate-term route to sustainable fuel and nonfuel commodities at enormous scale when paired with suitable fermentation infrastructure ( 3 ). In addition to quantities on the magnitude of fossil carbon, lignocellulose addresses issues such as food-fuel competition and arable land use that beset present-generation feedstocks such as corn ( 4 ). Fermented fuel products, notably ethanol, can be blended directly into the gasoline supply at 15 to 85% or chemically dehydrated to ethylene and upgraded into jet fuel ( 5 ). However, the severe pretreatments needed to deconstruct the highly recalcitrant plant fibers into fermentable sugars typically result in feedstocks toxic to microorganisms ( 6 – 8 ). Partly saddled by these technical challenges, the U.S. cellulosic ethanol industry has dwindled sharply (a single preproduction plant run by POET-DSM remains), and pretreatment research has refocused on conversions that yield clarified, biocatalyst-friendly substrates ( 9 , 10 ). Even then, the greater complexities required by these processes have generally increased costs (estimates as high as 30¢ per gallon ethanol) as well as eroded scalability and competitiveness ( 11 ). Engineering elevated microbial tolerance to the inhibitors released in simpler, but more aggressive, hydrolytic methods would, therefore, address one of the major obstacles impeding greater utilization of cellulosic feedstocks ( 12 – 15 ). Here, we show that a targeted combination of genetic and feedstock modifications is sufficient to enhance a single strain to tolerate a wide variety of highly toxified biomass hydrolysates and deliver cellulosic ethanol with performance comparable to current clean sugar ethanol. Our rationally designed approach is, additionally, highly modular: With introduction of a single gene and no further engineering, feedstock-agnostic hydrolysate tolerance is conferred on previously engineered metabolic chassis strains (including one synthesizing a biodegradable plastic) to enable cellulosic products beyond ethanol.",
"discussion": "DISCUSSION Our results describe a functionally independent, lightweight platform that both endows yeast with general lignocellulosic hydrolysate tolerance and integrates harmoniously with preexisting metabolically engineered chassis strains. Through systematic characterization of the three dominant toxicities released from biomass pretreatment, we have demonstrated that tolerance to each inhibitor can be realized through standard neutralization (for acetic acid) or conversion of the aldehydes to alcohols, which are subsequently ameliorated by the prior-identified elevated K + and pH treatment (for furfural and HMF). The general practice of hydrolysate tolerance can, therefore, be reduced to two specific and readily modifiable parameters: In a genetic background enhanced by GRE2 evol for the accelerated reduction of furfural and HMF, a large diversity of feedstocks, regardless of plant source and/or pretreatment process, can be accommodated via tailored adjustment of K + and pH. This substrate robustness indicates that the spectrum of hydrolytic by-products other than furfural, HMF, and acetic acid (for example, the various acidic and phenolic inhibitors shown in table S3) may be qualitatively immaterial. These benefits, collectively, renew and boost the value proposition of cellulosic fermentation. Wide feedstock compatibility can reduce the dependence on specific crop types or pretreatments and, consequently, ameliorate the supply variability (e.g., from seasonality, storage stability) and cost uncertainties surrounding biomass ( 13 ). Similarly, heightened tolerance, in addition to harnessing toxic sugar streams or transport-friendlier concentrates, enables production conditions of minimized contaminant growth that would otherwise require the standard-practice, but public health-concerning, use of antibiotics ( 55 ). Last, the targeted specificity of our detoxification approach underlies the high decoupling with metabolism and straightforward integration with previously engineered pathways such as those for xylose consumption and lactic acid synthesis. This underscores the notion of a “drop-in” tolerance phenotype extensible to even more non-native capabilities and high-volume biofuels and biochemicals."
} | 1,317 |
28076447 | PMC5226831 | pmc | 3,634 | {
"abstract": "Globally methane (CH 4 ) emissions from ruminant livestock account for 29% of total CH 4 emissions. Inherited variation about CH 4 emissions of different animal species might provide new opportunity for manipulating CH 4 production. Six rumen-simulating fermenters (Rusitec) were set up for this study lasting for 16 d. The diet consisted of forage to concentrate ratio of 50:50 with barley straw as the forage. Treated vessels were supplied with rumen fluid from yak or cattle (3 vessels per animal species). Microbial growth was measured using 15 N as a marker. The microbial community structure from liquid- and solid-fraction of each vessel was determined based on the 16S rRNA genes targeting both bacteria and archaea with MiSeq platform. CH 4 yield was lower when the inoculum used from yak than that from cattle (0.26 and 0.33 mmol CH 4 /g dry matter intake, respectively). Lower H 2 production was observed in Rusitec fermenters with rumen fluid from yak compare with that from cattle (0.28 and 0.86 mmol/d, respectively). The apparent digestibility of neutral detergent fiber, the isovalerate percentage with respect to the total amount of volatile fatty acids, the hydrogen recovery, and the proportion of liquid-associated microbial nitrogen derived from ammonia-nitrogen were higher in Rusitec fermenters incubated with rumen fluid from cattle than that from yak. The relative abundances of methanogens were no difference between two animal species. We hypothesize that more H 2 production contributes to the higher methane emissions in cattle compare with yak.",
"introduction": "Introduction Methane (CH 4 ) accounts for 11% of total greenhouse gas (GHG) emissions in China. Almost 21% of CH 4 emissions are due to enteric fermentation in ruminant livestock industry [ 1 , 2 ]. The Global Warming Potential (GWP) of CH 4 for a time horizon of 100 years is 28-fold that of CO 2 [ 3 ]. Enteric CH 4 emissions also represent a 2 to 12% loss of gross energy intake [ 4 ]. Many ways to manipulate enteric CH 4 emissions have been developed, including 4 broad categories: feeds and feeding management, rumen modifiers, genetics and other management strategies [ 5 ]. Investigating animals that produce lower CH 4 might lead to improving livestock systems through modifying rumen fermentation and reducing CH 4 emissions from other livestock [ 6 – 8 ]. Yak produced 1.7 g of methane /kg W 0.75 under grazing conditions, which was lower compared with published data about cattle (3.2~4.2 g of methane /kg W 0.75 ) [ 9 , 10 ]. Over 15 million yaks grazed in the Qinghai-Tibetan Plateau account for approximately 90% of the world's total number of yak [ 11 ]. Due to the harsh environment in the Qinghai-Tibetan Plateau, which is characterized by hypoxia, strong ultra-violet (UV) radiation, severe cold and deficiencies of forage resources, yak has evolved special abilities on the metabolism of certain nutrients, morphology, and genetics [ 12 – 14 ]. Therefore, yak shows great potential as a “low carbon” animal, which calls for systematic comparative studies of “low-CH 4 emissions” from yak. However, in a previous study, they conducted an investigation regarding yak without a control group [ 9 ]. The level of intake, type and quality of feed, and environmental temperature might contribute to great variation in CH 4 production [ 15 ]. Rumen-simulating fermenters, like the Rusitec fermenters, could be useful tools to evaluate methane emissions from different animals or under different additives treatment because major advantage of this system is the ability to remove fermentation liquid and maintain for relatively long periods of time [ 16 , 17 ]. The information would be useful using this type of fermenter before conducting expensive and time-consuming in vivo study to confirm the difference of methane emissions between yak and cattle. Thus, we performed a comparative study using a Rusitec system to investigate the difference of CH 4 emissions between yak and cattle under same conditions. The first aim of this study was to confirm if yak is lower CH 4 producer than cattle under the same conditions. The second aim was to explore the possible link between CH 4 production and liquid-/solid-associated microbes.",
"discussion": "Results and Discussion Average values of effluent volume, pH, and substrate digestibility were summarized in Table 2 . Because our main aim was to investigate the difference between yak and cattle species, a similar pH under two treatments would maintain a stable environment ( Table 2 ). No differences in effluent volume, apparent disappearance of dry matter (DM), organic matter (OM), and neutral detergent fiber (NDF) were found between the two species ( P >0.05). 10.1371/journal.pone.0170044.t002 Table 2 Effect of animal species on the level of the pH and amount of apparent disappearance under same low nitrogen diet in Rusitec fermenters. Item Cattle Yak SEM P -value Effluent, L/d 0.775 0.781 00024 0.776 pH before feeding 6.84 6.82 0.02 0.383 Apparent ruminal digestibility, % DM 42.2 40.9 1.10 0.245 NDF 68.9 69.6 1.95 0.69 OM 41.9 40.4 1.09 0.192 DM-dry matter; NDF- neutral detergent fiber; OM-organic matter. To our knowledge, this is the first study in which yak and cattle species were compared using Rusitec fermenters to investigate differences in methane production. The total amounts of gas and methane were higher in fermenters used rumen fluid from cattle, showing 27% and 32% increases in production, respectively ( Table 3 ). However, the results were still lower than those of other studies in Rusitec fermenters [ 18 , 29 ]. Under the same diet conditions, the donor animal was the main reason for the differences in gas and methane production [ 30 ]. 10.1371/journal.pone.0170044.t003 Table 3 CH 4 production and volatile fatty acids (VFA) in Rusitec fermenters (means for the whole experimental period). Item Cattle Yak SEM P -value Total gas production, mmol/d 50.6 39.7 3.990 0.010 CH 4 , mmol/d 6.2 4.7 0.602 0.023 CH 4 per g DM, mmol/g DM intake 0.33 0.26 0.026 0.032 CH 4 per g OM intake, mmol/g OM intake 0.35 0.27 0.026 0.034 CH 4 per g DM, mmol/g digestible DM 1.11 0.84 0.109 0.019 CH 4 per g DM, mmol/g digestible OM 0.85 0.68 0.082 0.049 H 2 , mmol/d 0.86 0.28 0.099 0.017 Total VFA production, mmol/d 38.1 46.6 5.420 0.127 Molar proportion (mol/100 mol) Acetate 50.1 49.4 0.864 0.424 Propionate 32.3 31.8 0.994 0.650 Butyrate 12.1 13.7 0.468 0.002 Isobutyrate 0.253 0.258 0.091 0.956 Valerate 3.81 3.76 1.012 0.965 Isovalerate 1.46 1.06 0.158 0.019 Acetate: Propionate 1.55 1.57 0.062 0.681 H recovery, % 85.56 73.91 3.703 0.004 DM-dry matter; OM-organic matter. Total volatile fatty acids (VFA) and VFA profiles (acetate, propionate, isobutyrate, valerate, and acetate to propionate ratio) showed no observable differences between the two species ( P >0.05, Table 3 ). Butyrate and isovalerate were affected by the two species. Butyrate was higher with yak species than cattle species. In contrast, isovalerate was lower in yak species ( P <0.05, Table 3 ). Previously, the propionate, butyrate, and acetate to propionate ratios were affected by the donor animals of cow and sheep [ 30 ]. The different results could be due to several reasons, including the microbial composition of the rumen ecosystem [ 31 ]. Hydrogen recovery was higher in cattle species than yak species, meaning that more hydrogen was used by microbes in cattle species ( P <0.05, Table 3 ). Microbial growth is one of the key measurements in an in vitro system because of the important roles played by microbes in host health and energy supply [ 32 ]. A suitable marker is very important for differentiating microbial nitrogen from different parts of feed degradation. In the present study, 15 N was selected as marker due to its accuracy [ 33 ]. As shown in Table 4 , no differences were observed in ammonia production, daily production of non-NH 3 N, flows of microbial nitrogen, and SAM nitrogen derived from ammonia-N ( P >0.05). Although 15 N enrichment of NH 3 -N and LAM were lower in cattle species than in yak species, the proportion of LAM nitrogen derived from ammonia-N was greater for cattle species ( P <0.05). However, a lower enrichment of the SAM was found in both cattle and yak species compared with the LAM, which was in agreement with the previous comparative report regarding Merino sheep and Rusitec fermenters [ 32 ]. The 15 N enrichment was lower in the SAM than in the LAM (the mean values were 0.5210 and 1.5298 atom% in excess in cattle species, respectively, and 0.5842 and 1.7620 atom% in excess in yak species, respectively). The LAM was located in free rumen fluid and SAM was loosely/tightly attached to feed particles and associated with feed surfaces, where the ammonia concentration may be lower on the surfaces of feed particles than in the rumen fluid. As a consequence of the differences in 15 N enrichment between the LAM and the SAM, the percent of microbial nitrogen derived from NH 3 -N was lower in the SAM than in the LAM in both cattle and yak species ( Table 4 ). The results of the nitrogen differences in the SAM and LAM suggested that these two different parts should be taken into account to obtain a more reliable result in determining microbial protein production in further studies. 10.1371/journal.pone.0170044.t004 Table 4 Nitrogen content and daily production of ammonia-N, NAN, and microorganisms, the proportion of microbial N derived from ammonia in Rusitec fermenters. Item cattle yak SEM P -value Ammonia-N, mg/d 59.4 48.8 7.119 0.178 NAN, mg/d 96.7 102.2 3.984 0.204 Microbial N flow, Total microorganisms, mg/d 72.1 76.4 4.044 0.338 LAM, mg/d 35.6 34.4 3.422 0.737 SAM, mg/d 36.5 42.0 5.021 0.318 SAM % of total 50.6 54.5 4.750 0.444 N content of LAM, mg/g DM 59.8 62.7 2.786 0.394 N content of SAM, mg/g DM 62.4 64.8 0.973 0.114 15 N enrichment, atoms % excess LAM 1.5298 a 1.7620 a 0.053 0.015 SAM 0.5210 b 0.5842 b 0.043 0.276 Ammonia-N 3.2416 4.0361 0.067 0.001 Proportion of microbial N derived from ammonia-N, % LAM 47.2 a 43.7 a 1.25 0.049 SAM 16.1 b 14.5 b 1.41 0.367 DM-dry matter; LAM-liquid-associated microorganisms; SAM- solid-associated microorganisms; NAN- non-ammonia nitrogen. a-b Means within a column without common superscript letters differ between LAM and SAM ( P < 0.05). Considering the observed differences in gas production, digestibility, and microbial growth, the microbial composition may be the main reason for these results. Thus, we isolated genomic DNA from the LAM and SAM. A culture-independent method, next generation sequencing technology using Miseq-250 platform, was used to assess the microbial community structures. After quality control, a total of 962,074 reads were obtained for the V4 16S rRNA sequences, with an average of 80, 172 ± 6392 (SD) per sample. The average length of the sequence reads was 273 bp. The number of OTUs observed in this study reached 9, 794 based on a similarity threshold of 97% at species level. No differences were found for all alpha-diversity index between yak and cattle species from same fraction in Rusitec fermenters ( P >0.05, Fig 1 ), which was consistent with the previous comparison study of low and high methane production cattle [ 34 ]. However, alpha diversity index of the solid-associated microbes from the same animal group was higher than that of liquid-associated microbes. The lowest values of alpha diversity were found in the rumen liquid-associated microbes of yak species. The PCoA analysis using the UnWeighted Unifrac metric indicated that the samples clustered according to the different parts in Rusitec fermenters with different animals ( Fig 2 ). In total, 14 phyla were identified as being distributed across all the samples in Rusitec fermenters ( Fig 3 ). Bacteroidetes , Firmicutes , Spirochaetes , and Proteobacteria were dominant phyla, regardless of the group ( Fig 3 ), but their proportions varied among the groups, as has been found by many others regarding rumen microbe studies [ 34 ]. Six phyla were affected by different animal fractions ( Proteobacteria , Spirochaetes , Lentisphaerae , Verrucomicrobia , Fibrobacteres and Cyanobacteria ) ( P <0.05, S1 Table ). The phylum Proteobacteria was the highest in the LAM and lowest in the SAM of yak species. Cyanobacteria , Spirochaetes , and Lentisphaerae were the highest in LAM of yak species, SAM of yak species, and LAM of cattle species, separately. The Fibrobacteres was higher in SAM parts than in LAM regardless of different animals. Because Fibrobacteres was most important fiber degradation microbiomes [ 35 ]. Verrucomicrobia was higher in cattle than yak species regardless of the fractions. 10.1371/journal.pone.0170044.g001 Fig 1 Changes in alpha diversity values among different groups. 10.1371/journal.pone.0170044.g002 Fig 2 Principal coordinate analysis (PCoA) of the microbiota community based on UnWeighted Unifrac distance. 10.1371/journal.pone.0170044.g003 Fig 3 The impact of different fractions in Rusitec fermenters on the microbiota composition. In our study, yak produced less methane than cattle with the same diet in vitro semi-continuous culture system (Rusitec). Thus, we want to know the difference at genus level between yak and cattle with the same fractions. The fold2changes were showed in Figs 4 and 5 . RFN20 , vadinCA11 , Fibrobacter , Asteroleplasma , Succiniclasticum , Campylobacter , p−75−a5 , Coprococcus , Pseudobutyrivibrio , Odoribacter , Bifidobacterium , and Selenomonas were higher in LAM of cattle and only Pediococcus and Ruminobacter were higher in LAM of yak species. Methanoplanus , Clostridium , Pseudomonas , Moryella , and Shuttleworthia were higher in SAM of cattle species and BF311 , Agrobacterium , Methanobrevibacter , and L7A_E11 were higher in SAM of yak species. Higher abundances of bacteria as H 2 producers existed in cattle species, such as Coprococcus , Succiniclasticum , and Clostridium [ 36 ]. However, the dominant methanogen genus Methanobrevibacter was higher in the SAM of yak than cattle species [ 37 ]. While the new order of methanogen Methanomassiliicoccales (Genus vadinCA11 ) was higher in LAM of cattle species. The difference between cattle and yak species was not linked regularly with the methane production. That might be the reason of more diversity of bacteria existed in the cattle fermenters, which contributed to the substrate needed by methanogen Methanomassiliicoccales . There were a few higher abundance bacteria in the yak species, which would produce less H 2 , in agreement with the less H 2 production of yak compare with cattle ( Table 3 ). In order to confirm the function difference between yak and cattle species, we used PICRUSt to predict the function of the microbiomes based on 16S RNA sequenced data (Figs 6 and 7 ). The result showed that many functions were different between two species (Figs 6 and 7 ), however, we just focused on the ones about nutrition and energy metabolism to discuss. Energy metabolism, vitamin B6 metabolism, and methane metabolism were higher in LAM of yak species. The higher methane metabolism function in LAM of yak might be the consequence of less diversity of microbiome existed ( Fig 1 ). That difference might be the consequence of the in vitro study limits. In future, in vivo study should be conducted to compare the methane difference between yak and cattle. However, the predicted functions involved in sugar, fat, protein and amino acid were higher in LAM of cattle species. Vitamin B6 metabolism was higher in the yak species, which might be the consequence of the higher relative abundance of phylum Cyanobacteria [ 38 ]. Twenty-six predicted functions were higher in SAM of yak than cattle species, indicating more microbiome pathways existed to help yak to adapt the harsh environment. The further study should be conducted to confirm this hypothesis. 10.1371/journal.pone.0170044.g004 Fig 4 OTUs counts significantly different ( P < 0.05) between cattle and yak of rumen liquid. The size of the dot represents the value of the OTU, the higher value of the OTU number, the bigger size of the dot. The P value was modified by p.adjust value. 10.1371/journal.pone.0170044.g005 Fig 5 OTUs counts significantly different ( P < 0.05) between cattle and yak of rumen solid. The size of the dot represents the value of the OTU, the higher value of the OTU number, the bigger size of the dot. The P value was modified by p.adjust value. 10.1371/journal.pone.0170044.g006 Fig 6 Predicted KEGG functions count significantly different ( P < 0.05) between cattle and yak of rumen liquid. The size of the dot represents the value of the function genes, the higher value of the function genes number, the bigger size of the dot. The P value was modified by p.adjust value. The predicted data was generated by PICRUSt, the analysis and figure were conducted by R (3. 2. 3). 10.1371/journal.pone.0170044.g007 Fig 7 Predicted KEGG functions count significantly different ( P < 0.05) between cattle and yak of rumen solid. The size of the dot represents the value of the function genes, the higher value of the function genes number, the bigger size of the dot. The P value was modified by p.adjust value. The predicted data was generated by PICRUSt, the analysis and figure were conducted by R (3. 2. 3). In conclusion, the results showed that yak was potential “low carbon” ruminant. The different microbe compositions correlated with methane emissions. The data might be used to manipulate or provide useful information to reduce the environmental effects of other ruminants."
} | 4,425 |
35800325 | PMC9208393 | pmc | 3,635 | {
"abstract": "2,5-Furandicarboxylic acid (FDCA) is a monomer that can be used to produce bioplastic, which has gained increasing interest worldwide. The production of FDCA from catalytic oxidation of 5-hydroxymethylfurfural (5-HMF) is regarded as the major route of the utilization of 5-HMF. In this work, FDCA was produced in a tubular reactor packed with Pt/C. The effects of operating parameters including reaction temperature, molar ratio of 5-HMF/NaOH, volumetric flow rate of oxygen, pressure, and catalyst amount on the yield of product were investigated. High oxygen supply provided by high pressure or high oxygen flow rate effectively decreased the yield of FDCA. The optimal conditions were at the reaction temperature of 90 °C, the molar ratio between 5-HMF and NaOH of 1 : 8, the oxygen flow rate of 33.9 mL min −1 and atmospheric pressure, providing the yield of FDCA of 86.4% and full conversion of 5-HMF. The productivity of 0.1266 kg FDCA kg cat. −1 h −1 was achieved with the residence time of 0.285 s, which was considerably shorter than the reaction times reported in the literature. These results can be used as a foundation for further investigation and for developing a compact processing unit for the production of FDCA.",
"conclusion": "Conclusion The synthesis of FDCA via catalytic oxidation of 5-HMF using a continuous packed-bed reactor was performed in this work. Platinum supported on activated carbon was used as catalyst. The reaction system was operated with different reaction parameters including reaction temperature (75–90 °C), molar ratio of 5-HMF/NaOH (1 : 2 to 1 : 12), volumetric flow rate of oxygen (33.9–90.9 mL min −1 ), pressure (1–5.1 bar) and the amount of catalyst (0.05–0.09 g). Only trace amount of DFF was detected. The system was relatively stable except for conditions with oversupply of oxygen such as high oxygen flow rate or high pressure. Under these conditions, the yield of FFCA was much higher than that of FDCA. The yield of FDCA at the optimal conditions (reaction temperature of 90 °C, molar ratio of 5-HMF/NaOH of 1 : 8, oxygen flow rate of 33.9 mL min −1 , and atmospheric pressure) was 86.4%, corresponding to the productivity of 0.1266 kg FDCA kg cat. −1 h −1 . The short residence time of 0.285 s and no requirements on pressurized equipment are the benefits that can be further utilized effectively for the production of FDCA.",
"introduction": "Introduction The synthesis of bio-based chemicals and polymers from renewable biomass is the sustainable alternative choice to reduce dependence on fossil-fuel resources. Biomass-derived 2,5-furandicaboxylic acid (FDCA) is a high value-added chemical that can be used to substitute the petrochemical-derived PTA (pure terephthalic acid), a principal precursor to polyethylene terephthalate (PET). 1 FDCA can be used for the production of polyethylene furanoate (PEF). 2,3 It can be synthesized from the oxidation of 5-hydroxymethylfurfural (5-HMF) which is derived from dehydration of carbohydrate. 4,5 Catalytic oxidation of 5-HMF to produce FDCA is a complex reaction involving several intermediate molecules such as 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), 2,5-diformylfuran (DFF) and 5-formyl-2-furancarboxylic acid (FFCA) in the system. 6–8 The synthesis of FDCA has been performed in the conventional batch process. 1,9–13 Davis et al. investigated the oxidation of 5-HMF over supported Pt, Pd, and Au catalysts in the semi-batch reactor. 10 The selectivity of FDCA was 79% and 71% over Pt and Pd, respectively, at 22 °C, 690 kPa O 2 with the reaction time of 6 h. The selectivity of FDCA was about 31% and 72% over Au catalyst when the reaction time was 6 h and 22 h, respectively. Schade et al. studied the oxidation of 5-HMF using supported gold- and silver-based catalysts in an autoclave as a batch reactor. 13 The yield of FDCA of approximately 75% was obtained over Au/ZrO 2 at 100 °C, under a pressure of 10 bar with the reaction time of 5 h. Similarly, Casanova et al. reported the FDCA yield of 78% with the reaction time of 24 h for Au/C catalyst at the reaction temperature of 65 °C and the pressure of 10 bar. 14 Although the full conversion of HMF has often been reported in the literature, improving the reaction time and the yield of FDCA still requires further development. Long reaction time was due to the gas–liquid film resistance in the system. Davis et al. mentioned that the flux of O 2 through the gas–liquid film was the limiting step. 10 In order to improve the productivity, the development of a continuous system for the production of FDCA is necessary. Recently, a continuous electrochemical oxidation of 5-HMF to produce FDCA using Ni/NiOOH foam electrodes was demonstrated. A two-step process was required to achieve full conversion of 5-HMF, providing the yield of FDCA of 82–90%. 15 However, scaling up the production to the industrial capacity can be a challenge. Packed-bed reactor can also be applied for the production of FDCA as a continuous process; however, no literature data is available. Based on the development using batch processing, it is possible to shorten the time required for the chemical reactions using a flow reactor, as the mixing is not the major effect. For instance, Chueluecha et al. shortened the reaction time from 4–8 h for batch reactor to 15 min for biodiesel production in a packed microchannel reactor. 16 Sonsiam et al. drastically shortened the reaction time (residence time) for the dehydration of fructose to 5-HMF in a continuous flow reactor compared to the batch reactor. 17 The synthesis of FDCA can be performed using either homogeneous or heterogeneous catalysts. Partenheimer and Grushin demonstrated the synthesis of FDCA from 5-HMF using metal/bromide catalysts, achieving 60% of FDCA yield in a batch system at 125 °C and 3 h of reaction time. 18 Although the use homogeneous catalyst is possible, the separation and recycling of catalyst can be difficult. Heterogeneous catalysts have been developed for this process to avoid the separation of catalyst and product. Among heterogeneous catalysts, noble metal catalysts such as Pt, Pd and Au on various supports have been used in many studies for the synthesis of FDCA, providing good activity under mild conditions. 1,10,12 Sahu and Dhepe achieved 96% of FDCA yield under normal oxygen pressure using Pt/γ-Al 2 O 3 as catalyst at the optimal reaction conditions of a stepwise increase in the temperature (75 and 140 °C for 12 h each). 19 Therefore, in this contribution, we developed a continuous system to enhance the productivity for the production of FDCA via partial oxidation of 5-HMF in a packed-bed reactor. Pt supported on activated carbon (Pt/C) was used as catalyst to improve the production of FDCA. The conversion of HMF, yield of FDCA and reaction time were monitored to evaluate the reactor performance. The effects of liquid and gas flow rates, the ratio of HMF/NaOH, temperature and pressure were investigated. The performance of our system was also compared to that of the conventional batch and semi-batch reactors."
} | 1,769 |
37682994 | PMC10491213 | pmc | 3,637 | {
"abstract": "Liquid entrainment with a solid architecture passing through the fluid-fluid interface is ubiquitous and widely used in industrial processes as a liquid transfer method. Besides liquid properties, solid structures play a core role in entrainment. Although the influence of its macroscopic curvatures and microscale roughness has attracted years of research, the effect and potential of the commonly seen millimetric structures have not been sufficiently explored and exploited. Here, we demonstrate enhanced liquid entrainment on the millimetric structured surface by the co-effect of viscosity and capillarity for sustained liquid transfer of small deviation, including high-quantity uptake and practically operational drainage with small and relatively uniform droplet dripping time of varied liquid viscosities. With the overall process of viscous-capillary entrainment, we achieve stable cyclical arrayed liquid transport, showing its potential for sustained liquid transfer in intractable situations in laboratory, industry, and even daily life.",
"introduction": "INTRODUCTION Liquid entrainment occurs when a solid architecture passes through the interface of two fluids, forming either liquid droplets or film on the solid surface. However, from the practical perspective, the whole entrainment process, including the liquid uptake with a solid surface withdrawn from a liquid bath and the liquid drainage after the withdrawal, still lacks overall concern, understanding and control. As entrainment phenomenon is ubiquitous in nature, such as nectar feeding by dipping with tongues ( 1 – 5 ) and water entraining while aquatic animals are leaping out ( 6 ), and widely applied in industrial processes across various scales, such as film coating on devices and products ( 7 – 9 ), the investigation and control of the liquid entrainment is important for industrial applications and, therefore, has raised broad interests for scientific research ( 10 – 15 ). Since Goucher and Ward’s pioneering work of the liquid entrainment dynamics on metal strips and wires in 1922 ( 10 ), extensive fundamental studies have been made to understand the liquid uptake process by entrainment on solid surfaces. The classic works of Landau and Levich ( 16 ) and then Derjaguin ( 17 ) quantified the thickness of the liquid film entrained by a flat plate withdrawn at a constant speed, which is now referred to as the LLD theory ( 18 ). Modifications were then made to consider different regimes and factors, such as expansion and correction terms ( 19 ), inertial effect ( 20 ), shape and confinement of liquid bath ( 21 , 22 ), and rheological properties of liquid ( 23 ). However, the consideration of the different aspect of the process, the liquid drainage after uptake, was relatively few and rarely integrated with the liquid uptake process in the practical view. Besides liquid properties, surface macroscopic curvatures influence the overall pressure gradient during the liquid entrainment ( 10 ), and microscale roughness can trap the liquid and change the friction state ( 24 – 26 ). However, the role of millimetric structure in the liquid entrainment process has not received adequate attention, although natural creatures have been found to take advantage of millimetric structure to enhance the liquid entrainment in some everyday situations, such as nectar feeding by bees and bats with their hairy tongues ( Fig. 1A ) ( 27 , 28 ). The lack of research is partly due to a lack of fabrication method and the complexity of dynamic effects involved at the millimetric scale. With the development of advanced manufacturing in recent years, such as three-dimensional (3D) printing, chemical and optical etching, and precision sculpting, the millimetric structured surfaces are designed and applied to open fluidics ( 29 – 31 ), thermal control ( 32 , 33 ), and wet state adjustment ( 34 , 35 ) to achieve higher efficiency than macro- or microstructures. Therefore, an in-depth study on the dynamic properties of the bio-inspired millimetric structure–modified entraining will lead to insight into the one-century-old question of liquid entrainment and provide possibilities for liquid transfer. Fig. 1. Liquid entrained by rod with arrayed millimetric structures. ( A ) Honey bee and glossophagine bat feed on nectar (left) with the help of their millimetric hairy tongues (right). ( B ) Schematic of a rod with periodic millimetric spiky plates. ( C ) Micro-CT images of a 3D-printed rod in the front (i) and cross-section (ii) views, with its axial plate distance D , plate angle α, axis radius R , and plate height H being marked out (1.5 mm, 30°, 1.0 mm, and 1.0 mm shown here, respectively). ( D ) Experimental setup for observation and measurement of liquid entrainment. The rod is withdrawn by the lifter from the liquid bath at a constant speed U , and the process is recorded by the high-speed camera. The balance measures the mass before the immersion and after the withdrawal of the rod. The coordinate origin is placed at the surface of the liquid bath ( z = 0). ( E ) Typical process: a rod with α of 30° or 150° was withdrawn from the liquid bath (silicone oil with a viscosity of 100 mPa·s) at a constant speed of 18.00 mm s −1 and entrained the oil. Sequence images go as four typical states from left to right: initial state, liquid uptake state, rod-bath separated state, and liquid drainage (droplet dripping) state. ( F ) Liquid uptake capacity of the millimetric structured rod (red) or bare rods (black) represented by the average volume of the entrained liquid per unit length along the rod’s axial direction (inset). Photo credit of bee image in (A): Z. Dong, Chinese Academy of Sciences. Bat image in (A) reproduced with permission from ( 1 ). The data are shown as mean ± SD ( N = 3), and the error bar represents SD. Note that some errors are smaller than the size of the symbol. Here, we demonstrate the whole entrainment process and the millimetric structure enhanced liquid entrainment. We perform theoretical analysis and reveal a co-effect of viscosity and capillarity at the millimetric scale that improves the quantity and robustness of the entrainment. The co-effect dominates the entraining processes during and after the withdrawal of solid surface, leading to liquid uptake with high quantity and small deviation, as well as liquid drainage and droplet dripping with operationally high frequency and small differences between liquids with varied viscosities. The fundamental understandings of the whole uptake and drainage process are integrated and applied to stable cyclical fetching, transporting, and depositing of viscous liquids from confined space, showing its potential for sustained liquid transfer in intractable situations and providing broader guidance for the assessment and design of liquid clinging and coating on the uneven surfaces in daily life, laboratory operation, and industrial production.",
"discussion": "DISCUSSION The liquid entrainment enhanced by the millimetric structure has unique superiority as the liquid transfer method in some intractable situations. A modularized apparatus for sustained liquid transfer was performed ( Fig. 5B ). A programmable multi-axis displacement table held a detachable array of structured rods, the core component, and could move vertically and horizontally at the set speeds and distances. The liquid sources consisted of well-arranged 10-ml centrifuge tubes with tested liquid inside. The liquid acceptors were installed on a single-axis mobile platform and could move in ± y direction to imitate conveyor belts in the industry. The rod array fetched, transported, and deposited liquid from the sources to the acceptors ( Fig. 5C ). First, we took glycerol as the example liquid, which is a high-viscosity liquid frequently used for standard study ( 41 ). Controlled by the multi-axis displacement table, the rod array cyclically transferred the dyed glycerol from the liquid sources to the liquid acceptors ( Fig. 5D and movie S2). The liquid acceptors (1.5-ml centrifuge tubes) were controlled by the single-axis mobile platform, moving in − y direction in a square-wave mode to match the liquid transfer period, ensuring that the following next cycle droplets drip just right into the successive liquid acceptors. Glycerol dyed in different colors was steadily deposited together by the programmed transfer ( Fig. 5D ). As shown in Fig. 5E , the cycling process was stable, with a dripping time deviation of ~9% among rods in one cycle and ~3% among cycles, and the mass of transferred liquid in one cycle of 0.30 ± 0.04 g (0.05 ± 0.01 g for each rod on average in this experiment). This transferred liquid quantity (~2.4 × 10 2 μl) matches the quantity range of the widely used commercial positive displacement pipettes ( 42 ) and can be easily adjusted by changing the number of detachable rods. The residual amount on the rod array was of a constant value of 0.74 ± 0.04 g, which would not grow in the cycling process and, therefore could be neglected after enough cycles. With aforesaid verification of method feasibility and stability, we then dispensed photosensitive resin in place of glycerol by the same apparatus plus an ultraviolet (UV) lamp ( Fig. 5F ). A substrate made of the same resin (3D-printed) ( Fig. 5G ) or polydimethylsiloxane (PDMS; thermal cured) (fig. S8) was placed on the single-axis mobile platform in advance. The resin droplets dripped on the substrate and were quickly cured under the UV light at the short time interval between two transfer cycles. An array of cured resin droplets was achieved and increased the transmittance of the rough resin substrate by ~50% for UV and visible light and ~20% for infrared light ( Fig. 5H ). Thus, the cured droplets could perform as an array of lenses to see the deformed images of objects behind (natural scenery in Fig. 5G and light matrix in fig. S8D). One step further and we placed the unpackaged blue light-emitting diode (LED) chips in advance. The resin droplets straightly dripped on and covered the LED chips, and were then cured under the UV lamp. LED chips covered with droplets of different sizes resulted in different lighting effects. Without cured resin covering, the main lighting area was localized around the chip ( Fig. 5I and fig. S9N). By contrast, the light was refracted when passed through the cured resin, and the main lighting area expanded ( Fig. 5I and fig. S9, S, M, and L). The apparatus can easily manipulate the droplet size (Materials and Methods). Therefore, batch fabrication, packaging, and modification of lenses, chips, or other similar devices can be achieved in the way of facile operation with the enhanced liquid entrainment method ( 43 , 44 ). In conclusion, we investigate the influence of the millimetric structure on the liquid entrainment by a withdrawn solid. The co-effect of the enhanced viscous friction and reversed capillary pressure gradient enhance the entrainment's quantity and robustness. The retreat and secondary entrainment on the structure modifies the shape and behavior of the liquid, resulting in a moderate and self-adapted entraining mode to manipulate the drainage and dripping of the entrained liquid. The viscous-capillary entrainment provides enhanced liquid uptake with high quantity and small deviation, as well as liquid drainage and droplet dripping with operationally high frequency and small differences between liquids with varied viscosities. As a proof of concept, the millimetric structured surface is arrayed and applied to fetching, transporting, and depositing viscous liquid from a confined space in a facile and sustained way. These results will contribute to the design of liquid collecting and dispensing methods, and potentially to the liquid control technology for various applications ( 45 , 46 ). In contrast with routine methods, the viscous-capillary entrainment on the withdrawn millimetric structured surface ensures sustained liquid transfer with high quantity, frequency, and robustness, and is suitable for some intractable situations or even special operations. For example, one could collect liquid with unknown composition from an opaque and confined space for further detection. In a broader view of understanding, controlling, and using liquid entrainment dynamics, uneven surfaces with millimetric morphology are ubiquitous, from toys being painted and nasopharyngeal swabs for virus testing ( 47 ), to tracks with lubricant and tires through a puddle ( 48 , 49 ), and even intestinal villi absorbing nutrients ( 50 ). Relative motion often occurs between these surfaces and liquid droplet, flow, bath, and water body, requiring analysis from the perspective of liquid entrainment. Therefore, our work can provide broader inspiration for the assessment of liquid clinging or material coating on uneven surfaces in daily life, laboratory operation, and industrial production."
} | 3,235 |
35108295 | PMC8809616 | pmc | 3,638 | {
"abstract": "Organic fertilizers and especially microbial biomass, also known as microbial fertilizer, can enable a paradigm shift to the conventional fertilizer-to-food chain, particularly when produced on secondary resources. Microbial fertilizers are already common practice (e.g. Bloom ® and Synagro); yet microbial fertilizer blends to align the nutrient release profile to the plant’s needs are, thus far, unexplored. Moreover, most research only focuses on direct fertilization effects without considering added value properties, such as disease prevention. This study has explored three promising types of microbial fertilizers, namely dried biomass from a consortium of aerobic heterotrophic bacteria, a microalga ( Arthrospira platensis ) and a purple non-sulfur bacterium ( Rhodobacter sphaeroides ). Mineralization and nitrification experiments showed that the nitrogen mineralization profile can be tuned to the plant’s needs by blending microbial fertilizers, without having toxic ammonium peaks. In a pot trial with perennial ryegrass ( Lolium perenne L.), the performance of microbial fertilizers was similar to the reference organic fertilizer, with cumulative dry matter yields of 5.6–6.7 g per pot. This was confirmed in a pot trial with tomato ( Solanum lycopersicum L.), showing an average total plant length of 90–99 cm after a growing period of 62 days for the reference organic fertilizer and the microbial fertilizers. Moreover, tomato plants artificially infected with powdery mildew ( Oidium neolycopersici ), a devastating disease for the horticultural industry, showed reduced disease symptoms when A . platensis was present in the growing medium. These findings strengthen the application potential of this novel class of organic fertilizers in the bioeconomy, with a promising match between nutrient mineralization and plant requirements as well as added value in crop protection.",
"conclusion": "Conclusions This study shows that microbial biomass is an excellent means for innovative and sustainable nutrient recycling, yielding microbial fertilizers very well capable of replacing conventional organic fertilizers. Depending on the specific fertilization needs of a particular crop, a tailor-made slow-release organic fertilizer blend can be composed and supplied gradually, maximally aligning with the plant’s needs throughout its growing period. Improved plant product quality and disease resistance can also be accomplished by microbial fertilizers, rendering their application interesting not only in the context of sustainability but also from an economic point of view.",
"introduction": "Introduction Our industrialized global society is heavily dependent on synthetic inorganic fertilizers for primary crop production in both agriculture and horticulture. It is estimated that over 110 million tons of N fertilizer (e.g. ammonium nitrate, ammonium sulfate and urea), 10.3 million tons of P and 15.6 million tons of K are annually consumed [ 1 , 2 ]. Intensive fertilizer use has, however, serious environmental and economic repercussions [ 3 ]. When fertilizers are applied to the soil or in growing media (GM) for agricultural and horticultural applications, they are not completely consumed by plants, yet suffer from inefficiencies such as leaching, runoff and volatilization. For every 100 units of fertilizer applied to the land, only 4–14 units of nitrogen and 17 units of phosphorus are eventually consumed by humans [ 4 – 6 ]. A considerable portion of these fertilizers ultimately end up in the environment, resulting in detrimental effects on water quality (e.g. eutrophication), air quality (e.g. emissions of ammonia and nitrous oxide), the greenhouse gas balance (e.g. nitrous oxide), biodiversity and soil quality (e.g. acidification of soils) [ 7 , 8 ]. Recycling nutrients from secondary resources may offer an improvement to the overall efficiency of the fertilizer-to-food chain [ 9 ]. Solid and liquid by-products and residues from plant and animal origin, such as animal manures, animal slurries, blood meal, cocoa shells, soybean meal and organic waste from restaurants and supermarkets play a central role [ 10 ]. These so-called organic fertilizers slowly release nutrients through decomposition or decay imposed by the microbiome present in the growing medium or the soil [ 11 ]. Organic fertilizers may provide benefits for the soil, such as better water retention, improved nutrient retention and an increase in the organic matter content of the soil, which on its account buffers the soil against salinity, pH changes and pesticides [ 12 – 15 ]. These organic fertilizers contribute to 5% of the total fertilizer market, yet their share will become more significant, as their annual growth rate was around 14% in 2019 compared to only 4% for the total fertilizer market [ 16 ]. Microbial fertilizers constitute a novel and promising class of organic fertilizers based on using microbial biomass as a source of plant nutrients [ 17 , 18 ]. The microbes can be produced on secondary resources such as industrial wastewater, sewage, manure, etc [ 19 , 20 ]. Microbial biomass can serve as a multi-nutrient fertilizer mainly rich in nitrogen (7–9 g N 100 g -1 dry matter) with an elemental composition of C 4.2 O 1.8 H 0.8 NP 0.2 S 0.1 K 0.1 [ 18 , 21 ]. In principle, dried microbial biomass produced on secondary resources is mixed with a growing medium or soil [ 17 , 18 ]. The microbiome present in the growing medium or soil will then mineralize the microbial biomass, thereby making nutrients available for plant growth [ 11 ]. Three types of microbial fertilizers are generally considered for microbial fertilizer production on secondary resources: (i) a consortium of aerobic heterotrophic bacteria (AHB); [ 22 ], (ii) photoautotrophic microalgae (MA); [ 23 , 24 ] and (iii) photoheterotrophic purple non-sulfur bacteria (PNSB); [ 17 , 18 , 25 , 26 ]. AHB are probably the most widely applied and explored microbial fertilizers. A very common type of AHB is treated and stabilized sewage sludge a.k.a. biosolids (Bloom® and Synagro). In Europe, for example, around 10 million tons dry matter is annually produced [ 27 ]. Extensive research has been performed on the mineralization and fertilization properties of biosolids. A review of 32 studies by Rigby, Clarke showed that the mineralizable nitrogen decreased with increasing biological stabilization [ 28 ]. In terms of fertilization, increased crop yields have been reported for different plants such as rice, radish, wheat and barley [ 29 ]. Added value properties such as plant protection have, according to the authors’ knowledge, not yet been reported. MA have also been explored as a microbial fertilizer, yet not as extensively as AHB. The mineralization of different types of MA has been studied, showing different final plant-available N fractions for Nannochloropsis biomass (31% after 95 days); [ 24 ], microalgal bacterial flocs (25% after 95 days); [ 8 , 30 ], Arthrospira platensis biomass (72% N after 77 days); Spanoghe, Grunert (18) and algal biomass grown on manure effluents (41% after 63 days). In terms of fertilization, several plants have been explored such as cucumber and cord seedlings, parsley ( Petroselinum crispum ), petunia and tomato ( Solanum lycopersicum L.), showing an equal or improved performance compared to commercial inorganic and/or organic fertilizers [ 8 , 18 , 24 , 30 ]. Relatively to AHB, added value properties of MA and their extracts have extensively been reported in terms of (a)biotic plant protection (e.g. antifungal, bacterial and antinematodal activity, alleviation of drought and salt stress) and biostimulation (e.g. increase in germination rate, salt tolerance, nutritional value, etc.) [ 31 ]. PNSB are probably the most novel type of microbial fertilizer. Mineralization studies on PNSB biomass are, however, limited to our previous research on Rhodobacter sphaeroides . A final plant-available N fraction of 70% was observed after 77 days [ 18 ]. Also, as fertilizer, only a few plants were tested such as pasture ryegrass ( Lolium rigidum Gaudin ), mandarine tree ( Citrus reticulata ) and parsley ( Petroselinum crispum ) [ 18 , 20 , 32 ]. Nonetheless, all results showed a positive or comparable effect of PNSB biomass on plant growth or fruit quality and production relative to the control [ 18 , 20 , 32 ]. In terms of added value properties of PNSB for plants, multiple researchers have reported growth promotion and alleviation of environmental stress [ 17 ]. Although AHB, MA and PNSB have been studied as a source of microbial fertilizers, research is mainly limited to their individual contribution to plant growth. Our previous research is one of the first to study blends and indicated that it is economically sensible because it can provide cheap nutrients and added value for plants [ 18 ]. The goal of this study was to gain a broader understanding of the applicability across horticultural and agricultural for microbial fertilizers based on AHB, MA and PNSB, both individually as well as in blends. First, nitrogen mineralization and nitrification profiles were determined in a commercially relevant growing medium. Second, the fertilization effect of the microbial fertilizers was evaluated on the plant growth performance in pot trials with perennial ryegrass ( Lolium perenne L.) and tomato. Third, disease susceptibility towards a biotrophic fungus ( Oidium neolycopersici ) and two necrotrophic fungi (i.e. Alternaria solani and A . alternata ) was assessed for tomato, to explore added value properties of the three microbial fertilizers.",
"discussion": "Results and discussion Fertilizer mineralization and nitrification During this experiment, the release pattern of ammonium and nitrate for the individual and blended microbial fertilizers was characterized ( Fig 2 ). Minor nitrogen mineralization was observed in the negative control, whereas divergent mineralization patterns were found for the different fertilizer treatments. The conventional organic fertilizer SF2 was characterized by relatively low ammonium levels and rapidly increasing nitrate levels. The AHB fertilizer triggered a gradual release of both ammonium and nitrate, as did the blends containing this microbial fertilizer. MA showed high and almost unchanging ammonium concentrations, while the nitrate concentration showed a greater increase with passing incubation period. PNSB produced minor nitrate concentrations, while ammonium concentrations were ascending steadily during the first 30 days of the incubation period and descending behavior thereafter. 10.1371/journal.pone.0262497.g002 Fig 2 Nitrogen mineralization and nitrification: Profiles of total ammoniacal nitrogen, nitrate nitrogen and pH. (A) organic growing medium without fertilizer (different primary y-axis), (B) conventional organic fertilizer SF2, (C) aerobic heterotrophic bacteria (AHB), (D) microalga (MA), (E) purple non-sulfur bacterium (PNSB), (F) 85% AHB and 15% MA, (G) 85% AHB and 15% PNSB and (H) 85% AHB, 12.5% MA and 12.5% PNSB. All individual or fertilizers blends were supplied at 352 mg-N L -1 . Error bars represent the standard deviations. Significant differences among treatments are indicated per sampling point by letter code for ammonia nitrogen, nitrate nitrogen and pH, resp. in italic, bold and grey. Mineralization started after an incubation period of 14 days. After 42 days, biological conversion of organic nitrogen to ammonium and nitrate was approximately 88% for MA, around 68% for SF2, AHB, AHB+MA 85/15 and AHB+PNSB 85/15, around 62% for the blend AHB+MA+PNSB 85/7.5/7.5 and 34% for PNSB. The final ammonium to nitrate ratio for the conventional organic fertilizer was 1:5, while AHB, MA, PNSB, and the three blends had a final ammonium to nitrate ratio of respectively 1:0.6, 1:0.6, 1:0.8, 1:0.2, 1:1.4, 1:0.7 and 1:1. Research indicates that the highest plant yields are achieved by a combined supply of both ammonium and nitrate [ 37 , 38 ]. Crops may be classified into four types: (i) preference to ammonium; (ii) preferences to nitrate; (iii) equal effect of ammonium and nitrate; (iv) combinative use of the two nitrogen sources being superior to either ammonium or nitrate alone. Recent results indicate that aerobes, phototrophs and combinations thereof could be used as organic fertilizer for plants with different nutrient requirements and nitrogen preferences [ 39 ]. pH differences in the growing medium/fertilizer mixtures were minor and did not show an effect over time. The organic fertilizers SF2, on the other hand, had a pH-lowering effect. When organic nitrogen is converted to ammonia it directly withdraws an H + to form ammonium, so an OH - remains and triggers an increase of the pH of the growing medium, as observed during the first 14 days of the experiment. The pH decline for SF2 during days 28–42 was due to microbial conversion of ammonium to nitrate (196 mg NO 3 — N/L produced, see Fig 2 ). Fertilization pot trial with perennial ryegrass Perennial ryegrass pot trials were performed to study the effect of the microbial fertilizers on plant growth. The cumulative aboveground dry matter yields of all fertilized treatments outperformed the negative control in both pot trials from the fourth and the fifth cut onwards ( Fig 3 ). The microbial fertilizers performed equally well as the conventional organic fertilizer SF2. No ammonium toxicity symptoms were detected. A similar performance was reported for an enriched PNSB biomass ( Rhodopseudomonas sp.) produced on pig farm wastewater in a pot trial with pasture ryegrass ( Lolium rigidum Gaudin ), showing an equal performance in terms of shoot dry weight compared to an inorganic fertilizer. MA biomass composed of Chlorella sp. and Scenedesmus sp., on the other hand, resulted in a lower shoot dry weight than the inorganic fertilizers. 10.1371/journal.pone.0262497.g003 Fig 3 Perennial ryegrass: Cumulative aboveground dry matter (DM) yield. A non-fertilized negative control (Neg control) was compared to a conventional organic fertilizer (SF2) and three microbial fertilizers, i.e. aerobic heterotrophic bacteria (AHB), microalga (MA) and purple non-sulfur bacteria (PNSB) individually or in blends (according to nitrogen supply). Significant differences among treatments are indicated by letter code per sampling point, while error bars represent the standard deviation. Plants have an impact on their growing medium by exudation of organic substances in the rhizosphere, influencing microbial community composition and activity, nutrient cycling and pH [ 40 ]. Therefore, the nitrogen mineralization pattern observed in crop-free growing medium can differ somewhat from the pattern observed in practice, when plants are grown in the growing medium. Regarding plant performance, the ryegrass pot trials demonstrated that the microbial fertilizers performed equally well as the reference organic fertilizer in aboveground dry matter production. Fertilization and plant protection pot trial with tomato The tomato pot trial was performed to study the effect of individual and blends of microbial fertilizers on plant growth and explore tolerance towards plant-pathogenic fungi. Plant lengths are summarized in Table 3 . After 35 days of growth, no significant difference in plant length was observed between the negative control and the fertilizers, except for AHB+MA 85/15, AHB+MA+PNSB 85/7.5/7.5 and AHB+MA 75/25 ( Table 3 ). Plant length after 45, 55 and 62 days, however, was significantly higher for all fertilized treatments than the negative control. Significant differences in plant length among fertilized treatments were only found after a growing period of 45 days. At 55 and 62 days, all fertilizers performed equally, thereby, showing that the nutrient release pattern of the microbial fertilizers fits the actual need of the plant. 10.1371/journal.pone.0262497.t003 Table 3 Pot trial with tomato: Total plant length (mean ± standard deviation per treatment). A non-fertilized negative control (Neg control) was compared to a conventional organic fertilizer (SF2) and three microbial fertilizers, i.e. aerobic heterotrophic bacteria (AHB), microalga (MA) and purple non-sulfur bacteria (PNSB), individually or in blends (according to nitrogen supply). Treatment code Total plant length (cm) after 35 days 45 days 55 days 62 days Neg control 15.7 ± 2.0 a 23.3 ± 2.9 a 39.6 ± 6.1 a 47.3 ± 5.0 a SF2 15.7 ± 2.0 a 33.3 ± 2.9 b 77.8 ± 7.6 b 98.9 ± 8.0 b AHB 16.5 ± 2.2 ab 36.2 ± 3.6 bc 77.2 ± 8.3 b 90.1 ± 9.9 b MA 17.2 ± 2.0 ab 38.2 ± 2.5 c 82.2 ± 7.4 b 96.9 ± 3.3 b AHB+MA 85/15 17.4 ± 1.7 b 37.9 ± 3.0 c 82.1 ± 7.9 b 94.9 ± 7.8 b AHB+PNSB 85/15 16.4 ± 2.3 ab 35.8 ± 3.3 bc 79.4 ± 6.4 b 90.7 ± 7.2 b AHB+MA+PNSB 85/7.5/7.5 17.5 ± 1.5 b 37.6 ± 3.2 c 84.3 ± 2.3 b 96.2 ± 11.3 b AHB+MA 75/25 17.5 ± 2.3 b 36.7 ± 3.9 bc 80.2 ± 7.5 b 98.1 ± 9.8 b AHB+PNSB 75/25 16.8 ± 1.8 ab 36.3 ± 3.3 bc 80.5 ± 6.1 b 98.6 ± 2.8 b AHB+MA+PNSB 75/12.5/12.5 17.1 ± 1.6 ab 36.6 ± 3.6 bc 80.4 ± 6.9 b 95.4 ± 7.6 b AHB+MA 95/5 16.5 ± 1.6 ab 37.2 ± 3.0 c 79.9 ± 8.8 b 97.4 ± 10.3 b AHB+PNSB 95/5 16.1 ± 1.2 ab 36.2 ± 2.3 bc 79.8 ± 6.1 b 97.6 ± 0.6 b Significant differences among treatments are indicated by letter code. Table 4 presents the leaf greenness index, tomato yield and powdery mildew infestation. The leaf greenness index did not vary much among treatments, but the negative control had a significantly lower chlorophyll content than MA, AHB+MA+PNSB 85/7.5/7.5 and AHB+PNSB 75/25. The negative control did not develop quickly enough during the experiment to facilitate tomato harvesting, resulting in significantly higher tomato yield for all fertilized treatments except for AHB+MA 85/15. Susceptibility towards powdery mildew did not differ significantly among treatments. However, the presence of MA in the growing medium significantly reduced powdery mildew infection (p-value 0.000). No significant differences among treatments were observed regarding Alternaria leaf spot, nor was there a significant influence of the presence of particular microbial fertilizers ( Fig 4 ). Ammonium toxicity, detectable by leaf chlorosis and stunted growth [ 41 ], was not observed for any of the treatments. 10.1371/journal.pone.0262497.g004 Fig 4 Pot trial with tomato: Alternaria leaf spot (mean ± standard deviation) for a selection of treatments. A conventional organic fertilizer (SF2) was compared to three microbial fertilizers, i.e. aerobic heterotrophic bacteria (AHB), microalga (MA) and purple non-sulfur bacterium (PNSB) individually or in blends (according to nitrogen supply). Significant differences among these treatments are indicated by letter code. 10.1371/journal.pone.0262497.t004 Table 4 Pot trial with tomato: Leaf greenness index (LGI) after 94 days, tomato yield and disease index upon infection with powdery mildew (mean ± standard deviation per treatment). A non-fertilized negative control (Neg control) was compared to a conventional organic fertilizer (SF2) and three microbial fertilizers, i.e. aerobic heterotrophic bacteria (AHB), microalga (MA) and purple non-sulfur bacteria (PNSB), individually or in blends (according to nitrogen supply). Treatment code LGI Tomato yield Powdery mildew (SPAD units) (g plant -1 ) (disease index) Neg control 33.4 ± 3.2 a 0 ± 0 a n.d. SF2 35.7 ± 2.1 ab 576 ± 154 b 1.00 ± 0.00 a AHB 36.8 ± 2.6 ab 429 ± 237 b 1.00 ± 0.00 a MA 37.9 ± 3.5 b 516 ± 107 b 0.58 ± 0.49 a AHB+MA 85/15 34.9 ± 2.9 ab 283 ± 255 ab 0.75 ± 0.27 a AHB+PNSB 85/15 36.3 ± 3.2 ab 339 ± 184 b 0.83 ± 0.26 a AHB+MA+PNSB 85/7.5/7.5 37.7 ± 1.0 b 451 ± 98 b 0.75 ± 0.42 a AHB+MA 75/25 37.5 ± 3.0 ab 459 ± 234 b 0.75 ± 0.27 a AHB+PNSB 75/25 37.6 ± 2.8 b 545 ± 201 b 1.00 ± 0.00 a AHB+MA+PNSB 75/12.5/12.5 36.1 ± 3.3 ab 494 ± 183 b 0.75 ± 0.42 a AHB+MA 95/5 37.3 ± 2.7 ab 627 ± 156 b 0.58 ± 0.49 a AHB+PNSB 95/5 35.7 ± 2.1 ab 529 ± 165 b 1.00 ± 0.00 a Significant differences among treatments are indicated by letter code. This pot trial demonstrated similar plant performance of the microbial fertilizers compared to SF2, with no significant differences in tomato yield. Therefore, it can be stated a posteriori that the conventional fertilizer and microbial fertilizers had a nutrient release pattern well aligned to the plant’s needs. An explanation for the finding that the presence of MA in the growing medium significantly lowered powdery mildew infestation could be attributable to the presence of plant growth-promoting substances (e.g. phytohormones, vitamins and carotenoids; [ 31 , 42 , 43 ]. Several authors have detected phytohormones like indole acetic acid and jasmonic acid in MA, promoting plant growth and supporting plants to cope with biotic as well as abiotic stress [ 44 – 49 ]. Kępczyńska and Król [ 42 ] demonstrated that the jasmonic acid derivative methyl jasmonate can elicit induced systemic resistance towards Alternaria in tomato. They applied methyl jasmonate by seed soaking or fumigation of seedlings, yet providing this plant hormone using a MA containing fertilizer would be a practical alternative strategy. In a greenhouse trial with tomato, Coppens, Grunert [ 24 ] compared three different organic fertilizers incorporated in an organic growing medium: a conventional organic fertilizer and two MA-based organic fertilizers. No significant differences were encountered on plant growth parameters among the three organic fertilizers, yet the MA-based fertilizers significantly reduced tomato yield compared to the conventional organic fertilizer. Fruit quality and thus also market value were, however, enhanced by the application of MA-based fertilizers. Therefore, a tailor-made fertilizer blend would allow combining excellent tomato fruit quality with satisfactory fruit yield, while preserving sustainable cultivation practices (28, 29). Moreover, a gradual increase of the organic nitrogen supply has been demonstrated to improve yields [ 50 ]. The use of PNSB as organic fertilizer has also been studied and improved fruit quality [ 51 , 52 ]."
