pmid stringlengths 8 8 | pmcid stringlengths 8 11 ⌀ | source stringclasses 2
values | rank int64 1 9.78k | sections unknown | tokens int64 3 46.7k |
|---|---|---|---|---|---|
37649327 | PMC10832561 | pmc | 9,222 | {
"abstract": "Abstract \n Clostridium luticellarii is a recently discovered acetogen that is uniquely capable of producing butyric and isobutyric acid from various substrates (e.g. methanol), but it is unclear which factors influence its (iso)butyric acid production from H 2 and CO 2 . We aimed to investigate the autotrophic metabolism of C. luticellarii by identifying the necessary growth conditions and examining the effects of pH and metabolite levels on product titers and selectivity. Results show that autotrophic growth of C. luticellarii requires the addition of complex nutrient sources and the absence of shaking conditions. Further experiments combined with thermodynamic calculations identified pH as a key parameter governing the direction of metabolic fluxes. At circumneutral pH (~6.5), acetic acid is the sole metabolic end product but C. luticellarii possesses the unique ability to co‐oxidize organic acids such as valeric acid under high H 2 partial pressures (>1 bar). Conversely, mildly acidic pH (≤5.5) stimulates the production of butyric and isobutyric acid while partly halting the oxidation of organic acids. Additionally, elevated acetic acid concentrations stimulated butyric and isobutyric acid production up to a combined selectivity of 53 ± 3%. Finally, our results suggest that isobutyric acid is produced by a reversible isomerization of butyric acid, but valeric and caproic acid are not isomerized. These combined insights can inform future efforts to optimize and scale‐up the production of valuable chemicals from CO 2 using C. luticellarii .",
"conclusion": "CONCLUSIONS This study investigated the autotrophic growth of C. luticellarii by identifying the necessary growth conditions and examining the effects of metabolite levels and pH on butyric and isobutyric acid titers and selectivity. Our experiments concluded that C. luticellarii requires the addition of complex nutrient sources and the absence of shaking conditions for autotrophic growth. To steer product selectivity, pH was identified as a key parameter governing the direction of metabolic fluxes. At circumneutral pH (>5.5), acetic acid is the sole metabolic end product and, additionally, C. luticellarii shows the particular ability to co‐oxidize organic acids such as valeric acid under high H 2 partial pressures (>1 bar). Conversely, mildly acidic pH (<5.5) stimulates the production of butyric and isobutyric acid. Additionally, elevated acetic acid concentrations stimulated butyric and isobutyric acid production up to a maximum combined selectivity of 53 ± 3%. Finally, our results suggest that isobutyric acid is produced by a reversible isomerization of butyric acid, but valeric and caproic acid are not isomerized by C. luticellarii . While this study positions C. luticellarii and its uniquely flexible metabolism producing butyric and isobutyric acid from H 2 and CO 2 as a promising candidate for industrial biotechnological applications, further research should focus on improving yields and production rates on reactor‐scale. Overall, the combined insights reported here can inform the future efforts to optimize and scale‐up the production of valuable chemicals from CO 2 using C. luticellarii .",
"discussion": "DISCUSSION \n Clostridium luticellarii requires complex nutrients and a low‐shear environment to grow on H 2 and CO 2 \n While C. luticellarii was capable of growing autotrophically, it required tryptone or yeast extract and static conditions to grow reproducibly across sub cultures. The necessity of complex nutrients (e.g. yeast extract and/or tryptone) to enable autotrophic growth of acetogens has been reported earlier. For example, Martin et al. ( 2016 ) tested the dependence of six different autotrophic strains of acetogens on yeast extract and only managed to grow three of them without a complex nutrient source. This can be connected to the fact that the metabolism of acetogens often runs close to the thermodynamic limit of life due to the low availability of energy under autotrophic conditions, potentially also impacting their capacity to synthesize all cell constituents de novo (Schuchmann & Müller, 2014 ). Therefore, many studies focusing on medium development for acetogens and studies characterizing the autotrophic growth of acetogens have used yeast extract as a complex nutrient source providing vitamins, trace elements and amino acids (Arslan et al., 2019 ; Chang et al., 2007 ; Groher & Weuster‐Botz, 2016a , 2016b ; Litty & Müller, 2021 ). However, regardless of the source of yeast extract, at least 7%–13% of it is composed of readily degradable carbon sources such as carbohydrates (Tomé, 2021 ), which can be converted to organic acids. This may distort the results of some studies where, comparatively, the electron equivalents potentially stemming from the carbohydrates in yeast extract are higher than the electron equivalents coming from the actual substrates under study. We advocate the use of moderate concentrations of tryptone over yeast extract whenever possible in studies characterizing production capacities of microorganisms since tryptone is a lean source of amino acids that does not usually support growth in absence of other substrates as shown in \n Clostridium luticellarii requires strict conditions for autotrophic growth . Additionally, where the use of yeast extract is inevitable, a close eye should be kept on whether electron and carbon balances close (i.e. equal amounts of carbon and electrons consumed as produced) to minimize the interference of yeast extract on the studied metabolism. Here, we chose to supplement growth media with 0.5 g L −1 tryptone instead of 1 g L −1 yeast extract, and managed to reproducibly grow C. luticellarii without any addition of carbohydrates (Figure 1 ). Results also indicated that C. luticellarii consumes gas much slower and to a lesser extent under shaking conditions than when statically incubated. This result is counterintuitive, since shaking increases the gas–liquid mass transfer rate and thus the substrate availability for the bacteria but it grows better under passive diffusion. One potential hypothesis may be that C. luticellarii is sensitive to the shear stress induced by shaking conditions. Although this is not uncommon in anaerobic bacteria, the underlying mechanism remains unclear (Jonczyk et al., 2013 ). Scaling up gas fermentation with shear sensitive C. luticellarii may prove a challenge since classic approaches such as CSTRs may introduce too much shear due to mixing, and bubble column or gas lift reactors may induce too much shear due to bubble flow. Attractive alternatives are low‐shear reactor designs such as the bubble‐less hollow fibre membrane biofilm reactors or the exclusive use of CO 2 ‐derived liquid molecules (e.g. methanol) (Elisiário et al., 2021 ; Petrognani et al., 2020 ). Autotrophic isobutyric and butyric acid production is stimulated by elevated acetic acid concentrations Autotrophic growth of C. luticellarii at pH 6.5 was mainly characterized by acetic acid production, which is the dominant product in the majority of acetogens (Bengelsdorf et al., 2018 ), accompanied by traces of butyric and isobutyric acid. These results are similar to those reported for other acetogens capable of producing traces of butyric acid when grown on H 2 and CO 2 , such as B. methylotrophicum (Lynd & Zeikus, 1983 ), E. limosum KIST612 (Litty & Müller, 2021 ), B. hydrogenotrophica DSM 10507, C. magnum DSM 2767 and E. aggregans DSM 12183 (Groher & Weuster‐Botz, 2016a ). It also aligns well with the findings of González‐Cabaleiro et al. ( 2013 ), where chain elongation to butyric acid with H 2 as electron donor was found to be kinetically bottlenecked in the first step of the reverse β‐oxidation pathway due to the unfeasibly low internal acetoacetyl‐CoA levels required to drive the condensation reaction of two acetyl‐CoA molecules. Increasing the initial acetic acid levels from 0 to 50 mM increased the selectivity of C4 compounds (sum of butyric and isobutyric acid) from 5 ± 1% to 53 ± 3%. Similar selectivity shifts towards butyric acid at elevated levels of acetic acid have been reported for methanol‐based studies with C. luticellarii , but also with E. limosum and B. methylotrophicum (Lynd & Zeikus, 1983 ; Pacaud et al., 1986 ; Petrognani et al., 2020 ). It has been hypothesized that high acetic acid concentrations exert a feedback inhibition on its production, thereby redirecting carbon flux from acetyl‐CoA to reverse β‐oxidation instead of acetic acid (Kremp & Müller, 2020 ), but it remains unclear how this alleviates the kinetic bottleneck proposed by González‐Cabaleiro et al. ( 2013 ). Isobutyric acid is likely formed by a selective isomerization of butyric acid In our experiments, isobutyric and butyric acids were co‐produced during the autotrophic growth of C. luticellarii at pH 5.5 and at elevated concentrations of acetic acid (Figure 2 ). Additionally, autotrophic growth at elevated butyric acid concentrations resulted in its consumption and the concomitant production of isobutyric acid at nearly equimolar amounts. This aligns with earlier work by Petrognani and co‐workers, who showed that isobutyric acid production from methanol and CO 2 is stimulated by elevated butyric acid concentrations, suggesting an interconversion of both compounds (Petrognani et al., 2020 ). The most plausible hypothesis is that C. luticellarii isomerizes butyric acid to isobutyric acid. This mechanism has been suggested before for C. luticellarii growing on methanol and CO 2 by Petrognani et al. ( 2020 ) and by Liu et al. ( 2020 ), who isolated a closely related Clostridium sp. BL3 producing butyric and isobutyric acid from lactic acid. However, neither study reported net butyric acid consumption alongside isobutyric acid production to fully support this hypothesis. The suggested pathway in these studies consists of isomerization of n‐butyryl‐CoA (coming from n‐butyric acid activation or from in situ production through reverse β‐oxidation) to isobutyryl‐CoA by a butyryl‐CoA:isobutyryl‐CoA mutase (BM) and final conversion of isobutyryl‐CoA to isobutyric acid by an isobutyryl‐CoA:acetate CoA transferase (Liu et al., 2020 ; Petrognani et al., 2020 ). Homologues for all enzymes involved are present in the C. luticellarii genome (Liu et al., 2020 ). All steps of the isomerisation pathway seem to be reversible, as supported by the observed consumption of isobutyric acid alongside production of butyric acid at nearly equimolar amounts (Figure 2 and Figure S3 ). However, C. luticellarii showed a much longer lag phase when grown on isobutyric acid compared to butyric acid, which is counterintuitive assuming the same enzymes are used in both directions. The mechanism behind the long lag phase thus remains unclear. To check whether C. luticellarii could also convert valeric acid or caproic acid to their respective isomers, autotrophic growth at elevated concentrations of valeric acid and caproic acid was tested. Neither of the added organic acids resulted in occurrence of their respective isomers. This aligns with the myriad of chain elongation studies that reported significant iso‐butyric acid production (but not iso‐caproic acid production) in mixed and pure culture systems engineered for butyric and caproic acid production (Huang et al., 2020 ; Liu et al., 2020 ; Mariën et al., 2022 ), and the lack of studies reporting significant iso‐valeric acid production in similar studies aiming for valeric acid production (Allegue et al., 2022 ; de Smit et al., 2019 ; Ganigué et al., 2019 ). This could be either due to the (iso‐)acids not being transported through the cell membrane, the acids not being activated to their respective CoA form, and/or the BM of C. luticellarii being specific for (iso)butyric acid. The transport into the cell and activation of the acid is likely not preventing production of isovaleric acid (or isomers of other carboxylic acids), since oxidation of valeric acid to propionic and acetic acid was observed, for which valeric acid needs to enter the cell and requires activation to valeryl‐CoA. The likely bottleneck is the specificity of the BM enzyme since several studies report acyl‐CoA mutases to have a strict substrate specificity (Cracan & Banerjee, 2012 ). Purification and characterization of the C. luticellarii BM could further prove this hypothesis. The ecological reason behind the interconversion of butyric acid to isobutyric acid (and vice versa) remains obscure since no free energy is released during the isomerization. Liu et al. ( 2020 ) and Allison ( 1978 ) argued that bacteria may convert n‐butyric acid to isobutyric acid to maintain a high isobutyric acid pool for the synthesis of valine in media poor in amino acids. Here, the medium C. luticellarii was grown on contains 0.5 g L −1 tryptone, providing a readily accessible source of amino acids (albeit at a low concentration). This suggests that either the mechanism may not be directly regulated by amino acid availability or another ecological reason is behind the mechanism. Others have suggested that isomerization is a way to deal with high concentrations of toxic n‐butyric acid by converting it to less toxic isobutyric acid, and thus allowing microorganisms to continue generating energy through butyric acid production (Chen et al., 2017 ). The production of isobutyric acid was not only observed here at high butyric acid concentrations but was also observed at low butyric acid concentrations (<10 mM) concomitantly with butyric acid production, casting doubt on the toxicity hypothesis. Additionally, neither of the proposed hypotheses explains the benefit of reverse isomerisation (i.e. isobutyric acid to butyric acid). Exploration of the proteome and metabolome of C. luticellarii under different environmental conditions may provide further answers as to why this isomerisation occurs. pH drives metabolic flexibility between the reductive and oxidative direction of the autotrophic reverse β‐oxidation pathway In our work, pH was identified to be a controlling parameter between reductive and oxidative processes in C. luticellarii : (1) circumneutral pH (~6.5) increased the fraction of valeric acid oxidized, and (2) mildly acidic pH (≤5.5) increased the selectivity for butyric and isobutyric acid production by stimulating reverse β‐oxidation. While both processes are seemingly different, they are in fact the same pathway running in opposite directions. Thermodynamic calculations and the results presented in this article show that C. luticellarii is unequivocally able to oxidize valeric acid to acetic and propionic acid at circumneutral pH under a hydrogen partial pressure of 120 kPa, contradicting at first glance established literature reporting the blockage of organic acid oxidation at H 2 partial pressures as low as 1 Pa (Ge et al., 2015 ; Stams, 1994 ). Organic acid oxidation is usually considered to be coupled with H 2 formation. This implies that the NAD + reduced to NADH (NAD + /NADH E′ = −280 mV (Buckel & Thauer, 2013 )) during oxidation must be reoxidised with protons as electron sink and hence the production of H 2 , which is thermodynamically unfeasible at high H 2 partial pressures (H + /H 2 E′ = −340 mV at 250 Pa H 2 which is even considerably lower than H 2 pressures used in this study (Schuchmann & Müller, 2014 ; Stams, 1994 )). A similar case of oxidation in seemingly unfavourable conditions has been reported in the model chain elongator C. kluyveri . For years, scientists had hypothesized on how C. kluyverii is able to oxidize ethanol (which yields NADH) during reverse β‐oxidation given that the strain internally accumulates H 2 up to a partial pressure of 1 bar (Li et al., 2008 ; Seedorf et al., 2008 ). The answer to this question was that NADH can be reoxidised in the reverse β‐oxidation pathway without generation of H 2 (Li et al., 2008 ; Seedorf et al., 2008 ). We argue that C. luticellarii may be able to carry out organic acid oxidations in a similar way where NAD + is likely regenerated by NADH oxidation in one of the reducing steps of the methyl branch in the WLP (i.e. reduction of CO 2 to formate, or the sequential reductions of methenyl‐THF to methylene‐THF and methyl‐THF) and not by H 2 production. However, currently very little is known about the electron carriers used in the Wood–Ljungdahl Pathway of C. luticellarii , making it difficult to prove this hypothesis at this point. Further characterization of the cofactors used in the individual reduction steps would allow performing a thermodynamic analysis over the entire metabolism (i.e. combination of the Wood–Ljungdahl pathway and organic acid oxidation) and elucidating the mechanism. Oxidation of organic acids simultaneously with autotrophic metabolism may also offer another advantage to C. luticellarii by providing an efficient alternative for ATP generation from reduced ferredoxin, which is produced during H 2 oxidation. During autotrophic metabolism, acetogens usually generate ATP from the excess reduced ferredoxin produced during H 2 oxidation by oxidizing it in the Ferredoxin:NAD + oxidoreductase complex (RnF), which uses the released energy to generate an ion motive force that drives ATP synthesis via an ATPase (Schuchmann & Müller, 2014 ). The oxidation of valeric acid offers an alternative to this because the enzyme complex catalysing the oxidation of valeryl‐CoA needs reduced ferredoxin in an electron confurcation process, where electrons from reduced ferredoxin and valeryl‐CoA reduce NAD + to NADH. Hence, the investment of two reduced ferredoxin (assumed monovalent here) in this case yields 3 NADH and one full ATP (Equation 5 ), while the classic RnF–ATPase route only produces one NADH and ~0.50–0.67 ATP from the same ferredoxin investment (Schuchmann & Müller, 2014 ). In this sense, we argue that those acetogens with substrate flexibility (i.e. the ability to (co‐)oxidize organic acids even at high H 2 partial pressures) may have a competitive advantage against those who cannot, since they may use more efficiently the available reduced ferredoxin and conserve more energy in the form of ATP. The use of valeric acid oxidation as an alternative way to regenerate ferredoxin is also supported by the fact that valeric acid is not oxidized in absence of H 2 , since in that case there is no direct source of reduced ferredoxin. This suggests that valeric acid oxidation is a secondary metabolism, depending on other substrates (here H 2 that drives the WLP) to support it. However, reduced ferredoxin could theoretically be indirectly generated through investing ~0.5–0.67 of the ATP formed in valeric acid oxidation via the reversal of the RnF–ATPase route, thereby consuming one NADH for the production of two reduced ferredoxin. In this way, valeric acid could be oxidized with the remaining two NADH being reoxidized to NAD + through H 2 production, maintaining a final ATP yield of 0.33–0.5 ATP per valeric acid oxidized. It is unclear why C. luticellarii is unable to do so to sustain growth in absence of H 2 . While the experiments presented here focused on valeric acid oxidation for which the degradation product propionic acid is easily detectable, a similar case can be made for the oxidation of other organic acids such as butyric acid. However, we were unable to prove butyric acid oxidation here due to (i) its degradation product being acetic acid which is also the main metabolic end product from the WLP, and (ii) butyric acid is also consumed for the production of isobutyric acid. Both processes occurring simultaneously with butyric acid oxidation make it difficult to accurately follow the carbon flux, hence, follow‐up studies using isotope‐labelled butyric acid may shed further light on this. The thermodynamic analysis identified pH as key parameter in Impact of the additions of valeric and caproic acid on the metabolism of C. luticellarii \n : at mildly acidic pH (<~5.5), C. luticellarii shifts its metabolism towards producing more of the chain elongation products butyric and isobutyric acid. Similar shifts towards chain elongation products have been reported by Worden et al. ( 1991 ) who observed a four‐fold increase in butyric acid production from CO by Butyribacterium methylotrophicum when shifting pH from 6.8 to 6.0. To explain the effect of pH on chain elongation and oxidation, we suggest the following hypothesis (Figure 6 ): at circumneutral pH (~6.5), C. luticellarii can leverage its substrate flexibility to oxidize H 2 and organic acids (e.g. butyric or valeric acid) simultaneously, thereby generating reduced electron carriers and allowing a higher ATP yield than when oxidizing H 2 alone due to the mechanism explained above. As long as pH remains high enough (>~5.5), C. luticellarii can run the Wood–Ljungdahl Pathway and organic acid oxidation simultaneously. At mildly acidic pH (≤~5.5), the oxidation of organic acids becomes less favourable and upon surpassing a thermodynamic threshold, the oxidation reverses and operates in the reductive direction (i.e. chain elongation) via reverse β‐oxidation (i.e. condensing and reducing two acetyl‐CoA molecules to produce butyric and isobutyric acid). This process represents a net consumption of electron equivalents coming from the oxidation of H 2 and is driven by the production of reduced ferredoxin (via electron bifurcation in the crotonyl‐CoA reduction) which can be used for additional ATP generation through its oxidation in the RnF complex. This remarkable metabolic flexibility shows how adapted C. luticellarii is to operate at the limit of thermodynamic viability since it is able to reverse the direction of carbon flux in its metabolism to generate additional ATP from the products of its autotrophic metabolism depending on the prevailing environmental conditions. Additionally, to the best of our knowledge this represents the first report of an acetogen with the capacity to use the (reverse) β‐oxidation in opposite directions according to a change of pH while conserving energy regardless of the direction. FIGURE 6 The hypothetical direction of electron flow in the metabolism of C. luticellarii driven by changes in pH. (A) C. luticellarii oxidizes hydrogen to generate reducing equivalents that contribute to the reduction of the electron carrier pool. (B) Reduced electron carriers are used in the WLP and for energy generation. The WLP serves as the central carbon fixation pathway regardless of pH. (C) Whenever pH is high enough (>~5.5) and butyric acid ( n = 4) or valeric acid ( n = 5) is present, they may be oxidized in parallel with the WLP to generate additional electron equivalents that contribute to the reduction of the electron carrier pool, as well as 1 ATP per organic acid oxidized. (D) When pH is low (≤~5.5) oxidation of organic acids is thermodynamically unfavourable and the pathway reverses to the direction of reverse β‐oxidation, now becoming an electron‐consuming pathway instead of an electron‐generating pathway. Fd, ferredoxin; Hdr, hydrogenase."
} | 5,814 |
36808979 | PMC9996816 | pmc | 9,223 | {
"abstract": "Reducing methane\nfrom livestock slurry is one of the quickest ways\nto counteract global warming. A straightforward strategy is to reduce\nslurry retention time inside pig houses by frequent transfer to outside\nstorages, where temperature and therefore microbial activity are lower.\nWe demonstrate three frequent slurry removal strategies in pig houses\nin a year-round continuous measurement campaign. Slurry funnels, slurry\ntrays, and weekly flushing reduced slurry methane emission by 89,\n81, and 53%, respectively. Slurry funnels and slurry trays reduced\nammonia emission by 25–30%. An extended version of the anaerobic\nbiodegradation model (ABM) was fitted and validated using barn measurements.\nIt was then applied for predicting storage emission and shows that\nthere is a risk of negating barn methane reductions due to increased\nemission from outside storage. Therefore, we recommend combining the\nremoval strategies with anaerobic digestion pre-storage or storage\nmitigation technologies such as slurry acidification. However, even\nwithout storage mitigation technologies, predicted net methane reduction\nfrom pig houses and following outside storage was at least 30% for\nall slurry removal strategies.",
"introduction": "Introduction Slurry from livestock animals is a considerable\nsource of global\nmethane emission. In many regions with intensive livestock production,\nmanure from pigs and cattle is managed as the liquid slurry in pits\nand channels underneath the barn floor. In such systems, the slurry\nenvironment is anaerobic, and anaerobic microbes transform organic\nmatter to methane and carbon dioxide. 1 , 2 The biological\nprocesses driving organic matter transformation depend strongly on\ntemperature. 3 − 5 Therefore, strategic removal of slurry from pig houses\nwith a temperature around 20 °C to cold outside storages or anaerobic\ndigesters can reduce net methane emission from the whole management\nchain. In Danish finisher pig houses, slurry pits underneath the floor\nare typically 40–60 cm deep, and the slurry is removed with\na vacuum flushing system when the pit is nearly full (after 5–6\nweeks). However, removing slurry more frequently on a weekly basis\nwas reported by Jørgensen et al. 6 to\nreduce in-house slurry methane emission by 45%. Methane emission rate\nis not directly proportional to slurry mass because slurry residence\ntime is crucial for the development of a methanogenic community, and\neffects of methanogenic adaptation on methane emission have been reported\nin multiple studies. 7 − 9 The complexity associated with a dynamic methanogen\ninoculum makes emission prediction difficult but presents an opportunity\nfor reducing emission through simple management changes. To avoid\ngrowth and adaptation of a methanogen community, it is necessary to\nremove the slurry frequently and reduce the amount of residual slurry\nin the pits. Vacuum flushing systems in conventional pig houses leave\na significant fraction of slurry behind (5–15%), and therefore\nnew slurry removal techniques must be developed. These should be tested\nand documented experimentally, but it is of increasing importance\nto also use modeling tools for (i) designing management strategies\nand (ii) integrating knowledge on methane production from animal slurry.\nThese management changes may increase methane emission from outside\nstorage due to increased transfer of organic matter from the barn.\nTo ensure a reduction in overall emission, the effects of this transfer\nmust be considered. With a higher slurry removal frequency,\nthe methanogen growth rate\nbecomes the limiting factor for methane emission but most currently\navailable models do not explicitly account for this. Instead, most\ncurrent models assume a fixed methanogenic community, for which the\nactivity is determined by temperature through an Arrhenius equation. 10 , 11 The anaerobic biodegradation model (ABM) 1 was developed to fill this gap, taking into account a dynamic community\nof methanogens and a separate step for conversion of organic matter\nto methanogen substrates (volatile fatty acids). The model processes\nrespond to changes in temperature, chemical environment, and management\npractices. However, reducing uncertainty in parameter estimates is\nessential and requires a large data set with repeated slurry filling-emptying\ncycles. The objectives of this work were to (i) evaluate three\ntypes of\nslurry handling systems with frequent removal of slurry for effectiveness\nin reducing methane, ammonia, and odor emission using full-scale experiments\nand (ii) refine, test, and apply a mechanistic model of slurry methane\nemission (ABM) to experiments and use it for predicting emission reduction\neffects in the complete manure management chain (in-house and outside\nstorage). The hypotheses of the study were that (i) new slurry removal\ntechniques focusing on increased removal frequency and removal of\nresidual slurry can reduce in-house slurry methane emission by at\nleast 50%, and (ii) a model that explicitly considers microbial growth\ncan accurately predict reductions in methane emission from barn systems\nwith frequent slurry removal. We consider an agreement in the overall\nemission reduction within ±10% of reference (control) emission\nto be accurate. Here, we report in situ continuous and year-around\nemissions of\nmethane, ammonia, carbon dioxide, hydrogen sulfide, and odor from\nfour pig sections with (i) weekly vacuum flushing, (ii) slurry trays,\n(iii) slurry funnels, and (iv) standard vacuum flushing. A best-fit\nparameter set of the ABM model was developed and used to evaluate\nthe model as well as extrapolate results of the manure removal strategy\neffects in barns and outside storages.",
"discussion": "Results and Discussion Emissions of Methane, Ammonia,\nCarbon Dioxide, and Odor Table 1 shows that\nin-house methane emission from the slurry was reduced considerably\nin all experimental sections compared to the control (C) section:\n89% ( p = 3 × 10 –6 ) for slurry\nfunnels (SF), 81% ( p = 3 × 10 –4 ) for slurry trays (ST), and 53% ( p = 0.044) for\nweekly flushing (WF). It is important to recognize that these reductions\ncould be negated in the outside storage due to increased transfer\nof organic matter, and this dilemma is addressed below. The dynamic\nemission rates in Figure 2 show that while emission peaks are higher in section C than\nin other sections, the first methane emission peak during period 1\nwas comparatively low, although it occurred simultaneously with the\nhighest slurry temperatures (in July). Temperature is a key factor\ncontrolling methane emission, which is inconsistent with this observation.\nIt is likely that during this period, the microbial inoculum was still\nnot established as the pig house had been absent of pigs and fresh\nmanure 1 year prior to this measuring campaign. During period 4, the\nslurry temperature was consistently higher in section WF than in section\nC, possibly explaining the similarity in methane emission during this\nperiod. Consequently, a Dunnett test finds no significant effect of\nWF (Dunnett test, p = 0.106), and the effect of WF\nis less clear than the others. Jørgensen et al. 6 reported a 45% reduction in slurry methane emission with\nweekly slurry removal, but this was with manure removal in weeks 5\nand 9 in the reference scenario, which could explain the small discrepancy\nfrom the present study. The simple model applied here for enteric\nmethane production was at least slightly inaccurate as evidenced by\nperiodic negative slurry methane emission ( Figure 2 ). Therefore, measured methane emission including\nenteric emission is included in Table 1 . Ammonia emission was reduced to the highest extent\nin section SF, by 29% ( p = 0.035), and measurements\nmay suggest a reduction in hydrogen sulfide and odor as well. Effects\non individual VOCs that contribute to odor are available in the Supporting\nInformation, Table S5 . For ammonia, the\nreductions in sections SF and ST (27%, p = 0.025)\ncan be attributed to the larger fraction of emitting surfaces being\neither dried out or urine running off the funnel and tray surfaces,\nthereby reducing net urea hydrolysis and release of ammonia. Unlike\nthe other manure removal strategies, WF did not provide a reduction\nin ammonia emission compared to C, and the overall mean was actually\n26% higher ( p = 0.12). Ammonia emission depends on\nthe emitting surface area, 33 and pit dimensions\nand flooring were nearly identical for the WF and C sections. Crust\nformation on the slurry surface was less pronounced for the WF section\ncompared to the C section, and this could contribute to higher ammonia\nemission. Reduced crust formation probably resulted from the more\nfrequent disturbance and mixing of slurry during flushing. Figure 2 Emission dynamics\nof four gases for four different slurry removal\nsystems over four production periods, along with slurry temperature\n(bottom panel). For slurry temperature, only measurements from the\ncontrol (C) and weekly flushing (WF) sections are shown. Methane and\ncarbon dioxide emissions are corrected for enteric contributions and\nrepresent emission from the slurry only. The start and end of the\nperiods are indicated with gray dashed lines. Hydrogen sulfide rates\nwere averaged over 5 days and other gases over 1 day. Table 1 Period-Averaged and Overall Mean Emission\nRates a Values are based\non n = 4 periods except for odor and H 2 S ( n = 2). b s is the sample\nstandard deviation. c P -values are from\nthe t -test of differences in period-averaged emissions\nbetween individual treatments and the control section from a linear\nmodel, and confidence intervals are from the same model. Statistical\ntests were omitted for H 2 S and odor due to the insufficient\nnumber of replicates. The\nfact that simple changes in slurry management drastically reduce\nmethane emission and simultaneously reduce emission of ammonia is\nan important and promising result, providing hope for low climate\nimpact pig production. These systems would be relatively simple to\nincorporate in new and some existing pig houses. Methane emission\nfrom outside slurry storages is expected to increase by the in-house\ntreatments presented here as they increase the transfer of degradable\norganic matter to outside storage. In northern regions, microbial\nactivity may be much lower in outside storages than in the pig house\ndue to lower temperatures, partially counteracting increases in substrate\ntransfer. In Denmark, the estimated yearly average temperature of\nan outside stored slurry is 9–10 °C, 34 which is substantially lower than within pig houses where\navailable measurements show temperatures of 16–22 °C with\nan average of ∼19 °C. 34 VanderZaag\net al. 26 measured methane emission rates\nfrom cold stored pig slurry in Canada that were lower than the current\nIPCC algorithms predict. 35 A reduced methanogenic\nactivity could, however, be outweighed by a much longer storage time,\nwhich may also change the chemical slurry properties drastically.\nDownstream methane emission scenarios resulting from frequent slurry\ntransfer are therefore important to consider in this context, and\nwe address these scenarios using the ABM below. Danish slurry\nmethane emission from finisher pigs is estimated\nto 1.3 kg CH 4 pig –1 produced (including housing and storage emissions), which is 0.91 kg CH 4 pig –1 produced from housing only\n(assuming that 70% comes from the barn). 34 Here, we measured 0.38 kg CH 4 pig –1 produced from in-house slurry in section C. Slurry methane\nemission from dirty washing water remaining in the pits between the\nbatches of pigs was not measured but assessed to have only a minor\neffect on emission estimates. The difference between measured emission\nand the Danish emission factor may be due to inaccurate emission factors\nand good management practices at the test facility (e.g., thorough\nwashing of the pig houses) or because the slurry was more effectively\ndrained with one plug per pen versus one plug per two to four pens\nat commercial farms. Regardless, the poor correlation highlights the\nneed for improving models and conducting farm scale emission measurements\nfor understanding drivers of methane emission from livestock slurry.\nA literature survey of reported methane emission from reference pig\nhouses clearly indicates high variation in methane emission and that\nthe current study had relatively low methane emission levels (Supporting\nInformation, Table S3 ). The treatment effect\nin cases with the low or high baseline methane emission level is discussed\nbelow. Ammonia emission levels were 1.90 kg NH 3 -N m –2 year –1 in section C (calculated\nfrom the full year emission value in Table 1 , pig count in Table 2 , and an area of 22 m 2 ), which\nis slightly below a Danish normative system reference value for partly\nslatted floor of 2.2 kg NH 3 -N m –2 year –1 . 36 Table 2 Experimental Conditions, Animal Growth,\nand Slurry Characteristics a control (C) weekly flushing (WF) slurry funnels\n(SF) slurry trays (ST) Temperature and\nventilation air temperature, °C 20.1 ± 1.2 20.3 ± 1.2 20.0 ± 1.2 20.0 ± 1.2 slurry temperature, °C 18.5 ± 1.5 18.6 ± 1.5 ventilation rate, m 3 h –1 pig –1 60.2 ± 24.2 63.7\n± 24.7 57.7 ± 24.8 57.0 ±\n25.3 Growth pigs\nin section 28.8 ± 0.8 28.8 ±\n1.5 29.5 ± 0.6 28.8 ± 0.4 daily growth, kg d –1 pig –1 1.10 ± 0.02 1.08 ± 0.06 1.07 ± 0.03 1.08 ± 0.02 feed consumption, kg d –1 pig –1 2.84 ± 0.19 2.85 ± 0.15 2.87 ± 0.34 2.86 ± 0.19 N ingested, g d –1 pig –1 156 ± 15 157 ± 15 162 ± 20 158 ± 11 N excreted, g d –1 pig –1 87 ± 11 86 ± 10 89 ±\n18 86 ± 11 Slurry slurry production, kg pig –1 418 ± 25 411 ± 22 459 ±\n43 NA b average slurry mass in section, ton 4.37 ± 1.19 1.27 ± 0.12 0.59 ± 0.02 0.78 ± 0.07 average slurry retention\ntime in section, d c 19.2 ± 2.3 4.79 ± 0.33 1.77 ± 0.07 3.67 ± 0.20 DM, g kg –1 69.6 ±\n10.9 81.5 ± 5.1 86.3 ± 8.9 85.1 ± 9.8 VS, g kgDM –1 768 ± 21 787 ± 10 794 ± 12 796 ± 15 pH 6.88 ± 0.15 6.82 ± 0.09 6.75 ± 0.09 6.83 ± 0.04 Kjeldahl-N, g kg –1 4.63\n± 0.54 4.85 ± 0.22 5.19 ±\n0.23 5.04 ± 0.20 NH 4 -N, g kg –1 3.37 ± 0.26 2.88 ± 0.10 3.15 ± 0.21 3.09 ± 0.17 VFA, g L –1 13.2 ± 1.6 11.2 ± 1.6 12.4 ± 1.4 11.8 ± 1.4 a Data given as mean ± standard\ndeviation of periods ( n = 4). b Not available due to systematic\nincorrect measurements. c Calculated as the average time the\nslurry spends in the pit before being removed. The molar CH 4 /(CH 4 + CO 2 ) fraction\nfrom the slurry increased with the average amount of slurry present\nin the sections. In section SF, the fraction was 0.7 ± 0.3%,\nand in section ST, it was 1.0 ± 0.2% (omitting period 3 for section\nST due to the negative carbon dioxide emission resulting from uncertainty\nin estimation of enteric carbon dioxide production, see Table 1 ). In section WF, the fraction\nwas 1.6 ± 0.6%, and in C, it was 5.2 ± 2.2%. Stoichiometry\ndictates that methanogenesis of pig manure with elemental composition\nC 13.2 H 22.2 O 6.5 N produces a CH 4 /(CH 4 + CO 2 ) fraction of at least 60%,\nsuggesting that the main source of carbon dioxide from the manure\nwas instead fermentation, surface respiration, or urea hydrolysis.\nCarbon dioxide emission from the manure was not significantly different\nin any of the sections, and differences in CH 4 /(CH 4 + CO 2 ) fractions between treatments hence reflect\nprimarily differences in methanogenic activity. A CH 4 /(CH 4 + CO 2 ) fraction in the range of 20–40%\nwas recently reported from the slurry sampled from the same measuring\ncampaign as this study. 4 In that study,\nan assay eliminated slurry surface respiration by incubation of the\nslurry with nitrogen in the headspace, 4 explaining the large difference from our in situ measurements. Experimental Conditions and Slurry Composition The\nDM, VS, and Kjeldahl-N contents of the slurries were on average lower\nin section C followed by WF, ST, and SF, monotonically increasing\nwith the estimated average slurry retention time in these sections\n( Table 2 ) and hence\nincreased time for organic matter transformation to occur. Notice\nthat the average slurry retention times are higher than half of the\nemptying interval due to the residual slurry being carried over in\nthe next filling/emptying interval. These transformation trends are\npartially consistent with the average emission, i.e., higher loss\nof VS in section C is consistent with higher methane emission, but\nthis could not be confirmed for carbon dioxide due to the larger uncertainty\nin measured emission. Conversely, TAN was slightly higher in section\nC, which could indicate more conversion of organic nitrogen to TAN.\nKjeldahl-N and TAN consistently increased during each batch and for\nall treatments ( Figure 3 ) with increases of 0.028 g Kjeldahl-N kg –1 slurry d –1 ( p ≤ 2\n× 10 –16 , r = 0.69) and 0.028\ng TAN kg –1 slurry d –1 ( p ≤ 2 × 10 –16 , r = 0.78). The pig diet was changed at a body weight of\n60 kg (around 40 days), but the N increase in the slurry was continuous\nand shows that the N excretion increases as the pigs grow. 37 Ammonia emission increased during each period\nand for periods 2–4 with an exponential increase toward the\nend. Ammonia emission responds linearly to the TAN concentration,\nand the exponential behavior could therefore be attributed to increasing\nventilation rate (data not shown) that affects mass transfer of ammonia\nacross the slurry–air interface. Another explanation could\nbe increased fouling on the slatted floor or depositing of urine puddles\non the solid floor toward the end of the batches, leading to an increased\nsurface ammonia emitting area. 33 , 38 Slurry production by\nthe pigs ( Table 2 )\nwas lower than national estimates for finisher pigs of 0.50 ton pig –1 produced . 36 Since\nslurry production in this study was inferred from slurry levels in\nthe pits, water evaporation may have reduced apparent slurry production\nrates by perhaps 5–15%. Slurry production in section ST is\nomitted as it was systematically underestimated (during all periods)\ndue to incorrect accounting of the slurry in the piping system during\nslurry removals. Figure 3 Concentration of nitrogen in slurry samples. TAN is the\ntotal ammoniacal\nnitrogen, and Kjeldahl-N is the total nitrogen as by the Kjeldahl\nmethod. Model Parameter Estimation\nand Validation A model sensitivity\nanalysis was conducted with different α opt and q max,opt reference values. With a low α opt reference value, the model was most sensitive to q max,opt and vice versa. This shows that the\nrate-limiting parameter depends on other parameter settings. The sensitivity\nanalysis results are shown in the Supporting Information, Figure S3 . From previous performance tests, 1 the q max,opt fit\nwas best at 40% of the reference values and therefore the parameter\nselection for optimization was based on the model sensitivity analysis\nwith q max,opt = 40% of reference values\n(equivalent to q max,opt of 0.6, 1.44,\nand 2.24 g COD-S g COD-B –1 d –1 for methanogen groups m0, m1, and m2, respectively)\nand α opt = 0.02 d –1 . Model sensitivity\nanalysis suggested that methanogenesis (mainly controlled by q max,opt ) will be rate-limiting for methane production,\nand q max,opt , α opt , K Scoef , and C Xi,in were chosen as optimization parameters. Other model input\nvariables are shown in Table 3 . The VFA and VSd inputs were calculated from the composition\nof the fresh slurry (Supporting Information, Table S6 ), whereas TAN and SO4 inputs were default ABM values. The\nmass of the slurry in the sections and the slurry temperature changed\nover time and was different for the four sections. Table 3 Model Input Variables for Control\nand Experimental Sections ABM input description unit value a VSd degradable VS in fresh slurry gCOD kg –1 73.8 b VFA conc. VFA in excreted\nslurry gCOD kg –1 2.83 TAN conc. total ammonia in excreted\nslurry gN kg –1 3.0 SO4 conc. sulfate in excreted slurry gS kg –1 0.2 pH pH in the slurry 6.88, 6.82, 6.75, 6.83 area surface area of the slurry m 2 22 resid_enrich enrichment\nfactor 0.9, 0.9, 0 c , 0.9 slurry_prod_rate slurry\nproduction rate kg d –1 Var d temp_C slurry temperature °C Var d a Multiple values\nindicate values\nfor different sections in the order of control (C), weekly flushing\n(WF), slurry funnels (SF), and slurry trays (ST). b Calculated from the fresh slurry\ncomposition (Supporting Information, Table S6 ). c The enrichment factor\n(resid_enrich)\nof SF was set to 0 as the slurry was exclusively handled in a piping\nsystem with high turbulence. d The model input was given as a vector\nof actual observations rather than a fixed value. As expected, the model showed good\nperformance for section C, which\nwas used for parameter estimation. Model-predicted methane emission\nreductions were accurate within ±10% of measured emission reductions\n( Table 4 ). The predicted\naverage methane emission rates were similar to measured emissions,\nbut the model did not always capture emission rate dynamics, particularly\nfor sections SF and ST ( Figure 4 ). The measured changes in emission rates over time are likely\nassociated with more frequent slurry disturbances and methane releases\nthrough ebullition, and they become increasingly apparent with lower\nmethane emission and different scaling of the Y axis\nin Figure 4 . This transport\nmechanism does not affect methane production and is not included in\nthe model. As previously mentioned, measurements periodically resulted\nin negative slurry methane emissions due to the subtraction of estimated\ncontribution from enteric methane production ( Figures 1 and 3 ) and obviously\nthis affects the performance of the model, which cannot predict production\nbelow 0. Further, the model algorithms for removal of slurry and enrichment\nof the residual slurry do not account for non-flat surfaces or spatial\ndifferences in slurry levels throughout the pit and piping system.\nThe VFA dynamics matched well in section WF but with poor performance\nin sections SF and ST ( Figure 4 ). The high VFA measurements in sections SF and ST, with short\nslurry retention times, suggest that hydrolysis rates of some organic\nmatter components are considerably larger than the average hydrolysis\nrate of degradable VS. Hence, improvements to the model should focus\non segregating degradable VS into smaller VS pools with considerable\ndifferences in hydrolysis kinetics. In fact, multiple VS pools with\ndifferent degradability might be the key to understanding the high\nvariability in reported methane emissions ( Figure 5 A), together with the methanogenic adaptation\nrate. In turn, VS pools and degradability are linked to feed composition\nand feed digestibility in the pig, implying that an integrated analysis\nmust be conducted to truly understand variation in reported emissions. Figure 4 Measured\n(black dots) and predicted (red lines) methane emission\nrates (to the left) and VFA concentration (to the right) of the different\ntreatments. Dashed gray lines indicate measuring periods and breaks\nin between. Y -axis scales vary for methane emission\nrates. Figure 5 (A) In situ measured methane emission from pig\nhouses (enteric\nemission + slurry emission) as gathered from a literature survey and\nfrom different pig categories. Red dots indicate averages. (B) ABM-predicted\nemission of methane emission from the control (C) section in the barn,\nstorage, and total (Net) with high and low baseline methane emissions\nas inferred by changing the hydrolysis rate (α opt ) and substrate conversion rate by methanogens ( q max,opt ). In the storage, the exported slurry volume is\nincluding washing water, whereas slurry volume in the barn is as excreted\nby the pigs and therefore excluding washing water. (C) Emission reduction\nof methane emission with the treatments, weekly flushing (W. flush.\n(WF)), slurry funnels (S. funnels. (SF)), and slurry trays (ST), in\nthe barn, storage, and total (Net), when changing parameter values\nof hydrolysis rate (α opt ) and substrate conversion\nrate by methanogens ( q max,opt ). Emission\nreduction was calculated as (1 – emission treatment /emission control ) × 100%; thus, negative reductions\nreflect higher emission from the treatment than the control. Table 4 Model Results and Comparison for Periods\n1–3 parameter\nvalues target parameters q max,opt K S,coef α opt C XI,fresh best-fit values 0.47, 1.13, 1.75 (m0, m1, m2) 1.17 0.049 0.063 performance control (C) weekly flushing\n(WF) slurry funnels (SF) slurry trays (ST) CH 4 measured, g d –1 162.3 63.6 18.1 31.5 CH 4 modeled, g d –1 157.9 65.6 16.9 24.8 CH 4 reduction measured, % 60.8 88.8 80.6 CH 4 reduction modeled, % 58.5 89.2 84.3 VFA conc. measured, gCOD kg –1 slurry 10.8 10.3 11.7 11.1 VFA modeled, gCOD kg –1 slurry 13.6 9.2 5.0 5.3 To increase certainty in parameter estimates, the\nobjective function\nshould be based on more output variables, i.e., carbon dioxide emission\nrate, and concentration of organic matter components in the slurry\nbut that would require a more complex algorithm for prediction of\ncarbon dioxide from fermentation processes. Using concentrations of\norganic matter components would on the other hand require intensive\nslurry sampling at multiple depths, which could disturb slurry processes\nor trigger the release of trapped methane. Measurement of individual\nparameters in separate experiments would reduce the number of variables\nto be optimized but introduce uncertainty due to differences in slurry\nproperties between in-lab and in situ slurry. In general, the\nABM model successfully captured methane emission\ndynamics from frequent slurry removal systems but had more difficulties\nin capturing emission dynamics from the unconventional slurry management\nsystems where slurry disturbances and transport phenomena affect measured\nmethane emission peaks relatively more. Recalling that large variations\nin methane emissions have been reported (Supporting Information, Table S3 ), model parameterization would benefit\nfrom not only multiple large measuring campaigns at pig houses with\ndifferences in management practice (including feeding, cleaning, and\nanimal type) but also from outside storage so that the effect of temperature\ncan be accurately tuned. Modeling of Management Implications in the\nBarn and Storage\nScenarios Measurements made in this single facility cannot\ndefinitively show that similar reductions would be observed in other\nlocations. A literature survey on methane emission from pig houses\n(Supporting Information, Table S3 ) showed\na mean of 17.2 ± 11.1 g pig –1 d –1 for finisher pigs including enteric- and slurry-derived methane\nemission ( Figure 5 A),\nand such large variation implies that factors controlling the production\nof methane vary substantially among facilities. The potential relative\nreductions from changes in slurry management may similarly vary. This\nuncertainty was partially assessed by predicting reductions in methane\nemission with high and low baseline methane emission, inferred by\nchanging the two rate-limiting processes: methanogenic activity (through\nchanges in q max,opt ) and hydrolysis rate\n(through changes in α opt ) ( Figure 5 B,C). Predicted slurry methane emission reductions\nfrom the barn (upper Figure 5 C) match well with 95% CI reported in Table 1 . For the three treatments, the barn-predicted\nranges of reductions were WF (40–70%), SF (84–96%),\nand ST (75–94%), strengthening confidence in measured reduction\neffects of the treatments in the barn. In all cases, considerably\nhigher q max,opt and α opt values reduced the treatment effects due to the depletion of degradable\nVS in the control section and rapid substrate consumption in the treatment\nsections, despite short slurry retention times. Increasing α opt , while keeping q max,opt fixed,\ndid not increase methane emission due to the VFA inhibition of the\nmethanogens—a phenomenon also observed in anaerobic digestion\nstudies. 27 , 39 , 40 The values\nof q max,opt and α opt applied\nin Figure 5 B,C are\nextreme but could theoretically result from changes in pig diet that\nare known to significantly affect the intestine microbiome 41 as well as chemical composition of excreted\nmanure. 42 As previously discussed,\nstorage methane emission could be significantly affected by the frequent\nslurry removal due to increased organic matter transfer to the outside\nstorage. In the model simulations, it is important to recognize that\nthe hydrolysis rate of VSd in barn simulations may not represent average\nhydrolysis rates in the storage, with an expected larger fraction\nof recalcitrant but still degradable VS remaining. Nevertheless, simulations\nof storage emissions ( Figure 5 B,C) suggest that storage methane emissions from section C\ncontribute significantly more than barn emission with 2.17 kg CH 4 m –3 slurry (exported from barn)\nversus 0.96 kg m –3 slurry (excreted by\nanimal) with the best-fit parameter set ( Figure 5 B, gray circles). The average retention time\nof slurry in the simulated storage was ∼4.8 months, equivalent\nto 5.4 kg CH 4 m –3 slurry (present\nin storage) year –1 , which is close to a baseline\nemission of 6 kg CH 4 m –3 slurry (present in storage) year –1 that Kupper et al. 43 reported in a recent review on storage emissions.\nHusted 44 measured year-around storage methane\nemission at a Danish farm of 8.9 kg CH 4 pig –1 year –1 and with emissions ranging from 0.15 to\n13.1 kg CH 4 m –3 slurry (present\nin storage) year –1 during cold and warm seasons.\nThis is roughly equivalent to ∼3.9 kg CH 4 m –3 slurry (exported from barn). Petersen et\nal. 45 measured methane emission from Danish\npilot storages with pig slurry during winter time at 0.11–0.23\nkg CH 4 m –3 slurry (present\nin storage) year –1 and during summer time at 14.0–17.4\nkg CH 4 m –3 slurry (present\nin storage) year –1 , which are consistent with the\nrange of emissions reported by Husted. 44 These figures suggest that the simulations underestimate storage\nemissions from section C, but on the contrary, Danish emission inventories\npredict 0.57 kg CH 4 m –3 slurry (exported from barn), 46 being considerably\nlower than model simulations presented here. This further suggests\nthat storage methane emissions are associated with considerable uncertainty\nand that more continuous measuring campaigns are needed. Figure 5 B indicates\nthat in the case of extremely high methanogenesis and hydrolysis activity,\nbarn slurry-methane emission will dominate due to the substrate depletion\nin the storage (flat curve in the storage plot). The simulations indicate\nthat there is a risk of losing reduction gains from the barn in the\nstorage when frequent slurry removal is applied, with higher risks\nfor sections SF and ST than WF (negative reductions in Figure 5 C, storage plots). However,\nsimulations suggest that with some parameter values, including the\nbest-fit parameter set, there is no increase or even reduction of\nmethane emission from the storage ( Figure 5 B, storage plots), despite more organic matter\nbeing transferred to the outside storage. This results from multiple\nmechanisms being triggered or enhanced with more frequent slurry removal:\n(i) lower concentrations of the dominant m1 methanogen in the slurry\nexported to the storage due to a limited growth period in the barn,\nwhich is consistent with Feng et al., 4 measuring\nlower specific methanogenesis activity in the frequently removed slurry;\n(ii) hydrolysis of degradable VS is not rate-limiting and hence increased\ntransfer of organic matter has little influence on methane emission;\n(iii) stronger inhibition of methanogens in the storage, due to increased\nconcentrations of hydrogen sulfide, but reductions were also estimated\nwithout inhibition from hydrogen sulfide (not shown). One should be\ncautious with model interpretations from storage simulations since\nthe datasets used for obtaining the best-fit parameter estimates did\nnot change much in temperature and therefore do not reflect well the\ntemperature effect on microbial activity at low temperatures. With\nthese reservations in mind, simulations (using optimized α opt and q max,opt ) predict net methane\nreductions of all the treatments of 48% for scenario SF (range, 25–52%)\nand 44% for scenario ST (range, 24–48%). The net effect of\nWF was smaller at 32% emission reduction and ranging from 26 to 41%\ndepending on the rate-limiting parameters ( Figure 5 B,C). Further model development will benefit\nfrom full-scale measuring campaigns focusing on outside storages at\na representative temperature range to adjust and increase confidence\nin parameter values under these conditions. Combining Frequent Slurry\nRemoval with Storage Mitigation Technology The risk of increasing\nemission from outside storages emphasizes\nthe importance of combining in-house frequent slurry removal with\nstorage mitigation technologies. Anaerobic digestion is an obvious\nchoice of technology since the frequently removed slurry contains\nmore VS, potentially increasing biogas output from anaerobic digestion.\nMøller et al. 47 assumed that 90% of\ndegradable VS is converted by anaerobic digestion, and Baldé\net al. 48 reported 88% reduction in methane\nemission from degassed slurry compared to normal storage. When neglecting\nfugitive methane emissions, which can be minimized by proper plant\nmaintenance, 49 an ∼90% reduction\nin methane emissions from the digestate storage as compared to normal\noutside slurry storage is a reasonable estimate. 48 Therefore, combining slurry funnels, which we here show\nto reduce in-house slurry-methane emission by 89%, with subsequent\nanaerobic digestion, a net methane reduction of ∼90% from the\ntotal manure management chain (sum of in-house and outside storage\nemission), is realistic. Early summer acidification with sulfuric\nacid was reported to reduce slurry methane emission from the storage\nof pig slurry by 95% 50 and is an alternative\nto anaerobic digestion. Ma et al. 51 also\nreduced methane emission by >95% using low-dose acidification with\n6 kg sulfuric acid m –3 slurry . However,\nMa et al. discussed that the inhibition effect might be overcome by\nthe methanogenic community at such low doses of sulfuric acid and\nit is unclear whether the slurry needs to be repeatedly acidified\nmultiple times during a season to sustain the inhibitive effect. 51 By combining slurry funnels with acidification\nmethods that reduce methane by >95% in the storage, net methane\nemission\nreduction (in-house and storage emission) could exceed 90%. Net methane\nreductions may vary slightly from the mentioned figures, depending\non how methane emissions distribute between barn and outside storage\nin a baseline emission scenario, as suggested in Figure 5 B,C, but these management combinations\nclearly show a large potential for reducing pig production’s\nclimate impact."
} | 8,707 |
35423570 | PMC8695775 | pmc | 9,224 | {
"abstract": "Hydrogels are a popular class of biomaterial that are used in a number of commercial applications ( e.g. ; contact lenses, drug delivery, and prophylactics). Alginate-based tough hydrogel systems, interpenetrated with acrylamide, reportedly form both ionic and covalent cross-links, giving rise to their remarkable mechanical properties. In this work, we explore the nature, onset and extent of such hybrid bonding interactions between the complementary networks in a model double-network alginate–acrylamide system, using a host of characterisation techniques ( e.g. ; FTIR, Raman, UV-vis, and fluorescence spectroscopies), in a time-resolved manner. Further, due to the similarity of bonding effects across many such complementary, interpenetrating hydrogel networks, the broad bonding interactions and mechanisms observed during gelation in this model system, are thought to be commonly replicated across alginate-based and broader double-network hydrogels, where both physical and chemical bonding effects are present. Analytical techniques followed real-time bond formation, environmental changes and re-organisational processes that occurred. Experiments broadly identified two phases of reaction; phase I where covalent interaction and physical entanglements predominate, and; phase II where ionic cross-linking effects are dominant. Contrary to past reports, ionic cross-linking occurred more favourably via mannuronate blocks of the alginate chain, initially. Evolution of such bonding interactions was also correlated with the developing tensile and compressive properties. These structure–property findings provide mechanistic insights and future synthetic intervention routes to manipulate the chemo-physico-mechanical properties of dynamically-forming tough hydrogel structures according to need ( i.e. ; durability, biocompatibility, adhesion, etc. ), allowing expansion to a broader range of more physically and/or environmentally demanding biomaterials applications.",
"conclusion": "5. Conclusions This study explored mechanistic insights into the stages and factors involved in tough hydrogel formation and its correlation with the subsequent mechanical properties. A model double network alginate–acrylamide polymeric system was explored, in the presence of both ionic (CaSO 4 ), and covalent ( N , N ′-methylenebisacrylamide) cross-linkers. The study offered an alternative reaction route to that predicted in the hypothesis; specific to ionic binding, and contrary to past reports, we observed a particular favouring of M-group co-ordination by Ca 2+ -linkers during gelation. More broadly, in the overall gelation reaction, two formation stages were identified spanning physical entanglement, dimerisation and covalent cross-linking in phase I, prior to the predominance of ionic cross-linking bonding interactions in phase II, which in turn seems to impact the exhibited mechanical properties. The detailed insights and broader findings offered in this paper may facilitate practical new synthetic routes to enhance the intrinsic gel chemo-physico-mechanical properties via direct intervention at various bonding stages, during the evolving cross-linking processes. We expect formation mechanisms in similar systems to be governed by the time-dependent bond formation, near-surface viscoelasticity of the bulk material, as well as the molecular architecture and composition of the gel relative to the crosslinking dynamics. These foundational insights will also aid understanding of tissue failure, tissue repair therapies, and design principles for future biomaterials and functional polymer gels. Such investigations will be the focus of future studies.",
"introduction": "1. Introduction Hydrogels are water-encased gels (typically >90% H 2 O) composed of molecular chain networks, which find use, in part or whole, across various biomedical applications ( e.g. ; contact lenses, drug delivery carriers, tissue engineering scaffolds and prophylactics, in consumer products). 1–11 Reports of tough hydrogels, especially Sun et al. 's 2012 account of double network (DN) tough hydrogels with remarkable mechanical properties ( i.e. ; DN of alginate–(poly)acrylamide(PAAm)), have further catalysed research into systems that can withstand (and recover) from various large mechanical forces, e.g. ; for potential use in physiological load-bearing applications. 12–14 Alginates, amongst the most widely used gel-forming components, are algae-sourced polysaccharides possessing hydrocolloid properties, and are biocompatible, biodegradable, immunogenic, and non-toxic. 13,15 Alginates are randomly 1-4-linked copolymers of repeating β- d -mannuronic acid (M-block) and α- l -guluronic acid (G-block) units – the acid block content, molecular weight and conformations, as well as the form and extent of cation-mediated ionic cross-linking, are crucial for alginates' gel-forming capacity and the resultant hydrogel chemo-physico-mechanical properties. 16–18 However, these classic covalent single network alginate hydrogels are mechanically weak ( e.g. ; break at low strain (∼120% for alginate)), rendering them unsuitable for mechanical loading. 12 DN approaches can markedly improve hydrogel toughness. Such DN tough hydrogels comprise two contrasting ( i.e. ; combinations of stiff/rigid but brittle networks with soft/ductile but mechanically weak networks) and interpenetrating block copolymer conjugated networks bound by myriad physico-chemical interactions over different length scales ( i.e. ; physical entanglement, ionic and covalent cross-linking), which also prevent dissolution of hydrophilic chains in hydrated polymer networks. 1,12,19–23 These viscoelastic networks yield mechanical properties often orders of magnitude greater than their discrete components because the loosely cross-linked DN systems allow molecules to slightly pull apart over large areas efficiently distributing stress throughout the material bulk. 1,24 Furthermore, the tough hydrogel materials properties ( i.e. ; permeability, stimuli responsivity, elastic modulus, fracture toughness and/or shear-thinning) are easily regulated through control of the preparation method and gel composition ( i.e. ; polymer volume fraction, temperature, and/or swelling agent). 13,23,25–30 Thus, structure–function relationships can be gauged and tuned through variation in cross-link type and density within hydrogels. Resilience ( i.e. ; ability to recover from elastic deformation), strength ( i.e. ; ability to bear a mechanical load) and toughness ( i.e. ; ability to resist fracture) are inherently contradictory material properties and so, hard to combine. High strength requires low mechanical dissipation ( i.e. ; suppression of dislocation and plastic deformation) while high toughness requires high mechanical dissipation during deformation ( i.e. ; large amounts of work before fracture). 31 DN tough hydrogels ( e.g. ; alginate–PAAm networks) can offer both high toughness and resilience via delayed stiffening and mechanical dissipation due to broad physico-chemical bonding and varying but complementary physico-mechanical properties between the gel components. 12,32 Physically cross-linked ( i.e. ; reversible) gels are bound by attractive non-covalent forces ( i.e. ; H-bonds, ionic cross-links and protein–ligand associations) between polymer chains. 13 Thus, absent covalent cross-links, linear chains form networks via topological ( i.e. ; entanglement) interactions, so exhibiting viscoelastic rheology. 21 Physical hydrogels, ( e.g. ; as formed by alginate), exhibit good toughness, but lack the creep resistance of covalent gels. 13 Ionic cross-links ( e.g. ; via Ca 2+ ) complexed with polyelectrolyte anions offer strong bonding and cross-link density, but also provide points of detachment/re-attachment, leading to, e.g. ; self-healing activity and modified de-/swelling behaviour of hydrogels. For example, Ca 2+ co-ordinates to M- and G-blocks on alginate during gelation, acting as junctions between blocks on adjacent chains ( i.e. ; egg-box model). 33–36 Thus, the distribution of M and G units along the alginate chain, as well as the changing M/G ratio value of the alginate system, as preferential ionic cross-linking occurs to one uronic acid block over another, determines many physico-chemical properties of the gel structure that are closely related to their functionality, i.e. ; high ratios results in a more elastic, flexible (although more fragile) gel whilst low M/G provide brittle, water-insoluble, more rigid gels. 18,37,38 This is because the semi-rigid chains of G-rich alginates strongly electrostatically interact with Ca 2+ via (G)–COO − groups, leading to either charge neutralisation of a single chain, or cross-linking across separate chains and so, possess greater durability due to the higher shear rate necessary to induce the chain orientation. 39–42 Conversely, for M-rich alginates, the electrostatic interactions are less significant, especially for low molecular mass, and the viscosity decrease starts at a lower shear rate. 43 However, there have been reports that low molecular weight and low M/G alginates produce the strongest, most well organised alginate structures. 44 Chemically cross-linked ( i.e. ; permanent) gels are facilitated through various functional groups in the polymer backbone, ( e.g. ; hydroxyl, amine and hydrazide) often via specific chemical cross-linking agents ( e.g. ; N , N ′-methylenebisacrylamide). The resultant permanent structure is usually more stable than for physically cross-linked, but may exhibit poor mechanical strength and toughness. 13 The degree of covalent cross-linking is usually the most important factor in determining the resultant macroscopic properties ( e.g. ; mechanical strength, swelling and encapsulant release). 45,46 Alternative chemical cross-linking methods include enzymatic linking and free-radical polymerisation. 13,47 Both physically- and chemically-cross-linked hybrid hydrogel systems undergo entanglement as well as ionic and covalent cross-linking of multi-component polymer networks; the extent of cross-linking dependent on the polymer functional groups, as well as the size and type(s) of cross-linking agent used. Such systems are the focus of this paper. A systematic exploration into the dynamic formation processes of model DN alginate–PAAm tough hydrogels should yield a fuller understanding of the myriad dynamic bonding interactions that occur within the tough hydrogel polymer structure. While past ( e.g. ) NMR studies have reported on chemical structure, bonding and internal mobility of constituents, the nature of solution NMR processing, ( i.e. ; gel dissolution prior to analysis, as well as the fact that sample spectra need to be acquired at high-temperature to decrease the viscosity of the gel) confounds data outputs and may introduce inaccuracies into systems; the various required interventions may affect the bonding environment and so give erroneous interpretations of structural changes and relationships in cross-linking processes. 16,48–51 Other methods ( e.g. ; mass spectrometry, XPS analysis etc. ), also suffer from potential process inaccuracies for tough hydrogels. This paper offers a detailed, time-resolved, investigation into tough hydrogel physico-chemical bonding effects and dynamic changes over a reaction, using myriad complementary analytical techniques ( i.e. ; spectroscopy (FTIR, Raman, UV-vis and fluorescence), microscopy (optical and video fluorescence), TGA-DSC, DLS, tensile- and compression-testing). Mechanistic insights into the formation routes and their correlations with resultant viscoelastic properties, are afforded using a non-intrusive continuous monitoring approach to give a clearer understanding of the structural changes over time, removing the need for guesswork or post-rationalisation with snapshot data, in order to highlight routes towards improved versatility and control over the resultant polymer physico-mechanical properties. 48,51–55 We explored a model DN alginate–PAAm system, with good molecular affinity between the two networks, as aided by both ionic (CaSO 4 ), and covalent ( N , N ′-methylenebisacrylamide) cross-linkers. Past reports into alginate hydrogels (theoretical and experimental) have explained the favourability of uronic acid block co-ordination for ionic bonding, where G-block co-ordination has frequently reported as most favourable in accordance with the egg-box model. However, the same such in-depth reports have not yet been made for alginate-based tough hydrogel systems, to the best of our knowledge. We hypothesise that, in line with past reports, G-block alginate co-ordination will be favoured during ionic cross-linking over the course of a gelation process. In such a G-block co-ordination, progress in a dynamically evolving reaction will result in rising M/G ( i.e. ; a positive correlation of M/G with time).",
"discussion": "4. Discussion The internal arrangement of the copolymer structure materially impacts the favourability of subsequent co-ordination to the alginate backbone and so, the favoured reaction path in a tough hydrogel formation process. Knowledge of such rearrangements is important, since pre-dominant co-ordination through one backbone unit-type preferentially over the other manifests marked differences in the resultant mechanical properties. Furthermore, the agglomeration route, as facilitated by the cross-linkers, and through electrostatic repulsion minimisation, indicates the mechanism by which aggregation is favoured. 168–171 Specifically, for the model DN Alg–PAAm tough hydrogel system explored in this study, time-resolved investigations identify two broad stages in tough hydrogel formation, where different bonding interactions seem to predominate across the macromolecular chains in solution, which has a critical impact on the resultant mechanical properties. Initially, during phase I, physical entanglement and covalent cross-linking are the major drivers, as monomers and dimers cross-link into longer polymer chains, signifying gelation onset. 73 Covalent bonding is thought to proceed via both the unsaturated carbon bonds on the Alg–PAAm frameworks, as well as the Alg-carbonyl groups which bind to the AAm-amide groups, via a MBA bridge. 14,107,114,172,173 After ∼8–12 min (and not appreciably earlier), in phase II, ionic cross-linking effects come to the fore, as Ca 2+ -mediated cross-linking initially proceeds via both M-block (slightly more favourable) and G-block units. 15,110 M-block co-ordination is thought to proceed via a bidentate bridging co-ordination, with the Alg-carboxylate groups playing a key role. 92,117,172 This also corresponds with increasing stiffness of the structure. The delayed onset of ionic contributions observed experimentally, may be partly due to the slower dissolution of the CaSO 4 precursor, which in turn delays the ionic contributions to bonding. This seemingly results in a more effective tough hydrogel system with more beneficial mechanical properties. Past empirical reports indicate that when insufficient durations are spent at each cross-linking stage ( e.g. ; faster onset of ionic cross-linking contribution that bypasses the dimerization phase), there is incomplete diffusion and entanglement of molecular chains, non-ideal cross-linking effects, and void formation, resulting in weaker and unsatisfactory mechanical properties. 34,74,75,174 Equivalent molar ratio variations of alternative cationic cross-linker ( i.e. ; CaCl 2 , MgSO 4 , Na 2 SO 4 , and Ca-lignosulfonate) were also explored, in place of CaSO 4 (see ESI † ). Thus, ideal mechanical properties ( i.e. ; ultimate strength and toughness), need optimisation of cross-link density and type in polymer systems, so as to improve ( e.g. ) the energy dissipation mechanism. 175–178"
} | 3,970 |
34740967 | PMC8609452 | pmc | 9,225 | {
"abstract": "Significance The physiological and ecological importance of natural products often remains obscured. Here, we report that Streptomyces -derived marginolactones, a distinct group of soil-borne natural products, specifically trigger the formation of gloeocapsoids, previously undescribed protective aggregate structures produced by the unicellular green alga Chlamydomonas reinhardtii. Gloeocapsoids are distinct palmelloids differing in their protective capability toward azalomycin F. The presence of marginolactone biosynthesis gene clusters in numerous streptomycetes, their ubiquity in soil, and our observation that three different members of this natural product group trigger the formation of gloeocapsoids suggest a cross-kingdom competition with ecological relevance. In the course of evolution, the polysaccharide matrix may have developed from a transient protective feature into the foundation of true multicellularity because of sustained marginolactone stress.",
"discussion": "Discussion Multicellular Gloeocapsoids Represent a Defense Strategy against Both Alkaline pH and Azalomycin F. Microorganisms face everchanging conditions in their respective habitats. Because of acids or bases released by microorganisms or farming, the environmental pH values can vary considerably between 3 and 10 ( 27 , 41 ). By the uptake of acetate, C. reinhardtii increased the pH around its colonies. Under these conditions, the natural product azalomycin F produced by S. iranensis displayed reduced algicidal activity ( SI Appendix , Fig. 2 ). This explains why a higher azalomycin F concentration was needed to induce gloeocapsoids in C. reinhardtii at pH 8 than at pH 7. It is remarkable that gloeocapsoids protect C. reinhardtii both against alkaline pH and azalomycin F since streptomycetes can increase their surrounding pH by the secretion of ammonia ( 27 ). Thus, an increase in pH in combination with algicidal marginolactones are indicators for the presence of competing streptomycetes. As a response, C. reinhardtii forms gloeocapsoids that represent a protective response to the presence of these harmful bacteria ( Fig. 5 ). Fig. 5. Schematic overview of the interaction of Streptomyces spp. with C. reinhardtii . C. reinhardtii induces the release of the marginolactone azalomycin F as well as the overproduction and release of desertomycin A and monazomycin in cocultures with the bacteria Streptomyces iranensis , Streptomyces macronensis , and Streptomyces mashuensis . Lethal concentrations of these marginolactones lead to bleaching and death of C. reinhardtii . Sublethal concentrations trigger the formation of multicellular gloeocapsoids in which the individual cells are surrounded by multiple cell walls, membranes, and a spacious polysaccharide matrix. These aggregates protect the individual cells against otherwise lethal azalomycin F concentrations as well as alkaline pH, indicating the presence of competing streptomycetes. We hypothesize that the additional membranes and the acidic polysaccharide matrix sequester the positively charged azalomycin F, and hence, the algicide cannot reach the individual algal cells. After stress is relieved, the gloeocapsoids disassemble and motile single cells emerge. We hypothesize that sustained sublethal stress leads to the evolution of a multicellular organism exhibiting differentiated cell types and division of labor. We have shown that C. reinhardtii required up to 7 d to develop gloeocapsoids, and therefore concentrations of 10 and 15 μg ⋅ mL −1 of azalomycin F killed non-pretreated algal cells within 24 h ( Fig. 5 ) ( 24 ). C. reinhardtii only formed resistant gloeocapsoids at sublethal azalomycin F concentrations. This reflects an ecological scenario in which C. reinhardtii is not instantly confronted with high concentrations of an algicide but rather with a concentration gradient. It is worth noting that this might be a reason for the unusual release mechanism of azalomycin F. S. iranensis steadily produces the compound but only secretes it in the presence of C. reinhardtii ( 24 ). Production, storage, and massive release of the algicide upon contact with C. reinhardtii might be a strategy of the bacteria to circumvent the algal resistance mechanism by instantly creating a high concentration of azalomycin F. Multicellularity Is a Widespread Protective Mechanism against Environmental Stressors. We provide evidence that aggregation facilitates the protection of single cells against environmental stressors. This is in accordance with observations made for, e.g., Pseudomonas aeruginosa , that produces biofilms in contact with antibiotics ( 42 ) or when experiencing DNA replication stress ( 43 ). A number of stressors induce the formation of palmelloids and actively formed aggregates in C. reinhardtii ( 14 ). Most likely, they play a role in protecting C. reinhardtii against these stressors. Since the main difference of NaCl- and sodium citrate–induced palmelloids and azalomycin F–induced gloeocapsoids is their protection against azalomycin F ( Fig. 4 and SI Appendix , Figs. 11–14 ), we assigned gloeocapsoids to the family of palmelloids as specialized structures for defense against azalomycin F. The matrix of gloeocapsoids is composed of acidic polysaccharides, which may have the capacity to buffer alkaline pH and might help to maintain an internal pH near 7. Similarly, sulfate-reducing bacteria buffer the surrounding pH by secretion of exopolysaccharides ( 44 ). The acidic polysaccharide matrix could also explain the resistance of gloeocapsoids against azalomycin F ( Fig. 4 C and D and SI Appendix , Fig. 10 B ). We hypothesize that the guanidyl moiety of azalomycin F is positively charged at physiological pH and thus may be sequestered by the negatively charged acidic polysaccharides of the extracellular matrix. Furthermore, membranes are known targets of azalomycin F ( 24 ). Since there are several membrane layers present in gloeocapsoids ( SI Appendix , Fig. 4 A ), this might promote sequestration of azalomycin F over a larger surface area ( Fig. 2 B ). Possibly, the additional surrounding membranes sequester marginolactones and thus protect the cytoplasmic membranes of cells lying underneath ( Fig. 5 ). A similar protective mechanism appears to be realized in the tripartite system examined recently. The mycelium of A. nidulans is proposed to provide an increased surface area composed of polar lipids that function to sequester free azalomycin F, thereby protecting the cohabitating C. reinhardtii ( 24 ). The formation of multicellular gloeocapsoids represents a protective strategy that does not require any other microbial partner organism. From an evolutionary point of view, such a strategy is in accordance with the theory that multicellularity initially evolved for survival in a hostile environment ( 45 ). Multicellularity is the fundamental step in the evolution of all higher organisms such as animals ( 46 , 47 ) and enables maintenance of internal homeostasis even in the presence of osmotic stress, extreme pH, toxins, or desiccation ( 45 , 48 ). This is also evident in the multicellular volvocine algae of the order Chlamydomonadales , Pleodorina and Volvox ( 49 ). These algae produce extracellular matrices, which enable the differentiation of soma and germ cells and also protect offspring developing inside the parent spheroid ( 49 ). Here, we report that C. reinhardtii produces a similar type of polysaccharide matrix. We highlight that the ability to form polysaccharide matrices is already present in the most basal member of the volvocine algae and might have been fundamental for the evolution of true multicellularity. This is also reflected in the low number of species-specific genes in Volvox carteri compared to C. reinhardtii ( 50 ). However, an essential difference found between Volvox and C. reinhardtii gloeocapsoids is the missing cytoplasmic connection between the individual cells, which enables Volvox cells to synchronize their behavior ( 51 ). Additionally, the polysaccharide matrix found in gloeocapsoids is linked to specific bacterial triggers and thus is formed for protection against harmful bacteria and could have later resulted in an evolution of multicellularity as suggested by similar experiments from Khona et al. ( 6 ). We hypothesize that multicellular green algae might have evolved by the sustained presence of bacterial marginolactones granting protection against these widespread algicides. Algicidal Marginolactones Are Widespread in Nature. The marginolactones azalomycin F, desertomycin A, and monazomycin exhibit similar structures and target the cell membrane ( 37 , 38 , 52 , 53 ). Sublethal doses of these natural products induce the formation of gloeocapsoids in C. reinhardtii . By contrast, amphotericin B and daptomycin, which also target the membrane by binding to ergosterol ( 39 ) or interfering with membrane microdomains ( 54 ), respectively, did not induce formation of gloeocapsoids ( SI Appendix , Figs. 7–9 ). Thus, it is not merely membrane stress that leads to gloeocapsoids but rather the specific activity of marginolactones. For desertomycin A, it was shown that its activity depends on the primary amine on the side chain ( 55 ). This is supported by the observation that inactive members of the desertomycin family, the oasomycins, lack a primary amine in their side chain. The substitution of this side chain with side chains containing a primary or secondary amine created oasomycin derivatives with antibacterial activity ( 55 ). Our data further support the notion that the activity of marginolactones relies on their positively charged side chain. Since marginolactones have been isolated from a number of actinobacteria, it can be expected that these compounds are widespread in soil habitats ( 56 , 57 ). Bioinformatic analyses show that the azalomycin F biosynthetic gene cluster is present in a number of streptomycetes isolated from diverse soil samples across the world ( 58 – 61 ). We previously showed that azalomycin F was specifically released by S. iranensis in the presence of C. reinhardtii and possesses algicidal activity against several algal species ( 24 ). Similarly, we demonstrated here that S. macronensis and S. mashuensis overproduced their respective marginolactones in the presence of C. reinhardtii ( Fig. 3 ). It is tempting to speculate that the streptomycetes create a spatially confined lethal concentration of marginolactones in soil to kill neighboring algae. It is worth noting that desertomycin G was isolated from a marine streptomycete present on the surface of a macroalga ( 62 ). Our data suggest that marginolactones may have evolved to target algae. The widespread distribution of marginolactone-producing bacteria makes it likely that the compounds affect C. reinhardtii in its natural habitat. The alga in turn developed a particular strategy for defense against these compounds that includes the aggregation of cells, production of a polysaccharide matrix, and maintenance of multiple cell membranes. This structure, which we named gloeocapsoid, allows C. reinhardtii to cope with marginolactones and alkaline pH stress."
} | 2,816 |
34489412 | PMC8421358 | pmc | 9,230 | {
"abstract": "Phenotypic plasticity represents a capacity by which the organism changes its phenotypes in response to environmental stimuli. Despite its pivotal role in adaptive evolution, how phenotypic plasticity is genetically controlled remains elusive. Here, we develop a unified framework for coalescing all single nucleotide polymorphisms (SNPs) from a genome-wide association study (GWAS) into a quantitative graph. This framework integrates functional genetic mapping, evolutionary game theory, and predator-prey theory to decompose the net genetic effect of each SNP into its independent and dependent components. The independent effect arises from the intrinsic capacity of a SNP, only expressed when it is in isolation, whereas the dependent effect results from the extrinsic influence of other SNPs. The dependent effect is conceptually beyond the traditional definition of epistasis by not only characterizing the strength of epistasis but also capturing the bi-causality of epistasis and the sign of the causality. We implement functional clustering and variable selection to infer multilayer, sparse, and multiplex interactome networks from any dimension of genetic data. We design and conduct two GWAS experiments using Staphylococcus aureus , aimed to test the genetic mechanisms underlying the phenotypic plasticity of this species to vancomycin exposure and Escherichia coli coexistence. We reconstruct the two most comprehensive genetic networks for abiotic and biotic phenotypic plasticity. Pathway analysis shows that SNP-SNP epistasis for phenotypic plasticity can be annotated to protein-protein interactions through coding genes. Our model can unveil the regulatory mechanisms of significant loci and excavate missing heritability from some insignificant loci. Our multilayer genetic networks provide a systems tool for dissecting environment-induced evolution.",
"introduction": "Introduction Most organisms are equipped with a capacity to produce multiple phenotypes in response to environmental change 1 – 6 . This so-called phenotypic plasticity is coded by a plastic developmental program that allows the organisms to sense environmental cues in early stages of life and develop phenotypes better adapted to environments encountered later in life 7 . Phenotypic plasticity is thought to be under genetic control 8 – 12 ; i.e., specific genes may occur to govern its pattern, sign, and magnitude. In modern ecological, evolutionary, and medical genetics, the genetic mechanisms underlying phenotypic plasticity and its evolutionary novelty have emerged as an important topic of research in both theory and application 4 , 13 – 18 . Genetic mapping and genome-wide association studies (GWASs) have been used to map the genome-wide distribution of quantitative trait loci (QTLs) responsible for phenotypic plasticity in a variety of species 19 – 25 . With the advent of high-throughput genotyping techniques, there is a pressing demand to chart a more detailed and precise genetic atlas of this phenomenon. Existing approaches are founded on reductionist thinking, which can only identify individual key QTLs at a time. However, it is becoming increasingly clear that a deeper genetic understanding of phenotypic plasticity as a complex trait requires not only a detailed characterization of its underlying individual genes, but also of their interactions as a cohesive whole 26 – 31 . A recently emerging theory, known as omnigenic theory, states that complex traits are controlled by all genes carried by an organism 32 . Taken together, to better interpret phenotypic plasticity, we need to coalesce all genes into genetic interaction networks and trace the roadmap of each gene toward this phenomenon. Networks are mathematical graphs, connecting nodes (e.g., genes) via edges (genetic interactions), which are particularly powerful for disentangling complex systems 33 , 34 . However, there is a formidable challenge in reconstructing genetic networks. First, for ordinary GWAS, hundreds of thousands of SNPs genotyped for each sample are not uncommon, making it computationally infeasible to reconstruct an omnigenic network. Second, the detection of genetic interactions requires a sample size that is hardly met in practice. Computer simulation shows that as many as 10,000 samples are needed to detect genetic interactions between a pair of genes in a GWAS 35 , let alone millions of millions of gene pairs. Third, in classic quantitative genetics, genetic interactions are defined as epistasis, which describes how much the effect of one gene reciprocally depends on the other gene 34 , 36 – 38 . Statistical approaches currently used in the literature cannot reveal the casualty of epistasis and the sign of casualty. Understanding the direction of regulation from one gene to next is of utmost significance to disentangle the genetic architecture of complex traits and design an efficient and effective gene editing program for trait improvement and disease control. Fourth, genetic variation in phenotypic plasticity is displayed as gene–environment interactions of which most studies are based on linear additive models 39 . However, the organism often responds to environmental change in a nonlinear manner owing to substantial uncertainties and random effects across timescales 40 – 42 . Nonlinearities between genes and environment can help trait phenotypes maintain their robustness to random perturbations in environmental exposures 43 . Here, we develop a computational model to infer omnigenic interactome networks underlying phenotypic plasticity. The model absorbs and integrates the elements of multiple previously disjointed disciplines into a unified framework. Association analysis based on individual SNPs estimates the marginal (net) genetic effect of each SNP on a complex trait from a pool of genome-wide markers. Evolutionary game theory 44 allows us to interpret how the genetic effect of a SNP is determined by its own intrinsic capacity and the epistatic influences of other loci on it. The integration of evolutionary game theory with predator-prey theory 45 – 49 makes it possible to derive a generalized nonlinear Lotka–Volterra (nLV) equations that decompose the net genetic effect of a SNP into its two components, the independent effect due to its intrinsic capacity and the dependent effect resulting from regulation by other SNPs 50 . According to network theory, we encapsulate independent effects of individual SNPs as nodes and dependent effects of SNP pairs as edges into a graph, in which edges represent bidirectional, signed, and weighted (bDSW) genetic interactions (i.e., epistasis). Viewing all SNPs together as a quantitative system, the distribution of bDSW interactions portrays a picture of the biological role of epistasis. By incorporating developmental modularity theory 51 and high-dimensional statistical models, we can reconstruct multilayer, multiplex, large-scale, and sparse genetic networks from any number of SNPs. We design and conduct two independent GWAS experiments of phenotypic plasticity to abiotic and biotic factors, respectively, using bacterial species S. aureus . As one of the most important members of the Firmicutes, S. aureus often cause skin infections, but can also lead to pneumonia, heart valve infections, and bone infections 52 . The abiotic GWAS experiment is to study how the growth of S. aureus responds to vancomycin, a glycopeptide antibiotic designed to control and treat S. aureus -caused diseases by inhibiting cell wall biosynthesis 53 , 54 . In clinical practice, the use of this drug has incurred vancomycin-intermediate S. aureus (VISA) isolates, with the minimal inhibitory concentration (MIC) of 4–8 µg/mL 55 . Vancomycin non-susceptibility in S. aureus involves a genetic component. Several nucleotide substitutions that distinguish the vancomycin-susceptible S. aureus from VISA isolates have been identified by a whole-genome sequencing approach 56 – 58 . Some studies further used GWAS to characterize common genetic variants associated with antibiotic resistance in S. aureus 55 , 59 . However, despite these advances, a comprehensive portrait of the genetic control mechanisms underlying vancomycin resistance is still unclear. The biotic GWAS experiment aims to investigate how S. aureus changes its growth when encountered with another species, Escherichia coli . Studying the plastic response of one bacterial species to its co-existing species has become increasingly interesting to researchers from many fields including the gut microbiomes 60 , 61 , but the genetic architecture of this phenomenon has been little explored. Jiang et al 60 . characterized a set of specific QTLs in S. aureus that mediates the change of microbial growth from a socially isolated environment to a socialized environment. It is likely that phenotypic plasticity to biotic factors includes a complex, still unknown genetic machinery. We apply our new model to systemically dissect the genetic architecture of how S. aureus responds to an abiotic factor (drug) and biotic factor (species coexistence). Phenotypic plasticity can be quantitatively defined as the difference of phenotypic values for the same genotype expressed in two different environments 9 – 11 . We define the difference of microbial growth for the same S. aureus strain expressed in vancomycin-free and vancomycin-exposed media or in isolated and E. coli -socialized media as the abiotic or biotic phenotypic plasticity of this strain, respectively. First, we globally view the genetic landscape of each type of phenotypic plasticity, filled by a complete set of bDSW interactions, from hundreds of thousands of SNPs and delimit this landscape into functionally distinct network communities. Second, we characterize the detailed roadmap through which individual key genes determine phenotypic plasticity directly and/or through multiple indirect pathways. Our model opens up a new avenue to unveil the genetic complexities of how genes interact with the environment to help the organism better adapt to environmental and biological cues.",
"discussion": "Discussion Phenotypic plasticity involves a complex genetic component 8 – 12 , but its genetic analysis is based on reductionist thinking of individual gene identification. As a composite trait derived multiple phenotypes expressed in different environments, phenotypic plasticity is likely to comply with omnigenic theory 30 ; i.e., it may be controlled by all genes an organism may carry. By integrating mapping theory and elements of multiple disciplines, we develop a computational model to encapsulates all SNPs from a GWAS into a phenotypic plasticity-driven genetic network. Different from gene regulatory networks from expression data 86 , 87 , our genetic networks unravel a direct link from DNA genotype to phenotype, providing a broader picture of genetic architecture. One major advantage of our model lies in its capacity to chart a comprehensive roadmap of how each SNP flows its genetic signal towards phenotypic plasticity, whether a SNP exerts its effect either in its own capacity (independent effect) or through regulation by promotors or inhibitors (dependent effect), or both, and how the intrinsic effect of a SNP is masked by other loci. To validate the biological relevance of our model, we design and conduct two independent GWAS experiments by culturing each panel of S. aureus strains in two contrast environments. These two experiments are complementary, one focusing on the genetic mapping of phenotypic plasticity to an abiotic environment (vancomycin-free vs. vancomycin-exposed) and the second on the genetic mapping of phenotypic plasticity to a biotic environment (isolation vs. socialization). We use our model to reconstruct a large-scale, multilayer, and maximally informative genetic network underlying abiotic phenotypic plasticity and biotic phenotypic plasticity, respectively. The two networks unravel many previously undetected genetic mechanisms by traditional mapping approaches and, also, produce findings that are reciprocally supported from the two GWAS experiments. From the genetic networks reconstructed from two GWAS experiments, we identify common causes of the significance of QTLs detected by genetic mapping based on marginal effects of single SNPs. coFunMap finds 16 and eight key QTLs of significant marginal effects that mediate S. aureus ’ response to vancomycin and to E. coli coexistence, respectively, in the abiotic and biotic GWAS. We postulate three mechanisms underlying the significance of these QTLs. A QTL is significant mainly because of its intrinsic capacity to express a strong independent genetic effect. Q550323 is a QTL detected in the abiotic GWAS experiment. Its tremendous independent effect on vancomycin susceptibility estimated by our model (Fig. S 4C ) is in good agreement with the direct role of gene SAOUHSC_00544 ( SdrC ) in mediating vancomycin resistance 63 , 64 at which this QTL is located (Table S2 ; Fig. S 8 ). Q550323 is so highly expressed in its own capacity that its net effect is still remarkably large even if it receives a negative regulation from S119306 (in a non-coding region). Q1831526 is a QTL located at gene SAOUHSC_01926 coding a hypothetical protein (Table S3 ), which is detected to affect the plastic response of S. aureus to the coexistence of E. coli in the biotic GWAS experiment (Fig. 7 ). This QTL displays a sizeable independent effect, which can be observed (net effect) even if it is down-regulated by a negative regulator S563832 (in a non-coding region). In practice, by silencing the expression of S119306 and S563832 vis gene editing, the effect of Q550323 on abiotic phenotypic plasticity and the effect of Q1831526 on biotic phenotypic plasticity can be amplified, respectively. Most of the QTLs detected from both GWAS experiments affect phenotypic plasticity through this mechanism, i.e., making use of favorable regulation from other regulatory genes rather than their own intrinsic capacity. For example, Q2728755 located at SAOUHSC_02967 of the S. aureus genome (detected from the abiotic GWAS experiment) does not affect vancomycin susceptibility in its own intrinsic capacity, but it does exert a remarkable effect, especially in the early stage of vancomycin susceptibility, through positive regulation from S2085718 (at the SAOUHSC_02250 location), S2773077 and S2773290 (both at the SAOUHSC_03000 location) (Figs. 4c , S 8 ). SAOUHSC_03000 ( capA ) gene is involved in the biosynthesis of outer capsules of cell walls 88 , which provides material support for Q2728755 to mediate S. aureus colonization, pathogenesis, and bacterial evasion of the host immune defenses. The net effect of Q2728755 decreases considerably with time because it receives negative regulation by S2748895 (at SAOUHSC_02982 ). Thus, by inhibiting the expression of S2748895 and/or promoting the expression of S2085718, S2773077, and S2773290, it is possible to amplify and maintain the genetic effect of Q2728755 in response to vancomycin. The expression of Q193712 detected in the abiotic GWAS experiment is promoted jointly by three SNPs and also inhibited, but to a lesser extent, by one SNP (Figs. S 4 C, S 6 ), leading the net effect of this QTL to be larger than its independent effect. One of the promotors is S550323 (at SAOUHSC_00544 ) that is proximal to the fibrinogen-binding protein SdrD gene encoding serine-aspartate repeat-containing protein D. This protein can promote the adhesion of bacteria to host cells, help resist the killing of innate immune components, such as neutrophils in the blood, and, thus, weaken bacterial clearance 89 . Q2783126 (detected from the abiotic GWAS experiment), located at gene SAOUHSC_03008 coding multiple phosphatase-related processes, is not expressed in its own capacity when its growth is at a linear stage (Fig. S 4C ; Fig. S 8 ), but because of accumulative up-regulation from many SNPs, it displays a tremendous net effect. When entering a stationary growth stage, Q2783126 is largely inhibited by S220967 (at SAOUHSC_00199 ) and S183439 (at SAOUHSC_00169 ), making its net effect below its increasing independent effect. As an inhibitor, SAOUHSC_00169 encodes ABC transporter permease that reduces intracellular antibiotic concentration through MDR proteins 90 , 91 , thereby leading to change in drug resistance. Some QTLs are significant because of both intrinsic capacity and extrinsic regulation. For example, Q183488 detected from the abiotic GWAS experiment exerts a large independent effect on vancomycin susceptibility (Fig. S 5C ) probably because of its proximity to gene SAOUHSC_00169 ( oppA ) that plays a critical role in antibiotic resistance through encoding peptide ABC transporter permeases 83 (Table S2 ). This QTL receives a mix of positive and negative regulation from three SNPs, making its net effect change with time in a cyclic pattern. SNP S213345, located at SAOUHSC_00192 encoding coagulase, promotes or inhibits the effect of Q183488, depending on stages of vancomycin response. Staphylocoagulase ( Coa ) is involved in the formation of biofilms 92 and can promote blood coagulation by activating prothrombin to convert fibrinogen into fibrin 93 . By changing the expression of regulators, the role of Q183488 in mediating vancomycin resistance can be maximized. QTL400804 detected from the biotic GWAS experiment is a QTL locate at gene SAOUHSC_00397 encoding type I restriction-modification system subunit M (Table S3 ). This QTL has a large independent genetic effect on S. aureus ’ response to the coexistence of E. coli , which is amplified by its altruistic Q407377 in a proximity to gene SAOUHSC_00405 encoding a hypothetical protein (Table S3 ; Fig. 7 ). Q407377 is strongly expressed in its own capacity and, meanwhile, receives up-regulation from S1456366, leading to its increasing net genetic effect, despite a negative regulation by the parasitic QTL400804. We provide a complete picture of how the QTLs detected by traditional approaches affect phenotypic plasticity directly or through indirect regulation from other genes. In each mechanism described above, the pattern, strength, sign, and number of regulations differ from QTL to QTL. A detailed understanding of these mechanisms is of great help to best utilize these QTLs for improving phenotypic plasticity, i.e., drug resistance or species interaction as described in this study, through marker-assisted selection or gene editing. In the past decades, GWAS has been criticized on the identification of only a small portion of genetic variance, leaving its main part as missing heritability 94 , 95 . There have been a body of literature on retrieving missing heritability by finding new genetic variants 96 . We can methodologically explore this issue through multilayer interactome networks. By dissecting the net genetic effect of these SNPs into their independent and dependent effects, our model characterizes novel interpretations. For example, S964411, located at gene SAOUHSC_00994 , codes bifunctional autolysin (Fig. S 9 ), but it is detected to be insignificant for vancomycin susceptibility by coFunMap (Fig. 5c ). However, this SNP would display a pronounced effect if it was in a socially isolated circumstance, because its insignificance is largely due to negative regulation by a non-coding S2201074. Although S964411 also receives positive regulation from a third SNP, this regulation does not adequately cancel the negative regulation. Thus, through inhibiting the expression of regulator S2201074, we expect that S964411 can fully execute its intrinsic capacity to deliver an impactful effect on S. aureus ’ response to vancomycin. Similarly, S784955 located at SAOUHSC_00802 encoding carboxylesterase is observed to be insignificant for S. aureus ’ reaction to E. coli, but it displays a large independent effect. Thus, if we silence the expression of its inhibitor, non-coding SNP595377, S784955 can greatly contribute to biotic defense. Taken together, those insignificant SNPs either with strong independent effects or regulated by promotors or inhibitors, or both, can be used to retrieve a certain amount of missing heritability is hidden in regulatory networks by uncoupling unfavorable regulation and/or strengthening favorable regulation. Despite its importance for trait control, epistasis is difficult to estimate without a considerable large sample size that is often hardly met in practice. For example, Zuk et al 34 . suggested that as many as 10,000 samples are required to estimate a genetic interaction in GWAS. However, as have been shown above, the characterization of dependent effects is not dependent on sample size but on net genetic effect curves. The estimation of main genetic effect curves does rely on sample size, but both simulation and experimental studies suggest that a sample size of 100 to 400 can reasonably well estimate such curves if the number of time points is two or more times the number of curve parameters 97 , 98 . Furthermore, traditionally defined epistasis is the genetic interaction between different loci, without knowing the direction and sign of epistasis. Our model can capture the full properties of epistasis, i.e., strength, signed, and direction by estimating dependent effects in our ODE model (equation (9)). Taken together, our model can detect any epistasis and identify its mechanism from any number of SNPs from a GWAS, but much less reliant upon sample size compared to traditional approaches. Our model integrates two emerging theories in quantitative genetics, i.e., omnigenic control and nonlinear gene–environment interactions 32 , 39 . By borrowing ingredients of various disciplines, these two theories are coalesced into a unified framework by which the genetic architecture of phenotypic variation and plasticity can be systemically characterized. However, our model has several limitations. First, it needs dynamic data of traits, which may not be widely available in ordinary general GWAS. Yet, Wu and Jiang’s 50 strategy for extracting dynamic information from static snapshots may be incorporated into our model to expand its use to static phenotypic data. Second, our model can only be used for a single phenotype, but the organism can be better described by its multifaced correlated features. However, how to reconstruct genetic networks for the phenotypic plasticity of multiple phenotypes is not a simple extension of the current model because these phenotypes may be interdependent of each other in a complex, likely nonlinear, manner. Third, our model only considers two environments. To extract the complete genetic mechanisms underlying how complex traits function, vary, and evolve along a spatiotemporal gradient, a series of environments should be sampled, but joint modeling of multiple environments cannot be made possible without the implementation of sophisticated spatial statistical methods. Despite these areas remaining to be explored, the precise mapping of SNP-SNP interactions identified by our model to the interactions between coding proteins, confirmed with two independent S. aureus GWAS experiments, implies its biological relevance and potential promise to be used as a general quantitative genetic approach at a new but higher standard. If this is a case, our model can likely make a paradigm shift in trait mapping from reductionist thinking of individual genes to a holistic approach that portrays a complete picture of genetic mechanisms."
} | 5,905 |
33972433 | PMC8157956 | pmc | 9,231 | {
"abstract": "Significance Multicellular bacterial communities called biofilms and pellicles significantly influence medical infections and industrial biofouling. Biofilms and pellicles often act as reservoirs of toxigenic bacteria. Here, we report a fractal wrinkling morphogenesis program that underlies pellicle formation in the model biofilm former and global pathogen, Vibrio cholerae . Pellicle morphogenesis is marked by the emergence of a cascade of self-similar structures with fractal-like scaling in wavelength, which increases surface area and presumably enhances nutrient transport and signaling among community members. This morphogenesis program can be altered by varying the spatial heterogeneity of the community. Thus, bacterial pellicles could provide tractable model systems to understand overarching principles driving morphogenesis and for engineering of functional soft biomaterials.",
"conclusion": "Conclusion In this study, we report the morphological progression of V . cholerae pellicles at a fluid–fluid interface. As a model soft biomaterial, a bacterial pellicle consists of biocomponents and living cells that display active metabolism and growth. In contrast to the merging of wrinkles into folds that occur in the shape evolution of smooth passive films on a fluid bath, we find that bacterial pellicles undergo a hierarchy of morphological transitions culminating in a cascade of wrinkles with increasingly smaller wavelengths. The origins of the structural complexity and the fractal scaling in surface dimensionality reside in the founding bacterial microcolonies that form the basic unit of the pellicle. We linked the dynamics of macroscopic morphological transitions directly to the microcolony structures and showed that distinct developmental modes can proceed depending on the inoculum seeding density. The basic elements of bacterial pellicle morphogenesis, such as cell growth, matrix production, the accumulation of mechanical stresses, and morphogenic transformations are ubiquitous in both prokaryotic and eukaryotic multicellular systems. For example, in eukaryotes, the folding of the gut, the wrinkling of skin, and the fractal branching of capillary blood vessels share many of the structural and mechanical features we observe here for bacterial pellicles. By controlling the matrix constituents, nutrient acquisition, signaling gradients, mechanics, or flow perturbations, bacterial biofilms and pellicles could provide tractable model systems to understand the overarching principles underlying morphogenesis and for engineering of functional soft biomaterials."
} | 646 |
39206688 | PMC11406059 | pmc | 9,233 | {
"abstract": "Abstract Microorganisms play vital roles in sulfur cycling through the oxidation of elemental sulfur and reduction of sulfite. These metabolisms are catalyzed by dissimilatory sulfite reductases (Dsr) functioning in either the reductive or reverse, oxidative direction. Dsr-mediated sulfite reduction is an ancient metabolism proposed to have fueled energy metabolism in some of Earth’s earliest microorganisms, whereas sulfur oxidation is believed to have evolved later in association with the widespread availability of oxygen on Earth. Organisms are generally believed to carry out either the reductive or oxidative pathway, yet organisms from diverse phyla have been discovered with gene combinations that implicate them in both pathways. A comprehensive investigation into the metabolisms of these phyla regarding Dsr is currently lacking. Here, we selected one of these phyla, the metabolically versatile candidate phylum SAR324, to study the ecology and evolution of Dsr-mediated metabolism. We confirmed that diverse SAR324 encode genes associated with reductive Dsr, oxidative Dsr, or both. Comparative analyses with other Dsr-encoding bacterial and archaeal phyla revealed that organisms encoding both reductive and oxidative Dsr proteins are constrained to a few phyla. Further, DsrAB sequences from genomes belonging to these phyla are phylogenetically positioned at the interface between well-defined oxidative and reductive bacterial clades. The phylogenetic context and dsr gene content in these organisms points to an evolutionary transition event that ultimately gave way to oxidative Dsr-mediated metabolism. Together, this research suggests that SAR324 and other phyla with mixed dsr gene content are associated with the evolution and origins of Dsr-mediated sulfur oxidation.",
"conclusion": "Conclusion As an ancient metabolism, sulfur cycling, including Dsr metabolism, is intricately connected to the redox state of Earth ( 8 , 37 ). During and post-oxygenation of Earth, microbial metabolism underwent evolution to adapt to the changing environment ( 98 ). The ability of certain bacteria belonging to transitionary phyla to exhibit remarkable metabolic diversity ( 65 , 67 , 99 ) suggests they may have adapted and shifted in response to evolutionary changes, making them intriguing subjects for further investigation from an evolutionary perspective. Previous studies have also pinpointed transitionary phyla as being subject to HGT ( 32 , 39 ). As HGT can be a major driver of evolution ( 100 ), this supports the idea that these phyla became adaptable in changing environments in Earth’s past. Unraveling the mechanisms behind this metabolic switch becomes particularly intriguing in the context of our evolving climate.",
"introduction": "Introduction Sulfur is pervasive in the Earth’s atmosphere, lithosphere, and hydrosphere ( 1–4 ). Sulfur is essential to life and comprises key components of amino acids, such as cysteine and methionine, organic compounds, and cofactors ( 5–7 ). The sulfur cycle constitutes a major biogeochemical cycle on our planet and actively interacts with other biogeochemical cycles, such as those of carbon, nitrogen, and iron ( 8–11 ). For example, transformations of sulfur compounds, such as sulfate, are closely associated with greenhouse gases, such as methane and carbon dioxide ( 12 ), which are vital to monitor in the face of climate change. Microorganisms play a critical role in the sulfur cycle, primarily by using sulfur compounds as electron donors or acceptors for energy metabolism ( 13–17 ). Nevertheless, the accumulation of some reduced sulfur compounds, such as hydrogen sulfide, can have toxic effects on environments ( 18 , 19 ), highlighting the significance of microbial sulfur metabolism. The dissimilatory reduction of sulfate to sulfide is a widespread and influential microbial metabolism, particularly in anoxic or low-oxygen environments ( 20–22 ). The sulfate reduction pathway ( Fig. 1A ) begins with the reduction of sulfate to adenosine phosphosulfate (APS) by sulfate adenylyltransferase (Sat) ( 23 , 24 ), followed by the reduction of APS to sulfite by adenosine 5′-phosphosulfate reductase (Apr) ( 25 ), with electrons provided by the quinone-interacting membrane-bound oxidoreductase (Qmo) complex ( 20 , 26 ). Finally, sulfite is reduced to sulfide by dissimilatory sulfite reductase (Dsr) proteins ( 27 , 28 ). While some microorganisms carry genes to complete every step of this pathway ( 29 , 30 ), data have shown that many microorganisms lack the genes for the complete reduction of sulfate to sulfide and possess genes for partial pathways ( 31 , 32 ). Enrichment studies have also revealed the existence of microorganisms that can produce sulfide in the presence of sulfite but not in the presence of sulfate ( 33 ). Observations such as these have led to hypotheses that microorganisms capable of partial biogeochemical pathways can interact with other microorganisms that complete different reactions, fueling biogeochemical cycling through community interactions ( 31 , 34 , 35 ). Therefore, even though the entire sulfate reduction pathway is of great interest, understanding each step in isolation can provide valuable insights. Figure 1 \n Simplified schematic representation of sulfate reduction and Dsr-mediated sulfur transformations. A) the sulfate reduction pathway (to sulfide) is depicted, with key enzymes/complexes involved at each step. The reaction highlighted in the manuscript, the reduction of sulfite to sulfide via Dsr, is enclosed within the dashed box. B) Overview of the forward and reverse Dsr-mediated reactions. Enzymes typically involved in each reaction, as discussed in text, are indicated. Figure generated with BioRender.com . The final step in the dissimilatory reduction of sulfate to sulfide, the catalyzation of sulfite to sulfide by Dsr proteins, is a pathway that has been given significant attention by the scientific community ( 32 , 36–40 ). In literature, the acronym “Dsr” is sometimes used to refer to the broader pathway of “dissimilatory sulfate reduction”; however, in this manuscript, the acronym Dsr will refer to “dissimilatory sulfite reductase” and the associated metabolic processes. Predicted to have evolved over 3 billion years ago, dsr genes are thought to have been present in some of Earth’s most ancient microbial lifeforms ( 36 , 37 ). In recent years, the number of microorganisms known to possess genes for Dsr-mediated metabolism has grown substantially ( 32 , 41 , 42 ), implying Dsr is influential in the numerous niches inhabited by these microorganisms. In the forward direction, Dsr proteins catalyze sulfite reduction. In contrast, Dsr proteins can also function in the reverse direction. Reverse-dissimilatory sulfite reductase can oxidize elemental sulfur to sulfite ( 43 , 44 ) and is a crucial part of sulfur oxidation metabolism. Therefore, this pathway has also been termed the oxidative Dsr pathway. Dsr is predicted to have an originally reductive function, with oxidative Dsr-mediated metabolism having evolved after ( 36 , 38 , 41 ). Numerous Dsr proteins exist, some of which have been associated with the direction of the sulfur transformation reaction that an organism catalyzes ( Fig. 1B ). Dsr proteins involved in both pathways include the core complex DsrAB ( 45 , 46 ) and DsrC, which acts as a physiological partner and “redox hub” in Dsr activity ( 28 , 47 ). DsrMK, which are often present in conjunction with DsrJOP, are crucial for both pathways and play a role in electron transport ( 48 , 49 ). Proteins associated with directionality of the reaction include DsrD which has recently been identified as an allosteric activator of DsrAB sulfite reductase activity and is associated with the reductive pathway ( 32 , 50 ). DsrD is deemed to have a nonessential function and is absent in some sulfate-reducing archaea ( 32 , 50 ). DsrEFH have been shown to act as a sulfur carrier to DsrC and are associated with microbial sulfur oxidation ( 32 , 51 ). DsrL possesses oxidoreductase activity and is thought to be essential for sulfur oxidation ( 52 , 53 ). Other dsr genes have also been identified and include dsrR , dsrS , and dsrT ( 32 , 54 ). Phylogenetic analyses, along with gene content, have also been used to assess Dsr pathway directionality, with three main divisions, archaeal reductive, bacterial reductive, and bacterial oxidative, typically being described ( 32 , 39 , 41 , 43 ). Likewise, sulfur-reducing microorganisms (SRMs) and sulfur-oxidizing microorganisms (SOMs) are often discussed as separate groups ( 9 , 55 ). Common examples of SRMs, including those possessing dsr genes, are organisms from phyla Desulfobacterota ( Deltaproteobacteria and Thermodesulfobacteria ), Bacillota ( Firmicutes ), Nitrospirota , Halobacteriota ( Euryarchaeota ), and Thermoproteota ( Crenarchaeota ) ( 39 , 41 , 56 , 57 ). Examples of SOMs, including those that have oxidative dsr genes, are organisms from the phyla Bacteroidota ( Chlorobi ) and Pseudomonadota ( Proteobacteria ) ( 39 , 41 , 58–61 ). Recent research has challenged traditional classifications of sulfur-metabolizing organisms, revealing organisms from diverse phyla with both reductive and oxidative dsr genes in the same genome ( 29 , 32 , 38 , 39 , 62 , 63 ). For example, though thought to be essential for, and therefore associated with sulfur oxidization, dsrL has been described in organisms with otherwise reductive-looking metabolisms ( 29 , 63 ). Moreover, a function for DsrL in sulfite reduction has been elucidated ( 62 ). The presence of dsrD and dsrEFH genes in the same genome has been noted in various phyla like SAR324, Actinomycetota, and Nitrospirota ( 32 , 38 , 39 ). Additionally, multiple copies of genes encoding for DsrAB, one suggested to function in the oxidative direction, and the other in the reductive, have been reported in phyla including Ca . CG2-30-53-67 (hereafter CG2-30-53-67) and Desulfobacterota ( 39 , 62 ). Reductive and oxidative gene combinations in the same genome have made it difficult to discern whether organisms possessing these genomic repertoires participate in the reductive or oxidative Dsr pathway. It has been suggested these organisms have the potential to switch between reductive and oxidative pathways ( 32 , 39 , 62 ). Alternatively, the incomplete nature of many genomes, recovered as Metagenome-Assembled Genomes (MAGs) and Single Cell-Amplified Genomes (SAGs), has also raised concerns about misbinning or misassembly as an explanation for unique gene combinations ( 39 ). With only a handful of these genomes being described in the literature, it has remained unclear if these unique gene combinations are a biological phenomenon (as opposed to misassembly in the case of MAGs and SAGs) and, if so, why these organisms might carry them. We delved further into the exploration of genomes encoding both reductive and oxidative Dsr proteins, zeroing in on an intriguing phylum: Candidate phylum SAR324 hereafter referred to as SAR324. Thriving in diverse environments, SAR324’s metabolically versatile nature adds complexity to its role in sulfur cycling ( 31 , 64–67 ). Not only do select SAR324 genomes exhibit unique dsr gene combinations ( 32 , 39 ), but SAR324 dsrA sequences have been found to stand apart phylogenetically from other bacterial sulfur oxidizers and reducers ( 65 ). These distinctive features make SAR324 a compelling subject for unraveling the intricacies of Dsr-mediated sulfur transformation. In this study, we undertake a comprehensive analysis of Dsr-mediated metabolism in SAR324, leveraging all publicly available SAR324 genomes. We describe both Dsr and oxidative Dsr pathways in SAR324 and shed light on the unique phylogenetic position of SAR324 DsrAB sequences positioned between oxidative and reductive types. Our findings reveal genomic and phylogenetic patterns in SAR324 that extend to related phyla, linking these organisms to the evolution of oxidative Dsr in organisms from the well-known sulfur-oxidizing phylum, Bacteroidota . Finally, we reveal undescribed combinations of reductive and oxidative genes in genomes of these same lineages.",
"discussion": "Discussion Conducting comparative genomic studies is vital for pinpointing key contributors to global biogeochemical cycles. As interest in utilizing sulfur-metabolizing bacteria for industrial and bioremediation applications grows ( 16 , 55 , 56 ), identifying versatile players within the sulfur cycle becomes crucial. With a significant portion of bacteria remaining uncultured ( 96 ), large-scale metagenomics-based investigations are urgently needed to uncover essential targets for cultivation and experimentation. This study successfully achieves this objective by highlighting the uncultured candidate phylum SAR324 as a significant participant in sulfur cycling. Specifically, we demonstrate the widespread global and phylogenetic distribution of Dsr-encoding SAR324, suggesting influential metabolic impacts across diverse ecological niches. The conclusions drawn here are significant because, although attention has been paid to the evolution of reductive Dsr ( 36 , 38 , 97 ), less is known about the transition to the oxidative type. Through a more in-depth examination of SAR324’s dsr gene content and phylogenetic position, we have identified SAR324 as a potential key player in the evolution of novel sulfur metabolism, specifically in the transition to oxidative Dsr-mediated sulfur metabolism. Specifically, we associate SAR324 with the evolution of Dsr in well-known sulfur-oxidizing phyla Bacteroidota . Additionally, by demonstrating the patterns seen in SAR324 (i.e. the unique phylogenetic position in a DsrAB phylogenetic tree and select genomes carrying genes for both oxidative and reductive Dsr), we showed that members from related phyla, such as Nitrospirota , Nitrospinota , Actinomycetota , and CG2-30-53-67, also are likely important in the evolution of the oxidative Dsr pathway. The proposed evolutionary scenario outlined here would explain, from an evolutionary perspective, phenomena regarding Dsr-encoding organisms that have previously puzzled researchers. For example, genomes with a mixed dsr gene repertoire can be explained as being remnants of the evolutionary transition event. DsrL being retained in genomes with otherwise reductive-looking dsr repertoires can be explained by the fact that these genomes gained genes for DsrL in the evolutionary transition event and maintained dsrL as it ended up serving a role in reductive Dsr metabolism. The presence of chimeric type dsr genes in Bacteroidota and the presence of dsrT in Bacteroidota can also be explained by Bacteroidota being related to transitionary phyla. Further studies should focus on the ecological drivers behind the split into these three different groups. It should also be noted that an evolutionary scenario for the acquisition of dsr genes by Bacteroidota , similar to the one described here, has previously been briefly suggested to be a possibility ( 63 ). The analyses presented here offer expanded data to support these hypotheses. In addition to narrowing down the phyla associated with the evolution of oxidative Dsr, these findings have implications for future genomics studies. The complexity revealed here underscores the need for thorough metabolic analyses to understand biogeochemical processes. Unanswered questions persist, necessitating follow-up studies on microorganisms in transitionary phyla to unravel their metabolic potential. Addressing these questions will contribute to a deeper understanding of the selective pressures governing microbial sulfur metabolism. Additionally, continued sequencing of microbial communities from diverse environments is essential, especially for phyla like CG2-30-53-67 and Actinomycetota , where only a few genomes with dsr were discovered."
} | 4,002 |
24903088 | PMC4047535 | pmc | 9,237 | {
"abstract": "Ocean acidification driven by rising levels of CO 2 impairs calcification, threatening coral reef growth. Predicting how corals respond to CO 2 requires a better understanding of how calcification is controlled. Here we show how spatial variations in the pH of the internal calcifying fluid (pH cf ) in coral ( Stylophora pistillata) colonies correlates with differential sensitivity of calcification to acidification. Coral apexes had the highest pH cf and experienced the smallest changes in pH cf in response to acidification. Lateral growth was associated with lower pH cf and greater changes with acidification. Calcification showed a pattern similar to pH cf , with lateral growth being more strongly affected by acidification than apical. Regulation of pH cf is therefore spatially variable within a coral and critical to determining the sensitivity of calcification to ocean acidification.",
"discussion": "Discussion Our estimates of the effects of ocean acidification on the extension of branch tips (apical growth) are similar to those of other studies on S. pistillata colonies 9 30 31 , being fairly stable over a wide range of pH T conditions ( Fig. 1b ). In contrast, extension of the lateral part of the colonies (lateral growth) was more strongly affected by ocean acidification, similar to other studies that have examined lateral growth in S. pistillata 30 . Boron isotope-based estimates of pH cf exhibited patterns similar to calcification data. Lateral growth, which showed the greatest decline in calcification in response to acidification also showed a greater decline in pH cf in response to acidification than adjacent apical growth. Thus spatial differences in the regulation of pH cf may account for changes in calcification, consistent with the IpHRAC 19 model in which internal pH cf regulation controls abiotic calcification rates. Variation in the ability of the overlying tissue layer to control the pH at the site of calcification, as suggested by pH cf estimates ( Fig. 2 ), may be linked to a number of differences existing between the apically and basal laterally growing regions 32 33 . The growth of these two regions is fundamentally different: apical growth occurs in a largely unrestricted environment whereas basal lateral growth occurs at an interface between the coral and a substrate. Thus the growing edge (initial basal lateral growth) faces a number of potential challenges not generally encountered by the rest of the coral tissue, including competing with other organisms for substrate, and isolating new substrates from the surrounding environment to allow crystal growth to occur. The devotion of resources to compete for substrate could limit the energy available for calcification and in-turn reduce the ability of the tissue to up-regulate pH when faced with a more acidified environment. Or the isolation of new substrates from the surrounding environment may not be as complete as the isolation of existing skeletal regions, thus allowing higher rates of seawater ingress resulting in a more pronounced effect of acidification. Regardless of the underlying mechanism(s), pH cf and thus calcification, was more strongly affected by acidification in the basal, laterally growing regions than the apical growth. Thus stages of coral growth that involve extension over new substrates are likely to be more strongly affected by ocean acidification. This may be particularly relevant to the larval-stage of coral growth when all calcification is occurring on a new substrate, to coral fragments that must cement themselves to a new substrate, as well as to damaged corals attempting to re-grow over exposed skeleton; all represent growth stages in which lateral growth over a substrate plays an important role. These stages are therefore least likely to be able to maintain pH cf under acidified conditions and thus are likely to be more adversely affected by ocean acidification. Since calcification within a coral differs spatially in its sensitivity to ocean acidification, and that variations in pH cf appear to correspond to these differences in calcification, pH cf may help to predict how calcification will respond to ocean acidification. Measurements of pH cf thus represent an important tool for identifying stages of coral growth (e.g. colonization of new substrates) and particular species that will most likely be adversely affected by ocean acidification. Measurements of pH cf using boron isotopes can allow pH cf to be estimated over long time scales and allow variations in biologically controlled pH up-regulation to be linked to events (e.g. bleaching, storms, etc) in the natural environment which may further impact the ability of corals to regulate pH cf . Collectively such data can help to better predict how corals will respond to the range of conditions they face."
} | 1,212 |
39231844 | PMC11489214 | pmc | 9,238 | {
"abstract": "Salt marshes represent a unique ecosystem at the marine-terrestrial boundary of shallow protected coastlines. Microarthropods form an essential component of soil food webs, but how they colonize new intertidal habitats is little understood. By establishing two experimental systems without animals, we investigated microarthropod colonization (1) at the seashore from the pioneer zone to the lower and upper salt marsh and (2) at the same tidal height on artificial islands 500 m from the seashore. Potential source populations of microarthropods in the respective zones were also investigated. Colonization of microarthropods after 5 years was consistently faster on the seashore than on the artificial islands. Collembola and Mesostigmata colonized all the zones both on the seashore and on the artificial islands, with colonization being faster in the upper salt marsh and in the pioneer zone than in the lower salt marsh. Oribatida colonized the new habitats on the seashore, but only little on the artificial islands. Variations in species composition were more pronounced between salt marsh zones than between experimental systems, indicating that local environmental conditions (i.e., inundation frequency) are more important for the assembly of microarthropod communities than the distance from source populations (i.e., dispersal processes). Variations in community body size of Oribatida and Mesostigmata indicated environmental filtering of traits, with smaller species suffering from frequent inundations. Notably, Mesostigmata most successfully colonized the new habitats across salt marsh zones on both systems. Overall, the results document major mechanisms of colonization of intertidal habitats by microarthropods with different life histories and feeding strategies. Supplementary Information The online version contains supplementary material available at 10.1007/s00442-024-05615-x.",
"conclusion": "Conclusions Our unique experimental systems allowed new insights into the patterns and processes of colonization of intertidal salt marshes by microarthropods, and comparisons between the importance of dispersal and body size-related filtering processes. Differences in the colonization patterns of microarthropod taxa between salt marsh zones indicate stronger dispersal limitation in Oribatida than in Collembola and Mesostigmata. Greater variation in species composition between salt marsh zones than between experimental systems indicates that local environmental conditions (i.e., inundation frequency) are more important for microarthropod community assembly than distance (i.e., dispersal) from source populations. This is further supported by the increase in CWM body size of Oribatida and Mesostigmata from the USM to the PZ. Overall, the colonization processes of intertidal habitats by microarthropods are mainly driven by environmental filtering in addition to dispersal, with the latter varying between taxa.",
"introduction": "Introduction Colonization involves the dispersal and establishment of organisms in new habitats. This fundamental process shapes the composition of ecological communities and biodiversity across landscapes. Colonization success is a product of two community assembly processes, dispersal and environmental filtering, with the former being a prerequisite for the latter. Colonization has been extensively studied in different habitats, including freshwater (Anderson and Smith 2004 ; Tavares et al. 2008 ; Incagnone et al. 2015 ), deep-sea (Grassle and Morse-Porteous 1987 ; Kelly et al. 2007 ; Priede and Froese 2013 ) and terrestrial ecosystems (Marsh et al. 2004 ; Bröring and Wiegleb 2005 ; Rosenberger et al. 2017 ). A wide range of taxa have been studied, including microbes (Wynn-Williams 1990 ; Vieira et al. 2020 ; Malard and Pearce 2022 ), plants (Wang et al. 2011 ; Lõhmus et al. 2014 ; García-Cervigón et al. 2017 ), vertebrates (Boyd and Pletscher 1999 ; Kerth and Petit 2005 ; Banaszek et al. 2010 ) and insects (Harrison 1989 ; Neve et al. 1996 ; Solbreck et al. 2023 ). However, information on colonization and community assembly processes of one of the most dynamic habitats, the interface between marine and terrestrial ecosystems, i.e. salt marshes, has received little attention. Salt marshes occur at the marine-terrestrial interface along protected shallow coastlines throughout the world’s oceans. These unique habitats are characterized by halophytic plants (Allen 2000 ) and are influenced by regular tidal inundations (i.e., submergence due to water coverage during high tide). Formed by sediment accretion, salt marshes provide a gradient of bank elevation with associated differences in inundation frequency (Roozen and Westhoff 1985 ; Bockelmann et al. 2002 ; Caçador et al. 2007 ). Abiotic environmental conditions shape the system into distinct vegetation zones, such as the pioneer zone (PZ), the lower salt marsh (LSM) and the upper salt marsh (USM) (Petersen et al. 2014 ). Regular exposure to water level changes and episodic storms, especially in winter, make salt marshes a heterogeneous dynamic system. The sediment matrix, with its pore spaces, aggregates and organic matter, serves as an initial colonization site for terrestrial organisms. Plant and animal communities assemble along this dynamic gradient (Haynert et al. 2017 ; Winter et al. 2018 ; Lõhmus 2020 ). In salt marsh systems, microarthropods with body sizes of a few millimetres (Karg 1993 ; Hopkin 1997 ; Weigmann 2006 ) are highly diverse and abundant (Polderman 1974 ; Sterzynska and Ehrnsberger 2000 ; Haynert et al. 2017 ). They play a critical role in ecosystem functions such as decomposition of organic matter, release of nutrients, formation of soil structure and stabilization of organic carbon (Rusek 1998 ; Grandy et al. 2016 ; Trentini et al. 2018 ). The dominant taxa of microarthropod communities in salt marshes are springtails (Collembola) and mites (Acari: Oribatida and Mesostigmata) (Polderman 1974 ; Sterzynska and Ehrnsberger 2000 ; Haynert et al. 2017 ). Collembola and Oribatida are decomposers that regulate microbial biomass and community composition, thereby indirectly influencing ecosystem functions (Griffiths and Bardgett 1997 ; Erktan et al. 2020a , b ; Erktan et al. 2021 ). In food webs, Collembola contribute to the propagation of basal resources to higher trophic levels (Terborgh and Estes 2010 ; Sofo 2020 ). By contrast, Mesostigmata are dominant predators in salt marsh habitats and are likely to stabilize food webs against disturbance (Haynert et al. 2017 ). Collembola generally comprise r-selected taxa that disperse and reproduce rapidly (Petersen 2002 ), whereas Oribatida predominantly comprise K-selected taxa that develop and reproduce slowly (Crossley 1977 ; Maraun et al. 2003 ; Minor and Cianciolo 2007 ). The colonization of terrestrial habitats by Collembola, Oribatida and Mesostigmata typically follows successional changes and contributes to the formation of terrestrial ecosystems, which is particularly evident along the marine-terrestrial boundary (Scheu and Schulz 1996 ; Perez et al. 2013 ; Haynert et al. 2017 ). The colonization of salt marsh sediments by these microarthropod taxa is likely to play a crucial role in forming salt marsh food webs and stabilizing animal communities against disturbances. Colonization of terrestrial habitats by microarthropods is a complex and dynamic process influenced by many factors, including individual dispersal ability, distance from source habitats and local environmental conditions. In salt marshes, microarthropod density and diversity have been shown to be reduced by tidal inundation (Haynert et al. 2017 ). Environmental conditions such as salinity, oxygen and food availability can influence colonization processes. However, different taxa may have specific tolerance ranges or traits for these environmental factors, resulting in different colonization patterns. Furthermore, as most microarthropods rely on passive water or wind dispersal (Hawes et al. 2007 ; Lehmitz et al. 2011 ; Schuppenhauer et al. 2019 ), smaller taxa may disperse longer distances than larger ones due to the lower mass and slower landing speed of the former (Jung and Croft 2001 ). Body size is therefore likely to be an important trait for microarthropod colonization of salt marsh habitats. However, few studies have investigated microarthropod colonization processes, with most studies conducted in the laboratory (Coleman and MacFadyen 1966 ; Hågvar and Abrahamsen 1980 ; Chauvat and Ponge 2002 ) and only few in the field (Dunger et al. 2002 ; Wanner and Dunger 2002 ; Stefano et al. 2007 ). Results suggest that microarthropods typically appear shortly after new land is exposed and act as foundation taxa in these habitats, facilitating vegetation succession and colonization by other organisms (Hodkinson et al. 2004 ). However, information on microarthropod colonization processes in the highly dynamic marine-terrestrial interface of salt marshes is limited. Our study fills this gap by investigating species composition, colonization success and body size of three dominant microarthropod taxa with different life histories and feeding strategies, i.e. Collembola, Oribatida and Mesostigmata, in salt marsh habitats of the Wadden Sea of Spiekeroog, Germany. To understand colonization processes in salt marsh habitats with different exposure to tidal inundation and distance from the source habitats, two experimental systems were established: (1) new mudflat sediments devoid of microarthropods exposed across salt marsh zones of different inundation frequency, i.e. the USM, LSM and PZ, on Spiekeroog Island, and (2) artificial islands erected in the tidal flats about 500 m from the seashore, where the same experimental design was established (i.e., new mudflat sediment exposed at the level of the USM, LSM and PZ). We compared the colonization patterns of the three microarthropod taxa in the different salt marsh zones between the two experimental systems, allowing us to differentiate between short- and long-distance dispersal (i.e., effects of dispersal limitation), while the different inundation frequencies of the USM, LSM and PZ allowed us to compare the strength of environmental filtering by inundation between these taxa. We hypothesized that (1) microarthropod colonization of newly exposed mudflat sediments is faster on Spiekeroog Island than on artificial islands; (2) assuming that long-distance dispersal is mainly due to wind rather than water, microarthropods of the less inundated USM disperse most easily; (3) differences in microarthropod community composition are more pronounced between salt marsh zones than between experimental systems. In addition, we hypothesized that (4) microarthropod colonization success is mediated by body size, but the pattern varies between taxa.",
"discussion": "Discussions We studied the colonization patterns of mudflat sediments by Collembola, Oribatida and Mesostigmata in experimental salt marsh systems consisting of natural seashore (Spiekeroog Island) and artificial habitats (man-made islands about 500 m from the seashore). We hypothesized that (1) microarthropod colonization of newly exposed mudflat sediments is faster in the SI system than in the AI system, (2) species of the less frequently inundated USM disperse most easily, (3) variation in microarthropod community composition is more pronounced between salt marsh zones than between experimental systems, and (4) colonization success varies between microarthropod taxa and is related to body size. Mudflat sediment colonization by microarthropods was consistently faster and more complete in the SI than in the AI system, regardless of microarthropod taxa, supporting the first hypothesis. The experimental mudflat plots in the SI system were surrounded by intact salt marsh, which served as a source habitat, allowing microarthropods to colonize the mudflat material by both active movement and passive dispersal by tide and wind. By contrast, the artificial islands were placed in the Wadden Sea tidal flats about 500 m from the source habitats on Spiekeroog Island. Active dispersal of flightless microarthropods from the source habitats to the AI system is unlikely; rather, they are passively dispersed over large distances by tides, wind and birds. Floating in water and rafting on plant debris or wrack during tidal surges may facilitate long-distance dispersal of microarthropods, as demonstrated for Oribatida across distant oceanic islands or even continents (Schatz 1991 , 1998 ; Schuppenhauer et al. 2019 ; Lindo 2020 ). Microarthropods are also dispersed by wind, allowing them to colonize new habitats (Lindroth 1973 ; Thornton et al. 1988 ; Schneider et al. 2011 ). In addition, phoresy on birds is a way for Oribatida to bridge long distances (Krivolutsky and Lebedeva 2002 ; Krivolutsky et al. 2004 ; Lebedeva 2012 ). Although birds heavily colonize salt marshes (Dierschke 2002 ; Dierschke and Bairlein 2004 ), it is unlikely that they transported Oribatida to the AI system as no Oribatida were found in the USM. The results suggest that long-distance dispersal is relatively slow, but could mediate local colonization of new intertidal habitats. Of the three microarthropod taxa, Oribatida were the least effective colonizers in both the SI and AI systems. In particular in the AI system, colonization by passive dispersal of Oribatida was less pronounced than that of Collembola and Mesostigmata, suggesting that Oribatida require longer periods of time (more than 5 years in our study) to colonize new intertidal habitats. The less effective colonization of the SI system by Oribatida compared to the other two taxa further indicates slow active dispersal of Oribatida. They move in the range of about 5 cm per day (Ojala and Huhta 2001 ; Lehmitz et al. 2012; Cameron et al. 2013 ), whereas Collembola can actively bridge distances > 3 m per day (Hågvar 1995 ; Zettel et al. 2000 ; Åström and Bengtsson 2011 ; Ponge 2020 ). Similarly, Mesostigmata are surface-living taxa that actively hunt for prey and may bridge long distances by active movement. Phoresy via ground beetles is also likely to contribute to the rapid colonization of Mesostigmata in both the SI and AI systems (Gwiazdowicz and Gutowski 2012 ). Furthermore, the high mobility and surface-living of Collembola and Mesostigmata suggests that they are more susceptible to displacement by water and wind than Oribatida which more intensively colonize the soil. To drift on the water surface during passive dispersal, many Collembola species are osmoconformers and can survive at high salinity (Witteveen et al. 1987 ; Witteveen and Joosse 1988 ). Many Mesostigmata species can also survive prolonged submersion. By contrast, most intertidal Oribatida species have large tarsal claws, allowing them to withstand tidal forces by clinging to plants or other substrates (Pfingstl 2023 ) presumably contributing to low dispersal by water. Although Oribatida can be dispersed up to a few hundred meters by wind, most wind dispersed species are arboreal (Lehmitz et al. 2011 ). Very limited wind dispersal of Oribatida is supported by the fact that they did not colonize the USM in the AI system. The lower colonization of the PZ by Oribatida in the AI system, however, may also have been due to their lower density in the source habitats on Spiekeroog Island than that of the other two taxa. Microarthropod colonization was faster in the USM and PZ than in the LSM for both SI and AI systems, supporting our second hypothesis. The USM is similar to terrestrial systems and is only inundated during storm events. Microarthropods in the USM can intensively and actively colonize adjacent habitats. Microarthropods in the USM are also more intensively exposed to wind and are likely to be passively dispersed by wind rather than water. By contrast, the PZ resembles aquatic habitats with frequent inundations, which limits active movement of microarthropods but increases passive dispersal by water. The general decrease in Oribatida density from the USM to the LSM to the PZ reflects the increased disturbance by tides, which fill the habitat by salt water. In our study, only few Oribatida species colonized the most frequently inundated PZ, such as Ameronothrus schneideri , Zachvatkinibates quadrivertex and Liebstadia similis . These species are halobionts able to tolerate salinity and survive submersion (Schulte 1973 ; Polderman 1974 ; Pfingstl 2013 ). More species of Collembola and Mesostigmata than Oribatida in the PZ indicate a greater number of species in the former groups that can cope with saline conditions and frequent inundations. Most of these Collembola ( Archisotoma besselsi , Halisotoma maritima , Thalassaphorura debilis , Mesaphorura krausbaueri , Mesaphorura macrochaeta and Parisotoma notabilis ) and Mesostigmata species ( Dendrolaelaps halophilus , Cheiroseius necorniger , Vulgarogamasus cf . trouessarti and Gaeolaelaps praesternalis ) are semi-aquatic, salt-tolerant and trophic specialists (Salmane 2000; Haynert et al. 2017 ). In support of our third hypothesis, microarthropod community compositions generally varied more between salt marsh zones than between experimental systems, although variation was best explained by both factors. Mesostigmata species compositions varied significantly between salt marsh zones and between experimental systems. Collembola communities overlapped most between experimental systems across salt marsh zones, with more heterogeneous composition in the AI system than in the SI and control systems. Community composition of Oribatida varied significantly between salt marsh zones and, as indicated by only few records, colonization of the artificial islands by Oribatida was very limited. These results suggest that differences in local environmental conditions due to tidal inundation frequency (suggesting habitat filtering processes) between salt marsh zones are more important than geographical distance to source populations (suggesting dispersal limitation) for the assembly of salt marsh microarthropod communities, particularly for Oribatida. As the frequency of tidal inundation increases from the USM towards the PZ, the pore space in the mudflat matrix in the twice-daily inundated PZ remains largely filled with water likely affecting the abundance and composition of microarthropod communities (Zaitsev and Wolters 2006 ; Nielsen et al. 2008 ; Erktan et al. 2020a , b ). Limited pore space is likely to more strongly affect small-sized taxa colonizing soil pores than larger taxa predominantly colonizing the soil surface. Supporting this assumption, the CWM body size of Oribatida and Mesostigmata increased from the USM to the PZ, although the CWM body size was not significantly correlated with colonization success. This suggests that inundation frequency favours larger individuals after they have colonized new habitats. In contrast to Oribatida and Mesostigmata, Collembola CWM body size did not vary between salt marsh zones, suggesting that body size is not a key trait for Collembola in their post-colonization processes. At our study sites, most Collembola species are larger than mites and live predominantly on the surface rather in soil pores, and this likely contributed to their high density in the PZ. Collembola may use resources on the soil surface more efficiently than the other two taxa (Sterzynska and Ehrnsberger 2000 ; Mertens et al. 2007 ; Haynert et al. 2017 ). Further, they may recover faster from flood-related mortality due to rapid reproduction (Adis and Junk 2002 ; González‐Macé and Scheu 2018 ) and the production of flood-tolerant eggs (Tamm 1984 ; Marx et al. 2012 ). In addition, Collembola are physiologically adapted to waterlogged habitats and can adjust their metabolism under anoxia (Zinkler and Platthaeus 1996 ). They may also trap air bubbles as a potential behavioural trait (Zinkler et al. 1999 ). The observed difference in CWM body size between the two mite taxa and Collembola reflect differences in trait-based assembly processes after colonization in salt marshes, partially supporting our fourth hypothesis. Colonization by microarthropods varied significantly between zones of the experimental salt marsh systems and was higher in the SI than in the AI system. This suggests that both distance from the source habitats and changes in the local environment related to inundation frequency structure microarthropod communities. In terms of community assembly processes, dispersal limitation (suggested by the distance of the AI system from the seashore) is most pronounced in Oribatida, less in Mesostigmata and least in Collembola. However, for all the three taxa, post-colonization filtering processes (mainly by inundation) are more important than dispersal per se."
} | 5,234 |
25024744 | PMC4094786 | pmc | 9,239 | {
"abstract": "Biosynthesis of liquid fuels and biomass-based building block chemicals from microorganisms have been regarded as a competitive alternative route to traditional. Zymomonas mobilis possesses a number of desirable characteristics for its special Entner-Doudoroff pathway, which makes it an ideal platform for both metabolic engineering and commercial-scale production of desirable bio-products as the same as Escherichia coli and Saccharomyces cerevisiae based on consideration of future biomass biorefinery. Z. mobilis has been studied extensively on both fundamental and applied level, which will provide a basis for industrial biotechnology in the future. Furthermore, metabolic engineering of Z. mobilis for enhancing bio-ethanol production from biomass resources has been significantly promoted by different methods (i.e. mutagenesis, adaptive laboratory evolution, specific gene knock-out, and metabolic engineering). In addition, the feasibility of representative metabolites, i.e. sorbitol, bionic acid, levan, succinic acid, isobutanol, and isobutanol produced by Z. mobilis and the strategies for strain improvements are also discussed or highlighted in this paper. Moreover, this review will present some guidelines for future developments in the bio-based chemical production using Z. mobilis as a novel industrial platform for future biofineries.",
"conclusion": "Conclusions Based on the previous and our reviews, Z. mobilis is firstly being developed as an effective ethanologenic by engineering strain improvement, including utilization of xylose and arabinose in addition to glucose. Undoubtedly, Z. mobilis has showed desirable characteristics for its special metabolic pathway. The scientific and technological progress of Z. mobilis have also made a significant contribution to the bioethanol industry. Compared with E. coli , Z. mobilis has high restriction-modification enzyme activity, and cannot be contaminated by bacteriphages [ 13 ]. It is fairly osmo-tolerant and can hence tolerate very high sugar concentrations, which is an advatange in fermentation in a high-sugar medium. Its smaller genome and simple metabolic pathway, also lead to less byproducts formation. On the other hand, its desirable characteristics will also make it a novel platform for future biorefineries, which will make a significant contribution to green or sustainable chemistry (as shown in Figure 5 ). Figure 5 General process of fuel or chemical production by \n Z. mobilis \n . Although extensive studies, such as general genetic tools, strategies of metabolic engineering, value-added bio-product production, genomic and transcriptomic, et cetera, have been developed in Z. mobilis since 1980s, non-commercialization of the Zymomonas process for ethanol production from sugar, starch-based or lignocellulosic biomass has developed successfully. Moreover, an increased range of higher-value product generation has also been restricted by its fundamental research. Especially, it is more difficult to engineere Z. mobilis than E. coli or yeast. Despite the extensive studies on general genetic tools and omics data available for Z. mobilis , it is necessary to further develop advanced technologies that can be used in metabolic engineering. Therefore, to realize the industrial potential of Z. mobilis for future biorefineries, considerable efforts should be focused on the following points in the future: developing universal tools for deletion of several genes in one round, controlling metabolic flux and optimizing regulatory networks to improve the yield of desired products, and developing a highly express system, et cetera; these novel technologies are necessary for further strain improvement or redirection of the metabolic pathway for fuel and chemical production. Moreover, different systems of metabolic engineering approaches are becoming powerful tools in developing engineered E. coli or S. cerevisiae [ 3 , 5 ], which should also be highlighted in engineered Z. mobilis strains. In particular, other biotechnological approaches, such as genome sequencing, functional genomics, genome engineering and omics will also provide a basis for pathway or genome reconstruction to improve its fitness and robustness for environmental stress [ 160 , 161 ]. Representitive biotechnologies, such as CRISPR/Cas systems [ 162 ], site-specific recombinases [ 163 , 164 ], genome shuffling [ 165 ], global transcription machinery engineering (gTME) [ 166 ], and Zinc-finger nucleases [ 167 ], which will also be used for enhancing cellular traits of Z. mobilils. Presumably, their potential will be further implemented with a promising future in developing or optimizing the metabolic pathway for the production of fuels as well as commodity and specialty chemicals.",
"introduction": "Introduction There have been growing concerns about biosynthesis of fuels, desired chemicals and materials from renewable biomass resources for limited fossil resources and associated environmental issues in the past few decades [ 1 , 2 ]. As model industrial or laboratory organisms, Escherichia coli and Saccharomyces cerevisiae were selected as important platforms for the purpose of desired biofuels and chemicals production via metabolic engineering [ 3 - 5 ]. Currently, strain optimization to utilize various feedstocks (for example, starch, sugarcane, agricultural residues, industrial waste, forest residues, energy crops, et cetera) [ 6 , 7 ], desired products spectrum (for example, biofuels and building block chemicals), and higher yields, which have made great progress in the past decades and provided a basis for industrial applications [ 1 - 5 ]. As a candidate bio-ethanol producer, Zymomonas mobilis showed some advantages, for example, higher specific rate of sugar uptake, high ethanol yield, lower biomass production, non-requirement of controlled addition of oxygen during fermentation, et cetera [ 8 - 13 ]. Extensive fundamental studies on Z. mobilis over the last 30 years have also made this strain a promising ethanologenic organism for large-scale bio-ethanol production. On the other hand, extensive studies on different genetic techniques (including plasmid vector, expression system, transposon system, gene knockout, gene transformation, and gene function, et cetera) will help Z. mobilis are amenability to genetic improvement for industrial biotechnology [ 13 ]. Furthermore, strategies of strain improvement (such as conventional mutagenesis, transposon mutagenesis, adaptive laboratory evolution, and metabolic pathway engineering, et cetera), and different value-added bio-products have also been paid more and more attention in the past 20 years. Importantly, genomics and transcriptomic of Z. mobilis have also been developed since 2005, which will aid future metabolic engineering and synthetic biology in strain improvement for industrial applications [ 14 ]. Selected milestones in Z. mobilis research are summarized in Figure 1 . Figure 1 Selected milestones in \n Z. mobilis \n research. Currently, three subspecies (subsp.) of Z. mobilis have been found, including Z. mobilis subsp. mobilis , Z. mobilis subsp. pomaceae and Z. mobilis subsp. Francensis [ 15 - 19 ]. All strains have also been summarized in the Ph D thesis of So Lok-yan (University of Hong Kong) and other review articles [ 19 ]. Among these strains, ATCC 31821 (ZM4), ATCC 10988 (ZM1), ATCC29191 (ZM6), CP4, and NCIMB 11163 from Z. mobilis subsp. mobilis , ATCC 29192 from Z. mobilis subsp. pomaceae , which were well-charcterized by previous studies on the level of physiology, biochemical, fermentation, genetics, metabolism, and omics . These strains are regarded as a model organism in Z. mobilis research or industrial applications. In general, Z. mobilis may play a critical role as a novel platform in industrial biotechnology for the development of a green replacement for petrochemical products. In this paper, we review some critical research progress on Z. mobilis for its use as a platform for the production of ethanol and other buck chemicals from biomass."