} | 5,519 |
35890572 | PMC9321720 | pmc | 3,639 | {
"abstract": "The paper addresses the synthesis of a nano-fibre network by coaxial electrospinning, embedding the healing agent dicyclopentadiene (DCPD) in polyacrylonitrile (PAN) fibres. Compared to other encapsulation methods, the use of nano-fibres filled with healing agent have no effect on the mechanical properties of the matrix and can address a larger healing area. Additionally, carbon nanotubes were added as nanofillers to enhance the reactivity between DCPD and the epoxydic matrix. The self-healing capability of the nano-fibre network was carried out by flexural tests, at epoxy resin level and composite level. Results obtained from Fourier transform infrared (FTIR) spectrometry, thermogravimetric analysis (TGA) and scanning electron microscopy (SEM) confirmed the successful encapsulation of DCPD healing agent in PAN fibres. Flexural tests indicate that after 48 h, the epoxy resin has recovered 84% of its flexural strength while the composite material recovered 93%.",
"conclusion": "5. Conclusions To pursue the self-healing properties of microvascular systems incorporated in polymeric composite materials, specific specimens were fabricated and subjected to three-point mechanical bending tests. Also, nanofiller elements (MWCNTs) were used to increase the molecular reactions between the free radicals of the repair agent (DCPD) and the epoxydic matrix. The use of carbon nanotubes also improves the electrical conductivity of the material, which once transformed into thermal conductivity, speeds up the repair process. This could represent a major benefit for aerospace structures, especially at aircraft control surfaces (slats, spoilers, aileron, flaps) and fuselages. As most of these structures are electronically manoeuvred, electric impulses can trigger the healing mechanism within the composite. The use of a microvascular self-healing system can be extremely beneficial in manufacturing thermosetting polymeric composite materials, as it has no negative effect on the nominal properties of the epoxydic matrix material and can be applied between each layer of the laminate, without causing delamination. This is due to the fact that the matrix encloses the microvascular system during the curing process, and it can also attach to the reinforcing fibres, as seen in the present paper.",
"introduction": "1. Introduction The polymeric composite materials (PCM) took a lot of attention from important industries (space, aerospace, automotive, military) as a great alternative for existing structural metallic materials, mainly by using thermoset epoxydic matrix reinforced with carbon fibres. The obtained carbon fibre reinforced polymers (CFRP) have a lightweight, high strength structure, considering its physical and mechanical properties (high strength-to-weight ratio, high modulus, design adaptability, flexibility to various geometries, great fatigue and corrosion resistance, good thermal expansion properties). As the CFRP provides many advantages their matrix has a brittle nature, and when different mechanical loads are applied (vibrations, stress, tension, etc.), it promotes cracks and delaminations that lead to losing their structural integrity. Generally, the failure modes occur at the nanoscale level and propagates to micro and macro levels, and this affects more or less in time the rate at which the material yields to the loads to which it is subjected to [ 1 ]. To this, different methods were used to aid the healing of the material, starting from the nano and micro level, thus stalling or stopping the crack from propagating further and affecting the integrity of the structure. Considering that the budget for repairing damaged composites is growing year by year [ 2 ], repair processes are time- and cost consuming. Having this, the addition of self-healing methods is appraised by many industrial companies and scientific community. Basically, by using their self-healing ability, polymeric materials can transform the physical energy (from the damage) into chemical energy required to heal the affected area [ 3 , 4 , 5 ]. In this regard, different healing agents and encapsulation methods have been studied and reviewed [ 6 , 7 , 8 , 9 ]. As the microcapsules are the most used and versatile self-healing method, they have an extremely tedious preparation process and also limit the potential healing applications. In comparison to microcapsule healing process, the use of microvascular self-healing systems permits multiple healing processes, allowing the healing agents to be spread and reach distant points, covering a larger area [ 10 ]. Also, as seen in the previous work performed by the authors [ 11 , 12 ], even though the microencapsulation of healing agents leads to an 80–85% recovery of mechanical properties, the integration of microcapsules within the matrix and the composite material acts as an induced damage and thus reduces the mechanical properties of the specimen, as compared to the reference ones. This is mainly due to the microcapsule dimensions and volume used, and the fact that after breakage, the microcapsule shells are left within the material, further acting as a stress concentrator for other mechanical loads to which the structure can be subjected to. To reduce these shortcomings, the authors have proposed the use of nano-fibres fabricated by coaxial electrospinning process, that act as a microvascular network within the matrix and the composite material, respectively. Initially demonstrated in 2010 by Braun et al. [ 13 ], the production and integration of nano-fibres self-healing materials within the composite material have shown that it has little to none influence on the mechanical properties of the matrix or the composite material. Having a tangled aspect, the nano-fibre mats can deliver the encapsulated agent into the affected area in a more rapid way. Depending on the fibres used as shell materials, these nano-fibres have the possibility to increase the mechanical properties of the matrix or the composite material itself [ 14 ]. Comparing the three known encapsulation methods, the development of nano-fibres through the electrospinning process is shown to be more cost-efficient solution to cover a larger, multiple healing area due to its random entanglement of nanofibers. Another major advantage of the electrospinning process relies in the controllability of the process to obtain a higher surface-area-to-volume ratio and the formation of a homogenous structure with no defects [ 15 ]. Different core-shell nano-fibres have been conducted to extend the lifetime and mechanical performance of different composite laminates, by incorporating healing agents into the core, inducing the healing ability to the composite structure. In 2012, Sinha-Ray et al. [ 16 ] have encapsulated dicyclopentadiene (DCPD) as a healing material in polyacrylonitrile (PAN) shell material using emulsion electrospinning. The authors evaluated the self-healing ability by introducing the core-shell nano-fibres within the composite layers and subjected the specimens to interlaminar fracture tests, demonstrating the release of the healing agent and facilitating its solidification within the crack. Wu et al. [ 17 ] have encapsulated DCPD in PAN nano-fibres by means of coaxial electrospinning, incorporating them together with Grubbs catalyst within CFRP layers. To assess the self-healing ability of the core-shell nano-fibres, the authors performed three-point bending tests and the results showed flexural stiffness values after healing of about 89% in average. In the work carried out by Neisiany et al. [ 18 ], the authors have studied the self-healing ability of a CFRP composite by incorporating an epoxydic resin in Styrene acrylonitrile (SAN) nanofibers structure by means of coaxial electrospinning method, and placing the core-shell nanofibers between CFRP layers. Having a high encapsulation yield (>90%), mechanical tests showed no reduction of mechanical properties for the composite embedding the core-shell nanofibers, and allowed the composite to have up to three repair cycles (97%, 94%, and 89% recovery), until exhaustion of healing agent. Other studies have been performed by Neisiany et al. [ 19 , 20 , 21 ] and successfully shown the releasing and solidifying of the healing agents using PAN and poly(methyl methacrylate (PMMA) as shell material. Mohammadi et al. [ 22 ] studied the self-healing capability of a glass fibre reinforced polymer composite (GFRP) having an epoxydic resin as healing agent within the microvascular channels and embedded them in the composite. Mechanical tests and microstructural analyses have shown that the healing ability was confirmed at different periods of time, namely after 4, 7, and 11 days, receiving a healing efficiency of maximum 69%, compared to the reference specimen. Polymer nanofibres have a great potential in many applications, as filtration material, biomedical, material reinforcement and electronics. To ease the manufacturing process of nanofibres, electrospinning processing technique has been widely used. Depending on the application, different nanofibres surface properties are desired, including wettability, adsorption and adhesion. One of the methods to improve the surface properties of the electrospun nanofibres is the cold gas plasma treatment, that modifies the surface of polymer nanofibres without affecting their bulk properties [ 23 , 24 ]. In the work carried out by Wei et al. [ 23 ], cold gas plasma treatment was used to analyse its effect on surface properties of polyamide 6 (PA6) nanofibers. It was observed that by changing the PA6 surface properties it has enhanced the surface roughness of the nanofibers and also, the contact angle between water and nanofibers has been reduced significantly. The authors concluded that the nanofibres surface modification has a great potential in biomaterials, sensors and medical devices applications. Another method of improving the surface properties of the electrospun nanofibres is the air plasma functionalization, which improves the hydrophilicity of the nanofibers. Considering bio-medical applications, Mozaffari et al. [ 25 , 26 ] have studied the effect of atmospheric air plasma method and Argon-Oxygen plasma surface modificaiton on improving the bio-functionality of the electrospun naofibres. As also reported by [ 23 ], the surface roughness has been significantly increased due to ionization and chemical degradation process. Also, the hydrophilicity, biocompatibility and bioactivity were improved. Although the electrospinning process of developing nano-fibres are mostly used in medical and biomedical applications, the nano-fibres developed through electrospinning process has taken large interest for other applications, as in the self-healing polymers. This is due to the large variety of materials that can be used to develop nano-fibre, as thermoplastic and thermosetting polymers, conjugated polymers, natural polymers, metals, metal oxides, nitrides and carbides, etc. There are many types of electrospinning processes depending on the final application, single fluid process, multiaxial, coaxial, triaxial, multi-needle, etc. [ 2 , 27 , 28 ]. A single-fluid blending electrospinning process was used together with casting film method to develop nanofibres for drug-controlled release application [ 29 ]. This combination between electrospinning and traditional casting method for producing functional hybrid films has been demonstrated successfully, providing a dual-phase drug-controlled release. This technique can be applied also in the self-healing application for polymer composites, where a controlled release of healing agent can benefit the healing efficiency. A side-by-side electrospinning technique has been used by [ 30 ] to develop a nanofiber wound dressing to promote the wound healing process, comprising three layers. The technique can be used in self-healing polymers applications, where both the healing agent and the catalyst can be embedded in the same nanofibre, separated by a middle layer. Same as side-by-side process, a tri-fluid electrospinning techniques can also be applied to deliver the healing agent into the polymer crack, to prevent propagation. Such technique was performed to develop tri-section nanofibres, to deliver a poorly water-soluble medicine [ 31 ]. By introducing nanofiller reinforcements such as carbon nanotubes (CNTs) in the polymeric matrix, not only it enhances the mechanical properties of the material, but it also offers structural, electronic and thermal properties which, in the end, may be beneficial for the healing ability of the self-healing system. More precisely, depending on the industry and application, the electronic advantage given by the CNTs may activate the healing agents when trapped within the crack, generated by small electric impulses. As aircrafts have many electronic controls that act on different composite components, the use of these impulses together with the CNT nanofillers may increase the healing ability of the corresponding components. Different self-healing polymer nano-composites containing MWCNT nano-fillers have been reported, providing great healing proficiency and recovery of physical properties [ 32 , 33 ]. Within this paper, the authors have synthesised by means of coaxial electrospinning, a microvascular network (MVN) of PAN nano-fibres filled with DCPD healing agent, in order to evaluate the healing ability on an aerospace matrix and composite material, by means of flexural tests. In addition, multiwalled carbon nanotubes (MWCNTs) were added to increase the healing capacity of the MVN.",
"discussion": "4. Discussion Given that most studies aim to investigate the use of microcapsules and hollow tubes as encapsulation methods for healing agents in polymer composites, this paper addresses the development and evaluation of a microvascular nano-fibre network, fabricated by coaxial electrospinning. Thus, a PAN/DCPD nano-fibre network has been obtained and used for the evaluation of its healing ability for aerospace graded materials, an epoxydic matrix and a pre-impregnated carbon fibre, respectively. The morphological studies by SEM revealed that the obtained nano-fibre network present a high density of beads on fibres. As this is not ideal for the electrospinning processing, the fabrication parameters were modified in accordance with suggestions provided in [ 17 ], which led to a smooth, uniform and bead-free nano-fibres, which are comparable to the ones described in other studies [ 1 , 10 , 16 , 21 , 39 , 40 , 41 ]. FT-IR analysis confirm the presence of PAN nano-fibres specific peaks and the presence of DCPD healing agent, as also identified in [ 42 ] and [ 11 ], respectively. The healing ability was appraised by means of three-point bending tests for specimens containing the MVN nano-fibres and MVN nano-fibres with MWCNTs nanofillers to improve the healing mechanism. It was observed that the addition of 0.5% MWCNT nanofillers has improved the healing ability of the epoxydic matrix, as also validated by SEM, illustrating the healing crack interface and bridging of the MVN nano-fibres. CFRP composites specimens were also investigated using the same nano-fibres and nanofillers, concluding the same healing process ability. Also, the addition of the MVN nano-fibres increased the flexural strength of the composite specimen, concluding that besides the healing ability, the nano-fibres network acts also as a reinforcement. Future work envisages the use of GFPR composite to have a better overview of the crack area and healing process, due to the less carbon content. Moreover, the analysis of an electric impulse trigger to initiate the healing process is also foreseen, due to the electro-mechanical properties of the MWCNT nanofillers."
} | 3,943 |
40075086 | PMC11903764 | pmc | 3,640 | {
"abstract": "When attempting to replicate the same biological spiking neuron model actions of the human brain, the spiking neuron model methodology and hardware realization design for the nervous system of the brain are crucial considerations. This work provides a modified neural modeling of complete Digital Spiking Silicon neuron model (DSSN4D). This model is capable for regenerating the basic attributes of the original model using a simplified power-2 based modeling technique. The suggested spiking neuron model is based on the fundamental power-2 based operations that can be implemented similar to the basic attributes of the main model. Removing the nonlinear parts of the main model (original one) and replacing them with modified ones leads to achieving a low-cost, low-error, and high-frequency digital system rather than the original modeling. A Xilinx Virtex-7 XC7VX690T FPGA board has been thought of and utilized for hardware realization and of the proposed model (this can validate the proposed system). The original and proposed models (in terms of neural activities) exhibit a significant degree of resemblance, according to hardware results. Additionally, greater frequency and low-cost conditions have been attained. Results of implementation indicate that overall savings are higher than for other papers and the original approach. Additionally, the new neural model’s frequency, which is roughly 502.184 MHz, is much greater than the original model’s frequency, which was 224 MHz. Also, results in hardware level shows that the proposed model takes a maximum 0.01% of the available resources of a Virtex-7 FPGA board.",
"conclusion": "Conclusion The appealing research field involves simulation and implementation of neural networks, which calls for an understanding of the central nervous system and its parts. Modeling neural activities must thus be extremely important when applied to the neuromorphic field. These spiking neuron model models may be implemented using a variety of methods, but the ideal method must address every component of an effective digital design (without any nonlinear and high-cost terms implementation such as: multipliers, dividers, exponential units, quadratic terms, etc.). As a result, a power2-based DSSN4D spiking neuron model model is provided in this study. The suggested design can efficiently and with excellent similarity recreate fundamental kinds of spiking signals. Virtex-7 FPGA board can be employed and taken into consideration for verifying and validating the proposed method design. Hardware outcomes demonstrate that this new model is capable of simulating the same behaviors as the previous neural modeling in this fashion. In the event of high-frequency and also low-cost realization conditions, the new hardware proposal can follow the original paradigm. Results of implementation reveal an improvement in FPGA total savings as well as a higher frequency for the suggested model, which is 502.184 MHz instead of the old model’s 224 MHz. The Implemented spiking neuron models (IN) for our proposed model in the better state compared by the original and other similar models.",
"introduction": "Introduction Spiking neural networks (SNNs) are a fascinating area of study that combines Artificial Intelligence (AI) with Neuromorphic Engineering (NE). This discipline has real-world applications in areas like memory, learning, medicine, data processing, etc 1 – 6 . The human brain is divided into several sections, and each section is made up of fundamental components: spiking neuron model cells connected by synapses. In the study of neural systems, it is crucial to recognize the diversity of neuron types that exist. Each type of neuron plays a unique role in neural circuitry, contributing to different functions such as sensory processing, motor control, and cognitive functions. For instance, excitatory neurons, such as pyramidal cells, facilitate signal transmission, while inhibitory neurons, such as interneurons, regulate and balance excitatory activity, ensuring overall network stability. Understanding the interactions between these various neuron types can provide insights into the complex dynamics of neural systems. Ordinary Differential Equations (ODEs) are typically used in mathematical applications to represent the functioning of spiking neuron models. To mimic genuine spiking neuron models, a variety of models at various biological levels have been published 7 – 13 . They are computationally expensive since they are based on biologically specific spiking neuron model models. As a result, the amount of biological complexity and computing expense must be balanced when selecting a suitable model for the simulation and implementation of spiking neural networks. In our paper, the Complete Digital Spiking Silicon spiking neuron model (DSSN4D) model is applied for hardware implementation process 7 . The goal of this model was to imitate several kinds of spiking neuron models using straightforward digital arithmetic circuits. In case of hardware implementation, there are two basic approaches that can be considered: Analog and Digital. The analog case has several advantages over digital ones, such as being easier to communicate with real-world signals and using less power and area. On the other hand, this technique has several drawbacks such being difficult to use, susceptible to process variability, and noise. Every change in its parameters results in a modification in the design. Different papers in this field have been presented in case of analog realization of models 14 – 18 . On the other hand, high power consumption, a large silicon need, and discretization over time are drawbacks of digital system architecture. Moreover, they have simple, multiple spiking neuron model simulations, flexibility in design, and are noise-resistant (and process variability). Recently, a reconfigurable system using an FPGA presented a high resource case to observe the activity of the spiking neuron models in neural networks as well as dynamical behaviors 19 – 28 . In this work, a hardware implementation of complete version of DSSN modeling (DSSN4D) for digital systems is proposed. The main challenge in this implementation is the quadratic term of the original spiking neuron model model. In most cases, the quadratic term slows down the final system because of its multiplier action. In other words, the Central Nervous System (CNS) depends greatly on the rate of neural activity. The CNS is comprised of the brain and spinal cord and is responsible for integrating sensory information, processing cognitive functions, and directing motor responses. It plays a critical role in regulating vital bodily functions such as breathing, heart rate, and coordination of movement. CNS is also central to higher cognitive functions, including memory, language, and decision-making. Neurons within the CNS communicate through complex signaling pathways, utilizing neurotransmitters and electrical signaling to relay information. The study of the CNS is essential in understanding various neurological disorders and developing therapeutic strategies. The neurological system is affected if the final system’s frequency is unacceptable. This nonlinear phrase must thus be eliminated or changed to another simple function. To create basic mathematical equations, many methods might be applied. The best method among them would be to transform the quadratic terms into power-2 based functions. Indeed, while conserving all behaviors of the original model, By converting the nonlinear components of the original model into a set of power-2 terms, we have a new model that converts all multiplier operations into a sequence of digital SHIFTs and ADDs (this method will be explaied). The suggested new model provides a low-cost, quick, and efficient system that can execute and trace the original spiking neuron model model with a high degree of resemblance. However, compared to original model and also earlier efforts 29 , 30 , the provided DSSN4D model is substantially more accurate and does not have the aforementioned flaws. The investigation shows that the original forms and dynamics can be successfully recreated using our approach. Because all nonlinear components are eliminated when employing base-2 functions, there is a high matching similarity and multiplierless implementation. This results in the creation of a digital system that is quick, inexpensive, and capable of large-scale digital implementation. Biological neural systems comprise a wide variety of neuron types, each exhibiting distinct electrophysiological behaviors that support cognitive functions, motor control, and sensory processing. These neurons, classified based on their spiking patterns and synaptic connectivity, include regular spiking (RS), fast spiking (FS), intrinsically bursting (IB), low-threshold spiking (LTS), and chattering (CH) neurons, among others. Simulating and implementing such diversity is a key challenge in neuromorphic engineering and computational neuroscience. While our study primarily focuses on the Digital Spiking Silicon Neuron (DSSN4D) model, this approach is not restricted to a single neuron type. The DSSN4D model is a generalized digital neuron framework that can replicate various neuronal behaviors through parameter tuning and structural adjustments. The core dynamical properties-membrane potential evolution, recovery mechanisms, and spike generation-allow DSSN4D to serve as a flexible template for multiple neuron types. By modifying key model parameters, different firing patterns can be generated, demonstrating its adaptability beyond a single neuron. Additionally, the power-2 based approximation technique introduced in this study is a universal optimization method that can be applied to various neuron models. This method reduces computational complexity while maintaining fidelity to biological neural behaviors, making it suitable for large-scale neuromorphic systems. In this work, we validate the feasibility of DSSN4D by demonstrating its efficiency in digital implementation; however, its potential extension to other neuron types is a natural progression of this research. This study lays the groundwork for scalable hardware-friendly neuron implementations, paving the way for the future realization of heterogeneous neural populations on digital platforms. The rest of the paper is organized as follows. In Section II, complete DSSN4D spiking neuron model model is briefly presented. In Section III, the proposed approach is explained. Proposed model simulation and validation are described in Section IV. In Section V, a population test is evaluated. The proposed model design and implementation process are presented in Section VI. Finally, FPGA results and conclusion are given in Sections VII and VIII, respectively.",
"discussion": "Discussion In biological neural networks, different neuron types coexist and interact to perform complex cognitive and sensory functions. Although this study primarily implements DSSN4D with a specific parameter set, our proposed digital framework allows for the simulation of multiple neuronal classes. By varying intrinsic model parameters such as time constants, adaptation dynamics, and trigger currents, DSSN4D can replicate a broad spectrum of spiking behaviors observed. The power-2 based approximation method introduced in this study further enhances the scalability and adaptability of DSSN4D. Since this transformation removes high-cost nonlinear terms while preserving the fundamental neuronal properties, it can be integrated into other neuron models such as: Izhikevich neuron model (efficient for large-scale networks), Morris-Lecar model (useful for modeling oscillatory behaviors), Adaptive Exponential Integrate-and-Fire (AdEx) model (bridging simple and biophysically detailed approaches). Future work will explore parameter tuning strategies to systematically classify DSSN4D variations corresponding to different biological neuron types. Additionally, we aim to extend the hardware implementation to heterogeneous neural populations, demonstrating real-time neural dynamics in a large-scale digital neuromorphic system. This flexibility bridges the gap between biologically plausible neuron modeling and efficient digital hardware implementation, making DSSN4D a versatile tool for neuromorphic computing applications. While the primary focus of this study is on optimizing the DSSN4D model for digital implementation, it is important to highlight that the model itself is not static. By modifying key parameters, DSSN4D can reproduce multiple neuronal firing behaviors, making it a flexible model for neuromorphic applications. To validate this adaptability, we conducted additional simulations using three parameter variations of DSSN4D, demonstrating how different neuronal behaviors can be obtained. The parameter variations can be adopted based on different values of parameters presented in Table 1 , but in this approach, we have tested some selected parameters that variations of their values can generate different spiking behaviors. As can be illustrated in this Table, DSSN4D model contains large number of fixed parameters that can be varied to have different spiking patterns. In this approach, we have selected some parameters to be varied to generate different behaviors of spiking. In this case, at first, we have tested the variation of parameter \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} . As can be depicted in Fig. 14 , by variation of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} , different spiking patterns such as bursting can be generated. Moreover, as can be seen in Fig. 15 , by variation of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} , different spiking patterns in another forms can be generated. Finally, as can be seen in Fig. 16 , by variation of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_0$$\\end{document} , different spiking patterns in other formats are generated. Fig. 14 The voltage signal generation in case of different trigger between \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{stim}=0.5$$\\end{document} to \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{stim}=1.6$$\\end{document} by rate of 0.1. Fig. 15 The voltage signal generation in case of different \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} between \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau =0.0003$$\\end{document} to \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau =0.0008$$\\end{document} by rate of 0.0001. Fig. 16 The voltage signal generation in case of different \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_0$$\\end{document} between \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_0=-10$$\\end{document} to \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_0=-6$$\\end{document} by rate of 0.5. Although this study presents a hardware implementation of the DSSN4D model, it is essential to highlight that the contributions extend beyond a mere technological demonstration. The core advancements introduced in this research include a novel power-2 based approximation for digital neurons, as traditional digital neuron models suffer from computational inefficiency due to the presence of high-cost nonlinear functions (e.g., quadratic terms and multiplications). This study introduces a power-2 based transformation, replacing multiplications with simple digital shift and add operations, thereby reducing computational complexity while maintaining biological fidelity. Unlike previous approaches that rely on approximations with substantial accuracy loss, our method preserves the essential dynamical properties of the DSSN4D model with minimal error. In terms of experimental validation of model accuracy and computational efficiency, we conducted rigorous computational error analysis using RMSE, MAE, and correlation metrics to ensure the proposed model closely matches the original DSSN4D. The model was further tested in a large-scale neural population, demonstrating its ability to simulate realistic neural network dynamics. The results confirm that the proposed approximation maintains biological accuracy while significantly improving computational efficiency, making it viable for large-scale neuromorphic applications. Additionally, regarding scalability and adaptability for large-scale neuromorphic systems, the DSSN4D model, implemented with our optimizations, achieves a high operational frequency (502.184 MHz), which is more than double the original model’s speed (224 MHz). By reducing FPGA resource utilization (eliminating DSP blocks and minimizing LUTs), we enable the deployment of a significantly larger number of neurons per FPGA core compared to previous implementations. This scalability is crucial for real-time neuromorphic computing, enabling applications in brain-inspired AI systems, cognitive computing, and biomedical simulations. Finally, our study bridges the gap between computational neuroscience and digital hardware implementation, as computational neuroscience often relies on biophysically detailed models that are computationally expensive. Our study provides a bridge between biologically realistic neural modeling and efficient digital hardware implementation, offering an approach that is both computationally efficient and biologically meaningful. The proposed model can be integrated into neuromorphic processors and real-time AI systems, making it highly relevant to next-generation computing architectures. These contributions establish this work as a significant advancement in neuromorphic engineering and computational neuroscience rather than merely a technical demonstration. The reliability of the proposed DSSN4D model was thoroughly evaluated by comparing its responses to those of the original model under diverse stimulus conditions. This analysis focused on the relationships between spike frequency ( F ) and time to first spike (TtFS) with variations in external stimulus current ( \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} ) and the time constant parameter ( \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} ). The results indicated that the DSSN4D model closely mirrored the behavior of the original model, displaying similar trends and confirming its accuracy in replicating neural dynamics. Minor differences in responses at higher stimulus levels were observed but remained within acceptable limits, indicating that the proposed model effectively represents the original model’s reactions to external stimulation. Further assessments revealed that both models exhibited consistent behavior in response to changes in \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} , with an increase leading to decreased spike frequency and increased TtFS. The close alignment between the two models reinforces the validity of the proposed modifications, showing that the DSSN4D model captures the essential characteristics of the original while introducing enhancements in numerical stability. A comparative visualization of membrane potential dynamics demonstrated that the proposed model maintained a high degree of alignment with the original model’s waveform patterns, with only slight variations in spike amplitude and timing. Overall, the analysis confirms that the DSSN4D model is a reliable alternative, preserving key dynamical properties of neuronal activity while improving responsiveness to parameter changes. As can be seen in Fig. 17 , our proposed model can follow the original DSSN4D modeling in case of patterns. Also, Table 9 shows the similarity between original and proposed models in case of frequency and TtFS parameters. Moreover, Table 10 shows this similarity for variations of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} . Table 9 Spike Frequency and time to first spike (TtFS) for original and proposed DSSN4D models by variations of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} . \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$F_{original}$$\\end{document} \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$TtFS_{original}$$\\end{document} \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$F_{proposed}$$\\end{document} \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$TtFS_{proposed}$$\\end{document} 0.5 1.50 0.0076 1.50 0.0081 0.6 2.50 0.0060 2.60 0.0065 0.7 6.50 0.0050 6.50 0.0055 0.8 10.00 0.0043 10.20 0.0048 0.9 13.50 0.0038 13.50 0.0043 1.0 17.00 0.0034 17.00 0.0039 1.1 21.00 0.0031 21.00 0.0036 1.2 25.00 0.0028 25.00 0.0033 1.3 29.50 0.0026 30 0.0031 1.4 34.00 0.0024 34.00 0.0029 1.5 38.50 0.0022 39 0.0027 1.6 43.00 0.0021 42.00 0.0026 \n Fig. 17 Comparison of membrane potential over time for the original and proposed DSSN4D models under different stimulus currents \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$I_{Stim}$$\\end{document} . \n \n Table 10 Spike frequency and time to first spike (TtFS) for original and proposed models by variations of \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} . \n \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\tau$$\\end{document} \n \n \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$F_{original}$$\\end{document} \n \n \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$TtFS_{original}$$\\end{document} \n \n \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$F_{proposed}$$\\end{document} \n \n \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$TtFS_{proposed}$$\\end{document} \n 0.0003 29.00 0.0017 29.00 0.0022 0.0004 24.00 0.0023 24.00 0.0028 0.0005 19.50 0.0029 20 0.0034 0.0006 17.00 0.0035 17.00 0.0040 0.0007 20 0.0041 19.50 0.0046 0.0008 26.50 0.0047 26.00 0.0052"
} | 7,056 |
31294170 | PMC6604995 | pmc | 3,642 | {
"abstract": "Common bean is one of the most important crops for human feed, and the most important legume for direct consumption by millions of people, especially in developing countries. It is a promiscuous host legume in terms of nodulation, able to associate with a broad and diverse range of rhizobia, although the competitiveness for nodulation and the nitrogen fixation capacity of most of these strains is generally low. As a result, common bean is very inefficient for symbiotic nitrogen fixation, and nitrogen has to be supplied with chemical fertilizers. In the last years, symbiotic nitrogen fixation has received increasing attention as a sustainable alternative to nitrogen fertilizers, and also as a more economic and available one in poor countries. Therefore, optimization of nitrogen fixation of bean-rhizobia symbioses and selection of efficient rhizobial strains should be a priority, which begins with the study of the natural diversity of the symbioses and the rhizobial populations associated. Natural rhizobia biodiversity that nodulates common bean may be a source of adaptive alleles acting through phenotypic plasticity. Crosses between accessions differing for nitrogen fixation may combine alleles that never meet in nature. Another way to discover adaptive genes is to use association genetics to identify loci that common bean plants use for enhanced biological nitrogen fixation and, in consequence, for marker assisted selection for genetic improvement of symbiotic nitrogen fixation. In this review, rhizobial biodiversity resources will be discussed, together with what is known about the loci that underlie such genetic variation, and the potential candidate genes that may influence the symbiosis' fitness benefits, thus achieving an optimal nitrogen fixation capacity in order to help reduce reliance on nitrogen fertilizers in common bean.",
"conclusion": "5. Concluding Remarks The natural diversity of rhizobia nodulating common bean has been widely studied, but, because of the promiscuity of this crop, novel symbionts of this legume should be expected and need to be screened. Information about rhizobia diversity in common bean serves to define host preferences and predominance of strains, to study the dynamics of exchange of genetic material, and provides a basis for the proposal of evolutionary trends. The diversity studies also reveal that each of the common bean-rhizobium associations coevolved independently after geographical separation, as did their genetic pools. Investigation about the structure of the indigenous rhizobial populations and their coevolution with the host plant could greatly contribute to better understanding and overcoming the frequent reports of nodulation failure. Molecular markers associated with nodulation genes are available in common bean, and QTL mapping studies showed that genes with varying effects seem to control N fixation. To date, few major QTLs and candidate genes have been reported in this legume. However, nodulating genes in model legumes have been cloned and several of their orthologs determined in common bean. Clearly, the evaluation of natural rhizobia diversity associated with common bean, making use of its well-characterized common bean biodiversity and feature-rich genomic tools, is becoming a powerful strategy of investigation, as are breeding cultivars for high symbiotic efficiency.",
"introduction": "1. Introduction Common bean ( Phaseolus vulgaris L.) is cultivated worldwide, constituting a staple in developing countries in East Africa and South America, and also in some regions of Asia, where it is the main source of protein [1] . In some of them, common bean can account for up to 20% of the total daily protein intake per person [2] , [3] . In Europe, there has been a notable increase in common bean consumption in past years, due to a greater demand for healthy and functional food [4] , and the current trend of vegetarian diets in Central Europe and the United Kingdom, in which beans and other pulses are included as meat substitutes. It is accepted that domestication of common bean from wild forms took place several thousand years ago in two main and independent centers of diversification, Mesoamerica (Mexico and Central America) and the Andes (Peru, Bolivia, and Northwest Argentina), resulting in two highly differentiated gene pools [5] . From these centers, the crop was spread all over the American continent, and, after the first voyages of Columbus (1492), common bean was brought to Europe. Both common bean gene pools spread widely in all parts of Europe with very complex pathways of dissemination that included several introductions from various regions of the Americas, combined with direct exchanges between European and other Mediterranean countries [6] and adaptation to European soils and climate conditions [3] , [7] , [8] , [9] . During the five centuries since the introduction of common beans into Europe, many varieties evolved under diverse environments and farmer preferences, to provide dry seeds or fresh pods [10] , thus the loss of variation might have been less than previously suspected. What is more, hybridization that occurred in Europe between the Andean and Mesoamerican gene pools probably had a significant impact on the maintenance of the overall level of genotypic diversity [9] . Common bean can establish symbiotic interactions with both rhizobia (the Rhizobium -legume symbiosis, RL) and arbuscular mycorrhizal (AM) fungi, leading to the formation of nitrogen-fixing nodules and phosphate acquiring mycorrhiza [11] . Both symbiotic interactions play a vital role in ecosystems and sustainable crop production, and are central for efforts to decrease dependence on commercial fertilizers. The intensive application of N fertilizers over the last century has perturbed the N cycle, by leaching excess N fertilizers to watercourses and the emission of pollutant NO x gases to the atmosphere [12] , [13] . Common bean is generally known as a weak Nitrogen (N) fixer in comparison with other grain legumes [14] , [15] . Therefore, application of N fertilizers in bean fields is recommended to achieve higher yields. Selection in the 20 th century based on common bean varieties with the best performance in highly mechanized monoculture systems is often reported to be the cause of the current relatively low Symbiotic Nitrogen Fixation (SNF) [16] . However, it has been reported that the climbing and indeterminate common bean varieties consistently have higher nodulation and SNF abilities, compared with most bush-type cultivars. These greater abilities are attributed to the relatively longer period of fixation during the growth cycle in climbing type cultivars [17] – [20] . Miranda and Bliss [21] reported that selection for high levels of SNF, especially when performed in low-fertility soils, might result in genetic gains in common bean breeding populations. Bliss [14] also discussed that the level of N fixation can vary significantly among common bean genotypes, and argued that reports of insufficient levels of N fixation were often based on observations with only a few genotypes, and were conducted with unsuitable N fixation measurement assays. Thus, proper characterization and evaluation of common bean germplasm collections as sources of adaptive alleles, and their utilization in breeding for enhanced SNF, are often limited or neglected. In addition to genetic background, several abiotic factors can greatly influence the SNF ability of common bean. Deficiencies of phosphorus (P), potassium (K), and sulfur (S) have been reported as environmental SNF-limiting factors, which may influence number and weight of nodules [22] , [23] . Direct impacts of P, K and (or) S deficiencies on nodules might be due to their influence on physiological and metabolic processes in nodules [23] , [24] . Other important environmental factors affecting SNF are salinity and different soil water conditions [25] , [26] . As a consequence, selection of best adapted common bean cultivars and most effective fixing rhizobial strains in each association is a must in order to maximize SNF [19] . Therefore, the study of rhizobia's natural diversity is a source of ecological information about symbioses, as it allows for the definition of host preferences and strain predominance, but, most importantly, since it provides the source for efficient strains to be used as inoculants in agricultural fields [18] . Breeding programs for improved SNF in common bean have been developed, resulting in the release of high N fixing Mesoamerican cultivars, promoting the development of cropping systems that are less dependent on N chemical fertilizers [27] . However, sustained success in developing Andean cultivars with enhanced SNF has been elusive. The availability of superior genotypes with higher N-fixation ability supports the idea that SNF in common bean may be improved through breeding efforts. In this sense, the advances in genetics and genomics resources of common bean and the high degree of synteny between this crop and its legume crops relatives can be exploited so as to understand complex traits associated with SNF, leading to the discovery of new genes or Quantitative Trait Loci (QTLs), as well as to improve genetic maps and develop molecular markers for Marker Assisted Selection (MAS). In this perspective, rhizobial biodiversity resources and the genes or QTLs of adaptive importance for RL interaction will be discussed, in order to accelerate the development of common bean cultivars with enhanced SNF."
} | 2,393 |
37242000 | PMC10223422 | pmc | 3,643 | {
"abstract": "This paper proposes two different approaches to studying resistive switching of oxide thin films using scratching probe nanolithography of atomic force microscopy (AFM). These approaches allow us to assess the effects of memristor size and top-contact thickness on resistive switching. For that purpose, we investigated scratching probe nanolithography regimes using the Taguchi method, which is known as a reliable method for improving the reliability of the result. The AFM parameters, including normal load, scratch distance, probe speed, and probe direction, are optimized on the photoresist thin film by the Taguchi method. As a result, the pinholes with diameter ranged from 25.4 ± 2.2 nm to 85.1 ± 6.3 nm, and the groove array with a depth of 40.5 ± 3.7 nm and a roughness at the bottom of less than a few nanometers was formed. Then, based on the Si/TiN/ZnO/photoresist structures, we fabricated and investigated memristors with different spot sizes and TiN top contact thickness. As a result, the HRS/LRS ratio, U SET , and I LRS are well controlled for a memristor size from 27 nm to 83 nm and ranged from ~8 to ~128, from 1.4 ± 0.1 V to 1.8 ± 0.2 V, and from (1.7 ± 0.2) × 10 −10 A to (4.2 ± 0.6) × 10 −9 A, respectively. Furthermore, the HRS/LRS ratio and U SET are well controlled at a TiN top contact thickness from 8.3 ± 1.1 nm to 32.4 ± 4.2 nm and ranged from ~22 to ~188 and from 1.15 ± 0.05 V to 1.62 ± 0.06 V, respectively. The results can be used in the engineering and manufacturing of memristive structures for neuromorphic applications of brain-inspired artificial intelligence systems.",
"conclusion": "4. Conclusions In conclusion, we optimized the scratching probe nanolithography of the photoresist for the intended pinholes and grooves array. For intended pinholes, the depth and diameter are well controlled at different average loads varied from 2 µN to 8 µN and ranged from 12.3 ± 1.2 nm to 41.6 ± 2.5 nm in depth and 53.4 ± 4.7 nm to 86.2 ± 5.8 nm in diameter. Furthermore, depth and diameter are well controlled at load times that ranged from 0.1 to 1.5 s and ranged from 22.1 ± 1.4 nm to 34.6 ± 2.5 nm in depth and 74.1 ± 5.5 nm to 91.3 ± 7.2 nm in diameter. The Taguchi method is employed to optimize AFM parameters (including normal load, scratch distance, probe speed, and probe direction) for the grooves array. As a result, we formed windows in the photoresist to full film depth (40.7 ± 1.6 nm) and roughness at the bottom of less than 1 nm. The results of scratching probe nanolithography were used for two memristive structure approaches. The first approach (prototype 1) was used to study the size of the memristor on resistive switching. Prototypes with memristor sizes ranging from 27 nm to 83 nm showed stable bipolar resistive switching with HRS/LRS ratios ranging from ~8 to ~128. Additionally, U SET and I LRS are well controlled and ranged from 1.4 ± 0.1 V to 1.8 ± 0.2 V and (1.7 ± 0.2) × 10 −10 A to (4.2 ± 0.6) × 10 −9 A, respectively. The retention test showed that R HRS and R LRS persist for at least 10,000 s. The second approach (prototype 2) was used to study the top contact thickness on the resistive switching. Prototypes with H ranging from 8.3 ± 1.1 nm to 32.4 ± 4.2 nm showed a stable bipolar resistive switching at 3V with an HRS/LRS ratio varying from ~24 to ~186. In addition, U SET was well controlled and ranged from 1.15 ± 0.05 V to 1.62 ± 0.06 V. The retention test also showed that R HRS and R LRS persisted for at least 10,000 s. The results can be used in engineering and manufacturing memristive structures for neuromorphic applications of brain-inspired artificial intelligence systems.",
"introduction": "1. Introduction Today, the Von Neumann architecture is the basis for computing systems whose main principle is the physical separation of arithmetic logic and memory units [ 1 , 2 , 3 , 4 ]. Here, each stage of information processing requires several steps where data are transferred between the processor and memory [ 5 , 6 ]. The processing steps are carried out sequentially, imposing initial limitations on the speed of such computing devices [ 7 , 8 , 9 , 10 , 11 , 12 ]. This creates time and energy problems, as the information must be transferred repeatedly between different system parts. This Von Neumann bottleneck limits the future development of computing systems [ 13 , 14 , 15 , 16 , 17 ]. Some parallelism can be introduced, but it is not enough. A promising solution is a neuromorphic architecture miming brain neuronal circuits, in which neurons represent computational units and synapses are local storage devices connected by communication channels [ 18 , 19 , 20 , 21 , 22 ]. In such an architecture, data processing is distributed throughout the computing system rather than concentrated in a central processor [ 23 , 24 , 25 , 26 ]. Processor and memory are integrated into a single unit, and the processing steps are performed in parallel rather than sequentially [ 27 , 28 ]. Although semiconductor-based computing systems have certain advantages, such as fault tolerance, power consumption, and handling of large data sets, they are significantly inferior to the biological brain [ 29 , 30 , 31 , 32 ]. Thus, there is a need for technology that has the advantages of biological and semiconductor materials but does not have drawbacks [ 33 , 34 , 35 , 36 ]. The biological brain supports various intellectual functions such as memory, learning, and decision-making [ 37 , 38 , 39 , 40 ]. One of the main ways of technical implementation of the biological brain is to manufacture ICs based on memristor structures, which are memory elements in the form of transition metal oxide film cells (neurons) that change their electrical resistance (between low-resistance (LRS) and high-resistance (HRS) states) under the action of an external electric field, connected by cross-synapses of data [ 41 , 42 , 43 , 44 , 45 , 46 ]. In doing so, ReRAM has a small cell size of a few nanometers, high integration density, high performance, and low power consumption, allowing it to mimic massive parallelism and low-power computing previously seen in the human brain [ 47 , 48 , 49 , 50 , 51 , 52 ]. Thus, the ReRAM technology satisfies all the basic requirements of neuromorphic systems [ 53 , 54 , 55 ]. An analysis [ 56 , 57 , 58 , 59 , 60 , 61 , 62 ] has shown that structures based on binary metal oxides are promising, especially zinc oxide (ZnO) obtained by pulsed laser deposition (PLD). To produce neuromorphic systems based on ZnO films on an industrial scale, fabrication regimes, as well as various control parameters on the resistive switching of ReRAM elements, are needed [ 63 , 64 ]. Therefore, it is necessary to study the effect of the ReRAM size element (the oxide film simulates the computing part of the biological brain, the neuron) and the contact thickness (the metal/oxide transition simulates the memory part of the biological brain, the synapse) on the memristive effect. This raises the need to develop new approaches for such studies, allowing local rapid prototyping of individual ReRAM elements and in situ diagnostics of their electrical and morphological parameters. A promising technique to form nanoscale structures is scratching probe nanolithography using an atomic force microscope (AFM) [ 65 , 66 , 67 , 68 ]. This technique involves the modification of thin polymer films by forming profiled nanoscale windows using the tip of an AFM probe. One of the ways to improve the quality of the windows is the Taguchi method, which is a statistical method developed by Genichi Taguchi, and more recently also applied to engineering, biotechnology, marketing, and advertising [ 69 , 70 ]. The essence of the method is to evaluate the quality performance of products and to determine whether quality losses are within the tolerance limits, as they increase as the current values of a parameter deviate from the nominal value [ 71 ]. Based on the Taguchi methods, the difference between the ideal and real objects is calculated, and the aim is to reduce it to a minimum, thereby providing an improvement in quality. So, a top contact can be formed through the windows to study the effect of different geometrical parameters on the resistive switching effect. In this paper, we propose two different approaches to studying resistive switching in thin zinc oxide films. For this purpose, we investigated scratching probe nanolithography regimes using the Taguchi method, which is known as a reliable method for improving the reliability of the result. Based on the obtained results, prototypes of memristor structures were fabricated, and the effect of memristor size and top contact film thickness on the resistive switching was investigated.",
"discussion": "3. Results and Discussion Figure 2 a shows the 3D AFM image of the photoresist surface that contains one of the indented pinholes. All formed pinholes were well defined. To make sure that we are not deforming ZnO film, we experimentally determined a critical normal load (at which the deformation of the ZnO film is observed) to be 9 µN. Consequently, we worked at a normal load value of up to 8 µN. Depth and diameter were well controlled at normal load from 2 to 8 µN and ranged from 12.3 ± 1.2 nm to 41.6 ± 2.5 nm in depth and 53.4 ± 4.7 nm to 86.2 ± 5.8 nm in diameter ( Figure 2 b). As expected, the load time ranging from 0.1 to 1.5 s had a lesser effect than the normal load and ranged from 22.1 ± 1.4 nm to 34.6 ± 2.5 nm in depth and 74.1 ± 5.5 nm to 91.3 ± 7.2 nm in diameter ( Figure 2 c). The non-linearity of the obtained dependencies can be explained by the viscoelastic properties of the photoresist. When the normal load was increased to 7 µN (load time at 0.5 s), the CAFM signal was silent, but at a force of 7 µN, peaks were recorded, indicating contact of the probe with the underlying ZnO film ( Figure 2 d). At a normal load of 8 μN, the signal broadened, which refers to an increase in the probe/ZnO size spot ( D ). At different load times ranging from 0.1 to 1.5 s ( Figure 2 d inset), spots were well controlled and ranged from 25.4 ± 2.2 nm to 85.1 ± 6.3 nm (normal load of 8 μN) due to the viscoelastic properties of the photoresist when it is pierced through the thickness of the photoresist film. Figure 3 a shows a 3D AFM image of a photoresist surface that contains a groove. All formed grooves were well defined. Figure 3 b shows the relationship between the width and depth of the grooves and the normal load. The depth and width were well controlled at different normal loads varying from 1 µN to 8 µN and ranging from 7.8 ± 0.8 nm to 40.5 ± 3.7 nm in depth and 47.3 ± 4.1 nm to 90.1 ± 7.6 nm in width. Figure 3 c shows the relationship between the width and depth of the grooves and the probe speed. Depth and width were also well controlled at probe speeds in the range from 2 µm/s to 9 µm/s and ranged from 38.3 ± 2.4 nm to 24.1 ± 2.0 nm in depth and 91.7 ± 8.2 nm to 68.2 ± 4.7 nm in width. The decrease in width and depth with increasing probe speed can be due to the viscoelastic properties of the photoresist, as well as the fact that the probe–sample contact is weakened at high speeds, which leads to an elevation of the probe. The speed of the probe affects depth and width, albeit to a lesser extent than the normal load. Figure 3 d shows the 3D AFM image of the photoresist surface carrying a window made by forming grooves parallel to each other at a certain distance. As can be seen, there is a residue of photoresist on the bottom of the window (surface roughness Ra ). It is important to exclude or at least minimize its presence, because we plan to deposit this window with metal, and it will have a negative effect on the result. This will lead to a scattering of resistive switching parameters from device to device. To achieve this purpose, the Taguchi method is used to optimize AFM parameters, including normal load, scratch distance, probe speed, and probe direction ( Figure 4 a). Here, the general task was to find the AFM parameters at which the photoresist was pierced to its full depth with a minimal surface roughness Ra . Therefore, the desirable results for depth and Ra are the largest and smallest possible values, respectively. The Taguchi method allows us to greatly reduce the number of experimental tests [ 68 ]. For our case, the orthogonal array is used; thus, only nine tests are needed. It follows from Figure 3 b that a normal load of 8 µN is required to pierce the photoresist film to the full depth; however, in the Taguchi method, we set 2 µN as the minimum value, because other AFM parameters also contribute to the depth. The levels of parameters are summarized in Figure 4 b, which also contains experimental results for depth and Ra . The depth is in the range from 7.3 ± 0.8 nm to 40.3 ± 1.5 nm, and Ra is in the range from 2.2 ± 1.4 nm to 14.3 ± 6.2 nm. As can be seen, the Ra results have a large confidence interval, which can explain the strong randomness of the photoresist ejection. Based on the formulas presented in Figure 4 c,d, the values of signal-to-noise (S/N) ratios for depth and Ra were determined ( Figure 4 e,f). According to the Taguchi method, the required values of the AFM parameters correspond to the maximum S/N ratio. Therefore, we must find the values with the maximum S/N ratio within each parameter A-D, while the values of other parameters with lesser S/N ratios are not considered. Therefore, the maximum depth is achieved with the normal load of 8 μN (A1), the scratch distance of 50 nm (B1), the probe speed of 9 µm/s (C3), and the direction of the probe being perpendicular (D1(D2)). Similarly, the minimum surface roughness Ra is achieved with the normal load of 2 μN (A3), scratch distance of 100 nm (B2), probe speed of 6 µm/s (C2), and probe direction being perpendicular (D1(D2)). Referring to our general task of finding the AFM parameters at which the photoresist was pierced to its full depth and with minimal Ra , we fixed the AFM parameters at which the minimum Ra is observed, except the normal load, and increased it. The general task was achieved at the normal load of 5 μN (A2), scratch distance of 100 nm (B2), probe speed of 6 µm/s (C2), and probe direction being perpendicular (D1(D2)). As a result, we formed windows in the photoresist to the full film depth (40.7 ± 1.6 nm) and a roughness at the bottom of less than a few nanometers. Tables S1 and Table S2 give the results of the analysis of variance (ANOVA) of the depth and Ra , respectively. Figure 5 shows the effect of memristor size on resistive switching in ZnO films. A feature of this approach (prototype 1 in Materials and Methods) is that we can localize resistive switching quite precisely, unlike other approaches, where there is a breakdown problem across the entire top contact due to its massive dimensions [ 41 , 42 , 50 , 51 , 59 ]. Therefore, we can land an AFM probe in any area of the top contact ( Figure 5 a) and ensure that resistive switching occurs at a single location. Prototypes with spot sizes ranging from 27 to 83 nm showed stable bipolar resistive switching at 3V sweep voltage ( Figure 5 b). The cumulative probability shows quite robust HRS and LRS resistances for all prototypes ( Figure 5 c). The HRS/LRS ratio varied from ~8 to ~128. It is curious that the R HRS is in a rather narrow range from (4.2 ± 0.3) × 10 10 Ω to (7.3 ± 0.5) × 10 10 Ω, whereas the R LRS is in a wide range from (0.41 ± 0.05) × 10 8 Ω to (7.2 ± 0.4) × 10 9 Ω. This can be explained by the fact that the memristor size without oxygen vacancies (HRS) does not strongly affect the resistance, based on previous studies of the resistive switching mechanism in zinc oxide films [ 57 , 58 , 59 ], according to which nanoscale conduction channels of oxygen vacancies are located along the grain boundary (filamentary mechanism); however, after resistive switching (LRS), the memristor size is proportional to the number of oxygen vacancies, and, accordingly, in this case, the resistance deviation increases. The SET voltage (U SET ) and the LRS current (I LRS ) are well controlled for the memristor size from 27 nm to 83 nm and ranged from 1.4 ± 0.1 V to 1.8 ± 0.2 V and (1.7 ± 0.2) × 10 −10 A to (4.2 ± 0.6) × 10 −9 A, respectively ( Figure 5 d). LRS decrease with the increase in memristor size can be linked to the filament increase in width (like a dendrite) with an increase in the memristor size, and hence the LRS decreases [ 57 , 59 ]. It is also worth considering that with an increase in the memristor size, the probability of the occurrence of several filaments increases [ 55 ]. The retention test showed that R HRS and R LRS persist for at least 10,000 s for all prototypes. ( Figure 5 e). Figure 6 shows the effect of top contact thickness ( H ) on the resistive switching in ZnO films. In this approach (prototype 2 in Materials and Methods), we cannot claim exactly where the resistive switching takes place, but we can quite precisely determine the thickness of the top contact using the AFM method ( Figure 6 a,b). The ripped edges of the TiN structures are the result of the lift-off process and do not affect the study of the impact of H on resistive switching. Prototypes with H ranging from 8.3 ± 1.1 nm to 32.4 ± 4.2 nm showed a stable bipolar resistive switching at 3V sweep voltage ( Figure 6 c). The HRS/LRS ratio varied from ~24 to ~186. The cumulative probability was quite robust, with the best result in the prototype with a TiN film thickness of H = 32.4 ± 4.2 nm. R LRS , R HRS , and R HRS /R LRS , in this case, are (3.1 ± 0.3) × 10 8 Ω, (1.1 ± 0.4) × 10 11 Ω, and 192, respectively ( Figure 6 d). The U SET and HRS/LRS ratios are well controlled at TiN top contact thickness from 8.3 ± 1.1 nm to 32.4 ± 4.2 nm and ranged from 1.15 ± 0.05 V to 1.62 ± 0.06 V and ~22 to ~188, respectively ( Figure 6 d). This can be explained by the fact that TiN is a reservoir of oxygen vacancies [ 72 , 73 ]. Therefore, an increase in the TiN top-contact thickness leads to an increase in the number of oxygen vacancies involved in resistive switching. This leads to an increase in the conductive paths and, consequently, a decrease in the I LRS ( Figure 6 c). However, when the oxygen vacancies are concentrated near the ZnO/TiN interface as well as in the TiN volume (HRS), the resistance of the ZnO film should not differ significantly at different H values, as confirmed by the experimental results. It is also worth noting that a higher number of vacancies requires a higher switching voltage. The retention test showed that R HRS and R LRS persist for at least 10,000 s for all prototypes. For H = 32.4 ± 4.2 nm, the HRS/LRS ratio varied from 192 to 181 ( Figure 6 e). Thus, increasing the size of the memristor ( D ) and the top contact thickness ( H ) leads to an increase in the HRS/LRS ratio, an increase in the U SET , and an increase in the I LRS . Unfortunately, a ‘universal’ memory does not exist, so we should proceed from the issue: when designing a neuromorphic system, in one case, multilevel characteristics may be a priority, and in another case, power consumption may be a priority. Hence, they should be based on the task: if the HRS/LRS ratio is a priority (for multilevel), D and H should be increased; if power consumption is a priority, they should be decreased. The results obtained from this study of resistive switching parameters correlate well with the results presented in [ 74 ], which present ZnO-based RRAM devices fabricated by pulsed laser deposition methods."
} | 4,901 |
37818249 | PMC10560873 | pmc | 3,644 | {
"abstract": "Abstract The soil fungal community plays an important role in forest ecosystems and is crucially influenced by forest secondary succession. However, the driving factors of fungal community and function during temperate forest succession and their potential impact on succession processes remain poorly understood. In this study, we investigated the dynamics of the soil fungal community in three temperate forest secondary successional stages (shrublands, coniferous forests, and deciduous broad‐leaved forests) using high‐throughput DNA sequencing coupled with functional prediction via the FUNGuild database. We found that fungal community richness, α‐diversity, and evenness decreased significantly during the succession process. Soil available phosphorus and nitrate nitrogen decreased significantly after initial succession occurred, and redundancy analysis showed that both were significant predictors of soil fungal community structure. Among functional groups, fungal saprotrophs and pathotrophs represented by plant pathogens were significantly enriched in the early‐successional stage, while fungal symbiotrophs represented by ectomycorrhiza were significantly increased in the late‐successional stage. The abundance of both saprotroph and pathotroph fungal guilds was positively correlated with soil nitrate nitrogen and available phosphorus content. Ectomycorrhizal fungi were negatively correlated with nitrate nitrogen and available phosphorus content and positively correlated with ammonium nitrogen content. These results indicate that the dynamics of fungal community and function reflected the changes in nitrogen and phosphorus availability caused by the secondary succession in temperate forests. The fungal plant pathogen accumulated in the early‐successional stage and ectomycorrhizal fungi accumulated in the late‐successional stage may have a potential role in promoting forest succession. These findings contribute to a better understanding of the response of soil fungal communities to secondary forest succession and highlight the importance of fungal communities during the successional process.",
"conclusion": "5 CONCLUSIONS Our results suggest the presence of shifts in the structure and function of soil fungal communities during temperate forest secondary succession. Soil NO 3 \n − ‐N and AP contents are important environmental filters in this process. Changes in fungal function in temperate forest secondary succession may potentially have a driving role in the succession process. The accumulation of fungal plant pathogens in early‐successional forests may be detrimental to seedling regeneration, thereby contributing to forest succession. At the same time, the significant accumulation of ectomycorrhizal fungi in late‐successional forests may help to maintain the dominance of established late‐succession species, suggesting that the soil microbial community shifts from promoting species turnover to promoting stability as succession progresses. These findings provide useful information to further our understanding of the response of soil fungal communities to secondary forest succession and highlight the importance of fungal community function during forest succession. It should however be taken into account that the tree species selected in this study were all symbiotic with ectomycorrhizal or arbuscular mycorrhiza and did not include species that are widely distributed in temperate zones and symbiotic with ericoid mycorrhizal; thus, our results may be limited by the mycorrhizal association of trees. Future studies should further explore the effects of differences in fungal communities on plant–soil feedback and competitive ability of different successional stage tree species through control experiments and consider including a more array of comprehensive mycorrhizal types to improve representativeness.",
"introduction": "1 INTRODUCTION Forest ecosystems play an important role in global biochemical cycles, but natural disturbances and human activities have resulted in the expansion of secondary forest succession, which has become more common on a global scale (Chazdon et al., 2016 ; Millar & Stephenson, 2015 ; Pugh et al., 2019 ). It is therefore becoming increasingly important to investigate secondary forest succession in temperate forest ecosystems. Forest succession can significantly affect the composition and function of soil fungal communities (Cline & Zak, 2015 ; Zhou et al., 2017 ). Since they constitute important participants in ecosystem soil processes, changes in the function of fungi may exert feedback on aboveground vegetation and thereby affect succession processes (Geisen et al., 2022 ; Knoblochová et al., 2017 ; Liao et al., 2018 ). However, much remains unknown about the changing patterns of fungal communities during temperate forest succession and their potential impacts. Forest succession can indirectly affect soil properties by changing the input of plant litter and root exudates, and by significantly affecting soil microbial community structure and function. This effect has been widely demonstrated in studies on bacterial communities (Dai et al., 2021 ; Zhang, Wang, et al., 2022 ). Soil pH, soil organic carbon (SOC), soil total nitrogen (TN), soil available nitrogen (AN), and soil available phosphorus (AP) exert significant influences on bacterial communities during secondary forest succession (Dai et al., 2021 ; Qu et al., 2020 ; Zhang et al., 2021 ). Qu et al. ( 2020 ) found that the increase in soil pH during forest secondary succession changed the structure of soil bacterial community and caused the dominant functional groups involved in the carbon cycle to be replaced by the groups involved in the nitrogen and sulfur cycles. The increase in soil carbon and nitrogen contents during secondary succession resulted in a significant decrease in the abundance of nitrification, and aerobic ammonia oxidation bacteria adapted to the oligotrophic environment had a considerable impact on nutrient cycling in forest soil (Zhang et al., 2023 ). Compared with bacteria, soil fungi are more sensitive to forest secondary succession, which is often accompanied by changes in the soil microbial community from bacteria‐ to fungi‐dominated (He et al., 2022 ; Li et al., 2022 ; Susyan et al., 2011 ). In forest ecosystems, the effect of the soil fungal community on ecosystem function is stronger than that of bacteria (Jiang et al., 2021 ). However, knowledge of the effects of soil properties on the structure and function of soil fungal communities’ change during the succession of secondary forests is limited. Therefore, further research is needed to deepen our understanding of the driving factors of the soil microbial community in secondary succession of temperate forests. Soil fungal pathogens and mutualists are key driving factors of plant community succession (Nara & Hogetsu, 2004 ; Richard et al., 2009 ; Van Der Putten & Peters, 1997 ). The accumulation of soil fungal pathogens during the early succession of dune vegetation significantly inhibited the growth of conspecific seedlings, promoting the replacement of early‐ by late‐successional species (Van Der Putten et al., 1993 ). Soil fungal pathogens can significantly increase seedling mortality near their parent trees and could be an important contributor to tropical forest population regulation (Bagchi et al., 2010 ; Liang et al., 2016 ). Symbiotic fungi can also influence the succession process through positive feedback with plants (Nara, 2006 ; Roy‐Bolduc et al., 2016 ). The accumulation of ectomycorrhizal and arbuscular mycorrhizal fungi is an important mechanism for host species to establish monodominant communities (Bennett et al., 2017 ; Laliberté et al., 2015 ; Liu et al., 2021 ). Although soil fungi thus play an important role in plant community dynamics, much of the potential impact of functional changes of fungal communities on temperate forest succession remains yet to be elucidated. This study was conducted on Kunyu Mountain, a temperate forest nature reserve located in northern China. We selected three typical forest types representing the succession process, namely shrublands, coniferous forests, and deciduous broad‐leaved forests. We used high‐throughput sequencing techniques combined with fungal functional taxon prediction (FUNGuild version 1.0) for the analysis. We aimed to answer the following questions: (1) How do the soil fungal communities and function structure respond to temperate forest succession? (2) What are the potential impacts of changes in soil fungal community function on forest succession processes?",
"discussion": "4 DISCUSSION This study illustrates the response pattern of fungal communities during the secondary succession of temperate forests, from shrublands and coniferous forests to deciduous broad‐leaved forests. The change in soil properties caused by succession was likely the driver behind the changes in the fungal community. Soil AP and NO 3 \n − ‐N were strongly associated with fungal community structure and function. In the early stage of succession, the shrub biomes significantly accumulated fungal pathogens, while in the late stage, the deciduous broad‐leaved forest significantly accumulated ectomycorrhizal fungi. The changes in fungal community function may have a potential influence on the succession process of temperate forests. We found that secondary succession in temperate forests had significant effects on soil nutrients, with AP and NO 3 \n − ‐N being significantly higher in early‐ than in later‐successional stages. Nitrogen and phosphorus availability showed a decreasing trend during forest succession. In previous studies on subtropical and tropical forest ecosystems, succession was often accompanied by an increase in soil nutrient availability (Bauters et al., 2022 ; Sullivan et al., 2019 ; Zheng et al., 2020 ). In contrast, in temperate forest ecosystems, plants in the middle‐ and late‐successional stages were predominantly ectomycorrhizal trees from the Pinaceae and Fagaceae families. These are characterized by slowly decomposing litter and lower nutrient cycling rates, which may explain the reduced availability of nitrogen and phosphorus during temperate forest succession (Lin et al., 2017 ; Phillips et al., 2013 ; Schilling et al., 2016 ). The tendency for fungal community diversity to decrease across successional stages may be explained by changes in soil nutrients (Dini‐Andreote et al., 2015 ; Zheng & Song, 2022 ). In the early stages of secondary succession, disturbances often result in the liberation of nutrients. This can facilitate the rapid colonization of soils by a diverse microbial population (Zhang et al., 2016 ). Plants appearing during the secondary stage of succession have been shown to grow faster and consume more nutrients (Lajtha, 2019 ). As succession proceeds, the use of nutrients by plants can therefore lead to a decrease in soil nutrient availability (Huang et al., 2012 ), thereby reducing the diversity of the soil fungal community. NO 3 \n − ‐N and AP were the main predictors of soil fungal community structure during this succession process and were significantly correlated with fungal functional groups. As important decomposers of soil organic matter, saprophytic fungi are sensitive to changes in soil properties. The abundance of saprotrophic guilds was positively correlated with NO 3 \n − ‐N and AP, and the high abundance of saprotrophs led to a faster turnover of soil nutrients and facilitated the maintenance of high nutrient availability (Liu et al., 2022 ; Wu et al., 2019 ). High nitrogen and phosphorus availability also promoted the colonization of fungal saprotrophs in the soil (Zhang, Dong, et al., 2022 ). Different guilds of saprotrophs showed different responses to total nutrients. This may reflect variation in the preference of different saprotroph guilds in temperate forest soils for total nutrients in the substrate (Cao et al., 2022 ; Cline et al., 2018 ; Guo et al., 2022 ). Furthermore, NO 3 \n − ‐N and AP contents during the succession process were not only negatively correlated with the abundance of ectomycorrhizal fungi but also positively correlated with those of arbuscular mycorrhizal fungi. This can be explained by the difference in plant nutrient acquisition between these two groups (Cheeke et al., 2016 ; Genre et al., 2020 ; Tedersoo et al., 2012 ). Ectomycorrhizal fungi have the ability to access nitrogen and phosphorus from organic material and transfer them to the host plant (Smith & Read, 2008 ; Tedersoo & Bahram, 2019 ), whereas arbuscular mycorrhizal fungi have relatively limited capacity for enzymatic degradation and mainly take up nutrients in mineral form (Chen et al., 2018 ; Tisserant et al., 2013 ). Therefore, soils with high nitrogen and phosphorus availability are more conducive to colonization by arbuscular mycorrhizal fungi. Given that ectomycorrhizal fungi incur a considerable carbon cost from cooperation, host plants in high nutrient availability environments often reduce their symbiosis with these fungi (Guo et al., 2021 ; Nilsson et al., 2005 ; Peng et al., 2022 ). Ectomycorrhizal fungi are thus more abundant in nutrient‐deficient than in nutrient‐rich environments (Bai et al., 2019 ). This phenomenon is consistent with the findings of our study. One emerging pattern in natural systems analysis is that pathogens often thrive in resource‐rich environments (Revillini et al., 2016 ; Reynolds et al., 2003 ). A series of studies in grassland ecosystems and agroecosystems have also shown that high nitrogen and phosphorus availability often leads to an increase in fungal pathogens (Ebeling et al., 2021 ; Lekberg et al., 2021 ). We observed that the abundance of pathotrophic guilds was positively correlated with NO 3 \n − ‐N and AP. This suggests that higher soil nutrient availability may be the reason for the accumulation of fungal pathogens in the early stage of succession. However, the correlation between plant pathogen fungi and soil nutrients is the opposite of that observed in ectomycorrhizal fungi, as the association of plant roots with ectomycorrhizal fungi also protects the host from pathogens (Bennett et al., 2017 ; Liang et al., 2020 ; Tedersoo et al., 2020 ). The correlation between plant pathogens and soil nutrients may thus be indirectly driven by ectomycorrhizal fungi. We found that during secondary succession in temperate forests, early‐successional forest types accumulated fungal plant pathogens, while late‐successional forest types accumulated ectomycorrhizal fungi. The positive driving role of microbial pathogens in community succession and the accumulation of pathogenic fungi in the soil can cause reduced seedling recruitment and survival around a conspecific adult. This in turn may facilitate forest succession (Domínguez‐Begines et al., 2020 ; Van Der Putten et al., 1993 ). The accumulation of ectomycorrhizal fungi is an important mechanism for maintaining dominant communities of host trees in temperate forests (Chen et al., 2019 ; Liang et al., 2020 ). As an important plant symbiont, ectomycorrhizal fungi can expand the absorption range of the root system and transport nutrients from organic matter to the host (Genre et al., 2020 ; Tedersoo et al., 2020 ). Given that we observed that nutrient availability decreased with succession, ectomycorrhizal fungi may play a more important role in nutrient acquisition in late‐successional forest types. We also observed that the relative abundance of ectomycorrhizal fungi increased between middle‐ and late‐successional stages. The abundance of ectomycorrhizal fungi often increased significantly in the late‐successional stage of temperate and boreal forests (Clark, & St. Clair, 2011 ; Jiang et al., 2021 ; Zhang, Zhao, et al., 2022 ). This phenomenon is consistent with the findings of our study. The accumulation of fungal plant pathogens in the early‐successional stage may lead to increased negative plant–soil feedback, while the accumulation of ectomycorrhizal fungi in the late‐successional stage may lead to increased positive plant–soil feedback (Liang et al., 2016 , 2020 ; Van Der Putten & Peters, 1997 ). Therefore, we suggest that changes in fungal functions during secondary succession in temperate forests potentially have a driving role in the succession process."
} | 4,101 |
23102412 | null | s2 | 3,645 | {
"abstract": "Fatty acid metabolism has received significant attention as a route for producing high-energy density, liquid transportation fuels and high-value oleochemicals from renewable feedstocks. If microbes can be engineered to produce these compounds at yields that approach the theoretical limits of 0.3-0.4 g/g glucose, then processes can be developed to replace current petrochemical technologies. Here, we review recent metabolic engineering efforts to maximize production of free fatty acids (FFA) in Escherichia coli, the first step towards production of downstream products. To date, metabolic engineers have succeeded in achieving higher yields of FFA than any downstream products. Regulation of fatty acid metabolism and the physiological effects of fatty acid production will also be reviewed from the perspective of identifying future engineering targets."
} | 214 |
29075325 | PMC5651578 | pmc | 3,646 | {
"abstract": "Background Economical conversion of lignocellulosic biomass into biofuels and bioproducts is central to the establishment of a robust bioeconomy. This requires a conversion host that is able to both efficiently assimilate the major lignocellulose-derived carbon sources and divert their metabolites toward specific bioproducts. Results In this study, the carotenogenic yeast Rhodosporidium toruloides was examined for its ability to convert lignocellulose into two non-native sesquiterpenes with biofuel (bisabolene) and pharmaceutical (amorphadiene) applications. We found that R. toruloides can efficiently convert a mixture of glucose and xylose from hydrolyzed lignocellulose into these bioproducts, and unlike many conventional production hosts, its growth and productivity were enhanced in lignocellulosic hydrolysates relative to purified substrates. This organism was demonstrated to have superior growth in corn stover hydrolysates prepared by two different pretreatment methods, one using a novel biocompatible ionic liquid (IL) choline α-ketoglutarate, which produced 261 mg/L of bisabolene at bench scale, and the other using an alkaline pretreatment, which produced 680 mg/L of bisabolene in a high-gravity fed-batch bioreactor. Interestingly, R. toruloides was also observed to assimilate p -coumaric acid liberated from acylated grass lignin in the IL hydrolysate, a finding we verified with purified substrates. R. toruloides was also able to consume several additional compounds with aromatic motifs similar to lignin monomers, suggesting that this organism may have the metabolic potential to convert depolymerized lignin streams alongside lignocellulosic sugars. Conclusions This study highlights the natural compatibility of R. toruloides with bioprocess conditions relevant to lignocellulosic biorefineries and demonstrates its ability to produce non-native terpenes. Electronic supplementary material The online version of this article (doi:10.1186/s13068-017-0927-5) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusions \n Rhodosporidium toruloides is emerging as a promising new production platform for the conversion of lignocellulose into biofuels and bioproducts. Much effort has focused on its oleaginous properties (high lipid proportions; > 60% w/w of cell mass), and it has been engineered to produce several lipid derivatives [ 8 , 39 ]. It has also been examined for its production of potentially valuable native carotenoids: β-carotene, torularhodin, and torulene [ 40 ]. In this study, we demonstrate that this organism is a versatile production host that possesses many features critical to reducing CAPEX and OPEX in a biorefinery: (1) it can be used to make a variety of bioproducts, including non-native terpenes with biofuel and pharmaceutical applications, (2) heterologous production of bioproducts does not require inducers or antibiotics and is stable through multiple generations, (3) it can efficiently utilize components of both the polysaccharide and lignin fractions of inexpensive, carbon neutral, and renewable lignocellulosic feedstocks, (4) it is compatible with single-unit operation pretreatment, saccharification, and fermentation bioprocessing configurations, and (5) bioproduct productivity is not inhibited in lignocellulosic hydrolysates. No other microbial production platform has been demonstrated to harbor all these properties, and R. toruloides sets a new standard for biotechnological applications that support a green economy.",
"discussion": "Results and discussion R. toruloides as a platform for terpene production Terpenes are produced by a variety of organisms and have a wide range of applications from flavors, fragrances, and pharmaceuticals to biofuels and chemical feedstocks [ 24 ]. In this study, we selected two terpenes, amorphadiene and bisabolene, to examine the suitability of R. toruloides as a lignocellulosic conversion host. Amorphadiene, a precursor of the antimalarial drug artemisinin, was chosen as an example of a commercially relevant bioproduct [ 25 ] and bisabolene, an immediate precursor of the D2 diesel alternative bisabolane, was chosen as an example of an advanced “drop-in” biofuel [ 26 ]. Codon-optimized expression cassettes for bisabolene (BIS) and amorphadiene (ADS) synthases were constructed and separately integrated into the genome of R. toruloides IFO0880 using A. tumefaciens -mediated transformation (ATMT) [ 8 ]. A number of transformants were confirmed to produce either bisabolene (Fig. 2 a) or amorphadiene (Fig. 2 b), with variance in titer between strains most likely due to copy number and integration site effects [ 27 , 28 ]. Terpene titers for selected strains in synthetic defined (SD) medium containing 2% (w/v) glucose, reached 294 mg/L for bisabolene and 36 mg/L for amorphadiene. These bisabolene and amorphadiene titers attained in R. toruloides are highly encouraging considering that they exploit the natural flux of carbon through this yeast’s native terpene biosynthetic pathway. In comparison, the yeast S. cerevisiae transformed with high copy plasmids harboring the BIS and ADS genes and grown in equivalent media attained significantly lower bisabolene and amorphadiene titers—approximately 20 and 10 mg/L, respectively ([ 29 ] for ADS, unpublished data for BIS). Another notable feature of the R. toruloides BIS strain is that bisabolene titers show remarkable stability over extended periods of serial cultivation, varying by less than 16% over the course of four cultures spanning 24 days (Fig. 3 ). It should be noted that this reproducibility was also achieved without the need for a heterologous inducer or antibiotic selection, since the BIS gene is stably integrated into the genome and its expression is under control of a constitutive GAPDH promoter [ 8 – 10 ]. Both of these features reduce OPEX in a biorefinery. In comparison, the bisabolene titer from an engineered strain of S. cerevisiae grown under similar conditions was found to decline by more than 75% over 14 days [ 30 ]. The strain stability we observed in engineered R. toruloides is an important industrial phenotype and a critical factor for large-scale economical production of any bioproduct. Fig. 2 Terpene titers of R. toruloides transformants. a Bisabolene and b amorphadiene titers in selected strains grown in SD medium with 2% glucose. 5 mL cultures in test tubes were set up at a starting OD of 0.1 with a 20% dodecane overlay. At day 7, the dodecane layer was sampled and analyzed for bisabolene or amorphadiene. ( n = 3, data shown as average ± standard deviation, representative from two independent experiments) \n Fig. 3 Stability of bisabolene production in serial cultures. Cultures in SD medium with 2% glucose were passaged consecutively every 6 days. ( n = 3, data shown as average ± standard deviation, from a single experiment) \n We found that the pH of the growth medium is an important factor for efficient sugar utilization by R. toruloides . After examining a range of starting pH values in unbuffered medium (3–8) in batch cultures, a starting pH of 7.4 was determined to be optimal to achieve complete glucose utilization (Table 2 ) and the highest bisabolene titer (Fig. 4 ). Interestingly, R. toruloides grew and produced bisabolene at a pH as low as 3.4, suggesting that the host may be amenable to the production of organic acids or other bioproducts that require low pH. One potential explanation for the decline in pH is that R. toruloides is producing native organic acids of potential value, a topic that merits further investigation. However, once the pH declines to 2.5 (in unbuffered medium starting at pH 7 or below), sugar utilization is strongly inhibited, suggesting that the pH must remain above this level to enable efficient carbon conversion. Therefore, all subsequent experiments in unbuffered media were performed with a starting pH of 7.4. Table 2 Percent utilization of glucose in SD media starting at various pH pH % utilization 3.4 52.6 ± 5.4 4.4 67.0 ± 4.7 5.4 69.3 ± 3.5 6.4 76.4 ± 1.3 7.4 100.0 ± 0.0 8.4 96.8 ± 2.4 Cultures were carried out as described above, the aqueous layer was sampled for glucose analysis ( n = 3, data shown as average ± s.d, from a single experiment) \n Fig. 4 The effect of pH on bisabolene titers. ( n = 3, data shown as average ± standard deviation, from a single experiment in SD medium with 2% glucose) \n R. toruloides can convert multiple carbon sources into bisabolene in defined media To demonstrate the capability of engineered R. toruloides to utilize different carbon sources to produce non-native terpenes, we cultivated the bisabolene-producing strain BIS3 with the most abundant sugars present in lignocellulosic hydrolysates: glucose and xylose. In addition, we observed the liberation and consumption of p -coumaric acid in the IL-pretreated corn stover hydrolysate described in detail below (the hydrolysate in Fig. 7 ), and therefore we examined this compound as a carbon source along with the sugars. This phenolic compound is associated with grass lignins through an ester linkage to lignin monomers formed prior to lignin polymerization [ 12 ]. This ester linkage can easily be cleaved under alkaline conditions [ 31 ], like those found in the pretreatments used to generate hydrolysates in this study. Initially, these three carbon sources were provided individually and growth, carbon utilization, and bisabolene production were monitored (Fig. 5 a–c). Glucose was completely consumed at the fastest rate, followed by p -coumaric acid, then xylose (in 1, 3, and 4 days, respectively). The highest bisabolene titers were observed in the p -coumaric acid cultures, likely due to its higher percentage of carbon relative to the sugars (Fig. 5 a–c). Remarkably, when combined, all three carbon sources were completely utilized within 4 days (Fig. 5 d). The p -coumaric acid was completely utilized earlier in the presence of the other sugars (2 vs 3 days), while complete utilization of glucose and xylose took slightly longer when present in the mixture (glucose: 2 vs 1 day; xylose: both day 4 but less consumed by day 2 in the mixture). Fig. 5 Conversion of glucose, xylose, and p -coumaric acid, both individually and mixed, into bisabolene by R. toruloides . Bisabolene titers, growth, and carbon utilization of strain BIS3 grown in SD medium supplemented with different carbon sources: a 0.5% glucose, b 0.5% xylose, c 0.5% p -coumaric acid, and d 0.5% glucose, 0.5% xylose, 0.5% p -coumaric acid. Left panels: lines represent ODs, bars represent bisabolene titers. Right panels: glucose (red), xylose (black), p -coumaric acid (blue). 5 mL cultures in test tubes were set up at a starting OD of 0.1 with a 20% dodecane overlay. At each time point, the dodecane layer was sampled and analyzed for bisabolene and the aqueous layer was sampled for OD measurement and carbon utilization analysis ( n = 3, data shown as average ± standard deviation, representative from at least four independent experiments) \n Much effort has been expended on metabolic engineering of common microbial host organisms such as E. coli and S. cerevisiae for simultaneous utilization of multiple carbon sources, such as glucose and xylose [ 21 , 32 ]. The ability of R. toruloides to efficiently utilize multiple carbon sources, particularly hexose and pentose sugars combined with aromatic compounds, is something that even extensively engineered strains of S. cerevisiae and E. coli have been unable to accomplish. However, the decrease in glucose and xylose consumption rates in R. toruloides cultures grown on mixed sugars merits further investigation to determine if there is competitive sugar transport, catabolite repression, or other mechanisms affecting the kinetics. Consumption of lignin-related monoaromatics The ability of R. toruloides to consume p -coumaric acid indicates that it may also have the natural ability to consume other lignin-related aromatic compounds. To better understand this potential, R. toruloides was tested for its ability to consume compounds related to possible lignin degradation products derived from the H, G, and S subunits. Compounds were tested with the same aromatic motif as the 4-hydroxyphenyl H units ( p -coumaric acid and p -hydroxybenzoic acid), the 4-hydroxy-3-methoxyphenyl G units (ferulic acid and vanillic acid), as well as the non-substituted aromatic compound benzoic acid. Complete consumption of each substrate occurred within 72 h in single carbon cultures, except for vanillic acid, which may have been toxic at the level tested (Fig. 6 ). We also tested a compound with the same aromatic motif as the 4-hydroxy-3,5-dimethoxyphenyl S units (sinapic acid), but the results were inconclusive due to apparent oxidation and precipitation of the substrate during the cultivation. Overall, these results indicate that R. toruloides has the metabolic potential to consume lignin degradation products derived from depolymerization processes that produce compounds similar to those tested. This capability highlights the potential of R. toruloides to be used as a conversion host of monoaromatic lignin degradation products, a characteristic that will become more important as biomass deconstruction technologies advance to provide more extensive lignin depolymerization. However, lignin depolymerization technologies can produce a very diverse and heterogeneous mixture of aromatic and non-aromatic compounds, so although R. toruloides has the potential to convert certain compounds derived from depolymerized lignin, much work would need to be done to tailor both the structure of lignin and lignin depolymerization strategies to bias the product range toward compounds that can be readily consumed by this organism. For example, one possible strategy to do this would be to engineer bioenergy feedstocks to generate lignin more heavily acylated with p -coumarate, which could then be readily liberated from the lignin under alkaline pretreatment conditions and converted into bioproducts [ 33 ]. Fig. 6 Utilization of several lignin-related aromatic compounds by R. toruloides. Carbon source utilization of R. toruloides grown in SD medium supplemented with 2 g/L of either p -coumaric acid, p -hydroxybenzoic acid (4-HBA), ferulic acid, vanillic acid, or benzoic acid ( n = 3, data shown as average ± standard deviation, from a single experiment) \n Bisabolene production and carbon source utilization in a corn stover hydrolysate pretreated with the novel biocompatible ionic liquid choline α-ketoglutarate The performance of R. toruloides grown on purified substrates indicates that it may be an excellent biocatalyst for the conversion of deconstructed lignocellulose into valuable bioproducts. To test this premise, we examined how R. toruloides performs on substrates derived from actual lignocellulosic biomass. There are many technologies that have been developed to efficiently depolymerize biomass into intermediates suitable for microbial conversion, and those based on ionic liquid (IL) pretreatment and enzymatic saccharification have been demonstrated to be some of the most efficient and effective [ 34 – 36 ]. Recently, biocompatible ILs that do not inhibit commercial cellulase enzyme mixtures or microbial growth have been developed, enabling single-unit operation biomass pretreatment, saccharification, and fermentation, potentially reducing both CAPEX and OPEX in a biorefinery [ 37 , 38 ]. Therefore, to test the performance of R. toruloides on a biomass hydrolysate, a corn stover hydrolysate containing glucose, xylose, and p -coumaric acid was generated using pretreatment with a novel biocompatible IL, choline α-ketoglutarate ([Ch][α-Kg]), followed by enzymatic saccharification (see “ Methods ” for details). It was chosen in this study because it was found to be biocompatible with R. toruloides. This IL falls into a recently developed class of ILs based on dicarboxylic acids such as choline glutamate ([Ch][Glu]) [ 37 ]. Compositional and X-ray diffraction data suggest that [Ch][α-Kg] is functionally equivalent to [Ch][Glu]: pretreatment with [Ch][α-Kg] reduced recalcitrance to enzymatic saccharification by removing high amounts of lignin from the biomass and reducing the cellulose crystallinity (Table 3 and Additional file 1 ). R. toruloides was able to grow in the [Ch][α-Kg] hydrolysate, completely consume glucose, xylose, and p -coumaric acid, and produce 261 mg/L of bisabolene (Fig. 7 a). In fact, it produced higher titers of bisabolene in the hydrolysate than it did in a control medium with matching concentrations of the IL, sugars, and p -coumaric acid (127 mg/L) (Fig. 7 b). Table 3 Chemical composition of dominant components in the dry corn stover before and after pretreatment Solid recovery/% Glucan/wt% Xylan/wt% Lignin/wt% Untreated corn stover / 38.90 ± 0.04 24.77 ± 0.01 18.42 ± 0.27 Pretreated corn stover 60.0 48.35 ± 0.06 28.92 ± 0.05 13.12 ± 0.03 Pretreatment conditions: 20 wt% biomass loading, 80 wt% [Ch][α-Kg] (40 wt% in H 2 O), 120 °C, 4 h \n Fig. 7 Conversion of biomass-derived glucose, xylose, and p -coumaric acid into bisabolene by R. toruloides. Bisabolene titers, growth, and carbon source utilization of strain BIS3 grown in ( a ) corn stover hydrolysate and ( b ) SD medium supplemented with individual components at the same concentration as those found in the corn stover hydrolysate: glucose (17.1 g/L), xylose (9.1 g/L), p -coumaric acid (383 mg/L), alpha-ketoglutarate (254 mM), and choline (586 mM). A low level of arabinose (0.98 g/L) was also detected in the hydrolysate and included in the control medium. Left panels: lines represent ODs, bars represent bisabolene titers. Right panels: glucose (red), xylose (black), p -coumaric acid (blue). 5 mL cultures in test tubes were set up at a starting OD of 0.1 with a 20% dodecane overlay. At each time point, the dodecane layer was sampled and analyzed for bisabolene and the aqueous layer was sampled for OD measurement and carbon utilization analysis ( n = 3, data shown as average ± standard deviation, representative from at least two individual experiments) \n R. toruloides converts corn stover hydrolysate into bisabolene in high-gravity fed-batch bioreactors In order to examine the spectrum of hydrolysates that R. toruloides can utilize and determine the impact of optimized cultivation conditions on bisabolene titers, a corn stover hydrolysate generated from an alkaline pretreatment was also tested. This pretreatment method generates very high concentrations of glucose and xylose, so it can be used for high-gravity fed-batch cultivation, which enables the addition of significantly more carbon than in the batch cultivations conducted with the [Ch][α-Kg] hydrolysate. The drawback to this approach is that lignin depolymerization products are removed during the process, so only glucose and xylose utilization can be examined. R. toruloides was cultivated in a controlled, high-gravity fed-batch bioreactor using the alkaline corn stover hydrolysate or a glucose-only medium set at a glucose concentration similar to those found in industry, and produced 680 and 521 mg/L bisabolene, respectively (Fig. 8 a, b). The lower titer in the glucose-only medium may be due to the lower amount of sugars added to the cultivation (alkaline: 73.8 g/L and glucose-only: 61.5 g/L), resulting in a slightly lower dry cell weight (alkaline: 27 g/L and glucose-only: 25 g/L) and lower bisabolene production. It is interesting to note that R. toruloides produced high titers of bisabolene in both hydrolysates relative to defined medias. In many instances, the opposite has been observed for organisms like S. cerevisiae and E. coli , further demonstrating the greater suitability of R. toruloides as a lignocellulosic conversion host. Fig. 8 High-carbon fed-batch fermentation of R. toruloides . Bioreactor cultivation of strain BIS3 in ( a ) alkaline hydrolysate ( b ) SD medium with glucose. Left panels: lines represent dry cell weight (DCW), bars represent bisabolene titers. Bisabolene titers were measured three times per time point, average value is shown. Right panels: measured glucose (red) and xylose (black) concentration profiles. At each time point, 10 mL of the culture was sampled. After separation, the dodecane layer was used for bisabolene measurement. 5 mL of the aqueous layer was used for the measurement of DCW \n These results show that R. toruloides is amenable to high-carbon fed-batch fermentations, which is another important feature when considering organisms for use in industry. The bisabolene titer of 680 mg/L achieved in this study is impressive relative to titers reported for strains of S. cerevisiae and E. coli that have undergone extensive genetic engineering [ 25 , 26 , 29 ]. In addition, no significant reduction in the native pools of lipids or carotenoids was observed in the bisabolene-producing strains compared to wild type, suggesting that significant increases in titer can be achieved by further strain engineering to divert carbon flux away from these native molecules (both of which are derived from acetyl-CoA) toward bisabolene (Additional file 2 )."
} | 5,338 |
36641786 | PMC10364308 | pmc | 3,647 | {
"abstract": "Abstract Our planet teeters on the brink of massive ecosystem collapses, and arid regions experience manifold environmental and climatic challenges that increase the magnitude of selective pressures on already stressed ecosystems. Ultimately, this leads to their aridification and desertification, that is, to simplified and barren ecosystems (with proportionally less microbial load and diversity) with altered functions and food webs and modification of microbial community network. Thus, preserving and restoring soil health in such a fragile biome could help buffer climate change's effects. We argue that microorganisms and the protection of their functional properties and networks are key to fight desertification. Specifically, we claim that it is rational, possible and certainly practical to rely on native dryland edaphic microorganisms and microbial communities as well as dryland plants and their associated microbiota to conserve and restore soil health and mitigate soil depletion in newly aridified lands. Furthermore, this will meet the objective of protecting/stabilizing (and even enhancing) soil biodiversity globally. Without urgent conservation and restoration actions that take into account microbial diversity, we will ultimately, and simply, not have anything to protect anymore.",
"conclusion": "CONCLUSION The United Nations have declared the decade starting with 2020 as the “UN Decade of Ecosystem Restoration” and even highlighted the role of BSCs in this pursuit. However, the recent outcome of COP27 shows that there is still much work to do to convince political leaders to take effective decisions against climate change. In this context of the political status quo , very little attention is given to drylands, which constitute one of the largest and certainly among the most vulnerable biomes on Earth, and even less to the application of their adapted/selected edaphic microbial communities to prevent and mitigate soil erosion and fertility loss. Here, we put forward an opinion, guided by an important set of recent advancements in understanding the dryland systems and in the potential applications for restorations and by the need for nature‐based solutions in our fight against climate change. We consider that the endemic dryland soil microbiota is revealed to be a concrete and viable solution for protecting vulnerable soil ecosystems under the threat of climate change and for sustaining dryland agriculture and food production. As discussed above, desertification and soil erosion should be counteracted by applying microbial Nature‐Based solutions. There is evidence that microorganisms already adapted to arid conditions, microbial communities and metaorganisms, such as BSCs and xerophytic plants, have the potential to be exploited for such nature‐based‐solutions and should be at the forefront of the strategies implemented in ecosystem restoration of degraded landscapes. The biodiversity and life‐history traits of arid/desert ecosystems and their biota is paramount and offer a natural advantage ahead of introducing foreign microorganisms or man‐made processes. We have described microbes‐centred strategies that could be used to mitigate the effect of aridification in drylands: (i) the protection and conservation of edaphic dryland microbiome from total disruption (complete desertification) as well as the restoration of its biodiversity (e.g. BSCs and plants) when already impacted, (ii) support adaptation of crops in drylands by using beneficial microbes (probiotics). Combined, these approaches will limit soil erosion and they should be used to improve current practices of arid land management and have the potential to be used either individually or in concert. In other words, if we do not act now, we will not have anything left to protect in the future as arid lands are expanding and therefore, their BSCs and vegetation risk disappearing (Figure 1 ). Restoring microbial biodiversity and multifunctionality in the already lost soil is critical for ecologically and socially responsible landscape restoration attempts, particularly as plants accelerate their biomass production when their endemic soil microbiome is healthy . Thus, we strongly advocate for an urgent awareness that we no longer have the time to “wait and see”. Actions should be made , and decisions should be taken . Within the scale of ecological catastrophes, it is indeed more pressing to conserve and restore drylands than to monitor off‐target effects while BSC cover erodes globally and 2 billion people are on the verge of starvation. FIGURE 1 The need to modify our perspective on global drylands and their protection. With climate change, drylands are expanding and becoming drier. In this context, it is necessary to “act now”, and we argue that dryland conservation, restoration and management can—and must—be done using “microbe‐based solutions”. Indeed, with probably one more billion human beings living in drylands in the next decade, and to conserve their diversity and food production potential, we have to consider the entire dryland ecosystem, including their indigenous microorganisms. We advocate for the use of dryland microbial communities in any environmental initiatives strategies and global diversity framework designed to mitigate the effects of climate change in drylands globally, giving them the attention deserve."
} | 1,344 |
35059153 | PMC8694393 | pmc | 3,648 | {
"abstract": "DNA tweezers have emerged as powerful devices for a wide range of biochemical and sensing applications; however, most DNA tweezers consist of single units activated by DNA recognition, limiting their range of motion and ability to respond to complex stimuli. Herein, we present an extended, tripodal DNA nanotweezer with a small molecule junction. Simultaneous, asymmetric elongation of our molecular core is achieved using polymerase chain reaction (PCR) to produce length- and sequence-specific DNA arms with repeating DNA regions. When rigidified, our DNA tweezer can be addressed with streptavidin-binding ligands. Full control over the number, separation, and location of these ligands enables site-specific streptavidin recognition; all three arms of the DNA nanotweezer wrap around multiple streptavidin units simultaneously. Our approach combines the simplicity of DNA tile arrays with the size regime normally provided by DNA origami, offering an integrated platform for the use of branched DNA scaffolds as structural building blocks, protein sensors, and dynamic, stimuli-responsive materials.",
"conclusion": "Conclusions We have developed a versatile method for the assembly of a large DNA nanotweezer with multiple, asymmetric arms. Using DNA strands with chemically conjugated branched units, we imbue our construct with flexibility and asymmetry that is propagated by the sequential growth of all three unique DNA arms. The dynamic behavior of our core junction, coupled with the rigidity of DNA arms, is critical to ensure the folding of our tripodal nanotweezer in response to protein stimuli, yielding a new DNA nanostructure: a 3D protein–DNA hybrid nanotube. Arm rigidity also allows precise nanomaterial organisation with defined separation, and the unique sequences in each arm allow independent and selective asymmetric streptavidin patterning. This method is complementary to DNA origami, but it is advantageous when larger wireframe structures are desired, whose size is not limited by folding of a viral scaffold strand; because it requires significantly fewer component strands, it is also valuable when a structure needs to be built from a minimal number of starting strands, such as in vivo applications. The covalent nature of our tweezer fulcrum endows robustness of the structure for biological applications, facilitates PCR manipulations and is highly tunable because of the variety of small molecules that could be used as corners of different strands. Our methodology will find broad applicability in generating more complex DNA-hybrid materials, especially in conjunction with other types of multivalent proteins or nanoparticles. By applying our new “printing-elongation-folding” methodology to more diverse small molecule cores with further branching degrees and trigger stimuli, we will export the utility of hybrid DNA-small molecule motifs to the construction of extended DNA nanotweezers with more diverse and complex guest sensing behaviors.",
"introduction": "Introduction Molecular tweezers are synthetic hosts with open cavities, often composed of multiple branching arms connected by a junction moiety. 1–5 The rigidity of this hinge is a strong determinant of tweezer properties: rigid spacers result in pre-organized concave binding sites, while more flexible spacers lead to a mechanical pincer-like motion that responds to the spatial and chemical requirements of guest binding. 1,6,7 DNA tweezers have emerged as powerful nanomechanical devices, with the ability to sense nucleic acids, proteins, and cellular processes, to control enzymatic activity, and measure biological distances. 8–18 Most DNA tweezers translate molecular recognition (typically another DNA strand) into short-range mechanical motion, resulting in a signal or function. 13,19–27 Extended DNA tweezers with arms containing multiple binding sites along their lengths would be able to translate multiple guest binding events into large-scale mechanical motion, reminiscent of octopus or squid tentacles around prey. This large-scale motion may find applications in molecular robotics, biosensing schemes with amplification, and cellular probes that interrogate and influence large sections of the cell membrane. 26,28–39 Herein, we report the synthesis of an extended DNA tweezer with a trivalent synthetic molecule as the core and three rigid and long DNA arms of different sequences ( Fig. 1 ). Site-specific, equidistant placement of biotin moieties on the three arms causes them to fold together when a streptavidin target is added, in a large-scale motion over hundreds of nanometers. The result is a multivalent, hybrid protein–DNA nanotube-like structure, where the nanotube “rungs” are streptavidin, and the arms are DNA. The three-way DNA scaffold is built using a “printing” process developed in our laboratory, 40,41 in which a pattern of three DNA arms is covalently transferred from a DNA nanostructure onto a small molecule, which then acts as the tweezer fulcrum. The sequence uniqueness of the three arms allowed their simultaneous extension by PCR using three different long DNA strands as templates, and the protein-binding three-way DNA nanostructure is formed when the arms are rigidified and addressed with binding ligands. Because of the ability to change the nature of the small molecule spacer on demand, we show that a more flexible molecular spacer gives a significantly higher yield of the folded streptavidin–DNA nanotube than a small and rigid aromatic spacer. Our methodology can be used as a new tool for the construction of DNA-minimal, stimuli-responsive architectures. Fig. 1 Schematic representation of three-way DNA nanostructure synthesis via (A) a “printing” strategy from a DNA junction, (B) arms elongation and rigidification, and (C) protein-binding tweezer formation. Biotinylated sequences were placed at an equidistant location from the branching core to favour binding to the same streptavidin. Upon incubation with streptavidin, the three-way DNA nanostructure recognizes the multivalent protein, causing the 3 arms of DX-T3 to fold into a 3D nanotube-like structure.",
"discussion": "Results and discussion Design parameters and tweezer formation To form large multivalent DNA tweezers using a size-defined three-way DNA nanostructure, there are four key design requirements. First, the long arms of the structure must be rigid, to translate molecular recognition into a large, directed motion about the junction, and to avoid intramolecular binding around the protein. Second, the binding sites on each arm must be equidistant from the branched building block so that multiple arms of the structure bind to the same guest molecules. As a third criterion, we were interested in generating this structure from a small number of strands, which requires the use of identical DNA sequences in strategic positions. Fourth, the tweezer junction must be covalently connected to DNA arms of different sequences, so that the tweezer arms can be elongated using the polymerase chain reaction (PCR) which requires cycles of heating and cooling, and its structure and flexibility must be tunable. Our group has recently developed methods for DNA “printing” – transferring a pre-defined pattern of DNA sequences onto other types of materials. 40 This simple process was used to generate synthetic vertices attached to multiple DNA arms with controllable valency, different sequences, and directionalities. 40 We applied this printing process to generate a small molecule aromatic core connected to 3 different DNA arms (Trimer or T3; DNA arms R1, R2, R3: 42, 39, and 41 bases respectively) (Fig. S1, ESI-VI). † We were interested first in producing the long tripodal tweezer. We designed each of the three short arms of T3 to act as a forward primer for elongation by PCR. Three long DNA strands were required as PCR templates. It is possible to use the viral single-stranded DNA scaffolds that are normally employed for DNA origami, 42,43 but this would then require a large number of complementary strands to rigidify the structure and substitute it with protein-binding units. To reduce the number of DNA components, we needed custom-made, long sequences with repeating DNA regions, as well as unique regions, and with full control over their number and placement. These sequences are too long to be built on an automated DNA synthesizer. We generated these template strands using a “temporal growth” method previously developed by our group. 44 In this method, the strands are built by sequentially adding complementary DNA building blocks. Each building block is composed of a strand with the desired sequence, and a complementary strand is used to form sticky end overhangs at each end of the duplex. DNA building blocks are progressively added to a seed unit that can only hybridize in one direction with simultaneous enzymatic ligation to covalently attach each additional building block as it is introduced. PCR is then used to isolate and enrich the full-length product. Each building block can be designed with different sequence domains, giving different DNA patterns. This resulted in three strands with alternating repeating domains: (AB)5, (CD)3 and (EF) 4 , where A–F are 42 base–pair building blocks of different sequences (ESI-VII). † Another important advantage of this strategy over DNA origami is that structural dimensions can be larger, as they can take advantage of the full length of the template strands, rather than relying on folding of a scaffold strand. We first elongated each of the three arms of T3 separately, using PCR with one of the three long strands as template, resulting in a clean product ( Fig. 2A , native agarose gel electrophoresis (AGE) lanes 4, 5 and 6, ESI-VII † ). Simultaneous extension of all three unique arms resulted in the formation of the desired product possessing three different arms with their expected lengths, in addition to side-products from elongation of one or two arms ( Fig. 2A , native AGE lane 9). Band excision from this gel led to the tripodal branched DNA structure with elongated arms. Characterization by AFM showed formation of the correct product as a monodisperse, star-like structure with three long arms and varying angles between them. In this structure, the double-stranded arms have a high degree of flexibility ( Fig. 2B ) which may favor the intramolecular folding of the arms on themselves, impinging the tweezer effectiveness. Fig. 2 Native agarose gel electrophoresis (AGE) characterization of mono-, bi- and tri-elongated “printed” junction (T3) (A). Lanes 1, 2, 3: (CD) 3 (252 bps), (EF) 4 (336 bps) and (AB) 5 (420 bps) respectively; lanes 4, 5 and 6: T3 separately elongated by PCR with (CD) 3 (315 bps), (EF) 4 (399 bps) and (AB) 5 (483 bps), respectively. A–F are 42 base–pair building blocks of different sequences. Lanes 7 and 9: simultaneous PCR performed in the presence of (CD) 3 , (EF) 4 and (AB) 5 without T3 and with T3 (1071 bps), respectively. Lane 8: PCR product of (EF) 4 and (AB) 5 with T3 (819 bps). GeneRuler DNA ladder mix is used. (B) AFM characterization of “printed” junction with elongated arms in its double-stranded form. Note that the spherical structures are salts resulting from the surface drying, while the branched structures correspond to the extended junction. To rigidify our construct for subsequent tweezer formation and materials patterning, periodic double-crossover (DX) tile structures were designed to assemble on each of the arms (ESI-III. B2). † These tiles have an even number of helical half-turns between crossover points (DAE) and they consist of 1 helical turn (10 bps) between the crossover points and 1.5 helical turns (16 bps) on the outer arms of each domain. They also feature a specific 4 bp sticky-end interaction between the tiles, as this interaction provides further robustness and rigidity to the construct. 45 First, we verified the clean formation of individual DX tile assemblies on the three template strands by annealing the component strands and templates from 95 to 4 °C over 4 h (non-denaturing AGE, Fig. S12, ESI-IX). † AFM images confirmed the formation of 1D structures with increased rigidity in comparison to double-stranded backbones and contour lengths that corresponded well to the expected values (130 ± 30 nm for DX-(AB) 5 , 102 ± 28 nm for DX-(EF) 4 , and 85 ± 24 nm for DX-(CD) 3 ) (ESI-IX). † We then assembled these DX structures onto the tripodal branched structure in its single-stranded form ssT3 (Fig. S11, ESI-VIII † ) by annealing it with all DX staple strands. Stepwise assembly of DX tiles on each arm ( Fig. 3A ) showed the formation of clean products in each case, and simultaneous assembly of DX tiles on the three arms of ssT3 was nearly quantitative, yielding a monodisperse product DX-T3. AFM images ( Fig. 3B ) revealed a higher rigidity of each individual arm in comparison to the double stranded version ( Fig. 2B ). The enhanced rigidity allows precise patterning of nanomaterials, such as nanoparticles or proteins with defined and constant separation, and the different sequences in each arm allow independent and selective hetero-patterning. Enhanced arm rigidity also favors the controlled folding of this large tweezer in response to an external stimulus. Fig. 3 Stepwise assembly of three-way DNA nanostructure, using elongated single-stranded “printed” junction (ssT3) as a scaffold, checked by native AGE (A). Lanes 1, 2 and 3 correspond to DX tile staples assembled on (AB) 5 , (CD) 3 and (EF)4, respectively. Lanes 4, 5, 6, 7 and 8 correspond to the assembly of DX tiles staples on (AB) 5 /(EF) 4 , (AB) 5 /(CD) 3 , (AB) 5 /(EF) 4 /(CD) 3 , core and (AB) 5 /(EF) 4 /(CD) 3 /core (DX-T3), respectively. Lane 9 corresponds to ssT3. GeneRuler DNA ladder mix is used. (B) AFM characterization of three-way DNA nanostructure in its DX-tile form (DX-T3). The population of fully assembled three-way DNA nanostructures calculated from AFM images is around 75% ( N = 190). This percentage includes the interconnected structures resulting from sample deposition on the mica surface in addition to structures with one- and two-arms. Contour lengths of DX-tiles assembled on ssT3 corresponded well to the expected values (141 ± 26 nm for DX-(AB) 5 , 104 ± 27 nm for DX-(EF) 4 , and 86 ± 15 nm for DX-(CD) 3 ) (ESI-IX). † Selective streptavidin patterning We demonstrated selective material organization on each arm of the elongated structure using biotin–streptavidin interactions. 46,47 Each individual arm consists of two alternating, different building blocks ( e.g. , A and B in (AB) 5 ), and the repeating blocks have different sequences on the three arms. This allows the selective functionalization of biotin molecules at the building block of interest. Building block C was first functionalized with biotin, resulting in 3 periodic biotin units on the DX-(CD) 3 arm. Upon addition of streptavidin, structures with exactly three proteins on the DX-(CD) 3 arm were observed by AFM imaging, with more than 75% yield ( Fig. 4A ). Similarly, AFM images revealed that we were able to selectively pattern four streptavidin on DX-(EF) 4 with a 77% yield ( Fig. 4B ). This approach can be further expanded to aperiodic and periodic patterning of proteins or nanomaterials within any elongated structure ( e.g . 3-arm, 4-arm or 5-arm molecules) and with any user-defined sequence and length. Fig. 4 AFM characterization of three-way DNA nanostructure patterned with periodic streptavidin proteins. Patterning of (A) DX-(CD) 3 arm and (B) DX-(EF) 4 arm. The bright dots in the AFM images correspond to the patterned streptavidin. The population of structures with exactly three (A) and four (B) streptavidin calculated from AFM images is 76% ( N = 186) and 77% ( N = 120), respectively. This percentage excludes the interconnected structures resulting from sample deposition on the mica surface. Streptavidin induced tweezer folding Finally, we explored our three-way DNA nanostructure as a nanotweezer capable of recognizing a multivalent protein. Streptavidin – as a tetrameric protein model – is capable of binding four different biotin molecules. We hypothesized that the addition of this multivalent protein would fold the 3 arms of DX-T3 into a 3D nanotube-like structure. DX tiles with biotinylated A, C, and E domains were thus assembled onto our tweezer scaffold and the entire construct was incubated with streptavidin. Biotinylated moieties were equidistant from the branching core to favor binding to the same streptavidin. AFM images revealed the formation of populations with no streptavidin bound, streptavidin bound on multiple arms or crosslinking different structures, and the product of interest which is the linear/tubular structure (ESI-XI. Fig. S16 † ). The yield of the latter population was low (around 10%). We hypothesized that the junction spacer of our structure might be too rigid and small to allow the efficient folding of the tweezer onto the protein units. We thus generated another asymmetric DNA-small molecule trimer where the tri-functionalized phenyl core was substituted with a more flexible tertiary amine core, and connected to the DNA arms via hexaethylene linkers to provide further flexibility (f-T3 for flexible T3, ESI-IV.B and ESI-XII † ). 48 This more flexible core played a major role in improving the bending degree of the arms in the DNA tweezer. As shown in Fig. 5A and B , the population of the tubular structures drastically increased from 10%, reported for the rigid core, to 85% ( N = 120) with the flexible core. In the current design, the first biotin moiety is located at an approximate distance of 27 nm (84 bps), while the flexible spacer only adds around 2 nm. The flexibility of the linker is thus most likely a major contributor to the enhancement in the tweezer folding, although we do not exclude the possibility of the additional spacer playing a role. We speculate that the lower limit for the distance between the biotin moiety and the branching core unit is 5 nm (the approximate size of streptavidin), but sterics may result in a higher value. To measure the core flexibility and estimate the degree of motion of the DNA arms as they close around the target proteins, we built an automated counting software that analyzes the angles between the arms from the AFM images of all the constructs (ESI-X † ). Since the initial, rigid small molecule core consists of 1,3,5-tris(azidomethyl)benzene, the ideal angle between each of the arms would be 120°, and deviation from this angle may be used as an indirect measure of core flexibility. This assumption is complicated by deposition and drying on the AFM substrate, but we reasoned that the histograms of the angles in the structures, compared to each other, would inform on the rigidity/flexibility of the core (ESI-XII † ). The flexible core resulted in a slightly larger standard deviation than the rigid core. Even though the standard deviation difference is low, the core's flexibility had a drastic effect on the efficiency of tweezer folding. Assuming a junction angle of ∼120° between the arms, the average distance between the ends of DX-AB and DX-EF arms is ∼213 nm, therefore the tweezer arms would have travelled a large distance of around 106 nm between the open and closed forms (ESI-XII). † Fig. 5 Schematic representation and AFM images of the tweezer folding. Folding of all arms, featuring periodic biotin moieties, in the presence of streptavidin into a tubular-like structure (schematic representation (A) and AFM images (B) (check ESI-XI, † for more AFM images)). Folding of DX-(CD) 3 and DX-(EF) 4 arms, featuring periodic biotin moieties on DX-(CD) 3 and DX-(EF) 4 , in the presence of streptavidin (schematic representation (C) and AFM images (D)). The population of fully folded structures calculated from AFM images is 85% ( N = 120) while that of the partially folded structure (DX-(CD) 3 and DX-(EF) 4 arms is 75% ( N = 90). This percentage excludes the interconnected structures resulting from sample deposition on the mica surface. To further investigate the folding of our DNA tweezer, assembled using the flexible core, we functionalized only building blocks C and E (and not A) with biotin, and incubated the construct with streptavidin. AFM revealed the periodic pattering of four streptavidin units, with merging of two arms and an unbound DX-AB arm ( Fig. 5C and D ). This indicates that our strategy can be applied to specifically fold any arm of the three-way DNA nanostructure depending on the position of biotin units and the needed application. On the other hand, and as detected by AFM, we have a small population of the partially folded structure, DX-(CD)3 and DX-(EF)4 arms with 2 streptavidin units, indicating that 2 streptavidin units may be enough to lock the tweezer in its folded form. The flexibility of the core, coupled with the rigidity of the arms are major determinants of proper tweezer folding, as they will bring biotin units on different arms in close enough proximity for streptavidin to bind to them. Once a single streptavidin is bound, the other biotin moieties will be close to each other and thus it may be easier for the next streptavidin to simultaneously bind to the other arms as the entropic cost is reduced. Despite the pre-organization afforded by the first binding event, folding the tweezer is entropically costly, and the strength of biotin–streptavidin binding enthalpically offsets this cost. Future protein–ligand pairs will require binding that overcomes this entropic cost. This can be controlled by increasing the number of binding sites to the protein. Examination of the height profile of the DNA segment located between two consecutive streptavidin molecules revealed an average height of 1.1 ± 0.2 nm for the structures having one streptavidin-templated arm ( Fig. 4 ). Similar DNA heights (1.2 ± 0.2 nm) were observed for branched structures having two arms DX-CD and DX-EF held together by streptavidin ( Fig. 5D ). In contrast, an average DNA height of 2.2 ± 0.4 nm was observed between two consecutive streptavidin units in the structures that have all three arms folded together ( Fig. 5B ). This 80% height increase in the latter case supports the folding of the three tweezer arms into a 3D-protein/DNA nanotube, where the streptavidin moieties are surrounded by DNA tiles. Our DNA arms thus fold around their protein targets, providing stimuli-responsive behavior that propagates large-scale motion and assembly reconfiguration. We envision that the appropriation of our technology to diverse and multiple multivalent proteins will proffer applications in sensing and biological diagnostics, wherein several identical or different proteins are recognized simultaneously to generate sensitive read-outs."
} | 5,712 |
28057830 | null | s2 | 3,649 | {
"abstract": "The concerted responses of eusocial insects to environmental stimuli are often referred to as collective cognition at the level of the colony. To achieve collective cognition, a group can draw on two different sources: individual cognition and the connectivity between individuals. Computation in neural networks, for example, is attributed more to sophisticated communication schemes than to the complexity of individual neurons. The case of social insects, however, can be expected to differ. This is because individual insects are cognitively capable units that are often able to process information that is directly relevant at the level of the colony. Furthermore, involved communication patterns seem difficult to implement in a group of insects as they lack a clear network structure. This review discusses links between the cognition of an individual insect and that of the colony. We provide examples for collective cognition whose sources span the full spectrum between amplification of individual insect cognition and emergent group-level processes."
} | 265 |
36370454 | PMC9923381 | pmc | 3,652 | {
"abstract": "Abstract 2-Phenylethanol (2- PE) is an aromatic alcohol with wide applications, but there is still no efficient microbial cell factory for 2-PE based on Escherichia coli . In this study, we constructed a metabolically engineered E. coli capable of de novo synthesis of 2-PE from glucose. Firstly, the heterologous styrene-derived and Ehrlich pathways were individually constructed in an L-Phe producer. The results showed that the Ehrlich pathway was better suited to the host than the styrene-derived pathway, resulting in a higher 2-PE titer of ∼0.76 ± 0.02 g/L after 72 h of shake flask fermentation. Furthermore, the phenylacetic acid synthase encoded by feaB was deleted to decrease the consumption of 2-phenylacetaldehyde, and the 2-PE titer increased to 1.75 ± 0.08 g/L. As phosphoenolpyruvate (PEP) is an important precursor for L-Phe synthesis, both the crr and pykF genes were knocked out, leading to ∼35% increase of the 2-PE titer, which reached 2.36 ± 0.06 g/L. Finally, a plasmid-free engineered strain was constructed based on the Ehrlich pathway by integrating multiple ARO10 cassettes (encoding phenylpyruvate decarboxylases) and overexpressing the yjgB gene. The engineered strain produced 2.28 ± 0.20 g/L of 2-PE with a yield of 0.076 g/g glucose and productivity of 0.048 g/L/h. To our best knowledge, this is the highest titer and productivity ever reported for the de novo synthesis of 2-PE in E. coli . In a 5-L fermenter, the 2-PE titer reached 2.15 g/L after 32 h of fermentation, suggesting that the strain has the potential to efficiently produce higher 2-PE titers following further fermentation optimization.",
"introduction": "Introduction Due to its pleasant rose-like odor and antibacterial properties, 2-phenylethanol (2-PE) is a high-value compound widely used in the cosmetic, food, and flavor industries (Zhan et al., 2022 ). Various wild-type yeasts, including Saccharomyces cerevisiae, Kluyveromyces marxianus , and Y arrowia lipolytica have the capacity of de novo synthesis of 2-PE (Etschmann et al., 2002 ; Wang et al., 2019 ; Zhang et al., 2014 ). Three biosynthesis pathways for 2-PE production have been reported to date, including the Ehrlich pathway (Gu et al., 2020 ), styrene-derived pathway (Machas et al., 2017 ) and phenylethylamine pathway (Masuo et al., 2015 ). Since these pathways start from L-phenylalanine (L-Phe) or phenylpyruvate (PPA) from the aromatic amino acid biosynthesis pathway, which is subject to strict feedback regulation, the carbon flux from the central metabolic pool toward 2-PE is quite weak and supports only comparatively low 2-PE titers. Therefore, high 2-PE titers are usually achieved by the biotransformation of externally added L-Phe, which is a more effective approach but with a high cost. As L-Phe is a relatively costly feedstock, direct 2-PE production from cheaper, renewable biomass sugars represents an attractive alternative to L-Phe conversion. With recent advances in metabolic engineering, 2-PE produced by microbial systems is a potentially sustainable alternative to fossil fuel-based production and isolation from native plants, which has attracted considerable interest (Kong et al., 2020 ; Li et al., 2021 ; Zhan et al., 2022 ; Zhan et al., 2020 ). In consideration of its rapid growth, clear genetic background and highly efficient genome editing, E. coli has been used as a versatile platform strain to produce various biochemicals by applying genetic engineering methods (Guo et al., 2020 ; Guo et al., 2022 ; Lai et al., 2022 ; Sheng et al., 2021 ; Zhang et al., 2022 ), including 2-PE (Guo et al., 2017 ; Kang et al., 2014 ; Wang et al., 2019 ). Although various efforts have been made to improve the synthetic ability of E. coli , the titer, yield and productivity were still less than satisfactory. In this study, we engineered E. coli W3110 for increased flux along the Ehrlich pathway and achieved the highest titer of 2-PE from glucose reported to date.\n\nIntroduction of a heterologous styrene-derived pathway for 2-PE biosynthesis Wu et al . reported a multienzyme cascade biocatalysis strategy for converting styrene into 2-PE with a particularly high conversion (>99%) by using styrene monooxygenase (to convert styrene to S-styrene oxide), styrene oxide isomerase (to convert styrene oxide to phenylacetaldehyde) and phenylacetaldehyde reductase (to convert phenylacetaldehyde to 2-PE) (Wu et al., 2017 ). A 2-PE biosynthesis pathway, derived from styrene, was constructed based on the extension of the styrene biosynthesis pathway. This process includes five steps starting from endogenous L-Phe (Machas et al., 2017 ). The thermodynamic driving force of the styrene-derived pathway was approximately 10 times greater than that of the Ehrlich pathway, and it was considered a better 2-PE biosynthesis route with high stability and efficiency in E. coli (Machas et al., 2017 ; Wu et al., 2017 ). This pathway is encoded by four heterologous genes, including PAL2 (from Arabidopsis thaliana , encoding phenylalanine ammonia-lyase), FDC1 (from Saccharomyces cerevisiae , encoding phenylacrylic acid decarboxylase), and styAB (from Pseudomonas putida S12, encoding styrene monooxygenase). It converts L-Phe to (S)-styrene oxide, after which styC -encoded styrene oxide isomerase (from Pseudomonas putida S12) converts (S)-phenylene oxide to phenylacetaldehyde. Finally, native NADPH-dependent alcohol dehydrogenases reduce phenylacetaldehyde into 2-PE (Fig. 1 ). Therefore, we attempted to assess the styrene-derived pathway in E. coli M4. Four heterologous genes PAL2 (GenBank: AAM12956.1 ), FDC1 (GenBank: DAA12368.1), styAB (GenBank: AJA17113.1 and AJA17114.1) and styC (GenBank: AJA17115.1) were codon-optimized for E. coli and synthesized. The corresponding coding sequences were placed under the control of the strong P T7 promoter and integrated into the chromosome of E. coli M4, resulting in the strains E. coli PE1 ( E. coli M4, gapC:: P T7 -PAL2-FDC1 ) and E. coli PE2 ( E. coli M4, gapC:: P T7 -PAL2-FDC1, yeeP:: P T7 -styAB-styC ). Compared with the parent strain, the L-Phe titer of E. coli PE2 decreased in shake flask after 72 h of fermentation, from 3.12 ± 0.18 g/L to 2.18 ± 0.22 g/L ( Supplementary information, Figure S1 ). Surprisingly, only trace amounts of 2-PE were detected in PE2. RT-PCR results confirmed that all four genes were successfully overexpressed in the cells ( Supplementary information, Figure S2 ). Taken together, the results indicate that the enzyme activity obtained by placing the gene under the control of a strong promoter was unsatisfactory, probably due to abnormal folding caused by excessively fast transcription.\n\nIntroduction of a heterologous ehrlich pathway for 2-PE biosynthesis We attempted to introduction the Ehrlich pathway, which was relatively simple. The Ehrlich pathway consists of three reactions. A transaminase first converts L-Phe into phenylpyruvate, which is subsequently converted into phenylethylaldehyde and 2-PE by phenylpyruvate decarboxylase and alcohol dehydrogenase, respectively (Qian et al., 2019 ). The decarboxylation reactions can be catalyzed by many enzymes, such as those encoded by pcd, kind, ipdC and ARO10 (Wang et al., 2019 ). We employed ARO10 from S. cerevisiae S288c, which encodes a phenylpyruvate decarboxylase with relatively high activity (Yin et al., 2015 ). In the beginning, three plasmids with different copy number were selected as backbone (high copy number pTrc99a, medium copy number pSTV28 and low copy number pWSK29), while the original ARO10 and optimized ARO10 * sequences were cloned by POE-PCR ( Supplementary information, Figure S3 ). The plasmids were then introduced into E. coli M4 resulting in E. coli PE3 ( E. coli M4; pTrc99a- ARO10 ), PE4 ( E. coli M4; pSTV28- ARO10 ), PE5 ( E. coli M4; pWSK29- ARO10 ) and PE6 ( E. coli M4; pTrc99a- ARO10 *). As shown in Fig. 2 , E. coli PE3 with the high copy number plasmid produced the highest 2-PE titer of 0.76 ± 0.02 g/L after 72 h of fermentation. This suggested that strong overexpression of phenylpyruvate decarboxylase plays a crucial role in 2-PE production. By contrast, the 2-PE titer decreased significantly in E. coli PE6 (much lower than PE3), we speculated that the optimized gene might lead to mRNA instability or abnormal folding, which finally resulted in low enzymatic activity of phenylpyruvate decarboxylate. Consequently, the engineered strain E. coli PE3 was chosen as the starting strain for further optimization. Fig. 2 Initial assessment of using the Ehrlich pathway to produce 2-PE from glucose.\n\nImproving 2-PE tolerance by introducing the LexA_E45I mutation and adaptive laboratory evolution Considering that the 2-PE titer of E. coli PE8 and E. coli PE9 reached over 2.0 g/L, we were highly aware of the problem of 2-PE toxicity to E. coli , which greatly limited the 2-PE titer. Liang et al. applied an iCREATE strategy to design, construct, and test a library of transcriptional regulators targeting 54 genes with 85 420 mutations for increased styrene resistance and production in E. coli . The best mutant, ST05 LexA_E45I, not only exhibited improved styrene tolerance but also produced a 3.45-fold higher styrene concentration than the parental strain (Liang et al., 2020 ). Therefore, we attempted to introduce the LexA_E45I mutation into E. coli PE3 to improve its 2-PE tolerance. The experimental results indicated that the LexA_E45I mutation improved the 2-PE tolerance to a certain extent, but the strain nearly completely lost its ability to produce 2-PE (data not shown). This finding suggests that the 2-PE tolerance mechanism is different from that of styrene tolerance. In addition, we used the classical tool of adaptive laboratory evolution to improve the 2-PE tolerance. Using E. coli PE8 as the starting strain, 90 days of continuous culture were carried out with increasing 2-PE concentrations (0.5, 1.0, 1.5 to 2.0 g/L, and each concentration was applied for 30 days). It is important to note that the evolved strain grew well in the presence of 2.0 g/L of 2-PE, but still almost completely loses its capacity for 2-PE synthesis. These results indicate that 2-PE tolerance and 2-PE biosynthesis are conflicting evolutionary objectives. As a consequence of our failed attempts, improving the 2-PE tolerance of E. coli by adaptive laboratory evolution still demands more elaborate directed evolution strategies, such as adding both 2-PE and p-fluorophenylalanine (structural analog of L-Phe) to the medium, which might improve its 2-PE biosynthesis ability while obtaining higher 2-PE tolerance. However, this hypothetical strategy needs to be further tested in future studies.",
"discussion": "Discussion The increasing demand for 2-PE has inspired great interest in its biotechnological production. Many studies have successfully implemented the synthesis of 2-PE in microbial strains expressing a heterologous Erich pathway, phenylacetaldehyde synthesis pathway, or styrene derivative pathway (Table 2 ). In this study, an Ehrlich pathway was successfully introduced into E. coli to de novo synthesize 2-PE from glucose and the titer was further improved by metabolic engineering. The plasmid-free chromosomally engineered strain E. coli PE12 was able to produce 2-PE from glucose and the culture process no longer requires antibiotics. In addition, a 5-L laboratory-scale fermentation of the best strain was preliminarily explored. Under 5-L fermentation conditions, 2-PE biosynthesis started already at 4 h, and the titer continuously increased in the first half of the fermentation process, reaching a maximum of 2.15 g/L at 32 h. Notably, this is the highest titer reported for the de novo synthesis of 2-PE by E. coli to date. Table 2. 2-PE production in rationally engineered strains using glucose as substrate Microorganism Characteristics Carbon source (g/L) Titer (g/L) Cultivation mode Reference \n Bacillus licheniformis PE23 DWc9n- xkdG::yugJ, Δpyk ; pHY- kivD-aroD * glucose (20) 6.24 Shake flask (Zhan et al., 2022 ) \n Pichia pastoris SK004 \n ARO10; ADH6; ARO8; aroG fbr ; pheA fbr \n glucose (20) 1.17 Shake flask (Kong et al., 2020 ) \n Kluyveromyces marxianus CBS 6556 glucose (20) 1.94 Shake flask (Li et al., 2021 ) \n K. marxianus BY25569 \n aro10; adh2; aroG fbr \n glucose (20) 1.3 Shake flask (Kim et al., 2014 ) \n Yarrowia lipolytica po1fk7P \n po1fk4 − 2 ΔylDGA2 ΔylDG-A1::loxP harboring pYLXP′-citB-ylIDP2-ylODC \n glucose (40) 2.67 Shake flask (Gu et al., 2020 ) \n Saccharomyces cerevisiae BY4741 \n aro10; ∆aro8 \n glucose (20) 0.10 Shake flask (Shen et al., 2016 ) \n S. cerevisiae BY4741 \n aro4 fbr ; aro7 fbr ; ∆aro8; ∆tyr1; ∆aro3 \n glucose (20) 0.41 Shake flask (Romagnoli et al., 2015 ) \n Escherichia coli NST74 \n PAL2; FDC1; styAB; styC; ∆feaB; ∆pykA; ∆pykF; ∆crr \n glucose (50) 1.94 Shake flask (Machas et al., 2017 ) \n E. coli DG02 \n aroG fbr ; pheA fbr ; kdc; yjgB; aro8 \n glucose (20) 1.02 Shake flask (Guo et al., 2017 ) \n E. coli MG1655 \n aro8; kdc; yjgB \n glucose (20) 0.18 Shake flask (Guo et al., 2017 ) \n Enterobacter sp. CGMCC 5087 \n pheA fbr ; DAHP \n glucose (20) 0.34 Shake flask (Zhang et al., 2014 ) \n E. coli BW25113(DE3) \n ipdC; yahK; aroF fbr ; pheA fbr ; ΔfeaB \n glucose (10) 0.94 Shake flask (Koma et al., 2012 ) As we summarized in Table 2 , many studies have successfully implemented the synthesis of 2-PE in microbial strains. However, microbial production of heterologous organic compounds is challenging, as the biosynthetic pathways are often complex and produce metabolites that are toxic to the host. Although 2-PE production has been significantly improved through diverse engineering strategies, industrial production of 2-PE is still a challenge. In fact, compared with highly tolerant hosts (for example, Y. lipolytica or B. licheniformis ), a serious disadvantage for 2-PE bioproduction in E. coli is its low tolerance to 2-PE, which greatly limited the 2-PE titer so far. New strategies to overcome the toxicity of 2-PE are expected to significantly advance the 2-PE titer and would further develop E. coli into a suitable platform for 2-PE biosynthesis. However, additional research in this direction is still needed in the future. This limitation has been partially addressed by strategies such as in situ product recovery (ISPR) (Hua & Xu, 2011 ; Wang et al., 2019 ). In terms of process engineering, in-situ separation technology is an effective means to improve the biosynthesis of 2-PE. Studies have applied this technology to continuously separate 2-PE from the fermentation broth, so that cells can grow normally and thereby increase the yield of 2-PE. Other approaches, such as pervaporation and solid-phase extraction (i.e. using hydrophobic resins) have resulted in up to 10-fold improvements of 2-PE production (Achmon et al., 2011 ). When S. cerevisiae Giv2009 was used to biosynthesize 2-PE, oleic acid was added to the fermentation broth, and the final yield of 2-PE reached 12.6 g/L in fed-batch fermentation (Stark et al., 2002 ). Hua et al. used PPG1500 as extractant to conduct supplementary fermentation with S. cerevisiae . The concentration of 2-PE in PPG1500 reached 22.0 g/L, and the total output was 7.5 g/L (Hua et al., 2013 ). Chreptowicz et al. reported that the yield of 2-PE extracted from the fermentation broth of S. cerevisiae with rapeseed oil was 9.79 g/L (Chreptowicz & Mierzejewska, 2018 ). While these methods considerably increased the titer of 2-PE, they result in higher production costs. Accordingly, future research studies should focus on improving the 2-PE titer by increasing its tolerance of E. coli to 2-PE."