} | 2,029 |
34073010 | PMC8227142 | pmc | 9,240 | {
"abstract": "This study reports for the first time the preparation of an electrospun microfibrous mat of PIM-EA-TB. The electrospinning was carried out using a chloroform/n-Propyl-lactate (n-PL) binary solvent system with different chloroform/nPL ratios, in order to control the morphology of the microfibres. With pure chloroform, porous and dumbbell shape fibres were obtained whereas, with the addition on n-PL, circular and thinner fibres have been produced due to the higher boiling point and the higher conductivity of n-PL. The electrospinning process conditions were investigated to evaluate their impact on the fibres’ morphology. These microfibrous mats presented potential to be used as breathable/waterproof materials, with a pore diameter of 11 μm, an air resistance of 25.10 −7 m −1 and water breakthrough pressure of 50 mBar.",
"conclusion": "4. Conclusions This paper reports for the first time the fabrication of PIM-EA-TB fibres with controlled morphology using a chloroform/n-Propyl lactate solvent solution with different volume ratios. Different surface morphologies (porous surface), shapes (cross-sectional dumbbell shape) and diameter were obtained according to the solvent system due to their physical properties (conductivity, boiling point). Firstly, the use of a volatile solvent, such as chloroform, induced the formation of a dumbbell shape and a surface porosity. With the addition of n-PL, which presents a higher boiling point, circular fibres were obtained. Secondly, formation of thick fibre was observed with low conductive solvent systems, with diameter up to 7.9 μm. With the addition of a more conductive solvent, a decrease in the diameter fibres was noticed, with diameter down to 4.7 μm. In terms of performances, the fibrous mats obtained with pure chloroform presented smaller pore size, higher air and water resistance than with n-PL due to the dumbbell shape.",
"introduction": "1. Introduction Electrospinning is a straightforward method to produce self-standing microfibre membranes presenting high porosity and pore size ranging from ten nanometres to several micrometres [ 1 ]. Microfibres are defined as a continuous filament with an average fibre diameter of 25 microns or smaller and have been mainly used for wastewater treatment [ 2 , 3 ], smart responsive surface [ 4 , 5 ], and bioengineering application [ 6 ]. They present larger pores which can allow or facilitate cellular infiltration and/or diffusion of nutrients in vitro culture [ 7 ]. The generation of fibres is based on the formation of a jet from a charged polymeric system under an electrical field. The solvent evaporation and the stretching of the jet, caused by the repulsive forces of the charged molecules within the jet, are responsible for the formation of the polymer fibres [ 8 , 9 , 10 , 11 , 12 ]. Thus, the final mat fibrous morphology depends on the polymer–solvent system (such as solvent nature, viscosity, conductivity) and on the electrospinning process parameters (such as feed flow rate, voltage, distance between the tip and the collector). Therefore, various nanostructures, from beads to bead-free fibres with different fibre diameter can be produced by tuning these parameters. Polymers of Intrinsic Microporosity (PIMs) are a class of macromolecule which have generated considerable interest in the field of gas separation [ 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 ], hydrogen storage [ 21 , 22 ], sensors [ 23 ] or liquid separation [ 22 , 24 ] thanks to a high surface area (typically 300–1500 m 2 /g) and interconnected micropores (<2 nm) [ 15 ]. Another advantage of PIMs is their solution processability in common organic solvents which allow the optimisation of the macroscopic formation of the microporous polymer, for example as electrospun fibres, coatings on woven fabrics, etc. Among others, research on electrospun PIMs fibres, mainly on PIM-1 and on modified PIM-1, has been gaining momentum in recent years [ 6 , 12 , 25 , 26 , 27 , 28 ]. Recently, a new type of PIM, derived from ethanoanthracene and Tröger’s base (PIM-EA-TB), has been developed using a polymerisation reaction based on the formation of the bridged bicyclic diamine called Tröger’s base (TB: 6H,12H-5,11-methanodibenzo[b,f][1,5]diazocine) [ 14 , 29 ]. PIM-EA-TB contains only benzene rings fused together via rigid bridged bicyclic units composed of TB and ethanoanthracene EA and demonstrates an apparent higher BET surface area than PIM-1 [ 14 ]. Its high rigidity results in enhanced gas separation performance with a superior vapour sorption capability [ 15 ]. Moreover, thanks to its basicity with the Tröger’s base, PIM-EA-TB can act as an ion-conducting material and can be used to construct ionic diodes [ 30 ]. To the best of our knowledge, no research on the electrospinning of PIM-EA-TB has been published. Here, we report for the first time the fabrication of PIM-EA-TB fibres with controlled morphology by an electrospinning process using a binary solvent system. The aim of this study is to investigate the impact of the different solvent systems properties and the electrospinning process conditions (feed flow rate, voltage, distance between tip and collector) on the surface morphology and diameter of electrospun PIM-EA-TB fibres. The resulting fibres were further characterised to assess their potential for application in terms of waterproof resistance, breathability and air permeability."
} | 1,345 |
36671654 | PMC9854848 | pmc | 9,241 | {
"abstract": "Ammonia is an important chemical that is widely used in fertilizer applications as well as in the steel, chemical, textile, and pharmaceutical industries, which has attracted attention as a potential fuel. Thus, approaches to achieve sustainable ammonia production have attracted considerable attention. In particular, biological approaches are important for achieving a sustainable society because they can produce ammonia under mild conditions with minimal environmental impact compared with chemical methods. For example, nitrogen fixation by nitrogenase in heterogeneous hosts and ammonia production from food waste using microorganisms have been developed. In addition, crop production using nitrogen-fixing bacteria has been considered as a potential approach to achieving a sustainable ammonia economy. This review describes previous research on biological ammonia production and provides insights into achieving a sustainable society.",
"conclusion": "4. Conclusions This review mainly focused on ammonia production using genetic engineering methods ( Table 1 ). Energy production with low environmental burden is considered essential for the formation of a sustainable society. In addition to studies in this review, for example, anaerobic culture is also one of the methods currently in practical use for energy production from food wastes [ 118 ]. These anaerobic cultures usually do not use genetic engineering methods, such that this method has room for optimization by using genetic engineering introduced in this review. To achieve a sustainable society, we have to solve some problems other than the development of biological methods such as the storage of food waste [ 119 , 120 ]. In addition, consumption of ammonia is important for creating sustainable environment. For example, ammonia is only 30–50% utilized for crop growth [ 121 ]. The remainder eutrophicates the oceans, contaminates drinking water, and, in some areas, changes the composition of vegetation and encourages the propagation of non-native species [ 122 , 123 , 124 , 125 ]. Ammonia is converted to N 2 O by microorganisms in the environment, which is also a well-known greenhouse gas, and its emission is also a problem [ 74 , 126 ]. Microorganisms such as E. coli and yeast contain nitrogen in their bodies, with about 20% of the total amount as proteins [ 127 ]. The nitrogen in cells comes from assimilation of amino acids or proteins in a medium. Therefore, the amount of produced nitrogen must be larger than the input nitrogen for effective biofuel production. To solve this problem, cell surface engineered yeast could be repeatedly used for production because the enzyme on the yeast cell surface is stable [ 128 ]. Recently, some studies attempted to estimate the metabolism of microorganisms so that it might be possible to calculate the potential of microorganisms to efficiently produce ammonia in the future [ 129 , 130 , 131 ]. In this review, we introduced various attempts to improve the Haber–Bosch process, one of the sources of carbon dioxide generation, which uses the largest amount of fossil fuels, in order to realize a sustainable society. The use of genetically optimized microorganisms to produce ammonia instead of the Haber–Bosch process is a potential and environmentally friendly approach to solve the global ammonia demand problem and to develop a sustainable carbon-free society ( Table 2 ). Therefore, in the future, it may be possible to produce ammonia more efficiently by using it concurrently with such methodologies.",
"introduction": "1. Introduction In recent years, rapid population growth and industrial development have increased the use of several fossil fuels and have increased the amount of waste and environmental pollutants, leading to destruction of the global environment and causing global warming, ocean acidification, and so on [ 1 ]. In alleviating this situation, researchers, governments, companies, and other organizations are greatly interested in improving the global environment, and the Sustainable Development Goals have been set forth [ 2 , 3 ]. Fossil fuels are the most significant energy source in the modern world, which are converted to electrical energy primarily through thermal power generation. Various natural energy sources have been used as alternatives to fossil fuels, including wind, hydro, solar, and geothermal energy [ 4 , 5 , 6 ]. In addition to these renewables, biomass as fuel has also received increasing attention. Biomass includes perennial plants, forestry waste, algae, and food waste (municipal solid waste), which are also recognized as carbon-neutral fuels. However, the supply of biomass is greatly affected by weather, time, location, and economics, making it unstable [ 7 , 8 , 9 , 10 , 11 , 12 , 13 ]. In addition, biodegradable materials could be used as sources of biofuels, and ammonia is expected to be used as a biofuel ( Figure 1 ) [ 14 ]. 1.1. Industrial Uses and the Need for Ammonia Ammonia is an important compound in a variety of industries [ 15 ]. Fixed nitrogen, such as ammonia, is essential for crop growth, and increasing the amount of nitrogen circulating on the planet allows for population growth [ 16 ]. The production of ammonia is highly important to sustain life, and about 80% of ammonia is used in the production of fertilizers. Ammonia is also used for refrigerant gas or the synthesis of various chemicals such as plastics, explosives (trinitrotoluene, nitroglycerin, and nitrocellulose), textiles (rayon and nylon), agricultural chemicals, dyes (for cotton, wool, silk, etc.), and so on [ 17 , 18 , 19 ]. In recent years, ammonia-fueled batteries have also been devised [ 18 , 20 , 21 , 22 , 23 , 24 ]. The potential application of ammonia as fuel has received increasing attention [ 8 , 25 , 26 , 27 , 28 , 29 ]. Hydrogen has low transport efficiency because of its low volumetric energy density (3 W h·L −1 ) and higher calorific value per liquid wight (141.9 MJ/kg) [ 30 ]. Therefore, approaches to converting hydrogen to a more transport-efficient substance have been investigated. Ammonia has a flammable range of 16% to 25% ( v / v ), which can be transported more safely than hydrogen [ 31 , 32 ]. Liquid ammonia contains hydrogen atoms per volume and energy density (MJ L −1 ) that are 1.7 and 1.5 times higher than that of liquid hydrogen, respectively [ 33 , 34 ]. Hydrogen has a low boiling point of −253 °C, and it requires considerable energy to liquefy. By contrast, ammonia has a boiling point of −33.4 °C, and it can be easily liquefied by using general-purpose refrigeration equipment and can be handled easily. Thus, it has been considered for use as a carrier for hydrogen in a hydrogen society [ 32 , 35 , 36 , 37 ]. In addition, the hurdle to the industrial application of ammonia is lower than that of hydrogen because the infrastructure for storage and transportation of ammonia has already been established. 1.2. Chemical Method for Ammonia Production The Haber–Bosch process is a typical example of ammonia nitrogen fixation, and 55% of the world’s ammonia is produced by this method [ 38 ]. This method requires the cleavage of the triple bond of the nitrogen molecule, which uses a large amount of energy [ 15 , 39 ]. The energy used by the Haber–Bosch process is equivalent to 2–3% of the world’s annual fossil fuel use, and it accounts for 1.4% of the global annual carbon dioxide emissions [ 38 , 40 ]. Carbon dioxide is a well-known greenhouse gas (GHG), and its reduction is strongly desired because of environmental issues of growing concern in recent years. Therefore, the Haber–Bosch process has been improved upon in recent years, particularly in catalyst improvements. Moreover, the development of catalysts that allow reactions to proceed under conditions closer to ambient temperature and pressure has been considered. Compared with iron catalysts, Ru-based catalysts enable the fixation of ammonia at lower pressure (90 atm), and they have about 20 times higher catalytic efficiency [ 41 ]. However, ruthenium has become increasingly expensive in the last decade, thereby hindering its industrial use [ 42 , 43 , 44 ]. The performance of catalysts has been greatly studied and further developed, including the development of several molybdenum-based catalysts that mimic the active center of nitrogenase (a nitrogen-fixing enzyme) for the synthesis of ammonia at ambient temperature and for pressure in microorganisms [ 45 , 46 ]. At present, most of the hydrogen required by the Haber–Bosch process is obtained by electrolysis. In general, the operation of a water electrolysis unit requires a continuous supply of highly purified water. Furthermore, 9 tons of highly purified water is required to produce 1 ton of hydrogen. Based on these data, 233.6 million tons of water per year is required to produce 1 ton of ammonia using hydrogen obtained from water electrolysis [ 47 ]. With the progression of global warming, improving the global environment is necessary [ 2 ]. Therefore, attempts are being made to supply ammonia in a sustainable manner using natural energy (green energy) such as wind and solar power generation [ 48 , 49 ]. Ammonia using hydrogen produced from green energy is known as “green ammonia” [ 50 ]. In addition to wood biomass, the amount of food waste has continued to grow excessively in recent years, particularly in developed countries [ 51 , 52 ]. Food waste produces considerable amounts of GHGs and environmental pollutants through landfilling and incineration [ 53 ]. Thus, many attempts have been made to obtain energy from food waste [ 13 ]. In particular, okara is an abundant food waste, and its various uses are being considered [ 54 , 55 ]. Attempts to produce ammonia from food waste by physicochemical methods have also been studied [ 56 , 57 , 58 ]. For example, considering that glutamic acid is a common source of nitrogen in food wastes, researchers have targeted glutamic acid contained in sewage sludge and meat and bone meal, and they succeeded in producing ammonia with 35–51% efficiency at 800 °C and 0.5–1.0 g g −1 carbon content by automated gasification [ 56 ]. This reaction is dependent on the catalytic metal ions in the food waste, which may need to be optimized for each feedstock. The addition of LaFeO 3 as a catalyst resulted in an efficiency of 54 vol% [ 57 ]. Therefore, this approach could potentially produce 10% of the ammonia used in Europe [ 56 ]. Several studies have developed sustainable methods to produce ammonia, most of which include green power generation methods, chemical synthesis methods, and sustainable ammonia production using biological methods [ 59 , 60 , 61 ]. Various methods of recovering ammonia from food wastes have also been used, and research is progressing toward practical application [ 62 ]. Alternatively, we focus on sustainable ammonia production methods such as biological methods—bacterial and yeast methods—and explore their application potential ( Figure 2 )."
} | 2,745 |
29062956 | PMC5625732 | pmc | 9,243 | {
"abstract": "The genomic era has revolutionized research on secondary metabolites and bioinformatics methods have in recent years revived the antibiotic discovery process after decades with only few new active molecules being identified. New computational tools are driven by genomics and metabolomics analysis, and enables rapid identification of novel secondary metabolites. To translate this increased discovery rate into industrial exploitation, it is necessary to integrate secondary metabolite pathways in the metabolic engineering process. In this review, we will describe the novel advances in discovery of secondary metabolites produced by filamentous fungi, highlight the utilization of genome-scale metabolic models (GEMs) in the design of fungal cell factories for the production of secondary metabolites and review strategies for optimizing secondary metabolite production through the construction of high yielding platform cell factories.",
"introduction": "1 Introduction Microbial secondary metabolites are widely exploited for their biological activities to ensure the well-being of humans. Secondary metabolites are used as antibiotics, other medicinals, toxins, pesticides, and animal and plant growth factors [1] . Although the antibiotic effects of certain molds have been reported earlier, it was Flemings' persistence in the usability of the antimicrobial activity of penicillin, which initiated what is known as the golden era of antibiotic discovery [2] . Despite the fungal origin of penicillin, produced by several members of the Penicillium genus [3] , most research on secondary metabolites has focused on bacteria, mainly soil isolates of actinomycetes with the majority of compounds originating from the Streptomyces genus [4] . Some of the pioneering work that paved the way for antibiotic discovery was conducted by Nobel laureate Selman Waksman, who's systematic screening of Streptomyces isolates, led to the identification of several antibiotics, including streptomycin and neomycin which have found extensive applications in the treatment of infectious diseases. However, to ensure translation of these findings for commercial production it was necessary with further product optimization and fermentation characterization of microbial physiology, and this resulted in the birth of industrial microbiology as a discipline, with Arnold Demain as one of the founding fathers. Today we know that although most living organisms can produce secondary metabolites, the ability to produce them is unevenly distributed. Among all known microbial antibiotics and similar bioactive compounds (altogether 22,500), 45% are from actinomycetes, 38% are from fungi and 17% are from unicellular bacteria [4] . Among this wealth of compounds, only about a hundred are in practical use for human therapy, with the majority being derived from actinomycetes [4] . However, it is worth mentioning that in addition to penicillin, several other fungal secondary metabolites have successfully reached the pharmaceutical market, including cholesterol lowering statins [5] , the antifungal griseofulvin [6] and the immunosuppressant mycophenolic acid [7] . Biosynthesis of secondary metabolites takes place from a limited number of precursor metabolites from the primary metabolism ( Fig. 1 ). In fungi, these precursors are mainly short chain carboxylic acids (e.g. acetyl-CoA) or amino acids, which are linked together by backbone enzymes such as polyketide synthases (PKSs), non-ribosomal peptide synthetases (NRPSs), dimethylallyl tryptophan synthetases (DMATSs) or terpene cyclases (TCs). The resulting oligomers are then subject to chemical modification by tailoring enzymes which are often controlled under common transcriptional regulation as the backbone enzyme [8] . A hallmark trait of the genes involved in a secondary metabolite pathway is that they, in most record cases, physically cluster in the chromosome in biosynthetic gene clusters (BGCs) [9] . Fig. 1 Biosynthesis of secondary metabolites from precursors of the central carbon metabolism. PPP: Pentose Phosphate Pathway. ETC: Electron Transport Chain. TCA: Tricarboxylic Acid. AAs: Amino Acids. Fig. 1 The characteristic clustering of genes as well as the conserved motifs of backbone genes can be exploited for computational detection of BGCs from sequence data. Tools like SMURF [10] , antiSMASH [11] , PRISM [12] and SMIPS/CASSIS [13] utilize these features to reliably and with a high accuracy detect BGCs of known compound classes in fungi. Other algorithms detects BGCs without relying on specific motifs or the presence of backbone genes, which enables identification of BGCs beyond PKS, NRPS, DMATS and TCs [14] , [15] , [16] , [17] . Tools and implementations of BGC mining algorithms have been extensively reviewed [18] , [19] , [20] , [21] , [22] , [23] . A limitation of secondary metabolite production is the low yields that are naturally achieved in most microbes, partly since many secondary metabolites are favored under suboptimal growth conditions [8] , [24] and because their biosynthesis compete with essential pathways of metabolism, involved in growth related processes ( Fig. 1 ). Applying metabolic engineering to circumvent these limitations can be greatly assisted by utilization of the mathematical representation of metabolism in genome-scale metabolic models (GEMs), which concepts and applications have been reviewed elswhere [25] , [26] , [27] . These models, however, often neglect secondary metabolite biosynthesis, hence their potential in studying secondary metabolism has not been fully tapped. Additionally, with the efficient gene editing tool CRISPR-Cas9 being developed for a number of fungal model organisms [28] , [29] , [30] , a great potential exists for implementing the necessary genetic modifications for the development of improved secondary metabolite producers. In this review, we will describe methods for linking BGCs to compounds and show how metabolic modeling can aid in translating the improved secondary metabolite discovery rate into metabolic engineering strategies for the development of fungal platform strains for the production of secondary metabolites."
} | 1,552 |
36876066 | PMC9978112 | pmc | 9,244 | {
"abstract": "Microbial ammonia oxidation is the first and usually rate limiting step in nitrification and is therefore an important step in the global nitrogen cycle. Ammonia-oxidizing archaea (AOA) play an important role in nitrification. Here, we report a comprehensive analysis of biomass productivity and the physiological response of Nitrososphaera viennensis to different ammonium and carbon dioxide (CO 2 ) concentrations aiming to understand the interplay between ammonia oxidation and CO 2 fixation of N. viennensis . The experiments were performed in closed batch in serum bottles as well as in batch, fed-batch, and continuous culture in bioreactors. A reduced specific growth rate (μ) of N. viennensis was observed in batch systems in bioreactors. By increasing CO 2 gassing μ could be increased to rates comparable to that of closed batch systems. Furthermore, at a high dilution rate ( D ) in continuous culture (≥ 0.7 of μ max ) the biomass to ammonium yield (Y (X/NH3) ) increased up to 81.7% compared to batch cultures. In continuous culture, biofilm formation at higher D prevented the determination of D crit . Due to changes in Y (X/NH3) and due to biofilm, nitrite concentration becomes an unreliable proxy for the cell number in continuous cultures at D towards μ max . Furthermore, the obscure nature of the archaeal ammonia oxidation prevents an interpretation in the context of Monod kinetics and thus the determination of K S . Our findings indicate that the physiological response of N. viennensis might be regulated with different enzymatic make-ups, according to the ammonium catalysis rate. We reveal novel insights into the physiology of N. viennensis that are important for biomass production and the biomass yield of AOA. Moreover, our study has implications to the field of archaea biology and microbial ecology by showing that bioprocess technology and quantitative analysis can be applied to decipher environmental factors affecting the physiology and productivity of AOA.",
"introduction": "Introduction Nitrification is the oxidation of ammonia (NH 3 ) to nitrate (NO 3 − ) via the intermediate nitrite (NO 2 − ). Ammonia oxidation is the first and usually rate limiting step in nitrification and is therefore important for the global nitrogen cycle. For over a century ammonia oxidation was thought to be performed by the few bacterial genera Nitrosomonas , Nitrosococcus , and Nitrosospira, until 20 years ago when evidence started to accumulate that archaea might play an important role in this process as well ( Venter et al., 2004 ; Könneke et al., 2005 ; Treusch et al., 2005 ). Nitrosopumilus maritimus was the first isolate from a marine aquarium ( Könneke et al., 2005 ; Qin et al., 2017 ) followed by Nitrososphaera viennensis ( Tourna et al., 2011 ; Stieglmeier et al., 2014 ) from garden soil. Since then multiple isolates and enrichments of ammonia oxidizing archaea (AOA) have been established and characterized to further improve our understanding of these ubiquitously abundant organisms ( De La Torre et al., 2008 ; Blainey et al., 2011 ; Lehtovirta-Morley et al., 2011 , 2016 ; Lebedeva et al., 2013 ; Jung et al., 2014 ; Zhalnina et al., 2014 ; Santoro et al., 2015 ; Qin et al., 2017 ; Sauder et al., 2017 , 2018 ; Abby et al., 2018 ; Daebeler et al., 2018 ; Alves et al., 2019 ; Bayer et al., 2019 ; Nakagawa et al., 2021 ). Viable habitats include oceanic crust ( Nunoura et al., 2013 ; Jørgensen and Zhao, 2016 ; Zhao et al., 2020 ), deep sea sediments ( Francis et al., 2005 ; Park et al., 2008 ; Nunoura et al., 2018 ; Vuillemin et al., 2019 ; Zhao et al., 2019 ; Kerou et al., 2021 ), marine water column ( Qin et al., 2017 ; Bayer et al., 2019 ; Santoro, 2019 ), oxygen minimum zones ( Bristow et al., 2016 ), various kinds of soils ( Lehtovirta-Morley et al., 2011 , 2016 ; Tourna et al., 2011 ; Jung et al., 2014 ; Zhalnina et al., 2014 ; Alves et al., 2019 ), fresh water ecosystems ( French et al., 2012 , 2021 ; Sauder et al., 2018 ), waste water treatment plants ( Mußmann et al., 2011 ; Sauder et al., 2017 ), terrestrial hot springs ( De La Torre et al., 2008 ; Reigstad et al., 2008 ; Dodsworth et al., 2011 ; Abby et al., 2018 ; Daebeler et al., 2018 ; Luo et al., 2020 ), and human skin ( Probst et al., 2013 ; Moissl-Eichinger et al., 2017 ). AOA outnumber their bacterial counterparts, ammonia oxidizing bacteria (AOB), by orders of magnitude in many habitats ( Karner et al., 2001 ; Leininger et al., 2006 ; Adair and Schwartz, 2008 ; Nicol et al., 2008 ; Hollibaugh et al., 2011 ), but their contribution to the nitrification process is still not completely resolved. While all known AOA and AOB are confined to oxidize NH 3 to NO 2 − , another group of bacteria was recently identified which are able to perform the complete oxidation of NH 3 to NO 3 − , thus given the name Comammox ( Daims et al., 2015 ; van Kessel et al., 2015 ). In bacteria, NH 3 is oxidized to hydroxylamine (NH 2 OH) by the enzyme ammonia monooxygenase (AMO) ( Hollocher et al., 1981 ; Hyman and Wood, 1985 ) which is then further oxidized to nitric oxide (NO) by the hydroxylamine oxidoreductase (HAO) ( Hooper and Terry, 1979 ; Caranto and Lancaster, 2017 ). The enzyme responsible for the oxidation of NO to NO 2 − is still unknown. Unlike the bacterial NH 3 oxidation pathway, the archaeal one is still very poorly characterized. Only the oxidation of NH 3 to NH 2 OH is inferred to be catalyzed by an AMO ( Vajrala et al., 2013 ) which is very distantly related to bacterial AMO and all other enzymes of the copper membrane monooxygenase superfamily (CuMMO) ( Alves et al., 2018 ). However, a counterpart to the bacterial HAO is still missing in archaea. Two models are currently proposed, one that mimics the bacterial pathway ( Lancaster et al., 2018 ) and another one where ammonia is oxidized to hydroxylamine which would then be further oxidized to NO 2 − with NO as a co-substrate ( Kozlowski et al., 2016 ). NO would be produced by the reduction of NO 2 − by a proposed nitrite reductase (NirK), which is highly expressed in most AOA, but whose role is still ambiguous as AOA do not perform nitrifier denitrification, unlike AOB ( Wrage-Mönnig et al., 2018 ). One important factor in niche differentiation of organisms is their substrate affinity, which is described either as reaction rate ( v ) based on the K m value ( Eq. 1 ) or as specific growth rate (μ) based on the K S value ( Eq. 2 ). In a steady state the residual substrate concentration ( S ) remains constant over time. \n Eq. 1 \n v = v max · S K m + S \n \n Eq. 2 \n μ = μ max · S K S + S \n AOA are notoriously difficult to grow and produce only very little biomass and therefore most information about AOA is provided in the form of the apparent substrate affinity ( K m (app) ), which is based on whole cell activity measurements of molecular oxygen (O 2 ) consumption in micro-respiratory chambers ( Kits et al., 2017 ; Jung et al., 2022 ). Most AOA are considered oligotrophs and their K m (app) values range from 0.1–1 μmol L −1 NH 3 and NH 4 + for marine strains, 1–10 μmol L −1 for soil or thermophilic strains to 0.1–10 mmol L −1 for the soil clade Nitrosocosmicus. Comammox have K m (app) in the range of 0.1–10 μmol L −1 NH 3 and NH 4 + and are thus considered also oligotrophs while AOB are rather considered eutrophs with K m (app) between 0.1 and 100 mmol L −1 NH 3 and NH 4 + ( Jung et al., 2022 ). In an attempt to measure the K S value of a fresh water AOA enrichment in batch cultures, no effect of the initial substrate concentration on μ was observed, suggesting that the K S is much lower than the lowest tested concentration of about 15 μmol L −1 NH 4 + ( French et al., 2021 ). In a chemostat experiment, N. maritimus was grown with 150 μmol L −1 NH 4 + and different dilution rates ( D ) of 0.010, 0.023 and 0.032 h −1 (μ max is 0.036 h −1 ) to investigate the influence of μ on the lipid composition of the organism. NO 2 − concentration only varied by a maximum of 7% ( Hurley et al., 2016 ) but no information was given about the residual substrate concentration. Nitrososphaera viennensis was isolated from garden soil and grows optimal at 42°C ( Stieglmeier et al., 2014 ) with the addition of pyruvate to scavenge reactive oxygen species (ROS) that are endogenously produced ( Kim et al., 2016 ). Cell concentrations and μ are usually approximated by NO 2 − concentrations, as they have been shown to correlate well ( Tourna et al., 2011 ; Stieglmeier et al., 2014 ). However, N. viennensis produces far too little biomass to measure optical density or dry cell weight. As the different forms of cultivation systems are rarely discussed in the field of nitrification (except for waste water treatment plants), a short overview of the most commonly used systems shall be given here. In general, systems can be distinguished by the level of which they allow energy and matter to be transferred through them. Closed systems (transfer of energy but not matter) and open systems (transfer of energy and matter) are the extreme cases of reality and the isolated system (no transfer of energy or matter) serves as an important theoretical construct that can only be asymptotically approached by closed systems. Cultivation systems are characterized by the level of transfer of matter and in analogy, closed batch (e.g., serum flask with rubber stopper) is a closed system ( Taubner and Rittmann, 2016 ; Mauerhofer et al., 2019 ), continuous culture (e.g., bioreactor with gassing and in- and outflow of medium) is an open system, with the openness of the system depending on the transfer rates. Open batch (e.g., Erlenmeyer flask with cotton plug, bioreactor with gassing) and fed-batch (e.g., bioreactor with gassing and inflow of medium) are open system intermediates in between the two extremes ( Mauerhofer et al., 2019 ). Batch systems are probably the most common cultivation systems used in microbiology, because they are very easy to set up and require little technological infrastructure. Closed batch is usually only used if an atmosphere different from air is required ( Mauerhofer et al., 2019 ; Hanišáková et al., 2022 ). A major disadvantage of batch systems are the changing substrate concentrations that lead to a very heterogeneous biomass which can complicate analysis. Fed-batch systems consist of a shorter batch phase followed by a feed phase, where usually a concentrated feed medium is used to increase the biomass concentration but at the same time avoid substrate inhibition. By using an exponential feeding strategy μ can be kept constant and a relatively homogeneous biomass can be produced, as long as no product inhibition or other limitations hamper growth. In continuous culture systems a stable flow of medium is maintained after an initial batch phase and by changing the dilution rate ( D , Eq. 3 ), different steady states can be established by changing the flow rate ( Q ) or the volume of the culture ( V ). \n Eq. 3 \n D = Q V A system is usually considered to be in a steady state after five volume exchanges (99.3% of medium exchanged) while all parameters are kept constant. As a result, the produced biomass is very homogeneous, because in a steady state μ is equal to D . Technically speaking, productivities of organisms that are grown in continuous culture are up to tenfold higher compared to batch systems ( Herbert et al., 1956 ), because the system can be stably operated near μ max and downtime for disassembly, sterilization and reassembly of the bioreactor becomes increasingly negligible with longer operation times. Continuous cultures are also excellent tools to study the physiology of microorganisms due to the high level of control and the extended periods of time a steady state can be maintained. A typical application is the determination of the K S value by varying D and measuring the residual substrate concentration ( S ), thereby relating μ to S and thus allowing to calculate K S when assuming Monod kinetics. However, biofilm formation, genetic adaptation or other factors might flaw the determination of such values with extended process runtimes. In stirred tank reactors gas is usually supplied by a fumigation tube at the bottom of the reactor. The gas transfer rate, which is often a limiting factor for fast growing organisms, can be increased by, e.g., the gassing rate, stirrer speed and operating pressure ( Rittmann et al., 2018 ; Pappenreiter et al., 2019 ). Due to the low μ max and biomass concentration ( X ) of AOA the gas transfer rate is not a limiting factor but rather needs to be considered because of the possibility to strip important metabolic intermediates (like NO) from the system. All cultivated AOA are chemolithoautotrophs and fix CO 2 by a modified version of the hydroxypropionate/hydroxybutyrate cycle ( Könneke et al., 2014 ). For the cultivation of AOA and AOB in bioreactors CO 2 is usually supplied by air and NaHCO 3 that is used to titrate the pH and act as C-source ( Hurley et al., 2016 ; Sedlacek et al., 2020 ; French et al., 2021 ). Given the fast reaction kinetics of the aqueous carbonate system, the liquid phase in a gassed reactor will tend to be in equilibrium with the supplied gas mix independently of NaHCO 3 titration ( Wang et al., 2010 ). The aim of this study was to investigate the growth behavior of N. viennensis in continuous culture systems to develop a biomass production process that would enable biochemical studies of the organism to eventually elucidate the energy metabolism of AOA. Growth conditions were optimized to ensure that only NH 3 acts as a limiting substrate while characterizing process parameters such as K S , biomass to substrate yield (Y (X/NH3) ) and the critical dilution rate ( D crit ) at which the organism will be washed out, for the bioprocess development. To further increase the biomass productivity per reactor volume we also established a continuous culture at a higher substrate concentration.",
"discussion": "Discussion The substrate affinity, together with μ, are very important parameters for understanding the ecological strategy of an organism ( r - and k -strategists). Due to their very high substrate affinity and low μ, AOA are regarded as typical k-strategists but the growth behavior of N. viennensis observed in this study can not sufficiently be described with Monod kinetics or more sophisticated models like the Briggs-Haldane model. The strong increase of S at D = 0.035 h −1 would indicate that D was already close to μ max , but then S should have increased right after the start of the continuous culture and further increases of D should have resulted in even stronger increases of S . Instead S plateaued around 800 μmol L −1 NH 4 + while the cell number decreased with D increasing up to 0.046 h −1 . An explanation for this behavior would be the formation of a biofilm that would retain cells in the reactor, but at the same time decrease planktonic cell concentrations. Thus, the activity of the whole system would increase while at the same time the measured cell concentrations would decrease, which describes the results. The stable S concentrations and sposntaneous but isochronal activity increases are difficult to integrate into this explanation. Only very strong increases of D like from 0.050 h −1 to 0.060 h −1 and further to 0.070 h −1 caused S to increase temporarily while cell concentrations remained surprisingly stable – considering a μ max of 0.048 h −1 for N. viennensis . At D > μ max cells should usually be washed out over time, but due to the seeding of cells by the biofilm this phenomenon was not observed. The linear decrease of S at D = 0.060 h −1 might be the result of an increase in active biofilm biomass, that reaches its limits at D = 0.070 h −1 . This would explain the very stable concentrations of S after the spontaneous activity increase. These increases of activity might be induced by quorum sensing and be dependent on cell density in the biofilm, which should gradually increase with time until a maximum is reached. This would provide a robust principle, which could explain the synchronistic nature of this very unusual physiological phenomenon. The mechanistic principle of how the cells are able to increase their substrate affinity abruptly could be explained by the expression of multiple amo C genes, of which the genome of N. viennensis contains six, which is likely to be the subunit that contains the catalytic center of the AMO protein complex. This assumption is based on recent findings of a bacterial particulate methane monooxygenase ( Ross et al., 2019 ). In addition, there seems to be another regulatory element in the growth behavior of N. viennensis that was observed with 2 mmol L −1 and 10 mmol L −1 NH 4 + continuous cultures. At a certain D the organism consumes NH 3 at a slightly lower rate than provided, thus slowly increasing S until it finally stabilizes at roughly 50% of S i . Once stabilized, S only marginally increases with D unless very strong changes are induced. This phenomenon appears to be linked to an increase in Y (X/NH3) , which seems to be highly elevated at D > 0.030 h −1 . However, we must state that we analyze the biomass productivity in relation to the energy metabolism, assuming that energy is limiting the biomass formation and not carbon, which was provided in excess to N. viennensis during chemostat and biofilm formation experiments. In a steady state Y (X/NH3) might be increased because the enzymatic machinery and metabolic networks can be fine tuned to a stable environmental condition. A batch culture needs to adapt constantly to the changing environment which requires energy. However, this phenomenon occurs in every organism and can not describe the vast increase of Y (X/NH3) in continuous cultures of up to 81.7% compared to batch cultures. This might only be explained by taking into account the unique nature of archaeal ammonia oxidation, which produces significant amounts of ROS that can destroy the cells if not taken care of by the environment (e.g., by catalases, alpha keto-acids). Thus, there is a strong selection pressure on the organism to regulate its activity in accordance to its environment. Given the wide distribution of AOA, it seems that these organisms might have evolved elaborate metabolic regulations to enable them to thrive. The production of ROS by AOA might thus be key to understanding their ecological success, because it could also be used by the organisms to generate substrates from the environment by oxidative decarboxylation ( Kim et al., 2016 ) or oxidative deamination ( Akagawa et al., 2002 ) of organic matter. This would supply the organism with both CO 2 and NH 3 and could explain why some AOA, like the members of the Nitrosocosmicus clade, are highly abundant in very organic rich soils. However, the conditions in soil, the primary habitat of Nitrososphaera spp., are not close to the conditions inside a bioreactor that operates in continuous culture mode. Hence, biofilm formation might be a preferred form of life for soil AOA, or could confer some resistance to selection via μ. In addition, this kind of metabolism, however, requires a high degree of metabolic regulation which might be enabled by the transcription apparatus of archaea, which is usually summarized as a simplified version of the eukaryotic machinery, even though this observed simplicity is under constant revision as new insights are accumulated ( Gehring et al., 2016 ). The unique nature of archaea and their gene regulation could therefore be responsible for this form of NH 3 oxidation. It appears that the guiding principle behind the growth dynamics of N. viennensis , and probably other AOA, is not substrate affinity but the maximization of Y (X/NH3) and therefore optimal utilization of a usually very limited substrate. These insights into the regulation of the energy metabolism of N. viennensis have important implications for further bioprocess development to optimize biomass productivity. Due to the increase of Y (X/NH3) with higher D , biomass productivity also increases despite high S concentrations. CO 2 concentration in the in-gas plays a crucial role in the way that it does not effect Y (X/NH3) but μ max and S . An optimal CO 2 concentration should not reduce μ max but at the same time minimize S and therefore maximize the cell concentration at a given D and S i . For higher concentrated feed medium it is probably worth in terms of biomass productivity to accept higher S concentrations as long as D positively affects Y (X/NH3) and biomass productivity. It is important to note that the NH 4 Cl to pyruvate ratio should always be at least in a 2:1 to prevent self toxification by endogenously produced ROS. Depending on the scientific question, higher NH 4 + concentrations might not be desirable as it certainty influences biomass composition and gene expression. For compounds of interest that cannot be recombinantly produced, higher substrate concentrations might very well be the method of choice as it looks promising to obtain good biomass productivities if further improved. For the production of biomass, biofilm formation is not favorable as it reduces the amount of cells that can be harvested and thus the effective Y (X/NH3) . Elevated temperatures are known to increase biofilm formation ( Qureshi et al., 2005 ), therefore lowering the temperature might be a solution but at the cost of reducing μ max . Reducing the CO 2 concentration will also likely reduce biofilm formation, because carbonate precipitate will be reduced and therefore also available surface that can induce biofilm formation. On the other hand, to understand the ecological function and behavior of N. viennensis , it would be very important to study the organism in biofilms, as this is more likely to be its prevalent form in soils. Gene abundances based on 16S rRNA or amo A are often used to infer NH 3 oxidizing activity of AOA in soils, but discrepancies with activity measurements of soil incubations are known. Estimates from this study show that only 14.67% of the cells in the biofilm would have been active at their maximum capacity. These results were obtained from two calculations with different assumptions in two different reactors in two different experiments. However, independent of the discrepancy of the results, this high ratio of inactive cells could also explain earlier findings of inactive cells that started the hypothesis of mixotrophic AOA together with the growth enhancing effect of alpha keto acids ( Mußmann et al., 2011 ; Tourna et al., 2011 ). As biofilms have the potential for very complex cell interactions, it might also be that “inactive” cells simply perform different tasks in the biofilm beside NH 3 oxidation. Regardless, careful validation and/or correction of these results with future experiments would be needed to have a more accurate picture of the behavior of AOA biofilms."
} | 5,803 |
36075935 | PMC9458709 | pmc | 9,245 | {
"abstract": "Light harvesting is fundamental for production of ATP and reducing equivalents for CO 2 fixation during photosynthesis. However, electronic energy transfer (EET) through a photosystem can harm the photosynthetic apparatus when not balanced with CO 2 . Here, we show that CO 2 binding to the light-harvesting complex modulates EET in photosynthetic cyanobacteria. More specifically, CO 2 binding to the allophycocyanin alpha subunit of the light-harvesting complex regulates EET and its fluorescence quantum yield in the cyanobacterium Synechocystis sp. PCC 6803. CO 2 binding decreases the inter-chromophore distance in the allophycocyanin trimer. The result is enhanced EET in vitro and in live cells. Our work identifies a direct target for CO 2 in the cyanobacterial light-harvesting apparatus and provides insights into photosynthesis regulation.",
"introduction": "Introduction Oxygenic photosynthesis evolved in the cyanobacteria about 2.5 Gyr ago and provided much of our current atmospheric O 2 1 . Cyanobacteria gather photons using a harvesting antenna known as the phycobilisome (PBS). Phycobiliproteins (PBPs) constitute PBSs to produce one of the largest protein complexes known. The PBPs include phycocyanin, phycoerythrin and allophycocyanin, together with their associated linker proteins 2 . PBS light absorption is achieved by open-chain tetrapyrrole (bilin) chromophores covalently attached to the PBPs and their associated linker proteins 3 . Two subunits of a PBP, an α- and a β-subunit, form an αβ heterodimer known as an (αβ) monomer, which is assembled into an (αβ) 3 trimer. Electronic energy transfer (EET) within the (αβ) 3 trimer is fast and efficient (on the sub-ps to ps timescales) 4 . The PBP (αβ) 3 trimers form the ordered supramolecular light-harvesting complex via the linker proteins 5 . A typical cyanobacterial hemidiscoidal PBS consists of a central core surrounded by peripheral chromophore-containing rods that capture photons 6 . The PBS core transfers excitation energy via a terminal emitter to the photosynthetic reaction centres with the subsequent synthesis of ATP and NADPH 7 . CO 2 assimilation into sugars requires NADPH and ATP. There is evidence from non-cyanobacterial model systems that CO 2 fixation and availability are coupled to light-harvesting 8 , 9 . For example, CO 2 -deficient conditions in algae reduce photosystem (PS) II antenna size, and PSI/PSII fluorescence intensity increases 10 , 11 . A light-harvesting chlorophyll–protein complex II migrated from PSII to PSI under CO 2 -deficient conditions in Chlamydomonas 12 . Insufficient CO 2 can also result in over-reduction of the plastoquinone pool, resulting in singlet oxygen species and photoinhibition 13 . The addition of a suitable carbon source under CO 2 -deficient conditions restored photosynthetic activity 14 . The mechanism(s) that couples inorganic carbon availability to light-harvesting is uncertain. In higher plants, bicarbonate binding to PSII provides a redox tuning mechanism regulating and protecting PSII 15 . However, questions remain as HCO 3 − is also proposed to be tightly bound to PSII under physiologically relevant environmental CO 2 , meaning this binding site’s role as an authentic regulatory site at limiting CO 2 awaits further experimentation 16 . Nevertheless, there is considerable evidence that CO 2 /HCO 3 − availability can regulate light-harvesting and subsequent EET. How might CO 2 regulate light-harvesting? CO 2 and protein can interact through carbamates on neutral N -terminal α-amino- or lysine ε-amino groups. For example, carbamylation regulates the activities of Rubisco 17 , haemoglobin (Hb) 18 , and ubiquitin 19 . In addition, several proteins carry a stable carbamate required for catalysis, e.g., urease, alanine racemase, transcarboxylase 5 S, class D β-lactamase and phosphotriesterase 20 . Therefore, we and others have hypothesized that reversible carbamylation of neutral N -terminal α-amino groups and/or lysine ε-amino groups could form a widespread mechanism for protein regulation by CO 2 17 , 20 . We developed triethyloxonium tetrafluoroborate (TEO) as a chemical proteomics tool to identify carbamate post-translational modifications (PTMs) 20 . We hypothesised the TEO chemical proteomics tool could be used to discover CO 2 -binding proteins in the cyanobacterial PBS to determine the mechanistic basis for coupling CO 2 availability to EET. Here we describe the discovery of a CO 2 -binding site in the PBS and its role in regulating EET.",
"discussion": "Discussion The Calvin–Benson–Bassham (CBB) cycle rate depends on multiple factors. However, PCO 2 is particularly important as it determines the relative rate of RuBisCO carboxylation compared to the competing oxygenation reaction. Following ribulose-1,5-biphosphate carboxylation, the CBB cycle requires ATP and NADPH, whose availability is controlled by light-dependent reactions. Therefore, although the light-dependent reactions might be independent of substrate availability, reduced PCO 2 could represent a bottleneck for the ensuing light-independent reactions. It follows that an organism should optimise the light-dependent reaction rate to fit the changing environmental conditions that, under specific adverse scenarios, could lead to damage of the photosystems. Therefore, it seems paramount that CO 2 can have a regulatory effect at the level of light-harvesting and/or light-dependent reactions for ‘sink to source’ regulation. The finding of a CO 2 -mediated PTM in the PBS might represent such a mechanism."
} | 1,388 |
25101064 | PMC4101577 | pmc | 9,246 | {
"abstract": "Intracellular habitats have been invaded by a remarkable diversity of organisms, and strategies employed to successfully reside in another species' cellular space are varied. Common selective pressures may be experienced in symbioses involving phototrophic symbionts and heterotrophic hosts. Here I refine and elaborate the Arrested Phagosome Hypothesis that proposes a mechanism that phototrophs use to gain access to their host's intracellular habitat. I employ the economic concept of production possibility frontiers (PPF) as a useful heuristic to clearly define the trade-offs that an intracellular phototroph is likely to face as it allocates photosynthetically-derived pools of energy. Fixed carbon can fuel basic metabolism/respiration, it can support mitotic division, or it can be translocated to the host. Excess photosynthate can be stored for future use. Thus, gross photosynthetic productivity can be divided among these four general categories, and natural selection will favor phenotypes that best match the demands presented to the symbiont by the host cellular habitat. The PPF highlights trade-offs that exist between investment in growth (i.e., mitosis) or residency (i.e., translocating material to the host). Insights gained from this perspective might help explain phenomena such as coral bleaching because deficits in photosynthetic production are likely to diminish a symbiont's ability to “afford” the costs of intracellular residency. I highlight deficits in our current understanding of host:symbiont interactions at the molecular, genetic, and cellular level, and I also discuss how semantic differences among scientists working with different symbiont systems may diminish the rate of increase in our understanding of phototrophic-based associations. I argue that adopting interdisciplinary (in this case, inter-symbiont-system) perspectives will lead to advances in our general understanding of the phototrophic symbiont's intracellular niche.",
"conclusion": "Conclusion The purpose of this perspective is to focus attention on significant trade-offs that exist for phototrophic symbionts residing in heterotrophic host cells. Constraints exist on investment strategies involving energy that is represented by fixed carbon produced through photosynthesis. The possible phenotypic responses to the trade-offs have significant evolutionary implications. To study these trade-offs, we must understand the cellular environment that the symbionts reside in because important symplesiomorphies likely exist among the various organisms that engage in this type of ecological interaction. One way to achieve success in this area is to increase the dialog that occurs among scientists working with different symbioses. By using names unique to specific hosts to describe endomembranous spaces that phototrophs live in, we may be missing important clues to how symbionts establish stable populations within a particular host. Finally, if we shift attention away from host “control” of the associations, and instead think about the role the symbiont might play in shaping the interactions, we may discover novel theoretical and empirical approaches that have broad explanatory power. Conflict of interest statement The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest."
} | 847 |
32388815 | PMC7306483 | pmc | 9,247 | {
"abstract": "Calderihabitans maritimus KKC1 is a thermophilic, carbon monoxide (CO)-utilizing, hydrogen-evolving bacterium that harbors seven cooS genes for anaerobic CO dehydrogenases and six hyd genes for [NiFe] hydrogenases and capable of using a variety of electron acceptors coupled to CO oxidation. To understand the relationships among these unique features and the transcriptional adaptation of the organism to CO, we performed a transcriptome analysis of C. maritimus KKC1 grown under 100% CO and N 2 conditions. Of its 3114 genes, 58 and 32 genes were significantly upregulated and downregulated in the presence of CO, respectively. A cooS–ech gene cluster, an “orphan” cooS gene, and bidirectional hyd genes were upregulated under CO, whereas hydrogen-uptake hyd genes were downregulated. Transcriptional changes in anaerobic respiratory genes supported the broad usage of electron acceptors in C. maritimus KKC1 under CO metabolism. Overall, the majority of the differentially expressed genes were oxidoreductase-like genes, suggesting metabolic adaptation to the cellular redox change upon CO oxidation. Moreover, our results suggest a transcriptional response mechanism to CO that involves multiple transcription factors, as well as a CO-responsive transcriptional activator (CooA). Our findings shed light on the diverse mechanisms for transcriptional and metabolic adaptations to CO in CO-utilizing and hydrogen-evolving bacteria. Electronic supplementary material The online version of this article (10.1007/s00792-020-01175-z) contains supplementary material, which is available to authorized users.",
"introduction": "Introduction Carbon monoxide (CO) is used as an energy source by CO-oxidizing microbes (carboxydotrophs) because of its low redox potential (Ragsdale 2004 ; Oelgeschläger and Rother 2008 ; Sokolova et al. 2009 ; Diender et al. 2015 ). Carboxydotrophs harness CO dehydrogenases (CODHs) for CO utilization by catalyzing the reaction CO + H 2 O ⇌ CO 2 + 2H + + 2e − (Ragsdale 2004 ; Oelgeschläger and Rother 2008 ). CODHs are divided into two families: anaerobic Ni-containing CODHs (Ni-CODHs) and aerobic molybdenum- and copper-containing CODHs (Can et al. 2014 ; Hille et al. 2015 ). Unlike aerobic CODHs, Ni-CODHs can reduce ferredoxin and thereby utilize various types of terminal electron acceptors, such as protons, CO 2 , sulfate, and ferric iron [Fe(III)] (Oelgeschläger and Rother 2008 ; Sokolova et al. 2009 ; Diender et al. 2015 ). Because of this unique feature of Ni-CODHs, physiologically diverse anaerobic carboxydotrophs have been described, such as hydrogenogens, acetogens, methanogens, sulfate reducers, and Fe(III) reducers (Oelgeschläger and Rother 2008 ; Sokolova et al. 2009 ; Diender et al. 2015 ). Hydrogenogenic carboxydotrophs couple CO oxidation with proton reduction to produce hydrogen (H 2 ), during which the proton- or sodium-motive force is generated with residual energy via a Ni-CODH/energy converting hydrogenase (ECH) complex (Singer et al. 2006 ; Schut et al. 2016 ; Schoelmerich and Müller 2019 ). CO-dependent H 2 production by hydrogenogenic carboxydotrophs is considered a “safety valve” to reduce toxic CO and supply H 2 , which is an energy source for H 2 -utilizing microbial communities (Techtmann et al. 2009 ). Hydrogenogenic carboxydotrophs are generally divided into three groups in their phylogeny: Firmicutes, Proteobacteria, and Archaea (Diender et al. 2015 ; Inoue et al. 2019 ). In Firmicutes, the Clostridia includes various types of thermophilic, hydrogenogenic carboxydotrophs that harbor multiple cooS genes and feature the Wood–Ljungdahl pathway (WLP) for carbon fixation (Techtmann et al. 2012 ; Shin et al. 2016 ; Inoue et al. 2019 ). The functions of these cooS genes have been predicted from their genomic contexts such as ECH, WLP, and ferredoxin–NAD(P)H oxidoreductase (Techtmann et al. 2012 ; Inoue et al. 2019 ), which are presumed to be regulated by CO-responsive transcription factors, such as CooA and RcoM (Shelver et al. 1995 ; Komori et al. 2007 ; Kerby et al. 2008 ). However, recent studies of two hydrogenogenic carboxydotrophs, Carboxydothermus pertinax and Thermoanaerobacter kivui , show that the enzymatic coupling of cooS and ech genes that are distantly encoded in their respective genomes enables CO-dependent H 2 production (Fukuyama et al. 2018 , 2019a ; Schoelmerich and Müller 2019 ). Moreover, cooS expression is upregulated in the presence of CO, despite the fact that no sequence motif is recognized by the CO-responsive transcriptional activator CooA in C. pertinax (Fukuyama et al. 2018 , 2019a ). Similarly, ech expression is upregulated in the presence of CO, although the genome does not encode any previously described CO-responsive transcription factors in T. kivui (Schoelmerich and Müller 2019 ). These studies suggest that there are previously unknown transcriptional response mechanisms to CO. Therefore, the direct observation of transcriptional, proteomic, and metabolic changes under CO is required to understand the metabolisms of hydrogenogenic carboxydotrophic bacteria. Calderihabitans maritimus KKC1 is a thermophilic, obligately anaerobic, hydrogenogenic, carboxydotrophic bacterium in the Clostridia, closely related to Moorella , and was isolated from a core sample taken from marine sediment in the Kikai Caldera, Japan (Yoneda et al. 2013 ). C. maritimus KKC1 can grow anaerobically on 100% CO while producing H 2 and CO 2 with or without additional electron acceptors such as thiosulfate, sulfite, ferric citrate, amorphous Fe(III) oxide, Fe 2 O 3 , or fumarate, and on organic compounds, such as pyruvate, in the presence of thiosulfate. C. maritimus KKC1 requires yeast extract for CO-dependent growth unlike chemolithoautotrophic, hydrogenogenic carboxydotrophs, such as Moorella stamsii and Carboxydothermus hydrogenoformans , and is unable to grow with H 2 and CO 2 unlike chemolithoautotrophic acetogens, such as Moorella thermoacetica (Svetlichny et al. 1991 ; Drake and Daniel 2004 ; Alves et al. 2013 ; Yoneda et al. 2013 ). The draft genome of C. maritimus KKC1 encodes six cooS genes that are categorized by their genomic contexts, as follows: cooS1 (KKC1_RS04465), WLP; cooS2 (KKC1_RS06675), ECH; cooS3 (KKC1_RS06585), ferredoxin-NAD(P)H oxidoreductase; cooS4 (KKC1_RS12505), a cysteine synthase and ABC transporter; cooS5 (KKC1_RS04925), 2-oxoglutarate:ferredoxin oxidoreductase (Kor); and cooS6 (KKC1_RS10495), CooA (Omae et al. 2017 ). These six cooS genes harbor the complete sequence motifs that form three types of metal clusters for catalysis, although cooS1 is frame-shifted like other hydrogenogenic, carboxydotrophic Moorella and Carboxydothermus species possibly as a result of cultivation at high CO concentrations (Wu et al. 2005 ; Omae et al. 2017 ; Poehlein et al. 2018 ; Fukuyama et al. 2018 , 2019b ). To the best of our knowledge, the genome contains the highest number of cooS genes (Omae et al. 2017 ; Toshchakov et al. 2018 ) and encodes six hydrogenase gene clusters that include two ech gene clusters, a coo -type gene cluster ( ech1 , KKC1_RS06640–KKC1_RS06665) with cooS2 , and a hyc / hyf -type gene cluster ( ech2 , KKC1_RS01155–KKC1_RS01200) with putative formate dehydrogenase genes (Omae et al. 2017 ). The genome only harbors one cooA gene as a CO-responsive transcription factor. The multiplicity of Ni-CODHs and hydrogenases and the broad usage of electron acceptors in C. maritimus KKC1 suggest genomic adaptation for its carboxydotrophic growth (Yoneda et al. 2013 ; Omae et al. 2017 ); however, its metabolic and transcriptional responses to CO remain largely unknown. In this study, we performed a transcriptome analysis of C. maritimus KKC1 grown in the presence or absence of CO using RNA sequencing (RNA-seq). Under CO conditions, we found the evidence of transcriptional changes for Ni-CODHs and hydrogenases, and for anaerobic respiration, carbon and nitrogen metabolism, and transcription factors. We suggest that the transcriptional response mechanism to CO involves multiple transcription factors.",
"discussion": "Discussion We examined transcriptomic changes in C. maritimus KKC1 growing in the presence or absence of CO using RNA-seq analysis. Our data showed that of the seven cooS genes and six hydrogenase genes studied, the cooS2 – ech1 ( coo -type) gene cluster (not the hyc / hyf -type gene cluster), the “orphan” cooS6 gene, and bidirectional hyd-3b genes were upregulated under CO condition (Fig. 2 and Table 1 ). In hydrogenogenic carboxydotrophic Moorella species, genes for the hyc / hyf -type hydrogenase form a gene cluster with cooS and are responsible for CO-dependent H 2 production (Poehlein et al. 2018 ; Fukuyama et al. 2019b ), while in Carboxydothermus species, the coo -type works with cooS (Wu et al. 2005 ; Fukuyama et al. 2019a ), suggesting that C. maritimus KKC1 utilizes a strategy similar to Carboxydothermus species rather than Moorella species for CO-dependent energy conservation. Upregulation of the “orphan” cooS6 gene under CO conditions could induce the reduction of the ferredoxin pool and perturbation of cellular redox balance in C. maritimus KKC1. The bidirectional hyd-3b genes might balance such redox perturbation. Moreover, unlike C. pertinax , H 2 -uptake hyd-1a genes were downregulated under CO conditions in C. maritimus KKC1 (Fig. 2 and Table 1 ) (Fukuyama et al. 2019a ). These data suggest that an H 2 -evolution system is predominant in the presence of CO because of the excessive reducing equivalents from CO in C. maritimus KKC1. Our data support the broad usage of electron acceptors in anaerobic respiration of C. maritimus KKC1 (Fig. 2 and Table 2 ). We found that three mhc genes presumed to utilize Fe(III) were upregulated in the presence of CO. In Thermincola potens , MHCs are responsible for respiratory electron transfer to Fe(III) (Carlson et al. 2012 ), and genomic analysis of Carboxydocella thermoautotrophica indicated the possible involvement of MHCs in CO-dependent Fe(III) reduction (Toshchakov et al. 2018 ). Similar to these hydrogenogenic carboxydotrophs, C. maritimus KKC1 produces Fe(II) with ferric citrate, amorphous Fe(III) oxide, and Fe 2 O 3 in the presence of CO, suggesting that Fe(III) is used as an electron acceptor under CO conditions (Yoneda et al. 2013 ). Therefore, these MHCs might be responsible for extracellular electron transfer to Fe(III) in C. maritimus KKC1. On the contrary, dsr genes that utilize sulfite in thiosulfate respiration were downregulated in the presence of CO although C. maritimus KKC1 is considered to couple thiosulfate reduction with CO or pyruvate oxidation (Yoneda et al. 2013 ). Thiosulfate respiration is conducted by a concerted way of thiosulfate/polysulfide reductase (Phs) and Dsr (Stoffels et al. 2011 ; Venceslau et al. 2014 ). We could not identify any phs genes in C. maritimus KKC1; therefore, a complete thiosulfate reduction pathway in this organism remains unknown. C. pertinax, which can also utilize Fe(III) and thiosulfate with CO or pyruvate oxidation, shows that expressions of the mhc -like genes, the dsr gene cluster (cpu_17430–17,360), and the phs gene cluster (cpu_06910–06,930) remain unchanged between CO and N 2 conditions (Fukuyama et al. 2019a ), suggesting that transcriptional responses in the anaerobic respiration pathway upon CO would differ between these two hydrogenogenic carboxydotrophs. Oxidoreductase-like genes encoding Kor, GOGAT, and GS involved in carbon and nitrogen metabolisms were also upregulated under CO conditions in C. maritimus KKC1, strongly suggesting metabolic adaptation to cellular redox change upon CO oxidation (Fig. 2 and Table 3 ). Moreover, kor genes are upregulated in Thermococcus onnurineus under CO conditions, suggesting the possible involvement of Kor in carbon fixation (Moon et al. 2012 ), whereas the GS/GOGAT cycle is a redox-buffering system involving the consumption of NAD(P)H in Caldicellulosiruptor bescii and Clostridium thermocellum (Sander et al. 2015 , 2019 ). It is possible that the enzymatic cycle involving Kor, GOGAT, and GS would assimilate carbon and nitrogen using excessive reducing equivalents from CO. Additionally, we identified putative transcription factors as DEGs (Table 4 ) and putative CooA- and Rex-binding motifs in the upstream regions of upregulated DEGs (Fig. 3 ). As noted, the expression of cooA remained unchanged under CO conditions, suggesting that the basal expression level of cooA would be adequate to adapt to CO (Table S2). This phenomenon is different in C. pertinax and Desulfovibrio vulgaris , where cooA genes are upregulated in the presence of CO (Rajeev et al. 2012 ; Fukuyama et al. 2019a ). Moreover, Rex is a redox-sensing transcriptional repressor and conserved among various bacteria, regardless of whether they are anaerobes or aerobes, and conformational changes from NAD + - to NADH-bound states induce transcription by releasing Rex from its recognition sequence (McLaughlin et al. 2010 ; Ravcheev et al. 2012 ). C. maritimus KKC1 possesses one Rex homolog (KKC1_RS10865), which exhibits 39% and 41% amino acid identities with those from C. bescii and Clostridium acetobutylicum , respectively. As described here, C. bescii Rex regulates various types of hydrogenases including ECH, and pyruvate:ferredoxin oxidoreductase belonging to the same family as Kor (Sander et al. 2019 ). In Clostridium species, Rex regulates the expressions of genes encoding various types of oxidoreductase, including nitrate reductases, hydrogenases, and WLP enzymes (Zhang et al. 2014 ). Therefore, cellular redox changes under CO conditions in C. maritimus KKC1 might result in a high NADH/NAD + ratio to drive Rex-dependent transcriptional activation of oxidoreductase-like genes. These findings suggest a possible multi-step transcriptional response to CO in C. maritimus KKC1 as follow: 1) upon CO exposure, CO-sensing CooA activates transcription of the cooS2 – ech1 gene cluster and cooS6 , and these two Ni-CODHs catalyze CO oxidation to supply reducing equivalents within the cell; 2) excessive reducing equivalents from CO result in a high NADH/NAD + ratio, followed by Rex-dependent transcriptional activation of the cooS2 – ech1 , hyd-3b , kor , fdh, and nap gene clusters, and three mhc genes; and 3) changes in the metabolisms or expression of transcription factors induce alterations in the transcription of other DEGs. Our data highlight the diversity of CO-responsive transcriptional regulation in thermophilic, hydrogenogenic, carboxydotrophic bacteria. In C. pertinax , of the nine gene clusters upregulated in the presence of CO, including cooS , only the ech gene cluster is directly regulated by CooA (Fukuyama et al. 2019a ), whereas the Rex-binding motif or those of other transcription factors are not found in upstream regions of these nine gene clusters. Additionally, in hydrogenogenic, carboxydotrophic Moorella strains, whose genomes encode no known CO-sensing transcription factor homologs, genes for RocR-like transcriptional activators (MOST_RS16225 in M. stamsii and MOTE_RS04420 in M. thermoacetica DSM 21,394) are located in the upstream regions of their Ni-CODH–ECH gene clusters. Because two genes for RocR-like proteins were upregulated under CO condition in C. maritimus KKC1 (Table 4 ), RocR-like proteins related to Ni-CODH–ECH gene clusters in these two Moorella species might be involved in response to CO. Moreover, a recent comparative genomics study of Parageobacillus thermoglucosidasius , a hydrogenogenic carboxydotroph also lacking known CO-sensing transcription factors, has found a transition-state regulator Hpr-binding sequence in the upstream region of its Ni-CODH–ECH gene cluster (Mohr et al. 2018 ). These findings imply previously undescribed transcriptional response mechanisms to CO. There could be various ways to respond to CO, including directly sensing CO, via stress caused by CO, or through cellular redox or metabolic changes via CO oxidation. Therefore, the diverse strategies for adaptation to CO-dependent metabolism would have been evolved in thermophilic, hydrogenogenic, carboxydotrophic bacteria."
} | 4,117 |
25450012 | PMC4406151 | pmc | 9,249 | {
"abstract": "Microorganisms can be engineered for the production of chemicals utilized in the polymer industry. However many such target compounds inhibit microbial growth and might correspondingly limit production levels. Here, we focus on compounds that are precursors to bioplastics, specifically styrene and representative alpha-olefins; 1-hexene, 1-octene, and 1-nonene. We evaluated the role of the Escherichia coli efflux pump, AcrAB-TolC, in enhancing tolerance towards these olefin compounds. AcrAB-TolC is involved in the tolerance towards all four compounds in E. coli . Both styrene and 1-hexene are highly toxic to E. coli . Styrene is a model plastics precursor with an established route for production in E. coli (McKenna and Nielsen, 2011). Though our data indicates that AcrAB-TolC is important for its optimal production, we observed a strong negative selection against the production of styrene in E. coli . Thus we used 1-hexene as a model compound to implement a directed evolution strategy to further improve the tolerance phenotype towards this alpha-olefin. We focused on optimization of AcrB, the inner membrane domain known to be responsible for substrate binding, and found several mutations (A279T, Q584R, F617L, L822P, F927S, and F1033Y) that resulted in improved tolerance. Several of these mutations could also be combined in a synergistic manner. Our study shows efflux pumps to be an important mechanism in host engineering for olefins, and one that can be further improved using strategies such as directed evolution, to increase tolerance and potentially production. Biotechnol. Bioeng. 2015;112: 879–888. © 2015 The Authors. Biotechnology and Bioengineering Published by John Wiley & Periodicals, Inc.",
"introduction": "Introduction Microbes can be engineered to produce a large variety of chemicals. Many of these chemicals impose varying degrees of toxicity on the microbial strain being engineered for their production. Host engineering to improve strain tolerance towards a target compound is essential to optimize the production levels. Different compounds may also require different host engineering approaches, as the nature of toxicity is compound dependent. Hydrophobic chemicals are hypothesized to impact membrane permeability and fluidity, diminishing energy transduction, interfering with membrane protein function, and affecting a range of essential cellular processes (Jarboe et al., 2010 ; Sikkema et al., 1995 ). Tolerance and defense mechanisms range from induction of chaperones, modification of membrane composition and cellular morphology, to induction of efflux pumps that export the compounds out of the cell membrane and the cell. Of these, export pumps have emerged as an important mechanism for use in host engineering (Chen et al., 2013 ; Doshi et al., 2013 ; Dunlop et al., 2011 ) and also good targets for directed evolution (Fisher et al., 2013 ; Foo and Leong, 2013 ). In this study we focus on the role of an Escherichia coli transporter in improving tolerance and production of olefin compounds that are valuable precursors in the plastic and polymer industry (e.g., bioacrylic, biostyrene) (McKenna and Nielsen, 2011 ; Schubert et al., 1988 ). Transporters provide a general mechanism for the export of toxic compounds (Nikaido, 2009 ; Takatsuka et al., 2010 ) thus also improving tolerance. For chemicals produced in the microbial cytoplasm, transport out of the cell may also improve production levels. In gram negative bacteria, tolerance towards solvent-like compounds is conferred by the hydrophobe/amphiphile efflux family of resistance-nodulation-division (RND) pumps (Nikaido and Takatsuka, 2009 ; Ramos et al., 2002 ; Tseng et al., 1999 ). RND efflux pumps are composed of three subunits. An inner membrane unit binds the substrate and with a proton motive force, exports it through the outer membrane subunit. The periplasmic subunit connects and stabilizes the inner and outer membrane units ( Fig. 1 A). Figure 1 Schematic representation of the AcrAB-TolC efflux pump (A), and assay developed to determine the impact of 1-hexene on cell survival (B). A: Organization of the three proteins forming the tripartite efflux pump complex and their localization in the E. coli membrane is represented. The trimer of the AcrB subunit is shown in color. The newly discovered AcrZ domain (Hobbs et al., 2012 ) is not depicted in this diagram. B: Assay with a saturated atmosphere using the volatile compound to impose toxicity. The arrows represent the inoculation cell culture on the agar surface; the open lid of the petri-dish is used to place small receptacles containing a volatile compound. The chamber is assembled by inverting the petri-dish on the lid, cells facing the receptacles and sealed using parafilm (See methods). This method was used for 1-hexene survival assays as well as to screen AcrB mutants. The saturated atmosphere assay was also used with other alpha-olefins like 1-octene and 1-nonene. In E. coli , the AcrAB-TolC efflux pump ( Fig. 1 A) composed of AcrB (inner membrane protein), TolC (outer membrane protein), and AcrA (periplasmic) has been extensively studied (Du et al., 2014 ; Murakami et al., 2002 ; Tikhonova et al., 2011 ) and has broad substrate specificity, ranging from detergents to antibiotics and solvents. This pump is reported to play a major role in the secretion of various alkanes such as hexane, heptane, octane, and nonane (Takatsuka et al., 2010 ) and also in imparting tolerance to various terpene based biofuel compounds (Dunlop et al., 2011 ). Several reports focus on the mechanism of this pump, especially that of AcrB, describing its rotational conformation changes (Seeger et al., 2006 ; Seeger et al., 2008 ; Sennhauser et al., 2007 ; Takatsuka and Nikaido, 2009 , 2010 ), its binding pockets and the amino acids involved in substrate recognition (Eicher et al., 2012 ; Husain and Nikaido, 2010 ; Vargiu and Nikaido, 2012 ). In this study we test the role of this RND efflux pump in providing tolerance towards various olefins and evaluate its impact on the production of styrene. We use a random mutagenesis based strategy to successfully identify variants of the AcrB protein for improved 1-hexene tolerance. We discuss the potential modes of action of the AcrB mutations that lead to the improved function.",
"discussion": "Discussion In this study we show that the E. coli efflux pump, AcrAB-TolC, plays a major role in tolerance to styrene and alpha-olefins, and may provide a significant advantage in reaching high production levels of toxic compounds. One of the first steps to improve the tolerance to alpha-olefins was to determine the benefit from increasing the levels of the AcrAB-TolC efflux pump. We observed that increasing the expression of this pump improved tolerance to 1-hexene but this strategy is limited by the toxicity known to be linked to overexpression of membrane proteins (Wagner et al., 2007 ). We thus decided to improve the pump function using directed evolution. We show that engineering the AcrB subunit of this efflux pump improves tolerance to 1-hexene. Our goal was to find variants that are able to improve the tolerance to 1-hexene and to characterize the primary domains responsible for 1-hexene export. It has been shown that the entrance of substrates into the efflux pump and the amino acids involved in the substrate binding depends of the properties and the size of the compounds (Eicher et al., 2012 ; Takatsuka et al., 2010 ). To better understand the domain involved in 1-hexene export we analyzed several beneficial mutations obtained after a round of evolution. We identified several mutations (A279T, Q584R, F617L, L822P, F927S, and F1033Y) that resulted in improved tolerance to 1-hexene. We found that these mutations have an additive effect, but also established that some changes may be disruptive for protein solubility. Since all tested AcrB variants in our study resulted in decreased protein levels, it is more likely that the improvement in the tolerance conferred by these variants is due to an alternate mechanism, such as higher pump efficiency. Crystallographic and site-directed mutagenesis studies of AcrB have determined the role of domains and of several amino acids in the protein (Husain and Nikaido, 2010 ; Murakami et al., 2002 ). Among the beneficial amino acid changes we identified, F927S and especially F1033Y were the most unexpected. F927S is localized in the transmembrane domain ( Fig. 6 B and C), at the top of the TM10 α-helix. This helix contains a crucial amino acid (K940) involved in proton transport (Su et al., 2006 ) and a mutation in this helix could impact the rotational movements leading to compound export. The mechanism of the F1033Y mutation is unclear but this amino acid could potentially interact with the newly discovered AcrZ protein (Hobbs et al., 2012 ). The improved tolerance with F1033Y and F927S could also be due to an indirect effect such as altering the membrane structure. However the mutations A279T, Q584R, F617L, and L822P were localized in positions known to be important for pump function. L822 is positioned between the two β-sheets Cβ13 and Cβ14 belonging to the pore subdomains PN1 and PC2, respectively ( Fig. 6 B and C). The mutation of a leucine to a proline is located between these 2 β-sheets at the “ceiling” of the vestibule, suggested to be a highly probable substrate entrance point, possibly altering the flexibility and/or the opening of the vestibule facilitating 1-hexene entrance into the pore. A279 is located in the binding pocket in which residues E273, N274, D276, I277, play important roles (Husain et al., 2011 ). Why the introduction of a polar amino acid at this position would improve the efflux of hydrophobic compound is unclear, but every substrate must eventually leave the binding site in order for its efflux to occur, so the substitution A279T (hydrophobic to polar residue) may help substrate release to the gate and the funnel. Q584 is located in a position potentially involved in trimer assembly. The AcrB subunit is reported to fold independently, and then assemble into a trimer (Yu et al., 2011 ). It has been shown that the P223 from one AcrB interacts with Q584 from another AcrB polypeptide, and is required for the assembly and stability of the trimer. The mutation Q584R could therefore impact trimer assembly and stability ( Fig. 6 B and C). This mutation also had a minor but significant improvement in E. coli tolerance to bile salts but not to styrene. Finally, the amino acid F617 has been shown to be located in the switch loop of the hydrophobic binding pocket. Reported crystal structures suggest that this amino acid could directly interact with various known substrates (Bohnert et al., 2008 ; Eicher et al., 2012 ; Vargiu et al., 2011 ). Mutating F617 to an alanine has been reported to have a direct impact on substrate uptake and was responsible for a substantial decrease in transport of novobiocin, but had a minor effect on the transport of oxacillin and various other macrolides (Bohnert et al., 2008 ). In our study, the mutation to a leucine at this position may have also improved transport of 1-hexene. Our study, as well as other reports in the literature (Bohnert et al., 2008 ) suggest that AcrB is destabilized when more than three mutations are simultaneously introduced. However, additional rounds of evolution, starting with variants containing fewer mutations, could stabilize the protein and allow the introduction of more mutations to further improve function. Heterologous pumps from other organisms have been shown to bestow solvent tolerance in E. coli (Dunlop et al., 2011 ) and may be good candidates for future studies as well as directed evolution. Use of dynamic control systems to regulate transporter expression has been shown to enhance the final tolerance phenotype (Frederix et al., 2014 ) and provides another engineering route to explore. Additionally, analyses of more AcrB variants with increased tolerance to 1-hexene could provide a better understanding of the path(s) of this compound through the pump, allowing targeted engineering of the protein. Directed evolution of AcrB has also been conducted by other groups. AcrB variants with improved tolerance to n-butanol (Fisher et al., 2013 ) as well as α-pinene and n-octane are known (Foo and Leong, 2013 ), and the authors propose the improvements to have resulted from changes in the AcrB-TolC interaction, in the enlargement of the entrance to the cleft, in the facilitation of conformational changes or in the improvement of affinity for the substrate. Taken together with these previous studies, our results show that, within this pump's complex mode of action, several domains can be mutated and lead to a more efficient pump. Our study emphasizes the importance of an AcrAB-TolC type system for improved tolerance and export in host engineering, and shows that directed evolution remains a powerful tool to obtain improved variants of such systems. We thank Huu Tran for his skillful technical assistance with the liquid handler robot, Dominque Loque and Fan Yang for providing the A. thaliana cDNA, James Kirby for help designing the GC-MS method, Thomas Ruegg for the Omnilog utilization, Marijke Frederix for helpful discussion and advice on various aspects of this study, and Nathan Hillson and Daniel Liu for reviewing the manuscript. Zosia Rostomian prepared Figure 1, 6B and C. The portion of the work conducted by the Lawrence Berkeley National Laboratory and the Joint BioEnergy Institute (JBEI) was supported by the U.S. DOE under Contract no. DE-AC02-05CH11231. This work, funded by Total, was performed as part of a collaborative program between JBEI and Total. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H. 2006. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Molecular Systems Biology 2:2006 0008. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proceedings of the National Academy of Sciences 97(12):6640-6645. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, 3rd, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6(5):343-5.McKenna R, Nielsen DR. 2011. Styrene biosynthesis from glucose by engineered E. coli. Metabolic engineering 13(5):544-54."
} | 3,626 |
26104694 | null | s2 | 9,250 | {
"abstract": "Since its initial discovery as an allosteric factor regulating cellulose biosynthesis in Gluconacetobacter xylinus, the list of functional outputs regulated by c-di-GMP has grown. We have focused this article on one of these c-di-GMP-regulated processes, namely, biofilm formation in the organism Pseudomonas aeruginosa. The majority of diguanylate cyclases and phosphodiesterases encoded in the P. aeruginosa genome still remain uncharacterized; thus, there is still a great deal to be learned about the link between c-di-GMP and biofilm formation in this microbe. In particular, while a number of c-di-GMP metabolizing enzymes have been identified that participate in reversible and irreversible attachment and biofilm maturation, there is a still a significant knowledge gap regarding the c-di-GMP output systems in this organism. Even for the well-characterized Pel system, where c-di-GMP-mediated transcriptional regulation is now well documented, how binding of c-di-GMP by PelD stimulates Pel production is not understood in any detail. Similarly, c-di-GMP-mediated control of swimming, swarming and twitching also remains to be elucidated. Thus, despite terrific advances in our understanding of P. aeruginosa biofilm formation and the role of c-di-GMP in this process since the last version of this book (indeed there was no chapter on c-di-GMP!) there is still much to learn."
} | 346 |
23596447 | PMC3622252 | pmc | 9,253 | {
"abstract": "Plant growth-promoting rhizobacteria (PGPR) are increasingly appreciated for their contributions to primary productivity through promotion of growth and triggering of induced systemic resistance in plants. Here we focus on the beneficial effects of one particular species of PGPR ( Pseudomonas fluorescens ) on plants through induced plant defense. This model organism has provided much understanding of the underlying molecular mechanisms of PGPR-induced plant defense. However, this knowledge can only be appreciated at full value once we know to what extent these mechanisms also occur under more realistic, species-diverse conditions as are occurring in the plant rhizosphere. To provide the necessary ecological context, we review the literature to compare the effect of P. fluorescens on induced plant defense when it is present as a single species or in combination with other soil dwelling species. Specifically, we discuss combinations with other plant mutualists (bacterial or fungal), plant pathogens (bacterial or fungal), bacterivores (nematode or protozoa), and decomposers. Synergistic interactions between P. fluorescens and other plant mutualists are much more commonly reported than antagonistic interactions. Recent developments have enabled screenings of P. fluorescens genomes for defense traits and this could help with selection of strains with likely positive interactions on biocontrol. However, studies that examine the effects of multiple herbivores, pathogens, or herbivores and pathogens together on the effectiveness of PGPR to induce plant defenses are underrepresented and we are not aware of any study that has examined interactions between P. fluorescens and bacterivores or decomposers. As co-occurring soil organisms can enhance but also reduce the effectiveness of PGPR, a better understanding of the biotic factors modulating P. fluorescens –plant interactions will improve the effectiveness of introducing P. fluorescens to enhance plant production and defense.",
"introduction": "INTRODUCTION Plant growth-promoting rhizobacteria (PGPR) are a diverse group of microorganisms that are increasingly appreciated for their contributions to primary productivity through promotion of growth and triggering of induced systemic resistance (ISR) in plants. By triggering plant defense, PGPR can make an important contribution to biocontrol of pests and pathogens of plants. However, the effectiveness of PGPR-triggered plant defense depends on a variety of genetic and biotic/abiotic environmental factors. PGPR naturally occur within a complex community of soil organisms inhabiting the rhizosphere. Hence, in order to understand the role of PGPR in influencing a plant’s defense against pests and pathogens, it is important to understand how biotic interactions with these rhizosphere organisms will affect the ability of PGPR to enhance plant defense. The aim of this review is to examine how the impact of PGPR on plant defense is modulated by the presence of other organisms in the rhizosphere. Other reviews have focused on particular interactions, e.g., between PGPR and aboveground insects ( Pieterse and Dicke, 2007 ; Pineda et al., 2010 ) or type of defense, e.g., volatiles ( Dicke and Baldwin, 2010 ). Those reviews have taken a plant centric approach (but see Whipps, 2001 ). In this paper, we will review how biotic interactions between PGPR and other rhizosphere- or plant-associated organisms affect the ability of PGPR to enhance plant defenses. We use Pseudomonas fluorescens , a very common and well-studied PGPR, as a model species. To test the dependence of PGPR–plant interactions on direct and indirect biotic interactions with other rhizosphere biota, we compare studies in which effects of P. fluorescens on plant defense are examined for a single P. fluorescens isolate with studies in which these effects are examined for a P. fluorescens isolate in combination with other isolates and/or species. We will discuss these interactions in increasing order of complexity, starting with single introductions of P. fluorescens with introductions of multiple P. fluorescens isolates, then with other PGPR, with other plant growth-promoting fungi, bacterivores, and finally with decomposing organisms. The basic interaction in all these studies is formed by a plant, P. fluorescens and a herbivore or pathogen. The latter is necessary to judge whether plant defense was changed. In addition, studies without herbivore or pathogens but that measure plant defense genes are included. Before we review these interactions we provide a brief introduction to PGPR and P. fluorescens in particular. Moreover, as we argue that the effect of PGPR on induced plant defense cannot be considered in isolation from the effects of other organisms that are also present in the soil such as nematodes, fungi, earthworms, or protozoa on the PGPR or on the plant, we also provide a brief overview of interactions between bacteria and other soil dwelling organisms in the rhizosphere."
} | 1,255 |
19900279 | PMC3091320 | pmc | 9,256 | {
"abstract": "A high-resolution genetic linkage map for the coral Acropora millepora is constructed and compared with other metazoan genomes, revealing syntenic blocks.",
"conclusion": "Conclusions A genetic linkage map, predominantly based on SNP markers derived from the transcriptome, has been constructed for a reef-building coral, Acropora millepora . This map has ample resolution for QTL analysis (3.4 cM) and represents the first linkage map for a coral, as well as for any non-bilaterian multicellular organism. The map will become the foundation for QTL analysis of adaptive traits and population genomics in the coral, to address the problem of coral evolution response to climate change, as well as for coral genome assembly. Comparative genomic analysis based on this map revealed a few statistically significant synteny blocks, which may reflect the features of ancestral metazoan genome organization. The specific mechanisms underlying such preservation are not yet clear, but represent an exciting area for future studies.",
"discussion": "Discussion SNP marker development in coral Molecular markers are useful tools for assessing important ecological and evolutionary issues such as connectivity, local adaptation, range shifts, biodiversity depletion, speciation, and invasion. Despite widespread concerns about the future of reef-building corals in the changing climate, genetic resources for corals remain scarce. The traditional ways of developing microsatellites or SNP markers are quite costly and time-consuming. Moreover, due to technical problems and low abundance in the genome, it has been shown that development of a large number of microsatellite markers in acroporid corals is particularly difficult based on the traditional microsatellite-enriched genomic library method [ 29 ]. Despite the advantages of SNP markers for a variety of tasks [ 30 ], their use in non-model organisms such as corals has been hampered primarily due to the costs of high-throughput SNP discovery and genotyping. With the introduction of the next-generation 454 sequencing technology, high-throughput SNP discovery is now feasible for any non-model organism. Our previous study [ 11 ], as well as others recently published [ 31 - 33 ], demonstrates a cost-effective way to produce a large number of gene-associated SNPs from transcriptome data obtained by 454 sequencing. Such gene-derived SNPs are particularly useful for non-model organisms, since they stand a better chance of identifying causal genes underlying complex traits in these organisms in the absence of genome sequence data [ 12 , 13 ]. The criteria that we used for SNP mining (at least 3× occurrence of the minority allele and at least 6× read coverage) are more stringent than those typically used (2× occurrence of the minority allele, and 4× or 5× read coverage) [ 11 , 31 , 32 ]. In our experience, the use of these stringent criteria enhances the success rate of marker development from 454 sequencing data. SNP genotyping via high resolution melting analysis Among the methods available for high-throughput SNP genotyping, the simple, fast and cost-effective HRM method is especially suitable for non-model organisms. The original HRM method requires one fluorescently labeled probe for each assay [ 34 , 35 ]. Later, this method was simplified by using an unlabeled probe in the fluorescent dye solution, but the 3' end of the probe still required costly chemical modification to prevent extension of the probe [ 36 , 37 ]. In this study, we further decrease HRM genotyping cost simply by adding two mismatched bases to the 3' end of an unlabeled probe instead of chemical modification. SNP marker transferability between populations Transferability of the assays to different populations is arguably the most important problem that may arise when trying to apply SNP markers to broad-scale population studies. The markers developed for one population may turn out to be appreciably polymorphic only in populations well connected to the original one, while being essentially homozygous in other, more isolated populations. The degree of connectivity between A. millepora populations between three reefs in the Great Barrier Reef (representing northern, middle, and southern regions) has been previously evaluated using allozyme markers [ 38 ]. Similar to nearly all coral species in that analysis, A. millepora demonstrated genetic subdivision among sampled sites (high F st values), although not without some connectivity (an estimated 5 to 30 exchanged migrants per generation). Oliver and Palumbi [ 39 ], on the other hand, detected strong barriers to connectivity over longer spatial scales (across Pacific archipelagoes) in two closely related species, A. cytherea and Acropora hyacinthus , using several intron- and mitochondrial DNA-derived markers that were developed for phylogeography applications. The study of the natural genotypic diversity and connectivity between A. millepora populations is of great interest for understanding the evolutionary responses of reef-building corals to ongoing climate change, and is among our high-priority research areas for the future. This emphasizes the importance of determining whether our SNP markers are polymorphic in other populations, or mostly represent 'private alleles' specific to the Magnetic Island (and perhaps even more specifically, Nelly Bay) population. Fortunately, in our interpopulation transferability test, most (65 to 75%) of the SNP markers we tested were polymorphic in just seven A. millepora colonies from Orpheus Island and Great Keppel Island, which are 80 km and 570 km away from Magnetic Island, respectively. Although this result suggests that the detected SNPs represent relatively common alleles in these populations, the distance between these populations is just a fraction of what was assayed in the Ayre and Hughes study [ 38 ], and so it remains to be seen how far this allele sharing extends. Still, this result is quite promising and suggests the potential for application of these SNP markers to inter-population studies of local adaptation in A. millepora . Mapping population For animals and plants with short generation times, very efficient mapping populations (second generation (F 2 ), backcross, recombinant inbred lines, double haploid, and so on) can be generated from the crosses among homozygous paternal strains or recombinant inbred lines, which usually requires multiple generations of sib-mating or self-fertilization. Despite several advantages of those methods, it would be very difficult, if not impossible, to produce such mapping populations in corals because most corals have long generation times (approximately 5 to 10 years in some corals, and 3 to 5 years in most acroporids), and the adult colonies are rather difficult to maintain. Last but not least, to our knowledge, synchronized coral mass spawning, an essential requirement for making genetic crosses, has never been recreated in laboratory-raised corals. In short, corals make poor laboratory models; however, this does not diminish the value of ecological and evolutionary questions pertaining to these organisms. Fortunately, previous studies have shown that A. millepora , like many other corals, is a highly heterozygous species [ 8 , 9 ]. Because of this, an outbred full-sib family would be a suitable mapping population for constructing a linkage map [ 40 - 45 ]. Although marker configurations are more complicated in such a family, they can be deduced after analyzing the parental origin and genetic segregation of the markers in the progeny (for a review, see [ 46 ]). In particular, coral larvae offer several key advantages over adult colonies for linkage mapping in that they are easy to obtain in great numbers, and, in this species, they do not contain algal symbionts, which would be a potential source of DNA contamination. Map density and recombination rate In the consensus map, marker density is dramatically variable across linkage groups, indicating that the protein-coding genes in A. millepora , like in human [ 47 ], are distributed very unevenly among chromosomes. This also suggests that including anonymous genetic makers into the current map will likely increase marker density in less populated linkage groups. The current genetic map covers 93% of the A. millepora genome and has a resolution of 3.4 cM, which should be sufficient for QTL mapping [ 48 , 49 ]. The average recombination rate across all linkage groups is approximately 7.5 cM/Mb in A. millepora , which is much higher than human (1.20 cM/Mb [ 50 ]), mouse (0.5 cM/Mb [ 50 ]), D. melanogaster (2 cM/Mb [ 51 ]), and even the plant Arabidopsis thaliana (5 cM/Mb; calculated based on data from The Arabidopsis Information Resource website [ 52 ]). This suggests that QTLs, if identified, can be narrowed down to rather small genomic regions in this coral species. Nine putative stress-related genes were mapped in the consensus map (markers colored red in Figures 1 , 2 and 3 ), and it would be interesting to see whether any of these are highlighted in future QTL mapping of adaptive physiology traits, such as heat tolerance. Moreover, SNPs in these genes might also prove useful for the study of allele-specific gene expression [ 53 ]. Last but not least, the high-resolution genetic linkage map would be invaluable for assembling the A. millepora genome, the sequencing of which is imminent (DJ Miller, personal communication). Gamete-specific recombination rates Differential recombination rates between sexes are widespread in animals and plants, with females often having more recombination and longer genetic maps than males [ 54 ]. Similar observations have also been reported in hermaphrodites, with greater recombination in female than male gametic tissue [ 41 , 55 , 56 ]. The underlying mechanism remains the subject of much debate, although several models have been proposed (for a review, see [ 57 ]). In this study, the length of the female map is 30% longer than that of the male map, suggesting that sex difference in recombination does exist in A. millepora . However, this difference seems attributable to only a few (that is, L4, L5, L6, L10 and L11), but not all, linkage groups. The 'haploid selection' model proposed by recent studies [ 58 , 59 ] seems to be the most plausible explanation for our observation. In the 'haploid selection' theory, sex differences in recombination result from a male-female difference in gametic selection. In coral Acropora spp., like in most animals, there is no female haploid phase, because meiosis is completed only after fertilization [ 60 ]. Since some genes (for example, genes responsible for meiotic drive systems) are expressed and under selection during the male haploid phase [ 61 , 62 ], this would tend to reduce recombination in males. If such genes were located in only a few chromosomes, this would be expected to reduce the amount of recombination observed in those chromosomes. Haploid selection might also explain the low polymorphism level of linkage group 8 in the male parent. Because the male parent was genotyped based on the sperm sample, it is possible that genotypes of some loci inferred from sperm mixtures are different from genotypes of adult tissues if these loci are subject to haploid selection. The significant low polymorphism level in L8 of the male parent may reflect strong haploid selection (for example, one of the homologous chromosomes corresponding to L8 might produce functional sperm, while the other might contain deleterious alleles that would produce non-functional sperm). Direct validation of this hypothesis would require tissue samples from the male parent, which are not available. However, the finding that more than half of annotated genes in L8 have putative roles in sexual reproduction supports the idea that this linkage group may be a target for haploid selection. Synteny analysis and permutation tests Synteny is defined as consistent linkage between certain genes across species. In the most general case, the definition does not require conservation of gene order or orientation. Previous comparative genomics studies have revealed synteny between distantly related metazoan taxa [ 63 , 64 ]. Most studies of genome evolution in animals have focused on bilaterian taxa for which extensive genomic resources are available [ 16 , 65 - 67 ]. More recently, the draft assemblies of the sea anemone and placozoan genomes have revealed substantial synteny between more distantly related metazoan taxa [ 68 , 69 ]. Our development of a genetic map for coral, which, to our knowledge, constitutes the first genetic map for a non-bilaterian metazoan, reveals the conservation of genomic organization among distantly related animal taxa. As the simplest free-living animals, placozoans represent a primitive metazoan form. A recent comprehensive phylogenetic study suggests that Placozoa are basal relative to all other non-Bilaterian animals ([ 70 ], but see [ 71 ]). Whole genome analysis of placozoan T. adhaerens shows that the placozoan genome has the lowest amount of local rearrangement relative to the common placozoan-cnidarian-bilaterian ancestor [ 69 ]. Previous comparative genome analysis revealed synteny blocks shared between placozoan and human genomes, which likely reflect ancestral features of the metazoan genome. In our study, we also found extensive synteny between coral and placozoan genomes (despite the incomplete assembly of the placozoan genome), suggesting that the coral genome also preserves many features of ancestral genome organization. Our preliminary synteny analysis identified numerous synteny blocks in each comparison between the coral map and other metazoan genomes. However, because of the number and positions of markers and their matching sequences within the two genomes, a substantial number of synteny blocks could be expected to arise by chance. Several methods to test for significant evidence of synteny between two genomes, based on randomly shuffled permutations of the real data, have been previously described [ 72 - 75 ]. The existing implementations of these methods are not well suited for our data (comparison of genetic maps and genome sequences across distantly related taxa), but are more applicable to comparative genome analysis of closely related species [ 19 ], because they require marker colinearity (that is, conserved marker order), and/or assume chromosome homology between chromosomes in comparison (for example, randomize markers only within a chromosome to evaluate significance of identified synteny). We followed a similar approach for our analysis, by randomly shuffling marker positions across the entire map and evaluating the likelihood that the number of synteny blocks, and the number of markers in each block, could have arisen by chance. Without any statistical tests, a simple analysis of synteny could be easily misinterpreted; for example, the large number of synteny blocks found in comparisons between the coral map and the worm and fly genomes (12 to 13 blocks in each comparison, with 3 to 10 markers per block) might suggest that the coral genome shared more structural similarities with worm and fly than with other animal genomes. However, permutation tests revealed that, in fact, neither of those comparisons found more synteny blocks than expected by chance (Table 3 ). There are several characteristics of genomic structure that would obviously be expected to affect the detection of synteny blocks by our criteria, including genome size, chromosome numbers, and the completeness of the assembly. Because the genomes considered in this study differed widely in these characteristics, this posed an important caveat for any conclusions drawn from these comparisons. Importantly, each of the comparisons between the coral map and another metazoan genome included at least one block that was significantly larger than expected by chance, based on permutation tests of block size (the number of markers within each block). Maintenance of synteny across great evolutionary distances If not maintained by natural selection, synteny would be expected to break down between distantly related taxa. One obvious factor that would affect this is the rate of genome rearrangement. Recent studies have shown that rates of chromosomal rearrangement are much higher in invertebrates than vertebrates [ 76 - 78 ]. For example, the rearrangement rates of Drosophila and Caenorhabditis are 350 to 850 and 1,400 to 17,000 times higher than those of mammals, respectively [ 77 ]. Our finding that the coral map and the worm and fly genomes share very little conserved synteny is consistent with these previous reports. Still, the worm and fly genomes do contain a small number (one and two, respectively) of synteny blocks (each including nine to ten genes), and these are significantly larger than expected by chance (Table 3 ). In general, eukaryotic genomes evolve by random micro- and macro-rearrangements such as indels, inversions and translocations [ 79 ]. Nevertheless, gene distribution in eukaryotic genomes is not random [ 80 ]. Several hypotheses have been put forward to explain synteny. Early research in genomic evolution and synteny assumed no selection for synteny, and suggested that synteny resulted from ancestral linkage groups that had not yet been disrupted by random chromosomal rearrangements [ 81 ]. The subsequent discovery that certain groups of co-regulated genes showed strict conservation of both gene order and linkage across taxa [ 82 ] refined this model by demonstrating that the co-regulation of a group of genes by local regulatory elements can drive conservation of synteny blocks containing those genes and their corresponding regulatory elements [ 83 ]. Recent studies have suggested an additional mechanism driving the conservation of synteny: the interdigitation of regulatory elements and their target genes by other genes with unrelated functions and regulatory pathways [ 84 , 85 ]. None of those proposed mechanisms provides a clear explanation for our findings. Several metazoan genomes showed more synteny blocks than expected by chance, but the gene functions suggested by sequence similarity for these syntenic markers were not linked in any obvious way. For example, the map includes one pair of genes that is linked in three species: LG5 of coral, chromosome 5 of worm, and chromosome 14 of human ([GenBank: EZ001917 ] and [GenBank: EZ012107 ]; Additional data file 1). There is no clear functional relationship between the genes associated with these markers (serine palmitoyltransferase 2, and enhancer of rudimentary homolog). Obviously this does not preclude the possibility of unknown functional relationships among the mapped genes, or of functional relationships between the other genes not included in the coral map. The list of syntenic markers associated with known genes also did not include any known examples of co-regulated genes (Additional data file 1). The identification of synteny blocks from the coral genetic map therefore provides no support for either explanation, but raises a number of interesting questions. Synteny blocks were distributed differently among taxa; for example, both fly and placozoan genomes showed conserved synteny with regions of LG2, but only the placozoan genome did with LG4 (Figure 5 ). The extent to which these differences are explained by selective pressures versus rates of genome rearrangements (for example, [ 77 ]) is not clear from our data, but this will probably become a more tractable question as genome sequences become available for a broader sampling of metazoan taxa. The extensive rearrangements evident within synteny blocks in the coral map (Figure 5 ) prompt questions about what mechanisms might account for conserved linkage but highly variable order. We speculate that selection might promote linkage between genes that must be modified in a correlated fashion to achieve an adaptive advantage (in other words, exhibit epistatic interactions). Linkage between epistatically interacting loci would allow for selection to operate on haplotypes rather than individual alleles [ 86 ], which would substantially improve the heritability of the evolving trait and hence the efficiency of selection. There are several pan-metazoan systems that can be viewed in terms of many correlated (or anti-correlated) traits determined by genes with otherwise unrelated functions. Examples include epithelial functions (rigidity, across-epithelial transport, along-epithelial connectivity, cuticle secretion, ciliation), cell-cell communication and nutrient exchange, and organism-wide transport and excretion. Future comparative analysis of genome sequence and function in the basal metazoans like A. millepora may help to elucidate the evolutionary origin of the pan-metazoan synteny."
} | 5,230 |
39483769 | PMC11523686 | pmc | 9,257 | {
"abstract": "This concept paper gives a narrative about intelligence from insects to the human brain, showing where evolution may have been influenced by the structures in these simpler organisms. The ideas also come from the author’s own cognitive model, where a number of algorithms have been developed over time and the precursor structures should be codable to some level. Through developing and trying to implement the design, ideas like separating the data from the function have become architecturally appropriate and there have been several opportunities to make the system more orthogonal. Similarly for the human brain, neural structures may work in-sync with the neural functions, or may be slightly separate from them. Each section discusses one of the neural assemblies with a potential functional result, that cover ideas such as timing or scheduling, structural intelligence and neural binding. Another aspect of self-representation or expression is interesting and may help the brain to realise higher-level functionality based on these lower-level processes.",
"conclusion": "8. Conclusions This paper gives a narrative that outlines structural components of simpler organisms that may have helped the human brain to evolve. More than that, the structures are so basic, they can be included in a computer model for Artificial Intelligence and are consistent with the author’s own cognitive model. The design may not be 100% accurate, but there appears to be a consistency about it and some biological and mathematical evidence can help to validate the theories. An early idea about scheduling through nesting may be seen in action in worms, for example, but in a simpler form. Then, one idea may be that intelligence can be realised automatically by converting from ensemble input to type-based output. This would occur automatically in the neuron network, where the realisation of types will produce some understanding and therefore intelligence. Amoebas are able to learn single types. The stigmergic processes of termites or ants, for example, have become interesting to explaining the neural structures for several reasons. Firstly, it is suggested that the neural microcircuitry is constructed primarily from the alignment of morphology or structure, rather than signal type and this includes synapse alignment and preparation. Although, the chemical signal will still change the type emitted by the cell. Secondly, the relationship between neurons and the substrate of glial cells, for example, also suggests stigmergic processes. It would be interesting if there is an underlying global memory structure to the brain, which is this perineuronal substrate and if it can abstract and even re-structure input signals. The uniformity of the substrate would allow it to communicate this to other modules and a computer model would be able to simulate it to some level. When modelling the biological structure, images may be stored as whole representations in the short-term memory, but when they are moved into long-term memory, they become tokenized and abstracted. One final idea is that the neural binding problem is constrained by current thinking about a holistic conscious and if it can be made more orthogonal and receive help from other organs, the problem will become much easier to solve. Most interesting then may be the idea that a cell or organism evolves, not only to survive, but also by expressing itself, where the expression is a result of its own internal structures and processes. In this respect, the memory substrate would be a precursor to our own natural language and this might also be seen in bees. The structural transformation from input to tokenized ensemble results in a communication process that is akin to a common language. The higher cognitive processes, if you like, have built themselves on the lower-level structures and processes.",
"introduction": "1. Introduction This paper describes some neural representations that may be helpful for realising intelligence in the human brain. The ideas come from the author’s own cognitive model, where a number of algorithms have been developed over time. Through developing and trying to implement the design, ideas like separating the data from the function have become architecturally appropriate and there have been several opportunities to make the system more orthogonal. Similarly for the human brain, neural structures may work in-sync with the neural functions, or may be slightly separate from them. Having more than 1 information flow actually makes the problem of how the human brain works much easier to solve. Another aspect of self-representation or expression is interesting and may help the brain to realise higher-level functionality based on these lower-level processes, maybe even natural language itself. The cognitive model is still at the symbolic level and so the neural representations are also at this level. The neuron discussion is therefore at a statistical or biophysical level rather than a biological one. The rest of the paper is organised as follows: Section 2 describes some related work. Then, the other sections discuss one of the neural assemblies with a potential functional result. Section 3 describes earlier work on a timer or scheduler. Section 4 describes how intelligence may be inherent in the neuron structure. Section 5 describes how the neural binding problem can be simplified. Section 6 describes some aspects of the author’s own cognitive model that have influenced the writing of this paper and Section 7 describes how natural language may have evolved naturally from similar structures. Finally, Section 8 gives some conclusions on the work."
} | 1,412 |
32257302 | PMC7062081 | pmc | 9,258 | {
"abstract": "Biofilms offer an excellent example of ecological interaction among bacteria. Temporal and spatial oscillations in biofilms are an emerging topic. In this paper, we describe the metabolic oscillations in Bacillus subtilis biofilms by applying the smallest theoretical chemical reaction system showing Hopf bifurcation proposed by Wilhelm and Heinrich in 1995. The system involves three differential equations and a single bilinear term. We specifically select parameters that are suitable for the biological scenario of biofilm oscillations. We perform computer simulations and a detailed analysis of the system including bifurcation analysis and quasi-steady-state approximation. We also discuss the feedback structure of the system and the correspondence of the simulations to biological observations. Our theoretical work suggests potential scenarios about the oscillatory behaviour of biofilms and also serves as an application of a previously described chemical oscillator to a biological system.",
"introduction": "1. Introduction Development of a complex biofilm provides several benefits to bacteria, including efficient nutrient distribution, defence from chemical attacks or, in the case of a floating pellicle on the surface of liquids, better gaseous exchange [ 1 ]. Biofilms are thus complex communities of bacteria and as such, many types of social dynamics come into play [ 2 , 3 ]. One of these is the division of labour [ 4 , 5 ]. The core of the biofilm growing on a solid surface shows a different metabolic state than the periphery. The periphery can freely access the nutrients from the surrounding environment. The interior, however, faces hindrance in obtaining a stable inflow of nutrients because the peripheral cells use up the nutrients that diffuse towards the interior. An experimental set-up to simulate that situation is provided by a microfluidics chamber [ 4 ]. An example of such a nutrient gradient is the production and diffusion of ammonia in the biofilm. Every cell in the biofilm has the ability to produce ammonia [ 4 , 6 ]. However, this small chemical compound is highly diffusive and therefore escapes into the environment as soon as it is produced by the cells in the periphery, thus leading to waste of nitrogen. In the interior, the ammonia produced by the cells diffuses out into the periphery. Thus, the interior cells monopolize ammonia production for the entire biofilm. Ammonia being an essential component of glutamine metabolism could be used to control the growth rate of the periphery by limiting its supply. The interplay between the inner and outer cells is required for glutamine synthesis and therefore the growth of the biofilm [ 4 , 6 , 7 ]. To understand biofilms more closely and make predictions based on empirical data, several models have been developed [ 4 , 7 – 12 ]. Liu et al. [ 4 ] observed oscillations in the biofilm, which they explained by different metabolic roles performed by the different compartments in the biofilm. They also established a model based on six differential equations. They defined two regions: the interior and periphery. Each of the regions has a variable representing glutamate and another representing the concentration of housekeeping proteins like ribosomal proteins. Ammonia and the active form of the enzyme glutamate dehydrogenase are also variables of the model. Since many biological oscillators have been described by less than six variables [ 13 , 14 ], a simpler model could be established for biofilm oscillations as well. Our ultimate aim was to develop a minimal model to describe the metabolic oscillations happening in a biofilm. Minimal models are the simplest way to describe a certain phenomenon with the least number of parameters [ 15 ] and this is in agreement with Occam's razor. For example, minimal models were established for glycolytic oscillations by Higgins [ 16 ] and Sel'kov [ 17 ] and for calcium oscillations by Somogyi & Stucki [ 18 ]. Here, we employ the smallest chemical reaction system showing a Hopf bifurcation [ 19 ], which was further analysed [ 20 , 21 ] and used to describe p53 oscillations [ 22 ]. At a Hopf bifurcation, damped oscillations turn into undamped oscillations [ 15 , 23 ]. In particular, Wilhelm & Heinrich [ 19 , 20 ] performed a thorough stability analysis of the model. We test to what extent the terms in this model match the processes in a biofilm system. In this analysis, we focus on the Hopf bifurcation, discuss the feedback structure and point out the correspondence of the simulations to biological observations. In our model, we use three variables only: ammonia, and interior and peripheral glutamate. Besides the quest for minimality, a reason for not considering the concentrations of housekeeping proteins as variables is that they change on a longer time-scale than metabolites. A similarity to the larger Liu model [ 4 ] is that, among the various amino acids, we focus on the metabolism of glutamate since glutamate and ammonia are both involved in the production of various amino acids through trans -amination, which is then equated to growth. To study the effect and possible benefit of oscillations, it is of interest to compute the average values of variables, as was done for several oscillators [ 24 – 29 ]. For linear differential equation systems showing oscillations (such as the system describing the harmonic pendulum), the average values equal the values at the marginally stable steady state. For nonlinear differential equation systems, the average values often differ from the values at the unstable steady state surrounded by the oscillations. However, there are some types of nonlinear systems for which equality holds, for example, Lotka–Volterra systems of any dimension [ 25 ]. The equality property has also been proved for some models of calcium oscillations [ 27 , 30 ] and the Higgins–Selkov oscillator [ 27 ]. Here, we probe the model employed for describing biofilm oscillations for the above-mentioned property.",
"discussion": "4. Discussion Here, we have used the smallest chemical system showing a Hopf bifurcation to model metabolic oscillations in B. subtilis biofilms. That model had been used earlier to describe p53 oscillations [ 22 ]. Here, we have applied the model to describe another biological phenomenon. We have specifically selected the parameter values to describe biofilm dynamics, which makes the model more relevant in the light of the biological observation. In our system, the diffusion of ammonia is critical for biofilm oscillations. All the terms in the model are linear, except k 2 AG p , which is bilinear. The model describes metabolic and diffusion processes as outlined above. As an output, the growth of the biofilm (consisting of incremental and halting phases) was also computed ( figure 4 ). A major reason of the observed oscillations was demonstrated to be the division of labour between the central and peripheral zones of the biofilm. While the release of usable ammonia is mainly delimited to the former, the production of biomass and, thus, growth, is mainly delimited to the latter. We have presented bifurcation diagrams, which clearly show supercritical Hopf bifurcations ( figure 5 and 6 , electronic supplementary material, figures S1–S3). Earlier, Wilhelm & Heinrich [ 19 , 20 ] had analysed that bifurcation and had presented one bifurcation diagram. Here, we have added some mathematical analysis. For example, we show the maxima and minima of oscillations, the knowledge of which gives us a quantitative insight into the biofilm dynamics. Moreover, we performed a QSSA and probed for the equality property of the average values. We analysed the Hopf bifurcation by changing not only the external glutamate G E but alternatively also all the rate constants except k 1 , since changing k 1 has the same effect as changing G E . Interestingly, a recent model [ 7 ] has shown a subcritical bifurcation in describing the behaviour of the stress levels in the biofilm periphery. However, they modelled the stress with a single delay differential equation and did not consider other molecular details, while we do not consider stress. Using a single differential equation meets the quest for minimality. However, a delay differential equation (meaning that the time derivative of a variable depends on that variable at a previous time point) is, from a mathematical point of view, very complex because it requires infinitely many initial values (from zero to the delay period, with a simplifying assumption being that they are all equal). Moreover, stability analysis is then considerably more complicated. Our model is complementary to their model. It is closer to Liu's original model [ 4 ] but much simpler because it involves only three rather than six variables, and thus only requires three initial values. We chose the parameters of the model such that they are in agreement with Liu's experimental results, namely the period and the amplitude of oscillations. In our model, peripheral glutamate exerts a positive feedback on itself. Mathematically, this has the form of a bilinear term involving peripheral and external glutamate concentrations. At very low values of G E , the feedback is not strong enough to enable a positive steady state. The system then tends to the TSS. In that state, G p is zero, so that growth is impossible. Biologically, this can be interpreted that the biofilm is too small to be viable. This is in agreement with observations in the recent study from the Suel group [ 7 ]. At a certain threshold value of G E (5.88 mmol l −1 ), the NTSS turns stable in a transcritical bifurcation. Beyond that value, the feedback is strong enough to enable growth. At high values of G E , the feedback becomes so strong that an overshoot occurs: more glutamate is taken up than needed, so that the G p level transiently exceeds the steady-state value. Then, more peripheral glutamate is consumed for release of ammonia or for growth, so that the concentration decreases again. This leads to oscillations. From a functional point of view, a steady state is quite appropriate [ 29 ]. Growth of the biofilm does not require oscillations. However, in this system, oscillations help in mitigating the chemical attack that challenges the biofilm [ 4 ]. This may have interesting clinical implications in view of treatment of biofilm-forming bacteria by antibiotics. Furthermore, another study [ 9 ] indicates that oscillations in growth actually help in sharing the nutrients among several biofilms more efficiently. However, not all biofilms show oscillations, indicating that it is not critical for biofilms. Our numerical calculations suggest that the average growth rate is lower when compared with growth at the metabolic steady state. By contrast, the individual concentration variables (ammonia, peripheral glutamate etc. but not biomass) show the equality property that their average during oscillations equals the value at the unstable state, as usual for Lotka–Volterra systems [ 25 ]. Here, we have shown that the bilinear input term k 2 AG p exhibits this equality property as well. This may come as a surprise because the ammonia and peripheral glutamate levels oscillate with a phase shift. In the paper by Liu et al. [ 4 ], the oscillations computed by their model have a sinusoidal shape. In our model, such a shape only occurs in the neighbourhood of the Hopf bifurcation. Further beyond it, the shape is more spike-like with the crests being sharper than the troughs. The question arises whether the model used and analysed here is minimal. On the basis of ODE systems (without delays), at least two variables are needed to generate oscillations [ 13 , 14 ]. However, when only linear and bilinear terms are included, at least three variables are needed, as was proved by Hanusse by an analysis of the Jacobian matrix [ 43 , 44 ]. As shown by Wilhelm & Heinrich [ 19 ], such a model requires at least five reactions. Thus, the above model is minimal in terms of the number of variables (criterion with highest priority) and number of reactions, given the type of kinetics. The famous two-variable Brusselator model [ 45 ] and the Higgins–Selkov oscillator involve a term of degree three each [ 16 , 17 ]. If the number of reactions is granted the highest priority, the model may look different. Thus, it depends on the criteria what a minimal model is. Note that a delay differential equation [ 7 ] is, from the viewpoint of the number of initial values, of infinite dimension. While our model is not necessarily the simplest, it provides a trade-off between simplicity and adequacy to match the observed oscillation in biofilms. As for any oscillatory system, it is interesting to elucidate the feedback structure. The term k 1 G E G p represents a positive feedback because peripheral glutamate stimulates its own uptake. This is because glutamate is a proxy for the concentration of various transport proteins embedded in the cell membranes. The higher the concentration of these proteins, the higher is the glutamate uptake rate. Since this positive feedback is the driving force for oscillations, at low values of k 1 G E , we observe a steady state rather than oscillations. In glycolytic and calcium oscillations, the cause of oscillations is also a positive feedback [ 13 , 14 , 17 , 46 ], while in a Goodwin oscillator, it is a negative feedback [ 41 , 42 ]. In addition to the positive feedback, there is also a negative feedback loop in the system ( figure 1 ). As seen from the differential equations, peripheral glutamate positively influences internal glutamate, which positively affects ammonia, which then negatively influences peripheral glutamate. Thus, the overall effect is inhibitory. This feedback structure of the Wilhelm–Heinrich model has been highlighted earlier [ 22 ]. By applying the QSSA, we have proved analytically that the limit cycle disappears if glutamate is degraded very fast or ammonia diffuses very easily. As mentioned in the Results section, the former case corresponds to a situation realized in experiments by overexpressing the glutamate dehydrogenase [ 4 ]. In that situation, no oscillations were observed. By contrast, in the case where oscillations occur, a description by a simple mass-action system requires three variables. Analyses in this direction may be relevant for clinical interventions via inhibition or activation of bacterial enzymes or changing diffusivity in the biofilm. The model analysed here has several pros and cons. In view of the mathematical analysis, its simplicity is certainly an advantage. In view of an adequate description of the biological and biochemical processes involved, the model may appear oversimplified. For example, describing growth by trilinear term is quite simplistic; usually, it is described by saturation kinetics (e.g. Michaelis–Menten kinetics). In addition, the assumption that glutamate uptake by the periphery is proportional to the glutamate concentration as explained above could be refined in future studies. Moreover, diffusion processes are usually reversible. In the above model, we neglected the backward processes in diffusion, which is justified if concentration differences are high. Many theoretical and experimental studies have been published on glycolytic oscillations [ 13 , 16 , 17 , 47 , 48 ]. However, these oscillations only occur under very special or even artificial conditions. In living cells, metabolic oscillations are rare, while being quite frequent in signalling systems [ 13 , 14 , 49 , 50 ]. The lights of a car are a helpful analogy: the headlights illuminate the street in a permanent way; there is no point in oscillations. By contrast, the side indicators (as signalling device) flash; that is, they emit oscillating light. Interestingly, in the case of biofilms, metabolic oscillations could provide advantages. While the work described here is quite theoretical, we consider it to be an appropriate basis for refined and more sophisticated models of biofilm oscillations."
} | 4,020 |
36303882 | PMC9590576 | pmc | 9,260 | {
"abstract": "Woody plant species represent an invaluable reserve of biochemical diversity to which metabolic engineering can be applied to satisfy the need for commodity and specialty chemicals, pharmaceuticals, and renewable energy. Woody plants are particularly promising for this application due to their low input needs, high biomass, and immeasurable ecosystem services. However, existing challenges have hindered their widespread adoption in metabolic engineering efforts, such as long generation times, large and highly heterozygous genomes, and difficulties in transformation and regeneration. Recent advances in omics approaches, systems biology modeling, and plant transformation and regeneration methods provide effective approaches in overcoming these outstanding challenges. Promises brought by developments in this space are steadily opening the door to widespread metabolic engineering of woody plants to meet the global need for a wide range of sustainably sourced chemicals and materials.",
"conclusion": "Concluding remarks and future perspectives Pivotal discoveries in herbaceous model systems such as Arabidopsis , tobacco, and tomato can help chart the course for research in woody species. Ongoing work in model systems is optimizing the delivery of multiple genes concurrently—which raises the possibility of circumventing multiple regeneration events for recalcitrant woody plants. Gene Assembly in Agrobacterium by Nucleic Acid Transfer using Recombinase Technology (GA A NTRY) uses multiple unidirectional site-specific recombinases to stack multiple genes in vivo. The GA A NTRY ArPORT1 strain of A. rhizogenes was demonstrated to be effective at stacking up to ten genes in rice (Hathwaik et al. 2021 ). As A. rhizogenes has proven effective at transforming woody plants (Gomes et al. 2019 ), the GA A NTRY system could prove useful for reconstituting metabolic pathways into woody plant chasses. Plant Artificial Chromosomes (PACs) are stable in plant cells and multiple genes on artificial chromosomes can be stably expressed. PACs have been the subject of research in non-woody plants such as barley, maize, and Arabidopsis , and promise remains for implementing PACs into woody plant species with established transformation and regeneration methods (Yu et al. 2016 ). Delivery of CRISPR-Cas9 systems is also being expanded, with viral delivery systems already developed and nanoparticle or nanotube delivery on the horizon (Demirer et al. 2021 ), facilitating the metabolic engineering effort in woody plants. Though gene regulatory networks have been extensively explored in herbaceous plants, considerable work remains to gain a more complete understanding of such regulatory networks in woody plants. The task of unraveling these networks in woody plant species is complicated by the diversity of tissue types present as well as the maturation time needed for such gene networks to emerge and be adequately explored. Poplar serves as a suitable woody model plant from which findings can be translated. Methods and knowledge developed in poplar can prove instrumental in exploring other woody plant systems to bolster our knowledge of their metabolic networks and to translate suitable approaches for metabolic engineering. Conquering the bottlenecks of transformation and regeneration in non-model woody plants will shorten the design-build-test-learn cycles in gene editing and transgene insertion to streamline their usage in metabolic engineering as well. Finally, a large amount of untapped diversity exists in domesticated cultivars across broad geographic ranges and non-domesticated genotypes. Comparative genomics can take advantage of the highly diverse genomes of woody species around the globe not only for breeding, but also to study the evolution of genes and gene families to facilitate gene mining for metabolic engineering (Tuskan et al. 2018 ). As woody plants are understudied, there is room to integrate current genetic and metabolic engineering strategies developed in other plant and microbial systems. Implementation of these strategies is finding increasing success in more studied woody plant systems such as poplar and citrus. Translating these tools into more woody plant systems not only leverages different native metabolic networks, but also expands the collection of ecosystem services available to the site of interest due to the broader species selection. These ecosystem services brought by woody plants are unique among plant chemical factories and include habitat, carbon sequestration, erosion reduction, and water retention. These ecosystem services can be viewed as the key benefits of a sustainable bioeconomy anchored in the used of long-lived woody plants. Low resource and labor demand further complement the vast ecosystem services of woody plant platforms. Metabolic engineering expedites our capability to simultaneously capitalize on the rich biochemical diversity of woody plants in tandem with the suite of environmental benefits associated with their usage. The transition from an oil-based economy to a sustainable bio-based economy will be facilitated by advances in this space.",
"introduction": "Introduction Woody plants are resilient perennials defined by their characteristic woody stems and large root systems. Woody plants are a highly diverse group that is polyphyletic in origin and have both flowering and non-flowering members. Many species of woody plants have evolved since their origins around 380 million years ago, and they have since come to dominate various landscapes around the globe (Wilson et al. 2017 ). In contrast to herbaceous plants that can employ the strategy of escaping and avoiding stresses by dispersing their seeds (Chelli-Chaabouni 2014 ; Sade et al. 2018 ), perennial woody plants must tolerate stresses in the areas they occupy. Woody plants invest a large amount of energy in vegetative growth (e.g. producing wood) and have well-developed xylem, phloem, and root system that enable them to survive diverse environments and stress conditions. In this regard, some monocot species such as palm, though lacking secondary growth characteristics of true wood, have substantial metabolic investment in vegetative growth that affords them resilience comparable to true woody plants. Approximately 80% of the biomass on Earth is stored in forests comprised of diverse woody plant flora. This immense genetic diversity remaining in woody plants serves as a precious deposit of untapped metabolic pathways that provides raw materials for metabolic engineering. The value and novelty of woody plant metabolism are paralleled in scale by their immense size and growth rate, raising promise for their use as a sustainable platform for metabolic engineering (Fig. 1 ). Because of the large biomass available from woody plants after establishment, high yield of desirable bioproducts can be achieved—yields that can be further expanded in some systems through short rotation planting or coppicing schemes (Ragauskas et al. 2006 ). Perennial woody plant cultivation can operate with less inputs, such as fertilizer, water, and labor, and can also adapt to different environments including land too marginal for conventional food crops, lowering their impact on the global food system. Woody plants are favored for biofuel production from lignocellulosic biomass as it is not derived from food crops, also lowering their strain on the global food system (Bryant et al. 2020 ; Choi et al. 2020 ). Moreover, the ecological benefits of woody plants are also considerable—contributing habitat, shade, erosion protection, and soil carbon sequestration. This suite of advantages makes woody plants a sound platform for metabolic engineering in the sustainable bioeconomy of tomorrow. Fig. 1 Prospects and current challenges of leveraging woody plants in metabolic engineering Despite the broad advantages of leveraging woody plant systems for metabolic engineering, they present unique challenges also (Fig. 1 ). The life cycle of woody plants is often long and prohibitive to engineering approaches that require multiple generations. In addition, many valuable phytochemicals produced by woody plants are low in accumulation—with some being synthesized only in response to certain stimuli due to their function in defense (Burlacu et al. 2020 ; Oleszek et al. 2019 ). The physiological and genetic understanding of woody plant species is also generally exceeded by that of more intensively studied herbaceous plants (Burlacu et al. 2020 ). Finally, some woody plants are difficult to transform and regeneration rates of many woody plants following transformation remains low, adding difficulty to engineering approaches that require genetic transformation. Nonetheless, there is current progress in the study of woody plant biology that alleviates these inhibitions to metabolic engineering. In this review, we will focus on current advances in woody plant metabolic engineering that has met some of the pressing challenges to their implementation; in addition, promising trajectories will also be discussed that chart the course for woody plant metabolic engineering in the near future."
} | 2,290 |
37722049 | PMC10523507 | pmc | 9,262 | {
"abstract": "Significance Long-term ecological stability plays a critical role in maintaining ecosystem functioning and sustainable delivery of ecosystem services in a varying world. While natural ecological communities can maintain stability through various mechanisms, little is known about how climate change might alter the effectiveness of these mechanisms and their consequences for ecosystem functions and services. Our study shows that the mechanisms historically governing community stability may be largely weakened as climate changes. Some may even flip to a destabilizing force and drive ecological communities to a state of much lower stability and higher sensitivity to environmental fluctuations. These mechanistic insights into community stability shifts can enhance our capacity to better forecast and prepare for upcoming community/ecosystem adjustments under climate change.",
"discussion": "Discussion Modification of Mechanisms Governing Community Stability. We found that as annual mean air temperature increased by 1°C and precipitation reduced by 40% over the past two decades, warming combined with prolonged, multiyear drought drove macroinvertebrate communities in a dryland stream to a new state of much lower stability and much higher sensitivity to environmental fluctuations. We show that historically, algae as stable primary producers providing a stable food source for consumers (macroinvertebrates), combined with asynchronous responses of consumers to climate variability ( Fig. 5 C and E ), gave rise to community stability as low variation in biomass and richness. The asynchronous responses occurred both at the family/genus level within the same order ( SI Appendix , Table S1 ) and across different orders/classes ( Fig. 2 A and B ). These mechanisms have also been commonly evoked to explain community stability in other ecosystems empirically ( 8 , 12 , 28 , 42 ) and theoretically ( 6 , 7 , 9 , 10 ). Furthermore, these mechanisms do not act independently. The stability of primary producers and their tight coupling with consumers support asynchronous responses of consumers to climate variability. Because of the stable and abundant supply of primary producers, the decreases of some groups of macroinvertebrates in a fluctuating environment could be compensated by the increases of other consumer groups, maintaining the overall community stability ( Fig. 5 E and G ). Fig. 5. Conceptual diagram of how various mechanisms buffer or amplify community sensitivity to climate variability. ( A and B ) Climate change alters the magnitude and frequency of climatic variability, creating new climatic and habitat niches in dry years by warming-drought feedbacks; ( C and D ) Shift from asynchronous species environmental responses to synchronous species responses; ( E and F ) Reorganization of species interactions and its amplifying effects by altered competition and delayed predators; and ( G and H ) transition from compensatory dynamics to synchronous dynamics as a result of new species-environment relationships and species interactions. In ( C – H ), blue lines and circles indicate taxa preferring wet years, red ones indicate taxa preferring dry years, and gray ones indicate taxa historically insensitive to the environment. Color arrows indicate the cascading effects of climate change. Within a new climate domain, however, these mechanisms became less effective or failed completely. We found cascading effects of climate change on the effectiveness of these stability mechanisms (color arrows in Fig. 5 ). The cascading effects were initiated by the synergetic effects of warming and drought, particularly the multiyear drought, which can dramatically alter habitats ( Fig. 5 B ). The extent of alteration likely exceeded tolerances of most species, triggering community-wide, nonlinear species responses to the environment beyond a threshold ( Fig. 5 D ). Such synchronized negative responses reduced species richness ( Fig. 5 F ) and total biomass ( Fig. 5 H ) in dry years. The reduction of species richness combined with the newly established wetland species led to reorganization of species interactions ( Fig. 5 F ). However, under the new climate regime, reorganized species interactions did not dampen but further amplified synchronized species responses to environment fluctuations. Given these cascading effects, long-term community dynamics shifted from compensatory dynamics to synchronous dynamics. Below, we discuss two key components of the cascading effects. The cascading effects are triggered by climate change altering habitats to the degree that most species might not thrive or survive in them, resulting in synchronized nonlinear species responses. Such dramatic habitat change can be created by pulse-type extreme events ( 21 ), synergetic effects of multistressors, and simultaneous effects of climate change from local and regional scales ( 43 ). As shown in our study, although the mean annual air temperature differences between a hot and a normal year are <2 °C ( Fig. 1 A ), water temperature differences can be amplified by up to 10 °C during the summer drought season if a hot year co-occurs with drought ( SI Appendix , Fig. S14 ). Drier years facilitate the growth of much more extensive macrophytes which slow down flows and increase evapotranspiration ( 36 , 37 ); thus, the rate of surface water loss by already high evaporation in hot years is exacerbated by rapid transpiration. This results in lower stream discharge and spatially intermittent flow of small, scattered patches, which can dramatically increase stream temperature and reduce dissolved oxygen concentration ( 40 , 44 – 46 ). Furthermore, at the watershed scale, the concurrence of warming and drought facilitated the expansion of shrubs/scrubs in replacement of evergreen forests ( SI Appendix , Fig. S7 and Table S3 ), likely reducing watershed water storage, groundwater level, and cold water refuge for species ( 35 , 47 ). In response to habitat change, all taxa were negatively affected by droughts ( Fig. 2 B ). While some taxa (e.g., Gastropoda and Coleoptera) increased their abundance in dry years when the degree of drought was within a certain range ( Fig. 2 A and SI Appendix , Fig. S11 C in 1989, 1999, 2011, and 2012), we found that above a threshold, their abundance started to decline remarkedly ( SI Appendix , Fig. S11 C in 2013 to 2016). This indiscriminate negative effect of drought substantially decreased taxa richness and promoted the establishment of macrophytes ( Fig. 3 ). With overall reduction in macroinvertebrate richness and the establishment of macrophytes, reorganization of species interactions further reduced community stability. We show that, in recent dry years, macrophytes outcompeted the historically stable algal community. The outcome of this competition dynamics might have induced the lowered abundance and richness of macroinvertebrate community in dry years (the right food web in Fig. 5 F ). More importantly, the decrease in macroinvertebrate taxa richness, induced by consecutive dry years, might have contributed to the dramatic increase in total biomass of macroinvertebrate community in wet years after dry periods. We observed that total biomass in recent dry years is much lower than that in similar dry years in the historical period, but the wet-year total biomass in the recent period is much higher than similar wet years in the historical period ( Fig. 2 C and D ). We speculate that such amplified biomass fluctuations are associated with reorganized species interactions. In consecutive dry years, species that prefer wet years disappeared ( SI Appendix , Fig. S11 A and B ). This contrasts with the historical period—while their abundance decreased in dry years, they persisted almost every year. Their persistence in dry years, although at low abundance, provides “storage” effects in the system ( 48 ), allowing these species to bounce up rapidly in wet years. However, such storage effects were weakened or even undermined under climate change, in which case, it might take longer for the removed species to return when conditions improve (e.g., a wet year). If these species played an important role in community stability, the loss of the storage effects can be consequential ( 1 , 49 ). For instance, we found that in the recent period, Odonata were first observed with a delay of ~1 mo compared to the historical period ( SI Appendix , Fig. S17 ). As large top predators, Odonata (>95% are Anisoptera) feed on many other macroinvertebrates (fish abundance is low, especially in the recent period and most are small minnows), including the often-abundant Diptera, Ephemeroptera, and Oligochaeta ( 50 , 51 ). Their delayed appearance might cause cascading trophic effects in wet years ( 23 , 40 , 50 )—the surge in the abundance of their prey, Diptera and Ephemeroptera, the two groups that are most sensitive to air temperature and hydrology ( SI Appendix , Table S4 and Figs. S8 and S10 and Fig. 2 B and , and the left food web in Fig. 5 F ). This is analogous to phytoplankton-zooplankton uncoupling as a result of delayed zooplankton observed in lakes with climate warming ( 30 ). Therefore, if changing climate significantly reduces species richness and abundance of functionally important species, the shrinkage of species storage effects might create a legacy effect on community composition in the following years, amplifying community fluctuation over a longer period. Comparing Long-Term Observations with Manipulative Experiments. Our long-term observations reveal cascading effects of climate change that significantly reduced community stability, mediated by decreased species richness and recently colonizing species. We show that reorganized species interactions further destabilized the community. This contrasts with findings from manipulative experiments that often show stabilizing effects of species interactions ( 24 , 25 ). Manipulative experiments and long-term observations are two powerful and complementary approaches to investigating how ecosystems work ( 21 ). Manipulative experiments excel at teasing apart confounding effects and examining causal effects of single or a few factors, while long-term observations provide the advantage of examining comprehensive and realistic effects. However, with regard to our research question—how climate change alters the effectiveness of stability mechanisms—so far, our understanding has been disproportionally derived from manipulative experiments. These experiments mostly manipulated one stressor and focused on local effects; therefore, they might not well represent synergetic effects of multidimensional climate change, the combined local and regional effects, or the colonization of newly arrived species. Our long-term study found that these factors working together could trigger the cascading effects of climate change on natural communities, which have rarely been captured in manipulative experiments. A common pitfall of long-term observations of natural communities is confounding factors. Particularly relevant to our study is the removal of cattle from the watershed in 2000 (see SI Appendix, Text S3 for detailed discussion on other potential confounding factors). The most significant response was the encroachment of macrophytes along the stream ( 37 ). We controlled this factor by including macrophytes in the causal networks of the recent period ( Fig. 4 B ); that is, the effects of the remaining variables were inferred after considering the confounding effect of macrophyte abundance. After controlling for the effect of macrophytes, the sensitivity of all other taxa to climate variability still increased markedly ( Figs. 3 B and 4 B , ⑨–⑪). In fact, we found that macrophyte abundance was influenced by hydroclimatic conditions—it was strongly, negatively correlated to the magnitude of winter floods ( Fig. 4 B , ④). Smaller and earlier winter floods in dry years ( Fig. 1 B and SI Appendix , Fig. S4 ) greatly facilitated seed germination and the expansion of macrophytes, consistent with previous findings in this system ( 36 ). The expansion of macrophytes and wetland plants under warmer and drier climate has also been reported in other streams around the world, particularly Mediterranean intermittent streams with dry summers like Sycamore Creek ( 52 , 53 ). Furthermore, the observed structural changes in macroinvertebrate communities such as the increase of predatory Diptera (e.g., Ceratopogonidae) and the decrease of Oligochaeta and Chironomidae ( SI Appendix , Table S4 ) in our study are remarkably consistent with whole-stream warming experiments ( 54 ), further confirming that the changes in community structure and dynamics reported here primarily resulted from climate change. With future climate projected to feature more frequent extreme events and nonanalogous regimes, findings from our study are particularly relevant and alarming. Taking advantage of the long-term ecological monitoring of a climate-sensitive ecosystem, we analyzed the mechanisms underlying community stability historically and the breakdown of these mechanisms under a changing climate. Multidimensional climatic stressors generated synergistic effects, and the resultant unprecedent habitat conditions invoked synchronous negative responses by taxa. This led to highly variable community dynamics from year to year and reduced taxa richness. While reorganizing species interactions might provide some buffer under species compositional change, we found that under a rapidly changing climate, when most species are already highly stressed, reorganized species interactions destabilized the community. In our study system, these multiple destabilizing mechanisms have maintained the community in a new, highly variable regime for a decade. However, it might be still too early to determine whether this new state constitutes a new stable state or a long transient state toward another state. As climate continues to change, the ecosystem might experience further loss of native species but will also receive new species that are better adapted to the drier and warmer climate projected for this region ( 55 ). Some of the newly arriving species might alter interactions and feedbacks, which can dramatically modify the ecosystem ( 56 ). In Sycamore Creek, recently colonizing macrophytes have increased sediment trapping and bank stabilization, resulting in conversion of some reaches (not our study reach) to riverine wetlands, and transforming braided channels in some other parts of the system into single-threaded ones ( 37 ). Such large-scale biogeomorphic changes might require decades to achieve, leading to significant modification of the physical habitat for the macroinvertebrate community ( 57 ). Over the course of continuing climate change at century timescales and with geomorphic–hydrologic–ecologic feedbacks occurring at multiple spatial and temporal scales, many ecosystems might undergo series of successive transient states before reaching a new, resilient, self-sustaining state. In the case of this stream, that state may be an ephemeral wash or arroyo with little or no aquatic life. Patterns emerging from our study system, such as synchronous, nonlinear species responses to environment fluctuations, and enhanced fluctuations of aggregate community properties, may serve as diagnostics of ecosystem transformation ( 6 ). Mechanistic insights from this study can help better anticipate and cope with ecosystem state changes in response to ongoing climate change worldwide."
} | 3,911 |
37288452 | PMC10243432 | pmc | 9,263 | {
"abstract": "Microbial electrosynthesis (MES) is a very promising technology addressing the challenge of carbon dioxide recycling into organic compounds, which might serve as building blocks for the (bio)chemical industry. However, poor process control and understanding of fundamental aspects such as the microbial extracellular electron transfer (EET) currently limit further developments. In the model acetogen Clostridium ljungdahlii , both direct and indirect electron consumption via hydrogen have been proposed. However, without clarification neither targeted development of the microbial catalyst nor process engineering of MES are possible. In this study, cathodic hydrogen is demonstrated to be the dominating electron source for C. ljungdahlii at electroautotrophic MES allowing for superior growth and biosynthesis, compared to previously reported MES using pure cultures. Hydrogen availability distinctly controlled an either planktonic- or biofilm-dominated lifestyle of C. ljungdahlii . The most robust operation yielded higher planktonic cell densities in a hydrogen mediated process, which demonstrated the uncoupling of growth and biofilm formation. This coincided with an increase of metabolic activity, acetate titers, and production rates (up to 6.06 g L −1 at 0.11 g L −1 d −1 ). For the first time, MES using C. ljungdahlii was also revealed to deliver other products than acetate in significant amounts: here up to 0.39 g L −1 glycine or 0.14 g L −1 ethanolamine. Hence, a deeper comprehension of the electrophysiology of C. ljungdahlii was shown to be key for designing and improving bioprocess strategies in MES research.",
"conclusion": "Conclusions The deeper understanding of the morphological state provides control over the MES performance of C. ljungdahlii . This is a key for further engineering and improvement in performance and to expand the range of products. In this study, glycine and ethanolamine were produced for the first time in MES from CO 2 , and by this bacterium. This highlights the importance of controlling and understanding electroautotrophic physiology not only as a production platform itself, but also as an exploratory tool to evaluate and extend the metabolic capacities of microbes capable of gas fermentation. 76 This study provides the first experimental indication of the RGP and the GSRP activity in C. ljungdahlii . Further biochemical investigations are now required to detail the metabolic source of these amino compounds and to explore, if a tailored as well as enhanced production of these metabolites is feasible. Additionally, the presented advances in understanding the drivers of microbial catalytic activity can now be combined with more traditional efforts for process improvement like reactor and electrode engineering such as the use of chemically or structurally modified electrodes 77–81 in various reactor scales and configurations. 82–86 Hence, the innovation cycle can be closed and detailed biological investigations, which are only possible in a well-controlled pure-culture BES, can pave the road for future process as well as reactor engineering.",
"introduction": "Introduction The important commitment towards decarbonization made during the 26 th United Nations Climate Change Conference of the Parties (COP26) represents a turning point. Countries must undertake political and economic changes to reduce greenhouse gas emissions, and double their efforts to mitigate climate change and its effects. From this situation an opportunity has emerged to create and improve key technologies enabling the transition to a bio-based and circular economy. 1–3 Thereby, the substitution of fossil hydrocarbons for chemical production is essential but remains a challenging task. Most of the current strategies to address the substitution of hydrocarbons in the chemical industry are based on the utilization of bio-derived feedstock that are transformed into commodity chemicals through chemical or biotechnological ways as well as combinations thereof. 4–6 Increasing attention is paid to the direct transformation of carbon dioxide (CO 2 ) into commodity chemicals, 7 as advancing such technologies will be key to a sustainable bioeconomy. 8,9 Microbial electrosynthesis (MES) 10 offers a biotechnological approach based on using electricity to provide the electrons needed for microbially catalyzed CO 2 reduction. MES from CO 2 is biochemically related to C 1 -gas fermentation, 11 but here reducing power is not delivered into the system as hydrogen gas (H 2 ) but directly as electric current using negatively poised electrodes ( i.e. cathodes) within a bioelectrochemical system (BES). Different microbes – being termed electrotrophs – are capable of consuming electrons from a cathode. This ability was first proven with Cupriavidus necator , 12 by electrolytically generating H 2 and oxygen, and since then many other electrotrophs have been found, as nicely reviewed here. 13,14 Some electrotrophs are also capable of using these electrons to reduce CO 2 and grow chemolithoautotrophically on CO 2 and current as the sole sources of carbon and energy. These can be denominated as electroautotrophs. The first reported electroautotrophic microbe was Sporomusa ovata . 15 In recent years, research on MES with electroautotrophs (mainly acetogens like Sporomusa spp. and Clostridium spp.) has shown the potential to produce commodity chemicals from CO 2 and electricity, but it has also highlighted the technological and biological limitations. 16 Drawbacks such as low product titers and coulombic efficiencies (CE), poor selectivity, challenging scale-up, and a very limited and economically unattractive product portfolio currently keep the field at an impasse. 17 Key to overcoming this overall lack of performance, is a much deeper understanding of the physiology of electroautotrophs. Very often the focus of developments of MES from CO 2 is on the technical side, developing new electrode materials, reactor configurations, or operational conditions. 18,19 Meanwhile, the biological side is often treated as a “black box” where undefined microbial communities (microbial mixed cultures or reactor microbiomes) are preferred, because they are easier to handle, more robust, and the associated costs are lower. 20 On the other hand, the use of pure cultures, although more challenging from the biological side, increases product specificity, reproducibility, and could potentially expand the range of products beyond the current limited portfolio. 21 Although there is also great interest in the production of methane (CH 4 ), 22–24 the main products obtained by MES from CO 2 are short-chain fatty acids and their corresponding alcohols (acetate/ethanol, butyrate/butanol, caproate/hexanol). 25–28 The conversion of CO 2 to acetyl-CoA, which then leads to the production of the aforementioned compounds, proceeds most often through the Wood–Ljungdahl pathway. Acetogens, as the main users of this pathway, utilize carbon monoxide (CO) or H 2 as the electron donor for reducing CO 2 during gas fermentation. 29 However, during MES, it is still not experimentally verified if the cathodic electrons are channeled into the cell via membrane-associated redox components like cytochromes, 30 nanowires 31,32 or unknown means of direct extracellular electron transfer (DET), or by using diffusible electron shuttles like H 2 , 33 flavins 34 or quinones 35 in mediated extracellular electron transfer (MET). In our study, we clarify the electrophysiology of Clostridium ljungdahlii , as a model acetogen, at electroautotrophic conditions. This bacterium was initially reported to be able to perform DET 25 but later studies rather suggested H 2 mediation 17 as the predominant mechanism of extracellular electron transfer (EET). More insight about the contribution of different EET mechanisms, as well as its influence in bacterial metabolism and growth, was essential to achieve a deeper comprehension about the physiology of this model microbe for MES from CO 2 . Ultimately, this allowed for higher throughput, better control over the biological aspects of the process and a broader range of products.",
"discussion": "Discussion We established a robust and controllable microbial electrosynthesis converting CO 2 and electrons to acetate ultimately outperforming previous studies ( Table 1 ). The maximum acetate titers (6.06 g L −1 ) were the highest obtained in MES using pure cultures, surpassing the best Clostridium spp. 49 and Sporomusa spp. 56 results, although it was around half of the best titers achieved with mixed cultures. 58,59 Noteworthy, this was achieved without any electrode modification or reactor optimization and only optimizing the biological part of the process. Although techno-economical evaluations of MES showed that its technology readiness level (TLR) corresponds to lab scale yet, 18,28,61,62 efforts are currently being made to increase the TLR of this technology by academia as well as the private sector. Framed in that effort, our research set the foundations for a better control and understanding of the biocatalysts used in MES, which combined with the available technological advances could reduce the costs and improve the performance of the technology. Nevertheless, the main goal of the study was to set the basis for the understanding of electroautotrophic metabolism of C. ljungdahlii , and related bacteria. Constant flushing with inert gas of the electrolyte solution avoided H 2 accumulation, which would have led to a process being similar to gas fermentation, and allowed the study of the electroautotrophic metabolism. The control of the microbial phenotypes of C. ljungdahlii by means of the cell density of the inoculum was the key to unveil the role of H 2 as the mediator for extracellular electron transfer, and to de-couple electroautotrophic growth from biofilm formation in MES. The phenotype control was highly robust being barely affected by stochastic factors, such as the electrode characteristics and high variance of currents and current densities, partially induced by random differences in the growth level of the inoculum. Acetate was produced electroautotrophically by C. ljungdahlii reaching a maximum titer of 6.06 g L −1 , with a productivity of 0.11 g L −1 d −1 –2.93 g m −2 d −1 . Acetate titers were significantly higher in the planktonic – than in the biofilm-dominant condition (Fig. S13 † ). Ethanol was likely also produced, but flushed out of the systems due to its volatility. Test experiments lowering the gas flow through the system allowed for ethanol accumulation in the liquid phase and prevented H 2 loss (data not shown), which could further increase the efficiency. In previous publications, 2-oxobutyrate was detected in low concentrations, 25 probably as a consequence of punctual redox imbalances linked to the activity of the glycine synthase. 63 But this compound was never detected in our experiments (<1 mg L −1 ). Glycine and ethanolamine were produced at higher titers in the biofilm-dominant condition (up to 0.39 g L −1 and 0.14 g L −1 , respectively), possibly due to a redox imbalance caused by a surplus of NADPH. RGP and GSRP are known to require more NADPH than reduced ferredoxin compared to the Wood–Ljungdahl pathway, making them less efficient but also more resilient under low energy conditions. 64 Our genome-scale modelling points to a redox imbalance between NADPH and reduced ferredoxin as the mechanism behind the observed metabolic shift (Fig. S12 † ). It seems similar to the metabolic shift during solventogenesis, in which ethanol production is promoted by NADH accumulation. 65 Formate was produced abiotically via CO 2 RR, which was found to proceed via in situ formed metal electrocatalysts by electroreduction of media components. Formate could be consumed in C. ljungdahlii only with further H 2 reducing equivalents bypassing the formate dehydrogenase and saving reduced ferredoxin, presumably leading to increasing performance. However, the rate of formate accumulation in the biofilm-dominant condition during starvation cannot only be explained by abiotic CO 2 RR, since the concentration exceeded the provided electrochemical reducing equivalents by 2.5-fold. Another biotic source for formate release to the solution must be involved here. Furthermore, formate accumulation in this case could be another sign of H 2 starvation, since C. ljungdahlii cannot grow solely on formate, 66 unlike other acetogens. 67 Combining different analytical approaches, H 2 was shown to be required for efficient MES from CO 2 using C. ljungdahlii . The growth of planktonic cells is physically incompatible with DET and the observed cease of activity when the potential was too positive for H 2 evolution, strongly indicates that H 2 mediates electron transfer in C. ljungdahlii , or at least plays a dominant role. In the case of biofilm-dominant conditions, activity followed the same pattern and also ceased at a similar cathode potential as in the planktonic-dominant reactor. Further, no indications pointing to other redox processes being relevant for EET were detected in CV analysis (Fig. S10 and 11 † ). Nevertheless, H 2 in the off-gas and dissolved in the catholyte was still always detected albeit at lower shares in the biofilm-dominant conditions compared to the planktonic-dominant conditions. This suggests that H 2 was still formed but was immediately consumed by the biofilm. An inconspicuous H 2 mediation by the activity of uncharacterized hydrogenases 68 able to efficiently channel H 2 into the cell immediately after being generated at the cathode or even to promote transient H 2 evolution by keeping a low H 2 partial pressure in immediate electrode proximity 69 are possible reasons for this observation. Yet, this immediate H 2 consumption should translate into higher CE for the biofilm-dominant conditions, since the H 2 would be metabolically utilized instead of “wasted”. However, this is not the case, as CE were always higher for the planktonic – than for the biofilm-dominant reactors. Besides, the higher H 2 production observed in the planktonic-dominated reactors compared to the abiotic ones indicates a more active promotion of H 2 evolution by the planktonic cells. A direct reduction of ferredoxin (with a similar redox potential to HER) is not possible, since ferredoxin is located in the cytoplasm. Unknown cellular components could potentially form a dedicated extracellular electron transfer machinery to ferredoxin, but there is no evidence for such a process. A DET process would likely result in a specific redox signature in the potential range more negative than the formal potential of ferredoxin in the CV ( E ′ 0 = −420 mV vs. SHE), 70 for which there is no indication. Other putative electron mediators like flavins or quinones 71 were examined but not found in the culture supernatants (data not shown). The electrochemical activity of media components such as resazurin, 72 cysteine 73 and yeast extract 74 have been reported, but involving more oxidative potentials than the ones used in this study, which is supported by our CV data. Nevertheless, there could be a minor influence of yeast extract in bacterial growth, while cysteine could be converted to acetate even if it cannot support growth. 75 Overall, the possible effect of these component was found negligible as evidenced in control runs at OCV (Fig. S14 † )."
} | 3,883 |
37122845 | PMC10139955 | pmc | 9,266 | {
"abstract": "Highlights • Methanobrevibacter boviskoreani JHI T , a rumen methanogen, can grow without pressurised hydrogen, using a variety of short chain alcohols. • Results demonstrate that in the absence of hydrogen, JHI T can use ethanol, propanol and butanol but not methanol, as a source of reducing potential for methanogenesis. • The ability of JHI T to use ethanol to drive methane production makes it a model rumen methanogen to use in high throughput culture-based bioassays. • Results demonstrate for the first time a direct inhibitory effect of the bacteriocin nisin on a rumen methanogen. • Methanobrevibacter boviskoreani JHI T is a useful methanogen culture for researchers looking to conduct high-throughput experiments with rumen methanogens.",
"introduction": "Introduction Methane emissions from ruminant livestock are the single largest source of agricultural greenhouse gas emissions globally and the development of methanogen inhibiting technologies that do not impact animal productivity are an active area of research ( Mizrahi et al., 2021 ). In the rumen, hydrogenotrophic methanogens use H 2 or formate as a source of reducing potential to reduce carbon dioxide (CO 2 ) to CH 4 , thereby generating membrane potential for ATP biosynthesis ( Attwood et al., 2020 ). Hydrogenotrophic Methanobrevibacter species are the dominant rumen methanogens ( Henderson et al., 2015 ) and their H 2 consumption ensures that fermentation remains thermodynamically favourable ( Janssen 2010 ; Ungerfeld 2020 ). Methanogens can be difficult to culture and rumen Methanobrevibacter spp are strictly anaerobic and dependent on H 2 for growth. Consequently, culture vessels are pumped with H 2 :CO 2 (4:1) and pressurised to 180 kPa to simulate the level of dissolved H 2 within the rumen ( Leahy et al., 2010 ). This makes it challenging to develop large scale or high-throughput culture-based assays that can be used to screen for anti-methanogen activity. Such assays are essential to test the ability of chemical compounds or microbial metabolites such as bacteriocins for the ability to inhibit methanogen growth, or to help select strains suitable for development as direct-fed microbial cultures that could limit CH 4 emissions ( Doyle et al., 2019 ). Methanogenic archaea that can grow on CO 2 and ethanol or isopropanol as sole energy sources have been described ( Berk and Thauer, 1997 ) and it has been shown that H 2 and formate may not be the only sources of reducing potential for some hydrogenotrophic Methanobrevibacter species ( Leahy et al., 2010 ; Weimar et al., 2017 ). Analysis of the genome sequence of the rumen methanogen Methanobrevibacter sp. AbM4 identified genes for alcohol and aldehyde dehydrogenases ( Leahy et al., 2013 ), which suggested that AbM4 may be able to use alcohols, such as methanol and ethanol, as an alternative source of reducing potential for methanogenesis. The supplementation of both ethanol and methanol at 20 mM allowed AbM4 to grow without pressurised H 2 ( Weimar et al., 2017 ). The presence of homologues of these genes in other methanogens has been shown to allow ethanol to be used as a source of reducing power for CH 4 production ( Hoedt et al., 2016 ; Kelly et al., 2019 ). The AbM4 culture is not available from a public culture collection and among the sequenced rumen methanogen genomes, Methanobrevibacter boviskoreani JH1 T ( Lee et al., 2013b ) has the highest degree of similarity to AbM4, with a 16 S rRNA identity of 99.8%, indicating that they belong to the same species. M. boviskoreani JH1 T was previously reported to use H 2 /CO 2 and formate/CO 2 for growth, but not ethanol or methanol ( Lee et al., 2013a ). However, the NADP-dependent alcohol dehydrogenase, aldehyde dehydrogenase and F 420 -dependent NADP reductase genes are present in the M. boviskoreani JH1 T genome and suggest it may have the ability to utilize alcohols for reducing potential in the absence of H 2 . The aim of this study was to explore and optimize the growth of M. boviskoreani JH1 T without H 2 to support the development of a high throughput bioassay enabling screening of this methanogen against compounds or microbial cultures with potential anti-methanogen activity.",
"discussion": "Discussion Methane resulting from enteric fermentation in ruminant livestock is a major source of agricultural greenhouse gasses. Consequently, the need to control CH 4 emissions has become a priority resulting in increased urgency to develop strategies that can practically mitigate CH 4 from ruminant animals ( Reisinger et al., 2021 ). Several different approaches have been used to specifically target methanogens in the rumen including the use of feed additives, certain seaweeds, and development of anti-methanogen vaccines ( Beauchemin et al., 2020 ). The use of microbial cultures (particularly lactic acid bacteria) or their metabolites has also been proposed ( Doyle et al., 2019 ). However, methanogens are difficult to culture, and a major limitation is the inability to perform high throughput screening of cultures or their metabolites to select those capable of methanogen inhibition. Therefore, new tools are needed that can contribute to the acceleration of CH 4 mitigation research ( Leahy et al., 2022 ). Here we have shown that M. boviskoreani JH1 T can grow without H 2 in the presence of ethanol, 1-propanol, or 1-butanol, and that supplementation with ethanol was able to greatly enhance the growth of M. boviskoreani JH1 T in the absence of pressurised H 2. Ethanol is a by-product of rumen fermentation, with the normal physiological concentration of ruminal fluid ethanol being 556 +/- 450 μM according to the Bovine Metabolome Database ( Foroutan et al., 2020 ). Therefore, it was surprising to find M. boviskoreani JH1 T could grow at ethanol concentration more than 200-fold above the physiological concentration, suggesting the M. boviskoreani JH1 T may have evolved as a specialized ethanol consuming methanogen or is able to partner with ethanol producing bacteria ( Ungerfeld, 2020 ). The ability of M. boviskoreani JH1 T to grow and produce CH 4 in the absence of pressurised H 2 enabled the development of a microtitre plate-based methanogen inhibition bioassay. We tested this bioassay using the bacteriocin nisin and showed that purified nisin was able to successfully inhibit the growth of M. boviskoreani JH1 T . Although nisin has previously been shown to inhibit a non-rumen methanogen ( Methanobacterium ) using an agar diffusion assay ( Hammes et al., 1979 ) this is the first time that such a direct effect of nisin on a rumen methanogen has been demonstrated. We plan to use this bioassay to screen metabolites from rumen microbes and lactic acid bacteria ( Doyle et al., 2019 ) for the ability to inhibit methanogens. Consequently, we believe that this Methanobrevibacter culture will be a useful tool to develop high-throughput screening approaches targeted towards inhibition of rumen methanogenesis."
} | 1,754 |
38352785 | PMC10861976 | pmc | 9,267 | {
"abstract": "Due to soils from arid regions with high lime and low organic matter content, farmers receive low yields along with high costs of agricultural inputs, which causes them to look for a solution. In this context, Arbuscular mycorrhizal fungi (AMF) have great potential to reduce fertilizer use by mediating soil nutrient cycles. However, little is known about studies of AMF inoculum on microbial biomass carbon (C), nitrogen (N), and phosphorus (P) cycling during vetch plant vegetation in calcareous areas. In this study, changes in soil biogeochemical properties related to soil C, N, and P cycling were investigated with five different AMF inoculations under vetch (common Vetch (CV; Vicia sativa L.) and Narbonne Vetch (NV; Vicia narbonensis L.) growing conditions. For the field study, a total of five different mycorrhizae were used in the experiment with the random plots design. AMF inoculation decreased the lime content of the soil, and the highest decrease was observed in NV with Glomus (G.) intraradices + G. constrictum + G. microcarpum inoculation (24.41 %). The highest MBC content was recorded in CV vetch G. intraradices (1176.3 mg C kg −1 ) and the highest MBN content in NV vetch G. intraradices + G. constrictum + G. microcarpum (1356.9 mg C kg −1 ). CAT activity of soils was highest in CV vetch G. intraradices (31.43 %) and lowest in NV vetch G. intraradices + G. constrictum + G. microcarpum (72.88 %), urease enzyme activity decreased in all treatments except G. constrictum + Gigaspora sp. and G. mosseae inoculations in CV. The highest DHG activity was detected in GF (15.72 %) AMFs in CV and GI (21.99 %) in NV. APA activity was highest in Glomus constrictum + Gigaspora sp. (23.33 %) in CV and Glomus fasciculatum (10.08 %) in NV. In CV plots, G. intraradices + G. constrictum + G. microcarpum (91.67 %) isolates had the highest and G. intraradices community had the lowest RC% (97.33 %) in mixed mycorrhiza species, while in NV plots G. fasciculatum inoculum had the highest and G. intraradices community had the lowest RC%. This study has important implications for the application of AMF for sustainable agriculture. When the results of the study were evaluated, the most effective AMF isolates in terms of C, N, and P cycles were G. constrictum + G. fasciculatum + Gigaspora sp. in Common vetch variety, and G. intraradices in Narbonne vetch variety.",
"conclusion": "5 Conclusion AMF inoculum to the soil under Common Vetch and Narbonne Vetch cultivated areas decreased the Lime and EC content of the soil but had no significant effect on pH due to the short-term cycle. Agricultural management practices based on AMF application can provide an economical, environmentally friendly, and sustainable way to improve soil fertility and yield. While positive effects on soil biochemical C, N, and P cycles were observed, it was observed that the rate of effect varied depending on the plant and AMF type. Especially G. constrictum + Gigaspora sp. in Common Vetch and GI AMFs in Narbonne Vetch were more effective than other treatments. The effects on soil enzyme activities (APA, CAT, DHG, Urease), MBC, and MBN were generally positive. The results of this study confirmed that AMF can release nutrients from complex materials by increasing soil enzyme activities. This study very clearly demonstrated that AMF could increase soil enzyme activity, which in turn can improve nutrient cycling. In our study to make suggestions for future studies, it is seen that the highest root colonization rates can be achieved with the inoculum of Glomus intraradices + G. constrictum + G. microcarpum AMF isolates for Common Vetch and G. fasciculatum isolates for Narbonne Vetch.",
"introduction": "1 Introduction Producing crops is a very laborious task for farmers, and they face the problems of being infertile poor soils, heat and drought stress, increasing production costs every year, and inadequate product prices. In this context, increasing fertilizer efficiency, reducing the amount of fertilizer, encouraging high rooting of the plant and easier and more efficient access to water are of vital importance. Mycorrhizae are the most widespread soil microorganisms forming a symbiotic relationship with more than 80 % of plants [ 1 ] and they can be found in diverse ecosystems worldwide [ 2 ]. Arbuscular Mycorrhizal Fungi (AMF) constitute the largest group among the different mycorrhizal groups, and they are AMFs that form mycorrhizal associations with plants colonized in the roots and establish a mutually beneficial relationship [ 3 ]. One of the benefits of symbiosis for plants is the resilience against stress factors such as metal toxicity, drought, and salinity, which enables some plant species to grow in harsh conditions. The enhanced water and nutrient availability provided by AMFs [ 4 ], even in nutrient-poor or dry soils, can positively affect other soil organisms, including bacteria and other fungi, by improving their nutrient supply and promoting their growth and activity, increasing the stress tolerance of plants, and extending root longevity [ 5 ]. Furthermore, AMF extraradical hyphae also play an important role in signaling and nutrient exchange by forming communal mycorrhizal networks between neighboring plants [ 6 ], and are therefore important functional groups for plant growth and soil quality maintenance [ 7 ]. AMFs such as G. intraradices increased the concentrations of organic acids such as proline and isocitrate [ 8 ]. G. constrictum had positive effects on photosynthetic pigments, gas exchange parameters, antioxidant enzymes, and nutrition of pepper plants grown under salt-stress conditions [ 9 ]. In banana plants inoculated with 2500 spores of G. microcarpum and F. mosseae, more than 80 % root colonization, increase in leaf chlorophyll content, leaf N, P, and K, significant decrease in soil pH, increase in soil available phosphorus and organic carbon were observed [ 10 ]. Organic matter positively affects the physical, chemical, and biological properties of soil [ 11 ], which is very important both because of this feature and because it is the largest terrestrial carbon reserve (about 0.68 Eg, i.e. 0.68 × 10 18 g organic carbon) [ 12 ]. AMF hyphae are involved in soil C translocation and provide a key link in the terrestrial C cycle and thus play a crucial role in the global C cycle [ 13 , 14 ]. Indeed, AMF is an effective agent to improve carbon sequestration in the mechanism of translocation of C from high respiratory activity around the root to soil aggregates. They play a very important role in the hydrolysis of high-molecular-weight N-containing organic compounds of plant litter and soils to NH 4 + in the regulation of N biogeochemical cycles in natural ecosystems [ 15 ]. They also exude large amounts of lytic enzymes and organic acids, which release recalcitrant organic and mineral nitrogen into the soil. These processes can bypass organic nitrogen mineralization by free-living decomposers, effectively short-circuiting soil–plant nitrogen cycling [ 15 ]. Also, AMFs are P activators that can accelerate the process of converting P into bioavailable forms through a series of chemical reactions and biological interactions [ 7 ]. In recent years, an increasing number of studies have recognized that the outcomes of plant-AMF interactions are on a continuum ranging from mutualism to parasitism, depending on the context in which interactions occur [ 16 , 17 ]. Vetches (Vicia sp.) is an important source of protein, minerals, vitamins, flavonoids, etc. in animal nutrition [ 18 ]. Narbonne vetch (V. narbonensis L.) ranks among the most important vetch species worldwide. Vetch crops affect the quantity and diversity of soil microorganisms due to their developed root system and low C:N ratio. Root secretions released by vetch plants, consisting of various organic compounds, can serve as a food source for soil microorganisms. These secretions promote the growth and activity of beneficial soil bacteria and fungi (such as nitrogen-fixing bacteria called rhizobia), increasing their populations in the rhizosphere [ 5 , 19 ]. Common vetch (Vicia sativa L.) is an annual legume grown as green manure and animal feed that provides rapid soil cover, and has the ability to increase soil moisture and organic matter content and reduce soil erosion [ 20 ]. These crops are particularly environmentally friendly, and their use is recognized as an important management practice with the potential to reduce dependence on mineral fertilizers and maintain soil organic matter content [ 21 ]. Since the availability of C substrates largely controls microbial growth in soil, green manure amendments promote microbial growth and activity in soil. Furthermore, legume-based green manures such as vetch and alfalfa are important sources of N in organic crop production [ 22 ]. Studying the response of soil microbial community and enzyme activities to warming provides a better understanding of soil biochemical processes under global warming [ 23 ]. The positive effects of AMF on improved soil fertility and plant community succession are well known [ 24 ]. However, the specific effects of AMFs on plant biomass, mycorrhizal colonization rate, soil microbial biomass, and soil enzyme activity may depend on factors such as soil conditions, environmental factors, and the cultivated varieties of plants. There are many studies on the effects of different AMF isolate treatments on soil and plant properties. However, studies on the use of AMFs with different vetch varieties in semi-arid climatic zones are very limited. This study aimed to investigate changes in soil physicochemical properties and biochemical C, N, and P cycles in the soil (Calcareous)-plant (Common and Narbonne vetch)-root-microorganism (AMFs) ecosystem.",
"discussion": "4 Discussion 4.1 Effect of AMF on some soil properties In this study, CV plant species were more effective than NV plant species in reducing the lime content of soils under both AMF and control conditions. These activities largely determine nutrient cycling and play an important role in the development of sustainable agriculture. However, related studies report largely different results for the effect of AMF on soil properties [ 42 , 43 ]. A decrease in the lime content of soils means a decrease in the pH value of the soil. Investigating how AMF stabilizes soil structure, Leifheit et al. [ 44 ] reported that AMF prefers soil pH close to neutral. Although the result of this study does not fully support the result of the present study due to the near-neutral soil pH, it shows that slightly alkaline soil pH can further enhance the functional performance of AMF in soil enzyme activity, and chemical and biological properties of soils. This means that there may be significant decreases in soil pH as a result of these applications for many years. This is because the decrease in the lime content of soils supports this, and the release of Ca +2 ions may also cause an increase in pH [ 43 ]. Therefore, the results of the present study are similar to those of Alguacil et al. [ 45 ]. Because they reported that soil pH may not be affected by AMFs. Akhzari et al. [ 46 ] provided scientific evidence supporting the correlation between soil pH, 10.13039/501100000780 EC , potassium, and AMF spore number, our results showed no significant positive correlation between soil pH and GICM in CV vetch, GCF + GS in NV vetch, 10.13039/501100000780 EC , and GI AMF in NV vetch. These differences may be related to AMF species and behavior since some AM fungi prefer acidic soils while others do not [ 47 ]. In calcareous soils, AMFs have proved to be an effective method that can be used to increase plant growth and yield. The reduction of soil lime and EC by AMF applications is important for arid and semi-arid calcareous zone soils. 4.2 Effect of AMF on bio-chemical C and N cycles of soils Fig. 1 shows that mycorrhizal roots create a sink demand for C and N. It is well known that increased populations of microorganisms in the soil increase the MBC and MBN content in the plant root zone, both due to the mineralization of organic wastes and because they are themselves a source of high-quality C and N. This increase in MBC and MBN increases plant nutrient availability by enhancing enzyme activity [ 48 ]. As increased plant biomass in the soil increases C availability, C allocation to AMF also increases and promotes AMF growth [ 49 ]. This C demand is supplied by C and transferred by the host plant through photosynthesis. Furthermore, AMF extramatric hyphae represent 10–80 % of soil microbial biomass, accounting for 15 % of soil organic C [ 50 ]. For efficient decomposition of organic matter by soil microorganisms, a soil C:N ratio of 10:1 is considered ideal. A balanced C:N ratio ensures that sufficient nitrogen is available to meet microbial demand during decomposition [ 51 ]. Generally, CV plots had lower C:N ratios and therefore higher mineralization (data not shown). However, the C:N ratio values of soils planted with NV cultivar vary in a significantly wider range (11–49), and thus the mineralization of the soil slows down the decomposition of organic matter [ 23 , 52 ]. In this study, this means that the NV plant is texturally more rigid and more resistant to decomposition than the CV plant. In addition, soils with a high C:N ratio have poor mineralization and nitrification, which is against nitrate formation and accumulation [ 53 ] because N mineralization and nitrification are positively correlated with soil NO 3 − -N content [ 54 ], which was supported by our results ( Table 4 ). 4.3 Effect of AMFs on enzyme activity Enzymes participate in many vital soil biochemical reactions and can have significant effects on soil fertility, strongly linked to AMF [ 55 ]. In our study, the effects of different AMFs on CAT, DGH, and urease enzymes were generally different from the control and these changes were statistically significant (p < 0.05). Urease is an agent that catalyzes the hydrolysis of soil-applied urea or existing urea to ammonia, causing the release of NH 3 and pH increase, and is also the first step of the nitrification process [ 56 ]. In our study, there was a statistically significant correlation between urease and NH 4 + (r = 0.38, weak positive), APA enzyme (r = 0.36, weak positive), and N plant (r = 0.51, moderate positive) ( Fig. 4 ). Based on this, it is thought that our study is consistent with the study conducted by Xiao et al. [ 57 ]. N soil significantly affected soil urease, indicating that AMFs can affect soil nitrogen and have a significant relationship with soil nitrogen supply capacity. The analysis (DSPC) showed that soil microbial activity was affected by soil characteristics and AMF inoculation ( Fig. 5 ). In our study, AMF(+) inoculation significantly improved urease activities under the conditions under which both vetch cultivars were grown, indicating that AMF(+) inoculation contributed to the improvement of soil enzyme activity under vetch growing conditions ( Fig. 2 ). It is consistent with AMF inoculation improving urease activities in the same soil and different plant species [ 58 ]. In particular, AMF enhanced urease activity by regulating extracellular enzymes [ 55 ] and facilitating the growth and development of microorganisms involved in soil N metabolism. The variation of urease activity by plant and AMF species has been demonstrated in many studies that vegetation types can modify the characteristics of soil microbial community structure and diversity [ 59 , 60 ]. When Co groups of CV and NV plants were compared in terms of CAT enzyme activity, NV had a higher activity. High CAT content in both plants was found in vetch plots where G I inoculants were applied. The fact that CAT activity was lower in CV plant variety plots compared to NV suggests that this vetch species may be more tolerant to stress, and under extreme stress conditions, it is thought that plants increase microorganism activities in the root region by secreting sugars and other compounds from the phloem and that these microorganisms, whose activities increase, secrete CAT enzyme substrate to the root rhizosphere region. This is because the high activity of enzymes catalyzing the breakdown of H 2 O 2 indicates that soil conditions are favorable for aerobic microorganism microflora [ 52 ]. CAT was significantly correlated with NH 4 (r = 0.40, moderate positive), plant Nt (r = 0.42, moderate positive), N soil (r = −0.55, moderate negative), DHG enzyme (r = 0.53 moderate positive) ( Fig. 4 ). Environments such as climatic factors, soil, and spatial patterns influence edaphic microbial richness and community structure [ 61 , 62 ]. Plant species are the major factor in determining microbial diversity and community in the rhizosphere, resulting in different microorganism compositions for various species growing in the same soil [ 63 ]. In some cases, however, host variety may have more influence on microbial composition than soil and plant species [ 64 ]. Among the enzymes tested, DHG is the most sensitive and this enzyme decreased in all tested areas except for partial increases. The highest DHG activity was detected in AMF inoculations of G F (13.59 %) in CV and G I (18.03 %) in NV plots ( Fig. 2 ). DHG activity was high in plots with high total C and total N utilization efficiency. The variability of DHG activity may be affected by soil C, N, lime, MBC, MBN, other enzyme activities, cultivation types, plant species, and genus. According to Burns [ 65 ], the effects of higher plants on soil enzymes depend on plant chemical composition, which can vary considerably between genera, species, and also between cultivars, even in the case of root exudates. Low DHG activities were observed in agriculturally cultivated soils, while Ostrowska and Porębska [ 66 ] reported higher activities of these enzymes in pasture soils. DHG activity of soils varies depending on the C:N ratio of soils. The relationship between enzyme activities and the C:N ratio confirms, among others, the importance of the quality of organic matter supplied by plants [ 66 ]. Statistically significant relationships were found between DHG activity and silt (r = −0.53 moderate negative), lime (r = 0.38, weak positive), N soil (r = −0.52, moderate negative), Mmic:Nmic (r = 0.45, moderate positive), and Cmic (r = 0.48, moderate positive) ( Fig. 4 ). Niemeyer et al. [ 67 ] state that the main negative impact on microbial indicators, and among them soil enzymes, is due to the limitation of plant re-establishment resulting in a low input of organic matter into the soil. In support of this, Patel and Patra [ 68 ] argue that the increase or decrease in DHG and APA activities is probably due to organic matter. DHG was estimated to vary because it is not always obvious in complex systems such as soils where the microorganisms and processes involved in the degradation of organic compounds are highly complex. From the above results and explanations, it is very clear that AMF can increase soil enzyme activity, which in turn can improve nutrient cycling. In the present study, AMF treatments positively affected APA enzyme, soil, and plant P content ( Fig. 3 ). This can be explained by the ability of AMF to convert inorganic phosphate into soluble forms through acidification, chelation, exchange reactions, and organic acid, H +, and metabolite production processes [ 52 ]. In our study, AMF applications significantly decreased the lime content of soils (p < 0.05, Table 3 ), and the organic acid produced by AMF converted insoluble mineral phosphate into a soluble form [ 69 ]. This event is believed to be due to AMF hydrolyzing organic P to inorganic P through a mechanism linked to the production of enzymes called phosphatases [ 70 ]. However, it is worth mentioning that bacteria known as phosphate-solubilizing bacteria, which mineralize organic P and produce phosphatase, also have very important potential [ 71 , 72 ]. The reason for the lower P content of the soils compared to the Co group can be attributed to physical fixation and uptake by the plant due to high activity. This is also explained by the fact that plant P content and plant biomass were high in most AMF treatments. The findings obtained are in agreement with the findings of Huo et al. [ 73 ]. It is thought that the high P content of the soil in the areas where Co groups and some AMF(+) inoculants were applied in the study may have inhibited APA activity [ 74 ], because the high available P content of the soil may slow down AMF activity. This was clearly seen in the Co groups, explaining the weak activity in the CV with G ICM and NV with G I species ( Table 4 ). In contrast to these views, Wei et al. [ 75 ] stated that there may be differences due to the variation in the regulatory gene system in the genotype. 4.4 Effect of AMF on some nutrients and plant biomass It was reported that AMF colonization improved plant nutrients and uptake as well as below- and above-ground biomass [ 76 ]. In terms of the development of plant roots, it was observed that the CV plant had better root development and better nutrition. As can be seen from this, plant roots have a profound effect on soil nutrient dynamics. Understanding these effects is important for sustainable agriculture and soil fertility management. The AMF symbiosis results showed strong colonization effects on nutrient uptake for most of the other nutrients we measured, especially N plant , N soil , NO 3 −, and NH 4 + ( Fig. 3 , Fig. 4 ). The study by Lehmann and Rillig [ 77 ] supports this. However, the fact that some of these improvements (P soil , N soil , etc.) were lower than the control does not mean that there was little or no synergistic effect, but rather that the nutrients were taken up by the plant and used in its metabolic functions. This could also mean that the CV plant takes more nutrients from the soil and uses them in metabolic activities than the NV plant. When the Co groups of the plants are compared, it is seen that the NV plant is more passive in the uptake of NO 3 − and NH 4 + from the soil. The high C content of the soil in which the CV plant was planted was found to improve soil N content ( Fig. 1 ) due to its ability to retain N and reduce N losses through leaching [ 78 ]. With this improvement, the available N increased compared to the Co group and this increase could be attributed to the increased amount of subsoil biomass. Previous research has shown that plant-available N (NH 4 + and NO 3 − ) may decrease due to high C content, which stimulates microbial N immobilization [ 79 ]. In this study, similar to the study by Hu et al. [ 23 ], plant roots and AMF symbionts increased the ability of plants to obtain inorganic N from soil, increased the biomass of host plants, and reduced the NH 4 + and NO 3 − content of soil ( Table 4 ). Fall et al. [ 80 ] reported that the addition of AMF(+) inoculants resulted in higher crop yields. The best results in our study were observed in the increase in the amount of subsoil biomass of both legume crops with AMF inoculant (data not shown). This increase was 48.55 % in CV and 43.60 % in NV plants. This result is significant in all respects that the yield increase of subsoil biomass was realized without the use of fertilizers. The results presented in this study demonstrate the environmental benefits of using AMF to increase the productivity of legume crops. Similar results were reported for pineapple, where AMF inoculation and application of half the fertilizer dose promoted the highest levels of fruit mass and organoleptic variables [ 80 ]. An increase in sorghum yield was also reported by Ramadhani and Widawati [ 81 ], showing that a significant reduction of fertilizer in combination with AMF can reduce soil degradation and improve its quality. In this study, when the combination of GCF + GS inoculant was used for CV-cultivated areas and GI inoculant for NV-cultivated areas, the results showed that legume crops responded better to the inoculant ( Table 4 ). AMF can therefore be seen as a good alternative to chemical fertilization, or at least reduce the need for large quantities of synthetic fertilizers (NPK) by half. Agricultural management practices based on AMF application can provide an economical, environmentally friendly, and sustainable way to improve soil fertility and yield. 4.5 AMF root colonization The lowest root colonization rate was found in the GI community applied to CV and NV planted plots. It is thought that there may be some complementarity when the mixed community with the highest root colonization rates in CV plots is inoculated with GICM and GF with the highest root colonization rates in NV plots. However, it was observed that the mixed community and GF inoculations had a very similar potential synergistic effect on soil enzyme activities, plant nutrients, and some physical and chemical properties of soil, RFW, or nutrient uptake. These results showed that GICM and GF successfully colonized the rhizosphere of common and big vetch and moved effectively in the soil [ 82 ] Van der Heijden et al. [ 83 , 84 ] found that plants respond differently to certain AM fungal species. Therefore, a similar pattern of seasonal mycorrhizal colonization index should not be expected from experimental plants along the topographic gradient, especially when their root systems, growth periods and dependence on mycorrhizae to grow on nutrient-poor soils differ. Differences in colonized root length and plant biomass between CV and NV plants could be attributed to interactions between the growth rates of both fungi within roots and roots within the soil. Although no fertilization including P was applied to the experimental plots, root colonization was high. Bars-Orak and Demir [ 85 ] investigated the effects of GI AMF species and different P doses in a similar region and climate. In their study, it was observed that AMF(+) plants were able to absorb more P than AMF(−) plants in foliar P analysis during the first flower formation period. Again, soil analysis showed that mycorrhizal plants were able to absorb sufficient P even when phosphorus was very low in the soil. Some studies have reported that high phosphorus levels inhibit the colonization of AMFs [ 86 ]. Most soils used for agricultural purposes contain excess amounts of organic and inorganic P [ 87 ]. Much of this P can come from fertilizers applied to the soil for agricultural production [ 88 ]. When applying AMF to the soil, rather than applying P fertilizer, it is very important to make available the inorganic P present in the soil and fixed in any way. Thus, AMF can contribute to higher soil fertility and health in the root rhizosphere regions of the soil in arid and semi-arid regions, depending on root colonization and plant diversity. In our study, the overall low root colonization of the NV legume plant was observed, which may be due to the low colonization rates of GI and GMS in the NV + AMF inoculant, which were similar in other plant varieties. For example, the average colonization percentages of a single G MS strain ranged from 2.6 % to 27.0 % in a range of tomato cultivars [ 89 ]. Interestingly, the cultivar with the lowest colonization percentage still showed a significant increase in root dry weight of 43.60 % for Intradies (GI) and 0.99 % for Mosseae (GMS) in response to inoculation, while the root weight of the CV cultivar with the highest colonization percentage was significantly affected by G I and G MS . This rather indicated that AMF community structure and mycorrhizal fungi were affected differently. This study shows that regardless of the cause of low colonization of some AMF (+) inoculants, low colonization percentages can significantly affect plant performance (RFW weight) and that the magnitude of AMF effects is not necessarily related to colonization percentage. The results of the study are in agreement with the study conducted by Wang et al. [ 90 ]. 4.6 Data grouping technique (network analysis-heatmap) The functions and performance of roots in plants provide valuable information about the overall health of the plant and its response to environmental conditions [ 91 ]. Our study showed that the growth-promoting effects of different AMF inoculants applied to the soil where CV and NV plants were grown varied between the fresh weight and dry weight of the roots (p < 0.05). These results indicate that the plant growth-promoting effects of all the tested AMF inoculants were consistent between fresh and dry weights and that there was a very strong positive relationship between them (r = 0.9344, p = 0.0000; r = 0.987, p < 0.001) and studies support this relationship [ 91 ]. The moderate correlation between RDW and C (r = 0.57, p=<0.001) and RFW and C (r = 0.64, p=<0.001) is due to the fact that soil organic carbon is formed from the decomposition and breakdown of plant and animal tissue residues [ 56 , [92] , [93] , [94] ]. Of course, the amount of water contained in the tissue of the plant in RFW may be the reason why the C ratio differs slightly in dry weight. Increasing the amount of subsoil biomass (RDW) increases the C content. Many studies have shown that the positive relationship between RDW and C is strong [ 95 ]. A strong negative relationship (r = −0.78, p=<0.001) was found between MBC and MBN and lime; high lime content in soils may affect soil biological activities through its effects on the amount, structure, and distribution of soil organic matter. The lime content of soils can affect fungal and bacterial biomass differently [ 11 , 96 ]. Shah et al. [ 97 ] determined fungal and bacterial biomass separately and showed that lime content increased bacterial biomass only for a short time, while fungal biomass was not affected. Zelles et al. [ 98 ] suggested that increased lime slightly increased bacterial biomass, but significantly reduced fungal biomass. There was a moderate negative correlation between soil N content and CAT enzyme activity (r = −0.55, p=<0.001). This suggests that the application of plant nutrients increases soil microbial biomass, but excessive fertilizer application reduces soil microbial biomass carbon [ 99 ], it has also been shown that fertilizer reduces microbial biomass by about −15 % [ 100 ]. Bargali et al. [ 101 ] showed that a moderate increase in nitrogen fertilizer application is beneficial to increase the carbon and nitrogen content of soil microbial biomass. There is a moderate positive correlation between CAT and urease enzymes (r = 0.53, p=<0.01). It is possible to see a similar moderate relationship between catalase and urease enzyme activities in other studies [ 102 , 103 ]. The presence of a moderate to positive correlation between C and MBN (r = 0.47, p=<0.01) is because microbial biomass acts not only as a C sink but also as an active driver of C and N transformation [ 104 ]. Although representing a small fraction of total soil C and N, microbial biomass plays a critical role in SOC mineralization [ 11 , 105 ]. There is a moderate positive correlation between urease and N plant (r = 0.51, p=<0.01), indicating that plant N and urease activity have a significant positive correlation. It is one of the most important enzymes involved in the nitrogen cycle and has been positively correlated with many nutrients [ 106 ]. There was also a significant correlation between soil NH 4 –N content and urease activity [ 107 ]. The moderate negative correlation (r = −0.53, p=<0.01) between MBC:MBN and APA can be explained by the increased amount of organic C in the soil and thus lower microorganism activity. On the other hand, Liu et al. [ 108 ] reported that N application or the presence of sufficient nitrogen in the soil slowed down APA activity. Saha et al. [ 109 ], and Böhme and Böhme [ 110 ] reported that organic fertilization stimulated alkaline phosphatase activity, while P fertilizers had a negative effect. These contradictory results and interpretations found in the literature may be due to soil properties and/or different plant species studied."
} | 8,051 |
34001948 | PMC8129112 | pmc | 9,270 | {
"abstract": "Plants must deal with harsh environmental conditions when colonizing abandoned copper mine tailings. We hypothesized that the presence of a native microbial community can improve the colonization of the pioneer plant, Baccharis linearis , in soils from copper mining tailings. Plant growth and microbial community compositions and dynamics were determined in cultivation pots containing material from two abandoned copper mining tailings (Huana and Tambillos) and compared with pots containing fresh tailings or surrounding agricultural soil. Controls without plants or using irradiated microbe-free substrates, were also performed. Results indicated that bacteria ( Actinobacteria , Gammaproteobacteria, and Firmicutes groups) and fungi ( Glomus genus) are associated with B. linearis and may support plant acclimation, since growth parameters decreased in both irradiated (transiently without microbial community) and fresh tailing substrates (with a significantly different microbial community). Consistently, the composition of the bacterial community from abandoned copper mining tailings was more impacted by plant establishment than by differences in the physicochemical properties of the substrates. Bacteria located at B. linearis rhizoplane were clearly the most distinct bacterial community compared with those of fresh tailings, surrounding soil and non-rhizosphere abandoned tailings substrates. Beta diversity analyses showed that the rhizoplane bacterial community changed mainly through species replacement (turnover) than species loss (nestedness). In contrast, location/geographical conditions were more relevant than interaction with the plants, to explain fungal community differences.",
"introduction": "Introduction Copper mining operations adversely affect the environment due to the deposition of large volumes of hard-rock waste materials in nearby areas, being mine tailings quite relevant sources of contamination 1 . Mine tailings are waste materials resulting from the mineral separation process, which are mostly composed of silt or sand-sized particles. They lack organic matter and nutrients while containing high quantities of heavy metals 2 and slightly alkaline to low pH 3 , 4 . Abandoned mining tailings may also result in secondary environmental impacts due to the dispersal of metal-rich particles and weathering of the material with the generation of acid mine drainage 3 . These characteristics of tailings limit microbial diversity 5 , as well as spontaneous plant colonization 6 and, therefore, affect the application of phytoremediation procedures.\n Conventional remediation technologies based on chemical/physical treatments are generally nonviable economic options in post-operative mining tailings generated by large size mine operations 6 . Application of organic amendments (alone or in combination with beneficial microorganisms, see below) to degraded soils has been successful in some cases 7 , 8 . In addition, plant-based technologies are generally cost-effective and environmentally sustainable 6 . Phytoremediation technologies allow metal immobilization, reduce metal contamination in the surrounding environments, and provide erosion control and a wildlife habitat 9 . In particular, the use of native plants in these technologies is also favored because they demonstrate tolerance to local environmental conditions and provide a foundation for natural ecological successions 1 , 10 . Spontaneous primary succession involves plants and, although initially limited to small patches of vegetation distributed mainly on their edges 11 , in the long-term, it leads to vegetation changes and transformation of mineral substrates into the soil 12 . The role that microorganisms have in growth, nutrition and health in plants is increasingly better known. Among other capacities, microorganisms play a key role in plant establishment as they degrade organic matter and recycle nutrients 13 and protect plants from stress 14 , 15 . However, the extent of the influence of microbial community structure, and especially microbial diversity, on colonizer plant growth on copper tailings has not been approached so far 5 , 16 . The role of native microbial communities in plant establishment in these degraded substrates is less understood. In this work, we investigated both the native microbial community structure and diversity and their contribution to early plant colonizers’ development in soils from copper mining tailings. We used a pioneer plant Baccharis linearis as a study model, and substrates originated from two abandoned copper mining tailings from a semiarid zone in the central part of Northern Chile (Huana and Tambillos). We first analyzed the physicochemical properties of these tailing substrates. Then, we analyzed their microbial communities by means of terminal restriction fragment length polymorphism (T-RFLP) and clone library analyses. Comparison with those found in a surrounding agricultural soil and contrasting different compartments according to their association with plants (non-rhizosphere, rhizosphere, and rhizoplane), was also carried out. Subsequently, we studied the effects of fresh tailings in the establishment and growth of B. linearis plants, and, finally, using gamma-irradiation we studied the effect of a reduction in the microbial community in the germination, growth, and establishment of this pioneer plant.",
"discussion": "Discussion Some of the challenges that plants face in colonizing tailings were analyzed in two copper mining tailings from operations located in northern-central Chile, using the pioneer plant B. linearis . The role that B. linearis associated microorganisms play in plant germination and establishment was especially assessed. We found that native microbiota was required to improve the pioneer plant B. linearis establishment and growth in copper tailings, and that microbial communities were more influenced by the pioneer plant’s presence than the substrate physicochemical properties. In contrast with reported acid/alkaline conditions in copper tailings, physicochemical analyses of tailings used in this work demonstrated that very little or no secondary acidification had occurred. This may be explained, at least in part, by high evapotranspiration rates occurring at the upper sulfide-rich horizon of tailings, due to arid and semiarid climate conditions, and/or high content of dissolved CO 3 2- due to high calcite/pyrite ratio, which is typical in these tailings [3,]. Since pyrite is the main source of sulfide in porphyry copper deposits, the latter does not only imply a high buffering capacity of the substrate solution, but also a lower susceptibility to iron sulfide oxidation and to the subsequent production of acid mine drainage 21 . On the other hand, the presence of organic residues coming from flotation reagents can explain the high organic matter content in the fresh tailings 22 . High soluble sulfate levels found in all these tailings could derive from the high amount of solubilized sulfate-containing minerals, like gypsum and jarosite. These minerals are commonly found in the Chilean copper mining area 23 , and not necessarily are related to sulfide oxidation, as for acidic tailings 24 . These high amounts of sulfate can also explain elevated concentrations of exchangeable cations, due to the formation of stable complexes with SO 4 2- , thus preventing precipitation after interaction with other anions 21 . Analysis of cation concentrations in rhizosphere substrates showed that both tailings had sodium ion levels equivalent to those of sodic soils 25 . This is relevant as such ions are transported to the surface of the tailings and, as water evaporates, precipitated as sodic salts. This salinization is evidenced by high electrical conductivity levels, and in fact, high levels were found in the three tailings when compared to surrounding soil, especially in fresh Tambillos tailings (Table 2 ). Electrical conductivity values of these tailings are within the range that usually limits the growth of terrestrial plants 18 . Comparatively low cation exchange capacity and total organic carbon values were obtained for both tailings when compared to surrounding soil. These parameters are slightly more favorable towards plant colonization in the older Tambillos substrate. This is consistent with longer time of weathering, surface deposition of soil particles from surrounding fields, and higher vegetation coverage. Differences between tailings and surrounding soils have also been reported for copper, zinc, and iron levels in Chilean copper mining tailings 26 . However, no differences could be detected among tailing locations, and total metal values were in the range of two (iron and zinc) to five (copper) times higher than those of surrounding soil. Microbial community compositions were studied to better understand B. linearis establishment in copper mining tailings. For logistic considerations, the selected culture-independent technique was chosen to make prospective comparisons of microbial communities in these soil substrates. It should be kept in mind that this molecular technique is not as powerful in coverage (richness) as high-throughput, next-generation sequencing approaches, since only dominant (abundant) members of the respective microbial community are detected. Therefore, future studies are required to get a deeper understanding of the B. linearis microbial community dynamics and composition in these substrates. The comparison of the most abundant members composition of the bacterial and fungal community suggested that substrate conditions have a strong influence in these communities (Fig. 3 ). This was mainly observed in fresh Tambillos tailings, which clearly differentiates from all other substrates, especially in terms of beta diversity. Huana and Tambillos tailings showed lesser but evident changes according to aging and/or compartments (Fig. 3 ). The clear differentiation of beta diversity in Tambillos tailings relative to Huana tailing and surrounding soil contrasts with the trend of physicochemical differences discussed above. Tambillos tailings were more like surrounding soil than to Huana tailings (Table 2 , Supplementary Table S1 online). This suggests that these microbial communities are less affected by differences in electrical conductivity, cation exchange capacity, and total organic carbon. The latter may be also true for metal concentrations in both tailings, as these microbial communities were previously reported not to be substantially variable among tailing locations, but significantly different in surrounding soil 26 . This would imply that metals have a low influence in the microbial diversity among tailings, but still affect plant growth 27 , although there are reports of contrasting evidence 28 . On the other hand, the microbial community similarity between Huana tailings and surrounding soil can also be explained by significant and continuous microbial inoculation from surrounding soils 29 , probably by soil erosive processes 30 . This is the first report approaching beta diversity analyses of microbial communities from abandoned copper mining tailings. In addition to the unexpected results just discussed, this analysis was also able to shed some light on what is driving beta diversity of both tailings, and surrounding soil B. linearis rhizoplane microbial communities. Differences in total beta diversity index (β SOR ) were found for Tambillos tailings with respect to Huana and surrounding soil (Fig. 3 ). Turnover patterns (β SIM ), as well as β SOR, mainly distinguished Tambillos from Huana tailings and surrounding soil (Fig. 3 B). This indicated that species replacement (turnover) was more significant than non-replacement (nestedness) in Tambillos tailings. However, high species richness was found for these tailings, which is in agreement with literature, indicating that bacterial and fungal diversity indices tend to be higher in old, reclaimed sites (15 to 20 –years old), as recovery time increased 31 . Therefore, higher turnover values for Tambillos tailings could be explained by a combination of factors as compared to Huana tailings, including a larger size of particle material (texture %). These factors increase water infiltration, together with high dissolved organic carbon, and low cation exchange conductivity, and electric conductivity values. Therefore, it may result in an improvement of micro-environmental conditions perceived by microbial communities. In any case, it is clear that different physicochemical conditions among substrates do not necessarily explain changes in B. linearis rhizosphere and non-rhizosphere microbial communities. This is further demonstrated when gross changes in microbial community structures are analyzed. While bacterial community structures were mainly affected by “root compartment” (microenvironment conditions), fungal communities were clearly influenced by “location” (geographical conditions) (Fig. 1 ). In addition, only rhizoplane bacterial community structures were significantly different from the rhizosphere, or non-rhizosphere environment. The latter suggests that close association with root plants provides a microenvironment different enough to influence B. linearis associated bacteria. Processes mainly operating at a few microns range (distinguishing rhizoplane from rhizosphere microhabitats), especially under conditions when fluid transport is limiting, may explain the preceding observation. Vegetation development causes the incorporation of organic carbon (rhizodeposition), stimulating microbial activity and element cycling processes 32 . In turn, root exudates production by pioneer plants may generate differential activation of rhizobacteria 33 . Thus, B. linearis may play a selective role at their rhizoplane, shaping bacterial diversity. This is in agreement with the proposed two-step selection model for root microbiota differentiation where plant ecotypes first restrict access to rhizoplane to certain specific microbiota members, and then the endophytic association may take place 29 . In contrast, fungal communities may overcome short-range rhizospheric microhabitat limitations by extensive filamentous growth, and therefore being relatively more affected by long-range space (location) and time (aging) conditions. Which bacterial taxa are differentially associated with B. linearis colonizing copper mining tailings? Previous reports had shown changes in relative abundances of Acidobacteria , Firmicutes, Nitrospira, Alphaproteobacteria, Gammaproteobacteria, and Deltaproteobacteria 5 , 9 , 10 , 34 , 35 . In partial agreement with these reports, this work revealed that Proteobacteria , Actinobacteria , Firmicutes , and Bacteroidetes taxa were found abundant in all tailings, although the composition of predominant phyla differed across the sites (Fig. 2 and Supplementary Fig. S1 online). In addition, this work further demonstrated that the rhizoplane and rhizosphere of B. linearis were clearly dominated by Proteobacteria, Firmicutes , and Actinobacteria in both tailings and in the surrounding soil. In rhizoplane and rhizosphere compartments, plants locally provide higher levels of C and N sources than levels that can be found in tailings, potentially favoring Proteobacteria , which can play a wide range of roles in the C, N and P cycles 31 . Gammaproteobacteria was a ubiquitous and abundant class in root compartments of B. linearis across all locations. Betaproteobacteria showed a marked presence at rhizoplane microhabitats, while Alphaproteobacteria were relatively less abundant, irrespective of the rhizosphere or non-rhizosphere microhabitats. Detection of members of the Firmicutes phylum in these substrates can be explained due to their tolerance to very low nutrient levels. Detection of Actinobacteria phylum was expected since it has long been recognized that several members within this phylogenetic group are capable to carry out biological nitrogen fixation among other microbially driven soil processes 34 . This work also reports a poor performance in plant growth and establishment when B. linearis individuals were grown in fresh Tambillos tailings, as compared with those grown on surrounding soil (Fig. 4 ). Besides obvious physicochemical differences between these substrates, plant growth disparities may also be explained by observed differences in bacterial and fungal community structures (Fig. 1 ). The same might apply to differences in plant growth between fresh and older tailings of the same mine (fresh vs. old Tambillos tailings), suggesting that ongoing processes of primary ecological succession are taking place 35 . Small particle sizes found in these tailings (Table 2 ) contribute to higher bulk density, mechanical compaction, and smaller pore size, hampering water infiltration, and therefore, plant growth. The role of associated microbiota on B. linearis growth and establishment was further explored in irradiated tailings with significantly depleted microbial activity. The irradiation method is recommended over the use of antibiotics or autoclaving, because it effectively reduces microbial community while producing low changes in the physicochemical properties of substrates 36 . Three consecutive doses of 25 kGy were used as they have been reported to reduce fungal, bacterial, and algal abundances in soils 36 , 37 . It has to be noted that irradiation treatment does not permanently eliminate microorganisms. The re-colonization process in sterilized/irradiated substrates has been studied, reporting an increase in microorganisms’ abundance over time, although with decreased diversity 37 , 38 . Although 18 weeks after irradiation total colony forming unit values for irradiated substrates were in fact higher than in non-irradiated ones, diversity of these colonies was significantly lower, and therefore, not affecting the outcomes of this work. In any case, irradiation of tailings had a severe effect on plant growth (Fig. 5 ), which can be explained by the absence (although transitory) of an active and diverse, normally associated, microbiota. In a similar way, plant growth on fresh Tambillos tailings (Fig. 4 ) was clearly affected by the presence of a quite different microbial community compared with those of abandoned tailings or surrounding soil (Fig. 3 ). As irradiated substrates contained significantly lower available N levels (Supplementary Table S4 online), this limitation can contribute directly to impaired plant growth, more than the sole absence of normal microbiota. The ability of plants, including B. linearis , to manage N limitation under conditions present in tailings is probably quite restricted, especially without the activity of N-fixing members of rhizosphere microbiota. Relatively high numbers of Alphaproteobacteria and Betaproteobacteria found in B. linearis rhizosphere individuals colonizing these tailings is consistent with a need for N-fixing proteobacterial classes 39 . An additional restriction for growth on irradiated substrates might come from P access, not because of decreased availability, as P levels were not reduced by irradiation (Supplementary Table S4 online). The absence of arbuscular mycorrhizae, otherwise abundant in B. linearis plants colonizing non-irradiated tailings (Supplementary Fig. S2 online) may explain this growth effect. Fungal spore isolation from tailings colonizing plants and further characterization clearly showed a massive presence of arbuscular mycorrhizae belonging to the Glomus genus. This is recognized as a phosphate solubilizing fungal group 40 , able to produce changes in rhizospheric microbial communities 41 . It is worth mentioning that interaction between mycorrhizal fungi and associated bacteria has been demonstrated to protect plant species from heavy metals effects 42 , besides helping in nutrient acquisition 43 , and phytostabilization of copper tailings 44 ."
} | 5,019 |
18831723 | PMC2776406 | pmc | 9,272 | {
"abstract": "Explaining the evolution of cooperative behavior is a long-standing problem for which much theory has been developed. A recent paper in BMC Biology tests central elements of this theory by manipulating a simple bacterial experimental system. This approach is useful for assessing the principles of social evolution, but we argue that more effort must be invested in the inverse problem: using social evolution theory to understand the lives of bacteria."
} | 113 |
31861482 | PMC7022246 | pmc | 9,273 | {
"abstract": "One of the prerequisites of successful address delivery is controlling the release of encapsulated drugs. The new method of bacterial spore encapsulation in polyelectrolyte microcapsules allows for degrading the nanoscale membrane shell of microcapsules. The possibility of encapsulating spore forms of Bacillus subtilis in polystyrenesulfonate sodium/ polyallylamine hydrochloride (PSS/PAH) polyelectrolyte microcapsules was demonstrated. The activation and growth on a nutrient medium of encapsulated bacterial spores led to 60% degradation of the microcapsules nanoscale membrane shell. As a result, 18.5% of Fluorescein isothiocyanatedextran was encapsulated into polyelectrolyte microcapsules, and 28.6% of the encapsulated concentration of FITC-dextran was released into the solution.",
"conclusion": "4. Conclusions We have found that the spores of B. subtilis subsp. can be encapsulated in the cavity of polyelectrolyte microcapsules. At the same time, the spores maintain their viability and, when they enter the nutrient medium, germinate, since the nanoscale shell of the microcapsules is semipermeable and capable of passing nutrients into the capsules. In the process of germination of bacterial spores, encapsulated in the cavity of the microcapsules, they deform and break the nanoscale shell of 60% of microcapsules, leaving the surrounding solution. Overall, 28.6% of encapsulated FITC-dextran was released into the solution after shell destruction of polyelectrolyte microcapsules. Thus, it can be said that the encapsulation of bacterial spores can be used to decapsulate and release the drugs contained inside the microcapsules under given conditions. The main advantage of this method of opening microcapsules is that it does not require the use of specialized expensive equipment, such as ultrasonic or laser generators. Given the widespread use in medicine and veterinary medicine of B. subtilis , as a probiotic culture with antibacterial and antifungal effects, the results can be a practical foundation for creating a new form of drug that will allow the delivery of not only the medicinal substance protected from the negative effects of an aggressive environment but also a bacterial culture that restores the resident microflora in certain parts of the digestive tract.",
"introduction": "1. Introduction The main problem in the use of drugs in medicine is the lack of selectivity. This can lead to many side effects, as well as a decrease in the effectiveness of drugs. One of the ways to solve the problem is to encapsulate medicinal substances in polyelectrolyte microcapsules (PMC). PMCs can be obtained by layer-by-layer adsorption of oppositely charged polyelectrolytes on the surface of microparticles and have a diameter of several hundred nanometers to several microns [ 1 ]. The PMC shell allows protection of the encapsulated substance from the effects of aggressive environmental factors. For instance, encapsulation of urease allows preservation of the enzyme activity in a solution with proteinase K, while the non-encapsulated enzyme under these conditions is rapidly inactivated [ 2 ]. B. Sukhorukov et al. demonstrated preservation of the activity of encapsulated urease in the presence of proteinase K in solution, while the free enzyme rapidly lost activity under these conditions [ 3 ]. Due to such protection, PMC can be used for drug delivery both to the focus of the disease [ 4 ] and to individual cells of the body [ 5 ], with the ability to control the speed of their release [ 4 ]. The release of the encapsulated substance is usually achieved by destroying the integrity of the microcapsule. For example, Borodina T.N. et al. [ 6 ] showed the destruction of capsules formed from biodegradable polyelectrolytes under the action of proteolytic enzymes. Authors of various articles have shown the controlled destruction of microcapsules with a change in the pH value of the medium [ 7 ], as well as the additional structural elements included in them [ 8 , 9 ]. For example, Demina P.A. et al [ 10 ] showed that microcapsules, in the shell of which TiO 2 nanoparticles are included, are destroyed by ultraviolet radiation. The research papers [ 8 , 11 , 12 ] showed the destruction of PMCs containing silver and gold or rhodamine nanoparticles by laser radiation, and the presence of Fe 3 O 4 particles in the capsule shell makes it possible to control them by microwave radiation [ 9 ]. The described methods involve the use of installations that generate the corresponding radiation. It leads to complication of the process of destruction and its rise in price, so we propose a fundamentally new approach that allows you to destroy the shell PMC. In this paper, we propose a fundamentally new approach to the destruction of the PMC shell through the germination of bacterial spores by aerobic endospore-forming bacterium Bacillus subtilis subsp. subtilis included in the microcapsule. Currently, researchers are actively working on the use of encapsulated forms of B. subtilis . For example, in [ 13 ], bacterial cells of Bacillus subtilis were encapsulated in alginate microcapsules, which improved the germination of cotton seeds. Shantanu S. Balkundi showed the possibility of encapsulating B. subtilis spores while maintaining their viability in polyelectrolyte layers of polyglutamic acid and polylysine, which were applied directly to the surface of the spores [ 14 ]. It is proposed to use the spores of B. subtilis not only as an object of encapsulation but also for breaking the nanoscale shell of a polyelectrolyte microcapsule. This method will allow the further release of medicinal substances contained in the cavity of PMC, with the destruction of the nanoscale shell by germinating bacterial spores when injected into the nutrient medium.",
"discussion": "3. Results and Discussion The bacterial spores that we used are ellipsoidal [ 18 ]. Spores are highly resistant to many agents (chemical, thermal, and radiation) for prolonged exposure [ 19 , 20 ]. Figure 1 b shows a micrograph of a spore of the strain VKM B-501 T in an optical microscope. Bacillus subtilis bacteria are non-pathogenic and members of the gut microflora, and for that reason, its spores are suitable for decapsulation of polyelectrolyte microcapsules in a digestive system [ 21 , 22 ]. Polyelectrolyte microcapsules were obtained with spores of the strain VKM B-501 T and included: particles in CaCO 3 in the process of their formation, followed by deposition of polyelectrolyte layers ( Figure 2 ). The method of light ( Figure 3 a) and fluorescence microscopy ( Figure 3 b) was used to analyze the products of encapsulation of spores, immediately after they were placed in a nutrient medium TSB, which indicated the absence of encapsulated spores. The capsules were particles with a diameter of 8 ± 1 μm, regular round shape, with a clearly defined undeformed shell ( Figure 3 ). The shell thickness was 37 ± 3 nm. B. subtilis growth in a liquid nutrient medium by seeding with a suspension of PMC, with the spores of the strain VKM B-501T, in a liquid nutrient medium tryptone soya broth (TSB) followed by incubation is shown in Figure 4 . After 4 h of cultivation, spores began to germinate in a liquid nutrient medium ( Figure 4 ). The picture showed that the spore started to destruct the nanoscale shell of the microcapsule and left the shell. After 24 h of cultivation, growth was observed in the form of uniform turbidity of the nutrient medium. The study of the culture fluid by optical microscopy showed the presence of the same types of vegetative cells (number 2 at Figure 5 a): whole microcapsules, deformed, germinating spores, and partially destroyed (number 1 at Figure 5 a). As can be seen from Figure 5 b, capsules have an irregular shape and are deformed. At the same time, the study, after 24 h of incubation of the control sowings in sterile distilled water PMC, with bacterial spores included, showed a complete absence of bacterial growth. Optical microscopy established the preservation of PMC, with bacterial spores included, unchanged. As can be seen in Figure, spores germinated (number 2 at Figure 6 ) and went out from polyelectrolyte microcapsules (number 1 in Figure 6 ). The shell of 60% of the microcapsules is destroyed. Also, it can be seen that some microcapsules are intact. It can be related to the size of the polyelectrolyte microcapsules. The size of whole microcapsules is around 2–3 microns, and the size of the destructed microcapsules is 5–7 microns. Thus, it is abundantly clear that when released into the nutrient medium, the spores encapsulated in the PMC are activated and germinate, thus destroying the microcapsule nanoscale shells. We encapsulated 2.22 mg FITC-dextran into polyelectrolyte microcapsules, and thus, 18.5% was encapsulated. FITC-dextran was decapsulated after the destruction of the microcapsule shell by the spores. Six-hundred-and-thirty micrograms of FITC-dextran were released into the solution (28.6% of the encapsulated concentration)."
} | 2,257 |
32310959 | PMC7192501 | pmc | 9,277 | {
"abstract": "Biological networks across scales exhibit hierarchical organization that may constrain network function. Yet, understanding how these hierarchies arise due to the operational constraint of the networks and whether they impose limits to molecular phenotypes remains elusive. Here we show that metabolic networks include a hierarchy of reactions based on a natural flux ordering that holds for every steady state. We find that the hierarchy of reactions is reflected in experimental measurements of transcript, protein and flux levels of Escherichia coli under various growth conditions as well as in the catalytic rate constants of the corresponding enzymes. Our findings point at resource partitioning and a fine-tuning of enzyme levels in E. coli to respect the constraints imposed by the network structure at steady state. Since reactions in upper layers of the hierarchy impose an upper bound on the flux of the reactions downstream, the hierarchical organization of metabolism due to the flux ordering has direct applications in metabolic engineering.",
"conclusion": "Conclusion Understanding hierarchies in biological networks not only can contribute to the discovery of underlying design principles, but also provide the means for effective control of these complex networks. Hierarchies in biological networks are multifaceted, and thus two principle approaches have been used to formalize and uncover them, namely, hierarchies of embedded subnetworks and hierarchies of individual components of the network. Since metabolic networks are composed of metabolites interconverted by biochemical reactions, the hierarchies following the second approach can be based on ordering of either metabolites of reactions. Here, we wanted to explore the extent to which the structure of the metabolic network alone imposes a hierarchy of reactions based on ordering of their fluxes. Such a view provides the means to discover a hierarchy that is present irrespective of the particularities of the reaction kinetics, for which there is still limited information at genome-wide scale. Moreover, this view can be readily cast in the constraint-based modelling framework and can contribute to the discovery of important functional limits of metabolic networks. The latter is due to the natural definition of the flux ordering relation, whereby reactions further up the hierarchy dominate reactions down the hierarchy with respect to their flux in every steady state supported by the network. Our formalization of the flux order relation facilitated the discovery of a hierarchy of reactions in the metabolic network reconstruction of the bacterium E. coli . Analysis of the graph representation pointed out that the hierarchy is branched and relatively shallow for three carbon sources. In addition, we found a complete agreement between the flux order relation and 13 C-based flux profiles. Interestingly, we also found a partial agreement when analysing other phenotypic profiles which correlate with reaction flux, i.e., transcript and protein levels, as well as enzyme catalytic constants. In the first case, these findings point at optimisation of resource allocation to minimise costs of enzyme synthesis. This hypothesis that was further supported by an increased agreement found in subsets of enzymes with increasing costs, particularly when glucose was the carbon source, in transcript and protein levels, and when glycerate was the carbon source, in protein levels. In the second case, the findings suggest network constraints acting upon the evolution of enzyme structural properties to match the flux order relation. The concept of flux order relation in metabolic networks can be readily extended to analyse the hierarchy outside of steady state dynamics, provided we have insights in the degree of metabolic concentration changes around a steady state [ 45 , 46 ]. In addition, due to the universality of a large part of metabolism, it is plausible that the flux order relation also holds in metabolic networks of other organisms. It would be particularly interesting to evaluate to what extent the hierarchy holds in eukaryotic organisms, in which metabolism is compartmentalized and more complex. Further, one can expand the concept for applications to metabolic networks with given analytically tractable kinetics, e.g. mass action [ 47 ]. Finally, since the flux order relation identifies reactions imposing a fine-tuned control over a target reaction, the applications of the hierarchy in metabolic engineering remain as a promising direction for future research.",
"introduction": "Introduction Hierarchical organization has been recognized as a salient feature of biological networks, spanning from the molecular to the ecosystem scale [ 1 – 6 ]. Biological networks do not operate in isolation but form interconnected layers. It is therefore important to identify the constraints that a hierarchical organization in one cellular layer impose on the others. For instance, metabolism is understood as an integrated outcome of biochemical reactions and transcriptional, translational, and post-translational changes with input from the environment. Therefore, elucidating metabolic hierarchies can contribute to understanding the constraints governing physiology as well as their implications for biotechnological applications. Two types of metabolic hierarchies have been identified: The first is based on nested subnetworks, while the second relies on pairwise relationships between components. Nested subnetworks can arise solely due to structural properties of the components in the network. For instance, it has been shown that metabolic networks can be decomposed into modules of highly connected metabolites, and that these modules form a nested hierarchy [ 5 , 7 ]. Further, a nested hierarchy of bow-tie structures, in which a few intermediate metabolites transform a large number of input and output metabolites, has also been identified in metabolic networks [ 8 ]. However, additional hierarchies of nested subnetworks arise when operational principles, like mass balance and steady-state constraints or optimality of cellular functions, are considered. Such hierarchies include the so-called self-maintained subnetworks (e.g. chemical organizations [ 9 , 10 ]) and subnetworks that operate as a unit when the network optimizes a cellular objective (e.g. flux-modules [ 11 – 13 ] and feasible coalitions [ 14 ]). Moreover, hierarchies that rely on pairwise relationships between components often consider operational principles in conjunction with the network structure. For instance, mass-balance and steady state constraints also induce hierarchies based on asymmetric, pairwise relations among reactions ( Fig 1c ). One such relation is the so-called directional flux coupling relation, in which one reaction controls the activation state of another in any steady state of the network [ 15 ]. The directionality (i.e. asymmetry) of this coupling relation induces a hierarchical organization, in which reactions situated at the top of the hierarchy control the activation state of all reactions accessible below [ 16 ]. 10.1371/journal.pcbi.1007832.g001 Fig 1 Illustration of the flux order relation and comparison to flux coupling relations. (a) In this toy metabolic network, reactions are depicted as arrows and named in the blue squares while metabolites are depicted as circles. Internal reactions are denoted with R, while exchange reactions with E. (b) Flux order graph corresponding to the toy metabolic network depicted in (a), here, a reaction is connected to another by a directional edge if it carries a greater of equal flux in any steady state. (c) Flux couplings graph for the toy metabolic network depicted in (a), directionally coupled relations are depicted as orange arrows, while partially and fully coupled relations are illustrated with double small and large arrows, respectively. Note that the directionally coupled relation is the only one that induces a directed acyclic graph (see main text). Several studies have already revealed the constraints of metabolic network structure and steady-state operation to the levels of associated genes and proteins. For instance, genes associated to stoichiometrically coupled reaction pairs tend to be co-expressed in E. coli [ 17 ] and maize [ 18 ] and to co-evolve in E. coli [ 19 ]. Further, genes associated to reactions in upper layers of the hierarchy based on directional flux coupling tend to be more regulated at the transcriptional level [ 16 ]. In addition, computational and experimental evidence has indicated that reactions in the core of a metabolic network (e.g. the citric acid cycle) generally carry higher flux than reactions in the network periphery [ 20 ], that flux distributions resemble a power law distribution in the case of E. coli [ 21 ] and that genes associated to the reactions carrying higher flux tend to show decreased evolutionary rates [ 20 ]. Yet, it is unclear whether there exists a hierarchy based on flux levels, whereby fluxes of reactions at the top would impose an upper bound to the fluxes of the reactions below in the hierarchy. Moreover, the extent to which such hierarchy would affect the levels of upstream components and their biochemical properties is unclear. Here, we define the flux order relation and characterize the flux order hierarchy found in E. coli ’s metabolic network. Further, we investigate the effects of the flux-based hierarchy on upstream components in the cellular organization. To this end, we use data from E. coli to evaluate if and to what extent the hierarchy induced by the flux order relation is manifested in transcript levels, protein abundances (for growth under three different carbon sources), measured fluxes (for growth on glucose) and enzyme catalytic constants. By comparative analysis with coupling relations, we show that the flux order relation induces a hierarchy in E. coli ’s metabolic network. Finally, we demonstrate that the flux order relation pinpoints reactions limiting the flux through target processes, which has direct biotechnological applications.",
"discussion": "Results & discussion We investigated the flux order relation in the iJO1366 model, a genome-scale metabolic reconstruction of E. coli ’s K-12 strain [ 37 ], and simulated growth in minimal medium under three different carbon sources: glucose, acetate and glycerate ( Methods , section: Model, growth media and data preparation). First, we characterized the flux order relation in the iJO1366 model for the three carbon sources by using techniques from constraint-based metabolic modelling and mathematical optimisation ( Methods , section: Background ad definitions). Next, we analysed the occurrence of metabolic pathways in iJO1366 along the hierarchy. Finally, we evaluated whether the flux order relation was respected by diverse experimental data sets ( Methods , section: Analysis of order relations in data). An online Jupyter Notebook containing the workflow followed in this study as well as additional interactive figures can be accessed from https://robaina.github.io/fluxOrders . Background ad definitions We followed a constraint-based approach to model the operation of the metabolic network at steady state [ 38 ]. In constraint-based metabolic modelling, the space of all feasible distributions of reaction fluxes v at steady state is determined by the linear system Sv = 0, and the flux bound constraints v min ≤ v ≤ v max , in which S , the stoichiometric matrix, captures the structure of the metabolic network ( Methods , section: Background ad definitions, Fig 1a ). We consider a pair of reactions to be flux-ordered if their fluxes satisfy v i ≥ v j in any steady state of the system ( Methods , section: Background ad definitions). The flux order relation can be represented by a directed acyclic graph (DAG), which we call here the flux order DAG ( Methods , section: Reconstruction of the flux order DAG). For instance, Fig 1a displays a toy metabolic network composed of nine reactions and five metabolites. In this example, reactions E 3 and R 2 are flux-ordered, i.e., v E 3 ≥ v R 2 in any steady state, which we denote as E 3 ≥ R 2 . We represent the order relation in the DAG in Fig 1b , where nodes depict reactions and directed edges the flux order relation. Hence, an edge connects reaction E 3 to reaction R 2 . Due to the transitivity of the flux order relation ( Methods , section: Background ad definitions), the flux order DAG contains chains of reactions which are all flux-ordered and in which the first reaction (a root) in the chain imposes and upper flux bound to the remaining reactions. In the example of Fig 1b , we have that R 2 ≥ E 2 and by transitivity E 3 ≥ E 2 . The triad E 3 , R 2 , E 2 forms a flux-ordered chain, E 3 ≥ R 2 ≥ E 2 , in which E 3 is the root. See Section 2.3 of the online Jupyter Notebook for an illustration of a flux-ordered chain in the iJO1366 model. Additionally, we provide a web application to explore the core section of the flux order DAG of iJO1366 in https://robaina.github.io/fluxOrders . The flux order relation in E. coli We identified the flux-ordered reaction pairs for the three carbon sources: glucose, acetate and glycerate in an scenario of aerobic growth, i.e., we constrained the minimum flux through the biomass reaction to be 95% of the maximum possible ( Methods , section: Identification of flux-ordered reaction pairs).The total number of flux-ordered reaction pairs was similar among the three carbon sources, with an average number of 119,019.3, i.e., 5.4% of the total number of reaction pairs in iJO1366 (online Jupyter Notebook, Section 1.3). However, there were 94,071 shared flux-ordered pairs among the three carbon sources, which represented as much as 91.8% of all flux-ordered pairs for glucose, and lower values of 75.33% for glycerate, and 72.49% for acetate. In fact, acetate imposed the largest number of 12,382 (9.54%) exclusive flux-ordered pairs (online Jupyter Notebook, Section 1.3). Further, the pairwise comparison, employing the Jaccard distance, indicated that acetate and glycerate shared more flux-ordered pairs in comparison to glucose (online Jupyter Notebook, Section 1.3). These findings show that flux-ordered pairs are condition-dependent and allow us to distinguish functionality under different carbon sources. A comparison between flux ordering and flux coupling Next, we asked if the flux order relation differed from flux coupling relations, which have already been used to discover hierarchies in metabolic networks [ 16 ]. To this end, we compared the flux order relation with the three types of flux coupling: full, partial and directional [ 15 ]. Flux coupling relations impose constraints on the activity of the reactions and on the ratio of their flux values. Specifically, for a pair of reactions, both full and partial coupling relations impose a bidirectional matching of the reaction activity, i.e. when one of the reactions carries non-zero flux, the other one must also carry non-zero flux (and vice versa). The fully coupled relation implies, in addition, that the ratio of the two fluxes is constant in every steady state, i.e., v i / v j = α , for α ∈ I R , while the partially coupled relation requires a positive, finite bound on the ratio i.e., 0 < v i / v j ≤ β , with β > 0 being a finite upper bound for the entire space of steady states. In contrast, the directional coupling relation only imposes a unidirectional match of the reaction activity, i.e., one (leading) reaction carrying non-zero flux causes another reaction to carry non-zero flux in any steady state. Given the constraints that the flux coupling relations impose in the flux ratio of certain reaction pairs, we investigated whether the flux-ordered pairs were fully represented by the flux coupling relations. First, we found that the total number of flux-ordered pairs was larger than that of coupled pairs. Specifically, while the average number of flux-ordered pairs for the three carbon sources was 119,019.3 (5.4%) (online Jupyter Notebook, Section 1.3), the number of all three types of flux-coupled pairs was 42114 (1.9%). Next, we evaluated the intersection of flux-ordered pairs with the flux-coupled pairs. We found that an average of 22.39% of the flux-ordered pairs over the three carbon sources was also flux-coupled. The largest fraction of flux-coupled pairs among the flux-ordered was found for glucose (25.23%), followed by glycerate (21.31%) and acetate (20.61%). Moreover, the majority (94.4%) of coupled and ordered pairs were directionally coupled ( Table 1 ). Finally, we evaluated the fraction of partially and directionally coupled pairs that were also ordered. Here we did not consider the fully coupled pairs since they are ordered by definition. Across the three carbon sources, we found that an average of 81.7% of the directionally and of 92.7% of the partially coupled were also flux-ordered ( Table 1 ). 10.1371/journal.pcbi.1007832.t001 Table 1 Overlap between flux-ordered and flux-coupled reaction pairs. Conditional probabilities between the sets of flux-ordered reaction pairs (O), the union of all three types of flux-coupled reaction pairs (C), directionally coupled (D), partially coupled (P) and fully coupled (F) pairs across the three carbon sources evaluated in this study. A majority of flux-ordered reaction pairs are not flux-coupled while most flux-coupled pairs are also flux-ordered, thus the flux order relation is not fully represented in all three coupling relations. Pr(C ∣ O) Pr(D∣O) Pr(P ∣ O) Pr(F ∣ O) Pr(O ∣ C) Pr(O ∣ D) Pr(O ∣ P) Glucose 0.3374 0.3187 0.0171 0.0017 0.8158 0.8314 0.9278 Glycerate 0.2710 0.2557 0.0140 0.0014 0.7990 0.8134 0.9252 Acetate 0.2597 0.2450 0.0134 0.0013 0.7955 0.8098 0.9236 Altogether, these results show that while most flux coupled pairs are also flux-ordered, the flux order relation is considerably richer than the three flux coupling relations, and thus can provide novel insights in the hierarchical organization of metabolism. Properties of the hierarchy implied by the flux order relation We next investigated how reactions and metabolic (sub)systems were distributed across the hierarchy implied by the flux order DAG. To this end, we first classified reactions in the DAG into different levels according to their distance to a root reaction. Specifically, we defined the level of a reaction as the number of reactions in the shortest chain to a root. Hence, the larger the level of a reaction, the more constrained the largest flux value it may take in any steady state ( Methods , section: Extracting levels in the flux order DAG). After reconstructing the flux order DAG for the three carbon sources, we found a maximum number of 20 levels for glucose and glycerate, while the DAG for acetate contained one additional level. Further, the majority of reactions were located in the first levels of the DAG in the three carbon sources, specifically, the first six levels contained over 90% of the reactions ( Fig 2a ). The number of sources, i.e., reactions in the first level (i.e., level 0 in Fig 2a ), was markedly greater than the number of sinks, i.e., reactions that do not carry a flux value greater than any other reaction in every steady state. Therefore, in line with observations of purely structural hierarchy in gene regulatory networks, the hierarchies based on flux orders for all three carbon sources is branching, but relatively shallow [ 39 ]. 10.1371/journal.pcbi.1007832.g002 Fig 2 Properties of the flux order directed acyclic graph (DAG). (a) Distribution of flux-ordered pairs per graph level and carbon source, levels one to six contain approximatively 90% of the reactions in the DAG in all carbon sources. (b) Distributions of metabolic macrosystems across graph levels and carbon sources. See the online Jupyter Notebook, Section 2.4, for a complete depiction of the metabolic systems across DAG levels. Next, we asked how metabolic pathways and systems were distributed across the levels of the DAG. The iJO1366 model contains a total of 32 metabolic subsystems. To facilitate the analysis, we grouped the subsystems into eight larger categories, here termed macrosystems ( S1 Table ). We first looked at the distribution of macrosystems among all reactions present in the DAGs of all three carbon sources as well as the sets of reactions which were exclusive to each carbon source. We found that among the set of shared reactions, Transport was the macrosystem with the largest number of reactions, followed by Amino acid and Lipid metabolism . However, Transport was not the dominant macrosystem in the set of acetate-exclusive reactions since this position was occupied by lipid metabolism ( Fig 2b ). We also analysed how the macrosystems were distributed in the levels of the DAGs. To this end, we computed the frequency of each macrosystem in each level of the DAG for each of the carbon sources. We found that macrosystems were not uniformly represented across levels. Instead, each macrosystem tended to be predominant in a particular region of the hierarchy. For instance, four macrosystems were only present within the first eight levels: Nucleotide metabolism in the first six levels, Amino acid metabolism and Lipid metabolism in the first seven levels and Carbohydrate metabolism in the first eight levels. Additionally, while Cell wall biosynthesis was also only represented in the first ten levels of the DAG, this macrosystem was predominant in levels six to eight, comprising between 52% and 82% of the reactions. Contrary to these macrosystems, Cofactor and vitamin metabolism was represented across all levels, with the exception of level eight. However, the latter macrosystem was markedly predominant in the last ten levels of the DAG, comprising over 86% of the reactions in the three carbon sources. Finally, Transport was not only the macrosystem with the largest overall representation in the DAG, but was also represented across a majority of levels, a result that reflects the utter dependence of reaction flux on import and export as well as intercellular transport processes (online Jupyter Notebook, Section 2.4). Altogether, these findings indicate that specific metabolic systems have a preferential position in the metabolic hierarchy. Therefore, the degree to which a reaction can control the flux of others depends on the specific metabolic system to which it belongs. The flux order relation in experimental data We next evaluated the extent to which molecular data profiles respected the flux order relation. First, we considered whether 13 C-estimated fluxes respected the predicted flux order relation. We also tested if the flux order relation was reflected in other data profiles which do not directly correspond to measurement of fluxes. Specifically, we also considered protein and transcript levels, which potentially correlate with the amount of active enzyme, as well as enzyme turnover number, i.e., k cat ( Methods , section: Model, growth media and data preparation). In all cases, we computed the distribution of differences of reaction-associated data values for all flux-ordered reaction pairs. In the cases where several data values were available for each reaction, i.e., flux, protein and transcript levels, we computed the average difference among all available data values for each flux-ordered reaction pair. We then employed the proportion of flux-ordered pairs with positive (averaged) data differences as our statistic to quantify the agreement of each data set with the flux order relation. Additionally, we performed a permutation analysis in which we randomized data values to evaluate the significance of our results ( Methods , section: Analysis of order relations in data). As commented in section: The flux order relation in E. coli , the sets of flux-ordered pairs previously analyzed were generated in an scenario of aerobic growth, in which we constrained the minimum flux through the biomass reaction to be 95% (controlled though a parameter, α , here α = 0.95) of the maximum possible. However, experimental measurements could correspond to a scenario where E. coli was growing at a larger or smaller rate. Thus, we also analyzed the flux order relation with a smaller, α = 0.925 and a larger, α = 1, minimum flux through the biomass reaction. As expected, the number of flux-ordered pairs decreased rapidly with α values smaller than 0.925 (online Jupyter Notebook, Section 6). The flux order relation is respected by flux data We found that all experimentally estimated flux values respected the flux order relation in the iJO1366 model. Specifically, all averaged mean differences were positive with an experimental p -value = 0.0015. However, experimental flux levels are obtained by fitting 13 C-labelling data to a metabolic network, typically considering core metabolic reactions. It is hence possible that the flux order relation observed in the flux data is imposed by the network used to estimate the fluxes and may thus represent artifacts. To investigate this possibility, we analysed the flux order relation in the small network employed to generate most of the samples contained in the flux database employed in this study [ 34 ], here termed experimental network ( Methods , section: Flux orders in 13 C-measured fluxes).We identified a total of 92 ordered reaction pairs in the experimental network, out of which 25 ordered pairs had available flux data. We found that, here too, flux data followed the order relation in all cases. Interestingly, the overlap between the sets of ordered reaction pairs of iJO1366 and the experimental network was small, with only five shared pairs with available data, a 20% of the ordered pairs with data in iJO1366 (online Jupyter Notebook, Section 3.1). Therefore, a majority of the flux-ordered reaction pairs followed by flux data in iJO1366 are genuine of the whole network and are not induced by the experimental network employed to estimate the flux data. Finally, we evaluated whether flux data also followed the flux order relation when the minimum flux through the biomass reaction was constrained to be 100% of the maximum (smaller values than 95% did not render any flux-ordered pairs with available flux data). We found that the previous results did not change qualitatively. In this case, 93% of the averaged mean differences were positive, with an experimental p-value = 0.0006 (online Jupyter Notebook, Section 3.1). Evaluation of the order relation in transcript and protein levels Any enzyme has an associated metabolic cost, i.e., the amount of metabolic resources that are invested to produce one unit of enzyme. We hypothesized that, in a scenario where cells minimise costs, regulating gene expression and protein levels of enzymes to match the flux order relation of the catalysed reactions could be economically advantageous due to avoiding overproduction of unneeded enzymes. Moreover, metabolic costs vary greatly among enzymes, which can be explained by both, the difference in protein size and also by an enrichment in costly amino acids [ 40 ]. Violation of the flux order relation could thus have a larger impact in reaction pairs catalysed by enzymes with larger metabolic costs. To test this hypothesis, we analysed to what extent transcript and protein data followed the flux order relation in the three carbon sources. In addition, we evaluated whether enzyme subgroups with increasing metabolic costs tended to better follow the flux order relation. We first obtained publicly available enzyme costs [ 30 ] and then evaluated the flux order relation in the data for several subsets of reactions with increasing enzyme costs. Specifically, we evaluated the following subsets: ( i ) all flux-ordered reaction pairs, ( ii ) flux-ordered pairs whose enzymes had a cost at least as large as the percentile P 70 , ( iii ) P 80 and ( iv ) P 85 of the distribution of all employed enzyme costs ( Methods , section: Model, growth media and data preparation). We found that the proportion of ordered reaction pairs whose transcript data matched the ordering increased across reaction groups with increasing protein costs with glucose as a carbon source ( Table 2 and online Jupyter Notebook, Section 3.2). Specifically, the proportion of positive average data differences increased from 62.8%, for P 0 , to 70.8%, for P 70 , 77.2% for P 80 and 83.5% for P 85 (for all p -value <5 x 10 −4 , permutation test). However, the tendency found for glucose was not observed for the other carbon sources. Specifically, while a similar result held to the one obtained for glucose at P 0 , i.e., a proportion of positive average data differences of 62.2% and 62.2% for acetate and glycerate, respectively, none of the alternative carbon sources showed a clear increase of the proportion for larger protein costs ( Table 2 and online Jupyter Notebook, Section 3.2). 10.1371/journal.pcbi.1007832.t002 Table 2 The flux order relation in experimental data. Fraction of positive average differences among flux-ordered pairs across the five data types employed, i.e. flux, transcript and protein levels and enzyme k cat values, for the three carbon sources and with a minimum biomass production of 95% of the maximum (i.e., α = 0.95). In the case of data on transcript and protein levels, results are shown for the four different protein cost percentile values, P 0 — P 85 (see main text). Since k cat values can be considered constant under different carbon sources, here we employ the same dataset for the three carbon sources. Experimental p -values from the permutation test are displayed within parentheses. Interactive figures displaying the actual distributions of average data differences are displayed in the online Jupyter Notebook, Section 3. Fluxes Transcript Protein k cat P 0 P 70 P 80 P 85 P 0 P 70 P 80 P 85 Glucose 1 (0.002) 0.628 (10 −4 ) 0.708 (10 −4 ) 0.772 (10 −4 ) 0.835 (10 −4 ) 0.597 (0.012) 0.621 (0.04) 0.707 (0.03) 0.759 (0.022) 0.792 (10 −4 ) Glycerate - 0.626 (10 −4 ) 0.624 (0.032) 0.665 (0.014) 0.663 (0.05) 0.598 (0.008) 0.626 (0.048) 0.727 (0.014) 0.779 (0.014) 0.79.5 (10 −4 ) Acetate - 0.622 (10 −4 ) 0.654 (0.002) 0.621 (0.07) 0.635 (0.068) 0.576 (0.036) 0.549 (0.194) 0.581 (0.138) 0.648 (0.098) 0.79.8 (10 −4 ) We also found a tendency towards increasing the agreement between protein abundance and the flux order relation with increasing protein costs for glucose and glycerate. Specifically, the proportion of positive average differences, with glucose as sole carbon source, increased from 59.7% for P 0 to 62.1% for P 70 , 70.7% for P 80 and 75.9% for P 85 , and from 59.8% for P 0 to 62.6% for P 70 , 72.7% for P 80 and 77.9% for P 85 with glycerate as a carbon source ( p -values < 0.04, permutation test, Table 2 and online Jupyter Notebook, Section 3.2). Similar to transcript data, protein abundance with acetate as a carbon source did not agree well with the flux order relation, and only the first case, at P 0 , with a proportion of 57.6% positive average differences was statistically significant ( p -value < 0.03, permutation test, Table 2 and online Jupyter Notebook, Section 3.2). Therefore, we found that the agreement between data profiles and the flux order relation increased for reaction pairs with higher protein costs, although this trend was only clear for transcript level data with glucose as a carbon source and for protein abundance data with glucose and glycerate as a carbon source. Finally, we evaluated both transcript levels and protein abundance under different minimum flux through the biomass reaction. In the case of transcript levels with α = 0.925 and glucose as carbon source the proportion of positive average differences increased from 62% at P 0 to 72% at P 80 ( p -values < 0.02, permutation test, online Jupyter Notebook, Section 3.2), while the other carbon sources did not render significant results. When setting α = 1, we found significant fractions of positive average differences among transcript levels only at P 0 in the three carbon sources, 65%, 63% and 64% for glucose, glycerate and acetate (online Jupyter Notebook, Section 3.2). In the case of protein abundances, we found proportions of positive average differences above 84% for the three carbon sources with α = 0.925 and α = 1, but only when protein costs were at least as large as the percentiles P 80 and P 85 (online Jupyter Notebook, Section 3.2). Evaluation of the order relation in enzyme k cat values We next investigated if the turnover numbers, i.e. k cat values, followed the flux order relation. k cat values are a fundamental, structural property of enzymes and, thus, independent of carbon source. Therefore, we evaluated the same dataset on the sets of ordered reaction pairs generated under the three carbon sources ( Methods , Section Model, growth media and data preparation). Here again, we found that data tended to follow the predicted flux order relation. Specifically, the proportion of positive average differences was 79.2%, 79.5% and 79.8% for glucose, glycerate, and acetate, respectively ( Table 2 and online Jupyter Notebook, Section 3.3, p -values = 10 −4 , permutation test). Additionally, k cat values followed the same trend for α = 0.925 and α = 1, with fractions of positive average differences above 76% for the three carbon sources (online Jupyter Notebook, Section 3.3). The previous results indicate that k cat values are the data type with the second-best agreement with the flux order relation, dominated only by 13 C-based flux estimations. Additionally, since k cat depends on enzyme structure, our results suggest optimisation of enzyme structure throughout evolution to match the ordering pattern imposed by the metabolic network at steady state. Further, since k cat values are measured under substrate saturating conditions [ 28 ], our results align with in vivo saturating conditions for a subset of enzymes, as already observed for the central carbon metabolism of E. coli [ 41 ]. Effect of the level in the flux order DAG on the number of gene regulatory interactions As previously discussed, reactions situated in the first levels of the flux order DAG impose an upper bound to the fluxes of a large number of reactions in the lower levels. Therefore, regulating the flux through these reactions can have a greater impact in controlling the overall flux distribution. A similar concept was explored by Hosseini et al. [ 16 ], where the DAG representing the directional coupling relation in a metabolic model of E. coli was explored with respect to the distribution of regulatory interactions at each level. This analysis found that genes associated to reactions situated in the first levels were more likely to be transcriptionally regulated. We conducted a similar analysis with the flux order DAG of E. coli , although this time we evaluated to what extent genes of reactions in prior levels exhibited more regulatory interactions than those in later levels. First, we computed the total number of regulatory interactions of the genes associated to each reaction with available gene regulatory data, which we obtained from the RegulonDB database [ 35 ]. Next, similar to Hosseini et al. [ 16 ], we partitioned the DAG levels into three sets—first, middle, last—of levels which had a similar number of reactions and compared the distributions of total regulatory interactions between the sets (Mann-Whitney U test, α = 0.05). Finally, we also analysed the flux coupling DAG of the iJO1366 model using the RegulonDB database ( Methods , section: Gene regulation in the flux order and flux coupling DAGs). We found that in both, the flux order and the flux coupling DAG, genes associated to reactions in the first and middle levels were significantly more regulated than those in the last level ( p -values < 0.04, Fig 3 and online Jupyter Notebook, Section 3.4). Additionally, we obtained identical results when we constrained the flux through the biomass reaction to be the maximum possible, i.e., with α = 1. In the case of α = 0.925, only genes of reactions in the middle levels were significantly more regulated than those in the last levels (online Jupyter Notebook, Section 6). Therefore, these results support the hypothesis that genes associated to reactions in higher positions, in both DAGs, are more regulated due to their greater control capability. 10.1371/journal.pcbi.1007832.g003 Fig 3 Evaluation of the flux order relation in E. coli ’s gene regulatory network. Relationship between the number of gene regulatory interactions of the enzyme-coding genes and the position in the hierarchy of the corresponding reactions. (a) The flux order DAG levels have been grouped into three categories: those occupying the first, middle and last levels (see main text). The distribution of the total number of gene regulatory interactions (both positive and negative) is represented for the three categories as box plots. The green line represents the median number of interactions while the red triangle represents the mean. (b) A similar analysis employing the directional flux coupling DAG. In both cases, the first and the middle and the middle and the last levels contain significantly larger numbers of gene regulatory interactions (p-value < 0.04, see main text). Flux order relation and reaction essentiality Identifying essential reactions—reactions that must carry non-zero flux to enable cellular growth—is crucial in both biotechnological and biomedical settings. For instance, in the first case, essential reactions and their genes form a set that cannot be removed during metabolic pathway optimisation approaches. In the second case, essential reactions and their enzymes provide potential targets to halt the growth of pathogens [ 42 – 44 ]. All reactions that are flux-ordered with the biomass (pseudo)reaction are also essential, due to the upper flux bound imposed on biomass production. However, it is unclear whether all essential reactions are also flux-ordered with the biomass reaction. To clarify this issue, we next investigated to what extent essential reactions were also flux-ordered with the biomass reaction ( Methods , section: The flux order relation in the context of reaction essentiality). We found a total of 27 reactions that were both, essential and flux-ordered with the biomass reaction under growth on glucose, i.e., were predecessors of the biomass reaction in the flux order DAG ( Fig 4 and online Jupyter Notebook, Section 4). Flux-ordered and essential reactions were organized in four levels in which four metabolic macrosystems were represented, in decreasing order of the number of reactions these were: Transport , Carbohydrate metabolism , Amino acid metabolism and Energy and maintenance . However, we found that most essential reactions (a total of 347) were not flux-ordered with the biomass reaction, demonstrating that reactions do not need to be flux-ordered with the biomass reaction to be essential (online Jupyter Notebook, Section 4). We conclude that the flux order relation can be used to identify other, subtle limitations to the biomass reaction flux, i.e., reactions that continuously constraint biomass production as opposed to the binary switch imposed by the unordered essential reactions. The latter can be readily used in relevant biotechnological applications, such as control and fine-tuning of growth-limiting genes. 10.1371/journal.pcbi.1007832.g004 Fig 4 Subgraph of the flux order DAG containing all predecessors of the biomass reaction for growth under glucose. All reactions represented in this DAG carry greater or equal fluxes than the biomass reaction. Hence, they are all essential reactions since they would impose an upper flux bound of zero to the biomass reaction if they were to be inactive. Only four reaction macrosystems are represented in the set of reactions: Amino acid metabolism, carbohydrate metabolism, Energy and maintenance and Transport, with transport being the one with the largest number of reactions. ATPS4rpp: ATP synthase, EX_h2o: H 2 O exchange, H2Otex_rev: H 2 O transport periplasm to extracellular, H2Otpp_rev: H 2 O transport cytoplasm to periplasm, GAPD: Glyceraldehyde-3-phosphate dehydrogenase, ENO: Enolase, PGK_rev: Phosphoglycerate kinase (reverse), PGM_rev: Phosphoglycerate mutase (reverse), CYTBO3_4pp: Cytochrome oxidase, CO2tpp_rev: CO 2 transporter cytoplasm to periplasm, CO2tex_rev: CO 2 transport periplasm to extracellular, EX_co2: CO 2 exchange, O2tpp: O 2 transport periplasm to cytoplasm, EX_o2: O 2 exchange, O2tex: O 2 transport extracellular to periplasm, CS: Citrate synthase, ACONTb: Aconitase, ICDHyr: Isocitrate dehydrogenase, ACONTa: Aconitase, ASAD_rev: Aspartate-semialdehyde dehydrogenase (reverse), ASPK: Aspartate kinase, GLNS: Glutamine synthetase, ASPTA_rev: Aspartate transaminase (reverse), NH4tpp: Ammonia transport periplasm to cytoplasm, NH4tex: Ammonia transport extracellular to periplasm, EX_nh4: Ammonia exchange, EX_glc_D: Glucose import. Conclusion Understanding hierarchies in biological networks not only can contribute to the discovery of underlying design principles, but also provide the means for effective control of these complex networks. Hierarchies in biological networks are multifaceted, and thus two principle approaches have been used to formalize and uncover them, namely, hierarchies of embedded subnetworks and hierarchies of individual components of the network. Since metabolic networks are composed of metabolites interconverted by biochemical reactions, the hierarchies following the second approach can be based on ordering of either metabolites of reactions. Here, we wanted to explore the extent to which the structure of the metabolic network alone imposes a hierarchy of reactions based on ordering of their fluxes. Such a view provides the means to discover a hierarchy that is present irrespective of the particularities of the reaction kinetics, for which there is still limited information at genome-wide scale. Moreover, this view can be readily cast in the constraint-based modelling framework and can contribute to the discovery of important functional limits of metabolic networks. The latter is due to the natural definition of the flux ordering relation, whereby reactions further up the hierarchy dominate reactions down the hierarchy with respect to their flux in every steady state supported by the network. Our formalization of the flux order relation facilitated the discovery of a hierarchy of reactions in the metabolic network reconstruction of the bacterium E. coli . Analysis of the graph representation pointed out that the hierarchy is branched and relatively shallow for three carbon sources. In addition, we found a complete agreement between the flux order relation and 13 C-based flux profiles. Interestingly, we also found a partial agreement when analysing other phenotypic profiles which correlate with reaction flux, i.e., transcript and protein levels, as well as enzyme catalytic constants. In the first case, these findings point at optimisation of resource allocation to minimise costs of enzyme synthesis. This hypothesis that was further supported by an increased agreement found in subsets of enzymes with increasing costs, particularly when glucose was the carbon source, in transcript and protein levels, and when glycerate was the carbon source, in protein levels. In the second case, the findings suggest network constraints acting upon the evolution of enzyme structural properties to match the flux order relation. The concept of flux order relation in metabolic networks can be readily extended to analyse the hierarchy outside of steady state dynamics, provided we have insights in the degree of metabolic concentration changes around a steady state [ 45 , 46 ]. In addition, due to the universality of a large part of metabolism, it is plausible that the flux order relation also holds in metabolic networks of other organisms. It would be particularly interesting to evaluate to what extent the hierarchy holds in eukaryotic organisms, in which metabolism is compartmentalized and more complex. Further, one can expand the concept for applications to metabolic networks with given analytically tractable kinetics, e.g. mass action [ 47 ]. Finally, since the flux order relation identifies reactions imposing a fine-tuned control over a target reaction, the applications of the hierarchy in metabolic engineering remain as a promising direction for future research."
} | 11,202 |
26878871 | PMC4824201 | pmc | 9,278 | {
"abstract": "Facilitating the evolution of new gene functions, gene duplication is a major mechanism driving evolutionary innovation. Gene family expansions relevant to host/symbiont interactions are increasingly being discovered in eukaryotes that host endosymbiotic microbes. Such discoveries entice speculation that gene duplication facilitates the evolution of novel, endosymbiotic relationships. Here, using a comparative transcriptomic approach combined with differential gene expression analysis, we investigate the importance of endosymbiosis in retention of amino acid transporter paralogs in aphid genomes. To pinpoint the timing of amino acid transporter duplications we inferred gene phylogenies for five aphid species and three outgroups. We found that while some duplications arose in the aphid common ancestor concurrent with endosymbiont acquisition, others predate aphid divergence from related insects without intracellular symbionts, and still others appeared during aphid diversification. Interestingly, several aphid-specific paralogs have conserved enriched expression in bacteriocytes, the insect cells that host primary symbionts. Conserved bacteriocyte enrichment suggests that the transporters were recruited to the aphid/endosymbiont interface in the aphid common ancestor, consistent with a role for gene duplication in facilitating the evolution of endosymbiosis in aphids. In contrast, the temporal variability of amino acid transporter duplication indicates that endosymbiosis is not the only trait driving selection for retention of amino acid transporter paralogs in sap-feeding insects. This study cautions against simplistic interpretations of the role of gene family expansion in the evolution of novel host/symbiont interactions by further highlighting that multiple complex factors maintain gene family paralogs in the genomes of eukaryotes that host endosymbiotic microbes.",
"introduction": "Introduction Gene family expansion is a key player in the evolution of innovation ( Ohno 1970 ; Arnegard et al. 2010 ; Deng et al. 2010 ; Kondrashov 2012 ; Voordeckers et al. 2012 ), but elucidating the factors maintaining paralogs in a genome can be tricky ( Innan and Kondrashov 2010 ). The vast majority of paralogs that arise in a population are lost through stochastic birth and death processes ( Lynch and Conery 2000 ), presenting an evolutionary conundrum when ancient paralogs are found. A complete understanding of paralog retention requires detailed knowledge of a number of traits, such as paralog expression and function, and the ecological and molecular mechanisms driving paralog diversification. While expression, function, ecology, and molecular traits facilitate explaining the maintenance of anciently acquired paralogs, obtaining data about all these traits is often intractable. In these cases, a combination of natural history and phylogeny of the taxa of interest can provide a framework for proposing the ecological and biological factors that underlie paralog retention. One trait that may drive paralog maintenance is the evolution of endosymbiosis between multicellular eukaryotes and microorganisms. In endosymbiosis, microbial symbionts reside intracellularly within hosts, commonly providing a nutritional benefit. New gene paralogs may facilitate the establishment and maintenance of a symbiotic partnership through expression diversification (e.g., from a nonsymbiotic tissue to a symbiotic tissue) or possibly through functional diversification (e.g., toward a function specialized for endosymbiosis). Indeed, genomic studies suggest a role for gene duplication in endosymbiosis—these studies have found lineage-specific duplication in genes functionally relevant to symbiotic interactions across systems as divergent as the legume /Rhizobia endosymbiosis ( Young et al. 2011 ), the endosymbiosis between corals or anemones and Symbiodinium ( Shinzato et al. 2011 ; Baumgarten et al. 2015 ), and the endosymbiosis between the pea aphid Acyrthosiphon pisum and its endosymbiont Buchnera aphidicola ( Huerta-Cepas et al. 2010 ; Price et al. 2011 ). A. pisum , a sap-feeding insect, has undergone expansions in over 2000 gene families ( Huerta-Cepas et al. 2010 ; International Aphid Genomics Consortium 2010 ), some of which may be mechanistically important for its relationship with Buchnera . Particularly intriguing in A. pisum are lineage-specific expansions in amino acid transporter genes ( Price et al. 2011 ; Duncan et al. 2014 ; Dahan et al. 2015 ; Wilson and Duncan 2015 )—genes whose membrane-bound protein products are crucial for nutritional exchange between A. pisum and Buchnera ( Price et al. 2014 , 2015 ). Indeed, amino acid transporters are over-represented among A. pisum gene families that underwent large expansions resulting in more than ten paralogs ( Huerta-Cepas et al. 2010 ). Similarly, we recently discovered that several sap-feeding insects (also with endosymbionts) experienced lineage-specific expansions in amino acid transporters. Interestingly, independent expansion of amino acid transporters in multiple sap-feeding insect lineages is a pattern that parallels other independently evolved signatures of host/endosymbiont genome co-evolution ( Wilson and Duncan 2015 ). The mechanistic importance of gene duplication in amino acid transporters for endosymbiosis in these insects is further supported by the observation that some lineage-specific paralogs in A. pisum and the citrus mealybug Planococcus citri are enriched in bacteriocytes ( Price et al. 2011 ; Duncan et al. 2014 ), the specialized insect cells where symbionts reside. Bacteriocyte enrichment of some paralogs implies paralog evolution to operate in a symbiotic context because bacteriocytes represent the interface between the insect host and the endosymbiont. This host/endosymbiont interface is made up of three membrane barriers separating host tissues from endosymbionts: (1) The plasma membrane surrounding the bacteriocyte, (2) the insect-derived membrane surrounding individual symbiont cells (symbiosomal membrane), and (3) the inner and outer bacterial membranes of each symbiont cell. While the symbiosomal membrane is the most immediate interface between host (bacteriocyte cytoplasm) and endosymbiont, the bacteriocyte plasma membrane is also an important part of the host/endosymbiont interface because of the role it plays in regulating metabolic output of the symbiont ( Price et al. 2014 ). In addition to independent recruitment of amino acid transporter paralogs to the host/symbiont interface of both A. pisum and P. citri , tests for signatures of selection in the pea aphid found, in one expansion, an elevated rate of evolution in the transition from high gut expression to enriched bacteriocyte expression. The elevated rate of evolution suggests functional evolution corresponding to a shift toward symbiotic expression ( Price et al. 2011 ). Lastly, expansions in sap-feeding insects are significantly associated with increased gene duplication rates and decreased gene loss rates ( Dahan et al. 2015 ), supporting an adaptive explanation for the retention of duplicate amino acid transporters—an explanation that may relate to shared traits among these insects, such as endosymbiosis with nutrient-provisioning bacteria. Despite the evidence supporting endosymbiosis as a factor influencing the retention of amino acid transporter paralogs in sap-feeding insects, transcriptomic and expression data indicate that other biological factors are also at play. For example, in aphids, some paralogs show biased expression in males ( Duncan et al. 2011 ), where symbionts are less abundant than in females ( Douglas 1989 ). In fact, accelerated rates of evolution also correlate with the evolution of male-biased expression in one male-biased paralog, suggesting a derived sex-biased function. Further, despite bacteriocyte enrichment in some citrus mealybug and pea aphid paralogs, most paralogs are not enriched in bacteriocytes ( Price et al. 2011 ; Duncan et al. 2014 ), suggesting that other features of these insects influence paralog retention. Lastly, while we found lineage-specific duplications in amino acid transporters of four sap-feeding insects—pea aphids ( Price et al. 2011 ), citrus mealybugs, potato psyllids, and whiteflies ( Duncan et al. 2014 )—we found no evidence for duplication in the sap-feeding cicada ( Duncan et al. 2014 ), indicating that gene duplication in amino acid transporters is not necessary for the evolution of endosymbiosis between sap-feeding insects and bacteria. These data do not rule out the possibility that gene duplication has played an important role in the evolution and maintenance of endosymbiosis in some sap-feeding insects. Even so, evidence that other factors influence paralog retention makes it unclear if endosymbiosis plays a primary or secondary role in selection for the maintenance of amino acid transporter paralogs. Here, using a comparative transcriptomic approach, we leverage the phylogeny and natural history of aphids and their close relative, the grape phylloxera Daktulosphaira vitifoliae (see table 1 and fig. 1 for taxonomic classifications and relationships among taxa in this study), to investigate life history traits underlying the retention of amino acid transporter paralogs. Aphids and phylloxera belong to different families of the hemipteran group Aphidomorpha—Aphididae and Phylloxeridae, respectively. The common ancestor of Aphididae established an endosymbiotic relationship with the bacterium B . aphidicola around 160–280 MYA ( Moran et al. 1993 ), and since that initial infection, Buchnera has been vertically inherited by nearly all extant aphids. In contrast, Phylloxeridae, including D. vitifoliae , lacks Buchnera or other intracellular symbionts ( Vorwerk et al. 2007 ; Medina et al. 2011 ). If endosymbiosis between aphids and Buchnera initially drove paralog retention in amino acid transporters, we expect to find that duplication took place, at least initially, in the aphid common ancestor. If, however, duplication predates the aphid/phylloxera split or postdates aphid diversification, then the initial trait influencing paralog maintenance is more likely not endosymbiosis with Buchnera , but another trait that evolved concurrently with gene duplication.\n F ig . 1.— Phylogenetic relationships among sampled taxa. Acquisition of the aphid endosymbiont Buchnera and relevant higher taxonomic classifications are mapped. Tree structure is based on phylogenetic analyses reported in Nováková et al. (2013) , Misof et al. (2014) and Dahan et al. (2015) . \n Table 1 Taxon Sampling Within Aphidomorpha Family Subfamily Tribe Genus species Phylloxeridae D. vitifoliae Aphididae Eriosomatinae Pemphigini P. obesinymphae Tamaliinae T. coweni Aphidinae Aphidini A. nerii Macrosiphini M. persicae A. pisum",
"discussion": "Discussion Recent studies on symbiotic organisms support a role for gene duplication in the evolution of endosymbiosis ( Price et al. 2011 ; Shinzato et al. 2011 ; Young et al. 2011 ; Duncan et al. 2014 ; Baumgarten et al. 2015 ; Dahan et al. 2015 ). However, most studies do not address the possibility of a gene duplication/endosymbiosis connection in more than one species and, as we have found in sap-feeding insects, nonsymbiotic traits may influence the retention of duplicate genes ( Duncan et al. 2011 , 2014 ). To gain insight into the role of endosymbiosis in retaining aphid-specific amino acid transporter paralogs, we used a comparative transcriptomic approach to pinpoint the timing of amino acid transporter duplications in the Aphidomorpha ( table 1 and fig. 1 ). Our results support a complex and dynamic evolutionary history of amino acid transporters in these insects—a history that was likely shaped by multiple biological and ecological factors. In support of a role for gene duplication in the evolution of endosymbiosis, we inferred several duplication events in the aphid common ancestor, corresponding to the acquisition of the aphid endosymbiont, Buchnera ( figs. 1–3 ). However, we also inferred duplication events both earlier and later than the evolution of endosymbiosis in aphids. Duplication at these earlier and later time scales implies that gene duplication and retention has also been driven by factors other than endosymbiosis. The slimfast Expansion Was Not Driven by the Evolution of Endosymbiosis We posit that the aphid/phylloxera slimfast expansion, at least initially, was not driven by endosymbiosis. Phylloxera lack an endosymbiont, and assuming that the shared ancestor of aphids and phylloxera also lacked an endosymbiont, the expansion predated the evolution of endosymbiosis. However, the relationship of the aphid/phylloxera lineage, derived from within Sternorrhyncha, suggests that endosymbiosis originated in the common ancestor of Sternorrhyncha and was secondarily lost in phylloxera. Even if endosymbiosis originated in the sternorrhynchan common ancestor, the slimfast expansion likely was not driven primarily by primary endosymbiosis because the expansion postdates aphid divergence from other major sternorrhynchan lineages ( Duncan et al. 2014 ). Furthermore, despite their shared ancestry, sternorrhynchan lineages may have evolved endosymbiosis independently, supported by ongoing discoveries of convergent patterns of host/symbiont genome coevolution (reviewed by Wilson and Duncan 2015 ). Given the timing of the slimfast expansion, its origin is more likely influenced by a trait shared by aphids and phylloxera, such as their complex life cycle that involves both sexual and asexual reproduction ( Blackman and Eastop 2000 ; Forneck and Huber 2009 ). Indeed, we previously reported on male-biased and asexual female-biased slimfast paralogs in aphids ( Duncan et al. 2011 ), supporting the notion that slimfast paralogs were retained as a result of selection for divergence to fulfill sex-specific roles. Dynamic Evolution of Amino Acid Transporters in Aphids APC Family Despite evidence that the slimfast expansion was driven by a shared trait between aphids and phylloxera, retention of slimfast paralogs in aphids was most likely influenced by more complex and dynamic factors and processes. For example, gene duplication within the slimfast expansion continued after aphids and phylloxera diverged ( fig. 2 ), implying that (1) additional selective pressures, perhaps shifting from ancestral selective pressures, influenced the retention of additional slimfast paralogs as they emerged, (2) additional slimfast paralogs emerged by chance and were retained through nonadaptive processes (e.g., the classic model of subfunctionalization known as Duplication, Degeneration, Complementation [ Force et al. 1999 ]), or (3) aphids are particularly prone to gene duplication. Notably, these three possibilities are not mutually exclusive and could all be operating. Another important aspect of paralog evolution in the slimfast expansion is highlighted by two pairs of A. pisum / M. persicae paralogs that have conserved bacteriocyte enrichment: ACYPI008904 / Mper-APC08 and ACYPI005118 / Mper-APC10 . A change in expression in these two orthologous pairs from the typical gut expression of slimfast ( Price et al. 2011 ) toward bacteriocyte enrichment implies that these slimfast paralogs were recruited to the aphid/ Buchnera symbiotic interface and retained for a role in endosymbiosis. In addition to being recruited to bacteriocytes, ACYPI008904 and Mper-APC08 may have diverged functionally from other members of their subclade ( fig. 2 ). Subclade 3 of the slimfast expansion experienced gene duplication in the common ancestor of Macrosiphini, resulting in multiple orthologous pairs for A. pisum and M. persicae while the other three aphid species each have only one gene. Gene duplication in subclade 3 provides an opportunity for functional divergence among paralogs while still fulfilling their ancestral function in two ways: (1) by evolving novel function and/or expression (neofunctionalization) or (2) by partitioning, and possibly optimizing or specializing, ancestral functions in sister paralogs (subfunctionalization). Indeed, supporting divergence following gene duplication in subclade 3, while ACYPI008904 and Mper-APC08 are both highly enriched in bacteriocytes, the closely related paralogs ACYPI002633 and Mper-APC09 are both bacteriocyte-depleted ( Price et al. 2011 ) ( fig. 3 ). In fact, our previous work in A. pisum revealed that ACYPI008904 and ACYPI002633 have very different expression profiles—while ACYPI008904 is enriched in bacteriocytes ( Price et al. 2011 ) as well as asexual adult females ( Duncan et al. 2011 ), ACYPI002633 is enriched in both gut ( Price et al. 2011 ) and adult male A. pisum ( Duncan et al. 2011 ). In addition, the branches leading to both ACYPI002633 and the clade containing ACYPI008904 and ACYPI008323 experienced accelerated rates of evolution ( Duncan et al. 2011 ; Price et al. 2011 ), supporting the possibility that gene duplication was followed by both expression and functional divergence. Without expression data for the other three aphid species, we cannot infer if the different expression profiles of the A. pisum and M. persicae APC paralogs result from neofunctionalization or subfunctionalization, both of which have implications for the role of endosymbiosis in the evolution of amino acid transporters. If bacteriocyte expression is novel in ACYPI008904 and Mper-APC08 , then amino acid transporter recruitment to bacteriocytes is an ongoing process in aphid evolution. In contrast, if single-copy aphid orthologs are also expressed in bacteriocytes, then we could infer that Buchnera infection coincided with recruitment of the single-copy, ancestral gene of subclade 3 to aphid bacteriocytes. In light of expression data for M. persicae ( supplementary file S6 , Supplementary Material online ) and A. pisum ( Hansen and Moran 2011 ; Price et al. 2011 ; Macdonald et al. 2012 ), we predict that the ancestral aphid gene of subclade 3 operated at the symbiotic interface, coincident with Buchnera infection. After all, the sister clade, subclade 4, contains the only additional slimfast paralogs with conserved bacteriocyte enrichment between A. pisum ( ACYPI005118 ) and M. persicae ( Mper-APC10 ). Thus, parsimony would predict that aphid slimfast members in both subclades 3 and 4 inherited bacteriocyte expression from their common ancestor—or, since subclades 3 and 4 appeared before the aphid/phylloxera split, their ancestral gene may have been expressed in the cells that gave rise developmentally to aphid bacteriocytes after the aphids and phylloxera diverged from a common ancestor ( Wilson and Duncan 2015 ). A scenario in which bacteriocyte expression evolved in single-copy aphid orthologs would imply that the ancestral aphid gene of subclade 3 operated in endosymbiosis in addition to other, previously defined functions in place prior to aphid/phylloxera divergence. Importantly, there is a precedent for genes with broad expression to play important roles in endosymbiosis. Indeed, the A. pisum AAAP member ACYPI001018 (also known as ApGLNT1 ) is both globally highly expressed and also a key regulator of Buchnera metabolic output ( Price et al. 2014 , 2015 )—a role that is very possibly conserved in M. persicae , given the conserved high bacteriocyte expression of the ortholog Mper-AAAP05 . AAAP Family Amino acid transporters in the AAAP family, like the slimfast paralogs, expanded at different time scales during aphid evolution. AAAP members resulting from duplication in the aphid common ancestor, including ACYPI000536 / Mper-AAAP1 1 and ACYPI001366 / Mper-AAAP03 , are prime candidates for facilitating transporter divergence toward a role in endosymbiosis. The fact that both these pairs of orthologs have conserved bacteriocyte enrichment while related paralogs have different expression profiles ( fig. 4 , supplementary file 6 , Supplementary Material online ) suggests that gene duplications in the aphid common ancestor were followed by functional divergence to symbiotic and nonsymbiotic roles. Our phylogeny also supports three gene duplications predating the diversification of the tribe Macrosiphini ( fig. 3 ). Retention of these sets of orthologs is unlikely to have been driven by acquisition of the primary endosymbiont Buchnera , given that they lack conservation of differential expression between A. pisum and M. persicae. Gene Duplication and Endosymbiosis in Aphids The factors driving gene duplication in the amino acid transporters of Aphidomorpha are complex—a pattern that may also influence our understanding of gene duplications in other host genomes and their role in endosymbiosis. Our data continue to point toward the importance of nonsymbiotic traits in driving selection for paralog retention. At the same time, our data support a role for endosymbiosis in maintaining duplicated amino acid transporters in aphids. Importantly, we find that the most parsimonious explanation for conserved bacteriocyte-enrichment in orthologous pairs is that amino acid transporters were recruited to bacteriocytes in the aphid common ancestor—coinciding with acquisition of the primary, obligate endosymbiont, Buchnera . Given the repeated support we have found for both symbiotic and nonsymbiotic roles for gene duplication in sap-feeding insects ( Duncan et al. 2011 , 2014 ; Price et al. 2011 ; Dahan et al. 2015 ), we caution against drawing strong conclusions about the role endosymbiosis plays in gene duplications found in the genomes of other symbiotic systems. An understanding of the natural history of a focal taxon can indeed help with making predictions about the evolutionary significance of interesting genomic patterns like gene duplication. However, by using a comparative approach, we have found that multiple complex factors maintain paralogs in a group of genes that are functionally critical to host/symbiont interactions. Moving forward, we advocate using a comparative framework with as much information as possible about the expression and function of duplicated genes. Differential expression analysis of symbiotic and nonsymbiotic tissues using RNAseq or qRT-PCR provides a valuable layer of information that contributes to our ability to infer the role of duplicated genes and the factors contributing to their retention in a genome. These genomic and transcriptomic approaches pave the way for the next phase in understanding why genomes evolved their particular architecture through application of functional genomic approaches."
} | 5,740 |
37693856 | PMC10487350 | pmc | 9,279 | {
"abstract": "SUMMARY Bioengineering devices and systems will become a practical and versatile technology in society when sustainability issues, primarily pertaining to their efficiency, sustainability, and human-machine interaction, are fully addressed. It has become evident that technological paths should not rely on a single operation mechanism but instead on holistic methodologies that integrate different phenomena and approaches with complementary advantages. As an intriguing invention, the ferroelectret nanogenerator (FENG) has emerged with promising potential in various fields of bioengineering. Utilizing the changes in the engineered macro-scale electric dipoles to create displacement current (and vice versa), FENGs have been demonstrated to be a compelling strategy for bidirectional conversion of energy between the electrical and mechanical domains. Here we provide a comprehensive overview of the latest advancements in integrating FENGs in bioengineering systems, focusing on the applications with the most potential and the underlying current constraints.",
"introduction": "INTRODUCTION As a multidisciplinary research field that integrates the principles of biology and engineering, bioengineering is constantly advanced by the most recent discoveries and novel ideas in electrical and mechanical engineering, material science, chemistry, biology, and other fields of science and technology. The main goal of advancements in bioengineering is to enhance the standard of living for humanity through a holistic approach that yields potent instruments for preserving wellness and combating illness. Bioengineering has become an important driving force behind the rapid advancement of disease prevention and patient treatment because of the ability of electrical and mechanical devices to precisely detect physiological signals coming from patient bodies and process them appropriately. 1 In the trend of the Internet of Everything (IoE), which expands the emphasis on machine-to-machine communication in the Internet of Things (IoT) to include people and processes as well, the concept of personalized healthcare is rapidly gaining ground to improve therapeutic efficacy, conserve medical resources, and reduce costs. 2 , 3 For healthy individuals, wearable bioelectrical devices can be used to identify chemical or physical reactions brought by biological behavior and convert the phenomena into electrical signals. 4 – 8 By continuously monitoring, collecting, and analyzing signals from the human body, personalized healthcare anticipates and possibly prevents disease onset. 9 Rapid progress in artificial intelligence (AI), Big Data, and 5G/6G provides a golden opportunity for fast development of lightweight, miniaturized, flexible, and multifunctional wearable devices that enable comfortable, safe, and novel communication between humans and external devices. The level of convenience in people’s daily lives could increase because of these developments. Over the past decade, the generation of electricity by exploiting different types of human body-related energies, particularly pervasive and continuous biomechanical energy, has attracted much enthusiasm from researchers. By this effort, numerous devices have been developed that target various application scenarios and have a range of functionalities. A popular and extensively investigated route for biomechanical energy conversion and biosensing is the use of triboelectric nanogenerators (TENGs), which take advantage of triboelectrification, a commonly occurring phenomenon that has been known for about 2,600 years. 10 The concept can be described as the process of moving static charge that has been generated as the result of the contact between surfaces of dissimilar materials. 11 In addition, piezoelectric effects, which were discovered about 150 years ago, have been used for several decades in energy harvesting and biosensing. 12 , 13 Wang and Song 14 defined the name “piezoelectric nanogenerators (PENGs)” after they discovered the piezoelectricity in zinc oxide nanowires in 2006. In general, the studies on PENGs focus on using nanostructured piezoelectric materials (e.g., zinc oxide [ZnO], lead zirconate titanate [PZT], and barium titanate) that have molecular dipoles for electricity generation. 15 – 17 Ferroelectret nanogenerators (FENGs) mark an important advancement in the piezoelectricity-based conversion of energy. The origin of FENGs can be traced back to a patent filed in 1984. 18 In contrast to PENGs, FENGs are completely nonpolar unless their internal voids are charged by dielectric barrier microdischarges. 19 – 21 Internal artificial voids are usually the hallmark of FENGs, and plasma discharge of these voids makes non-polarized materials behave like piezoelectric materials, where molecular dipoles are intrinsic to the material. 22 FENGs behave similarly to traditional, well-known piezoelectric polymers or piezocomposite at the macroscopic level, but their microstructural formation and operation principles are different. Ferroelectret materials are also distinguished from conventional piezoelectric materials by their high piezoelectric coefficient and low Young’s modulus, which are derived from the charged cellular structures that make them more sensitive to stress and effective at storing charge. Because of their special physical principles, FENGs combine the merits of the high piezoelectricity observed in inorganic piezoelectric compounds (e.g., PZT) and the flexible thin-film configuration featured by organic piezoelectric polymers (e.g., polyvinylidene fluoride [PVDF]). It is worth noting that, although the appearance of FENGs and TENGs in the vertical contact-separation mode of operation 9 might be similar under some circumstances, they possess distinct differences. FENGs mimic the properties of piezoelectric materials, while TENGs operate through the mechanism of triboelectrification. In terms of performance, FENGs demonstrate positive and negative piezoelectric effects with comparable efficiency. 23 The structural composition of FENGs is characterized by foam-like configurations with numerous microscopic voids, whereas TENGs usually feature two electrodes separated by a relatively larger gap. Despite the fact that some FENGs developed recently are self-polarized, most FENGs require a high-voltage polarization process, in contrast to TENGs, which do not require polarization. To date, the FENG has been proven to be an effective, affordable, lightweight, dependable, and compatible solution for mechanical energy harvesting and self-powered sensing at a broad range of frequencies. 23 – 30 As an emerging technology for electromechanical energy conversion, FENGs have been used to produce electricity from diverse biomechanical motions such as arm movement, breathing, vocal vibration, walking, heart beating, pulse wave, blood flow, pressure from blood vessels, stomach peristalsis, etc. 31 – 33 Not only can FENGs serve as energy transformers, but they can also be used as bioinformation carriers, bridging the communication gap between people and devices, systems, and machines. Especially with the help of the quick development of IoE and AI, advances in human-machine interface (HMI) offer FENGs an advantageous application scenario in bioengineering. Demonstrations of implementing FENGs in streaming bioinformation between machine and ser include keyboards, microphones, loudspeakers, fitness trackers, etc. 23 , 34 – 36 Additionally, the expansion of implantable and wearable electronics as well as biosensors in the context of IoE is indispensable for personalized healthcare and HMI. Battery-free (self-power) and flexibility are added features that increase the practicality of the implementation of various devices in bioengineering applications. To this end, FENGs’ energy conversion capability offers a method to effectively scavenge biomechanical energy from the human body itself or the environment. Beyond biomechanical energy generation and biosensing, FENGs’ intrinsic properties enable bidirectional conversion of energy between the electrical and mechanical domains, broadening the scope of their application in bioelectronics. The giant dipoles inside the FENG are reshaped by changes in the charge density on the surface electrodes or the electric field across the thickness, which allows the FENG to act as an actuator when extra charges are transferred to its surface electrodes (or when there is a potential difference between them). One of the represented applications utilizing FENGs’ actuating behavior is production of acoustic waves with a broad range of frequencies (up to megahertz), 37 – 39 which stimulates a series of applications in bioengineering. 27 , 40 – 42 While researchers are continuously delving into theoretical analysis, inventive manufacturing methods, novel system designs, enhanced stability, etc., FENGs have already proven their broad prospects as biomedical devices for real-world clinical applications for patients in different countries. 43 – 45 FENGs are poised to unleash their potential in the field of bioengineering that complements piezoelectric technologies and, in certain cases, may even serve as a substitute for them. Because the FENG is a relatively young technology, its name has evolved along with its development, and various groups/scholars frequently make reference to it using different terms. The name “FENG” was only coined recently, but it quickly gained popularity and academic acceptance. 23 , 34 , 46 – 53 There are two possible explanations for this. The first one is purely based on the history of the term; since the name “ferroelectret” was coined by Bauer et al. 19 and later published in 2004, ferroelectret has been widely accepted to refer to this technology. Second, the name “FENG” could be interpreted as a homage to “TENG,” a well-known technology with close overall functionality. Over the last decade, FENG research has maintained strong growth momentum in line with the maturity of this technology. This review provides a retrospective of FENGs’ latest developments for advancing bioengineering from various perspectives involving human-machine interaction, biomechanical energy harvesting, personalized healthcare, and animal and plant applications ( Figure 1 ). First, FENGs’ capability to function as different types of information exchangers between humans and machines during the ongoing information revolution is summarized. Then, the advancement of electricity production from various biomechanical motions using wearable or implantable FENGs as the energy supply is discussed. The applications of FENGs for personalized healthcare concerning the human heart, respiration, blood vessels, brain, bone, pulse, and medical imaging are introduced. Furthermore, the distinctive role that FENGs play in bionic technology, invasive animal detection, and plant protection is demonstrated. Finally, conclusions and viewpoints on the current challenges and potential research directions for FENGs are offered. This review aims to provide a thorough understanding of FENGs’ most recent contributions to the development of bioengineering and to raise awareness of the various applications of this versatile technology to improve the quality of human life."
} | 2,825 |
23209573 | PMC3510206 | pmc | 9,280 | {
"abstract": "Fishing can trigger trophic cascades that alter community structure and dynamics and thus modify ecosystem attributes. We combined ecological data of sea urchin and macroalgal abundance with fishery data of spiny lobster ( Panulirus interruptus ) landings to evaluate whether: (1) patterns in the abundance and biomass among lobster (predator), sea urchins (grazer), and macroalgae (primary producer) in giant kelp forest communities indicated the presence of top-down control on urchins and macroalgae, and (2) lobster fishing triggers a trophic cascade leading to increased sea urchin densities and decreased macroalgal biomass. Eight years of data from eight rocky subtidal reefs known to support giant kelp forests near Santa Barbara, CA, USA, were analyzed in three-tiered least-squares regression models to evaluate the relationships between: (1) lobster abundance and sea urchin density, and (2) sea urchin density and macroalgal biomass. The models included reef physical structure and water depth. Results revealed a trend towards decreasing urchin density with increasing lobster abundance but little evidence that urchins control the biomass of macroalgae. Urchin density was highly correlated with habitat structure, although not water depth. To evaluate whether fishing triggered a trophic cascade we pooled data across all treatments to examine the extent to which sea urchin density and macroalgal biomass were related to the intensity of lobster fishing (as indicated by the density of traps pulled). We found that, with one exception, sea urchins remained more abundant at heavily fished sites, supporting the idea that fishing for lobsters releases top-down control on urchin grazers. Macroalgal biomass, however, was positively correlated with lobster fishing intensity, which contradicts the trophic cascade model. Collectively, our results suggest that factors other than urchin grazing play a major role in controlling macroalgal biomass in southern California kelp forests, and that lobster fishing does not always catalyze a top-down trophic cascade.",
"introduction": "Introduction Trophic cascades, in which predator-prey interactions control the composition and structure of ecological communities across two or more trophic levels in a food web have been reported in terrestrial, aquatic, and marine ecosystems [1] , [2] . In a top-down cascade, changes in the abundances of predators act to alter the abundances of grazers, which in turn affect the biomass of primary producers [3] . The degree to which predators indirectly influence primary producers depends upon biotic and abiotic conditions that vary in space and time in response to physical disturbance, the availability of resources to primary producers, and the behavior of individual consumers [4] . As such, our understanding of how and why trophic cascades vary spatially and temporally is far from complete, which limits our ability to successfully manage and protect natural ecosystems in the face of increasing threats from anthropogenic disturbances and socio-economic pressures. In coastal marine ecosystems top-down trophic cascades have been linked to the removal of top predators through fishing [5] – [12] . Frequently cited examples of marine trophic cascades come from kelp forests, in which top predators, such as sea otters [10] , [11] , fishes [6] , [7] , [12] , and lobsters [5] , [7] , [13] – [15] , are reduced in abundance by humans, leading to a relaxation in top-down control on sea urchin grazers and a decline in macroalgal abundance due to enhanced herbivory. The trophic cascade triggered by fishing in kelp forests includes a fourth trophic level occupied by humans, and depends on strong top-down interactions involving: (1) humans capturing predators of sea urchins (e.g., lobsters, fishes, and sea otters), (2) predators consuming urchins, and (3) urchins grazing macroalgae. The importance of trophic cascades as the primary determinant of community structure in kelp forest systems has been challenged because macroalgal abundance can vary greatly across space and time for many reasons other than grazing intensity [16] , [17] . Therefore, the underlying cascade involving fishing, lobsters, urchins, and macroalgae may not be ubiquitous. Weak top-down control implies that macroalgal abundance is unrelated to the abundance of urchins and their predators, and to fishing pressure on them. Nutrient availability, wave disturbance, sedimentation, and interactions among these factors are widely recognized as other drivers of macroalgal population dynamics [18] , [19] . When nutrient supply is sufficiently high, kelp production can overwhelm the capacity of grazers to control kelp abundance [20] . Populations of grazers can similarly be affected by factors other than predation and fishing, as recruitment variability [21] , [22] , disease [23] , storm disturbance [24] , and hydrodynamic conditions [25] , [26] have all been shown to influence the local abundance of sea urchins. Larger-scale processes such as El Niño-Southern Oscillation events (ENSO) can have regional effects that permeate throughout the food web by altering species abundances and the interactions among species in different trophic levels [27] . Correlative evidence for the cascading effects of fishing in marine ecosystems [5] , [28] , [29] has fueled calls for more intensive conservation, including the establishment of marine protected areas that prohibit fishing [7] , [30] , [31] . Most studies examining the effects of marine reserves have shown increased biomass and diversity in no-fishing areas compared with fished areas, which has further validated pleas for increased conservation [30] . The vast majority of this work, although highly informative, did not explore the direct effects of fishing intensity on trophic cascades, but rather assumed that spatial variation in predator and grazer abundance, and therefore predation and grazing intensity, was due to the presence versus absence of fishing [23] , [26] , [32] . Assuming that comparisons of predator density inside versus outside of reserves provide a good estimate of fishing impacts can be problematic because unprotected areas often have large differences in fishing intensity, especially for lobster [33] , [34] . In addition, inherent differences in site-specific conditions may confound reserve-based assessments because factors such as depth, exposure, and sedimentation rates may help drive differences in the distribution and abundance of lobsters, urchins, and macroalgae between reserves and nearby fished areas [26] , [35] . Finally, the process of siting marine reserves tends to select areas of relatively high biodiversity, predator densities, and habitat quality for protection [36] , which limit the ability to distinguish between the effects of fishing on community structure versus those caused by other factors. Because much of the ocean's nearshore habitats remain open to fishing, a more thorough understanding of the extent to which fishing triggers trophic cascades is warranted. Identifying the conditions that promote cascades, and determining whether or not they are ubiquitous, may usefully inform the design of marine reserve networks, especially those established to protect kelp forest communities [30] . The California spiny lobster ( Panulirus interruptus ) is the target of one of the oldest commercial fisheries in southern California. Data on commercial landings date back to the early 1900s and have averaged approximately 325 MT in recent years [37] . Spiny lobster populations are considered relatively heavily fished [38] , although a recent stock assessment estimates that both total abundance and size structure have stabilized over the last decade [39] . Nevertheless, some believe that over the past century fishing has led to a decrease in overall abundance and individual size of spiny lobsters [38] . Such decreases have the potential to diminish the role of lobsters as effective sea urchin predators. The two main objectives of our study were to: (1) examine the patterns of abundance among lobster, sea urchin ( Stronglyocentrotus spp.), and macroalgae in southern California giant kelp forests to evaluate whether they are consistent with the hypothesis that lobsters control urchins through predation, and urchins control macroalgae through grazing; and (2) determine whether the biomass of macroalgae was inversely related to the intensity of commercial lobster fishing as predicted by a top-down trophic cascade involving lobsters, sea urchins, and macroalgae. We used a correlative statistical approach to compare the abundance of organisms within three trophic levels, specifically California spiny lobsters (predator), red and purple sea urchins (grazers), and giant kelp and understory macroalgae (primary producers). As such, we did not directly test for the presence of the trophic cascade nor for the impact of fishing on the cascade, which would have required a large-scale, long-term field experiment. However, unlike most studies involving marine reserves, our analyses used sites that were explicitly selected to represent the range of natural variation in the region's kelp forests [19] , [40] , [41] , which were subjected to varying levels of fishing intensity over an eight-year period. The results from our study provide a reasonable assessment of the strengths of the trophic relationships among lobsters, urchins, and macroalgae in southern California's giant kelp forests, as well as the extent to which lobster fishing triggers a top-down trophic cascade. Exploring whether ecological paradigms operate generally across space and time is necessary to advance ecology [16] , [42] – [44] , especially when conceptual models provide the framework for innovative marine resource management, including marine reserve and other spatial-based approaches [45] .",
"discussion": "Discussion Our results suggest that a trophic cascade caused by lobster fishing, in which lobster abundance is reduced leading to increases in urchins and subsequent decreases in macroalgae, is not ubiquitous in the Santa Barbara Channel marine ecosystem. While the density of urchins varied slightly with lobster abundance (as measured by lobsters caught), non-calcareous macroalgae biomass (which included giant kelp) remained largely unrelated to red and purple sea urchin density. Thus, the observed relationship between grazer and primary producer remained inconsistent with that expected in a trophic cascade. Sea urchin grazing was clearly evident at some of our sites but it accounted for relatively little of the observed spatial and temporal variability in macroalgal biomass. Variability in macroalgal biomass has been shown to be independent of urchin grazing in other temperate reef systems as well [55] , [56] . Variability between urchin abundance and macroalgal biomass in our data was undoubtedly driven by other unmeasured factors. Reed et al. [19] concluded that physical disturbance from waves was the major factor influencing the biomass of giant kelp, the dominant macroalgal species, at the same sites used in our study. Nutrient limitation and urchin grazing also have important influences on macroalgal abundance under some circumstances, including during ENSO events when nitrogen availability is low, and under conditions of severe urchin grazing, such as those experienced in urchin-dominated “barrens” [47] . What causes the development of urchin barrens in southern California appears to be complex interactions among several factors, including urchin density (as influenced by recruitment, predation, and disease), kelp detritus production, and oceanographic conditions that influence kelp recruitment, growth, and persistence [20] , [23] , [24] . Our analyses failed to detect strong evidence for the control of urchins by lobsters. However, urchin abundance tended to decline with lobster abundance across many sites ( Figure 2 ), although the relationship was not statistically significant in our regression model. In contrast, urchin density increased across all but one site with increasing fishing intensity ( Figure 3 ). Top-down control of urchins by lobsters has been reported in studies that compared communities inside versus outside marine reserves in New Zealand [52] and the Santa Barbara Channel Islands [23] , and from patterns observed in relatively long-term ecological data collected in Maine [5] and southern California [7] . Work in Alaska [29] also indicated that sea otters can control sea urchins. Results of our Models 2 and 3 indicated that if top-down control of urchins by lobster occurred, it was probably a context dependent relationship, a phenomenon first reported by Shears et al. [26] . Specifically, three of our sites, Mohawk, Carpinteria, and Naples Reefs, have topographically complex (or rugose) rock substrata, which is excellent habitat for both urchins [26] and lobsters [56] . We found that urchin density increased by 11±4.3 individuals m −2 for every 10-cm increase in rugosity m −1 length of substrate. This rather dramatic effect of reef topography implies that predation is probably of relatively minor importance in controlling urchin abundance in habitats with many reef cracks and crevices. Our results do not include estimates of small, sub-legal lobsters, which may prey preferentially upon small sea urchins. Had we included such data, the addition of small lobsters would have increased the density of lobsters at some sites, likely reducing the negative response of urchins to lobsters. In addition, most of the sites in our study are fished for red sea urchins, which may help explain why there were fewer red than purple urchins. If urchin fishing were not occurring at our sites, the negative relationship between lobsters and urchins may have been weaker as both urchins and lobster prefer reef habitats that provide similar types of shelter. Finally, prior studies that have reported strong top-down control of urchins by lobsters also report that urchin populations often display a bi-modal size structure, with many large and small urchins and relatively few medium-sized individuals, which are preferred by spiny lobster [51] . We found relatively few small urchins at all of our sites, which is not consistent with a bimodal size structure caused by lobster predation. Thus, explanations for the negative relationship between lobster and urchins that we observed should be made with caution, in part because like many ecosystems the Santa Barbara Channel is impacted by multiple anthropogenic disturbances. Our finding that lobster fishing did not trigger a cascade that reduced macroalgal abundance reflects our observation that both urchin density and macroalgal biomass increased with lobster fishing intensity ( Figure 3A,C ). This result is consistent with previous findings that urchin grazing is not the primary factor controlling giant kelp biomass at our study sites [19] . A similar result is found in kelp forests where urchins are not important grazers [54] , such as in southern Australia where kelp production is heavily influenced by anthropogenic nitrogen inputs [55] . Increases in macroalgae with increased fishing intensity of lobster would be expected if macroalgae were primarily controlled by sea urchins. Our interviews with Santa Barbara Channel fishermen indicated they usually target kelp forests for lobster fishing, which is supported by a quantitative assessment conducted by Guenther [C. Guenther, unpublished data ] indicating that lobster catch increased with the amount of kelp surface canopy. Lobster trap fishermen also assert that they target areas with consistently high kelp cover [57] . This makes ecological sense if macroalgal biomass is predominantly greater in less disturbed areas because disturbance also negatively impacts lobster populations [53] . Overall, our results found support for the hypothesis that lobsters have top-down control on urchins through predation, a trophic interaction that has been reported previously [5] , [9] , [13] . However, we found no evidence that lobster fishing indirectly impacts macroalgal populations through increases in the abundance of sea urchins. Instead, our results support the theory that trophic-cascades are context dependent [58] , and that although humans have profound impacts on the marine environment through fishing [10] , those impacts remain heterogeneous across space and time. Our study highlights an opportunity for long-term ecological monitoring programs to incorporate fishing data where appropriate towards improved understanding of fishing's role in community ecology. Campbell et al. [59] caution ocean managers and conservationists from continuing down the traditional path of treating human behavior as external agents in ecological processes. A better understanding of site-specific processes and identification of the critical variables that make a system resilient or vulnerable to certain activities remains necessary for fostering positive progress in area-based ecosystem management. As resource agencies develop spatial ecosystem-based management we may benefit from enhanced knowledge of when and where human activities most influence ecosystem processes."
} | 4,335 |
26485611 | null | s2 | 9,282 | {
"abstract": "Microalgae have reemerged as organisms of prime biotechnological interest due to their ability to synthesize a suite of valuable chemicals. To harness the capabilities of these organisms, we need a comprehensive systems-level understanding of their metabolism, which can be fundamentally achieved through large-scale mechanistic models of metabolism. In this study, we present a revised and significantly improved genome-scale metabolic model for the widely-studied microalga, Chlamydomonas reinhardtii. The model, iCre1355, represents a major advance over previous models, both in content and predictive power. iCre1355 encompasses a broad range of metabolic functions encoded across the nuclear, chloroplast and mitochondrial genomes accounting for 1355 genes (1460 transcripts), 2394 and 1133 metabolites. We found improved performance over the previous metabolic model based on comparisons of predictive accuracy across 306 phenotypes (from 81 mutants), lipid yield analysis and growth rates derived from chemostat-grown cells (under three conditions). Measurement of macronutrient uptake revealed carbon and phosphate to be good predictors of growth rate, while nitrogen consumption appeared to be in excess. We analyzed high-resolution time series transcriptomics data using iCre1355 to uncover dynamic pathway-level changes that occur in response to nitrogen starvation and changes in light intensity. This approach enabled accurate prediction of growth rates, the cessation of growth and accumulation of triacylglycerols during nitrogen starvation, and the temporal response of different growth-associated pathways to increased light intensity. Thus, iCre1355 represents an experimentally validated genome-scale reconstruction of C. reinhardtii metabolism that should serve as a useful resource for studying the metabolic processes of this and related microalgae."