} | 3,946 |
38818482 | PMC10989833 | pmc | 3,654 | {
"abstract": "Abstract Although the accomplishments of microbiome engineering highlight its significance for the targeted manipulation of microbial communities, knowledge and technical gaps still limit the applications of microbiome engineering in biotechnology, especially for environmental use. Addressing the environmental challenges of refractory pollutants and fluctuating environmental conditions requires an adequate understanding of the theoretical achievements and practical applications of microbiome engineering. Here, we review recent cutting‐edge studies on microbiome engineering strategies and their classical applications in bioremediation. Moreover, a framework is summarized for combining both top‐down and bottom‐up approaches in microbiome engineering toward improved applications. A strategy to engineer microbiomes for environmental use, which avoids the build‐up of toxic intermediates that pose a risk to human health, is suggested. We anticipate that the highlighted framework and strategy will be beneficial for engineering microbiomes to address difficult environmental challenges such as degrading multiple refractory pollutants and sustain the performance of engineered microbiomes in situ with indigenous microorganisms under fluctuating conditions.",
"conclusion": "4 CONCLUDING REMARKS In summary, researchers can rationally design an efficient, stable, predictable, and safe microbiome through microbiome engineering. With present mathematical models, high‐throughput culturing techniques, and quantitative analysis tools, the accuracy of predicting the structure and function of a microbial community has recently been improved. In the future, we should investigate how to rationally regulate microbiomes or de novo construct microbiomes to achieve desired functions through environmental factors, microbial interactions, and physical characteristics of involved strains. To this end, we need to test the feasibility of recently proposed principles of microbiome engineering under more complex conditions (e.g., in situ environments) and integrate different useful strategies. Benefiting from these new tests, novel quantitative principles can be proposed to increase our understanding of microbial ecology and guide future microbiome engineering. We expect that the combined framework of top‐down and bottom‐up approaches can be applied to address the enormous environmental challenges for bioremediation (Table 4 ). Table 4 Accomplishments and perspectives of microbiome engineering strategies. Approach Strategy What we have done What we can do Bottom‐up engineering With naturally occurring strains Successfully applied for construction of the microbiome to degrade specific compounds in the lab and practical bioremediation Maintaining long‐term stability and coexistence of the members With engineered strains Synthetic consortia were constructed by the DOL approach, such as dividing the pollutant mineralization pathway into different strains Maintaining long‐term stability and enhancing metabolite‐exchange efficiency. Mathematical models proposed by Lingchong You et al. can help guide the engineering of metabolic pathways by the DOL approach \n 54 \n . Developing population control techniques for GMO restriction Top‐down engineering Enrichment Successfully applied for construction of microbiomes to degrade specific compounds in the lab and practical bioremediation Optimization of imposed perturbations, including various environmental conditions, and dilution factors guiding the top‐down engineering process with the ecological models and mathematical models, such as bottleneck, resource shift, and species knock‐in \n 107 \n . If possible, using artificial selection and directed evolution before enrichment Artificial selection After dozens of selection cycles, the late generations are more efficient than the seeding microbiome Directed evolution Only in theory John Wiley & Sons, Ltd.",
"introduction": "1 INTRODUCTION Microorganisms live in communities and interact with their neighbors and environments. They drive global biogeochemical cycles, significantly change our living environments, and impact human health \n 1 \n , \n 2 \n , \n 3 \n . For example, approximately 60% of global wastewater is treated by microbial consortia involved in active sludge before being released into the natural water systems \n 4 \n , \n 5 \n . Inside the human body, the number of microbial cells is 10 times higher than that of human cells, and human microbiomes have fundamental roles in physiology and health \n 6 \n . Moreover, breakthroughs are constantly appearing in the bio‐industry because of the capabilities of microbiomes to synthesize valuable products \n 7 \n , \n 8 \n , degrade chemical pollutants \n 9 \n , \n 10 \n , and produce biofuels \n 11 \n . Currently, a large number of important industrial chemicals and medicines have been produced by microbiomes, including isobutanol \n 7 \n , taxanes \n 8 \n , and hydrogen \n 11 \n . Although these achievements have affirmed the enormous potential of microbiomes for human use, formidable challenges remain in manipulating microbiomes for controllable output. Specifically, the ability to control the microbiome structure to sustain its function is still lacking. To address this challenge, a concept called “microbiome engineering” has emerged recently \n 12 \n , \n 13 \n , \n 14 \n . “Microbiome engineering” is a process to enhance the performance of microbiomes through targeted manipulation of the composition of natural communities (top‐down) or rational design and construction of new synthetic consortia (bottom‐up). Microbiome engineering is typically driven by general principles derived from a mechanistic understanding of the ecology and evolution of microbiomes. Such principles are presented as quantitative frameworks that can accurately predict the dynamics and function of a given microbiome and thus guide the rational engineering of microbiomes. This process of microbiome engineering provides a significant opportunity to further unlock the large potential of microbial communities. Recently, an iterative “design‐build‐test‐learn” cycle (DBTL) was proposed as a general guideline for microbiome engineering \n 13 \n . Nevertheless, the development of microbiome engineering is still limited by knowledge and technical gaps. The major hurdle is the lack of quantitative theories and techniques to accurately measure, predict, and manipulate the structure and functions of microbiomes. Moreover, the interactions between many naturally occurring microbes are uncharacterized, and how these interactions are regulated by multiple environmental factors also remains poorly understood. Furthermore, tools to directly manipulate specific members of the microbial community have yet to be explored. In addition to these gaps, applying microbiomes to treat chemical pollutants in an open environment faces more challenges than applying microbiomes in a closed bioreactor. For example, degradation of multiple refractory pollutants (e.g., polycyclic aromatic hydrocarbons with high molecular weight, plastics, and halogenated compounds) in an open environment requires an engineered microbiome exhibiting diverse and high degrading abilities. In addition, fluctuating environmental conditions and indigenous microorganisms at contaminated sites affect the stability of engineered microbiomes. Inadequate tools for real‐time monitoring in the natural environment and policy restrictions on the use of genetically modified organisms (GMOs) limit the development of microbiome engineering for environmental use. Integrating the theoretical research on microbiome engineering and its application achievements in recent years would advance the development of microbiome engineering in environmental science. Here, a set of microbiome‐engineering strategies is summarized from recent cutting‐edge studies. We then discuss how these strategies can be applied to advance the development of microbiomes for bioremediation. Based on present theories and practical accomplishments, we summarize and propose an approach that combines bottom‐up and top‐down approaches for the applications of microbiome engineering in natural environments. Finally, an “avoidance of toxic intermediates” strategy is discussed, which specifically addresses the environmental applications of microbiome engineering to avoid the build‐up of toxic intermediates that pose a risk to human health and to the environment."
} | 2,105 |
31366907 | PMC6668475 | pmc | 3,655 | {
"abstract": "The ecological importance of common species for many ecosystem processes and functions is unquestionably due to their high abundance. Yet, the importance of rare species is much less understood. Here we take a theoretical approach, exposing dynamical models of ecological networks to small perturbations, to explore the dynamical importance of rare and common species. We find that both species types contribute to the recovery of communities following generic perturbations (i.e. perturbations affecting all species). Yet, when perturbations are selective (i.e. affects only one species), perturbations to rare species have the most pronounced effect on community stability. We show that this is due to the strong indirect effects induced by perturbations to rare species. Because indirect effects typically set in at longer timescales, our results indicate that the importance of rare species may be easily overlooked and thus underrated. Hence, our study provides a potential ecological motive for the management and protection of rare species.",
"introduction": "Introduction The ecological importance of common species is relatively well documented 1 – 3 . Yet, the ecological and functional role of rare species, which make up the overwhelming majority of species in the world’s ecosystems 4 , is less well known 5 . There are, however, studies clearly demonstrating the keystone role played by some rare species 6 – 8 and their significant contribution to the functional diversity of ecosystems 9 – 12 . Yet, the relative importance of rare versus common species for the stability of ecosystems in general is largely an open question. Theoretically, the relative ecological role of rare and common species has received little attention. So far, however, in dynamical models of ecological networks, i.e. networks describing who interacts with whom in an ecosystem, enforced extinctions of common species has been shown to lead to larger community wide effects than extinctions of rare species 13 , 14 . It has also been found that functional extinctions – where an increased mortality rate and subsequently reduced abundance of a given species first leads to an extinction of another species in the community rather than the species itself – are more likely to occur if the perturbation affects a ‘high biomass’ (i.e. common) species compared to a ‘low biomass’ (i.e. rare) species 15 . Similarly, harvesting a species at maximum sustainable yield has been shown to lead to larger community wide effects if harvesting is applied to common compared to rare species 16 . All these studies indicate that when large and sustained selective perturbations affect common species it will lead to larger community wide effects than if such perturbations affect rare species. However, the methods used in these studies are not directly suitable for evaluating the relative dynamical importance of rare and common species, since these type of perturbations may be considered large from a community point of view when applied to some species, and small when applied to others. For example, selective removal of a high biomass species may be considered a much larger perturbation to a community than removal of a low biomass species. Thus, thoroughly comparing the relative dynamical importance of rare and common species in an unbiased way requires a more sophisticated approach. To this end, an alternative is to investigate how small perturbations, assumed to be of equal magnitude across species, affects the stability properties of a community. Analytical methods for addressing the response to such perturbations have been derived 17 – 20 , yet the relative effect of perturbing the density of rare and common species has not been thoroughly investigated [but see 20 ]. Here we use a theoretical approach, combining analytical and numerical methods, to investigate the relative importance of rare and common species for the stability properties of ecological network models. First, we investigate the relative role of rare and common species in governing the rate of return to equilibrium following generic pulse perturbations; that is, temporary perturbations simultaneously affecting all species in a community. Second, we investigate the recovery of ecological networks exposed to selective pulse perturbations (i.e. temporary perturbations that affect one species at a time) to rare and common species, respectively. Third and finally, we investigate community wide responses of small selective press perturbations, that is, permanent perturbations, to rare and common species. Overall, our analytical and numerical modeling results suggest that rare species may be of particular importance for the stability of ecological communities as communities recover more slowly following pulse perturbations to rare than common species, and because stronger indirect effects are induced if rare rather than common species are permanently perturbed.",
"discussion": "Discussion The dynamical importance of common species, which make up the core of ecological communities, is unquestionably due to their high abundances 1 – 3 . Yet, the relative dynamical importance of common vs. rare species is less well understood 2 . Theoretically, we here investigated this fundamental ecological question, and found that the importance of rare species may easily be overlooked and thus underrated. Small selective perturbations (assumed to be of equal magnitude across species) to rare species generally led to larger dynamical effects than similar perturbations to common species. This contrasts with previous studies showing that large perturbations, such as species extinctions, led to larger dynamical effects when common species, rather than rare species, were eradicated 13 – 16 . These apparently conflicting results illustrate that the specific type of perturbation being imposed will affect which specific species that are identified as major contributors to ecological stability 23 . Moreover, the identification of keystones in a community also depends on the specific stability metric being studied 24 , 25 . Here we used the rate of return to locally stable fixed points as a measure of ecological stability. This definition of stability assumes a static equilibrium point, while empirically based stability metrics are often based on some form of population or community variability 24 , 25 . Thus, in order to relate our theoretical results to empirical results based on the latter view, it needs to be explored if and how our theoretical stability metrics can be related to population variability. Although this objective warrants future research focus, it is worth noting that there is a direct relationship between the dominating eigenvalue of interaction matrices and population variability in stochastic linear discrete models 26 . This points to the possibility that our results, based on deterministic models, may also hold information for other empirically based stability measures. We have further assumed that species trophic interactions can be described by linear functional responses. The extent to which our results are generalizable to systems with non-linear functional responses therefore remains to be explored. However, it is first worth noting that several theoretical studies have compared model results based on systems with linear and non-linear functional responses, and found that many aspects (such as the probability of observing pyramidal patterns of abundance 27 , the risk of secondary extinctions 15 , 28 , 29 , and the probability of local stability of ecological networks 30 ) were qualitatively similar for the two scenarios. Secondly, whether linear or non-linear functional forms is a more realistic model of species interactions, in the range of species densities observed in multispecies systems, is an open question 31 – 33 . Still, to validate if our results holds also under non-linear functional responses, linearization using numerical methods is straight forward 34 , 35 , and the methods used here could thus easily be adopted to other non-linear systems in future work. Another important aspect that should be acknowledged is the fact that equilibrium densities are functions of the interaction strengths in a model. One might thus argue that the findings presented here, that is, results illustrating the relative importance of rare vs. common species for the stability of ecological communities, may be confounded by a structured pattern of species interaction strengths (something that may emerge from body size scaling of parameters). Still, although we observe some weak relationships between species interaction strength and equilibrium biomass in the model realizations of the empirical networks (Supplementary Figs 8 – 10 ), these relationships are of variable signs for the different networks (positive relationships for some systems and negative for others), indicating that it is not differences in species interaction strengths that are driving the patterns we see in the model food webs. Moreover, for both the bipartite networks, where there are no relationships between equilibrium densities and interaction strengths (Supplementary Fig. 11 ), and the analytical results, which are to a large extent independent of the specific parametrization used, we found similar results with respect to the relative importance of rare vs. common species. Altogether, we thus think that the results presented here are fairly robust to the specific model parameterization being applied. Our contribution partly builds on the elegant and informative tools developed by Arnoldi et al . 20 to study short- to long-term recovery of communities following pulse perturbations. Although the major contribution by Arnoldi et al . 20 was to develop tools for studying recovery dynamics of communities following generic pulse perturbations, they also began to explore the relative role of abundant and less abundant species for ecosystem recovery. To this end, they derived an analytical result showing that the rare species in a two species community (where a negative unidirectional effect of a common species on a rare species is assumed) governs asymptotic recovery rate, and found numerically that the rarest species, in a system with competitive interactions only, governs asymptotic recovery rate. In contrast, our study disentangles the relative role of rare and common species for ecological stability at short- to long-time scales, in arbitrarily complex systems (i.e. specious systems with any interaction types), following generic pulse perturbations, selective pulse perturbations as well as selective press perturbations. Using a broad set of analytical as well as numerical methods we have developed the theoretical understanding of the potential relative role of rare and common species for ecological stability by significantly extending the findings of Arnoldi et al . 20 , specifically showing that: (i) the long term recovery rate of arbitrarily complex generalized Lotka-Volterra model systems (i.e. with any combination of interaction types) that are close to an instability is governed by rare species and that the contribution of rare species to the recovery process can be significant also at shorter time scales; (ii) press perturbations to rare species lead to an overall larger effect on ecological stability than what perturbations to common species do; (iii) communities are more sensitive to selective pulse perturbations to rare than common species since communities recover more slowly when rare rather than common species are perturbed. One might argue that our result, showing that communities recover slowly following selective pulse perturbations to rare species, only depends on the fact that rare species approaches equilibrium slowly. By extension, this might be interpreted as if perturbations to rare species only have negligible impact on community dynamics, and only at very long timescales as the rare species recover. Yet, in general we found this not to be true. In fact, (i) rare species have a strong effect on return rates already at intermediate time scales and (ii) stronger indirect effects (i.e. effects on other species mediated via chains of interactions) appear if rare species are perturbed compared to if common species are perturbed (See also Supplementary Information Section 7). Since indirect effects typically set in at intermediate to long timescales 36 , 37 , our results indicate that the effect of perturbations to rare species may be hidden in ecological dynamics (i.e. not immediately evident and manifested only when indirect effects unfold as time progresses), but the response of rare species perturbations may nevertheless be strong. Therefore, our study indicates that rare species may be just as ecologically important as common species, but the dynamical effects caused by perturbations to rare species may be hard to detect as well as difficult to predict (due to the indeterminacy of ecological interactions involving indirect effects 19 ). Thus, our results provides a potential ecological motive for the protection and management of rare species."
} | 3,282 |
33005314 | PMC7516209 | pmc | 3,658 | {
"abstract": "Termite mounds are fascinating because of their intriguing composition of numerous geometric shapes and materials. However, little is known about these structures, or of their functionalities. Most research has been on the basic composition of mounds compared with surrounding soils. There has been some targeted research on the thermoregulation and ventilation of the mounds of a few species of fungi-growing termites, which has generated considerable interest from human architecture. Otherwise, research on termite mounds has been scattered, with little work on their explicit properties. This review is focused on how termites design and build functional structures as nest, nursery and food storage; for thermoregulation and climatisation; as defence, shelter and refuge; as a foraging tool or building material; and for colony communication, either as in indirect communication (stigmergy) or as an information channel essential for direct communication through vibrations (biotremology). Our analysis shows that systematic research is required to study the properties of these structures such as porosity and material composition. High resolution computer tomography in combination with nonlinear dynamics and methods from computational intelligence may provide breakthroughs in unveiling the secrets of termite behaviour and their mounds. In particular, the examination of dynamic and wave propagation properties of termite-built structures in combination with a detailed signal analysis of termite activities is required to better understand the interplay between termites and their nest as superorganism. How termite structures serve as defence in the form of disguising acoustic and vibration signals from detection by predators, and what role local and global vibration synchronisation plays for building are open questions that need to be addressed to provide insights into how termites utilise materials to thrive in a world of predators and competitors.",
"introduction": "1 Introduction to termitology Termites are eusocial cockroaches [1] , many of which eat wood and show cryptic behaviours making them difficult to be detected. Consequently, termites have gained the reputation of notorious pests with an all-consuming appetite [2] . The reality is quite different: just 97 of more than 3,100 known species are considered to be economically relevant [3] , [4] , [5] , with most species providing important ecosystem functions and are considered to be ecosystem engineers [6] , [7] , [8] . Due to their sociality, their inter-dependency, their ability to communicate and their strict organisation, termite colonies are referred to as superorganisms [9] , [10] , [11] , [12] . Apart from having highly specialised direct communication based on vibrational information and pheromones [13] , [14] , termites build a variety of complex structures (underground nests, soil protruding mounds and nests high up on trees, cf. Fig. 1 ) as a product of cooperation [15] – presumably following simple sets of rules to produce a large diversity of shapes [16] through parameter tuning [17] . Fig. 1 Termite nests of (a) Coptotermes acinaciformis with commensurate termite species Macrognathotermes sunteri attached to its host mound; Berrimah, Northern Territories (photo credits: Sebastian Oberst, 2011); (b) Amitermes meridionalis , the ”magnetic” or ”compass” termite, mound-building, hypogeal species [3] (Arnheim, Northern Territories, Australia CSIRO picture collection; photo credits: Coppi, 1992), and (c) a tree-nest of Nasutitermes walkeri , arboreal, higher termites (dead-wood feeding [18] ), Warrambungle National Park, New South Wales, Australia (photo credits: Sebastian Oberst, 2018). Inserts show (b) a hard, outer shell and (c) a filigree inner structure. In 1954 Grassé [19] conceived that coordination during construction and excavation is achieved using stimulating patterns of matter for different regulatory responses including either (1) stigmergic stimuli, (2) responses to the environment or (3) nestmate interaction, factors which have largely been confirmed in research since then [20] , [21] , [22] . Small structures are designed based on the insect’s body size [23] while larger structures are built through collective interactions [15] . The environment, the state of the colony and the shape of the nest as well as the individual termite (caste, age, experience) determine individual and collective behaviours [24] . However, very little is known how these factors interact to affect mound size and variability, functional properties of different parts of the mound and among species and other probably important details, such as tunnel diameter and chamber size. 1 The complexity, utility and potential sustainability of biological morphogenesis [25] , especially nest construction, has inspired concepts of eco-friendly architectural designs [26] , [27] , [28] , [29] and ideas of generating sustainable biocemented materials [30] , [31] . Termite-built structures demonstrate how to protect the colony within a ’breathing’ shelter [32] ; how fluctuations of intensive environmental parameters could be used to passively climatise architecture ( homeostasis ) [32] , [33] of highest strength [19] , [34] to generate all-year-round ideal living conditions [15] , [35] . Noirot and Darlington [36] review termite nest architecture, climate regulation and defence, while Korb [37] studies similarly termite mound architecture, its function and construction with a focus on functional shapes of selected (mostly African) species. Consequently, past research was mostly concerned with autonomous nest constructions or building activities, the network structure of tunnels, and aspects of stigmergy and self-organisation [15] , [28] , [38] , [39] , [40] , [41] , [42] , [43] . Stigmergy , hereby defined as indirect communication [44] to exchange information through modification of the environment \n [40] , is a prerequisite of self-organisation and spontaneous order generation through local interactions of a seemingly erratic system [15] , [45] . Swarm behaviour or self-organisation is part of autonomous systems research with its emerging domain of swarm robotics and artificial intelligence [40] , [42] , [44] , [46] . In contrast, direct communication is provided by optical, pheromone, tactile and vibrational information [47] , [48] , [49] . Especially vibrational information (biotremology) has been a largely neglected communication modality, however, it is becoming increasingly clear that using vibrations is the dominant mode of communication in termites [13] , [14] , [50] . We hypothesise that biotremology is not only used to determine food size [50] , [52] or to drum alarm [53] , [54] , [55] , [56] , [57] but could also be essential for the construction of termite nests. However, as indicated by Darlington [58] , an explicit classification of various functional structural relationships between mound (nest, corridors and walls, material composition) and colony (individual, collective) is a neglected aspect in termite behaviour and ecology research. In addition to a brief review of well-known functionalities of termite nests, this paper is also aimed at discussing direct vibrational communications, as opposed to stigmergy, and identifying potential research areas which might offer insights into how termites interact within their mounds.",
"discussion": "4 Discussions As outlined above, there is variable depth of the state-of-the-art knowledge about the various functionalities of termite structures and there is a lack of systematic studies to allow general features and differences to be classified. Here we will discuss and identify key research topics that will potentially answer the question on the holistic picture of the interrelationship between termite structures, termites and their behaviour as a superorganism. While bees or ants would survive without their nest for some time, termites would be exposed to the twin dangers of desiccation or predation [13] . Similarly, without a termite colony, the mound would become brittle, and collapse like a ’house of cards’ [34] . Termite nest architecture is therefore an expression of innate insect behaviour, altered by contact with the environment as “morphological expression of the sum of behavioural patterns” [3] , [36] , [151] . Thus, past and contemporary research largely expanded on how termites (mostly African Macrotermiditae) climatise their mound; how colonies organise chores, decentralisticly and autonomously, assuming stigmergic and self-organisational mechanisms as root cause of complexity and collective building [15] , [42] , [46] . Pheromones and self-organisation revisited. The building process rather than the built structure has been the centre of interest [46] , [181] . Termite tunnelling has been mathematically modelled using (reaction-) diffusion systems, Laplacian growth models or Gaussian processes (diffusion system with randomised initial conditions), yet it is unclear to which scale these simulations are valid as no complete experimental validation is provided [15] , [46] , [190] . Corridor systems appear tree-like, as e.g. found in Cubitermes spp. [38] , [97] , containing only few loops. King et al. [33] , however, described corridors and conduits as well-connected , the essential enabler to successfully use gypsum in endocasts. The connectivity of the tunnels and nodes (chambers) is attributed to a sub-function of defence or climatisation – the assumption that the tunnel system and digging activities follow diffusive processes or resulting in tree-like shapes seems to be a convenient simplification, the truth might yet lie somewhere in-between, with stronger emphasis on determinism , i.e. defined functionality of engineered structures. While there is largely consensus that group level patterns emerge from interacting individuals following simple behavioral rules (individual-collective behaviour interaction), stigmergic building processes presumably originate from a cement pheromone. Yet, since termite mounds seem to grow in discrete stages, it has been argued, that a general continuous (global) growth model based on molecular diffusion of pheromones through the mound wall can be excluded [22] , [64] . A pheromone is assumed to be embedded within termite boluses and taken as main factor for diffusion processes with randomness being induced e.g. through termites walking off the construction side [15] , [16] . However, no cement pheromone has yet been identified [44] , [191] so that Green et al. [184] suggested a chemical signal other than a pheromone. What if a largely unknown mechanism, different to stigmergy but related to pheromones, is responsible for building? Biotremological signals. Recently, the action of digging and the aggregation of termite workers have shown a strong effect on recruiting termites for excavation and building work [184] . Aggregation alone as information , though, cannot be the only factor since termites within the nest walk and live next to each other which also leaves traces and signals [13] . It is also mentioned in [44] that termites act as physical obstacles and therefore limit the excavation. However, if termites are blind, cues other than aggregation and excavation could be the trigger for increased building activities. The application of Random walk or swarm behaviour models, widely applied in computer science, seem debatable in light of the eusociality of termites, which follow explicit cues and directed signals [13] , [15] , [46] , [48] , [192] . Some of the most prevalent direct signals termites are exposed to are those they use for biotremology, yet near to nothing is known about how termite colonies communicate in detail using vibrations such as using their mound as a communication channel, being adjacent to colonies of the same species or other species (intra- and interspecific communication among strongest inter-and intraspecific competition). Grohmann et al. [64] assigned regular mound distribution patterns and colony size of M. michaelseni to intraspecific competition for foraging areas; it may be assumed that communication and eavesdropping are significant in colony survival. Evans et al. [193] studied how the subordinate drywood termite Cryptotermes secundus eavesdrops on the dominant subterranean termite species Coptotermes acinaciformis to choose smaller pieces of wood to avoid competition; similar strategies – a preference for distinct diets to avoid conflicts – have been found in many neotropical termite cohabiting builder- and inquiline-species-relationships, cf. [120] . Oberst et al. [13] found that termites of the commensal species Macrognathotermes sunteri are very quiet, and their walking cause less vibrations than its host species, Coptotermes acinaciformis , resulting in the so-called disguise in the form of insignificance as a special mechanism of camouflage [148] , [194] . Similar relationships, whether they are host-commensurate/inquiline or parasite relationships, are known in many South-American species [195] , however, whether signalling is based on mainly chemical or vibrational signals/cues or on multimodal effects, needs to be yet studied for each relationship separately. Considering that biotremology plays a central role in termite communication [13] , [14] , [49] , [50] , and that the corridors within the mound are likely to be saturated with pheromones, and cannot work as a two-way communication system due to the directed airflow within the tunnels, the use of substrate borne vibrations and synchronisation seems advantageous. Synchronisation, as studied in nonlinear dynamics and mathematical physics [45] , [196] , [197] , as deterministic oscillatory (here: vibratory) motion, is observed in both the physical and biological world, ranging from mechanical oscillators and bio-acoustics to predator–prey cycles and ecosystem dynamics [196] , [197] , [198] , [199] , [200] . Synchronisation of vibrational information might be more important to building and nest growth, triggered initially via localised individual action which may lead to global collective behaviour. The termite nest would act as a both a communication network and a large vibration sensor; locally constrained, vibrational signals and cues would provide subnetworks with synchronised tasks clearly defined via transmission through the structure. Studying the mechanical composition (type of clay, silt and sand particles used) and their compound properties would enable a deeper understanding of how termites modify their surroundings – which should be crucial for their eusocial character and the information exchange (from nestmate interaction to synchrony) required to run a colony of several million individuals. Since biotremology has been shown to be significant in termite colony organisation, local and global synchronisation rather than stigmergy could be the prevalent trigger for building activities and the reason for group-level pattern emergence; this remains to be determined [14] , [34] , [70] . In order to study the mound and the structure of a termite nest, its wave propagation, its filter properties and its function as communication channel or even as a communication network, the material properties of the entire structure need to be determined. Understanding the structure . Until recently the galleries of ant and termite nests have been studied using endocasts, e.g. gypsum, dental plaster or lead [32] , [33] ; however, novel technology using X-ray and specifically micro-computed tomography ( μ CT, mm range) now allows non-destructive visualisation of tunnels and details of the structure [38] , [151] , [201] . The ventilation of the mound as well as the emergence of tunnel systems and their mathematical descriptions has received much attention followed by study of the coordination of individuals and their collective behaviour using conventional statistical and Fourier-based methods [15] , [35] , [43] , [202] . Yet, medical imaging lacks resolution and classical Fourier-based methods are linear and neither cover the spatial nor the temporal character of termite-built structures. Sophisticated measurement techniques such as ultra-high- or super-resolution X-ray tomography imaging, atomic force microscopy, alongside accurate granulometry, spectrometry, excellent computational resources, novel big data analysis techniques and computational intelligence methods would be required to capture the microscale of the walls including their porous structure and multiscale material characteristics and compositions. We know that variations in lignin characteristics and density fractions of termite nests reflect differences in feeding guilds of the studied taxa [170] . However, the exact composition of termite-built structures including the kind of lignin-based phenol used in different parts of the mound considering different functions remains unanswered. Cation-adsorbing capacity provides “expandable clays” as a surface chemical or surface complexation process to facilitate the exchange of chemical species between an aqueous solution and mineral surfaces present in geological porous formations [30] , [165] which could be related to micro-porosity, ventilation and natural evaporative cooling. Yet to-date, there is no clear understanding on which material composition can transmit signals efficiently, to carry loads, and to store which kind of food; there is virtually no knowledge on the geometry of the structures termites build (the tortuosity of the corridors or the porosity of the walls) and their effect on the ventilation; air-conduits are supposedly smoother than other parts of the nest [36] – however, different surfaces can cause the fluid’s boundary layer to change and the effect on ventilation should be quantified. The techniques used to visualise the nest as well as analysis methods applied to study complex structures have been identified as being problematic [15] . Since data of natural phenomena are inherently complex, nonlinear time series analysis (NTSA), particularly recurrence plots and their quantification measures as increasingly applied in science and engineering, could provide valuable insights into the physics of termite-built structures [203] , [204] , [205] , [206] , [207] . While these methods have been foremostly applied to the understanding of complex time-dependent behaviour, they are in general also applicable to discontinuous-discrete or continuous spatial and temporal-spatial structures [45] , [161] , [207] . Using NTSA measures to determine whether the wall composition and the tunnel geometry avoid being detected by ants in coexistence with termites as observed in [13] could be an interesting area of research. Using machine learning tools would allow features to be extracted and spatially different structures to be classified for species analysis so that evolutionary and ecological traits in their structures may be identified. However, as indicated by Korb (2011) [37] there are still too many open questions about the material properties of the walls, the multiple functions of structures and their connection to individual behaviour and communication, that can only be answered by highly multi-disciplinary studies."
} | 4,871 |
31405077 | PMC6789667 | pmc | 3,662 | {
"abstract": "With the advancement of synthetic biology, the cell-free protein synthesis (CFPS) system has been receiving the spotlight as a versatile toolkit for engineering natural and unnatural biological systems. The CFPS system reassembles the materials necessary for transcription and translation and recreates the in vitro protein synthesis environment by escaping a physical living boundary. The cell extract plays an essential role in this in vitro format. Here, we propose a practical protocol and method for Escherichia coli -derived cell extract preparation and optimization, which can be easily applied to both commercially available and genomically engineered E. coli strains. The protocol includes: (1) The preparation step for cell growth and harvest, (2) the thorough step-by-step procedures for E. coli cell extract preparation including the cell wash and lysis, centrifugation, runoff reaction, and dialysis, (3) the preparation for the CFPS reaction components and, (4) the quantification of cell extract and cell-free synthesized protein. We anticipate that the protocol in this research will provide a simple preparation and optimization procedure of a highly active E. coli cell extract.",
"conclusion": "5. Conclusions Over the past decade, the cell-free system has been revitalized as an essential toolkit for synthetic biology and biotechnology research. Moreover, with its unique non-living feature, the cell-free system provides a new insight to understand the cellular processes outside the shell. Preparation of a highly active cell extract is the first step to build this versatile CFPS system. In this study, we discussed a detailed cell extract preparation protocol in three major stages encompassing cell culture, sonication, and optional post-lysis steps for genomically engineered E. coli K12 MG1655 ΔprfA ΔendA Δrne as a model strain. We expect that this protocol will provide not only a practical guideline but also a foundation for the entire CFPS system.",
"introduction": "1. Introduction The technology involving the disruption of bacterial cells and collection of ribosomes for synthesizing proteins was first introduced when the fraction of ribosomes was identified as the core of the protein synthesis machinery of the cells [ 1 ]. The cell-free protein synthesis (CFPS) system has been developed for exclusive protein synthesis utilizing active ribosomes and other cellular machinery outside of the living cell [ 2 , 3 ]. Recent progress of synthetic biology highlights this versatile system as an essential toolkit for exploring and maneuvering complex cellular processes to accelerate technology advances [ 4 , 5 ]. The CFPS system offers many advantages over a cell-based system, such as ease of manipulating biochemical pathways [ 6 , 7 ], higher tolerance on chemicals and toxic compounds [ 8 , 9 ], utilization of PCR amplified linear template allowing for high-throughput preparation and protein synthesis of the gene of interest and breadboarding synthetic biological circuit [ 10 , 11 , 12 , 13 ], and the capability of highly efficient non-standard amino acid incorporation [ 14 , 15 ]. In addition, the CFPS system allows the benefits of unprecedented logistics along with freeze-drying paper-based format [ 16 ]. Since the cell extract carries most of the cellular machinery, its preparation is considered as the first important step for building a highly productive CFPS system. Many studies streamlined the overall procedure for cell extract preparation to improve overall extract performance in CFPS system [ 17 , 18 , 19 , 20 , 21 ]. Although recent progress in the preparation of Escherichia coli cell extract has resulted in an increase in protein productivity up to 1–1.5 mg/mL in a single batch cell-free system [ 22 ], the total protein yields are varied from strain to strain due to the necessary variations, dictated by the strain, of the three major cell extract preparation stages: pre-lysis, lysis, and post-lysis. The preparation step of the cell extract is crucial, as it contains key components for synthesizing proteins, so it is important to practice the optimized cell-free extract method for each E. coli strain for high protein yield. For example, Kim et al. [ 19 ] demonstrated that E. coli strain BL21(DE3)-derived cell extract prepared in a simpler procedure were more effective than the cell extracts prepared by the conventional method described by Pratt et al. [ 23 ]. However, the modified procedure was not as productive for the traditional host organism, E. coli A19 derived extract [ 19 ]. This evidence indicated that different preparation conditions are required depending on the E. coli strain of choice to maximize cell extract performance. However, the recent study from Kwon and Jewett was the first to introduce an optimized systemic cell extract preparation process for the non-commercial engineered E. coli strain K12 MG1655 (C495) which greatly advanced the potential for use in future biomedical/industrial applications [ 22 ]. In addition, the systemically optimized CFPS is inspiring novel ways to utilize the cell extracts from engineered E. coli strains for applications involving unnatural amino acid incorporation [ 15 ], patient-specific therapeutic vaccines [ 24 ], anticancer protein production [ 25 ], and more. Here we describe a procedure for cell extract preparation step-by-step for the genomically engineered E. coli strain ΔprfA ΔendA Δrne [ 15 ] to generate the optimal cell extract with the maximum protein productivity. This protocol also can be applied to other strains with slight modification. The protocol aims to clarify which processing variables are the most critical for the cell extract performance during CFPS and how the processing condition can be optimized for different E. coli strains."
} | 1,448 |
30687248 | PMC6334743 | pmc | 3,663 | {
"abstract": "Dairy wastes are widely studied for the hydrogen and methane production, otherwise the changes in microbial communities related to intermediate valuable products was not deeply investigated. Culture independent techniques are useful tools for exploring microbial communities in engineered system having new insights into their structure and function as well as potential industrial application. The deep knowledge of the microbiota involved in the anaerobic process of specific waste and by-products represents an essential step to better understand the entire process and the relation of each microbial population with biochemical intermediates and final products. Therefore, this study investigated the microbial communities involved in the laboratory-scale anaerobic digestion of a mixture of mozzarella cheese whey and buttermilk amended with 5% w/v of industrial animal manure pellets. Culture-independent methods by employing high-throughput sequencing and microbial enumerations highlighted that lactic acid bacteria, such as Lactobacillaceae and Streptococcaceae dominated the beginning of the process until about day 14 when a relevant increase in hydrogen production (more than 10 ml H 2 gVS -1 from days 13 to 14) was observed. Furthermore, during incubation a gradual decrease of lactic acid bacteria was detected with a simultaneous increase of Clostridia , such as Clostridiaceae and Tissierellaceae families. Moreover, archaeal populations in the biosystem were strongly related to inoculum since the non-inoculated samples of the dairy waste mixture had a relative abundance of archaea less than 0.1%; whereas, in the inoculated samples of the same mixture several archaeal genera were identified. Among methanogenic archaea, Methanoculleus was the dominant genus during all the process especially when the methane production occurred, and its relative abundance increased up to 99% at the end of the incubation time highlighting that methane was formed from dairy wastes primarily by the hydrogenotrophic pathway in the reactors.",
"conclusion": "Conclusion Anaerobic biosystem was strictly influenced by microbial communities structure and dynamics derived from the inoculum, feedstock and the operating conditions. It represented a sustainable management process for the valorization of abundant wastes and by-products recovered from dairy industry. Polyphasic approach highlighted the function of specific bacterial populations that drove the biohydrogen production. Besides, the inoculation in the reactors with pelleted manure allowed Archaea development, revealing that methane was primarily formed through the hydrogenotrophic pathway, since Methanoculleus was the dominant genus during the process.",
"introduction": "Introduction In the near future, novel bio-based technologies in waste management can be used to convert organic waste into valuable products such as renewable energy and/or biopolymers through biological processes ( Pagliano et al., 2017 ) with a goal to potentially replace fossil fuels with biomasses and reducing pollutant emissions. Several organic wastes are potentially suitable to be used as substrates for producing renewable energy vectors (e.g., biohydrogen, biogas, and biomethane) through anaerobic biosystems ( Raposo et al., 2012 ; Ghimire et al., 2015 ; Pagliano et al., 2017 ). Among them, cheese whey and buttermilk, residues from dairy factories as by-products of cheese, yogurt, milk, and butter production process are interesting substrates for their high content of soluble organic matter, i.e., chemical oxygen demand (COD) ranging from 0.1 to 100 g L -1 ( Prazeres et al., 2012 ). Besides substrates and operational conditions, microorganisms significantly affect the performance of the anaerobic process ( Panico et al., 2014 ). In fact, the efficiency and stability of this process is entirely dependent upon the syntrophic activity of microorganisms operating in different phases ( Li et al., 2009 ; Vanwonterghem et al., 2014 ). Actually, anaerobic digestion process can be conceptually divided into four stages defined by the primary catabolic reactions that occur at each stage: hydrolysis of complex polymers (I, hydrolysis), fermentation of the hydrolysis end-products to volatile fatty acids (VFAs) (II, acidogenesis), conversion of VFAs to acetate and hydrogen (III, acetogenesis), and finally the production of methane from acetate and hydrogen (IV, methanogenesis) ( Yu et al., 2010 ). Therefore, it is important to understand how the raw materials as well as environmental and physical conditions established in the system affect microbial growth and activity, and therefore, the performance of the anaerobic process. Numerous studies using different types of organic wastes have been conducted to better understand the role of the microorganisms involved in each stage and the microbiomes present in the anaerobic reactors ( Nelson et al., 2011 ; Li et al., 2017 ; Ros et al., 2017 ; Westerholm et al., 2017 ). For this purpose, various methods have been applied to investigate the microbial communities or targeted specific groups in anaerobic digesters, including clone library of 16S rRNA genes ( Rincón et al., 2008 ), denaturing gradient gel electrophoresis (DGGE) analysis ( Shin et al., 2010 ; Palatsi et al., 2011 ; Supaphol et al., 2011 ; Pagliano et al., 2018 ; Ventorino et al., 2018 ) and fluorescence in situ hybridization (FISH) ( Braguglia et al., 2012 ). All these methods, although are highly efficient, analyze a limited number of aspects if compared with the emerging metagenomic approaches based on high-throughput sequencing (HTS) ( Yang et al., 2014 ). Therefore, in this study, the use of a polyphasic approach including HTS in lab-scale batch tests, allowed to elucidate the dynamics of microbiota in different stages of the anaerobic process fed with a mixture of dairy waste from a mozzarella cheese factory. Culture-independent and culture-dependent approaches coupled with hydrogen and methane production can improve the knowledge concerning this specific anaerobic biosystem. In particular, it is important to elucidate how specific microbial populations can steer the hydrogen and methane production in order to control, also through traditional parameters (pH, TS, VS, COD) ( Pontoni et al., 2015 ), the efficiency of the anaerobic biosystems fed with different wastes and by-products from dairy industry.",
"discussion": "Discussion The microbial composition of the initial dairy waste mixture showed a high concentration of viable aerobic and anaerobic bacteria mainly belonging to LAB that commonly thrive in dairy waste ( Kasmi et al., 2017 ). Since LAB produce lactic acid by homolactic and heterolactic fermentation processes ( Palomba et al., 2012 ; Pradhan et al., 2017 ) they are well-adapted to the acidic environment ( De Candia et al., 2007 ), typical of the cheese whey used in this study collected after fermentation and addition of organic acids during the production chain ( Carvalho et al., 2013 ). Physical and chemical characteristics of the cheese whey are, actually, strictly related to the production chain, thus showing a wide range of organic matter concentration as reported in the literature ( Ghaly and Singh, 1989 ; Ghaly and Kamal, 2004 ; Farizoglu et al., 2007 ; Saddoud et al., 2007 ; Azbar et al., 2009 ). In this study, cheese whey characteristics were in accordance with Ghaly and Singh (1989) that reported the concentrations of COD and VS equal to 75.8 and 47.9 g L -1 , respectively. The high COD in cheese whey influenced the resulting COD in the mixture, indicating the potential of this substrate for feeding anaerobic process ( Carvalho et al., 2013 ) and producing H 2 and CH 4 . During incubation a relevant increase in hydrogen production was observed, simultaneously a decrease of the LAB concentration occurred until to reach a concentration less than 1 log CFU mL -1 at the end of the incubation time. The decrease in the LAB concentration was related to the increase in H 2 production proving that lactic fermenters were the main competitors with the H 2 producing microorganisms ( Perna et al., 2013 ). This result was also confirmed by HTS for the bacterial 16S rRNA gene that showed a noticeable decrease (from 67% at day 7 to 6% at day 14) in the relative abundance of the Lactobacillaceae family when relevant H 2 production occurred. Accordingly, it has been reported that Lactobacillus could not be related to high H 2 production rate ( Davila-Vazquez et al., 2009 ), although Yang et al. (2007) found higher abundance of Lactobacillus than Clostridium in anaerobic fermentation of cheese processing wastewater, thus reporting Lactobacillus related to H 2 production. On the other hand, in this work the increase in H 2 production occurred simultaneously with an increase in Clostridia load. Actually, Clostridia have been reported to convert lactate into butyrate, CO 2 and H 2 in the presence of acetate ( Barlow et al., 1991 ). Thus, the presence of these microorganisms could favor the production of hydrogen by fermentation of lactic acid followed by H 2 production ( Perna et al., 2013 ). For this reason, lactic acid was not detected at day 14 and an increase of acetic and propionic acids concentration was observed. This cultural approach allowed to acquire information about dynamics of viable bacterial populations which were able to live, grow, and die during the biodigestion process of this specific waste. Moreover, in order to obtain more information, a cultural-dependent approach was combined with a cultural-independent molecular method. According to Pandya et al. (2017) these techniques seem to be roughly equivalent and, when used in parallel, it is possible to obtain best results leading to major advances in the reliable knowledge of microbial populations living in an environment. The cultural results were confirmed by the HTS analysis that showed an increase in bacterial families belonging to Clostridiales order, such as Tissierellaceae and Clostridiaceae , which represented the most abundant bacterial taxa until the end of the incubation time. The selective pressure occurred in the ecosystem due to the presence of inoculum and the chemical–physical conditions of the anaerobic process selected these bacterial taxa, which are well-known to be involved in the H 2 production ( Navarro-Díaz et al., 2016 ; Alexandropoulou et al., 2018 ). In particular, in this study, the most dominant OTUs of Clostridiaceae and Tissierellaceae were affiliated with Clostridium spp. ( C. thermopalmarium, C. clariflavum , and C. tyrobutyricum ) and Sporanaerobacter acetigenes , respectively, which are commonly detected and isolated in many reactors for CH 4 or H 2 production ( Hernandez-Eugenio et al., 2002 ; Shiratori et al., 2006 ; Jo et al., 2008 ; Weiss et al., 2008 ; Kim et al., 2010 ; Sasaki et al., 2012 ; Xia et al., 2014 ; Kumar et al., 2015 ; Cibis et al., 2016 ). The genus Clostridium comprised a large number of species that were often used to produce H 2 ( Jiang et al., 2013 ). Among them, Clostridium tyrobutyricum has been widely reported to be able to produce significant quantities of H 2 from different sugars ( Jiang et al., 2018 ) as well as it is also capable to utilize lactate as the main substrate for producing H 2 ( Wu et al., 2012 ; Noblecourt et al., 2018 ). Furthermore, fermentation products of C. clariflavum are H 2 , CO 2 , acetate, lactate, ethanol and a small amount of formate ( Shiratori et al., 2009 ). Whereas C. thermopalmarium species are able to ferment sugars into butyric acid producing simultaneously H 2 , CO 2 , small amounts of acetate, lactate, and ethanol ( Soh et al., 1991 ). Geng et al. (2010) demonstrated that the inoculation of C. thermopalmarium strain BVP (DSM 5974) increased biohydrogen production rather than the monoculture of C. thermocellum from cellulose. In addition, Navarro-Díaz et al. (2016) , reported the positive relationship between the increased H 2 production with the presence of specific microbial families and genera, such as Tissierellaceae that may contribute also to complex substrate degradation because of its putative xylanolytic activity ( Niu et al., 2009 ). Sporanaerobacter acetigenes strain Lup 33 T , closely related to the OTU denovo78367, detected at the highest relative abundance, as well as related to the others representative OTUs affiliated to Tissierellaceae , was recognized as an acetogenic bacterium able to synthesize a mixture of VFAs, including acetate, isovalerate and isobutyrate, together with H 2 and CO 2 ( Hernandez-Eugenio et al., 2002 ). Han et al. (2016) reported that Sporanaerobacter acetigenes was one of the main contributors for the hydrolysis and acidogenesis stages during anaerobic digestion of food waste-recycling wastewater ( Han et al., 2016 ) as well as it was one of the primary species along with Clostridium during semi-continuous fermentation of C. pyrenoidosa biomass for H 2 production ( Xia et al., 2014 ). Moreover, the presence of Ruminococcaceae members in the samples taken at days 14 and 30 may be also correlated to H 2 production since they are recognized as hydrogen producers and important substrate hydrolyzers ( Niu et al., 2009 ). In addition, ethanol was also produced in high amount starting from day 14 likely causing an inhibiting effect on hydrogen production ( Hung et al., 2011 ) that therefore could have been higher than observed. This side effect can be related to the presence of Streptococcaceae , detected during incubation by using HTS for the bacterial 16S rRNA gene, since genera belonging to this family could produce ethanol thus inhibiting hydrogen production ( Ren et al., 2007 ). Furthermore, the environmental and physical conditions established in the system were also effective for the selection of the CH 4 producing archaea. First, the archaeal populations’ presence in the biosystem was related to inoculum since the initial samples of the mixture of cheese whey and buttermilk had a relative abundance of archaea less than 0.1%. Secondly, the anaerobic environment in the biodigesters selected archaeal genera causing the decrease of the aerobic Nitrososphaera genus also detected in other studies of anaerobic digestion process ( Li et al., 2014 ; Ventorino et al., 2018 ) even if it could be no able to produce methane ( Chen et al., 2012 ). During incubation, Methanoculleus genus percentage increased achieving very high relative abundance (99% at the end of the experiment) compared with that reported in other studies, such as Di Maria et al. (2017) (85%) and Leven et al. (2007) (18%), related to anaerobic digestion of organic waste. As reported by Poirier et al. (2016) , Methanoculleus cooperate with acetate oxidizing bacteria belonging to the Clostridiaceae family detected in this study by both cultural and molecular approaches. In agreement with Di Maria et al. (2017) , Methanoculleus abundance increased during the incubation time and became dominant, whereas Methanosarcina decreased, as they are usually dominant in process fed with organic fraction of municipal solid waste (OFMSW) and activated sludge as inoculum ( Lin et al., 2012 ). In this biosystem fed with a mixture of cheese whey and buttermilk, Methanoculleus was dominant using a hydrogenotrophic pathway to produce CH 4 causing no organic acids consumption when methane production occurred. This result was in agreement with a previous study in which hydrogenotrophic pathway was identified as the main driver for CH 4 production in batch reactors fed with dairy wastes, although, using different process conditions and a qualitative culture-independent method (DGGE) to microbial identification, Methanobrevibacter was found as the genus mostly related to the CH 4 production ( Pagliano et al., 2018 ). Interestingly, in this study, the dominant methanogenic archaeal OTU was affiliated with M. thermophilus which is able to produce CH 4 from H 2 or formate ( Rivard and Smith, 1982 ). Recently, this species, with M. beijingense , was found to be dominant in a full-scale thermophilic anaerobic digester treating food wastewater ( Lee et al., 2017 ). The other representative OTUs affiliated with Methanoculleus genus were closely related to M. sediminis, M. taiwanensis , and M. horonobensis previously isolated from sediments near a submarine mud volcano ( Chen et al., 2015 ), deep-sea sediment ( Shimizu et al., 2013 ) and deep subsurface groundwater from a diatomaceous shale formation ( Weng et al., 2015 ), respectively. Comparing the H 2 and CH 4 production with the literature, the production in this study (54.34 ± 0.15 L H 2 kg VS -1 and 16.74 ± 0.71 L CH 4 kg VS -1 ) were higher than that reported by Pagliano et al. (2018) (8.9 L H 2 kg VS -1 and 2.2 L CH 4 kg VS -1 ) using dairy waste as substrate and operating in batch mode. Different operating condition can promote the CH 4 production, as reported by Lateef et al. (2014) that studied an anaerobic sequencing batch reactor (ASBR) fed with dairy waste achieving 35.6 L H 2 Kg VS -1 and 627 L CH 4 Kg VS -1 . Overall, obtained results highlighted that culture-dependent and independent approaches provided evidence for examining the relationship between bacterial and archaeal populations and biogas production in this biosystem. Besides, metagenomics sequencing technology is important to quantify the different microbial populations occurred in the reactors as well as to better understand the microbial dynamic during the anaerobic process of dairy wastes."
} | 4,423 |
34667183 | PMC8526750 | pmc | 3,665 | {
"abstract": "Isoflavonoids comprise a class of plant natural products with great nutraceutical, pharmaceutical and agricultural significance. Their low abundance in nature and structural complexity however hampers access to these phytochemicals through traditional crop-based manufacturing or chemical synthesis. Microbial bioproduction therefore represents an attractive alternative. Here, we engineer the metabolism of Saccharomyces cerevisiae to become a platform for efficient production of daidzein, a core chemical scaffold for isoflavonoid biosynthesis, and demonstrate its application towards producing bioactive glucosides from glucose, following the screening-reconstruction-application engineering framework. First, we rebuild daidzein biosynthesis in yeast and its production is then improved by 94-fold through screening biosynthetic enzymes, identifying rate-limiting steps, implementing dynamic control, engineering substrate trafficking and fine-tuning competing metabolic processes. The optimized strain produces up to 85.4 mg L −1 of daidzein and introducing plant glycosyltransferases in this strain results in production of bioactive puerarin (72.8 mg L −1 ) and daidzin (73.2 mg L −1 ). Our work provides a promising step towards developing synthetic yeast cell factories for de novo biosynthesis of value-added isoflavonoids and the multi-phased framework may be extended to engineer pathways of complex natural products in other microbial hosts.",
"introduction": "Introduction Isoflavonoids constitute a diverse family of natural products that are primarily synthesized by leguminous plants 1 . In addition to playing significant ecological functions 2 , isoflavonoids exhibit various human health-promoting properties, such as antioxidant activity, cardioprotective activity, osteoporosis reduction, and cancer prevention, all of which have resulted in studies on exploiting these molecules as agents both in the pharmaceutical and nutraceutical industry 3 , 4 . The current production of isoflavonoids relies on direct plant extraction. However, the low phytochemical abundance, significant investment of time, energy, and capital, and huge requirement for potentially toxic solvents have excluded this approach from being used as it is neither economical nor environmental-friendly 5 , 6 . Moreover, the cultivation of legumes is geographically uneven and the amounts of isoflavonoids vary greatly from cultivars and climatic conditions 7 . All these facts introduce further risk and instability in the supply of these chemicals by means of plant extraction. Developing alternative sources of isoflavonoids is therefore a prominent challenge to be addressed, prior to being able to feasibly produce these chemicals at scale using standardized industrial processes. With the rapid advance in metabolic engineering and synthetic biology over the last few decades, microbe-based bioproduction has become increasingly pursued as an alternative to traditional chemical production techniques 8 , 9 . By re-engineering the cellular metabolism of fast-growing microorganisms, such as Escherichia coli and Saccharomyces cerevisiae , artificial cell platforms have been successfully constructed to produce high levels of chemicals ranging from biofuels to proteins 10 . In addition, grafting and optimizing plant biosynthetic pathways in microbial hosts is becoming a compelling route to supply plant natural products, as demonstrated by substantial biosynthesis of high-value-added alkaloids, stilbenes, and flavonoids, and terpenoids from simple sugar 11 – 15 . Based on this growing body of work, we speculated that microbial cell factories may also offer the potential for the production of commercially viable isoflavonoid as well. Structurally, isoflavonoids contain the common C6-C3-C6 flavonoid skeleton and are characterized by having the B-ring connected at C3 rather than C2 position of the C-ring, compared to other flavonoid subclasses 3 (Supplementary Fig. 1 ). The isoflavones genistein (GEIN) and daidzein (DEIN) constitute two basic scaffolds from which over thousands of isoflavonoids are derived as a result of diverse structural modifications, including hydroxylation, methylation, glycosylation, and molecular rearrangements 2 – 4 . Reconstruction of the isoflavone pathway for biosynthesis of these molecules, therefore, represents the entry point to microbial production of a large variety of different biologically active isoflavonoids. Previously, heterologous biosynthesis of GEIN and DEIN was demonstrated by introducing plant enzymes alongside feeding precursors, such as L-tyrosine, naringenin, or liquiritigenin, in both E. coli and S. cerevisiae 16 – 21 . Moreover, the expression of specific glucosyltransferase in E. coli also enabled the bioconversion of GEIN and DEIN to corresponding glucosides genistin (GIN) and daidzin (DIN) 22 , 23 , the primary form of stored isoflavones in leguminous plants 3 . While the reported low titers necessitate further improvement to support industrial-scale production, there have been rare efforts to engineer and optimize de novo microbial biosynthesis of isoflavones. Here we present the establishment of a de novo DEIN-producing yeast platform and its application for the biosynthesis of glycosylated isoflavonoids using a multi-phased metabolic engineering strategy (Fig. 1 ). In screening phase I, we first evaluated diverse plant enzymes to rebuild a functional DEIN pathway and extensively diagnosed exogenous and endogenous metabolic factors affecting the activity of key biosynthetic enzymes. Through pathway reconstruction in phase II, we improved the metabolic flux towards the DEIN pathway by implementing: (1) gene amplification to promote the expression of selected pathway genes; (2) protein fusion strategy to facilitate substrate trafficking; (3) additional genetic manipulations to increase the supply of metabolic cofactors (identified during phase I to be potential bottlenecks in pathway flux); (4) process development; and (5) fine-tuning of gene expression in the competing metabolic pathways. The systematic engineering enabled the production of 85.4 mg L −1 DEIN from glucose in shake flask cultivations. Finally, during application phase III, we demonstrated the efficient conversion of DEIN to bioactive glycosylated isoflavonoids by introducing plant glycosyltransferases. Supplementary Fig. 2 provides an overview of all strains constructed in the different phases of the development process. Fig. 1 Engineering the de novo biosynthesis of isoflavonoid chemicals in yeast. A multi-phased metabolic engineering strategy was implemented to enable de novo production of isoflavonoids, following the pipeline of screening-reconstruction-application. In screening phase I (green box), a synthetic DEIN pathway was established and metabolic factors improving the performance of rate-limiting reactions were identified using a moderate p -HCA producing platform strain QL11. In reconstruction phase II (orange box), the DEIN production was further optimized in a wild-type strain QL179 harboring the deletion of galactose utilizing genes ( GAL7 / 10 / 1) , through amplifying gene expression, enhancing substrate transfer, combining effective genetic targets identified in phase I and fine-tuning the expression of a key gene involved in competing metabolic pathway (orange triangle). For application phase III (magenta box), the generated DEIN platform strain was used as the starting point for the production of bioactive glucosides PIN and DIN through introducing plant glycosyltransferases and enhancing the supply of glycosyl group donor UDP-glucose. The selected plant biosynthetic genes and overexpressed yeast native genes for isoflavonoid production were highlighted in blue boxes. Magenta arrows, designed DEIN biosynthetic pathway. Blue arrows, reactions for generating glucosides; gray arrows, byproduct pathway. At4CL1 , 4-coumarate-coenzyme A ligase 1 from Arabidopsis thaliana ; GmCHS8 , chalcone synthase from Glycine max ; GmCHR5 , chalcone reductase from G. max ; GmCHIB2 , chalcone isomerase from G. max ; Ge2-HIS , 2-hydroxyisoflavanone synthase from Glycyrrhiza echinata ; GmHID , 2-hydroxyisoflavanone dehydratase from G. max ; AtPAL2 , phenylalanine ammonia lyase from A. thaliana ; AtC4H , cinnamic acid-4-hydroxylase from A. thaliana ; FAS1 , beta subunit of yeast fatty acid synthetase; CrCPR2 , cytochrome P450 reductase from Catharanthus roseus ; STB5 , yeast native transcriptional factor; EcyjfB , NAD + kinase from Escherichia coli ; GmUGT4 , isoflavone 7- O -glucosyltransferase from G. max ; PlUGT43 , isoflavone 8- C -glucosyltransferase from Pueraria lobate ; UGP1 , UDP-glucose pyrophosphorylase; PGM1/2 , phosphoglucomutase 1/2. In addition, yeast heme degradation was disrupted by deleting heme oxygenase-coding gene HMX1 . ER endoplasmic reticulum, E4P erythrose-4-phosphate, PEP phosphoenolpyruvate, L-Phe L-phenylalanine, 5-ALA 5-aminolevulinic acid, UDP-Glc uridine diphosphate-glucose, UTP uridine triphosphate, G-6-P glucose-6-phosphate, G-1-P glucose-1-phosphate.",
"discussion": "Discussion Isoflavonoids play important roles in the plant defense system and have many human health-related benefits. They, therefore, represent promising candidates in the development and engineering of agents for agricultural, nutraceutical, and pharmaceutical applications. Here we established a yeast-based de novo production platform for the efficient production of the isoflavonoid carbon skeleton DEIN as well as the high-value glucosides PIN and DIN. This was achieved by first identifying functional biosynthetic enzymes to generate DEIN (screening phase I), then by optimizing metabolic flux at enzyme and pathway levels to further increase DEIN titer (reconstruction phase II) and finally by introducing plant UGTs to convert DEIN to corresponding glucosides (application phase III). Gene duplication and diversification occur in the evolution of plant secondary metabolism to tackle the changing environment, creating a rich variability and complexity of plant products as a result 76 . However, this functional divergence poses an obstacle to identifying ideal candidate enzymes for reconstructing heterologous biosynthetic pathways for plant metabolite production. Most of the structural genes involved in isoflavonoid pathways have been characterized 25 , and here we exploited their genetic diversity, by performing a combinatorial evaluation of biosynthetic genes from both leguminous and non-leguminous plants, to enable DEIN production (Fig. 2b–d ). The P450s constitute the most versatile tailoring enzymes that catalyze irreversible and often rate-limiting reactions in the biosynthesis of plant-specialized products 31 . Though S. cerevisiae is generally identified as a superior host for the functional expression of membrane-bound plant P450s over its prokaryotic counterparts, extra efforts are required to maximize their catalytic efficiency. Two distinct P450s, the upstream C4H hydroxylating cinnamic acid and the downstream 2-HIS mediating the migration of aryl moiety of LIG, are involved in the biosynthesis of DEIN (Fig. 3a ). While the activity of AtC4H has been enhanced by co-expressing RP 77 in our screening strains providing excess precursor p -HCA (QL11 background), the selected Ge2-HIS still exhibited sub-optimal performance in converting LIG to DEIN (Supplementary Fig. 4 ). Starting with evaluating plant CPRs and artificial RP surrogates, which could impact the transfer of electrons required for P450 activity, we therefore proceeded with the optimization of Ge2-HIS activity by exploring other endogenous metabolic factors, including heme metabolism, ER homeostasis, and NADPH generation. These modifications increased DEIN titer to a level exceeding 12 mg L −1 (Fig. 4b ), accounting for a seven-fold improvement compared with the parental strain C33. Another challenge for isoflavonoid production lies in overcoming the intrinsically low catalytic efficiency and/or selectivity of enzymes participating in the biosynthesis of plant secondary metabolites 78 . Gene amplification, by for example promoter engineering, is one approach to enhance enzyme activity. Here, implementation of dynamic expression control using inducible GALps , which enable a higher level of gene transcription than constitutive promoters 79 , boosted LIG production to 37.6 mg L −1 (Fig. 5b ), a 284% increase relative to strain C09 having constitutive expression of the pathway genes. Spatial micro-compartmentalization via the formation of metabolons, which are ordered complexes of enzymes participating in sequential biosynthetic pathways, allows the effective formation of specialized metabolites and has shown to reduce metabolic crosstalk in plants 80 . To advance DEIN titers further, we therefore mimicked this natural phenomenon by bringing enzymes into proximity, using a linker-based fusion enzyme strategy, in turn greatly improving the metabolic flux through the LIG pathway and increasing its titer by 107% (Fig. 5b ). Besides the AAA-derived p -HCA, de novo isoflavonoid biosynthesis consumes malonyl-CoA, whose formation is predominately invested in FAs synthesis in S. cerevisiae 61 . By fine-tuning the expression of key enzymes involved in FAs synthesis, we were able to redistribute the cellular malonyl-CoA pool, resulting in a 20% further increase in DEIN titer (Fig. 6f ). In conclusion, as a proof-of-concept study, a final DEIN titer of 85.4 mg L −1 was achieved using glucose as the sole carbon source in shake flask cultivations (Fig. 6g ). This production level is comparable and, in some cases, higher than isoflavonoid levels produced by previous studies, which have additionally been aided with precursor feeding (Supplementary Table 2 ). Via further expression of different glycosyltransferases, approximately 80 mg L −1 of C - or O -glycosylated bioactive compounds PIN or DIN were produced (Fig. 7c ), showing the application potential of our platform strain. Moreover, our work sheds light on the complete microbial biosynthesis of value-added isoflavonoids such as DEIN-derived legume phytoalexins 81 and may be applied in characterizing novel metabolic enzymes for the production of isoflavonoid derivatives. Additional improvements on the catalytic efficiency and specificity of key isoflavonoid biosynthetic enzymes through protein engineering 78 , directed pathway evolution facilitated by biosensor-mediated high-throughput screening as well as engineering of extracellular transport of isoflavonoids 82 , 83 , may further optimize the phenotypes of our platform strains, including higher titer/productivity and reduction/elimination of byproducts, to meet industrial-scale production requirements in the future. Finally, the multi-faceted framework we herein present also offers the potential to be applied for engineering the biosynthetic pathways in other microbial hosts as well, for the production of complex natural products."