} | 467 |
36726672 | PMC9884966 | pmc | 9,284 | {
"abstract": "Introduction Wood formation is closely related to lignin biosynthesis. Cinnamoyl-CoA reductase (CCR) catalyzes the conversion of cinnamoyl-CoA to cinnamaldehydes, which is the initiation of the lignin biosynthesis pathway and a crucial point in the manipulation of associated traits. Liriodendron chinense is an economically significant timber tree. Nevertheless, the underlying mechanism of wood formation in it remains unknown; even the number of LcCCR family members in this species is unclear. Materials and Results This study aimed to perform a genome-wide identification of genes(s) involved in lignin biosynthesis in L. chinense via RT-qPCR assays and functional verification. Altogether, 13 LcCCR genes were identified that were divided into four major groups based on structural and phylogenetic features. The gene structures and motif compositions were strongly conserved between members of the same groups. Subsequently, the expression patterns analysis based on RNA-seq data indicated that LcCCR5/7/10/12/13 had high expression in the developing xylem at the stem (DXS). Furthermore, the RT-qPCR assays showed that LcCCR13 had the highest expression in the stem as compared to other tissues. Moreover, the overexpression of the LcCCR13 in transgenic tobacco plants caused an improvement in the CCR activity and lignin content, indicating that it plays a key role in lignin biosynthesis in the stems. Discussion Our research lays a foundation for deeper investigation of the lignin synthesis and uncovers the genetic basis of wood formation in L. chinense .",
"conclusion": "Conclusions In this study, 13 LcCCR genes were identified in the L. chinense genome, among which the LcCCR13 is speculated to potentially play a role in lignin synthesis in the stem as per the results of phylogenetic and bioinformatics analysis, gene expression profiling via RT-qPCR assays, and function verification via gene transformation in tobacco. In conclusion, this study lays a foundation to uncover the mechanism of wood formation in L. chinense .",
"introduction": "Introduction Lignin is an aromatic phenolic compound formed by the polymerization of three monolignols ( p -coumaryl alcohol, coniferyl alcohol, and sinapyl alcohol, also called the H, G, and S subunits) in plants. It is extremely abundant in nature and is only second to cellulose in the constituents of natural biomass ( Li et al., 2009 ). Lignin plays a pivotal role in maintaining the cell structure and imparting resistance to biotic and abiotic stresses ( Bhuiyan et al., 2009 ; Weng et al., 2010 ; Srivastava et al., 2015 ). Lignin content is also a critical factor in determining the specific applications of different types of wood. Wood with high lignin content is more rigid and usually used in furniture manufacture, while wood with low lignin content is easier to degrade and is often used in the pulp and paper industries ( Boerjan et al., 2003 ). Due to its agricultural and economic importance, the lignin biosynthesis pathway has been comprehensively investigated in the past. The phenylpropanoid pathway, an important pathway in lignin synthesis, requires the participation of a series of enzymes. Enzymes in this process are associated with 4-coumarate: CoA ligase (4CL), cinnamate-4-hydroxylase (C4H), phenylalanine ammonia-lyase (PAL), caffeic acid O -methyltransferase (COMT), hydroxycinnamoyl-CoA shikimate/quinate hydroxy-cinnamoyl transferase (HCT), cinnamate-3-hydroxylase (C3H), caffeoyl shikimate esterase (CSE), caffeoyl-CoA 3- O -methyl-transferase (CCoAOMT), cinnamoyl-CoA reductase (CCR), cinnamyl alcohol dehydrogenase (CAD), and ferulate-5-hydroxy-lase (F5H). Briefly, phenylalanine is converted to p -coumaroyl-CoA via PAL, C4H, and 4CL, where p -coumaroyl-CoA is the branching point in the flavonoid pathway ( Vogt, 2010 ). F5H and COMT convert coniferaldehyde to sinapaldehyde, and CCR triggers the reduction of feruloyl-CoA to coniferaldehyde ( Li et al., 2005 ; Leple et al., 2007 ; Vanholme et al., 2008 ; Yan et al., 2019 ). And CAD is involved in the last enzymatic step of monolignol biosynthesis by reducing the substrates, coniferaldehyde, and sinapaldehyde, to the G and S monolignols ( Yan et al., 2019 ). Moreover, it has been reported that most CAD enzyme-coding genes can influence plant growth by participating in lignin biosynthesis. In Gossypium hirsutum , Gh4CL -silencing and -overexpressing plants have a ~ 20% reduction and a ~ 10% increase in lignin content ( Sun et al., 2020 ). In Populus trichocarpa , monolignol biosynthesis is influenced by the PtrCAD1-PtrCCR2 protein complex ( Yan et al., 2019 ). The first step of the lignin reduction reaction requires cinnamyl-CoA reductase, which employs five hydroxyl cinnamic acid coenzyme CoA esters ( p -coumaryl-CoA, caffeoyl-CoA, feruryl-CoA, 5-hydroxyeruloyl-CoA, and sinapyl-CoA) as substrates to catalyze the production of cinnamaldehyde, leading to the generation of lignin ( Lewis and Yamamoto, 1990 ; Goujon et al., 2003 ). Therefore, the cinnamyl-CoA reductase encoded by the CCR gene is the point of initiation for the lignin synthesis pathway ( Lacombe et al., 1997 ; Humphreys and Chapple, 2002 ; Larsen, 2004) . The carbon flux toward lignin is regulated by the CCR -coding enzyme, thus, it is a suitable target to alter lignin levels ( Lacombe et al., 1997 ). Due to their key roles in monolignol biosynthesis, CCR s have been cloned and characterized in many plants, including the monocotyledons, Zea mays ( Pichon et al., 1998 ), Triticum aestivum ( Ma, 2007 ), Lolium perenne ( Tu et al., 2010 ), and Lilium Oriental Hybrids ( Li et al., 2009 ), as well as the dicotyledons, Caragana korshinskii ( Li et al., 2014 ), Salvia miltiorrhiza ( Wang et al., 2012 ), Arabidopsis thaliana ( Lauvergeat et al., 2001 ), Populus tomentosa ( Chao et al., 2017 ), and Betula platyphylla ( Wei, 2012 ), etc. \n CCR genes are preferentially expressed in the stems or roots of plants ( Pichon et al., 1998 ; Lauvergeat et al., 2001 ; Larsen, 2004 ), and have thus been thought to be involved in lignification. Moreover, manipulation of the CCR expression typically results in a large variation in the lignin content and composition. Plants with heavily down-regulated CCR expression often exhibit stunted growth and delayed development, along with altered carbon fluxes across lignin and other metabolic pathways ( Yin et al., 2021 ). The overexpression of BnCCR1 and BnCCR2 has been reported to increase lignin content in stems and roots of Brassica napus , which improved the lodging resistance in transgenic BnCCRox lines ( Yin et al., 2021 ). Su et al. reported that PbCCR1/2 are related to lignin biosynthesis in overexpression transgenic plants ( Su et al., 2019 ). A similar phenotype was observed in birch; Zhang et al. reported that the overexpression of BpCCR1 increases lignin content in transgenic plants ( Zhang et al., 2015 ). Giordano et al. demonstrated that the spatiotemporal expression pattern of CCR1 cDNAs from Paspalum dilatatum correlates with the developmental profile of lignin deposition ( Giordano et al., 2014 ). Compared to the wildtype (WT) plants, the lignin content in A. thaliana mutant irx4 plants is established to reduce significantly (50% of that in the WT plants), leading to abnormal growth of the irx4 plants ( Smith et al., 2017 ). Based on previous studies, regulating the expression of the CCR gene may be an effective way to change the lignin content in plants. As more plant genome resources become available, genome-wide surveys will enable systematic characterizations of key enzymes and their corresponding family members. Identification and functional analysis of lignin biosynthesis enzymes and their associated genes will lay a foundation for the systematic analysis of carbon flux through lignin metabolism ( Wang et al., 2019 ). \n Liriodendron , a tertiary relic genus, belongs to the Magnoliaceae family. At present, there are only two natural species in this genus, L. tulipifera L and L. chinense (Hemsl.) Sarg. L. tulipifera is distributed throughout eastern North America, while L. chinense is scattered in southern China and northern Vietnam ( Yang et al., 2021 ). Due to their fast growth rate, strong stress resistance, and good wood quality, Liriodendron trees are widely cultivated for use in timber, furniture, and paper-making industries. Lignin is closely related to wood quality, and CCR is a key enzyme-encoding gene in lignin synthesis ( Chao et al., 2017 ). Nevertheless, little is known about the key genes involved in lignin biosynthesis in L. chinense. Therefore, this study aims to identify LcCCR gene family members in L. chinense and analyze their potential roles in lignin synthesis. Our results revealed that the LcCCR13 gene participates in lignin synthesis and is useful for the elucidation of the mechanism of wood formation in L. chinense.",
"discussion": "Discussion Sequence characteristics and differential analysis of LcCCR s in L. chinense revealed functional differentiation Lignin synthesis starts from the common phenylpropanoid pathway and is regulated via a variety of enzymes, but the cinnamyl-CoA reductase encoded by the CCR is one of the key enzymes in lignin biosynthesis ( Lewis and Yamamoto, 1990 ; Raes et al., 2003 ). As for most forest tree species and associated metabolic pathways, genome-wide analysis is a crucial method to elucidate the biological functions of the CCR family in plants. Here, we report the phylogeny and genome structure of the CCR genes in basal angiosperms for the first time. After further characterization, 13 LcCCR genes were identified in L. chinense . The expansion of the CCR family members is important to plant evolution. After years of continuous research, studies have identified and cloned the full-length or partial coding sequence (CDS) of CCR genes from the xylem or other tissues in many plant species. For example, seven CCR genes involved in lignin biosynthesis have been identified in P. tomentosa ( Chao et al., 2017 ), 11 in Populus tremuloides ( Li et al., 2006 ), 4 in Boehmeria nivea ( Tang et al., 2019 ), and 11 in A. thaliana ( Raes et al., 2003 ). In general, there are more CCR genes in woody species. We hypothesize that the number of CCR family members in different species may be affected by genomic differences, the redundancy of gene functions, or species differences. In this study, 13 LcCCR genes were screened by proteomic mining of the complete sequence genome of L. chinense . The molecular weight of these LcCCR proteins was greater than 35 kDa each, which is the same as most CCR proteins reported in plants ( Prasad et al., 2010 ). Their isoelectric points were between 5.5 and 7.5, which is consistent with those for the BnCCR proteins ( Tang et al., 2019 ). Lauvergeat et al. examined the protein structures of the bona fide Arabidopsis CCR proteins, AtCCR5 (AT1G15950) and AtCC R 9 (AT1G80820) ( Lauvergeat et al., 2001 ), to reveal that the region of the N-terminal portion of AtCCR5/9 involved in the NADP(H) cofactor binding site was conserved. A very well-conserved motif, KNWYCY, which is thought to be involved in the catalytic site of this enzyme, also exhibits the signature of CCRs ( Lacombe et al., 1997 ). Previous studies have shown that the second and third amino acids (“W” and “Y”) are crucial for the binding of the enzyme to the substrate, and thus, they are rarely replaced ( Barakat et al., 2011 ). The conserved domain, “X-W-Y-X-X”, in the cinnamyl-CoA reductase family of proteins was found in all 13 LcCCR protein sequences here. A similar phenomenon was observed in Pyrus bretschneideri ( Chen et al., 2020b ) indicating that LcCCR s remained highly conserved during the evolution of L. chinense. \n Based on the results of MEME analysis with default parameters, the same group of LcCCR proteins shared common motifs in the phylogenetic tree, suggesting that these LcCCRs are highly conserved, strongly supporting the reliability of the group classifications ( Bailey et al., 2006 ). In addition, gene structure and motif analysis methods were employed to discover the potential features of genes ( Barakat et al., 2011 ). Phylogenetic analysis shows that LcCCR genes could be divided into four groups. The structures of exon-intron regions in the genes from the same group were similar. The change in the number of CCR genes and the diversification of characteristic motifs provides new insights to understand the evolution and gene function of the LcCCR gene family. Role of the LcCCR13 gene in the lignin biosynthesis The phenylpropanoid pathway, the main pathway in lignin biosynthesis that is regulated by a variety of enzymes, is well understood. Cinnamoyl-CoA reductase (CCR) is considered to be the first committed enzyme in the lignin-specific branch because it can catalyze the conversion of cinnamoyl-CoA to cinnamaldehyde in the monolignol biosynthetic pathway, which can further be transferred into three lignin monomers (G, S, and H) ( Gayoso et al., 2010 ). Besides, cinnamoyl-CoA can be synthesized into phenolic substances (such as anthocyanins and flavonoids) when the function of CCR has been lost. For this reason, the key role of the CCR gene in lignin synthesis has been confirmed in many plant species, and changing its expression levels might significantly affect the lignin content and growth of plants ( Liu et al., 2021 ). In our study, we measured the plant height and the lignin content in transgenic lines and observed that the overexpression of the LcCCR13 gene significantly increased lignin accumulation in the stems and decreased plant height. These results corroborate that the LcCCR13 is a key gene for lignin synthesis in the stems of L. chinense . The synthesis of lignin is tissue specific. In general, lignin synthesis occurs in tissues with higher lignification, and the CCR genes are strongly expressed in such tissues. For example, AtCCR5 is mainly concentrated in the tissues that are being lignified and participate in the process of tissue lignification ( Lauvergeat et al., 2001 ; Goujon et al., 2003 ); the activity of PtoCCR1/4/7 affects the accumulation of lignin ( Chao et al., 2017 ). The relative expression levels of the EgCCR gene are also the highest in the stems, followed by their expression in the roots and other tissues ( Lacombe et al., 1997 ). TaCCR2 is mainly expressed in the roots but also participates in the lignin synthesis in stems, indicating that it plays an important role in lignin synthesis in T. aestivum ( Lin et al., 2001 ). The expression analysis of 10 PoptrCCR s showed that all of them were expressed in the bark, leaves, and xylem, but only the bona fide PoptrCCR12/14 had the highest expression levels in the xylem, where they showed a significant difference from the expression levels in the leaf/bark ( Barakat et al., 2011 ). Our transcriptome data indicated that only LcCCR5/7/10/12/13 genes were predominantly expressed in the DXS (stem developmental xylem). RT-qPCR assays showed that the LcCCR13 had the highest expression level in the stem, which was significantly different from the expression level in other tissues. These observations were largely consistent with previous studies; For instance, the bona fide CCR genes are highly expressed in the tissues/organs with high lignification ( Zhang et al., 2015 ). In this study, only LcCCR13 was highly expressed in the stem, indicating that it might play potential roles in stem development. Therefore, we overexpressed LcCCR13 in tobacco to determine whether it was associated with lignin synthesis. Transgenic technologies have been widely used to control the lignin content and composition in various plants. Up-regulation or down-regulation of AtCCR expression in A. thaliana significantly affected the lignin content and other characteristics related to plant growth and development ( Goujon et al., 2003 ). As compared to the WT plants, the transgenic lines of B. napus over-expressing BnC.CCR2b had significantly higher lignin content in the stems ( Prasad et al., 2010 ). When BnCCR1 activity was increased in B. platyphylla , the lignin content also increased, and the height of the transgenic plants was reduced ( Zhang et al., 2015 ). The lignin content increased and the plant height decreased in the antisense transgenic tobacco plants, and their plant phenotypes (such as plant height, seed quality, and length of the leaves) was significantly altered ( Chabannes et al., 2001 ). These findings indicate that the manipulation of CCR gene expression affects lignin content, and the changes in lignin content in the transgenic plants is incompatible with normal phenotypes. In this study, we found that the over-expression of the LcCCR13 gene affects the growth and development of transgenic tobacco. As compared to other transgenic lines, transgenic plants with the highest CCR activity showed the highest lignin content, indicating that CCR might be the key gene involved in lignin biosynthesis. Hamedan et al. also reported that increased lignin content coincides with CCR activities in Gerbera jamesonii ( Hamedan et al., 2019 ). Moreover, according to the measurement of lignin content and its deposition sites in transgenic plants, we conclude that the LcCCR13 gene has a significant effect on lignin synthesis in the stem. Based on the fact that increasing the expression of LcCCR13 reduced the height growth and increased the lignin content in the stem, we assume that LcCCR13 is involved in the thickening of the cell wall by increasing the levels of all three types of subunits. It further regulates the ligin content in plants. Tu et al. found that the downregulation of CCR1 expression in transgenic L. perenne plants reduced the lignin content by lowering the levels of all three lignin monomers ( Tu et al., 2010 ). Interestingly, Zhang et al. found that BpCCR1 could control the height growth and lignin content by lignifying the cell wall ( Zhang et al., 2015 ), which was consistent with our results. All in all, these findings might be significant for understanding the roles of the LcCCR13 gene in lignin biosynthesis in the stems of L. chinense ."
} | 4,582 |
28544522 | PMC5552932 | pmc | 9,286 | {
"abstract": "Abstract The biological nitrogen cycle is driven by a plethora of reactions transforming nitrogen compounds between various redox states. Here, we investigated the metagenomic potential for nitrogen cycle of the in situ microbial community in an oligotrophic, brackish environment of the Bothnian Sea sediment. Total DNA from three sediment depths was isolated and sequenced. The characterization of the total community was performed based on 16S rRNA gene inventory using SILVA database as reference. The diversity of diagnostic functional genes coding for nitrate reductases ( napA ; narG ), nitrite:nitrate oxidoreductase ( nxrA ), nitrite reductases ( nirK ; nirS ; nrfA ), nitric oxide reductase ( nor ), nitrous oxide reductase ( nosZ ), hydrazine synthase ( hzsA ), ammonia monooxygenase ( amoA ), hydroxylamine oxidoreductase ( hao ), and nitrogenase ( nifH ) was analyzed by blastx against curated reference databases. In addition, Polymerase chain reaction ( PCR )‐based amplification was performed on the hzsA gene of anammox bacteria. Our results reveal high genomic potential for full denitrification to N 2 , but minor importance of anaerobic ammonium oxidation and dissimilatory nitrite reduction to ammonium. Genomic potential for aerobic ammonia oxidation was dominated by Thaumarchaeota . A higher diversity of anammox bacteria was detected in metagenomes than with PCR ‐based technique. The results reveal the importance of various N‐cycle driving processes and highlight the advantage of metagenomics in detection of novel microbial key players.",
"conclusion": "4 CONCLUSIONS AND OUTLOOK The results of our study indicate the importance of nitrogen cycling in the upper more oxidized Bothnian Sea sediment layers, where, based on genomic potential, full denitrification to N 2 dominated the N‐cycle driving processes. These findings corroborate previous activity‐based studies showing the dominant role of denitrification and only minor anammox or DNRA. Nitrification was dominated by aerobic ammonia oxidizing Thaumarchaeota and nitrite oxidizers belonging to the Nitrospira genus, pointing to a similar tight connection of denitrification and nitrification observed before in more eutrophic parts of the Baltic Sea (Thureborn et al., 2013 ). Unexpectedly, the peak of anammox bacterial community was detected below the oxidized layer, in a zone where no oxygen or nitrogen oxides were detectable and biogeochemistry was dominated by anaerobic methane oxidation with sulfate. Moreover, the anammox community composition seemed to be stratified between different layers with a potentially novel genus, which was not detectable by PCR with specific primers. Generally, identified microbial communities reflected well the biogeochemistry of the analyzed core with most abundant members probably being heterotrophs with multiple respiratory capabilities. Our findings also show that gene amplification‐based techniques might lead to underestimation or lack of detection of microbial key players of particular metabolic processes. This also shows that our understanding of microbial metabolic networks in coastal sediments is still far from being complete.",
"introduction": "1 INTRODUCTION Baltic Sea is a brackish basin which has been heavily impacted by eutrophication in the past decades (Jäntti & Hietanen, 2012 ). Intense agriculture and increases in populations have led to elevated input of reactive nitrogen and phosphorous via the riverine input. This eventually resulted in increased productivity, degradation, and expansion of hypoxic waters in wide areas of the Baltic Sea (Zillén, Conley, Andrén, Andrén, & Björck, 2008 ). Counteractive to the external input of reactive nitrogen are microbial processes of denitrification and anaerobic ammonium oxidation (anammox) which convert it back into N 2 and remove it from the system. They occur in both hypoxic water column and sediments, the latter is, however, of particular importance for N cycling as sediments represent hot‐spots of microbial activity due to excess availability of organic matter. The solid matrix of sediments limits the diffusion of substrates and so, through biotic and abiotic reactions, redox gradients establish and spatially separate aerobic and anaerobic metabolisms (Joye & Anderson, 2008 ). In a stark contrast, Bothnian Sea which is located in the northern part on the Baltic Sea, has been affected by eutrophication to a far lesser degree and is considered as an oligotrophic ecosystem (Lundberg, Jakobsson, & Bonsdorff, 2009 ). This was mainly attributed to physical factors such as sea bed topography, weak stratification due to low salinity, and water exchange dynamics (Lundberg et al., 2009 ). The main input of organic matter into the Bothnian Sea was calculated to be riverine and from occasional intrusions of eutrophic waters from the Baltic Proper (Algesten et al., 2006 ). Due to substantial differences in eutrophication status within the Baltic Sea basin, the reactive nitrogen balance and associated microbial processes are expected to differ too. Detailed studies have been performed on sediments of all major arms of the Baltic Sea with emphasis on denitrification, anammox, and dissimilatory nitrate reduction to ammonium (DNRA) (Deutsch, Forster, Wilhelm, Dippner, & Voss, 2010 ; Hietanen, 2007 ; Jäntti & Hietanen, 2012 ; Jäntti, Stange, Leskinen, & Hietanen, 2011 ; Stockenberg & Johnstone, 1997 ). The contribution of each of these processes to N‐oxide conversion is of particular importance as it would determine how much reactive nitrogen is being removed from the ecosystem (as N 2 gas via denitrification and anammox) or recycled into ammonium (via DNRA). Studies on highly eutrophic Gulf of Finland sediments revealed seasonal activity variability for anammox, DNRA, nitrification, and denitrification, with the latter two being closely interdependent (Deutsch et al., 2010 ; Jäntti et al., 2011 ). Denitrification was shown to be the dominant N‐oxide sink with anammox contribution under 20% (Hietanen, 2007 ; Jäntti et al., 2011 ). However, in events of persistent bottom water anoxia, nitrification is affected by low O 2 and DNRA begins to dominate over denitrification in N‐oxide reduction (Jäntti & Hietanen, 2012 ). Also in sediments of other parts of the Baltic Sea, including Baltic Proper and Gulf of Bothnia, denitrification was shown to be the dominant N‐oxide sink (Bonaglia et al., 2016 ; Deutsch et al., 2010 ; Stockenberg & Johnstone, 1997 ). The latest study on in situ N‐cycle activities in brackish sediments of the Bothnian Bay and Bothnian Sea showed that total rates of denitrification were lower than in eutrophic sediments of the southern Baltic Sea basin with anammox contributing at some sites up to 26% in N‐oxide reduction (Bonaglia et al., 2016 ). Interestingly, DNRA seemed to be of major importance at a coastal shallow, oligotrophic site in the Bothnian Bay, but was not measurable at an offshore site in the Bothnian Sea (Bonaglia et al., 2016 ). Despite extensive research on N‐cycle sediment activities, the responsible microbial communities remain largely unknown. Most molecular studies have focused separately either on the diversity of 16S rRNA or selected functional gene biomarkers (Edlund, Hårdeman, Jansson, & Sjöling, 2008 ; Falk et al., 2007 ; Glaubitz et al., 2009 ; Grote, Jost, Labrenz, Herndl, & Jürgens, 2008 ). A recent study investigated microbial community with emphasis on N‐cycle over a stratified redox gradient in the deepest and highly eutrophic part of the Baltic Sea by a metagenomics approach (Thureborn et al., 2013 ). Based on functional gene analysis, this site exhibited high potential for autotrophic sulfur‐based denitrification by ε‐Proteobacteria in possible syntrophy with ammonia‐oxidizing Thaumarchaeota (Thureborn et al., 2013 ). Similar trends have been observed at other eutrophic, anoxic sites throughout the Baltic Sea basin (Glaubitz et al., 2009 ; Grote et al., 2008 ). Most of the molecular analyses have, however, focused on eutrophic areas of the Baltic Sea and very little is known about the microbial community composition in oligotrophic sediments of the Bothnian Sea. Moreover, despite the existing activity studies, the underlying functional N‐cycle potential of these sediments remains unknown and thus comparisons to other parts of the Baltic Sea are not possible. In this study we investigated the phylogenetic composition and metagenomic potential of the in situ microbial community with respect to the biological N‐cycle in the Bothnian Sea sediment at three depths. Curated datasets of the diagnostic N‐cycle proteins (Figure 1 ) were used to estimate the abundance and diversity of the various reactions with specific emphasis on the anammox process. Figure 1 Biological N‐cycle illustrating major metabolic processes (number coded) with associated known key enzymes. Number‐coded metabolic processes: 1, nitrate reduction; 2, denitrification; 3, nitrogen fixation; 4, aerobic ammonia oxidation; 5, aerobic nitrite oxidation; 4+5, comammox; 6, dissimilatory nitrite reduction to ammonium ( DNRA ); 7, anaerobic ammonia oxidation. Nar/Nap, dissimilatory nitrate reductase; NirK/NirS, dissimilatory NO ‐forming nitrite reductase; Nor, nitric oxide reductase; Nod; nitric oxide dismutase; Nos, nitrous oxide reductase; Nif, nitrogenase; Amo, ammonia monooxygenase; Hao, hydroxylamine oxidoreductase; Nxr, nitrite:nitrate oxidoreductase; Nrf, dissimilatory ammonia‐forming nitrite reductase; Hzs, hydrazine synthase",
"discussion": "3 RESULTS AND DISCUSSION 3.1 16S rRNA gene‐based phylogenetic community composition The microbial community composition over the sediment core from the site US5B in the Bothnian Sea was analyzed with respect to 16S rRNA diversity and N‐cycle‐related diagnostic genes from three different depths: oxic/anoxic interface zone (OAZ), sulfate methane transition zone (SMTZ), and methanic zone (MZ) (Figure 3 ). Total SSU rRNA gene reads from each analyzed sediment sample comprised approximately 5%–6% of total raw reads in respective metagenomes. The majority was assigned to bacteria (80% in OAZ, 93% in SMTZ, and 80% in MZ) with archaea contributing 4% in OAZ, 5% in SMTZ, and 18% in MZ. Figure 3 16S rRNA gene distribution (normalized values) over the three analyzed depths in the Bothnian Sea sediment profile. Bacterial 16S rRNA is shown for groups with abundances over 2%, archaeal and proteobacterial over 5%. ANME , Anaerobic Methanotrophic Archaea; OAZ , oxic/anoxic interface zone; SMTZ , sulfate methane transition zone; MZ , methanic zone 3.2 Sediment in situ bacterial composition The most abundant bacterial phylum in OAZ and MZ was Proteobacteria with 55% and 32%, respectively. Other abundant phyla were comprised by Bacteroidetes (7% in both samples), Planctomycetes (5% in OAZ and 6% in MZ), and Chloroflexi (3% in OAZ and 9% in MZ). In SMTZ, Proteobacteria were as abundant as Planctomycetes (28% each), followed by Bacteroidetes (7%) and Chloroflexi (6%). The proportional distribution of the most abundant archaeal, bacterial, and proteobacterial groups in all samples is shown in Figure 3 . The distribution within the proteobacterial population varied substantially between all depth samples. Whereas the most dominant groups in OAZ were comprised by Methylococcales (14%), Campylobacterales (9%), and unclassified group Sh765B‐TzT‐29 (7%), the SMTZ was dominated by Desulfobacterales (18%), group Sh765B‐TzT‐29 (10%), group 43F‐1404R (8%), and Order Insertae Sedis/Family Insertae Sedis/Marine (7%). The MZ was dominated by group Sh765B‐TzT‐29 (13%), group 43F‐1404R (8%), Syntrophobacterales (7%), and Xanthomonadales (7%). The occurrence of Methylococcales in the upper depth corresponded to availability of both methane and oxygen in this depth which is essential for the methanotrophic lifestyle of this group. Campylobacterales , the second most abundant proteobacterial group in OAZ, comprised the genera Sulfurimonas and Sulfurovum belonging to the family Helicobacteraceae . Both Sulfurovum and Sulfurimonas spp. are commonly found in marine environments, they respire oxygen or nitrate with reduced sulfur species as electron donors (Inagaki, Takai, Kobayashi, Nealson, & Horikoshi, 2003 ; Inagaki, Takai, Nealson, & Horikoshi, 2004 ; Zhang, Zhang, Shao, & Fang, 2009 ). These findings are corroborated by profiles of nitrate and oxygen which were available only within the uppermost centimeter (cmbsf), but were not measurable below these depths (Egger et al., 2015 ; Slomp et al., 2013 ). Moreover, reduced sulfur‐fueled denitrification potentially coupled to archaeal ammonia oxidation was found to be a dominant process in the eutrophic parts of the Baltic Sea (Glaubitz et al., 2009 ; Thureborn et al., 2013 ). The coupling between biogeochemistry and community structure was also apparent in the SMTZ. Here, the dominance of Desulfobacterales corresponded to the availability of both sulfate and methane, which are removed during anaerobic methane oxidation by methanotrophic archaea (ANME) and sulfate‐reducing bacteria (SRB). Desulfobacterales are often observed as the SRB partner of ANME (Knittel, Lösekann, Boetius, Kort, & Amann, 2005 ; Michaelis et al., 2002 ; Siegert, Krüger, Teichert, Wiedicke, & Schippers, 2011 ). Interestingly, group Sh765B‐TzT‐29 falling within δ‐Proteobacteria (according to the applied SILVA taxonomy) was found to be abundant in all depths. After sequencing a single‐cell genome from one of the representatives of this group, a recent study reclassified it as Candidatus Dadabacteria which represents a novel phylum rather than a branch within δ‐Proteobacteria (Hug et al., 2016 ). The available genome information provided evidence for the use of organic carbon with complete glycolysis and TCA cycle pathways (Hug et al., 2016 ). 16S rRNA gene sequences from this group have been found in various anoxic environments and it was discussed that it might be involved in sulfur cycling (Hietanen, 2007 ; Jäntti & Hietanen, 2012 ; Siegert et al., 2011 ), nitrate reduction, or DNRA (Hug et al., 2016 ). Also a single‐cell genome from the δ‐proteobacterial group 43F‐1404R, abundant in both SMTZ and MZ, has been recently sequenced and provided more clues about its physiological potential (Hug et al., 2016 ). Members of this group seem to be heterotrophs with canonical electron transport chain, capacity for nitrate reduction, and nitrite reduction to ammonium (Hug et al., 2016 ). 16S rRNA gene sequences from this group have been detected previously in marine sediments with active sulfur cycling (Asami, Aida, & Watanabe, 2005 ), marine hydrothermal field sediments (Kato et al., 2009 ), and paddy soils (Itoh et al., 2013 ). In MZ, a relative increase in abundance of Syntrophobacterales was observed. Members of this order are closely related to SRB (McInerney et al., 2008 ). They were shown to be metabolically flexible and depending on environmental conditions and metabolic partners able to perform sulfate respiration or fermentation (Plugge, Zhang, Scholten, & Stams, 2011 ). Syntrophobacterales are often found in syntrophic partnerships with hydrogen‐consuming organisms in anoxic methanogenic environments (Lueders, Pommerenke, & Friedrich, 2004 ; Stams & Plugge, 2009 ). An abundant methanogen population would potentially act as a hydrogen sink in this depth. The Planctomycete population differed substantially within the sediment transect with a remarkable abundance of Brocadiales ‐related sequences (65% of all Planctomycete reads) in SMTZ versus 5% in OAZ and none in MZ. Members of Brocadiales are known for their capacity to oxidize ammonium with nitrite in the absence of oxygen (anammox) (Strous et al., 2006 ). All gene sequences (including functional) associated with anammox bacteria peaked in the putative SMTZ zone despite the apparent absence of available nitrite in this zone. 3.3 Archaeal in situ sediment composition The most dominant groups within the archaeal population at all depths were the thaumarchaeal Marine Group I (MG‐I), Deep Sea Hydrothermal Vent Group 6 (DSHVG‐6), Thermoplasmatales, and the ANME‐2. Group MG‐I comprised 65% of all archaeal 16S rRNA gene reads in OAZ, whereas in SMTZ and MZ its relative abundance was with 24% and 28% substantially lower, respectively. Previous studies have shown that the representatives of this group are capable of aerobic oxidation of ammonia (AOA) (Stahl & de la Torre, 2012 ). Furthermore, MG‐I seems to be dominant among archaea in various marine water and sediment habitats underlining its importance for biogeochemical element cycling (Dang, et al., 2010b ; Dang, et al., 2013a ; DeLong, 1992 ; Galand, Casamayor, Kirchman, Potvin, & Lovejoy, 2009 ). Although most gene reads belonging to MG‐I were detected in OAZ where both oxygen and nitrate co‐occurred, their relatively high abundance in deeper anoxic layers was puzzling. Similar observations were reported previously from deep oligotrophic sediment subsurface (Inagaki et al., 2006 ; Sørensen, Lauer, & Teske, 2004 ; Teske, 2006 ), estuarine and marine sediments of the South China Sea (Dang et al., 2013a ), and deep‐sea methane seep environments in the Okhotsk Sea (Dang, et al., 2010b ). The metabolic function of these deep‐sediment MG‐I group archaea remains unclear. A positive correlation between the occurrence of archaeal amoA and sediment organic matter content led to a speculation that they might possess a heterotrophic lifestyle (Dang, et al., 2010b ). Group DSHVG‐6 was second most abundant in the Bothnian Sea sediment and its relative abundance increased with depth (17% in OAZ, 27% in SMTZ, and 38% in MZ). Recent metagenomic efforts have obtained several single‐cell genomes of this group originating from anoxic aquifers, and reclassified it into separate Woesearchaeota and Pacearchaeota sister phyla (Castelle et al., 2015 ). The available genomic data point to anaerobic fermentative and/or symbiosis‐based lifestyles without a consistent metabolic signal within the phylum‐level radiations (Castelle et al., 2015 ). Interestingly, some member genomes encoded archaeal type III ribulose 1,5‐bisphosphate carboxylase/oxygenase (RuBisCO) indicating a possible function in the nucleotide salvage pathway (Castelle et al., 2015 ). 16S rRNA gene sequences of these archaea have been detected in various marine and freshwater environments. Investigations of methane seep sediments from Nankai Through have revealed similar trends in archaeal populations with groups MG‐I and Pace‐/Woesearchaeota being the most dominant and showing a decrease in MG‐I and increase in Pace‐/Woesearchaeota abundance with depth (Nunoura et al., 2012 ). \n Thermoplasmatales archaea were found in relatively constant abundance (8%–10%) at all depths over the sediment transect. The sequences clustered within five groups (SILVA phylogeny): ASC21, AMOS1A‐4113‐D04, Marine Group II, Terrestrial Miscellaneous Group (TMEG), and 2B5. Their metabolism remains unknown, but related 16S rRNA sequences have been detected in methane seeps of the North Sea (Wegener et al., 2008 ), subseafloor sediments in gas hydrate area, oxygen minimum zone in Pacific (unpublished), methanogenic estuarine sediments in Orikasa River (Kaku, Ueki, Ueki, & Watanabe, 2005 ), and deep‐sea anoxic sediments of the Okhotsk Sea (Dang, Luan, Zhao, & Li, 2009 ; Dang, et al., 2010b ). ANME archaea were only detected in SMTZ (24%) and MZ (8%) which is in correspondence to methane availability in these depths. These archaea are involved in anaerobic methane oxidation in cooperation with SRB from the order Desulfobacterales (Knittel et al., 2005 ). 3.4 N‐cycle diagnostic gene analysis The combined functional gene analysis was performed in order to assess the metabolic potential for various N‐cycle processes and to relate it to the population diversity based on 16S rRNA gene analysis at the site US5B. The overview of the proportional distribution of N‐cycle‐related functional gene reads in relation to normalized total 16S rRNA over the depth profile is shown in Figure 4 . Numbers of reads mentioned in the text refer to normalized read counts ( nrc ) (Table S2 ). Figure 4 Proportional distribution of N‐cycle related normalized gene reads in relation to total 16S rRNA normalized reads along the analyzed depth profile at site US 5B in the Bothnian Sea. cmbsf, cm below sediment surface; OAZ , oxic/anoxic interface zone; SMTZ , sulfate methane transition zone; MZ , methanic zone 3.5 Dissimilatory nitrate reduction: nitrate reductase ( narG/napA ) The nitrate‐reducing community harboring narG showed a clear stratification within the sediment transect. Most reads (55) were found in OAZ, where oxygen and nitrogen oxides were still available for respiration. Roughly half of all reads were assigned to phylum Proteobacteria with the most dominant groups belonging to Methylococcales (3.3), Desulfuromonadales (3.5), and Rhodocyclales (5.3). Other dominant narG ‐containing groups were assigned to candidate division OP3 ( Omnitrophicae , 9.9) and Haloarchaea (3.4). In SMTZ (34), the major groups containing narG were comprised by Rhodocyclales (4.7), Desulfobacterales (2.3), Deinococcales/Thermales group (1.8), candidate division OP3 ( Omnitrophicae , 3.6), and Haloarchaea (2.9). Notably, the change in biogeochemical parameters toward the absence of oxygen and dominance of sulfur cycle in this zone was accompanied by the shift in narG ‐harboring community toward sulfate‐reducing bacteria ( Desulfobacterales ) and decline in methane‐oxidizing bacteria ( Methylococcales ) and iron/iron/sulfur‐reducing bacteria ( Desulfuromonadales ). The deeper MZ layer was characterized by a stark decrease in overall narG read numbers (3) most of which were assigned to candidate division OP3 ( Omnitrophicae , 0.9) with the remaining reads being distributed among Proteobacteria . These results were congruent with all other observations and indicated the low potential and need for nitrate reduction in this depth. Conspicuous was the dominance of candidate division OP3 within the narG ‐harboring community in all depths. There are so far no cultured representatives from this group and their metabolic potential remains elusive. 16S rRNA gene information revealed that candidate division OP3 belongs to the Planctomycetes/Verrucomicrobia/Clamydiae (PVC) superphylum and it was suggested that members of this group are most likely anaerobes thriving in marine sediments, lakes, and aquifers (Glöckner et al., 2010 ; Ragon, Van Driessche, Garcia Ruiz, Moreira, & Lopez‐Garcia, 2013 ). Currently available genome information obtained from single cells and a waste water treatment plant for members of this group, provisionally named Omnitrophica , revealed the presence of genes coding for respiratory nitrate reductase, heme/copper‐type cytochrome/quinol oxidases, and nitric oxide reductases (Speth, In ‘t Zandt, Guerrero‐Cruz, Dutilh, Jetten, 2016 ). However, complete gene sets encoding the full denitrification pathway could not be detected. Another recent study found strong correlations between the occurrence of candidate division OP3 16S rRNA genes and the content of oxidized iron minerals at the discharge zone of an intertidal aquifer, speculating on the possible involvement of some members of this group in iron cycling (McAllister et al., 2015 ). The iron‐rich sediments of the Bothnian Sea might thus provide a suitable habitat for members of this group. \n napA gene reads were less abundant than narG . The distribution followed the pattern of narG with most reads detected in OAZ (30), where most abundant groups belonged to Flavobacteriia (1.9), Campylobacterales (4.1), and Alteromonadales (3.9). The distribution of napA in SMTZ (12) was more evenly spread with most abundant groups belonging to Planctomycetes (1.1), Burkholderiales (1.1), Desulfobacterales (0.9), Desulfuromonadales (0.9), and Campylobacterales (1.1). In MZ, napA was with one detected read only of minor importance consistent with the finding on narG and other N‐cycle genes. 3.6 Denitrification/anammox: Dissimilatory NO‐forming nitrite reductase (nirS;nirK) \n The distribution of both nirK and nirS gene reads followed the same trend as for other N‐cycle‐related genes with decreasing abundance with increasing depth. A large proportion (39.4) of detected nirK reads in OAZ (total of 92) was assigned to Thaumarchaeota . Next abundant groups were assigned to Methylococcales (4.1) and Rhizobiales (3.4). The most abundant nirK groups in SMTZ (total of 68) belonged to Thaumarchaeota (11.3), Actinobacteria (7), and Rhizobiales (2.7), indicating a clear community shift with the increasing depth. Interestingly, deeper in the sediment transect (MZ), the thaumarchaeal nirK seemed to increase in both relative abundance and the total nrc value (14.5) in comparison to SMTZ, despite the observation of total group MG‐I 16S rRNA gene reads being significantly lower in abundance in this depth. The available genome information of AOA has confirmed the presence of multiple gene copies of multicopper oxidoreductase‐type nitrite reductases, a trait which seems to be highly conserved among AOA (Lund, Smith, & Francis, 2012 ). This led to a conclusion that AOA might be able to perform nitrifier denitrification, and findings of marine N 2 O produced by AOA further corroborated this idea (Santoro, Buchwald, McIlvin, & Casciotti, 2011 ). Recent findings, however, point to the inability of Thaumarchaeota to enzymatic production of N 2 O via NO in the classical process of nitrifier denitrification (Kozlowski, Stieglmeier, Schleper, Klotz, & Stein, 2016 ), pointing to a different route. The physiological role of AOA NirK remains unclear, it might be involved in detoxification of nitrite or use of nitrite as the alternative electron acceptor to oxygen under hypoxia (Walker et al., 2010 ). Environmental studies have revealed wide distribution of thaumarchaeal nirK genes in marine water columns and sediments (Lund et al., 2012 ; Venter et al., 2004 ; Yakimov et al., 2011 ), also indicating distinct communities between water columns and sediments (Lund et al., 2012 ). The calculated ratios of nirK to group MG‐I 16S rRNA showed a decrease with increasing depth (3.5 in OAZ, 2.7 in SMTZ, and 0.9 in MZ). \n Proteobacteria was the most dominant bacterial phylum harboring nirS ‐like genes in OAZ (36.4), SMTZ (24), and MZ (1.4). Order Methylococcales represented with 8.1, the most abundant nirS ‐harboring group in OAZ. This finding corresponded to widespread occurrence of Methylococcales bacteria in this depth as inferred from 16S rRNA and findings of other genes of the denitrification pathway (e.g., narG , nor, nirK ). Available genome data of methane‐oxidizing bacteria confirm widespread presence of genes involved in denitrification and their contribution to N 2 O production (Stein & Klotz, 2011 ). Recent work has shown that some methanotrophic bacteria encode and express denitrification genes when exposed to nitrate and are able to link the reduction in nitrate to the oxidation of methane under hypoxia in a bioenergetically favorable manner (Kits, Klotz, & Stein, 2015 ). 3.7 Denitrification: Dissimilatory nitric oxide reductase ( nor )/nitric oxide dismutase ( nod ) \n nor ‐like genes were detected in all depths over the sediment transect with decreasing abundance with increasing depth (38 in OAZ, 29 in SMTZ, and 5 in MZ). In OAZ, no clear dominance of any taxonomic group could be observed. Here, among the most abundant groups to which the nor ‐like gene reads were assigned were Spirochaetales (2.2), Methylococcales (2.4), Desulfuromonadales (2.2), Myxococcales (2.9), Burkholderiales (2.8), and Planctomycetes (2.1). Notably, large proportion of nor‐ like reads showed nearest identity to Flavobacteriia (3.4) and strain HdN1 (2.9). Several nor ‐like reads from the Flavobacteriia order revealed highest identity to Muricauda ruestringensis and other organisms containing alternative nor ‐like genes with sequence features found in Nod proteins. Also, sequences resembling nearest identity to strain HdN1 nod sequence pointed to an abundant population of bacteria‐containing Nod‐like proteins. However, no 16S rRNA genes affiliated with either M. oxyfera or strain HdN1 could be detected pointing to novel nod ‐like gene‐harboring bacterial groups. The observation of abundant nod ‐like gene reads would be congruent with biogeochemical data indicating the presence of methane and nitrogen oxides in OAZ for N‐AOM. However, an alternative Nod‐catalyzed metabolism might be possible. Also, similar nor ‐containing phylogenetic groups were identified in SMTZ; however, the abundances differed to OAZ. In particular, the abundance of nor ‐like nrc assigned to Desulfuromonadales and Flavobacteriia increased to 4, respectively. The abundance of nod ‐like gene‐containing strain HdN1‐like population remained similar (2.2). This finding showed that despite the absence of nitrogen oxides in SMTZ, nod‐ like genes were still relatively abundant. This pointed to either a nonactive population of bacteria containing Nod proteins or an active population using Nod‐like proteins to perform an alternative metabolism. 3.8 Denitrification: Nitrous oxide reductase ( nosZ ) Our metagenome analysis revealed high abundance of nosZ‐ encoding gene reads (53 in OAZ, 40 in SMTZ, and 8 in MZ, respectively) which was approximately within the same range as the abundance of narG ‐ and nirS ‐encoding gene reads. Most dominant phylum harboring nosZ gene in OAZ was assigned to Bacteroidetes with Flavobacteriia (12), Cytophagales (3.5), and Bacteroidetes Order Insertae II Sedis (2.5) groups being the most dominant. Similar bacterial groups were dominating the SMTZ, however, the abundances changed. nosZ ‐like nrc assigned to Flavobacteriia decreased to 7.1 and those to Myxococcales increased to 1.8. In MZ, Flavobacteriia was with nrc of 2.1 the most dominant nosZ ‐harboring group. These results pointed to widespread capacity for nitrous oxide reduction in the Bothnian Sea sediment along the whole sediment transect with molecular nitrogen being the most likely product of denitrification. 3.9 Nitrogen fixation: Nitrogenase ( nifH ) Analysis for nitrogen fixation potential in the sediment transect revealed low abundance of nifH gene reads in all sediment depths (10 in OAZ, 6 in SMTZ, and 2 in MZ) as well as clear differences in microbial populations responsible for the process. The majority of detected nifH gene reads in OAZ was assigned to Methylococcales (7.3) with Methylobacter as the most dominant genus (up to 100% identity on protein level). This dominance of Methylococcales was congruent with 16S rRNA gene data. The remaining reads were mostly assigned to other γ ‐ Proteobacteria. The nifH inventory in SMTZ revealed a shift in nitrogen fixing population toward Methanomicrobia (3.1). The remaining potential nitrogen fixing population was represented by putatively SRB from δ ‐ Proteobacteria, Nitrospirae and Firmicutes . The phylogenetic affiliation of nifH reads within the sediment transect reflected the dominance of major functional microorganism groups in each particular depth observed from the 16S rRNA analysis. 3.10 Aerobic ammonium oxidation: ammonia monooxygenase ( amoA ) At our sampling site, amoA gene sequences were found in all three analyzed depths, with decreasing abundance with increasing sediment depth. The highest abundance was observed in OAZ (23), a sediment zone where oxygen and ammonium still co‐occurred, thus providing substrates for ammonia oxidizers. Taxonomic assignment revealed that the majority of amoA reads was assigned to Thaumarchaeota , which strongly pointed to their dominance in aerobic ammonia oxidation process in the Bothnian Sea sediment. All archaeal sequences fell within the Marine Group 1.1a. Reads revealed high similarity to sequences found in ecosystems ranging from fully marine over brackish to terrestrial. The closest cultured representatives were Nitrosopumilus and Nitrosoarchaeum spp. AmoA reads were also detected in SMTZ (9) and MZ (6), where electron acceptors other than sulfate or CO 2 were not available. Most sequences detected in the deeper sediment revealed high similarity to sequences found in marine, estuarine, and freshwater habitats. The analysis of metagenomes for bacterial amoA gene reads resulted in two reads in each OAZ and MZ assigned to amoA ‐like bacterial sequences, respectively. This showed that although bacterial ammonia oxidizers were not completely absent at site US5B, they were probably not of significant importance for N‐cycle transformations at this site. Based on these observations, it was evident that AOA were not restricted to sediment zones where oxygen was still present, but rather occurred in all analyzed sediment depths. This corresponded to 16S rRNA gene results of MG‐I Thaumarchaeota which were detected in all depths and decreased in abundance with increasing sediment depth. The ratio of amoA to 16S rRNA of MG‐I was approximately 2 for the upper two depths and 0.3 for MZ. Currently available genomic information of archaeal ammonia oxidizers shows amoA and 16S rRNA being single‐copy genes in sequenced genomes of thaumarchaeal ammonium oxidizers. However, several previous studies have reported amoA /16S rRNA ratios to be higher than 1 and speculated on several amoA copies in AOA genomes (Beman, Popp, & Francis, 2008 ; Lund et al., 2012 ; Santoro, Casciotti, & Francis, 2010 ), which might be the case for novel sedimentary AOA. The occurrence of group MG‐I Thaumarchaeota in anoxic sediment layers has been reported previously (Dang, et al., 2010b ; Dang et al., 2013a ; Jorgensen et al., 2012 ; Roussel et al., 2009 ; Sørensen et al., 2004 ). It has been speculated that electron acceptors other than oxygen might be used for ammonia oxidation in these organisms, or that ammonia monooxygenase might serve a different function than ammonia oxidation (Jorgensen et al., 2012 ; Mußmann et al., 2011 ). It has also been shown recently that not all Thaumarchaeota are capable of ammonia oxidation, but are metabolically more flexible and can grow with organic nitrogen substrates (Weber, Lehtovirta‐Morley, Prosser, & Gubry‐Rangin, 2015 ). 3.11 Aerobic ammonium oxidation/anammox: hydroxylamine oxidoreductase ( hao ) \n hao ‐like gene reads and its multiheme cytochrome c homologs were detected in all depths of the sediment transect; however, unlike other N‐cycle genes, more reads were detected in SMTZ than in OAZ (52 vs. 33). This higher abundance was mostly attributable to the dominance of anammox (order Brocadiales ) and sulfate‐reducing bacteria (orders Desulfobacterales and Desulfovibrionales ). Based on available genome information, anammox bacteria contain up to 10 divergent hao ‐like gene paralogs (Kartal, van Niftrik, Keltjens, Op den Camp, & Jetten, 2012 ; Strous et al., 2006 ; van de Vossenberg et al., 2013 ), at least one of which was designated as hydrazine dehydrogenase and shown to be responsible for the oxidation of hydrazine to N 2 (Maalcke et al., 2016 ). This observation was in congruence with 16S rRNA and hzsA gene data which all indicated higher abundance of anammox bacteria in SMTZ. The hao ‐like nrc assigned to anammox was 3.9 for OAZ and increased to 8.9 for SMTZ. Blast‐based analysis revealed that the majority (75%) of anammox‐like hao gene sequences originating from SMTZ were assigned to Kuenenia stuttgartiensis as the closest relative with amino acid identities ranging between 50 and 100%, the remaining reads were assigned to Scalindua spp. (14% with identities between 54 and 100%), Brocadia spp. (9%, identities between 52 and 70%), and Jettenia (2%, 58% identity). In contrast, the majority of reads assigned to anammox in OAZ resembled highest identity to Scalindua spp. (65%, accession nr. WP_034410018) with identities ranging between 54 and 100%, the rest was assigned to Kuenenia spp. (15%, identities between 47 and 65%), Brocadia spp. (7%, identities between 53 and 64%), and Jettenia (11%, identities between 76 and 78%). These results pointed to diversity differences in both depths and a presence of a potentially novel anammox genus, an observation which was supported by data derived from 16S rRNA and hzsA gene phylogeny. Potentially novel clades of anammox bacteria, deduced from amplified hzo gene sequences, have also been reported previously from sediments of the Jiaozhou Bay in China (Dang et al., 2010a ). However, as the phylogenetic resolution was limited due to short read length, the interpretation of current data should be treated with care and needs further investigation. Previous surveys revealed marine ecosystems to be dominated by anammox bacteria of the Scalindua genus and freshwater terrestrial habitats by the genera Kuenenia , Brocadia , Jettenia, and Anammoxoglobus (Galán et al., 2009 ; Humbert et al., 2009 ; Kuypers et al., 2003 ; Penton, Devol, & Tiedje, 2006 ). Molecular studies based on amplification of anammox‐specific 16S rRNA and functional genes have reinforced a hypothesis salinity being the major environmental factor shaping the community shifts between the dominance of either Scalindua or other anammox genera (Dale, Tobias, & Song, 2009 ; Hirsch, Long, & Song, 2011 ), the latter designated as “freshwater” anammox genera. However, physiological studies have shown that members of the genus Kuenenia can gradually be adopted to high salt concentrations in a bioreactor system (Kartal et al., 2006 ), thus indicating that environmental parameters other than salinity might play a role in anammox distribution. In general, the major proportion of hao ‐like gene reads in all depths was comprised by δ‐Proteobacteria: 8.9 in OAZ, 18.2 in SMTZ, and 1.4 in MZ, with Desulfobacterales , Desulfovibrionales, Syntrophobacterales, and Myxococcales being the most abundant orders. Due to short read length and thus limited sequence information, accurate function predictions of detected hao ‐like genes fragments were not possible. However, previous studies have shown several SRB, in particular within the δ‐subdivision, to possess multiheme cytochrome c proteins of the C 554 and other families (Pereira et al., 2011 ). In SRB, these proteins were speculated to be involved in respiration as in storage of electrons derived from periplasmic hydrogen oxidation (Heidelberg et al., 2004 ), enzymatic metal reduction (Lovley & Phillips, 1994 ; Lovley, Roden, Phillips, & Woodward, 1993 ; Michel, Brugna, Aubert, Bernadac, & Bruschi, 2001 ), regulation (Pereira et al., 2011 ), or detoxification (Greene, Hubert, Nemati, Jenneman, & Voordouw, 2003 ). To our knowledge, there is no evidence for bona fide hydroxylamine oxidoreductase proteins in these organisms. Interestingly, hao ‐like gene reads affiliated with AOB were detected in low abundance in all analyzed depths (between 0.5 and 1). The AOB‐like hao reads were assigned to Nitrosomonadales within β‐Proteobacteria. In addition to 16S rRNA and amoA gene data, this was another indication of AOB being of low importance in ammonium oxidation at the US5B sampling site. In contrast to the two upper depths, hao ‐like genes were of much less importance in MZ depth (6). This trend followed other N‐cycle genes, which pointed to a much lower importance of nitrogen cycling in this depth. 3.12 Nitrite oxidation/anammox: nitrite:nitrate oxidoreductase ( nxrA ) The community members performing nitrite oxidation in the Bothnian Sea were mainly dominated by anammox bacteria and representatives of the genus Nitrospira with low fractions of Nitrospina ‐related genes detected in the two upper sediment samples. The nxrA ‐like sequences were detected throughout the whole sediment transect with very similar overall abundance at OAZ (39) and SMTZ (41). Also, the phylogenetic distribution of the nxrA reads was very similar in both zones with a clear dominance of anammox (24.6 vs. 25.3) followed by Nitrospira (9.7 vs. 9.6). In contrast, the deeper methanic zone harbored significantly less nxrA ‐like sequences (4), most of which were assigned to Nitrospira (2.2) and anammox (1.4). Nitrospina ‐related nrc were at 1.5 in OAZ and 2.7 in SMTZ. Phylogenetically, Nxr is divided into two only distantly related phylogenetic groups: anammox‐ Nitrospira‐Nitrospina and Nitrobacter‐Nitrococccus‐Nitrolancea (Mogollón, Mewes, & Kasten, 2016 ). The blast analysis revealed that in all depths nitrifiers from the Nitrobacter‐Nitrococcus‐Nitrolancea group did not comprise a significant proportion of nxrA reads (4% in each sample). Previous studies on the ecology and substrate affinities of nitrite‐oxidizing bacteria (NOB) in different environments have pointed to niche differentiation between Nitrospira ‐ and Nitrobacter ‐like organisms, which was mainly attributed to available concentrations of nitrite (Bonaglia et al., 2016 ; Javanaud et al., 2011 ). Thus, natural ecosystems with limiting nitrite concentrations like the Bothnian Sea sediment would not favor NOB of the Nitrobacter/Nitrococcus type. It has been shown previously that marine environments are mainly dominated by NOB of the genus Nitrospina (Bonaglia et al., 2016 ). Their low abundance in the Bothnian Sea sediment could possibly be explained by the low salinity and brackish conditions. The high abundance of Nitrospira was not surprising as its dominance as the main functional NOB has been shown previously for various environments (Conley & Johnstone, 1995 ). This ubiquity has also partly been attributed to the metabolic flexibility of Nitrospira bacteria (Deutsch et al., 2010 ). The phylogenetic assignment of anammox nxrA revealed a diverse community with top blast hits assigned to various freshwater and marine species. The similarity on the amino acid level, however, did never exceed 85% which is in accordance with findings about other genes involved in the anammox metabolism. The Bothnian Sea sediment possibly harbors a new anammox genus. 3.13 Dissimilatory ammonia‐forming nitrite reductase ( nrfA ) \n nrfA ‐like gene reads were detected in all depths of the sediment transect, however, their abundance was considerably lower than of those involved in denitrification. Also, more nrfA ‐like gene fragments were detected in SMTZ (16) than in OAZ (12). The most abundant groups possessing nrfA in OAZ were assigned to Desulfuromonadales (2), Bacteroidetes / Chlorobi (3), and Verrucomicrobia (1.2). In SMTZ, Bacteroidetes / Chlorobi (4.1) and Desulfuromonadales (3.3) still comprised the most abundant nrfA ‐possessing groups. Also in MZ most of the nrfA ‐like reads were assigned to Bacteroidetes / Chlorobi (3.3). Thus, the nrfA ‐bearing Verrucomicrobia detected in OAZ might be adapted to higher sediment redox state and probably able to tolerate some oxygen. Previous studies investigating DNRA in estuarine environments reported its relative importance in comparison to denitrification in organic carbon‐ and sulfide‐rich sediments, speculating on inhibitory role of sulfide on denitrification (An & Gardner, 2002 ). Moreover, it has been speculated that salinity might play a crucial role for the fate of nitrate reduction pathway, where denitrification was inhibited at higher salinities while DNRA was not affected (Giblin, Weston, Banta, Tucker, & Hopkinson, 2010 ). Based on those previous observations, the combination of low‐sulfide, oligotrophic and hyposaline conditions in the Bothnian Sea sediment would likely favor denitrification over DNRA for nitrate reduction. In fact, the low overall abundance of nrfA in comparison to denitrification‐related gene reads supported this hypothesis. Reports on nrfA ‐bearing communities in sediments are scarce. So far, these have been analyzed in three estuary ecosystems exhibiting gradients in organic carbon, sulfide, and salinity parameters (Smith, Nedwell, Dong, & Osborn, 2007 ; Song, Lisa, & Tobias, 2014 ; Takeuchi, 2006 ). The majority of nrfA sequences detected in the Colne estuary, United Kingdom, were comprised by representatives of δ‐Proteobacteria most closely related to order Desulfuromonadales . Despite the very different biogeochemical properties of our sampling site in the Bothnian Sea with the hypernutrified sediment of the Colne estuary, we observed a similar trend in DNRA community toward the dominance of those particular δ‐proteobacterial groups. 3.14 Anammox: Hydrazine synthase ( hzsA ) Metagenome analysis revealed higher hzsA gene read abundance in SMTZ (2.4) than in OAZ (0.6). These results contradicted the assumption of anammox bacteria being more abundant in layers where they would have access to oxidized nitrogen oxides for respiration, and corroborated findings on other anammox‐specific genes in the analyzed core. Comparing the identity of hzsA reads between both depths, a clear difference was observed. Whereas all reads found in OAZ were assigned to Scalindua spp., all reads but one in SMTZ were assigned to Kuenenia spp. as top blast hit. Only one read in SMTZ was assigned to Scalindua . No hzsA gene reads could be identified in MZ. Also the majority of 16S rRNA reads assigned to Brocadiales from SMTZ were affiliated with Kuenenia and other freshwater anammox genera supporting the findings at the hzsA gene level. Despite low read numbers assigned to hzsA gene, these results pointed to a vertical stratification of the anammox community within the Bothnian Sea sediment transect. Findings of brackish sediments inhabited by anammox bacteria belonging to different genera including Scalindua have been reported before (Dale et al., 2009 ; Dang et al., 2010a ; Dang, et al. 2013b ). However, shifts in community structure between freshwater and marine genera have been investigated in horizontal gradients and attributed to changing environmental parameters like salinity, pH or C/N ratio (Dale et al., 2009 ; Dang et al., 2010a ). Slight changes in pH, nitrate/nitrite availability, C/N ratio, or interactions with different metabolic partners between OAZ and SMTZ might be responsible for observed vertical anammox community structure in the Bothnian Sea sediment at site US5B. Also, due to its geographical location the Bothnian Sea is strongly influenced by riverine input from mainland and occasional intrusions of saltier waters from the North Sea. This might have contributed to introduction and preservation of microbial communities from other locations including freshwater habitats of the mainland. PCR on extracted total DNA with primer pair combinations targeting either the Scalindua genus or other five known genera of anammox bacteria resulted in positive amplification only for Scalindua ‐specific hzsA gene. Positive amplification was observed for three samples between 0 and 7.5 cmbsf indicating significant presence of Scalindua ‐specific hzsA genes only in this upper layer. The absence on hzsA below 7.5 cmbsf is congruent with metagenome results for MZ where no reads could be assigned to hzsA and Brocadiales ‐specific 16S rRNA genes. However, as PCR product concentration was too low for ligation, a seminested PCR reaction was performed with Scalindua ‐specific primers for greater yield and further cloning procedure resulting in a final amplicon sequence length of 229nt. In total, 60 amplicon sequences could be retrieved: 21 for 0–2.5 cmbsf, 18 for 2.5–5 cmbsf, and 21 for 5–7.5 cmbsf. All sequences shared 97%–100% amino acid identity with uncultured Scalindua spp. originating from the marine sediments in Guyamas Basin (AGV76990) (Russ et al., 2013 ). However, due to short sequence length, the uncertainty in correct phylogenetic annotation is high and solid information can only be deduced at the genus level. Biogeochemical parameters of the sediment transect showed measurable nitrate only within the upper 0.5 cmbsf, below this depth the sediment was anoxic. Thus, it remains unclear where an abundant anammox bacterial population would derive nitrogen oxides for respiration below 2.5 cmbsf. The presence of other genes involved in aerobic processes ( amoA ) below 2.5 cmbsf might point to occasional fluxes of both oxygen and nitrogen oxides which would be rapidly consumed thus keeping concentrations below detection limits. An alternative explanation could be the presence of a dormant, not active community which was preserved during the sedimentation process. To our knowledge, deep bioturbation (below 2 cmbsf) which would introduce oxygen in deeper layers was not occurring at the US5B sampling site (Matthias Egger, personal communication)."
} | 12,295 |
24910493 | null | s2 | 9,288 | {
"abstract": "1. Classical studies of succession, largely dominated by plant community studies, focus on intrinsic drivers of change in community composition, such as interspecific competition and changes to the abiotic environment. They often do not consider extrinsic drivers of colonization, such as seasonal phenology, that can affect community change. 2. We investigated both intrinsic and extrinsic drivers of succession for dipteran communities that occupy ephemeral pools, such as those in artificial containers. By initiating communities at different times in the season and following them over time, we compared the relative importance of intrinsic (i.e., habitat age) vs. extrinsic (i.e., seasonal phenology) drivers of succession. 3. We placed water-filled artificial containers in a deciduous forest with 20 containers initiated in each of three months. Containers were sampled weekly to assess community composition. Repeated-measures mixed-effects analysis of community correspondence analysis (CA) scores enabled us to partition intrinsic and extrinsic effects on succession. Covariates of temperature and precipitation were also tested. 4. Community trajectories (as defined by CA) differed significantly with habitat age and season, indicating that both intrinsic and extrinsic effects influence succession patterns. Comparisons of AICcs showed that habitat age was more important than season for species composition. Temperature and precipitation did not explain composition changes beyond those explained by habitat age and season. 5. Quantification of relative strengths of intrinsic and extrinsic effects on succession in dipteran and other ephemeral communities enables us to disentangle processes that must be understood for predicting changes in community composition."
} | 444 |
35929783 | PMC9387530 | pmc | 9,289 | {
"abstract": "Conventional bioprocess models for wastewater treatment\nare based\non aggregated bulk biomass concentrations and do not incorporate microbial\nphysiological diversity. Such a broad aggregation of microbial functional\ngroups can fail to predict ecosystem dynamics when high levels of\nphysiological diversity exist within trophic guilds. For instance,\nfunctional diversity among nitrite-oxidizing bacteria (NOB) can obfuscate\nengineering strategies for their out-selection in activated sludge\n(AS), which is desirable to promote energy-efficient nitrogen removal.\nHere, we hypothesized that different NOB populations within AS can\nhave different physiological traits that drive process performance,\nwhich we tested by estimating biokinetic growth parameters using a\ncombination of highly replicated respirometry, genome-resolved metagenomics,\nand process modeling. A lab-scale AS reactor subjected to a selective\npressure for over 90 days experienced resilience of NOB activity.\nWe recovered three coexisting Nitrospira population\ngenomes belonging to two sublineages, which exhibited distinct growth\nstrategies and underwent a compositional shift following the selective\npressure. A trait-based process model calibrated at the NOB genus\nlevel better predicted nitrite accumulation than a conventional process\nmodel calibrated at the NOB guild level. This work demonstrates that\ntrait-based modeling can be leveraged to improve our prediction, control,\nand design of functionally diverse microbiomes driving key environmental\nbiotechnologies.",
"introduction": "Introduction Currently, full-scale wastewater treatment\nplants are typically\nengineered and designed using biological process models based on predicted\nbulk suspended biomass concentrations, and typically do not incorporate\nactual measurements of microbial community composition or metabolic\ndiversity. Models based on inferred biomass levels of broad functional\nguilds, such as the Activated Sludge Models (ASMs), 1 , 2 can\nfall short when capturing process dynamics imposed by physiological\nand/or metabolic diversity within such trophic groups. 3 , 4 Incorporating parameters of physiological diversity, such as maximum\ngrowth rates, substrate affinities, nutrient limitation, and cell\nsizes, into ecosystem-level models has proven critical to predict\nnutrient levels and cellular concentrations in environments like the\nglobal ocean. 5 , 6 In contrast to process models\nbased on homogeneous biomass, the calibration and validation of ecosystem-level\nmodels incorporating physiological diversity pose multiple challenges,\nas doing so requires direct measurements of the abundance, underlying\nmetabolic potentials, and growth strategies of diverse taxa that oftentimes\nremain uncultured. 7 Recently, gene-centric\nmodeling has been proposed as a means of\nutilizing culture-independent molecular data to connect the biokinetics\nof a particular metabolism to the measured concentration of the underlying\nfunctional gene(s), and was shown to capture nutrient profiles across\ndepth gradients within marine ecosystems 8 , 9 and pollutant\ndegradation in soils. 10 While the framework\nof gene-centric modeling is amenable to its incorporation into existing\nbiological process models by substituting bulk biomass concentrations\nwith measured levels of functional gene(s), its applicability to wastewater\ntreatment may be challenged by functional degeneracy often observed\nwithin such ecosystems. 11 − 13 Here, functional degeneracy refers\nto the case where multiple taxa share a common metabolic function\nbut also diverge with respect to other physiological traits, such\nas metabolic diversity, growth rates, substrate affinities, and/or\nstress tolerances. 14 Commonly, key ecosystem\nfunctions like nitrification, denitrification, and phosphorus uptake\ncan be performed by multiple coexisting but physiologically diverse\nmicroorganisms in wastewater treatment systems. 13 , 15 , 16 Therefore, it may benefit process models\nto parse the singular ‘biomass’ parameter into appropriate\nsubpopulations, particularly if those subpopulations have distinct\nbiokinetic growth parameters and/or ecophysiological traits that drive\nprocess dynamics. For instance, different species of acetoclastic\nmethanogens in anaerobic digesters can have different maximum growth\nrates and substrate affinities, 17 which\nhave important implications for modeling process performance and resilience. 18 The application of gene-centric process modeling\nmay therefore benefit from resolving distinct subpopulations based\non their physiological traits. The emergence of genome-resolved\nmetagenomics now enables an in-depth\nexploration of the taxonomic and metabolic diversity within trophic\nguilds driving biological processes. 16 , 19 − 21 Applied to wastewater treatment systems, genome-resolved approaches\nhave identified high levels of degeneracy in hydrolyzing and fermenting\nfunctional guilds within anaerobic digesters, 12 as well as the coexistence of multiple closely related strains of\npolyphosphate accumulating organisms within single activated sludge\n(AS) reactors. 22 Functional traits of such\nuncultured organisms may be inferred through combinations of metabolic\nreconstructions, gene expression profiles, time-series or gradient\nanalysis, and/or in situ physiology measurements. 21 − 26 Thus, genome-resolved data could be leveraged to incorporate functionally\ndegenerate populations into process models to better predict ecosystem\ndynamics by appropriately considering their physiological traits,\nparticularly when such diversity exists within trophic guilds. Among biological processes catalyzed by functionally degenerate\npopulations, nitrification plays a critical role in the global biogeochemical\nnitrogen cycle 27 and is essential for ammonium\n(NH 4 + ) and nitrogen removal in wastewater treatment\nsystems. 28 , 29 Many of the steps of nitrification can be\nperformed by physiologically diverse taxa, including chemolithoautotrophic\naerobic ammonia-oxidizing bacteria (AOB) and archaea (AOA) that oxidize\nNH 4 + to nitrite (NO 2 – ) 30 , 31 and aerobic nitrite-oxidizing bacteria (NOB)\nthat oxidize NO 2 – to nitrate (NO 3 – ). 32 The washout\nof NOB and promotion of shortcut nitrogen removal can provide up to\n25% aeration energy savings compared to conventional nitrification–denitrification\nprocesses, 33 or 60% aeration savings when\ncombined with anammox. 34 Consequently,\nmuch focus has been placed on developing operational selective pressures\nfor AS that wash out NOB by means of dissolved oxygen limitation, 35 , 36 cyclic aeration, 37 , 38 solids retention time (SRT) control, 39 or exposure of biomass to inhibitory agents\nlike free ammonia (FA) and free nitrous acid. 40 , 41 However, NOB have been shown to adapt to many of these selective\npressures, 42 − 44 likely due to their high metabolic flexibility and\nphysiological diversity, especially in Nitrospira -related genera 13 , 45 − 47 ( e.g. , Nitrospira _A, Nitrospira _C, Nitrospira _D, Nitrospira ,_E, and Nitrospira _F in the Genome Taxonomy Database 48 ). This diversity of Nitrospira spp. thus poses a hurdle for low-energy nitrogen removal from wastewater\nby obfuscating operational strategies for their out-selection. 49 , 50 Furthermore, such physiological diversity is not captured by conventional\nAS modeling frameworks. The ability to model process performance during\nshifts in physiologically diverse NOB could open doors to effective\ndesign and engineering of sustainable nitrogen removal AS systems. Here, we hypothesize that functionally degenerate Nitrospira within wastewater bioreactors can have different physiological traits.\nWe test this hypothesis by leveraging a combination of highly replicated\nrespirometry, genome-resolved metagenomics, activity-based cell sorting,\nand process modeling in an AS community that underwent a compositional\nshift in its dominant Nitrospira members during an\nimposed selective pressure. 44 Our findings\nindicate that the combination of genome-resolved metagenomic data\nwith biokinetic modeling can elucidate physiological traits of closely\nrelated organisms within wastewater treatment processes, and can capture\nprocess dynamics more accurately than gene-centric modeling. We posit\nthis approach can extend to other environments where physiological\ndiversity can drive ecosystem dynamics.",
"discussion": "Results and Discussion Nitrite Oxidation was Resilient to Out-Selection Pressure As reported previously, 44 NO 2 – accumulated in the treatment AS reactor up to\n11 mg N/L by day 37 of the treatment phase but stayed below detection\nin the control AS reactor ( Figure 1 ). At that time, the nitrite accumulation ratio (NAR)\nwas 41.9 ± 2.1% in the treatment reactor, while that in the control\nreactor stayed below 5% throughout the treatment phase ( Figure 1 ). Yet, by day 64, the effluent\nNO 2 – concentration decreased to below\n1 mg N/L in the treatment reactor, and its NAR decreased to 10.9 ±\n6.0% ( Figure 1 ). These\nobservations indicate that NO 2 – oxidation\nactivity was initially reduced by the applied out-selection pressure,\nbut the NO 2 – oxidation activity was resilient\nover time, likely due to the adaptation of the NOB community. 44 Figure 1 Nitrogen species profiles in the (A) treatment and (B)\ncontrol\nreactor throughout the duration of the study. The sampling times (day\n−30, day 31, day 61, and day 84) for metagenomic sequencing\nare indicated with dashed vertical lines. Samples for respirometry\nbatch tests were taken on days −30 and day 84. (C) Nitrite\naccumulation ratio during the press disturbance in both the experimental\nand control reactors. This plot was adapted with permission from Madill\net al. 44 under the terms of Creative Common\nAttribution 4.0 ( https://creativecommons.org/licenses/by/4.0/ ). Genome-Resolved Metagenomics Reveals a Population Shift in Nitrospira Genome-resolved metagenomics was used\nto recover a set of non-redundant population genomes, measure changes\nin microbial community structure, and assess the metabolic potential\nof community members within the two reactors. A total of 73 dereplicated\nMAGs with over 75% completion and less than 10% redundancy were recovered\nfrom the two reactors ( Figure S6, Supporting Data File 1 ). Principal component analysis showed a shift in community\ncomposition over time between the two reactors, highlighting the effect\nof the applied out-selection pressure on the AS microbiome ( Figure S7 ). Genomes of three putative NOB\nwere recovered, including two Nitrospira _A MAGs (>94%\ncompletion and <4% redundancy) and one Nitrospira _D MAG (95% completion and <2% redundancy). An HMM search of all\nrecovered dereplicated MAGs showed that these three Nitrospira MAGs contained the only detected positive hits for nxrB genes, indicative of the nitrite oxidation function ( Table S6 ). No amoA genes were\ndetected in the Nitrospira MAGs using HMM and BLASTP\nsearches, and thus they likely did not possess the comammox phenotype.\nWe confirmed the taxonomic placement of these MAGs within the Nitrospira _A and Nitrospira _D genera using\nalignments of single-copy marker genes, 16S rRNA genes, and nxrB genes ( Figure 2 ), as well as hierarchical clustering of ANI values with publicly\navailable Nitrospira genomes ( Figure S8 ). The ANI between the two Nitrospira _A MAGs was 85% ( Figure S8 ), indicating\nthat they represented two individual species at an ANI cutoff of 95%. 64 , 65 The highest ANI between the Nitrospira _A MAGs and\nother reference genomes was 79% with Nitrospira defluvii (FP929003.1) and Nitrospira sp. ND1 (FWEX0000000.1)\n( Figure S8 ), indicating that the Nitrospira _A MAGs also represent novel species. The recovered Nitrospira _D MAG shared a maximum ANI of 82% with Nitrospira lenta (OUNR00000000.1), and thus it also\nrepresents a novel species within the Nitrospira _D\ngenus ( Figure S8 ). Based on this ANI analysis,\nwe hereafter refer to the two Nitrospira _A MAGs as Candidatus ( Ca .) Nitrospira_A sp. UBC1\nand Ca . Nitrospira_A sp. UBC2, and the Nitrospira _D MAG as Ca . Nitrospira_D sp. UBC3. Figure 2 Taxonomic placement of\nrecovered Nitrospira MAGs\nwithin Nitrospira lineages by (A) concatenated alignment\nof single-copy marker genes, (B) 16S rRNA gene, (C) nxr B gene. Taxonomic genera are differentiated using colors for the\ntips and shaded regions behind labels ( Nitrospira _A in red, Nitrospira _C in blue, Nitrospira _D in yellow, Nitrospira _E in orange, and Nitrospira _F in purple). Both 16S rRNA gene and nxrB gene alignments were constructed using MUSCLE (v3.8.31), 68 and maximum likelihood trees were constructed\nusing iq-tree (v2.0-rc2) 69 with 1000 bootstraps.\nTrees were visualized using ggtree (v2.0.2). 70 The MAGs recovered in this study are shown in bold font. The relative\nabundance of Nitrospira-related organisms throughout reactor operation\nbased on (D) MAGs recovered in this study and (E) Nitrospira 16S rRNA gene ASVs recovered by Madill et al. 44 from the same DNA extracts. The values for 16S rRNA gene\nASV abundance reflect an average of three DNA extracts on each sampling\nday. Within the treatment reactor, a clear shift was\nobserved in the\nrelative abundance of the three Nitrospira MAGs during\nthe treatment phase ( Figure 2 D). In the treatment reactor, Ca . Nitrospira_A\nsp. UBC1 and Ca . Nitrospira_A sp. UBC2 increased\nin relative abundance between the start-up and treatment phases from\n0.78 and 0.39% to 3.5 and 2.4%, respectively, while that of Ca . Nitrospira_D sp. UBC3 decreased from 7.0 to 2.9% over\nthe same period ( Figure 2 D). In the control reactor, the two Nitrospira_A MAG abundances stayed\nrelatively stable below 1% relative abundance throughout the treatment\nphase, while Ca . Nitrospira_D sp. UBC3 was the dominant\nNOB and increased in abundance from 5.4 to 9.2% between the start-up\nand treatment phases ( Figure 2 D). These observed dynamics in Nitrospira MAGs closely agree with observed changes in 16S rRNA gene amplicon\nsequencing variants (ASVs) derived from the same DNA extracts from\nthese reactors, 44 in which one Nitrospira ASV (ASV_8) was washed out of the treatment reactor\nfollowing the implementation of the out-selection pressure while two\ninitially rare Nitrospira ASVs (ASV_32 and ASV_47)\nincreased throughout the treatment phase ( Figure 2 E). Madill et al. 44 also performed bioorthogonal noncanonical amino acid tagging (BONCAT)\ncombined with fluorescence-activated cell sorting (FACS) and 16S rRNA\namplicon sequencing throughout exposure of this AS community to FA,\nand found that Nitrospira ASV_8 was translationally\ninhibited by FA whereas Nitrospira ASV_32 and ASV_47\nwere unaffected. By leveraging consensus binning along with read mate\npair information and high-stringency paired-read mapping (see Text S2 ), similar to the approach used by Lesker\net al., 66 we could link each Nitrospira MAG with a corresponding Nitrospira 16S rRNA gene\nASV identified in these same reactors by Madill et al. 44 These findings indicate that our genome-resolved\nmetagenome analysis successfully recovered the predominant NOB populations\nwithin the study bioreactors and captured their compositional shift\nduring the applied out-selection pressure. These MAG-to-ASV linkages\nwere also leveraged to gain insights into the in situ physiologies of the Nitrospira populations based\non the BONCAT–FACS results reported by Madill et al. 44 Combining those reported results with the MAG-to-ASV\nlinkages found here, it can be inferred that Ca .\nNitrospira_D sp. UBC3 had reduced translational activity following\nthe side-stream 24 hr exposure to FA, while the translational activities\nof Ca. Nitrospira_A spp. UBC1 and UBC2 were unaffected\nby FA exposure. These findings collectively indicate that coexisting\nand physiologically resistant Nitrospira _A species\nconferred functional resilience to the NO 2 – oxidation process following the out-selection pressure induced by\nroutine FA exposure. Thus, despite previous reports of its short-term\nsuccess, 33 , 44 , 67 continual\nexposure of return sludge to high levels of FA may not be an effective\nstandalone strategy for achieving partial nitrification in mainstream\nactivated sludge, as we show that some Nitrospira genomic populations are physiologically resistant. In addition\nto the above-mentioned NOB populations, five MAGs of\nputative AOB were recovered and were all classified as Nitrosomonas (MAG_19, MAG_20, MAG_42, MAG_47, MAG_56) ( Figure S9 ). Similar trends of Nitrosomonas MAG relative\nabundances were observed in both reactors across the two phases ( Figure S9 ), indicating that Nitrosomonas MAGs were less affected by the out-selection pressure compared to Nitrospira -related MAGs, consistent with previous observations. 71 , 72 Metabolic Potential of Nitrospira MAGs Indicates\nNiche Partitioning of NOB The recovered Nitrospira _A and Nitrospira _D MAGs differed in their response\nto FA exposure based on their time-series abundance profiles during\nthe treatment phase ( Figure 2 D) and via linkages to previously reported changes in cellular\ntranslation with BONCAT–FACS during the applied out-selection\npressure. 44 We therefore investigated potential\nmetabolic adaptations that could confer their differential responses\nto the applied out-selection strategy. The elevated salinity\nlevels during the side-stream incubation in this study (∼2.5\ng NaCl/L) compared to the mainstream conditions could have induced\nboth osmotic and oxidative stress in the NOB. Along these lines, it\nwas recently reported that salinity impacted Nitrospira population structure in saline-alkaline lakes (1.3–12.8 g/L\nsalinity). 74 Previously, NOB have been\nreported to possess mechanisms for adaptation to such osmotic and\noxidative stresses. For instance, in Nitrobacter winogradskyi , high salinity conditions (∼10 g Na + /L) caused\nan increased expression of proteins involved in osmotic and oxidative\nstress responses. 75 Additionally, the marine Nitrospira marina can produce osmolytes, such as\ntrehalose, to tolerate osmotic stress. 45 Along these lines, three trehalose synthesis pathways were detected\nwithin the two Nitrospira _A MAGs recovered in this\nstudy, with high amino acid identities shared among many enzymes in\nthose pathways ( Text S6 ), indicating their\nsimilar potential for the production of trehalose as an osmolyte.\nIn contrast, no mechanism for trehalose synthesis was found in Ca . Nitrospira_D sp. UBC3. To tolerate oxidative stresses,\ncertain Nitrospira spp. ( e.g. , Nitrospira moscoviensis , N. sp.\nND1, N . sp. NJ1) can produce catalases and superoxide\ndismutases. 46 , 76 Superoxide dismutase catalyzes\nthe conversion of a superoxide radical of oxygen to peroxides and\nmolecular oxygen, and catalase peroxidase catalyzes the conversion\nof peroxides to molecular oxygen and water, thereby collectively reducing\noxidative stress. 77 , 78 We identified genes encoding\nfor reactive oxygen species (ROS) degradation in the Nitrospira _A MAGs, such as [Mn-Fe] superoxide dismutase, catalase, and catalase\nperoxidase ( Table S7 ). The genes encoding\nfor ROS degradation in the recovered Nitrospira _A\nMAGs shared high identities with similar genes in Nitrospira sp. ND1 (>77% AAI for both superoxide dismutase and catalase\nperoxidase\nsequences) ( Figure 3 ). In contrast, the genome of Ca . Nitrospira_D sp.\nUBC3 recovered in this study did not possess superoxide dismutase\nnor catalase peroxidase. Figure 3 Amino acid identity (AAI) of enzymes of interest\n(on x-axis) from\ndifferent nitrite-oxidizing bacterial genomes (y-axis) relative to\nreference genomes ( Nitrospira sp. ND1, N. defluvii , and Nitrospira inopinata ) obtained from GenBank (genome accession numbers in the parenthesis).\nBLASTP was used to calculate the AAI (color bar for AAI on the right\nof the heatmap). Genomes are colored according to their genus. Nitrospira _F in green, Nitrospira _D in\norange, Nitrospira _E in purple, Nitrospira _C in gray, and Nitrospira _A in blue. MAGs recovered\nin this study are shown in bold text. AMO = ammonia monooxygenase,\nHAO = hydroxylamine oxidoreductase, NXR = nitrite oxidoreductase,\nNIRA = assimilatory nitrite reductase, NIRC = nitrite transporter,\nNIRK = dissimilatory nitrite reductase, NARK = nitrite/nitrate transporter,\nFDH = formate dehydrogenase, and FOC = formate transporter. Additionally, organisms with metabolic capabilities\nfor energy\ngeneration from alternative electron sources could have gained a growth\nadvantage under the conditions used for out-selection. This could\nbe particularly true for the out-selection strategy employed here,\nas FA is thought to diffuse across cell membranes and increase cytoplasmic\npH, leading to increased cell maintenance costs for proton and potassium\nbalancing. 79 Specifically, the non-aerated\ncondition of the side-stream FA sludge treatment 44 could have created opportunities for anoxic growth on alternative\nelectron donors produced via fermentation of cell decay products,\nsuch as formate. 80 Previous studies have\nillustrated the ability of certain Nitrospira species\nto couple formate oxidation with nitrate reduction under anoxia. 46 , 47 , 81 Both of the recovered Nitrospira _A MAGs contained genes for formate metabolism,\nincluding formate dehydrogenase α, β, and γ subunits\nas well as formate transport ( Figure 3 ; Table S8 ). The amino acid\nsequences for the putative enzymes for formate oxidation and transport\nin the recovered Nitrospira _A MAGs were highly similar\n(>78% AAI) to those of Nitrospira defluvii ( Figure 3 ) that was\npreviously reported to perform mixotrophic metabolism on formate. 46 Consistent with previous reports of formate\noxidation coupled with nitrate reduction, 46 genes for nitrate reductase were also detected in Ca . Nitrospira_A sp. UBC1 (LZF62_280015) and Ca . Nitrospira_A\nsp. UBC2 (LZF60_290016). Conversely, Ca . Nitrospira_D\nsp. UBC3 did not encode any of the above mechanisms for formate metabolism\nnor nitrate reduction ( Figure 3 ). With the applied out-selection pressure established in\nthis study involving exposure to high FA levels under relatively high\nsalinity and nonaerated conditions for 24 h, these identified accessory\nmetabolic functions in the Nitrospira _A MAGs could\nhave provided additional growth strategies that contributed to their\nobserved increased abundance in the treatment reactor over time ( Figure 2 D). Along similar\nlines, the lack of such mechanisms in Ca . Nitrospira_D\nsp. UBC3 potentially contributed to its decrease in abundance within\nthe treatment reactor over time ( Figure 2 D) and its decreased translational activity\nafter FA exposure measured with BONCAT–FACS. 44 Trait-Informed Process Modeling Resolves Ecophysiologies of Nitrospira Populations As the dominant Nitrospira genera appeared to have different growth strategies\nbased on BONCAT–FACS and genome annotation, and their compositional\nshift in the treatment reactor corresponded with NO 2 – accumulation, 44 we hypothesized\nthat the different Nitrospira lineages had distinct\nbiokinetics of NO 2 – oxidation. To resolve\nsuch biokinetic growth parameters, highly replicated and high-resolution\nrespirometry data was combined with genome-resolved metagenomic abundance\ndata to calibrate a process model based on Monod biokinetics at the\ngenus-level resolution. As biomass yields were determined using\nreaction stoichiometry and measured cumulative oxygen uptake within\nthe respirometry tests, it was not possible to estimate the individual\nyields of the Nitrospira _A and Nitrospira _D MAGs separately. Rather, biomass yields were obtained for the\nbroader functional guilds of NOB and AOB and were then applied to\nall MAGs within those guilds. Experiments with isolated or enriched\ncultures of the individual species would be needed to estimate their\nrespective biomass yields. 82 The estimated\nbiomass yields for AOB ( Nitrosomonas ) of 0.13 ±\n0.05 mg COD/mg N and NOB ( Nitrospira _A and Nitrospira _D) of 0.08 ± 0.03 mg COD/mg N ( Table 1 ) were within the\nranges of previously reported values for these genera ( Tables S9 and S10 ). The obtained biomass yields\nwere then used to estimate the Monod biokinetic parameters, μ max (day –1 ) and K s (mg N/L), of the AOB and NOB community members at\ngenus-level resolution ( e.g. , Nitrospira _A and Nitrospira _D as separate NOBs) by incorporating\ngenome-resolved metagenomic abundance data into the model calibration\nusing the batch respirometry data from the start-up and treatment\nphases. Overall, the biokinetic parameters obtained for the Nitrosomonas genus and the two Nitrospira genera ( Table 1 )\nfell within the ranges reported for pure cultures or enrichments of\nthose groups ( Tables S9 and S10 ), indicating\nthat the genus-level process model was capable of resolving the behavior\nof these different populations within a mixed bioreactor community.\nThe predicted OUR values from the calibrated genus-level process model\nwere strongly correlated with observed OURs within the respirometry\nvials. The mean Spearman correlation was 0.85 ± 0.11 and 0.8\n± 0.07 between the observed and predicted OURs in the respirometry\nbatch tests with NH 4 + -N and NO 2 – -N as substrates, respectively ( Figure 4 ). The predicted OURs of the calibrated model\nwere, on average, 0.96 ± 0.05 and 0.86 ± 0.11 times that\nof the observed OURs in the NH 4 + -N and NO 2 – -N respirometry tests, respectively (mean\nR 2 values of 0.94 ± 0.01 and 0.86 ± 0.06 for\nNH 4 + -N and NO 2 – -N as substrates, respectively) ( Figure 4 ). In contrast to nitrifier biokinetic parameter\nestimation using unreplicated extant respirometry, 61 , 83 the highly replicated respirometry conducted here allowed for a\nglobal model fit spanning an order of magnitude of initial nitrogenous\nsubstrate concentrations in triplicate. Figure 4 Genus-level process model\nfit evaluation based on predicted and\nmeasured OUR values in the triplicate respirometry vials with (A)\nNH 4 + -N or (B) NO 2 – -N as the substrate for both the treatment and control reactors in\nboth operating phases. The linear regression equation indicating the\nslope and R 2 are shown on each faceted plot for both reactors\nduring both experimental phases. The best fit line is shown as a solid\nblack line. The Spearman correlation coefficients ( R ) between measured and predicted OUR values from the model for each\nbatch test are also provided. The OUR values are data points colored\nby the initial substrate concentration in the respirometry batch tests\nwith biomass from the corresponding reactors. Table 1 Biokinetic Parameters of Nitrifying\nGenera Recovered from the Study via Integration of Genome-Resolved\nMetagenomic Abundance Data during the Calibration of the Genus-Level\nProcess Model Parameter Nitrosomonas Nitrospira _A Nitrospira _D Biomass yield (g COD/g N) 0.13 ± (5 × 10 –2 ) 0.08 ± (3 × 10 –2 ) 0.08 ± (3 × 10 –2 ) μ max (day -1 ) 0.69 ± (3 × 10 –3 ) a 3.62 ± (6 × 10 –2 ) 0.05 ± (2 × 10 –3 ) 0.25 ± (2 × 10 –3 ) K s (mg N/L) 0.43 ± (7 × 10 –3 ) a 1.21 ± (4 × 10 –2 ) 0.1 ± (3 × 10 –2 ) 0.65 ± (1 × 10 –2 ) a Values correspond to the control\nreactor during the treatment phase (see Text S3 ). The calibrated process model makes predictions about\npopulation\ngrowth rates and substrate utilization, which was validated using\nmetagenomic sequence data and nitrogenous compound concentrations\nsampled in situ within the SBRs on two operational\ndays (days −30 and 84). Within the treatment reactor on day\n84, the genus-level process model predicted an accumulation of nitrite\nthroughout the SBR aerobic period, which was consistent with measured in situ values ( Figures 5 A; S10 ). For comparison,\na process model was also calibrated at the broader resolution of functional\nguilds by aggregating the biomass concentrations of all AOB and NOB\nMAGs, similar to gene-centric modeling 8 , 9 and conventional\nAS modeling 1 (see Text S4; Table S11; Figure S11 ). This ‘guild-level’\nmodel did not predict the observed nitrite accumulation within the\ntreatment reactor on day 84 ( Figures 5 B and S10 ). Similarly, long-term\nsimulations of reactor performance over the two experimental phases\n( Text S5 ) showed that the trait-based genus-level\nmodel effectively predicted the observed compositional shift in NOB\npopulations along with the rise and eventual decline of nitrite accumulation\nduring the treatment phase ( Figure S12 ),\nwhereas the guild-level model did not ( Figure S13 ). Figure 5 Comparison of observed nutrient dynamics and simulated\nvalues during\nthe aerobic period of the treatment reactor during day 84 of the treatment\nphase using (A) the calibrated model aggregated at the genus level\nand (B) aggregated at the functional guild level ( i.e. , “gene-centric”). Nutrient concentrations observed\nduring the 150 min aerobic period (after reactor feeding) are plotted\nas points. Simulated nutrient concentrations are indicated by the\nsolid lines. The shaded area represents 25 and 75% quantiles for the\npredicted values using Monte Carlo simulations with 1000 iterations. Conventional AS and gene-centric models would group\nthe two Nitrospira genera ( Nitrospira _A and Nitrospira _D) as one entity due to their\nhighly similar\nNXRB enzymes (>95% AAI) and shared functional guild of NOB ( Figure 3 ). However, based\non in situ physiology measurements ( e.g. , BONCAT–FACS) and genome annotations ( Figure 3 ), it is apparent that these Nitrospira lineages are likely functionally degenerate with distinct phenotypic\ntraits. We therefore posit that these NOB populations should be decoupled\nin a process model, as they had different growth strategies that manifested\ninto the observed changes in the NO 2 – oxidation capacity during the applied out-selection pressure. This\nfinding that trait-informed process modeling could predict nitrogen\ndynamics more accurately compared to guild-level modeling highlights\nthe utility of high-resolution genomic data for examining the impact\nof physiological diversity within shared metabolic niches. Here, incorporating\nmetagenomic data at the genus level into a process modeling framework\nprovided a basis for predicting system dynamics ( Figures 5 A, S12, and S14 ) during population shifts that were not captured by\nmodeling based on broader functional guilds ( Figures 5 B, S13, and S15 ). Obtaining total cellular biomass concentrations as VSS or\nCOD is\nessential to maintain stoichiometry in bioprocess models, 84 and inferring these values for functional guilds\nand/or subpopulations using molecular biology approaches represents\na current knowledge gap. Cellular biomass concentrations in AS processes\nhave been previously estimated using flow cytometry and fluorescence in situ hybridization. 84 − 86 However, these approaches\nare also limited by technical errors from microbial aggregation during\ndirect cell counting, as well as assumptions for the dry weight of\ncells and average biovolume of target microbes. 86 For instance, we attempted to determine biomass concentrations\nusing estimated cell biovolumes obtained with epifluorescence microscopy\n( Text S7 ). The biomass concentrations predicted\nby biovolume estimates were unreasonably low compared to steady-state\nmodeling predictions 44 ( Table S12 ), and the genus-level model calibrated with these\nbiomass concentrations ( Table S13 ) had\na poor fit to the respirometry data ( Figure S16 ). Since active biomass is a matrix of various substances, including\ncells and extracellular polymeric substances like enzymes, the epifluorescence\nmicroscopy may have underestimated active biomass by only measuring\ncellular biovolume. Here, we estimated the biomass of microbiome members\nby multiplying their relative abundance from metagenomics by the measurements\nof total VSS in the bioreactor. Active biomass concentrations of functional\nguilds may not be best represented by bulk VSS measurements in all\nsystems; however, in this current application, it was estimated that\nthe active biomass comprised roughly 70% of the total VSS in the AS\nsystem based on steady-state modeling. 44 Alternatively, it could also be possible to measure the relative\nabundance of microbiome members using metaproteomics 87 and to convert protein measurements into estimates of biomass\nconcentration. However, such metaproteomic approaches can be challenged\nwhen multiple closely related organisms are present, such was the\ncase in this study AS system, due to nonunique and/or ambiguous peptide\nmatches. 87 It is also important to note\nthat the DO levels were maintained above 3 mg/L throughout the aerobic\nphase, and thus we do not anticipate impacts of DO diffusion limitation\non the measured biokinetic parameters. 88 However, in systems with lower DO and/or more substantial aggregation\nof biomass ( e.g. , biofilms or granular sludge), diffusion\nlimitations and cellular densities can impact apparent substrate utilization\nkinetics. 88 In those applications, this\ntrait-based modeling approach could be augmented to account for diffusion\neffects using strategies to obtain genome-resolved spatial abundance\ndata. 21 , 89 Thus, this study provides a proof-of-concept\nfor incorporating biomass measurements of functional populations into\nbioprocess modeling, which could be further refined by progressing\nnew methods to accurately measure the active community fraction as\nwell as taxon-resolved biomass measurements and their spatial organization. Implications for Microbial Community Management, Control, and\nPrediction The integration of large microbial molecular datasets\nwith process rate measurements remains a hurdle that obfuscates linkages\nbetween microbial community composition and function with ecosystem\ndynamics in many environments. 23 Resolving\nmicrobial growth strategies and physiological traits and incorporating\nthat information into ecosystem models has been proposed as a solution\nto this grand challenge. 23 While demonstrated\napproaches to establish mechanistic linkages between measurements\nof microbial physiology, molecular datasets, and ecosystem dynamics\nfor mixed communities remain sparse, 8 − 10 the approach is gaining\npopularity. 23 Our genome-resolved\nmetabolic reconstructions indicated that the growth strategies of Nitrospira _A and Nitrospira _D were distinct\nin terms of resource acquisition ( e.g. , metabolic\nflexibility) and stress tolerance ( e.g. , osmolyte\nproduction and ROS degradation). We also leveraged measurements of\nphysiology by establishing linkages to BONCAT–FACS 16S rRNA\namplicon sequencing data that further supported the sensitivity of Nitrospira _D and the resistance of Nitrospira _A to the applied selective pressure. Further, the combination of\nrespirometry, process modeling, and metagenomic sequencing showed\nthat the biokinetic growth parameters of Nitrospira _A and Nitrospira _D were divergent, especially in\nregards to μ max . The slower μ max of Nitrospira _A MAGs (0.05 day –1 ) compared to Nitrospira _D (0.25 day –1 ) helps to explain the observed dominance of Nitrospira _D within the control reactor that did not have the selective pressure\napplied. We propose that the decrease in the relative abundance of Nitrospira _D and replacement by Nitrospira _A during the applied out-selection pressure in the treatment reactor\nwas attributed to the lack of mechanisms to abate osmotic and salinity\nstresses in Nitrospira _D, leading to their more rapid\ndecay, as well as potential growth on formate and nitrate by Nitrospira _A during the unaerated sludge treatment. The\nslower μ max of Nitrospira _A in comparison\nto Nitrospira _D in this study suggests that a combination\nof routine FA exposure with strict SRT control could be a successful\nstrategy to effectively wash out NOB for partial nitrification and\nenergy-efficient nitrogen removal. This hypothesis was supported by\nsimulating reactor dynamics using the calibrated genus-level model,\nwhich indicated that a 7-day SRT in the treatment phase would wash\nout both Nitrospira _A and Nitrospira _D due to the combinatorial selective pressures of FA exposure and\na short SRT, while also retaining the AOB necessary for nitritation\n( Figure S17 ). Therefore, the observed higher\nsensitivity to FA in Nitrospira _D and the lower maximum\ngrowth rates of Nitrospira _A are physiological traits\nthat could be exploited for their out-selection. This study therefore\ndemonstrates the utility in considering the ecological traits of genomic\npopulations when constructing and calibrating process models, which\ncan then be utilized to inform effective microbiome management and\nengineering strategies. Collectively, our findings highlight\nthe utility and importance\nof measuring and inferring physiological characteristics of populations\nso that they can be grouped based on shared traits within process\nmodels. This may be particularly true for wastewater treatment systems\nthat can harbor a large number of individual genomic populations, 90 as modeling each individual genome separately\ncould lead to issues with model overfitting and parameter identifiability. 91 Thus, a potential limitation of implementing\nthis trait-based modeling approach is that it requires experimentation\nto measure the genomic potential, biokinetics, and physiological characteristics\nof target populations, which could hinder its rapid uptake and implementation.\nHowever, we envision that the formulation of shared databases that\nlink microbial identity, genomic features, physiological traits, and\nbiokinetics could facilitate faster and cheaper implementation of\ntrait-based process modeling, particularly if marker genes can be\nlinked to such a database. Such efforts are currently underway, such\nas the MiDAS global wastewater 16S rRNA gene database, 92 along with studies probing the in situ physiology of high-quality MAGs within those systems. 21 Therefore, trait-based wastewater process modeling\nmay soon be more widely attainable via such collective efforts to\nmap the wastewater microbiome. This trait-based process model\nframework can be extended to model\na variety of environmental systems to describe biochemical processes\nand predict community composition and nutrient dynamics, given there\nare known or calibrated growth parameters for the community members\nand genome-resolved biomass concentration data. Such a modeling framework\ncould be advantageous in environmental biotechnology to design or\npredict processes that have narrow windows for operational success\ndue to metabolic or growth constraints, such as stable partial nitrification\nin AS. Another use-case could be to predict a taxon’s contribution\nto pollutant degradation to identify strategies for microbiome management\nand/or bioaugmentation. Therefore, this approach is a step toward\nintegrating growth strategies with microbiome engineering for desired\necological outcomes, particularly in environments with high functional\ndegeneracy like wastewater treatment systems."
} | 9,790 |
33870124 | PMC8041867 | pmc | 9,290 | {
"abstract": "Summary This article proposed a high-performance triboelectric-electromagnetic hybrid wind energy harvester (WEH). By adopting the revolution and rotation movements of tapered rollers, which serve as both the rotor of the electromagnetic generator (EMG) part and freestanding layers of the triboelectric nanogenerator (TENG) part, the WEH can work as a sustainable power source and a self-powered wind speed sensor. When the wind speed is 12 m/s, super-high open-circuit voltage peaks of 47.4 and 683 V can be achieved by the EMG and TENG, respectively, corresponding to the high-power outputs of 62 and 1.8 mW. It was demonstrated that the WEH can easily light up over 600 red light-emitting diodes and even a 5-W globe light. A self-powered wireless temperature and humidity sensing network was also systematically demonstrated. In summary, the proposed WEH exhibits bright future toward IoT applications, such as in border detection, smart buildings, and so on.",
"conclusion": "Conclusion In this article, a novel triboelectric-electromagnetic hybrid WEH based on the revolution and rotation movement of tapered rollers was proposed. Fundamentally, both theoretical model and simulation analysis were carried out, followed by a series of simulated analysis and contrast experiments to optimize the output performance of the EMG and TENG. At the wind speed of 12 m/s, the EMG and TENG of the optimized hybrid WEH can generate the maximum open-circuit voltages of 47.4 and 683 V, respectively. Under the same wind speed, the output power of the EMG reaches the climax of 62 mW at the external load of 660 Ω, which corresponds to a volume output power density of 72.1 W/m 3 , whereas the maximum output power of the TENG about 1.8 mW is achieved at the external load of 60 MΩ, which corresponds to a volume output power density of 0.27 W/m 2 . A supercapacitor of 1.5 F can be charged to 3 V in 10 min by the EMG, and the TENG can run for ∼35,000 cycles continuously without obvious decay. The WEH is demonstrated to power up arrays of LEDs and even a 5-W globe light, which indicates potential talent to be used as a supplementary power source for field lightening. The TENG can also be used as a wind speed sensor by analyzing the output frequency characteristics. Last but not least, combined with corresponding power management circuit, wireless sensor node, and terminal monitoring interface, the device can easily power a wireless environmental monitoring system, such as a wireless temperature and humidity sensing node and a self-powered anemograph, which shows bright application prospects in smart buildings, intelligent agriculture, border detection, and so on.",
"introduction": "Introduction With the wide spread of Internet of Things (IoTs) and further development of embedded technology, the application of wireless sensor networks is becoming more and more popular in recent years ( Li et al., 2016 ; Chen et al., 2018 , 2019 ; Ryu et al., 2019 ). Numerous sensor nodes are conventionally powered by batteries, which intrinsically suffer from the problem of limited lifetime ( Chen et al., 2015b ). Considering the challenges of worldwide energy crisis and related environmental issue ( Chu and Majumdar, 2012 ; Armstrong et al., 2016 ), power supply turns to be a bottleneck problem for the long-term application of IoTs. To crack this hard nut, remarkable efforts have been made over the last decades. Harvesting renewable energy, including but not limited to solar ( Oh et al., 2012 ), wind ( El-Askary et al., 2015 ; Wen et al., 2016 ; Kosunalp, 2017 ; Orrego et al., 2017 ), rain drop ( Zheng et al., 2014 ), and blue energy ( Li et al., 2019 ; Chen et al., 2020 ) from the ambient environment, and then converting it into electricity to provide power supply to IoT applications, is of utmost significance to the sustainable development of modern society. Wind energy is widely known as a kind of renewable energy, which is ubiquitous in the wildness ( Cheng and Zhu, 2014 ; Wang et al., 2016 ; Lai et al., 2019 ), and wind energy harvester (WEH) is considered as an ideal solution to sustainable power supply to smart sensors in IoT applications ( Bethi et al., 2019 ; Wang et al., 2020 ). Generally speaking, the existing WEHs can be divided into two types: vibrational and rotational. The vibrational WEHs convert wind energy into the vibration energy of an elastic structure by utilizing the effects of fluttering ( Li and Lipson, 2009 ; Zhao et al., 2016 ; Liu et al., 2018 ; Lin et al., 2019 ), galloping ( Fei et al., 2012 ; Zhou et al., 2018 ), vortex shedding ( Weinstein et al., 2012 ; Chen et al., 2016 ), and resonance ( Wang et al., 2014 ; Zhang et al., 2015 ). The rotational WEHs convert wind energy into the rotation energy of rotators based on the piezoelectric ( Fu and Yeatman, 2015a , 2015b ; Zhang et al., 2017 ), triboelectric ( Chen et al., 2015c ; Wen et al., 2015 ; Ahmed et al., 2017 ; Kim et al., 2018 ), and electromagnetic ( Weimer et al., 2006 ; Liu et al., 2019 ) effects, independently or in combination ( Zhong et al., 2015 ; Chen et al., 2015a ; Guo et al., 2016 ; Cao et al., 2017 ; Yang et al., 2018 , 2019 ; Hao et al., 2019 ). Compared with the vibrational WEHs, the rotational WEHs are inherent with the characters of higher power output, longer working life, and more stable performance. From the aspect of energy conversion effects, triboelectric nanogenerators (TENGs) ( Wang et al., 2012 ; Lin et al., 2013 , 2014 ; Li et al., 2015 ) show great advantages of high voltage, small size, light weight, and easy fabrication ( Wang, 2013 ; Wang et al., 2015 ; Liu et al., 2017 ; Cheng et al., 2019 ), whereas electromagnetic generators (EMGs) ( Spreemann et al., 2006 ; A. Bansal et al., 2009 ; Halim et al., 2018 ) hold superior in high currents, low cost, compact structure, and excellent stability ( Zhao et al., 2019b ; Hou et al., 2019 ). Hence, it is rational to combine TENG with EMG to draw on each other's strengths. Numerous researches show the hybridization road of energy harvesting from single effect to multi-effects. After a wind cup-driven rotational TENG was proposed by Xie et al. (2013 ), Zhao et al. proposed a triboelectric-electromagnetic hybrid nanogenerator ( Zhao et al., 2019a ). Based on the hybridization of triboelectric and electromagnetic effects, a well-packaged WEH was proposed by Fan et al. (2020 ) to power a commercial hygrothermograph or a wireless environmental monitoring system even in harsh environment. Zhang et al. demonstrated a rotating-disk-based hybridized nanogenerator to effectively harvest energy from wind generated by a moving vehicle through the tunnel ( Zhang et al., 2016 ), which can be utilized as a self-powered wireless sensor for traffic volume monitoring in the remote mountain area. Cao et al. reported a rotating-sleeve triboelectric-electromagnetic hybrid nanogenerator ( Cao et al., 2017 ), which was capable of lighting dozens of light-emitting diodes (LEDs) as well as powering an electronic watch under blowing wind. He et al. introduced a rotary cylinder-based nanogenerator, by hybridizing a TENG and an EMG ( He et al., 2019 ), which can operate as a self-powered counter and timer for potential speed detecting. Guo et al. designed a pinwheel-like nanogenerator by coupling triboelectric and electromagnetic effects ( Guo et al., 2019 ), which not only can charge capacitors but also can power many portable electronic devices even in a rather low wind speed. As energy harvesting technology is booming for decades, although some WEHs have the potential ability to convert the wind energy into electricity for powering the self-powered wireless sensor nodes in the wild, there is still a long way to go to enhance the output performance. Still on this purpose, a high-performance triboelectric-electromagnetic hybrid WEH for wind energy harvesting is proposed, which can be used as not only a high-output power source but also a self-powered anemometer. The hybrid WEH consists of an EMG part and a TENG part. Driven by the rotating wind cup, the nylon tapered rollers together with magnets begin to rotate, which not only induces an electric current in the surrounding coils but also enables the bottom fluorinated ethylene propylene (FEP) layer to generate electricity between electrodes.",
"discussion": "Results and discussion Device structure The device structure of the proposed WEH is schematically illustrated in Figure 1 A, which consists of a horizontal conical TENG part and a vertical cylindrical EMG part. Both the TENG and EMG are composed of stators and rotators. For the stator of the TENG, a copper foil was first cut into interdigital electrode pattern and then adhered to the conical surface of the acrylic (PMMA) substrate. A layer of FEP film was finally laminated on the electrodes to work as one triboelectric material. Four hollow tapered nylon rollers as the other triboelectric material were used as the rotator of the TENG. For the EMG, four penny-shaped polarized magnets were fixed inside the bottom hole of the hollow tapered rollers serving as the rotator. Aligned with the magnets, eight synclastic twined copper coils were uniformly embedded on the octagonal side wall of PMMA frame serving as the stator of the EMG. A wind cup was connected with a PMMA top cover through a shaft to convert wind flow into the revolution and rotation movement of the four tapered rollers. In order to enhance the movement of tapered rollers, a ball bearing was fixed at the end of the shaft. Once the relative movement between the rotators and stators occurs, both charge transfer of the TENG and flux change of the EMG will produce electrical output periodically. Figure 1 B shows the photographs of (1) the stator part, (2) the rotator part, and (3) the assembled WEH. Figure 1 Device structure of the WEH (A) Structural schematic of the WEH. (B) Photographs of the (i) rotator part, (ii) the stator part, and (iii) the assembled WEH. In order to further illustrate the advantages of tapered roller structure, the comparisons of potential distribution and equivalent stress between tapered, cylindrical, and spherical rollers are shown in Figures S1 and S2 , respectively. When the external conditions are consistent, the tapered roller shows better potential distribution and mechanical properties than the other two. In addition, the diameter of the inner and outer rings of the tapered roller can be adjusted to ensure only pure rolling of the roller with the base, as shown in Figure S3 . However, the contact between the cylindrical roller and the base is similar to that of the spherical roller, which could induce sliding friction with the base and affect the rotation of the device. Therefore, the tapered roller structure enables the device to work more efficiently. Working principle and simulation As shown in Figure 2 A, the working mechanism of the EMG is based on the fundamental principle of Faraday's law of electromagnetic induction ( Zhang et al., 2014 ). Briefly speaking, as long as the magnetic flux passing through the closed coil changes, the induced current will be generated and the corresponding open-circuit voltage and short-circuit current can be represented as: Equation (1) E E M G = d Ф d t = N S d B d t Equation (2) I E M G = E E M G R where N is the number of the coil turns, Ф is the magnetic flux, t is the time, B is the magnetic flux density, S is the area of the coil, and R is the resistance of the coil. Figure 2 Working mechanisms of the EMG and TENG (A) Basic model of the EMG. (B) Magnetic flux density contour of the whole volume: (i) original position; (ii) intermediate position. (C) Magnetic induction line distribution of the midsection: (i) original position; (ii) intermediate position. (D) The magnetic flux density with respect to the rotation angle of magnets simulated by JMAG software. (E) Induced circuit voltage against EMG simulated via JMAG software. (F) The numerical calculation results about induced voltage of the EMG with different arrangements of coils and magnet number. (G) Basic model of the TENG part and corresponding schematic illustrations of the charge distribution: (i) initial state; (ii) intermediate state; (iii) the third state. (H) Potential distribution between the paired electrodes simulated via COMSOL: (i) initial state; (ii) intermediate state; (iii) the third state. To gain a more quantitative understanding of the proposed working principle of the EMG, finite element analysis was employed to calculate the magnetic flux density contour of the whole volume and the magnetic induction line distribution of the midsection via JMAG, as displayed in Figures 2 B and 2C, respectively. There are two critical positions between the rotational magnets and fixed coils. One is the original position where each magnet is well-aligned with a coil as shown in Figures 2 B(i) and 2C(i), and the other is the intermediate position where each magnet is just located in the middle of two adjacent coils as shown in Figures 2 B(ii) and 2C(ii). Figure 2 D shows the magnetic flux density with respect to the rotation angle of magnets, and Figure 2 E displays the induced circuit voltage against EMG part. In the original position, the magnetic flux through the coil is maximum and there is no current in the coil. As the magnets rotate anti-clockwise with taper rollers, the magnetic flux in the coils declines first, which will lead to an induced positive current consequently. Until the magnets rotate to the intermediate position, the magnetic flux reaches a minimum. As the magnets continue rotating and recover to the original position eventually, the magnetic flux in the coil then increases gradually, which will lead to negative current subsequently. A more in-depth simulation analysis was also carried out to optimize parameters of the EMG part under the same rotation speed of 240 r/min. In this section, the arrangement of coil and the magnet number were supposed to play a considerable role in the output performance of the EMG, because the electromagnetic induction process relies on the number of the coil turns and the rate of the change in magnetic flux. For this purpose, the side wall of PMMA frame was evenly distributed with 4, 6, 8, 10, and 12 coils in different outer diameters of 60, 53, 46, 39, and 32 mm correspondingly, with a constant magnet number of 4. The magnetic flux density contour of the whole volume, the magnetic induction line distribution of the midsection, and the detailed induced voltage of the EMG are shown in Figure S4 . Figure 2 F(i) depicts the series-induced voltages of all these five arrangements of coils. It is also found that the induced voltage decreases as the coil number increases from 4, 8 to 12, which are the integral multiples of the magnet number 4, since the turns of coil are actually decreased from 2,220 to 980. When the coil number is 8, the induced voltage is lower than the case when the coil number is 4; however, the average output power is much higher than that case. Figure S6 is the output work curve of the EMG within 1 s, and its slope represents the average power of the EMG. So considering the power output, the arrangement of 8 coils is a favorable choice. It is worth mentioning that the induced voltage tends to be at a remarkable low level when the coil number is 6 and 10, which is because the direction of the induced current in different coils is inconsistent in these two arrangements. In other words, there is an integer proportional correspondence between the coil number and the magnet number. For the purpose of avoiding that inconsistency, the magnet number can only be a divisor of 8, which is 8, 4, 2, and 1. The numerical calculation results about induced voltage of the EMG with different magnet numbers are presented in Figure 2 F(ii). Figure S5 shows the corresponding magnetic flux density contour of the whole volume, the magnetic induction line distribution of the midsection, and the detailed induced voltage of the EMG. It can be observed that the induced voltage increases as the coil number increases from 1 to 8. However, in subsequent experiments, when the magnet number is 8, the magnetic force is so strong that the bottoms of the tapered rollers are attracted together, making the rollers unable to rotate. Therefore the magnet number was optimized as 4 in the device. Figure 2 G shows the basic model of the TENG part, and the working mechanism is based on the coupling effect of triboelectrification and electrostatic induction. Figures 2 G(i)–2G(iii) illustrates the following three consecutive states of a tapered roller between a pair of interdigital electrodes. When the tapered nylon roller and the FEP film are brought into contact, surface charge transfer takes place due to the triboelectrification effect and electrons are injected from the nylon roller into the FEP film because of the higher electron affinity of FEP than nylon. According to the charge conservation law, an equal amount of negative and positive charges would be generated on the FEP film and nylon roller, respectively. At this moment, the tapered nylon roller can be regarded as a uniform positive electric roller. Once the roller was approaching to and/or departing from the two interdigital electrodes, it would create an asymmetric charge distribution via induction in the media causing the electrons to flow between the two electrodes to balance the local potential distribution. The oscillation of the electrons between the paired electrodes in response to the motion of tapered rollers produces an AC output, which forms the fundamental processes of converting mechanical energy into electricity. According to the capacitor model ( Zhang et al., 2014 ), the open-circuit voltage and the short-circuit current of the TENG can be represented as Equation (3) E T E N G = Q C Equation (4) I E M G = d Q d t where Q is the triboelectric charge capacity and C is the capacitance between electrodes. The variations of electric potential distribution between the paired electrodes for three critical positions are also simulated via COMSOL as displayed in Figure 2 H. At the initial state in Figure 2 H(i), the nylon roller is on the left electrode spaced by the FEP film, where corresponds to a maximum electrical potential on the left electrode and a minimum potential on the right electrode. Thus a maximized electrical potential deference between the two electrodes is generated, resulting in a maximum open-circuit voltage. Then the nylon roller starts to rotate clockwise, the open-circuit voltage starts to diminish till the intermediate state between the two electrodes in Figure 2 H(ii), where it turns to zero. Further rotation beyond this position will result in a reversely established open-circuit voltage and is maximized at the third state in Figure 2 H(iii). Output characterization of the hybrid WEH To begin with, a series of experiments were carried out to optimize the output performance of the TENG part under the wind speed of 12 m/s, including the material of dielectric film, the segment number of each electrode, and the thickness of the dielectric film and the electrodes. First, the output performance of dielectric films made of different materials (PTFE, PVDF, FEP) was investigated systematically. Figures 3 A(i)–3A(iii) depict the open-circuit voltage, short-circuit current, and transferred charge of three different dielectric films. It can be noted from Figure 3 A that the dielectric film made of FEP film has the highest output, which is probably because that FEP has a lower dielectric constant than PTFE and PVDF, resulting in a lower capacitance between electrodes and eventually leading to a higher voltage output according to Equation 3 . Second, three types of TENGs with different segment numbers (4, 8, and 12) of electrodes were taken into consideration. According to the decreasing trend of the potential distribution as the roller size decreases in Figure S7 , the size of the roller was retained when increasing the number of electrode segments. Figures 3 B(i)–3B(iii) demonstrate the open-circuit voltage, short-circuit current, and transferred charge of the TENG with those different types. It can be found that the frequency of the output signals increases exponentially with the doubled or tripled segmentation. It can also be noted that the transferred charge of the TENG decreases obviously as the segment number increases from 4, 8 to 12, which may result from the lower magnitude of polarization due to more sharp edges for finer segments. As the open-circuit voltage of the TENG is directly related to the transferred charges between the electrode pairs according to Equation 3 , the open-circuit voltage of the TENG exhibits an obvious declining trend with the increase of segment number. On the contrary, the short-circuit current shows a slightly ascending trend as the segment number increases, mainly owing to the obvious increase of the charge transferring rate between the electrodes from fully contact to fully separation according to Equation 4 . Third, the effect of the thickness of FEP film (30, 50, and 80 μm) on the output performance of the TENG was studied subsequently. Figures 3 C(i)–3C(iii) illustrate the open-circuit voltage, short-circuit current, and transferred charge of the TENG with different thicknesses of FEP films, indicating that there is only a minor decrease in those three output signals of the TENG when the FEP film gets thicker, probably on account of the vertical gap between the triboelectric surface and the electrode plane. It is worth mentioning that when the thickness of FEP is 30 μm, the electric output of the film was significantly reduced due to the severe wear. Hence, the thickness of the FEP is determined as 50 μm after a trade-off between output performance and service lifetime. Last, we researched the output performance of the TENG with different thickness of the Cu electrodes (30, 50, and 80 μm). As displayed in Figures 3 D(i)–3D(iii), the open-circuit voltage, short-circuit current, and transferred charge of the TENG are relatively larger when the thickness of Cu electrode is 50 μm. Figure 3 Optimization of the output performance of the TENG (A) Output performance of the TENG with Cu electrode of different materials of dielectric film (FEP, PTFE, PVDF) at the wind speed of 12 m/s: (i) open-circuit voltage; (ii) short-circuit current; (iii) transferred charge. (B) Output performance of the TENG with Cu electrode of different segment numbers (4, 8. and 12) at the wind speed of 12 m/s: (i) open-circuit voltage; (ii) short-circuit current; (iii) transferred charge. (C) Output performance of the TENG with the FEP film of different thickness (30, 50, and 80 μm) at the wind speed of 12 m/s: (i) open-circuit voltage; (ii) short-circuit current; (iii) transferred charge. (D) Output performance of the TENG with Cu electrode of different thickness (30, 50, and 80 μm) at the wind speed of 12 m/s: (i) open-circuit voltage; (ii) short-circuit current; (iii) transferred charge. The influence of the wind speed on the output performance of the EMG and TENG is then investigated systemically. As the wind speed increases from 3.5 to 15 m/s, the open-circuit voltage of the EMG increases gradually from 3.14 to 47.4 V in Figure 4 A, which is due to the acceleration of change in magnetic flux as the increase of rotation speed. Figure 4 B shows the open-circuit voltage of the TENG increases gradually from 71 to 683 V with the wind speed increasing from 3.5 to 12 m/s and then remains saturated at 12 m/s. When the wind speed is high enough, the triboelectric friction will no longer be enhanced, and that is why contact surface area and triboelectric charge density will hardly increase and the open-circuit voltage approaches a constant. Figures 4 C and 4D depicted the output voltages and power against different external load resistances at the wind speed of 12 m/s. The output voltages of both EMG and TENG grow gradually with the increase of load resistance. The corresponding output power of both reach maximum values and then turn to decrease until the optimal matched load resistances. The output power of EMG reaches the climax of 62 mW at the external load of 660 Ω, which corresponds to a volume output power density of 72.1 W/m 3 , whereas the maximum output power of TENG is achieved as 1.8 mW at the external load of 60 MΩ, which corresponds to an output power density of 2.7 W/cm 2 . From the aforementioned data, it can be concluded that the output characteristics of TENG generates higher voltage but with larger matched load resistance, whereas the EMG generates lower voltage but with relatively higher power output. For a more intuitive comparison of high performance of our device, relevant information is listed as Table 1 . Figure 4 Output characterization of the WEH (A) Open-circuit voltage of the TENG under different wind speeds. (B) Open-circuit voltage of the EMG under different wind speeds. (C) Output voltage and power of the TENG with different external load resistances at the wind speed of 12 m/s. (D) Output voltage and power of the EMG with different external load resistances at the wind speed of 12 m/s. (E) Measured voltages of different supercapacitors in F-level charged by EMG at the wind speed of 12 m/s. (F) Measured voltages of the TENG run for ~35,000 cycles at the wind speed of 12 m/s. Table 1 Comparison and summary of the reported wind energy harvesters with various output performances. Reference Harvester Output power Rotation speed Power density (W/m 3 ·rpm) ( Xie et al., 2013 ) TENG 12 mW 180 rpm a 0.21 b ( Zhao et al., 2019a , 2019b ) EMG and TENG 10.8 mW 100 rpm 0.24 b ( Fan et al., 2020 ) EMG and TENG 18.96 mW 420 rpm a 0.05 b ( Zhang et al., 2016 ) EMG and TENG 17.5 mW 1,000 rpm 0.06 b ( Cao et al., 2017 ) EMG and TENG 12.7 mW 250 rpm 0.05 b ( He et al., 2019 ) EMG and TENG 1.8 mW 150 rpm 0.08 b ( Guo et al., 2019 ) EMG and TENG 3.61 mW 300 rpm a 0.04 b This work EMG and TENG 62 mW 267 rpm 0.27 a The rotation speed is estimated according to the frequency of AC voltage given in the article. b The volume is estimated according to the dimensional parameters given in the article. The charging capability of the EMG is investigated by connecting a full-wave rectifying bridge to charge different supercapacitors of 0.1, 0.47, 1, and 1.5 F, respectively. As the charging curves of the EMG shown in Figure 4 E, a supercapacitor of 1.5 F can be charged to 3 V within 10 min, which implies its excellent charging capacity. For testing the stability of TENG, the device with an optimized structure has been continuously run for ∼35,000 cycles, and the result indicates that the generated open-circuit voltage of ∼640 V did not have an obvious decay after those many cycles, as shown in Figure 4 F. Application demonstration of the hybrid WEH To demonstrate the potential application toward a practical power source, the WEH was first operated to power up lighting electronic, such as an array of LEDs and even a 5-W globe light. It can be seen from Figure 5 B(i) ( Video S1 ) that the individual EMG can easily light up (ii) 360 red LEDs connected in parallel, whereas the individual TENG can visibly light up (iii) 240 red LEDs connected in series. As demonstrated in Figure 5 C, a 5-W globe light was lit up instantly and capable of providing sufficient illumination in complete darkness. Together with solar power generation, the WEH shows potential talent to be used as a supplementary power source for wind-solar complementary generation mode in outdoor lighting as illustrated in Figure 5 A. Figure 5 Performance demonstration of the WEH aimed to field lighting application (A) Schematic illustration of the WEH used as a supplementary power source for wind-solar complementary circuit in field lightening. (B) Pictures of (i) experimental setup of LED demonstration experiment, (ii) 360 LEDs powered by the EMG, and (iii) 240 LEDs powered by the TENG. (C) A 5-W globe light supplied by the WEH for reading texts in the darkness. Video S1. LEDs powered by the WEH, related to figure 5 Owing to the development of fifth-generation (5G) and big data, smart buildings have become the ideal living and office spaces pursued by government agencies, companies, enterprises, and residential building. Each smart building has a number of sensing nodes that form a network in the city and feedback real-time sensing information including temperature, wind speed, and humidity. The hybrid WEH can not only convert wind energy into electrical energy as power supply to sensors, microprocessors, and wireless transmission modules but also as a self-powered anemometer due to the linear relationship between frequency of AC voltage of TENG and wind speed. As schematically illustrated in Figure 6 A, the WEH is applicable for smart building and other IoT applications. On the one hand, the device can be used to activate a temperature and humidity (T&H) sensing node and a hygrothermograph simultaneously in Figure 6 B(i), which can transmit signals to a computer interface via Bluetooth with the aid of a 6,800-μF capacitor. Figure 6 B(iii) shows the charging-discharging curve of the capacitor of 6,800 μF. The experimental setup of the WEH powering wireless T&H sensor node is shown in Figure 6 B(ii), and the corresponding experimental circuit is shown in Figure S8 A. It is worth mentioning that the capacitor that is charged for only 10 s can work as long as 55 s, and a demonstration video is shown in Video S2 . On the other hand, a self-powered anemograph is exploited as an IoT sensing node for outdoor appliance in smart buildings. As illustrated in Figure 6 C(i), the output of EMG is rectified to charge a supercapacitor of 1.5 F to be used as a power source, and the output of TENG provides AC sensing signals to indicate the wind speed according to the rotating speed of the WEH. The corresponding experimental circuit is shown in Figure S8 B. When the rotators turn a full circle, the TENG can generate four complete cycles of AC voltage waveform. That is to say, the frequency of the TENG output waveform can be an available access to obtain the rotation speed of the WEH. Relevant data were measured by an oscilloscope, and the result was depicted in Figure 6 C(iii). The frequency of the TENG increases from 2.2 to 29.8 Hz as the wind speed increases from 3.5 to 15 m/s. A systematic test to inspect the accuracy and effectiveness of our anemometer is demonstrated in Video S3 , and the experimental setup is shown in Figure 6 C(ii). The wind speed is classified into five levels: Calm (<3.5 m/s); Light breeze (3.5–6 m/s); Gentle breeze (6–9 m/s); Fresh breeze (9–12 m/s), and Gale (>12 m/s) according to the meteorological information. During the experiment, the relative position between the WEH and blower remained unchanged and the wind speed was adjusted by a programmable power supply. When the voltage at both ends of the capacitor reached 5 V after charging 20 min by the rectified EMG output, the switch is turned on and the TENG anemograph is charged. Corresponding voltage signal of TENG output was analyzed and processed by Arduino, and eventually the wind speed level was displayed on the computer interface. Considering the capability of powering IoT sensing nodes, the WEH exerts great potential in the application of smart buildings, to provide with temperature, humidity, and wind speed information. Figure 6 Performance demonstration of the WEH aimed to smart building application (A) Schematic illustration of the WEHs used as the power supplies for IoT nodes in smart buildings. (B) (i) Working process of the T&H sensing node powered by the WEH; (ii) experimental setup of the WEH powering wireless T&H sensor node; (iii) charging-discharging curve of the 6,800-μF capacitor when powering the hygrothermograph and the T&H sensor node simultaneously. (C) (i) Working process of the self-powered anemograph; (ii) experimental setup of the self-powered anemograph; (iii) the relationship curve between output frequency of AC voltage of TENG and wind speed. Video S2. The WEH supplies a hygrometer and a T&H sensing node simultaneously, related to figure 6 Video S3. Demonstration of the WEH utilized as a self-powered wind anemometer, related to figure 6 Conclusion In this article, a novel triboelectric-electromagnetic hybrid WEH based on the revolution and rotation movement of tapered rollers was proposed. Fundamentally, both theoretical model and simulation analysis were carried out, followed by a series of simulated analysis and contrast experiments to optimize the output performance of the EMG and TENG. At the wind speed of 12 m/s, the EMG and TENG of the optimized hybrid WEH can generate the maximum open-circuit voltages of 47.4 and 683 V, respectively. Under the same wind speed, the output power of the EMG reaches the climax of 62 mW at the external load of 660 Ω, which corresponds to a volume output power density of 72.1 W/m 3 , whereas the maximum output power of the TENG about 1.8 mW is achieved at the external load of 60 MΩ, which corresponds to a volume output power density of 0.27 W/m 2 . A supercapacitor of 1.5 F can be charged to 3 V in 10 min by the EMG, and the TENG can run for ∼35,000 cycles continuously without obvious decay. The WEH is demonstrated to power up arrays of LEDs and even a 5-W globe light, which indicates potential talent to be used as a supplementary power source for field lightening. The TENG can also be used as a wind speed sensor by analyzing the output frequency characteristics. Last but not least, combined with corresponding power management circuit, wireless sensor node, and terminal monitoring interface, the device can easily power a wireless environmental monitoring system, such as a wireless temperature and humidity sensing node and a self-powered anemograph, which shows bright application prospects in smart buildings, intelligent agriculture, border detection, and so on. Methods Fabrication of the WEH The WEH, including a rotor, a stator, a driver, and a frame, was mainly fabricated through machining. The device possesses a cylindrical structure with a total dimension of Φ148 mm diameter and 50 mm height. The rotor is mainly composed of four tapered nylon rollers, four magnets made of NdFeB, and an aluminum mounting ring with four threaded mounting holes. The stator consists of a pair of copper electrodes, a layer of FEP film, and an array of eight copper coils. An Al wind cup acts as the driving part, which connects with a PMMA end cover through an Al shaft anchored in the frame with a ball bearing. The frame is made of PMMA, which has a large tapered plane with high precision to ensure the smoothness of the rotation movement of rotors. The electrodes on the bottom of the stator were prepared by cutting plotter. Measurement system The output electric signals (open-circuit voltage, short-circuit current, and transfer charge) of the hybridized WEH were measured by a digital oscilloscope (Keysight DSO3032T), a low-noise current preamplifier (Stanford Research SR570), and a programmable electrometer (Keithley 6514), respectively. The wind was generated by a blower (HK130FLJ5), and an anemometer (Omega HHF11A) was utilized to measure the wind speed. The blower was controlled by a programmable power supply (Gwinstek APS-1102A) to provide different wind speeds. The electrical potential distribution of the TENG was simulated by the COMSOL Multiphysics software, and the distribution of magnetic flux density of the EMG was simulated by the JMAG software."
} | 8,932 |
33123518 | PMC7573125 | pmc | 9,291 | {
"abstract": "Aldehydes are a class of highly versatile chemicals that can undergo a wide range of chemical reactions and are in high demand as starting materials for chemical manufacturing. Biologically, fatty aldehydes can be produced from fatty acyl-CoA by the action of fatty acyl-CoA reductases. The aldehydes produced can be further converted enzymatically to other valuable derivatives. Thus, metabolic engineering of microorganisms for biosynthesizing aldehydes and their derivatives could provide an economical and sustainable platform for key aldehyde precursor production and subsequent conversion to various value-added chemicals. Saccharomyces cerevisiae is an excellent host for this purpose because it is a robust organism that has been used extensively for industrial biochemical production. However, fatty acyl-CoA-dependent aldehyde-forming enzymes expressed in S. cerevisiae thus far have extremely low activities, hence limiting direct utilization of fatty acyl-CoA as substrate for aldehyde biosynthesis. Toward overcoming this challenge, we successfully engineered an alcohol-forming fatty acyl-CoA reductase for aldehyde production through rational design. We further improved aldehyde production through strain engineering by deleting competing pathways and increasing substrate availability. Subsequently, we demonstrated alkane and alkene production as one of the many possible applications of the aldehyde-producing strain. Overall, by protein engineering of a fatty acyl-CoA reductase to alter its activity and metabolic engineering of S. cerevisiae , we generated strains with the highest reported cytosolic aliphatic aldehyde and alkane/alkene production to date in S. cerevisiae from fatty acyl-CoA.",
"conclusion": "Conclusion In this work, we successfully engineered an alcohol-forming FACR into one that produces aldehyde and validated the functions of the two domains in the enzyme as well as the catalytic residues. By expressing the engineered maFACR SYK in Saccharomyces cerevisiae and strain optimization through gene deletion to increase substrate availability and inactivate competing pathways, 2,005 μg/L of fatty aldehyde was produced. To our knowledge, this is the highest reported fatty aldehyde titer produced from fatty acyl-CoA in S. cerevisiae . Subsequently, we demonstrated the utilization of our engineered maFACR SYK for downstream application, namely, ALK production, in this work. In combination with culture optimization, we attained ALK titer of 1,540 μg/L and skewed the ALK production profile toward shorter chain length. The ALK titer is the highest achieved to date via cytosolic ALK production in S. cerevisiae from fatty acyl-CoA. We believe that our engineered maFACR SYK and yeast strains re-established the feasibility of aldehyde production from fatty acyl-CoA in S. cerevisiae for potential applications in biosynthesizing ALKs and other valuable aldehyde-derived compounds.",
"introduction": "Introduction Fatty aldehydes are a class of compounds with a wide range of applications, such as fragrances and flavorings ( Kohlpaintner et al., 2013 ). Importantly, due to the reactivity of the carbonyl functional group, they are versatile chemicals that can undergo a wide range of reactions, including oxidation, reduction, addition, imination, and amination ( Murray, 2019 ). Therefore, fatty aldehydes can be converted to a gamut of compounds and are important precursors in the chemical manufacturing industry ( Kohlpaintner et al., 2013 ; Murray, 2019 ). Conventionally, fatty aldehydes and their derivatives are synthesized chemically from fossil resources, which require harsh conditions and expensive and/or toxic catalysts ( Kohlpaintner et al., 2013 ). Alternatively, fatty aldehydes can be biosynthesized under ambient conditions from fatty acids or their acyl-CoA forms via enzymatic reactions in biological systems ( Reiser and Somerville, 1997 ; Koeduka et al., 2002 ; Schirmer et al., 2010 ; Akhtar et al., 2013 ). The aldehydes could also serve as precursors to concurrently produce their derivatives in vivo via other metabolic pathways ( Schirmer et al., 2010 ; Jin et al., 2016 ; Ladkau et al., 2016 ). Thus, metabolic engineering of microorganisms for biosynthesizing fatty aldehydes could provide a platform for sustainable and economical production of aldehydes from renewable resources. By introducing synthetic metabolic pathways, the aldehydes formed could also serve as substrates for conversion to a variety of valuable chemicals. Initial successes in microbial fatty aldehyde bioproduction were achieved in Escherichia coli by employing fatty acyl-CoA reductase (FACR) or fatty acyl-(acyl-carrier-protein) (ACP) reductase (FAAR) to transform endogenous fatty acyl-CoAs and/or fatty acyl-A CPs to aldehydes ( Reiser and Somerville, 1997 ; Schirmer et al., 2010 ). The aldehyde-producing microbes were applied in the context of biofuel production, as aliphatic and olefinic aldehydes can be transformed by aldehyde deformylating oxygenases (ADOs) or aldehyde decarbonylases (ADs) into alkanes and alkenes (ALKs) ( Schirmer et al., 2010 ; Marsh and Waugh, 2013 ), which are ideal biofuel candidates since they are major components in fossil fuels and have high energy density. Subsequently, there was much interest in employing similar fatty acyl-CoA-dependent pathways for fatty aldehyde and ALK production in the model yeast Saccharomyces cerevisiae because it is a robust industrial host able to withstand harsh fermentation conditions and does not succumb to phage contamination ( Hong and Nielsen, 2012 ; Foo et al., 2017 ). However, due to the poor activity of the aldehyde-forming FAARs and FACRs when used in the yeast strain ( Buijs et al., 2014 ; Zhou et al., 2016b ), fatty aldehyde production levels in S. cerevisiae were extremely low, leading to mediocre ALK production titers compared with those achieved in E. coli ( Choi and Lee, 2013 ). Consequently, free fatty acid (FFA)-dependent pathways were preferred for fatty aldehyde and ALK production in S. cerevisiae because carboxylic acid reductase (CAR) and fatty acid α-dioxygenase (DOX) show higher activity in S. cerevisiae and produced more aldehydes as substrates for conversion to ALKs ( Zhou et al., 2016b ; Foo et al., 2017 ). In this work, we sought to generate a catalytically efficient aldehyde-forming FACR to re-establish the feasibility of de novo fatty acyl-CoA-dependent ALK biosynthesis pathway in S. cerevisiae due to the merits of utilizing fatty acyl-CoAs as substrates. First, fatty acyl-CoA is readily available in S. cerevisiae for utilization in fatty acyl-CoA-dependent pathways without the need to overexpress thioesterases and delete fatty acyl-CoA synthetases to accumulate FFAs, which are required when implementing FFA-dependent pathways ( Runguphan and Keasling, 2014 ). Second, the coenzyme A moiety is hydrophilic and possesses both acidic and basic functional groups. Therefore, fatty acyl-CoAs are much more soluble over a wider range of pH to serve as substrates than FFAs ( Forneris and Mattevi, 2008 ), which are soluble only at high pH. Third, fatty acyl-CoAs are intracellular, while FFAs upon formation can diffuse or be transported out of the cells, often resulting in under-utilization of FFAs and resource wastage due to challenges in transporting extracellular FFAs back into the cells ( Teixeira et al., 2017 ). Although no catalytically efficient aldehyde-forming FACR has been identified for application in S. cerevisiae , high levels of fatty alcohols have been produced in S. cerevisiae using heterologous alcohol-forming FACRs ( Runguphan and Keasling, 2014 ; Feng et al., 2015 ; Zhou et al., 2016a ). Hence, we aim to repurpose alcohol-forming FACR for aldehyde production by protein engineering. Alcohol-forming FACRs possess two reductase functions: one for reduction of fatty acyl-CoAs to aldehydes and another for subsequent reduction of aldehydes to alcohols. Although many of these FACRs have only one active site for both reductase functions ( Hellenbrand et al., 2011 ), an alcohol-forming FACR from Marinobacter aquaeolei VT8, maFACR, was predicted to have two distinct domains, each putatively performing one reductase function ( Willis et al., 2011 ; Figure 1A ). Moreover, functional expression of maFACR in S. cerevisiae has been reported for fatty alcohol production ( d’Espaux et al., 2017 ). Therefore, maFACR is a good candidate for rational engineering into an aldehyde-forming FACR by inactivating the domain that reduces aldehyde to alcohol ( Figure 1B ). Herein, we described identification of the catalytic residues of maFACR and verification of the two domains’ functions. Subsequently, maFACR was engineered into an aldehyde-forming FACR by inactivating the aldehyde reductase domain through mutation of the corresponding catalytic residues. In vivo production of aldehyde in S. cerevisiae was demonstrated using the engineered maFACR, and the production host was optimized to improve the aldehyde titer by increasing fatty acyl-CoA availability and deleting competing pathways. To exemplify application of the engineered maFACR for pathway construction in S. cerevisiae , the engineered maFACR was co-expressed with a cyanobacterial ADO (cADO) to achieve de novo production of ALK ( Figure 1C ). Upon optimization of the culture condition and expression system, we attained the highest reported cytosolic production of fatty aldehyde and ALK from fatty acyl-CoA in S. cerevisiae reported to date. FIGURE 1 Schematic illustration of the maFACR engineering strategy and metabolic pathway for the fatty acyl-CoA-dependent production of alkanes and alkenes (ALKs) in engineered Saccharomyces cerevisiae . (A) maFACR converts fatty acyl-CoA to alcohol without releasing the aldehyde intermediate. It putatively has two distinct domains: a fatty acyl-CoA reductase (FACR) domain for reducing fatty acyl-CoA to aldehyde and a fatty aldehyde reductase (FALDR) domain to reduce aldehyde to fatty alcohol. The FALDR domain has catalytic residues A1, A2, and A3, which are found in this work to be Ser126, Tyr152, and Lys156, respectively. (B) Mutating the FALDR catalytic residues A1, A2, and A3 with the S126D, Y152F, and K156A modifications, respectively, inactivates the domain, thus allowing the release of aldehyde from the enzyme. (C) The engineered aldehyde-forming maFACR* is employed for conversion of endogenous fatty acyl-CoAs to aldehydes in S. cerevisiae and subsequent production of ALKs by deformylation of the aldehydes with cADO. To improve aldehyde titer, a transcription regulator (TR) was deleted to increase fatty acyl-CoA production, and ADHs were inactivated to diminish reduction of aldehydes to alcohols. INO1 , inositol-3-phosphate synthase.",
"discussion": "Discussion Aldehyde-forming bacterial enzymes have been employed with success for producing aldehydes in Escherichia coli , particularly the FACR from Acinetobacter baylyi and FAAR from Synechococcus elongatus ( Schirmer et al., 2010 ; Lehtinen et al., 2018 ). However, there has been limited success in converting fatty acyl-CoAs into aldehydes in Saccharomyces cerevisiae due to the low activity of the aldehyde-forming FAARs and FACRs when employed in the yeast strain ( Supplementary Figure S7 ; Buijs et al., 2014 ; Zhou et al., 2016b ). In contrast, alcohol-forming bacterial, mammalian, and avian FACRs have been functionally expressed in S. cerevisiae for high-level fatty alcohol production ( Runguphan and Keasling, 2014 ; Feng et al., 2015 ; d’Espaux et al., 2017 ). Analysis of the mammalian and avian FACRs shows only one distinct active site for reduction of both fatty acyl-CoA and aldehyde ( Hellenbrand et al., 2011 ); thus, these enzymes are difficult to engineer rationally to eliminate solely the FALDR activity. On the other hand, maFACR has two distinct domains that appear to correspond to FACR and FALDR domains. Therefore, we selected this enzyme for rational engineering because the FALDR activity can be inactivated independent of the FACR activity. Indeed, we successfully repurposed in this work the alcohol-forming maFACR into one that is aldehyde-forming, thus demonstrating the importance of protein engineering for synthetic biology and metabolic engineering applications ( Foo et al., 2012 ). Through in vivo enzymatic assay, we have verified the catalytic roles of the two domains of maFACR and identified the catalytic residues involved. The reduction of fatty acyl-CoA to fatty alcohol by maFACR was proposed to proceed via a reaction mechanism where a two-step reduction occurred within one active site or two highly cooperative active sites through a hemithioacetal intermediate covalently bound to maFACR ( Willis et al., 2011 ). Our results indicate that two active sites are involved, whereby fatty acyl-CoA is reduced to aldehyde in the C-terminal domain and further reduced to alcohol in the N-terminal domain. Additionally, structures proposed by homology modeling of maFACR with the Robetta server do not show any cysteine near the catalytic residues ( Supplementary Figure S8 ). Thus, an enzyme-bound thiohemiacetal intermediate appears to be unlikely. Since the two domains of maFACR belong to SDR families, we propose that each domain employs the general SDR catalytic mechanism involving the Ser-Tyr-Lys triad for substrate binding, hydride transfer, and co-factor binding ( Figure 7 ; Nobutada et al., 2001 ). It is unclear how the aldehyde intermediate is transferred from the FACR domain to the FALDR domain without releasing the aldehyde from the enzyme, but efficient substrate channeling between domains/protomers has been well-documented in enzymes ( Huang et al., 2001 ). The crystal structure of maFACR will be required to determine the exact mechanism for aldehyde transfer between the domains. FIGURE 7 Proposed reaction mechanism of maFACR. We propose a mechanism whereby fatty acyl-CoA is reduced in the C-terminal fatty acyl-CoA reductase (FACR) domain (in green) to aldehyde, which is transferred to the N-terminal fatty aldehyde reductase (FALDR) domain (in blue) for further reduction to fatty alcohol. Ser515 and Tyr527 in the FACR domain first activate the carboxyl group of the fatty acyl-CoA by hydrogen bonding. NADPH, which interacts with Lys532 via hydrogen bonding, donates a hydride to reduce the fatty acyl-CoA to a hemithioacetal upon proton donation by Try527. Subsequent elimination of CoASH forms a fatty aldehyde, which is channeled to the N-terminal. Similarly, Ser126 and Tyr152 activate the carbonyl group of the fatty aldehyde to facilitate reduction by a hydride donated by an NADPH bound to Lys156. Upon accepting a proton from Tyr152, a fatty alcohol is formed. APRPP, adenosine-2-phosphate ribose pyrophosphate moiety of NAPDH. We employed the engineered maFACR SYK for in vivo aldehyde production in S. cerevisiae BY4741 and was already able to produce 1,376 μg/L aldehyde without strain optimization. This contrasts with previous reports that deletion of the aldehyde dehydrogenase HFD1 was critical for aldehyde production in S. cerevisiae by preventing oxidation of aldehydes to fatty acids ( Buijs et al., 2014 ; Zhou et al., 2016a ). Interestingly, deleting this gene from ADH6 Δ was deleterious to aldehyde titer, drastically reducing the total aldehyde formed by 97.5–33 μg/L ( Supplementary Table S8 ). This could be due to differences in strain background, as HFD1 Δ in BY4741 also led to reduced titer when DOX from rice was used for aldehyde production ( Foo et al., 2017 ). Thus, HFD1 deletion was not investigated further. Nevertheless, we successfully improved BY4741 to increase aldehyde and reduce alcohol production by deleting several alcohol dehydrogenases and upregulating fatty acyl-CoA biosynthesis. Although our efforts in this work have enhanced aldehyde biosynthesis in S. cerevisiae , there is still much room for improvement. To further increase aldehyde production, directed evolution of maFACR SYK and other alcohol-producing FACRs may be explored to generate mutants with higher aldehyde-producing ability. Notably, TaFACR and MmFACR from owl and mouse, respectively, were shown to produce much more fatty alcohols than maFACR ( d’Espaux et al., 2017 ), suggesting that these avian and mammalian FACRs have higher activities in converting fatty acyl-CoAs to aldehyde intermediates. However, as aforementioned, avian and mammalian FACRs may share the same active site for reduction of both fatty acyl-CoAs and aldehydes ( Hellenbrand et al., 2011 ) and thus could not be easily engineered rationally like maFACR to eliminate the FALDR activity. Nevertheless, if directed evolution of TaFACR and MmFACR can significantly increase the affinity for fatty acyl-CoA over aldehyde, highly active aldehyde-forming variants can potentially be generated. Furthermore, despite deletion of several transcription regulators and ADHs, the improvement in aldehyde accumulation is still limited. One possible reason is the presence of several aldehyde dehydrogenases (ALDHs) in S. cerevisiae other than HFD1 . Deletion of the five other major ALDHs ( ALD2–6 ) ( Navarro-Avino et al., 1999 ) may be evaluated to determine if enzymatic oxidation could be reduced to aid aldehyde accumulation. Expression of efflux pumps and the use of solvent overlay may also be investigated to transfer the aldehydes out of the cells to drive the flux toward aldehyde production by minimizing in vivo enzymatic conversion of aldehyde to by-products ( Zhang et al., 2016 ; Zhou et al., 2016a ; Perez-Garcia and Wendisch, 2018 ). The use of efflux pumps and solvent overlay has been successfully employed to improve biochemical production and hence may also be applicable for improving the accumulation of aldehydes ( Zhang et al., 2016 ; Zhou et al., 2016a ). For ALK production, BYΔ6OYGA is the best host strain, although BYΔ6O and BYΔ6YGA are better host strains for producing aldehydes, suggesting a synergy between the deletion of OPI1 along with the four ADHs that benefits the deformylation of aldehydes to ALKs, particularly those of shorter chain lengths. The reason is unclear, although it is possible that OPI1 deletion increased the availability of fatty acyl-CoA, and deletion of the four medium-chain ADHs reduced competition from the ADHs with the cADO for the shorter chain-length aldehyde substrates, thus increasing the titer and skewing the ALK production profile toward shorter chain length. It is also noted that YDR541C , GRE2 , and ARI1 are NADPH-dependent ADHs ( Liu and Moon, 2009 ; Choi et al., 2010 ; Moon and Liu, 2015 ). Therefore, their absence may improve availability of NADPH to cADO, which requires two molecules of NADPH for each deformylation reaction, hence accelerating the deformylation step. Further experiments will be required to elucidate the roles of the deletions in BYΔ6OYGA that promote ALK production. Nonetheless, we have achieved ALK titer up to 1,540 μg/L, which is to our knowledge the highest cytosolic ALK production to date in S. cerevisiae from fatty acyl-CoA. Even without genetic modification of the parent strain BY4741, our ALK production pathway using our engineered maFACR attained 1,496 μg/L ALK, which is already approximately 40- and 10-fold higher than the ALK titers reported in wild-type ( Zhou et al., 2016a ) and engineered S. cerevisiae strains ( Zhou et al., 2016b ), respectively, using cytosolic fatty acyl-CoA-based pathways with a low-activity aldehyde-forming FAAR ( Table 2 ). In recent works on ALK production in S. cerevisiae , FFA-based pathways using CAR or DOX were favored for forming aldehydes toward ALK production due to the low activity of aldehyde-forming FACRs in S. cerevisiae ( Zhou et al., 2016b ; Foo et al., 2017 ). With our engineered maFACR SYK , we have achieved ALK titers that are comparable with those attained via FFA-dependent pathways, including those based on fatty acid decarboxylases ( Chen et al., 2015 ; Zhu et al., 2017 ; Table 2 ), thus re-establishing the viability of the fatty acyl-CoA-based ALK production pathway. By employing maFACR SYK in conjunction with novel strategies, such as organelle targeting of the ALK production pathway ( Xu et al., 2016 ; Zhou et al., 2016a ) and genetic circuit development ( Lo et al., 2016 ; Xia et al., 2019 ), and expressing maFACR SYK in non-conventional oleaginous host strains, such as Yarrowia lipolytica ( Xu et al., 2016 ), ALK production in yeast can potentially be further improved. However, the ALK titers obtained in yeast strains still pale in comparison with those achieved in E. coli ( Choi and Lee, 2013 ). More studies are required to identify the bottlenecks of ALK production pathways in yeast, such as competing pathways, co-factor availability, and low activity of cADO. As advances in synthetic biology and synthetic genomics for S. cerevisiae gain momentum ( Chen et al., 2018 ; Foo and Chang, 2018 ), new tools are increasingly available for improving characteristics of yeast to maximize the potential of yeast as a production host for fatty aldehydes and their derivatives. TABLE 2 Comparison of reported ALK production titers in yeast. Host ALK pathway enzymes Remarks ALK titer (mg/L) a References Saccharomyces cerevisiae Engineered FACR and cADO Wild-type S. cerevisiae 1.496 This work Engineered FACR and cADO Four ADHs were deleted. Fatty acyl-CoA biosynthesis was upregulated by deleting OPI1. 1.54 S. cerevisiae Cyanobacterial FAAR and cADO HFD1 aldehyde dehydrogenase gene was deleted. 0.14 Zhou et al., 2016b CAR and cADO POX1 was deleted to inactivate beta-oxidation. HFD1 and ADH5 were deleted to inactivate competing pathways. 0.8 S. cerevisiae Cyanobacterial FAAR and cADO Cytosolic ALK pathway in wild-type S. cerevisiae. 0.035 Zhou et al., 2016a CAR and cADO Cytosolic ALK pathway in wild-type S. cerevisiae. 0.06 CAR and cADO Cytosolic ALK pathway. POX1 was deleted to inactivate beta-oxidation. HFD1 , ADH5 , and SFA1 were deleted to inactivate competing pathways. 0.7 CAR and cADO Peroxisomal ALK pathway. POX1 was deleted to inactivate beta-oxidation. HFD1 was deleted to inactivate competing pathway. 1.2 CAR and cADO Peroxisomal ALK pathway. POX1 was deleted to inactivate beta-oxidation. HFD1 was deleted to inactivate competing pathway. Peroxisome biogenesis was increased by deleting PEX31–32 and overexpressing PEX34 . 3.55 S. cerevisiae DOX and cADO FAA1 and FAA4 were deleted to accumulate FFAs. 0.074 Foo et al., 2017 Yarrowia lipolytica Bacterial FACR and cADO Cytosolic ALK pathway in an oleaginous yeast host. 3.2 Xu et al., 2016 Bacterial FACR and cADO The ALK pathway was targeted to the endoplasmic reticulum of the oleaginous yeast. 16.8 CAR and cADO Cytosolic ALK pathway in an oleaginous yeast host. 23.3 S. cerevisiae Fatty acid decarboxylase, UndA The host was engineered to produce medium-chain fatty acids and inactivate the beta-oxidation pathway. The highest titer was achieved with 20 g/L of glucose. 3.35 Zhu et al., 2017 S. cerevisiae Fatty acid decarboxylase, OleT FAA1 and FAA4 were deleted to accumulate FFAs. HEM3 was overexpressed to increase the heme co-factor. CCP1 was deleted to accumulate H 2 O 2 . The highest titer was achieved upon gene expression tuning and bioreactor process optimization. 3.7 Chen et al., 2015 ALK, alkane and alkene; FACR, fatty acyl-CoA reductase; FAAR, fatty acyl-(acyl-carrier-protein) reductase; cADO, cyanobacterial aldehyde deformylating oxygenase; FFA, free fatty acid. a Titers of fatty acyl-CoA-dependent ALK pathway are shown in bold. Otherwise, the titers were from FFA-dependent ALK pathways."
} | 5,942 |
29606531 | PMC5980998 | pmc | 9,292 | {
"abstract": "Graphical abstract",
"conclusion": "4 Conclusion This study has demonstrated that the supplementation of appropriate GAC dosage could accelerate the syntrophic degradation of propionate and butyrate efficiently under heavy organic load. Specifically, the degradation rates of propionate and butyrate were sharply increased by 1.5 and 4.2 times at 5 g/L of GAC as compared to the control (GAC0), nevertheless minor increment was found for R max when further increasing GAC dosage to 25 g/L. Therefore, the lower dosage of GAC is recommended to use in anaerobic digester, and economics of this approach for improving digester performance would be favorable. GAC benefits the enrichment of syntrophic oxidation bacteria but need a period of cultivation, thus it is suggested to retain GAC within the continuous feeding digester.",
"introduction": "1 Introduction Biomethane production through anaerobic digestion is one of the most successful strategies utilizing bio-energy worldwide ( Ferguson et al., 2016 , Xiao et al., 2013 ). In general, anaerobic methanogenesis is carried out by several groups of microorganisms involved in the hydrolysis, acidogenesis, acetogenesis and methanogenesis processes. Fermentative bacteria and acetogens produce volatile fatty acids (VFAs) and other intermediates, such as lactate, ethanol and butanol and, etc., from the degradation of complex macromolecules ( Karthikeyan and Visvanathan, 2013 , Lee et al., 2016 ). Methanogens utilize simple organic substrates, such as acetate, CO 2 /H 2 , methanol, and formate to generate methane ( Mcinerney et al., 1981 , Pan et al., 2016 , Stuckey and David, 1999 ). As VFAs other than acetate can’t be directly used by methanogens and therefore propionic and butyric acids are mostly found in the effluent from digester with high loads ( Li et al., 2017 , Viggi et al., 2014 ). In fact, the oxidation of propionate and butyrate are highly endergonic under standard conditions and occurs only if methanogens keep the concentrations of these intermediate products low ( Müller et al., 2010 ). Propionate and butyrate are firstly converted to acetate and CO 2 /H 2 by acetogens, and then they are utilized by aceticlastic- and hydrogenotrophic- methanogens. Syntrophic interspecies H 2 transfer is essential to make the reaction energetically favorable ( Müller et al., 2010 , Schink, 1997 ). There are considerable studies aiming to strengthen the syntrophic metabolism within methanogenic conditions by supplementing conductive iron oxides, such as magnetite and Fe 0 ( De Vrieze et al., 2016 , Yamada et al., 2015 ) or conductive carbon materials, such as activated carbon ( Liu et al., 2012 , Xu et al., 2015 ), biochar ( Luo et al., 2015 ), carbon cloth and graphite ( Dang et al., 2017 , Lee et al., 2016 , Mumme et al., 2014 , Zhao et al., 2015 ) etc. The stimulated methane production in reactors with conductive materials might be attributed to the promotion of direct interspecies electron transfer (DIET) ( Li et al., 2017 , Liu et al., 2012 , Rotaru et al., 2014a , Rotaru et al., 2014b ). One potential reason for this is that the availability of non-biological conductive materials may save cells energy because they do not need to produce as extensive extracellular biological electrical connections, such as electrically conductive pili and c-type cytochromes ( Zhao et al., 2015 ). Carbon materials could also provide high specific area for the effective immobilization of syntrophic microorganisms ( Li et al., 2017 , Luo et al., 2015 , Kindzierski et al., 1992 ). Zhao et al. (2016) found that the abundance of Geobacter species, such as G. sulfurreducens and G. lovleyi increased in the propionate- and butyrate-fed reactors, accounting for 20% of the community attached to biochar, meanwhile Syntrophomonas and Smithella species declined. Nevertheless, Dang et al. (2017) reported that granular activated carbon (GAC) seemed to significantly increase the abundance of syntrophic bacteria such as Syntrophomonas, Symbiobacterium and Desulfotomaculum species, whereas Geobacter were not enriched in any of the OFMSW reactors supplemented with GAC. The distinct results might either attributed to the different inoculums or different carbon sources, e.g. single/mixed VFAs, ethanol or complex organic matters ( Kato et al., 2012 , Wang et al., 2016 ). Thus further investigations are in demand to understand the syntrophic communities for propionate and butyrate in digester with carbon materials. Furthermore, it is noted that only simple comparison between AC treated group and blank group has been reported in most studies, and the supplementing dosage of AC varied widely (e.g. from 0.005 to 50 g/L), as shown in Table S1 . Nevertheless, Chen et al. (2014) reported that the metabolism rates of ethanol in methanogenic reactor increased when the amount of carbon cloth was doubled from 10 g/L to 20 g/L, but without further interpretation. Therefore, it is also necessary to clarify whether there is a dose-dependent effect and provide a quantitative basis for related practices. Based on the above rationale, this study has investigated the degradation kinetics of acetate, propionate and butyrate, separately in methanogenic digesters supplemented with a series of GAC dosages (i.e. 0.5–25 g/L). Meanwhile, two different organic loads of substrate, i.e. 1 g/L and 5 g/L were compared. The rates of VFAs’ degradation and methane generation were evaluated by using first-order kinetics and Modified Gompertz model. The high throughput technique was used for 16s rDNA sequencing to detect the microbial community structure, and the alternation of syntrophic VFAs degrading bacteria and methanogens due to GAC addition was discussed in this study.",
"discussion": "3 Results and discussion 3.1 Profile of VFAs degradation and methane generation at low strength Syntrophic interaction is essential to overcome the thermodynamic barriers in the anaerobic oxidation of fermentation intermediates especially propionate and butyrate ( Hattori, 2008 ). In present study, we examined the methanogenic degradation of HAc, HPr and HBu, respectively with the supplementation of GAC at different organic loads, i.e. 1 g/L and 5 g/L, which profiles of cumulative methane production and VFAs declination are presented in Figs. 1 and 2 . Fig. 1 Degradation profiles of 1 g/L VFA (a–c) and corresponding methane production (e–h) with different dosages of GAC. Fig. 2 Degradation profiles of 5 g/L of VFA (a–c) and corresponding methane production (e-h) with different dosages of GAC. With the initial concentration of 1 g/L, three species of VFAs i.e. HAc, HPr and HBu were rapidly degraded, which almost vanished after 5 days. Similarly, the lag phase of methane generation could be neglected. It indicates that the enriched microbial consortia have strong metabolic ability for the specific substrate, i.e. acetate, propionate and butyrate. Furthermore, there is no obvious difference among the reactors with the same substrate and different dosage of GAC as shown in Fig. 1 . The calculated kinetic values of ultimate methane yield ( P CH4 ), maximum production rate ( R max ), and lag phase (λ) from Modified Gompertz model were presented in Fig. 3 . The P CH4 from acetate was 0.45–0.49 mmol-CH 4 /mmol-C added , which was close to the theoretical value of 0.5 mmol-CH 4 /mmol-C added . The average R max of HAc was 0.16 mmol-CH 4 /mmol-C added /d, which value was slightly higher than the previous study, i.e. 0.107–0.143 mmol-CH 4 /mmol-C added /d ( Lü et al., 2013 ). Meanwhile P CH4 from propionate and butyrate were slightly lower than their theoretical values, i.e. 0.48 and 0.68 mmol-CH 4 /mmol-C added . Thus the strengthening effect of GAC at low strength was not prominent. Fig. 3 The calculated kinetic values of lag phase λ (a–c), maximum production rate R max (d–f) and ultimate methane yield P CH4 , (g–i) from Modified Gompertz model. 3.2 Profile of VFAs degradation and methane generation at high strength With the initial concentration of 5 g/L, acetate-acclimated culture still showed high metabolic activity and the supplementation of GAC almost did not affect the methane production rate. The ultimate methane yielding P CH 4 from acetate was around 0.45 mmol-CH 4 /mmol-C added , which was close to the value obtained in low strength of acetate (e.g. 1 g/L). However, the methanogenic conversion of propionate and butyrate were obviously inhibited when increasing the substrate concentration from 1 g/L to 5 g/L. The lag-phase time of propionate and butyrate reactors was 4.2 days and 12.7 days for GAC0. It is like previous studies that methanogenesis was vulnerable to high concentration of VFAs, mainly attributing to the inhibition of growth and metabolism of methanogens by undissociated VFAs ( Pavlostathis and Giraldo-Gomez, 1991 ). Nevertheless, the addition of GAC in present study was found to accelerate the metabolism of propionate and butyrate significantly. As shown in Fig. 2 , when increasing GAC concentration from 0.5 g/L to 25 g/L, the lag-phase period reduced from 3.4 d to 0.9 d for propionate-fed reactors, and from 12.7 d to 7.8 d for butyrate-fed reactors. The degradation kinetics ( k ) of each VFAs species are calculated and presented in Table 1 . Basically, the R 2 is high, only the R 2 of HBu (5.0 g/L) was lower than others especially at low GAC dosage. The reason could be attributable to the inhibition effect, leading to the deviation from sigmoidal function. With 5 g/L of propionate as substrate, about 1.5 times of increment was found for the k value in digesters, i.e. from 0.0022 to 0.0056 h −1 when increasing GAC dosage from 0 to 25 g/L. The stimulating effect on degradation of butyrate was more significant, which kinetics value was enhanced by 7.1 times i.e. from 0.0043 to 0.0306 h −1 when GAC dosage increased from 0 to 25 g/L. The above results clearly indicated that the supplementation of GAC could accelerate methanogenesis from propionate and butyrate in a dose-dependent manner. Li et al. (2015) also reported about the stimulating effect of nano-Fe 3 O 4 might be limited by the supplementing concentration. It should also be noted that although the degradation rates of propionate and butyrate were significantly increased by supplementing GAC, their rates were still lower than that of acetate in the range of 0.0386–0.0431 h −1 . These results further confirmed the difficulty in conversion of propionate and butyrate and the importance to further explore the promotion mechanism. Table 1 First-order kinetics for the consumption rate of propionate and butyrate. GAC dosage HAc HPr HBu VFA k (h −1 ) R 2 k (h −1 ) R 2 k (h −1 ) R 2 0 g/L 0.0386 ± 0.0065 0.92 0.0022 ± 0.0002 0.97 0.0043 ± 0.0015 0.83 5.0 0.5 g/L 0.0371 ± 0.0049 0.95 0.0033 ± 0.0004 0.92 0.0036 ± 0.0009 0.77 g/L 5 g/L 0.0431 ± 0.0046 0.97 0.0029 ± 0.0041 0.75 0.0180 ± 0.0041 0.82 25 g/L 0.0393 ± 0.0078 0.90 0.0056 ± 0.0004 0.99 0.0306 ± 0.0021 0.98 0 g/L 0.0228 ± 0.0042 0.91 0.0328 ± 0.0013 0.99 0.0251 ± 0.0040 0.99 1.0 0.5 g/L 0.0279 ± 0.0044 0.93 0.0319 ± 0.0025 0.98 0.0247 ± 0.0022 0.98 g/L 5 g/L 0.0274 ± 0.0041 0.94 0.0329 ± 0.0017 0.99 0.0248 ± 0.0031 0.99 25 g/L 0.0252 ± 0.0036 0.94 0.0333 ± 0.0023 0.99 0.0268 ± 0.0025 0.99 3.3 Variations of intermediate products The dynamics of substrate and intermediate products were presented in Fig. 4 . Different accumulation of intermediate product i.e. acetate was found in digesters with different substrate and GAC dosages. Scholten and Conrad (2000) also found the accumulation of acetate in reactor fed with propionate or butyrate as the sole substrate. The syntrophic interaction is essential to overcome the thermodynamic barriers in the anaerobic oxidation of fatty acids. The main pathways involved in the syntrophic degradation of acetate, propionate and butyrate are shown as follows (3) CH 3 CH 2 CH 2 COO - + 2 H 2 O → 2 CH 3 COO - + H + + 2 H 2 Δ G 0 ′ = + 48.1 kJ / mol (4) CH 3 CH 2 COO - + 2 H 2 O → CH 3 COO - + CO 2 + 3 H 2 Δ G 0 ′ = + 76.0 kJ / mol (5) CH 3 COO - + H + + 2 H 2 O → 2 CO 2 + 4 H 2 Δ G 0 ′ = + 94.9 kJ / mol (6) CH 3 COO - + H 2 O → HCO 3 - + CH 4 Δ G 0 ′ = - 31.0 kJ / mol (7) HCO 3 - + 4 H 2 + H + → CH 4 + 3 H 2 O Δ G 0 ′ = - 135.6 kJ / mol The Gibbs free energy of Eqs. (3) , (4) , (5) are quite high, which turns to be exergonic reaction only at the low partial pressure of H 2 or low concentration of acetate ( Müller et al., 2010 ). Thus, the syntrophic degradation requires both hydrogenotrophic and aceticlastic methanogens to consume H 2 and acetate. Fig. 4 Variations of acetogenic intermediate during the anaerobic degradation of HPr (a–d) and HBu (e–h) at varied GAC dosages. In digesters with propionate as substrate, a slow declination rate of propionate was observed and the accumulation of intermediate product acetate was barely detected. Whereas, acetate was found to be accumulated in digesters with butyrate as substrate ( Fig. 4 e–h). As the inoculum is the enriched culture mixture of syntrophic acetogenic bacteria and methanogenesis, the detected concentration of acetate was the balance result between syntrophic acetogenesis and methanogenesis. By observing the status of acetate accumulation, results indicate that a slight depression of methanogenesis was occurred in digesters with propionate and butyrate, meanwhile the supplementation of GAC could trigger the acetogenic conversion of propionate and butyrate. The study of Viggi et al. (2014) had evaluated the maximum electron carrier flux occurred in digester with conductive magnetite particles, i.e. DIET which rate is 10 6 higher than that associated with interspecies H 2 transfer. Such a scenario could be applied to present study in degrading butyrate and propionate but with a slower rate of electron carrier flux as the conductivity of GAC was smaller than magnetite particles. In addition, it is noted that the substrate of butyrate tends to be exhausted in GAC25 within 8 days ( Fig. 4 h) whereas the methane generation continued to 14 days ( Fig. 2 f). On one side, apart from acetate other intermediates such as formate and H 2 might existed in the digesters to contribute for the methane generation. As evidenced by Fig. S1 , the remaining TOC concentration was about 211 mg/L when butyrate nearly consumed at Day 7. Nevertheless, the concentration of hydrogen has not been monitored in this study, which warrants further investigations. On the other side, the remaining methane production might be derived from the absorbed substrate on GAC at high dosage. The results of adsorption experiment showed that the adsorption capacity of GAC for HAc, HPr and HBu were 22, 25 and 38 mg/g, respectively. 3.4 Characteristics of microbial community This study compared microbial population enriched with different substrates, i.e. acetate, propionate and butyrate, as well as the effects of GAC supplementation by comparing GAC0 and GAC5. The relative abundance of bacterial and archaeal community at genus level are presented in Fig. 5 and Table S3 . Fig. 5 Bacterial (a) and Archaeal (b) community structure at genus level in the anaerobic sludge with 5 g/L GAC and without GAC. 3.4.1 Bacteria The bacterial community structure changed along with different incubation condition. In reactors fed with acetate, Aminicenantes and Thermovirga were the predominant bacteria ( Fig. 5 a). Until now there are only three Aminicenantes genomes have been sequenced and no cultured representatives of this lineage. Nevertheless, putative genes for formate dehydrogenase (i.e., hydrogenase-3 and formate hydrogenylase) have been identified in Aminicenantes species and could be used to convert formate to hydrogen and carbon dioxide as terminal products of fermentation ( Robbins et al., 2016 ). One OTU clustered to Aminicenantes was also reported to use Wood-Ljungdahl pathway in reverse to consume acetate and generate CO 2 in syntrophic association with a hydrogenotrophic methanogen of the order Methanomicrobiales ( Gies et al., 2014 ). As shown in Fig. S2 , the function of Aminicenantes is closely clustered to Syntrophobacter, which might make contribution to the syntrophic degradation of VFAs. Thermovirga accounted for 45% and 21% of the total OTUs in GAC0 and GAC5 fed with acetate. Thermovirga together with Aminivibrio , Acetothermia are attached to the family of Synergistaceae under the phylum of Synergistetes. Similarly, some identified acetate degrading bacteria belong to the Synergistetes, which is probably syntrophic acetate oxidation coupled with hydrogenotrophic methanogens ( Ito et al., 2011 ). It has found that the identified Synergistetes group had a lower affinity to acetate and a higher acetate utilization rate than Methanosaeta -like acetoclastic methanogen ( Ito et al., 2011 ). Thus, Methanosaeta and Synergistes group seem to be not competitive, but cooperative for fluctuating concentration of acetate in the anaerobic batch reactor used in this study. The composition of Thermovirga was relatively low in GAC0 fed with propionate and butyrate, whereas a higher proportion of Exilispira was detected, which is affiliated to the phylum S pirochae te. Bacteria within the Spirochaetes are frequently detected in anaerobic digestion systems that treat municipal sludge, livestock wastewater and synthetic organic matters ( Lee et al., 2013 ). The selective enrichment of Spirochaetes was reported in reactors accepting fatty acids especially acetate as substrate, suggesting the possible role of Spirochaetes in syntrophic acetate oxidation ( Lee et al., 2013 ). It also found that the relative abundance of Syntrophomonas and Smithella were higher in propionate reactors, which members are propionate and butyrate oxidizers ( Müller et al., 2010 , Mcinerney et al., 1981 ). Liu et al. (2011) had examined the organisms involved in syntrophic oxidation of butyrate in paddy soil using DNA based stable isotope probing, where Syntrophomonas spp. together with methanogens Methanosarcina and Methanocella were the most active. It seems that the syntrophic fatty acids oxidation community was susceptible to the inoculum source. 3.4.2 Archaea The relative abundance of Archaea in sludge samples ranged from 21% to 61% as seen in Table S3 . Comparatively, Methanosaeta, Methanobacterium, Methanosarcina and Methanolinea are the dominant archaea species in the enriched culture degrading VFAs as shown in Fig. 5 b. Methanosaeta is a typical acetoclastic methanogen, while Methanobacterium and Methanolinea belongs to the hydrogenotrophic methanogen species. Methanosarcina produce CH 4 through three pathways using H 2 /CO 2 , acetate and methylated one-carbon compounds ( De Vrieze et al., 2012 ). On one side, although the selectively enriched archaea varied among reactors with different substrates, Methanosaeta predominated in all the reactors ( Fig. 5 b), and the relative abundance of Methanosaeta followed the trend of acetate > butyrate > propionate ( Fig. 6 a). In acetate-fed reactors, Methanosaeta attributed for ∼80.7% of methanogens (HAc0), meanwhile in the propionate- and butyrate-fed reactors, it accounted for 42.8% and 44.8%. Zhao et al. (2016) also reported that Methanosaeta species were predominant with either butyrate or propionate as the substrate. The high abundance of Methanosaeta could be affiliated to the suitable concentration of acetate (max. 68 mM), and it could also consume electrons derived from the oxidation of propionate or butyrate to acetate. Fig. 6 Difference analysis of microbial community on genus level among the groups of HAc, HPr and HBu (a) and between the groups of 0 g/L and 5 g/L GAC (b), and representational difference analysis for the microbial community converging environmental factors of GAC, R max and VFAs concentration (c). On the other side, the relative abundance of Methanosaeta in the digester with GAC generally decreased as compared to the control without GAC. Similar trend was found for Methanobacterium . Whereas the abundance of Methanosarcina and Methanolinea increased. It has been reported that the presence of Methanosarcina could shortly eliminate the negative effect of acidity accumulation and produce methane in high performance ( Wang et al., 2017 ). To data, Methanosarcina and Methanosaeta are the only methanogens known to participate in DIET by directly receiving electrons to reduce CO 2 and produce CH 4 ( Rotaru et al., 2014a , Rotaru et al., 2014b ). The promoting mechanism of DIET has been well understood in ethanol metabolism with a co-culture of Geobacter metallireducens and Methanosarcina barkeri ( Liu et al., 2012 , Rotaru et al., 2014b ). In total, the proportion of three known hydrogenotrophic methanogens (i.e. Methanolinea, Methanobacterium and Methanosarcina ) increased in GAC5 than GAC0. Comparatively, the changes on the Archaeal community of in acetate reactors were less than propionate and butyrate reactors, which was in accordance to the similar reaction kinetics of methane generation with varied GAC dosages. 3.4.3 Difference and correlation analysis of microbial community The influences of various environmental factors such as GAC supplementation, maximum methane yield rate ( R max ) and the concentrations of VFAs on the dynamics of microbial community have been analyzed by the representational difference analysis (RDA) and presented in Fig. 6 . Firstly, results indicate that the supplementation of GAC was regarded as the major environmental factor influencing the microbial community. The distributions of HAc1 and HPr1 and HBu1 were in the same direction with GAC, whereas HAc0, HPr0 and HBu0 were located on the contrary coordinate. Furthermore, the abundance of syntrophic bacteria such as Aminicenantes , Thermovirga, Synergistaceae and Syntrophomonas were also clustered with the same direction of GAC, whereas Methanosaeta was on the contrary coordinate, which is in accordance to the above observation. To data, only Methanosarcina and Methanosaeta are known methanogens to participate in DIET by directly receiving electrons to reduce CO 2 and produce CH 4 ( Rotaru et al., 2014a ). Geobacter is also known as one important bacterial genus to participate in DIET, which accounted for ca. 20% of the community attached to biochar ( Zhao et al., 2016 ) or GAC ( Lee et al., 2016 ). Lee et al. (2016) found the enrichment of exoelectrogens e.g. Geobacter and hydrogenotrophic methanogens (e.g. Methanospirillum and Methanolinea ) from the biomass attached to GAC. However, in this study, the OTUs clustered to Geobacter detected in sludge with GAC was quite low (< 1% of total bacteria). Similar result has been previously reported ( Xu et al., 2015 , Dang et al., 2017 , Barua and Dhar, 2017 ). Thus, it’s a possibility that other organisms rather than Geobacter may also participate in DIET with methanogens as suggested ( Kato et al., 2012 ). Last but not the least, the relative concentration of syntrophic oxidation bacteria and methanogens to the substrate should also be carefully attention ( Ferguson et al., 2016 ). As shown in Fig. 4 e and f, the accumulation of acetate was relatively low when GAC concentration was very low, whereas when GAC concentration increased, the conversion rate of butyrate into acetate was accelerated and the accumulation of acetate accordingly elevated. It seems that with the low GAC dosage, microorganism enriched on the GAC surface and associated DIET could eliminate the resistance to acetogenesis, and the bottleneck turns to be the methanogenesis. However, this deduction requires further demonstration as the intermediate products are not fully identified in present study."
} | 5,958 |
17784941 | PMC2375017 | pmc | 9,294 | {
"abstract": "A global analysis of Pseudomonas putida gene expression performed during the interaction with maize roots revealed how a bacterial population adjusts its genetic program to the specific conditions of this lifestyle.",
"conclusion": "Conclusion The current study constitutes, to our knowledge, the first report on bacterial genomics in the rhizosphere. The main functions identified in this transcriptional profiling study as being specifically expressed in the rhizosphere are integrated in the scheme shown in Figure 4 . Future work should aim at unveiling the regulatory mechanisms that control such reprogramming of transport, metabolic and stress-related functions. We have also demonstrated that RNA samples of good quality and in enough quantity can be obtained from a bacterial population growing in this complex environment so that parameters of great interest in the plant- Pseudomonas interaction, such as the physical contact between the root and the bacteria and also the constant supply of root exudates, can be considered in gene expression studies. This work opens a challenging perspective to the study of mutualistic plant-microbe associations where, besides other determinants, energetic balances between nutrient availability and stress resistance should be considered to explain the success of these interkingdom relationships. Figure 4 Integrated scheme showing relevant bacterial functions induced in the rhizosphere- Pseudomonas interaction. Functions related to genes included in Additional data file 3 have also been included in this figure. See text for details.",
"discussion": "Results and discussion Analysis of the Pseudomonas putida genetic program in the rhizosphere To investigate how Pseudomonas populations readjust their genetic program upon establishment of a mutualistic interaction with plants, we have performed a genome-wide analysis of gene expression of the root-colonizing bacterium Pseudomonas putida KT2440 in the rhizosphere of corn ( Zea mays var. Girona), using microarrays (ArrayExpress repository for microarray data, accession number A-MEXP-949). Among other relevant characteristics, this strain is an excellent root colonizer of plants of interest in agriculture [ 7 ] and activates induced systemic resistance against certain plant pathogens (Matilla et al ., in preparation). Different experiments were designed in order to obtain as broad a picture as possible, comparing rhizosphere populations with three alternative controls: planktonic cells growing exponentially in rich medium (LB medium); planktonic cells in stationary phase in LB medium; and sessile populations established in sand microcosms (defined medium), under the same conditions used to grow inoculated corn plants (see Materials and methods). The combination of these diverse growth conditions balances the contribution of parameters such as growth phase, nutrients and life style to any observed changes in gene expression. Unveiling differentially expressed genes common to all the studies would minimize noise and allow us to identify genes with a reliable and specific change in their expression level in the rhizosphere, likely to be important for survival in this environment. RNA samples were obtained from bacterial cells recovered from the rhizosphere six days after inoculation of gnotobiotic seedlings, and from each of the control settings. Microarrays were hybridized with equal amounts of differentially labeled cDNA and examined for up- and down-regulated genes. Data were processed in two separate ways. The first consisted of evaluating every single experiment (consisting of three biological replicas each) independently, followed by the imposition that genes showing significant changes in gene expression did so in the three experiments, each with a different control. The second analysis evaluated these three experiments through a combined examination of the nine microarrays altogether, followed by a P value adjustment. Finally, the results from both data treatments were compared. Two general observations can be highlighted when rhizospheric KT2440 bacteria are compared to their control counterpart by analyzing each experiment individually. The first is that gene activation is more conspicuous than gene repression in the bacterial rhizospheric life style, as reflected by the fact that over 50 genes were induced more than 6-fold in the three experiments (Figure 1 ). In total, 90 genes appeared consistently up-regulated in the rhizosphere versus all three controls (fold change >2, P value < 0.05), and none down-regulated (fold change <-2, P value < 0.05) (Figure 2 ; Additional data file 1). A selection of up-regulated genes is listed in Table 1 . The second relevant result was that sessile P. putida growing in sand microcosms and stationary phase cells exhibited the most comparable and the most dissimilar gene expression pattern, respectively, with respect to rhizosphere cells (Figure 1 ). It is worth noting that many genes encoding ribosomal proteins are induced in the rhizosphere after six days of colonization compared to stationary phase (Additional data file 1), indicating the existence of active growth and metabolism, at least in a subpopulation of cells. These results offer a view of P. putida life in association with plant roots as a situation where metabolically active bacterial cells grow in a biofilm-resembling state [ 11 ], although with their genetic program adjusted to the presence of the plant. Table 1 Rhizosphere up-regulated ( rup ) genes Fold change Locus - TIGR annotation Log* St. † Sand ‡ Combined § Cytochrome biosynthesis PP0109 - membrane protein putative P > 0.05 10.3 7.9 7.7 PP0110 - cyoE-1 -protoheme IX farnesyltransferase 13.2 51.5 10.2 - PP3183 - SCO1/SenC family protein/cytochrome c 2.9 5 3 - Metabolism PP0326 - soxG -sarcosine oxidase gamma subunit 4.8 7.9 5.8 6 PP1403 - bglX -periplasmic beta-glucosidase 2.5 2.9 2.5 2.6 PP2694 - aldehyde dehydrogenase family protein 9 8.3 LS 10.2 PP2847 - ureJ -urease accessory protein UreJ 22.9 29.6 21.9 24.6 PP3281 - phenylacetic acid degradation protein PaaI putative 6.2 8.1 8.6 7.5 PP3352 - arylsulfatase putative 36.5 17.6 49.4 31.5 PP3746 - glcE -glycolate oxidase subunit GlcE 3.6 3.4 3.6 3.5 PP3923 - phosphoglycerate mutase family protein 4.5 4.3 2.8 3.8 PP4588 - nitroreductase family protein 2.6 2.3 3.3 2.7 PP4782 - thiD -phosphomethylpyrimidine kinase 8.5 5.8 5.5 6.5 PP5076 - gltB -glutamate synthase large subunit 3.4 5.5 2.1 - PP5197 - ubiF -2-octaprenyl-3-methyl-6-methoxy-1,4-benzoquinol hydroxylase 6.9 3.5 5.8 5.2 Secondary metabolism PP3786 - aminotransferase 5.3 4.3 2.1 - Chemotaxis and motility PP4331 - conserved hypothetical protein 4.9 5.6 4.5 5 PP4359 - fliL -flagellar protein FliL 4 4.6 2.8 3.7 PP4391 - flgB -flagellar basal-body rod protein FlgB 3 5.2 5.2 4.3 PP4987 - chemotaxis protein putative 6.4 7.3 4.4 5.9 Regulators and sensor proteins PP1066 - sigma-54 dependent response regulator 13 41.9 10.1 17.6 PP3640 - transcriptional regulator AraC family 19.7 28.9 10.3 - PP4295 - transcriptional regulator TetR family 8.9 8.7 6 7.7 PP4508 - transcriptional regulator AraC family 3.2 2.7 3.2 3 PP0700 - transmembrane sensor putative 21.9 50.4 26.2 30.7 PP2127 - sensor histidine kinase 31.8 14.1 17.3 19.8 PP4959 - sensory box protein/response regulator 14.7 5.1 9.4 8.9 PP5321 - phoR -sensory box histidine kinase PhoR 10.5 9.1 9.7 9.8 Stress adaptation and detoxification PP0373 - Pmp3 family protein 8.1 8.4 8 8.1 PP1874 - glutathione peroxidase (GSH_peroxidase) 4.3 7.9 5.4 5.7 PP2376 - cti -esterified fatty acid cis / trans isomerase 2.4 3.6 2.3 - PP3535 - ggt-1 -gamma-glutamyltransferase 2.2 2.2 2.1 2.2 ABC transporters PP0196 - ABC transporter ATP-binding protein putative 2.4 6.7 3.7 - PP2669 - outer membrane protein putative 9.5 11.6 4.7 8 PP3210 - ABC transporter pernease protein 3.4 4.3 3.9 3.8 PP3223 - ABC transporter periplasmic binding protein (dipeptide) 36.9 25.4 66.4 39.5 PP3802 - cation ABC transporter ATP-binding protein putative 13.8 20.1 5.2 - PP4305 - periplasmic thiosulfate-binding protein 3.2 3.8 2.7 3.2 PP4483 - basic amino acid ABC transporter ATP-binding protein 3.5 4.6 2.6 3.5 Efflux pumps PP0670 - transporter bile acid/Na+ symporter family 5.7 11.1 3.7 - PP0906 - multidrug efflux RND transporter putative 3.5 8.5 2.6 - PP1271 - multidrug efflux MFS transporter putative 11.3 25.6 18.1 17.4 PP2817 - mexC -multidrug efflux RND membrane fusion protein MexC 3.8 6 2.2 - PP3583 - RND efflux transporter permease protein 2.7 4.2 3 - Other transporters PP2385 - azlC -branched-chain amino acid transport protein AzlC 4 4.1 3.2 3.8 PP3132 - polysaccharide transporter putative 3.1 3.9 2.6 3.2 PP5297 - amino acid transporter putative (polyamines) 6.6 8.8 6.7 7.3 DNA replication, recombination and repair PP1476 - conserved hypothetical protein 17.3 76.2 29 33.7 PP2565 - helicase putative 5.5 12.6 5.1 - PP3966 - ISPpu14 transposase Orf1 17.3 11.4 8.1 11.7 Others PP2076 - hypothetical protein 4.4 P > 0.05 7.5 5.2 PP2155 - lolD -lipoprotein releasing system ATP-binding protein 4 2.4 5.2 - PP2560 - transport protein HasD putative 23.3 60 9 - PP3184 - hypothetical protein 6.6 3.9 3.9 4.6 Proteins with predicted general function and hypothetical proteins not mentioned in the text are not included (Additional data file 1). Although the P. putida KT2440 genome is sequenced and annotated [49], the locus functions listed in this table were one by one re-confirmed by comparing the amino acid sequences with those in the databases. The complete list of rup genes (genes with fold induction >2, P value < 0.05, and average signal-to noise A >64) is available in Additional data files 1-3. *Control with LB log cells; † control with LB stationary phase cells; and ‡ control with sessile cells from microcosm without plant. § Genes passing the Bonferroni cutoff after a combined analysis of the nine microarrays altogether. A dash is used to mark those rup genes not passing the Bonferroni cutoff, although they did pass the Benjamini and Hochberg adjustment. LS, low signal (below cutoff). Figure 1 Microarray profiling of bacterial gene expression in the rhizosphere. (a-c) Global gene expression of P. putida KT2440 was analyzed in the rhizosphere versus that of LB log bacteria (OD 660 = 0.7) (a, a'), LB stationary phase cells (OD 660 = 3.3) (b, b'), and sessile bacterial cells incubated in sand microcosms (c, c'). Experimental set up and cDNA preparation are described in detail in Materials and methods. Genes induced (red) and repressed (green), with a P value < 0.05 and A >64 were clustered according to their fold change (>2 to >10) and (≤2 to ≤10) and the number is plotted. Figure 2 Venn diagrams showing the overlap between differentially expressed genes in the rhizosphere. (a) Down-regulated and (b) up-regulated genes resulting from the individual analysis of each experiment: rhizosphere versus LB exponentially growing cells (Rhi/Log), rhizosphere versus LB stationary phase cells (Rhi/St), and rhizosphere versus sessile cells in microcosm without plant (Rhi/Microcosm). Common genes were clustered automatically with a freely available informatics tool [48]. (c) The result of the combined analysis of the three experiments before and after applying adjustments in the P value. Out of the 57 Bonferroni genes, 54 are included in the 90 overlapping rup genes. The combined analysis of the three experiments as a group (nine microarrays) identified a larger number of genes induced and repressed in rhizospheric cells than when the independent analysis, followed by clustering, was done as described above. This was an expected statistical consequence of the increased number of tests in the analysis. Table 2 shows the numbers before and after applying corrections of the P value. Even with the strictest procedure, the Bonferroni correction [ 12 ], which is scarcely used in microarray studies due to its stringency, 57 genes appeared as up-regulated in P. putida in the rhizosphere. Of therse, 54 were part of the group of 90 mentioned above as a result of the independent analysis (Figure 2 ). No repressed gene passed the cutoff with the Bonferroni method (Additional data file 2). The remaining 36 up-regulated genes identified with the independent analysis were part of the group of over 300 obtained after applying a less strict correction [ 13 ] to the combined analysis (Figure 2 ; Additional data file 3). With this method, a substantial number of genes appear as rhizosphere-repressed, although the majority show fold changes close to the -2 cutoff (Additional data file 3). Table 2 Number of differentially expressed genes in the rhizosphere Induced Repressed Non-adjusted P value 376 119 Benjamini and Hochberg 349 85 Bonferroni correction 57 - Result of the combined analysis of nine microarrays, three biological replicates per experiment and three experiments (each with a different control). See the text for details. Dash indicates no gene passed the cutoff. Real time RT-PCR confirmation of changes in the mRNA levels of rhizosphere differentially expressed genes To validate our microarray data by an independent technology, gene expression of six genes among those identified as rhizosphere up-regulated was examined in the rhizosphere versus microcosm by real time RT-PCR. Other neighbor genes not identified in the microarrays as rhizosphere induced were also included in the RT-PCR analysis, that is, PP1477 and PP4988, which likely form part of same transcriptional units as PP1476 and PP4987 (Table 1 ), respectively, and PP3744, the glc transcriptional activator of the rhizosphere induced gene encoding PP3746 (Table 1 ). All these genes were differentially expressed with a fold change higher than three (Table 3 ), confirming the microarray results, and indicating that if genes are listed in Table 1 it does not necessarily mean they were not induced in the rhizosphere. Restrictive conditions imposed to pass the cutoff might be the cause of this underestimation. Gene expression variability within the same operon may also occur, due to factors such as different mRNA stability or the existence of internal promoters. Table 3 Differential gene expression of rup genes (real time RT-PCR) Gene Fold change a PP1476 - conserved hypothetical protein 5.07 ± 0.4 PP1477 - recJ -single-stranded-DNA-specific exonuclease RecJ 5.48 ± 0.7 PP2076 - hypothetical protein 7.13 ± 0.4 PP2560 - transport protein HasD putative 3.84 ± 0.3 PP3744 - glc operon transcriptional activator 3.80 ± 0.2 PP3746 - glycolate oxidase, subunit GlcE 21.30 ± 1.5 PP4987 - chemotaxis protein putative 6.51 ± 0.9 PP4988 - chemotaxis protein putative 5.02 ± 0.8 PP5321 - phoR -sensory box histidine kinase PhoR 4.58 ± 0.1 a Rhizosphere versus microcosm. Average of three samples and standard deviation are shown. Reliable rhizosphere up-regulated ( rup ) genes Following the initial premise of identifying genes with a reliable and specific change in their expression levels, we focused our attention primarily on the 93 genes showing increased expression in the rhizosphere (90 obtained from the independent analysis and the additional 3 that passed the Bonferroni adjustment of the combined analysis; Table 1 ) with respect to any other condition. About one-third of these encode hypothetical proteins whose specific functions have yet to be determined (Additional data file 1). The remaining genes with increased expression in the rhizosphere provide an ample view of the determinants at play in this plant-bacterial interaction (Table 1 ), some confirming previous data about the participation of elements such as flagella or thiamine (vitamin B1) biosynthesis. One conclusion to be drawn is that aside from interspecific competition, which is not contemplated in these experiments, two opposing forces act simultaneously, driving bacterial adaptation to life in the rhizosphere. On one hand, nutrient availability is reflected by the increased expression of genes involved in the uptake of certain carbon and nitrogen sources (in particular amino acids, dipeptides and polyamines), some of them, like glycolate, excreted by plants, as well as metabolic and degradative functions (degradation of aromatic compounds such as phenylacetic and/or phenylalkanoic acids, sarcosine oxidase, plant exopolymers β-glucosidase, urease). The high induction of an urease accessory protein may be related to the limiting nitrogen source, since the nitrogen is likely being used by the plant. Alternatively, compounds other than urea in the root exudates of corn plants might be responsible for this induction, since urea is not produced by this plant [ 14 ]. On the other hand, genes coding for stress response and detoxification proteins also show increased expression in the rhizosphere. These include glutathione peroxidase, a protein of the Pmp3 family, or the fatty acid cis-trans isomerase Cti, which has been related to stress adaptive mechanisms, in particular to membrane-toxic organic compounds [ 15 , 16 ], as well as several putative efflux transporters of toxic compounds. This indicates the necessity to cope with oxidative stress and other damaging agents, for instance, secondary metabolites present in seed and root exudates that show antimicrobial activities [ 17 , 18 ]. In this context it has been shown that the TtgABC efflux pump of P. putida recognizes a wide set of flavonoids [ 19 ]. However, these results should also be interpreted from a second perspective, and not just as defense mechanisms required by rhizosphere colonizers in order to benefit from the nutrients released by the plant. Reactive oxygen species produced in root tissues have been implicated not only in stress but also in signaling processes in legume-rhizobia symbioses [ 20 ]. Such a potential role has recently started to be investigated in other mutualistic associations [ 21 ]. Plant-bacterial signaling may be reflected by another relevant group of genes induced in the rhizosphere, comprising signal transduction sensors and response regulators, as well as three transcriptional regulators of the AraC and TetR families. In this group is included the sensor histidine kinase PhoR, which participates in the global response to inorganic phosphate limitation. Autophosphorylation of phoR in Bacillus has been reported to be modulated by the redox state, so that terminal oxidases are required for the Pho system's full induction [ 22 ]. Interestingly, two genes predicted to be involved in aa 3 -type cytochrome c oxidase assembly (PP0109 and PP0110) are induced in the rhizosphere. In other microorganisms, the Pho regulon has been implicated, among other processes, in biofilm formation and includes genes for the synthesis of antibiotics and other secondary metabolites [ 23 - 25 ]. The importance of secondary metabolism in mutualistic plant- Pseudomonas interactions is being studied from the point of view of the plant [ 17 ] and of the microorganism [ 26 ]. In our study, PP3786, predicted to participate in the synthesis of an as yet undefined secondary metabolite, was induced in the rhizosphere. Finally, it is worth noting the expression of genes pointing towards potential DNA transfer and rearrangements that may take place in the rhizosphere. These include a helicase and the insertion sequence IS Ppu14 transposase, of which the six copies present in the genome of KT2440 were induced in the rhizosphere (five at a significant level; Additional data file 3). Role of rhizosphere up-regulated genes in colonization fitness To give an ecological significance to our results, we analyzed the role of several rup genes in competitive colonization and found that mutants in some rup genes are hampered in their survival in the rhizosphere in competition with the wild type, while they are indistinguishable from it under laboratory conditions. Nine transposon insertion mutants were chosen to test the relevance of rhizosphere expressed genes for the establishment of the Zea mays - P. putida association and microbial fitness in the rhizosphere. The mutants were selected to represent various classes of physiological roles (respiratory chain, transport, metabolism, stress adaptation, motility, transcriptional regulation and also hypothetical proteins). In some cases where the rhizosphere-induced gene is part of an operon, available mutants in genes included in the transcriptional unit were used. Competitive rhizosphere colonization assays were performed, and the proportion of each strain in the rhizosphere population was assessed after 12 days (Figure 3 ). In five cases, the wild type had displaced the mutant to a significant extent, so that the later represented less than 30% of the total population, supporting the idea that genes differentially expressed in the rhizosphere versus all the controls are important for bacterial fitness in this environment. The identification of KT2440 genetic determinants with a specific role in rhizosphere fitness constitutes a relevant result, since previously identified mutants of this strain hampered in colonization were also affected in growth under laboratory conditions [ 27 ] (our unpublished results). Most of the mutants identified here are also hampered in root colonization of the model plant A. thaliana (our unpublished results). Two of these mutants, PP0110 and PP1477, are also affected in adhesion to corn seeds (data not shown), indicating that their role is directly related to life on plant surfaces. The open reading frame encoding PP1477 is located 9 bp downstream to the rup gene encoding PP1476, so that polar mutations in PP1476 likely affect the expression of PP1477. The most noticeable result was obtained with mutant PP0906, which is affected in a putative multidrug efflux transporter, in agreement with the notion that the ability to cope with toxic compounds is one of the key traits for survival in the rhizosphere. However, this mutant was hampered in growth under laboratory conditions and was not further considered. Mutant PP3279 is affected in aromatic compound metabolism, specifically in CoA activation of phenylacetic acid [ 28 ], perhaps its reduced fitness being a consequence of its inability to remove toxic compounds from the exudates. Another mutant in the phenylacetyl-CoA pathway, PP3283, which forms part of the same functional unit as the rup gene ecoding PP3281 [ 28 ], was also affected in competitive colonization (not shown). The fifth mutant is affected in a type I secretion system of an exported protein with animal peroxidase and calcium binding domains. Two other mutants also showed reduced competitive colonization capacity, although to a lesser extent. These genes are nevertheless interesting. PP4959 codes for a response regulator containing a signal receiving domain and GGDEF/EAL domains, which have been implicated in regulating the transition from planktonic to sessile life styles through secondary messenger c-di-GMP levels [ 29 ]. Mutant PP4988 is affected in a sensor histidine kinase that forms part of a chemotaxis signal transduction operon (comprising loci PP4990 to PP4987), which in P. aeruginosa controls twitching motility mediated by type IV pili [ 30 ]. A role for type IV pili in biofilm formation as well as in attachment to legume roots has been reported [ 31 , 32 ], but this is the first indication that signal(s) present or absent in the root environment may trigger type IV pili functionality. Figure 3 Rhizosphere fitness of mutant strains in competition with KT2440RTn 7 -Sm. The knocked-out open reading frame in the mutant strains is indicated by the locus name. Proportion of mutant (grey) and wild type (white), which was 50% ± 2% in the initial inocula, is plotted after 12 days of colonization. Data are the averages and standard error for six plants. KT2440RTn 7 -Sm, a streptomycin resistant derivative of KT2440R (see Materials and methods), was used as the wild-type strain in the experiments. KT2440RTn 7 -Sm and KT2440R are equally competitive in root colonization (not shown). Sm resistance marker of the wild-type bacteria allowed their specific selection against the mutants, which were kanamycin resistant derivatives of KT2440R. Statistical analysis was carried out using SPSS software (version 12.0.1 for Windows, SPSS Inc., Chicago, IL, USA). The linear model univariate analysis of variance rendered significant differences for the mutants shown in the figure ( P value < 0.05) in comparison with the wild type. Seed adhesion rate was similar for mutants and KT2440R (0.5%), with the exception of PP0110 and PP1477 (0.1%). The growth of the mutants under laboratory conditions (rich and defined medium) was indistinguishable from that of KT2440RTn7-Sm. The transcriptional organization of mutated genes is shown in the bottom. The space between the 3' end of PP1476 and the 5' end of PP1477 is 7 bp. Translational coupling between PP3281 and PP3280 (8 bp) was observed. PP3279 is probably in an independent operon; however, PP3279 and PP3281 code for enzymes in the same degradative pathway [28]. Translational coupling between PP4987 and PP4988 was also observed (8 bp). Arrows indicate direction of transcription. Transposon insertion is indicated by inverted triangles. Transcriptional profiling in vitro [ 8 ] serves to pinpoint relevant components of exudates and how these influence bacterial physiology, but it obviates some of the conditions characteristic of the actual situation in the rhizosphere, in particular the association between bacterial cells and the plant root surface. A comparison of the data obtained here with that work shows limited overlap. Six genes identified in P. aeruginosa with increased expression in the presence of sugar beet exudates are homologs of P. putida genes induced in the corn rhizosphere, such as the helicase PP2565, or functionally related to them, like soxB (encoding sarcosine oxidase β-subunit). These are likely to reflect common characteristics of the root exudates of both plant species and/or compounds causing equivalent responses. With respect to genes previously identified in P. putida by in vivo expression technology, it is worth mentioning the PP1476/PP1477 operon, which as described above, is required for efficient colonization of seeds and roots. PP1476 encodes a homolog to E. coli YaeQ, which compensates for the loss of RfaH [ 33 ], a specialized transcription elongation protein. PP1477 corresponds to RecJ, an exonuclease involved in recombination and DNA repair after UV [ 34 ] or oxidative damage [ 35 ], again supporting the view of the rhizosphere as an environment where nutrient availability comes at an extensive cost in terms of the battery of protection mechanisms that have to be kept active. This work opens a challenging perspective to the study of mutualistic plant-microbe associations where, besides other determinants, energetic balances should be taken into account as part of the factors that define the success of these cross-kingdom interactions."
} | 6,799 |
29416559 | PMC5784704 | pmc | 9,295 | {
"abstract": "Background Styrene is a versatile commodity petrochemical used as a monomer building-block for the synthesis of many useful polymers. Although achievements have been made on styrene biosynthesis in microorganisms, several bottleneck problems limit factors for further improvement in styrene production. Results A two-step styrene biosynthesis pathway was developed and introduced into Escherichia coli BL21(DE3). Systematic optimization of styrene biosynthesis, such as enzyme screening, codon and plasmid optimization, metabolic flow balance, and in situ fermentation was performed. Candidate isoenzymes of the rate-limiting enzyme phenylalanine ammonia lyase (PAL) were screened from Arabidopsis thaliana (AtPAL2), Fagopyrum tataricum (FtPAL), Petroselinum crispum (PcPAL), and Artemisia annua (AaPAL). After codon optimization, AtPAL2 was found to be the most effective one, and the engineered strain was able to produce 55 mg/L styrene. Subsequently, plasmid optimization was performed, which improved styrene production to 103 mg/L. In addition, two upstream shikimate pathway genes, aroF and pheA , were overexpressed in the engineered strain, which resulted in styrene production of 210 mg/L. Subsequently, combined overexpression of tktA and ppsA increased styrene production to 275 mg/L. Finally, in situ product removal was used to ease the burden of end-product toxicity. By using isopropyl myristate as a solvent, styrene production reached a final titer of 350 mg/L after 48 h of shake-flask fermentation, representing a 636% improvement, which compared with that achieved in the original strain. Conclusions This present study achieved the highest titer of de novo production of styrene in E. coli at shake-flask fermentation level. These results obtained provided new insights for the development of microbial production of styrene in a sustainable and environment friendly manner. Electronic supplementary material The online version of this article (10.1186/s13068-018-1017-z) contains supplementary material, which is available to authorized users.",
"conclusion": "Conclusion In this study, a series of rational and systematic optimizations of styrene biosynthesis was described (Table 1 ). The PAL gene sequences from A. thaliana , P. crispum , F. tataricum , and A. annua were collected, optimized, and synthesized; these genes were based on the preferred codon usage of E. coli . The engineered strain E. coli BL01 produced 55 mg/L styrene from glucose in shake-flask cultures. Subsequently, the plasmids were optimized by evaluating the effects of copy numbers and promoters on styrene production. E. coli BL05 harboring plasmid pTrc- AtPAL2 - FDC1 strain produced 103 mg/L styrene in shake-flask fermentation, which represented a 187% increase in styrene production, when compared with that produced by E. coli BL01 harboring plasmid pACYC- AtPAL2 - FDC1 . Then, the upstream shikimate pathways genes aroF and pheA were overexpressed in E. coli BL0501 strain, which led to a final styrene production at titers of 210 mg/L in shake-flask fermentation, which represents a 203% improvement in styrene production, when compared with that achieved in E. coli BL05 strain. Moreover, when combined with the overexpression of central metabolic pathway genes tktA and ppsA , the engineered E. coli BL0801 produced 275 mg/L styrene. Fermentation time optimization and toxicity assays confirmed that the endogenously synthesized styrene had inhibitory effects on cell growth. To alleviate the toxic effects of styrene on the host cells, ISPR was applied. When isopropyl myristate was used as the solvent, the styrene titers reached 350 mg/L after 48 h of shake-flask fermentation, representing 127 and 636% improvements in styrene production, when compared with that achieved in cultures without solvent (275 mg/L) and the original E. coli BL01, respectively. To the best of our knowledge, this is the highest styrene titer produced by de novo production of styrene in E. coli in shake-flask cultures. Table 1 Results of the systematic optimization of styrene biosynthesis in E. coli BL21(DE3) Step Optimization strategy Strain Plasmid Yield (mg/L) Improvement over previous step (%) Improvement over original step (%) 1 Enzyme screening and codon optimization E. coli BL01 pACYC- AtPAL - FDC1 55 – – 2 Plasmid optimization E. coli BL05 pTrc- AtPAL - FDC 1 103 187 187 3 Co-expression of aroF and pheA E. coli BL0501 pTrc- AtPAL - FDC1 and pACYC- aroF - pheA 210 203 382 4 Co-expression of aroF , pheA , ppsA , and tktA E. coli BL0801 pTrc- AtPAL - FDC1 - ppsA - tktA and pACYC- aroF - pheA 275 131 500 5 Expression time optimization E. coli BL0801 pTrc- AtPAL - FDC1 - ppsA - tktA and pACYC- aroF - pheA 275 100 500 6 In situ extraction E. coli BL0801 pTrc- AtPAL - FDC1 - ppsA - tktA and pACYC- aroF - pheA - tktA - ppsA 350 127 636",
"discussion": "Results and discussion Screening PAL isoenzymes and codon optimization In a previous study, Nielsen et al. confirmed that PAL was the rate-limiting enzyme in the styrene biosynthesis because the tCA titers remained low throughout [ 11 ]. In this study, to screen for a more efficient PAL, three candidate isoenzymes from Petroselinum crispum (PcPAL), Fagopyrum tataricum (FcPAL), and Artemisia annua (AaPAL), were screened with AtPAL2 as a control. It has been reported that recombinant PcPAL exhibits activity towards its natural substrate l -Phe, with a Km value of 116 ± 4 and Kcat value of 1 ± 0.05, and might have better catalytic properties [ 26 ]. Both FtPAL and AaPAL come from medicinal and nutrient-rich plants with high levels of flavonoids, and recombinant FtPAL protein has been found to be specific to l -Phe, with an activity of up to 35.7 IU/g [ 27 , 28 ]. In the present study, to improve expression level, codon-optimized versions of all the above-mentioned genes were synthesized. FDC1 from Saccharomyces cerevisiae and optimized PALs were cloned into the pACYCDuet-1 vector and introduced into E. coli BL21(DE3), and the obtained transformants were inoculated into M9 medium for 24 h. As shown in Fig. 2 a, the expression of FDC1 with AtPAL2 led to production of 55 mg/L styrene, which was higher than that achieved with FtPAL (44 mg/L) or PcPAL (36 mg/L), respectively, whereas the strains containing AaPAL rarely produced styrene. Although no enzyme with higher activity was detected, both FtPAL and PcPAL were confirmed to be specific to L-Phe because gas chromatography–mass spectrometry (GC–MS) did not detect 4-Vinylphenol in the fermentation products (Additional file 1 : Figure S1). Based on these results, the expression of AtPAL2 combined with FDC1 was used for the production of styrene in subsequent experiments. Fig. 2 Effect of PALs and different plasmids on styrene production. a Styrene production with the expression of different PAL candidate genes from P. crispum (PcPAL), F. tataricum (FcPAL), A. annua (AaPAL), and A. thaliana (AtPAL) in recombinant E. coli BL21(DE3). b Styrene production in recombinant E. coli BL21(DE3) harboring different plasmids (pTrcHis2B, pET-28a, pACYCDuet-1, and pColADuet-1) in recombinant E. coli BL21(DE3). Error bars represent one standard deviation from triplicate experiments Selection of a suitable plasmid for increasing styrene production Production of recombinant proteins in E. coli cells is affected by the number of plasmids, as well as their structural and segregational stability, which have essential impacts on productivity [ 29 ]. To achieve a high styrene producing capability, a two-step pathway was incorporated into four plasmids with different copy numbers and different promoters to assess any effect on styrene production. All the strains produced styrene from an initial glucose concentration of 15 g/L under aerobic conditions in 600-mL flasks. As shown in Fig. 2 b, the strains with high-copy-number plasmids achieved higher styrene titers. E. coli BL06 harboring high-copy-number plasmid pET-28a- AtPAL2 - FDC1 produced 70 mg/L styrene, which had a 123 and 507% improvement compared with that achieved in E. coli BL01 harboring the medium-copy-number plasmid pACYCDuet- AtPAL2 - FDC1 and E. coli BL07 harboring the low-copy-number plasmid pColADuet- AtPAL2 - FDC1 , respectively. In contrast, plasmid copy number had little effect on cell growth during the process of styrene fermentation, with OD 600 of all strains were around 2.0 (Fig. 2 b). In addition to the effect of plasmid copy number, the promoter also had a significant influence on styrene production. E. coli BL05 harboring trc promoter plasmid pTrcHis2B- AtPAL2 - FDC1 produced 103 mg/L styrene under the same conditions, which was a 146% higher than that produced by E. coli BL06 harboring T7 promoter plasmid pET-28a- AtPAL2 - FDC1 . Based on these results, E. coli BL05 harboring plasmid pTrcHis2B- AtPAL2 - FDC1 was chosen as the parent strain for further styrene production optimization. Effect of co-expression of upstream genes aroF and pheA on styrene production Aromatic amino acids such as l -Phe are naturally produced mainly from the shikimate pathway. The first rate-limiting step in this pathway is the condensation reaction between PEP and E4P to form 3-deoxy- d -arabino-heptulosonate 7-phosphate (DAHP). This reaction is catalyzed by three DAHPS isoenzymes encoded by three genes aroF , aroG , and aroH , respectively. The second rate-limiting step is the conversion of chorismate into phenylpyruvate, via prephenate, catalyzed by CM-PDT, which is encoded by pheA . In a previous study, Backman et al. utilized a genetically modified E. coli strain with pheA fbr and aroF fbr genes to improve the metabolic flux towards l -Phe biosynthesis and achieved 50 g/L l -Phe with a yield of 0.25 (mol l -Phe/mol glucose) [ 30 ], which was the highest l -Phe production reported thus far. In our study, to enhance upstream pathway flux, aroF and pheA , were overexpressed in E. coli BL05 and the resultant strain E. coli BL0501 was evaluated for its ability to produce styrene in shake flask. After 24 h of cultivation, the styrene titers produced by E. coli BL0501 reached 210 mg/L, while the control strain E. coli BL0500 produced 100 mg/L styrene (Fig. 3 a). These results confirmed that aroF and pheA genes are the key factors determining the biosynthesis of endogenous l -Phe and co-expression of CM-PDT and DAHPS could significantly improve l -Phe and l -Phe derivatives produced by E. coli , similar to that reported in previous studies [ 31 , 32 ]. Fig. 3 Effects of overexpression of key upstream genes on styrene production. a Effect of overexpression of aroF and pheA genes on styrene production. E. coli BL0501 and BL0500 cells were cultivated and their cell growth (OD600) and styrene production titers were compared. b Effect of overexpression of aroF , pheA , ppsA , and tktA genes on styrene production. E. coli BL0801, BL08, and BL0500 cells were cultivated and their cell growth (OD600) and styrene production titers were compared. c Optimization of induction length. E. coli BL0801 cells were induced for 24, 36, and 48 h and their cell growth (OD600) and styrene production titers were compared. Error bars represent one standard deviation from triplicate experiments Effect of co-expression of central metabolic pathway genes tktA and ppsA on styrene production To produce one molecule of l -Phe, two molecules of PEP and one molecule of E4P, which are both involved in the central metabolic pathways, are required. PEP is predominantly utilized in the phosphotransferase system (PTS), which is responsible for the translocation and phosphorylation of glucose, converting one PEP molecule to pyruvate (Fig. 1 ) [ 33 ]. Enhancing the expression level of PEP synthase (encoded by ppsA ), which recycles pyruvate generated by PTS-mediated glucose transport to PEP, is an important approach for increasing the carbon flux from PEP to the aromatic amino acids pathway [ 33 ]. E4P can be directly produced by transketolase (encoded by tktA ) or transaldolase (encoded by talB ), and tktA have been demonstrated to be more effective in directing the carbon flux to the aromatic pathway than talB [ 34 ]. In our study, to investigate the effect of overexpression of ppsA and tktA on styrene production, styrene production of engineered E. coli strains harboring different constructs in shake-flask fermentation was investigated. As shown in Fig. 3 b, E. coli BL0801 (harboring pTrc- AtPAL2 - FDC1 - ppsA - tktA and pACYC- aroF - pheA ) accumulated 275 mg/L styrene after 24 h of fermentation with the consumption of 6.7 g/L glucose, which represents a 131 and 268% improvement over styrene production by E. coli BL08 (harboring pTrc- AtPAL2 - FDC1 and pACYC- aroF - pheA ) and E. coli BL05 (harboring pTrc- AtPAL2 - FDC1 ), respectively. When compared with the previous study, which overexpressed AtPAL2 and FDC1 in an L-Phe overproduction strain E. coli NST74 ( aroH 367, tyrR 366, tna -2, lacY 5, aroF 394(fbr), malT 384, pheA 101(fbr), pheO 352, aroG 397(fbr)) and the resulting strain was able to produce 260 mg/L styrene from 15 g/L glucose [ 11 ], it seemed that the styrene titer achieved in E. coli BL0801 was not improved as much as expected. Several reasons might be responsible for this result. Different genetic backgrounds between E. coli NST74 (K-12) and E. coli BL21(DE3) may result in different styrene yields. Furthermore, aroF fbr , aroG fbr , tyrR , pheA fbr , and pheO were overexpressed in E. coli NST74, while aroF wt , pheA wt , tktA , and ppsA were introduced into E. coli BL21(DE3). Multiple isozymes encoding genes, aroF and aroG , pheA , and pheO were co-expressed, which may significantly increase the flow of central metabolic carbon to phenylalanine biosynthesis. In the case, our engineered strain only had a slightly higher styrene production than the strain using E. coli NST74 as a host. Therefore, different host including E. coli NST74 (K-12) and more genes would be considered to increase the production of styrene in the subsequent research. During the fermentation, the growth of engineered E. coli strains was very slow, with OD 600 reaching 2.0 after 24 h of induction. Therefore, the effect of induction time on styrene production was examined. As shown in Fig. 2 c, styrene production and OD 600 did not increase along with the elongation of induction time, which could possibly be owing to the inhibition of cell growth by styrene produced by the host strains. This result reconfirmed that product toxicity is a limiting factor that must be addressed in addition to metabolic regulations. Styrene toxicity assay To evaluate styrene toxicity on E. coli BL21(DE3), the effect of exogenous addition of styrene at different concentrations (100, 200, 300, and 400 mg/L) on growing cultures was investigated. As shown in Fig. 4 a, an OD 600 of 3.4 was reached in the absence of styrene. When the concentration of styrene was less than 300 mg/L, no significant growth inhibition was observed. However, when styrene concentration was increased to more than 300 mg/L, a remarkable cell growth inhibitory effect to cell growth was detected at 0–10 h. Interestingly, after 10 h, the values of OD 600 started to increase, which could possibly be due to the adaptation of cells to cultivation conditions with styrene or evaporation of styrene due to the insolubility of styrene in water. These findings are consistent with the previously reported styrene toxicity thresholds for E . coli NST74 [ 11 ]. Fig. 4 Tolerance of E. coli BL21(DE3) cells to styrene toxicity and ISPR with different solvents. a Growth response of E. coli BL21(DE3) cells to 0, 100, 200, 300, and 400 mg/L styrene in LB medium. b Effect of different solvents on styrene production and cell growth. c Time course of biomass by E. coli BL0801 cells in biphasic and single-phase cultures. Error bars represent one standard deviation from triplicate experiments \n After systematic optimization, the styrene output achieved 275 mg/L in the present study, which was close to the inhibitory threshold (300 mg/L). However, for economically viable and sustainable production of microbial-derived renewable styrene, the styrene titers and productivity must be ultimately improved, in other words, styrene toxicity must be overcome or effectively circumvented [ 11 ]. The accumulation of de novo synthesized biofuels or other solvent-like compounds within the cytoplasmic membrane has been shown to disrupt membrane integrity, resulting in the leakage of ions, metabolites, lipids, and proteins, as well as affecting the cells ability to maintain its internal pH and an appropriate trans-membrane proton gradient [ 35 , 36 ]. Efforts have been made to improve host strains for better solvent tolerance, including introduction of efflux pumps or transporters, heat shock proteins, membrane modifications, genome engineering, random mutation, adaptive evolution, and approaches that integrated multiple-tolerance strategies [ 24 ]. In addition, ISPR and medium supplements can help to ease the burden of end-product toxicity and may be used in combination with genetic approaches. Therefore, in the present study, ISPR was used in combination with genetic approaches to increase styrene tolerance and production capacity of the host strains. In situ product removal (ISPR) and solvent selection For successful application of this approach, the selection of a suitable solvent is important. Ideal solvents should be biocompatible yet non-bioavailable and display high equilibrium partitioning of the target compound over water [ 37 ]. Various kinds of solvents, such as oleic acid, oleyl alcohol, miglyol, isopropyl myristate, and polypropylene glycol have been tested for their ability to improve the production of aromatic compounds [ 23 – 25 , 38 ]. In our study, the effect of oleyl alcohol, n -dodecane, and isopropyl myristate on cell growth and styrene production of the engineered strain E. coli BL0801 was investigated, with no-solvent culture as a control. After 48 h of shake-flask fermentation, in the presence of n -dodecane, oleyl alcohol, and isopropyl myristate, the styrene titers reached 304, 340, and 350 mg/L, representing 110, 124, 125% improvements, respectively, when compared with that achieved in single-phase cultures (275 mg/L) (Fig. 4 b). In addition, the cell growth curves of the E. coli BL0801 in biphasic and in single-phase cultures were also examined. The results revealed that the cell growth was significantly increased in biphasic culture when compared with that in single-phase culture (Fig. 4 c), it is presumed that isopropyl myristate could selectively remove the styrene from the reaction system, thereby maintaining the styrene concentration around the cells below the inhibitory threshold, and allowing the strains to continue styrene production, resulting in higher biphasic culture than that in single-phase culture. Furthermore, maximum theoretical yield coefficients max Y Phe /Glc were calculated from the known stoichiometry of l -Phe biosynthesis from glucose, in an engineered strain where either the PTS was inactive or PYR was being recycled back to PEP, and the maximum theoretical yield( eng. max Y Phe /Glc ) was 0.55 g/g [ 20 , 39 , 40 ]. Moreover, based on the hypothesis that complete conversion of all endogenously produced l -Phe to styrene is possible (e.g., if the pathway engineered in the present study could achieve a particularly high flux), the maximum theoretical yield ( eng. max Y styrene /Glc ) was calculated to be 0.35 g/g. According to this value, in single-phase culture, E. coli BL0801 strains reached yields 0.041 g/g, corresponding to 12% of the eng. max Y styrene /Glc , while in biphasic culture, E. coli BL0801 strains reached yields 0.048 g/g, corresponding to 14% of the eng. max Y styrene /Glc . The above data indicate that it is possible to achieve higher yield efficiency. Major challenges may come from low enzymatic activity and flux imbalance. In the present study, methods to improve activity of PAL, which is a rate-limiting enzyme in the styrene biosynthesis, were first considered. Significant achievements were made by enzyme engineering, such as screening of enzymes with high activity and specificity [ 11 , 41 , 42 ], mutating the coding sequence in the regulatory domain [ 12 , 43 , 44 ], and family shuffling recombines natural proteins with high sequence identity [ 45 – 47 ]. Currently, these strategies have not yet been applied to attain PAL with higher activity. Feedback inhibition is one of the fundamental mechanisms that regulates the synthesis of amino acids and avoids their excessive accumulation which may cause imbalanced metabolism [ 48 , 49 ]. To improve the metabolic flux towards l -Phe biosynthesis, overexpression of feedback-resistant pheA ( pheA fbr ) and aroF ( aroF fbr ) genes may be effective strategies. However, mutant enzymes may decrease thermostability and catalytic efficiency. For example, the use of the aroF wt produced much more l -Phe than aroF fbr (Asn8-Lys) due to the decreasing thermostability of aroF fbr [ 50 , 51 ]. To overcome flux imbalances, rational strategies to regulate gene expression were developed, such as application of inducible promoters, use of non-native RNA polymerase [ 52 ], the replacement of the ribosome binding site [ 53 ], as well as multivariate modular metabolic engineering [ 54 ]. Recently, biosensors have been employed to regulate metabolic flux. Biosensor application is an effective strategy that could dynamically detect pathway flux or the levels of pathway intermediates or products and regulators that respond to sensor input and accordingly regulate enzyme expression [ 55 ]. In a report, transcription factor (TF)-based sensor, a mutated transcriptional activator NahR from Pseudomonas putida , was used to detect benzoate and 2-hydroxybenzoate accumulation in E. coli [ 56 ]. In another report, researchers utilized a lysine riboswitch (RNA-based) to regulate the expression of citrate synthase and control the metabolic flux of the tricarboxylic acid cycle in a lysine-producing strain Corynebacterium glutamicum LP917, which increased the lysine production by 63% [ 57 ]. However, most applications have been limited to natural sensor-regulators [ 55 ]. Modular scaffold strategies are also effective approaches to improve metabolic flux. A scaffold protein carrying multiple protein–protein interaction domains is used to co-localize sequential pathway enzymes that have been tagged with peptide ligands specific for the domains on the scaffold [ 58 ]. The combined use of these multi-functional enzymes might increase the yield and titer of aromatic compounds from glucose. However, a major drawback of this method is that it often results in decreased activity of one enzyme or both the enzymes [ 59 ]. Based on the above reason, our subsequent work would focus on finding feedback-resistant pheA ( pheA fbr ) and aroF ( aroF fbr ) genes with improving thermostability and catalytic efficiency. Furthermore, screening an efficient method to overcome flux imbalances would lay the foundation for industrialized production of styrene."
} | 5,869 |
39417445 | PMC11579442 | pmc | 9,297 | {
"abstract": "Summary \n Forest soils play a pivotal role as global carbon (C) sinks, where the dynamics of soil organic matter (SOM) are significantly influenced by ectomycorrhizal (ECM) fungi. While correlations between ECM fungal community composition and soil C storage have been documented, the underlying mechanisms behind this remain unclear. Here, we conducted controlled experiments using pure cultures growing on naturally complex SOM extracts to test how ECM fungi regulate soil C and nitrogen (N) dynamics in response to varying inorganic N availability, in both monoculture and mixed culture conditions. ECM species dominant in N‐poor soils exhibited superior SOM decay capabilities compared with those prevalent in N‐rich soils. Inorganic N addition alleviated N limitation for ECM species but exacerbated their C limitation, reflected by reduced N compound decomposition and increased C compound decomposition. In mixed cultures without inorganic N supplementation, ECM species with greater SOM decomposition potential facilitated the persistence of less proficient SOM decomposers. Regardless of inorganic N availability, ECM species in mixed cultures demonstrated a preference for C over N, intensifying relatively labile C compound decomposition. This study highlights the complex interactions between ECM species, their nutritional requirements, the nutritional environment of their habitat, and their role in modifying SOM.",
"introduction": "Introduction Forests constitute a significant reservoir of carbon (C), the majority of which is stored belowground, primarily in the form of soil organic matter (SOM) (Pan et al ., 2011 ; Schmidt et al ., 2011 ). The decomposition of SOM in forests is integral to the global cycling of C and nitrogen (N), underpinning diverse and critical forest ecosystem services such as climate regulation, biomass production and habitat provision for forest species (Deluca & Boisvenue, 2012 ). Within temperate and boreal forests, evidence increasingly suggests that ectomycorrhizal (ECM) fungi are involved in the decomposition of SOM (Phillips et al ., 2014 ; Lindahl et al ., 2021 ) mainly to capture and immobilize N into their tissues, which they can then exchange with their plant hosts for photosynthetically derived C (Lindahl & Tunlid, 2015 ; Baldrian, 2017 ). However, our understanding of how SOM decomposition differs across ECM fungal species and environmental contexts is in its infancy. These fundamental gaps pose challenges to the refinement of strategies aimed at optimizing C sequestration within the context of climate change. ECM fungi originate from multiple phylogenetic groups and their ability to decompose SOM exhibits considerable variation across evolutionary lineages (Kohler et al ., 2015 ; Pellitier & Zak, 2018 ). For example, Amanita muscaria , which evolved within a clade of brown rot saprotrophs, has undergone a genetic loss resulting in a reduced capacity for decomposing SOM (Kohler et al ., 2015 ). By contrast, Hebeloma cylindrosporum , descended from a white‐rot ancestor that used class II fungal peroxidases to oxidize SOM, has retained three manganese peroxidase genes for SOM decomposition (Kohler et al ., 2015 ). Furthermore, the genome of Cortinarius glaucopus contains 11 peroxidases, a number comparable to that observed in numerous white‐rot wood decomposers, underscoring their likely significant contribution to the decomposition of SOM within forest ecosystems (Bödeker et al ., 2009 ; Miyauchi et al ., 2020 ). Given the inherent functional heterogeneity of ECM fungi, shifts in their community composition are likely to drive distinct and profound effects on C and N cycling within forest ecosystems (Sterkenburg et al ., 2018 ; Lindahl et al ., 2021 ). An important driver of ECM fungal community composition is the availability of inorganic N (Zak et al ., 2019 ), which can also act as a regulator of ECM‐mediated SOM decomposition (Bogar et al ., 2021 ; Argiroff et al ., 2022 ). Recent findings demonstrated that ECM fungal communities thriving in environments characterized by limited inorganic N content manifest an elevated genomic capacity for SOM decomposition (Mayer et al ., 2023 ). These communities are often characterized by the prevalence of genera such as Cortinarius and Hebeloma (Pellitier & Zak, 2021 ). By contrast, ECM communities in soils with high inorganic N concentrations are typically dominated by genera such as Scleroderma and Russula , which have a weaker capacity for SOM decay (van der Linde et al ., 2018 ). Other studies have revealed significant positive correlations between lignin‐derived SOM and soil C content with inorganic N availability (Argiroff et al ., 2022 ). This association is attributed to the presence of ECM fungi equipped with peroxidase enzymes, which exhibit diminished occurrence with increasing inorganic N availability (Clemmensen et al ., 2015 ; Argiroff et al ., 2022 ). Interactions between soil N availability, ECM fungal community composition, and soil C sequestration have been demonstrated within natural forest ecosystems, but the intricate mechanisms underpinning these relationships remain unresolved. Along with variations in soil chemistry, interspecific interactions among ECM fungi exert significant influence on the structure of entire ECM communities, consequently impacting SOM dynamics (Kennedy, 2010 ; Fernandez & Kennedy, 2016 ). Studies have demonstrated that competition for N resources between ECM fungi and free‐living decomposers can slowdown overall soil C cycling and increase soil C storage (Averill & Hawkes, 2016 ; Fernandez et al ., 2020 ). However, how interactions between ECM fungal species might also alter rates of C cycling remains unclear, even though interspecific competition for resources within this group has been widely demonstrated (Koide et al ., 2005 ; Kennedy, 2010 ; Smith et al ., 2023 ) and is recognized as a key determinant shaping their community composition (Kennedy, 2010 ) and structure (Pickles et al ., 2012 ). Interspecific interactions between ECM species may lead to similar inhibition, or alternatively could enable facilitation, whereby those species possessing more powerful decomposition strategies free up nutrients from recalcitrant soil compounds, enabling poorer decomposers to persist, in turn accelerating soil C cycling (Tiunov & Scheu, 2005 ; Lindahl & Tunlid, 2015 ). The lack of studies employing natural composite SOM extracts in competition experiments involving ECM fungi contributes to the uncertainty surrounding the impact of species interactions on SOM decomposition. Here, we conducted controlled pure culture experiments to address the knowledge gaps surrounding the mechanisms and context‐dependency of SOM decomposition by ECM fungi, strengthening our understanding of C and N cycling in forest ecosystems. We identified and cultured ECM species that thrived under conditions of low inorganic N availability and that were expected to demonstrate enhanced capacity to decompose SOM, alongside ECM species typically from soils with high inorganic N availability. We used these isolates to explore the context‐dependency of SOM decomposition, testing the hypothesis that an increase in inorganic N availability would lead to reduced ECM fungal decomposition of N compounds, while potentially enhancing their decomposition of C compounds as a regulatory mechanism to offset the C and N imbalance induced by inorganic N supplementation. Additionally, we tested the hypothesis that interspecific interactions might negatively affect ECM fungal growth, with the magnitude of this impact varying by fungal identity and environmental contexts, consequently shaping SOM decomposition patterns. Analyses of fungal growth and extracellular enzyme production were paired with pyrolysis gas chromatography mass spectrometry (Py‐GC‐MS) to examine the SOM dynamics at the molecular level.",
"discussion": "Discussion Variability in SOM decomposition capabilities among ECM species Our results reveal significant variations in the ability of different ECM fungal species to decompose complex and ecologically realistic SOM extracts from a Scots pine forest. Overall, Hebeloma velutipes and Suillus variegatus demonstrated significant ability to decay SOM. By contrast, Amanita rubescens and Lactarius rufus showed a distinct preference for ammonia‐N utilization (Fig. S4 ), coupled with a limited capability for SOM decomposition. This is inherently linked to the morphological characteristics and evolutionary histories of these species (Kohler et al ., 2015 ; Pellitier & Zak, 2018 ). To contextualize these findings, it is important to note that the soil extracts employed in our study were collected from Swedish forests characterized by abundant SOM but with little inorganic N (Fig. S5 ). Hebeloma velutipes and Suillus variegatus were able to grow on SOM extracts, whereas Amanita rubescens and Lactarius rufus could not persist under these conditions (Fig. 1a,b ). Furthermore, our findings demonstrate that the extent of SOM extract modification was significantly greater for Hebeloma velutipes and Suillus variegatus compared with Amanita rubescens and Lactarius rufus (Fig. 1c,d ). Our analysis is not able to disentangle the direct effects of the ECM fungi on SOM decomposition vs the contribution of organic matter to the growth media from hyphal exudation. ECM fungi are known to release a range of molecules as hyphal exudates in pure culture, including organic acids (Sun et al ., 1999 ), notably oxalate (Johansson et al ., 2008 ). However, we emphasize that the molecules that changed most in our experimental systems (Table S1 ) were generally more recalcitrant and complex than those measured in previous characterizations of hyphal exudates, and therefore, we suggest that the modifications in SOM composition in our study more likely arise from decomposition of organic matter. In many pristine forest ecosystems, inorganic nutrients derived from both mineralization and atmospheric deposition are scarce, so nutrient availability may primarily be driven by changes in the concentrations of organic molecules, such as amino acids. Simple organic forms of N may modify competitive interactions because some ECM fungi can readily utilize N from amino acids (Hazard et al ., 2017 ) and even take the molecules up intact (Näsholm & Persson, 2001 ). Modification of the C : N ratio of the growth substrate as a result of C in organic forms of N may also alter species interactions (Fransson et al ., 2007 ). We predict that ECM fungi with higher SOM decomposition capacities, such as Hebeloma velutipes and Suillus variegatus , might become more dominant on the root systems of plants exposed to elevated organic N inputs, potentially displacing less efficient SOM‐decomposing ECM fungi. Inorganic N availability governs ECM fungal decomposition of SOM \n Increased inorganic N availability contributes to restructuring of ECM fungal communities, favouring species with diminished SOM decay potential over those with greater capabilities (Pellitier & Zak, 2021 ; Argiroff et al ., 2022 ; Mayer et al ., 2023 ). A consensus among these studies is that this ecological shift is largely attributed to alterations in host plant nutritional needs, whereby rising soil inorganic N prompts plants to rely less on ‘costly’ ECM fungi for SOM decomposition and prefer ‘low‐cost’ inorganic N (Franklin et al ., 2014 ; Bogar et al ., 2019 ). Our study introduces a critical dimension by highlighting the dynamic interplay between the nutritional demands of ECM fungi for their own growth and the nutritional environment within their habitats (Plett et al ., 2020 ). Our findings indicate that the decomposition of SOM by ECM species facilitates their growth predominantly in the absence of inorganic N, with this benefit diminishing upon the introduction of inorganic N (Fig. 2a ). This phenomenon is likely associated with the alteration in their mechanism for the decomposition of organic compounds subsequent to the introduction of inorganic N. For example, in pure culture conditions, we observed a correlation between increased ECM fungal biomass and N compound decomposition without inorganic N supplementation; however, this relationship was absent when inorganic N was added (Fig. 2b ). This observation highlights that inorganic N addition alleviates N limitation in ECM fungi themselves (Högberg et al ., 2021 ), which may be a critical factor contributing to the reduction in SOM decomposition following fertilization. This finding also provides valuable insights for interpreting field‐based observations, where it has been demonstrated that, under low‐N conditions, trees allocate more C to their roots (Högberg et al ., 2010 ; Corrêa et al ., 2011 ) but this does not appear to correspond to an increase in N gain from the fungus (Corrêa et al ., 2011 ; Näsholm et al ., 2013 ). Soil microorganisms, including ECM fungi, are considered to be C‐limited (Zak et al ., 2019 ). We found that in the absence of plant photosynthesis‐derived C supply, ECM species, particularly Hebeloma velutipes and Suillus variegatus , have the capacity to decompose organic C compounds to sustain their growth. This is evidenced by a significant correlation observed between increased ECM biomass and the decomposition of C compounds (Fig. 2c ). Notably, our results reveal that inorganic N addition enhances the decomposition of C compounds by ECM species, partly because ECM fungal growth becomes primarily limited by C availability when N supply is abundant (Högberg et al ., 2021 ). To maintain the balance between C and N, ECM species expend additional energy by releasing more enzymes for the decomposition of organic C compounds (Fig. S6 ). Given that the enzymatic breakdown of C compounds by ECM fungi is a resource‐intensive process, these fungi primarily depend on the photosynthates provided by their tree hosts in forest ecosystems (Zak et al ., 2019 ). In comparison to species that lack the capacity to decompose SOM, those capable of SOM decomposition typically exhibit greater C demands (Fig. S7 ), largely due to their extensive mycelial production (Clemmensen et al ., 2015 ) or greater production of extracellular enzymes. The shifts in dominance from ECM taxa typified by Suillus or Hebeloma to those represented by Amanita or Lactarius , as observed in the transition from N‐poor to N‐rich soil, may offer insights into field observations indicating that limited available inorganic N in soils is associated with increased plant investment of C into ECM fungi (Hasselquist et al ., 2012 ), as well as ECM fungi tending to produce less aboveground and belowground biomass in N‐polluted forests (Ekblad et al ., 2016 ; Lilleskov et al ., 2019 ). Interspecific interactions regulate ECM fungal decomposition of SOM \n In natural ecosystems, individual trees are typically colonized by a diverse array of ECM fungi, leading to complex interspecific interactions among these fungal species (Kennedy, 2010 ). Such interactions exert a significant influence on the spatial distribution of ECM fungi and play a crucial role in shaping the structure and functions of their communities (Kennedy, 2010 ). To investigate how interspecific interactions among ECM fungi affect SOM decomposition, we established experimental conditions that simulated both independence and interaction of ECM species. Our findings demonstrate that in scenarios where two ECM fungi with differing SOM decomposition capacities are present, the species with limited SOM decomposition potential (i.e. Amanita rubescens and Lactarius rufus ) are advantaged, particularly under conditions of inadequate inorganic N (Fig. 1b ). This advantage is presumably attributable to the small‐molecule substances released by Hebeloma velutipes and Suillus variegatus during SOM decomposition, which serve as resources utilized by the co‐cultured Amanita rubescens and Lactarius rufus to facilitate their growth. Additionally, our study reveals that ECM species exhibit a heightened demand for C relative to N under interactive conditions, particularly following the addition of inorganic N (Fig. 3 ). This C preference serves to support the extension of their hyphal networks and the generation of biomass (Bogar, 2023 ), enhancing their competitive advantage in colonizing soil resources and the root tips of their host plants within forest ecosystems (Kennedy, 2010 ). However, this elevated C demand, in the absence of photosynthetic C contributions, imposes some constraints on ECM species involved in SOM decomposition (Fig. 1b ). In response, these species adapt their strategies for degrading different organic C forms (Fig. 4a–d ). These adaptations can be attributed to changes in the extracellular enzymes released by these species under varying environmental conditions (Fig. 4e,f ), potentially aimed at optimizing energy expenditure during the decomposition process. Mechanisms underlying ECM fungal control of SOM dynamics and its contextual dependence We applied SEMs to elucidate the underlying mechanisms by which ECM fungi influence SOM dynamics in response to increasing inorganic N availability, considering both species‐independent and interactive conditions. Our analysis encompassed SOM dynamics, including N and C cycling. Overall, we found that with increased inorganic N availability, ECM fungi play a more important role in regulating C dynamics than N dynamics, particularly in interspecific interaction environments. Specifically, inorganic N addition significantly disrupts the C and N balance of ECM fungal habitats, intensifying the C limitation experienced by ECM species and escalated their demand for C (Ekblad et al ., 2016 ). Consequently, ECM species release more extracellular enzymes for the decomposition of soil organic C compounds to sustain their growth (Kennedy et al ., 2007 ). While these results reflect fungal behaviour under culture conditions, they raise questions about how ECM fungi acquire and utilize C when in symbiosis with a plant. Interestingly, the ability of some ECM fungi to grow in culture without a host plant may reflect important differences in their biology. These culturable fungi may possess traits that allow them to more flexibly source C, such as a greater ability to use SOM or stored C reserves when plant‐derived photosynthates are unavailable. By contrast, unculturable fungi could be more specialized for direct symbiotic interactions, depending heavily on their host for C acquisition (Tedersoo et al ., 2010 ). In symbiotic conditions, C allocation from the plant may differ, with plants potentially favouring fungi that deliver more N for less C investment (Bogar et al ., 2021 ). This effect could lead to shifts in ECM fungal biomass, species richness, and community composition in response to changes in photosynthate availability (Corrales et al ., 2017 ; Averill et al ., 2018 ). Such changes could alter the role of ECM fungi in SOM dynamics, especially if fungi turn to SOM as an alternative C source when plant‐derived C is limited (Talbot et al ., 2008 ). Our results suggest that the utilization of SOM by ECM fungi is closely linked to both fungal identity and environmental contexts, but further research is needed to explore how these dynamics are shaped by symbiotic associations with plants and whether culturable fungi exhibit different ecological strategies compared with their unculturable counterparts. Genome sequencing and gene expression studies on an increasingly large range of ECM fungal species will help unpick relationships between genetic potential for SOM decomposition and responsiveness to environmental conditions, but again expansion to include less easily culturable species is critical. In conclusion, our study indicates that ECM fungi significantly influence SOM dynamics, with their contributions intricately linked to environmental contexts. Increased inorganic N availability effectively alleviates N limitations in ECM fungi but exacerbates their C constraints. This is particularly notable in species with pronounced SOM degradation capabilities, leading to an acceleration in soil C cycling. Additionally, C limitation among ECM fungi is heightened in the presence of interspecific interactions. Such interactions between ECM species with varying degradation capacities confer advantages upon nondegrading species, as species proficient in SOM decomposition release small molecular organic substances during SOM degradation, particularly in the absence of inorganic N supplementation. However, this process prompts an increase in labile organic C compound decomposition by ECM species with significant SOM degradation abilities, aiding in energy conservation to a certain extent. These findings help shed light on the underlying mechanisms governing the observed correlations among ECM fungal community composition, soil C storage dynamics, and inorganic N availability."
} | 5,290 |
39823332 | PMC11740958 | pmc | 9,298 | {
"abstract": "Many bacteria live in polymeric fluids, such as mucus, environmental polysaccharides, and extracellular polymers in biofilms. However, laboratory studies typically focus on cells in polymer-free fluids. Here, we show that interactions with polymers shape a fundamental feature of bacterial life—how they proliferate in space in multicellular colonies. Using experiments, we find that when polymer is sufficiently concentrated, cells generically and reversibly form large serpentine “cables” as they proliferate. By combining experiments with biophysical theory and simulations, we demonstrate that this distinctive form of colony morphogenesis arises from an interplay between polymer-induced entropic attraction between neighboring cells and their hindered ability to diffusely separate from each other in a viscous polymer solution. Our work thus reveals a pivotal role of polymers in sculpting proliferating bacterial colonies, with implications for how they interact with hosts and with the natural environment, and uncovers quantitative principles governing colony morphogenesis in such complex environments.",
"introduction": "INTRODUCTION Many bacteria live in polymeric fluids, such as mucus that lines the airways, gut, and cervico-vaginal tract in the body ( 1 , 2 ), exopolymers in the ocean ( 3 ), and cell-secreted extracellular polymeric substances (EPSs) that encapsulate biofilms ( 4 ). However, laboratory studies of bacteria typically focus on cells in polymer-free fluids. As a result, despite their prevalence, how extracellular polymers influence bacterial behavior remains poorly understood. Recent work hints that interactions with polymers can dramatically alter how individual motile cells swim ( 5 – 8 ) and aggregate with other cells ( 9 – 13 ). Nevertheless, the possible influence of polymers on another fundamental characteristic of bacterial life—spatial proliferation in a multicellular colony ( 14 , 15 )—remains unknown. Colony morphology can be biologically important. For example, studies of bacteria proliferating on planar surfaces show that their resulting colonies can exhibit a variety of morphologies ( 14 , 16 – 25 ), which in turn can affect cell-cell signaling ( 26 ), genetic diversity ( 27 – 30 ), and colony resilience and susceptibility to external stressors ( 9 , 31 , 32 ). Hence, we ask: Do interactions with polymers sculpt the morphology of proliferating bacterial colonies? Here, we demonstrate that this is indeed the case, and we elucidate the underlying mechanisms. In nature, many bacteria are nonmotile or lose motility ( 33 – 56 ), but still continue to proliferate in colonies; indeed, the loss of flagella is often associated with pathogenesis and bacterial adaptation to diseased mucosal environments, such as in cystic fibrosis. Therefore, we focus on the spatial proliferation of rod-shaped, nonmotile mutant cells that lack flagella. By performing experiments with these bacteria in polymer solutions, we find that when polymer is sufficiently concentrated, cells form large-scale “cables” as they proliferate in a colony—in stark contrast to forming a random dispersion, as in the conventionally studied polymer-free case. This characteristic cable morphology arises independent of variations in cell type and polymer composition across three different species of bacteria and seven different polymers, including mucins, a key component of mucus in the body. By combining experiments, theoretical modeling, and agent-based simulations, we trace the origin of cable formation to an interplay between polymer-induced entropic attraction between neighboring cells and their hindered ability to diffusively separate from each other after growth and division in a viscous polymer solution. Our work thus reveals a pivotal role of polymers in shaping proliferating bacterial colonies and provides quantitative principles to predict and control these morphodynamics more broadly.",
"discussion": "DISCUSSION Despite their prevalence in natural habitats, little is known about how polymers influence one of the most fundamental aspects of bacterial life—their proliferation in colonies. By combining experiments, simulations, and theory, we have shown that extracellular polymers can shape bacterial colonies through purely physicochemical interactions, as broadly suggested by previous simulations ( 82 ). In particular, we found that nonmotile bacteria proliferating in sufficiently concentrated polymer solutions—both biological and synthetic—nematically align to form an intertwined network of long, serpentine, multicellular cables. This characteristic colony morphology arises due to the combined influence of polymer-induced entropic attraction holding cells together and the enhanced solution viscosity hindering cells from diffusively separating after dividing. Hence, this phenomenon arises generically and predictably across diverse bacterial species and polymer compositions; indeed, not only did our experiments directly demonstrate cable formation in gut mucus, but also our analysis (detailed in the Supplementary Materials) indicates that, more broadly, biological polymers like mucus have physicochemical properties that promote cable formation, as shown by the gray region of Fig. 6 . Our work thus uncovers quantitative principles governing the morphogenesis of bacterial colonies—and potentially other microbial systems—in their complex environments in the real world. It also opens up a new direction for research in soft matter physics: While polymer-induced entropic attraction is well studied for passive particulate systems, our work builds on previous simulations of biofilms ( 82 ) to highlight that fascinating new behaviors can emerge when the constituent particulates can also proliferate ( 15 ). Cable-like structures with smectic, i.e., layered, internal ordering have been observed in systems of passive nanorods under strong depletion attraction ( 83 ); however, in the bacterial cables reported here, while there appears to be a slight amount of such ordering (e.g., in Fig. 1D ), we have not observed appreciable smectic ordering over large scales—possibly because it is disrupted by the natural variability in cellular lengths within a cable. Whether such ordering may arise during cable growth under other conditions, such as those induced by nutrient starvation ( 84 ), will be interesting to investigate in the future. Another interesting direction for future research will be to explore the generality of cable formation in complex environments with nonpolymeric depletants, as described further in the Supplementary Materials. Colony morphologies reminiscent of cables—often called “chains” or “cords”—have been previously observed in a variety of bacterial systems. However, their origin is likely fundamentally different from the cables reported here; while cables arise generically due to entropic forces induced by polymers on cells, chains/cords are thought to form instead due to specific biochemical interactions or cellular processes. One prominent example arises when host-secreted antibodies cross-link proliferating cells of Salmonella enterica together ( 85 ). Another example is thought to arise for Mycobacterium tuberculosis due to hydrophobic interactions between mycomembrane lipids that strongly adhere cells together ( 86 ); cell surface hydrophobicity can similarly mediate aggregate formation in P. aeruginosa during pathogenesis ( 87 ). Similarly, under certain conditions, E. coli surface adhesins can adhere cells together in chains during biofilm formation ( 88 ). In all three of these cases, specific biochemical interactions prevent cells from separating after division—unlike in cable formation, which arises due to a fundamentally different physicochemical mechanism across a broad range of polymer chemistries. A useful direction for future work will be to investigate how cable formation may be altered by the added influence of such specific biochemical interactions. A final example is the chains of cells that arise during Bacillus subtilis biofilm formation; unlike our cables, in which cells fully divide and separate from each other during proliferation, these chains form because cells do not completely divide and separate, and are therefore inseparably retained end-to-end ( 89 – 91 ). Our work thus reveals a distinct, more general mechanism by which proliferating bacteria can form such long, multicellular cables upon exposure to polymers. Extensive work has focused on the ability of polymers to enhance the mobility of swimming bacteria ( 5 – 8 ). By contrast, their influence on nonmotile cells is understudied, despite the fact that many bacteria in natural polymeric environments (e.g., mucus) are nonmotile or lose motility ( 33 – 56 )—an important virulence factor that often correlates with pathogenesis and colonization/biofilm formation. Our study therefore focused on the proliferation of nonmotile cells. However, a natural extension of our work is to investigate how cellular motility may alter cable formation. Cables are not formed when we repeat our experiments with swimming E. coli (movies S11 and S12)—presumably because the hydrodynamic force generated by swimming cells exceeds the attractive depletion force induced by polymers, as supported by calculations in the Supplementary Materials ( 12 , 92 ). Moreover, in the case of mucus, additional biochemical interactions can further promote dispersal of swimming cells ( 93 ). Further investigating the interplay between cellular motility and polymer-induced cable formation will thus be an interesting direction for future research. Also, given that the entropic forces underlying cable formation depend sensitively on cell size and shape, another natural extension of our work is to investigate how cables form for cells of other shapes (e.g., curved) and in mixtures of different founder cell types. Our study also focused on cable formation by non–biofilm-forming strains of bacteria when exposed to exogenous polymers. Another interesting direction for future research is to investigate how cable formation may be altered in the case of biofilm formers; it may even be possible that the polymeric EPS secreted by the cells drives cable formation on its own, without requiring exposure to exogenous polymers. What are the biological implications of cable formation? This phenomenon could be beneficial to nonmotile bacteria by giving their colonies a way to extend outward and explore new environmental niches, including on surfaces or within host tissues ( 86 , 94 ). Proliferating in a cable could also help bacteria counter host immune responses against them, either mechanically by impeding phagocytosis ( 95 ) and compressing host cell structures ( 86 ) or geometrically by reducing the amount of cell surface that is exposed to the surroundings—compared to the case of freely dispersed cells—potentially helping to protect cells within a cable from antimicrobials ( 86 ). Alternatively, it could be that hosts secrete polymers, such as mucins, to force cells to proliferate in cables, potentially enhancing clonal extinction rates and pathogen clearance ( 85 ), in addition to specific biochemical interactions that down-regulate pathogenic genes ( 1 , 64 , 96 ). Such effects may also be induced by dietary polymers transiting through the gut ( 97 , 98 ), potentially providing an indirect link between host diet, the physiology of the gut microbiome, and host health. Moreover, by localizing cells together, cable formation could alter the dynamics and extent of infection by bacteriophages throughout a colony, as suggested by simulations ( 99 ). Building on our findings to investigate these possible consequences of cable formation will be an exciting direction for future research."
} | 2,949 |
30742725 | null | s2 | 9,299 | {
"abstract": "In Vibrio species, quorum sensing controls gene expression for numerous group behaviors, including bioluminescence production, biofilm formation, virulence factor secretion systems, and competence. The LuxR/HapR master quorum-sensing regulators activate expression of hundreds of genes in response to changes in population densities. The mechanism of transcription activation by these TetR-type transcription factors is unknown, though LuxR DNA binding sites that lie in close proximity to the -35 region of the promoter are required for activation at some promoters. Here, we show that Vibrio harveyi LuxR directly interacts with RNA polymerase to activate transcription of the luxCDABE bioluminescence genes. LuxR interacts with RNA polymerase in vitro and in vivo and specifically interacts with both the N- and C-terminal domains of the RNA polymerase α-subunit. Amino acid substitutions in the RNAP interaction domain on LuxR decrease interactions between LuxR and the α-subunit and result in defects in transcription activation of quorum-sensing genes in vivo. The RNAP-LuxR interaction domain is conserved in Vibrio cholerae HapR and is required for activation of the HapR-regulated gene hapA. Our findings support a model in which LuxR/HapR bind proximally to RNA polymerase to drive transcription initiation at a subset of quorum-sensing genes in Vibrio species."
} | 342 |
35423682 | PMC8693375 | pmc | 9,304 | {
"abstract": "The slippery liquid infused porous surface has developed into a potential technology to solve the problem of poor durability in corrosion resistance. Herein, a kind of slippery liquid infused porous surface is created on 7075 aluminum alloy by wire electrical discharge machining for corrosion resistant applications. The hardness of the constructed porous microstructure is similar to the aluminum alloy substrate material, which ensures the stability of the slippery liquid infused porous surface. The modification of low surface energy substance fluorosilane avoids the direct contact between corrosive liquid and porous surface, and improves the lyophobic performance of the porous microstructure surface. The corrosion resistance of the porous microstructure surface is enhanced by the injection of perfluorinated lubricating oil. The experimental results show that the created slippery liquid infused porous surface can display super-slippery properties and durable corrosion resistance. The average sliding velocity of a water droplet is 0.48 ± 0.05 mm s −1 at a sliding angle of 5°. The corrosion current density of the surface is 3.116 × 10 −6 A cm −2 , which is 2 orders of magnitude lower than that of the polished surface. And the impedance radius reaches 90 kΩ cm 2 , which is about 20 times that of the polished surface.",
"conclusion": "4. Conclusions A novel WEDM technology was used to construct high strength porous microstructures on 7075 aluminum alloy surfaces. Perfluorinated lubricating oil was injected into the surface after fluorosilane treatment to obtain the SLIPS. Evaluated by potentiodynamic polarization curve measurement, electrochemical impedance spectroscopy test and immersion experiment in 3.5 wt% NaCl solution, SLIPS has been proved to have better stable and durable corrosion resistance than other three surfaces PS, WEDMS and FWEDMS. The corrosion current density of the SLIPS is 2 orders of magnitude lower than that of the polished surface. And the impedance radius is about 20 times than that of the polished surface. The results of wettability analysis show that the prepared SLIPS can achieve super-slippery property, and the SA is less than 5°. Meanwhile, the self-healing ability of lubricating oil with fluid properties can significantly improve the surface defect induction and corrosion problems. Furthermore, the processing of planar and various complex shapes of cavity can be achieved, and WEDM has been proved to be effective in fabricating porous microstructure on metal substrates, which provides a new idea for fabricating porous microstructure with large area and low cost, and widens the potential application of the SLIPS.",
"introduction": "1. Introduction The high-strength aluminum alloy, including the 2XXX, 6XXX, and 7XXX series, is one of the most widely used non-ferrous metal materials in industry. 1 Because of its good mechanical and physical properties, it is widely used in the fields of aviation, aerospace, automobiles and mechanical manufacturing. 2–4 However, due to its active chemical properties and poor corrosion resistance in harsh environments, the application of aluminum alloys in corresponding fields is restricted. 5,6 Therefore, it is of great significance to improve the corrosion resistance and prolong the service life of aluminum alloy by surface treatment technology. Corrosion resistance technology mainly includes electrochemical protection, 7 surface nano crystallization protection 8 and preparation of a superhydrophobic surface. 9–12 Slippery liquid infused porous surfaces (SLIPS) use lubricating oil to replace the air in the microstructure gap on the surface of the substrate, with the help of capillary action to lock the lubricating oil in the microstructure and form an oil film on the substrate. The oil film blocks the contact between corrosive liquid and substrate, so as to achieve the purpose of corrosion resistance. In recent years, SLIPS has gradually become one new technology to replace the above corrosion resistance technologies. 13,14 The fluidity of lubricant provides the SLIPS with an ability to repair itself, which is an advantage of SLIPS similar to superhydrophobic surfaces, 15–17 the SLIPS has broad application prospects in many other fields including anti-icing, 18,19 self-cleaning 20 and bioengineering. 21 The SLIPS is usually developed based on superhydrophobic substrates and there are many excellent works in the field of superhydrophobic substrates. Dr Song 22 proposed a mold replication technology to realize the large-scale fabrication of superhydrophobic conical pillars with high mechanical strength Boinovich 23 considered the role of different mechanisms of corrosion protection, related to the superhydrophobic state of the surface. The corrosion current density of the superhydrophobic coatings contacted with 3 M solutions of KCl during 2 h can reached 4.0 × 10 −11 A cm −2 . Up to now, scholars have done a lot of research on the construction of surface porous microstructure required by SLIPS with various methods including chemical reaction, 24 spraying, 25,26 self-assembly, 27,28 electrochemical coating 29 and laser irradiation. 30 For example, Kim 29 reported a direct fabrication method of SLIPS by electrochemical coating and studied the effect of surface structure size and layer distribution on lubricant retention under high shear conditions by comparing the loss of lubricants, contact angle lag and sliding angle of water and ethanol droplets on different surfaces. Yeong 30 prepared a flexible superhydrophobic silica gel material by copying the micro-texture of laser-irradiated aluminum substrate to polydimethylsiloxane, by injecting silicone oil into the microstructures, the material achieved the anti-icing property. Although the preparation of porous microstructure has been made significant progress, the problems of high cost, low efficiency, poor stability and durability have not been solved. For example, the raw materials used in spraying method are expensive, some of which are organic and harmful to health. The biggest disadvantage of self-assembly technology is that the stability is slightly poor, and the structure is easy to be destroyed under the action of some solvents. The problem of preparing porous structure by chemical reaction is low efficiency. In this paper, the porous microstructure is constructed on the surface of 7075 aluminum alloy by a novel wire electrical discharge machining (WEDM). The principle of WEDM is to use high temperature melting and vaporizing materials produced by pulse discharge between tool electrode and workpiece electrode, which enables precision machining of porous microstructure on metal surface. 31 Due to the rapid cooling of coolant, a recast layer with porous microstructure is formed on the surface, and its hardness is generally greater than that of the substrate material. 32 This is beneficial to the stability of the porous microstructure required by SLIPS. Furthermore, the processing of planar and various complex shapes of cavity can be achieved by WEDM, it provides conditions for large-area, high-efficiency precision construction of porous microstructure on metal surface. After modification with a low surface energy substance fluorosilane, the prepared porous microstructure surface is injected with perfluorinated lubricating oil to obtain the SLIPS. The corrosion resistance of prepared sample surface treated by different methods is evaluated by potentiodynamic polarization curve measurement, electrochemical impedance spectroscopy test and immersion experiment in 3.5 wt% NaCl solution. The results indicate that the prepared SLIPS provides stable and durable corrosion resistance.",
"discussion": "3. Results and discussion 3.1. Surface micromorphology and wetting performance analysis \n Fig. 3 shows the SEM images of aluminum alloy surface treated by WEDM with different cutting times. Fig. 3a and b are the micro morphology of the treated surface after first cutting. It can be seen that many craters with a diameter ranging from 60 to 100 μm were formed on the surface. The craters are surrounded by some porous structures due to the accumulation of spattered melt during WEDM. When the sample surface is cut for the second time, the number of craters on the aluminum alloy surface per unit area increases significantly due to the decrease of crater size, ranging from 20 to 60 μm (as shown in Fig. 3c ). The increase of the number of craters also results in the increase of the number of porous structures around the craters, as shown in the Fig. 3d . After cutting the sample for the third time (as shown in Fig. 3e and f ), the crater size continues to decrease (15–20 μm), while the number of crater and porous structure increases correspondingly. Following the forth cutting (as shown in Fig. 3g ), the size of craters reached the minimum, ranging from 5 to 15 μm, displaying the largest number of craters. At the same time, the number of the porous structure increases significantly (as shown in Fig. 3h ), which is very conducive to the storage of lubricating oil. Fig. 4 is the 3D morphology of the treated surface by WEDM with different cutting times. It can be observed that with the increase of cutting times, the size of surface crater decreases but the number increases, which also leads to the increase of the number of porous microstructure. Fig. 3 SEM images of aluminum alloy surface treated by WEDM with different cutting times. (a and b) First cutting. (c and d) Second cutting. (e and f) Third cutting. (g and h) Fourth cutting. (b, d, f and h) are the enlarged views of (a, c, e and g) respectively. Fig. 4 3D morphology of treated surface by WEDM with different cutting times. (a) First cutting. (b) Second cutting. (c) Third cutting. (d) Fourth cutting. \n Fig. 5 shows the wetting state of water droplets on the SLIPS. As shown in Fig. 5a , the CA of water droplet on the SLIPS is 110.65 ± 1.14°, exhibiting excellent hydrophobicity. Sliding process of water droplet on the SLIPS was also investigated in this paper as shown in Fig. 5b1 and b2 . The average sliding velocity of water droplet (10 μL) is 0.48 ± 0.05 mm s −1 at a sliding angle (SA) of 5°. In addition, the sliding state of the dyed acid droplet (pH = 3), alkali droplet (pH = 13), salt droplet (3.5% NaCl solution) and water droplet on the SLIPS are shown in Fig. 5c–f . It can be seen that the droplets can easily slide on the surface, although the surface tilt angle is very small, it still shows a smooth sliding. It indicates that the WEDM aluminum alloy surface can be injected lubricating oil to achieve super-slippery property. Fig. 5 Wetting state of water droplets on SLIPS. (a) CA of water droplet on SLIPS. (b1) and (b2) Sliding state of water droplet on SLIPS. (c–f) Motion results of different kinds of droplets on SLIPS inclined at an angle of 10 degrees: (c) acid droplet, PH = 3. (d) Alkali droplet, PH = 13. (e) Salt droplet, 3.5% NaCl solution. (f) Water droplet. 3.2. Corrosion current density, corrosion potential and impedance analysis \n Fig. 6 illustrates the polarization curves of the PS, WEDMS, FWEDMS and SLIPS after immersion in 3.5 wt% NaCl solution. Table 2 shows the corrosion potential ( E corr ), corrosion current ( I corr ) and corrosion protection efficiency (IE) of the sample surface with different treatment methods. Compared to PS, the corrosion current density of the WEDMS decreases slightly from 1.119 × 10 −4 A cm −2 to 7.175 × 10 −5 A cm −2 . After modified by fluorosilane, the corrosion current density of the WEDMS further reduces to 5.606 × 10 −5 A cm −2 . The SLIPS has the smallest corrosion current density of 3.116 × 10 −6 A cm −2 , decreased by about two orders of magnitude compared with the PS. Additionally, the corrosion potential of the WEDMS, FWEDMS and SLIPS shifts to the right relative to the PS, and the offset of the SLIPS is the largest, increased by 0.1639 V. Fig. 6 Polarization curves of the PS, WEDMS, FWEDMS and SLIPS after immersion in 3.5 wt% NaCl solution. Corrosion potential ( E corr ), corrosion current ( I corr ) and corrosion protection efficiency (IE) of aluminum alloy sample surface with different treatment methods after immersion in 3.5 wt% NaCl solution Samples \n E \n corr (V) \n I \n corr (A cm −2 ) IE PS −1.1710 1.119 × 10 −4 — WEDMS −1.1304 7.175 × 10 −5 35.88% FWEDMS −1.1593 5.606 × 10 −5 49.90% SLIPS −1.0071 3.116 × 10 −6 97.22% The increase of corrosion potential and the decrease of corrosion current density indicate that the SLIPS effectively inhibits the anodic solubility of the aluminum alloy. This is similar to the results of Jeong 33 and Dong 34 discussed in reducing the corrosion current density by using electrochemical anodizing technique and sand peening method. The oil film formed by the injection of lubricating oil plays an important role in preventing the direct contact between the chloride ions and the surface of the sample, thereby improving the corrosion resistance of the aluminum alloy surface in a corrosive environment. The corrosion protection efficiency of aluminum alloy sample surface treated by different methods can be calculated by the corrosion current density of the surface. The equation is as follows: 1 where, I 0 is the corrosion current density of the PS and I is the corrosion current density of the WEDMS, FWEDMS and SLIPS. According to the equation, the corrosion protection efficiency of WEDMS, FWEDMS and SLIPS is 35.88%, 49.90% and 97.22%, respectively, as shown in Table 2 . \n Fig. 7 presents the Nyquist plots of the PS, WEDMS, FWEDMS and SLIPS. The PS has an impedance radius of approximately 4 kΩ cm 2 . Compared to the PS, the impedance of the WEDMS is increased, and the impedance radius is about 2 times higher than that of the PS. The reason is that the air filled in the porous microstructures of the WEDMS forms an air layer, which prevents corrosive chloride ions from coming into contact with the metal substrate. But this kind of air layer is unstable and will be destroyed under certain conditions, such as high liquid pressure. In the process of WEDM, as the cutting fluid is oil organic matter, a large amount of free carbon (the weight percentage of carbon element increases 14.9% by WEDM) is produced by decomposition at high temperature, which is attached to the surface layer of substrate after cooling. So the carbon layer also plays an important role in the corrosion resistance of the material. In addition, the improvement of corrosion resistance of WEDMS compared to PS may be related to the forming microstructure due to the re-melting of the surface during the WEDM. For the FWEDMS, the impedance further increases, and the impedance radius reaches 25 kΩ cm 2 . It can be explained that the hydrophobicity of the surface is enhanced after modification, approximating to superhydrophobicity (CA is about 149°), and a more stable air layer is formed between the metal substrate and the corrosion solution. The impedance of the SLIPS is the largest, and the impedance radius reaches 90 kΩ cm 2 , which is about 20 times than that of the PS. This high impedance characteristic is due to the fact that the liquid lubricanting oil completely covers the entire sample surface and the formation of stable oil film inhibits the electron transfer between the substrate and the corrosion solution. Fig. 7 Nyquist plots of aluminum alloy samples surface treated by different methods. 3.3. Analysis of corrosion resistance based on salt solution immersion experiment \n Fig. 8 is the optical image of aluminum alloy samples surface treated by different methods before and after immersion in 3.5 wt% NaCl solution for 21 days (from left to right are PS, WEDMS, FWEDMS and FLIPS). As shown in the first line, the PS, WEDMS and FWEDMS are initially smooth and clean, while the SLIPS is covered with an oil film before immersion. It can be observed that a layer of loose white substance appears on the PS after 21 days of immersion, which are salty contaminants and corrosion products deposited on the surface of aluminum alloy samples. We can see there are fewer loose white substance on the WEDMS than that of the PS, and the FWEDMS and SLIPS are as clean as before immersion. Salt solution immersion experiments show that both the FWEDMS and SLIPS provide strong corrosion resistance. This also confirms the previous results of potentiodynamic polarization curve measurement and electrochemical impedance spectroscopy test, SLIPS has stronger corrosion resistance. Fig. 8 Optical image of aluminum alloy samples surface treated by different methods before and after immersion in 3.5 wt% NaCl solution for 21 days. From left to right are PS, WEDMS, FWEDMS and SLIPS. The first line is the sample surface before immersion and the second line is the sample surface after immersion. \n Fig. 9 shows the SEM image of aluminum alloy surface treated by different methods after 21 days of immersion in 3.5 wt% NaCl solution. A large number of corrosion voids are observed on the PS, with a diameter of about 10 μm ( Fig. 9 , SEM images of the PS), that is, pitting of aluminum alloy. The minimum number of corrosion voids are found on the SLIPS ( Fig. 9 , SEM images of the SLIPS), and the results indicate that the SLIPS has good corrosion resistance in the corrosion environment of 3.5 wt% NaCl solution. Fig. 9 SEM images of aluminum alloy surface treated by different methods after 21 days of immersion in 3.5 wt% NaCl solution. The first line are the SEM images of the PS at lower and higher magnification. The second line are the SEM images of the SLIPS at lower and higher magnification. In order to further study the corrosion state of different surfaces in salt solution, the four groups of samples after immersion were taken out from the salt solution, and then cleaned in alcohol and deionized water by ultrasonic vibration in turn, the surface elements were observed using EDS after drying. Table 3 shows the changes of Al, O, F, Na and Cl elements on the PS, WEDMS, FWEDMS and SLIPS before and after immersion in 3.5 wt% NaCl solution for 21 days. It is found that the weight percentage of Al element on PS decreases from 98.9% to 92.1%, while O element increases from 1.1% to 5.2% after immersion. The corrosion of aluminum alloy makes Al 3+ ions enter into the solution to reduce the weight percentage of Al element. Meanwhile, part of the generated Al 3+ ions diffuse outside the pitting, usually forming an aluminum hydroxide layer which increases the weight percentage of O element on the sample surface (see Section 3.4 for details). Furthermore, Na and Cl elements on the surface are observed, indicating that corrosion occurred on the PS in NaCl solution. A small amount of Na and Cl elements are observed on the WEDMS and FWEDMS after 21 days of immersion in NaCl solution, and the weight percentage of O element increased compared with that before immersion (WEDMS: from 15.1% to 18.6%, FWEDMS: from 16.1% to 16.4%). The increase of weight percentage O element on FWEDMS is less than that on WEDMS. This means that the corrosion degree of the WEDMS is lower than that of the PS, and the corrosion degree of the FWEDMS is the lowest among them. For the SLIPS, Na and Cl elements are not detected on the surface after 21 days of immersion in NaCl solution, indicating that the surface exhibited strong corrosion resistance to the NaCl solution. On the other hand, F element is observed on the FWEDMS and SLIPS, with weight percentage of 1.1% and 1.2%, respectively. It is found that there is no significant change in the weight percentage of F element before and after immersion, indicating that the low surface energy film is not corroded in NaCl solution. It is further shown that the FWEDMS and SLIPS have good corrosion resistance stability in 3.5 wt% NaCl solution. Changes of elements Al, O, F, Na and Cl on aluminum alloy surface treated by different methods before and after immersion in 3.5 wt% NaCl solution for 21 days Sample types Elemental composition and content (wt%) Al O F Na Cl Day 0 Day 21 Day 0 Day 21 Day 0 Day 21 Day 0 Day 21 Day 0 Day 21 PS 98.9 92.1 1.1 5.2 — — — 0.9 — 1.8 WEDMS 84.9 81.5 15.1 18.6 — — — 0.3 — 0.6 FWEDMS 82.7 82.6 16.1 16.4 1.2 1.1 — 0.3 — 0.5 SLIPS 82.7 82.7 16.1 16.1 1.2 1.2 — — — — The wettability of SLIPS was tested to further study the corrosion resistance stability and durability of the SLIPS in the salt solution during the immersion process. Fig. 10 shows the wettability of SLIPS immersed in 3.5 wt% NaCl solution for different days. It can be seen that the CA of water droplets on the SLIPS without immersion is 110.35°. With the increase of immersion days, the CA has a downward trend. In the first three days of immersion, the CA decreases from 110.25° to 108.12°, demonstrating an obvious downward trend. This may be related to the desaturation of lubricating oil because excessive oil will saturate the porous surface. During the 3–21 days of immersion, the CA varied from 108.13–105.47° with a small change ( Fig. 10a ). However, the SA increases gradually with increasing immersion days. In Fig. 11b , it can be seen that the SA of water droplets on the surface without immersion is about 5°, after 21 days of immersion in NaCl solution, the SA increases to about 20°, but it still shows good hydrophobicity. The reason for the stable wettability of the SLIPS may be that the lubricating oil film becomes thinner but not completely destroyed during immersion in NaCl corrosion solution. Therefore, the SLIPS can achieve good corrosion resistance stability and durability in 3.5 wt% NaCl solution environment. Fig. 10 The wettability of SLIPS immersed in 3.5 wt% NaCl solution for different days. (a) CA. (b) SA. Fig. 11 The schematic diagram of corrosion resistance mechanism of solid–liquid two-phase contact interface between aluminum alloy surface and 3.5 wt% NaCl solution under different treatment conditions (a) FWEDMS, with air layer. (b) SLIPS, with oil layer. 3.4. Surface corrosion characteristics and corrosion resistance mechanism In fact, the surface of the aluminum alloy will undergo the following corrosion reaction in the NaCl solution. When the NaCl solution is in contact with the aluminum alloy substrate, the reaction formula is as follows: 2 Then, the boehmite produced in the previous step will be dissolved by chloride ions. The reaction formulas are as follows: 3 Al(OH) 3 + Cl − = Al(OH) 2 Cl + OH − 4 Al(OH) 2 Cl + Cl − = Al(OH)Cl 2 + OH − 5 Al(OH)Cl 2 + Cl − = AlCl 3 + OH − Meanwhile, the direct contact between chloride ion and aluminum alloy substrate will aggravate the anodic dissolution of the substrate, resulting in pitting, as shown in formula (6) . Part of the generated Al 3+ ions diffuse outside the pitting, usually forming an aluminum hydroxide layer at the edge of the pitting, which hinders the further outward dispersion of Al 3+ ions. The concentration of metal ions in corrosion voids is higher than that outside due to the aggregation of Al 3+ ions, that is, excess positive charge. Therefore, the external chloride ions will continue to enter pitting to maintain charge balance. At the same time, the hydrolysis of Al 3+ ions in pitting increases the acidity of the local solution, resulting in a decrease in the pH value, as shown in formula (7) . The hydrolysis of chloride caused by the large amount of chloride ions will further aggravate the acidification of corrosion solution, thus promoting the continuation of corrosion. 6 Al = Al 3+ + 3e − 7 Al 3+ + 3H 2 O = Al(OH) 3 + 3H + During immersion in 3.5 wt% NaCl solution, first, an air layer will be formed on the FWEDMS to prevent the corrosion liquid from wetting the surface ( Fig. 11a ). Second, for the hydrophobic surface of FWEDMS, it is negatively charged in neutral solutions. Among the negative ions, corrosive chloride ions has the lowest saturation coverage. The negative charging of a hydrophobic surface causes a redistribution of ions inside the double electric layer leading to the depletion of chloride anion concentration in the vicinity of a solid surface. 35 Finally, the well-ordered layer of a hydrophobic agent acts as a barrier for charge transfer, corrosive chloride ions cannot penetrate into the substrate, and the possibility of corrosion reaction on the FWEDMS is lower than that on the PS. For SLIPS, a barrier to isolate corrosion solution is formed on the surface due to the injection of high density and low surface energy lubricating oil, as shown in Fig. 11b . Moreover, the high strength porous microstructure produced by WEDM can lock the lubricant firmly on the surface of the sample, which makes the barrier more stable and extremely difficult to penetrate by corrosion solution. In addition, lubricating oil with fluid properties spontaneously flow to the defective areas on the surface through capillary action driven by surface energy, resulting in self-healing ability, which significantly improves the surface defect induction and aggravation of corrosion problems. In conclusion, the SLIPS has better corrosion resistance in NaCl corrosion solution environment."
} | 6,299 |
39830389 | PMC11737406 | pmc | 9,305 | {
"abstract": "ABSTRACT The enormous computational requirements and unsustainable resource consumption associated with massive parameters of large language models and large vision models have given rise to challenging issues. Here, we propose an interpretable ‘small model’ framework characterized by only a single core-neuron, i.e. the one-core-neuron system (OCNS), to significantly reduce the number of parameters while maintaining performance comparable to the existing ‘large models’ in time-series forecasting. With multiple delay feedback designed in this single neuron, our OCNS is able to convert one input feature vector/state into one-dimensional time-series/sequence, which is theoretically ensured to fully represent the states of the observed dynamical system. Leveraging the spatiotemporal information transformation, the OCNS shows excellent and robust performance in forecasting tasks, in particular for short-term high-dimensional systems. The results collectively demonstrate that the proposed OCNS with a single core neuron offers insights into constructing deep learning frameworks with a small model, presenting substantial potential as a new way for achieving efficient deep learning.",
"introduction": "INTRODUCTION Deep learning (DL), particularly through deep neural networks (DNNs), has proven to be highly effective in terms of learning data representations. The DNNs, which are derived from conventional neural networks but significantly outperform their predecessors, excel at capturing complex patterns and features within data. Recently, empowered by DNNs, DL has been widely incorporated into artificial intelligence (AI) infrastructure as an indispensable component and has achieved distinguished performance in a variety of fields, including time-series forecasting [ 1 ], computer vision [ 2 ], natural language processing [ 3 , 4 ], bioinformatics [ 5 ] and recommendation [ 6 ]. Nevertheless, the training and deployment expenses of DNNs are substantial since DNNs, especially the large language models (LLMs), e.g. LlaMA [ 7 ] and LaMDA [ 3 ], typically contain vast numbers of parameters or neurons and require a huge amount of training data to be trained on high-performance computing platforms for extended periods. For example, the LaMDA [ 3 ] has up to 137B parameters and needs to be pretrained on 1024 TPU-v3 chips for a total of ∼57.7 days. The massive parameter sizes in the ‘large models’ lead to rapidly growing carbon footprints and unsustainable energy consumption levels, which may result in serious impacts on the environment and climate in the near future [ 8 ]. To address the challenges mentioned above, researchers have developed many DNN compression techniques [ 9 , 10 ] to diminish the sizes and complexity levels of DNNs without substantially compromising their performance to an extent. In broad categorization, these methods encompass network pruning [ 11 ], quantization [ 12 ], low-rank factorization [ 13 ] and knowledge distillation approaches [ 14 ]. Deep compression [ 15 ] and the method detailed in [ 16 ] yield significant compression rates in neural networks—ranging from 35× to 49× on basic CNN architectures, like AlexNet [ 17 ] and VGG [ 18 ], and achieving a maximum of 4.7× on complex ResNets [ 19 ]. Despite recent developments in DNN compression, notable challenges persist. For example, identifying a pruning strategy that avoids performance degradation can be daunting [ 20 ]. Additionally, quantization introduces quantization errors [ 9 ], leading to performance deterioration in specific scenarios. In conclusion, most model-compression methods still require training accurate models with a considerable number of parameters beforehand, and do not address the issue of excessive parameters from a mechanistic or intrinsic perspective. Due to these limitations, challenges remain with regard to bridging the gap between DNNs and various applications, e.g. the Internet of Things (IoT) [ 21 ], embedded systems [ 22 ] and edge computing [ 23 ]. The emergence of machine learning applications [ 24 , 25 ] that incorporate the delay models derived from time-delay dynamical systems [ 26 , 27 ] shed light on the design of new frameworks. This innovation has the potential to intrinsically reduce the number of parameters required to achieve optimal model performance. From a mathematical perspective, continuous time-delay systems possess a noteworthy property wherein their state spaces can be extended to infinite dimensions. In practice, the dynamics of these delay systems with finite dimensions still display the characteristics of short-term memory and high dimensionality [ 26 ]. Recently, delay-based reservoir computing [ 27 ] has replaced interconnected nodes in the conventional reservoir structure with virtual nonlinear nodes subjected to delayed feedback. Although the prediction accuracy of this method is unsatisfactory, the efficacy of delay-based concepts in reservoir computing has spurred subsequent applications within DNNs. For instance, Fit-DNN [ 28 ] was proposed to emulate a full DNN by using only a single neuron with multiple delayed and modulated feedback, which empowers DNNs with sparse connectivity and reduces memory consumption costs. However, this method neither fully explores the temporal dynamics of a single neuron nor provides data representation for dynamical systems. By assuming that the steady states of a high-dimensional system are constrained on a low-dimensional manifold, which is generally satisfied by dissipative real-world systems, the spatiotemporal information (STI) transformation has theoretically been derived from the delay and non-delay embedding theory [ 29–32 ]. The STI transformation equation converts the high-dimensional/spatial data into the temporal information of a latent or target variable, i.e. a high-dimensional state/vector topologically corresponds to a one-dimensional sequence and vice versa. Several methods have recently been developed for predicting short-term time series within the STI transformation framework, with an explicit target variable, e.g. randomly distributed embedding (RDE) [ 32 ], the auto-reservoir neural network (ARNN) [ 33 , 34 ] and the spatiotemporal information conversion machine (STICM) [ 35 ]. However, none of these methods can simultaneously forecast multivariate states for a dynamical system. If we consider each variable from a high-dimensional system as a target and train the models independently, the computational cost would sharply increase. While existing DL methods generally represent complex high-dimensional data by ‘a latent vector’ (plenty of neurons), the STI equation, in contrast, is able to represent such data by the time series of ‘one latent variable’ rather than a vector, thus providing a new way for DL even with just a single neuron. Notably, time-delayed reservoir computing [ 36 ] has recently demonstrated that a few delayed neurons are sufficient to represent the dynamics of a high-dimensional dynamical system, further supporting the feasibility of using a single neuron for deep learning. In this work, inspired by both the STI transformation equation and machine learning approaches with time-delay models, we introduce a novel small neural network framework (Fig. 1a ) called the one-core-neuron-system (OCNS), which is actually a recurrent neural network (RNN [ 37 ]) of one neuron with two linear layers. We theoretically demonstrate that the one-core-neuron (OCN) generically is an embedding of the original nonlinear dynamical system (see Theorem 1 of Methods). In addition to the theoretical background, the versatility of the OCNS is demonstrated in tasks such as multivariate time-series forecasting (Fig. 1b ) and classification ( Fig. S3 ). In contrast to the existing ‘large models’, such a ‘small model’ OCNS is able both theoretically and computationally to reconstruct the states of the entire high-dimensional dynamical system with a single latent variable, thus enabling data representation learning with only one neuron. Specifically, as a small model, the OCNS achieves comparable or superior performance using on average only 0.035% of the parameters necessary for large-model-based methods [ 38 ], while also requiring just 1.16% of the parameters needed by the typical transformer-based approaches [ 1 , 39 , 40 ]. Intuitively, the OCNS consists of one neuron RNN \\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${{\\bf \\Phi }}$\\end{document} with two linear layers (i.e. two weight matrices A, B ), simultaneously representing both the primary and conjugate STI equations in an autoencoder form (Fig. 1a ). The OCN \\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${{\\bf \\Phi }}$\\end{document} is specifically designed to achieve DL using only a single core neuron. In particular, the input weight A and the OCN \\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${{\\bf \\Phi }}$\\end{document} encode the spatial information of original multi-variables (or a state) as the temporal information of a single latent variable (i.e. a one-dimensional time series), while the output weight B decodes the latent temporal information as the high-dimensional/spatial information (or state) of the original dynamical system. Based on the STI transformation [ 32 , 33 , 35 ], the dynamics of the original system can be topologically reconstructed from one latent variable by the OCNS. Once the latent one-dimensional delay dynamical system (derived from the OCN \\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${{\\bf \\Phi }}$\\end{document} ) is constructed, the latent temporal information can iteratively predict future spatial information through output weight B in the conjugate STI equation. In other words, a solid theoretical foundation enables the OCNS to represent the whole states/dynamics of an original system and, furthermore, to forecast multivariate time series in a multistep-ahead manner. Moreover, by fully exploiting the advantages of the one-dimensional delay dynamical system and nonlinear STI transformation, the OCNS only requires a much smaller number of parameters, specifically including A, B , and a small subset in the OCN \\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${{\\bf \\Phi }}$\\end{document} . And the OCN can almost maintain a small and constant parameter size even when processing a large-scale system. Therefore, the OCNS, as a small model, has a broader range of applications than conventional DL, especially for scenarios with strict restrictions on the number of parameters and specific requirements for learning capacity. Figure 1. Overview of the OCNS as a ‘small model’ for time-series predicting. (a) The framework of the OCNS is similar to that of an autoencoder. For an observed high-dimensional vector \\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${{{\\boldsymbol{X}}}^t}$\\end{document} , a latent delay vector \\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${{{\\boldsymbol{Z}}}^t}$\\end{document} , comprised of the dynamics of a one-dimensional delay dynamical system \\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${{z}^t}$\\end{document} , is constructed via the input weight A and OCN \\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${{\\bf \\Phi }}$\\end{document} through a delay-embedding scheme. The delay vector \\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${{{\\boldsymbol{Z}}}^t}$\\end{document} corresponding to time t contains the latent temporal information from the delay dynamical system \\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${{z}^t}$\\end{document} , which can topologically reconstruct all the dynamics of the original system \\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${{{\\boldsymbol{X}}}^t}$\\end{document} . With the output weight B , the original spatial information \\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${{{\\boldsymbol{X}}}^t}$\\end{document} of the original system can be recovered from \\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${{{\\boldsymbol{Z}}}^t}$\\end{document} . The OCN \\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${{\\bf \\Phi }}$\\end{document} , which generates a delay dynamical system \\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${{z}^t}$\\end{document} with D delays feedback in a single neuron-based fashion, is the core of the OCNS. (b) Derived from the solid theoretical foundation of delay dynamical systems [ 42 , 43 ] and the delay embedding theorem [ 29–31 ], the information flow of the OCNS is dictated by the OCNS-based STI equations, which encompass both the primary and conjugate STI equations (Eq. 3 ). Here, we build the delay vector \\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${{{\\boldsymbol{Z}}}^{t + i}} = {{[ {{{z}^{t + i - S + 1}},{{z}^{t + i - S + 2}},\\ldots,{{z}^{t + i}}} ]}^{\\prime}} \\in {{\\mathbb{R}}^S} $\\end{document} at time \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$t + i$\\end{document} , where \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$i = 0,1,2\\ldots $\\end{document} and \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$S $\\end{document} is the delay-embedding dimension. Specifically, the input weight A and OCN \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n${{\\bf \\Phi }}$\\end{document} transform the spatial information in the original attractor \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$\\mathcal{A}$\\end{document} into the temporal information of the delayed attractor \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$\\mathcal{N}$\\end{document} corresponding to the primary STI equation, while the conjugate STI equation represents the reconstruction and prediction of the original system constrained on attractor \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$\\mathcal{A}$\\end{document} from the delayed attractor \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n$\\mathcal{N}$\\end{document} through the output weight B . In this way, the OCNS effectively consists of an RNN with one neuron and two linear layers. To evaluate the performance of the OCNS, we conduct extensive experiments on the numerical Lorenz datasets [ 41 ] and eight real-world datasets. The experimental results numerically demonstrate that the OCNS, as a ‘small model’ with only a single core neuron, can outperform or achieve comparable performance to current forecasting benchmark methods, proving that it can match the performance of ‘large models’ with many more neurons, particularly for short-term high-dimensional systems. Moreover, we extend and generalize the OCNS to image categorization tasks or cross-sectional datasets, and these results also demonstrate the excellent performance of the OCNS. All of the theoretical and computational results show that from a dynamical perspective, the OCNS is powerful and flexible enough to offer a new way to achieve DL with a small model.",
"discussion": "DISCUSSION In this study, we proposed a novel deep learning framework called the OCNS. Unlike existing deep learning models that typically contain many neurons and layers to represent complex high-dimensional data using a latent vector, the OCNS, as a small model, comprises a single core neuron with merely two linear layers to represent high-dimensional time-series data using a latent variable (rather than a vector). One prominent advantage is that the OCNS can fully exploit the complexity of the dynamics in the delay system, ensuring high capacity. Meanwhile, it significantly reduces the parameter scale of a deep learning model, requiring only a fraction (e.g. 0.00017%) of the parameters in LLM-based methods. Based on the delay embedding theorem, the OCNS is proposed to represent the whole states of a dynamical system even with only a one-dimensional latent variable, and we derive the OCNS-based STI equations (Eq. ( 3 )), i.e. one high-dimensional state corresponds to one-dimensional sequence and vice versa. Specifically, the OCNS is an autoencoder-based framework that can solve both primary and conjugate OCNS-based STI equations. From this perspective, the input weight A and the OCN \\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${{\\bf \\Phi }}$\\end{document} collaboratively serve as the encoder, transforming the spatial information of dynamical systems into temporal dynamics, while the output weight B functions as the decoder, reconstructing the original system from the latent time series. In other words, the STI equations enable the OCNS, as a small model, to create a latent one-dimensional delay dynamical system that can encode and topologically reconstruct the original dynamical system. Generally, explicitly reconstructing the original dynamical system \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n${{{\\boldsymbol{X}}}^t}\\ $\\end{document} from the nonlinear observable \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n${{{\\boldsymbol{Z}}}^t}$\\end{document} (temporal dynamics) requires a nonlinear function f , i.e. \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n${{{\\boldsymbol{X}}}^t} = f( {{{{\\boldsymbol{Z}}}^t}} )$\\end{document} . Using a linear matrix B for decoding raises a concern that it might limit the model's ability to decode highly nonlinear information. However, our empirical evaluations show that the nonlinear OCN \\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n${{\\bf \\Phi }}$\\end{document} effectively compensates for this limitation, thus producing high-quality forecasting results, in particular for short-term high-dimensional data. One candidate for the conjugate OCNS-based STI equation can be a multilayer nonlinear neural network, as demonstrated by the superior performance of OCNS + . Depending on the task complexity, it is feasible to choose a larger decoder, or even some larger models. Our approach differs from a recent, promising time-delayed reservoir computing (RC) model [ 36 ], which introduces time delays only in the output layer. The delay mechanism allows the delayed RC to match the performance of larger RCs while reducing the reservoir's physical size, providing flexible memory capacity to handle more complex dynamic reconstruction tasks than a standard RC of the same size. While these two methods share some similarities, such as both featuring evolution dynamics and a linear time-delayed readout, OCNS uniquely combines the strengths of autonomous evolution (delay dynamical systems) and delay embedding schemes. Additionally, OCNS trains parameters for both linear decoding and nonlinear forward propagation, enhancing its ability to capture the nonlinear characteristics of the original systems. A detailed comparison is provided in Section S14. From the perspective of reusing small models systematically and thoroughly, the recently proposed learnware paradigm [ 61–63 ] addresses challenges that remain unsolved by prevailing singular large models, such as the lack of training data and skills, catastrophic forgetting, and the complexities of continual learning. Simultaneously, learnware obviates the necessity for users to construct a model from scratch. Although the current learnware paradigm is in a preliminary stage and poses numerous issues for future exploration, such a concept enables existing trained small models to be competent in most tasks, thus effectively reducing carbon emissions by minimizing redundant model deployments. In contrast to the learnware paradigm, we propose a minimal and concise OCNS that utilizes an OCN to represent the entire dynamical system, thereby inherently ensuring good-enough performance for small models. At the same time, the concept of the learnware paradigm offers profound inspiration for the future design of a more refined and generalized OCNS framework. For instance, we envision adopting a divide-and-conquer approach by incorporating multiple interacting OCNs to represent an exceedingly complex dynamical system, with each OCN responsible for representing a portion of the system. Several ongoing challenges and potential research directions remain and motivate future work. First, the integration of the OCNS with deep learning is limited to the MLP architecture so far, while deep learning includes several other foundational network models, such as CNNs and transformers. Designing a more general delay system architecture based on these frameworks might unlock more groundbreaking methods. Besides, it is still critical to further explore efficient gradient computation methods that are tailored for delay dynamical systems. In conclusion, in contrast to the existing ‘large models’, the proposed OCNS paves a new way for achieving efficient deep learning and AI through a small model framework, offering significant potential for a wide range of real-world applications."