} | 3,762 |
29146959 | PMC5691036 | pmc | 3,670 | {
"abstract": "The current paradigm, widely incorporated in soil biogeochemical models, is that microbial methanogenesis can only occur in anoxic habitats. In contrast, here we show clear geochemical and biological evidence for methane production in well-oxygenated soils of a freshwater wetland. A comparison of oxic to anoxic soils reveal up to ten times greater methane production and nine times more methanogenesis activity in oxygenated soils. Metagenomic and metatranscriptomic sequencing recover the first near-complete genomes for a novel methanogen species, and show acetoclastic production from this organism was the dominant methanogenesis pathway in oxygenated soils. This organism, Candidatus Methanothrix paradoxum, is prevalent across methane emitting ecosystems, suggesting a global significance. Moreover, in this wetland, we estimate that up to 80% of methane fluxes could be attributed to methanogenesis in oxygenated soils. Together, our findings challenge a widely held assumption about methanogenesis, with significant ramifications for global methane estimates and Earth system modeling.",
"introduction": "Introduction Modeling and biological studies investigating methane flux from wetlands discount microbial methane production in surface, oxic soils 1 , 2 . The basis of this assumption is that critical methanogen enzymes are inactivated by oxygen and methanogens are poor competitors with other microorganisms for shared substrates 3 , 4 . Because of the assumed physiological constraint that oxygen has on methanogens, global terrestrial biogeochemical models limit soil methane production in the presence of dissolved oxygen (DO) 5 . Recent reports present an alternative view that in some ecosystems methanogenesis also occurs in oxic environments, known as the methane paradox. In freshwater lakes, isotopic and molecular biology techniques provided evidence for the presence and activity of methanogens in well-oxygenated portions of the water column 6 – 8 . Similarly, isotopic signatures in oxygenated soils and activity measurements from soil laboratory enrichments have provided intriguing evidence for methanogenesis in soils with up to 19% oxygen 9 , 10 . Despite this mounting, indirect evidence, comprehensive genomic investigations that link methanogens to methane production in any oxic habitat in situ are lacking. Here we analyze observations from the Old Woman Creek (OWC) National Estuarine Research Reserve, a freshwater wetland at the shore of Lake Erie in Ohio. In this study, we experimentally assess biological methane production and emission in freshwater wetland soils across multiple spatial and temporal gradients. The results presented here provide the first ecosystem-scale demonstration of methane production in bulk-oxic soils, its microbial drivers, and the global significance of this currently underappreciated process."
} | 713 |
39805819 | PMC11731047 | pmc | 3,673 | {
"abstract": "Recently, the biologically inspired intelligent artificial visual neural system has aroused enormous interest. However, there are still significant obstacles in pursuing large-scale parallel and efficient visual memory and recognition. In this study, we demonstrate a 28 × 28 synaptic devices array for the artificial visual neuromorphic system, within the size of 0.7 × 0.7 cm 2 , which integrates sensing, memory, and processing functions. The highly uniform floating-gate synaptic transistors array were constructed by the wafer-scale grown monolayer molybdenum disulfide with Au nanoparticles (NPs) acting as the electrons capture layers. Various synaptic plasticity behaviors have been achieved owing to the switchable electronic storage performance. The excellent optical/electrical coordination capabilities were implemented by paralleled processing both the optical and electrical signals the synaptic array of 784 devices, enabling to realize the badges and letters writing and erasing process. Finally, the established artificial visual convolutional neural network (CNN) through optical/electrical signal modulation can reach the high digit recognition accuracy of 96.5%. Therefore, our results provide a feasible route for future large-scale integrated artificial visual neuromorphic system.",
"conclusion": "Conclusions In summary, we design a highly uniform artificial visual neural network based on wafer-scale single-layer MoS 2 floating-gated field effect transistors array, which is demonstrated as a 28 × 28 devices array within a 0.7 × 0.7 cm 2 area. Each device exhibits stable and superior optoelectronic performance to simulate the plasticity of the visual synapse, thus allowing the integrated array of artificial synaptic devices to simulate human visual neural network. The biomimetic processes of perceiving, remembering, and forgetting visual signals are efficiently replicated through these artificial synaptic devices. The constructed artificial synaptic neural network offers high integration, stable uniformity, outstanding parallelism, and high efficiency. Through programming of optical signals and erasing images via electrical signals, it is anticipated that the ability to process optoelectronic signals in parallel will significantly augment the performance of future-generation computers. Furthermore, the synaptic weight updates regulated through the light signals has been leveraged for handwritten image recognition, achieving a high recognition accuracy of up to 96.5%. In order to avoid connection instability caused by complex circuits, we adopted the probe station to measure the properties of the devices. However, designing appropriate supporting circuits can help to make the testing more convenient and efficient, which is the direction we will further optimize. In short, the accomplishment underscores the network’s potential applicability in deep learning scenarios, highlighting its utility in advancing computational technologies.",
"introduction": "Introduction In the human visual system, information acquisition and processing are carried out within the same framework. However, current advancements in neuromorphic vision technology are impeded by challenges such as high circuit complexity, increased power consumption, low efficiency, and device miniaturization, primarily due to the physical separation between signal devices and processing units. Prior studies indicate that employing a single device to simulate the biomimetic vision systems falls short of the requirements for parallel processing of visual information 1 – 10 . Therefore, there is a pressing need for biologically inspired strategies, such as the smarter visual perception and computational processing design solutions 11 . The biological visual system has its natural advantages, optical information perceived by the eyes is transmitted between neurons through synapses, ultimately reaching the neural network in the cerebral cortex for further processing and memory learning. Synapses characterized by low-power consumption (approximately 10 fJ per synaptic event), high speed and parallelism can surpass the properties of the supercomputers 12 – 26 . As a result, the rapid development of the bio-inspired optoelectronic synapses that combine light detection and synaptic plasticity learning capabilities is necessary to achieve an advanced artificial vision system with intrinsic visual perception and neuromorphic computational behavior. Recently, transition metal dichalcogenides (TMDs) have shone brightly in semiconductor materials. Especially, single-layer MoS 2 is an emerging two-dimensional (2D) TMDs material with great potential in future artificial intelligence electronic devices due to its excellent electrical and optical properties, which has aroused great interest 26 – 33 . Moreover, its extraordinarily high surface-to-volume ratio enhances the device’s sensitivity to charge presence 34 – 36 . Previously, artificial synaptic devices based on 2D materials and their heterostructures only focus on a single device as a-proof-of-concept, and large-scale synaptic devices array was rarely investigated, which was restricted by the size of the materials 5 , 37 – 41 . Furthermore, the design and manufacturing of 2D synaptic devices usually involve complex processes, especially the floating-gate devices, posing challenges to achieving highly uniform and scalable synaptic arrays. Unlike heterojunction transistors, which start from the perspective of channel materials, floating-gate transistors aim to achieve storage performance through substrates. And the high-k material HfO 2 was used to adjust devices with lower gate voltage, which is expected to achieve lower power consumption 33 , 42 – 48 . Nonetheless, it is necessary to promote the development of artificial neural networks (ANNs) through constructing large-scale high-density integrated synaptic devices array to simulate the function of the biological vision at the hardware scale. Large-scale 2D material integrated device arrays based on floating-gate transistors can simulate biological vision systems from mechanism to effect, promoting the development of artificial vision neural networks. In this article, we report a highly uniform artificial visual neural network based on a wafer-scale monolayer MoS 2 floating-gate field-effect transistors array (28 × 28 devices, 0.7 × 0.7 cm 2 ), enabling the integrated application for simulating human visual neural networks. According to the Au NPs floating-gate layer capture and slowly release charges, thereby effectively realizing synaptic plasticity, including EPSC and PPF through the single device. Immediately after, the artificial synaptic device simulated the biomimetic processes of visual signal perception, memory, and processing functions. After testing and analyzing 784 devices, the uniform performance of the devices array revealed a stable on/off rate of ~10 6 and the mobility of ~8 cm 2 V −1 s −1 , highlighting the exceptional optoelectronic synaptic performance. In addition, by applying optical spikes to the device array, the emblem of Beijing Institute of Technology was successfully encoded into a 28 × 28 synaptic devices array. The ability to preserve image information of badges over a long period of time is attributed to the slow discharge of floating gate layer charges. By using an array of optical pulse scanning devices, we were able to sequentially program letter images into the device array, demonstrating the array’s ability to fast write and erase process, and the ability to process photo/electric signals in parallel of the device. Finally, the fabricated artificial synaptic neural network, based on the devices array, leverages both light and electrical spikes for synaptic weight adjustments, achieving a high digit recognition accuracy of 96.5%, proving the network’s potential for image recognition applications. The exceptional performance of the integrated artificial synapses array based on MoS 2 devices marks a significant stride towards practical applications within a device-to-system level simulation framework, and heralds promising prospects for applications in brain-inspired learning and memory, artificial visual nervous systems, and neural morphology computing.",
"discussion": "Results and discussion Visual perception is a pivotal sense for humans, with approximately 80% of the external information we receive coming through our eyes 29 , 49 , 50 . The process of visual perception begins in the retina, where neural cells detect and convert light signals into electrical spikes. These spikes are then transmitted through the optic nerve to the visual cortex located at the back of the human brain, as illustrated in Fig. 1a . The intricate process allows humans to perceive, interpret, and interact with their surrounding environment. Mirroring the biological process, the artificial optic-neural network composed of the 28 × 28 synaptic devices array (~0.7 × 0.7 cm 2 ) can simply simulate the process of image recognition, as depicted in Fig. 1b . The detailed floating-gate structure of the device manufactured on the Si substrate is presented in Fig. 1c . Initially, a 20 nm HfO 2 as the insulating layer was deposited on the substrate using atomic layer deposition (ALD), followed by the thermal deposition of 2 nm discontinuous Au film. The average size of the Au NPs was approximately 15 nm, as shown in Supplementary Fig. S1 . The subsequent 10 nm HfO 2 was deposited as the tunneling layer. Then, the monolayer MoS 2 grown on the sapphire was transferred onto the pre-prepared silicon substrate using wet etching technology and transferred onto a prefabricated silicon substrate. As shown in the Supplementary Fig. S2 , we used large-area wafer grade MoS 2 , which exhibited excellent uniform characteristics. The Raman spectra of the transferred MoS 2 film in Supplementary Fig. S3 shows two characteristic peaks E 1 2g at ~384 cm −1 and A 1g at ~404 cm −1 , corresponding to the in-plane vibration and out-of-plane phonon coupling modes of MoS 2 , respectively. The ~20 cm −1 difference between two peaks indicates the monolayer properties of the MoS 2 film prepared by the CVD method 36 , 51 – 53 . Additionally, the atomic force microscopy (AFM) image reveals a smooth MoS 2 surface without significant defects or contamination, with the channel material thickness measured to be ~0.48 nm (Supplementary Fig. S4 ). After the standard UV lithography, reaction ion etching and electrodes (Cr/Au, 5/35 nm) deposition, the MoS 2 devices array with source-drain terminals were patterned. The meticulously designed artificial synaptic network not only embodies the principles of human visual perception but also paves the way for advanced applications in artificial intelligence and neuromorphic computing. Fig. 1 Human visual system and artificial synaptic devices array. a Schematic diagram of the human visual perception system, including neural networks in the human lens, hemispherical retina, optic nerve, and visual cortex. b Optical image of the 28 × 28 device array. c 3D schematic diagram of the artificial synaptic device structure based on the MoS 2 floating-gate device Synapses represent the critical junctions within the human nervous system that connect two neurons for information transmission. Figure 2a depicts a schematic diagram illustrating the synaptic operation. Considering the hysteresis window shown in Supplementary Fig. S5 , the Au NPs layer plays a pivotal role in electron capture in contrast to the devices without the floating-gate layer in Supplementary Fig. S6 . There is no gate voltage applied in Fig. 2 . Figure 2b portrays the generation of the postsynaptic current by the artificial synapse under the stimulation of the optical spike (λ = 520 nm, spike width t = 0.2 s, source/drain voltage V ds = 1 V, optical power density P = 20 mW/cm 2 ). The EPSC reached ~11 nA after a single learning session. The channel’s response to light serves as the analog to the input terminal of the presynaptic membrane, while the current in the channel between the source-drain electrodes can be defined as the EPSC of the postsynaptic membrane. Upon stimulation by the light spike, the MoS 2 channel exhibited an increase in current due to photo-generated charges. A portion of these charges were subsequently captured by tunneling into the floating-gate layer. The removal of light facilitated the gradual release of captured charges, resulting in a progressive return of the current to its original state. Analogous to the biological neural signal transduction, the modulation of channel conductivity and the capture of charges by Au NPs facilitate synaptic weight and neurotransmitter transmission alteration, respectively. The variation in postsynaptic current with different spike widths (P = 20 mW/cm 2 , t ranging from 0.1 s to 0.5 s) was examined, revealing an increase in EPSC from 0.9 to 58 μA with incremental spike width, corresponding to the enhanced excitatory postsynaptic potential with prolonged stimulation, indicative of increased neurotransmitter release, as illustrated in Fig. 2c . Additionally, the postsynaptic current was measured across varying optical power densities (P ranging from 0.02 to 20 mW/cm 2 ) in Fig. 2d , with the EPSC evidently increased from ~0.03 to ~5 μA with different optical density. Supplementary Fig. S7 shows the output curves of the device gated by different optical power densities ranging from 40 to 0 mW/cm 2 with a step of −8 mW/cm 2 . The output curves show the characteristic of the device current decreasing as the optical power density weakens. Supplementary Fig. S8 displays the corresponding transfer curves of the floating-gate transistor under assorted optical power densities, corroborating the charges capture characteristics of the Au NPs floating-gate layer. The applied optical spike with different power can modulate the charge capture effect between the gate dielectric and the channel, thereby modifying the channel conductance of the device. Further, the frequency dependence of the EPSC was explored through the application of 20 optical spikes, revealing a gradual increase in EPSC with frequency escalation, as depicted in Supplementary Fig. S9 . The experimental phenomena indicate the artificial synapse’s capability to simulate enhanced learning processes observed in biological organisms. Fig. 2 Plasticity of the artificial optoelectronic synapses. a Schematic diagram of biological synapses. b EPSC generated by pulsed laser-induced electrical artificial synapses, with a spike width of 0.2 s and a source drain voltage of 1 V. c EPSC generated by different widths of optical spikes. d EPSC generated by different optical power densities. e The synaptic transistor triggered by a pair of optical spikes with a duration of 0.4 s. f The PPF index of the EPSC plotted as a function of the spike interval (ΔT). g Simulation of the “learning-forgetting-relearning-forgetting” process using different optical spikes stimulation In the biological nervous system, PPF reflects the dynamic enhancement of postsynaptic current, closely related to learning, memory, and information-processing functions. 1 \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$${PPF\\; index}={A}_{2}/{A}_{1}=\\left[1+{C}_{1}{e}^{\\left(-\\frac{\\triangle T}{{\\tau }_{1}}\\right)}+{C}_{2}{e}^{\\left(-\\frac{\\triangle T}{{\\tau }_{2}}\\right)}\\right]\\times 100 \\%$$\\end{document} PPF index = A 2 / A 1 = 1 + C 1 e − ∆ T τ 1 + C 2 e − ∆ T τ 2 × 100 % Here, A 1 and A 2 represents the peak intensities of EPSCs for the first and second spikes, respectively. C 1 and C 2 is separately defined as the original facilitation magnitudes of two phases, while τ 1 and τ 2 denote the corresponding relaxation times for C 1 and C 2 . The PPF index is calculated to be ~160%, a value exceeding 100%, which signifies an augmented number of electrons induced by photons during the stimulation by the second spike, as illustrated in Fig. 2e, f . Moreover, the PPF index rapidly decrease from 160%, and then gradually approach the value of 100% as the inter-spike interval (ΔT) increases, which can be accurately modeled by a specific formula. The relaxation times, τ 1 and τ 2 , were calculated ~87 and 4506 ms, respectively, values that are in alignment with the relaxation times observed in biological synapses. The PPF effect, which amplifies the excited photocurrent and extends the decay period, enables the high-performance photoelectric synapses to simulate the typical “learning-memory-forgetting” behavior by two continuous optical spikes sequences with a 200 ms interval, as depicted in Fig. 2g . Initially, the device was irradiated with 40 spikes (λ = 520 nm, t = 0.2 s, ΔT = 0.2 s, V ds = 1 V, P = 6 mW/cm 2 ) to mimic the first learning process. The synaptic current started to increase to be saturated (EPSC ≈ 1.3 μA), mirroring the phenomenon of the human brain tending to become fatigued after repeated learning processes. Subsequently, the current decayed spontaneously to a moderate level after removing the light stimulus, which is consistent with the forgetting behavior over time. During the second learning process, only 20 identical spikes can reach the same current (synaptic weight) as in the first learning process, indicating that the relearning process requires less time. In addition, the decrease of synaptic weight during the second forgetting process (~0.29 μA) is weaker than the current during the first forgetting process (~0.22 μA), akin to the long-term memory ability of humans after repeated learning experiences. In the first learning process, some holes in the Au NPs layer are firmly occupied by electrons from the MoS 2 channel through the Fowler-Nordheim tunneling under light excitation. The bounded electrons cannot quickly return to the channel in the first forgetting process, leading to that the current cannot decay to the original station. Subsequently, when the second light pulse was applied, due to the bounded electrons in the Au NPs layer coming from the first learning process, the better memory state (higher current) can be obtained (~0.29 μA > ~ 0.22 μA) 33 , 54 , 55 . Figure 3 shows the detailed working principle of the optical synaptic device. Initially, the charges in the device were in a state of equilibrium, with the Fermi level of the MoS 2 channel remained unaffected. Upon exposure to light, a significant number of electron-hole pairs were generated on the surface of the MoS 2 , leading to an accumulation of electrons within the channel and a consequent downward bending of the energy band 33 , 54 , 55 . Simultaneously, some induced electrons were captured by Au NPs floating-gate layer via the Fowler-Nordheim tunneling effect. After removing the optical spike, even though the dissipation of the enhancement effect, the electrons captured in the floating-gate layer were unable to immediately return back, resulting in a gradual depletion of electrons within the MoS 2 channel. Subsequently, the electrons detained in the floating-gate layer slowly diffused back to the MoS 2 channel layer over time, culminating in gradual recovery of PSC. At this stage, the PSC cannot quickly recover to its initial position due to electrons being bound by the floating-gate layer. Ultimately, the device reverted to its initial state following the complete depletion of the captured electrons. The electrical signals, sequentially pass through Au NPs floating-gate layer, HfO 2 tunnel layer, and MoS 2 channel, simulate presynaptic, synaptic, and postsynaptic process, respectively. And the entire process is similar to the behavior of EPSC in biological synapses. Fig. 3 Schematic diagrams of the working mechanism of the optical synaptic transistor. The band energy of the device is regulated by the laser. Panels a – d separately demonstrates the working state of the device before applying laser, under 520 nm laser irradiation, after removing the laser, and slow recovery to the initial state Building on the exceptional synaptic performance exhibited by the single device, we fabricated a 28 × 28 artificial synaptic devices array on a single chip and undertook a comprehensive statistical analysis to assess their stability and uniformity. Figure 4a illustrates a schematic diagram of the artificial photoelectric synaptic devices array being illuminated by a 520 nm laser. Subsequent to the measurement of transfer curves across all devices, we derived statistical switch ratios and mobilities as shown in Fig. 4 b, d. The consistent on/off ratios of the 28 × 28 devices array can exceed ~10 6 , with the average mobility of ~8 cm 2 V −1 s −1 , demonstrating the uniformity and indicating the potential for optic-neural network applications. Therefore, we selected a 5 × 7 synaptic devices array to simulate the process of biological synaptic learning and forgetting process. Figure 4c demonstrates the devices array to stimulate synaptic characteristics by applying varying numbers of consecutive laser spikes to induce multi-pulse EPSCs, further affirming the high homogeneity of the synaptic device array through statistical analysis. We calculated the means and the standard deviations of the on/off ratio, mobility, and EPSCs in Supplementary Table S1 . Through numerical analysis of the means and standard deviations, it can be demonstrated that the device array has highly uniform performance. Fig. 4 Excellent and uniform performance of 784 optoelectronic synapses. a Schematic diagram of the optical synaptic devices array illuminated by 520 nm light. b Statistical results of on/off ratios for 784 synaptic devices. c Statistical analysis of EPSC generated by the multi-pulse stimulation. d Statistical analysis of the mobilities of the devices array In our study, we successfully encoded the emblem of Beijing Institute of Technology into a 28 × 28 array of synaptic devices by applying 50 spikes with a power density of 20 mW/cm 2 , as illustrated in Fig. 5a . The spike width and interval time were uniformly set at 0.2 s and 0.2 s, respectively. By individually scanning the synaptic devices using optical signals, the image information of the emblem can be precisely stored in the synaptic devices array. The EPSCs recorded at time intervals of 0, 20, 40, and 60 s were approximately 3.3, 1.8, 0.6, and 0.1 μA, respectively. Furthermore, the results demonstrate the capability of the synaptic devices array to retain the image information for durations extending beyond 60 seconds is attributed to the slow discharge of charges from the floating-gate layer. The feature facilitates the perception and preservation of image information within the synaptic devices array. Moreover, to simulate the enhancement and inhibition of biological stimuli more intuitively, we applied promotion (50 optical spikes under the same conditions as those used for encoding the emblem, as shown in Fig. 5b (t = 0.2 s, ΔT = 0.2 s, V ds = 1 V, P = 20 mW/cm 2 ), and inhibition spikes (V g = −5 V for 10 s) were applied, and the corresponding EPSCs of synaptic devices were recorded as ~3.2 μA. By varying the location of the applied light pulse, we were able to sequentially program the images of the letter “B”, “I”, and “T” into the device array, demonstrating the array’s ability to acquire and distinguish between different images. The capability underscores the potential of our synaptic devices array in the domain of optical information processing and storage, mirroring the dynamic and adaptable nature of biological synaptic functions. Fig. 5 Synaptic devices array used for iconic memory and analog recognition. a Implemented 28 × 28 synaptic arrays as the trainable memory. The image of the emblem of Beijing Institute of Technology was input into the memory array using 50 spikes of laser irradiation with P = 20 mW/cm 2 , t = 0.2 s and ΔT = 0.2 s. b Applying different voltage spikes to write and erase the letter “B”, “I”, and “T”. c Schematic diagram of a simulated CNNs with an input layer that captures a 28 × 28-pixel image, a convolution layer with a 3 × 3 kernel, an average pooling layer followed by a fully connected layer and output layer. d The recognition accuracy of the visual signal stimuli for different training iterations under different light illumination To further explore the impact of light illumination on the learning capability of the device artificial synaptic array, the recognition of the National Institute of Standards and Technology (MNIST) handwritten digit database was simulated using a multilayer perceptron network, drawing on the measurements in Supplementary Fig. S10 56 , 57 . The schematic diagram in Fig. 5c can demonstrate CNNs based on model designed for recognizing a bunch of images of handwriting numbers from the MNIST dataset. The artificial visual neural network, which draws inspiration from the human retina, is structured in three layers: the photoreceptor layer, the intermediate neural cell layer (encompassing bipolar, horizontal, and amacrine cells), and the ganglion cell layer. The network architecture includes an input layer that captures a 28 × 28-pixel image, a convolution layer with a 3 × 3 kernel, a max pooling layer followed by a fully connected (FC) layer and an output layer (Fig. 5c ). As shown in Fig. 5d , a recognition accuracy only of 83.1% was achieved under light illumination at a optical power density of 5 mW/cm 2 . Previous studies suggested that higher recognition accuracy correlates with larger ratios of maximum to minimum conductance (G max /G min ) and good linearity 58 – 60 . The synaptic device under light illumination with the optical power density of 10 mW/cm 2 and 15 mW/cm 2 achieved larger G max /G min and smaller nonlinearity (NL) (calculated from the long term depression/potentiation (LTD/LTP) curves (LTP: 40 optical spikes (t = 0.2 s, ΔT = 0.2 s), LTD: 40 electrical spikes (V g = -3 V, t = 0.2 s, ΔT = 0.2 s)) and fitting of equations in Supplementary Fig. S10 ), culminating in enhanced recognition accuraciy of 91.3% and 96.5%, respectively. Therefore, this study demonstrates that adjusting the power density of light illumination can significantly enhance recognition accuracy, thereby offering a viable method for optimizing the performance of artificial visual neural networks."
} | 6,658 |
28844883 | PMC5790576 | pmc | 3,674 | {
"abstract": "Increased environmental pollution has necessitated the need for eco-friendly clean-up strategies. Filamentous fungal species from gold and gemstone mine site soils were isolated, identified and assessed for their tolerance to varied heavy metal concentrations of cadmium (Cd), copper (Cu), lead (Pb), arsenic (As) and iron (Fe). The identities of the fungal strains were determined based on the internal transcribed spacer 1 and 2 (ITS 1 and ITS 2) regions. Mycelia growth of the fungal strains were subjected to a range of (0–100 Cd), (0–1000 Cu), (0–400 Pb), (0–500 As) and (0–800 Fe) concentrations (mgkg −1 ) incorporated into malt extract agar (MEA) in triplicates. Fungal radial growths were recorded every three days over a 13-days’ incubation period. Fungal strains were identified as Fomitopsis meliae , Trichoderma ghanense and Rhizopus microsporus . All test fungal exhibited tolerance to Cu, Pb, and Fe at all test concentrations (400–1000 mgkg −1 ), not differing significantly ( p > 0.05) from the controls and with tolerance index >1. T. ghanense and R. microsporus demonstrated exceptional capacity for Cd and As concentrations, while showing no significant ( p > 0.05) difference compared to the controls and with a tolerance index >1 at 25 mgkg −1 Cd and 125 mgkg −1 As. Remarkably, these fungal strains showed tolerance to metal concentrations exceeding globally permissible limits for contaminated soils. It is envisaged that this metal tolerance trait exhibited by these fungal strains may indicate their potentials as effective agents for bioremediative clean-up of heavy metal polluted environments.",
"conclusion": "Conclusion Indigenous filamentous fungal species from gold and gemstone mine sites exhibited remarkable tolerance in heavy metal-rich media. Exposures of F. meliae , T. ghanense and R. microsporus to elevated Cu, Pb and Fe levels revealed high tolerance with index values >1. Furthermore, T. ghanense and R. microsporus demonstrated extraordinary tolerance for As and Cd concentrations with a tolerance index >1 at 25 mgkg −1 Cd and 50 mgkg −1 As. These exceptional traits displayed by these fungal species to elevated heavy metal levels may indicate the bioremediative potentials inherent in the indigenous filamentous fungal species.",
"introduction": "Introduction Increased heavy metal contamination of soil and water environments 1 has necessitated the need for clean-up strategies. Recently, diverse eco-friendly remediation options have been explored for the restoration of contaminated environments. These remediation options, among others, include the use of plants (phytoremediation), 2 bacteria (bacterial bioremediation) 3 and fungi (mycoremediation). 4 The employability of these bio-resources (plants, bacteria and fungi) for effective bioremediation has been well reported. 2 , 3 , 4 At present of these options, mycoremediation strategy has received increased attention in the bioremediation of contaminated/polluted environments due to its reasonably low cost implications and significant success outcomes. 5 , 6 , 7 , 8 Filamentous fungal species have been identified for their distinct attributes (ability to thrive under extreme pH, temperature and nutrient variability conditions, as well as tolerance to high metal concentrations) 9 , 10 , 11 and hence their effective remediation traits of contaminated sites. Metal tolerance/resistance has been defined as the ability of an organism to survive metal toxicity by means of one or more mechanisms devised in direct response to the metal(s) concerned. 7 , 12 Metal tolerance by filamentous fungi has been associated with their sites of isolation, toxicity of the metal tested, its concentration in medium, and on the isolate's competence. 10 Contaminated sites are known as principal sources of metal-resistant species 18 , 19 , 20 , 21 , 22 with indigenous fungal strains isolated from heavy metal contaminated sites exhibiting notable tolerance for high heavy metal concentrations. 9 , 21 , 23 , 24 , 25 However, of more importance is the specific and nonspecific heavy metal tolerance mechanisms adopted by fungal species. According to Vadkertiova and Slavikova 13 the introduction of heavy metals into the environment has induced physiological and morphological adaptation strategies in the microbial community. Specifically, fungal species adopt one or more metal tolerance strategies which include extracellular metal sequestration and precipitation, suppressed influx, enhanced metal efflux, production of intracellular/extracellular enzymes, metal binding to cell walls, intracellular sequestration and complexation. 14 , 15 , 16 , 17 Several metal-tolerant filamentous fungi ( Rhizopus , Trichoderma , Aspergillus , Penicillium , and Fusarium ) have been isolated from multiple heavy metal contaminated soils. 7 Zafar et al. 7 reported that Rhizopus sp., isolated from metal-contaminated agricultural soils tolerated Cd and Cr concentrations. In addition, Volesky 26 observed that the mycelium of a Rhizopus specie was biosorbent towards Pb, Cd, Cu and Zn. Trichoderma species have also been known to exhibit tolerance to a range of toxicants 27 , 28 , 29 and Cu, Cd, As and Zn heavy metals in in vitro conditions. 8 , 23 , 27 , 30 , 31 , 32 , 33 , 34 However, there is a dearth of knowledge of the growth response and heavy metal tolerance of filamentous fungal species isolated from gold and gemstone mining sites. This study was therefore designed to isolate, identify and assess the growth response and tolerance/resistance of filamentous fungi isolated from gold and gemstone mining sites to varied concentrations of selected heavy metals associated with mining sites.",
"discussion": "Discussion The occurrence of F. meliae , T. ghanense and R. microsporus on heavy metal contaminated gold and gemstone mining sites was confirmed in this study. The presence of fungal species in various contaminated/polluted sites with elevated heavy metal concentrations has been well documented. Specifically, Zafar et al. 7 and Fazli et al. 43 reported the occurrence of fungal strains in soils with elevated Cd, Cu, As and Zn concentrations. In addition, Anand et al. 9 and Karcprzak and Malina 29 confirmed the presence of fungi in heavy metal polluted soils. Iram et al., 12 Iskandar et al., 21 and López and Vázquez 23 also affirmed the occurrence of fungal species in sewage sludge water plants, heavy metal contaminated freshwater ecosystem and sewage and industrial waste waters respectively. Furthermore, Mo et al., 44 Srivastava et al. 45 and Babu et al. 46 confirmed the existence of fungal strains in Pb and As polluted sites and mine tailings soils. Of more importance is the marked tolerance displayed by these fungal species to heavy metals. Fungal species tolerate metals 6 , 15 , 47 and thrive at elevated metal concentrations. 9 , 24 , 48 In particular, indigenous filamentous fungi isolated from contaminated sites have shown tolerance to heavy metals. 12 , 18 , 19 , 49 This exceptional trait may be attributed to the isolates’ tolerance strategies to elevated heavy metal contaminations. Fomina et al., 14 Turnau et al., 15 Gadd 16 and Vala and Sutariya 17 reported that these tolerance mechanisms include metal binding to cell walls, production of intracellular/extracellular enzymes, intracellular sequestration, extracellular metal sequestration and precipitation, suppressed influx, enhanced metal efflux, and complexation. Remarkable heavy metal tolerance was demonstrated by R. microsporus and T. ghanense species. Trichoderma and Rhizopus species have been widely reported for their notable tolerance to various heavy metals at varied concentrations. 7 , 23 , 31 , 50 Some strains of Rhizopus and Trichoderma revealed high resistance to a range of heavy metals, such as Cd, 11 , 23 , 26 , 44 , 50 , 51 Cu, 21 , 26 , 46 Pb 26 , 52 and As. 17 , 45 Vala and Sutariya 17 reported that Rhizopus species were highly tolerant to 25 and 50 mgkg −1 As concentrations, which confirms the findings of this study. In addition, strains of Trichoderma tolerated Cd at 100 and 125 mgkg −1 53 , 54 and Cu at 300 mgkg −1 , 23 500 mgkg −1 , 9 800 mgkg −1 21 and 1000 mgkg −1 45 concentrations. Furthermore, a strain of Trichoderma was found to tolerate Pb concentrations of 1000 mgkg −1 in medium. 21 All three fungal species demonstrated extraordinary preference for Fe at all concentrations as observed in their tolerance index values. This may be ascribed to the fact that iron serves as a micronutrient and is crucial in many metabolic processes. 55 In addition, Kosman, 56 Philpott 57 and Johnson 58 found that fungal species have a high affinity and capacity to take up Fe in various forms and variety. Aznar and Dellagi 55 and Neilands 59 stated that most fungal strains synthesize and secrete siderophores (small organic compounds that bind ferric Fe with high affinity and specificity) which they utilize to extract Fe from their environment. Furthermore, according to Kosman 56 microorganisms including fungi basically deploy three main strategies to increase iron solubility by acidifying the environment, reducing ferric iron to a more soluble ferrous form and secreting soluble iron-chelating molecules. Overall, T. ghanense and R. microsporus exhibited higher tolerance in Cd, Cu, Pb and As-enriched media compared to F. meliae which specifically displayed sensitivity to all Cd and As concentrations. Studies confirm that differing levels of metal resistance have been demonstrated by different fungal species isolated from the same source of metal-contaminated sites. 12 , 60 , 61 , 62 , 63 , 64 This may be ascribed to variations in the tolerance mechanism utilized by the fungal species 7 which is individually dependent. 65 In addition, the evident sensitivity to all Cd and As concentrations displayed by F. meliae may be attributed to the known toxicity of these heavy metals as reported. 66 , 67 , 68"
} | 2,513 |
39503721 | PMC11539682 | pmc | 3,675 | {
"abstract": "Abstract Malonyl‐coenzyme A (CoA) is a key precursor for the biosynthesis of multiple value‐added compounds by microbial cell factories, including polyketides, carboxylic acids, biofuels, and polyhydroxyalkanoates. Owing to its role as a metabolic hub, malonyl‐CoA availability is limited by competition in several essential metabolic pathways. To address this limitation, we modified a genome‐reduced Pseudomonas putida strain to increase acetyl‐CoA carboxylation while limiting malonyl‐CoA utilization. Genes involved in sugar catabolism and its regulation, the tricarboxylic acid (TCA) cycle, and fatty acid biosynthesis were knocked‐out in specific combinations towards increasing the malonyl‐CoA pool. An enzyme‐coupled biosensor, based on the rppA gene, was employed to monitor malonyl‐CoA levels in vivo. RppA is a type III polyketide synthase that converts malonyl‐CoA into flaviolin, a red‐colored polyketide. We isolated strains displaying enhanced malonyl‐CoA availability via a colorimetric screening method based on the RppA‐dependent red pigmentation; direct flaviolin quantification identified four engineered strains had a significant increase in malonyl‐CoA levels. We further modified these strains by adding a non‐canonical pathway that uses malonyl‐CoA as precursor for poly(3‐hydroxybutyrate) biosynthesis. These manipulations led to increased polymer accumulation in the fully engineered strains, validating our general strategy to boost the output of malonyl‐CoA–dependent pathways in P . putida .",
"conclusion": "CONCLUDING REMARKS In this work, we identified and implemented gene deletions within different metabolic pathways that produce or consume malonyl‐CoA, obtaining engineered strains with an increased pool of intracellular malonyl‐CoA. While several metabolic pathways had to be modified and rewired simultaneously towards boosting the thioester levels, the one that seemed most influential is fatty acids synthesis. Interestingly, the engineered P . putida strains generated in this study had growth features comparable to the parental strain, yet a relatively small growth compromise was observed in glucose‐dependent cultivations—a consequence expected when genes involved in pathways essential for growth are deleted (Nogales et al., 2020 ). Yet, all strains reached comparable final OD 600 values, making them suitable for testing PHB biosynthesis. We demonstrated that these strains could support enhanced PHB production, with SEM11 Δ glta Δ fabF Δ pha , SEM11 Δ hexR Δ gcd Δ fabF Δ pha , and SEM11 Δ glta Δ fabF Δ pha P \n \n J23108 \n \n → fabF‐2 as the most promising candidates for subsequent rounds of metabolic engineering to improve malonyl‐CoA–dependent production. Several strategies have been previously explored in the context of metabolic engineering for PHA production in P. putida . Some examples include overexpressing phaJ in P. putida KCTC1639, which enhanced medium‐chain‐length PHA biosynthesis from octanoate by ~9% (Vo et al., 2008 ). Introducing PHA synthases from various microorganisms into P. putida KT2442 Δ pha led to the accumulation of structurally diverse PHAs (Chung et al., 2009 ). Other approaches exploited manipulating β‐oxidation; specifically, strains with deletions in both fadBA1 and fadBA2 , along with a knock‐out in phaZ (the PHA depolymerase gene), had a 20% and 100% increase in the PHA yield from p ‐coumarate and lignin, respectively, as compared to the wild‐type strain (Salvachúa et al., 2020 ). Promoter engineering has also proven effective to boost PHA production, as demonstrated in P. putida KT2440 with modified transcription levels of phaC1 and phaC2 (the PHA synthase genes). When combined with the overexpression of genes encoding components of the pyruvate dehydrogenase complex and deletion of Gcd, the PHA yield on biomass was increased by 90% (Zhang et al., 2021 ). Unlike medium‐chain‐length PHAs, PHB biosynthesis in engineered P . putida has only been reported in a few publications (Ackermann et al., 2024 ; Didion et al., 2024 ; Kozaeva et al., 2021 ), and our present study indicates that malonyl‐CoA could be harnessed as a precursor for biopolymer production. However, the scalability of these strains for industrial production has yet to be demonstrated, as is the feasibility of achieving high‐yield biopolymer production under large‐scale operational conditions. To this end, multiple factors will have to be evaluated and optimized, e.g., the robustness of the strains under changing operating conditions, potential bottlenecks in metabolic fluxes and oxygen transfer, and the overall economic viability of the bioprocess (Manikandan et al., 2021 ). Additionally, further studies are required to understand the systems‐level impact of the genetic modifications on bacterial metabolism (Tokic et al., 2020 ), which could affect the productivity across extended cultivation periods and in bioreactors of different configurations. Despite these shortcomings, this study provides a foundation for optimizing malonyl‐CoA–dependent bioproduction in P . putida , a bacterial host that is gaining attention for industrial applications (de Lorenzo et al., 2024 ).",
"introduction": "INTRODUCTION Malonyl‐coenzyme A (CoA) serves as a hub metabolite for the biosynthesis of lipids and acts as an attractive building block for the microbial production of biologically active polyketides and fatty acid‐derived compounds, including biofuels (Li et al., 2024 ). In most organisms, malonyl‐CoA is produced through acetyl‐CoA carboxylation by acetyl‐CoA carboxylase (ACC). Bacteria and plant chloroplasts possess a multi‐subunit ACC enzyme, which consists of three distinct domains responsible for catalyzing two separate reaction steps (Cronan, 2021a , 2021b ; Cronan & Waldrop, 2002 ). The primary function of malonyl‐CoA in bacterial metabolism is serving as an extender unit for fatty acids synthesis (Polyak et al., 2012 ), while its role in the production of native secondary metabolites appears to be relatively minor (Cronan & Thomas, 2009 ; McNaught et al., 2023 ). Several studies have focused on engineering model microorganisms to enhance the biosynthesis of malonyl‐CoA–derived compounds (Fowler et al., 2009 ; Liu et al., 2023 ; Milke et al., 2019 ; Milke & Marienhagen, 2020 ; Valdehuesa et al., 2013 ). In most cases, however, metabolic and regulatory bottlenecks have been encountered in the engineered strains, and productivity limitations have been identified in the associated bioprocesses. Typically, the relatively low availability of intracellular malonyl‐CoA is a primary constraint limiting product yield (Milke & Marienhagen, 2020 ), exposing the need for emerging metabolic engineering approaches to boost malonyl‐CoA levels. Redirecting fluxes through central carbon metabolism to enhance the availability of acetyl‐CoA, the immediate precursor of malonyl‐CoA, has been shown to promote the synthesis of malonyl‐CoA–derived products (Milke et al., 2018 ). Consequently, various metabolic engineering approaches have been implemented to enhance the synthesis of malonyl‐CoA while reducing the consumption of acetyl‐CoA. Some of the key metabolic and regulatory targets for manipulation are indicated in Figure 1A . FIGURE 1 Genetic modifications to boost malonyl‐CoA availability in Pseudomonas putida . (A) P. putida SEM11 was modified by knocking‐out various combinations of the four genes shown in red and reactivating a dormant FabF‐2 via promoter insertion in order to increase the malonyl‐CoA pool. In addition, the native gene cluster encoding the enzymes for PHA production and degradation was deleted to eliminate endogenous biopolymer production. Abbreviations: G6P, glucose‐6‐phosphate; 6PG, 6‐phosphogluconate; KDPG, 2‐keto‐3‐deoxy‐6‐phosphogluconate; F6P, fructose‐6‐phosphate; FBP, fructose‐1,6‐bisphosphate; G3P, glyceraldehyde‐3‐phosphate; 3PG, 3‐phosphoglycerate; DHAP, dihydroxyacetone phosphate; PEP, phosphoenolpyruvate; Cit, citrate; Acon, aconitate; Icit, isocitrate; 2‐KG, 2‐ketoglutarate; Suc‐CoA, succinyl‐CoA; Succ, succinate; Fum, fumarate; Mal, malate; and Oaa, oxaloacetate. (B) Growth curves of the resulting engineered strains. Cell density was estimated as the optical density measured at 600 nm (OD 600 ). QurvE software was used to analyze growth curves (Wirth, Funk, et al., 2023 ; Wirth, Rohr, et al., 2023 ), and the maximum specific growth rate (μ max ) was derived from the OD 600 measurements over time. GraphPad Prism 9 (GraphPad Software, Inc.) was used to perform all statistical analyses; the levels of significance are indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. The error bars represent standard deviations; n = 3. Modulating the endogenous fatty acid synthesis has been proposed as a promising approach to enhance the availability of malonyl‐CoA for polyketide synthesis (Cronan & Thomas, 2009 ). This strategy was used in a study where synthetic antisense RNAs were employed to reduce fatty acid biosynthesis, resulting in an enrichment of the malonyl‐CoA pool in Escherichia coli . Such interventions led to a relatively modest increase in the production of 4‐hydroxycoumarin, resveratrol, and naringenin (Yang et al., 2015 ). Cress et al. ( 2015 ) used a CRISPathBrick tool to modulate FadR in E. coli , a transcriptional regulator that represses β‐oxidation and positively regulates fatty acid synthesis (Cronan, 2021a , 2021b ), supporting a high malonyl‐CoA turnover and improved naringenin titers. In another examples based on engineered E. coli , RNA interference was directed against the transcription of fabB and fabF , encoding the β‐ketoacyl‐acyl carrier protein (ACP) synthases (KAS) I and II (Wu et al., 2014 ). CRISPR interference for malonyl‐CoA accumulation has also been adopted in E. coli to increase flavonoid production by targeting all possible malonyl‐CoA–related genes, with the best effects achieved via altering fabF transcription (Wu et al., 2015 ). Other engineering approaches have been successfully applied to various microbial hosts. For instance, several transcriptional regulators of phospholipid synthesis were manipulated in Saccharomyces cerevisiae to increase the production of 3‐hydroxypropionic acid (Chen et al., 2017 ), a platform chemical derived from malonyl‐CoA. Another elegant example is the combined engineering of Corynebacterium glutamicum by deregulating the expression of genes encoding ACC components, reducing the acetyl‐CoA entry into the tricarboxylic acid (TCA) cycle, and eliminating anaplerotic pyruvate carboxylation, which resulted in improved production of the pentaketide noreugenin (Milke et al., 2019 ). E . coli and Pseudomonas taiwanensis have also been engineered using similar strategies; increased production of phloroglucinol, resveratrol, and 3‐hydroxypropionic acid, respectively, was achieved through these manipulations (Chen et al., 2024 ; Schwanemann et al., 2023 ; Zha et al., 2009 ). Developing non‐traditional microbial hosts as platform strains with enhanced malonyl‐CoA supply could enable the efficient bioproduction of compounds that are challenging to produce in E . coli and other model species. Pseudomonas putida is a Gram‐negative soil bacterium (Belda et al., 2016 ; Calero & Nikel, 2019 ), which, over the years, became a biotechnological chassis (Martínez‐García & de Lorenzo, 2019 ; Schwanemann et al., 2020 ; Weimer et al., 2020 ). Owing to its robust and adaptable metabolism, P. putida can use a wide variety of structurally diverse molecules as carbon and energy sources (D'Arrigo et al., 2019 ; Fernández‐Cabezón et al., 2022 ; Turlin et al., 2022 , 2023 ). This bacterium can also handle the stress induced by toxic molecules or environmental conditions (Bitzenhofer et al., 2021 ; Nikel & de Lorenzo, 2018 ; Wirth et al., 2022 ). P. putida has already been successfully employed as a host for the production of natural products, e.g., rhamnolipids, terpenoids, polyketides, non‐ribosomal peptides, and biopolymers—as well as other bulk and specialty chemicals (Batianis et al., 2020 ; de Lorenzo et al., 2024 ; Kozaeva et al., 2024 ; Prieto et al., 2016 ; Weimer et al., 2020 ; Wirth & Nikel, 2021 ). Pseudomonas species are efficient biopolymer producers (Mezzina et al., 2021 ), especially polyhydroxyalkanoates (PHAs). PHAs comprise a large family of natural polymers that includes poly(3‐hydroxybutyrate) (PHB), the most widespread example of short‐chain‐length PHAs, and copolymers containing 3‐hydroxyvalerate (Steinbüchel et al., 1992 ) and longer carbon structures (Anderson & Dawes, 1990 ; Suriyamongkol et al., 2007 ). These bio‐based polymers emerged as an alternative to conventional plastics because they present many of the same characteristics and provide some benefits over petrochemical materials, e.g., a lower carbon footprint and more alternatives for waste disposal (Choi et al., 2020 ; Koller et al., 2017 ; Meng & Chen, 2018 ). As an intermediate derived from acetyl‐CoA, malonyl‐CoA directly participates in the synthesis of PHAs by supplying the key two‐carbon units that are ultimately incorporated into the polymer structure (Aduhene et al., 2021 ; Mitra et al., 2022 ). Hence, some engineering strategies to enhance PHA production have focused on increasing the intracellular concentration of malonyl‐CoA. For example, overexpressing the genes fabH \n \n Ec \n , fabD \n \n Ec \n , and fabD \n \n Ps \n in an engineered E. coli strain led to an increased PHA content (Taguchi et al., 1999 ). PHA production was also enhanced by increasing the fluxes through the CoA biosynthetic pathway, introducing a gene encoding the prokaryotic type III pantothenate kinase and supplementing pantothenate or β‐alanine as CoA precursors (Kudo et al., 2023 ), and by overexpressing ACC components (Wang et al., 2012 ). Based on these examples, we reasoned that altering malonyl‐CoA availability in P . putida could multiply the value of this host as a platform for PHA production. In this study, we adopted a metabolic engineering approach that started by knocking‐out genes associated with glycolytic pathways, the TCA cycle, and fatty acid biosynthesis to increase the levels of malonyl‐CoA in a genome‐reduced derivative of P . putida KT2440, strain SEM11 (Wirth, Funk, et al., 2023 ; Wirth, Rohr, et al., 2023 ). A colorimetric assay, based on a repurposed polyketide synthase, was used to screen for isolates with high malonyl‐CoA content. These strains were then employed to assess PHB production from glucose through a non‐canonical PHB biosynthesis pathway, revealing a substantial enhancement in biopolymer accumulation compared to the non‐engineered, parental strain.",
"discussion": "RESULTS AND DISCUSSION Design, engineering, and characterization of P . putida strains with enhanced malonyl‐ CoA availability We engineered strain SEM11, a genome‐reduced variant of wild‐type P. putida KT2440, through several gene deletions in order to increase the intracellular pool of malonyl‐CoA (Figure 1A ), using a well‐established method for genome engineering in Pseudomonas species based on the I‐SceI meganuclease (Martínez‐García & de Lorenzo, 2011 ; Wirth et al., 2020 ). Two genes were deleted to increase the catabolic fluxes through glycolysis: (i) hexR , encoding the transcriptional repressor that controls the expression of several glycolytic genes, including zwf‐1 (glucose 6‐phosphate dehydrogenase), the genes encoding enzymes in the Entner‐Doudoroff pathway that ultimately yield glyceraldehyde‐3‐phosphate and pyruvate ( edd , eda , and glk ), and gap‐1 , glyceraldehyde‐3‐phosphate dehydrogenase (del Castillo et al., 2007 , 2008 ; Nikel et al., 2015 ), and (ii) the gcd gene, encoding the membrane‐bound glucose 2‐dehydrogenase, responsible for sugar processing via the periplasmic oxidative route (Sudarsan et al., 2014 ; Volke et al., 2023 ). The deletion of gcd has been shown to have a positive effect on both PHA accumulation (Poblete‐Castro et al., 2013 ) and the free CoA pool in P. putida (Gläser et al., 2020 ). This gene has been also targeted to increase the malonyl‐CoA content in P . taiwanensis (Schwanemann et al., 2023 ). In order to reduce acetyl‐CoA consumption, the gltA gene (encoding citrate synthase) was also targeted for deletion, which would prevent acetyl‐CoA from entering the TCA cycle (Figure 1A ). In a previous study from our laboratory, gltA was identified as a promising candidate for CRISPR interference (CRISPRi) towards increasing the availability of acetyl‐CoA in P. putida (Kozaeva et al., 2021 ). Furthermore, deleting this gene increased malonyl‐CoA levels in P . taiwanensis (Schwanemann et al., 2023 ). Malonyl‐CoA is used in the cell mainly for fatty acid biosynthesis, and we altered the flux through this pathway by activating the transcriptionally dormant fabF‐2 gene, encoding a 3‐ketoacyl‐ACP synthase (KAS), via promoter engineering. Specifically, the medium‐strength synthetic P J23108 promoter was integrated upstream the fabF‐2 gene (PP_3303) via homologous recombination (Martínez‐García & de Lorenzo, 2011 ), thereby enabling its constitutive expression. FabF‐2 is expected to display a reduced activity compared to the very active variant encoded by fabF (Dong et al., 2021 ). Sequence and structure comparison between FabF and FabF‐2 of P. putida KT2440 indicated that the two proteins share only 48.7% identity (Figure S1 in the Supporting Information). However, the relatively high structural conservation suggests that their mechanisms may still be similar. In addition to the promoter engineering approach to activate fabF‐2 , fabF was deleted to reduce the flux of malonyl‐CoA towards fatty acid synthesis (Figure 1A ). Various combinations of gene deletions and transcriptional engineering led to ten engineered strains, obtained by pairing modifications in fatty acid synthesis (Δ fabF, P \n \n J23108 \n \n → fabF‐2 ) with deletions affecting glycolytic pathways (Δ hexR and Δ gcd ) or the TCA cycle (Δ gltA ). These strains are SEM11 P \n \n J23108 \n \n → fabF‐2 , SEM11 Δ fabF , SEM11 Δ gltA , SEM11 Δ gltA Δ fabF , SEM11 Δ gltA P \n \n J23108 \n \n → fabF‐2 , SEM11 Δ gltA Δ fabF P \n \n J23108 \n \n → fabF‐2 , SEM11 Δ hexR Δ gcd , SEM11 Δ hexR Δ gcd Δ fabF , SEM11 Δ hexR Δ gcd P \n \n J23108 \n \n → fabF‐2 , and SEM11 Δ hexR Δ gcd Δ fabF P \n \n J23108 \n \n → fabF‐2 . The native operon encoding the enzymes involved in PHA synthesis and depolymerization (de Eugenio et al., 2010 ) was knocked out in all these strains to avoid competition for precursors that support the accumulation of medium‐chain‐length PHAs. We first tested whether these gene deletions were deleterious for the P . putida strains by growing the modified strains in mineral salt medium with glucose as a carbon source (Hartmans et al., 1989 ). Growth parameters, including maximum growth rate (μ max ), extension of the lag phase, and cell density (estimated as the optical density at 600 nm, OD 600 ), were computed with the QurvE software (Wirth, Funk, et al., 2023 ; Wirth, Rohr, et al., 2023 ). The strains containing only changes in the fatty acid synthesis and those containing a lower number of deletions had minimal variations in the growth parameters when compared to the parental SEM11 strain, whereas strains with more gene deletions showed a decrease in μ max and an increase in the extension of the lag phase (Figure 1B ). The growth curves of SEM11 Δ pha overlapped with that of the reference strain (SEM11, Figure 1B ), suggesting that eliminating the gene cluster for PHA metabolism in P . putida has no substantial effect on the cell physiology under these conditions. All strains reached a maximum OD 600 comparable to SEM11 Δ pha (adopted as the reference strain henceforth, since its behavior was practically indistinguishable from that of P . putida SEM11); only SEM11 Δ gltA Δ fabF Δ pha , SEM11 Δ gltA Δ fabF Δ pha P \n \n J23108 \n \n → fabF‐2 , and SEM11 Δ hexR Δ gcd P \n \n J23108 \n \n → fabF‐2 had a slight (but statistically significant) decrease in final OD 600 compared to the reference strain (~3%). The most affected strains in terms of μ max were SEM11 Δ gltA Δ fabF Δ pha and SEM11 Δ hexR Δ gcd P \n \n J23108 \n \n → fabF‐2 , with a reduction of 41% and 43% compared to SEM11 Δ pha , respectively. These results align well with the expected perturbation of essential biochemical processes, i.e., the TCA cycle, fatty acid synthesis, and sugar processing. In particular, the gltA deletion was expected to result in several metabolic consequences that can negatively affect fitness, including reduced fluxes through the TCA cycle, an altered redox and energy metabolism, and a possible shortage of critical building‐blocks needed for synthesizing amino acids, nucleotides, and other biomass constituents (Zhou et al., 2024 ). A previous study from our laboratory analyzed the effect of knocking‐down the gltA gene in P. putida through a CRISPRi strategy (Kozaeva et al., 2021 ). A combined metabolomic and proteomic analysis showed a substantial impact on central carbon metabolism, negatively affecting some components of the EDEMP cycle (Nikel et al., 2015 , 2021 ), the glyoxylate shunt, and enzymes involved in aromatic amino acid biosynthesis (Volke et al., 2021 , 2022 ). The gcd deletion was also proven to cause a low rate of total carbon consumption and a reduced growth rate (Bentley et al., 2020 ), largely due to the role of Gcd in mediating sugar catabolism with energy conservation (Volke et al., 2023 ). Finally, deleting fabF in P. putida has been shown to cause a deficiency in unsaturated fatty acid synthesis (Dong et al., 2021 ), as FabF is responsible for extending palmitoleic acid ([9 Z ]‐hexadec‐9‐enoic acid, C 16 ) to cis ‐vaccenic acid ([11 E ]‐octadec‐11‐enoic acid, C 18 ), a key process that affects the membrane lipid composition (Do et al., 2018 ). Considering these observations, the metabolic effects of the gene modifications tested herein likely account for the impact on the growth of the modified P . putida strains. Despite the slight deleterious effect of some of the gene deletions, the resulting strains are not significantly impaired under the conditions tested. After 18 h of incubation, for instance, all strains reached a comparable final OD 600 , confirming their suitability for applications under standard growth conditions. Semi‐quantitative analysis of malonyl‐ CoA levels through an enzyme‐coupled biosensor We employed an enzyme‐coupled biosensor, based on 1,3,6,8‐tetrahydroxynaphthalene synthase (RppA), for a semi‐quantitative screening of malonyl‐CoA availability in the engineered strains. RppA converts malonyl‐CoA to flaviolin (a red‐colored compound), offering a direct colorimetric readout of malonyl‐CoA levels (Figure 2A ). The biosensor has been previously tested in E. coli , P. putida , and C . glutamicum , simplifying malonyl‐CoA detection through a rapid colorimetric assay (Yang et al., 2018 ). We integrated the rppA gene in the genome of the engineered strains using a mini‐Tn 7 transposon system that allows for the incorporation of DNA constructs in the chromosome at the att Tn 7 attachment site that is conserved in many bacterial species (Zobel et al., 2015 ). The P . putida strains were transformed with two plasmids, one containing the mini‐Tn 7 transposon system itself and a helper plasmid encoding the Tn 7 transposase required for transposon mobilization (Schweizer & de Lorenzo, 2004 ). The rppA gene was constitutively expressed from the strong P 14f promoter followed by the translational coupler BCD2 (Mutalik et al., 2013 ). We performed the integration in all engineered strains and in the reference strain SEM11 Δ pha . After growing the cells under the same conditions indicated above, culture supernatants were visually inspected as a preliminary step prior to flaviolin quantification, and some of the samples displayed an intense red color, indicative of increased intracellular malonyl‐CoA (Figure 2B ). FIGURE 2 Analysis of malonyl‐CoA levels with a biosensor. (A) A semi‐quantitative analysis of malonyl‐CoA levels in the engineered strains was carried out with an enzyme‐coupled biosensor based on the rppA gene. This gene encodes RppA, a type III polyketide synthase that converts five molecules of malonyl‐CoA into one molecule of flaviolin, which displays a red color. Malonyl‐CoA is first converted to THN by RppA (THNS), and the metabolite undergoes non‐enzymatic oxidation to flaviolin. In our design, the rppA gene was placed under the control of the constitutive P 14f promoter and the module was integrated in the genome at the att Tn 7 site. A workflow for assessing malonyl‐CoA levels in the engineered strain is shown; this illustration was created with BioRender.com . (B) Engineered Pseudomonas putida strains endowed with enhanced malonyl‐CoA turnover were identified through a colorimetric screening method, isolating clones that displayed increased red pigmentation in the supernatant. (C) HPLC analysis was used to quantify the flaviolin produced by the different engineered strains. Arbitrary units (A.U.) indicate the intensity of the peak corresponding to flaviolin, measured at 310 nm. GraphPad Prism 9 (GraphPad Software, Inc.) was used to perform all statistical analyses; the levels of significance are indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. The error bars represent standard deviations; n = 3. For a more precise assessment of malonyl‐CoA levels, HPLC analysis was used to directly quantify flaviolin produced by the different P . putida strains. The quantification of flaviolin is based on HPLC intensity as no flaviolin standard is available for direct quantification, limiting the possibility of reporting absolute yields. The analysis was performed after 24 h of growth, when the cultures had reached the same final OD 600 . As shown in Figure 2C , four strains showed significantly higher flaviolin levels than the others. The strains with the highest flaviolin production were SEM11 Δ hexR Δ gcd Δ fabF Δ pha , SEM11 Δ gltA Δ fabF Δ pha , SEM11 Δ gltA Δ pha P \n \n J23108 \n \n → fabF‐2 , and SEM11 Δ gltA Δ fabF Δ pha P \n \n J23108 \n \n → fabF‐2 . These strains had an increase in the flaviolin levels of 85%, 78%, 76%, and 63%, respectively, when compared to the reference strain. These strains harbor several gene manipulations, showing that multiple pathways need to be altered to increase malonyl‐CoA formation. Except for SEM11 Δ gltA Δ pha P \n \n J23108 \n \n → fabF‐2 , all strains with high flaviolin production contain the fabF deletion, suggesting that targeting fatty acid synthesis has a major effect on the phenotype—yet eliminating FabF alone is not sufficient. A comparable outcome was reported for a platform P. taiwanensis strain; in this case, enhancing malonyl‐CoA availability involved a reduced demand of the thioester for fatty acid synthesis (Schwanemann et al., 2023 ). Enhanced malonyl‐ CoA availability supports PHB accumulation through a non‐canonical biosynthesis pathway We employed a malonyl‐CoA shunt‐based pathway for PHB biosynthesis to evaluate the engineered P. putida strains for biopolymer accumulation, utilizing malonyl‐CoA as a substrate. The canonical PHB biosynthesis pathway begins with de novo acetoacetyl‐CoA formation, typically catalyzed by acetoacetyl‐CoA thiolase via a thioester‐dependent Claisen condensation between two acetyl‐CoA molecules (Choi & Lee, 1999 ; Nikel et al., 2006 ). The non‐canonical biosynthesis pathway used in our study differs in that its first step is catalyzed by NphT7, an acetoacetyl‐CoA synthase from Streptomyces sp., which irreversibly condenses acetyl‐CoA and malonyl‐CoA to produce acetoacetyl‐CoA and CoA (Okamura et al., 2010 ). The next two steps are mediated by PhaB and PhaC from the canonical pathway of Cupriavidus necator (Figure 3A ). The topology of this non‐canonical pathway for PHB biosynthesis offers improved control over precursor availability at the acetyl‐CoA and malonyl‐CoA metabolic nodes (Orsi et al., 2021 , 2022 ). This pathway was shown to mediate PHB accumulation from glucose in engineered P. putida (Kozaeva et al., 2021 ; Martínez‐García, Aparicio, et al., 2014 ; Martínez‐García, Nikel, et al., 2014 ; Nikel & de Lorenzo, 2013 ). A synthetic operon, comprising the genes encoding the three enzymes of the pathway, was expressed from plasmid pSEVA2311·PHAS, which carries the cyclohexanone‐inducible ChnR/P \n chnB \n expression system (Benedetti et al., 2016 ). FIGURE 3 Adopting engineered Pseudomonas putida strains for malonyl‐CoA–dependent PHA production. (A) A non‐canonical PHA biosynthesis pathway, based on NphT7 (an acetoacetyl‐CoA synthase from Streptomyces sp.), was used in this study. PhaB catalyzes the NADPH‐dependent reduction of acetoacetyl‐CoA, followed by polymerization by PhaC, a PHA synthase; both PhaB and PhaC are enzymes from C. necator . The genes encoding these three enzymes are encoded in plasmid pSEVA2311·PHAS under control of the cyclohexanone‐inducible ChnR/P \n chnB \n expression system. (B) Maximum cell density (estimated as the optical density at 600 nm, OD 600 ) reached by the strains transformed with plasmid pSEVA2311·PHAS, in the presence or absence of the inducer (cyclohexanone). (C) Nile Red staining for visualization of PHB granules. Nile Red is a lipophilic fluorescent dye that binds to PHB granules and can be readily detected through fluorescence microscopy, offering qualitative evidence of biopolymer accumulation. (D) Nile red staining for semi‐quantitative assessment of PHB levels in engineered P . putida . Nile red was added to the cultures after 24 h of growth, when the cultures had reached stationary phase, and the fluorescence was read after 30 min of incubation. Nile red fluorescence values for each strain were normalized to the OD 600 of the corresponding culture; normalized fluorescence values were compared with those of the parental strain, SEM11 Δ pha . The relative fluorescence of SEM11 Δ pha containing an empty pSEVA2311 vector is indicated by the dotted gray line. GraphPad Prism 9 (GraphPad Software, Inc.) was used to perform all statistical analyses; the levels of significance are indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. In all cases, the error bars represent standard deviations; n = 3. After all strains were transformed with plasmid pSEVA2311·PHAS, we characterized their growth patterns in the presence and absence of cyclohexanone. By analyzing the growth curves of these engineered P . putida strains containing the plasmid for PHB accumulation (Figure 3B ), we noticed a different physiological response to the addition of the inducer, probably due to the burden caused by the production of the pathway enzymes. In particular, virtually all strains with modifications in the TCA cycle showed a higher final cell density (OD 600 ) in the presence of cyclohexanone. Since PHA accumulation affects light scattering and results in higher absorbance values (Martinez & Déziel, 2020 ), we ascribed the higher final OD 600 readings to enhanced PHB accumulation. Culturing SEM11 Δ pha , the reference strain, under the same conditions with increasing cyclohexanone concentrations resulted in no substantial changes in the final OD 600 (data not shown). These results exclude a potential effect of the inducer on the absorbance. Most of the strains with gene deletions affecting glycolysis exhibited a lower maximum OD 600 , suggesting that PHB accumulation may act as a stress factor in these genetic backgrounds (Ankenbauer et al., 2020 ). PHB granules within the cells were stained with Nile red, a fluorescent lipophilic dye that binds hydrophobic inclusion bodies and enables their visualization by fluorescence microscopy (Martínez‐García, Aparicio, et al., 2014 ; Martínez‐García, Nikel, et al., 2014 ; Nikel et al., 2009 ; Nikel, Pettinari, Galvagno, & Méndez, 2008 ; Nikel, Pettinari, Ramírez, et al., 2008 ; Spiekermann et al., 1999 ). Using this protocol, we obtained semi‐quantitative evidence of PHB accumulation in all engineered strains (Figure 3C ). PHB granules were visualized as intensely fluorescent intracellular dots; in the negative controls, where the corresponding P . putida strain had been transformed with the empty pSEVA2311 vector (Martínez‐García et al., 2023 ), only the membranes were stained. Nile red can also be used to compare PHA accumulation levels by adding the stain to cultures upon they reach stationary phase (Nikel, Pettinari, Galvagno, & Méndez, 2008 ; Nikel, Pettinari, Ramírez, et al., 2008 ). Fluorescence and OD 600 are measured after a short incubation, providing a semi‐quantitative estimation of PHB content on biomass. Almost all engineered strains had significantly higher normalized fluorescence values than SEM11 Δ pha harboring the empty vector (used as a negative control), except for SEM11 Δ pha P \n \n J23108 \n \n → fabF‐2 , SEM11 Δ gltA Δ pha and the reference strain SEM11 Δ pha (Figure 3D ). In particular, cultures of SEM11 Δ gltA Δ pha P \n \n J23108 \n \n → fabF‐2 , SEM11 Δ glta Δ fabF Δ pha , and SEM11 Δ hexR Δ gcd Δ fabF Δ pha had an increase in the normalized fluorescence of 51%, 49%, and 141%, respectively, when compared to SEM11 Δ pha . We observed that the strains that exhibited high flaviolin production also performed well in PHB accumulation experiments; in fact, the two parameters showed a strong correlation across all experimental conditions and strains (Figure S2 in the Supporting Information). In general, we concluded that deleting gltA in combination with an altered fatty acid synthesis seems to favor malonyl‐CoA turnover and PHB accumulation. In contrast, deleting genes involved in glycolysis, coupled with fatty acid synthesis, appears to introduce a stress factor if combined with PHB accumulation—although the PHB content in these strains remained largely unaffected. Although the increase in Nile red fluorescence observed in the best‐performing strains validated our engineering strategy to boost malonyl‐CoA, we wanted to explore a more direct quantification method for PHB content. To this end, strain SEM11 Δ glta Δ fabF Δ pha P \n \n J23108 \n \n → fabF‐2 was grown in shaken‐flask cultures using glucose as the main substrate, and the biomass and PHB content was determined after a 24 h incubation by methanolysis and detection of the 3‐methyl esters of 3‐hydroxybutyrate by GC‐FID (Ruiz et al., 2006 ). Under these conditions, this engineered strain accumulated PHB to 25.1 ± 3.9% (on a cell dry weight basis). While further experiments are necessary for a detailed comparison of all engineered strains, our findings set the basis for developing improved bioprocess for PHB production via malonyl‐CoA. Additionally, these engineered strains could serve as versatile platforms for other malonyl‐CoA‐dependent pathways and products."