} | 6,519 |
37406763 | null | s2 | 9,308 | {
"abstract": "Engineering the utilization of non-native substrates, or synthetic heterotrophy, in proven industrial microbes such as Saccharomyces cerevisiae represents an opportunity to valorize plentiful and renewable sources of carbon and energy as inputs to bioprocesses. We previously demonstrated that activation of the galactose (GAL) regulon, a regulatory structure used by this yeast to coordinate substrate utilization with biomass formation during growth on galactose, during growth on the non-native substrate xylose results in a vastly altered gene expression profile and faster growth compared with constitutive overexpression of the same heterologous catabolic pathway. However, this effort involved the creation of a xylose-inducible variant of Gal3p (Gal3p"
} | 189 |
35480020 | PMC9037623 | pmc | 9,310 | {
"abstract": "Through the special chemical structure of dopamine (DA), superhydrophobic polyester (PET) fabric was fabricated by introducing the low surface energy substance hexadecyltrimethoxysilane (HDS) into the PET fabric and chelating Fe ions with phenolic hydroxyl groups of polydopamine (PDA) to form a rough surface. The water contact angle (WCA) of the prepared PDA/Fe/HDS PET fabric was higher than 160° and the scrolling angle (SA) was lower than 2.09°. The excellent adhesion property of polydopamine (PDA) on the substrate is helpful to improve the stability of superhydrophobic PDA/Fe/HDS PET fabric. The tests results showed that the modified PET fabric maintains excellent mechanical properties. Its superhydrophobic property had good stability and durability in the harsh environment of washing, mechanical friction, UV irradiation, seawater immersion, acid–base and organic reagents erosion. The PDA/Fe/HDS PET fabric also had good self-cleaning and oil–water separation properties. It still had good oil–water separation performance after repeated use for 25 times, and the separation efficiency was more than 95%. The preparation method was facile, the treatment time can be shortened, the cost of the modified substrate was low, and fluorine-free substances were used in the process. This work provides a new way to expand the added value of PET fabrics and develop durable superhydrophobic fabrics in practical application.",
"conclusion": "Conclusions In this work, superhydrophobic PDA/Fe/HDS PET fabric was successfully fabricated. The excellent adhesion property of PDA contributes to the fastness of HDS on PET fabric, thereby greatly reducing the surface tension of the fabric, resulting in a superhydrophobic PET fabric with WCA up to 161.7° and SA about 2.09°. Compared with the traditional preparation method of oil–water separation material, this method is simple, the reagents used are environmentally friendly, and the reaction time is greatly shortened. The amount of HDS and the cost of the substrate were both reduced, and the prepared superhydrophobic fabrics had good stability against various harsh environments and excellent reusability. This work opens up a new way to expand the value of PET fabrics and inspires the practical the practical application of durable superhydrophobic fabrics.",
"introduction": "Introduction In the 21st century, the PET industry has made great progress. Due to its low price, excellent performance, easy washing and quick drying, it has become one of the most popular textile raw materials. Therefore, the functional finishing of PET fabric has become a research hotspot for researchers. In general, the solid–liquid WCA on the surface of this material is greater than 150°, and the SA is less than 10°. 1–3 Superhydrophobic PET fabric has outstanding self-cleaning, anti-fouling, anti-fogging, anti-icing, drag reduction and anti-corrosion functions, 4–7 and has important application values in industry, military, biomedicine and other fields. The idea to prepare superhydrophobic fabrics is to introduce low surface energy substances to reduce the surface tension of the fabric, and to generate nanoparticles on the surface of the fabric or etch them chemically or physically to construct a structure with a micro/nano rough structure similar to the lotus leaf surface. Therefore, researchers have adopted different strategies, 2 for example, by chemical vapor deposition, 3 sol–gel coating 4,5 and other methods to introduce fluorinating agents 8–13 and other low surface energy substances into the fabric, or using nanoparticle in situ growth, 14 chemical etching 15 or plasma etching 16,17 to endow the fabric surface with a micro/nano-level rough structure to achieve the superhydrophobic aim. However, the operations of these methods are complicated, or the reagent used is not environmentally friendly, and the prepared superhydrophobic material has poor stability, which seriously hinders practical application. 18 Marine mussels have attracted widespread attention due to their excellent bio-adhesion. Mussels can adhere to almost all types of substrates in humid environments by secreting adhesion proteins. 19–22 Mucus proteins secreted by mussels mainly include mfp-1, mfp-2, mfp-3, mfp-4, mfp-5, and mfp-6. 23–25 Among them, mfp-3 and mfp-5 are the keys to interface adhesion. When contacting with a hydrophobic surface, mucins will expose an indole or benzene ring structure to form strong hydrophobic interactions with the surface, and studies show that their interaction with hydrophobic surfaces is much greater than with hydrophilic surfaces. 26–28 The modified substrate can be firmly combined with dopamine polymer by simple dip-coating in dopamine aqueous solution, which greatly improves the mechanical stability of the modified material. 29–32 At the same time, dopamine has active functional groups such as amino and phenolic hydroxyl groups, which can produce a variety of chemical reactions, providing a reaction platform for the second modification of the substrate. Dopamine is a non-toxic, environmentally friendly and highly adaptable substance. There are many reports about the application of dopamine in preparation of superhydrophobic fabrics. The traditional self-polymerization of dopamine under alkaline environment was used to modify the fabric, 33,34 and the introduction of fluorine-containing low surface energy substances can rapidly endow the fabric with superhydrophobic property. 35 The problem of these methods is that the preparation process is time consuming or the chemical used is not environmentally friendly. Therefore, a simple and pollution-free superhydrophobic preparation method with good stability is an urgent requirement. In this paper, a simple, rapid and environmentally friendly preparation method of superhydrophobic PET fabrics was introduced. By using the excellent adhesion property of dopamine, low surface energy substance HDS was introduced to the surface of PET, and metal ions were chelated on the surface to form a rough convex and waxy “lotus leaf” like surface, and the superhydrophobic PET fabric was prepared. The basic principle of the reaction is as follows: dopamine acts as a ligand to form a stable dopamine/metal complex with metal ions as the center ( Fig. 1a ), 36–39 which forms nanoparticles on the surface of PET fabric. At the same time, the alkoxy group at one end of the molecular structure of silane coupling agent is hydrolyzed to silanol group, which is dehydrated and condensed to form oligosiloxane containing silanol group. It can form a hydrogen bond with a phenol hydroxyl group in the molecular structure of DA, and forms a covalent bond with PDA during the heating process ( Fig. 1b ). Under the catalysis of oxidants, DA molecules rapidly polymerize to form a network structure of polydopamine macromolecules ( Fig. 1c ), 40,41 which covers the surface of the fabric. Therefore, nano-scale particles were formed on the surface of PET fabric under the action of PDA, and the surface energy of the fabric was greatly reduced due to the introduction of HDS, and PDA/Fe/HDS PET fabric was prepared. The schematic of preparation of superhydrophobic PET fabric is shown in Fig. 1 . Fig. 1 Schematic of (a) dopamine chelated ions, (b) dopamine formed covalent bonds with long-chain alkyl siloxane and (c) the possible oxidative polymerization mechanism of PDA on the PET fabric. The surface morphology and chemical compositions of the PDA/Fe/HDS PET fabric were observed by SEM, AFM, EDS, XPS. The mechanical stability and stability in seawater, acid, alkali and organic solutions were further tested to simulate the durability of superhydrophobic PET fabric in complex marine environment. The results show that PDA/Fe/HDS PET fabric has excellent and stable superhydrophobic properties, indicating that this method has broad application prospects in the modification of superhydrophobic polyester, and providing a research basis for the industrial preparation of superhydrophobic PET fabric.",
"discussion": "Results and discussion Surface morphology characterization The surface morphology of superhydrophobic material is one of the key factors affecting its wettability. Therefore, it is of great significance to study the surface morphology of polyester fabric. A field emission scanning electron microscope (FESEM) was used to observe the surface morphology of PET fabrics. It can be seen from Fig. 3a that the surface of the original PET is smooth and has a certain degree of hydrophobicity in a short period of time. The WCA is 115.1° and the water droplet completely penetrates into the original fabric ( Fig. 3a ) after 4 min. The surface of PET fabric remains smooth after HDS finishing ( Fig. 3b ), and the WCA reached 145.2°. After finishing with PDA/Fe ( Fig. 3c ) and PDA/HDS ( Fig. 3d ), a layer of fine nano-particles appear on the surface of the fabric, with WCA of 135.5° and 151.6°, while WCA and SA of PDA/Fe/HDS PET fabric can reach 161.7° ( Fig. 3e(i) ) and 2.09° ( Fig. 3e(ii) ), respectively. The results show that the material with low surface energy has a greater influence on the preparation of superhydrophobic polyester fabric, and the roughness can further improve the superhydrophobic property. Fig. 3 SEM images of original PET fabric (a), HDS PET fabric (b), PDA/Fe PET fabric (c), PDA/HDS PET fabric (d) and PDA/Fe/HDS PET fabric (e) with different multiples, the insets are WCA (a, b, c, d and e(i)) and SA (e(ii)). Atomic Force Microscopy (AFM) was used to further characterize the surface roughness of different PET fabrics. It can be seen from Fig. 4 that the original PET fabric and HDS PET fabric have smooth surfaces, the root mean square (RMS) value of the smooth original PET fabric surface and HDS PET fabric were approximately 4.16 nm and 4.04 nm. The higher RMS value represents larger roughness. 44 Compared with the original PET fabric, the rough structure of PDA/Fe PET fabric covered by the nano particles can be clearly seen from Fig. 4c , and the RMS value is about 18.63 nm. After modified with HDS, the RMS value of PET fabric is further increased to 44.38 nm ( Fig. 4d ). The PDA/Fe PET fabric and PDA/Fe/HDS PET fabric have rough surfaces, which are consistent with the results of SEM. Therefore, the presence of PDA and metal ions can effectively increase the surface roughness of the fabric. By increasing the surface roughness of the PET fabric while reducing the surface energy, a superhydrophobic PET fabric with good hydrophobic properties can be prepared. Fig. 4 AFM image of original PET fabric (a), HDS PET fabric (b), PDA/Fe PET fabric (c), and PDA/Fe/HDS PET fabric (d). Chemical analysis of PDA/Fe/HDS PET fabric The chemical structure and chemical compositions of the fabric are mainly determined by FTIR and XPS analysis. Table 1 lists the element compositions of different PET fabrics obtained by XPS. The original PET fabric contains carbon (68.34%) and oxygen (31.66%) elements. After HDS finishing, a layer of long-chain alkyl siloxane is deposited on the surface of the PET fabric, the carbon element content is increased to 79.13%, and 3.36% silicon element is introduced. The N element on the surface of the fabric treated with PDA/HDS and PDA/Fe comes from PDA. Because a large number of phenolic hydroxyl groups in PDA molecular structure can couple with HDS, the content of silicon element is increased. PDA/Fe/HDS PET fabric contains nitrogen, iron and silicon. Among them, Fe comes from metal salt and silicon comes from HDS. Fig. 5 shows the EDS distribution mapping of surface elements of PDA/Fe/HDS PET fabric. It can be seen that C and O elements are widely distributed on the surface, while N, Fe and Si elements are less, but the distribution is very uniform, 45,46 indicating that metal and HDS were successfully introduced to the PET fabric surface through dopamine. Surface elemental compositions of PET fabrics Samples Atomic percentage of surface element (%) C N O Fe Si Original PET fabric 68.34 — 31.66 — — HDS PET fabric 79.13 — 17.51 — 3.36 PDA/Fe PET fabric 73.07 2.68 23.68 0.56 — PDA/HDS PET fabric 75.34 1.38 18.7 — 4.57 PDA/Fe/HDS PET fabric 81.02 1.09 10.98 0.29 6.62 Fig. 5 Surface elemental compositions map of PDA/Fe/HDS PET fabric. \n Fig. 6a shows the FTIR spectra of PET fabrics. The peak at 3311 cm −1 is the stretching vibration of –OH and the stretching vibration of –NH 2 . For the spectra of HDS PET fabric, PDA/HDS PET fabric and PDA/Fe/HDS PET fabric, 2917 cm −1 and 2850 cm −1 are the symmetrical vibrations and asymmetrical stretching vibrations of –CH 3 and –CH 2 groups. The stretching vibration at 1711 cm −1 corresponds to 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, which increases after the introduction of PDA due to the oxidation of the phenolic hydroxyl group in DA molecular structure to carbonyl groups. The –O C–O vibration absorption bands of aromatic esters appear at 1245 cm −1 and 1094 cm −1 , and the peak at 1016 cm −1 (ref. 47 ) corresponds to the in-plane bending vibration of the benzene ring and the antisymmetric stretching of Si–O–Si, and the C–C and Si–O–Si symmetric stretching vibration of the aromatic rings appear at 872 cm −1 (ref. 48–50 ), indicating that the long-chain alkyl groups were successfully introduced to the surface of PET. Fig. 6 The FTIR patterns (a) and wide scan XPS (b) spectra of PET fabrics. \n Fig. 6b shows the wide scan of XPS spectra of different PET fabrics, and the introduction of Si element can be seen. Due to the low content of Fe and N, there is no peak in the wide scan spectra. Fig. 7a, b and c correspond to the C1s spectra of the original PET fabric, HDS PET fabric and PDA/Fe/HDS PET fabric, and the peaks at 285.1 eV, 286.7 eV, 288.9 eV and 290.1 eV correspond to C–C/C–H, C–O, C O and O–C O bonds, 50,51 respectively. The new peak at 285.7 eV corresponding to C–N bond appears in Fig. 7c . Fig. 7d and e are the N1s spectrum and Fe2p spectrum of PDA/Fe PET fabric, respectively. In Fig. 7d , the peaks at 399.7 eV, 402.4 eV and 398.5 eV correspond to –NH–, –NH 2 and –N bonds. The –N bond may be derived from the intramolecular rearrangement of PDA. In Fig. 7e , Fe2p peaks appear at 727.3 eV and 713.4 eV. Fig. 7f shows the Si2p spectrum of PDA/Fe/HDS PET fabric, and three peaks at 103.8 eV, 102.6 eV and 101.9 eV belong to C–Si–O–Si, C–Si–O–H and C–Si–O–C bonds. 52,53 The results of XPS and FTIR confirmed that Fe 2+ and HDS were successfully co-deposited on the surface of PET fabric through the strong adhesion of PDA. Fig. 7 XPS C1s spectra of the original PET fabric (a), HDS PET fabric (b) and PDA/Fe/HDS PET fabric (c); XPS N1s (d) and Fe2p (e) spectra of PDA/Fe PET fabric; XPS Si2p spectra of PDA/Fe/HDS PET fabric (f). Mechanical stability of PDA/Fe/HDS PET fabric The biggest advantage of PET fabric is its good wrinkle resistance and shape retention. Therefore, the change of strength before and after modification is also one of the important indexes of its stability. Table 2 lists the breaking strength and elongation at break before and after superhydrophobic finishing. The results show that the strength of PET fabric before and after superhydrophobic finishing has almost no change, indicating that this modification method is mild and has little effect on the mechanical properties of the fabric, and the fabric is not easy to be damaged during use. Mechanical properties of original PET fabric and PDA/Fe/HDS PET fabric Property Original PET fabric PDA/Fe/HDS PET fabric Breaking strength (N) 586.99 ± 12.68 (warp) 577.48 ± 16.67 (warp) 419.49 ± 12.49 (weft) 421.44 ± 14.13 (weft) Elongation at break (%) 14.18 ± 1.01 (warp) 14.19 ± 0.48 (warp) 16.75 ± 0.53 (weft) 16.99 ± 0.66 (weft) Stability tests of the PDA/Fe/HDS PET fabric The complexity of the marine environment requires the superhydrophobic materials used for oil–water separation to have certain physical/chemical stability and durability in harsh environments. Accordingly, the physical stability of fabrics was tested, including washing fastness, abrasion resistance and UV aging resistance, and the chemical stability mainly includes seawater resistance, acid and alkali resistance and organic reagent resistance ( Fig. 8 ). Fig. 8 The stability of PDA/Fe/HDS PET fabric with washing resistance (a), abrasion resistance (b), UV aging resistance (c), seawater resistance (d), acid/alkali resistance (e) and organic reagents resistance (f). The washing fastness is one of the most important tests for fabric finishing. The washing stability test was done according to AATCC 61-2006 standard method, which is equivalent to 5 laundering cycles of commercial and domestic washing. 21 WCA and SA of the superhydrophobic PET fabric prepared after washing at 0 min, 45 min, 90 min, 135 min, 180 min, and 225 min were tested, respectively. After washing, WCA of the PDA/Fe/HDS PET fabrics are still stable above 150°, and SA remains at about 10° ( Fig. 8a ). Due to the strong adhesion of PDA, after 5 laundering cycles, the staining fastness of cotton and polyester is about level 5, and the fading fastness of modified fabric remained level 4.5. After 0, 200, 400, 600, 800 and 1000 cycles of abrasion, WCA and SA of the fabrics were tested. The results showed that with the increase of abrasion cycles, the fluff on the surface of the fabric increased, WCA of the fabric gradually decreased, and SA gradually increased ( Fig. 8b ). However, WCA of the washed fabrics are all greater than 150° and SA are less than 10°, which remained in the superhydrophobic range. WCA and SA of PDA/Fe/HDS PET fabric were measured after being exposed to the Accelerated Weathering Tester UV for 4 h, 8 h, 12 h, 16 h, 20 h and 24 h at 60 °C to simulate the effect of sunlight on superhydrophobic properties of PDA/Fe/HDS PET fabric in outdoor oil–water separation or outdoor storage. The results are shown in Fig. 8c , the WCA of PET fabric has little change after 24 h of UV irradiation, indicating that the fabric has good stability and resistance to ultraviolet radiation. Due to the potential application of superhydrophobic PET fabric in marine oil spill accidents and the complexity of marine environment, it is very important to test its stability against seawater, acid/alkali and organic reagents. The PDA/Fe/HDS PET fabric was placed between acrylic-resin plates to withstand 12.5 ± 9 kPa pressure for several hours to test its durability in the sea. The results are shown in Fig. 8d , the lowest WCA decreased to 155.1°, and the fabric shows good stability against seawater erosion. The fabric was soaked in a solution of pH 1, 3, 5, 7, 9, 11, 13 prepared by HCl and NaOH for 24 hours to test the acid and alkali resistance of the superhydrophobic properties of PDA/Fe/HDS PET fabric. The data ( Fig. 8e ) shows superhydrophobic property of the fabrics change little in different pH solutions. The WCAs were above 155°, and SA were about 10°. The organic solution stability tests can simulate the stability of superhydrophobic materials to various adsorbed marine oils. The PDA/Fe/HDS PET fabrics were soaked in CCl 4 , AT, n -H, PE, DCM and THF for 72 h, respectively. The changes of WCA and SA were tested to characterize the stability of the fabric against organic reagents. The result shows that WCA of the soaked fabrics are above 150°, and the SA are about 10° ( Fig. 8f ). The WCA decreased by about 10° compared with that before soaking, which may be due to the fact that part of HDS was dissolved in the organic solvents, which increased the surface energy of the fabric and destroyed the superhydrophobic property. 54,55 It can be seen from the stability test results that the prepared superhydrophobic PET fabric has excellent resistance to washing, abrasion, UV irradiation, sea water, acid/alkali and organic reagents, which provides basis for its application in oil/water separation. Self-cleaning property and antifouling performance test of PDA/Fe/HDS PET fabric The methylene blue dye was sprayed randomly on the surface of the fabric, and the water droplets mixed the dye quickly on the surface of original PET fabric and contaminated the sample ( Fig. 9a ). The water droplets on the surface of PDA/Fe/HDS PET fabric quickly rolled off and took away the dye, and the mixed solution flowed into the culture dish, leaving a clean and flawless sample surface ( Fig. 9b ). Fig. 9 Self-cleaning performance of original PET fabric (a), and PDA/Fe/HDS PET fabric (b). The fabricated superhydrophobic PET fabric shows excellent self-cleaning performance like a lotus leaf. Water, vinegar, soy sauce and sesame oil were dropped on the surface of the prepared superhydrophobic PET fabric. After testing, the WCA of water, vinegar and soy sauce on the surface of the sample are 161.7°, 158.6° and 161.0°, respectively ( Fig. 10a(ii)–c(ii) ), while the sesame oil penetrates directly into the sample ( Fig. 10d(ii) ). The results show that PDA/Fe/HDS PET fabric has excellent antifouling, hydrophobic and lipophilic properties. Fig. 10 Droplets of water, vinegar, soy sauce and sesame oil on the original PET fabric (a(i), b(i), c(i) and d(i)) and PDA/Fe/HDS PET fabric (a(ii), b(ii), c(ii) and d(ii)). Oil–water separation properties With the advance of industrialization, a large amount of oily wastewater generated in the production process. To explore the oil absorption function of the superhydrophobic fabric, the modified fabric (4 cm × 4 cm) was used to adsorb organic reagents ( n -H and DCM) from water to prove that the superhydrophobic fabric can remove oil stains from water. As shown in Fig. 11 , the modified PET fabric can adsorb a small amount of light oil n -H ( Fig. 11a ) and heavy oil DCM ( Fig. 11b ) in water. Fig. 11 Selective absorption of modified fabric for (a) n -H and (b) DCM (dyed with Oil Red O) in water. In the separation process of light oil/water mixture, 80 mL of n -H or PE was mixed with 100 mL of water to simulate the state of floating oil at sea. To improve the oil adsorption capacity of PDA/Fe/HDS PET fabric, a self-made oil absorption bag containing three-dimensional porous PU nanosponge was made. The results show ( Fig. 12a ) that the experimental oil is quickly absorbed and the volume of water remains unchanged. The prepared oil adsorption bag has excellent adsorption capacity, which can absorb light oil about 14 times of its own weight ( Fig. 12d ). The adsorption capacity depends on the size of the oil adsorption bag and sponge, and its size is adjustable. Fig. 12 The adsorption bag for light oil/water separation (a), and gravity-driven oil–water separation device (b); the separation efficiency ( η ) of PDA/Fe/HDS PET fabric or adsorption bag for different organic solvent (c), and the recycle numbers of recyclability (d). The heavy oil/water mixture was separated by gravity driven method ( Fig. 12b ). The target heavy oil such as CCl 4 , DCM and CB were marked with Red Oil O, and the water was marked with methylene blue. The PDA/Fe/HDS PET fabric was fixed on the glass device, and then 200 mL of oil/water mixture was poured. Due to the excellent hydrophobic and lipophilic properties of the fabric, the oil soaked the fabric and entered the flask and the water was retained. The separation efficiency is about 99% ( Fig. 12c ). After 5, 10, 15, 20, and 25 times of separation, the separation efficiency is still above 95% ( Fig. 12d ), therefore, it is considered that PDA/Fe/HDS PET fabric has good application value in oil–water separation."
} | 5,999 |
40173217 | PMC11963962 | pmc | 9,312 | {
"abstract": "Cortical neuronal activity varies over time and across repeated trials, yet consistently represents stimulus features. The dynamical mechanism underlying this reliable representation and computation remains elusive. This study uncovers a mechanism for reliable neural information processing, leveraging a biologically plausible network model incorporating neural heterogeneity. First, we investigate neuronal timescale diversity, revealing that it disrupts intrinsic coherent spatiotemporal patterns, induces firing rate heterogeneity, enhances local responsive sensitivity, and aligns network activity closely with input. The system exhibits globally input-slaved transient dynamics, essential for reliable neural information processing. Other neural heterogeneities, such as nonuniform input connections, spike threshold heterogeneity, and network in-degree heterogeneity, play similar roles, highlighting the importance of neural heterogeneity in shaping consistent stimulus representation. This mechanism offers a potentially general framework for understanding neural heterogeneity in reliable computation and informs the design of reservoir computing models endowed with liquid wave reservoirs for neuromorphic computing.",
"introduction": "INTRODUCTION Cortical neurons exhibit substantial spiking time irregularity and trial-to-trial variability in both spontaneous activities and evoked responses to repeated stimuli ( 1 , 2 ). Despite this notable variability, neural representations of stimuli remain functionally consistent ( 3 – 6 ). Stimulus onset widely quenches neural variability ( 7 ) and leads to reliable sensory coding ( 6 ). Neural population dynamics has been suggested to straightforwardly produce resilient movement patterns, even when confronted with highly unpredictable neural responses ( 8 , 9 ). Thus, a central challenge is to understand the neural mechanisms that reconcile inherent cortical variability with reliable representations of external inputs. Randomly connected recurrent neural networks (RNNs) with excitation-inhibition (E-I) balance ( 10 ) are often used to model irregular and asynchronous activity with reliable macroscopic dynamics, but they mainly track low-dimensional inputs ( 8 , 10 , 11 ). Spatially extended spiking neural networks (SNNs), incorporating the distance-dependent connection probability ( 12 ), limit dynamical complexity but can also generate intricate spatiotemporal dynamics through Turing-Hopf bifurcation, producing coherent spatiotemporal patterns ( 13 , 14 ). These destabilized networks can be used for reservoir computing ( 15 ), whose reliability can be achieved by breaking the symmetry of network dynamics via nonuniform input connections ( 13 ), referred to as neural heterogeneity in terms of heterogeneous input connections. This suggests that neural heterogeneity essentially contributes to neural computation, as previously demonstrated, e.g., through the heterogeneity of cell spiking thresholds ( 16 ) or membrane time constants ( 17 ). However, the underlying dynamical mechanism remains elusive. Biological neural populations are highly heterogeneous, varying in structure, gene expression, and electrophysiological properties, such as membrane capacitance and resistance, resting potential, or spiking threshold. These variations capture the difference in the structural composition of the cell membrane across neurons of both the same and distinct types ( 18 – 20 ). Emerging research suggests that neuronal diversity plays a pivotal role in information processing ( 16 , 17 , 21 – 23 ). Excitability heterogeneity serves as a homeostatic control mechanism that enriches neural dynamics and enhances network resilience by stabilizing responses to modulatory input and preserving robust brain function ( 22 ). This heterogeneity in cell spiking thresholds further enables computational specialization in brain circuits by differentially regulating the gating, encoding, and decoding of signals in E and I neurons, in turn expanding the functional repertoire of local networks ( 16 ). On the other hand, neuronal heterogeneity in terms of membrane and synaptic time constants [observed in the brain ( 18 – 20 )] enhances robust learning ( 17 ). Furthermore, a reservoir computing model with diverse timescales achieves superior prediction accuracy and flexibility for multiscale chaotic dynamics by dynamically selecting appropriate timescales, outperforming standard reservoir computing models with identical neurons in both short- and long-term forecasting tasks ( 23 ). However, despite these advances, a dynamical mechanism to understand the reliability of input representation and computation remains absent, and a unified framework to elucidate the roles of various neural heterogeneities has yet to be established. Here, we reveal a potentially general dynamical mechanism, starting by investigating two specific neural diversities: heterogeneous leakage time constants and gain time constants ( 18 , 24 ). We explore their roles in the reliability of computation regarding an input-output mapping task, using a biologically plausible SNN model. Our findings show that either diversity can enhance reliable computation, playing a role similar to that of nonuniform input connections ( 13 ). It suggests a unique dynamical mechanism, relying on consistent representation: The diversity disrupts intrinsic coherent spatiotemporal patterns ( 13 , 25 ) and induces firing rate heterogeneity, leading to local sensitivity and globally input-slaved transient dynamics ( 26 ). This dynamics shapes the high-dimensional representation, inducing a liquid-like spatiotemporal activity pattern and forming input-slaved trajectories for reliably representing input information ( 26 ). This mechanism is robust across networks with varying connection ranges and connectivity randomness, and explains the similar roles of other heterogeneities, such as nonuniform input connections, spike threshold heterogeneity, and network in-degree heterogeneity. It provides a potentially uniform framework for understanding the significance of various forms of neural heterogeneities in ensuring reliable neural information processing. Thus, our work sheds light on reservoir computing models endowed with this mechanism.",
"discussion": "DISCUSSION This study elucidates the dynamical mechanism underlying reliable representation and computation in SNNs with neural heterogeneity. We start by exploring the roles of two biologically plausible neuronal timescale diversities and demonstrate that neural heterogeneity disrupts intrinsic coherent spatiotemporal patterns ( 13 , 25 ), induces firing rate heterogeneity, enhances local responsive sensitivity, and aligns neural network activity closely with inputs. This process induces input-related dynamical modes, expands representation space, and forms high-dimensional input-slaved trajectories that reliably represent the input and ensure robust neural information processing against noise, despite neuronal spikings being variable over time and across repeated stimulation trials ( 35 – 38 ). Thus, we have uncovered a dynamical mechanism that reconciles inherent cortical variability with reliable representation of external inputs. This mechanism also explains the similar roles of other neural heterogeneities, such as nonuniform input connections, spike threshold heterogeneity, and network in-degree heterogeneity, serving as a potentially general framework for understanding the role of various forms of neural heterogeneities in reliable representation and computation. Roles of neural heterogeneity Neural heterogeneity, which includes diversity in morphology, type, excitability, connectivity, ion channel distributions, and timescales, is ubiquitous in the brain ( 18 – 20 ). Neuron-to-neuron variability in molecular, genetic, and physiological features is increasingly recognized as a critical aspect of brain function. This biophysical diversity enriches the neural system’s dynamical repertoire, fostering individual variability and complex neural dynamics. Complexity and heterogeneity are fundamental to neurobiological systems, manifesting at every level and process, and are intricately linked to the systems’ emergent collective behaviors and functions. Numerous hypothetical roles of neural heterogeneity ( 39 ) have been suggested, including the following: (i) population coding: enhancing the neuron population’s ability to encode information efficiently ( 40 , 41 ); (ii) reliability: increasing the reliability of neural computations ( 13 , 42 ); (iii) working memory: supporting complex cognitive functions ( 43 – 45 ) and functional specialization in neural computation ( 24 ); and (iv) reduction in pathological synchronization: preventing the onset of pathological states such as epilepsy ( 22 , 46 ). Neural heterogeneity profoundly influences the functional repertoire and adaptability of neural circuits. Recent computational studies underscore the importance of spike threshold heterogeneity in enhancing neural network capabilities ( 24 ), such as efficient and robust encoding of stimuli ( 16 , 47 , 48 ), improving information transmission and learning ( 49 ), and maintaining robustness and persistence of brain function over time and in the face of changing environments (resilience) ( 22 ). On the other hand, neural heterogeneity of the membrane time constants of neurons makes learning complex tasks more stable and robust ( 17 ), indicating that this observed heterogeneity is a vital component of adaptive neural systems. Moreover, reservoir computing models equipped with diverse timescales outperform homogeneous models in both short- and long-term predictions of multiscale chaotic dynamics, illustrating the value of neural heterogeneity in optimizing temporal processing ( 23 ). Despite these advancements, a comprehensive dynamical mechanism that unified the roles of various forms of neural heterogeneities and explains their contributions to reliable representation and computation remains to be established. Our study addresses this gap by first investigating the influence of timescale diversity on network dynamics ( 50 ) and its role in shaping reliable neural representation ( 24 ). Our findings highlight the substantial role of timescale diversity in reliable representation and computation, achieved by dissociating two key timescales: leakage time constant τ L and responsive time constant τ Γ ( 18 , 24 ), as their correlation deteriorates their individual contributions to the reliability. We further uncover the underlying dynamical mechanism: local responsive sensitivity and globally input-slaved transient dynamics, which can also account for the roles of other forms of neural heterogeneities: nonuniform input connections, spike threshold heterogeneity, and network in-degree heterogeneity. Thus, our study provides a cohesive framework that integrates diverse forms of heterogeneities, shedding light on the variability across multiple neuronal features, opening a new window and providing deeper insights into the role of neural heterogeneity, and potentially unifying the underlying dynamical mechanism for their roles in reliable representation and computation. Input-slaved transient dynamics The brain continuously receives and integrates vast amounts of environmental information, from sensory perception to task-driven behaviors and decision-making processes. Initial studies on neural encoding focused on individual neurons, revealing that certain neurons selectively respond to specific stimulus features such as spatial or temporal frequency, orientation, position, or depth, leading to sparse encoding. However, advancements in recording technologies have shifted research toward the collective dynamics of neural populations ( 51 ). The trajectories of these populations are typically constrained within low-dimensional manifolds in the high-dimensional space of neural activity ( 9 ), suggesting that the entire population dynamically encodes stimulus variables in a reduced or coarse-grained neural state space. Robustness and reproducibility of sequential spatiotemporal responses are essential features of many neural circuits in the sensory and motor systems ( 9 , 52 ). Traditional dynamical regimes, such as fixed points, limit cycles, chaotic attractors, and continuous attractors, do not adequately describe reproducible transient sequential neural dynamics. Instead, a stable heteroclinic sequence (SHS) ( 53 , 54 ), although not an attractor, has been proposed to model these dynamics. SHS trajectories are highly sensitive to input but remain globally stable, with small deviations corrected by distributed stabilizing effects within the neural population. When the input is withdrawn, the system relaxes back to its baseline state. However, SHS is merely an illustrative model, requiring manual design rather than self-organization, and it lacks a solid biophysical basis. Besides, RNNs trained on cognitive tasks have shown that low-dimensional subspaces naturally emerge to support flexible computations at the population level ( 55 , 56 ). However, the internal connectivities of these artificial networks do not adhere to biological constraints such as Dale’s principle, which states that neurons must be either excitatory or inhibitory. Here, we found a self-organized dynamical mechanism to reliably and robustly produce such input-slaved trajectories for representing the input, based on biologically plausible E-I recurrent SNNs with neural heterogeneity. With no need for predesigned SHS-like dynamical structures ( 57 ), neural heterogeneity disrupts intrinsic coherence between neurons, rendering them sensitive to input for flexible computations. The formed trajectories are transient and nonstationary ( Fig. 7C ), entirely input induced, and embedded in the complex spatiotemporal neural activities. Despite its distributed coding of the input ( Fig. 6C ) and substantial variability of neuronal activities, neural heterogeneity effectively reduces the trial-to-trial variability of these trajectories in the noisy system. Thus, ubiquitous neuronal diversity provides a natural biophysical basis for reliable neural information processing. Impacts on neuromorphic computing Machine learning, particularly through artificial neural networks, has revolutionized computational applications across various fields by excelling in tasks such as regression, classification, prediction, and generation. However, optimizing these networks demands substantial computational resources and energy. Neuromorphic computing ( 58 , 59 ), which uses spiking potentials and waves instead of digital bits, offers a promising alternative by leveraging the inherent efficiencies of biological systems. Recent advancements ( 60 – 62 ) have demonstrated the potential of using nonlinear waves, such as rogue waves and solitons, for neuromorphic computing. These waves’ complex interactions make them particularly suitable for designing reservoir computing. By encoding information onto these waves, it is expected to substantially enhance the development of machine-learning devices leveraging wave dynamics. Our study adds to this innovative field by elucidating the dynamical mechanism that underpins reliable representation and computation in biologically plausible SNNs characterized by neural heterogeneity. We found that neural heterogeneity disrupts coherent spatiotemporal patterns, thereby enhancing local sensitivity. This mechanism offers neurons liquid-like superior sensitivity, making them prone to capturing input information efficiently. It also renders the system to generate input-related dynamical modes that are intrinsically nonlinear waves ( Fig. 7 ), thus expressing the complexity necessary for learning from a dataset. This understanding of neural heterogeneity and its role in reliable computation provides a new avenue for designing reservoir computing models. By incorporating liquid wave reservoirs, one can leverage the benefits of neuromorphic computing while maintaining the complexity and efficiency needed for high-dimensional data representation and processing. Our discoveries, integrated with the principles of wave dynamics in neuromorphic computing, pave the way for developing advanced reservoir computing that leverages the efficiency and robustness of biological neural information processing. Future work Reliable representation is fundamental for subsequent neural processes such as memory, learning, and decision-making. In this study, basic E-I recurrent SNNs with neural diversity are used to investigate the underlying dynamical mechanisms of reliable representation and computation. This mechanism, characterized by local sensitivity and globally input-slaved transient dynamics due to neural heterogeneity, is expected to persist when the model incorporates more biological details, such as synaptic timescales and plasticity. Specifically, with synaptic timescales mediated by N -methyl- d -aspartate receptor currents, the input-slaved trajectories observed in our study can form the dynamical basis for transient memories, allowing for the temporary storage of information and its flexible use in decision-making ( 63 – 65 ). In addition, short-term plasticity can induce traces of hidden states essential for working memory ( 66 – 70 ). Therefore, our mechanism, which induces high-dimensional input-slaved trajectories, can serve as a substrate for working memory. We plan to investigate these issues further in the future."
} | 4,392 |
31477760 | PMC6718658 | pmc | 9,313 | {
"abstract": "Rhodopseudomonas palustris CGA009 is a purple non-sulfur bacterium that can fix carbon dioxide (CO 2 ) and nitrogen or break down organic compounds for its carbon and nitrogen requirements. Light, inorganic, and organic compounds can all be used for its source of energy. Excess electrons produced during its metabolic processes can be exploited to produce hydrogen gas or biodegradable polyesters. A genome-scale metabolic model of the bacterium was reconstructed to study the interactions between photosynthesis, CO 2 fixation, and the redox state of the quinone pool. A comparison of model-predicted flux values with available Metabolic Flux Analysis (MFA) fluxes yielded predicted errors of 5–19% across four different growth substrates. The model predicted the presence of an unidentified sink responsible for the oxidation of excess quinols generated by the TCA cycle. Furthermore, light-dependent energy production was found to be highly dependent on the quinol oxidation rate. Finally, the extent of CO 2 fixation was predicted to be dependent on the amount of ATP generated through the electron transport chain, with excess ATP going toward the energy-demanding Calvin-Benson-Bassham (CBB) pathway. Based on this analysis, it is hypothesized that the quinone redox state acts as a feed-forward controller of the CBB pathway, signaling the amount of ATP available.",
"conclusion": "Conclusion In this study, a genome-scale metabolic network ( iRpa 940) was used to propose a system-wide mechanistic model of the interactive system that includes photosynthesis, carbon dioxide fixation, and the quinone redox state. The model was validated using experimental genome essentiality data 31 (84% accuracy) and flux measurement data 13 , 14 for four carbon sources (5–19% prediction error). Model simulations predicted the presence of an unidentified quinol sink. Predictions also indicated that the extent of CO 2 fixation is dependent on the amount of ATP present, with the quinone redox state acting as a feed-forward signal to the CBB system. Going forward, the proposed mechanism can be used to generate strategies for engineering strains capable of more efficiently harnessing photosynthetic energy, and that have the ability to reroute energy towards bio-production and lignin valorization. Future experimental work will be conducted to measure the electron transport rate, intracellular ATP concentration, and RuBisCO gene expression across different quinone redox states to strengthen the proposed hypothesis and further refine the model.",
"introduction": "Introduction Purple non-sulfur bacteria (PNSB) are considered to be among the most metabolically versatile groups of bacteria 1 , 2 . Within this class, Rhodopseudomonas palustris CGA009 (hereafter R . palustris ) demonstrates this elasticity through its ability to survive in a myriad of diverse environmental conditions 3 . It can grow either aerobically or anaerobically, utilize organic (heterotrophic) or inorganic (autotrophic) carbon sources, and exploit light to obtain energy when growing anaerobically 3 . Several interesting features have been observed in this bacterium, such as its consumption of fatty acids, dicarboxylic acids, and aromatic compounds including lignin breakdown products (LBPs) 4 – 6 . It is also one of two known bacteria that can express three unique nitrogenases, each with a different transition-metal cofactor 7 . Furthermore, this metabolically versatile strain’s genome encodes the aerobic and anaerobic pathways for three of the four known strategies that microbes use to break down aromatic compounds, such as LBPs 8 . Harnessing R . palustris ’ unique metabolic versatilities for the conversion of plant biomass to value-added products, such as polyhydroxybutyrate (PHB) 9 , n-butanol 10 , and hydrogen 11 , 12 , has garnered increasing interest. However, lack of a systems-level understanding of how the bacterium’s complex web of metabolic modules operates in response to environmental changes is hindering the development of the PNSB as a biochemical chassis. Several studies conducted on R . palustris showed that in addition to the Calvin-Benson-Bassham (CBB) cycle’s role of carbon assimilation during autotrophic growth, the pathway plays a major role in maintaining redox balance under heterotrophic conditions 10 , 12 – 14 . It was shown that heterotrophic growth of the PNSB on substrates that are more reduced than biomass, such as LBPs, is dependent on the availability of an electron sink 13 . CO 2 -fixation using the enzyme ribulose-1,5-biphosphate carboxylase/oxygenase (RuBisCO), nitrogen-fixation through the enzyme nitrogenase 12 , and supplementation with an electron acceptor (e.g., trimethylamine-N-oxide (TMAO)) 15 all prevent the inhibitory accumulation of excess reducing agents. Therefore, the use of CO 2 as a redox balancing strategy for the conversion of plant biomass to value-added products is an attractive approach that could increase profitability while improving sustainability. However, the complex interplay between the electrons supplied by the catabolism of different carbon sources, CO 2 fixation, and the cyclic electron flow during photosynthesis is not fully understood, thus diminishing the ability to engineer this promising bacterium. A Genome-Scale Metabolic Model (GSMM) provides a mathematical representation of an organism’s metabolic functionalities 16 , 17 by translating the repertoire of biochemical transformations into a stoichiometric matrix 18 . Due to the underdetermined nature of metabolic networks, optimization tools are used to predict reaction rates for a pre-specified objective function, such as the maximization of biomass 19 . One of the most common optimization tools used to model metabolism is Flux Balance Analysis (FBA). FBA performs a pseudo-steady state mass balance for each metabolite in the network to predict the maximum growth rate and corresponding reaction fluxes during the cell’s exponential growth phase 20 – 24 . Due to the high dimensionality of the network, other tools such as Flux Variability Analysis (FVA) are used to determine the sensitivity of growth rate as a function of each reaction flux 25 . Finally, a modified FBA formulation can be used to predict the set of essential genes under a specified growth condition 26 . Thus far, a limited number of small-scale metabolic reconstructions have been developed for PNSB, examining either the central carbon metabolism 27 or the electron transport chain 28 . However, these models are limited in scope, as they consider less than 4% of the organism’s metabolic functionality and are therefore incapable of capturing system-wide interactions between different metabolic modules. Very recently, a GSMM of the bacterium was reconstructed and used to test an array of cellular objectives during phototrophic growth. Anaerobic growth on acetate, benzoate, and 4-hydroxybenzoate was simulated using eight different biologically relevant objective functions 29 . The model predicted that the organism primarily optimized for growth, ATP production, and metabolic efficiency 29 . However, the model could be improved further by integrating recently annotated metabolic pathways for lignin monomer degradation 30 , as well as making use of experimental data on gene essentiality 31 and metabolic flux analysis for growth under different carbon sources 13 , 14 to validate and refine the network. In this work, a GSMM of R . palustris ( iRpa 940) was constructed to model the bacterium’s metabolic functionality under different environmental conditions. The model was used to simulate growth on different carbon sources and showed excellent agreement with experimentally measured fluxes 13 , 14 . Gene essentiality analysis was also performed for aerobic and anaerobic growth on acetate. The predicted essential genes were compared with available trans-mutagenesis data 31 , and an accuracy of 84% was achieved. After the model indicated the presence of an unidentified quinol sink, in silico simulations were combined with published in vivo flux measurements 13 , 14 to study the effect (and the extent) of the quinone redox state on cellular growth, electron transport rate, and CO 2 fixation. It was observed that an increase in the quinol oxidation rate resulted in an increase in the electron transport rate, and therefore ATP generation. These results suggest that redox state acts as a feed-forward controller of the highly energy-demanding CBB cycle by regulating the rate of light-generated ATP. Overall, an understanding of the metabolic control points of this interconnected system constitutes the first step towards engineering strains capable of more efficiently harnessing photosynthetic energy and rerouting this energy towards bio-production and lignin valorization.",
"discussion": "Results and Discussion Model Reconstruction and validation A summary of the iRpa 940 model’s major statistics is shown in Fig. 1A . Overall, the 940 genes associated with 1393 model reactions account for 62% of the genes involved in energy metabolism, biosynthesis, carbon & nitrogen metabolism, and cellular processes in R . palustris’ genome 3 . Figure 1B shows the relative molar abundance of each macromolecular class in R . palustris 13 . This data was used to calculate the stoichiometric coefficients of components in the model’s biomass equation (see Methods). Thus, an initial high-confidence model containing 540 genes and 915 reactions with no orphan reactions was constructed. The gap-filling procedure was carried out next in KBase 32 using reactions from the ModelSEED database 33 . Out of the 478 reactions added during gap-filling, 368 were annotated using information from organism-specific databases (see Methods). A breakdown of the number of GPR relationships established from each annotation source is shown in Fig. 1C . This resulted in the addition of 328 annotated and 110 unannotated (orphan) reactions. The inclusion of these reactions was necessary to ensure biomass production. Figure 1 Summary of the iRpa 940 model statistics and validation. ( A ) Overall model statistics. (B) Model biomass component compositions. (C) Sources of gene annotation. (D) Gene essentiality analysis results. G: Growth (non-essential gene), NG: No Growth (essential gene). pFBA was used to simulate growth on a number of different carbon sources, including carboxylic acids (acetate, fumarate, succinate and butyrate) and lignin monomers. pFBA is analogous to FBA but adds an outer objective that minimizes the sum of all reaction fluxes (see Methods). This is justified by the assumption that cells synthesize the minimum amount of cellular machinery required to maintain the maximal growth rate 37 . Simulating growth using pFBA has two main advantages over FBA. First, pFBA avoids unrealistic flux predictions for reactions participating in thermodynamically infeasible cycles (TICs) 43 . TICs are usually removed from GSMMs to avoid false predictions; however, when analyzing highly connected networks like that of R . palustris , removing these cycles can lead to the model missing certain functionalities and metabolic modes utilized by the organism. pFBA avoids these false predictions by the additional constraint that reaction fluxes should be minimized. Second, the pFBA formulation results in a significantly reduced set of optimal solutions compared to FBA. Flux Balance Analysis usually results in a large number of alternate optimal solutions (especially in highly connected networks), most of which are not biologically relevant, and can therefore lead to false conclusions 44 . pFBA’s additional objective greatly restricts the solution space and leads to more biologically insightful conclusions 37 . In silico gene essentiality analysis identified 368 essential reactions, out of which 249 were associated with gene annotations in the model. These essential reactions were then compared with in vivo gene essentiality data for aerobic growth on acetate 31 to check the model accuracy (Fig. 1D ). The calculated sensitivity and false negative rate (FNR) were consistent with recently published GSMMs 45 , 46 . Moreover, given that this is a non-model organism with no well-characterized close relatives, high-confidence annotation was not available for the less-studied pathways. Therefore, an automated pipeline like GrowMatch 47 could not be implemented with justifiable accuracy to further improve essentiality predictions. The effect of the quinone pool on light uptake, carbon dioxide fixation, and growth During initial phototrophic growth simulations, growth on any of the four carbon sources (acetate, fumarate, succinate, and butyrate) was observed to be hindered due to the accumulation of excess quinols formed in the TCA cycle. Flux analysis of the electron transport chain (ETC) revealed that the rate of quinol oxidization through the cytochrome bc1 complex was equivalent to the rate of quinone reduction in the Reaction center (RC). This result is consistent with previous studies in PNSB 28 , and is necessary for steady-state flow of electrons through the cyclical chain. Furthermore, previous studies on the activity of the ETC concluded that the thermodynamically unfavorable process of reverse electron transfer through NADH dehydrogenase had very low activity compared to the rate through the RC 28 , 48 . Therefore, this reaction could not account for the oxidation of the excess quinols produced in the TCA cycle. Since no other high-confidence reaction was found to consume quinols in R . palustris , a quinol “sink reaction” was added to the iRpa 940 model. Sink reactions are often incorporated into metabolic models when a metabolite is known to be produced during metabolism but for which no means of consumption have been identified 49 , or to describe the accumulation of a storage compound 49 (e.g. glycogen). Furthermore, recent experimental work with R . palustris TIE-1 reported the presence of an unidentified quinol-oxidizing reaction that had not been accounted for previously 48 , giving further support to this prediction. pFBA simulations were conducted under different quinol sink rates to qualitatively predict how changes in the quinone redox state affected the rest of the metabolic network. The quinol sink reaction was treated as a parameter in the model and pFBA simulations were conducted at varying quinol oxidation (sink) rates to determine how light uptake ( i . e . Electron Transport Rate or ETR), growth, and CO 2 fixation are affected by changes in the quinone redox state (Fig. 2 ). Carbon uptake was restricted to a maximum value of 100 mmol/gDW/hr for acetate and 50 mmol/gDW/hr for fumarate, succinate, and butyrate to ensure the same number of carbons were being taken up. MFA values were scaled to the same carbon uptake rates 13 , 14 . For growth on butyrate, the supplementation of CO 2 is required for growth, as the substrate is more reduced than biomass and requires an electron sink 14 . The media was supplied with CO 2 at a maximum uptake rate of 32.1 mmol/gDW/hr to match MFA observations. Since steady-state GSMMs cannot capture metabolite concentrations, the redox state cannot be quantified directly. Instead, the qualitative behavior of the redox state was predicted by varying the rate of the quinol sink. As the quinol oxidation rate increases, the quinone pool becomes more oxidized. Using experimental MFA data 13 , 14 , the quinol oxidation rate was predicted for each of the four substrates (Table 1 ). These values were calculated by minimizing the sum of errors between the in silico- generated pFBA fluxes and the MFA flux values. The table also shows the quinone reduction rate through the TCA cycle for each carbon source. The percentage of CO 2 fixed was defined as the rate of CO 2 fixation divided by the total rate of CO 2 produced metabolically. Figure 3 shows the resulting flux predictions obtained at the predicted quinol oxidation rates for growth on acetate (Fig. 3A ), and the calculated percent errors of these predictions for each carbon substrate (Fig. 3B ). A comparison of flux predictions with MFA values for the other three carbon sources is provided in Supplementary File 2 (see Figs S2 – S4 ). Figure 2 Effect of the Quinol sink rate on: ( A ) Light uptake rate, ( B ) Growth rate, ( C ) Carbon source uptake rate, and ( D ) Carbon fixation rate for growth on four carbon sources. ace: acetate, but: butyrate, suc: succinate, fum: fumarate. Table 1 Predicted reaction rates for growth on four different carbon sources. Carbon source QH 2 oxidation rate (mmol/gDW/hr) Q reduction rate a (mmol/gDW/hr) Electron transport rate (dmol/gDW/hr) CO 2 fixation rate (mmol/gDW/hr) % CO 2 fixed b Net CO 2 excretion rate (mmol/gDW/hr) Acetate 52.5 39.1 5.3 29.7 73.2 10.9 Butyrate 54.9 37.4 5.4 57.8 — −18.6 c Succinate 49.8 49.2 3.0 35.6 50.6 34.8 Fumarate 0 0 2.3 17.3 25.1 51.5 a The rate of quinone reduction in the TCA cycle. b The rate of CO 2 fixation divided by the rate of total CO 2 produced. c CO 2 was supplied in the media during growth on butyrate. Figure 3 Comparison of model-predicted vs MFA-generated flux values for reactions involved in central carbon metabolism. ( A ) Metabolic flux map showing reaction rates for growth on acetate (B) Percentage error between model predictions and MFA flux values for growth on four carbon sources. For growth on acetate and butyrate, light uptake ( i . e . ETR) showed two distinct regions based on the extent of quinol oxidation (Fig. 2 ). Under low oxidation rates, flux through the quinol-producing succinate dehydrogenase reaction was avoided by using the glyoxylate shunt and subsequently the CBB cycle. Therefore, both light uptake and CO 2 fixation increased rapidly in this region. In the second region, at high quinol oxidation rates, flux shifted toward the oxidative TCA cycle. Therefore, in this region, both the Electron Transport Chain (ETC) activity and the rate of CO 2 fixation decreased with increasing quinol oxidation. Furthermore, as can be seen from Table 1 , the ratio of quinol oxidation rate to quinone reduction rate was similar for both carbon sources. Due to the supplementation of CO 2 during growth on butyrate, the percentage of CO 2 fixation could not be calculated. During growth on succinate, the production of quinols through succinate dehydrogenase could not be avoided, therefore light uptake rate increased linearly with the rate of quinol oxidation. Moreover, the rates of quinol oxidation and quinone reduction were equivalent, indicating that the quinone pool was more reduced when compared to the redox state during growth on acetate and butyrate. This led to a reduced electron flow through the ETC, and subsequently lower ATP generation. Finally, the model predicted that during growth on the highly oxidized (compared to cell biomass) carbon source fumarate, the rate of the quinol sink did not affect the flux distribution. A similar parameter sampling procedure was performed to determine the effect of light uptake on growth. Light uptake rate was set as a parameter and the quinol oxidation rate was fixed to the value predicted based on MFA fluxes (Fig. 4 ). Again, there were two distinct growth regions: (i) a low-light (LL) energy-limited region, and (ii) a high-light (HL) carbon-limited region. In the LL region, growth was highly dependent on the amount of light available and the model predicted that all of the ATP produced was used to convert the carbon source into biomass precursors. Therefore, no ATP remained for the energy-intensive CBB pathway. In the HL region, when the maximum substrate uptake rate was reached, the carbon source could not be incorporated any faster. The additional energy produced from light was then directed towards CO 2 fixation. Although the model predicted that the rate of CO 2 fixation increased linearly with light uptake rate, kinetic and thermodynamic constrains on the highly inefficient CO 2 -fixing RuBisCO enzyme 50 hinders this process at high light uptake. Figure 4 Effect of the light uptake rate on ( A ) Growth rate, ( B ) Carbon source uptake rate, ( C ) Carbon fixation rate, and ( D ) Carbon dioxide excretion rate for growth on four carbon sources. ace: acetate, but: butyrate, suc: succinate, fum: fumarate. In A, B, and D, the lines for succinate and fumarate lie on top of each other. Proposed mechanism for the interplay between the quinone redox state, the electron transport rate, and CO 2 fixation Based on how the quinol oxidation rate effected the light uptake and the model’s flux distribution, a mechanistic explanation of the system-wide metabolic interactions can be postulated. During steady-state operation of the cyclic ETC, the flux through the quinone reducing RC and quinol oxidizing cytochrome bc1 complex are coupled to ensure a constant rate of electron flow through the cycle 28 . Therefore, as shown in Fig. 5 , increased flux through the oxidative TCA cycle leads to the accumulation of reduced quinols. This in turn leads to a restriction in the flow of electrons through the ETC and consequently in the amount of ATP produced. The CBB system thus lacks the energy required to fix CO 2 . Therefore, the quinone redox state is predicted to act as a feed-forward controller to the energetically expensive CBB pathway, indicating how much ATP is available at a given condition. Figure 5 Schematic of a proposed mechanism for the interaction between the quinone redox state, electron transport rate, and carbon fixation. ( A ) High rate of quinol oxidation. ( B ) Low rate of quinol oxidation. Comparison of pFBA-generated growth simulations with MFA data led to the hypothesis that an unidentified quinone oxidoreductase reaction has to occur to obtain the observed flux distribution. A previous study on the PNSB R . capsulatus suggests that complex I, the NADH:quinone oxidoreductase enzyme, is responsible for the observed quinol oxidation through reverse electron flow 51 . However, the model predicted that the rate of quinol oxidation required cannot be accounted for through complex I only, which showed low activity. Furthermore, based on the high thermodynamic cost of reverse electron flow, it appears unlikely that it can account for the predicted rate of quinol oxidation 28 . Although the source of quinol oxidation (sink) is yet to be identified, there are a number of candidate reactions that could perform this role. Primarily, the malate:quinone dehydrogenase (MDH) appears to be a potential reaction for oxidizing excess quinols. In the forward direction, this reaction converts malate into oxaloacetate and produces ubiquinol in the process. A second NAD-dependent malate dehydrogenase is also coded for by R . palustris and could perform the same function. Knocking out and over-expressing these enzymes could be employed to investigate their role in ETR, ATP production, and CO 2 fixation."
} | 5,752 |
30613467 | null | s2 | 9,314 | {
"abstract": "Adaptive laboratory evolution (ALE) has emerged as a powerful tool in basic microbial research and strain development. In the context of metabolic science and engineering, it has been applied to study gene knockout responses, expand substrate ranges, improve tolerance to process conditions, and to improve productivity via designed growth coupling. In recent years, advancements in ALE methods and systems biology measurement technologies, particularly genome sequencing and "
} | 119 |
21435029 | null | s2 | 9,315 | {
"abstract": "The adapter protein MecA targets the transcription factor ComK for degradation by the ClpC/ClpP proteolytic complex, thereby negatively regulating competence in Bacillus subtilis. Here we show that MecA also decreases the frequency of transitions to the sporulation pathway as well as the expression of eps, which encodes synthesis of the biofilm matrix exopolysaccharide. We present genetic and biophysical evidence that MecA downregulates eps expression and spore formation by directly interacting with Spo0A. MecA does not target Spo0A for degradation, and apparently does not prevent the phosphorylation of Spo0A. We propose that it inhibits the transcriptional activity of Spo0A∼P by direct binding. Thus, in its interaction with Spo0A, MecA differs from its role in the regulation of competence where it targets ComK for degradation. MecA acts as a general buffering protein for development, acting by two distinct mechanisms to regulate inappropriate transitions to energy-intensive pathways."
} | 249 |
35849638 | null | s2 | 9,318 | {
"abstract": "The automation of liquid-handling routines offers great potential for fast, reproducible, and labor-reduced biomaterial fabrication but also requires the development of special protocols. Competitive systems demand for a high degree in miniaturization and parallelization while maintaining flexibility regarding the experimental design. Today, there are only a few possibilities for automated fabrication of biomaterials inside multiwell plates. We have previously demonstrated that streptavidin-based biomimetic platforms can be employed to study cellular behaviors on biomimetic surfaces. So far, these self-assembled materials were made by stepwise assembly of the components using manual pipetting. In this work, we introduce for the first time a fully automated and adaptable workflow to functionalize glass-bottom multiwell plates with customized biomimetic platforms deposited in single wells using a liquid-handling robot. We then characterize the cell response using automated image acquisition and subsequent analysis. Furthermore, the molecular surface density of the biomimetic platforms was characterized "
} | 279 |
28591840 | PMC5812545 | pmc | 9,320 | {
"abstract": "Abstract Deeper sequencing and improved bioinformatics in conjunction with single-cell and metagenomic approaches continue to illuminate undercharacterized environmental microbial communities. This has propelled the ‘who is there, and what might they be doing’ paradigm to the uncultivated and has already radically changed the topology of the tree of life and provided key insights into the microbial contribution to biogeochemistry. While characterization of ‘who’ based on marker genes can describe a large fraction of the community, answering ‘what are they doing’ remains the elusive pinnacle for microbiology. Function-driven single-cell genomics provides a solution by using a function-based screen to subsample complex microbial communities in a targeted manner for the isolation and genome sequencing of single cells. This enables single-cell sequencing to be focused on cells with specific phenotypic or metabolic characteristics of interest. Recovered genomes are conclusively implicated for both encoding and exhibiting the feature of interest, improving downstream annotation and revealing activity levels within that environment. This emerging approach has already improved our understanding of microbial community functioning and facilitated the experimental analysis of uncharacterized gene product space. Here we provide a comprehensive review of strategies that have been applied for function-driven single-cell genomics and the future directions we envision.",
"introduction": "INTRODUCTION Metagenomics has evolved through the need to better characterize our world by studying the composition and coding potential of complex microbial communities where a majority of the diversity is not yet cultivated. Facilitated by next-generation sequencing platforms, metagenomics can provide incredible sequence output from a minimally disturbed complex environmental sample, making it the gold standard for recovering DNA sequences from samples of interest. Despite the magnitude of the output, the logistics of metagenome assembly are complicated and many community features can be lost within the sequence data. Furthermore, annotating assembled sequences depends on homology-based computational tools to infer function with varying degrees of ‘homology creep’ (Woyke and Jarett 2015 ), and provides no information regarding the activity of that organism within the environment unless coupled with experimental approaches like metatranscriptomics, proteomics or stable isotope probing. Even when an ‘all of the above’ approach is taken, the final result is the average of a bulk population, masking the large variability in functional activity due to spatial microenvironments or phenotypic noise known to drive divergent functions even in clonal populations (Ackermann 2013 ). Thus, the strength of metagenomics for its indiscriminate capture of all DNA from a heterogeneous sample is also its pitfall in that the complex output can confound the recovery of complete genomes, obscure genomic population heterogeneities (Engel, Stepanauskas and Moran 2014 ), overlook rare organisms (Ainsworth et al. 2015 ), miss ecological relationships, blend spatially divergent populations and diminish perspective for the interpretation of recovered sequences of unknown function. Single-cell genomics offers a complimentary approach to metagenomics by capturing and amplifying DNA from a single isolated cell, as opposed to many cells in metagenomics, and is routinely capable of producing partial (Zhang et al. 2006 ), and even complete (Woyke et al. 2010 ), non-composite genomes. This approach thus addresses some of the caveats of metagenomics at the sacrifice of throughput. Although technical limitations and inherent biases currently exist for single-cell approaches (Lasken 2012 ; Gawad, Koh and Quake 2016 ), within microbiology single-cell genomics has demonstrated itself as a powerful tool by generating genomes from rare (Martijn et al. 2015 ), symbiotic (Siegl et al. 2011 ) and previously uncharacterized microbial lineages (Marcy et al. 2007 ; Blainey et al. 2011 ; Rinke et al. 2013 ). In addition, auxiliary DNA from viruses, organelles, plasmids, transformed DNA and symbionts is also captured and sequenced along with the host DNA due to its colocalization with the cell (Stepanauskas 2015 ). Thus, associations between these genetic elements and the host organism of interest are maintained with high resolution, allowing characterization of subtle ecological interactions (Yoon et al. 2011 ; Roux et al. 2014 ). When coupled with microscopy, single-cell genomics enables spatiotemporal characterization of microscale environments, improving the resolution at which our understanding of microbial ecology takes place (Landry et al. 2013 ). Single-cell genomics has identified a novel genetic code (Campbell et al. 2013 ), uncovered unexplored protein sequence space relevant for biotechnology and human medicine (Wilson et al. 2014 ), and facilitated the expansion of novel branches in the bacterial and archaeal tree of life and improved phylogenetic read anchoring for metagenomic data sets (Rinke et al. 2013 ). With these efforts, pipelines for the single-cell isolation and sequencing of environmental microbes have optimized their efficiency and throughput (Hutchison et al. 2005 ; Woyke et al. 2011 ; Rinke et al. 2014 ), thereby advancing single-cell genomics methods to the next level. The major outstanding limitation in metagenomic and single-cell approaches is the overreliance on using predicted protein function as a proxy for functional activity in the environment. These approaches lack the ability to discern which recovered populations are active, potentially overestimating the importance of abundant organisms while overlooking significant ecological contributions of lower abundance organisms (Martinez-Garcia et al. 2012 ). This caveat becomes even more pronounced when trying to study ‘microbial dark matter’ i.e., microbes that lack characterized cultured representatives (Rinke et al. 2013 ; Kamke et al. 2014 ; Brown et al. 2015 ) and whose genes often lack functional prediction (Youssef et al. 2011 ; Kantor et al. 2013 ; McLean et al. 2013 ). Thus, although both metagenomic and single-cell genomic approaches can recover novel genes, neither approach currently provides insight into the function of the gene or activity level of the organism. Identifying the function of a gene has traditionally relied on classical genetic knockout/complementation studies, and more recently high-throughput relative fitness studies (Wetmore et al. 2015 ). Because these approaches require compatible genetic toolkits and cultivable organisms, they are unsuitable for studying the function of unknown genes from the uncultivated majority. Function-based screens of metagenomic DNA sequences have previously utilized clone libraries to heterologously express environment-derived DNA fragments in a gain-of-function approach (Daniel 2005 ). Escherichia coli , the workhorse of heterologous screening approaches, was computationally determined to be able to transcribe ∼40% of genes from well-known cultivated lineages of microbes with the typical expression library approach (Gabor, Alkema and Janssen 2004 ). Considering downstream incompatibilities such as codon bias (Tuller et al. 2010 ), strategic rare codon utilization (Komar et al. 1999 ), required metabolite pools, post-translational modification, accessory proteins, apoprotein activation, secretion and even the availability of a compatible read-out assay, it becomes apparent that only a small fraction of functional space is accessible through this approach. Furthermore, when introducing more divergent DNA sequences, such as those from candidate phyla, the success rate for accessing the vast functional diversity that exists with current tools is expected to rapidly diminish. Thus, the majority of unknown functions from uncultivated organisms remain obscured within their native expression hosts. Due to the increasing recovery and accumulation of sequences of unknown function, a context-driven approach to investigate the roles of these genes within their native hosts and environment is required. This will facilitate improved protein annotation, interpretation of microbial networks and understanding of microbial influences on biogeochemistry (Hicks and Prather 2014 ; Woyke and Jarett 2015 ). Thus, an increased focus on characterizing the functional roles and activity levels of uncultivated organisms in conjunction with downstream genomic sequencing has motivated the development of methods to ascertain both functional trait and/or phenotype, and corresponding coding potential from these organisms at the resolution of a single cell. Here we present the approaches and limitations of pioneering studies focused on developing a suite of methods for the function-driven single-cell identification, isolation and sequencing of uncultivated environmental microbes. Though this review focuses on these approaches from a largely bacterial and archaeal perspective, many of the same strategies can be adopted to small environmental eukaryotic cells, and some are being applied in human cells to address human health issues (Gao et al. 2004 ; Yu et al. 2011 )."
} | 2,321 |
29643847 | PMC5882818 | pmc | 9,321 | {
"abstract": "This study implements temporal and spatial appraisals on the operational performance and corresponding microbial community structure of a full-scale advanced anaerobic expanded granular sludge bed (AnaEG) which was used to treat low organic loading starch processing wastewater. Results showed stable treatment efficiency could be maintained with long-term erratic influent quality, and a major reaction zone located at the bottom of the AnaEG, where the main pollutant removal rate was greater than 90%. Remarkably, high-throughput sequencing of 16S rRNA gene amplicons displayed that the predominant members constructed the major part of the overall microbial community and showed highly temporal stability. They were affiliated to Chloroflexi (16.4%), Proteobacteria (14.01%), Firmicutes (8.76%), Bacteroidetes (7.85%), Cloacimonetes (3.21%), Ignavibacteriae (1.80%), Synergistetes (1.11%), Thermotogae (0.98%), and Euryarchaeota (3.18%). This part of microorganism implemented the long-term stable treatment efficiency of the reactor. Simultaneously, an extraordinary spatial homogeneity in the granule physic properties and microbial community structure along the vertical direction was observed within the AnaEG. In conclusion, the microbial community structure and the bioreactor’s performance showed notable spatial and temporal consistency, and the predominant populations guaranteed a long-term favorable treatment performance of the AnaEG. It provides us with a better understanding of the mechanism of this recently proposed anaerobic reactor which was used in low organic loading wastewater treatment.",
"conclusion": "Conclusion In this study, the temporal and spatial evaluation of the operation performance and microbial community structure of an AnaEG revealed that, the reactor operated with high efficiency and high temporal stability. The predominant populations, including Bacteroidetes, Chloroflexi, Firmicutes, and Proteobacteria, and Methanosaeta , carried out the metabolism of anaerobic hydrolysis, acidification, and methanogenesis. And this part of functional microorganisms determined the overall microbial community characteristics. No temporal variety and spatial hierarchy of microbial community structure had been detected in the AnaEG reactor. And the microbial community dynamics displayed remarkable coherence with the reactor’s performance.",
"introduction": "Introduction Starch serves as the important source for humans and industrial production, thus, it is used extensively around the world ( Avancini et al., 2007 ). Notably, more than 20 million tons of starch processing wastewater (SPW) were generated in China per year ( Lu et al., 2009 ). Unfortunately, the direct discharge of SPW would cause serious water pollution due to its high concentration of organic and chemical reagents. Consequently, numerous methods are applied for SPW treatment, especially, anaerobic biological treatment is the enormously popular one because of its many advantages. Such as high organic loading rate (OLR), low energy consumption, and renewable energy production. Among the different alternatives of anaerobic biotreatment, the expanded granular sludge bed (EGSB), an improved anaerobic treatment technology based on an upflow anaerobic sludge blanket (UASB), has attracted more attention in SPW treatment in recent years ( Fang et al., 2011 ; Fettig et al., 2013 ). However, its treatment performance stability is still unsatisfactory ( Zhao et al., 2012 ). Recently, an advanced anaerobic expanded granular sludge bed (AnaEG) which combines the advantages of UASB and EGSB technology was invented by Li et al. (2014) . In this new type reactor, the effluent does not need to be recycled back to the influent to maintain a high upward velocity. And well-dispersed upward influent flow increases the expansion grade of AnaEG reactor by the plug flow pattern. Furthermore, almost no surplus sludge need to dispose of the reactor and its adaptability to shock-load are higher than other anaerobic reactor. Consequently, the AnaEG overcomes the shortcomings of the existing anaerobic reactors. Its high efficient decomposition of organic pollutant plus remarkably reduced energy consumption displays the big advantage of this reactor. Anaerobic digestion is characteristic for its application on high concentration organic wastewater ( Li and Yu, 2016 ). However, some researchers found that the anaerobic bioreactor also could complete excellent treatment performance for the low OLR [less than 1 kg chemical oxygen demand (COD) m -3 d -1 ] industrial wastewater processing ( Zhou et al., 2009 ). And the AnaEG had been efficiently operated for treating low organic loading wastewater in lab-scale experiment (0.806 kg COD m -3 d -1 ; Li et al., 2014 ), and also has been implemented in various industrial wastewater treatment, including SPW, ethanol wastewater, pharmaceutical and chemical wastewater. The factories’ monitoring data demonstrated that high treatment efficiency could be achieved by AnaEG (data not published). Thus, the AnaEG has potential for prevail usage in various organic wastewater treatment. It might represent a new generation of anaerobic wastewater treatment processes. More important, a comprehensive knowledge to the microbial community structure of anaerobic reactor is crucial for the industrial application. Because it is not only paramount to understanding the potentials and limitations of the metabolic reactions ( McKeown et al., 2012 ), but could also be used to take precautions to prevent inhibitory processes that have the tendency to significantly disturb anaerobic bioreactor ( Niu et al., 2016 ). In the existed microbiological studies, some of them indicated that the bacterial and archaeal populations demonstrated a temporal succession when the psychrophilic anaerobic reactor was operated at increasing loading rate and VFA accumulation condition ( Connaughton et al., 2006 ). Other researchers indicated that different microbial populations could be observed in the primary and secondary metabolism of quinoline degradation ( Wang et al., 2017 ). However, others found that the bacterial diversity was quite chaotic but the archaea remained relatively stable over the whole operation period ( Perendeci et al., 2013 ). Furthermore, because anaerobic digestion is a successive metabolic procession, and different microbial communities are responsible for these metabolisms. Hence, researches revealed that microbial spatial distribution existing in anaerobic reactor. For example, a stratified microbial structure was detected in the anaerobic reactor ( Antwi et al., 2017 ) and the methanogenic, acetogenic, and hydrolytic fermentative microorganisms were predominant in the upper, middle, and bottom compartments of the reactor, respectively ( Xing et al., 2014 ). In addition, researchers found that some groups could be enriched in the special environment. For example, in a study on EGSB for sucrose wastewater treatment, Firmicutes and a H 2 -utilizing methanogen, Methanospirillum , significantly increased when the temperature gradually decreased from 20 to 10°C ( Tsushima et al., 2010 ). And others indicated that the shock load reduced the abundance of hydrogenotrophic methanogenic Methanomicrobiales, but dramatically increased the abundance of Methanobacteriales ( Bialek et al., 2013 ). In the case of AnaEG reactor, there is few study for the microbial community. Only a briefly description in the research of Li et al. (2014) , when AnaEG was used to treat coal gasification wastewater, they found that micrococci were the main microorganisms and filamentous bacteria intertwined randomly throughout the cross-section. It is regrettable that they did not make an in-depth analysis of the specific composition and distribution of these microorganisms. To the best of our knowledge, there is no other comprehensive investigation of microbial community structure on both temporal and spatial dimension of the anaerobic bioreactor, especially on the AnaEG bioreactor. Considering the unparalleled merits of AnaEG, it is worth investigating the microbial community when it was used to low organic loading wastewater processing. Hence, in this study, we investigated the performance of a full scale AnaEG reactor. The main goals in this study were to (i) assess the temporal and spatial degradation performance of a long-term operational full-scale AnaEG for SPW treatment and (ii) clarify the temporal and spatial microbial community structure that was responsible for the bioreactor’s treatment performance.",
"discussion": "Discussion The treatment performance of the AnaEG reactor indicated a high efficiency both in temporal and spatial scale. For the temporal treatment performance evaluation, many studies showed that the quality of wastewater could affect the treatment efficiency of anaerobic bioreactor. For example, in a long-term study, when the mesophilic biofilm-based EGSB was used to starch-containing wastewater treatment and simultaneous hydrogen recovery, only low treatment performance was completed (average COD removal rate was 31.1%) when it operated at the OLR of 0.125–1 g starch L -1 d -1 , pH of 3.68–4.95, and HRT of 4–24 h. Methanogens is known as a group of microorganisms who are sensitive to pH changes, low pH may be the important reason of their low COD removal rate ( Guo et al., 2008 ). In another report, short pH shock (pH decreased from 9.0 to 7.0 for 24 h) could lead a deterioration of EGSB ( He et al., 2016 ). But in our study, the AnaEG reactor could always implement desirable treatment efficiency with the average COD removal rate higher than 95%, although with variable influent, such as shock pH (5.7–8.5) and OLR (1–3.3 kg COD m -3 d -1 ). The results indicated the microbial community in our bioreactor was a resilient ecosystem and could maintain a strong functional stabilization when it treated the varied influent. Moreover, it was worth considering why the pollutants was mainly degraded at the bottom of this AnaEG reactor. In previous research ( Lu et al., 2015 ), the main reaction zone for starch hydrolysis, located between the bottom and middle areas of the UASB and the organic compounds gradually attenuated along with the upflow. The different main reaction zones in these researches may due to the structural design difference of the reactor. Because AnaEG is a plug flow pattern reactor which ensures the sufficient expansion of sludge bed. This particular design improved the contact between substrate and microorganism, and reinforced the transfer effect. In addition, VFAs analysis revealed that starch was directly degraded to acetate, and very few intermediates such as propionate and butyrate produced, which is consistent with previous research ( Schmidt and Ahring, 1996 ). Consequently, organic matter could be degraded efficiently at the bottom area, and high treatment performance could be completed even though if with the transient influent. Furthermore, in the study previously described, smaller granules had higher nitrogen removal rate than the larger ones ( Wang et al., 2014 ), the former offered more sufficient contact area between substrate and microorganism to complete preferable treatment efficiency. The granule size in our study, obviously smaller (0.9 mm) than that of ever previous described SPW anaerobic treatment reactors (2–5 mm) ( Nizami and Murphy, 2011 ; Lu et al., 2015 ). This characteristic of the AGS further strengthen the AnaEG treatment efficiency. More important, the AnaEG offered a possibility to the industrial application for a high efficient low organic loading wastewater processing, although the anaerobic reactor was traditionally recognized as especially applicable for high organic loading wastewater treatment. In term of microbial community of AnaEG, the species diversity was much higher than other anaerobic bioreactor. For example, in a thermophilic anaerobic digestion of pig slurry, the Shannon index only ranged from 3.34 to 4.76, and the maximum COD removal rate was less than 60% ( Cerrillo et al., 2017 ). In a continuous stirred tank reactor which fed with distiller’s grains and supplemented with trace elements, the Shannon index of the reactor was only 4.47–4.56 ( Wintsche et al., 2016 ). Although other study reported a relatively higher species diversity in starch wastewater, such as in the description of Antwi et al. (2017) , they found the Shannon index in different section of UASB ranged from 4.61 to 5.26 when it used to potato starch wastewater treatment. But all of them were much lower than the species diversity of AnaEG reactor in this study, the Shannon index of this reactor were ranged from 6.28 to 6.63, and its COD removal rate kept above 90%. Therefore, the higher species diversity might be one important reason of the high efficiency of AnaEG reactor. The temporal stable and spatial uniform microbial community structures in AnaEG reactor showed remarkable agreement with the reactor’s performance. But an obvious spatial distribution was observed in a similar experiment of lab-scale EGSB ( Ambuchi et al., 2016 ), in which Chloroflexi, Euryarchaeota, and Firmicutes were the dominant phyla in the reactor’s bottom, middle, and upper segments, respectively. However, no operation performance data were available to illustrate the reasons for the microbial community spatial distribution. Moreover, substantial stratification of the microbial community was also observed when the UASB in potato starch wastewater processing had a shorter HRT (36 h; Antwi et al., 2017 ). It appeared that shortening the HRT may lead to microbial spatial hierarchical distribution, and this was also shown in a previous study ( Lu et al., 2015 ). Nevertheless, the organic matter concentration was low and kept stable from the first port to the upper port of the AnaEG in this study, and also the bacterial composition was unchanged along the vertical direction. The possible reason of this phenomenon may be caused by efficiently gas stirring which produced in anaerobic digestion under the special plug flow pattern of AnaEG. Thus, the interior of the reactor could be a uniform system. Therefore, no spatial hierarchical distribution could be observed. Because the SPW is a common biodegradable wastewater, it is reasonable that we regarded the predominant populations as the authentic functional performers and it should be responsible for the temporal treatment stabilization of the AnaEG. In our study, we comprehensively investigated the temporal and spatial distribution of microbial community structure on OTU level. All the results implied that high abundant microorganisms consist of the main part of the overall microbial community. This part microorganisms should be functional crucial members. And thanks to their temporal stability, the AnaEG reactor could implement temporal stable and high efficient treatment performance. Furthermore, some of the predominant phyla detected in this study, such as Bacteroidetes, Chloroflexi, Euryarchaeota, Firmicutes, and Proteobacteria were also considered as predominant microorganisms in sugar industrial wastewater treatment ( Ambuchi et al., 2016 ). And because of the high biodegradable ability of starch, many hydrolytic and acidification reactions need to attend the anaerobic biodegradation. As expected, all the dominant phyla detected in this study were generally recognized as common anaerobic fermentation phyla ( Zhou et al., 2016 ). Chloroflexi ( Lynd et al., 2002 ) and Firmicutes ( Bertin et al., 2012 ) were often involved as critical anaerobic hydrolytic and acidifying phyla. The bacteria of the dominant genus of Mesotoga in this study had been considered to be hydrolytic bacteria ( Imachi et al., 2014 ), and Syntrophobacter seemed to possess a higher affinity to degrade propionate ( Gomes et al., 2016 ). These literatures further supports our point that the dominant community was composed of hydrolytic and acidifying microorganisms. Notably, the microorganisms of Methanosaeta which comprised most of the methanogens investigated in this work, was recognized as aceticlastic methanogens ( Imachi et al., 2014 ; Rotaru et al., 2014 ). Methanosaeta was also reported to be the predominant methanogens in a study that used lab-scale UASB to treat potato SPW ( Antwi et al., 2017 ). Considered that acetic acid was one of the major component in the SPW, and the predominant role of Methanosaeta in the reactor, we may concluded that the aceticlastic pathway predominated the methanogenesis in AnaEG reactor. Besides, many multifunctional degradable microorganisms that are common in wastewater were detected in the AnaEG reactor, such as Thauera , it was not only widely regarded as a kind of functional bacteria with denitrification, but also could degrade aromatic compounds ( Wu et al., 2015 ). In other words, the reactor may have the potential to other complex compounds degradation."
} | 4,238 |
37338270 | PMC10469762 | pmc | 9,322 | {
"abstract": "ABSTRACT Microbes commonly organize into communities consisting of hundreds of species involved in complex interactions with each other. 16S ribosomal RNA (16S rRNA) amplicon profiling provides snapshots that reveal the phylogenies and abundance profiles of these microbial communities. These snapshots, when collected from multiple samples, can reveal the co-occurrence of microbes, providing a glimpse into the network of associations in these communities. However, the inference of networks from 16S data involves numerous steps, each requiring specific tools and parameter choices. Moreover, the extent to which these steps affect the final network is still unclear. In this study, we perform a meticulous analysis of each step of a pipeline that can convert 16S sequencing data into a network of microbial associations. Through this process, we map how different choices of algorithms and parameters affect the co-occurrence network and identify the steps that contribute substantially to the variance. We further determine the tools and parameters that generate robust co-occurrence networks and develop consensus network algorithms based on benchmarks with mock and synthetic data sets. The Microbial Co-occurrence Network Explorer, or MiCoNE (available at https://github.com/segrelab/MiCoNE ) follows these default tools and parameters and can help explore the outcome of these combinations of choices on the inferred networks. We envisage that this pipeline could be used for integrating multiple data sets and generating comparative analyses and consensus networks that can guide our understanding of microbial community assembly in different biomes. IMPORTANCE Mapping the interrelationships between different species in a microbial community is important for understanding and controlling their structure and function. The surge in the high-throughput sequencing of microbial communities has led to the creation of thousands of data sets containing information about microbial abundances. These abundances can be transformed into co-occurrence networks, providing a glimpse into the associations within microbiomes. However, processing these data sets to obtain co-occurrence information relies on several complex steps, each of which involves numerous choices of tools and corresponding parameters. These multiple options pose questions about the robustness and uniqueness of the inferred networks. In this study, we address this workflow and provide a systematic analysis of how these choices of tools affect the final network and guidelines on appropriate tool selection for a particular data set. We also develop a consensus network algorithm that helps generate more robust co-occurrence networks based on benchmark synthetic data sets.",
"introduction": "INTRODUCTION Microbial communities are ubiquitous and play an important role in marine and terrestrial environments, urban ecosystems, and human health ( 1 \n - \n 7 ). These microbial communities, or microbiomes, often comprise several hundreds of different microbial strains interacting with each other and their environment, often through complex metabolic and signaling relationships ( 8 \n - \n 11 ). Understanding how these interconnections shape community structure and function is a fundamental challenge in microbial ecology and has applications in the study of microbial ecosystems across different biomes. With the advancement in DNA sequencing technologies ( 12 \n - \n 14 ), more information can be extracted from these microbial community samples than ever before. In particular, high-throughput sequencing, including metagenomic sequencing and sequencing of 16S ribosomal RNA (16S rRNA) gene amplicons (hereafter referred to as 16S data) of microbial communities, can help detect, identify, and quantify a large portion of the constitutive microorganisms of a microbiome ( 15 \n - \n 18 ). These advances have led to large-scale data collection efforts involving terrestrial ( 2 , 4 , 19 ), marine ( 1 , 3 ), and human-associated microbiota ( 7 , 20 , 21 ). This wealth of information has the potential to help us understand how communities assemble and operate. In particular, a powerful tool for translating microbiome composition data into knowledge is the construction of association (co-occurrence) networks, in which microbial taxa are represented by nodes, and frequent co-occurrences (or negative co-occurrences) across data sets are encoded as edges between nodes. While the relationship between directly measured interactions ( 22 \n - \n 24 ) and statistically inferred co-occurrence is still poorly understood ( 25 , 26 ), a significant amount of effort has gone into estimating co-occurrences from large microbiome sequence data sets ( 27 \n - \n 30 ). The importance of these networks is twofold: first, they can serve as maps that help identify hubs of keystone species ( 26 , 31 ) and the community response to environmental perturbations or underlying host conditions ( 32 ); second, they can serve as a bridge toward building mechanistic models of microbial communities, greatly enhancing our capacity to understand and control them. For example, multiple studies have shown the importance of specific microbial associations in the healthy microbiome ( 7 , 21 , 33 ) and their role in dysbiosis ( 32 , 34 , 35 ). In the context of terrestrial biogeochemistry, co-occurrence networks were shown to help understand microbiome assembly ( 36 ) and the response of microbial communities to environmental perturbations ( 37 ). One of the most frequently used avenues for inferring co-occurrence networks is the parsing and analysis of 16S sequencing data ( 26 , 38 ). Numerous software tools and pipelines have been developed to analyze 16S sequencing data, with a strong emphasis on the known limitations of this method, including resolution, sequencing depth, compositional nature, sequencing errors, and copy number variations ( 39 , 40 ). Popular methods for different phases of the analysis of 16S data include tools for (i) quality checking and trimming the sequencing reads, (ii) denoising and clustering the trimmed reads ( 41 \n - \n 43 ), (iii) assigning taxonomy to the denoised reads ( 44 ), (iv) processing and transforming the taxonomy count matrices ( 45 ), and (v) inferring the co-occurrence network ( 46 \n - \n 48 ). Different specific algorithms are often aggregated into popular online platforms (such as MG-RAST ( 49 ) and Qiita ( 50 )) and software packages (such as Quantitative Insights Into Microbial Ecology 2 (QIIME2) ( 51 )). The different methods and tools can lead to vastly different inferences of community compositions and co-occurrence networks ( 52 , 53 ), making it difficult to reliably compare networks across different publications and studies. This difference is partially due to the focus of existing platforms on operational taxonomic unit (OTU) or exact sequence variant (ESV) generation and not on the effects of upstream statistical methods on the inferred co-occurrence networks. Furthermore, no organized framework currently exists that can systematically analyze and compare each step in the pipeline for processing amplicons into co-occurrence networks. In this study, we present a standardized 16S data analysis pipeline called Microbial Co-occurrence Network Explorer (MiCoNE) that produces robust and reproducible co-occurrence networks from 16S sequence data of microbial communities and enables users to interactively explore how the network would change upon using different alternative tools and parameters at each step. Our pipeline is coupled to an online integrative tool for the organization, visualization, and analysis of inter-microbial networks called Microbial Interaction Network Database (MIND) ( 54 ), which is available at http://microbialnet.org/ . Through a systematic comparative analysis, we determine which steps of the MiCoNE pipeline have the largest influence on the final network and which choice seems to have the optimal agreement with the tested mock and synthetic data sets. These steps together with our default settings ensure better reproducibility and easier comparison of co-occurrence networks across data sets. We expect that our tool will also be useful for benchmarking future alternative methods and ensuring a transparent evaluation of the possible biases introduced by the use of specific tools.",
"discussion": "DISCUSSION Why MiCoNE? A myriad of tools and methods have been developed for different parts of the workflow for inference of co-occurrence networks from 16S rRNA data. Our analyses have shown that networks generated using different combinations of tools and approaches can be substantially different from each other, highlighting the need for a clear evaluation of the source of variability and tools that provide the most robust and accurate results. Our newly developed software, MiCoNE, is a customizable pipeline for the inference of co-occurrence networks from 16S rRNA data that enable users to compare networks generated by multiple possible combinations of tools and parameters. Importantly, in addition to revisiting the test cases presented in this work, users will be able to explore the effect of various tool combinations on their own data sets of interest. The MiCoNE pipeline has been built in a modular fashion; its plug-and-play architecture enables users to add new tools and steps, either using existing packages that have not been examined in the present work or those developed in the future. The MiCoNE Python package provides functions and methods to perform a detailed analysis of the count matrices and the co-occurrence networks. The inferred networks are exported to a custom JSON format (see Supplementary) by default but can also be exported to Cytoscape ( 77 ), GML ( 78 ), and many other popular formats via the Python package. While several tools/workflows such as QIIME2 ( 51 ) and NetCoMi ( 79 ) can be used to generate co-occurrence networks from 16S sequencing data, no single tool exists that integrates the complete process of inferring microbial interaction networks from 16S sequencing reads. MiCoNE is unique as it offers this functionality packaged in a workflow that can be run locally, on a computer cluster, or in the cloud. The default pipeline and recommended tools Through MiCoNE, in addition to transparently revealing the dependence of co-occurrence networks on tool and parameter choices (see discussion in Supplementary text for details on the DC, TA, and OP steps), we have taken advantage of our spectrum of computational options and the availability of mock and synthetic data sets to suggest a default standard setting that streamlines comparisons across data sets. Additionally, we have developed a consensus approach that can reliably generate fairly robust networks across multiple tool choices. Even if, in the current analysis, we have shown the relevance of our approach to two very different types of microbiome data sets (human and vole stool), it is important to remember that there is no universal standard for microbial interaction data, and our conclusions are based on the specific data sets used in our analysis. While our analysis is based on several mock and synthetic data sets that cover a diverse range of abundance distributions and network topologies, data sets with drastically different distributions may require a re-assessment of the best settings. However, the MiCoNE pipeline provides a platform for easy evaluation of accuracy, variance, and other properties at each workflow step for any other data set of interest. The networks generated by different network inference methods show considerable differences in edge-density and connectivity, partially due to the underlying assumptions regarding sparsity, distribution, and compositionality. To address this issue, we have developed two consensus algorithms (simple voting and SS method) that generate networks whose links have evidence based on multiple inference algorithms. We find that the SS method performs the best on synthetic data sets and is, therefore, chosen as the default for the NI step of the pipeline. Notably, the consensus network displays a higher precision and returns a concise list of robust associations representing a valuable set for experimental validation follow-up. Future directions Future work building upon our current results could enhance the network inference process in multiple ways. The current analyses make use of a FMT data set with control and ASD samples, stool samples from radiation-exposed bank vole, three mock community data sets, and several data sets generated by two synthetic interaction methods. Incorporating data sets from a broad spectrum of biomes with varying microbial distributions into MiCoNE will likely increase the robustness and generalizability of the results from these analyses. The network analyses in this study are primarily at the Genus level, wherein the lowest resolution of a node is a Genus, and if an entity cannot be resolved to the Genus level, the next taxonomic level is used (e.g., Family). Consequently, two entities belonging to the same lineage where one entity is resolved to the Genus level and another is resolved to the Family level are treated as two different nodes in the network. Thus, developing an overlap metric to compare nodes with shared lineages within and across networks could enable more biologically and phylogenetically relevant comparisons. In thinking about the possible biological interpretations of the co-occurrence network computed by us and others, it is important to remember that there is no solid basis for assuming that these networks carry information about physical or metabolic interactions ( 80 , 81 ). Comparing co-occurrence networks and directly measured interactions remains a major unresolved challenge ( 82 , 83 ), which needs to be investigated further. Understanding this connection will be beneficial for predicting interactions in systems where direct interaction measurements cannot be taken. Further, benchmarking of co-occurrence networks could also be pursued through the use of literature-based interactions ( 1 ) or biological benchmark interaction data ( 84 ). Additionally, MiCoNE could be extended to enable the processing of metagenomics sequencing data, facilitating the analysis of a much larger and diverse range of data sets and domains of life. Although in the current analysis, we have only used default parameter values recommended by the tool creators, the MiCoNE pipeline could be used in the future to explore any combinations of parameters and optimize these values for improved network inference. Overall, there likely is no “best method” for the various steps of 16S data analysis, and hence, MiCoNE is intended to help researchers to identify the methods and algorithms that are most suitable for their data sets in an easy-to-use and reproducible manner. We envision that MiCoNE, and its underlying tools and databases, will be increasingly useful for building large comparative analyses across studies. It enables rapid, configurable, and reproducible inference of microbial networks and furthers the formulation of hypotheses about the role of these interactions on community composition and stability. These comparative analyses will require coupled network analysis and visualization tools (such as MIND ( 54 )) and need systematic access to data sets, shared in accordance with FAIR standards ( 81 )."
} | 3,866 |
23713139 | PMC3656021 | pmc | 9,323 | {
"abstract": "Microbial populations adapt to environmental fluctuations through random switching of fitness-related traits in individual cells. This increases the likelihood that a subpopulation will be adaptive in a future milieu. However, populations are particularly challenged when several environment factors change simultaneously. We suggest that a population can rapidly adapt to multiple environmental changes if individual members stochastically flip a hub-switch that controls a set of adaptive phenotypes in a single event. This mechanism of coupling phenotypic outcomes via a hub-switch can protect a population against large fluctuations in size. Here we report that the general amino acid transporter Gap1 is a potential hub-switch. The GAP1 gene is flanked by two direct repeats that can lead to GAP1 deletions (∆ gap1 ) and a self-replicating GAP1 circle. Thus, an isogenic GAP1 population can differentiate into two variant, reversible genotypes, ∆ gap1 or GAP1 circle . These subpopulations have different phenotypic advantages. A ∆ gap1 population has a selective advantage on allantoin or ammonium as a nitrogen source and high stress tolerance. Advantages of the GAP1 population include amino acid uptake, fast energy recruitment by trehalose mobilization, and in some cases, adherent biofilm growth. Our proposed model of a hub-switch locus enhances the bet-hedging model of population dynamics."
} | 354 |
36449969 | PMC9940878 | pmc | 9,324 | {
"abstract": "Abstract Vascular plants reinforce the cell walls of the different xylem cell types with lignin phenolic polymers. Distinct lignin chemistries differ between each cell wall layer and each cell type to support their specific functions. Yet the mechanisms controlling the tight spatial localization of specific lignin chemistries remain unclear. Current hypotheses focus on control by monomer biosynthesis and/or export, while cell wall polymerization is viewed as random and nonlimiting. Here, we show that combinations of multiple individual laccases (LACs) are nonredundantly and specifically required to set the lignin chemistry in different cell types and their distinct cell wall layers. We dissected the roles of Arabidopsis thaliana LAC4, 5, 10, 12, and 17 by generating quadruple and quintuple loss-of-function mutants. Loss of these LACs in different combinations led to specific changes in lignin chemistry affecting both residue ring structures and/or aliphatic tails in specific cell types and cell wall layers. Moreover, we showed that LAC-mediated lignification has distinct functions in specific cell types, waterproofing fibers, and strengthening vessels. Altogether, we propose that the spatial control of lignin chemistry depends on different combinations of LACs with nonredundant activities immobilized in specific cell types and cell wall layers.",
"introduction": "Introduction Lignin is a complex, heterogeneous phenolic polymer that is deposited in the cell walls of specialized cell types ( Meents et al., 2018 ). Associated with the evolutionary emergence of the plant vasculature and the transition to terrestrial habitats, lignin confers structural rigidity, and hydrophobicity to the vascular system ( Eriksson et al., 1991 ; Ménard et al., 2022 ). Lignin deposition proceeds in three steps: biosynthesis of phenolic monomers, mostly phenylpropanoids, in the cytoplasm ( Barros et al., 2015 ); their export into the apoplast ( Perkins et al., 2019 ); and their subsequent oxidative polymerization by radical coupling catalyzed by laccases (LACs) and class III peroxidases (PRXs) in the cell wall ( Blaschek and Pesquet, 2021 ). Among the phenoloxidases associated with lignin, specific paralogs of LACs in Arabidopsis ( Arabidopsis thaliana ) and poplar ( Populus sp.) are the main enzymes required to accumulate lignin in vascular tissues, but their effect on lignin chemistry is not known (reviewed in Blaschek and Pesquet, 2021 ). Although defined in the singular, the term lignin describes a multitude of chemically diverse polymers. This chemical diversity includes variation in both phenolic ring substitution (hydroxyphenyl, H; caffeyl C; guaicayl, G; syringyl, S) and aliphatic tail function (alcohol, X CHOH ; aldehyde, X CHO ) of the canonical phenylpropanoid monomers. In addition to phenylpropanoids, other monomers incorporated into lignin include benzaldehydes ( Ralph et al., 2001 ; Chen et al., 2012 ), flavonoids ( Lan et al., 2015 ; Rencoret et al., 2022 ), stilbenoids ( del Río et al., 2017 ; Rencoret et al., 2019 ) as well as other residues containing phenyl (P) structures ( Faix and Meier, 1989 ; Kawamoto, 2017 ; del Río et al., 2022 ). The main lignified vascular cell types in plants are tracheary elements (TEs), which act as both structural support and sap conduits for long-distance water transport. Lignin is essential for structural support at the organ level, as drastic reductions in lignin content lead to dwarf plants unable to stand upright and lodging ( Bonawitz and Chapple, 2010 ; Muszynska et al., 2021 ). Lignin is equally crucial at the cellular level, as reductions in lignin levels impair the biomechanical capacity of TEs to withstand the negative pressures required to transport water ( Ménard et al., 2022 ). In addition, most angiosperms and some gnetales also develop lignified xylem fibers located within (xylary fibers, XFs) or between (interfascicular fibers, IFs) vascular bundles, which fine-tune the mechanical properties of plant organs ( Zhong and Ye, 1999 ). To support their specific cellular functions, the concentration, composition, and structure of lignins are tailored to each cell type and their different cell wall layers—chemistries that are conserved between plant species ( Pesquet et al., 2019 ). Lignin concentration is generally higher in the secondary cell walls (SCWs) of TEs than in the SCWs of fibers, as well as generally higher in primary cell walls/middle lamella (CML) and cell corners (CCs) than SCWs ( Serk et al., 2015 ). However, large variations are also observed between different TE morphotypes, with early forming protoxylem (PX) TEs being far less lignified than later forming metaxylem (MX) TEs ( Ménard et al., 2022 ). Lignin composition, depending in part on the proportion of monomers with different phenolic ring substitutions, varies drastically between cell types and cell wall layers, with fiber SCWs enriched in S residues, TE SCWs enriched in G residues, and CML/CCs at the interface of these two cell types enriched in H residues ( Terashima and Fukushima, 1988 ; Fukushima and Terashima, 1990 ). The proportion of monomers with different aliphatic tail functions is also specifically controlled between cell types and cell wall layers, as well as between morphotypes, with high levels of X CHO in MXs compared to PXs, and high levels of X CHO in CMLs compared to SCWs ( Peng and Westermark, 1997 ; Blaschek, Champagne, et al., 2020 ). The lignification of each cell type progresses by incorporating distinct lignin residues at specific phases of the development and maturation of each cell type, as exemplified by S and X CHO residues that are mostly incorporated in the late maturation phases of TEs compared to the early incorporation of H and G residues ( Kutscha and Gray, 1972 ; Terashima and Fukushima, 1988 ; Blaschek, Champagne, et al., 2020 ; Ménard et al., 2022 ). The temporal control of cell wall lignification is specific to each cell type. TEs lignify their SCWs mostly after having committed programmed cell death (PCD) ( Pesquet et al., 2010 , 2013 ; Smith et al., 2013 ) whereas fibers gradually and centripetally lignify their SCWs as they are being deposited, both before and after their cell death ( Fukushima and Terashima, 1991 ; Terashima et al., 1996 ; Barros et al., 2015 ). The mechanisms enabling such a specific spatiotemporal control of lignin chemistry in different cell types and cell wall layers are still unknown. The current mechanisms proposed to explain this cell type/cell wall layer-specific control of lignification mainly rely on the regulation of monomer biosynthesis and/or their export to cell walls, while their polymerization is mostly viewed as random, nonlimiting, and nonspecific. Our current view of lignin formation thus posits that the oxidation of lignin monomers and their subsequent cross-coupling is mostly guided by thermodynamics and steric hindrance ( Blaschek and Pesquet, 2021 ). However, once in the apoplast, lignin monomers are extremely mobile due to autocrine, paracrine ( Pesquet et al., 2013 ; Smith et al., 2013 , 2017 ), and potentially endocrine secretions ( Aoki et al., 2019 ; Blaschek, Champagne, et al., 2020 ). This cell–cell transport is essential for TEs, which lignify postmortem ( Pesquet et al., 2010 , 2013 ; Derbyshire et al., 2015 ; Ménard et al., 2022 ) by polymerizing the monomers secreted by adjacent living cells. Neighboring fibers also use these externally supplied monomers semi-cooperatively in complement to their own export during and after the deposition of SCW polysaccharides ( Barros et al., 2015 ; Blaschek, Champagne, et al., 2020 ). The high mobility of lignin monomers in the cell wall contrasts with the spatially restricted differences of lignin chemistry in each cell wall layer of each cell type. This discrepancy suggests that additional mechanisms control lignification at the level of the cell wall itself. Such spatial control may depend on different LACs and/or PRXs with distinct catalytic activities that are targeted and immobilized in specific cell types and cell wall layers ( Derbyshire et al., 2015 ; Blaschek and Pesquet, 2021 ). Lignin-associated LACs are indeed immobilized in the cell walls of lignified tissues by tight ionic bonds ( Bao et al., 1993 ; Ranocha et al., 1999 ) highly limiting their mobility ( Chou et al., 2018 ). In fact, different LACs show highly distinct binding pocket and active site topologies, suggesting differences in catalytic efficiency, pH optimum, and/or substrate specificity ( Blaschek and Pesquet, 2021 ). LACs involved in the formation of specialized lignins in seed coats and compression wood have recently been shown to exhibit a certain degree of substrate specificity for monomers with different ring substitutions ( Wang et al., 2020 ; Hiraide et al., 2021 ; Zhuo et al., 2022 ). The role of LAC specificity to set distinct chemistries in the main developmental lignification of xylem remains, however, unknown. Here, we provide experimental evidence demonstrating that the lignin polymerization capacity of the different cell wall layers of each xylem cell type spatially controls its lignin chemistry. Using higher-order loss-of-function mutants of LAC s involved in vascular lignification, we show that the isolated function of different LAC paralogs is specific to each cell type and cell wall layer, and discriminately incorporates specific monomers. The resulting differences in lignin chemistry also affected its function in specific cell types, distinctly altering the cell wall mechanical resistance of TEs to negative pressure and the hydrophobicity of fiber cell walls. Altogether, we show that different immobilized combinations of LAC paralogs with specific activities control the lignin chemistry in each cell wall layer and cell type to support their different functions.",
"discussion": "Discussion The precise spatiotemporal control of lignification is pivotal for normal development ( Zhao et al., 2013 ), drought resistance ( Lima et al., 2018 ; Ménard et al., 2022 ), and defense against herbivores and pathogens ( Whitehill et al., 2016 ; Joo et al., 2021 ). The importance of lignin monomer biosynthesis in each cell type ( Dixon and Barros, 2019 ; Blaschek, Champagne, et al., 2020 ) and their transport from the cytosol into the apoplast ( Perkins et al., 2019 ; Väisänen et al., 2020 ) are essential to control lignin but insufficient to regulate the strict spatial distribution of lignin in between cell types and cell wall layers. In addition to their extreme mobility in cell walls, alcohol and aldehyde phenylpropanoids have been shown to diffuse freely across biological membranes ( Vermaas et al., 2019 ) making lignin oxidative polymerization by LACs a potential main driving force controlling the metabolic gradient-dependent transport of lignin ( Perkins et al., 2022 ). Specific LAC paralogs had previously been shown to be important for lignin accumulation in xylem cells when knocked out or knocked down in Arabidopsis, purple false brome ( Brachypodium dystachion ), rice ( Oryza sativa ), maize ( Zea mays ), and poplar but without clearly distinguishing between cell types or morphotypes for their lignin chemistry (reviewed by Barros et al., 2015 and Blaschek and Pesquet, 2021 ). As specific LAC paralogs are exclusive to lignifying conditions ( Figure 1 ) and have been suggested to drive the transport of phenylpropanoids ( Perkins et al., 2022 ), LACs potentially represent the central regulating components capable of channeling extracellular phenolics toward lignin. Here, we provide clear evidence that specific LAC combinations are required for depositing distinct lignin chemistries in the different cell wall layers for the specific function of each cell type. In contrast to the random oxidation model of lignin only considering thermodynamics, steric hindrance, and substrate availability as limiting factors, our results show that the identity of the oxidizing LAC establishes an important parameter for the spatial control of lignin chemistry. Our study showed that LAC4 and LAC17, previously shown to be involved in vascular lignification together with LAC11 ( Berthet et al., 2011 ; Zhao et al., 2013 ), exert nonredundant functions in the lignification of specific vascular cell types. Far from being the only players, we also revealed novel and specific functions for LAC5, LAC10, and LAC12 in vascular lignification. In higher-order lac mutants, the loss of different LAC combinations resulted in specific and nonredundant changes of lignin chemistry affecting specific cell type and/or cell wall layer ( Figures 4–6 ). Table 1 summarizes the different morphological and biochemical aspects observed in our higher lac mutants, which confirmed the substrate-specific, layer-specific, and cell type-specific activities of each of these different LACs. The drastic reduction in lignin contents in higher-order lac mutants, especially in fiber cell walls, demonstrated the implication of the five tested LAC s for vascular lignification ( Figure 4B ). By contrast, the stable LAC activity in MX TEs in our higher-order lac mutant series ( Figure 3 ) suggests that the remaining LAC11 , shown to be required for vascular lignification in the absence of LAC4 and 17 ( Zhao et al., 2013 ), suffices to ensure some lignification, although both composition and function of TEs are impaired due to their inward collapse ( Figure 7 ). This observation confirms recent results showing that TE biomechanical properties do not only depend on lignin concentrations but also on its chemistry ( Ménard et al., 2022 ). The resulting changes in lignin chemistry thus depend on the combined effects of reduced/abolished substrate specific-oxidation together with modifications due to nonspecific oxidation by the remaining LACs and PRXs ( Blaschek and Pesquet, 2021 ), reduction in diffusion of some lignin precursors to cell walls due to altered metabolic gradients created by oxidation ( Perkins et al., 2022 ), and feedback modifications of intracellular phenolic metabolism ( Vanholme et al., 2012 ). Our results also suggest that LACs discriminate both different ring structures and aliphatic tails, as LAC4 favored the incorporation of G CHOH , whereas LAC17 accumulated more G CHO ( Figures 4 and 5 ). Table 1 Effects of single active LACs in the Q background (e.g. the effect under LAC4 describes the difference between Q and Q-4 ) Structure Parameter \n LAC4 \n \n LAC5 \n \n LAC10 \n \n LAC12 \n \n LAC17 \n Plant Stem height WT — ↑ — WT IF Activity — — — — ↑ Lignin ↑ — — — ↑ \n G \n CHO / G CHOH — — — — ↑ Swelling WT — — — WT CC Activity ↑ — — ↑ ↑ Lignin ↑ — — ↑ ↑ CML Activity ↑ — — ↑ ↑ Lignin ↑ — — — ↑ XF Activity ↑ ↑ — ↑ ↑ Lignin ↑ ↑ — ↑ ↑ \n G \n CHO / G CHOH ↑ ↑ — ↑ ↑ PX Activity — — — — — Lignin — — — — — \n G \n CHO / G CHOH — ↑ ↑ ↑ WT Convexity ↑ ↑ ↑ ↑ WT MX Activity — — — — — Lignin ↑ — ↑ ↑ ↑ \n G \n CHO / G CHOH — — — — — Convexity WT ↑ ↑ — WT —, no effect; ↑, partly compensated; WT, reaching wild-type level. Preferential substrate for distinct LAC paralog was shown recently by the gain-of-function of spider flower ( Cleome hassleriana ) LAC8 in Arabidopsis that enabled increased accumulation of the noncanonical C residue, with two hydroxyl groups in its aromatic ring ( Wang et al., 2020 ; Zhuo et al., 2022 ). This nonredundant activity of LAC paralogs between cell types and cell wall layers ( Figure 3 ) is moreover supported by recent protein modeling results showing that LACs have distinct protein structures affecting the position of key catalyzing amino-acids, active site binding pocket volume, shape, and accessibility ( Blaschek and Pesquet, 2021 ). Moreover, our results clearly localized the activity of different LAC paralogs in distinct cell wall layers and cell types, such as LAC12 in the cell corners of IFs ( Figures 3 and 6 ). The localization of specific LAC paralogs using our activity measurements showed many overlaps with previous studies using transcriptional, translational reporter, and immunolocalization ( Berthet et al., 2011 ; Turlapati et al., 2011 ; Schuetz et al., 2014 ; Derbyshire et al., 2015 ; Hoffmann et al., 2020 ). This differential cell wall layer localization between LAC paralogs was reflected by the expression of LAC12 in both TEs with lignified SCWs as well as in xylem parenchyma (XP) with nonlignified PCW once TEs had died ( Figure 1 ). Additionally, differences in localization could help support the formation of specific lignin in the “sequential intervention model” proposed by Barros et al. (2015) , where phenolics are sequentially oxidized by specific LACs from their site of secretion at the plasma membrane to their site of accumulation, passing across the different layers of the cell wall. In addition to the spatial restriction of lignification, specific cell types in vascular tissues maintain distinct lignin composition and amounts across species. For example, TE enrichment in G residues is conserved among all vascular plant species ( Pesquet et al., 2019 ). This cell type-specific function of lignin was recently demonstrated for the different TE morphotypes that require a tight regulation of G CHO to G CHOH ratio to maintain the balance between SCW stiffness and flexibility ( Ménard et al., 2022 ). Here, we expand on the importance of distinct lignin chemistries for each cell types by showing that specific LACs are required for the accumulation of specific lignins to reinforce structurally TEs and limit cell wall swelling in fibers ( Figures 7 and 8 ). Altogether, our data provide the missing link between lignin monomer biosynthetic control and their polymerization into distinct lignin polymers in specific cell types for plants to adapt to the numerous environmental and developmental stresses faced by their cell walls."
} | 4,475 |
27774389 | PMC5057325 | pmc | 9,326 | {
"abstract": "A piezoelectric paper based on BaTiO 3 (BTO) nanoparticles and bacterial cellulose (BC) with excellent output properties for application of nanogenerators (NGs) is reported. A facile and scalable vacuum filtration method is used to fabricate the piezoelectric paper. The BTO/BC piezoelectric paper based NG shows outstanding output performance with open‐circuit voltage of 14 V and short‐circuit current density of 190 nA cm −2 . The maximum power density generated by this unique BTO/BC structure is more than ten times higher than BTO/polydimethylsiloxane structure. In bending conditions, the NG device can generate output voltage of 1.5 V, which is capable of driving a liquid crystal display screen. The improved performance can be ascribed to homogeneous distribution of piezoelectric BTO nanoparticles in the BC matrix as well as the enhanced stress on piezoelectric nanoparticles implemented by the unique percolated networks of BC nanofibers. The flexible BTO/BC piezoelectric paper based NG is lightweight, eco‐friendly, and cost‐effective, which holds great promises for achieving wearable or implantable energy harvesters and self‐powered electronics.",
"conclusion": "3 Conclusion In this work, we demonstrate a piezoelectric paper by incorporating BaTiO 3 nanoparticles into bacterial cellulose nanofiber networks to realize lightweight, flexible, and cost‐effective NGs. The piezoelectric paper is fabricated by a vacuum filtrating method, which is facile and scalable. The BTO/BC piezoelectric paper based NG shows outstanding output performance with the open‐circuit voltage of 14 V and short‐circuit current density of 190 nA cm −2 . The maximum power density generated by this unique BTO/BC structure is 0.64 μW cm −2 , which is more than ten times higher than BTO/PDMS structure. This enhancement can be ascribed to homogeneous distribution of piezoelectric BTO nanoparticles in the BC matrix, which is implemented by the percolated networks of BC nanofibers. The BTO/BC piezoelectric paper based NGs also represent potential applications as flexible energy harvesters. In the cyclic bending condition, the device can generate a peak voltage of 1.5 V with high stability and durability. A commercial LCD screen can be driven by the cyclic generated power. The flexible BTO/BC piezoelectric paper based NG is lightweight, eco‐friendly, and cost‐effective, which holds great promises for achieving wearable or implantable energy harvesters and self‐powered electronic devices.",
"introduction": "1 Introduction With the increasing demand for sustainable and reliable energy for personal electronics and wireless nanosystems, it is envisioned as a promising approach to harvest energy from the ambient sources, such as body movements, air flow, acoustic waves, and even thermal fluctuations which are available in most of the circumstances. 1 , 2 , 3 , 4 Nanogenerators (NGs) are emerging as novel devices which can convert various kinds of ambient energy into electric power via piezoelectric, 5 triboelectric, 6 , 7 or pyroelectric effect 4 both in nanoscale and macroscale. As a typical piezoelectric nanomaterial, ZnO nanowire was found to demonstrate piezoelectric potential distributing along its c ‐axis when subjected to lateral bending or axial tension and compression, which is the basic mechanism of piezoelectric NGs. 8 , 9 Up till now, ZnO nanowires based piezoelectric NGs not only have been used for harvesting mechanical energy with much enhanced output performance 10 but also functioned as self‐powered sensors to detect motion and vibration signals. 11 , 12 \n In the pursuit of higher piezoelectric output performance, it is desirable to utilize piezoelectric materials with higher piezoelectric coefficient. Conventional piezoelectric ceramics including Pb(Zr 1− x Ti x )O 3 (PZT), BaTiO 3 (BTO), and (1− x )Pb(Mg 1/3 Nb 2/3 )O 3− x PbTiO 3 (PMN‐PT) has been outstanding candidates for high‐output NGs in view of their excellent piezoelectric properties. 13 , 14 , 15 , 16 , 17 In particular, BTO is the most attractive material for its lead‐free property as well as high piezoelectricity. 16 However, as a major problem for practical application, these bulk piezoelectric ceramics are intrinsically very brittle and thus hardly compatible with the irregular mechanical deformations. To address this problem, it has emerged as an effective methodology to incorporate piezoelectric nanomaterial into soft polymer matrix resulting in a flexible piezoelectric nanocomposite to improve the mechanical durability of the piezoelectric components. 18 Polydimethylsiloxane (PDMS) is the most commonly used polymer matrix because of its flexibility and robustness. Piezoelectric polymer such as poly(vinylidene fluoride) (PVDF) 19 and its copolymers 20 have also been utilized as matrix because they themselves are piezoelectric materials which can provide enhancement for the overall output performance. For the design of piezoelectric composite film, it is believed that piezoelectric nanomaterial must be dispersed uniformly in the composite to obtain optimized piezoelectric output. 21 However, it is challenging to get such a well‐distribution because most of the flexible matrix materials used are often sticky and hydrophobic, which makes it hard for ceramic nanoparticles to separate sufficiently. In view of these difficulties, adding dispersants such as graphitic carbons 18 , 22 or metal nanorod 23 into the composite and biosynthesis of piezoelectric nanostructure by virus templates 24 to realize entangled network structures have shown promising enhancement. Despite the enormous performance improvement that has been achieved, more facile, inexpensive, and scalable approaches are still needed. As a kind of natural material, cellulose has become an attractive candidate for paper based devices due to its flexibility, biocompatibility, and low cost. Recently, various techniques have been explored to implement novel paper based devices. Compared to other flexible materials, cellulose possess a much lower coefficient of thermal expansion, which is an advantage for the thermal stability of devices. 25 In addition, suitability for printed electronic device is another promising merit for paper materials. 26 Bacterial cellulose (BC), a kind of biopolymer produced by Gluconacetobacter strains, has higher mechanical strength and better chemical stability than regular paper due to its high purity and crystallinity. Due to its intrinsic textured nanofibrillated structure, BC can present itself as either reinforcement or matrix for functional materials such as flexible transparent film, 27 conductive polymers, 28 and antibacterial textiles. 29 It is also necessary to verify BC as building materials for flexible and cost‐effective NGs. In this work, we fabricated a piezoelectric paper with piezoelectric BTO nanoparticles and BC through a facile vacuum filtration method. The BTO/BC piezoelectric paper based NGs showed an enhanced output performance compared to traditional BTO/PDMS based devices. The unique entangled BC nanofiber networks enable BTO nanoparticles to be dispersed in the piezoelectric paper uniformly, which makes the excellent piezoelectric property of the paper. The BTO/BC piezoelectric paper based NG can generate voltage of 1.5 V in bending conditions, which can drive a commercial liquid crystal display (LCD) screen. The flexible BTO/BC piezoelectric paper based NG is lightweight, eco‐friendly, and cost‐effective, which holds great promises for achieving wearable or implantable energy harvesters and self‐powered electronic devices.",
"discussion": "2 Results and Discussion \n Figure \n \n 1 \n a–d shows the fabrication process of the BTO/BC piezoelectric paper. As can be seen in Figure 1 a,e, raw BC are transparent gel‐like pellicles which are composed of cellulose nanofibers with the diameter of ≈10–30 nm and all the nanofibers are found to aggregate densely due to the loss of water. The X‐ray diffraction (XRD) pattern of the BC membrane is shown in the inset of Figure 1 e. The diffraction peaks of 14.2°, 16.5°, and 22.5° are assigned to diffraction planes of (101), (10‐1), and (002) for native cellulose I, respectively. 30 An aqueous suspension of BC fibers can be obtained by mechanically breaking the interconnected nanofiber networks with a high speed homogenizer. BTO nanoparticles were synthesized by hydrothermal method. 18 As illustrated in Figure 1 f, the average diameter of BTO nanoparticles is about 100 nm. Raman spectrum (inset of Figure 1 f) shows peaks positioned at 250 cm −1 [ A \n 1 (TO)], 302 cm −1 [ E , B \n 1 (TO+LO)], 511 cm −1 [ E , A \n 1 (TO)], and 710 cm −1 [ E , A \n 1 (LO)], which represent a high piezoelectric tetragonal phase of BTO. 31 The as‐synthesized BTO nanoparticles were ultrasonically dispersed in distilled water and then mixed with BC aqueous suspension under vigorous stirring to guarantee sufficient blending. BTO/BC piezoelectric paper can be formed by vacuum filtrating the blended suspension with a microporous membrane followed by a pressing and drying process. Figure 1 g demonstrates the field‐emission scanning electron microscopy (FE‐SEM) image of the BTO/BC piezoelectric paper. It is found that the disintegrated BC nanofibers get associated again due to the strong interaction of hydrogen bonds and BTO nanoparticles are uniformly bounded within the BC matrix. Thermogravimetry analyses (TGA) in the inset of Figure 1 g exhibits that pristine BC undergoes decomposition in two steps with degradation temperatures at around 350 °C (weight loss of 70%) and 450 °C (weight loss of 99%). Compared with the pure BC which has almost no mass residue after calcination, the BC mixed with 0.5 g BTO nanoparticles presents a mass residue of nearly 80% even above 500 °C, which indicates that the mass fraction of BTO nanoparticles in the paper is very high. This loading percentage is much higher than that achieved by a layer‐by‐layer approach (48 wt%). 32 This distinct feature makes the BTO/BC piezoelectric paper as light as possible because most of the weight is concentrated on the piezoelectric component, which is an important merit for integration applications. Figure 1 a–d) Fabrication process of the BTO/BC piezoelectric paper. e) An SEM image and X‐ray diffraction pattern (inset) of the pristine BC membrane. f) An SEM image and Raman spectrum (inset) of the BTO nanoparticles synthesized by hydrothermal method. g) An SEM image and thermal degradation behavior (inset) of the piezoelectric paper. BTO/BC piezoelectric paper was used as an active layer to fabricate NGs, as shown in Figure \n \n 2 \n a. First, a thin layer of PDMS was spin‐coated on both side of the piezoelectric paper to provide a smooth surface for the electrodes and to protect the paper from moisture in the air. A layer of Ti/Au (10 nm/100 nm) was then deposited on both sides of the paper as electrodes. Two conductive tapes were then connected to the top and bottom electrodes, respectively. All the devices are characterized with the same active size of 3 × 2 cm 2 . Basically, poling process is of great importance for the BTO/BC piezoelectric paper because disordered dipoles in ferroelectric BTO domains need to be aligned by an external electric field so that the piezoelectric potential can be enhanced in a specific direction. To reveal the necessity of this process, a range of electric field were applied to pole the devices. As shown in Figure 2 b,c, the unpoled device showed output voltage of only ≈1 V. As the poling electric field increased from 50 to 200 kV cm −1 , the output voltage can be enhanced to more than 12 V. This can be ascribed to the rearrangement of the ferroelectric dipoles under high electric field. Furthermore, the output performance of the NG device can be affected by the amount of BTO nanoparticles contained in the BTO/BC piezoelectric paper. As shown in Figure 2 d,e, when no BTO was contained in the paper, negligible output signals were detected which may originate from the capacitance change of the device. As the amount of BTO nanoparticles increasing from 0.2 to 0.5 g, the output voltage raised from ≈4 to ≈13 V. However, when BTO was further added to 0.8 g, the output voltage decreased to ≈8 V. This can be understood by the trade‐off between the density of piezoelectric points and the total permittivity of the piezoelectric paper. The increasing amount of BTO nanoparticles will undoubtedly provide more piezoelectric points in the paper which is helpful to generate high piezoelectric output. However, excessive amount of BTO nanoparticles can result in very high dielectric constant of the composite, which can weaken the electromechanical coupling effect of the piezoelectric paper. 33 \n Figure 2 a) Fabrication process of the BTO/BC piezoelectric paper based NG device. b) The output voltage of the BTO/BC piezoelectric paper based NG with different poling electric field. c) Average output voltage dependence on poling electric field. d) The output voltage of the BTO/BC piezoelectric paper based NG with different BTO content. e) Average output voltage dependence on BTO content. The output performance of the device is investigated by measuring the open‐circuit voltage and short‐circuit current when the NGs are subjected to cyclic compressive stress in the normal direction. As shown in Figure \n \n 3 \n a,b, the open‐circuit voltage and short‐circuit current density can reach as high as 14 V and 190 nA cm −2 , respectively. The value discrepancy of each peak can be attributed to the different strain rate of the device during compressing and releasing process. 34 To verify the signals are induced by the piezoelectric potential in response to the deformation of the piezoelectric paper, it is essential to conduct switching‐polarity test. 9 In the reverse connection, the V – t signal exhibits a negative pulse followed by a positive pulse in response to a pressing and a releasing action with the average output voltage and current density maintained at similar magnitude with the forward connection. Therefore, the possible artifacts from triboelectricity and the measurement system can be ruled out. Figure 3 c illustrates the dependence of the output characteristic on external load resistance. With the increment of the load resistance from 1 MΩ to 1 GΩ, the output voltage increases gradually from about 0.16 to 14 V, while the current density decreases from 160 to 14 nA cm −2 . The output power density can be calculated by V \n 2 / R , where V and R represent the output voltage and the corresponding external load resistance, respectively. As plotted in Figure 3 d, the output power reaches the maximum value of 0.64 μW cm −2 at a matched resistance value of 60 MΩ. Figure 3 a) Open‐circuit voltage and b) short‐circuit current density of the BTO/BC piezoelectric paper based NG both in forward and reverse connections. The output voltage and c) current density and d) power density of the BTO/BC piezoelectric paper based NG with different external load resistance. The output voltage and e) current density and f) power density of the BTO/PDMS based NG with different external load resistance. To understand the enhancement effect of this BTO/BC structure for the performance of NGs, we have also fabricated devices with commonly used BTO/PDMS structure. As exhibited in Figure 3 e, both of the output voltage and current density are much lower than the BTO/BC based devices. The maximum output power generated by BTO/PDMS based devices is only 0.04 μW cm −2 , which is more than ten times lower. To reveal the remarkable merit of using BC as matrix for BTO nanoparticles to disperse, SEM characterization is used to compare the top, bottom, and cross‐sectional structure of both kinds of the film. As shown in Figure \n \n 4 \n a–c, BTO nanoparticles are uniformly embedded in the whole BC matrix without any obvious aggregation. However, for BTO/PDMS film, limited amount of BTO nanoparticles are existed near the top surface of the film while extensive nanoparticles are accumulated at the bottom of the film. From the side view of the PDMS based film in Figure 4 e, it is clear that most of the nanoparticles have aggregated into large clusters and distribute at the bottom of the PDMS body. Uniform dispersion of piezoelectric nanoparticles within a polymeric matrix is one of the key issues for piezoelectric film to yield high output. The piezopotential distributions inside the piezoelectric films are simulated by COMSOL software. As calculated by simplifying the film and BTO nanoparticles as a rectangular model and six piezoelectric circles in Figure 4 g,h, it is predicted that homogeneous dispersion of BTO nanoparticles can lead to higher piezoelectric potential than the case that all the particles are distributed at the bottom of the matrix with other conditions unchanged. However, it is intrinsically difficult to disperse BTO nanoparticles into sticky PDMS for inevitable aggregation and settlement will happen before the whole body has fully cured. The percolated network of the BC nanofibers enables well dispersion of piezoelectric nanoparticles in the film; thus, the piezoelectric output can be optimized. Furthermore, cellulose is reported to have a Young's modulus in the range 78 ± 17 GPa, 35 which is properly higher than PDMS (1–2 MPa). 18 The stiff nanofibers can transfer stress to the localized piezoelectric BTO nanoparticles effectively. As a result, the BTO nanoparticles will be deformed more significantly in the BC matrix and enhanced output signals can be yielded. Figure 4 SEM image of the a) top surface, b) cross‐sectional, and c) bottom surface of the BTO/BC piezoelectric paper. SEM image of the d) top surface, e) cross‐sectional, and f) bottom surface of the BTO/PDMS film. COMSOL simulation results of the output voltage of the g) BTO/BC piezoelectric paper and h) BTO/PDMS film with the compressive stress of 0.1 MPa. To demonstrate the potential applications requiring flexible properties of the NG device, the output performance was measured in bending/releasing conditions. The NG device was mounted on a Kapton film which was bent and released repeatedly by a bending stage. As shown in Figure \n \n 5 \n a, a peak output voltage of 1.5 V was achieved with the deformation frequency of 1 Hz. Furthermore, the durability of the device was also examined by applying bending/releasing repeatedly (Figure 5 b). It was found that at the beginning of the test, minor degradation of output voltage was observed, which was caused by small relative displacement between fixture and the device. The output voltage was maintained at about 1.2 V afterward and no obvious performance decline can be seen for up to 3000 cycles, indicating excellent robustness of the device. The BTO/BC piezoelectric paper based NGs can be applied to harvest mechanical energy and operate small electronics. A commercial LCD can be directly triggered by cyclic bending/releasing the device. Since the LCD is a nonpolar device, it can be driven by AC source without a rectifier unit. Both the bending and releasing action of the finger can generate the flash of number “8” on the screen, while no number is displayed during the holding process (shown in a video, Supporting Information). Figure 5 a) Output voltage of the BTO/BC piezoelectric paper based NG under the bending frequency of 1 Hz. b) Cyclic bending test for 3000 cycles of bending/releasing motions. c) Four states when a commercial LCD screen is driven by the NG under bending conditions."
} | 4,896 |
30429841 | PMC6220076 | pmc | 9,327 | {
"abstract": "Fertilizer application has contributed substantially to increasing crop yield. Despite the important role of soil fungi in agricultural production, we still have limited understanding of the complex responses of fungal taxonomic and functional groups to organic and mineral fertilization in long term. Here we report the responses of the fungal communities in an alkaline soil to 30-year application of mineral fertilizer (NP), organic manure (M) and combined fertilizer (NPM) by the Illumina HiSeq sequencing and quantitative real-time PCR to target fungal internal transcribed spacer (ITS) genes. The results show: (1) compared to the unfertilized soil, fertilizer application increased fungal diversity and ITS gene copy numbers, and shifted fungal community structure. Such changes were more pronounced in the M and NPM soils than in the NP soil (except for fungal diversity), which can be largely attributed to the manure induced greater increases in soil total organic C, total N and available P. (2) Compared to the unfertilized soil, the NP and NPM soils reduced the proportion of saprotrophs by 40%, the predominant taxa of which may potentially affect cellulose decomposition. (3) Indicator species analysis suggested that the indicator operational taxonomic units (OTUs) in the M soil occupied 25.6% of its total community, but that only accounted for 0.9% in the NP soil. Our findings suggest that fertilization-induced changes of total fungal community were more responsive to organic manure than mineral fertilizer. The reduced proportion of cellulose decomposition-related saprotrophs in mineral fertilizer treatments may potentially contribute to increasing their soil C stocks.",
"conclusion": "Conclusion In the alkaline soils used in this study, long-term mineral and organic fertilizer applications not only increased fungal abundance and diversity, but also altered fungal community structure. Such changes were more pronounced in the organic manure treated soils (M and NPM) than in the mineral fertilizer NP soil. In particular, mineral fertilizer application (NP and NPM) could select, promote or reduce specific groups that may have positive impacts on soil C and N cycling. Furthermore, fungal genera known as plant pathogens were better promoted by mineral fertilizer than by organic manure. The discrepant observations on fungal community between the alkaline soil in this study and previous reports from neutral and acidic soils jointly highlight the necessity to further distinguish the possibly different responses of soil fungi among different soil types.",
"introduction": "Introduction Fertilization has contributed substantially to increasing crop yield (Zhang et al., 2012 ). Long-term fertilizer inputs, especially nitrogen (N) inputs, can result in reducing soil pH (Guo et al., 2010 ) and increasing soil organic carbon (C) and total N content (Guo et al., 2011 ; Jian et al., 2016 ), which potentially influence soil fungal community composition. While soil fungi play important roles as decomposers (Fontaine et al., 2011 ; Ma et al., 2013 ), plant symbionts (Clemmensen et al., 2006 ) and pathogens (Ohm et al., 2012 ), the effects of fertilizer application on fungal communities have not yet well understood. Previous research on N fertilization and fungal community structure mostly focused on total fungal community (Weber et al., 2013 ; Yuan et al., 2013 ; Wang et al., 2017 ) or specific subsets of fungi, e.g., mycorrhizal fungi (Gryndler et al., 2006 ; Sheng et al., 2013 ; Ekblad et al., 2016 ) and saprotrophic fungal taxa (Allison et al., 2007 ). Only few studies comprehensively describe the fungal community in a given ecosystem (Sterkenburg et al., 2015 ; Morrison et al., 2016 ), and most of them mainly focused on N fertilization in acidic or neutral soils. For example, in addition to descriptions of entire community and phyla over discussions of species, some reports have shown that mineral N application has a positive impact on soil saprotrophic fungi in soils of pH 4.5–6.8 (Sterkenburg et al., 2015 ; Morrison et al., 2016 ) and promotes fungal genera with known pathogenic traits (Hartmann et al., 2015 ; Paungfoo-Lonhienne et al., 2015 ; Zhou et al., 2016 ). While soil pH in acidic and neutral soils is very susceptible to N application and in turn affects fungal community composition (Zhang et al., 2016 ) and decreases fungal diversity (Zhou et al., 2016 ), soil pH in alkaline soil is less affected by N fertilization due to its higher pH buffering capacity (Guo et al., 2010 ). Thus, in alkaline soil, fungal community might be more linked to changes of soil nutrients rather than soil pH (Lauber et al., 2008 ). However, the potential effects of fertilizer on total fungal community and major fungal functional groups in alkaline soil have been less explored (Rousk et al., 2010 ). In alkaline soil, N application has been found no influence on fungal community structure (Mueller et al., 2015 ; Chen et al., 2016 ) but increasing fungal class Sordariomycetes (Mueller et al., 2015 ). In fact, alkaline soil is widely distributed in northern China, especially on the Loess Plateau with great relevance for regional and national food security (Kuhn et al., 2016 ). Furthermore, organic amendments have been reported to increase soil nutrient status and organic C content (Jian et al., 2016 ), but the response of fungal communities and functional groups to manure application has not been thoroughly studied (Francioli et al., 2016 ). This therefore calls for a systematic investigation on the possible impacts of mineral fertilizer and organic manure treatments on soil fungal community structure and functional population dynamics in alkaline soils. In this study, we collected alkaline soil samples from a 30-year fertilization experiment in northwest China and determined the abundance and community composition of fungi under mineral and organic fertilization. The Illumina HiSeq sequencing and quantitative real-time PCR of fungal internal transcribed spacer (ITS) genes were used to quantitatively and qualitatively assess changes in fungal communities. We hypothesized that long-term fertilization in alkaline soil not only shifts fungal community structure, but also changes the proportions of fungal functional groups, which may help us to advance our current understanding of fertilization introduced increase in soil fertility.",
"discussion": "Discussion Fungal communities in unfertilized vs. fertilized soils Long-term mineral and organic fertilizer application to the alkaline soil in this study generally increased both fungal abundance (quantified by qPCR) and fungal Simpson's diversity (Table 3 ). This is partly consistent with previous findings in neutral and acidic soils, where fungal abundance was enhanced but fungal diversity was reduced with N fertilization (Allison et al., 2007 ; Wang et al., 2015 ; Zhou et al., 2016 ). Such discrepancy is probably because the soil pH in this study was inherently different from those in previous reports. In previous studies with acid and neutral soils, N application led to soil acidification, which further changed soil fungal community (Zhou et al., 2016 ). However, in the alkaline soil of this study, fungal abundance and diversity showed strong correlations with soil total organic C, total N and available P concentrations (Supplementary Table 3 ), all of which are factors known to influence the soil microbiota (Lauber et al., 2008 ). Moreover, in a previous study, fungal community composition was significantly related to soil fertility (Sterkenburg et al., 2015 ). This indicates that fertilizer-induced increase in soil fertility might be the strong driver of fungal community in alkaline soils. Ascomycota, who is a key decomposer in agricultural soils (Ma et al., 2013 ), showed no difference between mineral fertilized and unfertilized soils. This is not consistent with previous findings that Ascomycota increased proportions in N application soils (Weber et al., 2013 ; Zhou et al., 2016 ). This might be due to the different responses of fungal taxa at lower phylogenetic levels to fertilizer applications. Most members of Ascomycota were affiliated to class Sordariomycetes, which increased proportion in manure application soils compared to the unfertilized soil. The second most dominant class in Ascomycota in this study was Dothideomycetes, which in together with class Leotiomycetes increased proportions in mineral fertilized soils (Figure 2 ). Sordariomycetes and Leotiomycetes has been shown to increase proportion (Freedman et al., 2015 ; Mueller et al., 2015 ) but Dothideomycetes decreased with N fertilization (Freedman et al., 2015 ; Zhou et al., 2016 ). The changes of class Dothideomycetes in this study were mainly driven by the genus Pseudogymnoascus , which is known as saprotrophic fungi (Małecka et al., 2015 ). Moreover, the genus Tetracladium drove the changes of class Leotiomycetes in mineral fertilized soil, and is known as common root fungi (Sati et al., 2009 ) which has potential to benefit the growth and nutrient acquisition of their host plants. These results indicate that mineral fertilizer application in long term might enhance these taxa related litter decomposition and plant-fungal symbioses in the studied soil. The proportion of Basidiomycota decreased in the mineral fertilizer and manure application soils and such decline in proportion was more pronounced in the M and NPM than in the NP (Figure 2 ). This may be related to the organic carbon composition in the M and NPM treatments. Basidiomycota was dominated by the class Agaricomycetes which are particularly important during the later stage of litter decay (Purahong et al., 2016 ). The high organic matter quality and input rate in the manure fertilized soils may have reduced the competitiveness of late-stage fungi. While not consistent with the increasing pattern observed in high fertility forest soils in Sterkenburg et al. ( 2015 ), the reduced Basidiomycota proportion in this study agrees well with the recent findings from soils with high N application (Weber et al., 2013 ; Paungfoo-Lonhienne et al., 2015 ; Zhou et al., 2016 ). Although the proportion of several putative saprotrophic fungal genera increased in the fertilized soils, such as Staphylotrichum and Trichocladium in the M soil, Kernia in the NP soil and Microascus in the NPM soil (Supplementary Figure 2 ), the overall putative saprotrophs proportion was decreased in the mineral fertilized soils compared to the unfertilized soil (Figure 3 ). This was mainly driven by the genera Chaetomium and Penicillium whose primary niche appears to be cellulose decomposition (Li et al., 2013 ; Sharma et al., 2013 ). Since cellulose is the dominant form of carbon entering in arable soils (Jin and Chen, 2007 ; Thomsen et al., 2008 ), the decrease in putative saprotrophs proportion may contribute to the increase of soil C storage (Table 1 ). This is in agreement with a previous finding that mineral fertilizer reduced soil cellobiohydrolase activity in an arable soil (Fan et al., 2012 ). Furthermore, several saprotrophic fungal genera such as Chaetomium and Myrothecium in the class Sordariomycetes and genera Penicillium and Talaromyces in the class Eurotiomycetes, were strongly reduced by mineral and organic fertilizer applications. These genera were potential N 2 O-producing fungal denitrifiers (Mothapo et al., 2013 , 2015 ), their overall decrease in proportion (Figure 3 ) indicates that fungal induced N 2 O emission might be decreased in the fertilized soils. In contrast to putative saprotrophs, proportion of putative plant pathogens significantly increased after long-term fertilizer application (Figure 3 ). Its community structure was also altered, with a greater change in mineral fertilizer soil than in organic manure soil (Table 2 ). This is in line with several recent studies that the pathogenic taxa increased proportion after chronic fertilizer application in agricultural and forest soils (Paungfoo-Lonhienne et al., 2015 ; Morrison et al., 2016 ). Moreover, the increased proportion of putative plant pathogens did not have negative effects on plant biomass in the fertilized soil in this study (Table 1 ). This agrees well with other studies in agricultural soils (Morrison et al., 2016 ; Zhou et al., 2016 ), and also stays consistent with the findings in natural soils that proportion of plant pathogens was not related to plant richness (Tedersoo et al., 2014 ). However, it has also been suggested that plant pathogens at least partly coevolve with their hosts, as they usually attack a phylogenetically limited set of host plants (Gilbert and Webb, 2007 ). The specific role of plant pathogens in agricultural and natural soils still requires further investigations. It is important to understand that we can only speculate on the ecological role of the detected taxa based on previous descriptions in other studies. Moreover, we found several management-sensitive fungal taxa which we have little or no information about their lifestyle or even their taxonomic information at lower levels. Therefore, our data should not be overgeneralized and the observations need to be confirmed in other agricultural systems. Management-sensitive fungal taxa in mineral fertilizer vs. organic manure application soils The indicator OTUs in the M soil were distinctively different from those in the NP soil (Table 4 , Figure 4 ), indicating that manure application can select or promote fungal taxa significantly different from mineral fertilizer application. Furthermore, the indicator OTUs in the M soil were mainly classified into putative saprotrophs, whilst the taxa of putative plant pathogen and animal parasite were also promoted by the NP and NPM soils. The predominance of putative saprotrophic indicator OTUs in the M soil is probably because fungal genera related to degradation processes of organic materials was closely associated with manure-based systems (Hartmann et al., 2015 ). The increase of putative plant pathogen in the NP and NPM soils is in line with previous findings that fungal genera with known pathogenic traits tend to increase proportions in mineral fertilizer application soils (Morrison et al., 2016 ; Zhou et al., 2016 ). Nevertheless, for most of taxa promoted by the NP and NPM soils, no information about their lifestyle or taxonomic information at lower levels is available. Hence, the ecological importance of the promoted taxa in mineral fertilizer-based soils remains to be determined. Furthermore, the varied taxa promoted by different fertilization regimes observed in this study is also in good agreement with previous findings, implying that only very few taxonomic groups responded uniformly to management practices (Hartmann et al., 2015 ). Patterns of fungal communities in the alkaline soils and possible predictors Similar to the findings from neutral and acidic agricultural soils (Lentendu et al., 2014 ; Francioli et al., 2016 ; Zhou et al., 2016 ), Ascomycota and Basidiomycota were the most abundant phyla and overall proportion of putative plant pathogen increased after fertilizer application in the alkaline soils investigated in this study (Figure 2 ). However, the responses of fungal taxa to fertilizer application were substantially different in the alkaline soils. For instance, while Basidiomycota decreased and Ascomycota remained stable in the fertilized soils of this study, Basidiomycota remained stable and Ascomycota decreased by mineral fertilizer application in neutral and acidic agricultural soils (Francioli et al., 2016 ; Zhou et al., 2016 ). Unlike the dominant influences of soil pH in the neutral and acidic soils (Lentendu et al., 2014 ; Zhou et al., 2016 ), we found that available P and respiration rate had the strongest influence on the total abundance and community structure of soil fungi in the alkaline soil of this study (Table 5 , Supplementary Figure 3 ). The cellulose decomposition-related saprotrophs, such as Chaetomium and Penicillium , decreased proportion in the mineral fertilizer application soils in this study, but increased in an acidic forest soil reported in Morrison et al. ( 2016 ). However, the discrepant changes of cellulose decomposition-related saprotrophs in two soils were both related to the increase of soil organic C content. This is probably because cellulose is the main form of plant input in agricultural soil (Jin and Chen, 2007 ; Thomsen et al., 2008 ) while lignin is the major source in forest soil (Frey et al., 2014 ). Moreover, based on observations from an agricultural soil of pH 4.6–6.4, Zhou et al. ( 2016 ) reported that mineral fertilizer application induced changes of soil fungal community had a potential negative impact on soil C storage. The discrepant observations between our study and Zhou et al. ( 2016 ) jointly highlight the necessity to further distinguish the possibly different responses of soil fungi between alkaline soil and acidic and/or neutral soil."
} | 4,278 |
38107563 | PMC10725675 | pmc | 9,328 | {
"abstract": "An increased abundance of macroalgae has been observed in coral reefs damaged by climate change and local environmental stressors. Macroalgae have a sublethal effect on corals that includes the inhibition of their growth, development, and reproduction. Thus, this study explored the effects of the macroalga, Caulerpa taxifolia , on the massive coral, Turbinaria peltata , under thermal stress. We compared the responses of the corals’ water-meditated interaction with algae (the co-occurrence group) and those in direct contact with algae at two temperatures. The results show that after co-culturing with C. taxifolia for 28 days, the density content of the dinoflagellate endosymbionts was significantly influenced by the presence of C. taxifolia at ambient temperature (27 °C), from 1.3 × 10 6 cells cm −2 in control group to 0.95 × 10 6 cells cm −2 in the co-occurrence group and to 0.89 × 10 6 cells cm −2 in the direct contact group. The chlorophyll a concentration only differed significantly between the control and the direct contact group at 27 °C. The protein content of T. peltata decreased by 37.2% in the co-occurrence group and 49.0% in the direct contact group compared to the control group. Meanwhile, the growth rate of T. peltata decreased by 57.7% in the co-occurrence group and 65.5% in the direct contact group compared to the control group. The activity of the antioxidant enzymes significantly increased, and there was a stronger effect of direct coral contact with C. taxifolia than the co-occurrence group. At 30 °C, the endosymbiont density, chlorophyll a content, and growth rate of T. peltata significantly decreased compared to the control temperature; the same pattern was seen in the increase in antioxidant enzyme activity. Additionally, when the coral was co-cultured with macroalgae at 30 °C, there was no significant decrease in the density or chlorophyll a content of the endosymbiont compared to the control. However, the interaction of macroalgae and elevated temperature was evident in the feeding rate, protein content, superoxide dismutase (SOD), and catalase (CAT) activity compared to the control group. The direct contact of the coral with macroalga had a greater impact than water-meditated interactions. Hence, the competition between coral and macroalga may be more intense under thermal stress.",
"conclusion": "Conclusions The shift from coral dominance to algal dominance that has been observed in many reefs due to global climate change and overfishing. Coral dominance is sensitive to key algal groups and other benthic groups, and shifts in ecosystem phases have a noticeable impact ( Tebbett et al., 2023 ). The results of this study showed that C. taxifolia negatively affected the endosymbiont density, chl a content, feeding rate, growth rate and protein content of T. peltata and increased the antioxidant activity at 27 °C. The combination of elevated temperature and macroalgae interactions may further exacerbate the adverse effects on corals. Future studies are needed to explore the interactions of multiple coral-macroalgal species under climate change. Because of the vulnerability and sensitivity of coral reef ecosystem, relevant entities should take urgent steps to prevent CO 2 emissions that exceed the goals of the Paris Climate Agreement. Herbivorous fish populations should also be restored to improve macroalgae management in reefs.",
"introduction": "Introduction The recent effects of climate change and other anthropogenic impacts have caused the severe degradation of coral reefs worldwide ( Leggat et al., 2022 ). The first mass coral bleaching event was observed in 1998 and it killed approximately 8% of the world’s coral; an additional 14% of corals were lost between 2009 and 2018 ( Souter et al., 2021 ). Many studies have asserted that ocean warming is a major factor in the reduction of coral cover ( Hughes et al., 2017 , 2019 ; Lough, Anderson & Hughes, 2018 ; Leggat et al., 2022 ). For example, the successive bleaching events in 2016–2017 devastated Australia’s Great Barrier Reef and resulted in an 89% decline in larval recruitment in 2018 compared to historical levels ( Hughes et al., 2017 , 2019 ; Lough, Anderson & Hughes, 2018 ). During this period, 31% of reefs experienced 8–16 degree heating weeks (DHWs, °C-weeks). A decline in coral cover may lead to an increase in the cover of other benthic organisms in the reefs, such as macroalgae ( Fulton et al., 2019 ). The impact of herbivorous fishes is mostly ignored, although some studies assert that overfishing and nutrient pollution is the main cause of phase shifts towards macroalgae ( Barott et al., 2012 ). Research had shown that prior to 2011, the estimated global average cover of algae was low (~16%) and stable for 30 years. Since 2011, the amount of algae on the world’s coral reefs has increased by about 20% ( Souter et al., 2021 ). Thus, the coral reef ecosystem is undergoing an ecological phase transition to that of an ecosystem dominated by macroalgae. Macroalgae are functional communities that are important for stabilizing reef structure ( Fulton et al., 2019 ), generating primary productivity ( Fulton et al., 2014 , 2019 ), maintaining nutrient cycling in reef areas, and providing food sources for herbivores ( Dubinsky & Stambler, 2011 ). However, there is competition between macroalgae and corals. Macrolgae harm corals through direct contact ( Coyer et al., 1993 ; Manikandan et al., 2021 ) and allelopathy ( Bonaldo & Hay, 2014 ; Fong et al., 2020 ), weakening the photosynthetic performance of symbiodiniaceae ( Rasher et al., 2011 ), causing the retraction of polyps ( Jompa & Mccook, 2003 ), increasing the number of pathogenic microorganisms ( Clements et al., 2020 ; Rasher & Hay, 2010 ), triggering coral bleaching ( Bonaldo & Hay, 2014 ), and resulting in the reduced calcification and coral growth, fecundity, survival rate, and settlement rate ( Fong et al., 2020 ; Tanner, 1995 ; Leong et al., 2018 ; Rasher & Hay, 2010 ). Specifically, macroalgae affect coral feeding, endosymbiont function, tissue recovery, and oxidative stress response. Morrow & Carpenter (2008) found that Dictyopteris undulata weakened the particle capture rates of Corynactis californica by redirecting particles around polyps and causing contraction of the feeding tentacles. The dissolved organic carbon (DOC) and terpenoids released by macroalgae decreased photosynthesis and the density of endosymbionts ( Rasher et al., 2011 ; Smith et al., 2006 ; Diaz-Pulido & Barrón, 2020 ). Bender, Diaz-Pulido & Dove (2012) asserted that the green filamentous macroalga, Chlorodesmis fastigiate , significantly reduced tissue recovery in Acropora pulchra and led to the infection of A. pulchra with ciliates. High levels of reactive oxygen species (ROS, a toxic byproduct of biological aerobic metabolism) could cause damage to cells ( Blanckaert et al., 2021 ). Shearer et al. (2012) found that the oxidative stress response of Acropora millepora was activated in response to ROS by altering the transcription factors after contact with the macroalga Chlorodesmis fastigiata and its hydrophobic extract over a short-term period (1 and 24 h). The oxidative imbalance results in rapid protein degradation and eventually to apoptosis and/or necrosis when compensatory transcriptional action by the coral holobiont insufficiently mitigates damage. In addition, the combined effects of ocean warming, acidification, and macroalgae contact could significantly alter the physiological response of corals ( Chadwick & Morrow, 2011 ; Kornder, Riegl & Figueiredo, 2018 ; Brown et al., 2019 ; Rölfer et al., 2021 ). Rölfer et al. (2021) have shown that light enhanced calcification (LEC) rates of Porites lobata were negatively affected after contact with Chlorodesmis fastigiata in an ocean warming and acidification scenario, compared to coral under ambient conditions. Typically, the coral-algal competition is related to seasonal and temporal cycles. These, in turn, may be related to the abundance, biomass, and composition of macroalgae, as well as the seasonal dynamics of temperature, p CO 2 , and light intensity ( Brown et al., 2019 , 2020 ). The sensitivity of various macroalgae to environmental stressors is also different. For example, intermediate levels of ocean warming could enhance the growth and production of Laurencia sp. and Lobophora sp., which was not the case for Sargassum sp. ( Fulton et al., 2014 ; Hernández et al., 2018 ). Additionally, overfishing and eutrophication have been shown to lead to an increased growth rate of some kinds of macroalgae ( Lapointe & Bedford, 2010 ), which may indirectly enhance the competitive ability of macroalgae. Therefore, to better understand the resilience of coral reef ecosystems in the future, it is necessary to determine how coral-algal interactions are impacted by global and local stressors. According to the China Ocean Climate Monitoring bulletin ( www.oceanguide.org.cn ), the average sea surface temperature (SST) in the Xuwen Sea area was 27–30 °C from May to September in 2020. In 2021–2022, the average SST that caused coral bleaching in the Great Barrier Reef from December to April was 28–30 °C ( Spady et al., 2022 ). It is plausible that the physiological responses of corals in the Xuwen Coral Reef National Nature Reserve of China may be affected by thermal stress when SST reaches 30 °C. During thermal stress, coral feeding rates are drastically reduced and more energy is needed to maintain biological processes (DNA repair etc .,) to resist heat stress ( Ferrier-Pagès et al., 2010 ; Chakravarti et al., 2020 ). Triggered by thermal stress, ROS may be produced by the endosymbionts mainly due to PS II dysfunction caused by damage to the D1 protein ( Warner, 1999 ) and host cells ( Nii & Muscatine, 1997 ). The increase in the production of ROS is a stress signaling mechanism that can potentially trigger an oxidative stress and apoptotic cascade in coral cells ( Hensley et al., 2000 ; Drury et al., 2022 ). The relationship between macroalgae and corals under climate change conditions remains inconclusive. To investigate the effects of macroalgae on hermatypic coral under ocean warming, the massive coral, Turbinaria peltata , and macroalga, Caulerpa taxifolia , which are common species with frequent interactions in the Xuwen Sea, were selected as study species. C. taxifolia is a multinucleate siphonous green alga and is known to have great invasive potential worldwide ( Zubia et al., 2020 ). Furthermore, it has been found that C. taxifolia can produce potential allelochemicals, such as monoterpenes and sesquiterpenes ( Guerriero et al., 1992 , 1993 ). Given that C. taxifolia usually grows on various hard substrates close to large numbers of live coral colonies, the physical and chemical impact has to be better understood. To evaluate the effect of its chemical and physical effects, an indirect contact group was used to investigate chemical effects, and the direct-contact group was used to explore the combined effects of physical and chemical processes. In this study, we show that increased temperature, representing the ocean-warming range projected for this century ( Spady et al., 2022 ), enhances the ability of seaweeds to impact the physiology of corals, potentially shifting the competitive interaction between corals and seaweeds in favour of seaweeds. It provides a reliable basis for the evolution of competition between corals and macroalgae under future global changes.",
"discussion": "Discussion This study explored the crucial issue of how the physiology and oxidative stress response of a hermatypic coral are affected by macroalgae at elevated temperatures. We set up three treatments of the macroalga C. taxifolia (direct contact, indirect, water-mediated presence, no alga) to act on the coral T. peltata at ambient temperatures (27 °C) and elevated temperature (30 °C). The results demonstrated that macroalgal presence increased the antioxidant activity at 27 °C. In addition, combined with elevated temperature, there was a remarkably synergistic effect that macroalgae impacted the feeding rate, protein contain and further increased the oxidative stress of the coral, in which contact with algae had a more severe effect than indirect interaction. Effects of C. taxifolia on endosymbiont of T. peltata Algae was found to influence the average endosymbiont density and chl a content of T. peltata , however, no bleaching occurred. A number of studies have reported that coral’s photosynthetic efficiency decreased (Fv/Fm), or bleaching occurred, when there was direct or indirect contact with macroalgae. However, not all coral species are equally susceptible to algae and not all algae will have deleterious effects on corals ( Smith et al., 2006 ; Rasher & Hay, 2010 ; Fong et al., 2020 ). Rasher & Hay (2010) suggested that Padina perindusiata and Sargassum sp. did not inhibit photosynthetic efficiency or induce bleaching of Porites porites , which might be explained by the fact that the 20-day interaction period was too short to impact the coral health. Additionally, T. peltata is a massive coral that could resist environmental pressure by increasing its basic metabolism ( Loya et al., 2001 ). This may explain why there was no significant bleaching effects of macroalgae on endosymbiont density. There was a stronger effect for direct contact compared to co-occurrence in chl a concentration at 27 °C, suggesting that physical mechanisms are still the main way in which macroalgae affect corals. Effects of C. taxifolia and thermal stress on the physiology of T. peltata The feeding rate was affected by elevated temperature. Johannes & Tepley (1974) also found that the feeding rate of coral decreased in heat stress because of the polyp contraction or a loss of nematocyst function. Our results suggest that a decrease in chl a content and endosymbiont density was the reason why the feeding rate was impacted in elevated temperature. Endosymbionts provide photosynthate to host cells ( Van Oppen & Blackall, 2019 ). A decrease in the endosymbiont density at high temperatures may result in reduced energy expenditure to maintain normal physiological functions and reduce resistance to predation. This study showed that contact with C. taxifolia resulted in the greatest reduction in the feeding rate at 30 °C. In summary, thermal stress might a crucial factor affecting the feeding ratio of T. peltata , which became more severe when in contact with macroalgae. Macroalgae can induce reduced protein content in corals. Damage to coral tissue by contact with macroalgae has been documented in many studies. Bender, Diaz-Pulido & Dove (2012) asserted that Acropora sp. lost tissue and decreased its growth rate due to allelopathy mechanisms after coming into contact with Chlorodesmis fastigiata . In fact, macroalgae may transfer many allelopathic substances to corals, altering the structure of the microbial community and impacting the physiological processes of corals ( Fong et al., 2020 ). This damage may ultimately reduce the protein content. Under stress, massive corals with thicker tissues may overcome the effects of endosymbiont loss through catabolism ( DeCarlo & Harrison, 2019 ). Macroalgae may affect coral tissue by creating anoxic zones. Barott et al. (2009) demonstrated that after the interactions between corals ( Pocillopora verrucosa , Montipora sp.) and some species of macroalgae ( e.g ., Gracilaria sp., Bryopsis sp., and various turf algae), the characteristic patterning of coral pigments and polyps was altered and the tissue appeared damaged. The growth rate of coral was altered by the macroalgae and temperature, the results were consistent with previous studies ( Tanner, 1995 ; Rölfer et al., 2021 ; Vega Thurber et al., 2012 ; Vermeij et al., 2009 ). At 27 °C, the growth rate of the direct contact group was lower than that of the co-occurrence group. Therefore, the effects of seaweed on coral growth may require direct contact at ambient temperatures ( Clements et al., 2020 ). Brown et al. (2019) also demonstrated that coral growth was reduced or even negative at 30 °C when in contact with algae. Longo & Hay (2015) determined that corals showed signs of stress at 30 °C, and contact with Halimeda heteromorpha further contributed to a decreased growth rate and increased mortality rate. These results may be due to the simultaneous decline in the autotrophic and heterotrophic activities of corals under the impacts of thermal stress or macroalgae, resulting in a drop of protein content and ultimately affecting the growth rate. Effects of C. taxifolia and thermal stress on oxidative stress of T. peltata Corals under thermal stress may produce ROS ( Blanckaert et al., 2021 ). Downs et al. (2002) documented that when exploring the varied oxidative stress response of coral under seasonal change, the SOD in summer was three times higher than that in winter. This study determined that SOD in corals was higher when macroalgae were present (the effect was stronger for direct contact compared to co-occurrence group), the increased temperature, or there was the synergistic effect of both. Thus, weakened corals were found to be more vulnerable to competition from algae, which was also supported by the results of Diaz-Pulido et al. (2010) . The level of both antioxidant enzyme activities was similar to thermal stress alone when C. taxifolia indirectly contacted T. peltata . These results indicate that the stress triggered by macroalgal allelochemicals on coral was equivalent to that induced by increased temperature. The temperature effect was stronger in CAT activity compared with SOD activity under direct contact, which may be related to the reduced protein content in coral tissues caused by elevated temperature under the direct contact treatment. Due to the evident decrease in the protein content of coral tissues in the direct contact group, the amount of antioxidant enzymes produced by coral is not enough to resist the damage of ROS."
} | 4,549 |
29263352 | PMC5738337 | pmc | 9,329 | {
"abstract": "Photosynthesis of microalgae enables conversion of light energy into chemical energy to produce biomass and biomaterials. However, the efficiency of this process must be enhanced, and truncation of light-harvesting complex (LHC) has been suggested to improve photosynthetic efficiency. We reported an EMS-induced mutant (E5) showing partially reduced LHC in Chlorella vulgaris . We determined the mutation by sequencing the whole genome of WT and E5. Augustus gene prediction was used for determining CDS, and non-synonymous changes in E5 were screened. Among these, we found a point mutation (T to A) in a gene homologous to chloroplast signal recognition particle 43 kDa (CpSRP43). The point mutation changed the 102nd valine to glutamic acid (V102E) located in the first chromodomain. Phylogenetic analyses of CpSRP43 revealed that this amino acid was valine or isoleucine in microalgae and plants, suggesting important functions. Transformation of E5 with WT CpSRP43 showed varying degrees of complementation, which was demonstrated by partial recovery of the LHCII proteins to the WT level, and partially restored photosynthetic pigments, photosynthetic ETR, NPQ, and growth, indicating that the V102E mutation was responsible for the reduced LHC in E5.",
"introduction": "Introduction Microalgae are photosynthetic eukaryotes (excluding plants) known for their high photosynthetic efficiency and potential for use in the production of biomass and biomaterials. However, they are far less efficient relative to their theoretical photosynthetic potential partly due to an oversized photosynthetic antenna, which functions as a light-harvesting complex (LHC) 1 . Microalgae have evolved to cope with low light levels and competition with other organisms, which have led to an increase in the LHC size. This results in shading effects in an artificial high-density culture, contributing to the reduced photosynthetic efficiency 2 . LHCs are composed of various photosynthetic pigments and pigment-binding proteins. In higher plants and green algae, LHCI associated with PSI and LHCII associated with PSII consist of LHCA and LHCB proteins, respectively 3 . The proteins needed for the assembly of LHCs and photosystems are guided and integrated into the developing thylakoid membrane by the chloroplast signal recognition particle (CpSRP) pathway 4 , 5 , which is similar to the signal recognition particle (SRP) pathway in bacteria 6 . Nuclear-encoded LHC proteins are targeted to chloroplasts via their transit peptide and are imported into the chloroplast stroma. CpSRP43 and CpSRP54 form a transit complex together with imported LHC proteins 4 , 7 , and is recognized by the signal recognition receptor CpFTSY. This LHC-CpSRP43-CpSRP54-CpFTSY complex is guided to the chloroplast SRP insertase ALB3 protein 8 , 9 , and the LHC protein is integrated into the thylakoid membrane. The CpSRP complex is disassembled for another cycle of LHC protein integration. Chloroplast-encoded proteins such as PSII reaction-center proteins are integrated into the thylakoid membrane via the co-translational pathway, which involves CpSRP54, ALB3, CpSECY and potentially CpFTSY 4 , 10 , 11 . Recent studies of green algae show that the functions of the proteins involved in the CpSRP pathway differ from those of higher plants according to analyses of knockout mutants of CpSRP pathway-related genes in Chlamydomonas reinhardtii \n 4 , 9 , 12 – 14 . These knockout mutants have phenotypes of truncated light-harvesting chlorophyll antenna (TLA) mutations, which show reduced levels of chlorophylls and LHC proteins. These TLA mutations have been suggested as a strategy for the genetic engineering of microalgae to improve photosynthetic efficiency 1 , 15 . Microalgae have evolved to possess extensive light-harvesting chlorophyll antenna to maximize light absorption in light-limiting environments 16 . This process is advantageous for the survival of cells in an environment in which they must compete with other organisms. However, in high-density monocultures for the production of biomass and other target products, the maximized light absorption by oversized LHCs leads to the wasteful dissipation of light energy into heat via non-photochemical quenching (NPQ) and uneven light distributions in cell cultures 17 , 18 . TLA mutants showed improved photosynthetic activity and accelerated growth under high light conditions by alleviating the over-absorption of light and by allowing greater light penetration 14 , 15 , 19 , 20 . Moreover, recent studies have demonstrated that the application of a TLA strategy can enhance crop yields 21 , 22 . We previously reported a TLA mutant (E5) generated by ethyl methanesulfonate (EMS) mutagenesis in Chlorella vulgaris which showed improved biomass productivity under high light conditions 20 . Identification of the responsible gene in E5 can provide a valuable genetic resource for potential target genes to improve biomass production in Chlorella and other microalgae. By sequencing the whole genome of WT and E5, and with Augustus gene modeling, we were able to identify non-synonymous mutations in E5. Among these, a point mutation was found in CpSRP43, and complementation analyses by transforming E5 with the WT gene confirmed that this mutation was responsible for the TLA phenotype in E5. To the best of our knowledge, the determination and complementation of the random mutation is the first in Chlorella , except for the complementation of the nitrate reductase marker 23 , 24 .",
"discussion": "Discussion Photosynthetic organisms usually have large arrays of LHC and photoprotection mechanisms to survive under fluctuating light conditions 16 , 17 , 37 . However, these mechanisms can be a main bottleneck and thus reduce photosynthetic efficiency under artificial cultivation conditions. Truncation of LHC has been suggested as a solution for these problems, and previous studies have shown improvements of photosynthetic efficiency and biomass productivity by alleviating the over-absorption of light on the culture surface and achieving a better light distribution over the whole culture in the model microalga Chlamydomonas reinhardtii \n 1 , 4 , 18 , 19 , 38 , 39 . We demonstrated that the antenna truncation strategy also worked in a commercial microalgal species, C . vulgaris \n 20 . This study determined a SNP at the CvCpSPR43 locus which was responsible for the TLA mutant phenotype in E5. Nucleotide transversion from T to A was induced by an EMS treatment to generate the E5 mutant 25 , 26 , resulting in an amino acid change from V to E (V102E) located in CD1 of CvCpSRP43. The reduction of the LHCPs in E5 was limited to LHCB1- and LHCB4-like proteins, which are major and minor peripheral LHCPs in PSII, respectively. It is interesting that the partial reduction of LHCPs in E5 is different from the CpSRP43 knockout mutant of C . reinhardtii , which shows severe reductions of all major and minor LHCPs in PSII, including LHCB1/LHCB2, LHCB3, LHCB4 and LHCB5 13 . The discrepancy may be caused by the nature of the mutations, as the point mutation in E5 is located in V102E in CD1, while that of C . reinhardtii is disrupted by insertional mutagenesis leading to complete knockout of the gene functions 13 . The valine residue together with the isoleucine found mostly in plants is conserved and may be critical for certain functions. It is tempting to speculate that this amino acid or CD1 is involved in interactions with certain LHCPs. Thus far, it is known that the deletion of the CD1 domain from CpSRP43 increases the GTP hydrolysis activity 40 , suggesting that CD1 negatively regulates GTP hydrolysis by the CpSRP43/CpSRP54/CpFtsY complex 41 – 45 . Guanine nucleotide binding by CpSRP43/CpSRP43/CpFtsY releases LHCP from the CpSRP complex, and GTP hydrolysis is required for the recycling of CpSRP from CpFtsY 46 . Given these findings, it is conceivable that the V102E mutation in E5 somehow reduced the GTPase activity and disturbed the complete assembly of the light-harvesting antenna. It has also been reported that CD1 is rigidly fused to the neighboring ankyrin repeats that bind LHCPs 47 . It is therefore possible that the negative charge in V102E might have negative effects on interaction with the LHCB1- and LHCB4-like proteins, though this possibility requires further studies. It is also interesting to note that the amounts of the CvCpRRP43 protein and RNA were increased in E5, whereas they were decreased (or partially recovered) in the complemented strains (Fig. 3 ). This is reminiscent of a type of feedback regulatory mechanism known as positive feedback loop regulation, where defective downstream processes can activate the expressions of the upstream cognate genes to compensate for defective functions. Such an example can be found in the increased expression of a forkhead transcription factor (FOXO3a) in patients diagnosed with Huntington’s disease 48 . This type of auto-regulatory mechanisms has not been reported in relation to microalgae, and is worthwhile to study this in more detail. In complementation tests, recovery of the phenotypes occurred consistently at the intermediate level between WT and E5. Transformation of E5 with the WT CpSRP43 gene would generate complemented strains harboring both mutated Cpsrp43 and normal CpSRP43 , which may interact negatively. In addition, the 29B HSP70 promoter was used to drive the expression of WT CpSRP43 instead of the endogenous promoter in C . vulgaris . This transgenic expression of CpSRP43 may not be sufficient to replace the mutated Cpsrp43 completely in E5. However, all complemented strains showed consistent restoration of the TLA phenotypes (with small standard error), including photosynthetic pigments, LHCPs and photosynthetic parameters although small sample size ( n = 2) was used to calculate P-value. This indicates that the V102E mutation in E5 was responsible for the TLA phenotypes of E5. This also suggests that the CpSRP43 homologs are good targets for knock-down to improve biomass productivity in Chlorella \n 20 . It should also be noted that there could be additional mutation(s) contributing to the phenotypes in E5, but this possibility cannot be pursued currently due to incomplete genomic information. SNPs in E5 should be reevaluated when genome sequencing is completed for Chlorella vulgaris . Even though the TLA strategy has been demonstrated as a promising concept for improving microalgal productivity in lab-scale experiments under high light irradiances, potential caveats should be considered and studied for further application on outdoor large scales where the light irradiance is fluctuating. Under the low light conditions, reduced ability to absorb light in TLA mutants lowers photosynthetic productivity per cell. Considering light-limiting conditions, where photo-damage is not severe, photosynthetic growth would decrease as shown in this study (Fig. 5a and b ). Besides, the TLA strategy is considered advantageous only under narrow range of conditions 49 . In this aspect, the TLA strategy might show better performance in continuous cultivation system because growth conditions can be easily controlled. For example, chemostat can be incorporated into the continuous cultivation system to maintain proper cell density resulting in maximum productivity, which can avoid light attenuation due to extremely dense culture. In fact, the potential for large scale application appears to be valid, since it has been reported that a TLA mutant can improve biomass productivity under the outdoor light condition 15 . Conclusively, a single nucleotide change resulting in the V102E mutation in CD1 of CvCpSRP43was responsible for the TLA phenotype in E5. This was confirmed by the complementation experiments, where the complemented strains showed the partial recovery of the TLA phenotypes. In the identification and complementation experiments, we also made several interesting observations, including that of the LHCP-specific role for the 102 nd amino acid in CD1 and the possible positive feedback regulatory mechanism of the gene expression of CvCpSRP43 , all of which deserve further studies. It should also be noted that the homologs of CvCpSRP43 would be good knock-down targets to improve biomass production in other industrial microalgae."
} | 3,086 |
27582299 | PMC5007514 | pmc | 9,330 | {
"abstract": "The NC10 phylum is a candidate phylum of prokaryotes and is considered important in biogeochemical cycles and evolutionary history. NC10 members are as-yet-uncultured and are difficult to enrich, and our knowledge regarding this phylum is largely limited to the first species ‘ Candidatus Methylomirabilis oxyfera’ ( M. oxyfera ). Here, we enriched NC10 members from paddy soil and obtained a novel species of the NC10 phylum that mediates the anaerobic oxidation of methane (AOM) coupled to nitrite reduction. By comparing the new 16S rRNA gene sequences with those already in the database, this new species was found to be widely distributed in various habitats in China. Therefore, we tentatively named it ‘ Candidatus Methylomirabilis sinica’ ( M. sinica ). Cells of M. sinica are roughly coccus-shaped (0.7–1.2 μm), distinct from M. oxyfera (rod-shaped; 0.25–0.5 × 0.8–1.1 μm). Notably, microscopic inspections revealed that M. sinica grew in honeycomb-shaped microcolonies, which was the first discovery of microcolony of the NC10 phylum. This finding opens the possibility to isolate NC10 members using microcolony-dependent isolation strategies.",
"discussion": "Results and Discussion Activity determination of the culture The denitrifying methanotrophic culture was originally enriched from paddy soil with natural freshwater medium for 18 months 9 and artificial inorganic medium 31 for the next 42 months. In the last 6 months, the concentrations of the trace elements iron and copper (important components of the key enzymes in the central metabolism) in the medium were increased to 20 and 10 μM, respectively, to accelerate the growth of the methanotrophs 32 . To assess the denitrifying methanotrophic activity of the culture, batch activity tests were performed, and the results are shown in Fig. 1 . As expected, the methane oxidation and nitrite reduction were coupled in the culture, showing good denitrifying methanotrophic activity with a rate of 0.084 ± 0.004 μmol CH 4 per hour, and there was no activity in the control. The nitrite reduction rate of AOM coupled to nitrite reduction process was 0.21 ± 0.01 μmol NO 2 − per hour, calculated according to He et al. 31 . The ratio of the methane oxidation rate to the nitrite reduction rate was 3.20 ± 0.03:8, close to the stoichiometric ratio of 3:8 ( Eq. 1 ). Phylogenetic analysis of the NC10 phylum Both 16S rRNA and pmoA gene sequences of the NC10 bacteria in the culture were phylogenetically analyzed, and the results are shown in Fig. 2 . The phylogenetic analysis of the 16S rRNA genes ( Fig. 2a ) indicated that the representative sequence in the culture (indicated as ‘ Candidatus Methylomirabilis sinica’, M. sinica ) belonged to group A of the NC10 phylum but was in a distinct cluster with M. oxyfera . Several NC10 sequences from China were in the same cluster as the representative sequence. The sequence similarity between the representative sequence and the 16S rRNA gene sequence of M. oxyfera was 96.9%. The phylogenetic analysis of the pmoA genes ( Fig. 2b ) also indicated that the representative pmoA gene sequence (also indicated as ‘ Candidatus Methylomirabilis sinica’) was in a distinct cluster with M. oxyfera . The representative pmoA sequence from this culture had a low sequence similarity of 85.3% to the pmoA sequence of M. oxyfera . Moreover, both 16S rRNA and pmoA phylogenetic trees ( Fig. 2 ) suggested the existence of the third cluster that contained the sequences from Lake Biwa sediments 17 and a peatland enrichment culture 23 . To analyze the correlations between the NC10 gene sequences obtained in this work and the sequences in the previous studies, 2,478 16S rRNA and 1,314 pmoA sequences of the NC10 phylum were retrieved from NCBI GenBank (date: 26-Jun-2016). The sequence similarities with M. oxyfera and M. sinica are shown in Fig. 3 . 154 16S rRNA sequences not only have high similarity (>97 %) with M. sinica but also higher than those with M. oxyfera (marked in Fig. 3a ); 28 pmoA sequences have high similarity (>93 %) with M. sinica (marked in Fig. 3b ). According to the sequence descriptions in NCBI GenBank, these sequences were all obtained from Chinese ecosystems, including lake sediment, swamp sediment, paddy soil, forest soil, coastal sediment, estuary sediment and bay sediment. Microscopic observation of the culture Fluorescence in situ hybridization (FISH) images ( Fig. 4a–f ) revealed that NC10 bacteria grew in a large numbers of microcolonies (clusters of the identical cells). The bright field images of the confocal laser scanning microscope (CLSM) ( Fig. 4g–l ) present the structure of the microcolonies clearly. The microcolonies are dense and appear in round or oval shapes with sizes of 10–30 μm. All cells of the NC10 bacteria in this culture were roughly coccus-shaped with sizes of 0.7–1.2 μm, whereas the previous NC10 bacteria (cluster M. oxyfera and the third cluster in Fig. 2 ) enriched in other laboratories were rod-shaped 5 6 23 with a polygonal appearance under electron microscopy 33 . Close observation of the bright field images ( Fig. 4g–l ) suggested that the cells of our study were also polygonal ( e.g., pentagon, hexagon, and heptagon). Due to the polygonal shapes of the single cells and the dense structure of the microcolonies, these microcolonies resemble honeycombs, especially the microcolony in Fig. 4h . Moreover, there was some dense matter on the surfaces of the microcolonies, which is particularly clear in Fig. 4k (black line surrounded the microcolony, indicated by a white arrow), which might be important for the stability of the microcolonies. All the NC10 bacteria were observed in microcolonies in the culture, and all the other organisms were detected in free cells ( Fig. 5a ). The microcolonies of NC10 bacteria, the free cells of other organisms and the abiotic matters together formed the flocs, and the flocs were all similar in the culture. Similar phenomena (one species of microorganism in microcolonies and the others in free cells) were also observed in other active sludge systems 30 . Based on this feature of the culture, a conceptual model of the floc was proposed that the dense microcolonies of NC10 bacteria and the free cells of other bacteria were embedded individually in the flocs, as shown in Fig. 5b . NC10 bacteria can be isolated on the basis of this feature of the culture, and the microcolonies could be selected based on the different particle sizes or settling velocities 30 34 . Microcolony formation is a common behavior of microorganisms but was not described in previous studies on NC10 bacteria 5 6 8 23 24 35 . In environmental microbiology, microcolony formation has attracted attention due to its importance in the structure of activated sludge 36 37 and the isolation of uncultured bacteria 28 29 30 . From the CLSM images, the microcolonies of NC10 bacteria are roughly spherical, dense and strong. It might be attributed to the intensive shear caused by high-rate magnetic stirring in the bioreactor. Due to poor settleability, single cells were easily withdrawn from the system with the medium exchange. Therefore, the formation of microcolony benefited NC10 bacteria “stay” in the reactor, whereas other microorganisms (single cells) were washed out when the culture was settled and the supernatant was replaced with fresh medium. Previous research indicated that extracellular DNA 37 and other extracellular polymeric substances (EPS) 36 were important for microcolony strength in microbial flocs and biofilms. In this work, it seemed that the dense matter (like inorganic precipitants) on the surface of the microcolonies was also important and could protect microcolonies from disintegration. The microcolony formation was long regarded as a life strategy of microorganisms under the nutrient-poor or adverse conditions 38 , and it might benefit NC10 bacteria in the competition with other microorganisms 39 , such as heterotrophic denitrifiers 24 . Moreover, the dense aggregation of cells enhanced the interactions (material, signal, gene, etc. ) among cells 40 41 , and it might stimulate the growth of NC10 bacteria. New denitrifying methanotrophs of the NC10 phylum The similarity of the 16S rRNA gene sequences between the M. oxyfera and the representative sequence in this work (positions 28 to 1,511) was 96.9%. According to the species delineation of 97% similarity and genus of 95% of the 16S rRNA gene for bacteria 42 43 , the representative sequence in this work represented a new species within the genus ‘ Candidatus Methylomirabilis’. This species was first obtained in China and has only been detected in Chinese habitats, so we tentatively proposed the name ‘ Candidatus Methylomirabilis sinica’ ( M. sinica ). The geographic distribution of this species may not be true to its name because most previous studies on NC10 bacteria in natural environments were performed in China 44 . More ecological investigations on NC10 bacteria should be performed in other countries to verify whether M. sinica exists in other regions. The representative sequence of the pmoA genes in this work (85.3% similarity to M. oxyfera ) also showed that a new species was obtained in the culture, according to the species boundary of 93% of the pmoA gene for methanotrophs 45 . The activity tests ( Fig. 1 ) demonstrated that the culture had the activity of AOM coupled to nitrite reduction. Therefore, the dominant species M. sinica should be a novel denitrifying methanotroph, affiliated to the genus ‘ Candidatus Methylomirabilis’ in the NC10 phylum. The FISH primer S-*-DBACT-1027-a-A-18 could be completely aligned to the target positions of the 16S rRNA gene sequence of M. sinica ( Table S2 ), and M. sinica was the only NC10 bacteria in the culture (detected by 8F/1492R), which indicated that the cells hybridized by this NC10-specific primer in FISH images ( Fig. 4 ) should be M. sinica . So far, only denitrifying methanotrophs in the new cluster M. sincia (see Fig. 2 ) were observed as coccus and those in other clusters are rod-shaped ( Table S3 ). It further indicated that a new species was obtained. Key physiology of M. sinica The important physiological parameters of M. sinica were determined in this study and in our previous works with the same culture. The optimal temperature and pH ranges were measured by batch experiments, and the values were 30 to 40 °C and 7.0 to 8.0, respectively 31 . M. sinica can grow in both freshwater 9 and saline environments 10 . In the previous work, we obtained a halophilic NC10 culture that was also dominated by M. sinica 10 , and its reference sequences, KM888211 for 16S rRNA and KM979292 for pmoA , are shown in Fig. 2 , respectively. The doubling time of M. sinica was approximately 25.0 days 46 , longer than that of M. oxyfera (1–2 weeks 11 ), and the growth rate was estimated to be 0.028 ± 0.002 d −1 \n 46 . The apparent substrate affinity constants for methane and nitrite were measured in this work, and they were 7.8 ± 1.2 μM and 8.9 ± 2.9 μM, respectively, similar to the results from the previous halophilic NC10 culture (9.8 ± 2.2 μM for methane and 8.7 ± 1.5 μM for nitrite 10 ). The specific cell activity of M. sinica was approximately 0.3 fmol CH 4 day −1 cell −1 in freshwater 47 and 0.14 fmol CH 4 day −1 cell −1 in saline water 10 , higher than that of M. oxyfera (0.09 fmol CH 4 day −1 cell −1 \n 5 ). It may be explained by the size of the cell; M. sinica is significantly larger than M. oxyfera (0.7–1.2 × 0.7–1.2 μm vs. 0.25–0.5 × 0.8–1.1 μm). M. sinica bacteria in natural habitats The phylogenetic trees ( Fig. 2 ) and sequence similarity analyses ( Fig. 3 ) indicated that the species M. sinica is widely distributed in natural environments. These M. sinica sequences were retrieved from freshwater systems (freshwater lake, swamp, wetland, and paddy soil) 9 19 and low saline water environments (estuary, coast, and bay) 10 21 48 , but M. sinica sequences have not been detected in high saline water environments (such as saline lakes and deep sea) 18 22 . These findings suggested that M. sinica exists in various aquatic environments with low salinities and may be ecologically important in these ecosystems. The existing primers for the NC10 phylum were designed based on M. oxyfera , and they may have bias for M. oxyfera . A mismatch was discovered between the sequences of the most widely used primer qP1F 5 and M. sinica in this work ( Table S2 ). This mismatch was at the last base of primer qP1F (at the 3′ end), which might influence the PCR amplification of M. sinica sequences. The last base of qP1F is guanine (G), but the corresponding position in the sequence of M. sinica is adenine (A). Therefore, the primer qP1F should be modified or redesigned to remove the PCR bias. In previous studies, the abundance and the diversity of M. sinica in natural habitats may have been underestimated due to this mismatch."
} | 3,268 |
26484255 | PMC4583632 | pmc | 9,331 | {
"abstract": "The soil-mousse surrounding a geothermal spring was analyzed for bacterial and archaeal diversity using 16S rRNA gene amplicon metagenomic sequencing which revealed the presence of 18 bacterial phyla distributed across 109 families and 219 genera. Firmicutes, Actinobacteria, and the Deinococcus-Thermus group were the predominant bacterial assemblages with Crenarchaeota and Thaumarchaeota as the main archaeal assemblages in this largely understudied geothermal habitat. Several metagenome sequences remained taxonomically unassigned suggesting the presence of a repertoire of hitherto undescribed microbes in this geothermal soil-mousse econiche."
} | 162 |
28602523 | null | s2 | 9,332 | {
"abstract": "Engineered biological systems such as genetic circuits and microbial cell factories have promised to solve many challenges in the modern society. However, the artisanal processes of research and development are slow, expensive, and inconsistent, representing a major obstacle in biotechnology and bioengineering. In recent years, biological foundries or biofoundries have been developed to automate design-build-test engineering cycles in an effort to accelerate these processes. This review summarizes the enabling technologies for such biofoundries as well as their early successes and remaining challenges."
} | 152 |
37593528 | PMC10432169 | pmc | 9,333 | {
"abstract": "Microbial electrochemical technologies have been extensively employed for phenol removal. Yet, previous research has yielded inconsistent results, leaving uncertainties regarding the feasibility of phenol degradation under strictly anaerobic conditions using anodes as sole terminal electron acceptors. In this study, we employed high-performance liquid chromatography and gas chromatography-mass spectrometry to investigate the anaerobic phenol degradation pathway. Our findings provide robust evidence for the purely anaerobic degradation of phenol, as we identified benzoic acid, 4-hydroxybenzoic acid, glutaric acid, and other metabolites of this pathway. Notably, no typical intermediates of the aerobic phenol degradation pathway were detected. One-chamber reactors (+0.4 V vs. SHE) exhibited a phenol removal rate of 3.5 ± 0.2 mg L −1 d −1 , while two-chamber reactors showed 3.6 ± 0.1 and 2.6 ± 0.9 mg L −1 d −1 at anode potentials of +0.4 and + 0.2 V, respectively. Our results also suggest that the reactor configuration certainly influenced the microbial community, presumably leading to different ratios of phenol consumers and microorganisms feeding on degradation products.",
"conclusion": "4 Conclusion In contrast to inconclusive literature, this work provides evidence for strict anaerobic phenol degradation using anodes as sole terminal electron acceptors. Our results demonstrated the presence of typical intermediates of anaerobic degradation, such as 4-hydroxybenzoic acid and pimelic acid, whereas specific metabolites for aerobic phenol degradation, including catechol and muconic acid, were absent. The removal rate was comparable in all tested reactor configurations, with the highest observed in one-chamber reactors at 0.4 V, achieving a phenol removal rate of 3.6 ± 0.1 mg L −1 d −1 . Syntrophorhabdus , a genus containing a syntrophic phenol degrader, was identified in all experiments, and the inoculum with a stable relative abundance of 7.1 ± 2.0%, suggesting a key role for anaerobic phenol degradation, also in bioelectrochemical systems. Subsequently, the consumption of intermediates, such as acetate by Geobacter (relative abundance of 26.6 ± 4.9% and 15.7 ± 1.1% in biofilm and planktonic phase, respectively) and Arcobacter (relative abundance of 9.7 ± 1.7% and 28.2 ± 4.8% in biofilm and planktonic phase, respectively) in two-chamber reactors, resulted in current production. In one-chamber reactors, the lower relative abundance of known electroactive microorganisms is accompanied by a strong emergence of genera involved in sulfur cycling, with a relative abundance of 52.7 ± 1.2%, suggesting their contribution to phenol oxidation or parallel metabolic processes. However, further investigations employing more detailed experiments with pure cultures, defined co-cultures, and multi-omics approaches are necessary to reveal a potential mutualistic metabolism and microbial interactions associated with anaerobic phenol degradation.",
"introduction": "1 Introduction Phenol is a common pollutant discharged by many industries, such as chemical, petrochemical, pharmaceutical, and pesticide sectors [ 1 ]. Given its harmful effects on human health, including toxicity, mutagenicity, and carcinogenicity, and its ability to extensively diffuse and persist in the environment, effective treatment methods for phenol are crucial in order to mitigate its associated adverse impacts [ 1 , 2 ]. Various methods are employed for the removal from wastewater including abiotic, such as adsorption and advanced oxidation, and biological methods, including aerobic and anaerobic microbial phenol degradation [ 3 , 4 ]. Among these methods, microbial degradation is often favored due to its cost-efficiency and environmental-friendly properties [ 4 ]. Moreover, because of the widespread occurrence of phenol in the environment, phenol-degrading microorganisms are ubiquitous in numerous habitats [ 5 ]. However, one apparent bottleneck of microbial phenol degradation is the need for a sufficient supply of electron acceptors, such as oxygen under aerobic conditions or nitrate, sulfate, Fe(III), and Mn(IV) under anaerobic conditions [ 5 , 6 ]. Furthermore, phenol could also be degraded under fermenting conditions, resulting in methane production [ 7 ]. One opportunity to overcome this limitation is the provision of electrodes as inexhaustible sources of electron donors and acceptors within the scope of microbial electrochemical technology (MET), an emerging platform for pollutant removal [ 8 , 9 ]. The corresponding reactors are termed bioelectrochemical systems (BES) that harness electroactive microorganisms (EAM), being capable of extracellular electron transfer (EET) and coupling their metabolism with the electric current flow at electrodes [ 10 ]. Multiple studies have reported phenol degradation in BESs under anaerobic [ [11] , [12] , [13] ] and micro-aerobic conditions [ 14 ]. Previous studies on microbial electrochemical phenol degradation analyzed different inoculum sources [ 15 ], phenol-degrading microbial communities [ 16 ], various electrode materials [ 17 , 18 ], and the influence of the applied anode potential [ 14 ]. Phenol degradation utilizing an anode as the sole terminal electron acceptor (TEA) is anticipated to follow the common anaerobic phenol degradation pathway. Here, phenol is initially carboxylated, forming 4-hydroxybenzoate as a stable intermediate. This intermediate is then further processed through the reduction and opening of the aromatic ring by the benzoyl-CoA reduction pathway [ 19 ] ( Fig. 1 b). However, corresponding literature concerning phenol degradation in BESs is inconsistent. For instance, muconic acid and catechol, which are typical intermediates for aerobic phenol degradation [ 5 ] ( Fig. 1 a), were reported during assumed anaerobic phenol degradation in BESs [ 20 , 21 ]. Hassan and colleagues reported that chlorophenol and phenol were transformed into benzene in an anaerobic BES [ 22 ], which has never been observed before and is incompatible with the known anaerobic phenol degradation pathway [ 19 ]. Two studies presented genetic evidence for an anaerobic phenol degradation pathway, but one study suggested a combined anaerobic and aerobic phenol degradation as genes for both pathways were identified [ 23 ]. Espinoza-Tofalos and colleagues reported a decreased coverage of the marker gene for the anaerobic phenol degradation (encoding for 4-hydroxybenzoate decarboxylase) for the cultivated microbiome compared to the inoculum [ 24 ], questioning the main pathway being responsible for phenol degradation. In another study, 4-hydroxybenzoic acid, a key intermediate for the anaerobic phenol oxidation, was identified, but significant amounts of nitrate, sulfate, Fe(III), and oxygen — all representing alternative TEA — were present in the experiments [ 25 ]. Therefore, it is still doubtful and lacks conclusive evidence if the anaerobic phenol degradation in BESs by mixed cultures and using anodes as sole TEAs is feasible and follows the well-known anaerobic phenol degradation pathway. Fig. 1 Simplified anaerobic ( a ) and aerobic ( b ) phenol degradation pathways. During anaerobic phenol degradation, phenol is first converted by phenylphosphate synthase and phenylphosphate carboxylase to 4-hydroxybenzoic acid, which is subsequently activated to benzoyl-CoA [ 19 ]. The aromatic ring is opened by a hydrolase in the benzoyl-CoA degradation pathway resulting in 3-hydroxypimelyl-CoA being further transformed to acetyl-CoA [ 22 ]. Benzoic acid is an artificial decay product of benzoyl-CoA which is commonly detected in corresponding measurements. The dashed line arrows represent multi-step reactions. During aerobic phenol degradation, phenol is first activated by monooxygenases resulting in catechol, followed by cleavage of the benzene ring by dioxygenases either between the two hydroxyl groups (ortho-cleavage) or next to one of the hydroxyl groups (meta-cleavage) [ 26 ]. The products are further oxidized into acetyl-CoA utilized in the citric acid cycle [ 27 ]. Fig. 1 This study aims to address this knowledge gap by investigating the feasibility of a strict anaerobic phenol degradation process using anodes as sole TEAs. In addition, we evaluated the influences of reactor configuration (one-chamber and two-chamber BESs) and applied anode potential on phenol degradation. Applying high-performance liquid chromatography (HPLC), gas chromatography-mass spectrometry (GC-MS), and amplicon sequencing showed strong evidence that phenol was solely anaerobically degraded. To our knowledge, this is the first report about anaerobic phenol degradation in BESs using anodes as sole TEAs.",
"discussion": "3 Results and discussion 3.1 Influence of reactor configuration and applied anode potential on phenol degradation Phenol ( c p h e n o l , 0 = 90.6 ± 1.9 mg L −1 ) was degraded entirely in all bioelectrochemical systems (BESs) within 44 days ( Fig. 2 ). Only 2.7 ± 0.8% of phenol was removed during the first seven days, indicating an adaption phase after inoculation ( Fig. 2 ). The phenol removal rate ( r p h e n o l ) increased after seven days in all experiments with an average r p h e n o l of 3.6 ± 0.1, 3.5 ± 0.2, and 2.6 ± 0.9 mg L −1 d −1 in TC4, OC4, and TC2, respectively, without significant differences ( Table 2 ). Phenol was fully removed within 28 days in TC4 and OC4. In TC2, complete depletion was achieved after 44 days, presumably due to the lower applied anode potential. Moreover, the replicates of TC2 showed a heterogeneous behavior indicated by the high 95% confidence intervals, which were not observed for TC4 and OC4 ( Fig. 2 ). Abiotic and open circuit voltage control reactors showed no phenol removal demonstrating a clear relation between anaerobic phenol degradation and current production (SI-5). The observed r p h e n o l was considerably lower than previously reported values [ [29] , [30] , [31] ]. However, the present study did not aim at high performance but to study the feasibility and mechanisms of anaerobic phenol degradation using anodes as sole TEAs. For instance, an r p h e n o l of 12.2–37.0 mg L −1 d −1 were reported for a BES poised to 0.2 V using an anode area-to-reactor volume ratio (determining the availability of the electron acceptor) of 0.4 cm 2 mL −1 [ 31 ], which is eight times higher than in this study (0.05 cm 2 mL −1 ). Fig. 2 Time course of phenol concentrations ( c p h e n o l ) during chronoamperometric cultivation. The black, red, and blue lines represent TC4 (two-chamber reactors, anode potential 0.4 V), OC4 (one-chamber reactors, 0.4 V), and TC2 (two-chamber reactors, 0.2 V), respectively. The data are mean values, and error bars represent 95% confidence interval ( n = 3). Fig. 2 Table 2 Overview of performance parameters of BESs. Table 2 Reactor configuration Anode potential vs. SHE (V) Abbreviation Average current density (mA cm −2 ) Coulombic efficiency a (%) Phenol degradation rate b (mg L −1 d −1 ) pH Two-chamber 0.4 TC4 0.005 ± 0.001 52.9 ± 9.1 3.55 ± 0.08 6.8 ± 0.2 One-chamber 0.4 OC4 0.019 ± 0.005 180.7 ± 38.8 3.46 ± 0.20 7.3 ± 0.1 Two-chamber 0.2 TC2 0.006 ± 0.003 66.4 ± 27.8 2.59 ± 0.93 6.8 ± 0.2 a Independent two-sample t -tests indicated significant differences between TC4/OC4 ( p = 0.01) and OC4/TC2 ( p = 0.02) but not between TC4 and TC2 ( p = 0.50). b Independent two-sample t -tests indicated no significant differences between TC4/OC4 ( p = 0.72), OC4/TC2 ( p = 0.14), and TC4/TC2 ( p = 0.06). Although TC4 and OC4 exhibited no significant differences in r p h e n o l , their microbial composition differed noticeably (Section 3.3 ). The achieved average current density and C E were considerably higher in OC4 ( Table 2 , Fig. S3 ) than in TC4 and TC2, which can be assigned to the microbial utilization of cathodically produced H 2 and thus recurring recycling of electrons as it was demonstrated before [ 32 , 33 ]. Intriguingly, the C E of two-chamber BES is high ( Table 2 ) when compared to the literature ( C E of 4–23% [ 21 ]). Considering the share of electrons required for biomass formation and maintenance, the high C E illustrates that phenol was most likely fully mineralized and indicates a negligible role of other terminal electron acceptors (including oxygen) than the anode. 3.2 Identification of phenol degradation intermediates in bioelectrochemical systems During reactor operation, minor peaks of benzoic acid and 4-hydroxybenzoic acid were observed by HPLC in all BESs ( Fig. S4 ), suggesting no differences in the phenol degradation pathway between one-chamber and two-chamber reactors as well as between the different applied anode potentials. This also indicated that phenol was anaerobically degraded in the BES, as these compounds are typical metabolites of the anaerobic phenol degradation pathway [ 34 ]. To verify the identity and occurrence of these two intermediates, an additional one-chamber BES was inoculated using the same inoculum. After being repeatedly fed with phenol, this long-term BES showed an accumulation of the intermediates mentioned above ( Fig. S5 ). Furthermore, GC-MS analysis ( Fig. S6 ) revealed the presence of glutaric acid, ( E )-pent-2-enedioic acid, ( E )-but-2-enoic acid, and 3-hydroxybutanoic acid, which are typical downstream metabolites of the anaerobic phenol degradation pathway [ 19 , 26 , 35 , 36 ]. Fig. 3 highlights the metabolites annotated using HPLC and GC-MS measurements of the long-term BES. 10.13039/100014337 Furthermore , the absence of metabolites being typical for the aerobic phenol degradation pathway strongly supports the hypothesis that microorganisms solely degraded phenol via the anaerobic degradation pathway using anodes as sole TEAs. Although the literature about characteristics of phenol degradation at controlled microaerobic conditions is unavailable to our knowledge, for degradation of other monoaromatics at microaerobic conditions, it has been observed that metabolites of the initial oxygen-dependent mono- or dioxygenation reactions at the aromatic ring are usually accumulating in the external medium, e.g., (chloro)catechols [ [37] , [38] , [39] ]. Fig. 3 The proposed route of the anaerobic phenol degradation in BESs using anodes as sole terminal electron acceptor based on compounds annotation by HPLC and GC-MS analysis [ 22 ]. Benzoic acid is an artificial decay product of benzoyl-CoA, which is commonly detected in corresponding measurements, as in the here performed HPLC analysis. The dashed arrows indicate potential one- and multi-step reactions. Fig. 3 3.3 Microbial community composition At the end of the experiments, anode biomass and planktonic cells were analyzed via 16S rRNA gene amplicon sequencing to reveal structure-function relationships within the bacterial community. Analyzing the genetic data with principal coordinates analysis (PCoA) indicated that each experimental condition led to different microbial communities. One-chamber and two-chamber reactors differed more than two-chamber reactors poised at different anode potentials ( Fig. S7 ). Whereas different anode potentials influence the EET rate and, presumably, the abundance of weak and strong electricigens (see discussion below), the reactor configuration determines the availability of cathodically produced H 2 . The presence of this additional substrate certainly influenced the microbial composition, as it was already demonstrated for acetate-fed electroactive biofilm anodes [ 40 , 41 ]. 3.3.1 Microorganisms directly involved in anaerobic phenol degradation using anodes as sole terminal electron acceptors Proteobacteria dominated all BESs with a relative abundance of 63.1 ± 11.2% (combined value of anode biomass and planktonic biomass), representing a considerable enrichment compared to the inoculum (39.3%) ( Fig. 4 a). The relative abundance of Epsilonbacteraeota significantly increased in both two-chamber reactors, especially in the planktonic phase (28.4 ± 4.7%), while they were nearly absent in one-chamber reactors. Notably, Firmicutes (i.e., Bacillus ) were less abundant in BESs than in the inoculum ( Fig. 4 a and b). Overall, the results confirm previous studies, showing that Proteobacteria are the dominant phylum in BES treating phenol [ 13 , 42 ] and phenolic compounds [ 23 ]. Fig. 4 Community composition of microbial samples derived from the inoculum (SI-3), biofilm, and planktonic cells at the phylum ( a ) and genus ( b ) level. R1, R2, and R3 are independent replicates of each experimental condition. Fig. 4 On the genus level, the relative abundances of phylotypes belonging to Geobacter , being one model genus for direct EET, were significantly higher (see SI-6 for an overview of statistical tests on microbial community composition) in two-chamber reactors (biofilm: 26.6 ± 4.9%, planktonic cells: 15.7 ± 1.1%) than one-chamber reactors (biofilm: 12.1 ± 0.6%, planktonic cells: 7.1 ± 0.7%) ( Fig. 4 b). In one-chamber reactors, the presence of hydrogen as an additional electron donor, alongside acetate, led to a higher microbial diversity, as hydrogen utilization is a widespread ability of anaerobes [ 43 , 44 ]. Geobacter species can directly interact with electrodes leading to current production [ 45 ]. The involvement of two different Geobacter phylotypes in the anaerobic phenol degradation is hypothesized to occur through two distinct mechanisms. First, complete phenol oxidation is performed by Geobacter metallireducens, which anaerobically degrades phenol by performing EET [ 34 ]. Second, Geobacter phylotypes performed EET based on oxidizing acetate and other small carbon molecules that originated from fermentative anaerobic phenol degradation performed by, for instance, Syntrophorhabdus . This genus hosts syntrophic phenol degraders producing, e.g., acetate and butyrate [ 46 , 47 ], and its occurrence in phenol-degrading BESs was already reported [ 48 ]. However, its relative abundance was comparably low (5.6 ± 2.1% in all experiments), with a slightly higher abundance in planktonic biomass (7.1 ± 2.0%). Syntrophorhabdus aromaticivorans were reported to exhibit slow growth under syntrophic conditions with phenol as the substrate, likely due to the ATP-consuming activation of phenol by phenylphosphate synthase [ 7 ]. Its appearance in the inoculum (6.3%), supplemented with phenol but with different TEA (i.e., sulfate), suggests the particular role of phylotypes affiliated to Syntrophorhabdus in anaerobic phenol degradation, providing metabolites to further members of the microbial metabolic network (i.e., mutualistic metabolism). Furthermore, a phylotype related to Arcobacter (i.e., Epsilonbacteraeota), a genus known for containing members capable of manganese oxide reduction coupled to acetate oxidation [ 49 ] and previously identified at the anode of BESs [ 50 ], could contribute to the degradation of intermediates and current production. It was enriched in the two-chamber BES (biofilm: 9.7 ± 1.7%, planktonic cells: 28.2 ± 4.8%) but not in the one-chamber BES. Furthermore, phylotypes affiliated with the genus Pseudomonas could contribute to phenol degradation. They were present in two-chamber reactors (8.3 ± 4.0%) but exhibited a negligible relative abundance in one-chamber reactors (0.3 ± 0.0%). Pseudomonas- affiliated phylotypes were described as phenol degraders in the anaerobic BES [ 15 , 51 ] and for aerobic conditions [ 52 ]. The closely related genus Thauera was also reported for anaerobic phenol degradation [ 19 ] and was slightly increased in TC2 (2.8 ± 1.4%). Notably, Pseudomonas and Thauera -affiliated phylotypes from activated sludge were shown to assimilate phenol under anaerobic conditions [ 53 ]. However, it remains unclear if Pseudomonas oxidized phenol or only consumed secreted small carbon molecules derived from phenol oxidation by other microbial community members and then performed mediated EET [ 54 ]. Pseudomonas is also considered a weak electricigen, exhibiting low currents typically during redox stress and having limited growth while performing EET [ 55 ]. In our study, the weak electricigenity of Pseudomonas is indicated by the observation that its relative abundance was significantly higher (e.g., biofilm: 13.3 ± 0.3% in TC2 and 3.6 ± 0.1% in TC4, p = 1 × 10 −6 ) at a lower anode potential (i.e., less driving force for EET). At the same time, the relative abundance of Geobacter , a strong electricigen, was significantly lower in TC2 compared to TC4 (e.g., biofilm: 31.4 ± 0.7% in TC4 and 21.7 ± 0.4% in TC2, p = 4 × 10 −5 ). At lower anode potentials, weak electricigens could occupy an ecological niche alongside the strong electricigens from the genus Geobacter . But at higher anode potentials, Geobacter outcompeted weak electricigens due to their superior EET capabilities, high affinity, and high uptake rate for acetate, one central intermediate of the anaerobic phenol degradation [ 26 , 33 , 56 ]. 3.3.2 Microorganisms involved in non-bioelectrochemical processes One-chamber reactors exhibited a higher relative abundance of genera involved in sulfur cycling, including Desulfovibrio , uncultured Desulfobulbaceae , Desulfuromonas , Desulfurivibrio , Desulfoprunum , and Desulfomicrobium (52.7 ± 1.2% averaging biofilm and planktonic cells) compared to two-chamber reactors (12.4 ± 4.1%, p = 1 × 10 −13 ). For instance, the sulfate-reducing family Desulfobulbaceae , comprising physiologically versatile members who, e.g., can degrade aromatics like toluene [ 57 ], was detected in a high relative abundance in one-chamber reactors (biofilm: 21.5 ± 1.1%, planktonic cells: 12.4 ± 0.3%). Their presence indicates a contribution to phenol degradation and other processes not directly linked to phenol oxidation. Most of these genera contain members described to utilize H 2 , produced at the cathode, as the electron donor for sulfate reduction [ 58 ], which was present in small amounts (1.4 mM) in the medium. Considering this low sulfate concentration, it seems conceivable that the produced sulfide was abiotically oxidized at the anode, as demonstrated previously [ 59 ]. Consequently, sulfide oxidation and sulfate recycling could contribute to the high CE in one-chamber reactors and a continual enrichment of genera involved in sulfur cycling, respectively. An additional trophic layer could be constituted by Spirochaetaceae, which exhibited a relative abundance of 7.6 ± 2.6% and 7.8 ± 1.4% in one-chamber and two-chamber reactors, respectively. Spirochaetaceae , frequently identified in anaerobic hydrocarbon-contaminated environments, were supposed to recycle necromass secreting small electron donors and other nutrients to the microbial metabolic network [ 60 ]. The absence of any gas overpressure in all reactors, the emergence of Geobacter and Arcobacter (in two-chamber reactors), and genera involved in sulfur cycling (in one-chamber reactors) indicate a minor role of methanogens. Presumably, they are outcompeted by the genera mentioned above, as using an electrode and sulfate as the TEA is energetically more favorable than CO 2 ."
} | 5,802 |
31998352 | PMC6965313 | pmc | 9,335 | {
"abstract": "Plants manipulate their rhizosphere community in a species and even a plant life stage-dependent manner. In essence plants select, promote and (de)activate directly the local bacterial and fungal community, and indirectly representatives of the next trophic level, protists and nematodes. By doing so, plants enlarge the pool of bioavailable nutrients and maximize local disease suppressiveness within the boundaries set by the nature of the local microbial community. MiSeq sequencing of specific variable regions of the 16S or 18S ribosomal DNA (rDNA) is widely used to map microbial shifts. As current RNA extraction procedures are time-consuming and expensive, the rRNA-based characterization of the active microbial community is taken along less frequently. Recently, we developed a relatively fast and affordable protocol for the simultaneous extraction of rDNA and rRNA from soil. Here, we investigated the long-term impact of three type of soil management, two conventional and an organic regime, on soil biota in fields naturally infested with the Columbian root-knot nematode Meloidogyne chitwoodi with pea ( Pisum sativum ) as the main crop. For all soil samples, large differences were observed between resident (rDNA) and active (rRNA) microbial communities. Among the four organismal group under investigation, the bacterial community was most affected by the main crop, and unweighted and weighted UniFrac analyses (explaining respectively 16.4% and 51.3% of the observed variation) pointed at a quantitative rather than a qualitative shift. LEfSe analyses were employed for each of the four organismal groups to taxonomically pinpoint the effects of soil management. Concentrating on the bacterial community in the pea rhizosphere, organic soil management resulted in a remarkable activation of members of the Burkholderiaceae, Enterobacteriaceae, and Pseudomonadaceae. Prolonged organic soil management was also accompanied by significantly higher densities of bacterivorous nematodes, whereas levels of M. chitwoodi had dropped drastically. Though present and active in the fields under investigation Orbiliaceae, a family harboring numerous nematophagous fungi, was not associated with the M. chitwoodi decline. A closer look revealed that a local accumulation and activation of Pseudomonas, a genus that includes a number of nematode-suppressive species, paralleled the lower M. chitwoodi densities. This study underlines the relevance of taking along both resident and active fractions of multiple organismal groups while mapping the impact of e.g. crops and soil management regimes.",
"conclusion": "Concluding Remarks The development of a time-efficient and affordable protocol to extract total DNA and RNA from soil ( Harkes et al., 2019 ) allowed us to monitor the effect of a legume, pea, on both resident and active communities of primary decomposers as well as primary consumers of bacterial and fungal assemblages. Pea was shown to exert a large effect on the rhizobiome, and this was not only true for the primary decomposers but also for the protist and metazoan community. For all four organismal groups, and irrespective of the algorithm used to assess community shifts, the variables “Nucleic Acid” and “Sample Type”—representing respectively the differences between the resident and the active communities, and the effect of pea on the rhizobiome—had the highest impact on the soil microbiome. Notwithstanding this conclusion, soil management (“Treatment”) had also a significant effect on both the primary decomposers and the two primary consumer groups. A number of taxonomic groups (mostly at family level) were identified as contributors to these contrasts. In some cases, these taxa could be linked to treatment or crop identity, but in other cases such families were highly speciose or barely characterized from a soil ecological point of view. In essence, this was also true regarding our efforts to find possible biological explanations for the remarkably low levels of the RKN M. chitwoodi under the prolonged organic management regime. Our data suggest that Pseudomonadaceae—here members of the genus Pseudomonas —could have played a role in the biological suppression of this notorious RKN species. It should be underlined that biological associations have been identified in this research, and it was by no means proven that one or more Pseudomonas species were actually responsible for the observed decline in RKN levels in fields under organic soil management. In this study a broad approach was used to characterize shifts in the soil microbial community under various soil management regimes with a legume—pea—as main crop. In our analyses we mainly focused on the active fractions, and this allowed us to pinpoint target taxa associated with the various treatments for each of the four organismal groups. One of the main remaining hurdles for the fundamental understanding of shifts in soil microbial communities is our fragmented and often poor knowledge about the ecologies of soil inhabitants.",
"introduction": "Introduction For decades, conventional soil management has resulted in consistent and high level of crop production by external inputs such as chemical fertilizers and pesticides. However, it is widely acknowledged that intensive monocropping has a number of downsides including soil degradation, leaching of nutrients, and biodiversity loss ( Tsiafouli et al., 2014 ). Organic farming, an umbrella term for a wide range of management regimes having the abstinence of the use of mineral fertilizers and chemical pesticides in common, is a possible alternative that might alleviate the negative impact of crop production on soil ecosystems. In organic farming, most often organic manure is used to replenish the nutrient levels in the top soil and to maintain or increase the overall soil organic matter content. In addition, grain legumes are frequently part of the crop rotation because of their nitrogen binding capability. However, especially in Europe a wider application of grain legumes is currently hampered, by the relatively high level of variability in yield. This variation is thought to be due to the sensitivity of these crops to biotic and abiotic stressors ( Cernay et al., 2015 ). One of the key characteristics of sustainable soil management regimes should be the preservation of a relatively high level of soil biodiversity. In terms of biomass, bacteria and fungi are the most important biotic constituents of soils. Depending on soil type, cultivated soils typically harbor 0.2–0.7 mg of bacteria per g of dry soil, whereas the fungal community is represented by 0.01–0.2 mg per g ( Kaczmarek, 1984 ). Protists and nematodes are major consumers of bacteria and fungi in soil ecosystems. Although the biomass of protists and nematodes is small compared to the primary decomposers ( Bar-On et al., 2018 ), their impact on the turnover of bacteria and fungi is enormous. Protists alone are typically consuming >50% of the bacterial productivity ( Foissner, 1999 ). Though it is a simplification of the biological reality, one could argue that the bacterial and fungal communities are shaped by (1) the quantity and nature of external C and energy inputs into the soil ecosystem, and (2) the activity of protist and nematode communities. Being present in the soil ecosystem does not imply that a given organism is actively participating in the soil food web. On the contrary, many soil inhabitants are able to reduce their metabolic activity when unfavorable conditions occur, such as food scarcity or drought. This is especially relevant for bulk soils, where typically 80% of the cells, and 50% of the Operational Taxonomic Units (OTUs) are inactive ( Lennon and Jones, 2011 ). Hence, it is essential to take both the resident and the active fractions into account when assessing the biological functioning of a soil ecosystem. A range of studies underlined the relevance of the distinction between resident and active soil biota ( Baldrian et al., 2012 ; Nunes et al., 2018 ; Schostag et al., 2019 ). For taxonomic profiling, 16S or 18S ribosomal DNA (rDNA) is often used as molecular marker. Ribosomal RNA is frequently used to map the active microbial fractions ( Ofek et al., 2014 ; De Vrieze et al., 2016 ). By the molecular characterization of both the resident and the active fractions of the bacterial, fungal, protist, and metazoan community, it is possible to assess the impact of soil management regimes on the soil food web ( Harkes et al., 2019 ). The rhizosphere of plants creates a center of high metabolic activity in soils. At the interface between the plant root and soil, the plant releases primary and secondary metabolites ( Hinsinger et al., 2009 ; Reinhold-Hurek et al., 2015 ). With this blend of plant-derived components, the plant boosts a specific fraction of the soil biota. In return, stimulated microbiota increase the bioavailability of plant nutrients and/or they may contribute to the protection of the plants against pathogens ( Lugtenberg and Kamilova, 2009 ; Berendsen et al., 2012 ; Turner et al., 2013 ). Especially in agricultural soil, the microbial community structure was shown to be distinct from the surrounding bulk soil ( Sharma et al., 2005 ). Due to the application of fertilizers, root exudation is enhanced which on its turn affects the microbial community in the rhizosphere ( Zhu et al., 2016 ). Next to bacterivores and fungivores, the nematode community harbors a wide range of plant parasites. Most of them are relatively harmless root hair feeders and ectoparasites ( Quist et al., 2015 ). Only a small subset may have a high impact in crop production. Root-knot nematodes (RKN), members of the genus Meloidogyne , are number one in terms of global crop damage by plant-parasitic nematodes ( Jones et al., 2013 ). The highly polyphagous Columbian RKN Meloidogyne chitwoodi has a global distribution in temperate climate zones ( Santo, 1989 ). In this study we investigated the long-term effects of three soil management regimes, conventional, integrated and organic, on the soil microbiome in fields naturally infested with M. chitwoodi. The legume Pisum sativum was used as main crop in these fields. Illumina MiSeq sequencing was used to characterize the active (rRNA) as well as the resident (rDNA) communities of bacteria, fungi, protozoans and metazoans both in bulk soil and in the rhizosphere. The main objectives of this study were (i) to characterize the resident and active microbial community in the rhizosphere of pea with the underlying hypothesis that—besides being present—microbiota need to be active in order to be able to contribute to local changes in food web functioning, (ii) to map the effects of pea on the active and resident fractions of the four organismal group in the rhizosphere compared to the bulk soil under different soil management systems, and (iii) to identify microbial taxa which activities changed in parallel with distinct infestation levels of the root-knot nematode M. chitwoodi",
"discussion": "Discussion Mapping of resident and active fractions of the primary decomposer community—bacteria and fungi—as well two major primary consumer groups—protists and nematodes—under three distinct soil management regimes revealed that pea exerts a large effect on the soil microbiome. Below we will discuss (1) how the current characterisation of the pea rhizobiome relates to other studies, (2) how our observations regarding the effect of soil management relate to previous findings, and (3) whether we can find plausible biological explanations for the observed sharp decline in RKN densities in fields under prolonged organic soil management. How Does the Current Characterisation of the Pea Rhizobiome Relate to Previous Studies? As exemplified by the impact of pea, lentil and chickpea, legumes have been shown to exert a large influence on the soil microbiome as compared to other crops such as cereals ( Turner et al., 2013 ; Hamel et al., 2018 ). N rhizodeposition has been shown to comprise 13% of the total plant N for pea ( Mayer et al., 2003 ), and presumably this has contributed to this large impact. The large overall effect of legumes could be corroborated by comparing the current study with a recent study on the barley rhizobiome that made use of the same experimental fields ( Harkes et al., 2019 ). In case of barley, the compartment effect (bulk vs rhizosphere) explained a smaller percentage of the observed variation than impact of the soil management regime. For pea, on the contrary, all four organismal groups indicated the compartment effect to be larger than the soil treatment effect. Hence, under similar experimental conditions the compartment effect on the soil microbiome induced by pea (a legume) is stronger than the effect induced by barley (a cereal). Recently, the effect of various frequencies pulse crop cultivation (including pea) on resident soil bacterial communities was mapped ( Hamel et al., 2018 ). Increased frequency of pulse cultivation resulted in higher abundances of α-Proteobacteria in the rhizosphere, and a decrease in γ-Proteobacteria (although the latter was accompanied by an increased presence of Pseudomonas ). Keeping in mind that the Rhizobiales (in our study Rhizobiaceae and Labraceae) belong to the subclass α-Proteobacteria, an overall increase of this subclass was to be anticipated. Moreover, Hamel et al. (2018) detected an increase Pseudomonas read in rotations involving pea. This might correspond to the increased activity of Pseudomonadaceae we observed in the pea rhizosphere ( \n Figure 2 \n , panel 1). We could not confirm the increased presence of Actinobacteria in the pea rhizosphere as observed in rotation systems with frequent inclusion of pulses (referred to as “3-pulse systems”). This phenomenon might only be observable after repeated cultivation of legumes. We conclude that a number of parallels can be discerned between studies on the effect of pea on the rhizobiome. It is noted that differences in experimental approach (focus on resident or active soil biota) and set up (soil type, soil management practices) complicates the identification of generic effects of legumes on the soil living community. How Does the Current Characterisation of the Effects of Soil Management on the Soil Microbiome Relate Other Studies? In a long-term (>10 years) greenhouse experiment the effect of organic, integrated and conventional farming systems on the soil rhizobiome was investigated ( Li et al., 2019 ). The authors identified a bacterial hub, a small number of highly interconnected taxa, consisting of Bacillus (Bacillaceae), Sporosarcina (Planococcaceae), Hyphomicrobium (Hyphomicrobiaceae), Gaiella (Gaiellaceae), as well as Pirellula and Blastopirellula (both Planctomycetaceae) that were significantly more abundant in soil from the organic management regime ( Li et al., 2019 ). Another hub comprising of the genera Rhizobium (Rhizobiaceae), Sphingobium (Sphingomonadaceae), Pseudoxanthomonas (Xanthomonadaceae), and Dyadobacter (Cytophagaceae) was present in higher densities in the conventional and the integrated treatments. These findings show very little resemblance with the bacterial taxa that were shown to be activated under the organic or one of the two conventional soil managements systems in the present study ( \n Figure 4 \n ). From this, we conclude that the plant effects can be stronger than the effect of soil management (variable “treatment” in \n Tables 1 \n and \n 2 \n ). Moreover, it is noted that the active bacterial community can be quite distinct from the resident bacterial communities mapped by Li et al. (2019) ( \n Figure 1 \n ). In another long-term field experiment (running for ≈ 15 years at time of sampling) fields were continuously exposed to either conventional or organic farming practices, and the impact of the practices on bulk soil were determined ( Bakker et al., 2018 ). These authors showed a remarkable contrast between bacterial phyla with regard to the extent by which they were affected by the contrasting farming practices. More taxa showed higher abundances in organic as compared to conventional farming. Moreover, some bacterial phyla such as Chloroflexi, Firmicutes, and Gemmatimonadetes seemed to be unaffected by farming practice while others such as Proteobacteria, Acidobacteria, and Verrucomicrobia were. Our data only partly support this observation. The families Burkholderiaceae and Hyphomicrobiaceae (Proteobacteria) and the Peptostreptococcaceae (Firmicutes) were both more abundant and more active in bulk soil in organic fields ( \n Supplementary Figure S3 \n ). Peptostreptococcaceae are one of the dominant family in the gut microbiome of earthworms (see e.g. \n Zeibich et al., 2018 ), and as such this result could point at an elevated presence of earthworms in fields under organic management. Our analysis of resident bacterial community in bulk soil under the organic regime, also showed an increase of members of the Acidobacterial family Blastocatellaceae ( \n Supplementary Figure S3 \n ). In the most conventional soil management system (ConMin), the Verrucobacterial family Pedosphaeraceae was both abundant and highly activated. At high taxonomic level, this is in line with the observations presented by Bakker et al. (2018) . Hence, despite the fact that plant identity may have a stronger effect on the rhizobiome than soil management practices, the effect of these practices could be pinpointed at taxon level. Our data suggest that the working hypothesis saying that only a subset of the soil bacterial phyla is affected by conventional or organic soil management practices might be correct. Can We Pinpoint Nematode-Suppressive Bacterial or Fungal Taxa That Might Underlie the M. chitwoodi Decline in Fields Under Organic Management? In this study we investigated the soil microbiome of pea in fields naturally infested with M. chitwoodi under three different soil management regimes, conventional, integrated, and organic. M. chitwoodi is a highly polyphagous plant parasite infecting numerous mono- and dicotyledonous crops, including pea ( P. sativum ) ( Oepp/Eppo, 1991 ), and it has a reputation as a major pest in potato. In all studied fields here, M. chitwoodi was already present for multiple years ( Visser et al., 2014 ). As potato—a highly suitable host—was the main crop in the previous growing season, we expected the M. chitwoodi population to be physiologically fit at the onset of the pea growing season. At the end of the growing season, M. chitwoodi densities in the two conventional soil management systems harbours ≈ 60 individuals per 100 g soil, whereas about two individuals per 100 g were detected in the organically managed system. We investigated whether a biological explanation could be found for this difference. Within the bacterial and fungal rhizosphere communities, families were detected that are known to harbor multiple nematode-trapping species. As shown in \n Figure 2 \n (panel 3), the Orbiliaceae were shown to be active in the pea rhizosphere. This family comprises genera such as Arthrobotrys , Dactylella , and Monacrosporium. These genera are essentially saprophytic fungi but can become predatory under e.g. low nutrient conditions ( Gray, 1987 ). As a response, the fungi will form traps ( e.g. constricting rings, adhesive networks) which allow them to prey on nematodes ( Xie et al., 2010 ). We verified whether Orbiliaceae activity was upregulated in the fields under organic management. This was not the case, and even a non-significant trend towards lower activity in organic fields was observed ( \n Supplementary Figure S2 \n ). Elevated activity of another fungal family, the Olpidiaceae, typified fields under organic soil management. A member of this family, Olpidium vermicola , has been reported to parasitize eggs and females of endoparasitic nematodes such as M. chitwoodi ( Esser and Schubert, 1983 ; Askary, 2015 ). However, other Olpidium species are virus-transmitting plant pathogens. Zoospores of Olpidium virulentus colonize roots of various plant species, and were demonstrated to accumulate in crop rotation with multiple pulses including pea ( Niu et al., 2018 ). The bacterial family Pseudomonadaceae was also identified as an indicator species for organic farming ( \n Figure 4 \n ). Further analyses identified the genus Pseudomonas as main contributor to the indicator status of Pseudomonadaceae. The Pseudomonas species P. aeruginosa, P. fluorescens, P. protegens, and P. chlororaphis belong to ecologically most relevant nematode-suppressive bacteria in soil ( Li et al., 2014 ). Pseudomonas species produce toxins which may inhibit hatching, survival and M. chitwoodi ’s ability to penetrate plant roots ( Thiyagarajan and Kuppusamy, 2014 ; Nandi et al., 2015 ; Kang et al., 2018 ). The increased densities of bacterivorous nematode families might form an indirect explanation for the decrease of M. chitwoodi in organic soil management systems ( \n Figure 5 \n ). As bacteria-grazing nematodes in the immediate vicinity of plant roots could locally improve nutrient availability via the excretion of easily uptakeable N and P. Plants could benefit from this in terms of improved growth and vitality, possibly making them less susceptible to plant-parasitic nematodes ( Thoden et al., 2011 ). Presumably multiple factors have contributed to significantly lower M. chitwoodi levels in the organic fields. This might have included nematode parasitic members of fungal genus Olpidium and/or the elevated activity of members of the Pseudomonadaceae. These results should be seen as potential leads for more detailed studies on the effect soil management regimes on the activity levels of nematode-suppressive bacteria and/or fungi. Concluding Remarks The development of a time-efficient and affordable protocol to extract total DNA and RNA from soil ( Harkes et al., 2019 ) allowed us to monitor the effect of a legume, pea, on both resident and active communities of primary decomposers as well as primary consumers of bacterial and fungal assemblages. Pea was shown to exert a large effect on the rhizobiome, and this was not only true for the primary decomposers but also for the protist and metazoan community. For all four organismal groups, and irrespective of the algorithm used to assess community shifts, the variables “Nucleic Acid” and “Sample Type”—representing respectively the differences between the resident and the active communities, and the effect of pea on the rhizobiome—had the highest impact on the soil microbiome. Notwithstanding this conclusion, soil management (“Treatment”) had also a significant effect on both the primary decomposers and the two primary consumer groups. A number of taxonomic groups (mostly at family level) were identified as contributors to these contrasts. In some cases, these taxa could be linked to treatment or crop identity, but in other cases such families were highly speciose or barely characterized from a soil ecological point of view. In essence, this was also true regarding our efforts to find possible biological explanations for the remarkably low levels of the RKN M. chitwoodi under the prolonged organic management regime. Our data suggest that Pseudomonadaceae—here members of the genus Pseudomonas —could have played a role in the biological suppression of this notorious RKN species. It should be underlined that biological associations have been identified in this research, and it was by no means proven that one or more Pseudomonas species were actually responsible for the observed decline in RKN levels in fields under organic soil management. In this study a broad approach was used to characterize shifts in the soil microbial community under various soil management regimes with a legume—pea—as main crop. In our analyses we mainly focused on the active fractions, and this allowed us to pinpoint target taxa associated with the various treatments for each of the four organismal groups. One of the main remaining hurdles for the fundamental understanding of shifts in soil microbial communities is our fragmented and often poor knowledge about the ecologies of soil inhabitants."
} | 6,081 |
33441960 | PMC7806881 | pmc | 9,336 | {
"abstract": "Connectivity is fundamentally important for shaping the resilience of complex human and natural networks when systems are disturbed. Ecosystem resilience is, in part, shaped by the spatial arrangement of habitats, the permeability and fluxes between them, the stabilising functions performed by organisms, their dispersal traits, and the interactions between functions and stressor types. Controlled investigations of the relationships between these phenomena under multiple stressors are sparse, possibly due to logistic and ethical difficulties associated with applying and controlling stressors at landscape scales. Here we show that grazing performance, a key ecosystem function, is linked to connectivity by manipulating the spatial configuration of habitats in microcosms impacted by multiple stressors. Greater connectivity enhanced ecosystem function and reduced variability in grazing performance in unperturbed systems. Improved functional performance was observed in better connected systems stressed by harvesting pressure and temperature rise, but this effect was notably reversed by the spread of disease. Connectivity has complex effects on ecological functions and resilience, and the nuances should be recognised more fully in ecosystem conservation.",
"conclusion": "Conclusions Key ecosystem functions in unstressed systems were positively linked to connectivity in our experimental microcosms. While stressors reduced function across all connectivity levels, the benefits of better connectivity were maintained under spatially explicit stressors (e.g. heat-stress, animal harvesting), but reversed in disease scenarios where morbid individuals interact more frequently in better-connected systems. Such contrasting effects of connectivity may present a challenge for conservation and management if trends are consistent with those in more complex natural systems. Our experimental technique and findings provide a platform for future studies to better understand how connectivity affects resilience in more complex systems and at larger spatial scales.",
"introduction": "Introduction Multiple stressors are affecting multiple ecosystems on Earth 1 – 6 . Ecosystem resilience describes the capacity for an ecosystem to maintain a stable state in the face of disturbance, either by resisting change or by rapidly recovering from disturbance effects, thus avoiding a regime shift to an alternative stable state 7 . How readily ecosystems resist, or recover, from the consequences of external pressures depends in part on the type of perturbations 8 , the responses of organisms within them, and the connectivity of habitat networks that surround them 7 , 9 . Connectivity also shapes how disturbance spreads through ecosystems and modifies the strength of ecological processes, affecting ecosystem resilience 10 , 11 . Animals perform many ecological functions that can stabilise or maintain ecosystem structure and function when disturbed, but disturbances themselves can also reduce the efficacy of these mitigating functions. In fragmented habitats, for example, plant production may be limited when pollination declines, but the movement of animal pollinators between patches can counteract this 12 . External stressors may impede travel between plants resulting in declines in pollination rates 13 . This affects ecosystem resilience by reducing the reproductive capacity of certain plants, potentially triggering changes in the species composition and functional capacity of ecosystems 13 . Similarly, fishing on coral reefs can remove herbivorous fish that control algal growth via grazing e.g. 5 , 14 , 15 . Phase-shifts can occur when herbivory rates are reduced because coral settlement and expansion are inhibited by algal overgrowth 16 , 17 . In spatial habitat mosaics, theory posits that the number and architecture of connections plays a role in constraining and shaping the way disturbance events modify ecological functions 18 , 19 . Yet, there is scant empirical data to test these predictions, especially how connectivity and disturbance types interact to control key ecological processes. We expect that highly connected networks will be more resilient to most individual stressors that affect only part of a system (e.g. harvesting near to protected areas). We refer to ‘part of a system’ in the respect that there be some capacity for unstressed areas to provide support to the stressed areas in resisting or recovering from disturbance. For animal-derived functions (e.g. pollination, grazing), the capacity for unaffected areas to provide this support will depend on the dispersal capacity and behaviour of the animals performing the function. For some animals, this may be many thousands of kilometres, while for others it may be metres. Thus, dispersal capacity should be considered when assessing connectivity and disturbance to ensure appropriate spatial scales are considered. When external stressors impede the performance of animals in parts of the system, better-connected habitat networks will be less affected because animals can be re-supplied more readily via unperturbed alternative pathways e.g. 20 – 22 . By contrast, stressors like disease, which rely on interactions between organisms to spread, may have the opposite effect in better-connected systems e.g. 23 . Despite the clear benefits for ecosystem management that might come from understanding these dynamics, the nature of the connectivity-resilience relationship remains poorly understood. Here we aim to assess the role that connectivity plays in shaping animal functional responses to single and multiple disturbance events of different types. Specifically, whether the type of stressor(s) determines how connectivity shapes the resilience of ecosystems."
} | 1,432 |
38561814 | PMC10983722 | pmc | 9,338 | {
"abstract": "Understanding the characteristics and structure of populations is fundamental to comprehending ecosystem processes and evolutionary adaptations. While the study of animal and plant populations has spanned a few centuries, microbial populations have been under scientific scrutiny for a considerably shorter period. In the ocean, analyzing the genetic composition of microbial populations and their adaptations to multiple niches can yield important insights into ecosystem function and the microbiome's response to global change. However, microbial populations have remained elusive to the scientific community due to the challenges associated with isolating microorganisms in the laboratory. Today, advancements in large-scale metagenomics and metatranscriptomics facilitate the investigation of populations from many uncultured microbial species directly from their habitats. The knowledge acquired thus far reveals substantial genetic diversity among various microbial species, showcasing distinct patterns of population differentiation and adaptations, and highlighting the significant role of selection in structuring populations. In the coming years, population genomics is expected to significantly increase our understanding of the architecture and functioning of the ocean microbiome, providing insights into its vulnerability or resilience in the face of ongoing global change. \n Video Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s40168-024-01778-0.",
"conclusion": "Conclusions Beginning in the 90 s with the onset of the “molecular revolution” and continuing into the 2000s with the advent of High-Throughput Sequencing technologies, omics approaches have significantly advanced our understanding of the ocean microbiome, revealing the various lineages it harbors, their distributions, and metabolisms. Specific markers, such as the rRNA gene, provided a clearer dimension of the diversity that is contained in the ocean microbiome. Yet, the rRNA gene normally underestimates or misses the dimension of diversity that is found within individual species (Fig. 4 ). So far, only a limited number of studies have delved into the population-level diversity of environmental microbes. Understanding the population diversity of microbes is fundamental for a better comprehension of ecosystem function and the adaptation of microbes to different niches. Isolating and culturing environmental strains has been one of the main obstacles in accessing the species-level diversity of microbes. Today, the use of metagenomics and metatranscriptomics allows us to investigate the diversity that is present within species, bypassing the need for culturing. Population-level studies have the potential to open a new chapter in environmental microbiology, deepening our understanding of the ocean microbiome's composition, configuration, ecological interactions, and intricate relationships with ecosystem functioning. This new knowledge will also be pivotal in the context of global change as we seek to comprehend the ocean microbiome’s resilience or vulnerability, as well as its potential impact on broader Earth system processes."
} | 793 |
33478162 | PMC7835748 | pmc | 9,340 | {
"abstract": "Flexible electronic devices have gained significant interest due to their different potential applications. Herein, we report highly flexible, stretchable, and sensitive sensors made of sprayed CNT layer, sandwiched between two polymer layers. A facile fabrication process was employed in which the CNT solution was directly sprayed onto a patterned bottom polymer layer, above which a second polymer layer was casted to get a sandwiched composite structure. Varying amounts of CNT solution (i.e., 10, 25, 40, 70, and 100 mL) were sprayed to get conductive CNT layers of different thicknesses/densities. The physical characteristics of the conductive CNT layers were studied through SEM and optical images. The starting electrical resistance values (without strain) as well as the changes in electrical resistance against human body motions were monitored. The synthesized samples exhibited good response against finger and wrist bending. The conductivity of the samples increased with increase of CNT solution volume while the sensitivity followed the inverse relation, suggesting that the sensors with controlled sensitivity could be fabricated for targeted strain ranges using the proposed method.",
"conclusion": "4. Conclusions In summary, the efficiency of the facile fabrication process was demonstrated. The conductive CNT layers were directly deposited onto the polymer. Using the proposed method, increasing amounts of CNTs could be easily spray-deposited, while avoiding issues like the non-uniform dispersion of CNTs in polymers, stiffening of the polymer itself, etc. The synthesized samples showed good sensing capabilities against different human motions. Furthermore, sensitivity of the samples was found to be dependent on the amount of CNT solution used. Hence, a specific amount of CNT solution could be used to achieve the desired sensitivity to target specific human motions.",
"introduction": "1. Introduction Wearable electronic devices are getting increasing attention due to their several applications [ 1 , 2 , 3 ]. For such devices, sensors should possess high stretchability and sensitivity. In this regard, conductive fillers are normally incorporated into soft polymers to attain desired functionality [ 4 , 5 , 6 ]. Among others, CNTs are regarded as favorable conductive fillers possessing excellent sensing capabilities [ 7 ]. Nevertheless, dispersion of CNTs in polymers using conventional solution dispersion or a melting process is still a cumbersome task. Recently, efforts were made to deposit conductive fillers onto polymer substrate using hot press [ 8 ], inkjet printing [ 9 ], or spray-deposition [ 10 , 11 ]. By doing so, a complete conductive layer/film could be obtained that generates signals upon stretching/releasing of the polymer. Mishra et al. [ 8 ] and Amjadi et al. [ 10 ] deposited a CNT film onto donor substrate which was then transferred to a recipient substrate. The samples showed good sensing response, however, it is difficult to ensure the integrity of the film as well as its complete transfer to the recipient substrate during the film transfer. Moreover, limited geometries could be created in the film transfer technique [ 9 ]. Lipomi et al. [ 11 ] spray-coated CNTs onto two PDMS layers and then joined them together using Ecoflex TM silicone polymer, which served as a dielectric layer. In this work, we report a facile fabrication process in which CNT solution was directly sprayed onto patterned polymer substrate, using a simple spray gun. Since quantity of CNTs is of vital importance in terms of electrical conductivity and sensitivity, varying amounts of CNT solution were sprayed to obtain conductive CNT layers with different thicknesses/densities. The deposited CNT layers were characterized by electrical resistance measurements and SEM/optical image analyses. Following this, the sensors were tested for detecting different human gestures."
} | 970 |
34430808 | PMC8365361 | pmc | 9,341 | {
"abstract": "Summary Soil-borne diseases cause serious economic losses in agriculture. Managing diseases with microbial preparations is an excellent approach to soil-borne disease prevention. However, microbial preparations often exhibit unstable effects, limiting their large-scale application. This review introduces and summarizes disease-suppressive soils, the relationship between carbon sources and the microbial community, and the application of human microbial preparation concepts to plant microbial preparations. We also propose an innovative synthetic microbial community assembly strategy with synergistic prebiotics to promote healthy plant growth and resistance to disease. In this review, a new approach is proposed to improve traditional microbial preparations; provide a better understanding of the relationships among carbon sources, beneficial microorganisms, and plants; and lay a theoretical foundation for developing new microbial preparations.",
"conclusion": "Concluding remarks The soil microbiome plays a crucial role in plant protection. Through the development of synthetic microbial communities, a simple microbial community can be used to efficiently promote ecological functions. However, the mechanism of the nutrient-plant-microbial community interaction network and the effects of nutrition on regulating the microbial community remain unclear. Thus, an understanding of the regulation of the soil microbiome by soil microbial carbon sources will help us build better, more durable synthetic microbial communities. At present, microbial preparations must overcome many issues, such as unstable effects and poor rhizosphere competitiveness. Synthetic microbial communities represent a feasible approach to solve these problems. Although the mechanisms of disease suppression in soils are complex and diverse, studies have shown that the ability of disease-suppressive soils to inhibit disease is related to the enrichment of specific beneficial microbial communities. Microbial separation and culture techniques have been used to design a synthetic microbial community that inhibits the occurrence of soil-borne diseases. Simultaneously, we can also directionally shape the microbiome to serve agriculture through nutritional restriction or microbiome genetic modification. This synthetic microbial community assembly process not only provides useful insights into the development of composite microbial preparations but also helps us better understand the interactions among microbial communities, the mechanisms of microbial community and plant interactions, and soil microbial regulation by organic matter and root exudates in microbial communities. Synthetic microbial communities can help us understand the mechanisms of action of disease-suppressive soils. The composition of carbon sources in the soil plays an important role in regulating the microbiome in the soil. The stability of synthetic microbial communities can be improved by adding specific carbon sources, enabling to clarify the causal relationship between root microbiota and plant phenotypes and analyze the interactions between microbiota members under natural soil conditions. Inspired by synbiotics and prebiotics, we propose that the Biolog-ECO plate method, high-throughput sequencing, and microbial isolation and culture technology will be useful to synthesize a more stable microbial community or directionally manipulate the microbial community based on the mode of action of the microbiome in the soil. Figure 2 roughly describes our idea of the process for assembling this synthetic microbial community. This idea will allow us to better understand the mechanism of action between the microbiome and the carbon source. Meanwhile, sequencing technology, synthetic community analysis and modeling, and functional joint analysis have important reference value for future research on rhizosphere interactions. The mechanism of plant-beneficial microorganism-harmful microorganism interactions is mostly unknown. More knowledge of the nutritional needs of the members of the microbiome, including pathogens and their social networks, may achieve the development of reasonable interventions and may produce new tailor-made disease prevention strategies. Figure 2 Process of synthetic microbial community assembly and community regulation by carbon sources The letter (A) represents the use of high-throughput sequencing and bioinformatics analyses such as multilevel species linear discriminant analysis effect size (Lefse) analysis and cooccurrence network analysis to analyze the microbial communities corresponding to each particular function in the susceptible soil and the disease-suppressive soil; (B) represents the use of microorganism separation and culture techniques, in which microorganism strains are separated and cultivated and functional strains are screened using screening tests such as antagonism tests, etc.; (C) represents the use of the results of a microbiome analysis to select and test synthetic microbial community members; (D) represents the use of differential Biolog-ECO plate carbon sources in disease-sensitive and disease-suppressive soils and analyzing the results; (E) represents the addition of synthetic microbial communities and specific carbon sources to the susceptible soil to test whether the ability of the susceptible soil to suppress soil-borne diseases is further enhanced; and (F) represents that synthetic microbial communities are abstractions of natural systems that allow the detailed study and analysis of the fundamental building blocks and processes that compose a microbial community.",
"introduction": "Introduction The interactions of pathogens, hosts, and the environment determine the occurrence of plant diseases. The environment, especially soil characteristics, determines the source of microorganisms recruited by plant roots, which mainly affects disease outbreaks ( Chiaramonte et al., 2021 ; Zheng et al., 2021 ). Diseases that spread through soil are called soil-borne plant diseases, and they adversely affect crop production worldwide, whereas some other types of organic matter might exert beneficial effects that reduce the presence of pathogens such as Fusarium , Phytophthora , Pythium, and Rhizoctonia ( Jambhulkar et al., 2015 ). If this organic matter is composed of recalcitrant carbon resources, then the ability of the soil to suppress soil pathogens is enhanced, while microbial activities in the soil are stimulated by the addition of labile carbon resources as a result of priming effects. Owing to the decomposition of microbial communities, which in turn promotes plant productivity, studies exploring a method for modifying microbial communities to improve the environment of plant root growth, which ultimately enhances plant resistance and yields, are important. Disease-suppressive soils depend on the uniformity, composition and abundance of the core microbial community ( Mendes et al., 2011 ). Therefore, the intensified competition between microbial communities and plant pathogens in disease-suppressive soils may be a disadvantage in niche allocation in the soil. These microbial communities have the ability to adapt to the various environmental conditions established for the purpose of sustainable crop production ( Keswani et al., 2019 ). Therefore, an understanding of and the targeted modification of soil microorganisms may provide new strategies for sustainable agricultural system management ( Wang and Li, 2019 ). The diversity and richness of soil bacteria and fungi determine many important aspects of ecosystem function to some extent ( Singh and Gupta, 2018 ; Sintim et al., 2019 ; Wagg et al., 2014 ). However, the imbalance of nutrients and the presence of a large number of plant pathogens in the soil might exert a negative effect on plant health ( Steffan et al., 2018 ). Soil management practices involving farming, crop rotation, and burning affect the quantity and quality of organic matter, which in turn affects soil health ( Noble and Coventry, 2005 ). Microbiome engineering technology for core microorganism inoculation directly adjusts the relationship between microorganisms, inhibits harmful microorganisms, and recruits functional microorganisms, thereby improving soil health ( Morais et al., 2019 ). Although the effects of chemical control are stable, this method often produces environmental pollution and promotes resistance in plant pathogens. In recent years, researchers have achieved outstanding progress in biological agent research and development ( Compant et al., 2019 ; Fitzpatrick et al., 2020 ). Rhizosphere growth-promoting bacteria are also important biological control resources ( Martins et al., 2013 ; Santiago et al., 2015 ). Therefore, the research and development of control methods that are harmless to the environment and exert stable effects have become increasingly important. The effects of biological control are often unstable; however, the synthetic microbial community represents a new approach to address common problems associated with current microbial fertilizers ( Qin et al., 2016 ). At the same time, disease-suppressive soil is a natural biological control agent (bca) that potentially affects the survival, infection or reproduction of pathogens. The core microbial community and its structural characteristics and the nutrients in the soil are closely related to the disease-suppressing ability of disease-suppressive soils ( Carrion et al., 2018 ; Jaiswal et al., 2017 ; Xiong et al., 2017 ). Moreover, the disease-suppressing ability of disease-suppressive soils is also related to the mode of the nutrient-microbiome interaction ( Mendes et al., 2011 ; Thakur and Geisen, 2019 ). In this review, we will discuss disease-suppressive soils, the application of human microbial preparation concepts to plant microbial preparations, and the relationship between carbon sources and the microbial community. This review aims to understand the relationship between carbon sources and microbial communities in this process."
} | 2,514 |
36234108 | PMC9572252 | pmc | 9,343 | {
"abstract": "Recently, an unprecedented growth in the internet of things (IoT) is being observed, which is becoming the main driver for the entire semiconductor industry. Reliable maintenance and servicing of the IoT is becoming challenging, knowing that the IoT nodes outnumber the human population by a factor of seven. Energy harvesting (EH) can overcome those difficulties, delivering the energyautonomous IoT nodes to the market. EH converts natural or waste energies (vibrations, heat losses, air flows, light, etc.) into useful energy. This article explores the performance of ZnO nanowires under mechanical actuation to characterize their piezoelectric performance. ZnO nanowires were fabricated using ALD and a subsequent chemical bath growth. AISI 301 steel was used as a substrate of the EH device to better fit the mechanical requirements for the piezoelectric generator. We determined that a thin layer of another oxide below ZnO provides outstanding adhesion. The samples were submitted under repetitive mechanical stress in order to characterize the output piezovoltage for different conditions. They exhibited a piezoelectric signal which was stable after hundreds of actuations. This shows good promise for the use of our device based on ZnO, an Earth-abundant and non-toxic material, as an alternative to the conventional and popular but harmful and toxic PZT. The designed measurement setup demonstrated that a AISI 301 steel substrate coated with ZnO deposited by ALD and grown in a chemical bath has promising performance as a piezoelectric material. Characterized ZnO samples generate up to 80 nJ of energy during 55 s runs under matched load conditions, which is sufficient to supply a modern IoT node.",
"conclusion": "4. Conclusions In this work, we demonstrated the growth of ZnO nanowires from a chemical bath on a novel type of substrate—AISI 301 steel. The substrate was chosen with mechanical energy harvesting applications in mind. After its deposition, the measured voltage upon repeated mechanical actuation provided the means of verifying its applicability in internet-of-things applications. The nanostructured active layer was deposited directly on the substrate, which was first prepared by seeding process using atomic layer deposition of thin oxides. A typical procedure in which pure zinc oxide is used as a seed layer to catalyze the nanowire growth was modified to account for the particular surface properties of the AISI 301 substrate. The seed layer modification, by inserting a 20 nm-thick layer of another oxide Al 2 O 3 , was found to eliminate the adhesion issues of the ZnO ALD layer. The measured voltage signal was stable upon thousands of actuations of each sample, confirming the resilience of the piezo layer to sustained mechanical stress. Moreover, the piezoactive layer generated a considerable amount of power, enough for typical operations performed by IoT devices. The power density was comparable to other reported values (see Table 1 ). In particular, it was on the same order of magnitude as the ZnO nanowires fabricated on another type of flexible substrate. These two factors, when combined, show high promise for the final application of the device as a part of a power generation component for a IoT wireless sensor node. The presented fabrication technology of the ZnO nanorods is economically attractive and can be easily upscaled. Moreover, the presented ZnO nanogenerators exhibit an interesting piezoresponse which can be boosted using topological and material optimization. Additionally, the characterization of the harvesting performance in the function of frequency will add important information of how to mechanically tune the harvester to fixed vibrations in order to reach the optimal output energy. Beyond its technological and economical attractiveness, the ZnO nanowire fabrication technology can be easily upscaled, enabling NW fabrication upon very large surfaces. This particularity, besides the high harvested output power, enables countless possibilities in sensing applications. A characterization of the inverse piezoelectric effect (excitation of vibrations when applying an AC voltage) in the presented ZnO nanorods is an inspiring and interesting research path. Such research will enable the cheap and large-scale production of crack detectors in long and big objects, such as airplane wings or train rails.",
"introduction": "1. Introduction For more than five decades, the semiconductor industry has continually astonished the market with rising fabrication capacity (delivering more than 100 bn units in 2022) [ 1 ] under unprecedented cost effectiveness [ 2 ], dumping the average transistor price more than ×1000. This has resulted in an average compound annual growth rate (CAGR) of 8.6% over a 43-year time span [ 3 ], a growth rate that has never been seen before (see Figure 1 a). Recently, a new driver for the entire semiconductor industry has appeared, which is the emerging market of the internet of things (IoT) [ 4 ]. IoT nodes already outnumber the human population 6.4× and are expected to reach 70 bn units in 2030 [ 5 ] (see Figure 1 b). The IoT market represents already a branch with a size of ∼$1.6 bn [ 6 , 7 ] and exhibited an outstanding average CAGR of ∼39% over 2015–2021 [ 8 ]. Surprisingly, the IoT growth could have been even faster, but it has been impaired by the lack of alternatives for battery and wire supply. The principle of IoT is to establish communication between small electronic devices, called Things, e.g., temperature sensors, presence detectors, and pressure sensors, distributed upon a large area. Communication between Things omits human intervention [ 9 ]. Current challenges for IoT include the size, portability, location spots and quantity of devices, and providing power supply for each IoT device node. Conventional supply (batteries or wires) is burdened with important drawbacks. Batteries require periodic checks and replacements if needed. This requires direct access to each of the nodes, and without significant breakthroughs, would soon give full-time jobs for all of humanity. Additionally, batteries have a negative impact on the environment in the form of tons of used-batteries trash. Wire supply is expensive and difficult to modify, which significantly reduces the portability and mobility of Things. Therefore, energy autonomy would be a huge relief for IoT, paving the way to future growth and expansion to the market. The fabrication capabilities boost, and cost reduction illustrated in Figure 1 a is accompanied by the galloping miniaturization (transistor’s gate length reduced by factor 200×) and performance boost (clock frequency increased 21,000×) (see Figure 2 a). Owing to this, modern electronic devices are requiring less and less energy to operate. Over the last three decades, the energy consumption has decreased by a factor of almost 42,000 for the same load. These two industrial circumstances, IoT market growth and the development of less energy-demanding devices, are driving the research toward alternative power supplies, enabling the elimination of wires and/or batteries. The development of micro- and nano-technologies has opened new possibilities to produce sufficient energy from wastes or naturally available energy sources. This approach defines the branch known as energy harvesting (EH) [ 12 ]. EH is nowadays gaining enormous attention from both industrials and academics, which is manifested by the number of filled patent applications and scientific publications (see Figure 2 b). A decrease in publications and patents after 2019 can be observed, which can be attributed to the pandemic of COVID-19 that confronted industrial and academic sectors with huge difficulties: cutting costs, production stoppages, temporary lab closure, remote work, confinements, etc. This effect is visible particularly in the share of conference papers which diminished greatly in 2020 and 2021. Nevertheless, the raising trend from before 2019 is expected to be reconstituted soon when the pandemic situation no longer affects industrial and academic working conditions. Figure 3 summarizes decades of research and thousands of scientific publications highlighting main conventional EH techniques and illustrating the physics and input energy source. Commenting on Figure 3 , it can be underlined that different energy sources require an adapted EH technique. Different physical effects and topologies are capable of producing different amounts of power. Confronting power densities in the range of μ W or mW with power needed to supply modern electron devices (see Figure 2 a gray line) or modern IoT nodes requiring ∼10–100 μ J/cycle of operation [ 18 , 29 , 30 ], it can be concluded that a well-selected EH generator can fully cover the supply requirements of a modern IoT node, offering energy-autonomous devices. This paper focuses on one of the most common EH techniques, the piezoelectric (PZ). A more focused literature review of PZ harvesters is provided in Table 1 . PZ converts mechanical energy (vibrations, stress) into electric voltage, owing to the piezoelectric effect discovered in 1880 by Jacques and Pierre Curie [ 31 ]. This effect was observed in crystals, such as tourmaline, topaz, calamine, and quartz, but not in amorphous materials. The first application of this effect was the ultrasonic submarine detector constructed in France in 1917 [ 32 ]. Since their discovery, huge improvements in PZ materials have been obtained. This market of piezoelectric devices was estimated at USD14.3 billion in 2010. The use of PZ materials is very wide ranging, from well-known high voltage generators in lighters to infertility treatment [ 33 ]. Piezoelectrics are mainly used as sound sensors [ 34 ], as cantilevers in atomic force microscopes [ 35 ], noise and vibrations attenuators [ 36 ], motors [ 37 ], and, recently, as energy harvesters [ 19 , 38 , 39 , 40 ]. The piezoelectric effect is reversible. A mechanical deformation can be achieved by applying an electric potential to the material [ 40 ]. The reverse piezoelectric effect is very useful in various diagnostic applications, e.g., wrist blood pressure or heart beat rate monitoring [ 41 ]. It is also used to detect fatigue cracks in materials [ 42 ], e.g., wind turbines rotor blades [ 43 ]. PZ offers single-step conversion by directly transforming mechanical vibrations into electric potential. The material is therefore the crucial point of this conversion. The most popular piezoelectric material is the lead zirconium titanate (PZT) [ 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 ]. Its high cost and lead (Pb) content have motivated research on alternative materials. A widely studied example is zinc oxide, which is composed of abundant and non-toxic elements. Of particular note is the possible synthesis in the form of unidirectional nanowires [ 59 , 60 , 61 , 62 , 63 ] which enhances the piezoelectric effect in the preferred axis. The directionality of the nanowires highly depends on the substrate on which they are grown, which is also an important issue concerning the current results described further in the text. In recent years, focused research effort has been directed toward components fabricated on flexible substrates [ 58 , 64 , 65 , 66 ] that greatly expand the possible range of applications served by EH components. The current research is in line with this trend. materials-15-06767-t001_Table 1 Table 1 Summary of reported piezoelectric harvesters and their parameters. f Excitation Mass Volume P Power Density Material Reference (Hz) (m/ s 2 ) (g) (cm 3 ) ( μ W) ( μ W/cm 3 ) 0.5 N/A 1.2 0.101 0.25 2.47 PZT-5H [ 52 ] 1.1 N/A N/A 25 8400 336 PZT [ 44 ] 100 72.7 0.96 0.20 35.5 16.3 PZT [ 48 ] 120 2.5 9.2 1 375 375 PZT [ 46 ] 120 0.98 N/A N/A 500 N/A PMNZT [ 67 ] 13.9 106 N/A 27·10 − 6 1 37.04·10 3 PZT [ 49 ] 1500 3.92 9·10 − 4 0.005 0.03 60 AlN [ 68 ] N/A N/A N/A 8.19·10 − 4 9.24·10 − 6 11.28 AlN [ 65 ] 230 9.8 N/A N/A 0.27 N/A PZT [ 55 ] 40 2.5 52.2 4.8 1700 700 PZT [ 50 ] 462.5 19.6 N/A N/A 2.15 N/A PZT [ 56 ] 50 N/A N/A 9 180 20 PZT [ 51 ] 56 N/A 228 113 1·10 5 2650 P1-89 PZT [ 54 ] 608 9.8 0.0016 0.0006 2.16 3600 PZT [ 53 ] 67 4 2.8 0.987 240 243.1 PZT [ 57 ] 7000 N/A N/A N/A 1600 N/A PZT-PIC255 [ 47 ] 80 2.3 0.8 0.128 2.1 16.4 PZT [ 45 ] 1 N/A N/A 0.04 5.6 140 PZT/PVDF [ 58 ] 3 N/A N/A N/A 1.296·10 − 6 2.47 PVDF micro wall [ 64 ] N/A N/A N/A 0.045 1.03 22.8 ZnO NWs [ 63 ] N/A N/A N/A 0.13·10 − 3 0.1·10 − 3 1.28 ZnO NWs on paper [ 61 ] N/A N/A N/A N/A N/A 1·10 − 3 ZnO NWs [ 59 ] N/A N/A N/A N/A N/A 10 ZnO NWs [ 60 ] 3 N/A N/A N/A N/A 144 ZnO NWs [ 62 ] 1.63 N/A N/A 4.51·10 − 4 1.38·10 − 3 3.1 ZnO NWs on steel (this work) PZT—Lead zirconium titanate, Pb[Zr x Ti 1 − x ]O 3 , PMNZT—PZT co-doped with Mg and Nb, ZnO NWs—zinc oxide nanowires, PVDF—polyvinylidene difluoride, N/A—data not available.",
"discussion": "3. Results and Discussion The ALD and CBD processes were performed on samples including the full size AISI 301 steel substrates, as well as on smaller pieces which were used for SEM imaging. The processes listed in Table 2 were carried out with heating provided by a hotplate. Two types of ALD layers were used as seed layers. The first type, labeled A , comprised ∼200 nm ZnO layers. The results of the XRD analysis are presented in Figure 9 . XRD patterns of all obtained materials show well-defined diffraction peaks of the crystalline ZnO phase with hexagonal wurtzite structure (JCPDS card # 36-1451) [ 72 ]. In the case of the A-100 and A-30 samples, recorded XRD patterns have an additional diffraction peak at 2 θ of 29.55° of relatively high intensity. Its occurrence may indicate the presence of microstrains in the samples, which are usually the results of crystal imperfection and distortion. On the other hand, this additional peak may indicate the presence of phase impurity in these samples [ 73 , 74 ]. This is also supported by SEM images of these layers ( Figure 10 ), showing two different morphologies of ZnO structures formed on the surface. In turn, XRD patterns of the B-75 and B-50 samples show peaks corresponding to pure wurtzite structure. Moreover, XRD patterns of these samples have a much more pronounced relative intensity of the diffraction peak at 2 θ of ca. 34.50°, corresponding to the (002) plane. This indicates that the growth orientation of the formed ZnO nanowires in this case is along the c-axis of the hexagonal crystals [ 75 ]. The diffractogram of the B-50 sample showed the broadest (002) peak, corresponding to the formation of ZnO structures with the smallest crystallite size. The crystallite size (D) of the formed ZnO structures was calculated using the Debye–Scherrer equation: D = 0.9 λ β cos θ , \nwhere D is the mean size of the crystallite, 0.9 is the dimensionless shape factor, λ is the X-ray wavelength, β is the line broadening at half of the maximum intensity, and θ is the Bragg angle. The obtained parameters for the (100) and (002) planes are summarized in Table 3 . The calculated values for the A-100, A-30, and B-75 samples are almost identical, suggesting good reproducibility and versatility of the used technique. The B-50 sample has the smallest value of crystallite size among the examined samples. The presence of such nanocrystallites significantly increases the surface area in electrical measurements. Figure 10 shows magnified images of each of the four samples (A-100, A-30, B-75, and B-50) which display the different growth mode depending on the type of ALD layer. It reveals that the growth of nanowires on the type A seed layer was minimal and scattered in very few randomly distributed sites on the surface. On the other hand, the ALD layer labeled B provided adequate conditions for the uniform growth of the nanowires. SEM images were used to calculate the volume of the active layer, which was then used to estimate the generated power density in the piezoelectric effect generation measurements. ZnO grown in the form of nanowires features considerable space gaps between each individual nanowire. Using the cross-section image presented in Figure 6 b, the typical length of a single nanowire was estimated to be ∼2 μ m, amounting, together with the seed layer of 200 nm, to the total possible thickness of ∼2.2 μ m. As evidenced by SEM images (e.g., in Figure 5 a), in contrast to the crystalline silicon substrate, ZnO nanowires on AISI 301 substrate grow in a more disoriented manner. The nanowires appear to grow at an angle to the surface. This is attributed to the lack of a uniform crystal structure of the substrate. Similar to the Si substrate, there is a considerable amount of gap space between the wires. Taken these aspects into account, the filling factor of the active layer was estimated at 40% for power density calculations. Thus, the volume of the piezoactive layer used for power density calculations was equal to 16 mm × 32 mm × 2 μ m × 0.4 = 4.51·10 − 4 cm 3 . Using the voltage recorded for each sample, instantaneous power was computed according to Ohm’s law, P ( t ) = V 2 ( t ) / R L O A D , where R L O A D is the resistive load set on the resistance box. Then the value of energy accumulated from the beginning of the measurement up to t = T was computed by numerical integration of power with respect to time, E ( T ) = ∫ 0 T P ( t ) d t = ∫ 0 T V 2 ( t ) d t / R L O A D (see Figure 11 a). Recorded waveforms included the signal generated by piezoelectric samples as well as a certain amount of noise. Both of these types of waveforms were integrated when computing the generated energy. For higher values of R L O A D , the contribution of noise was negligible. However, when lower values of R L O A D were used, even the low voltage values contributed significantly to the integrated energy. This was alleviated by introducing a cutoff value V C U T O F F . Below this value, the recorded voltages were ignored as if V(T) was equal to 0. Cutoff values between 0 and 5 mV were tested. It was found that V C U T O F F = 5 mV was the appropriate setting, taking into account that the undesired influence of noise at low R L O A D values was mitigated while keeping energies computed at high R L O A D essentially unchanged. The high signal-to-noise ratio shows that the samples and experimental setup were well prepared. Waveforms were recorded at each R L O A D value for a total of 55.2 s. During this period, the samples were displaced 90 times by the emulator, resulting in actuation frequency equal to 1.63 Hz. This procedure was repeated for 30 different values of R L O A D , amounting to a total of 2700 mechanical actuations delivered to each sample. Across the measurements, the voltage signal was stable. The amount of accumulated energy is displayed in Figure 11 b as a function of the resistive load connected to each sample. Energy generation parameters are listed also in detail in Table 4 . For each of the measured samples, a maximum value of accumulated energy was obtained at a certain resistive load R L O A D value. This is the value in which the internal impedance of the sample is matched to the external impedance of R L O A D . Depending on the sample, the matched impedance value was between 0.81 and 8.11 M Ω . It should be noted, however, that the energy-load profile is the most spread out for the sample that appears to have the lowest impedance, A-30. The power was just 0.02 nW lower at R L O A D = 7.11 M Ω . This places it much closer to impedance-matched conditions measured for other samples. Other samples, in particular type- B have a more clearly defined maxima. This leads us to believe that for these two samples, the highest energy points indeed correspond to the impedance-matched conditions. The highest amount of energy is generated by sample B-75—76.37 nJ—at the optimal setting of the resistive box. This corresponds to the amount of power equal to 1.38 nW. Taking into account the volume of the active layer, the power density can be calculated at 3.1 μ W/cm 3 . Compared to other reported piezoelectric harvesters (see Table 1 ), this value places it in the middle of reported power densities for harvesters based on ZnO nanowires. It is slightly higher but comparable to ZnO synthesized on a paper substrate, which is also a flexible substrate, similar to deposition on the AISI 301 steel reported here, and may serve similar applications in mechanical energy harvesting. Devices which draw few nanowatt sleep power and 16 pJ per transmitted bit are realized in the research of wireless sensors [ 52 ]. These could be powered easily by a small array of piezoelectric devices described here. The growth methods also allow for the deposition of a thicker layer (ALD) or vertically stacked structures [ 76 ]. Both of these options might be engineered to contribute to an increased piezovoltage in future applications. Error Analysis The characterization of 2D materials is challenging because of the big impact of leakages and parasitic effects introduced by the measurement circuit. Therefore, in order to minimize the impact of the measurement circuit on the characterized ZnO nano-rods piezoresponse, large-surface metallic electrodes were used. Additionally, the access electrodes are fixed using screws, which provide small contact resistivity under significant mechanical vibrations. Figure 7 presents the characterization circuit used during this experiment. It can be observed that the ZnO nanowires piezovoltage is measured by the oscilloscope. According to the technical documentation of the oscilloscope [ 77 ], the maximal error along the Y-axis (voltage error) induced by the oscilloscope is 4.25%. In order to minimize the impact of this error, a statistical data treatment upon the population of at least three measurements was implemented."
} | 5,463 |
37454187 | PMC10349831 | pmc | 9,344 | {
"abstract": "Predicting ecosystem function is critical to assess and mitigate the impacts of climate change. Quantitative predictions of microbially mediated ecosystem processes are typically uninformed by microbial biodiversity. Yet new tools allow the measurement of taxon-specific traits within natural microbial communities. There is mounting evidence of a phylogenetic signal in these traits, which may support prediction and microbiome management frameworks. We investigated phylogeny-based trait prediction using bacterial growth rates from soil communities in Arctic, boreal, temperate, and tropical ecosystems. Here we show that phylogeny predicts growth rates of soil bacteria, explaining an average of 31%, and up to 58%, of the variation within ecosystems. Despite limited overlap in community composition across these ecosystems, shared nodes in the phylogeny enabled ancestral trait reconstruction and cross-ecosystem predictions. Phylogenetic relationships could explain up to 38% (averaging 14%) of the variation in growth rates across the highly disparate ecosystems studied. Our results suggest that shared evolutionary history contributes to similarity in the relative growth rates of related bacteria in the wild, allowing phylogeny-based predictions to explain a substantial amount of the variation in taxon-specific functional traits, within and across ecosystems.",
"introduction": "Introduction In soils, microorganisms participate in many ecological processes that are critically important to the maintenance of ecosystems, such as organic matter decomposition, nitrogen fixation, and nutrient immobilization [ 1 , 2 ]. These ecosystem processes are determined by the aggregated traits of the individual taxa that make up microbial communities [ 3 – 5 ]. Unfortunately, most studies of soil bacteria characterize communities using marker gene sequencing which provides little information beyond phylogenetic community composition. To understand how community composition influences ecosystem processes we must characterize the traits of microbial taxa. Trait-based approaches have proven useful to connect the composition of plant and animal communities with ecosystem functions for modeling [ 6 – 8 ]. However, the diversity of microorganisms and the difficulty associated with measuring the traits of microbial taxa in natural communities has made connecting microbial community structure and function challenging. Most environmental bacteria cannot be isolated and the few organisms that are culturable outside of their natural environments fail to adequately represent prokaryotic diversity [ 9 , 10 ]. Metagenomic sequencing can provide functional ‘potential’ [ 11 ] and can be used to estimate bacterial replication rates [ 12 ]. However, genome-based indicators of functional potential often fail to predict observed traits. For instance, rRNA copy number [ 13 ] and genome size [ 14 ], are predictive of maximum growth rates in pure culture, but these traits do not correlate with growth under natural soil conditions [ 15 ]. Most molecular methods of community analysis do not distinguish active populations of microorganisms from dormant, although the latter may constitute the majority of observed taxa in a community [ 16 ], which may contribute to the disparity between growth rates in culture and natural conditions. Quantitative stable isotope probing (qSIP) enables the measurement of key microbial traits, such as relative growth rate, by measuring the amount of heavy isotope incorporation into taxon-specific DNA sequences in their natural environments [ 17 ]. Measurements of microbial function reflect the contributions active populations under specific environmental conditions, and quantifying the effects of environmental factors (such as temperature) on the traits of individual microbial taxa is an important step toward connecting microbial community composition with function. Experiments using qSIP have begun to quantify soil bacterial traits in an increasing number of ecosystems [ 18 , 19 ] and in response to a variety of experimental treatments [ 20 – 23 ]. However, the direct measurement of bacterial traits for all taxa in all ecosystems would be an insurmountable feat. Consequently, characterization of microbial processes across all ecosystems will require methods for inferring functional processes from microbial community composition. A phylogenetic signal in microbial functional traits (i.e. greater similarity in the traits of close relatives than expected by chance [ 21 ]) may permit trait predictions for uncharacterized taxa from phylogenies. Evolutionary processes such as rapid evolution, gene loss, and horizontal gene transfer can disrupt the phylogenetic signal in microbial functional traits. For instance, traits associated with carbohydrate metabolism in bacteria are only weakly phylogenetically clustered. In contrast, complex functional traits, such as methanogenesis and photosynthesis, that are controlled by multiple genes are more phylogenetically conserved [ 24 , 25 ]. Work with simulated community and trait data suggests traits that exhibit an adequate phylogenetic signal may be amenable to phylogeny-based trait prediction [ 26 ]. Phylogeny-based trait prediction maps trait variation to a phylogenetic tree based on observed trait measurements for members of the phylogeny, then predicts trait values for ancestors and unobserved taxa based on their position within the phylogeny [ 27 , 28 ]. When measured via qSIP, bacterial growth rates as well as carbon and nitrogen assimilation rates exhibit phylogenetic signals [ 18 , 21 , 28 , 29 ], but it is unclear if the phylogenetic signal in bacterial traits is sufficient for phylogeny-based predictions. A strong phylogenetic signal is likely to permit phylogeny-based predictions of the traits of unobserved taxa using the traits of related species measured within the same ecosystem. However, differences in phylogenetic community composition (a lack of closely related species) and trait plasticity in response to environmental variation could hamper trait prediction across ecosystems. Despite these challenges, phylogeny-based prediction across ecosystems may be possible if related organisms are present in both ecosystems and there is consistency in the estimated traits of ancestors (nodes in the phylogeny) present in both ecosystems. Additionally, challenges associated with trait plasticity may be diminished by measuring traits in experiments that manipulate environmental conditions—such as temperature, which is a principal regulator of microbial activity [ 4 , 5 ]. While warming generally increases microbial activity and decomposition rates, soil carbon responses to warming temperatures remain challenging to predict. This may be because the temperature sensitivity of individual taxa varies [ 19 ], which is not currently accounted for in ecosystem models [ 5 ]. Community composition and taxon-specific temperature responses to warming were found to improve predictions of soil carbon mineralization in a controlled experiment [ 19 , 30 ], phylogeny-based trait prediction could help make this possible on a larger scale. Phylogenies constructed from hundreds of thousands of genomic sequences provide a robust model of prokaryotic evolutionary history [ 31 , 32 ], and widespread surveys of prokaryotic community composition provide data from diverse environments and soil conditions [ 33 , 34 ]. Predicting traits from phylogeny could harness this data to estimate taxon-specific and community level function, but the accuracy of phylogeny-based trait prediction using empirical data is currently unknown. Here we aimed to determine if the phylogenetic signal in bacterial relative growth rate is sufficient to support phylogeny-based trait prediction and examine the accuracy of phylogeny-based trait prediction within and across distinct ecosystems. Our first objective was to assess the accuracy of phylogeny-based trait predictions of bacterial relative growth rates and determine how this accuracy varied with the phylogenetic signal within ecosystems. Our second objective was to determine if there was covariation in relative growth rates for taxa and ancestral nodes shared between ecosystems that could provide a foundation for phylogeny-based trait prediction across ecosystems. Our third objective was to assess the potential for, and accuracy of, phylogeny-based prediction of bacterial relative growth rates across dissimilar ecosystems. To address these objectives we used qSIP measurements of bacterial relative growth rates, from a previously published study of Arctic, boreal, temperate, and tropical soils ( n = 5) [ 19 ]. As bacterial relative growth rate is highly sensitive to temperature, we tested phylogeny-based prediction of relative growth rates across a range of temperatures to gain insight into how environmental conditions may influence the utility of phylogeny-based trait predictions of bacterial activity. Bacterial phylogenies were constructed for the communities of each ecosystem incubated at each temperature and used to assess phylogenetic signals, estimate ancestral relative growth rates, and assess phylogeny-based trait prediction within and across ecosystems.",
"discussion": "Discussion Taken together our results suggest that evolutionary history imprints a phylogenetic signal on the traits of bacteria that can allow phylogeny-based predictions within and across ecosystems. Specifically, bacterial relative growth rates in soil from four ecosystems displayed phylogenetic signals sufficient for trait prediction with a meaningful level of accuracy across a range of temperatures (15°–35 °C). Consequently, phylogeny-based trait prediction may help facilitate the inclusion of biodiversity informed microbial growth and turnover rates into ecosystem models [ 1 , 2 ]. In soil, bacterial growth rates are the emergent product of many genes expressed in response to spatially and temporally heterogeneous environmental conditions [ 15 , 42 , 43 ]. In our study, these environmental conditions, including pH, soil texture, and mean annual temperature and precipitation, varied widely between ecosystems, resulting in distinct bacterial communities with little taxonomic overlap [ 19 ]. However, there was a significant phylogenetic signal in bacterial relative growth rate in nearly all the ecosystem-temperature combinations (Table 1 , Fig. 1 ) demonstrating the importance of evolutionary history in shaping this ecologically important trait. The strength of phylogenetic signals we observed for bacterial growth rates are consistent with previous qSIP experiments that have shown bacterial growth and assimilation of carbon and nitrogen to be evolutionarily constrained across environmental gradients [ 18 , 28 , 29 ]. Although the phylogenetic signal for plant and animal traits varies widely, many functional traits have comparable phylogenetic signal values to bacterial growth rates and could be suitable for comparing results of phylogeny-based analyses [ 37 , 44 ]. Bacterial genes related to complex ecological functions such as nitrogen fixation, methanogenesis, and photosynthesis are often phylogenetically conserved [ 24 ] and our results suggest the genetic basis of bacterial growth rate may follow similar patterns of vertical inheritance. The accuracy of phylogeny-based trait prediction using the traits of related species measured within the same ecosystem increased linearly with the strength of phylogenetic signal (Fig. 2 ). Our results are consistent with past in-silico work with simulated bacterial communities and trait data, which showed the accuracy of phylogeny-based trait predictions increases with stronger phylogenetic signal and decreases with the proportion of the community missing trait data and the mean phylogenetic distance to a taxon with a described trait [ 26 ]. Many clades, even among animals and plants, still lack sufficient observations of functional traits and consequently phylogeny is used to predict traits values [ 45 – 47 ]. Physiological and life strategy traits have been estimated for amphibian and mammalian species, generally with similar or higher accuracy than the best predictions of our analysis (i.e., R 2 > 0.58); this could reflect the use of larger trait datasets in these studies or selection for traits that exhibit stronger phylogenetic signals [ 48 , 49 ]. Phylogenetic analyses of plant and animal communities often benefit from more samples contributing to trait datasets and use phylogenies that represent a relatively narrow taxonomic clade, such as a single order, which complicates a comparison to our analysis which includes taxa from across an entire domain [ 45 – 49 ]. Accurate estimates of complex traits from phylogeny, such as body mass or longevity for animals and leaf area or wood density for plants, can increase the accuracy of models combining trait and environmental data to predict ecological range or threat status [ 45 – 49 ]. Despite greater diversity in the bacterial phylogeny and smaller trait datasets many animal and plant traits exhibit similar phylogenetic signal values to bacterial growth rates, and the accuracy of phylogeny-based predictions for these traits are similarly comparable to bacterial growth rate predictions [ 45 – 49 ]. The spatial distance and environmental dissimilarity of the ecosystems studied was reflected in differences in bacterial community composition [ 50 ], with relatively few taxa observed in more than one ecosystem (Supplementary Fig. S1 ). However, many of the ancestral nodes were shared across ecosystem pairs (Supplementary Fig. S2 ), and bacterial growth estimates for ancestral nodes shared between ecosystem pairs were correlated at almost every temperature. However, the strength of significant relationships, measured as Pearson’s correlation coefficient, varied drastically (Fig. 3 ). The relationships of ancestral character state estimates for nodes shared by communities incubated at 5 °C were the most variable and included the strongest correlation (e.g. between the Arctic and temperate soil communities), but also some of the poorest correlations (e.g. those involving the tropical ecosystem). The correlation we observed in ancestral trait values across very different ecosystems indicates a significant and consistent influence of evolutionary history on bacterial growth rates, strong enough to persist across great variation in biotic and abiotic conditions. For example, some clades had below average relative growth rates across ecosystems, these included Rhizobiales (node 1398), and Gaiellales (node 1556), while other clades, including Xanthomonadaceae (1513) and Sphingobacteriales (node 1765), had above average relative growth rates across ecosystems (Fig. 4 ). These consistent patterns may provide a foundation for relating phylogenetic community composition to ecosystem function across space and time. Correlation in ancestral growth estimates between distinct ecosystems suggests phylogeny may aid in predicting functional traits, even for bacterial communities with very limited overlap in taxonomic identity. The accuracy of relative growth rate prediction, using phylogeny and trait measurements from distinct ecosystems, varied with temperature and ecosystem, and accuracy increased linearly with correlation of ancestral growth estimates at shared nodes (Fig. 5 ). Estimates were generally less accurate than predictions within an ecosystem (Fig. 2 ), which is unsurprising because trait measurements reported in these experiments are not independent from ecosystem-specific biotic and abiotic conditions. Variation in relative growth rates across the different environments and temperatures is the product of both environment and genetics, but only the latter affects trait prediction based on phylogenetic analyses. The relationship between relative growth rate prediction accuracy and consistency in ancestral relative growth rate estimates between two communities indicates that phylogeny-based trait prediction across ecosystems is only possible when phylogenetic coherence is strong enough to persist across differences in biotic and abiotic conditions. For pairings of communities with strongly correlated growth estimates across their shared ancestry up to 38% of trait variation could be explained by phylogeny alone, without accounting for environmental factors (Fig. 5 ). Given the great differences between the ecosystems included in this study (Supplementary Table 1 ), more accurate cross-ecosystem predictions may be possible for ecosystems pairs with higher biotic and abiotic similarity. The effect of temperature on growth rate can vary across individual taxa depending on their genes, physiology, and interactions in the ecosystem. Overall, bacterial growth tended to increase with temperature, with growth in the 5 °C incubations substantially lower than in the other incubations. At 5 °C many bacteria may have been dormant with relative growth rates too low for reliable quantification, resulting in low phylogenetic signals and inconsistent correlations between ancestral relative growth rate estimates for shared nodes (Fig. 5 ). Thus, some experimental or environmental conditions, such as low temperatures, might prevent the application of phylogeny-based prediction of traits. At the higher incubation temperatures, the poorest ancestral growth estimate relationships were generally observed for ecosystem pairs that included the boreal soil (Fig. 5 ), which exhibited the highest cumulative growth rates and lowest community diversity among the four ecosystems [ 19 ]. Additionally, the strength of phylogenetic signals were generally lower in the Boreal ecosystem relative to other systems, which could be a product of decreased community diversity and more significant influence of environmental factors on relative growth rates in this ecosystem (Table 1 ). Our results suggest that the utility of phylogeny-based trait prediction may vary in response to biotic (e.g. diversity) and environmental (e.g. temperature) factors. Consequently, additional research may be needed to identify the circumstances under which phylogeny-based trait prediction can provide reliable estimates of microbial functional traits. Looking forward, phylogeny-based trait prediction would benefit from trait databases that include environmental context, especially for abiotic factors, such as pH, temperature, and soil texture, that are known to explain biogeography of soil microbiomes [ 1 , 20 , 30 ]. In our study, phylogeny explained 38% of variance at best, and averaged just 14%, when predicting relative growth rate across ecosystems. As traits are a function of gene expression (phenotypes), a modeling effort that includes basic environmental parameters (e.g., pH and temperature) may be able to greatly improve our predictive power of phylogenetically conserved traits, like relative growth rate. Determining which microbial traits are appropriate for these methods will require substantial testing, but patterns of phylogenetic organization in both trait values and the abundance of genes associated with traits of interest indicates that phylogeny can inform ecologically relevant microbial functions [ 24 , 51 ]. Modeling the interaction of environmental factors and phylogeny was beyond the scope of this project, but the data from this and similar experiments is ideal for developing such a model. As quantitative trait measurement is applied to more diverse ecosystems and processes, the increase in data will provide more reliable trait estimates. Experiments that measure bacterial traits in situ are particularly important, as results from microcosm experiments may not adequately represent ecosystem processes as they naturally occur. Increased understanding of the influence of evolutionary history on trait distribution under different environmental conditions will help determine the traits and ecosystems that would be most suitable for phylogeny-based trait prediction. In summary, our results suggest that bacterial growth, a complex trait influenced by many heritable features, exhibits phylogenetic organization and phylogeny-based prediction can explain a substantial amount of the variation in this trait within and across ecosystems. Microbial traits such as growth rate impact how microbes transform elements within ecosystems, indeed estimates of microbial growth are often tied to rates of carbon mineralization [ 19 , 52 ]. Given this, phylogeny-based predictions of microbial traits such as growth rates may help bridge the divide between phylogenetic microbial community composition and ecosystem function."
} | 5,182 |
33671140 | PMC7923039 | pmc | 9,346 | {
"abstract": "Genome-scale metabolic models are of high interest in a number of different research fields. Flux balance analysis (FBA) and other mathematical methods allow the prediction of the steady-state behavior of metabolic networks under different environmental conditions. However, many existing applications for flux optimizations do not provide a metabolite-centric view on fluxes. Metano is a standalone, open-source toolbox for the analysis and refinement of metabolic models. While flux distributions in metabolic networks are predominantly analyzed from a reaction-centric point of view, the Metano methods of split-ratio analysis and metabolite flux minimization also allow a metabolite-centric view on flux distributions. In addition, we present MMTB (Metano Modeling Toolbox), a web-based toolbox for metabolic modeling including a user-friendly interface to Metano methods. MMTB assists during bottom-up construction of metabolic models by integrating reaction and enzymatic annotation data from different databases. Furthermore, MMTB is especially designed for non-experienced users by providing an intuitive interface to the most commonly used modeling methods and offering novel visualizations. Additionally, MMTB allows users to upload their models, which can in turn be explored and analyzed by the community. We introduce MMTB by two use cases, involving a published model of Corynebacterium glutamicum and a newly created model of Phaeobacter inhibens .",
"introduction": "1. Introduction Genome-scale metabolic models are important tools for systems biology. They are used in various fields, e.g. in metabolic engineering, the identification of potential drug targets, and as knowledge libraries to understand the behavior of biological systems in detail [ 1 ]. In these applications, mathematical methods are essential for the calculation of flux distributions in large networks. Flux balance analysis (FBA) allows the prediction of the steady-state condition of the metabolic network of an organism under different environmental conditions [ 2 ]. Metabolic networks typically have more reactions than metabolites, which results in an under-determined solution space. The standard approach to solve the under-determined system is to apply linear programming. In this approach, constraints are added to reduce the solution space. The optimal solution is one that maximizes or minimizes a given objective function, e.g. biomass production. The underlying assumption is that the organism has been optimized through evolution for a biological goal, such as optimal growth or minimal nutrient uptake [ 3 ]. In contrast, minimization of metabolic adjustment (MOMA) is used to calculate the flux distribution for perturbed networks, such as in knock-out mutants [ 4 ]. Due to the perturbation of the network, the assumption of FBA that the organism is evolved for a biological goal may not be valid. The underlying assumption of MOMA is that the organism undergoes a minimal redistribution of the network. For this reason, MOMA employs quadratic programming to find a solution that is closest to the wild-type flux distribution. Flux variability analysis (FVA) is an approach for analyzing the robustness of a metabolic network [ 5 ]. FVA is used to find the minimal and maximal flux for each reaction in the network while constraining some states of the flux distribution, e.g. limiting the biomass production flux to at least 95% of the FBA maximum. As most applications for flux optimization focus on a reaction-centric point of view, they pay little attention to metabolite fluxes. Only a small number of previous studies showed the benefit of analyzing models from a metabolite-centric point of view, e.g. for the discovery of new drugs [ 6 , 7 ]. There are also algorithms that maximize specific metabolite fluxes for the calculation of metabolic changes in order to predict the effect of certain drugs [ 8 ]. However, these methods were never integrated into a comprehensive modeling toolbox and can only be reproduced by experts in programming and metabolic modeling. Metabolic modeling is used for a variety of applications, ranging from biotechnological applications to ecological questions [ 9 ]. A well-characterized marine model organism is Phaeobacter inhibens DSM 17395, an ideal model organism for basic research and elucidation of stress responses because of its metabolic versatility. Experimental data for degradation pathways of different carbon sources based on metabolome and proteome measurements have been determined for P. inhibens DSM 17395 in previous studies by us and others [ 10 , 11 , 12 , 13 , 14 ]. The organism carries three extrachromosomal elements (65 kb, 78 kb and 262 kb) [ 15 , 16 ]. Proteins required for the biosynthesis of the antibacterial compound tropodithietic acid (TDA) are encoded on one of them [ 15 , 17 , 18 ]. Deletion of the plasmids leads to large variations in the growth efficiency with amino acids as a carbon source [ 19 ]. More recently, we could show that the main reason for higher growth yields of the Δ262-kb plasmid-cured mutant strain are higher non-growth-associated energy requirements of the wild-type strain because its antibiotic TDA disturbs the membrane proton gradient [ 20 , 21 ]. Thus, wild-type and TDA-negative mutant strains of P. inhibens DSM 17395 are applicable for metabolic modeling. The second model system of this study is Corynebacterium glutamicum DSM 20300, a widely used biotechnological production strain that was originally isolated as an l -glutamate-producing bacterium [ 22 ] and is used to produce about 2.5 million tons of glutamate per year [ 23 ]. Furthermore, C. glutamicum has been metabolically engineered for the production of other amino acids, organic acids, alcohols, and polymers [ 23 ]. In this study, we present Metano, an open-source command-line toolbox for metabolic modeling, and MMTB, a web-based interface to Metano that comes with an integrative database of biochemical reactions and genome annotations. While Metano was developed for a more versed audience, MMTB was designed especially for inexperienced users or users without a strong background in bioinformatics. We applied Metano methods to the published model of C. glutamicum to verify the accuracy or capability of the toolbox. In a second step, we used the TDA-negative mutant strains of P. inhibens DSM 17395, with optimal growth on three different amino acids to reconstruct a metabolic model for P. inhibens DSM 17395, based on genomic and experimental data. FBA was performed to analyze the carbon flux in the wild-type strain compared to the TDA-negative mutant strains with respect to carbon loss and growth inhibition by TDA. MMTB methods provided a metabolite-centric view on the model to determine changes in the intracellular carbon flux under different growth conditions.",
"discussion": "3. Discussion In this study, we presented Metano, a standalone toolbox for metabolic modeling. Metano provides a large number of analysis tools, is easily extensible, and supports different model formats. The toolbox includes several computationally efficient analysis methods, an interactive GUI for network visualization, a module for batch FBA simulation, and a metabolite-centric view on flux distributions that is the first of its kind. The Metano method of metabolite flux minimization uses a variability analysis approach to minimize the fluxes through metabolite nodes. While other studies already pointed out the importance of metabolite fluxes, e.g., for drug discovery, we implemented this approach in a comprehensive toolbox. Metano is extended by MMTB, a web-based toolbox for metabolic modeling. MMTB integrates EnzymeDetector annotations and biochemical reactions from different databases to assist during model creation. Web-based analyses, for example, FBA, FVA, MFM, and split-ratio analysis, allow users to analyze models without a local installation. An intuitive user interface introduces inexperienced users to stoichiometric modeling. In addition, a modeling platform drives the reuse of models and guides new users through the wide landscape of existing models. Conversion tools for SBML and JSON models further support this aspect. In contrast to well-established toolboxes, such as COBRA that provide many advanced methods especially for metabolic engineering, MMTB focuses on usability and a less experienced audience that want to effortlessly explore metabolic modeling, without refraining from the accuracy or capability of a full-fledged modeling toolbox. We confirmed the latter by applying Metano methods to the model iMG481 and comparing the results with published analyses. Moreover, we applied Metano methods to the published model of C. glutamicum DSM 20300. We were able to evaluate the accuracy and reliability of Metano methods by reproducing results from the primary literature. Furthermore, we applied Metano methods to create and elucidate a metabolic model of P. inhibens DSM 17395. We were able to simulate both the metabolic versatility and the growth-inhibiting effect of TDA. Furthermore, metabolite-centric analyses and visualizations highlighted metabolic fluxes that contribute to carbon loss and growth inhibition."