} | 8,821 |
39204482 | PMC11359911 | pmc | 3,676 | {
"abstract": "Biodegradable biobased polymers derived from biomass (such as plant, animal, marine, or forestry material) show promise in replacing conventional petrochemical polymers. Research and development have been conducted for decades on potential biodegradable biobased polymers such as polylactic acid (PLA), polyhydroxyalkanoates (PHAs), and succinate polymers. These materials have been evaluated for practicality, cost, and production capabilities as limiting factors in commercialization; however, challenges, such as the environmental limitations on the biodegradation rates for biodegradable biobased polymer, need to be addressed. This review provides a history and overview of the current development in the synthesis process and properties of biodegradable biobased polymers, along with a techno-commercial analysis and discussion on the environmental impacts of biodegradable biobased polymers. Specifically, the techno-commercial analysis focuses on the commercial potential, financial assessment, and life-cycle assessment of these materials, as well as government initiatives to facilitate the transition towards biodegradable biobased polymers. Lastly, the environmental assessment focuses on the current challenges with biodegradation and methods of improving the recycling process and reusability of biodegradable biobased polymers.",
"introduction": "1. Introduction Biodegradable biobased polymers (more commonly known as biodegradable bioplastics) are substances entirely or partially derived from biomass or biological products, including plant, animal, or forestry materials. The advent of biobased polymers that are both biodegradable and renewable provides a green alternative to petroleum-based plastics; however, there are limitations and concerns with the current life cycle of biodegradable biobased polymers that need to be addressed before widespread adoption. Figure 1 below illustrates the general life cycle of biodegradable biobased polymers. Some major industries that use biodegradable biobased polymers are pulp and paper, agricultural, beverage, and medical supply manufacturers [ 2 ]. Specifically, within the medical industry, poly-( ε -caprolactone) (PCL) is primarily used for implantable composites, bone fixation, and medicine release systems [ 3 ]. Furthermore, certain types of polyhydroxyalkanoates (PHAs) have also been shown to have applications within the medical industry. For instance, poly(4-hydroxybutyrate) (P(4HB)) was first used in 2007 to create “TephaFLEX”, a type of absorbable suture [ 4 ]. Research into P(4HB) showed that the polymer can cross the blood–brain barrier, leading to further development of using PHAs to develop various medical devices [ 4 ]. Cellophane (a cellulose-based biopolymer) is a common material used in food packaging [ 5 ]. Other polymers of interest in packaging are Polylactic acid (PLA) and Poly (3-hydroxybutyrate) [ 5 ]. The first considerable interest in biobased polymers was during the mid-1800s when celluloid was accidentally discovered to be a thin, flexible film that could be molded into shapes and was relatively resilient [ 6 ]. In 1888, microbiologist Martinus Willem Beijerinck observed light-refractive inclusions in microorganisms, which were determined to be polymer granules [ 7 ]. Later, in the 1900s, French chemist Maurice Lemoigne discovered the ability of Bacillus megaterium to produce an intracellular polyester called poly(3-hydroxybutyrate) (PHB), from which current research has found over 100 different types of polyhydroxyalkanoates (PHAs) [ 8 ]. However, the onset of World War II stifled research into biobased polymers as society recognized the cost-effectiveness of utilizing petroleum to produce plastics. After the war era ended, research into bioplastics gradually returned. For instance, PLA emerged as another possibility for mass produced bioplastics as it could be derived naturally; however, mass production was cost- and energy-intensive [ 9 ]. As different alternatives were discovered and developed further, each material seemed to have particular use cases, and experimentation beginning in the late 1980s proved fruitful. The mid-20th century also saw the beginnings of research around degradation methods of petro-based and bio-based polymers. An example is the isolation of a PLA-producing enzyme, a thiolase, which can be used to degrade PHAs [ 8 ]. The consequence of climate change and plastic pollution [ 2 , 10 , 11 ] has continued to stimulate growth in research into the 21st century, as shown in Figure 2 . This review paper aims to provide a comprehensive overview of the current research landscape concerning biodegradable biobased polymers. It will delve into the advancements in synthesis methods and material properties, offering a detailed analysis of the technological and commercial aspects of biodegradable biobased polymers currently employed in various industries, as well as those that are undergoing active research and development. Additionally, the paper will evaluate the potential end-of-life options for these materials, contributing to a holistic understanding of their environmental impact and sustainability.\n\n5.1. Introduction to Techno-Commercial Analysis Biodegradable biobased polymers are impactful in the environmental view of plastic production and plastic waste management, but also the techno-commercial view, as well as the global economic view. In this section, the impact, current and future, of biodegradable biobased polymers is discussed in the context of a case study analysis of one country, global market trends, government initiatives, and life-cycle analysis."
} | 1,405 |
37578580 | PMC10482783 | pmc | 3,679 | {
"abstract": "In this study, response surface methodology (RSM) was applied with a Box–Behnken design to optimize the biosorption (removal and bioconcentration) of rare earth elements (REEs) (Y, La, Ce Eu, Gd, Tb) by living Ulva sp. from diluted industrial wastewaters (also containing Pt and the classic contaminants Hg, Pb, Zn, Cu, Co, and Cd). Element concentration (A: 10–190 μg/L), wastewater salinity (B: 15–35), and Ulva sp. dosage (C: 1.0–5.0 g/L) were the operating parameters chosen for optimization. Analysis of the Box–Behnken central point confirmed the reproducibility of the methodology and p- values below 0.0001 validated the developed mathematical models. The largest inter-element differences were observed at 24 h, with most REEs, Cu, Pb and Hg showing removals ≥ 50 %. The factor with the greatest impact (positive) on element removal was the initial seaweed dosage (ANOVA, p < 0.05). The optimal conditions for REEs removal were an initial REEs concentration of 10 μg/L, at a wastewater salinity of 15, and an Ulva sp. dosage of 5.0 g/L, attaining removals up to 88 % in 24 h. Extending the time to 96 h allowed seaweed dosage to be reduced to 4.2 g/L while achieving removals ≥ 90 %. The high concentrations in REE-enriched biomass (∑REEs of 3222 μg/g), which are up to 3000 times higher than those originally found in water and exceed those in common ores, support their use as an alternative source of these critical raw materials. Supplementary Information The online version contains supplementary material available at 10.1007/s11356-023-29088-2.",
"conclusion": "Conclusion The present study has demonstrated the potential of widely available seaweed as the basis of a simple, efficient, and low-cost technology to remove a myriad of elements from diluted industrial effluents and contaminated waters, especially those with high salinity that hamper the efficiency of other methods. The use of real saline water and a contamination scenario that mimics real effluent (e.g., from lamp production and dismantling) are also innovative aspects of this study. Among the parameters studied, the initial seaweed dosage was the most impactful factor for REEs removal, with higher dosages resulting in greater removals (up to 88 % in 24 h). The time extension from 24 h to 96 or 144 h proved significant. Optimized conditions for REEs removal, obtained from the model optimization by RSM, were determined as a concentration of 10 μg/L at salinity 15 and seaweed dosage of 5.0 g/L. The high REE concentration in the enriched biomass (∑REEs of 3222 μg/g) obtained under optimal conditions, which is approximately 3000 times higher than originally in the water and higher than in ordinary ores, encourages its use as an alternative source of these critical raw materials.",
"introduction": "Introduction Since China’s position in the global rare earth elements (REEs) market changed in 2009, anxiety befell among manufacturers of high-tech equipment. The unique properties of REEs led to ever-increasing demand and to severe constraints on its supply—critical raw materials (European Commission 2023 ). The recovery of REEs from waste electric and electronic equipment (e-waste) such as lamp/TV phosphors and batteries, or from wastewater can be an additional source of REEs and help reduce the serious environmental impacts associated with REEs extraction and improper e-waste management (Deng et al. 2022 ). With the ban on incandescent lamps, fluorescent lamps (FLs) and light-emitting diodes (LED) greatly expanded (Machacek et al. 2015 ). Both use REEs phosphors to produce visible light (red, blue, and green) (Jiang et al. 2022 ), although FLs contain higher concentrations of REEs compared to LED (0.005 vs. 1.2 g/unit) (Machacek et al. 2015 ; Rebello et al. 2020 ) and are classified as hazardous waste due their Hg content. Millions of FLs are sold and discarded each year, making them an interesting holdover from a REEs recovery perspective (Zhang et al. 2017 ). Spent NiMH batteries are another potential source of REEs and other critical elements such as Co, Ni, and Mn, which poses an environmental hazard if not properly managed (Rasoulnia et al. 2021 ). Nevertheless, despite the high value of the recoverable materials in e-waste, the recycling rates have not kept pace with e-waste generation rates (1 – 12 %) (Balaram 2019 ). Several methods have been described to recover REEs from e-waste and wastewater, including ion exchange (Felipe et al. 2021 ), solvent extraction (Xie et al. 2014 ), chemical precipitation (Silva et al. 2019 ), ionic imprinted polymers (Hu et al. 2018 ), among others. Nevertheless, all share disadvantages such as high operational costs, high reagent/energy consumption, production of secondary metabolites (Pereao et al. 2018 ), and low efficiency at relatively low REEs concentrations. Biosorption has been highlighted as a greener, and cheaper tool to remove elements from aqueous solutions (Giese 2020 ). A highly efficient, economical, and environmentally friendly solution for water remediation, with low nutrient requirements, lower volume of sludge to be disposed of, and the potential for element recovery are important advantages pointed to biosorption. Available research reports the use of a variety of biosorbents such as crab shell, orange peel (Liu et al. 2021 ), animal (fish scales) and plant origin (neem sawdust) (Das et al. 2014 ), micro- and macroalgae, the latter also known as seaweed when they are marine (Kucuker et al. 2017 ; Viana et al. 2021 ). However, most studies are still conducted under ideal conditions, i.e., in distilled water, focusing on one or a few elements in unrealistically high concentrations, and also not considering the effect of ionic strength, although high salt concentrations are common in industrial effluents and decrease biosorption performance in general (Torres 2020 ). In general, the influence of the operating conditions is assessed individually, but to understand the complexity of the biosorption process, it is essential to know the interactive effects of the operating variables. Response surface methodology (RSM) allows to obtain of a set of relevant information with a small set of trials (Bowden et al. 2019 ). The purpose of this methodology is to set up a series of experiments (design) for adequate predictions of a response and to optimize conditions to reach the highest efficiency (Whitford et al. 2018 ). In the present study, the effects of three factors, initial element concentration, wastewater salinity and biosorbent dosage, on the removal and preconcentration of six rare earth elements (Y, Eu, La, Ce, Tb, and Gd) by living Ulva sp. were studied using the Box–Behnken design, which provided mathematical models showing the influence of each variable and their interactions. To bridge the gap between laboratory biosorption studies and real industrial effluents, a complex mixture of contaminants simulating industrial effluents (Y, Eu, La, Ce, Tb, Gd, Hg, Pb, Zn, Cu, Co, Cd, Pt) was examined.",
"discussion": "Results and discussion Quality control and influence of time on removal Although control experiments are not preconized by BBD, four control assays were used in the present work to ensure the quality of the results (Fig. 1 ). A maximum acceptable variation over time of 20 % was stipulated (upper and lower limits in Fig. 1 ). All C t /C 0 values were within the defined intervals with no relevant changes over time, indicating that losses (vessel adsorption, co-precipitation, or volatilization in the case of Hg), or contamination were negligible. Exceptions are two occasional cases for Hg (Fig. 1 A and B) and a punctual case for REEs (Fig. 1 D) which may attributed to uncertainties related to equipment quantification or contamination during sample collection. Fig. 1 Control charts for the elements normalized concentration in control experiments over the exposure time (UCL, upper control limit; LCL, lower control limit) The BBD implies that central point replicates, where all the factors are in their central values, are done to assess whether the data is within the normal dispersion and repeatability is ensured. Except for Cd (for all contact times), global between-replicate coefficients of variation (CV) were ≤ 10 % for all elements at all contact times, validating the experimental design (Fig. 2 ). Fig. 2 Removal (%) of the 13 elements from the mixture mimicking the elemental composition of industrial effluent, for the condition corresponding to the BBD central point (mean value ± standard deviation; n = 3). The black line indicates 50 % of removal The contact time is known to influence the efficiency of the sorption process, so its assessment is paramount to introduce the process into a real company scenario (Azubuike et al. 2016 ). Shorter contact times are beneficial for industries as they reduce the overall costs associated with the process. To assess difference among contact times regarding both removal and bioconcentration ( p -value < 0.05), a non-parametric statistical test was applied (Friedman’s test with Dunn’s multiple comparison test) (Ali and Bhaskar 2016 ). The results are summarized in Table S 2 of the Supplementary Material. Time extensions from 24 to 48 h, 48 to 96 h, and 96 to 144 h were not significant for any of the responses evaluated. In contrast, increasing the time from 24 h to either 96 h or 144 h proved to be significant for both responses. The extension from 48 to 144 h was also significant. Based on Table S 2 , no significant benefits are seen in running the process for times longer than 96 h, which can be considered the most appropriate exposure time. Box–Behnken design data modelling The removal efficiency (%) and the corresponding bioconcentration (μg/g) of Y, La, Ce, Eu, Gd, and Tb in Ulva sp. for the 15 conditions studied and times analyzed are summarized in Table 1 . In terms of removal (%), the highest values for all REEs (> 94 %) were achieved in trials 3 and 9 (A: 10 μg/L; B: 25; C: 5.0 g/L and A: 10; B: 15; C: 3.0), after144 h. Cerium was the REEs with the fastest kinetic, similarly to Hg, with removals of 80 and 75 % for trials 3 and 9, respectively, after 24 h. In contrast, the lowest removals observed after 144 h exposure were 39, 71, 80, 69, 67, and 61 % for Y, La, Ce, Eu, Gd and Tb, respectively, in trial 6 (A: 100 μg/L; B: 35; C: 1.0 g/L). Costa et al. (Costa et al. 2020a ) studied the removal of REEs (Y, La, Ce, Pr, Nd, Eu, Gd, Tb, Dy) from a solution in the presence and absence of some PTEs (Cr, Ni, Cu, Cd, Hg, Pb), in equimolar concentration (1 μmol/L), and observed that the presence of PTEs altered the kinetic profile, leading to an improvement in REEs removal. Table 1 Removal (%) and bioconcentration (μg/g) of rare earth elements by Ulva sp. from the mixture mimicking the elemental composition of an industrial effluent over the exposure time Removal (%) Bioconcentration (μg/g) Trial Time (h) Y La Ce Eu Gd Tb Y La Ce Eu Gd Tb 1 24 26 37 48 38 38 35 12 23 26 23 22 15 48 32 45 69 47 43 44 15 28 37 28 25 19 96 45 60 87 65 60 63 21 37 48 40 35 27 144 55 68 88 73 72 72 26 43 48 45 43 31 2 24 20 29 35 36 38 34 201 322 393 363 383 302 48 32 43 49 48 49 47 322 473 544 484 494 413 96 50 60 62 60 61 61 504 665 685 604 614 534 144 70 76 76 76 77 77 705 846 846 766 776 675 3 24 49 69 85 75 73 70 6 9 12 10 9 8 48 58 78 95 81 76 77 7 10 13 11 10 8 96 - 90 96 95 94 94 - 12 13 12 12 10 144 - 95 97 96 96 96 - 13 13 13 12 10 4 24 53 58 64 70 72 68 129 154 172 172 176 143 48 53 68 78 73 74 70 129 181 207 178 181 147 96 78 80 83 80 82 82 192 214 221 196 201 172 144 88 87 88 86 87 87 216 232 234 212 214 183 5 24 29 33 40 41 38 38 165 201 267 252 216 190 48 36 43 53 53 50 49 206 267 355 324 283 247 96 62 68 72 71 70 70 350 417 484 437 396 355 144 76 78 80 80 79 80 432 478 535 494 448 401 6 24 12 20 27 26 22 18 61 111 152 142 111 81 48 14 33 45 33 29 25 71 182 253 182 152 111 96 22 55 71 54 49 43 111 304 405 294 253 192 144 39 71 80 69 67 61 203 395 456 375 344 273 7 24 61 68 73 75 76 73 70 86 101 95 96 77 48 59 75 84 82 82 78 68 95 116 103 103 83 96 86 85 88 88 89 88 99 108 122 111 112 93 144 92 90 91 91 92 92 105 114 125 115 116 97 8 24 29 53 66 63 60 54 33 65 85 76 69 54 48 24 56 80 57 55 48 26 69 102 69 63 48 96 61 77 87 79 77 76 67 95 112 95 89 76 144 91 88 91 90 90 90 101 109 116 109 104 90 9 24 49 65 75 70 68 65 9 14 16 15 13 11 48 62 77 93 81 79 78 12 17 20 18 16 14 96 91 90 96 94 94 94 18 19 21 20 19 16 144 - 95 97 97 96 97 - 21 21 21 19 17 10 24 23 45 65 48 44 38 4 9 14 9 9 5 48 27 50 87 52 49 46 4 10 18 10 10 6 96 75 79 94 83 71 84 12 16 20 16 14 11 144 - 90 94 93 93 94 - 19 20 18 19 13 11 24 50 50 52 54 55 55 206 225 244 244 225 206 48 66 69 72 73 72 73 274 309 336 328 298 274 96 86 86 86 87 87 88 356 386 405 392 358 328 144 92 91 92 92 91 92 381 411 430 413 377 345 12 24 23 41 49 48 46 43 82 145 174 170 156 128 48 34 58 66 60 58 55 121 206 234 209 195 163 96 76 78 78 79 79 80 266 277 277 277 266 238 144 80 78 78 79 79 80 280 277 277 277 266 238 13 24 27 49 58 52 51 47 47 95 117 98 91 73 48 42 60 73 62 61 58 73 117 146 117 109 91 96 79 77 80 79 80 79 138 149 160 149 142 123 144 86 84 85 84 84 84 149 162 171 158 149 131 14 24 27 49 61 54 51 47 48 96 122 103 92 74 48 44 64 76 65 63 63 78 126 151 126 114 100 96 81 81 83 81 80 81 143 159 165 155 145 128 144 85 84 86 84 84 84 151 164 171 161 152 133 15 24 27 46 59 55 52 47 50 96 127 112 100 77 48 35 59 79 62 60 56 66 123 170 127 116 92 96 76 80 84 81 81 80 143 166 182 166 156 133 144 85 85 87 86 85 85 161 177 188 174 165 140 The highest bioconcentration (μg/g) in Ulva sp. was achieved in trial 2 (A: 190 μg/L; B: 25; C: 1.0 g/L), reaching values > 650 μg/g after 144 h of exposure. Such values reveal the high capability of Ulva sp. to preconcentrate REEs from complex effluents that enter the ecosystem. The bioconcentration factor (BCF, DW), determined as the ratio between the REEs concentration in Ulva sp. at 144 h and initially in solution, varied within the broad interval of 900 (Tb) and 5350 (Ce) (Supplementary Material, Table S 3 ), with trial 5 achieving higher BCF values (lowest value of 4010 for Tb). The capability of Ulva sp. to concentrate REEs was previously assessed in mono (Ferreira et al. 2020 ) and multi-element scenarios (Costa et al. 2020a ). The higher value attained in this study is 3.1 and 1.4 times higher than the reported by Ferreira et al. and Costa et al., respectively. Thorough multiple regression analysis, mathematical correlations between factors and removal or bioconcentration were established. The goodness of fit of the models developed can be seen in the Supplementary Material (Table S 4 to Table S 9 ). ANOVA confirmed that the models are significant and suitable to predict REEs removal (Tables S 4 – S 9 ). For removal (%), the linear and quadratic terms B, C, A 2 , and C 2 were significant for most of the quadratic models. Based on the F-value, the main effects of the independent variables followed the order of seaweed dosage, salinity, and initial concentration. Non-significant terms, such as A, AB, AC, BC, and B 2 , had limited impact on the response and were excluded from the study to improve the model (reduced model). Since Ulva sp. is a euryhaline seaweed, found in brackish water conditions and estuarine substrates (Pereira 2015 ), it is tolerant to different salinities, being perfectly adapted to different habitat conditions, resulting in extreme resilience. Thus, the significance that the RSM attaches to salinity may be related to the different speciation of the elements in saline water rather than to the adaptability of Ulva sp.. This seaweed can also grow in sewage-contaminated areas, and is an opportunistic species that can form massive blooms that reflect fluctuations in environmental quality faster than slow-growing seaweed like Fucus (Christiansen 2018 ; Coelho et al. 2005 ). This may justify why the initial concentration of the element is not an impactful factor according to the RSM. For bioconcentration (μg/g), the terms A, B, C, AC, and C 2 were significant for most of the quadratic models. Based on the F -value, the main effects of the independent variables followed the order of initial concentration, seaweed dosage, and salinity. Non-significant terms, such as AB, BC, A 2 , and B 2 , had a limited impact on response and were excluded from the study to improve the model. The polynomial equations developed considering only the significant factors for the responses considered (reduced models, uncoded variable values) are shown in Table 2 (the goodness of fit can be found in Table S 10 ). The best-fitted models for the removal and bioconcentration at 24 h were found for Tb ( R 2 adj = 0.998) and Y ( R 2 adj = 0.968), respectively, while at 96 h the best adjustments were observed for Gd ( R 2 adj = 0.941) and Tb ( R 2 adj = 0.968), for removal and bioconcentration, respectively. Table 2 Reduced models of the removal (%) and bioconcentration (μg/g) responses as a function of the significant variables ( p- value < 0.05) for Ulva sp. Removal (%) Bioconcentration (μg/g) 24 h 96 h 24 h 96 h Y y = 61.6 − 2.07B + 4.91C + 8.24 × 10 −4 A 2 + 2.66 × 10 −2 B 2 + 8.14 × 10 −1 C 2 + 1.28 × 10 −2 AC − 1.78 × 10 −1 BC - y = 116 + 1.55A – 2.58B – 57.3C – 3.31 × 10 −2 AB – 9.17 × 10 −2 + 8.37× 10 −1 BC + 1.85 × 10 −3 A 2 + 5.49C 2 - La y = 51.7 − 2.24 × 10 −1 A − 1.05B + 13.1C + 4.43 × 10 −4 A2 − 8.39 × 10 −1 C2 + 3.28 × 10 −3 AB y = 65.1 − 5.06 × 10 −1 B + 5.67C − 1.20 × 10 −3 A 2 y = 103 + 1.74A – 2.45B – 36.4C – 2.14 × 10 −1 AC + 6.06C 2 y = 192 + 3.80A – 2.97B – 99.1C – 5.92 × 10 −1 AC + 16.0C 2 Ce y = 29.7 − 1.96 × 10 −1 A + 1.02B + 16.1C + 4.72 × 10 −4 A 2 − 2.89 × 10 −2 B 2 − 1.24C 2 y = 75.1 − 1.88 × 10 −1 A + 11.3C + 4.87 × 10 −4 A 2 − 1.24C 2 y = 132 + 2.13A – 2.53B – 52.6C – 2.87 × 10 −1 AC + 8.68C 2 y = 199 + 3.85A – 141C – 5.96 × 10 −1 AC +21.4C 2 Eu y = 60.9 − 2.16 × 10 −1 A − 1.14B + 8.89C + 3.57 × 10 −4 A2 + 4.51 × 10 −3 AB y = 66.3 − 1.34 × 10 −1 A − 5.67 × 10 −1 B + 19.6C + 4.56 × 10 −4 A 2 − 2.32C 2 y = 132 + 1.98A – 2.61B – 53.5C – 2.47 × 10 −1 AC +8.58C 2 y = 210 + 3.50A – 3.47B – 90C – 5.28× 10 −1 AC + 13.9C 2 Gd y = 59.3 − 1.26B + 9.10C + 4.99 × 10 −4 A 2 + 4.48 × 10 −3 AB y = 68.0 − 1.22B + 21.4C − 2.22C 2 + 4.25 × 10 −3 AB − 1.73 × 10 −2 AC y = -14.5 + 2.04A +3.07C – 2.69 × 10 −1 AC y = 182 +3.51A – 3.29B – 86.4C – 5.42 × 10 −1 AC + 14.2C 2 Tb y = 54.5 − 2.37 × 10 −1 A − 7.64 × 10 −1 B + 8.76C + 6.03 × 10 −4 A 2 − 1.25 × 10 −2 B 2 + 4.22 × 10 −3 AB y = 61.7 − 7.18 × 10 −1 B + 22.7C + 7.34 × 10 −4 A 2 − 2.71C 2 y = 101 + 1.66A – 2.7B – 38.1C – 2.11 × 10 −1 AC + 6.69C 2 y = 289 – 3.11A – 8.91B - 109C – 4.79× 10 −1 AC +1.82BC + 10.7C 2 Optimization studies focusing on the removal of La and Ce by seaweed have been reported in the literature (Keshtkar et al. 2019 ), however, with seaweed biomass after drying and grinding (not taking advantage of the bioaccumulation process), after functionalization treatment (increasing the cost and complexity of the process), and using central composite design (CCD) instead of BBD. Recent studies used RSM with a BBD to optimize Nd and Dy removal by two living seaweed ( Ulva sp. and Gracilaria sp.), although both studies were performed with mono-element solutions (Fabre et al. 2021 ; Ferreira et al. 2021 ). Optimizing REEs removal from complex solutions by living seaweed is a new approach of the present study. Three-dimensional (3D) surface responses Figures 3 and 4 show the three-dimensional (3D) response surfaces for Tb removal and bioconcentration by Ulva sp. at 24 and 96 h (the 3D response surfaces of the remaining REEs can be found in Supplementary material—Figure S 1 to Figure S 12 ). The interaction between salinity (B) and initial concentration (A) (Fig. 3 A), at constant seaweed dosage (C = 3.0 g/L), shows that a higher removal is obtained when lower initial concentrations and salinities are applied for both elements, reaching removals close to 65 % for Tb. A positive effect on removal (up to 60 % of removal) is observed when the seaweed dosage increases and the initial concentration of Tb decreases, at constant salinity (B = 25) (Fig. 3 B). Figure 3 C shows that low salinity combined with an increase in seaweed dosage, while the initial concentration is kept constant (A = 100 μg/L), leads to higher removals. Extending the time to 96 h (Fig. 3 D–F) increases the removal, but the way the variables interact is almost the same as discussed for 24 h. Fig. 3 3D response surface for the removal (%) of Tb by Ulva sp. at 24 h and 96 h Fig. 4 3D response surface for the bioconcentration (q t (μg/g)) of Tb by Ulva sp. at 24 h and 96 h In terms of bioconcentration (μg/g), the 3D surface plots for exposure times of 24 and 96 h are shown in Fig. 4 . Seaweed dosage only affected bioconcentration at high concentrations of Tb (Fig. 4 A), which may be because the number of active sites available on the seaweed surface is much higher than the number of REEs ions in solution (for low initial concentrations), resulting in low levels of REEs per g of seaweed, even if all the element in solution was bioaccumulated by Ulva sp. The results also showed that higher bioconcentration is achieved when a lower dosage of seaweed is applied at a higher concentration. Lower seaweed dosage in low salinity (Fig. 4 B) shows higher bioconcentration of Tb, which may be attributed to less competition among ions in solution since there are fewer salt cations in solution at low salinity. At constant seaweed dosage (C = 3.0 g/L; Fig. 4 C) bioconcentration increase is only dependent on the initial concentration, with a larger uptake occurring at higher concentrations at any salinity. Analyzing the uptake for 96 h (Fig. 4 D – F), it is also possible to observe an enhancement in bioconcentration, as was observed for the removal, but the variables still interact in the same way as previously discussed for 24 h. As with removal, the results for the bioconcentration of the remaining REEs at 24 h and 96 h were consistent with those obtained for Tb. Optimal conditions for the removal of rare earth elements from the complex mixture Optimizing the removal of some potentially toxic elements such as Pb, Cu, and Hg with Ulva sp. and Gracilaria sp. have been reported in the literature (Çetintaş et al. 2020 ; Isam et al. 2019 ), but all were performed after drying and grinding treatment of the seaweed biomass. For REEs, only studies for mono-solutions of Nd (Fabre et al. 2021 ) and Dy (Ferreira et al. 2021 ) using living seaweed are known. In the present study, RSM was applied to all REEs data to obtain the optimal conditions that maximize the removal of all REEs. The optimization (Table 3 ) leads to removals from 67 % of Y to 88 % of Ce by Ulva sp. after 24 h under the operating conditions of 10 μg/L of REEs mixture, salinity of 15, and seaweed dosage of 5.0 g/L. After 96 h of exposure, the optimized response has values ≥ 90 % for all REEs under the operating conditions of 10 μg/L of REE mixture, salinity of 15, and seaweed dosage of 4.2 g/L. The highest removal is observed with the highest seaweed dosage, indicating that the available sorption sites are still unsaturated. Both conditions (high seaweed dosage and low salinity) appear to be consistent with previously information, as it is well described that increasing biomass dosage provides greater availability of binding sites, and consequently increases removal efficiency, while lower ionic strength reduces competition for binding sites (Cao et al. 2021 ). Table 3 Optimal values for REEs removal (%) and bioconcentration (μg/g) with Ulva sp. and corresponding calculated bioconcentration (μg/g) or removal (%) REE Time (h) Initial concentration (μg/L) Salinity Seaweed dosage (g/L) Removal (%) Calculated bioconcentration (μg/g) Optimized variables for REE removal (%) Y 24 10 15 5.0 67 7.4 La 79 8.7 Ce 88 9.8 Eu 87 9.6 Gd 83 9.2 Tb 84 9.4 Y 96 10 15 4.2 - - La 90 10 Ce 99 11 Eu 98 11 Gd 100 11 Tb 100 11 REE Time (h) Initial concentration (μg/L) Salinity Seaweed dosage (g/L) Bioconcentration (μg/g) Calculated removal (%) Optimized variables for REE bioconcentration (μg/g) Y 24 190 15 1.0 253 24 La 326 31 Ce 401 38 Eu 377 36 Gd 325 31 Tb 304 29 Y 96 190 16 1.0 - - La 668 63 Ce 693 66 Eu 667 63 Gd 618 59 Tb 576 55 The natural REE concentrations in seawater are usually very low (10 −6 – 10 −9 g/L), increasing several order of magnitude in sites impacted by anthropogenic sources such as the discharge of industrial wastewater and e-waste leachates (Arienzo et al. 2022 ). For example, values of 130–152 μg REEs/L were reported in surface waters in areas affected by mining activities (Liang et al. 2014 ), which are in line with the range of concentrations of the present study, highlighting the potential of the approach to remediate contaminated environments. The use of living seaweed over non-living biomass can be valuable as the constant daily growth of seaweed creates new available binding sites, allowing for the removal of REEs from the complex solution even days after the onset of exposure (Costa et al. 2011 ; Henriques et al. 2015 ). Bioconcentration (μg/g) optimization resulted in an initial element concentration of 190 μg/L, a salinity of 15, and a seaweed dosage 1.0 g/L, with values ranging from 253 μg/g for Y to 401 μg/g for Ce (∑REEs of 1986 μg/g), after 24 h exposure. For 96 h of exposure, a maximum ∑REEs of 3222 μg/g (excluding Y) was obtained, using the optimized conditions (initial concentration of 190 μg/L, salinity 16, and seaweed dosage of 1.0 g/L). The total concentration of REE in Ulva sp. is circa of 3000-fold higher than that initially in water and also exceeds that found in common apatite ores (∑REEs 1098 – 1688 μg/g) (Xiqiang et al. 2020 ), supporting the use of REEs-enriched seaweed biomass as an alternative to mineral ores. It is possible to further increase these concentrations by processing the biomass, namely its pyrolysis which reduces its weight by about 87 %. The recovery of REEs from the biomass is also simpler than their extraction from ores which are much more difficult to solubilize. Diluted acids such as nitric and chloric acids or chelating agents have been used to regenerate various synthetic materials (Afonso et al. 2019 ; Lee et al. 2018 ; Ramasamy et al. 2017 ) and are also efficient in recovering REEs from seaweed. In some cases, the osmotic shock caused by placing the seaweed in ultrapure or distilled water, leading to disruption of the cell membrane, is enough for the elements to migrate into the solution (Henriques et al. 2019 ). It should be noted that in addition to REEs, Ulva sp. also incorporated other elements, the separation of which is crucial for the valorization of technology. In a previous work, for the same time of exposure, we verified that REEs removed from water by Ulva sp. were mostly bound on the seaweed surface and were easily desorbed with a 15 min wash with EDTA (Viana et al. 2023 ). Other elements, such as the PTE Hg remained in the biomass after EDTA washing."
} | 6,802 |
37002442 | PMC10066054 | pmc | 3,681 | {
"abstract": "Highlights \n A conductive, stretchable, adaptable, and self-healing, GaInSn/Ni--based composite hydrogel by incorporating a magnetic liquid metal into the hydrogel framework through crosslinking polyvinyl alcohol with sodium tetraborate. The multifunctional composite hydrogels showed outstanding performance for magnetic repair movement sensing, and EMI shielding. \n Supplementary Information The online version contains supplementary material available at 10.1007/s40820-023-01043-3.",
"conclusion": "Conclusion In summary, we developed a multifunctional PVA/liquid metal composite hydrogel with rapid self-healing ability, excellent stretchability, shape adaptability, magnetic prototyping, sensing capability, and good EMI shielding properties. The fluidity of liquid metal and reversible hydrogen bonds between PVA chains and borate ions enabled the PVA/liquid metal hydrogel to complete self-healing rapidly without external stimuli. The proposed PVA/liquid metal hydrogel could effectively repair broken wire joints when they were placed in a precision-sealed space under an applied magnetic field. The synergy between the moderate conductivity and inherent moisture-rich environment endowed the composite hydrogel with high-efficiency EMI shielding. The total SE values significantly increased to 65.8 dB (blocking 99.99997% of EM waves) for PVA/EGaInInSn–8Ni hydrogels, and the increase percentage reached as high as 83.0% in comparison with those of the pure PVA hydrogel. Introducing liquid metals to PVA could effectively increase the absorption loss and decrease the reflection loss, and absorption might dominate the EMI shielding mechanism. Significantly, the excellent EMI shielding performance was maintained after storage for one year, showing long-term stability. Moreover, liquid metal hydrogels could conform to objects of arbitrary geometry and recover rapidly from damage, demonstrating significant application potential in flexible electronics and artificial skin. The hydrogel served as a strain sensor to detect various body movements and a signature sensor for sensitive and rapid responses to external stimuli. Based on the above functions, the present study offers a novel strategy to develop intelligent hydrogels for multifunctional applications as well as a versatile method for fabricating liquid metal composites with extended performance.",
"introduction": "Introduction Liquid metals of Ga-based alloys have attracted intensive interest because of their deformability, nontoxicity, self-healing capability, high electric/thermal conductivity, and unique surface chemistry [ 1 – 3 ]. In particular, the fluidity and deformability of liquid metals in aqueous environments are promising for applications in various fields, such as flexible electronics [ 4 – 6 ], soft robotics [ 7 , 8 ], energy harvesting/storage [ 9 , 10 ], microfluidics [ 11 ], sensors [ 12 ], actuators [ 13 , 14 ], and wearable sensing [ 15 ]. Currently, miniaturized, integrated, and high-power electronic devices are being rapidly developed for wireless communication, and significantly large electromagnetic interference (EMI) is produced as an inevitable by-product [ 16 , 17 ]. EMI has a substantial effect on the nearby electronic apparatus and may result in the degradation and malfunctioning of electronics, particularly those working at high frequencies [ 18 – 20 ]. Furthermore, EMI pollution has severe adverse effects on the surrounding environment and human health. Therefore, the development of EMI functional materials has become a significant alternative to effectively alleviate this dilemma [ 21 – 23 ]. Compared with traditional rigid metals, MXenes and carbon nanomaterials [ 24 – 26 ], liquid metals have emerged as promising EMI shielding materials and attracted increasing attention because of their excellent processability, good fluidity, and high conductivity [ 27 – 29 ]. Zhu et al. fabricated a flexible multifunctional EM shielding film derived from an Ecoflex elastomer filled with magnetic liquid metal droplets [ 30 ]. The film exhibited strain-improved electrical conductivity and EMI shielding properties when subjected to uniaxial tensile stress. A three-dimensional (3D) liquid metal network was inserted in a stretchable polydimethylsiloxane composite foam, and the resulting material demonstrated significant EMI shielding enhancement under compression and stretching [ 31 , 32 ]. The 3D liquid metal configuration could regulate the electrical conductivity during stretching and compression, enhancing the reflection of EM waves and shielding the EMI. Liquid metal-based monoliths with a 3D continuous conductive network were successfully prepared using a constrained thermal expansion method [ 33 ]. The as-prepared monoliths possessed tunable architectures owing to the microcosmic fluidity of the liquid metal and demonstrated excellent EMI shielding performance. Huang et al. developed a self-standing thermostable composite film for EMI shielding by introducing a small quantity of aramid nanofibers into liquid metals [ 34 ]. However, the EMI shielding properties of liquid metal-based composites are mainly achieved by reflecting electromagnetic waves, which will cause secondary pollution. Moreover, their limited self-healing ability hampers their applications in intelligent, flexible devices. In contrast to metal liquid-filled elastomers, hydrogels are a type of engineering material composed of a crosslinked network of hydrophilic blocks surrounded by water [ 35 ], demonstrating the potential for absorption-based EMI blocking [ 36 ]. Moreover, compared to their dry counterparts, soft polymer hydrogels possess shape adaptability and self-healing capabilities, enabling stable and conformal interfaces with blocked targets for applications in artificial skin or wearable electronics [ 37 – 39 ]. In the past decades, fast self-healing, shape/size-tunability, and EMI shielding properties of hydrogels have been extensively studied. Accordingly, several hydrogels have been developed, such as sandwich-structured hydrogels consisting of a layer of silver nanowire and two layers of polyvinyl alcohol (PVA) hydrogels reinforced by aramid nanofibers [ 40 ], Ti 3 C 2 -MXene-functionalized poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) hydrogels [ 41 ], ionic liquid doped PEDOT:PSS hydrogels [ 42 ], polyacrylic acid/cellulose nanofibers/MXene/calcium ion composite hydrogel [ 43 ], MXene organohydrogel containing glycerol and water binary solvents [ 44 ], and multiwall carbon nanotubes/polyacrylamide/cellulose nanofiber hydrogels [ 45 ]. However, these hydrogels repair themselves by the aid of external manual operation. The fabrication of self-healing liquid metal-based soft hydrogels with no need for manual touch remains an open challenge. Herein, we present a simple ultrasonic method to fabricate PVA/GaInSn–Ni composite hydrogels with rapid self-healing capability and excellent stretchability and shape adaptability. The multifunctional composite hydrogels demonstrated high performance for magnetic repair and prototyping, body movement sensing, and EMI shielding. The abundant reversible hydrogen bonds between PVA and borate ions and the fluidity of liquid metals render the PVA/liquid metal hydrogel self-healing features in the absence of any external stimulus. In the presence of a magnetic field, broken wires can be repaired remotely by placing them in a sealed space with one end warped by the magnetic liquid metal hydrogel. The composite hydrogel could be used as a strain sensor to detect body motions and as a signature sensor that sensitively and rapidly responds to various external stimuli. The water-rich hydrogel with moderate conductivity provides absorption-dominated EMI shielding performance to the composite. Importantly, the composite hydrogel exhibited long-term stability for EMI shielding even after storage for 1 year.",
"discussion": "Results and Discussion Characterization of PVA–Liquid Metals Hydrogels Figure 1 a illustrates the fabrication process of the self-healing PVA/EGaInSn–Ni composite hydrogel. A eutectic GaInSn (EGaInSn) suspension was first prepared by the ultrasonic treatment of Ga, In, and Sn metals with a specific mass ratio. The composite PVA/EGaInSn–Ni was obtained by mixing the EGaInSn suspension with Ni particles, PVA, and sodium tetraborate. Herein, borate molecules acted as crosslinkers by forming abundant hydrogen bonds with PVA [ 46 , 47 ]. Previous studies showed that a thin oxide skin layer could be easily formed on the surface of gallium-based liquid metals in the air [ 48 ], leading to secondary crosslinking with PVA chains. The as-fabricated PVA/EGaInSn–Ni hydrogel exhibits excellent self-healing capability (Fig. 1 b), which is derived from the abundant H-bond sites. The self-healing properties of the prepared hydrogels are mainly attributed to two factors: (1) As a self-healing supramolecular adhesive, PVA polymer not only has certain viscosity, but also has complex network structure and good biocompatibility, which can promote fluidity, ductility and self-healing. Polymerization of PVA with borax by diol results in many network structures, in which the hydroxyl group in the PVA chain forms a weak hydrogen bond with the borax molecule, forming a “tenon-like” structure. After the material is broken, the exposed new surface is rich in PVA hydroxyl group and borax molecule. When the two damaged surfaces contact, the weak hydrogen bonds formed between the hydroxyl group in the PVA chain and borax molecule gradually recombine and connect, which leads to the material healing. (2) Hydrogen bonds between the hydroxyl group of the PVA chain and the thin metal oxide layer on the liquid metal surface provide abundant sites for molecular crosslinking during self-healing, which can be repaired quickly after interconnection. Additionally, the fluidity of liquid metal further facilitates the crosslinking of the PVA hydrogel and promotes self-healing. These PVA–liquid metal composite hydrogels are expected to have excellent EMI shielding properties (Fig. 1 c) because of the conductive networks of liquid metals, abundant dipoles in water, and numerous interfaces between PVA, EGaInSn, and Ni. Adding magnetic Ni particles allows the composite hydrogels to move under a magnetic force (Fig. 1 d). Fig. 1 Fabrication of the self-healing liquid metal hydrogel. a Scheme illustrating the preparation of the PVA/EGaInInSn–Ni composite hydrogel containing primary crosslinked networks. b Self-healing mechanism of the PVA/EGaInInSn–Ni hydrogel. c Schematic demonstrating the high-performance EMI shielding of the liquid metal-based hydrogel resulting from conductive loss, interfacial polarization, and dipole polarization. d Schematic showing external magnetic force can move composite hydrogel The EGaInSn particles were synthesized through an isothermal sonication process at 150 ℃ for 60 min. The long sonication process yields highly dispersed liquid metal droplets with uniform elemental distribution (Fig. S1). A more detailed study on the surface composition of EGaInSn was performed using XPS (Fig. S2). In the XPS spectra of Ga 3 d (Fig. S2a), the peak at 18.5 eV is associated with metallic gallium (Ga 0 ) [ 49 ], while two peaks at 19.6 and 20.6 eV correspond to the chemical states of Ga 1+ in Ga 2 O and Ga 3+ in Ga 2 O 3 [ 50 ], respectively. The O 2 s peak at 23.8 eV is also observed in the Ga 3 d spectrum, confirming the oxidation of Ga [ 51 ]. The functional shell on the liquid metal droplets had numerous oxygen-containing groups, which were expected to significantly enhance the interaction between the droplets and polymer matrix. In the In 3 d core-level spectrum, two distinct peaks at 443.1 (In 3 d 5/2 ) and 450.7 eV (In 3 d 3/2 ) are attributed to In 0 (Fig. S2b). Similarly, in the Sn 3 d XPS spectrum (Fig. S2c), the peaks centered at 484.8 and 493.2 eV are related to the metallic Sn 0 . These results suggest that the surface of the EGaInSn particles a mainly composed of Ga oxides and a small amount of the mixture of In and Sn oxides. The O 1 s XPS spectrum (Fig. S2d) is deconvoluted into two peaks at 531.7 and 533.3 eV, suggesting the presence of intact stoichiometric oxides and oxygen vacancies, respectively [ 52 ]. Figure 2 a displays the powder XRD patterns of pure PVA, PVA/EGaInSn, and PVA/EGaInSn containing 8 wt% Ni particle (PVA/EGaInSn–8Ni) hydrogels. No distinct peaks are observed in the PVA hydrogel, confirming its crosslinking state [ 53 ]. After PVA is mixed with EGaInSn liquid metal, a low-intensity broad peak at ~ 35° appears, indicating the presence of amorphous components (e.g., amorphous oxides and/or the liquid itself) [ 54 ]. With the introduction of Ni particles, three clear diffraction peaks at 44.6°, 52.0°, and 76.5° are observed, which are indexed to the (111), (200), and (220) planes of the face-centered cubic structure of Ni (JCPDS no. 04–0850). Similar to PVA/EGaInSn–8Ni, all XRD patterns of the hydrogels with various Ni contents show three diffraction peaks of Ni (Fig. S3). FTIR spectroscopy provides more detailed information for elucidating the crosslinked structure (Figs. 2 b and S4). Pure PVA, PVA/EGaInSn, and PVA/EGaInSn–Ni hydrogels exhibit a strong and broad peak at 3400 cm −1 due to the absorption characteristic of water and stretching vibration of the –OH group, which is a typical indication of hydrogen bonding [ 55 ]. The peak associated with C–O stretching vibration is observed at 1630 cm −1 . The carbonyl (C=O, the acceptors of hydrogen bonds) and hydroxyl groups (–OH, the donors of hydrogen bonds) can produce a high density of hydrogen bonds [ 56 ], which is consistent with the broad and strong peak at 3455 cm −1 . Fig. 2 Material characterizations of PVA/EGaInInSn–Ni composite hydrogels. a XRD patterns and b FTIR spectra of pure PVA, PVA/EgaInSn, and PVA/EGaInInSn–8Ni hydrogels. High-resolution XPS spectra of c Ga 2 p , d Ni 2 p , and e C 1 s in the PVA/EGaInInSn–8Ni hydrogel. f SEM images of PVA/EGaInInSn–8Ni and g corresponding element mapping of Ga, In, Sn, Ni, C, and O The chemical bonding states of the PVA/EGaInSn–8Ni composite hydrogels were confirmed using XPS (Fig. 2 c–e and S5). The Ga 2 p core-level spectrum (Fig. 2 c) is deconvoluted to three peaks related to Ga + , Ga 3+ , and metallic Ga. The high-resolution XPS spectrum of Ni 2 p (Fig. 2 d) shows the coexistence of Ni 0 and Ni 2+ species [ 57 ], indicating that oxidation occurred on the surface of the Ni particles. In the In 3 d region (Fig. S5a), the metallic indium exhibits two peaks centered at 443.7 (In 3 d 5/2 ) and 451.4 eV (In 3 d 3/2 ), with a separation of spin–orbit components \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\Delta_{{{\\text{metal}}}}$$\\end{document} Δ metal ≈ 7.7 eV [ 58 ]. In the Sn 3 d XPS spectrum (Fig. S5b), the peaks centered at 484.9 and 493.5 eV correspond to the metallic tin [ 59 ]. As shown in Fig. 2 e, the C 1 s core-level spectrum has three-peak components, i.e., C–C/C=C (284.9 eV) and C–O (286.5 eV). The strong C signal indicates the presence of abundant hydrophilic groups on the PVA chains, which offer abundant crosslinking sites in the hydrogel [ 36 ]. The O 1 s peaks are deconvoluted to metal oxide, hydroxide, and lattice oxide component peaks (Fig. S5c), demonstrating the formation of mixed liquid metal oxides and the Ni oxidation states on the shell [ 60 ]. SEM images illustrate the uniform dispersion of liquid metal particles containing Ni in the PVA polymer (Fig. 2 f, g). Corresponding elemental mapping using EDS reveals that Ga, In, Sn, and Ni elements are confined in the particles, while the C and O elements are observed in all the SEM images, which is in good accordance with the hydrogel composition. In these PVA–liquid metal composite hydrogels, the water contents are high, with slight differences among the PVA hydrogels with introduction of liquid metals. As shown in Fig. S6, the water contents of PVA hydrogels decreased with increase in liquid metal amounts, and the water content of the original PVA-EGaInSn–8Ni hydrogel sample can reach as high as about 74 wt%. Self-healing, Magnetic-Drag and Smart Responsible Properties of Hydrogels The rheological properties of the liquid metal hydrogels were studied using oscillatory rheology (Fig. S6). For pure PVA and PVA/EgaInSn hydrogels, their loss modulus ( G ″) values are smaller than the storage modulus ( G ) in the frequency range of 1–10 Hz, indicating their solid-like states. Significantly, the G ″ value of PVA/EGaInSn–8Ni exceeds G′ , exhibiting a liquid-like state (Fig. S7a). Therefore, Ni in the hydrogels facilitated the generation of a more liquid-like condition without damaging the complex network structure. Furthermore, for the liquid metal hydrogels with different Ni contents (Fig. S7b), the Gʹ of all composite hydrogels is less than G ″, confirming their liquid-like behaviors. Therefore, for the subsequent experiments, we chose PVA/EGaInSn–8Ni as the smart sensor hydrogel. The black adhesive PVA/EGaInSn–8Ni hydrogel is in a chuddy‐like state with multifunctional performance. Its stretchability easily reaches more than 800% (Fig. 3 a). The liquid metal hydrogel also demonstrated excellent plasticity and writability. It could adapt to the desired shape, and the formed shape could be broken and pasted onto various objects in different shapes, such as starlike, heart, face-like shapes, indicating the shape-controllable feature of composite hydrogels. During the deformation process, the hydrogel remained intact without cracks. Owing to its good liquid mobility, the hydrogel inks were used to write any words, including “FD” and “CUMTB.” The as-fabricated PVA/EgaInSn–8Ni also has significant self-healing capability (Fig. 3 b). When two separate parts touched each other in the natural environment, they joined seamlessly and rapidly. The high self-healing performances were derived from the intensive hydrogen bonding and liquid metal fusion that facilitated PVA crosslinking. Moreover, an emitting diode was used to illustrate the self-healing performance of the liquid metal composite hydrogel. When the power supply, wires, diode, and hydrogel were connected to form a loop, the diode was lit up. When the hydrogel was broken, the diode was turned off. Once the broken hydrogel was joined, the diode glowed again. Figure 3 c illustrates a typical hydrogel repair of wire joints guided by an external magnetic field because the PVA/EGaInSn–8Ni hydrogel possesses the specific saturation magnetization ( M s ) value of 12.2 emu g −1 (Fig. S8). The disconnected wire, in which one end is wrapped in the hydrogel, is placed in a plastic container. A permanent magnet is placed outside the plastic container. When the magnet is moved closer to the hydrogel, the entire hydrogel shifts toward the magnet. The strong adhesion between the liquid metal and PVA prevents the liquid metal droplets from leaving the hydrogels during magnetic field-driven movement. The magnet can drag the hydrogel to repair the wire conduction and finally lit the diodes up (Movie S1). In addition, the liquid metal hydrogel can be guided using the magnetic field to fill the custom-designed heart-shaped pattern (Fig. 3 d and Movie S2). Fig. 3 Flexible, self-healing, magnetic-responsive properties of the liquid metal composite hydrogel. Optical photographs demonstrating a stretchability, plasticity, and writability; b automatically self-healing capability; c magnetic field-driven movement for the junction circuit; d magnetic field-induced prototyping. Relative current changes of the liquid metal hydrogel used as a pressure sensor for the real-time detection of e finger bending, f wrist bending, and g swallowing. h PVA/EGaInInSn–8Ni hydrogel sandwiched between two layers of polyimide films to real-time monitor the handwriting of English letters “A/B/C” The high stretchability, plasticity, writability, self-healing, and magnetic field-induced repair and shaping properties enable the application of PVA/liquid metal hydrogels as strain sensors for detecting various body movements. This magnetic liquid metal hydrogel was significantly soft and flexible and could easily adhere to different joints of the human body. The PVA/liquid metal hydrogel was attached to the index finger to monitor finger flexion, as shown in Fig. 3 e. The uniform sensing signal and distinct signal change ( \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\Delta I/I0$$\\end{document} Δ I / I 0 , \\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym} \n\t\t\t\t\\usepackage{amsfonts} \n\t\t\t\t\\usepackage{amssymb} \n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$\\Delta I$$\\end{document} Δ I : change current under strain stimuli; I 0 : original current) indicate the good reliability of the magnetic PVA/liquid metal hydrogel monitoring. When the PVA/liquid metal hydrogel is attached to the wrist, the current increases when bending the wrist, and the current–time curve exhibits significant stability and reproducibility (Fig. 3 f). Therefore, the PVA/liquid metal hydrogel sensor demonstrates considerable potential for real-time human health monitoring [ 61 ]. Based on the compression effect generated by vocal cord vibration [ 62 ], when the hydrogel sensor is attached to the throat, swallowing can also be clearly identified (Fig. 3 g). The accurate detection of swallowing during water drinking suggests that the sensor can be used to create a pharyngeal motion recognition device. Additionally, the hydrogel sensor could accurately capture the frequency of the laryngeal vibrations. Moreover, when the hydrogel sensor was coated with two layers of polyimide film on its surface, it acted as a smart-writing keyboard to sense the signature on its surface. When the letters “A,” “B,” and “C” are written on the hydrogel sensor (Fig. 3 h), various forms of signals are generated. These results showed that magnetic PVA/liquid metal hydrogels have significant application potential in flexible wearable electronics. EMI Shielding Properties of PVA–Liquid Metals Hydrogels Electrical conductivity is crucial in determining the EMI shielding properties [ 63 ]. Therefore, the electrical conductivities of the PVA/liquid metal composite hydrogels were measured using a four-probe instrument. The electrical conductivity of the PVA/liquid metal composite is related to the migration of free ions and free water in the PVA network and the transport of electrons in the crosslinks containing EGaInSn–Ni particles (Fig. S9). As the Ni:EGaInSn ratio increased from 0.5:1 to 8:1, the conductivity of the PVA-based hydrogel gradually improved from 0.015 to 0.041 S m −1 . More Ni content in the hydrogel prompted electron migration and thereby increased electron conduction. However, the ion channels decreased with increasing Ni concentration, resulting in the reduction of ion conduction. Therefore, the final conductivity depended on the competition between ion and electron conduction. EMI between electronic devices often results in equipment failure. Self-healed conductive hydrogels with good EMI properties are ideal for soft robotic applications that integrate many electronic components. The EMI shielding capabilities of PVA–liquid metal composite hydrogels (1.0 mm in thickness) were calculated by testing the scattering parameters ( S 11 , S 12 , S 21 , and S 22 ) in the band (8.2–12.4 GHz) using a vector network analyzer. In everyday applications, an EMI shielding effectiveness (SE) of 20 dB can block approximately 99% of incident EM wave energy [ 64 ]. As shown in Fig. 4 a–c, the EMI shielding performances of PVA/EGaInSn and PVA/EGaInSn–8Ni hydrogels are significantly improved compared to that of the pure PVA hydrogel. An average total SE (SE T ) of 36 dB is obtained in PVA, which shields 99.975% of EM waves (Fig. 4 c). The SE T values significantly increase to 46.1 dB (blocking 99.9975% EM waves) and 65.8 dB (blocking 99.9999747% EM waves) for PVA/EGaInSn and PVA/EGaInSn–8Ni hydrogels, respectively; the increase percentage is 28.3% and 83.0%, respectively. These results indicate that introducing liquid metals into the PVA hydrogel can prompt EMI shielding. In comparison with EMI shielding properties and sensor ability of state-of-the-art hydrogels (Table S1), our PVA–liquid metal composite hydrogels demonstrate the competitive performance for EMI shielding, magnetic repair and prototyping. The contributions from absorption loss (SE A ) and reflection loss (SE R ) were analyzed to assess the EMI shielding mechanism of the hydrogels (Figs. 4 b, S10 and S11). The average SE A values of PVA, PVA/EGaInSn, and PVA/EGaInSn–8Ni hydrogels are 31.8, 42.2 (increase percentage of 32.9%), and 62.5 dB (increase percentage of 96.8%), respectively, while the SE R values decrease from 4.2 to 3.9 (decrease percentage of 7.1%) and 3.3 dB (decrease percentage of 22.0%), respectively. This demonstrates that adding liquid metals can effectively increase the absorption and decrease the reflection, and it seems that the EMI shielding mechanism mainly results from absorption. However, the calculated SE R was based on the total power of EM waves, whereas SE A was based on the power of incident waves [ 65 ], as reflection occurs before absorption. To further evaluate the actual shielding mechanism of the hydrogels, the power coefficients of absorption ( A ) and reflection ( R ) for the PVA, PVA/EGaInSn, and PVA/EGaInSn–8Ni composite hydrogels are shown in Fig. S12. The R values are higher than A values over the entire frequency range for the PVA and PVA/EGaInSn hydrogels, suggesting that the EMI shielding is dominated by reflection. However, in the case of PVA/EGaInSn–8Ni, the A value is higher than the R value in one-third of the X-band, indicating that introducing magnetic Ni is favorable for absorption. Although reflection dominates the EMI shielding mechanism, absorption plays an important role in the shielding contribution. Fig. 4 EMI shielding performances of PVA–liquid metal composite hydrogels. a EMI SE T plots, and b average SE R , SE A , and SE T values of PVA, PVA/EGaInSn and PVA/EGaInInSn–8Ni hydrogels at X-band. c EMI SE T increment in PVA/EGaInSn and PVA/EGaInInSn–8Ni hydrogels compared with that of the pure PVA hydrogel. d EMI SE T plots, and e average SE R , SE A , and SE T values of PVA/EGaInInSn–Ni hydrogels with various Ni contents. f Schematic of the reinforcement effect of Ni in absorption coefficients of PVA/EGaInInSn–Ni hydrogels. g EMI SE (SE R , SE A , and SE T ) values, and h SE R , SE A , and SE T values in the healed PVA/EGaInInSn–8Ni hydrogel. i SE R , SE A , and SE T values of PVA/EGaInInSn–8Ni hydrogels and j after storage for 1 year The EMI shielding properties of the PVA–liquid metal hydrogels with various Ni contents were investigated, and the results are shown in Fig. 4 d. As expected, the total EMI shielding performance of the PVA–liquid metal hydrogels increases with the Ni content. To further understand the EMI shielding mechanism, the ratio of SE R and SE A to the overall SE T were calculated, and the results are shown in Figs. 4 e and S13. Distinctly, a high Ni content in PVA–liquid metal hydrogels leads to an increase in SE A and a decrease in SE R , and the SE A values are significantly higher than the SE R values. However, as previously discussed, the shielding mechanism cannot be well understood based on the SE R and SE A values. In addition to the SE R and SE A values, A and R values were calculated to provide insights into the electromagnetic response [ 66 , 67 ]. As shown in Fig. S14, the power coefficient A increases with the Ni content, whereas the R values exhibit an opposite trend, indicating the reinforcement effect of Ni in the absorption. In general, reflection loss is depended on the unbalanced impedance at the interfaces between the air and shields, and the absorption loss is related to the EM energy converted from the generated current and polarization relaxation [ 18 , 68 ]. Figure 4 f shows a schematic of the enhanced absorption coefficients resulting from the increase in the Ni content. For the PVA–liquid metal composite hydrogel, conductive networks due to free ions and liquid metals dominate the ohmic loss [ 25 ], and the abundant dipoles in PVA and free water cause strong EM loss [ 69 ]. After introducing Ni particles, numerous interfaces between water, PVA, EGaInSn, and Ni particles are created; as a result, interfacial polarization significantly dissipates EM energy [ 70 ]. Importantly, adding Ni causes magnetic loss via the main natural resonance, significantly attenuating the EM energy [ 71 , 72 ]. In other words, the synergetic effects of conductive loss, dielectric loss, and magnetic loss are responsible for the enhanced EM absorption. The EMI shielding performance of PVA/EGaInSn–8Ni before and after self-healing was verified (Fig. 4 g, h). All SE T , SE R , and SE A values demonstrate a slight decrease owing to the minor damage to the conductive transport networks during the cutting process. Considering the power coefficients of A and R (Fig. S15), the R is higher than A in the frequency range of 8.2–10.7 GHz; however, the A is higher than R in the frequency range of 10.7–12.4 GHz. This illustrates that the reflection dominates in most measured frequency ranges. Most conventional EMI shielding materials show deteriorated shielding performance when placed in an air atmosphere for a long time [ 22 , 41 ]. Noticeably, compared with the freshly prepared hydrogel sample (about 74 wt%), the water content still maintains about 65.47 wt% after storage of 1 year (Fig. S16). Furthermore, the as-developed PVA–liquid metal composite hydrogel maintains a high EMI shielding performance (Figs. 4 i–j, S17, and S18). Significantly, when stored in air atmosphere for one year, SE T values increased from 65.8 to 75.2 dB, and SE A values increased from 62.5 to 72.9 dB. The increase percentage of SE T and SE A reached 14.3% and 16.6%, respectively. However, SE R values decreased from 3.3 to 2.3 dB, and the decrease percent was 28.5%. By analyzing power coefficients of A and R (Fig. S18) , PVA/EGaInInSn–8Ni presents an absorption-dominated EMI shielding mechanism after storage for one year. In an air atmosphere, hydrogels could absorb environmental water molecules to generate more free ions and dipoles, which was beneficial for EM energy dissipation and enhancing SE T and SE A values."
} | 7,805 |
35179537 | PMC8978807 | pmc | 3,682 | {
"abstract": "Surface icing is detrimental to applications ranging from transportation to biological systems. Soft elastomeric coatings can engender remarkably low ice adhesion strength, but mechanisms at the microscale and resulting ice extraction outcomes need to be understood. Here we investigate dynamic ice-elastomer interfacial events and show that the ice adhesion strength can actually vary by orders of magnitude due to the shear velocity. We study the detailed deformation fields of the elastomer using confocal traction force microscopy and elucidate the underlying mechanism. The elastomer initially undergoes elastic deformation having a shear velocity dependent threshold, followed by partial relaxation at the onset of slip, where velocity dependent “stick-slip” micropulsations are observed. The results of the work provide important information for the design of soft surfaces with respect to removal of ice, and utility to fields exemplified by adhesion, contact mechanics, and biofouling.",
"conclusion": "Conclusion In summary, we show that the velocity at which the force is imposed on ice during ice removal, is an important factor significantly affecting the ice adhesion strength on elastomers. By employing cTFM, we quantified the transient non-uniform elastic deformation fields of the elastomer and related it to the applied external shear stress. The maximum deformation increases significantly with shear velocity, and consequently τ ice increases with V . Further, we observe pulsations in the slip regime at high velocity, which agrees well with the reported observations in the literature. The capability to visualize the deformation of the elastomer at high resolution is expected to be useful in studies exemplified by bio-fouling control, viscoelastic adhesives, contact mechanics, and soft robotics.",
"introduction": "Introduction Soft viscoelastic solids (or simply elastomers) exhibiting both elastic and viscous behavior are ubiquitous in nature and practical applications. Biological tissues mainly comprise viscoelastic components with each tissue having a specific function. 1 From a practical point of view, elastomers are traditionally used in foams, adhesives, food additives, and so on. 2 With recent technological advances, the applications extend to the exciting domains of soft robotics, tissue engineering, flexible electronics, and so on. 3–5 One of the potentially important applications of elastomers is their passive icephobic performance during harsh environmental icing conditions. 6–9 Ice adhered to surfaces incurs malfunctioning or a drastic decrease in efficiency in many industrial settings. 10 It is previously shown that elastomers offer exceptionally low ice adhesion shear strength, τ ice = F* / A ≤ 10 kPa, either through interfacial fracture- or slippage mechanisms. 6,8,9,11–13 Here, F* indicates the peak ice removal force recorded by a force gauge in a standard ice adhesion shear test at a constant shear velocity, V , and A is the apparent substrate-ice interfacial area. 14 In contrast, rigid and liquid impregnated surfaces (which have an intervening lubricant layer in between ice and the base solid) have a lower limit of τ ice ≈ 50 and 15 kPa, respectively. 15–20 On this class of materials, τ ice can increase significantly due to the highly humid ambient conditions or lubricant depletion over time. 21–29 However, elastomers do not have such bottlenecks and are shown to be durable over several (10–20) icing/de-icing cycles. 6,8 Other strategies with weakly interacting surface layers using organogel materials to mitigate icing on surfaces are discussed in ref. 30 . Interfacial fracture on elastomers can be either stress- or toughness-limited de-bonding of ice, where the elastomer cavitation leads to easy ice removal. 6,8 Slippage occurs when the elastomer chains adhered to the ice are sufficiently mobile, which can be realized by either reducing the cross-linking density of the elastomer or infiltrating the elastomer network with a lubricant. 6,7,31,32 The latter is unlike liquid impregnated surfaces, because the lubricant is introduced into the bulk of the elastomer, and any lubricant layer on the interface interacting with ice is avoided. During slippage, ice remains adhered to the elastomer but continues to slide away from its initial location as long as shear is applied. Whether ice fractures or slips during an adhesion test is governed by a multitude of parameters: elastomer surface temperature, T , shear velocity, V , type of stress applied in the adhesion test (shear, mixed, and normal), elastomer thickness, h , and Young's modulus, E (also a function of temperature). This illustrates the complexity involved in comprehending the ice removal mechanisms on elastomers across a much broader spectrum of influencing conditions. The ice adhesion shear strength, τ ice , is the accepted defining metric for the performance of a surface, with respect to its resistance to ice detachment. In previous studies, τ ice on elastomers is typically reported for low shear velocities of V ≤ 0.1 mm s −1 . 6,8,13 Here, we show that τ ice on soft silicone elastomeric coatings ( E = 15.4 kPa at T = −20 °C) can increase by an order of magnitude, depending on the shear velocity, V , which can potentially define the domain of their applicability in practical settings. To understand the reason for this behavior, we first probe into the elastomer deformation using confocal traction force microscopy (cTFM) providing the necessary high-resolution visualization of spatio-temporal in-plane deformation of the elastomer, during shear mode experiments. Therefore, the cTFM technique provides unprecedented detail of the elastomer deformation at the microscopic scale. We then invoke stochastic models involving cyclic adsorption and desorption of elastomer chains near the interface to explain the observed trend of τ ice with V . Further, we investigate the effect of mode of adhesion test (shear, mixed, and normal) and elastomer temperature on the ice adhesion strength, and show that the trend of τ ice with V in mixed mode tests is similar to that of shear mode tests while τ ice decreases with T .",
"discussion": "Results and discussion Dynamic ice adhesion To first understand the effect of shear velocity, V , on the peak ice removal force, F *, we perform macroscopic shear adhesion tests using a standard setup (see Fig. S1, ESI † ) on a silicone coating with h = 35 μm, and T = −20 °C at different velocities (see Fig. 1a ). Here, the elastomer thickness, h , is sufficiently large to avoid the effects of the underlying rigid substrate. Clearly, F * increases significantly with V , and ice continues to slip as long as shear is applied. Nevertheless, beyond a critical velocity, we can have interfacial fracture even in the shear mode test, due to increase in the normal force at the interface. 11 This is because the external force from the finite size force pin (see earlier section) cannot be practically applied in exactly the same plane as the elastomer–ice interface, which results in non-zero torque. At high velocities, the magnitude of this torque becomes significant (implying higher normal stress) leading to interfacial fracture. In our case, we observe fracture at V = 10 mm s −1 for h = 35 μm, and T = −20 °C (see Movie S1, ESI † ). As an estimate, F * increases from 0.8 to 6.6 N (see Fig. 1a ) over the investigated velocity spectrum (0.01 to 10 mm s −1 ). An order of magnitude increase in F * (consequently τ ice ) with V emphasizes its importance while designing icephobic surfaces. In contrast, rigid surfaces do not show any such dependence (see Fig. S2, ESI † ). 14 A remark here is that, in a recent study 35 it has been shown that the prolonged application (∼100 s) of constant force (< F *) on ice leads to melting of interfacial ice near substrate which reduces τ ice significantly. However, in the present case, the condition imposed is a constant shear rate in a given experiment. Therefore, the possibility of local melting of ice and increase in quasi liquid layer thickness is low as the time taken to reach peak ice removal force, F * (thereby τ ice ), in a typical F vs . t plot (see Fig. 1a ) even at the lowest velocity, V = 0.01 mm s −1 , is typically of the order 10 s; an order of magnitude lower than the time required to melt the interfacial ice. Fig. 1 Icephobicity of elastomers depends on shear velocity. (a) Representative force, F , vs . time, t , plots for an ice adhesion test in shear mode at different shear velocity, V , with elastomer thickness, h = 35 μm, and surface temperature, T = −20 °C. The peak ice removal force, F *, increases with V . Time, t = 0 is taken just before the instance when the force has non-zero values. (b) Sketch (not to scale) showing the experimental setup of cTFM. (c) Sketch (not to scale) showing the elastomer printed with fluorescent quantum dots (QDs) grid in red with each dot having a diameter of ≈0.2 μm. The cuvette inner diameter, 2 R = 6.5 mm, is much larger than the grid dimension (290 × 290 μm). As ice is sheared, the deformation of the QDs is captured from the bottom view. (d and e) Selected bottom view snapshots of the QD grid (and with it the elastomer surface) during the shear mode test at V = 0.01 and 0.1 mm s −1 , respectively, with h = 35 μm, and T = −20 °C. Time, t = 0 in (d and e) represent the reference configuration at the instance just before the QDs start to translate. The orange line indicates the reference axis as a guide to the eye and the inset in yellow represents a zoomed in portion of the QD grid. The red arrows indicate the direction of motion of the QD grid. At t = 198.8 s, the region enveloped by the green line indicates the air pocket. Scale bar: (d), 100 μm (same in (e)). To understand the physics behind this behavior, we employ cTFM (see Fig. 1b, c and Section S2, ESI † ), which involves printing and tracking of fluorescent quantum dots (QDs) on the elastomer surface using an electrohydrodynamic drip printing technique reported in previous work. 34,36–39 In addition, we also perform bright field experiments to confirm that no undulations persist on the elastomer ( i.e. no interfacial fracture) and only interfacial slip occurs during the adhesion test under similar experimental conditions (see Fig. S3 and S4, ESI † ). Fig. 1d and e show exemplary bottom view snapshots of the QD grid printed elastomer during ice adhesion test with h = 35 μm, T = −20 °C at V = 0.01 and 0.1 mm s −1 , respectively (see Movies S2 and S3, ESI † ). The mechanical properties of the selected elastomer are known—established by previous work 34,36,39 —such that we can use cTFM to reveal the microscopic mechanisms responsible for the rate dependent τ ice . Time, t = 0 indicates the reference configuration i.e. just before the shear is applied. The grid dimension (290 × 290 μm) is significantly smaller than the diameter of the ice block, 2 R = 6.5 mm (see Fig. 1c ). The reference configuration grid on the elastomer is located approximately near the center of the initial ice location. In Fig. 1d , as the force pin pushes the ice, the entire QD grid starts to globally translate reaching a maximum displaced configuration around t = 27.2 s. To quantify the displacement of the QDs we track them in time (see Section S3, ESI † ), and define the instantaneous mean displacement, (see Fig. 2a ), as the mean of all the tracked QDs displacements in the X-direction (see Fig. 1b–d ) at a given time, t . Here, U x,i represents the X-direction displacement of an i th QD at a time, t , and n p represents the number of tracked QDs in each time step. During the adhesion experiments, some of the QDs can undergo out-of-plane motion and go out of focus as seen from t = 19.8 to 270.0 s. This can be attributed to the presence of non-zero normal stresses at the interface as discussed earlier. 11 For this reason, n p decreases with time. We therefore track the QDs until the time where Ū x approximately reaches a plateau. Fig. 2b shows the variation of Ū x at V = 0.01 mm s −1 with the green region in the plot indicating the deformation regime where Ū x increases approximately linearly with t reaching a peak, at t = 27.2 s. In addition to the global translation of the QDs, we have simultaneous local distortion i.e. the QD grid layout in each time step is different from the reference configuration. This qualitatively confirms that the transient local displacement and the traction fields are non-uniform. Beyond t = 27.2 s, we observe partial relaxation of the grid i.e. the QDs are displaced back towards the reference axis (orange line in Fig. 1d ) slightly. The red region in Fig. 2b indicates the slip regime where Ū x falls below . This explains the dip in τ macro after attaining the peak i.e. τ macro = τ ice . Here, τ macro = F / A , is the instantaneous shear stress estimated from the force gauge reading at time, t . The partial relaxation of the QD grid marks the beginning of ice slipping on the elastomer, for if interfacial fracture were to occur, the grid should restore completely back to the reference configuration, which does not take place. As mentioned earlier, tracking all the QDs for a long time is difficult. Therefore, in Fig. 2c , we track the displacement of five individual QDs (i = 1–5 in Fig. 2a ), U x,i , for an extended period of time. The partial relaxation of the elastomer is identified by the decrease in U x,i after attaining the peak. This relaxation continues to take place over a time, t r ≈ 5 s, which is of the same order as the macroscopic relaxation time of the elastomer, t relax = 11.7 s. 34 Eventually U x,i reaches a plateau, and is confirmed by the fact that the QD grid does not translate significantly from its local position (see Fig. 1d and Video S2, ESI † ). The edge of the ice starts to crossover the grid at around t = 270.0 s, and once the crossover is complete, the elastomer relaxes fully i.e. the QD grid restores back approximately to the reference configuration with only a slight offset from the reference axis, as shown at t = 322.8 s. The slight offset could be due to the permanent set, which leads to a negligible residual strain in the elastomer after the adhesion test. Fig. 2 Dynamic icephobicity behavior in shear mode test. (a) Sketch showing the transient QD grid deformation to estimate instantaneous individual displacement, U x,i , and mean displacement, Ū x in the X-direction. The box indicates the optical field of view during the experiment. (b) Variation of instantaneous shear stress, τ , and Ū x with time, t , at shear velocity, V = 0.01 mm s −1 , elastomer thickness, h = 35 μm, and surface temperature, T = −20 °C. The instantaneous macroscopic shear stress, τ macro , is defined as the ratio of the force, F , at t and the apparent elastomer–ice interfacial area, A . The calculated shear stress from 2 nd order Ogden model , τ Ogden , is obtained by tracking a total number of QDs, n p ≈ 4000, in each time step. The region in green indicates the elastic deformation regime where Ū x increases linearly with t approximately. The region in red indicates the slip regime where Ū x falls below the peak (at the transition between green and red regions in (b)). (c) Variation of displacement of five QDs (i = 1–5 in (a)), U x,i , with t for an extended period of time for the same experiment in (b). The partial relaxation proceeds over a t ime, t r ≈ 5 s, and U x,i reaches a plateau eventually. (d) Variation of τ with t at V = 0.1 mm s −1 , h = 35 μm, and T = −20 °C. The tracking algorithms fail to properly track the QDs in this case since the displacement in each time step is large. Thus, only the first three columns of the grid (starting from i = 1 in (a)) are tracked i.e. n p ≈ 200, but this increases the error in τ Ogden as the number of tracked particles is reduced, and the effect of any local distortions becomes amplified. (e) Variation of U x,i with t for an extended period of time for the same experiment in (c). Here, we observe stick-slip pulsations with A ≈ 10 μm, and f ≈ 0.1 Hz. Further, the partial relaxation proceeds over a t ime, t r ≈ 5 s, similar to that in (c). We observe a small air pocket developing at the interface at t = 198.8 s (region enveloped by the green line in Fig. 1d ) where only some of the QDs come into focus. The formation of the air pocket is due to the release of dissolved gases during the freezing of water, which is statistical in nature, and can occur anywhere in the bulk or at the interface. 40 Nevertheless, the presence of an air pocket is found to have negligible effect in the present experiments with respect to τ ice . As V increases (see Fig. 1e ), the maximum deformation of the elastomer, (see Fig. 2d ), at t = 4.2 s increases significantly compared to V = 0.01 mm s −1 . This explains the reason for the increase in τ ice with V . As the rate at which the external force applied on the ice increases, the elastomer deformation before slip is higher leading to an increase in adhesion strength. However, the qualitative behaviour of the elastomer i.e. global translation of the QD grid reaching maximum displacement followed by partial relaxation (at t = 15.5 and 25.0 s in Fig. 1e ) is similar in both the cases. Here again, we observe that the partial relaxation proceeds over a time, t r ≈ 5 s (see Fig. 2e ). However, in this case, we observe pulsations reminiscing of stick-slip behavior; a consequence of higher shear rate. 41 The amplitude, A ≈ 10 μm, and the frequency, f ≈ 0.1 Hz, correlates well with the relaxation time of the elastomer i.e. f ∼ 1/ t relax . This is in line with the observations reported in the previous works for polymer melts—rather than elastomers. 41 However, polymer melts are also viscoelastic, and therefore we believe that their behavior in the slip regime can be used to explain the behavior of our elastomer. Finally, as ice crossover is complete, the QD grid returns back to the reference configuration with only a slight offset at t = 41.1 s. Using the digital image correlation method, we estimate the in-plane deformation of the elastomer during the adhesion test at every time step (see Section S3, ESI † ). 42,43 We calculate the shear stress from a 2 nd order Ogden model, τ Ogden , at any given time, t . Despite the inherent complexities involved in the experiment, a good qualitative match is obtained between τ macro and τ Ogden (see Fig. 2b and d ). Further, the shear stress magnitude reduces after attaining the peak ( τ ice ) in our experiments, which indicates that ice already in motion is easier to remove compared to ice at rest. This result is intuitive as the kinetic friction coefficient on any surface is expected to be less than static friction coefficient. Owing to the difficulties involved in tracking QDs at V = 0.1 mm s −1 affecting accuracy, we do not estimate τ Ogden beyond t = 3.99 s. Experiments with V > 0.1 mm s −1 were not successful due to the limitation of the imaging system. 34,35 We also remark here that the alternative method of depositing fluorescent beads 44,45 is inferior to our approach as the beads are significantly less bright than QDs during ice adhesion tests, which results in poor image quality. Further, the intrusive effect of QDs is remarkably low as the thickness of each QD <30 nm. 36 Mechanism and dynamics of ice removal The physical reason for the increase in τ ice with the shear velocity, V , is due to the higher elastomer deformation and higher corresponding force before slip as elucidated with the help of the cTFM experiments. Going beyond this, we seek to obtain a more general relationship quantifying the dependence between τ ice and V . Intuitively, the slip of ice on an elastomer can be considered analogous to the kinetic friction problem of elastomers sliding on smooth solids. Owing to the complexities involved in the phenomenon, several theories are proposed to explain the observed behavior. 46 It has been shown that friction on elastomers is not due to the bulk viscoelastic dissipation, a fact supported by the increase in shear stress with the modulus of the elastomer to start the sliding motion. 6,31 Rather, stochastic processes i.e. the cyclic adsorption and desorption of elastomer chains at the interface control the frictional behavior as proposed by Schallamach 47 and later refined by Chernyak and Leonov. 48 According to these theories, the adhesion of the elastomer with the solid is governed by the van der Waals interactions. When an elastomer slides on a smooth solid, the active load-bearing elastomer chains at the interface stretch, detach, relax, and re-attach to the solid in a cyclic process. Ice stick-slipping on an elastomer is an interesting parallelism for this. In the Chernyak and Leonov model, the tension in each load bearing chain, S , increases monotonically with the shear velocity as S ∼ Vt . Therefore, the frictional force increases with V . The number of active load bearing chains is given by, N = N 0 t b [ t b + t d ] −1 . Here, N 0 is the total polymer chains within the ice–elastomer interfacial area, t b is the average bound time of a chain with ice, and t d is the average lifetime of the chain in the detached state. The energy required for the adhesive link of the chain to break-off from ice is of the form J = J 0 − CS . Here J 0 is the break-off energy when V = 0, and C is a constant. Thus, the probability of adhesive failure of a link at any time, t , increases with V , and consequently t b (thereby N ) reduces while t d remains constant as it is related to the detached state of the chain. The net frictional shear stress is therefore given by the product of the frictional force per unit area and the number of active load bearing chains, 48 1 τ friction = [ aG ×{ VB ×(1 − exp(−1/ VB ))−exp(−1/ VB )}][ m + 1 − exp(−1/ VB )] −1 Here, G is the shear modulus of the elastomer, a is the cotangent of the ultimate angle of a chain at adhesive break-off with respect to the initial position i.e. a = cot ψ , B = t b /( aδ ), δ is the molecular roughness, and m = t d / t 0b , where t 0b = t b when V = 0. The parameters a , and m are independent of elastomer surface temperature, T , while B and G vary. 48 For an incompressible isotropic material, we can estimate the shear modulus as G = E /3 where E is the Young's modulus. 49 We measured the Young's modulus of the material using an indentation technique (see Section S4, ESI † ) at different elastomer surface temperature, T ( Fig. 3a ). Interestingly, we found a non-monotonic variation of E with T as opposed to the widely accepted linear relation E ∝ T . This suggests that the elastomer does not behave as an “ideal rubber” and therefore both internal energy and entropic elasticity effects are important. Fig. 3 Effect of temperature and shear velocity on ice adhesion in shear mode test. (a) Non-monotonic behaviour of Young's modulus of the elastomer, E , with surface temperature, T . (b) Variation of τ ice with shear velocity, V , at an elastomer (without and with 20 wt% Silicon oil) surface temperature, T = −20 °C, and thickness, h = 35 μm. The solid and dashed black curves indicate the best fit of Chernyak and Leonov adhesive friction model ( eqn (1) ) to the experimental data without and with oil, respectively. (c) Variation of τ ice with shear velocity for different Chernyak and Leonov parameters for arbitrarily selected values of a , B , and m . When only a increases ( a 1 < a 2 < a 3 ), the peak value of τ ice increases but the velocity at which the peak is obtained remains constant (black solid lines). When only B increases ( B 1 < B 2 < B 3 ), the curve shifts to the left (blue dashed lines) but the peak value of τ ice remains the same. When only m increases ( m 1 < m 2 < m 3 ), the plateau width after attaining the peak value of τ ice reduces (red dashed lines). (d) Variation of τ ice with elastomer surface temperature, T , at V = 0.1 mm s −1 . The solid curves indicates the Williams–Landel–Ferry transformation to the experimental data for different elastomer thickness, h . Error bars represent standard deviation for n e ≥ 3 independent experiments in (a, b and d). The black data points in (a) represent individual data points. The variation of τ ice with V for h = 35 μm, at T = −20 °C is shown in Fig. 3b . Assuming τ friction = τ ice with G = 5.13 kPa at T = −20 °C, we obtain the parameters through non-linear curve fitting, a = 83.83, B = 241 s m −1 , and m = 2.3 × 10 −8 , all of which seem quite reasonable. 48 We remark here that the viscous retardation in the chains is neglected in our study; due consideration of this phenomenon would result in . Here, s represents the ratio of viscous retardation time of the chain with the average bound time in the rest state, t 0b . Since s ≤ m ≤ 1, 48 we have in our case as m ≪ 1. However, as s → 1, i.e. when the viscoelastic effects of each polymer chain are significant, the retardation effects can become significant. The parameter m correlates with the width of the plateau; increasing m reduces the plateau width. Since a correlates with the maximum stretch of the elastomer chain at the instance of detaching, relatively mobile chains will have lower magnitude of a (see Fig. 3c ). Reducing the parameter B (1/ B represents a molecular velocity scale) merely shifts the curve to right. Finally, τ ice ∝ G as described in previous works. 6,7 Although, controlling the parameters a , m , and B , is challenging experimentally, the insight their variation provides is valuable. The shear modulus, G , we can easily modify and observe the trend of τ ice with V . The infusion of oil into the elastomer reduces the cross-linking density (thereby reducing shear modulus, G ), 6 which leads to a reduction of τ ice . As an example, with all the experimental conditions being the same as in Fig. 3b , we show that infusing the polymer network with 20% silicone oil leads to a reduction in τ ice (dashed line in Fig. 3b ) with a = 117.7, B = 159.9 s m −1 , and m = 1.4 × 10 −7 . The modified shear modulus when oil is infused is estimated as (1 − ω ) 5/3 G = 3.54 kPa where ω corresponds to the oil weight fraction in the polymer network. 7 This indicates that the shear modulus (reduced by 30% in comparison to no oil case) has a dominant effect on τ ice compared to the interfacial chain mobility ( a increased by 40% in comparison to the no oil case). This is also supported from the mathematical form of τ friction ( eqn (1) ). The areal density of active load bearing chains N ∝ G , 7 and therefore we conclude that the areal density of chains has a dominant effect than the interfacial mobility of chains. Finally, the curve shifts to the right since B reduced while m increased which is also a desired consequence from oil addition as the plateau width reduces. We then investigate the effect of elastomer surface temperature, T , on τ ice at V = 0.1 mm s −1 , for h = 35 μm (see Fig. 3d ) where τ ice decreases with T . Employing an analogy from the variation of elastomer friction with temperature, we use the classic empirical Williams–Landel–Ferry (WLF) transformation to explain these results. 50,51 The transformation indicates that the temperature dependence of mechanical processes on elastomers can be estimated using a single empirical function. Accordingly, we have log( τ ice / τ ref ) = [− Y ( T − T ref )][ Z + T − T ref ] −1 where τ ref = τ ice at T ref = 253 K, and Y = 30 and Z = 900 K are empirical constants obtained from the data. The solid curves in Fig. 3d represents the estimate from WLF transformation which matches reasonably well with the experimental data for different elastomer thickness, h (discussed next in detail). To understand the effect of elastomer thickness, h , we performed experiments by varying it over three orders of magnitude (see Section S2, ESI † ). In Fig. 4 , we observe that τ ice is practically independent of h. As h is varied from 0.07 to 64 μm at different shear velocities, V , and elastomer temperature, T = −20 °C, τ ice remains practically constant. While we do observe a minor variation in τ ice for V ≤ 0.1 mm s −1 , the dependency is not as strong as compared to V and therefore can be neglected. This behavior has its origin in the presence of interfacial slip, and that only the adhesive links of the chains near the ice–elastomer interface (not in the much larger elastomer volume of thickness, h ) are expected to play a role on the value of τ ice . This is also reinforced by the fact that the friction model parameters a , m , and B for all the elastomer thicknesses does not vary significantly as shown in Fig. S5 and Table S1 (ESI † ). Fig. 4 Effect of elastomer thickness on ice adhesion in shear mode test. τ ice remains independent of the elastomer thickness, h , at different shear velocities and a surface temperature, T = −20 °C. Error bars represent standard deviation for n e ≥ 3 independent experiments. Mixed and normal mode tests To explore the effect of mode of testing, we also conducted macroscopic experiments in mixed (shear and normal), and pure normal modes. The ice removal mechanism in both the modes is due to interfacial fracture as ice clearly de-bonds from the elastomer (see Fig. S1b, ESI † ). The elastomer develops instabilities at the interface with a characteristic wavelength in the order equal to the thickness of the film, h. 11,12 Fig. 5a shows the variation of ice adhesion strength ( τ ice , σ ice in mixed mode) with V for h = 35 μm at T = −20 °C. The magnitude of τ ice in mixed mode is lower in comparison to shear mode due to the presence of significant normal stresses. Clearly, the trend in mixed mode is similar to the shear mode although the mechanisms are different. Previous studies reported τ ice ∼ 1/ l where l is the height at which the force pin contacts the ice (see Fig. S1b, ESI † ) at the critical shear velocity, where interfacial fracture is observed. 11 Therefore, we propose that the relation τ ice ∼ 1/ l holds across the range of shear velocities in the mixed mode when all the other experimental conditions remain the same. This leads to the relation τ ice,m ∼ ( l s / l m ) τ ice,s where the subscripts m and s indicate mixed and shear mode tests. From experiments, we have l s ≈ 0.5 mm, and l m = 6.1 mm obtaining τ ice,m ∼ 0.1 τ ice,s . The black curve in Fig. 5a represents τ ice,m = 0.45 τ ice,s where τ ice,s = τ friction (using eqn (1) ) with the model parameters a , B , and m being the same as in Fig. 3b (no oil case). The pre-factor of τ ice,s varies slightly in the experiments but is of the same order of the ratio l s / l m and more importantly, τ ice,m and τ ice,s are directly proportional at all values of V . This remarkably simple proposition also works well for different elastomer thicknesses, h (see Fig. S6, ESI † ). By definition, the parameters a , B , and m should not vary in the mixed mode test. We conjecture that the presence of elastic instabilities in the mixed mode test would reduce the number of active load bearing chains, N , as compared to the shear mode test, which leads to a reduction of τ ice . Fig. 5b shows the variation of ice adhesion strength in the normal mode test, σ ice , with V for h = 35 μm at T = −20 °C. In this case, V represents the pull up velocity of the force gauge. σ ice increases with V in a similar fashion as in shear and mixed mode tests. However, the error is significantly higher (see also Fig. S7, ESI † ) making it difficult to determine a clear dependency with V . We attribute this behavior to the partial cohesive failure of the elastomer during the adhesion test (see Fig. S8, ESI † ). Fig. 5 Effect of mode of ice adhesion test. (a and b) Variation of τ ice (filled) and σ ice (unfilled) with shear velocity, V , at surface temperature, T = −20 °C, and elastomer thickness, h = 35 μm in mixed and normal mode tests, respectively. The black solid curve in (a) indicates 0.45 τ ice,s where τ ice,s = τ friction (using eqn (1) ) is the ice adhesion strength in the shear mode when all the other experimental conditions are the same. (c and d) Variation of τ ice (filled) and σ ice (unfilled) with T , at V = 0.1 mm s −1 , and h = 35 μm in mixed and normal mode tests, respectively. The black solid curve in (c) indicates the Williams–Landel–Ferry transformation. Error bars represent standard deviation for n e ≥ 3 independent experiments. \n Fig. 5c and d show the variation of ice adhesion strength ( τ ice , σ ice in mixed mode and σ ice in normal mode) with T at V = 0.1 mm s −1 . Again, using WLF transformation in the mixed mode, we can estimate the temperature dependence of τ ice with T (see Fig. 5c and Fig. S9, ESI † ). σ ice in the normal mode remains constant in the entire temperature domain for a given h . Further, the magnitude of σ ice in the normal mode (see Fig. 5d ) is much higher than that in the mixed mode. Again, the error is significantly high to predict any conclusive trend (see also Fig. S10, ESI † )."
} | 8,433 |
19859812 | null | s2 | 3,683 | {
"abstract": "We describe a simple and reliable fabrication method for producing multiple, manually activated microfluidic control valves in polydimethylsiloxane (PDMS) devices. These screwdriver-actuated valves reside directly on the microfluidic chip and can provide both simple on/off operation as well as graded control of fluid flow. The fabrication procedure can be easily implemented in any soft lithography lab and requires only two specialized tools-a hot-glue gun and a machined brass mold. To facilitate use in multi-valve fluidic systems, the mold is designed to produce a linear tape that contains a series of plastic rotary nodes with small stainless steel machine screws that form individual valves which can be easily separated for applications when only single valves are required. The tape and its valves are placed on the surface of a partially cured thin PDMS microchannel device while the PDMS is still on the soft-lithographic master, with the master providing alignment marks for the tape. The tape is permanently affixed to the microchannel device by pouring an over-layer of PDMS, to form a full-thickness device with the tape as an enclosed underlayment. The advantages of these Tape Underlayment Rotary-Node (TURN) valves include parallel fabrication of multiple valves, low risk of damaging a microfluidic device during valve installation, high torque, elimination of stripped threads, the capabilities of TURN hydraulic actuators, and facile customization of TURN molds. We have utilized these valves to control microfluidic flow, to control the onset of molecular diffusion, and to manipulate channel connectivity. Practical applications of TURN valves include control of loading and chemokine release in chemotaxis assay devices, flow in microfluidic bioreactors, and channel connectivity in microfluidic devices intended to study competition and predator/prey relationships among microbes."
} | 476 |
28233813 | PMC5324122 | pmc | 3,684 | {
"abstract": "Microbial symbionts in sponges are ubiquitous, forming complex and highly diverse host-specific communities. Conspecific sponges display remarkable stability in their symbiont communities, both spatially and temporally, yet extreme fluctuations in environmental factors can cause shifts in host-symbiont associations. We previously demonstrated that the marine sponge Hymeniacidon heliophila displayed significant community-level differences in microbial symbiont diversity, structure and composition when sampled from intertidal and subtidal environments. Here, we conducted a 70-day reciprocal transplant experiment to directly test the effect of tidal exposure on the microbiome of H. heliophila , using next-generation Illumina sequencing of 16S rRNA gene sequences to characterize symbiont communities. While sponges transplanted between habitats displayed shifts in microbial communities after 70 days, temporal variation was the dominant factor affecting microbial community composition. Further, we identified core symbionts that persisted across these spatio-temporal scales and used a metagenomic approach to show that these dominant members of the microbiome of H. heliophila represent nitrogen cycling taxa that have the potential to contribute to a diverse array of nitrogen transformations in the sponge holobiont. Together, these results indicate that despite moderate spatio-temporal shifts in symbiont composition, core symbiont functions (e.g. nitrogen cycling) can be maintained in sponge microbiomes through functional redundancy.",
"discussion": "Discussion Previously, we demonstrated variation in the microbiome of H. heliophila from intertidal and subtidal environments, with significant differences in the composition of both dominant and rare members of the microbiome 39 . Herein, a reciprocal transplant experiment was conducted between the same intertidal-subtidal environments and demonstrated that transplantation of sponges between habitats induced shifts in their microbial communities. These treatment differences were driven by shifts in a few microbial taxa, including a greater abundance of the most common OTU1 (alphaproteobacterium, Novosphingobium resinovorum ) in subtidal treatments and a greater abundance of OTU12 (alphaproteobacterium, family Rhodobacteraceae) in intertidal treatments. While some inter-annual variability in the magnitude and assortment of compositional differences was detected between intertidal and subtidal sponges (July 2014 39 vs. July 2015, this study), the shifts in dominant symbiont taxa that we observed in experimental sponges indicate that tidal exposure does select for a unique consortium of microbial taxa in each environment. This study also revealed that the microbiome of H. heliophila displays temporal variation, with significantly different microbial symbiont community structure, composition, and diversity between sponges sampled in May and July 2015. In fact, temporal variation was the dominant factor affecting microbial community composition in the reciprocal transplant experiment. To date, few sponge species have been reported to demonstrate temporal variability in symbiont structure 14 and most represent low microbial abundance (LMA) sponges 20 21 22 , with the exception of the high microbial abundance (HMA) sponge Aplysina cauliformis 33 . Similarly, recent studies in the Mediterranean Sea reported species-specific bacteria in both sponge groups 42 , yet greater temporal variability of microbial communities associated with LMA than HMA sponge hosts over one year 43 . Our results support the hypothesis that some LMA sponges display significant temporal variability, as H. heliophila has been classified as a LMA sponge based on transmission electron microscopy observations and microbial species composition 44 45 . Interestingly, two sponge species for which temporal variation has been previously reported are closely related to H. heliophila : the congeneric sponge H. sinapium 22 and Halichondria panacea 21 , which belong to the same family (Halichondridae) as H. heliophila . Further, spatial variation in archaeal communities of H. heliophila was detected between sponges collected inside and outside of a polluted estuary in Brazil 41 . Together, these results suggest that certain taxonomic groups of LMA sponges exhibit more dynamic microbial symbiont communities than HMA sponges, thus representing important targets for understanding how host-symbiont interactions vary in response to changing environmental conditions. Temporal variation in the sponge microbiome may result from a physiological response of the microbial community to temporal fluctuations in abiotic factors, from differentially adaptive microbial symbionts during life history-associated seasonal changes in the sponge host, or a combination of both effects. While temperature represents an important driver of seasonal variation in free-living microbial communities 46 , greater differences in temperature regimes did not account for greater variation in the microbiome of H. heliophila . Rather, we report greater variation in sponge microbiomes over time (coinciding with a modest temperature increase) than between intertidal and subtidal habitats (where markedly different temperature regimes occurred), indicating that temperature may not play a major role in driving temporal variation in the microbiome of H. heliophila . Other environmental factors that could have affected the observed temporal variation include salinity, dissolved oxygen levels, or the availability of food resources to the sponge, but these variables were not measured in this experiment. Temporal variation in symbiont communities of H. heliophila may also be linked to seasonal cycles in host sponge reproductive processes, such as embryogenesis and gametogenesis, as reported for the congeneric intertidal sponge H. sinapium from the Yellow Sea 22 . H. sinapium exhibited clear seasonal variation in microbial symbiont communities over 1.5 years that coincided with four distinct developmental and reproductive stages of the host sponge 22 . Future studies incorporating host life cycle data alongside other temporally variable environmental factors are needed to further elucidate the drivers of variation in the sponge microbiome. Our study also revealed insight into the interactive effects of season and habitat on microbial symbiont community structure. During the summer season (July), significant differences were detected between intertidal and subtidal sponge microbial communities, consistent with previous results 39 ; however, these differences were not detected during the spring season (May). We hypothesize that annual re-colonization of intertidal habitat by subtidal sponge individuals may explain the lack of microbial community variation between intertidal and subtidal sponges in May. Intertidal populations of H. heliophila decrease markedly during the winter months in North Carolina 47 (B. Weigel pers. obs.), possibly due to exposure to low air temperatures as this is the northern limit for the intertidal form of H. heliophila 47 . As spring approaches, intertidal populations of H. heliophila increase in North Carolina and may include colonization by overwintering, subtidal sponges. While more detailed observations of the seasonal dynamics of H. heliophila are necessary, the greater differences between intertidal and subtidal hosts in summer compared to spring may reflect adaptation of microbial communities to the intertidal environment throughout the summer. In addition to investigating variation in microbial symbiont community structure, we examined functional genes involved in nitrogen cycling and how predicted gene counts varied over time. Nitrogen metabolism genes identified in both the predicted and actual metagenomes of H. heliophila represented a broad functional diversity of nitrogen transformations, including ammonia assimilation, ammonia oxidation, assimilatory nitrate reduction, dissimilatory nitrate reduction, and denitrification. Ammonia assimilation was the dominant nitrogen transformation pathway in the H. heliophila holobiont (ca. 60–80% of all predicted and actual nitrogen metabolism genes), consistent with studies showing very high rates of ammonium excretion by sponges 48 49 . While some sponge species host diverse ammonia-oxidizing symbionts 32 50 , ammonia oxidation genes accounted for only 2–3% of nitrogen metabolism genes in H. heliophila . Genes involved in denitrification and dissimilatory nitrate reduction pathways were detected, accounting for ca. 15% of all nitrogen metabolism genes. Denitrification has been quantified in multiple sponge species 6 7 and genes for nitrite ( nirK ) and nitrous oxide reductase ( nosZ ) have been detected in diverse sponge species 51 . The presence of the complete suite of denitrification genes in the H. heliophila holobiont suggests that nitrate can be reduced all the way to atmospheric N 2 gas. Nitrogen transformations were predicted to occur primarily in the bacterial phyla Proteobacteria, Planctomycetes, Actinobacteria, Cyanobacteria, Bacteroidetes and Verrucomicrobia . The core microbiome of H. heliophila consisted of 18 OTUs, which together comprise an average of 63% of the total microbial community of H. heliophila , and each member of the core microbiome was predicted to be associated with at least one nitrogen metabolism gene. These results suggest that nitrogen cycling taxa are dominant, persistent members of the microbiome of H. heliophila and may perform nitrogen transformations of ecological importance to the sponge host. Determining the extent of functional redundancy in microbial symbiont communities is a current research priority in the field of sponge microbiology, as this knowledge aids in understanding and predicting how the sponge holobiont will respond to changing environmental conditions 52 . While the symbiont community composition in H. heliophila displayed significant temporal variation, no corresponding differences in predicted functional gene counts were detected, suggesting that core nitrogen cycling processes were maintained through functional redundancy. The sole exception was a significant difference in assimilatory nitrate reduction genes over time; however, these genes comprised a small proportion of all nitrogen cycling genes in H. heliophila ( Supplementary Fig. S4 ). Previous work suggests that specific functions of the sponge microbiome are conserved across different host species, as the Mediterranean sponges Agelas oroides and Chondrosia reniformis harbored distinct microbial communities but displayed similar rates of nitrification and nutrient uptake 53 . Additionally, functionally similar nitrogen cycling genes were detected in six taxonomically diverse sponge species with divergent microbial communities 54 . Functional redundancy is likely to play a role in microbial communities with high genetic diversity, as nutrient cycling functions can be performed by diverse taxa and some microbial groups display flexibility in their metabolic functions 55 . Furthermore, selective pressures may maintain nutrient cycling functionality in sponge-microbe symbioses if these metabolic processes are important for host sponge physiology (e.g. nitrogenous waste removal). While we report stable nitrogen cycling functions in the microbiome of H. heliophila following temporal shifts in community composition and seasonal changes in temperature, exceeding thermal stress limits can cause drastic changes in microbial community structure and functional collapse of the sponge holobiont 56 . Future work should seek to understand how the functionality of the sponge holobiont is affected by temporal variation in microbial community structure, as well as how more extreme environmental perturbations and future climate regimes may impact the stability and health of the sponge holobiont."
} | 2,996 |
39322641 | PMC11424623 | pmc | 3,685 | {
"abstract": "Inspired by the starfish's unique ability to achieve flexibility and posture-holding with minimal energy expenditure, we present a novel bioinspired morphing structure. Our two-component design, consisting of a thermoplastic mesh and elastomeric jacket, effectively mimics the functions of the starfish's ossicles, mutable collagenous tissues, and derma. This structure exhibits a remarkable combination of self-healing, time-dependent shape memory, and self-posture-holding properties. Systematic variations in mesh geometry demonstrate precise control over structural stiffness and thermal response, enabling customization for specific applications. The structure's scalability and ease of fabrication further enhance its adaptability. We experimentally demonstrate the potential of our biomimetic morphing structure using several prototypes. This work lays the foundation for developing a new type of versatile morphing structures with applications in diverse fields, including robotics, biomedical devices, and adaptive structures.",
"introduction": "Introduction Morphing structures are load-bearing structures designed to alter their shape, allowing them to adapt to different operational requirements or external conditions 1 , 2 . These structures are increasingly gaining attention in research and application across various sectors, including automotive, aerospace, and power plants 3 – 7 . Traditionally, standard mechanical joints are used to design such morphing structures, where rigid separate elements are connected to kinematic chains using joints with low friction and predetermined degrees of freedom. Another approach involves using complaint monolithic mechanisms, where flexible elements are used to create a structure and through their deflections, the structure can adapt to new shapes 8 , 9 . The concept of 4D structures adds an exciting dimension to this field. In 4D structures, the transformation is triggered by stimuli over time, such as temperature, humidity, or other environmental changes 10 , 11 . Recent advancements have significantly extended the capabilities of 4D structures. For instance, voxelated machines have emerged as a novel approach, leveraging the precise control of material properties at the voxel level to create complex, adaptable structures with high degrees of functionality 12 . Similarly, bio-inspired micro-origami 4D tessellations have been developed, offering new ways to achieve self-morphing behaviors through intricate geometric designs 13 . The integration of liquid–metal-network actuators into an elastomer matrix combines the flexibility and responsiveness of liquid metals with the durability of polymeric networks, enabling the creation of structures that are both highly adaptable and responsive to external stimuli, capable of changing shape or function as needed 14 . Additional studies have further expanded the capabilities of 4D printing. For example, research on adaptive metamaterials by functionally graded 4D printing has demonstrated how gradient-based material properties can be programmed during fabrication to create self-folding or self-coiling structures that respond to environmental changes 15 . Similarly, geometric design pattern-driven single-material 4D printing for self-morphing actuators has shown that even single-material systems can achieve sophisticated morphing behaviors when designed with appropriate patterns 16 . These morphing structures, while effective in achieving desired shape changes, share a common limitation: they require continuous energy input and active control to maintain their posture. Conventional frictionless mechanical joints rely on external control systems to hold the structure in place; without this, the structure may revert to its original shape or become unstable. Similarly, compliant mechanisms, which depend on the deflection of flexible elements, need constant external forces to maintain their shape. Even in 4D structures, where shape change is triggered by environmental stimuli, an ongoing external force is often required to sustain the new configuration. To minimize the requirement for a continuous energy supply to uphold a structure's geometry, locking mechanisms are occasionally utilized to secure these structures in the desired configuration 17 – 19 . However, these structures are not able to hold their changed morphological position by themselves. Very few morphing structures are designed in such a way that they have more than one stable state and can hold their shape without requiring any external energy input 20 , 21 . However, still energy and control are essential for maintaining alternative morphological positions. In the last years a few bio-inspired solutions for the mechanical challenges of morphing structures have been suggested, mostly focusing on morphogenesis and actuation 22 – 24 . We here present a new bio-inspired approach to address the energy consumption issue faced by morphing structures. Our bio-inspired design is based on the starfish skeleton, an echinoderm belonging to class Asteroidea. Starfish can bend, move and twist their arms with great degrees of freedom 25 , 26 . Alongside its structural versatility, the starfish skeleton has the unique capability to be able to hold any body posture with minimal expenditure of energy 27 , 28 . Simply stated, starfish skeletons consist of thousands of small bone-like magnesium calcite structures called ossicles 29 , 30 . These ossicles are connected by small inter ossicular muscles (IOMs) and form a complex mesh with ridge-shaped geometrical patterns 31 , 32 . This whole assembly of ossicle-IOMs is connected and covered with mutable collagenous tissues 25 , 31 . Using neural control mechanisms, starfish can control the stiffness of this collagenous tissue to hold their posture for prolonged periods 27 , 28 . These unique capabilities make starfish an excellent example for a bio-inspired energy-efficient morphing structure in nature. To better understand the functional principles and detailed morphology of the complex starfish skeleton, we initially performed high-resolution X-ray CT imaging on Asterias rubens specimens. Based on our findings we then assigned specific technical roles to different skeleton components, depending upon their functions during morphing and posture-holding processes. Subsequently, we identified commercially available and easy to process materials capable of fulfilling these roles and designed a morphing structure inspired by the starfish skeleton. The resulting morphing structure exhibits the capability to maintain any posture without the need for external energy or control. Alongside, it exhibits a time-dependent shape memory effect (4D) which allows to design smart morphing structures. Our principle also allows for full customizability of mechanical and thermal properties, scalability in length, width, and depth, and shows self-healing properties.",
"discussion": "Discussion This study presents a novel bio-inspired morphing structure inspired by the self-locking ability of the starfish skeleton. The morphing structure design employs only two easily fabricated components: a thermoplastic mesh and an elastomeric jacket. The thermoplastic mimics both the load-bearing ossicles and the posture-locking mutable collagenous tissues of the starfish skeleton. The elastomer, analogous to the starfish's derma, provides encapsulation, and protection from the external environment and enables controlled shape change (Fig. 1 ). By systematically varying the geometric design of the thermoplastic mesh, we achieved precise control over structural stiffness, time-dependent thermal softening response, and ultimately, morphing behavior (Figs. 2 , 3 , and 4 ). This level of control over morphing behavior distinguishes our work from previous research, where 4D effects were typically induced through direct 3D printing of thermoplastic materials, either alone or in combination with geometric design 15 , 16 , 33 . Additionally, some studies have utilized hinges made from elastomer materials that store energy by accumulating strain or deformation during the folding process. When the external stimulus is removed or altered, the stored energy in these hinges is released, causing the structure to revert to its original shape or maintain a new configuration, depending on the design 34 . In our case, the time-dependent shape memory behavior of our prototypes stems from the interplay between the thermoplastic mesh geometry and the silicone rubber jacket. This allows us to precisely control morphing behavior by tailoring the mesh geometry. Meshes with greater porosity relative to structure thickness exhibited lower stiffness and faster thermal response, ideal for rapid shape change. Conversely, designs with thicker structures relative to porosity offered higher stiffness and slower thermal response, suitable for applications requiring robust shape retention. Interestingly, within the group of structures with comparable porosity and structure thickness (T1P1, T2P2, and T3P3), all displayed similar stiffness. However, the mesh with the finest structure (T1P1) exhibited a faster thermal response compared to T2P2 and T3P3 (Figs. 3 and 4 ). This could be attributed to the enhanced heat transfer through the finer mesh, allowing the thermoplastic to reach a higher temperature more quickly and thus soften more rapidly. Across all designs, we achieved a significant stiffness reduction of over 92% upon transitioning from the hard to soft state, highlighting the structure's adaptability (Fig. 3 ). In addition, our morphing structure demonstrates noticeable self-healing properties. X-ray CT scans confirmed complete crack closure following thermal treatment (Fig. 5 A–D). Although there are some variations in stiffness and load to failure in the healed structure during different healing cycles (Fig. 5 E), these can likely be attributed to the general behavior of the thermoplastic during the re-heating process. As the 3D-printed layers melt and fuse, slight geometric changes occur in the healed structure compared to the original. Furthermore, the reduction in the mechanical properties may be due to thermoplastic degradation from repeated thermal cycling 35 – 38 . Despite these slight variations, the self-healing capability of our structure substantially enhances its reusability and overall lifespan. Our morphing structure balances design simplicity with exceptional functionality. It demonstrates self-locking, continuous bending with a low radius of curvature, scalable and customizable design, self-healing, and time-dependent shape memory effects. These properties, combined with its ease of fabrication and low production costs, make it a highly promising candidate for diverse applications in fields like robotics, biomedical devices, prosthetics, aerospace, and beyond. To demonstrate the versatility of our morphing structure, we created a hand-shaped prototype. This design showcased our ability to tailor both the mechanical properties and shape memory behavior of individual \"fingers”, enabling complex and time-dependent programmed 4D actuation (see supplemental video 2 ). Our current prototypes open up several avenues for future research. Expanding the range of stimuli beyond heat to include light, electricity, or other triggers could enable novel actuation methods. Additionally, exploring alternative thermoplastic-elastomer combinations could yield tailored properties for specific applications. Integrating embedded actuators directly into the morphing structure would increase control complexity. Furthermore, developing advanced control strategies could enable precise shape changes and complex morphing sequences. Finally, investigating the tessellation of morphing structure units could lead to the design of large-scale morphing structures for countless applications."
} | 2,967 |
32499692 | PMC7243343 | pmc | 3,686 | {
"abstract": "While interacting with the world our senses and nervous system are constantly challenged to identify the origin and coherence of sensory input signals of various intensities. This problem becomes apparent when stimuli from different modalities need to be combined, e.g., to find out whether an auditory stimulus and a visual stimulus belong to the same object. To cope with this problem, humans and most other animal species are equipped with complex neural circuits to enable fast and reliable combination of signals from various sensory organs. This multisensory integration starts in the brain stem to facilitate unconscious reflexes and continues on ascending pathways to cortical areas for further processing. To investigate the underlying mechanisms in detail, we developed a canonical neural network model for multisensory integration that resembles neurophysiological findings. For example, the model comprises multisensory integration neurons that receive excitatory and inhibitory inputs from unimodal auditory and visual neurons, respectively, as well as feedback from cortex. Such feedback projections facilitate multisensory response enhancement and lead to the commonly observed inverse effectiveness of neural activity in multisensory neurons. Two versions of the model are implemented, a rate-based neural network model for qualitative analysis and a variant that employs spiking neurons for deployment on a neuromorphic processing. This dual approach allows to create an evaluation environment with the ability to test model performances with real world inputs. As a platform for deployment we chose IBM's neurosynaptic chip TrueNorth. Behavioral studies in humans indicate that temporal and spatial offsets as well as reliability of stimuli are critical parameters for integrating signals from different modalities. The model reproduces such behavior in experiments with different sets of stimuli. In particular, model performance for stimuli with varying spatial offset is tested. In addition, we demonstrate that due to the emergent properties of network dynamics model performance is close to optimal Bayesian inference for integration of multimodal sensory signals. Furthermore, the implementation of the model on a neuromorphic processing chip enables a complete neuromorphic processing cascade from sensory perception to multisensory integration and the evaluation of model performance for real world inputs.",
"introduction": "1. Introduction While interacting with the world our senses are exposed to a rich and constant flow of information. Making sense of this vast of information is one of the most important task of our brain and crucial for survival. It does this by combing complementary information about the same event from different senses into a single percept. This integration process leads to an enhancement of the combined signal, thus supports the detection of events or objects of interest, improves disambiguation and allows for faster and more accurate processing than could be derived by a mere linear combination of unimodal information streams (Stein and Stanford, 2008 ). Humans and other mammals are equipped with complex neural circuits to ensure fast, reliable and optimal combination of signals from various sensory organs (Marrocco and Li, 1977 ; Edwards et al., 1979 ; Cadusseau and Roger, 1985 ). This multisensory integration (MSI) process can be found already in the superior colliculus (SC) of the brain stem where auditory, visual and vestibular signals are combined to facilitate fast reflexive eye movements (Stein et al., 1983 ). This integration process is refined on ascending cortical pathways for higher level processing and decision making. The SC is a melting pot of information from various sensory modalities and neurons in the SC are the first multimodal processing units in ascending sensory pathways (Meredith and Stein, 1983 ; Wallace and Stein, 1997 ) with spatially aligned receptive fields to these modalities (Meredith and Stein, 1996 ). The superficial layers of the SC receive mainly retinotopic inputs from the visual system and respond only to visual signals (Wallace et al., 1998 ). However, neurons in deeper layers of the SC gradually receive ascending inputs from other modalities and exhibit receptive fields for these modalities. In addition, their responses are multi-modal, i.e., receiving input from two different modalities leads to response characteristics that are different than responses to uni-modal signals (Stein and Stanford, 2008 ). Inputs to neurons in deep layers come from a diverse set of sensory systems and range from auditory signals from the inferior colliculus to proprioceptive signals from the vestibular system. To create a common frame of references for these different signals and, thus, spatially align them the retinotopic visual input is used as a guidance signal. This has been demonstrated in neurophysiological studies as well as modeling investigations (Rees, 1996 ; Wallace et al., 2004 ; Oess et al., 2020a ). Despite of all the ascending sensory signals in the SC, neurophysiological studies in cats indicate that there are several descending projections from the association areas (AES) of the cortex. Unimodal cortical projections from anterior ectosylvian visual area (AEv) and the auditory field of the anterior ectosylvian region (FAEs) are observed (Meredith and Clemo, 1989 ; Wallace et al., 1993 ; Wallace and Stein, 1994 ). These projections seem to play an essential role for the integration ability of SC neurons. Studies demonstrate that when these projections are deactivated, the neurons in the SC loose all their multisensory response characteristics (Alvarado et al., 2007a ). These characteristics of SC neurons are complex and are the result of not just descending cortical projections but also neural circuitry and dynamics in the SC as we will describe later. One of such a response characteristic is the so called multisensory enhancement which describes an enhanced activity for multisensory input signals that is higher than the linear combination of all unisensory inputs (Stein and Stanford, 2008 ). Such multisensory enhancement changes with the intensities of the input signals and creates the commonly observed and described inverse effectiveness for multi-modal signals of MSI neurons (Perrault et al., 2003 ; Stein and Stanford, 2008 ). That is, low intensity multimodal stimuli in spatial and temporal register lead to an enhanced response of MSI neurons which is greater than the summed responses for separately presented unimodal stimuli (super-additivity). In contrast, for high intensity multimodal stimuli, responses tend to be smaller than the sum of unimodal responses (sub-additivity). As a consequence, the probability of detecting low intensity events registered by two or more senses is increased. Another important response characteristic of MSI neurons is the suppression for bimodal signals outside the receptive field of the neuron (Meredith and Stein, 1996 ). That is, the otherwise strong activity of SC neurons is suppressed by input signals of another modality with spatial or temporal offsets (spatial and temporal principle of multisensory integration). This suppression leads to a sub-additive combination of the two stimuli and thus can be seen as a means to prevent fusion of input stimuli that do not belong to the same event. All these response characteristics can only be observed for active descending cortical feedback from association areas to MSI neurons in the SC (Jiang et al., 2001 , 2007 ; Jiang and Stein, 2003 ; Alvarado et al., 2007a , 2009 ). When cortical projections or corresponding cortical areas are deactivated, multimodal response characteristics vanish (Rowland et al., 2014 ; Yu et al., 2016 ). Thereby, AES cortical feedback projections mediate multisensory integration abilities in SC neurons. The aim of such a complicated integration process is to infer a percept that is more reliable and robust than unimodal perceptions. In fact, studies pointed out that humans integrate signals from different modalities in a statistically optimal fashion (Ernst and Banks, 2002 ), so called Bayes optimal . That is, they weight each signal based on its reliability before linearly combining them. Thereby, increasing the certainty of the combined signal. In addition, it has been suggested that in order to integrate signals in such an optimal way, neural populations need to be able to encode and integrate sensory signals Bayes optimally (Deneve et al., 2001 ; Ma et al., 2006 ). Hence, one of the challenges in computational modeling of multisensory integration is to demonstrate that a model integrates its input signals in a Bayes optimal or at least near-optimal way and to explain how the variety of response characteristics can emerge from population dynamics. We introduce a neural network model of multisensory integration of audio-visual signals that exhibits such near-optimal Bayesian behavior, incorporates cortical feedback and demonstrates typical multi-sensory response characteristics. The contribution of this work is several-fold: First, we introduce a neural network model of conductance-based neurons in the superior colliculus that incorporates neurophysiological plausible cortical feedback connections. We investigate how this feedback alters the responses of multisensory neurons and enables them to integrate signals from multiple sensory streams. In addition, we examine what enables this process to integrate multimodal signals in a Bayesian optimal fashion and demonstrate that the introduced model does near-optimal Bayesian inference. This finding links the algorithmic mechanisms and representations to functionality. In a second part, we incorporate a spike-based output encoding of the model and deploy it on IBM TrueNorth neurosynaptic system (Cassidy et al., 2014 ), a neuromorphic processing chip with connections to neuromorphic sensory systems. Evaluations with this neuromorphic model demonstrate the performance for real world input data. This is a novel approach of testing a biological inspired architecture since it enables a complete neuromorphic processing cascade from sensory perception to multisensory integration and the evaluation of model performance for real world inputs.",
"discussion": "4. Discussion We introduced two implementations of a neural model simulating functions of SC neurons for integration of audio-visual signals. The model incorporates modulatory cortical feedback connections to facilitate enhancement of multisensory signals. The rate-based implementation of the model and its responses were evaluated in various simulation experiments and we demonstrated the importance of cortical feedback projections for near-optimal integration of signals. Furthermore, the spike-based model implementation on neuromorphic hardware showed its capability of integrating real world spike inputs from neuromorphic sensors. 4.1. Multisensory Integration Typical multisensory neurons show response enhancement for multimodal stimuli that arrive in temporal and spatial coincidence (Meredith and Stein, 1996 ). Previous studies report that this property only arise for enabled cortical feedback projections (Stein et al., 1983 ; Wallace and Stein, 1994 ; Jiang et al., 2001 ; Alvarado et al., 2007b ). Our model results replicate such observations and show that response enhancement can vary with the gain of the modulatory cortical feedback projections controlled by λ parameter (Equation 2) in the model (see Figures 3D,E ). Thus, the gain of how neurons integrate modulatory feedback could explain the observed variety of multisensory enhancement in responses of SC neurons as has been observed previously (Kadunce et al., 2001 ). Without cortical projections the response to multisensory input remains sub-additive (see Figure 3B ) even for high input intensities. Such cortical projections are only activated when both modality specific cortical signals are active. If only one cortical region is active multisensory response properties vanish. This is in line with findings in cats, where multisensory integration disappears for deactivated cortical areas (Meredith and Clemo, 1989 ; Alvarado et al., 2009 ). In our model, this is achieved with a specially designed cortical cross-modal forward inhibition circuit in the feedback projections (see Equation 9). Furthermore, the model follows the previously described spatial principle of MSI neurons (Meredith and Stein, 1996 ) (see section 1 for definition) by suppressing responses for bimodal stimuli with large spatial offsets. We would like to point out that this suppression is achieve merely by dynamic interactions between the pool normalization, the feedforward inhibition and excitation of sensory neurons, thus implicitly creating a center surround receptive field of MSI neurons. 4.1.1. Bayesian Inference Several investigations show that afferent connections from cortical regions to the SC are necessary for multisensory integration (Alvarado et al., 2007a , 2008 , 2009 ). However, the functional purpose of such feedback projections is still unclear. Our Simulation experiments show that multisensory integration of two input signals in a near-optimal Bayesian way appears only when cortical feedback projections are active. The variance of the integrated signal is substantially similar to the computed, optimal value for active projections than compared to responses without these projections (see Figure 6G ). This is especially true for larger spatial offsets of the two input stimuli (see Figure 6F ). Thus, we hypothesize that one purpose of cortico-collicular feedback is to facilitate optimal integration of multimodal signals and that such an optimal integration might already happen on the level of the SC. Presented model variance values ( Figure 6H ) exhibit an offset for higher variance values which we assume to result from the static size of the receptive field of model neurons and could be compensated with a dynamically changing receptive field depending on sensory certainty. 4.2. Neuromorphic Implementation We demonstrated that the proposed model architecture is suitable for robotic applications by implementing it on a real-time neuromorphic processing chip. Preliminary results for real world spike recordings obtained by neuromorphic sensory hardware suggest that the model is robust and capable of integrating real world multisensory signals. It was shown that the model's ability to fuse two modalities into a single percept changes with cortical feedback projections. This supports the hypothesis that cortex plays a crucial role in determining whether two stimuli belong to the same event or if they represent two separate events. This is further investigated in a last experiment in which the cortical feedback signals are fixed to the location of the auditory sensory input while the visual sensory input is spatially shifted. The response enhancement remains at the auditory location even if the sensory visual input is not present anymore. This can be interpreted as an increased cortical focus for this specific location. Thus, cortical projections might be controlling the mandatory fusion range of multisensory neurons and in addition serve as a spatial attention signal, as has been suggested by McDonald et al. ( 2001 ); Mozolic et al. ( 2008 ), and Talsma et al. ( 2010 ). In future experiments, we are planning to implement such a spatial attention mechanism in order to selectively choose which multisensory signals should be enhanced. We believe that this could be accomplished by a more sophisticated cortical feedback signal with spatial properties different than the perceived sensory inputs. 4.3. Comparison to Other Models for Multisensory Integration Several models that account for multisensory integration in the colliculus of different granularity and focus have been suggested over the years. Some of them try to explain the various response properties of MSI neurons (Anastasio and Patton, 2003 ; Ursino et al., 2017 ) whereas others focus more on the biological detailed architecture (Cuppini et al., 2011 , 2017 ; Casey et al., 2012 ). In the following, we will describe two of them and point out their strengths and weaknesses compared to our presented model. In Rowland et al. ( 2007 ), the authors presented an algebraic and compartmental model of multisensory integration that incorporate cortico-collicular projections and try to explain the existence of AMPA and NMDA receptors in MSI neurons. Their goal was to reproduce a variety of physiological findings without paying much attention to the underlying biological anatomy and structure. Like our model, the authors are able to reproduce several MSI characteristics like multisensory enhancement, inverse effectiveness and super- and sub-additivity. In addition to our presented results, they also demonstrate the MSI neuron dependence on NMDA receptors and the temporal window of integration of their model. However, they did not present any results that indicates a Bayesian optimal integration of the signals. Another approach is taken by Ohshiro et al. ( 2011 ) and their normalization model in which they show that many of the MSI response characteristics can be achieved by a pool normalization of the neuron output. Their model assumes MSI neurons that integrate signals according to a linear weighted sum with different input weights across modalities and neurons. In addition to the replication of MSI characteristics, the authors performed a virtual experiment of vestibular-visual integration task with their model and provided data that closely resembles findings in monkeys. Despite the profound analysis of their model and resemblance of experimental data, the authors neglect cortical projections to MSI neurons entirely. 4.4. Limitations of the Model As we have shown, the two proposed model implementations using rate-based and spike-based encoding are both able to replicate several physiological findings, predict the purpose of cortical modulatory projections and are capable of reliably processing real world spiking data. One of the drawbacks of the current implementations is the lack of any learning mechanism in the process. The model assumes that all connections are already established and inputs are spatially aligned, even though, studies show that multisensory integration emerges during maturation of the nervous system by a constant exposure to multimodal signals (Wallace and Stein, 1997 ). This long term exposure influences how and to what extent multisensory integration takes place. This limitation in our model could be tackled by incorporating a Hebbian correlation learning mechanism between the cortical feedback projections and MSI neurons as well as the inputs of the model. The current assumption that the two modalities are spatially aligned is a strong constraint and simplifies the model architecture but is not biologically plausible. We are confident that this can be overcome with a previously proposed architecture of spatial map alignment of visual and auditory inputs (Oess et al., 2020a ). 4.5. Outlook The proposed model implementations of MSI neurons set a solid basis for future investigations. One important question we are planning to investigate is the role of the cortical feedback. One plausible hypothesis is that the feedback projections can be controlled by an attention mechanism to set special focus on a particular region and thereby enhances signals at that spatial location. This is an essential mechanism when conflicting events are present. In addition, the spike-based implementation of the model on neuromorphic hardware is an important step toward a real-time capable robotic platform. This platform will be equipped with audio and visual sensory hardware which directly communicates with the neuromorphic processing chips via spike trains, thereby creating a complete neuromorphic system from the sensory perception to decision making and action execution."
} | 5,023 |
37458125 | PMC10909567 | pmc | 3,687 | {
"abstract": "Abstract The persistent exposure of coral assemblages to more variable abiotic regimes is assumed to augment their resilience to future climatic variability. Yet, while the determinants of coral population resilience across species remain unknown, we are unable to predict the winners and losers across reef ecosystems exposed to increasingly variable conditions. Using annual surveys of 3171 coral individuals across Australia and Japan (2016–2019), we explore spatial variation across the short‐ and long‐term dynamics of competitive, stress‐tolerant, and weedy assemblages to evaluate how abiotic variability mediates the structural composition of coral assemblages. We illustrate how, by promoting short‐term potential over long‐term performance, coral assemblages can reduce their vulnerability to stochastic environments. However, compared to stress‐tolerant, and weedy assemblages, competitive coral taxa display a reduced capacity for elevating their short‐term potential. Accordingly, future climatic shifts threaten the structural complexity of coral assemblages in variable environments, emulating the degradation expected across global tropical reefs.",
"conclusion": "Conclusions A limited understanding for the abiotic determinants driving the dynamics of coral assemblages inhibits our capacity to predict their future performance and, therefore, manage global coral community reassembly (Edmunds, 2020 ; Edmunds et al., 2014 ; Edmunds & Riegl, 2020 ). Here, we demonstrate how coral populations can adopt demographic strategies associated with enhanced short‐term potential to improve their viability when exposed to greater abiotic variability. However, strategies of enhanced short‐term potential come at a cost to the long‐term demographic performance characteristics of coral populations. Despite presenting a framework for quantifying population resilience (Capdevila et al., 2020 ), the short‐term demographic characteristics of coral assemblages remain largely overlooked (Cant, Cook, et al., 2022 ). Yet, our findings emphasize that the winners and losers in coral assemblages exposed to more variable environments cannot be predicted using measures of long‐term performance. Moreover, our observed heterogeneity in the short‐term demographic characteristics of coral assemblages can help to explain why different assemblages display varying responses to periodic disturbances (Kim et al., 2023 ). This insight will benefit future predictions into the compositional reassembly of reef communities worldwide under future global change scenarios. Consistent with documented shifts in the species composition of coral assemblages exposed to increased abiotic variability, our findings here highlight a key mechanism underlying the differential susceptibilities of coral species to periodic disturbance. Correlative assessments of community change over time illustrate how weedy and stress‐tolerant coral taxa in the Caribbean and tropical Atlantic have, in the past, fared better in periodically disturbed environments relative to competitive coral taxa (Cramer et al., 2021 ). Complementing these assessments, the variation we observed in the short‐term demographic potential of competitive, stress tolerant, and weedy coral taxa implies that enhanced short‐term demographic characteristics offer weedy and stress‐tolerant corals a greater capacity for enduring within frequently disturbed environments, relative to competitive coral species. With competitive coral taxa often considered paramount for supporting the structural complexity of coral reefs worldwide (Graham & Nash, 2013 ), this indictment compounds concerns for the future functioning and viability of global coral reef ecosystems. Crucially, by adopting a novel framework for quantifying demographic resilience, our work here contributes to disentangling the biotic and environmental drivers underpinning the diversity of coral responses to ongoing global change.",
"introduction": "INTRODUCTION Anticipating the resilience of natural communities requires an in‐depth understanding for the determinants underpinning their constituent populations' responses to recurrent disturbances (Vázquez et al., 2017 ; Williams et al., 2008 ). Changes in environmental regimes provoke spatial shifts in the performance and distribution of populations, which upscale to the compositional reassembly of biological communities (Pecl et al., 2017 ; Totland & Nyléhn, 1998 ). Exposure to more variable environments is expected to indirectly augment community resilience (Boyd et al., 2016 ; Rivest et al., 2017 ). However, nuanced relationships between population characteristics and biophysical conditions ensure inconsistent responses to climate shifts (Parmesan & Yohe, 2003 ); with differential population sensitivities to habitat change having both accelerated and reversed expected poleward range shifts in response to climate warming (Chen et al., 2011 ). By linking the mechanisms driving heterospecific variation across population responses to environmental change, one can predict the resilience of whole communities to increased climatic variability (Dawson et al., 2011 ; Foden et al., 2013 ; Williams et al., 2008 ). Located at the interface between tropical and temperate ecoregions, subtropical coral assemblages provide an opportunity for exploring the determinants of population resilience (Beger et al., 2014 ; Burt et al., 2020 ; Camp et al., 2018 ). Recently, subtropical coral assemblages have undergone transformation, with tropical coral taxa undergoing poleward range expansions in response to shifting thermal regimes (Baird et al., 2012 ; Booth & Sears, 2018 ; Precht & Aronson, 2004 ; Tuckett et al., 2017 ; Yamano et al., 2011 ). At higher latitudes, however, coral assemblages are exposed to enhanced seasonality and cooler temperatures and, thus, experience greater abiotic variability relative to their tropical counterparts (Sommer et al., 2018 ). Subsequently, corals in subtropical regions offer insight into how differing coral populations utilize strategies to mediate their performance in response to environmental stochasticity across community‐ and regional‐scales. Exploring the performance of populations exposed to environmental stochasticity requires a consideration of their transient (i.e., short‐term) dynamics (Cant, Cook, et al., 2022 ; Ezard et al., 2010 ; Hastings, 2004 ; Hastings et al., 2018 ). Asymptotic (i.e., long‐term) population growth rate (λ), which describes temporal changes in population size at stationary equilibrium (Caswell, 2001 ), is the predominant metric used to quantify population performance (Caswell, 2001 ; Crone et al., 2011 ). However, stochastic conditions maintain natural populations within a transient state, preventing the emergence of stationary equilibria (Hastings, 2001 , 2004 ; Hastings et al., 2018 ). Within stochastic environments, recurrent disturbances impose short‐term changes upon the structure of populations that can elevate (i.e., demographic amplification) or diminish (i.e., demographic attenuation) their growth rates, resulting in population performance characteristics deviating from long‐term expectations (Ezard et al., 2010 ; Stott et al., 2011 ). Quantifying how transient population performance (henceforth short‐term potential) deviates from long‐term expectations therefore is crucial for predicting the success or failure of natural populations (Koons et al., 2005 ), an approach that remains neglected within coral research (Cant, Salguero‐Gómez, & Beger, 2022 ). In species rich communities, evaluating ecological dynamics requires a trait‐based approach to condense vast quantities of demographic detail (Chalmandrier et al., 2021 ). Given the diversity of coral assemblages, exploring patterns across the demographic characteristics of co‐occurring species presents a logistical challenge (Madin, Anderson, et al., 2016 ). Yet, this is a challenge that can be navigated by pooling individuals based on shared trait characteristics. Morphological, physiological, and phenological functional traits influence the fitness of individuals and thus determine the demographic characteristics of their populations (Violle et al., 2007 ), their responses to disturbances (Grime & Pierce, 2012 ), and subsequently the assembly of biological communities (Cadotte et al., 2011 ; Falster et al., 2017 ; McGill et al., 2006 ). Indeed, functional trait characteristics impact upon the demographic properties of coral populations (e.g., colony growth and reproduction [Álvarez‐Noriega et al., 2016 ; Madin et al., 2012 ]), mediating their ability to respond to local abiotic patterns (Sommer et al., 2014 ). Given such strong links between coral traits and demographic performance, the categorization of coral taxa into competitive, stress tolerant, generalist, and weedy life‐history assemblages (sensu Darling et al., 2012 ) can be used to evaluate broadscale patterns in coral community reassembly (Darling et al., 2013 , 2019 ; Zinke et al., 2018 ). Trait‐based assessments of coral community assembly also better inform upon the wider implications of ongoing community shifts than taxonomic‐based assessments, thereby aiding the management of coral reef ecosystems (Darling et al., 2019 ). Here, we investigate how the demographic characteristics of competitive, stress‐tolerant, and weedy coral assemblages map onto patterns of abiotic variability associated with the transition between tropical and subtropical environments. Exploiting gradients across the differing dimensions of thermal variability (monthly mean sea surface temperature [SST], monthly SST variance, and monthly SST frequency spectrum) as a key measure of abiotic variability (McIlroy et al., 2019 ; Toth et al., 2021 ), we used integral projection models (IPMs; Easterling et al., 2000 ) to quantify the association between abiotic variability and the short‐term potential and long‐term performance characteristics of tropical and subtropical coral assemblages across both the northern and southern hemispheres (Figure 1 ). Describing how state‐specific patterns in individual survival, development, and reproduction translate into population‐level characteristics, IPMs offer an approach for quantifying how abiotic environments influence population viability (Merow et al., 2014 ). Specifically, we anticipate that, compared to their tropical counterparts, subtropical coral assemblages will prioritize short‐term potential over long‐term performance, corresponding with the need for subtropical coral populations to endure periodically disturbed environments. Thus, we expect that characteristics associated with enhanced short‐term potential will align with more variable abiotic conditions, whereas measures of long‐term performance will be greater in more consistent environments; a pattern that will persist irrespective of functional strategy and geographic location. FIGURE 1 Using repeated annual surveys of tagged individual colonies, conducted between 2016 and 2019, we quantified the influence of environmental stochasticity on the long‐term performance and short‐term potential of tropical and subtropical coral populations in southern Japan and eastern Australia. (A) As climate shifts induce range expansions in many coral species worldwide, their populations are increasingly exposed to a gradient in thermal regimes, illustrated here by mean monthly sea surface temperatures ( x̄ \n sst ; degree Celsius; sst, sea surface temperature) recorded between 1950 and 2019 (Rayner et al., 2003 ). (B) Between 2016 and 2019, we documented the survival, growth, fragmentation, and recruitment patterns of 3171 tagged coral individuals within the tropical reef communities (▲) of Okinawa (Japan) and Heron Island (Australia), and within the subtropical communities (●) of Kochi (Japan) and the Solitary Islands Marine Park (Australia). (C) Using these data, we parameterized integral projection models (IPMs) describing the dynamics of tropical and subtropical assemblages of competitive, stress‐tolerant, and weedy coral taxa. Combining outputs obtained from these models with measures of the thermal regimes experienced by each population, we then explored the relationships between the long‐term performance and short‐term potential of coral populations, and their exposure to gradients in abiotic variability.",
"discussion": "DISCUSSION Transient buffering in variable environments Principally, contrasting patterns between long‐term performance and short‐term potential imply that long‐term performance does not predict the capacity for populations to endure repeated disturbances. Also, although enhanced short‐term potential may enable natural populations to persist within variable environments, it comes at a cost to their long‐term performance. Historically, variability in population growth rate was thought to diminish individual fitness (Pfister, 1998 ), thus hindering the persistence of populations (Lande, 1993 ). This understanding formed the basis of the demographic buffering hypothesis, whereby populations can minimize the influence of environmental stochasticity on their long‐term performance by limiting temporal variability in crucial vital rates (e.g., survival, development, and reproduction [Morris & Doak, 2004 ]). Thus, variable environments were assumed to select for populations with the ability to buffer key vital rates, thereby reducing temporal variation in performance characteristics (Hilde et al., 2020 ; Morris & Doak, 2004 ; Pfister, 1998 ). More recently, however, enhanced short‐term potential has been presented as an adaptive mechanism allowing populations to exploit more stochastic environments (McDonald et al., 2016 ). Ellis and Crone ( 2013 ) demonstrated how increased short‐term potential can buffer the effects of stochastic conditions on population growth rates, an effect that was increasingly evident in populations possessing lower λ estimates. Thus, it is not unexpected that coral assemblages established within variable environments would possess enhanced short‐term potential (Figure 2 ), but the energetic cost associated with this strategy would likely inhibit their long‐term performance characteristics. Our finding that short‐term potential is greatest in coral assemblages displaying reduced long‐term performance contrasts with previous work on mammals and plants showcasing a positive association between population growth rates and short‐term potential (e.g., Gamelon et al., 2014 ; Morris et al., 2008 ). Higher population growth rates are assumed of populations characterized by faster individual development and high fecundity (Oli, 2004 ), with these populations also expected to exhibit greater variability in size following disturbances (Gamelon et al., 2014 ). While each of our surveyed assemblages are in, or close to, a state of long‐term decline (λ < 1; Table 1 ), projected long‐term performance was highest in the tropics, where relative capacities for demographic amplification were lowest (Figure 3A ). Populations exhibiting longer generation times typically display reduced temporal variability in size due to higher investment in individual survival reducing the need to counteract disturbances (Morris et al., 2008 ); a pattern that we show to be evident in our examined coral assemblages (Figure 3C ). Interspecific variation in short‐term potential Variation in short‐term potential across our assemblages of differing coral taxa (Figure 3 ) suggests that exposure to abiotic variability alone does not assure resilience towards future climatic variability. Using data focused on a single taxon, Cant, Cook, et al. ( 2022 ) suggested that a capacity for short‐term increases in population growth observed in a subtropical Acropora spp. assemblage may underpin its viability in more variable high‐latitude environments. Here, we present evidence that this compensatory strategy is not just isolated to competitive coral taxa, but that stress‐tolerant and weedy coral taxa appear to possess a more pronounced capacity for demographic amplification at higher latitudes (Figure 3A ). Weedy corals typically exhibit smaller colony sizes, faster growth rates, and brooding reproductive strategies, producing larvae that settle quickly after release (Darling et al., 2012 ; Knowlton, 2001 ). Together, these strategies support faster population turnover, enabling weedy coral species to proliferate within highly disturbed environments (Adjeroud et al., 2018 ). Conversely, stress‐tolerant corals display slower growth rates, longer life expectancies, high fecundity, and broadcast spawning strategies (Darling et al., 2012 ; Klepac & Barshis, 2020 ). The larger, more robust, morphologies associated with stress‐tolerant coral taxa maximize energy storage, promoting their persistence within challenging environments (van Woesik et al., 2012 ). Longer lifespans and elevated fecundity allow stress‐tolerant corals to endure stochastic conditions by taking advantage of sporadic improvements in local conditions (Darling et al., 2012 ). Consequently, our findings support existing projections that weedy and stress‐tolerant coral taxa are likely to become increasingly prevalent throughout disturbed coral assemblages (Cant et al., 2021 ; Loya et al., 2001 ). However, these projections herald the future loss of the structural complexity considered essential to the functioning of reef ecosystems (Graham & Nash, 2013 ). We note that, while long‐term performance was typically highest in the tropics across each of our tropical‐subtropical assemblage pairings, the Japanese weedy coral assemblages show the opposite trend (Table 1 ). One possible explanation is that, in contrast to all other tropical‐subtropical assemblage pairings, the taxonomic composition of the Japanese weedy coral assemblages changed little between the tropics and subtropics (Appendix S1 : Table S3 ). Consequently, we acknowledge that our observed patterns in the long‐term performance characteristics of each tropical and subtropical assemblage may result from their differing species compositions. However, it can also be argued, therefore, that the species compositions of subtropical coral assemblages allow them to exhibit an enhanced short‐term potential, relative to their tropical counterparts. Crucially, this scenario points to a potential driving mechanism responsible for the variation in species composition typically observed between tropical and subtropical coral assemblages. Abiotic stress generated by the more variable conditions associated with higher latitude environments selects for traits conferring a competitive advantage, filtering the species composition of subtropical coral assemblages (Sommer et al., 2018 ). From this perspective, our findings here support the environmental filtering hypothesis in subtropical coral assemblages, and present evidence that the selected traits relate to enhanced short‐term potential; a capacity that is not selected for in tropical environments. Crucially, our findings here do not wholly reflect the current reality for many coral assemblages within regions of high abiotic variability, suggesting that the composition of coral assemblages is not solely mediated by the interplay between their short‐term dynamics and abiotic variability. Despite the reduced capacity for demographic amplification seen in subtropical competitive corals compared to subtropical weedy and stress‐tolerant populations, competitive coral taxa dominate many subtropical coral assemblages (Harriott et al., 1995 ; Nozawa et al., 2008 ; Sugihara et al., 2009 ). Utilizing fast growth strategies, colonies of competitive coral taxa are capable of rapidly colonizing available substrate, quickly outcompeting heterospecifics for both space and light (Darling et al., 2012 ). Whereas this competitive nature explains their dominance across contemporary subtropical communities, the sensitivity of many competitive coral taxa to environmental shifts means that these assemblages are often regarded as early successional, dominating only within optimal environments, and receding as reef ecosystems approach climax states (Ohba et al., 2008 ; Wilson et al., 2019 ). Within subtropical environments, coral community composition is mediated by environmental pressures and dispersal barriers that filter the occurrence of species according to their trait characteristics (Mizerek et al., 2021 ; Sommer et al., 2014 ). As a result, subtropical coral assemblages typically consist of a subset of tropical species found on tropical coral reefs (Sommer et al., 2017 ), as well as subtropical specialists and endemics. The dominance of competitive coral taxa within subtropical coral assemblages, despite their reduced short‐term potential relative to other coral taxa, may therefore imply that competitive interactions profoundly influence the performance of coral populations (Brito‐Millán et al., 2019 ; Idjadi & Karlson, 2007 ). Certainly, further investigation into the influence of competitive interactions upon the short‐term dynamics of coral populations is needed to disentangle how coexistence between coral populations facilitates their persistence within variable environments. Conclusions A limited understanding for the abiotic determinants driving the dynamics of coral assemblages inhibits our capacity to predict their future performance and, therefore, manage global coral community reassembly (Edmunds, 2020 ; Edmunds et al., 2014 ; Edmunds & Riegl, 2020 ). Here, we demonstrate how coral populations can adopt demographic strategies associated with enhanced short‐term potential to improve their viability when exposed to greater abiotic variability. However, strategies of enhanced short‐term potential come at a cost to the long‐term demographic performance characteristics of coral populations. Despite presenting a framework for quantifying population resilience (Capdevila et al., 2020 ), the short‐term demographic characteristics of coral assemblages remain largely overlooked (Cant, Cook, et al., 2022 ). Yet, our findings emphasize that the winners and losers in coral assemblages exposed to more variable environments cannot be predicted using measures of long‐term performance. Moreover, our observed heterogeneity in the short‐term demographic characteristics of coral assemblages can help to explain why different assemblages display varying responses to periodic disturbances (Kim et al., 2023 ). This insight will benefit future predictions into the compositional reassembly of reef communities worldwide under future global change scenarios. Consistent with documented shifts in the species composition of coral assemblages exposed to increased abiotic variability, our findings here highlight a key mechanism underlying the differential susceptibilities of coral species to periodic disturbance. Correlative assessments of community change over time illustrate how weedy and stress‐tolerant coral taxa in the Caribbean and tropical Atlantic have, in the past, fared better in periodically disturbed environments relative to competitive coral taxa (Cramer et al., 2021 ). Complementing these assessments, the variation we observed in the short‐term demographic potential of competitive, stress tolerant, and weedy coral taxa implies that enhanced short‐term demographic characteristics offer weedy and stress‐tolerant corals a greater capacity for enduring within frequently disturbed environments, relative to competitive coral species. With competitive coral taxa often considered paramount for supporting the structural complexity of coral reefs worldwide (Graham & Nash, 2013 ), this indictment compounds concerns for the future functioning and viability of global coral reef ecosystems. Crucially, by adopting a novel framework for quantifying demographic resilience, our work here contributes to disentangling the biotic and environmental drivers underpinning the diversity of coral responses to ongoing global change."
} | 6,015 |
34098512 | null | s2 | 3,688 | {
"abstract": "Microbial communities and their functions are shaped by complex networks of interactions among microbes and with their environment. While the critical roles microbial communities play in numerous environments have become increasingly appreciated, we have a very limited understanding of their interactions and how these interactions combine to generate community-level behaviors. This knowledge gap hinders our ability to predict community responses to perturbations and to design interventions that manipulate these communities to our benefit. Dynamic models are promising tools to address these questions. We review existing modeling techniques to construct dynamic models of microbial communities at different scales and suggest ways to leverage multiple types of models and data to facilitate our understanding and engineering of microbial communities."
} | 214 |
37205326 | PMC10187175 | pmc | 3,689 | {
"abstract": "Microbial communities play a critical role in ecological processes, and their diversity is key to their functioning. However, little is known about if communities can regenerate ecological diversity following species removal or extinction, and how the rediversified communities would compare to the original ones. Here we show that simple two-ecotype communities from the E. coli Long Term Evolution Experiment (LTEE) consistently rediversified into two ecotypes following the isolation of one of the ecotypes, coexisting via negative frequency-dependent selection. Communities separated by more than 30,000 generations of evolutionary time rediversify in similar ways. The rediversified ecotype appears to share a number of growth traits with the ecotype it replaces. However, the rediversified community is also different compared to the original community in ways relevant to the mechanism of ecotype coexistence, for example in stationary phase response and survival. We found substantial variation in the transcriptional states between the two original ecotypes, whereas the differences within the rediversified community were comparatively smaller, but with unique patterns of differential expression. Our results suggest that evolution may leave room for alternative diversification processes even in a maximally reduced community of only two strains. We hypothesize that the presence of alternative evolutionary pathways may be even more pronounced in communities of many species, highlighting an important role for perturbations, such as species removal, in evolving ecological communities.",
"introduction": "Introduction Ecological diversification is the process by which a population or community of organisms evolves to occupy different ecological niches or habitats in a given ecosystem 1 . This diversification can occur in various ways, such as the development of different physical or behavioral adaptations that allow individuals to exploit different resources or tolerate different environmental conditions 2 – 4 . The potential for ecological diversification within a community typically hinges on factors such as environmental conditions, existing biodiversity, and the ecological interactions among resident species 4 – 7 . Microbial communities have proven particularly useful for studying the interplay of evolutionary and ecological processes underlying diversification due to their manageable time scales for reproduction and evolution 8 – 16 . Diversification may be influenced by the availability of unoccupied niches, often referred to as “ecological opportunities” 5 , 17 , which can become scarce when most niches are already occupied due to high diversity levels. Alternatively, the resident community can create new niches, enabling the establishment of novel species, suggesting that “diversity begets diversity” 18 . Cross-feeding exemplifies this latter scenario, where species release metabolites that can foster the emergence of new species by creating exploitable niches 13 , 15 , 19 – 21 . Ecological interactions within microbial communities have been shown to have negative 11 , 12 , positive 13 , 15 , and even mixed effects 22 on a community’s ability to diversify. However, the stable coexistence of a novel species and its ancestor is not guaranteed and may depend on various community properties, such as metabolic trade-offs 21 , 23 . In experimental settings, ecological differentiation of a diversified ecotype is often indicated when an ecotype’s fitness inversely correlates with its frequency, i.e. displaying negative frequency-dependent fitness effects. Stable coexistence between the diversified ecotype and its ancestor is implied if it can invade at small frequencies and cannot invade at large frequencies. Even when species can stably coexist, it does not guarantee that they will coexist indefinitely or at all locations. Species can migrate to new territories, potentially without other community members, or some species within the community may spontaneously go extinct. In either case, the community becomes perturbed, losing one or more members and potentially leaving ecological niches unfilled. Theoretical models suggest that perturbed communities may respond with a combination of ecological and evolutionary changes 24 – 27 . These evolutionary changes may include both directional and diversifying selection 27 , with newly evolved variants either replacing existing community members or coexisting alongside them. However, it remains unclear which communities have the potential to rediversify. Recently diversified communities may be more likely to rediversify following ecotype isolation, as they have recently arisen from an ancestor that underwent ecological diversification. However, as a community coevolves, the potential for rediversification might diminish, but this may not necessarily always the case. When rediversification does occur after species removal, there are two possible scenarios: (i) the community eventually rediversifies and returns to a state similar to the original community before the disturbance, or (ii) the perturbed community rediversifies and forms a community that is qualitatively different from the original one. Here, we investigate the aforementioned questions surrounding rediversification using a minimal microbial model community of only two, naturally diversified E. coli strains. Specifically, we employ two strains derived from the E. coli Long-Term Evolution Experiment (LTEE), which was started by Dr. Richard Lenski and has been running for over 30 years or more than 70,000 generations 28 . An initially isogenic strain of E. coli was split into 12 replicate populations and propagated through daily dilutions in glucose minimal media (DM25). At the outset of the LTEE around 6.5k generations, it was found that one lineage, ara-2, spontaneously diversified into two lineages– S and L –that coexist via negative frequency dependence 29 . The ecotypes were named for the sizes of their colonies on certain agar plates, either small ( S ) or large ( L ). The S and L lineages inhabit distinct temporal and metabolic niches in the LTEE environment. During exponential phase, L grows more quickly on glucose, while S specializes in stationary phase survival and utilizes acetate, a byproduct of overflow metabolism 30 , 31 . Since their diversification, the lineages have persisted and evolved over time, exhibiting genetic, transcriptional, and metabolic divergence 29 – 36 . The LTEE-derived communities are ideal for our plan to investigate the possibility and potential patterns of rediversification over evolutionary time. We can revive the S - L community at 6.5k generations to probe rediversification right after emergence of the community, and compare with rediversification at later stages of the evolution experiment. We found that when we isolated the S ecotype under certain conditions, it would spontaneously rediversify, giving rise to a new big colony ecotype S B , even if we used S clones separated by more than 30,000 generations of evolutionary time. The new ecotype, S B , displays hallmarks of ecological differentiation, including negative frequency-dependent fitness effects when with its ancestral S clone. We dissected the new, rediversified community, and found that while S B shares a number of traits with both L and S , it also behaves in entirely new ways. Our findings suggest that even in a maximally reduced community of only two strains, evolution may leave room for alternative diversification processes, suggesting a hidden adaptive potential only revealed by ecotype removal. This raises the possibility that perturbations, such as species removal, could play an important role in evolving ecological communities by creating opportunities for alternative evolutionary pathways.",
"discussion": "Discussion Our study explores the capacity of an evolved microbial community to quickly regenerate ecological diversity following the removal of a species. Our results suggest that even in the case of a community composed of only two strains in a minimal environment, evolution can leave room for alternative diversification processes. The rediversified ecotype, S B , demonstrates the robustness of microbial communities to perturbations by sharing several growth traits with the ecotype it replaces, L . For instance, both S B and L exhibit slower initial growth or longer lag times compared to S across all LTEE timepoints, which may be involved in a trade-off allowing for higher exponential growth rates, as observed in other systems 44 . However, differences between the rediversified and original communities suggest that the mechanism of ecotype coexistence has shifted. Notably, we observe variations in stationary phase responses and survival, as well as distinct patterns of gene expression. Together, these findings indicate that ecological rediversification in the S - L system may be influenced by a combination of constraints and opportunities. While adaptation may lead some traits to evolve nearly deterministically due to strong ecological or physiological constraints, other trait values may experience more freedom. The interplay between contingency and determinism mirrors patterns observed in various other evolving systems, including the LTEE 45 – 47 . Dissecting why some traits are more evolutionarily constrained during diversification compared to others could be a fruitful avenue for future investigation. We attempted to determine a potential genetic origin of the S B phenotype. However, we did not find any consistent mutations shared between the independent S B clones, relative to their S ancestor. Thus, the S B phenotype likely either has a large target size, such that many different mutations can cause the same phenotype 48 , 49 , or it is caused by a non-genetic heritable change. Despite the fact that we did not find any shared mutations, the transcriptional changes of two S B clones were targeted to the same handful of pathways, predominantly related to amino acid metabolism. This points to parallelism at least on the transcriptional level, if not on the genetic level. Additionally, while the differentially expressed pathways in S B and L relative to S were generally different, we saw decreased expression of ribosomal proteins in both ecotypes. The fraction of the proteome devoted to ribosomes is known to control many growth traits in bacteria 50 , 51 , so the similar changes in L and S B may help to explain the handful of observed similarities in growth traits. One might expect that ribosome expression should be lower in S , due to its slower exponential growth rate 52 , 53 ; so the fact that this is not the case may suggest that S B and L are both allocating their proteome not just to optimize exponential growth rate, but also other growth traits as well. While we saw that S could rediversify following isolation, we did not see any obvious ecological or phenotypic diversification when L was isolated. There may be several reasons for this. (i) S may have some amount of physiological/ genetic/ metabolic plasticity that allows it to diversify that L lacks. (ii) Diversification of L may happen slowly or rarely, or more quickly only under certain environmental conditions. (iii) Perhaps L can rapidly diversify, but cryptically, where no phenotypic changes are obvious without more extensive phenotyping. It is certainly the case that we would not have found S B without the obvious changes in colony size. It could be that rediversification is much more common than currently appreciated, but simply not detected. Sequencing technologies, including metagenomic 35 and DNA barcoding-based methods 54 , could help to better reveal the full extent of rediversification across microbial communities. In fact, through metagenomic sequencing, we now know that ecological diversification is much more common in the LTEE than previously thought 35 . Our study has implications for our understanding of the ecological consequences of species removal or extinction. The ability of microbial communities to rediversify following such perturbations may represent a crucial mechanism by which communities can maintain their functioning and stability over time. While one might think that evolution would be too slow compared to ecological processes, we see here that evolution is crucial for community recovery following a perturbation. The presence of alternative evolutionary pathways, even in a maximally reduced community of only two strains, suggests that such mechanisms may be even more pronounced in communities with greater species richness. In conclusion, our study provides insights into the capacity of microbial communities to regenerate ecological diversity and adapt to environmental perturbations. Further research into the mechanisms that govern these dynamics will be crucial for understanding the functioning and stability of microbial communities, as well as their response to environmental change."
} | 3,265 |
34691011 | PMC8527027 | pmc | 3,690 | {
"conclusion": "Conclusions More detailed insights into microbial community dynamics in polluted soils are a major achievement in recent research activities, which were benefitting from novel methodological advancements, such as high-throughput sequencing technologies. Concerning the effect of microbes on pollutant degradation, immobilization or plant uptake, the effectiveness was confirmed. Further improvements by e.g., selecting more efficient microbes should be the focus of future investigations. More importantly, the use of microbial consortia presents a new approach for efficient remediation of complex mixtures of contaminants and a viable option to share the undesired metabolic burden among various microbial strains. However, many uncertain factors, such as the little elucidated mechanism of cell-to-cell communications, the change in the structure of the microbial population, and the compromise of culture conditions, still exist when consortia of microorganisms are used. To solve these problems, in-depth studies focusing on the interaction mechanisms of natural microbial consortia, the introduction of functional support materials and the development of analytical and predictable computer models should be carried out in the future."
} | 310 |
29138413 | PMC5686229 | pmc | 3,691 | {
"abstract": "Synchronization occurs in many natural and technological systems, from cardiac pacemaker cells to coupled lasers. In the synchronized state, the individual cells or lasers coordinate the timing of their oscillations, but they do not move through space. A complementary form of self-organization occurs among swarming insects, flocking birds, or schooling fish; now the individuals move through space, but without conspicuously altering their internal states. Here we explore systems in which both synchronization and swarming occur together. Specifically, we consider oscillators whose phase dynamics and spatial dynamics are coupled. We call them swarmalators, to highlight their dual character. A case study of a generalized Kuramoto model predicts five collective states as possible long-term modes of organization. These states may be observable in groups of sperm, Japanese tree frogs, colloidal suspensions of magnetic particles, and other biological and physical systems in which self-assembly and synchronization interact.",
"introduction": "Introduction This year marks the fiftieth anniversary of a breakthrough in the study of synchronization. In 1967, Winfree proposed a coupled oscillator model for the circadian rhythms that underlie daily cycles of activity in virtually all plants and animals 1 . He discovered that above a critical coupling strength, synchronization breaks out spontaneously, in a manner reminiscent of a phase transition. Then Kuramoto simplified Winfree’s model and solved it exactly 2 , leading to an explosion of interest in the dynamics of coupled oscillators 3 – 5 . Kuramoto’s model in turn has been generalized to other large systems of biological oscillators, such as chorusing frogs 6 , firing neurons 7 – 11 , and even human concert audiences clapping in unison 12 . The analyses often borrow techniques from statistical physics, such as mean-field approximations, renormalization group analyses 13 , 14 , and finite-size scaling 15 , 16 . There has also been traffic in the other direction, from biology back to physics. For example, insights from biological synchronization have shed light on neutrino oscillations 17 , phase locking in Josephson junction arrays 18 , the dynamics of power grids 19 , 20 , and the unexpected wobbling of London’s Millennium Bridge on opening day 21 . A similarly fruitful interplay between physics and biology has occurred in the study of the coordinated movement of groups of animals. Fish schools, bird flocks, and insect swarms 22 – 26 have been illuminated by maximum entropy methods 27 , agent-based simulations 28 , and analytically tractable models based on self-propelled particles 29 , and continuum limits 30 – 33 . Studies of swarming and synchronization have much in common. Both involve large, self-organizing groups of individuals interacting according to simple rules. Both lie at the intersection of nonlinear dynamics and statistical physics. Nevertheless the two fields have, by and large, remained disconnected. Studies of swarms focus on how animals move, while neglecting the dynamics of their internal states. Studies of synchronization do the opposite: they focus on oscillators’ internal dynamics, not on their motion. In the past decade, however, a few studies of “mobile oscillators,” motivated by applications in robotics and developmental biology, have brought the two fields into contact 34 – 38 . Even so, the assumption has been that the oscillators’ locations affect their phase dynamics, but not conversely. Their motion has been modeled as a random walk or as externally determined, without feedback from the oscillators’ phases. We suspect that somewhere in nature and technology there must be mobile oscillators whose phases affect how they move. For instance, many species of frogs, crickets, and katydids call periodically, and synchronize in vast choruses 6 , 39 – 41 . The natural question is whether they tend to hop toward or away from others depending on the relative phases of their calling rhythms, and if so, what spatiotemporal patterns are produced. A clue comes from the physics of magnetic colloids 42 – 44 and microfluidic mixtures of active spinners 45 , 46 , both of which show rich collective behavior. In these systems, the particles or spinners attract or repel one another, depending on their orientations. Given that orientation is formally analogous to the phase of an oscillation (both being circular variables), a similarly rich phenomenology is expected for mobile oscillators whose phases affect their motion. We call these hypothetical systems ‘swarmalators’ because they generalize swarms and oscillators. One possible instance of a swarmalator system is a population of myxobacteria, modeled in 2001 by Igoshin and colleagues 47 . The movements of these bacteria in space are thought to be influenced by an internal, biochemical degree of freedom, which appears to vary cyclically. Igoshin et al. 47 modeled it as a phase oscillator. Experimental evidence suggests that the evolution of this phase is influenced by the spatial density of neighboring cells; thus there appears to be a bidirectional coupling between spatial and phase dynamics, as required of swarmalators. Tanaka and colleagues also made an early contribution to the modeling of swarmalators 48 , 49 . They analyzed a broad class of models in the hope of finding phenomena which were not system-specific. They considered chemotactic oscillators, whose movements in space are mediated by the diffusion of a background chemical. The oscillators’ consumption of this chemical depends on their internal states, thereby completing the bidirectional space-phase coupling. Tanaka et al. 48 , 49 began with a general model with these ingredients, from which they derived a simpler model by means of center manifold and phase-reduction methods. Here we take a bottom-up approach. We propose a simple model of a swarmalator system which lets us study some of its collective states analytically. We hope our work will draw attention to this class of problems, and stimulate the discovery and characterization of natural and technological systems of swarmalators.",
"discussion": "Discussion We have examined the collective dynamics of swarmalators. These are mobile particles or agents with both phase and spatial degrees of freedom, which lets them sync and swarm. Furthermore, their phase and spatial dynamics are coupled. By studying simple models, we found this coupling leads to rich spatiotemporal patterns which we explored analytically and numerically. These patterns were robust to modifications to the model, namely motion in one, two, and three spatial dimensions, distributed natural frequencies, noisy interactions, and alignment dynamics. We thus believe they could be realized in nature or technology. A pertinent future goal, then, is to investigate the behavior of real-world systems of swarmalators. As mentioned in the introduction, colloidal suspensions of magnetic particles 42 – 44 or active spinners 45 , 46 are promising candidates. For example, structures equivalent to the static phase wave state have been experimentally realized by Snezhko and Aranson, when studying the behavior of ferromagnetic colloids at liquid-liquid interfaces 43 (the particles comprising the colloids can be considered swarmalators if we interpret the angle subtended by their magnetic dipole vectors as their phase). As shown in Fig. 4 of ref. 43 , the colloids can form asters. These are structures composed of radial chains of magnetically ordered particles, which “decorate slopes of a self-induced circular standing wave” 43 , analogous to the annular pattern of correlated phases and positions of the static phase wave shown in Fig. 2c . Perhaps colloidal equivalents of the splintered and active phase wave states could also be realized. Aside from being theoretically interesting, the ability to engineer these states could have practical application. For instance, Snezhko and Aranson also show that asters can be manipulated to capture and transport target particles. The non-stationary behavior of the splintered and active wave states might also have locomotive utility. Tentative evidence for this claim is provided by populations of cilia, whose collective metachronal waves, similar to the motion of swarmalators in the aforementioned states, are known to facilitate biological transport 57 – 59 . Other plausible systems of real-world swarmalators are biological microswimmers, self-propelled micro-organisms capable of collective behavior 60 . One such contender is populations of spermatoza, which exhibit rich swarming behavior such as trains 61 , 62 and vortex arrays 52 , the latter of which is reminiscent of the active phase wave state, as mentioned in the Results section. The phase variable for each sperm is associated with the rhythmic beating of the sperm’s tail, which can synchronize with that of a neighboring sperm 63 , 64 . It has been theorized that this can induce spatial attraction 65 , leading to clusters of synchronized sperm, consistent with experimentally observed behavior 66 . There are also theoretical avenues to explore within our proposed model of swarmalators. For instance the curious stability properties of the static async state deserve further study. Another route would be to include more realism by including heterogeneity in the coupling parameters K , J , or by choosing more complicated interaction functions I \n att , I \n rep , G , H \n att . For example we chose H \n att ( θ ) = sin( θ ) to mimic the Kuramoto model, but as we saw, it led to just the trivial static sync state when K > 0. Perhaps choosing the more realistic Winfree model for the phase dynamics, which gives rise to richer collective behavior, would lead to more interesting swarmalator phenomena in this parameter regime. Perhaps the most important direction for future work is to more fully explore the interplay among aggregation, alignment, and synchronization—or put another way, to explore the collective behavior of particles with a position x , an orientation β , and an internal phase θ . The primary goal of our work is to draw attention to this class of problems, which we believe define a wide landscape of emergent behavior. In this work, we have started to map out this landscape by studying a simple model that contains a subset of these three effects, namely aggregation and synchronization. Others have considered the remaining subsets. For example, Leon and Liverpool have explored the interaction between alignment and synchronization 67 . They introduced a new class of soft active fluids whose units have an orientation and phase. They found this mixture can either enhance or inhibit the transition from disordered states to states with polar order. The latter states are roughly similar to the aligned static async states. They also found transitions from disordered states to states with phase order, which are analogous to unaligned static sync states. Yet counterparts of the static, splintered, and active phase waves were not reported. The final combination, aggregation and alignment, is perhaps the most well studied, in both new models and old. For instance, Starnini et al. 68 recently introduced a model of mobile particles capable of aggregating and aligning their opinions, and found the emergence of echo chambers. Even in the classic Vicsek model and its numerous extensions, new phenomena are still being found. For instance, Kruk et al. 69 found that delayed alignment in the Vicsek model produces self-propelled chimeras; perhaps delayed phase interactions could lead to similar states for swarmalators. Liebchen and Levis 70 considered units with an intrinsic rotation, and found ‘phase separated droplets’: clusters of rotation-synchronized particles surrounded by a sea of incoherent particles (multiple droplets are also possible). These droplets are similar to our static sync states, but they differ in the crucial respect that the entire population is synchronized in our static sync state. Here too, the counterparts of our static, splintered, and active phase waves were not seen. Thus, to the best of our knowledge, no other models display states analogous to the splintered phase waves and active phase waves found in our swarmalator model. In that sense, those two states are unprecedented."
} | 3,089 |
27435461 | PMC4958250 | pmc | 3,694 | {
"abstract": "ABSTRACT A fundamental question in microbial physiology concerns why organisms prefer certain nutrients to others. For example, among different nitrogen sources, ammonium is the preferred nitrogen source, supporting fast growth, whereas alternative nitrogen sources, such as certain amino acids, are considered to be poor nitrogen sources, supporting much slower exponential growth. However, the physiological/regulatory logic behind such nitrogen dietary choices remains elusive. In this study, by engineering Escherichia coli , we switched the dietary preferences toward amino acids, with growth rates equivalent to that of the wild-type strain grown on ammonia. However, when the engineered strain was cultured together with wild-type E. coli , this growth advantage was diminished as a consequence of ammonium leakage from the transport-and-catabolism (TC)-enhanced (TCE) cells, which are preferentially utilized by wild-type bacteria. Our results reveal that the nitrogen regulatory (Ntr) system fine tunes the expression of amino acid transport and catabolism components to match the flux through the ammonia assimilation pathway such that essential nutrients are retained, but, as a consequence, the fast growth rate on amino acids is sacrificed.",
"introduction": "INTRODUCTION Previous physiological studies demonstrated that glutamine serves as an internal sensor of external nitrogen availability in enteric bacteria ( 1 , 2 ). Under nitrogen-limiting conditions, the bacterial nitrogen regulatory (Ntr) system responds to the decrease in the internal concentration of glutamine and activates the expression of Ntr-regulated genes/operons ( 3 – 5 ), whose products facilitate the efforts of bacteria to scavenge nitrogenous compounds available in the environment ( 6 ). The Ntr system of Escherichia coli comprises a hierarchical regulatory network, including the bifunctional uridylyltransferase/uridylyl-removing enzyme (UTase/UR) GlnD, the two PII signal transduction proteins GlnB and GlnK, and the NtrBC two-component regulatory system ( 7 – 11 ). Under nitrogen-limiting conditions, GlnD covalently modifies GlnB, enabling NtrB to phosphorylate NtrC, which then activates Ntr-dependent genes, including the expression of glnK (which encodes GlnK). Although GlnD also covalently modifies GlnK under nitrogen-deficient conditions, there is evidence that GlnK feedback inhibits some Ntr-dependent promoters during nitrogen starvation ( 5 ), which is likely to result from the incomplete uridylylation of GlnK. This interaction of the non-covalently modified form of GlnK with NtrB represses the kinase activity and activates the phosphatase activity of NtrB ( 12 ). The consequent dephosphorylation of NtrC prevents activation of Ntr-dependent genes. However, the physiological function of GlnK during nitrogen-limiting exponential growth is unclear. Growth rate maximization is considered to be an important factor in the survival and fitness of unicellular organisms. Among various nitrogen sources, bacteria prefer ammonia, which supports a fast growth rate in E. coli compared with alternative nitrogen sources such as amino acids ( 13 ). It has been confirmed that the Ntr system maintains a fast growth rate across a wide range of ammonium concentrations, primarily by regulating both the expression and activity of the glutamine synthesis enzyme GS (glutamine synthetase) and also of the ammonium transporter AmtB ( 5 , 14 – 16 ). Alternative nitrogen sources, such as amino acids, support much slower growth rates and are considered to be “poor” nitrogen sources ( 1 ), although genes for the utilization of certain amino acids also belong to the Ntr regulon. This nitrogen preference is likely to reflect physico-chemical constraints on the transport and catabolism of amino acids. However, our studies suggest that this nitrogen dietary preference is deliberately maintained by regulatory constraints, enforced particularly by the Ntr system. We report a direct correlation between nitrogen influx ( J N ) and growth rate in E. coli , regardless of the nitrogen source used, and demonstrate that the slow growth rate on specific amino acids is limited by constraints on transport and catabolism, as anticipated ( 17 ). In contrast, when two transport-and-catabolism (TC)-enhanced (TCE) strains were constructed, they exhibited fast growth on cognate amino acids, comparable to the growth rates observed with ammonium as the sole nitrogen source. Remarkably, these engineered strains prefer to utilize the cognate amino acid rather than ammonia as the sole nitrogen source. However, this switch in nitrogen dietary preferences results in ammonium leakage from the TCE strains, benefiting the growth of competitors. Our quantitative analysis demonstrated that in wild-type E. coli , nutrient leakage is efficiently prevented by two negative-feedback loops in the Ntr system that downregulate the expression of the TC genes, which, in turn, determines the slow growth rate on amino acids.",
"discussion": "DISCUSSION The overall strategy for Ntr-mediated regulation of nitrogen assimilation is apparently to optimize the growth rate in relation to internal ammonium availability as determined by the glutamine concentration ( 1 ). Our study results highlight the importance of feedback loops mediated by both GlnK and the product (glutamine) in restricting the assimilation of alternative nitrogen sources. Although this has consequences for growth rate and optimal production of biomass, it fine tunes the level of nitrogen metabolites to prevent nutrient leakage. Therefore, we propose that the principle behind the preference for the nitrogen diet is that of trading off the benefits of fast growth on alternative nitrogen sources with the fitness penalty associated with excreting nitrogen that becomes available to competitors. This tradeoff is achieved through the highly complex Ntr regulatory circuitry, which prioritizes nutrient retention over fast growth, resulting in metabolic slowdown on alternative nitrogen sources. It was reported recently that when an alternative carbon source, such as glycerol, is used as the sole carbon source, enzymatic constraints on carbon uptake prevent acetate leakage from wild-type E. coli and, as a consequence, result in sacrifice of the enhanced growth rate ( 28 ). Therefore, nutrient containment could be a common strategy for both carbon and nitrogen metabolism in E. coli . From the application perspective, it should be feasible to metabolically engineer efficient utilization of other amino acids as nitrogen sources, using synthetic biology approaches similar to those described here. Our studies are thus likely to underpin more-efficient biorefinery processes that utilize waste amino acids as raw material to achieve sustainable biofuel or biochemical production and recover fertilizer nitrogen ( 29 , 30 )."
} | 1,715 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.