} | 2,308 |
35888885 | PMC9324494 | pmc | 9,347 | {
"abstract": "In recent years, microbubbles have been widely used in the field of microrobots due to their unique properties. Microbubbles can be easily produced and used as power sources or tools of microrobots, and the bubbles can even serve as microrobots themselves. As a power source, bubbles can propel microrobots to swim in liquid under low-Reynolds-number conditions. As a manipulation tool, microbubbles can act as the micromanipulators of microrobots, allowing them to operate upon particles, cells, and organisms. As a microrobot, microbubbles can operate and assemble complex microparts in two- or three-dimensional spaces. This review provides a comprehensive overview of bubble applications in microrobotics including propulsion, micromanipulation, and microassembly. First, we introduce the diverse bubble generation and control methods. Then, we review and discuss how bubbles can play a role in microrobotics via three functions: propulsion, manipulation, and assembly. Finally, by highlighting the advantages and current challenges of this progress, we discuss the prospects of microbubbles in microrobotics.",
"introduction": "1. Introduction Bubbles are attractive, magical, multifunctional, and ubiquitous in industrial production and our daily life. They have a variety of physical properties including large specific surface area, low density, and surface hydrophobicity. Applications based on bubbles have received extensive attention and research in recent decades [ 1 , 2 ]. Depending on their sizes, bubbles can be categorized into macrobubbles, microbubbles, and nanobubbles. Macrobubbles are 2–5 mm in diameter. At the microscopic scale, bubbles can be classified into microbubbles (diameters of 1–100 μm) and nanobubbles (diameters below 1 μm) [ 3 ]. Following the development of microfluidics and microrobotics in recent years, micro/nanoscale bubbles have been seen as an emerging tool for solving numerous challenges in various lab-on-a-chip (LOC) applications, and they are gaining increasing attention from researchers. In the micro/nano research field, bubble-based applications have attracted increasing attention because of their simplicity, controllability, and biocompatibility. They can be flexibly integrated with different microfluidic devices or microrobots. For example, in microfluidics, bubbles can be remotely excited by an acoustic field to act as micromixers [ 4 , 5 ], micropumps [ 6 , 7 ], or microvalves [ 8 , 9 ], and they can operate upon particles and cells [ 10 , 11 ]. In addition, in microrobotics, bubbles can act as the manipulating or transmission components of microrobots [ 12 , 13 ], and the bubbles themselves can act as microrobots [ 14 ]. However, the bubble generation methods and their applications in the microrobotics field (e.g., propulsion, micromanipulation, and microassembly) have not been classified and summarized in detail. In this review, we introduce and discuss the propulsion, manipulation, and assembly capabilities of bubbles in microrobotics. We demonstrate the importance, flexibility, and versatility of bubbles, as described in typical research papers. A schematic diagram of the generation and control methods of bubbles and their roles in microrobotics is shown in Figure 1 . In Section 2 , we introduce the methods of bubble generation (chemical reaction, direct acquisition, and optothermal effect) and control (acoustic oscillation, optothermal effect, and electrowetting-on-dielectric (EWOD) technology). In Section 3 , we discuss how bubbles can be used as propulsion mechanisms for microrobots (e.g., tubular micromotors, Janus particles, and self-propelled micromachines). In Section 4 , we demonstrate how bubbles act as the tools of microrobots and can be used for micromanipulation or transmission. In Section 5 , we introduce bubbles as microrobots to achieve micromanipulation and microassembly in two-dimensional (2D) and three-dimensional (3D) spaces. Finally, we summarize the current limitations of bubbles in microrobotics and discuss future developments."
} | 1,008 |
31354767 | PMC6640087 | pmc | 9,348 | {
"abstract": "Arbuscular mycorrhizal fungi (AMF) form symbioses with the roots of most plant species, including cereals. AMF can increase the uptake of nutrients including nitrogen (N) and phosphorus (P), and of silicon (Si) as well as increase host resistance to various stresses. Plants can simultaneously interact with above-ground insect herbivores such as aphids, which can alter the proportion of plant roots colonized by AMF. However, it is unknown if aphids impact the structure of AMF communities colonizing plants or the extent of the extraradical mycelium produced in the soil, both of which can influence the defensive and nutritional benefit a plant derives from the symbiosis. This study investigated the effect of aphids on the plant-AMF interaction in a conventionally managed agricultural system. As plants also interact with other soil fungi, the non-AMF fungal community was also investigated. We hypothesized that aphids would depress plant growth, and reduce intraradical AMF colonization, soil fungal hyphal density and the diversity of AM and non-AM fungal communities. To test the effects of aphids, field plots of barley enclosed with insect proof cages were inoculated with Sitobion avenae or remained uninoculated. AMF specific and total fungal amplicon sequencing assessed root fungal communities 46 days after aphid addition. Aphids did not impact above-ground plant biomass, but did increase the grain N:P ratio. Whilst aphid presence had no impact on AMF intraradical colonization, soil fungal hyphal length density, or AMF community characteristics, there was a trend for the aphid treatment to increase vesicle numbers and the relative abundance of the AMF family Gigasporaceae. Contrary to expectations, the aphid treatment also increased the evenness of the total fungal community. This suggests that aphids can influence soil communities in conventional arable systems, a result that could have implications for multitrophic feedback loops between crop pests and soil organisms across the above-below-ground interface.",
"conclusion": "Conclusion Aphids increased the evenness of the entire fungal community within plant roots, and also tended to increase the level of vesicles and abundance of the AMF family Gigasporaceae. Whether these increases are due to increased C allocation below-ground by plants attempting to increase nutrient uptake, or the active selection of fungal taxa in response to herbivory requires elucidation. Whether these changes in the below-ground soil community feed back into altered aphid performance is currently unclear, but the response of agriculturally relevant fungal communities to top-down effects of herbivory suggests that above-below-ground community feedback could occur in agricultural systems.",
"introduction": "Introduction Arbuscular mycorrhizal fungi (AMF) form obligate symbioses with the roots of c. two-thirds of land plant species, including agriculturally important cereals ( Smith and Read, 2008 ; Fitter et al., 2011 ). Enhancing this symbiosis has been proposed as an important tool for increasing food security and agricultural sustainability ( Gosling et al., 2006 ; Fitter et al., 2011 ; Jacott et al., 2017 ; Thirkell et al., 2017 ). Whilst the host plant provides a fixed carbon (C) source for AMF, AMF transfer nutrients such as nitrogen (N) and phosphorus (P) to the plant ( Hodge et al., 2001 ; Smith et al., 2009 ; Hodge and Fitter, 2010 ; Karasawa et al., 2012 ). AMF colonization affects multitrophic interactions between above- and below-ground herbivores ( Yang et al., 2014 ) and may also enhance the uptake of silicon (Si), which can alleviate the impact of both biotic and abiotic stress ( Dias et al., 2014 ; Garg and Bhandari, 2016 ; Frew et al., 2017 ). The bottom-up effect of below-ground AMF on the performance of above-ground herbivores such as aphids can range from positive to negative ( Gange and West, 1994 ; Wurst et al., 2004 ; Ueda et al., 2013 ; Simon et al., 2017 ; Wilkinson et al., 2019 ). These impacts on aphid performance likely occur because of alterations to plant defense and nutrition due to the AMF symbiosis ( Wurst et al., 2004 ; Meir and Hunter, 2018b ) and can depend on the level of AMF colonization of the host plant ( Tomczak and Müller, 2017 ; Maurya et al., 2018 ; Meir and Hunter, 2018a ). In turn, aphids may impose top-down effects on AMF colonization via the host plant ( Babikova et al., 2014 ; Meir and Hunter, 2018a ). Top-down and bottom-up effects can therefore modulate the outcome of each other, potentially resulting in above-below-ground multitrophic feedback loops ( Meir and Hunter, 2018a ). Thus, if aphids influence AMF colonization this could impact how AMF affect plant nutrient uptake and tolerance to abiotic stress in multitrophic systems. The AMF extraradical mycelium (ERM) phase is of also of key importance for interactions between plants and other rhizosphere organisms ( Perotto and Bonfante, 1997 ; Jones et al., 2004 ; Hodge and Fitter, 2013 ) and can be directly related to AMF derived plant nutrient acquisition ( Hodge et al., 2001 ; Barrett et al., 2011 ). Additionally, the ERM can be involved in plant defense, carrying signals of aphid attack to neighboring plants connected via ERM underground networks ( Babikova et al., 2013 ). Elucidating how intra- and extraradical structures of AMF respond to top down effects is therefore important in understanding their potential for use in complex agro-ecosystems. However, current knowledge of how AMF respond to aphids sharing the same host plant is limited to the impact on AMF colonization ( Babikova et al., 2014 ; Vannette and Hunter, 2014 ; Maurya et al., 2018 ; Meir and Hunter, 2018a ). The identity of the taxa within the AMF community colonizing the host plant can be important in determining the nutrient uptake or defense benefit gained from the symbiosis. In small, artificially selected AMF communities, AMF species identity determines the level of protection the AMF provides for the host plant against biotic stressors ( Pozo et al., 2002 ; Sikes et al., 2009 ; Malik et al., 2016 ), and certain AMF species may deliver more or less nutrients to their host plant ( Jansa et al., 2008 ; Leigh et al., 2009 ; Thirkell et al., 2016 ). Similarly, soil community transfer experiments suggests the AMF community structure can also be important in determining nutrient acquisition and plant growth responses ( Hodge and Fitter, 2013 ; Williams et al., 2014 ; Manoharan et al., 2017 ; Jiang et al., 2018 ). Large vertebrate grazing can affect AMF communities ( Ba et al., 2012 ; Guo et al., 2016 ), and insect herbivory can alter below-ground ectomycorrhizal ( Gehring and Bennett, 2009 ), non-mycorrhizal fungi ( Kostenko et al., 2012 ) and rhizosphere bacterial community characteristics ( Kong et al., 2016 ), but the impact of arthropod herbivory on AMF communities is currently unknown. The impact of herbivory on AMF structures is variable ( Barto and Rillig, 2010 ), and phloem feeding aphids can increase or decrease the intraradical AMF colonization of their host plant ( Babikova et al., 2014 ; Meir and Hunter, 2018a ). The C limitation hypothesis proposes that above-ground removal of fixed C by herbivory will result in less C available below-ground ( Wallace, 1987 ), although the subsequent allocation of this limited amount of C between roots and AMF contained within roots is unknown. The reduced C availability might result in changes to the AMF community because only a limited number of AMF species can be supported and thus the number and relative abundance of less competitive species might be reduced ( Gange, 2007 ; Ba et al., 2012 ). Alternatively, low levels of herbivory could lead to more C allocated below-ground in an attempt for the plant to take up more nutrients for regrowth ( Wamberg et al., 2003 ), which could increase fungal diversity ( Ba et al., 2012 ). There are also examples of herbivory affecting the composition of ectomycorrhizal communities rather than species richness, and thus altering the beta diversity of communities, making communities more distinct ( Gehring and Bennett, 2009 ). Arbuscular mycorrhizal fungi communities in conventionally managed agricultural systems are often distinct from AMF communities in other settings and often have low diversity due to tillage, chemical fertilizer, pesticide and fungicide regimes ( Jansa et al., 2002 ; Gosling et al., 2006 ; Wetzel et al., 2014 ; Hartmann et al., 2015 ; Manoharan et al., 2017 ). We selected a conventionally managed system to investigate agriculturally relevant AMF communities that are tolerant to such practices. Apart from AMF, other fungi also associate with plant roots, including other endophytic mutualists ( Murphy et al., 2015 ; Lugtenberg et al., 2016 ). As some of these other fungi and their community composition can influence aphid performance ( Hartley and Gange, 2009 ; Battaglia et al., 2013 ; Kos et al., 2015 ), the AMF community must be placed in the context of any changes in the wider fungal community. Here, we investigate the impact of aphids on the below-ground fungal community with a focus on AMF given their key role as ecosystem engineers. Specifically, we tested the following hypotheses: (1) As aphids will depress plant growth and nutrient status, the AMF will benefit less from the association with the plant, and consequently AMF structures, both internal and external to the root, will be reduced. (2) Aphids will cause a reduction in both the alpha diversity and evenness of the soil communities, which results in a distinct soil fungal community composition (increased beta diversity).",
"discussion": "Discussion This study aimed to investigate the impact of aphids on soil fungi in a conventionally managed agricultural system. It was hypothesized that by supressing plant nutrition and growth, aphid feeding would lead to negative impacts on AMF structures and species richness, as well as the evenness of the AMF and other fungi communities, measured here through analysis of an AMF specific amplicon and another less precise, but wider encompassing amplicon (total fungi). It was also proposed that the effects of aphids would impact the compositions of these communities ( Gehring and Bennett, 2009 ), and the relative abundance of AMF taxa within them. In fact, aphid presence had less of an impact on AMF community structure than the location of the host plant in the field, although there was a trend of the abundance of AMF vesicles, and the abundance of the Gigasporaceae family to increase when aphids fed on the host plant. Within the total fungi community, the relative abundance of AMF was also affected by location rather than aphid presence. However, aphid presence increased evenness across the total fungi community. Effects of Aphids on the Plant Biomass and Nutrition, and AMF Structures Contrary to our hypothesis, S. avenae had little effect on the above-ground nutrition of barley in the field, although aphid presence tended to increase the above-ground plant N:P ratio ( Table 3 ), possibly due to nutrient re-allocation caused by aphid feeding ( Sandstrom et al., 2000 ; Thompson and Goggin, 2006 ; Nowak and Komor, 2010 ), or differences between the requirement for N and P by aphids ( Tao and Hunter, 2012 ). Moreover, aphid presence did not reduce above-ground biomass in the present study. Since the aphids used in this experiment were cultured under controlled conditions it is likely that as the aphids did not vector any plant viruses that are a major contributor to aphid related yield loss in cereals ( McKirdy et al., 2002 ). While it is possible that the agro-chemical inputs of fungicides, herbicides and plant growth regulators in this conventionally managed system may have influenced aphid development, the total number of aphids per tiller, and thus, estimated in each “+ Aphid” plot, remained high. However, the aphid population in the current experiment was launched at an earlier date than in non-controlled systems ( Blackman and Eastop, 2000 ) which may have also influenced the results. AMF colonization of barley roots was high compared to that measured in field studies previously ( Boyetchko and Tewari, 1995 ), and in certain glasshouse studies ( Grace et al., 2009 ). Thus, while we originally hypothesized that aphids would have a negative impact on above-ground plant biomass and nutrition which in turn, would reduce both the internal and external phases of the AMF, no negative impact occurred. This may explain why both AMF RLC and hyphal length density were not affected by the presence of aphids in this study. However, there was a trend for aphids to increase the proportion of vesicles in plant roots. Mechanical defoliation has also been shown to influence the proportion of vesicles in plant roots when grown with a native AMF soil community ( Garcia and Mendoza, 2012 ). As vesicles are lipid storage organs in AMF, and AMF derive lipids from the host plant ( Keymer and Gutjahr, 2018 ), this might suggest that more fixed C is available to the AMF via the plant under aphid herbivory. Alternatively, AMF vesicle numbers may have increased in a response to less C flow from the plant in order to aid survival under stressed conditions, similar to the AMF response observed in cool conditions ( Hawkes et al., 2008 ), It is also possible that the frequency of AMF with a propensity to form more vesicles occurred within the Glomeracea community. However, the tendency for aphid presence to increase the relative abundance of the AMF family Gigasporaceae, which do not develop vesicles ( Smith and Read, 2008 ), suggests that this explanation is less likely. Effects of Aphids on AMF and Total Fungal Communities The second hypothesis proposed that the effects of aphid feeding on the host plant would reduce AMF and total fungal species richness and evenness. The low richness of AMF species identified in the current study is similar to that documented in other conventionally managed barley monoculture systems ( Manoharan et al., 2017 ), and as the read depth achieved for the AMF specific amplicon is sufficient to capture AMF diversity ( Vasar et al., 2017 ), it can be assumed that this is an accurate representation. The number of species of AMF were not impacted by aphid presence perhaps as there was no effect of aphid presence on above-ground plant biomass; however, this is similar to the lack of an effect of arthropod feeding on pinyon pine associated ectomycorrhizal communities ( Gehring and Bennett, 2009 ). This was also reflected in the number of species in the entire fungal community in the current study. Moreover, and counter to expectations, aphid presence increased the evenness of the entire fungal community in the present study. Nutrient flows below-ground were not measured in the current study, but aphids can affect below-ground respiration depending on plant growth stage, potentially due to alterations in C availability to soil microbes ( Vestergård et al., 2004 ). Aphids can also alter the profiles of compounds, released from plant roots ( Hoysted et al., 2018 ) and can also change the profiles of sugars found in AMF hyphae sharing the same host plant ( Cabral et al., 2018 ). Moreover, aphids excrete honeydew as a result of their C rich diet of phloem sap which can be utilized as a C source by soil microbiota, thus shaping community structure and biomass ( Katayama et al., 2014 ; Milcu et al., 2015 ). As more C sources become available in the root, it is possible that niches may enlarge allowing less competitive fungi to compete, reducing the dominance of abundant taxa. However, it should be noted that aphid induced alterations to soil organisms can occur independently of honeydew C inputs ( Sinka et al., 2009 ), and that soil microbes can be influenced by aphid induced changes to plant root exudates in systems where honeydew does not reach the microbe ( Kim et al., 2016 ). Above-ground herbivory generally stimulates the cycling of nutrients by decomposers in the soil ( A’Bear et al., 2014 ), which could partly explain increases in the N:P ratios of plant tissues infested by aphids in the current study. It was hypothesized that the abundance of AMF taxa within the AMF community would be impacted by aphid presence, and there was a marginal increase in the abundance of Gigasporaceae under aphid infestation of the host plant. A recent meta-analysis revealed that members of this family are tolerant to fertilizer input disturbances, suggesting a role aside from nutrient acquisition, perhaps in plant defense ( van der Heyde et al., 2017b ). Species indicator analysis may also identify taxa affected by treatments, however, as low abundance taxa typically score as poor indicators, the results of this method may differ from those investigating relative abundances ( Longa et al., 2017 ). None of the AMF VTs were indicators of either treatment, but several total fungi amplicon OTUs were significant indicators of aphid presence or absence. Currently, it is unclear whether these fungi are responding to changes in nutrient availability, whether the plant recruits them in response to aphid feeding to aid with defense ( Kong et al., 2016 ; Pineda et al., 2017 ), or whether the recruitment of specific soil microbes ultimately benefits the aphid ( Kim et al., 2016 ). No clear effects of aphid presence were found on community composition, perhaps as a longer period of top down pressure is required to impact this metric. For example, the effects of grazing by large vertebrates on AMF community structure are strongly linked to the length of the grazing ( van der Heyde et al., 2017a ). However, as plant communities are removed regularly in cereal systems, and aphid feeding is seasonal ( Blackman and Eastop, 2000 ), only a relatively short window is available for these interactions to occur. Plot location, grain P concentration and AMF abundance as well as community composition were tightly linked in the current study. AMF communities may be associated with environmental and nutritional gradients in soil systems ( Bainard et al., 2014 ; Horn et al., 2014 ; Guo et al., 2016 ; Yao et al., 2018 ) which also likely reflects the spatial differences measured in above-ground plant nutrition and AMF physiology here. Although care was taken to reduce the spread of plots across the slope of the site, these spatial differences highlight the importance of environmental heterogeneity within relatively small distances (less than 10 m) in field sites, and this could have contributed to masking top down effects of aphids on community composition in the current study. Whether the associations between AMF community composition and plant Si are due to AMF uptake of Si ( Garg and Bhandari, 2016 ; Frew et al., 2017 ), or an artifact of AMF responses to soil pH gradients or water availability ( Bainard et al., 2014 ) requires further study."
} | 4,735 |
21141029 | null | s2 | 9,352 | {
"abstract": "Marine microorganisms are presented with unique challenges to obtain essential metal ions required to survive and thrive in the ocean. The production of organic ligands to complex transition metal ions is one strategy to both facilitate uptake of specific metals, such as iron, and to mitigate the potential toxic effects of other metal ions, such as copper. A number of important trace metal ions are complexed by organic ligands in seawater, including iron, cobalt, nickel, copper, zinc, and cadmium, thus defining the speciation of these metal ions in the ocean. In the case of iron, siderophores have been identified and structurally characterized. Siderophores are low molecular weight iron-binding ligands produced by marine bacteria. Although progress has been made toward the identity of in situ iron-binding ligands, few compounds have been identified that coordinate the other trace metals. Deciphering the chemical structures and production stimuli of naturally produced organic ligands and the organisms they come from is fundamental to understanding metal speciation and bioavailability. The current evidence for marine ligands, with an emphasis on siderophores, and discussion of the importance and implications of metal-binding ligands in controlling metal speciation and cycling within the world's oceans are presented."
} | 333 |
34308012 | PMC8296010 | pmc | 9,356 | {
"abstract": "The biological reduction of ferrous ethylenediaminetetraacetic\nacid (EDTA-Fe II -NO and EDTA-Fe III ) is an important\nprocess in the integrated electrobiofilm reduction method, and it\nhas been regarded as a promising alternative method for removing NO x from industrial boiler flue gas. EDTA-Fe II -NO and EDTA-Fe III are crucial substrates that\nshould be biologically reduced at a high rate. However, they inhibit\nthe reduction processes of one another when these two substrates are\npresented together, which might limit further promotion of the integrated\nmethod. In this study, an integrated electrobiofilm reduction system\nwith high reduction rates of EDTA-Fe II -NO and EDTA-Fe III was developed. The dynamic changes of microbial communities\nin the electrobiofilms were mainly investigated to analyze the changes\nduring the reduction of these two substrates under different conditions.\nThe results showed that compared to the conventional chemical absorption-biological\nreduction system, the reduction system exhibited better performance\nin terms of resistance to substrate shock loading and high microbial\ndiversities. High-throughput sequencing analysis showed that Alicycliphilus, Enterobacteriaceae , and Raoultella were the dominant genera (>25% each) during the process of EDTA-Fe II -NO reduction. Chryseobacterium had the\nability to endure the shock loading of EDTA-Fe III , and\nthe relative abundance of Chryseobacterium under\nabnormal operation conditions was up to 30.82%. Ochrobactrum was the main bacteria for reducing nitrate by electrons and the\nrelative abundance still exhibited 16.11% under shock loading. Furthermore,\nhigher microbial diversity and stable reactor operation were achieved\nwhen the concentrations of EDTA-Fe II -NO and EDTA-Fe III approached the same value (9 mmol·L −1 ).",
"conclusion": "3 Conclusions This study revealed the key influencing factors and the structure\nof the microbial community in the reduction of EDTA-Fe II -NO and EDTA-Fe III by an electrobiofilm system. Microbial\nactivity was considered to be critical in the reduction of EDTA-Fe II -NO, and a rich microbial diversity in an electrobiofilm\nreactor is important in resisting shock loading and ensuring long-term\nstable operation. As an EDTA-Fe III -reducing bacteria, Chryseobacterium can endure shock loading well. Dysgonomonas is a type of autotrophic EDTA-Fe III -reducing bacteria that is sensitive to variation in shock loading. Ochrobactrum , which can reduce nitrate using electrons,\nis more stable under shock loading. Alicycliphilus , Enterobacteriaceae , and Raoultella , accounting for approximately 80% of the electrobiofilm community,\nare likely to be the dominant genera involved in EDTA-Fe II -NO reduction, suggesting that the chelated NO-reducing bacteria\nwere predominant in this system. Therefore, higher microbial diversity\nand stable reactor operation could be achieved when the concentrations\nof EDTA-Fe II -NO and EDTA-Fe III became comparable.",
"introduction": "1 Introduction As primary pollutants, nitrogen oxides (NO x ) not only directly affect human health but also combine with ozone\nand hydrocarbons to form photochemical smog in the troposphere. 1 , 2 In 2017, a total of 1258.8 × 10 4 t of NO x were emitted in China, according to the official\ndata from the China National Bureau of Statistics. 3 However, during the outbreak of coronavirus (COVID-19)\nin February 2020, NASA and the European Space Agency (ESA) detected\na significant decrease in airborne nitrogen dioxide (NO 2 ) over China. NO x emissions are related\nto industrialization in China, 4 and it\nis crucial to limit the release of NO x into the atmosphere. For decades, the main anthropogenic source of NO x has been emissions from industrial boilers (kilns). 5 , 6 Some technologies have been introduced for controlling flue gas\nfrom boilers (kilns) and reducing the release of NO x , such as selective catalytic reduction (SCR), low-NO x burners, absorption, adsorption, and selective\nnoncatalytic reduction (SNCR). 7 , 8 However, these methods\ncan have a high cost, low removal efficiency, and cause secondary\npollution. 9 The biological treatment of\nindustrial flue gas for NO x removal was\nproposed in the 1980s as a low-cost and environmentally sustainable\napproach, and related studies have focused mainly on isolating denitrifying\nbacteria and improving biological reactors. 10 , 11 Wang et al. 12 studied denitrifying bacteria\nin bioreactors for landfill leachate treatment and found that the\nmain bacteria in the bioreactor varied with changes in the hydraulic\nloading. Sposob et al. 13 analyzed the microbial\ncommunities involved in autotrophic sulfide denitrification with changes\nin temperature and found that Thauera sp. and Alicycliphilus sp. were predominant at 25 °C. Xing\net al. 14 studied the microorganisms involved\nin the micro-electrolysis and autotrophic denitrification processes\nby high-throughput sequencing and found that β-, γ-, and\nα- Proteobacteria were the dominant genera.\nHowever, biological approaches have been limited by their low efficiency,\nwhich is caused by the low solubility of NO in liquid and the higher\nproportion of NO in NO x from flue gas. 15 , 16 Therefore, a new integrated technology has been developed that combines\ncomplex absorption processes with biological reduction. 6 , 17 Ferrous ethylenediaminetetraacetic acid (EDTA-Fe II ) had\nbeen reported to rapidly form complexes with NO, which resolves the\nissues associated with the low gas–liquid mass transfer efficiency\nof NO. 18 An electrobiofilm was subsequently\nintroduced and has been demonstrated to further strengthen the regeneration\nrate of EDTA-Fe II , 19 as it not\nonly forms complexes with NO to generate EDTA-Fe II -NO but\nalso oxidizes into EDTA-Fe III by oxygen in the flue gas\n(approximately 9% content in the flue gas in industrial boilers). 20 Therefore, the biological regeneration of EDTA-Fe II was\nbelieved to be a key step to allow the greater application of this\nmethod. 21 This is depicted in Figure 1 , which describes\nthe principle of the electrobiofilm-integrated method for NO x removal. The electrobiofilm method integrates\nthe advantages of both electrochemical and complex absorption-bioreduction\n(CABR) processes. This method offers bacteria with two categories\nof electron donors, carbon sources and currents, thereby enhancing\nthe diversity and activity of the microorganisms. 22 Figure 1 Principle of electrobiofilm processing. Several microorganisms have been screened for L-Fe III and L-Fe II -NO reduction with high efficiency, where L\nrepresents complexes of citrate or EDTA. 23 , 24 Zhang et al. 25 studied the microbial\ncommunities in CABR-integrated systems by the polymerase chain reaction-denaturing\ngradient gel electrophoresis (PCR-DGGE) method and found that Pseudomonas sp. was the dominant microorganism related to\nthe NO x removal in the biofilm. Li et\nal. 26 also analyzed the microbial communities\nin CABR-integrated systems by high-throughput sequencing and found\nthat the dominant denitrifying bacteria varied from anaerobic to facultative\nanaerobic and aerobic denitrifying bacteria with an increase in the\ninlet oxygen loading. Wang et al. 27 analyzed\nthe microbial community structure of the BTF-ABR-integrated system\nby the real-time polymerase chain reaction and high-throughput sequencing\nmethod. The results showed that the cooperation of denitrifying bacteria\nand iron-reducing bacteria in the system was the key to the stable\nand efficient removal of NO x and the regeneration\nof EDTA-Fe II simultaneously. High-throughput sequencing,\nalso referred to as “deep sequencing” technology, involves\nthe parallel sequencing of millions of molecules at a time, allowing\nrapid, detailed, and comprehensive analysis of the transcriptome and\ngenome of a species or a microbial community. 28 High-throughput sequencing has a more rapid response, higher accuracy,\nand larger reaction scale than the previously widely used applications,\nsuch as PCR-DGGE, 29 and has become an efficient\nresearch method in the field of molecular biology. 30 However, electrobiofilm-integrated systems are typical\nmultiphase complexes, and their microbial communities have not yet\nbeen studied. Illuminating the microbial communities of such systems\ncould allow for a better understanding of the EDTA-Fe II regeneration mechanism. Additionally, the stability and capacity\nfor long-term operations are crucial indicators for evaluating a bioreactor. 21 The sensitivity of the microbial system is an\nimportant factor affecting the stable operation of a bioreactor. 20 Microorganisms are sensitive to changes in environmental\nfactors, such as temperature, process conditions, and load changes. 31 However, promoting the biofilm diversity in\nan electrobiofilm system can improve resistance to shock loading of\nNO x and EDTA-Fe III . The objective of this study is to describe the key factors affecting\nthe activities of an electrobiofilm and evaluate the changes in microbial\ncommunities of electrobiofilm-integrated systems under shock loadings\nof the main absorption product EDTA-Fe II -NO and the oxidation\nproduct EDTA-Fe III by the molecular biotechnology of high-throughput\nsequencing. Furthermore, the changes of dominant strains under different\nconditions and the regeneration of EDTA-Fe II under different\nelectron donor combinations are analyzed. Finally, approaches to achieving\nstable operation of electrobiofilm-integrated systems were explored\nbased on the variation of the microbial communities. This work will\nidentify the biological mechanism of EDTA-Fe II regeneration\nin the bio-electrochemical system, discuss the optimal control mechanism\nof microbial activities in this kind of system, and provide theoretical\nreference for engineering applications on NO x removal in the future.",
"discussion": "2 Results and Discussion 2.1 Biofilm Formation in the Reactor The formation of an electrobiofilm is vital in achieving the efficient\nreduction of EDTA-Fe II -NO and EDTA-Fe III in\nthe reactor. Batch experiments were conducted with a solution containing\nup to 2 g·L −1 glucose and 18 mmol total iron\nat the startup of the electrobiofilm reactor. The biofilm started\nbecoming visible on the surface of the cathodes from the tenth day.\nEDTA-Fe II -NO was gradually added after 22 days, followed\nby repeated batch reduction until its concentration was equal to the\ninitial EDTA-Fe III concentration. As shown in Figure 2 , when the reduction\nefficiency of EDTA-Fe II became stable at around 80%, it\nis considered that the reactor has adapted to a certain ratio of substrate\ncombinations. The current increased gradually as an electron donor\nafter the 38th day ( Table 1 ). That is, part of glucose (carbon source) was replaced by\nthe current. The composition of the electron donor was changed to\nadapt the microorganisms to a carbon source (glucose) concentration\nof 1 g·L −1 . At the end of the eighth week,\nthe biofilm on the cathode was highly dense, according to field emission\nscanning electron microscopic (FESEM) images, as shown in Figure 3 . The efficiency\nof EDTA-Fe II regeneration increased from 12 to 94% after\n55 days. During the stable operation of the reactor, the CO 2 produced at the anode dissolved in the liquid phase and formed a\nCO 2 –HCO 3 2– system that\nhad a buffering effect on the pH value, such that the pH value in\nthe reactor generally remained between 6.7 and 6.9. By contrast, the\nEDTA-Fe II regeneration efficiency reported by Gao et al. 22 was 76–85% at the end of the 90-day domestication\nperiod. Overall, the sequential biofilm formation method could accelerate\nthe domestication of microorganisms and biofilm formation due to the\nnegative effects of EDTA-Fe II -NO on the activities of microorganisms\nin electrobiofilm systems 32 and difficulties\nin the cultivation of microbial systems that relied on an electrical\ncurrent as an electron donor. Figure 2 Reduction efficiency of different EDTA-Fe II -NO and EDTA-Fe III concentrations (total Fe = 18 mmol, I =\n20 mA, U = 12 V, initial glucose = 0.2 g·L −1 , liquid flow rate = 1.2 L·min −1 , pH = 6.7–6.9)\n(solid black box: EDTA-Fe II -NO concentrations; solid red\ncircle: EDTA-Fe III concentrations; and blue star: glucose\nconcentrations). Figure 3 FESEM images of (a) the electrode before biofilm formation and\n(b) the electrobiofilm after full growth (×5000). Table 1 Conditions of Electron Donors in Different\nStages of Biofilm Formation stages current (mA) glucose (mg·L −1 ) 7-1 10 2000 7-2 15 1500 7-3 20 1000 2.2 Optimization of Electron Donor Combination Electrical currents and carbon sources (glucose) are the crucial\nelectron donors in electrobiofilm treatment and primarily impact the\nEDTA-Fe II regeneration rate. In our earlier studies, the\nrespective influence of each electron donor on the EDTA-Fe II regeneration rate was discussed. The carbon sources were more important\ndonors for EDTA-Fe II regeneration than the electrical current.\nTo investigate the interactions between current and glucose and their\ninfluences on the EDTA-Fe II regeneration rate, a factorial\nanalysis of the effects under different electron donor combinations\nwas also performed. When both variables influence the experimental\nresults, these can be used as a function to evaluate the interactive\neffects of cathode electrons and glucose during EDTA-Fe II regeneration. The results obtained under different currents\nand carbon sources were used to develop a prediction model equation\nusing Design-Expert software, which is as follows where V is the regeneration\nrate of EDTA-Fe II (mmol·L −1 ·h −1 ), I is the applied current (mA),\nand G is the glucose concentration (g·L −1 ). The P -value of the model obtained\nby factorial analysis was 0.0063 ( P ≤ 0.05),\nindicating that the obtained model was reliable and statistically\nsignificant. The coefficient of G was positive, suggesting\nthat the effect of glucose is positive, and could promote the regeneration\nof EDTA-Fe II . Meanwhile, the coefficient of I was negative, suggesting that the regeneration of EDTA-Fe II was reduced with an increased current. The I·G coefficient was positive, indicating that the interaction between\nthe two electron donors could promote EDTA-Fe II regeneration\n(either or both of the EDTA-Fe II -NO and EDTA-Fe III reduction). Glucose acted as an essential organic carbon source\nfor the growth of microorganisms and was an electron donor during\nthe EDTA-Fe III and EDTA-Fe II -NO reduction. The\nhydrogen produced by the cathode electrons could be used by microorganisms\nin situ. Thus, these two processes promoted the regeneration of EDTA-Fe II . 32 , 33 32 , 33 It had been\nspeculated that the microbial activity contributed more to the regeneration\nof EDTA-Fe II , while the current promoted other aspects\nof the electrobiofilm system. Therefore, the biofilm mechanism in\nthis system needs to be better understood. By comparing the actual value obtained from the experiment with\nthe predicted value obtained from the prediction equation under the\noperating conditions of a 20 mA current and glucose content of 1000\nmg·L −1 , it was found that the actual and predicted\nvalues were well correlated (i.e., coefficient of correlation ( R 2 ) of 0.84; Figure 4 ). Additionally, according to Table 2 , when the current was 20 mA\nand the glucose content was 1000 mg·L −1 , the\nprediction equation gave the smallest deviation with the experimental\nEDTA-Fe II regeneration rates. Moreover, when the concentration\nof the electron donor exceeded a certain value, the EDTA-Fe II regeneration rate generally stayed stable. Therefore, considering\nthe reduction efficiency and long-term stability of the reactor, this\nkind of electron donor combination was thought to be optimal and more\nbeneficial for maintaining stable operation of the system. Figure 4 Prediction curve of the EDTA-Fe II regeneration rate\n([EDTA-Fe II -NO] = 9 mmol, [EDTA-Fe III ] = 9 mmol, U = 12 V, G = 1 g·L −1 , liquid rate = 1.2 L·min −1 , pH = 6.7–6.9). Table 2 Factorial Analysis of Different Electron\nDonor Combinations batch electron donor\nglucose (mg·L −1 ) current (mA) EDTA-Fe II regeneration rate (mmol·h −1 ) experimental value predictive value 1 200 10 0.96 1.02 2 200 20 1.02 0.98 3 200 60 1.11 0.96 4 300 10 1.07 1.09 5 300 20 1.11 1.06 6 300 60 1.10 1.25 7 500 10 1.10 1.04 8 500 20 1.31 1.22 9 500 60 1.28 1.21 10 1000 10 1.71 1.67 11 1000 20 1.71 1.70 12 1000 60 1.67 1.65 2.3 Reduction of EDTA-Fe II -NO and EDTA-Fe III under Different Substrate Concentration Ratios It has been confirmed that EDTA-Fe II -NO and EDTA-Fe III can inhibit one another during the reduction of either\nsubstrate. 21 Therefore, to ensure the stable,\nlong-term operation of the electrobiofilm reactor, the EDTA-Fe II -NO and EDTA-Fe III reduction efficiencies were\nstudied under operating conditions that covered a range of different\nconcentrations. The reactor was operated for no less than 14 days\nunder each concentration ratio. As shown in Figure 5 , the reduction efficiencies\nof both EDTA-Fe II -NO and EDTA-Fe III became optimal\nunder a concentration ratio of 1:1. Moreover, the reactor remained\nsteady during the operating conditions presented in Figure 5 . There was no decrease in\nthe reduction efficiencies of EDTA-Fe II -NO and EDTA-Fe III at high EDTA-Fe II -NO concentrations (i.e., the\nconcentration ratio of 3:1). Both EDTA-Fe II -NO and EDTA-Fe III were fully reduced after 10 h of daily operation. However,\nthe reduction efficiencies of EDTA-Fe II -NO appeared to\nbe lower (i.e., 70%), under a concentration ratio of 1:5. A high concentration\nof EDTA-Fe III can inhibit the activities of microorganisms\nin the electrobiofilm, as the actual reduction of EDTA-Fe II -NO at an initial EDTA-Fe II -NO concentration of approximately\n3–4 mmol·L −1 was much lower than that\nunder other substrate concentration ratios. However, the EDTA-Fe II -NO reduction efficiency improved as the EDTA-Fe III concentration decreased. Therefore, microbial activity is considered\ncritical during the reduction of EDTA-Fe II -NO, in that\nEDTA-Fe III can inhibit the activity of the EDTA-Fe II -NO-reducing bacteria. This indicates that high concentration\nof chelated NO would have toxic effects on substrate-reducing bacteria,\nthus inhibiting the reduction of EDTA-Fe III . Previous studies\nby our research team found that EDTA-Fe II -NO was easier\nto be reduced than EDTA-Fe III in an electrode biofilm reactor\nunder the same experimental conditions, and the two substrates have\na competitive relationship during the reduction process. In the presence\nof EDTA-Fe II -NO, the reduction rate of EDTA-Fe III was initially inhibited, especially when the concentration of EDTA-Fe II -NO was 6 mmol·L −1 , and the reduction\nof EDTA-Fe III was almost quit in the first 3 h. Figure 5 Reduction efficiencies of EDTA-Fe II -NO and EDTA-Fe III under different substrate ratios ([EDTA-Fe II -NO] = 9 mmol, [EDTA-Fe III ] = 9 mmol I = 20 mA, U =12 V, liquid rate = 1.2 L·min −1 , pH = 6.7–6.9) (solid blue box: EDTA-Fe II -NO; solid red box: EDTA-Fe III ; solid green circle:\nEDTA-Fe II -NO under 1:1; and solid black box: EDTA-Fe III under 1:1). 2.4 Microbial Community Analysis The\nα-diversity can reflect the number of species in microbial communities,\nwhile the species abundance and diversity of communities can be evaluated\nthrough a series of statistical analysis of molecular biological indices. 43 The coverage index indicates the extent to which\nthe coverage of various sample libraries reflects the reliability\nof the sequencing results. 34 As shown in Table 3 , the coverage value\nof the samples under all experimental conditions is 1, implying that\nthe α-diversity index is reliable for sequencing and the samples\nwere all well tested. Table 3 Statistics of the α-Diversity\nIndex sample Shannon index ACE index chao1 index coverage biofilm formation 3.341 35.243 34.5 1 1:5 3.623 35 35 1 1:3 3.459 19 19 1 1:1 3.5 44 44 1 3:1 2.769 48.601 47.333 1 abnormal operation 3.208 42 42 1 2.4.1 Microbial Community during Biofilm Formation As mentioned above, the biofilm growth on the cathode was observed\nto be dense. Moreover, higher amounts of cocci than bacilli grew,\nand this was captured in FESEM images using a magnification of 5000×.\nTo further investigate the distribution of bacteria, the samples from\nthe electrobiofilm were analyzed by 16S rDNA high-throughput sequencing. Ten of the most abundant microbial species after electrobiofilm\nformation are shown in Figure 6 . Raoultella occurred at abundances of 27%\nand has a certain ability in terms of denitrification. 35 Dysgonomonas is a type of autotrophic\nEDTA-Fe III -reducing bacteria, 36 while Ochrobactrum can catalyze nitrate reduction\nby electrons. 37 Chryseobacterium occurred at relatively low abundances but may also be involved in\nEDTA-Fe III reduction. 32 Raineyella and Enterobacteriaceae are also\ndenitrifying bacterial genera. 33 The mature\nmicrobial community in the electrobiofilm changed greatly from that\nof the inoculated sludge from the wastewater treatment plant under\nan anaerobic environment in which the growth of autotrophic or heterotrophic\nanaerobic denitrifying bacteria was promoted. However, further studies\nare required to clarify whether the bacteria mentioned above can reduce\nchelated NO (EDTA-Fe II -NO). Nonetheless, microbial species\ndiversity greatly benefits the long-term stable operation of the reactor. 24 Figure 6 Microbial community under the conditions of biofilm formation. 2.4.2 Microbial Community under Different Substrate\nConcentration Ratios Variations in the substrate concentrations\nhad a great impact on the distribution of bacteria in the reactor.\nTherefore, the characteristics of the microbial communities in the\nelectrobiofilm under different substrate concentrations were investigated\nusing the high-throughput sequencing, as shown in Figure 7 . First, the EDTA-Fe III reduction efficiency reached\n95% under EDTA-Fe II -NO and EDTA-Fe III ratios\nof 1:5, 1:3, and 1:1, as shown in Figure 5 . Both Dysgonomonas and Chryseobacterium were present under these ratios and were\nalso predominant with abundances ranging from 10 to 30%, respectively.\nBoth of these bacteria can reduce EDTA-Fe III. 36 , 38 − 41 The reduction efficiency of EDTA-Fe II -NO exceeded 80%\nin ratios of 1:3, 1:1, and 3:1, and the reduction efficiency of EDTA-Fe III decreased to 80%, which is almost equal to that of EDTA-Fe II -NO under the ratio of 1:3. Therefore, the two types of reducing\nbacteria appeared to exhibit similar competitiveness toward the electron\ndonors at this ratio, resulting in a decrease in the abundance of\nthe microbial community. This was also indicated by the ACE and Chao1\nindex results presented in Table 3 . Falsochrobactrum , Ochrobactrum , and Raineyella were presumed to be the dominant\ndenitrifying-bacteria genera during the EDTA-Fe III reduction\nprocess. 33 , 34 , 37 Alicycliphilus is a genus of denitrifying bacteria that can reduce NO 2 – to N 2 , 35 35 and its abundance increased apparently\nwith increases in the EDTA-Fe II -NO concentration. It is\ninferred that Alicycliphilus is the main denitrifying\nbacteria for chelated-NO reduction. It has been confirmed that a high\nconcentration of NO has a toxic effect not only on EDTA-Fe III -reducing bacteria but also on EDTA-Fe II -NO-reducing bacteria. 35 When the ratio of EDTA-Fe II -NO and\nEDTA-Fe III was 3:1, the abundance of Enterobacteriaceae , a genus of denitrifying bacteria, increased notably to approximately\n28% and was considered to be one of the predominant bacteria for EDTA-Fe II -NO reduction. Additionally, a new genus of denitrifying\nbacteria, i.e., Raoultella , was observed in the electrobiofilm\nand accounted for approximately 30% of the community. 35 In summary, high microbial diversity and stable reactor\noperation could be achieved, when the concentrations of EDTA-Fe II -NO and EDTA-Fe III were similar. The microbial\ncommunities cultivated in the electrobiofilm reactor studied here\ndiffered significantly from those of the enhanced CABR system studied\nby Li et al., 24 and no autotrophic bacteria\nwere observed without current in the CABR system as a carbon source.\nThis is mainly because the hydrogen produced by cathode electrons\ncould be utilized by microorganisms in situ and promoted the growth\nof autotrophic reducing bacteria in the presence of an external current\n( Figure 8 ). Figure 7 Microbial communities at different ratios of the substrate. Figure 8 Microbial community of unsteady operation. 2.4.3 Evaluation of the Microbial Community under\nAbnormal Operation Conditions To study the shock-loading\nresistance of the reactor and its corresponding microbial diversity,\nthe initial concentration of EDTA-Fe II -NO was increased\nto 6 mmol, while the EDTA-Fe III concentration remained\nat 12 mmol. After 21 days of operation, the EDTA-Fe II -NO\nand EDTA-Fe III reduction efficiencies decreased, as shown\nin Figure 9 . The reduction\nof EDTA-Fe II -NO decreased by approximately 15%, while that\nof EDTA-Fe III decreased slightly by approximately 8%. At\nthis time, the biofilm on the cathode differed significantly from\nthat of the reactor during stable operation, as is apparent from the\nFESEM images shown in Figure 10 a,b, respectively. By comparing the microbial distribution\nat the same magnification, it can be seen that the cocci were reduced\nwhile the agglomeration phenomenon intensified under low reactor efficiency.\nThe microbial community is further illuminated in Figure 8 . Figure 9 Comparison of reduction efficiencies of EDTA-Fe II -NO\nand EDTA-Fe III under different operation conditions ([EDTA-Fe II -NO] = 6 mmol, [EDTA-Fe III ] = 12 mmol, liquid\nrate = 1.2 L·min −1 , pH = 6.7–6.9, U\n= 12 V, G = 1 g·L −1 ) (solid black box: EDTA-Fe II -NO and solid red box: EDTA-Fe III ). Figure 10 FESEM images of different operation conditions of the reactor (×3000).\n(a) FESEM image of reduction efficiencies of EDTA-Fe II -NO\nand EDTA-Fe III declined. (b) FESEM image of stable operation\nof the reactor. The abundance of Chryseobacterium was significantly\nhigher than that in Figure 6 , indicating that the EDTA-Fe III -reducing bacteria\non the electrobiofilm could resist the shock loading. However, the\nabundance of Dysgonomonas decreased to 11%, indicating\nthat autotrophic EDTA-Fe III -reducing bacteria were sensitive\nto variation in shock loading. Moreover, according to Figure 7 , the presence of EDTA-Fe II -NO may have inhibited the growth of Dysgonomonas . The proportion of Ochrobactrum in the microbial\ncommunity exhibited better stability under shock loading. 37 According to Figure 7 , the reduction of EDTA-Fe II -NO\nexceeded 80% under the EDTA-Fe II -NO and EDTA-Fe III ratios of 1:3, 1:1, and 3:1, while Alicycliphilus , Enterobacteriaceae , and Raoultella were presumed to be the dominant genera involved in EDTA-Fe II -NO reduction (>25% each). However, their abundance apparently\ndecreased to 3–9% each, as shown in Figure 8 , thereby inhibiting the ability of the electrobiofilm\nreactor to reduce EDTA-Fe II -NO. By comparing the microbial communities under all of the conditions\nin this study, the ACE and Chao1 indices were found to be suitable\nat describing the amount of microbial growth 34 and gave values that were higher at substrate concentration ratios\nof 1:1 and 3:1. This indicates that denitrifying bacteria genera were\ngenerally more abundant than EDTA-Fe III -reducing bacteria\nin the electrobiofilm reactor. The distribution of species on the\nelectrobiofilm under ratios of 1:1 and 1:5 was well balanced, while\nit was not under a ratio of 3:1 and abnormal operating conditions,\naccording to the Shannon index. The growth of the same genera of denitrifying\nbacteria was relatively concentrated. Therefore, high microbial diversity\nand stable reactor operation could be achieved when the concentrations\nof EDTA-Fe II -NO and EDTA-Fe III were almost equal."
} | 7,040 |
27866628 | null | s2 | 9,360 | {
"abstract": "The goal of this work was to synthesize gold nanoparticles (AuNPs) using electrode-respiring Geobacter sulfurreducens biofilms. We found that AuNPs are generated in the extracellular matrix of Geobacter biofilms and have an average particle size of 20nm. The formation of AuNPs was verified using TEM, FTIR and EDX. We also found that the extracellular substances extracted from electrode-respiring G. sulfurreducens biofilms reduce Au"
} | 108 |
36785930 | PMC9928423 | pmc | 9,361 | {
"abstract": "Antibiotic consumption and its abuses have been historically and repeatedly pointed out as the major driver of antibiotic resistance emergence and propagation. However, several examples show that resistance may persist despite substantial reductions in antibiotic use, and that other factors are at stake. Here, we study the temporal, spatial, and ecological distribution patterns of aminoglycoside resistance, by screening more than 160,000 publicly available genomes for 27 clusters of genes encoding aminoglycoside-modifying enzymes (AME genes). We find that AME genes display a very ubiquitous pattern: about 25% of sequenced bacteria carry AME genes. These bacteria were sequenced from all the continents (except Antarctica) and terrestrial biomes, and belong to a wide number of phyla. By focusing on European countries between 1997 and 2018, we show that aminoglycoside consumption has little impact on the prevalence of AME-gene-carrying bacteria, whereas most variation in prevalence is observed among biomes. We further analyze the resemblance of resistome compositions across biomes: soil, wildlife, and human samples appear to be central to understand the exchanges of AME genes between different ecological contexts. Together, these results support the idea that interventional strategies based on reducing antibiotic use should be complemented by a stronger control of exchanges, especially between ecosystems.",
"conclusion": "Conclusion and perspectives The present study provides a broad picture of the spatial, temporal and ecological distributions of AME genes as well as their association with MGEs and reveals contrasted patterns for the different gene families. It additionally establishes that the recent temporal variations of AME bacteria in Europe are explained first by ecology, second by human exchanges and last by antibiotic consumption. This means that selection by man-made antibiotics is not the only evolutionary force explaining the frequency of aminoglycoside resistance and its variation, such that interventional strategies based on prudent uses of aminoglycosides for humans, animals, and plants are likely to be a necessary but insufficient way to control and limit the spread of aminoglycoside resistance. The importance of ecology and human exchanges in shaping the patterns of AME gene prevalence is adding to the growing body of evidences that AR depends not only on clinical therapeutic guidelines, but also on the high interconnectivity of ecosystems, both locally and globally ( Hernando-Amado et al., 2019 ). Thus, although continuing AR monitoring in clinical and farms samples is crucial, current sampling methods hugely bias genomic datasets and insufficient standardization of data limits their exploitation. Understanding the big picture on AR will require a stronger sampling effort in natural ecosystems and as we have shown that resistomes tend to cluster by ecology rather than by geography, emphasis should be put on monitoring the resistome of all biomes with equal intensity. Although the conclusions of this study cannot be extended to AR genes other than AMEs, the methods implemented could easily be applied to other AR gene families (especially modifying enzymes). As a complement, this could help us clarifying the overview of the forces that shape resistance prevalences across large temporal and spatial scales. Finally, this study highlights that AME genes are frequently associated with MGE but also shows that this level of association strongly varies between gene families. It additionally reveals the role of MGE in the generation of within-genome duplications and even more importantly in functional diversification and resistance spectrum broadening. MGEs are known to be vehicles of HGT and are likely to participate in the spread of AME genes but the strong correlation between resistome composition and species composition established here within biome groups suggests that AME genes are also spreading by AME bacteria between biomes. The relative contributions of HGT and bacteria migration to AR propagation as well as the factors that shape and orient them should be investigated on large, integrative datasets in order to understand the antibiotic resistance traffic rules.",
"introduction": "Introduction Antibiotic resistance (AR) is a persistent global public health problem that has increased over the last decades, with resistances spreading faster and faster after antibiotic introduction in clinical use ( Witzany et al., 2020 ). Although first concerns about infections by antibiotic resistant bacteria (AR bacteria) were formulated in the 1940s, the discovery and development of new antibiotics allowed for treatment substitutions ( Podolsky, 2018 ) during the first decades of antibiotic use. The discovery that AR was frequently acquired by horizontal gene transfer (HGT, Watanabe, 1963 ) and the subsequent emergence of multi-resistant strains led the international health agencies to start raising the issue, at the end of the 1970s, that antibiotic resistance propagation was threatening to end the antibiotic golden area and jeopardize the huge progress made in the treatment of bacterial infectious diseases. In parallel, the discovery and design of new antibiotics had become more and more difficult ( Livermore et al., 2011 ). This trend persisted despite the bulk of information provided by recent advances in genomics ( Ribeiro da Cunha et al., 2019 ), thus decreasing the hope for potential treatment substitutions. A recent review ( Antimicrobial Resistance Collaborators, 2022 ) evaluated that in 2019, around 1.25 million deaths were directly attributable to bacterial antimicrobial resistance. Antibiotic consumption and its abuses have been historically and repeatedly pointed out as the major cause of AR genes propagation ( O’Neill, 2014 ; Podolsky, 2018 ; Schrijver et al., 2018 ): the frequency of AR increases in bacteria communities under the selective pressure of antibiotics. To fight this threat, most health agencies are thus focusing their policies on sanitation and mostly on a more reasonable use of antibiotics ( World Health Organization, 2015 ). This perception of the factors driving AR spread and its associated policies still prevail nowadays: for example, a meta-analysis over 243 studies found a positive correlation between antibiotic consumption and presence of AR ( Bell et al., 2014 ). However, even if antibiotic consumption decreases in most developed countries, AR does not always follow the same path: for example, a drastic reduction in sulfonamide consumption in the United Kingdom during the 1990s was not followed by a decrease in prevalence of sulfonamide-resistant Escherichia coli ( Enne et al., 2004 ). Antibiotic consumption therefore does not seem to be the only factor maintaining AR in pathogenic bacteria communities. Indeed, although the impact of human activities on the circulation and spread of AR genes is now well documented through the accumulation of specific examples, integrative studies identifying large-scale trends are lacking and this absence of general view has been pinpointed as a gap in our knowledge of drivers of antimicrobial resistance ( Holmes et al., 2016 ). Another strongly overlooked factor is ecology, and a growing number of studies has called for a more comprehensive analysis of AR outside of farms and hospitals (see e.g Bengtsson-Palme et al., 2018 ). Though data in natural ecosystems remain scarce, AR genes have probably always been natural members of the gene pools of environmental microbial communities ( D’Costa et al., 2011 ). Natural ecosystems may contribute to the spread of AR, both as sources and as vectors of propagation ( Bengtsson-Palme et al., 2018 ; Berendonk et al., 2015 ; Marti et al., 2014 ). This realization, combined with an increasing access to genomic data, led to bioinformatic studies where the goal was to extend our understanding of AR in ecological contexts that are often overlooked: for example β-lactam resistance in dairy industry ( Pitta et al., 2016 ), in slaughterhouses ( Lavilla Lerma et al., 2014 ), in wastewater treatment plants ( Karkman et al., 2016 ), or in natural fresh water ( Czekalski et al., 2015 ). However, even if these studies complement our knowledge on the presence, the frequency, the nature, and the circulation of AR in poorly documented environments (see e.g. Cuadrat et al., 2020 ; Zhang et al., 2020 ), descriptions of global distribution patterns and analyses of factors underlying them remain scarce. In this study, we investigate the relative importance of environmental and genomic factors in shaping the routes of antibiotic spread on a large scale, focusing on aminoglycoside resistance. Aminoglycosides are antibiotics that bind to the 30 S subunit of prokaryotic ribosomes and thus inhibit protein synthesis ( Mingeot-Leclercq et al., 1999 ). This antibiotic class was first introduced in 1944 with the successful use of streptomycin against tuberculosis, and the first antibiotic able to fight Gram-negative bacteria. Several years later, other drugs produced by Streptomyces spp . were discovered (kanamycin, spectinomycin, tobramycin, neomycin, apramycin) and introduced in clinics. They were followed in the 1970s by a series of new isolates or derivatives synthesized compounds (amikacin, netilmicin, isepamicin, dibekacin, arbekacin, see van Hoek et al., 2011 ). However, the emergence of resistant strains during the following years, combined with the requirement of administration by injection and possible nephro- and ototoxicity has reduced the use of aminoglycosides in therapies ( Murray and Murray, 1991 ). Nowadays, these drugs are only used in humans as a second-line or last-resort treatment for Gram-negative bacteria ( Garneau-Tsodikova and Labby, 2016 ). However, they remain frequently used in agriculture and veterinary medicine ( European Medicines Agency, 2019 ). As an example, in 2015, aminoglycosides still represented 3.5% of the total sales of antimicrobials for farm animals, and they are still important drugs for some pathologies, for example post-weaning diarrhoea in pigs ( van Duijkeren et al., 2019 ). Thus, aminoglycoside resistance is still highly prevalent among farm animals: in Denmark, 85% of Salmonella spp. serovar Kentucky isolated in turkeys are not susceptible to aminoglycosides, and in Italy, up to 67% of E. coli samples are resistant to at least one aminoglycoside drug ( van Duijkeren et al., 2019 ). Thus, although these drugs are no longer widely prescribed for patients, aminoglycoside resistance is still a major threat, at least toward food production and the treatment of patients infected by multi-resistant bacteria. Moreover, since most aminoglycosides originate from soil-dwelling bacteria, there are at least three ecological contexts (hospitals, farms, and soil) in which aminoglycoside resistance can evolve and spread to cause public health issues. The three main mechanisms of aminoglycoside resistance are ( Garneau-Tsodikova and Labby, 2016 ): (i) decrease in drug uptake (through modifications of membrane permeability or of the peri-membrane ion gradient) and/or increase in drug efflux (through efflux pump activation); (ii) drug inactivating enzymes; and (iii) modification of the drug target, for example by point mutations in the genes coding for ribosomal small subunit ( Finken et al., 1993 ; Sander et al., 1995 ; Toivonen et al., 1999 ). Aminoglycoside-modifying enzymes (AMEs) are a class of inactivating enzymes that catalyze the transfer of chemical groups on specific residues of the aminoglycoside molecules, leading to a modified drug which poorly binds to its target ( Jana and Deb, 2006 ). AMEs represent the most common mechanism of aminoglycoside resistance in clinical isolates and are well characterized biochemically ( Ramirez and Tolmasky, 2010 ). The classical nomenclature of AMEs is based on the group they transfer (i.e. acetyltransferases, AACs; nucleotidyltransferases, ANTs; and phosphotransferases, APHs), on the residue modified, and on the resistance profile they confer ( Ramirez and Tolmasky, 2010 ). However, AMEs emerged several times during evolution (see e.g. Salipante and Hall, 2003 for AAC(6’) enzymes), so biochemical nomenclatures do not reflect the evolutionary history of any class of AMEs. Finally, many AME genes are carried by mobile genetic elements, which give them the potential to be transmitted both vertically and horizontally ( Davies, 1983 ). They represent today a major threat for the treatment of multidrug-resistant bacteria, notably Mycobacterium tuberculosis ( Labby and Garneau-Tsodikova, 2013 ). Through a computational approach, more than 160,000 publicly available genomes were screened to identify the presence of AME-encoding genes (AME genes) across the phylogeny of Eubacteria. The present study intended (i) to describe the genomic, geographical and ecological distribution of AME genes; and based on these data, (ii) to quantify the relative contribution of several key factors (geography, ecology, genomic context, human activities) potentially driving the spread of AME genes.",
"discussion": "Discussion AME genes show a ubiquitous presence in sequenced bacteria genomes In this study, we defined and screened 27 clusters of genes encoding resistance to aminoglycosides across more than 160,000 genomes spread across phylogeny, ecology, geography, and time. About one quarter of these bacteria were found to carry a gene known to provide resistance to aminoglycosides. This very high prevalence of aminoglycoside resistance is likely biased by the much higher availability of genomic data for human, clinical, and agriculture isolates. Yet, the lowest prevalence found in a biome was actually 9%, which is consistent with previous studies on genomic ( Pal et al., 2015 ; Zhang et al., 2020 ) and metagenomic ( Zhang et al., 2020 ) datasets sampled from multiple biomes. Around 40% of the AR genes found were potentially mobile, that is associated with plasmids, integrons, prophages, ICEs or transposons. A previous study reported <20% of potentially mobile AR genes: 29.4% for AR genes conferring multidrug resistance, 15.5% for β-lactam resistance, and 10.5% for aminoglycoside resistance ( Zhang et al., 2020 ), but this study only focused on plasmids and integrons. AME bacteria are widely spread over space, time, and ecology. We were able to detect AME genes in all the biomes considered, and all over the world. The prevalence of AME bacteria ranged from 64% in Turkey to 0% in Arctic and Antarctic regions. AME bacteria were found as early as 1905, that is nearly 40 years before the first aminoglycoside was isolated and identified as a potential antibacterial therapeutic agent. This is not a surprise as aminoglycosides are naturally produced by bacteria from the Streptomyces and Micromonospora genera ( Durand et al., 2019 ) and the selection pressure for the evolution of resistance genes existed long before the clinical use of aminoglycosides, as shown for other antibiotic resistance genes ( D’Costa et al., 2011 ). AME bacteria prevalence strongly increased between the 1940s and the 1980s, and has plateaued around 30% since the 1990s ( Figure 4B ). The phase of increasing prevalence can easily be explained by the discovery and marketing for clinical use of most aminoglycosides. Their massive use must have strongly selected for the emergence and spread of AME genes. However, AG consumption has stabilized and/or decreased (at least since 1997 in Europe), so the pre-existing selection pressures have remained stable since the 1990s. Besides, most CHGs coexist over long time periods ( Figure 4C ), but they are unevenly distributed across space ( Figure 2B ). This apparent coexistence might therefore result from the combination of different local dynamics ( Figure 4—figure supplement 1 ). Interestingly, two very distant regions can display the same time trends: for example, the replacement of AACf1 by AACc and AACe1 in both Europe and Southern Asia and Oceania since the 1990s, or a surge in the frequency of ANTb in both Europe and North America during the 2000s. This suggests that local distributions of AME genes are driven by local parameters, but that local resistomes are also connected at the global scale. Finally, the ecological and phylogenetic distribution of AME genes varies to a great extent, from ubiquitous CHGs to some being restricted to a few biomes and species. Ubiquitous CHGs were first sampled earlier than the others, and the extent of ecological and phylogenetic spread for each CHG is positively correlated to the time since its first sampling occurrence. It is however impossible to determine whether a broader spread made it more likely to be detected earlier or an early emergence gave a CHG more time to spread across several biomes and phyla. Antibiotic resistance prevalence is shaped not only by antibiotic use but also by ecology and human exchanges Reduction of antibiotic consumption and fighting antibiotic pollution have historically been the most common public health recommendation to control or reduce the spread of antibiotic resistance (see e.g. Mölstad et al., 2008 ; Sabuncu et al., 2009 ). The predominance of this recommendation is driven by the idea, in evolutionary epidemiology, that selection is the major determinant in the emergence, accumulation, and propagation of resistance ( Blanquart et al., 2017 ; Spicknall et al., 2013 ). This idea is supported by correlative studies linking antibiotic consumption levels and resistance prevalence, mainly in environments where antibiotic consumption is high such as hospitals or farms (e.g. Goossens et al., 2005 ). In this study, one of the goals was to ask whether the impact of aminoglycoside use on aminoglycoside resistance prevalence could be inferred on a very broad geographical and temporal scale and also across biomes. The local concentrations of aminoglycosides are likely to strongly vary between biomes, as antibiotics are mainly used in hospitals, outpatients, farms and agrosystems and antibiotics are found as pollutant in other biomes. Data on the amount of antibiotics prescribed in human and veterinary clinics are generally available but antibiotic pollution data are globally sparse (although local databases have been set up in recent years, see Umweltbundesamt, 2016 ) and were not available on a sufficiently broad geographical and temporal scale to be integrated in our analysis. Thus, here we considered that aminoglycoside concentration is sufficiently correlated between biomes (as suggested for other antibiotics by e.g. Li et al., 2016 ), since a large part of the antibiotics can pollute other biomes ( Danner et al., 2019 ; Gothwal and Shashidhar, 2015 ; Kümmerer and Henninger, 2003 ; Tello et al., 2012 ), so that variation in aminoglycoside prescriptions have the same effects throughout the biomes. However, one could alternatively consider that aminoglycoside consumption data are only reliable in the biomes where they are directly used, because aminoglycoside concentrations differ between hospitals and farms on the one hand, and soil, freshwater, etc. on the other hand. Hence we performed another model selection in this case to control for this potential effect, by only including biomes where antibiotics are expected to exert a strong selection pressure that is clinical, human, and farms samples (for a total of 15 CHGs, see Supplementary file 4 ). The selected models accounted on average for a reduction of R 2 adj =28.5% of deviance. On this subset, even though a lower number of biomes was considered, the contribution of aminoglycoside consumption to the prevalence of AME genes is small in comparison to differences between biomes: ecology was kept as an explanatory factor for 15 CHGs and had an average contribution of 18.8% to R 2 adj , whereas aminoglycoside consumption and its interaction with ecology explained the prevalence of respectively 9 and 4 CHGs (with average contributions of respectively 2.1% and 1.0% to R 2 adj ). Thus, independently of the biomes considered, our results show that between 1997 and 2018 in Europe, aminoglycoside consumption is not the main factor explaining AME genes prevalence patterns. For all CHGs studied, aminoglycoside consumption was only a minor explanatory factor for the distribution of AR bacteria, with few positive effects and mainly non-directional effects on the probability to sample AR bacteria. There are even in this dataset examples of low ARB prevalence occurring in countries with high antibiotics consumption: for example over the period, the average prevalence of bacteria carrying AACf1 was 1.5% in France and 18.1% in Austria, whereas these two countries consumed on average respectively 0.22 and 0.02 g per inhabitant/human-equivalent of aminoglycosides each year. Our results clearly contrast with the previously established correlations between antibiotic consumption and antibiotic resistance prevalence as well as the obvious selective advantage conferred by resistance genes across large range of antibiotic concentrations ( Gullberg et al., 2011 ). One reason could be that the period studied (1997–2018) is likely to be a post-emergence period during which aminoglycoside consumption was on average stable in Europe. The relative weight of selection as evolutionary force driving the resistance gene frequency variation is likely to be lower than in the previous period. Additionally, several mechanisms can allow resistant strains to thrive even under decreasing antibiotic selection pressure. (i) Selective pressures on AR are multiple and man-made antimicrobial agents are just one of them. Many AMEs might have evolved as adaptations to antibiotics produced in natural communities, or as exaptations of metabolic enzymes selected for other functions (e.g. AACs from enzymes involved in the acetylation of ribosomal proteins, Rather, 1998 ). (ii) AME genes can be maintained in populations by hitchhiking on other genes (e.g. resistance to heavy metals, biocides) carried on the same MGE ( Zhai et al., 2016 ). (iii) Carrying an AR gene is costly only if the gene is expressed, but AR genes can be silenced in nonselective conditions and be reactivated in selective conditions only ( Kime et al., 2019 ). The results of our integrated study strongly suggest that reducing aminoglycoside consumption is a necessary but not a sufficient measure to limit the propagation of antibiotic resistances and that complementary measures based on the reduction of other selection pressures (such as heavy metal pollution, Seiler and Berendonk, 2012 ) maintaining AR gene frequency should be implemented. Trade and migration matter more than antibiotic use to explain antibiotic resistance prevalences In our dataset, human exchanges explain a significant part of the variation in the prevalence of AME bacteria for a higher number of CHGs than antibiotics consumption does. Over the 1997–2018 period, we observed this trend within Europe, and human exchanges also have an important effect worldwide ( Figure 5—figure supplement 1 and Supplementary file 2 ), although at this larger scale it was impossible to compare their effects to the effect of antibiotic consumption. Numerous examples of direct impact of human activities, outside of antibiotic use, on the emergence, retention, and propagation of AR have been documented: AR bacteria can be carried over continents by plant ( Zurfluh et al., 2015 ) and animal products ( Eltai et al., 2020 ; Le Hello et al., 2011 ), exchanged through international trade, as well as immigrants ( Nellums et al., 2018 ) and travelers in general (e.g. Lepelletier et al., 2011 ). The AR genes carried by imported strains may then spread to local strains and species by horizontal transfer, thus enriching the local resistome or increasing the local frequency of resistant genes. In the case of AME genes, the decomposition of human exchanges in different good categories revealed that no category, either specific (e.g. animal feedstuff imports) or general (e.g. animal products imports), contributes to explain the distribution of all CHGs. This could be due to different CHGs having different geographical origins, to different CHGs being carried by bacteria with different ecological niches, or to contingency in the different transmission pathways. Yet, these factors were identified by correlation only, so whether these imports are transmission pathways for these CHGs cannot be assessed without direct sampling of such products. Once this assessment will be done, the similarity in resistome composition between the exporting and importing countries could technically become a criterion for the choice of the origin of some imported products. The importance of migration of AR genes, through trade and human travels, as a factor explaining the variation in AR gene frequency, revealed by our analysis suggests that reducing the import of AR bacteria would be an efficient way of limiting AME gene propagation. Procedures of AR gene monitoring in traded goods have been set up but are probably insufficient. For example, as reported for the European Union ( Schrijver et al., 2018 ), the modalities for detection and characterization of AR genes in meat are not harmonized between countries, and only focus on clinically relevant species, thereby ignoring the risk of horizontal transmission to and from other unmonitored strains. Ecology matters most to explain antibiotic resistance prevalence Not only did we find a wide diversity of AME genes outside of hospitals and farms, where antibiotics are the most consumed (see Figure 3 ), but differences between biomes explained most of the variance when the frequency of AR bacteria was modeled over time, space, and ecology. Across ecosystems (defined as the intersection of a biome and a geographical unit), we found that the resemblance between resistomes depended more on the biomes in which they were sampled, than on their geographical location. This is consistent with previous studies: functional metagenomic selections show drastic differences in the resistance profiles between soil and human gut microbiota isolates ( Gibson et al., 2014 ), and <10% of AR genes sampled in wastewater treatment plants can also be found in other environments ( Munck et al., 2015 ). Yet, consistent with Forsberg et al., 2012 who found AR genes shared with 100% identity between soil bacteria and human pathogens, many AR genes were still shared between biomes, including between natural and anthropized biomes. Additionally, the study of certain biomes seems crucial to understand the spread of AR. Indeed, in the network of resistomes, ‘fauna and flora’ and ‘soil’ have the highest measures of centrality, which suggests that these two biomes are involved in most events of AME gene transfer across ecology (either as an origin or as a destination). This could be explained by both antibiotics and AR genes transferred from humans and anthropized habitat to natural ecosystems. Natural ecosystems have actually been suggested to constitute reservoirs for AR genes originating from hospitals (see e.g. Baquero et al., 2009 ; Tripathi and Cytryn, 2017 ). Also, most natural aminoglycosides are produced by two bacteria genera, Streptomyces and Micromonospora ( Durand et al., 2019 ), that live mostly in soil and decaying vegetation. It is therefore possible that many of the CHGs first evolved in soil and flora bacteria as defense mechanism against the chemical warfare of these two genera and remained at high frequency because of continuous selection pressure. Along the same lines, we found that, in our dataset, even though 19 times more genomes were sampled in hospitals than in soil, only 2.5 times more CHGs had their first documented occurrence in clinical samples than in soil. If not all CHGs emerged in soil and wildlife, these two biomes appear at least as a hub for the circulation of AR genes. Our biome network approach identified three major modules of resistome: (1) hospitals, humans, farms, and agrosystems; (2) soil, human habitat, waste, and freshwater; and (3) domestic and wild animals and plants. Previous metagenomic studies using abundance-based metrics found that resistome composition is correlated to bacteria community composition: either globally ( Pehrsson et al., 2016 ), or in specific biomes: soil ( Forsberg et al., 2014 ), wastewater treatment plants ( Ju et al., 2019 ), human feces ( Pehrsson et al., 2016 ). In agreement with these studies, while using presence-absence-based indices, we found substantial correlation between resistome similarity and phylogenetic beta-diversity in the first and the second modules. This correlation indicates that resistome similarity is at least partially due to species composition similarity, and so to the exchange of AR bacteria between biomes. Consequently, our results also indicate that the propagation of AME genes could result from the dissemination of AR bacteria: between ecosystems. The propagation could occur by abiotic dissemination ( Allen et al., 2010 ), fecal transmission ( Karkman et al., 2019 ), transmission by wastewater ( Vaz-Moreira et al., 2014 ), or transmission through the food chain ( Founou et al., 2016 ). Yet, the correlation between resistome and species composition was not found in the third module, even though we controlled for the lower sampling in this module. This surprising result could actually be due to a higher environmental fragmentation in these biomes (see e.g. Lu et al., 2014 for the gut microbiota of wild flying squirrels), which could result in biased sampling. Moreover, previous results showed that higher spatial structures increase the likelihood of HGT events to low-abundance strains ( Cairns et al., 2018 ), that might not be sampled and therefore decrease the likelihood to find a correlation between resistome composition and community composition. The absence of correlation between resistome and species composition leads to formulate the hypothesis that horizontal transfer might play a more important role in AR gene propagation in this module than in the other two. This hypothesis could be tested using HGT detection methods such as the one developed in Corel et al., 2018 . Combination of AME genes in a genome is shaped by diversifying selection and MGEs movements In our dataset, 40% of AME genes are carried by MGEs. MGE carriage is known to be a strong determinant of the capacity to propagate within and between genome and has been identified as key element for classifying a given antibiotic resistance in the highest risk category ( Martínez et al., 2015 ). The association with MGE frequently takes the form of an embedded structure: for example half of transposable elements carrying AR genes were also located on plasmids ( Figure 7 ). The embedded structures combine the intra- and inter-genome propagation potential. At a finer scale, the fraction of genes associated to MGE as well as the identity of the associated MGE strongly differ between CHGs, suggesting that these clusters have contrasted propagation probabilities and pathways. The MGE carriage of resistance genes also means that they can be acquired from different sources and combined within a genome. A second interesting finding is that the combination of AME genes within genomes is not random: the resistance spectrum widens at a higher rate with the number of AME genes than expected by chance, which suggests that the combination of AME genes in a genome is under selection for functional diversification. This broadening of the resistance spectrum is mainly driven by MGEs with inter-genomic mobility (i.e. plasmids, ICEs, and prophages). Finally, we also found that MGEs increase the likelihood for a genome to carry several copies of the same AR gene. It has to be noted here that the copy number accessible from whole genome sequence data is an underestimation of the actual number of copies of the gene for all genomes in which the AR gene is carried by a plasmid, because plasmids are usually present in more than one copy. The presence of several copies of a resistance gene, either because of its association with an intra-genomic MGE or its plasmid carriage not only increases the expression level of AR genes ( Depardieu et al., 2007 ; Sandegren and Andersson, 2009 ), but also allows bacteria to evolve new antibiotic resistance functions on duplicated sequences ( San Millan and MacLean, 2017 ; Sandegren and Andersson, 2009 ), thus participating in the functional diversification. Limitations The approach taken here allows to exploit a large amount of publicly available data to gain a broad scale vision of AME gene circulation and propagation. However, it suffers some drawbacks, mainly linked to the fact that it is based on available data and does not result from a dedicated sampling. On the one hand, our screening is likely to underestimate the frequency of resistance carrying genomes because some AME families probably remain unknown and aminoglycoside resistance can be conferred by other resistance mechanisms such as target change, hydrolysis, etc. ( Blair et al., 2015 ). On the other hand, resistance gene carrying genomes do not necessarily produce resistant bacteria. It is indeed nearly impossible to determine with certainty whether each of the genes identified is expressed and codes for a functional protein and thus actually confers aminoglycoside resistance. So far, the genotype-phenotype relationship is poorly understood for AR genes in general ( Hughes and Andersson, 2017 ), and particularly for AMEs: a single amino acid change is likely to change the enzyme’s target, and even to remove its resistance function ( Feldgarden et al., 2019 ; Zárate et al., 2018 ). However, our screening might still be considered as a decent approximation when working with a high spatio-temporal scale of genomic data. Other limitations come from the fact that our dataset consists of genome sequences in multiple research projects. Because of this, sampling was biased towards industrialized countries and towards phyla and biomes with clinical interest (see Figure 4—figure supplement 2 for a full overview). Thus, the genomes available are not representative of the species composition of the different biomes and some of the biomes or geographical location are over-represented. The frequencies of resistant bacteria established in this study are useful for comparisons between biomes, geographical location and time periods but cannot be taken as absolute estimates. In particular, the diversity of CHGs represented across phyla can be assumed to be explained by the differences in the number of genomes sampled for each phylum, which is itself linked to the presence of species of medical interest in some phyla. Along the same line, the absence of detected resistance in certain places in our data set is likely due to a lack of published genomic data from these places: for example, we did not detect any AME gene in genomes of bacteria from the Antarctic, when metagenomic studies have shown that AG resistance have evolved in polar communities ( Perron et al., 2015 ). However, the repetition of the analyses on several subsets of the dataset to take sampling bias into account showed that our main conclusions still hold on each of the subsets. Besides, environmental data were insufficiently standardized. Regarding sampling locations, many of them were unknown in NCBI Biosamples metadata; and others could only be determined to the precision of the country, sometimes by looking at unrelated columns. We thus chose to use the country scale as spatial grain, in order to consider as many of them as possible, but at the cost of geographical precision. Moreover, because sampling can be scarce in many geographical areas, countries outside Europe were grouped in larger entities. This choice of spatial scales, constrained by the metadata available, might have prevented to uncover dynamics that occur at finer spatial scales. In the same way, we had to rely on unstandardized data regarding the categorization of the biomes in which bacteria were sampled. Our categorization of biomes was an attempt to reproduce current categorization of metagenomes, but unlike metagenomic datasets, the ecosystems in which bacteria genomes are sampled are usually poorly described: for example the distinction between human samples and clinical samples is often very subtle. Some samples may therefore have been assigned to the wrong biome. And since sampling is greatly biased toward clinical, human, and farms samples, we chose to merge certain classical ecological categories in order to treat the widest diversity of samples. This categorization thus partly differs from other studies. Conclusion and perspectives The present study provides a broad picture of the spatial, temporal and ecological distributions of AME genes as well as their association with MGEs and reveals contrasted patterns for the different gene families. It additionally establishes that the recent temporal variations of AME bacteria in Europe are explained first by ecology, second by human exchanges and last by antibiotic consumption. This means that selection by man-made antibiotics is not the only evolutionary force explaining the frequency of aminoglycoside resistance and its variation, such that interventional strategies based on prudent uses of aminoglycosides for humans, animals, and plants are likely to be a necessary but insufficient way to control and limit the spread of aminoglycoside resistance. The importance of ecology and human exchanges in shaping the patterns of AME gene prevalence is adding to the growing body of evidences that AR depends not only on clinical therapeutic guidelines, but also on the high interconnectivity of ecosystems, both locally and globally ( Hernando-Amado et al., 2019 ). Thus, although continuing AR monitoring in clinical and farms samples is crucial, current sampling methods hugely bias genomic datasets and insufficient standardization of data limits their exploitation. Understanding the big picture on AR will require a stronger sampling effort in natural ecosystems and as we have shown that resistomes tend to cluster by ecology rather than by geography, emphasis should be put on monitoring the resistome of all biomes with equal intensity. Although the conclusions of this study cannot be extended to AR genes other than AMEs, the methods implemented could easily be applied to other AR gene families (especially modifying enzymes). As a complement, this could help us clarifying the overview of the forces that shape resistance prevalences across large temporal and spatial scales. Finally, this study highlights that AME genes are frequently associated with MGE but also shows that this level of association strongly varies between gene families. It additionally reveals the role of MGE in the generation of within-genome duplications and even more importantly in functional diversification and resistance spectrum broadening. MGEs are known to be vehicles of HGT and are likely to participate in the spread of AME genes but the strong correlation between resistome composition and species composition established here within biome groups suggests that AME genes are also spreading by AME bacteria between biomes. The relative contributions of HGT and bacteria migration to AR propagation as well as the factors that shape and orient them should be investigated on large, integrative datasets in order to understand the antibiotic resistance traffic rules."
} | 9,937 |
22124204 | null | s2 | 9,363 | {
"abstract": "We consider the stochastic patterns of a system of communicating, or coupled, self-propelled particles in the presence of noise and communication time delay. For sufficiently large environmental noise, there exists a transition between a translating state and a rotating state with stationary center of mass. Time delayed communication creates a bifurcation pattern dependent on the coupling amplitude between particles. Using a mean field model in the large number limit, we show how the complete bifurcation unfolds in the presence of communication delay and coupling amplitude. Relative to the center of mass, the patterns can then be described as transitions between translation, rotation about a stationary point, or a rotating swarm, where the center of mass undergoes a Hopf bifurcation from steady state to a limit cycle. Examples of some of the stochastic patterns will be given for large numbers of particles."
} | 229 |
21874096 | null | s2 | 9,364 | {
"abstract": "Precise flow control in microfluidic chips is important for many biochemical assays and experiments at microscale. While several technologies for controlling fluid flow have been implemented either on- or off-chip, these can provide either high-speed or high-precision control, but seldom could accomplish both at the same time. Here we describe a new on-chip, pneumatically activated flow controller that allows for fast and precise control of the flow rate through a microfluidic channel. Experimental results show that the new proportional flow controllers exhibited a response time of approximately 250 ms, while our numerical simulations suggest that faster actuation down to approximately 50 ms could be achieved with alternative actuation schemes."
} | 188 |
21390231 | PMC3048397 | pmc | 9,366 | {
"abstract": "The distribution of the number of links per species, or degree distribution, is widely used as a summary of the topology of complex networks. Degree distributions have been studied in a range of ecological networks, including both mutualistic bipartite networks of plants and pollinators or seed dispersers and antagonistic bipartite networks of plants and their consumers. The shape of a degree distribution, for example whether it follows an exponential or power-law form, is typically taken to be indicative of the processes structuring the network. The skewed degree distributions of bipartite mutualistic and antagonistic networks are usually assumed to show that ecological or co-evolutionary processes constrain the relative numbers of specialists and generalists in the network. I show that a simple null model based on the principle of maximum entropy cannot be rejected as a model for the degree distributions in most of the 115 bipartite ecological networks tested here. The model requires knowledge of the number of nodes and links in the network, but needs no other ecological information. The model cannot be rejected for 159 (69%) of the 230 degree distributions of the 115 networks tested. It performed equally well on the plant and animal degree distributions, and cannot be rejected for 81 (70%) of the 115 plant distributions and 78 (68%) of the animal distributions. There are consistent differences between the degree distributions of mutualistic and antagonistic networks, suggesting that different processes are constraining these two classes of networks. Fit to the MaxEnt null model is consistently poor among the largest mutualistic networks. Potential ecological and methodological explanations for deviations from the model suggest that spatial and temporal heterogeneity are important drivers of the structure of these large networks.",
"introduction": "Introduction Describing complex ecosystems as networks of interacting components and explaining the structure of those interaction networks is an essential part of understanding the role of biodiversity in the function and robustness of ecological communities [1] , [2] . While food webs, networks of antagonistic consumer-resource interactions, have a long history of study and are the most familiar example of ecological networks [3] , [4] , [5] , significant attention has recently been focused on networks of mutualistic interactions such as plants and their pollinators or plants and seed dispersers [6] , [7] . These networks provide a valuable overview of one type of mutualism occurring within a community and several apparently general patterns in the structure of mutualistic networks have been found [8] , [9] . Co-evolutionary processes are believed to play a strong role in shaping mutualistic communities [8] , [9] , though others have questioned whether such processes structure mutualistic networks [10] , [11] . The distribution of the number of links per species, or degree distribution, is a widely used summary of the topology of complex networks [12] that has been studied in both food webs and mutualistic networks [9] , [13] , [14] . Because of its central role in describing network topology, considerable importance has been placed on understanding the processes driving the form of the degree distribution in ecological networks [9] , [11] , [13] , [14] . Interest in the relative abundance of generalists and specialists motivated early studies of networks of mutualistic interactions [15] , [16] , and such networks were found typically to have strongly skewed degree distributions, with many species with few links and few species with many links [9] , [17] . Earlier work found that degree distributions in mutualistic networks are best-fit by a truncated power law [9] , but recent work does not support that finding [18] . Similar attempts to fit degree distributions to particular functional forms for food webs have also produced variable results [13] , [14] , [19] . The obvious difference between the observed skewed distributions and the binomial distributions of random networks [20] has driven the assumption that these skewed distributions are a result of ecological or evolutionary processes shaping species interactions [7] , [9] . From early ideas about succession [21] , [22] through more recent debates about community assembly [23] , [24] to current research into macroecological patterns [25] , the debate as to whether perceived patterns in ecosystem properties are the result of chance, biological processes or bias in the data has been an enduring theme in ecological research. The principle of maximum entropy [26] asserts that the least biased probability distribution satisfying a set of constraints is the maximum entropy distribution, and any other distribution would be assuming information not captured by the constraints. It has recently been recognized as a powerful tool in the search for explanations of ecological patterns and has been used to argue that a number of macroecological patterns can be predicted with minimal appeal to specific ecological processes [27] , [28] . Recently it was shown that a null model for degree distributions of food webs based on MaxEnt could not be rejected as a null model for 57% of the food web degree distributions studied [29] . This very simple MaxEnt model requires a minimal amount of ecological information: the number of species, the number of species with no prey (basal species) or predators (top species), and the number of links in the network. Since food webs and mutualistic networks are built primarily from antagonistic and mutualistic interactions respectively, it is interesting to consider whether the different types of interaction causes the structure of these two classes of networks to be significantly different. Mutualistic networks are bipartite networks, with interactions occurring between two groups of species, here plants and animals, but not within the groups. While food webs are not bipartite since they include taxa at many trophic levels and interactions can occur between animals, an obvious subset of a food web, the primary producers and their consumers, form a natural counterpart to the mutualistic plant-animal networks, one in which the interactions are primarily antagonistic. The different types of interaction cause different pressures on organism's behavior, and so it is reasonable to expect networks dominated by antagonistic or mutualistic interactions to have different structure. A study of 14 food webs included as part of a much larger study of mutualistic networks showed that mutualistic and antagonistic networks differed significantly in their nestedness [8] . Given that different ecological processes may shape the networks, it is possible that the degree distributions of these two different types of networks also have different forms. The goals of this work are three-fold. First, to test whether a MaxEnt model like that used to predict food web degree distributions [29] can predict the degree distributions of mutualistic networks; second, to compare the deviation of mutualistic and antagonistic networks from the MaxEnt model to better understand how the structure of these two classes of networks differs; and third, to explore how specific features of some mutualistic networks might influence their degree distributions and drive them away from the MaxEnt expectation.",
"discussion": "Discussion While degree distributions in mutualistic and antagonistic networks are strongly skewed, with many species having few connections and few species having many connections, the results here show that their shape can usually be explained by a simple statistical model and does not require a model involving specific ecological or evolutionary processes. The MaxEnt model is found to be a good model of the degree distributions of mutualistic and antagonistic networks more often than it was found to be a good model for food web degree distributions [29] , suggesting that ecological processes play a more important role in structuring multi-trophic level food webs than the bipartite networks considered here. Recently, models based on MaxEnt have also been used to explain a broad range of macroecological distributions, such as species-abundance and species area relationships [27] , [28] , [34] . Together, these findings show that a wide range of large-scale ecological patterns can be explained without turning to detailed descriptions of the ecological processes at work in the system. An earlier null model for degree distributions in mutualistic networks suggested that species' degree (number of species it interacts with) is a function of its frequency of interaction [11] . Other explanations relate species degree to specific trait combinations making certain links impossible (so-called “forbidden links”) [9] , [40] or to a combination of abundance and traits [41] . Evolutionary network models have also been explored as explanations for the structure of ecological networks and a range of degree distributions have been found [42] , [43] . These models suggest that the observed exponential-like degree distributions results from variation in the links passed from parent to child species during evolution. A recent analysis of the application of MaxEnt to species abundance distribution argues that it is common for distributions, each resulting from one or more mechanistic model, to also be found as a solution of an appropriately formulated entropy maximization problem [34] . The fact that the formulation used here is so often successful suggests that its formulation and constraints reflect simple constraints commonly operating on these systems. The existence of multiple mechanistic models giving similar degree distributions suggests that multiple mechanisms can place similar simple constraints on the degree distributions, whether through trait distributions or evolutionary processes. This in turn suggests that it will not be possible to determine which ecological or evolutionary processes are constraining the structure of mutualistic networks by studying their degree distributions alone. When deviations from the MaxEnt model do occur, it is necessary to question whether they are due to ecological processes or systematic sampling biases shaping the degree distributions. I have identified three deviations from the MaxEnt model in the degree distributions of the networks studied here. Importantly, these deviations are different in antagonistic and mutualistic networks, suggesting that different processes at work structuring networks with different types of links. First, plant distributions of the mutualistic networks are significantly better fit by the MaxEnt model than the plant distributions of the antagonistic networks. Second, plant distributions of antagonistic networks tend to be more broadly distributed than predicted by the MaxEnt model. This means that antagonistic networks generally have both more highly vulnerable plant resources and more relatively invulnerable plant resources than predicted by this simple null model. Third, there are opposite trends in the scale-dependence of the relative width of the animal distributions of mutualistic and antagonistic networks. The animal distribution of large mutualistic networks tends to be more broadly distributed than predicted by the MaxEnt model, while the animal distributions of larger antagonistic networks tend to be more narrowly distributed. Since pollinators and seed dispersers also consume the plants that they benefit reproductively, this suggests that highly generalist animals only occur if they are also conferring a reproductive benefit to their resource. In food webs, it has been suggested that generalist intermediate species are uncommon because of their destabilizing influence on the system [44] . The results presented here suggest that restricted relative generality of plant consumers is more common in larger networks. There are two sources of deviations from MaxEnt distributions in large mutualistic network animal distributions. First, the degree distribution can be strongly affected by the presence of a single highly connected species, causing a markedly high value of W 95 . Second, a larger than predicted fraction of species interacting with a single species can lead to the network having a distribution with a high value of W 95 . A detailed examination of two of these data sets helped reveal potential reasons for their broad animal distributions. A recent simulation study [45] suggests that spatial processes can have important effects on the structure of mutualistic networks, though did not specifically address their degree distributions. The simple heterogeneous-system degree distribution model suggests a biological explanation for the broad degree distributions seen in the large, low connectance pollination networks with a small number of super-generalist pollinators. Strong spatial compartmentalization within sub-networks, leading to networks that contain relatively high connectance sub-networks with MaxEnt degree distributions that are interconnected by one or a small number of highly general pollinators, could lead to the observed highly skewed distributions. The MULL network also has a large number of animals that pollinate a single plant species. Again, questions arise as to whether this phenomenon is determined by methodology or biology. It could be driven by the relative abundance of the species involved and the observation effort expended [11] . Alternatively, it might be the result of greater than expected specialization of the plant and animals in this system leading to relatively abundant but specialized species. Given the highly variable phenology of plants and the multiple seasons over which the PTND data [38] were collected, it is likely that the community is functioning as a set of sub-networks separated in time, with specialist pollinators active at different times of year or in different years, connected by common generalist pollinators that are much more regularly present. Here time rather than space is leading to heterogeneity in the community [46] , but with a similar effect on the network degree distribution. Other recent studies suggest that strong temporal heterogeneity is a common feature of pollination networks, and so the temporal sampling scheme must be considered when interpreting the relative degree of specialization among species [47] , [48] . Analysis of the deviations from the MaxEnt model in these two data sets demonstrates how the MaxEnt model can focus attention on the particular features of degree distributions which require further explanation. Here, it was found that spatial and temporal heterogeneity might play an important role in shaping the degree distributions and other features of the network's structure. This possibility was also highlighted in a number of recent studies [39] , [45] , [47] , [48] , [49] . Spatio-temporal heterogeneity is another mechanism which explains why some links cannot occur (“forbidden links”), caused by the lack of species co-occurrence at appropriate points in their life history. Forbidden links are often hypothesized to be an important driver of the structure of mutualistic networks assumed to arise from complementary traits in co-occurring species [9] , [40] – here those traits are the spatial or temporal domains in which the species occur. Some large systems are composed of loosely coupled small systems which are either, like MULL, highly spatially heterogeneous or, like PTND, temporally heterogeneous. The observed degree distribution will then depend on an observer's definition of the system's boundaries. While the MaxEnt null model is useful for understanding how ecosystem features such as spatial and temporal heterogeneity can affect network structure, methodological variability across the available data limit the ecological insight that can be drawn from analyses across a broad range of data sets. As noted in earlier studies, similar limitations driven by variability in data collection protocols still exist in the data describing antagonistic networks [29] , [50] , [51] . There is a clear need for more consistent data collection protocols and for systematic studies of the effects of variability in data gathering procedures and data collection effort on observed network structure. Despite these issues, the MaxEnt model successfully describes the degree distributions of bipartite ecological networks across a wide range of empirical data. Rather than requiring detailed understanding of the ecological or co-evolutionary processes at work in these systems, the relative abundance of specialist and generalist species in these networks can usually be explained by a simple statistical model."
} | 4,219 |
27652345 | PMC5026421 | pmc | 9,367 | {
"abstract": "Zeolite-catalyzed acylation is used for selective formation of C–C bonds and could have applications in bioenergy.",
"introduction": "INTRODUCTION Heteroaromatic functionalization via Friedel-Crafts acylation to attach a carbonyl group to the aromatic rings is important for the production of several specialty chemicals and drugs ( 1 ). Acylation is commonly carried out with acyl chlorides or anhydrides as acylating agents because of their ease of activation. Although efforts have been made to use carboxylic acids directly to minimize waste and improve efficiency, the more severe conditions that are required often lead to undesired side reactions with negative consequences for specialty chemical synthesis. Studies carried out with the direct use of carboxylic acids as acylating agents are therefore the vast minority, with a nearly nonexistent understanding of the consequential influence of water on the said reaction rates. However, to upgrade biomass streams, the direct activation of acetic acid, which is the most abundant compound present in many biomass-derived streams ( 2 , 3 ), may be a prerequisite to obtain adequate yields of high-value products. The efficient conversion of molecules present in streams derived from biomass to useful fuel and chemical precursors is a daunting task. Thermochemical routes, such as pyrolysis, torrefaction, and solvolysis that transform the polymers in lignocellulosic biomass to monomeric species yield a complex mixture of chemically incompatible compounds ( 4 – 9 ). Although mild thermal approaches, such as torrefaction, can selectively decompose hemicellulose and cellulose, this still yields a blend of furanic and carboxylic acid species that are difficult to separate ( 8 , 9 ). The acidity introduced by acetic acid facilitates the polymerization of furanic species at room temperature, creating obvious storage and transportation challenges ( 2 , 3 , 8 , 10 – 12 ). Stabilization through severe hydrodeoxygenation decomposes these acids, which consist predominantly of acetic acid, to low-value C 1 - and C 2 -containing species. Although this approach yields a stable product, the excessive requirement of hydrogen and the low yield of products that are liquid at room temperature hinder the economics of the overall process ( 13 – 15 ). Excessively high temperatures with acid catalysts, with the aim of producing deoxygenated products, suffer from similar consequences of excessive amounts of carbon wasted as coke and light compounds. One promising route to improve yields is to couple carboxylic acids, predominantly acetic acid, in the vapor phase to produce larger products. Decarboxylative ketonization is one route that has received a great deal of attention to stabilize these acids to form ketones that can undergo sequential coupling, such as aldol condensation, to generate useful compounds ( 16 – 20 ). Perhaps the most significant drawback of this reaction is the stoichiometric formation of carbon dioxide as a by-product, corresponding to 25% of the carbon in the acetic acid. An alternative approach to prestabilizing the blend of oxygenates via selective C–C coupling of acids and furanic compounds while yielding thermally stable intermediates would lead to a step change in the field. Direct dehydration of acetic acid to form acyl intermediates followed by coupling to a furanic substrate would be very appealing. Friedel-Crafts acylation has been the subject of several reviews, with the vast majority of studies concerning carboxylic acids focused on coupling with aromatic or phenolic compounds in the absence of cofed water for specialty chemical synthesis ( 1 , 21 , 22 ). Furanic species have been shown to couple with acetic anhydride under very mild anhydrous conditions ( 23 ), but direct coupling of furanic species, such as furan or 2-methylfuran (2-MF) with acetic acid has not been reported. Furanic species are known to undergo a wide range of side reactions over Brønsted acid sites, many of which can lead to uncontrolled polymerization and coke formation ( 24 ). Here, we report the direct vapor-phase coupling of acetic acid with furanic species over HZSM-5 zeolites, resulting in fuel range products without losing CO 2 in the process and increasing the overall amount of carbon in the biomass retained in transportation fuel precursors in the form of C–C bonds ( 25 , 26 ). 2-MF is chosen as a probe molecule because it can be generated from the selective C–O cleavage of furfural ( 27 ), which is abundant in biomass degradation streams. Under milder conditions than necessary for ketonization, acylation of 2-MF occurs with high selectivity under a wide range of concentrations. The kinetics of this reaction and the influence of water on the reaction rate are discussed. Temperature-programmed desorption (TPD) experiments and density functional theory (DFT) calculations are used to explain the mechanism of this promising reaction, highlighting the important role of confinement on the overall reaction and the acyl formation step. Finally, we demonstrate the important influence of MF on the stabilization of the charge of the acylium ion at the transition state, which facilitates C–C bond formation. These results have a broader influence on C–C coupling through Friedel-Crafts reactions over zeolites.",
"discussion": "DISCUSSION The results presented here illustrate the stable and selective C–C coupling of two biomass-derived species (acetic acid and 2-MF) that lead to polymerization and coke formation when introduced alone to Brønsted zeolite catalysts ( 28 , 34 ). This is a promising new route for the conversion of biomass-derived streams to higher-value fuels and chemicals without sacrificing carbon as CO 2 in the process. Alternative approaches to upgrading acetic acid, such as ketonization, exhibit a larger apparent barrier for C–C coupling. The high abundance of surface acyl groups present when acetic acid is introduced alone to a zeolite is in equilibrium with ketene species, leading to uncontrollable coke formation and catalyst deactivation ( 28 ). Furanic species are known to polymerize over Brønsted acid sites even under ambient conditions ( 34 ). In the scenario presented here, the low activation barrier for 2-MF coupling to acyl species allows for the stable production of C–C bonds while avoiding these undesired side reactions. Flow reaction studies show that the direct acylation of 2-MF with acetic acid over HZSM-5 is a relatively stable reaction, to produce predominantly 2-acetyl-5-MF. Acetone is observed as a side product with insignificant concentrations at temperatures below 300°C, under the reactant concentrations used here. Water inhibits the rate but does not detrimentally influence selectivity. Furthermore, as shown with high partial pressures of acetic acid, water may improve catalyst stability at higher temperatures ( 28 ). Streams consisting primarily of acetic acid, furfural, and water can be produced from the torrefaction of lignocellulosic biomass ( 25 ), and the conversion of furfural has been shown to selectively produce furan or 2-MF over a variety of catalysts ( 27 , 35 ). It should be noted that furfural conversion has been shown to not be inhibited by the presence of acetic acid ( 36 ). This implies that this selective acylation reaction could be applied to upgrade biomass-derived streams containing water. Results in Fig. 2E emphasize the importance of topology for this reaction. HZSM-5, with a smaller pore size than Hβ zeolites, exhibits a higher TOF and improved catalyst stability for this reaction. Upon studying the acylation of phenol with zeolites, Padró and Apesteguía ( 37 ) also reported the superior performance of HZSM-5 compared to larger-pored HY zeolites. This suggests that confinement plays an important role in this reaction. Charge analysis from DFT ( Fig. 4 ) implies that the ability of the MF molecule to compensate for the charge of the acylium ion in the transition state may significantly reduce the barrier for C–C coupling. In the case of 2-MF, the reduction in the barrier is significant enough to make the rate of acid dehydration a kinetically relevant step. This could explain the stability of the catalyst under these conditions, because the rapid coupling of 2-MF with surface acyl groups will decrease the population of surface acyl species that could otherwise convert to ketenes and, subsequently, carbonaceous deposits. In addition, low partial pressures of 2-MF and a high coverage of acetic acid will likely prevent the excessive furan polymerization over Brønsted zeolites. These calculations demonstrate the importance of the ability of 2-MF to stabilize the charge of the acylium ion at the transition state, which may be generally true for other C–C coupling reactions involving a charged species."
} | 2,201 |
23874569 | PMC3706605 | pmc | 9,369 | {
"abstract": "Fungi regulate key nutrient cycling processes in many forest ecosystems, but their diversity and distribution within and across ecosystems are poorly understood. Here, we examine the spatial distribution of fungi across a boreal and tropical ecosystem, focusing on ectomycorrhizal fungi. We analyzed fungal community composition across litter (organic horizons) and underlying soil horizons (0–20 cm) using 454 pyrosequencing and clone library sequencing. In both forests, we found significant clustering of fungal communities by site and soil horizons with analogous patterns detected by both sequencing technologies. Free-living saprotrophic fungi dominated the recently-shed leaf litter and ectomycorrhizal fungi dominated the underlying soil horizons. This vertical pattern of fungal segregation has also been found in temperate and European boreal forests, suggesting that these results apply broadly to ectomycorrhizal-dominated systems, including tropical rain forests. Since ectomycorrhizal and free-living saprotrophic fungi have different influences on soil carbon and nitrogen dynamics, information on the spatial distribution of these functional groups will improve our understanding of forest nutrient cycling.",
"introduction": "Introduction Ectomycorrhizal (ECM) and saprotrophic fungi are major contributors to nutrient cycling in forest ecosystems [1] . These functional groups are globally distributed and coexist in many forest ecosystems. Approximately 6000 tree species worldwide depend on ECM fungi for nutrient acquisition [2] , and the distribution of ECM trees spans the globe ranging from northern boreal regions to tropical rain forests. Strikingly, a disproportionate number of the dominant trees in temperate, boreal and certain tropical forests form ECM associations [3] – [6] , suggesting that ECM fungi are likely responsible for a significant quantity of C, N, and P cycling worldwide. In boreal forests, ECM fungi contribute up to 86% of total plant N [7] . Saprotrophic fungi are also critical to nutrient cycling, and are the major decomposers of complex, organic molecules such as lignin. Thus, understanding how ECM and saprotrophic fungi are distributed within and across ecosystems is critical for making inferences about nutrient cycling and related ecosystem functions in forest communities. It is well established that mycorrhizal fungi interact with other soil organisms such as bacteria and invertebrates, but interactions among mycorrhizal and decomposer fungi have been more challenging to evaluate [1] , [8] . There is evidence from boreal and temperate forests that ECM and saprotrophic fungal taxa vertically segregate in soils [9] – [11] , suggesting physiological specialization of fungi on organic substrates in various levels of decay [10] . However, there have been few studies of fungal spatial dynamics in tropical ECM forests, so it is unclear if the patterns detected in boreal and temperate forests are similar to those found in the tropics. While the majority of trees in temperate and boreal forests form ECM associations, most species of trees in lowland tropical rain forests form arbuscular mycorrhizal (AM) associations. When tropical trees do form ECM symbioses, they are more likely to become locally dominant [6] or in some cases regionally dominant (e.g., the Dipterocarpaceae in Southeast Asia). At this point, we do not know if generalizations can be made about ECM forests at a global scale or if tropical ECM forests contain unique fungal communities that function differently from ECM fungi at higher latitudes. From the data that have been collected, it seems that tropical forests have lower ECM diversity than temperate and boreal ecosystems [12] , [13] , although there is clearly a gap in our knowledge and a paucity of belowground studies in tropical ECM forests. Since tropical forests harbor 40% of all terrestrial biomass and are responsible for 32% of terrestrial net primary production [14] , [15] , understanding the dynamics of fungal distribution and function in tropical forests is important for making inferences about global nutrient cycles. In this study, we used sequence-based approaches to assess the distribution of fungal taxa in a tropical forest located in central Guyana and a boreal forest in Delta Junction, Alaska. The tropical forest site contained two types of rain forest: an ectomycorrhizal monodominant forest and a non-ectomycorrhizal mixed forest [16] , [17] . Our objectives were to: 1) examine the level of taxonomic similarity in fungal community composition across the two ECM forests in different biomes, 2) compare fungal community composition across organic and mineral soil horizons within each ecosystem, and 3) determine if patterns of functional group separation across soil horizons were analogous in the boreal and tropical forest. Since tropical ECM forest dynamics have been shown to be significantly different than non-EM forests within the same biome [6] , [18] , we predicted that the boreal forest and tropical ECM forest would exhibit more similar fungal community patterns than the ECM and non-EM tropical forest.",
"discussion": "Discussion While numerous sporocarp surveys have been done in tropical forests [e.g., 45,46,47], our study provides some of the first molecular evidence that confirms biogeographical separation of fungal communities across a tropical and boreal forest, despite the occurrence of dominant trees that form ectomycorrhizae in both ecosystems. As has been found in other tropical ECM forests, the major ECM fungal lineages reflect those already known to dominate temperate and boreal ecosystems [48] , [49] . Additionally, fungal communities were unique across soil and litter horizons within the same ecosystem, possibly due to fungal specialization on substrates in differing levels of decay [50] . Clone library sequencing and pyrosequencing showed analogous results in both ecosystems indicating that these patterns are robust to sequencing technology and gene region targeted, which has been a major concern among microbial ecologists [51] . While the clone library sequencing gave more reliable taxonomic information for the environmental DNA sequences (i.e., longer sequence reads), the OTUs generated from pyrosequencing aligned to similar taxa, probably as a result of incomplete coverage of fungal reference sequences in Genbank. However, because pyrosequencing allows for greater sequencing depth (for this study samples were rarified to 1000 sequences each), we can more reliably say that we have fully characterized the fungal community of a sample. Thus, the tandem use of these technologies provides strong support for our results in terms of fungal community characterization and taxonomic placement of environmental sequences. Another result supported by both pyrosequencing and clone library sequencing was that within the tropical and boreal ECM forests, ECM fungi were not prevalent in litter horizons from the forest floor, but rather occupied lower organic and mineral soil horizons. Findings that ECM fungi were more abundant in deeper soil depths have also been observed in temperate [9] and Swedish boreal forests [11] , indicating that vertical segregation of ECM and saprotrophic fungi in soils may be a widespread phenomenon in ECM-dominated forest ecosystems. The reasons for spatial segregation of these fungal groups are likely due to the distribution of C and nutrients in litter versus soil. Since ECM fungi have decomposer abilities [52] , [53] , but are not C-limited like other saprotrophs due to their access to plant photosynthate, ECM fungi may reside below the freshly-fallen litter layer in deeper horizons to target substrates richer in other nutrients [54] , [55] . An alternative, but not mutually exclusive explanation for the predominance of ECM fungi in the soil horizons may be due to antagonistic relationships between ECM and saprotrophic fungi [56] – [58] . Since these fungal groups compete for some of the same resources, they may vertically segregate to avoid competitive exclusion [10] . Within the tropical ecosystem, pyrosequencing showed that soil fungal communities were distinct between the ECM and the diverse, non-ECM forests, indicating that at a local scale, the presence of an ECM tree can dramatically alter the general fungal community. The magnitude of differentiation in soil fungal communities across these tropical forests was almost as dramatic as the differentiation observed across biomes, and previous research in this site has shown that soil physicochemical properties are not responsible for determining these patterns [18] . Fungi detected in forest floor litter were also clustered by forest type in the tropical system, although the magnitude of difference was much less. This result may be due to the fact that the same species of non-ECM trees are present in both tropical forests [16] , [21] , so the chemical composition of the leaf litter is somewhat similar [18] . However, there is an overwhelming abundance of litter in the ECM forest from the ECM, monodominant tree ( Dicymbe corymbosa ), which may explain the differences in the litter fungal communities across these forests. In another study, a reciprocal litter decomposition experiment has shown that leaf litter of Dicymbe and non-ECM trees decomposes slower in the ECM forest relative to the non-ECM forest [18] , indicating that these differences in fungal communities may result in altered nutrient cycling. In the boreal biome, the results of this study suggest that related fungal taxa may dominate the organic layers in boreal forest soils across different systems. For example, the genus Cortinarius was the most abundant in our boreal soil samples, and this genus also dominates Swedish boreal forest soil [11] . In an earlier study, Allison et al. [59] also found that the ECM genus Cortinarius was the dominant taxon from our Alaskan study site. Future work focusing on the function of Cortinarius in decomposition would be valuable, as it is a globally distributed genus and known to occupy litter at late stages of decomposition [60] . However, other than protease ability [61] , its complete enzymatic capabilities are still unknown. We also found Cortinarius taxa in the tropical samples, although our sequence analysis indicated that they are different genotypes than the boreal taxa. Some of the ECM genera we detected from the Agaricales in the tropical samples are known to associate with the dominant ECM tree, Dicymbe corymbosa , as they have been described by mycologists working in that region [62] , [63] . However, some of the sequences we generated are likely undescribed taxa. This is probably true for the numerous Clavulina species we observed from the Cantharellales in the tropical soil, which was the second most abundant order. Clavulina diversity is known to be high in this region [64] , which reflects what we detected in our environmental pyrosequencing data. The finding that fungal communities are distinct in litter horizons also has implications for environmental sampling of fungal communities. For a comprehensive understanding of microbial community composition, sampling should incorporate both the organic and underlying soil horizons. In addition, environmental changes that affect one soil layer more than another may have disproportionate consequences for the two fungal groups. For instance, forest fires primarily burn the upper soil horizons (depending on severity), so direct effects of fire may be stronger on saprotrophic fungi than on ECM fungi. Making inferences about fungal communities from only mineral samples may, therefore, underestimate diversity and provide an incomplete picture of community composition."
} | 2,948 |
25994183 | PMC4455696 | pmc | 9,370 | {
"abstract": "Background Lignin plays an important role in plant structural support and water transport, and is considered one of the hallmarks of land plants. The recent discovery of lignin or its precursors in various algae has raised questions on the evolution of its biosynthetic pathway, which could be much more ancient than previously thought. To determine the taxonomic distribution of the lignin biosynthesis genes, we screened all publicly available genomes of algae and their closest non-photosynthetic relatives, as well as representative land plants. We also performed phylogenetic analysis of these genes to decipher the evolution and origin(s) of lignin biosynthesis. Results Enzymes involved in making p -coumaryl alcohol, the simplest lignin monomer, are found in a variety of photosynthetic eukaryotes, including diatoms, dinoflagellates, haptophytes, cryptophytes as well as green and red algae. Phylogenetic analysis of these enzymes suggests that they are ancient and spread to some secondarily photosynthetic lineages when they acquired red and/or green algal endosymbionts. In some cases, one or more of these enzymes was likely acquired through lateral gene transfer (LGT) from bacteria. Conclusions Genes associated with p -coumaryl alcohol biosynthesis are likely to have evolved long before the transition of photosynthetic eukaryotes to land. The original function of this lignin precursor is therefore unlikely to have been related to water transport. We suggest that it participates in the biological defense of some unicellular and multicellular algae. Reviewers This article was reviewed by Mark Ragan, Uri Gophna, Philippe Deschamps.",
"conclusion": "Conclusions The widespread distribution of coumaryol biosynthesis gene homologs across various eukaryotic supergroups suggests an ancient origin for this pathway. Although we cannot dismiss the possibility of its presence in an ancient eukaryotic ancestor and subsequent loss in all lineages in which it is absent, an origin in archaeplastids is more parsimonious. The ancient pathway for p -coumaryl alcohol synthesis should contain one or more gene(s) that precede 4CL, CCR and CAD, as it requires a source of p -coumaric acid, but these have yet to be discovered. Since p -coumaric acid is found in the haptophytes E. huxleyi and D. lutheri , the diatom P. tricornutum and the green alga C. vulgaris [ 27 , 28 ], and none of these lineages contains any homologs of PAL and C4H, there is little doubt in the existence of enzyme(s) with an analogous function(s). It is not implied that any organisms carrying this ancient pathway can necessarily polymerize p -coumaryl alcohol (H monolignol) to form H lignin or even synthesize the more complex G and S monolignols, but they are likely able to synthesize at least p -coumaryl alcohol. As all the secondary photosynthetic organisms investigated as well as most green and red algae are marine organisms, it is intriguing to consider an authentic marine source of monolignols or lignin. As these compounds and their degradation products are used as biomarkers to calibrate for terrestrial carbon input into marine systems [ 58 , 59 ], marine sources of monolignols or lignin therefore have the potential to redefine our understanding of the marine carbon cycle. The function of p -coumaryl alcohol in unicellular marine photosynthetic eukaryotes such as diatoms, dinoflagellates, hapthophytes, cryptophytes and some green and red algae, is unclear. The two main roles of lignins derived from p -coumaryl alcohol and other monolignols in land plants are water transport and structural support. Water transport systems are absent in unicellular algae. If the p -coumaryl alcohol likely produced by unicellular and photosynthetic eukaryotes is polymerized as lignin or lignans, it could also contribute to their structural strength, although other compounds such as silica and cellulose are already known fulfil this function in such organisms [ 60 ]. More likely functions that can be fulfilled by p -coumaryl alcohol are UV protection and microbial defence. Phenolic compounds such as p -coumaric acid and its derivatives exhibit high UV absorptivity and could potentially protect an organism against the damaging effects of sunlight [ 61 ]. The lignin biosynthetic pathway has also been implicated in the defence system of plants, as individual enzymes (e.g., CAD, CCR and CCoAMT) have been shown to defend against microbial attacks [ 8 , 9 , 62 , 63 ]. Intermediates of the lignin biosynthesis pathway have also been shown to have antimicrobial properties [ 10 , 64 , 65 ]. Such a role in host defense or UV protection may have provided selection for the early evolution of lignin biosynthetic pathway in the ocean, which was then co-opted for water transport and structural strength in land plants faced with new selective pressures of an air-land environment.",
"discussion": "Results and discussion The lignin biosynthetic pathway has a conserved and taxonomically widespread core An extensive screen for homologs of the known lignin biosynthesis genes was performed across all domains of life, with a specific focus on eukaryotes (Tables 1 and 2 ). Previous research has focused on lignin biosynthesis in Arabidopsis [ 29 - 32 ] and other model land plants [ 2 , 19 , 33 - 35 ], so only representative species of this group have been included in our search. As expected, all the lignin biosynthesis genes were found in all land plants for which genomes are available, with the exception of ferulate 5-hydroxylase (F5H), which is absent from the only bryophyte (moss) in our dataset, Physcomitrella patens [ 7 ]. Surprisingly, four gene families had a wide distribution across the various eukaryotic supergroups and were not restricted to land plants. These include: 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductase (CCR); cinnamyl alcohol dehydrogenase (CAD); and caffeoyl-CoA O -methyltransferase (CCoAMT) (Tables 1 and 2 ). Table 1 \n Distribution of lignin biosynthesis genes in archaeplastid genomes \n \n Organism \n \n Phylum \n \n PAL \n \n C4H \n \n 4CL \n \n CCR \n \n CAD \n \n HCT \n \n C3H \n \n COMT \n \n CCoAMT \n \n F5H \n \n CSE \n \n sm \n F5H \n \n PER \n \n LAC \n \n ARCHAEPLASTIDS \n \n Land plants \n \n Arabidopsis thaliana \n Streptophyta + + + + + + + + + + + + + \n Oryza sativa \n Streptophyta + + + + + + + + + + + + + \n Physcomitrella patens \n Streptophyta + + + + + + + + + + + + \n Picea abies \n Streptophyta + + + + + + + + + + + + + \n Selaginella moellendorffii \n Streptophyta + + + + + + + + + + + + + + \n Green algae \n \n Bathycoccus prasinos \n Chlorophyta + G \n + G \n \n Botryococcus braunii \n Chlorophyta + E \n + E \n \n C. reinhardtii \n Chlorophyta + + + G \n + G \n \n Chlorella sp. NC64A \n Chlorophyta + G \n + G \n + G \n + G \n \n Chlorella variablilis \n Chlorophyta + G \n + G \n + G \n + G \n + G \n \n Coccomyxa subellipsoidea \n Chlorophyta + G \n + + + + G \n + G \n \n Dunaliella salina \n Chlorophyta + E \n \n Haematococcus pluvialis \n Chlorophyta + E \n \n M. pusilla CCMP1545 \n Chlorophyta + G \n + G \n \n Micromonas sp. RCC299 \n Chlorophyta + G \n + G \n \n O. lucimarinus \n Chlorophyta + \n Ostreococcus sp.RCC809 \n Chlorophyta + G \n \n Ostreococcus tauri \n Chlorophyta + G \n + G \n \n Polytomella sp. \n Chlorophyta + G \n \n Polytomella parva \n Chlorophyta + E \n \n Prototheca wickerhamii \n Chlorophyta + E \n \n Volvox carteri \n Chlorophyta + G \n + + G \n \n Red algae \n \n C. tuberculosum \n Rhodophyta + G \n + G \n + G \n + G \n + G \n \n Chondrus crispus \n Rhodophyta + G \n + G \n + G \n + G \n \n Cyanidioschyzon merolae \n Rhodophyta + G \n + G \n + G \n \n Galdiera sulphararia \n Rhodophyta + G \n + G \n \n Porphyridium cruentum \n Rhodophyta + E \n + E \n + E \n + E \n \n Glaucophytes \n \n Cyanophora paradoxa \n Glaucophyta + G \n + G \n + G \n Footnote = + Present in both genome sequence and EST library, + E Present in EST library only, + G Present in genome sequence only. Presence (+) is determined by a reciprocal BLASTP hit with an E-value < 1x10 -30 or less using the characterized land plant A. thaliana or S. moellendorffii gene as a query, searching the NCBI, JGI, and Congenie databases. The abbreviations used for the enzyme can be described as follows: phenylalanine ammonia-lyase (PAL); cinnamate 4-hydroxylase (C4H); 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductase (CCR); cinnamyl alcohol dehydrogenase (CAD); p -hydroxycinnamoyl-CoA (HCT); p -coumarate 3-hydroxylase (C3H); caffeic acid O-methyltransferase (COMT); caffeoyl-CoA O-methyltransferase (CCoAMT); ferulate 5-hydroxylase (F5H); caffeoyl shikimate esterase (CSE) and Selaginella moelledorfii F5H (smF5H); peroxidase (PER); and laccase (LAC). Table 2 \n Distribution of lignin biosynthesis genes in non-archaeplastid genomes \n \n Organism \n \n Phylum \n \n PAL \n \n C4H \n \n 4CL \n \n CCR \n \n CAD \n \n HCT \n \n C3H \n \n COMT \n \n CCoAMT \n \n F5H \n \n CSE \n \n sm \n F5H \n \n PER \n \n LAC \n \n STRAMENOPILES \n \n Diatoms \n \n Fragilariopsis cylindrus \n Bacillariophyta + + E \n + + \n Phaeodactylum tricornutum \n Bacillariophyta + + + + E \n \n T. pseudonana \n Bacillariophyta + G \n \n Pelagophyte \n \n A. anophagefferens \n Heterokontophyta + G \n + G \n \n Brown algae \n \n Ectocarpus siliculosus \n Phaeophyceae + G \n \n Eustigmatophytes \n \n N. gaditana \n Eustigmatophyceae + G \n + G \n \n Chrysophytes \n \n Ochromonas danica \n Chrysoophyceae + E \n \n Oomycetes \n \n Albugo laibachii \n Heterokontophyta + G \n + G \n + G \n + \n Phytophthora infestans \n Heterokontophyta + + + + \n Phytophthora sojae \n Heterokontophyta + + + \n ALVEOLATES \n \n Apicomplexa \n \n Cryptosporidium muris \n Apicomplexa + G \n + \n Cryptosporidium parvum \n Apicomplexa + G \n + \n Theileria parva \n Apicomplexa \n Toxoplasma gondii \n Apicomplexa \n Dinoflagellates \n \n Symbiodinium minutum \n Dinoflagellata + G \n + G \n + G \n \n Ciliates \n \n Paramecium tetraurelia \n Ciliophora + G \n + \n T. thermophila \n Ciliophora + G \n + G \n \n HAPTOPHYTES \n \n E. huxleyi CCMP1516 \n Haptophyta + + + + + G \n \n CRYPTOPHYTES \n \n Guillardia theta \n Cryptophyta + G \n + G \n + G \n + G \n \n Hemiselmis andersenii \n Cryptophyta Footnote = + Present in both genome sequence and EST library, + E Present in EST library only, + G Present in genome sequence only. Presence (+) is determined by a reciprocal BLASTP hit with an E-value < 1x10 -30 or less using the characterized land plant A. thaliana or S. moellendorffii gene as a query, searching the NCBI, JGI, OIST Marine Genomics Unit ( http://marinegenomics.oist.jp/genomes/gallery/ ) and Paramecium ( http://paramecium.cgm.cnrs-gif.fr/ ) databases. The abbreviations used for the enzyme can be described as follows: phenylalanine ammonia-lyase (PAL); cinnamate 4-hydroxylase (C4H); 4-coumarate:CoA ligase (4CL); cinnamoyl-CoA reductase (CCR); cinnamyl alcohol dehydrogenase (CAD); p -hydroxycinnamoyl-CoA (HCT); p -coumarate 3-hydroxylase (C3H); caffeic acid O-methyltransferase (COMT); caffeoyl-CoA O-methyltransferase (CCoAMT); ferulate 5-hydroxylase (F5H); caffeoyl shikimate esterase (CSE) and Selaginella moelledorfii F5H (smF5H); peroxidase (PER); and laccase (LAC). The first three of these enzymes (4CL, CCR and CAD) catalyze consecutive steps in lignin biosynthesis and are sufficient to produce p -coumaryl alcohol from coumaric acid (Figure 1 ). Homologs of all three enzymes have similar taxonomic distributions, being found mostly in marine photosynthetic algae in addition to land plants. Representatives of green algae, red algae, glaucophytes, diatoms, dinoflagellates, haptophytes and cryptophytes, as well as the non-photosynthetic oomycetes, harbor homologs of these three enzymes. Although oomycetes are non-photosynthetic, they are believed to share a photosynthetic ancestor with other stramenopiles such as brown algae and diatoms [ 36 ]. If the 4CL, CCR and CAD homologs in these diverse eukaryotes indeed catalyze the same biochemical reactions as their plant homologs, the capacity to make at least the precursor of the simplest form of lignin (H lignin) would be much more widespread than currently thought (land plants and red algae). We have performed functional prediction analysis on all homologs of 4CL, CCR and CAD homologs using the Argot2 [ 37 ] and ESG [ 38 ] packages, which are among the best performing functional annotation programs available [ 39 ] (Figure 2 ). Homologs of plant CCR and CAD were predicted to have a conserved function with moderate to high confidence in at least one representative of the green algae, red algae, diatoms, dinoflagellates, haptophytes and cryptophytes. For CAD, alcohol dehydrogenase function was often predicted along with cinnamyl alcohol dehydrogenase function with comparable confidence. The prediction for the specific function of 4-coumarate:CoA ligase was difficult for 4CL homologs, being weak even for enzymes that have been biochemically characterised as having that function (such as several of Arabidopsis thaliana paralogous 4CL enzymes) [ 7 ]. However, ligase function was predicted with high or very high confidence for all organisms with predicted CCR and CAD functions. Although homologs of all three enzymes are also found in glaucophytes and oomycetes, functional predictions were weak or absent for one or more of these enzymes in these two taxonomic groups. Figure 2 \n Functional prediction of the \n p \n -coumaryl alcohol biosynthesis pathway genes. Both programs used for functional prediction (Argot 2 and ESG) need to predict the correct function for it to be annotated as such. An \"F\" indicates that an enzyme from that taxonomic group has been biochemically characterized. No single fungi or bacteria harbour homologs for all three enzymes, but all the enzymes are found individually in some representatives of these taxonomic groups. An empty space indicates that no homolog is found in a particular species. Confidence values are derived from the ESG software. * for 4CL, as even enzymes with biochemically demonstrated function are not annotated as such with significant confidence by the softwares used, the figure indicates prediction of ligase activity. Most enzymes involved in lignin biosynthesis are multifunctional or have multiple, slightly divergent, paralogous copies with different functions, either with the direction of the reaction catalyzed reversed or a change in substrate affinity. This is especially true for CAD, which catalyzes the last step in monolignol biosynthesis. SAD (sinapyl alcohol dehydrogenases) cannot be differentiated from CAD phylogenetically and both enzymes display some of the other’s specific activity [ 40 , 41 ]. Also, there are many enzymes with CAD activity (oxidation of an aldehyde to an alcohol) as well as alcohol dehydrogenase activity (reduction of an alcohol to an aldehyde), such as yeast ADH6 and ADH7 enzymes [ 42 ]. Given the functional flexibility in these protein families, functional switches are likely to have occurred frequently throughout eukaryotic evolution, making exact functional predictions difficult without biochemical data. p -coumaryl alcohol synthesis is likely common in photosynthetic eukaryotes Homologs for all genes of the 4CL-CCR-CAD pathway, responsible for the synthesis of p -coumaryl alcohol from p -coumaric acid, occur in several widely divergent eukaryotic taxonomic groups (stramenopiles, haptophytes, cryptophytes, dinoflagellates and archaeplastids). This does not agree with a previously proposed origin in early land plants [ 3 ]. Several alternative hypotheses could explain this distribution. The genes of this pathway (which will be called the p -coumaryl alcohol biosynthesis pathway henceforth) could have been present in eukaryotic ancestors predating the origin of some or all of these groups, and lost in lineages in which the pathway is absent. The superfamilies to which 4CL (adenylate-forming enzymes), CCR (adenylate reductase) and CAD (medium-chain dehydrogenase/reductases) belong are most certainly ancestral to eukaryotes, being widespread in all eukaryotic supergroups [ 35 , 43 - 47 ]. However, close homologs to characterized plant 4CL, CCR and CAD enzymes with a correctly predicted function are not as ubiquitous. The complete p -coumaryl alcohol biosynthesis pathway is present almost solely in photosynthetic eukaryotes, making a scenario in which they evolve in an ancient eukaryotic ancestor and are subsequently selectively lost in non-photosynthetic taxonomic groups unlikely. A more parsimonious explanation could be that this pathway would have originated in an ancestor of green and red algae and was subsequently transferred to other taxonomic groups in the same way they acquired the capacity to photosynthesize: through taking up a red or green alga in a secondary endosymbiotic event (endosymbiotic gene transfer or EGT) [ 36 ]. Another possibility is that several genes in this pathway were acquired independently by lateral gene transfer (LGT) from bacteria or heterotrophic eukaryotes in various photosynthetic lineages. Phylogenetic analysis of the 4CL, CCR and CAD enzymes were performed to differentiate between these hypotheses on the origin(s) of the p -coumaryl alcohol biosynthesis pathway. p -coumaryl alcohol biosynthesis could have originated in an ancient archaeplastid The p -coumaryl alcohol biosynthesis pathway seems to be ancestral to green algae, but with frequent loss of this metabolic function throughout this taxonomic group (Figure 2 ). All green algal species are found within a single well-supported (97%) clade in the 4CL tree (mixed with red algae and secondarily photosynthetic organisms) (Figure 3 ) and in two clades in the CCR tree and three clades in the CAD tree (Figure 4 and 5 ). The multiple green algal clades in the CAD tree likely represent conservation of a different paralog in two major green algal lineages; prasinophytes and the core chlorophytes [ 48 ]. The core chlorophytes are monophyletic and group with land plants, while the prasinophytes are divided in two clades, the result of multiple divergent paralogs being present in Bathycoccus prasinos and Ostreococcus lucimarinus. In the CCR tree, the two green algal clades are both composed of core chlorophytes, as well as overlapping in their species content and are proximal to each other. They likely represent paralogy originating after the divergence of core chlorophytes from ancestral green algae. Despite paralogy being observed in green algae for all three genes, the presence of p -coumaryl alcohol biosynthesis gene homologs in various species from two major green algal groups suggests the presence of this pathway is ancestral. Green algae are never part of clades containing organisms from other taxonomic groups besides secondarily photosynthetic species and red algae, suggesting the absence of LGT for p -coumaryl alcohol biosynthesis gene homologs in this lineage. The fact that only two of fifteen green algal genomes or EST libraries screened contain homologs to all three p -coumaryl alcohol biosynthesis enzymes suggests this function has been frequently lost and is likely non-essential, as compared to its crucial role in most land plants. Figure 3 \n Maximum likelihood phylogeny of the 4-coumarate:CoA ligase (4CL) enzyme from the \n p \n -coumaryl alcohol biosynthesis pathway. Amino acid sequences were aligned with MUSCLE and the tree compiled using RaxML. Numbers above branches refer to bootstrap values above 50%. Luciferase, a 4CL homolog [ 79 ], was used as the outgroup. Gene names are included next to taxa when function could be predicted. * indicates that the enzyme has been biochemically characterized in this organism [ 29 , 72 , 73 ]. Figure 4 \n Maximum likelihood tree of cinnamoyl-CoA reductase (CCR) enzyme from the \n p \n -coumaryl alcohol biosynthesis pathway. Amino acid sequences were aligned with MUSCLE and the tree compiled using RaxML. Numbers above branches refer to bootstrap values above 50%. 3-hydroxysteroid dehydrogenase, a CCR homolog [ 63 ], was used as the outgroup. The various classes of CCR are shown based on previous research including bona fide CCR compared to CCR or CCR-like genes. In addition, genes showing high similarity to dihydroflavonol reductase (DFR) genes were included [ 19 , 45 ]. Gene names are included next to taxa when function could be predicted. * indicates that the enzyme has been biochemically characterized in this organism [ 30 , 69 , 70 ]. Figure 5 \n Maximum likelihood tree of cinnamyl alcohol dehydrogenase (CAD) enzyme from the \n p \n -coumaryl alcohol biosynthesis pathway. Amino acid sequences were aligned with MUSCLE and the tree compiled using RaxML. Numbers above branches refer to bootstrap values above 50%. Sorbitol dehydrogenase, a CAD homolog [ 80 ], was used as the outgroup. The various classes of CAD are shown based on previous research, including the sinapyl alchohol dehydrogenases (SAD), which share some specific activity with CAD [ 40 , 41 ]. Gene names are included next to taxa when function could be predicted. * indicates that the gene has been biochemically characterized in this organism [ 31 , 42 , 71 , 74 ]. The story is different for red algae. In this group, only 4CL seems to be ancestral, being found in a single clade, which also contains green algae and secondarily photosynthetic organisms (Figure 3 ). The only exceptions to this are two of the five 4CL paralogs in Calliarthron tuberculosum clustering in bacterial clades, one of them very strongly with a marine α-proteobacteria group known as the roseobacter clade, indicating a likely LGT from bacteria. For both CCR and CAD, red algae are polyphyletic, with some paralogs weakly clustering with homologs from bacteria or heterotrophic eukaryotes. It is therefore difficult to say whether the pathway was ancestrally found in red algae or some of its enzymes were acquired by LGT. Like in green algae, only a small proportion of red algal genomes investigated have homologs for all three enzymes (one in five). Although the fact that the red alga Calliarthron cheilosporioides, a close relative to C. tuberculosum, can synthesize p -coumaryl alcohol is well established [ 20 ], this function does not seem to be an essential feature of red algae, making the loss of it likely. This is perhaps visible in the specialized role of lignin in the uncommon lignified joints, genicula, of C. cheilosporioides [ 20 ]. These results suggest that all three genes of the p -coumaryl alcohol biosynthesis pathway are likely to have been present in at least the shared ancestor of plants and green algae, but possibly earlier, before the speciation of the red algal ancestor (Figure 6 ). A few additional features of the phylogenies support an earlier origin than the ancestor of land plants. The CAD homologs found in core chlorophytes cluster with land plants (albeit with modest 74% bootstrap support). Also, although both the CCR and CAD phylogenies likely contain hidden ancestral paralogy and differential loss, this is not the case for the 4CL tree. Paralogy is evident only for red and green algae and confined to a single clade (with the exception of copies likely acquired from bacteria by C. tuberculosum ). As red and green algal species are polyphyletic inside this clade, duplication(s) of the 4CL homolog likely occurred in their ancestor and therefore would predate these lineages. It is difficult to say if the pathway could have been present in the archaeplastid ancestor itself, before divergence of the glaucophytes. Glaucophyte is the earliest branching lineage in archaeplastids [ 49 ] and could potentially be quite informative on the origin of the p -coumaryl alcohol biosynthesis pathway in this group. Unfortunately, although it harbours homologs for all three enzymes of the pathway, none of them has strong functional prediction and their position in phylogenies is unresolved. It is therefore not possible to determine if the pathway is ancestral to archaeplastids with information currently available. Figure 6 \n Major evolutionary events hypothesized in the evolution of the lignin biosynthetic pathway across the eukaryotic tree. The tree is a consensus of current phylogenetic analyses of the eukarotic domain [ 1 , 36 , 49 ]. Major events indicated by labeled arrows on the tree are hypothesized from our genome survey and phylogenetic analyses of the putative p -coumaryl alcohol biosynthesis enzymes. Taxonomic groups in which three or more enzymes catalyzing consecutive steps in the lignin biosynthesis pathway were found are colored. The chemical detection of polymerized lignin is indicated in the margin for each taxonomic group, with the type of lignin (either H, G or S) specified. A question mark (?) indicates that some putative lignin biosynthetic enzymes were found and fuctionally predicted but that there is currently no biochemical evidence of polymerized lignin. Arrows indicate the origin and direction of putative EGT and LGT events (solid arrows are used for events that are conclusive, dashed arrows when events are hypothesized). The p -coumaryl alcohol biosynthesis pathway likely spread through endosymbiotic gene transfer Four very disparate groups of secondarily photosynthetic organisms have at least one representative with all three enzymes of the p -coumaryl alcohol biosynthesis pathway with the correct functional predictions: dinoflagellates, diatoms, hapthophytes and cryptophytes. The large majority of 4CL, CCR and CAD homologs found in these organisms cluster with each other (despite being from widely divergent taxonomic groups) or with red or green algae. There is no observed paralogy of the 4CL gene in secondarily photosynthetic organisms. Diatoms and the haptophyte weakly cluster together (51%), the cryptophyte clusters strongly (100%) with another secondarily photosynthetic organism ( Ectocarpus silicosis , a brown algae) and with green algae, while the dinoflagellate is found inside the same well-supported (97%) mixed green and red algal clade. For CCR, the haptophyte has two paralogs, one clustering strongly with the dinoflagellate (98%) and the other with both the dinoflagellate and the diatom Fragilariopsis cylindrus (100% support). Other diatom CCR paralogs weakly cluster with green algae (50% support), while the single CCR homolog found in the diatom Phaeodactylum tricornutum strongly groups with bacteria (91%) and was likely acquired from a member of this domain by LGT. The cryptophyte single CCR homolog position in the tree is unresolved. The CAD homologs of secondarily photosynthetic organisms show a similar pattern to 4CL and CCR. A dinoflagellate and a hapthophyte paralog strongly cluster together within a green algal clade that also includes the cryptophyte (100% support). The second dinoflagellate CAD paralog weakly clusters with the diatom F. cylindrus single CAD homolog (59% support) and the second haptophyte paralog groups strongly with green algae (100%). The underlying pattern is clear: a recurring clustering of secondarily photosynthetic organisms from disparate taxonomic groups with each other or green or red algae. Although the exact pattern of species clustering varies between phylogenies (a likely result of ancient paralogy and differential loss), it suggests that most 4CL, CCR and CAD homologs present in secondarily photosynthetic species have been acquired by EGT from a red or green algae or their ancestors. This also implies that the p -coumaryl alcohol biosynthesis pathway, or at the very least its component genes, are ancient, predating the diversification of various major eukaryotic taxonomic groups such as the dinoflagellates, hapthophytes, cryptophytes and diatoms. Although ancient paralogy coupled with differential loss can often make phylogenies misleading, it is very unlikely that it would result in similar patterns of taxonomically unrelated secondarily photosynthetic organisms clustering with each other or with green/red algae for three different genes. The recurrent co-clustering of dinoflagellates, hapthophytes and diatoms in 4CL, CCR and CAD phylogenies is more parsimoniously explained by a common origin. The cryptophyte, Gulliardia theta, on the other hand, does not cluster directly with these other secondarily photosynthetic organisms in any phylogeny, suggesting an independent origin. More evidence is needed to confirm the exact origin of these genes and the number of events in which they might have been acquired by secondarily photosynthetic organisms. LGT could have impacted the evolution of lignin precursors biosynthesis Previous studies have suggested that at least one gene in the lignin biosynthesis pathway, phenylalanine ammonia lyase (PAL) (Figure 1 ), was likely acquired through LGT from soil bacteria to an ancestor of land plants [ 50 ]. LGT is likely to also have influenced the evolution of p -coumaryl alcohol biosynthesis. Some of the 4CL homologs present in the red alga C. tuberculosum might have been affected by this phenomenon. Indeed, the 4CL phylogeny contains a strongly supported clade composed of the red alga C. tuberculosum grouping with bacteria, including many sequences derived from roseobacters (Figure 3 ). This suggests a horizontal 4CL gene transfer from a roseobacter to a red alga (Figure 3 ), giving C. tuberculosum extra copies of 4CL in addition to those it likely inherited from archaeplastid ancestors. Close physical associations have been shown to promote LGT, and two roseobacter species are known to live intracellularly and intercellularly within red macroalgae [ 51 , 52 ]. Some red algal species even depend on bacteria for growth or morphogenesis, highlighting the intimacy of this relationship [ 53 , 54 ]. This bacteria-red algae clade in the 4CL tree also includes a functionally characterized gene from the bacterium Streptomyces coelicolor , which has been shown to have 4CL activity [ 55 ]. Argot2 and ESG also predict all roseobacter and C. tuberculosum 4CL homologs in this clade to have 4CL function, a prediction only made for these bacteria, red alga and some land plants enzymes in our datasets. This makes it likely that these laterally transferred homologs have true 4-coumarate:CoA ligase function. For a LGT event from a bacterium to a eukaryote to be confirmed, bacterial genes need to be found inserted next to genuine eukaryotic genes. Unfortunately, the C. tuberculosum 4CL homologs are found on very small contigs in the alga’s genome and further upstream and downstream sequence data would be needed to determine if they have a bacterial or algal context. It is therefore not possible to exclude that the C. tuberculosum 4CL genes clustering with roseobacters represent bacterial contamination present in its genome sequence, despite systematic screening of sequence data to remove it [ 56 ]. Regardless of whether LGT between roseobacters and C. tuberculosum has taken place, the presence of 4CL in these bacteria raises the possibility that they could provide intermediates for the production of p -coumaryl alcohol to their algal host. If roseobacters can make p -coumaric acid (none of the known genes for doing so have yet been found in this group), the presence of 4CL would theoretically enable them to produce p -coumaroyl-CoA and potentially provide it to their algal host. Another likely case of LGT is the acquisition of a putative CCR (with very high confidence functional prediction) by the diatom P. tricornutum from bacteria. The former is found nested inside a bacterial clade with strong support (91%). As other diatoms’ putative CCRs cluster with green algae, it is likely that a bacterial CCR displaced the homolog from algal origin previously present in P. tricornutum . Although LGT did not bring a novel gene to either C. tuberculosum or P. tricornutum (both had an existing homolog prior to LGT which was either displaced or complemented), it likely had an effect on their secondary metabolism. The phenylpropanoid pathway is unique to land plants and fungi The phenylpropanoid pathway enzymes responsible for the production of p -coumaric acid from the amino acid phenylalanine, PAL and cinnamate 4-hydroxylase (C4H), were only found in land plants and fungi and are clearly missing from all other eukaryotic genomes screened (Tables 1 and 2 , Figure 1 ). How is it possible for organisms to synthesize monolignols without these enzymes? The red alga Calliarthron can produce all types of lignin [ 20 ] and we could not find these enzymes encoded in the C. tuberculosum genome sequence (Table 1 ). The product made by PAL and C4H from phenylalanine, p -coumaric acid, has been found in an axenic culture of the haptophyte E. huxleyi [ 27 ], which also lacks PAL and C4H (Table 2 ). Since both Calliarthron and E. huxleyi have 4CL, CCR and CAD homologs, there must be other, yet to be described, enzyme(s) capable of synthesizing p -coumaric acid and provide it as a substrate to the 4CL-CCR-CAD pathway to produce p -coumaryl alcohol. Furthermore, we could not find PAL or C4H genes in the genomes of any green alga or diatom (Tables 1 and 2 ), although p -coumaric acid has been found in species from both of these groups [ 28 ]. PAL was likely acquired from bacteria by the ancestor of land plants or fungi and later horizontally transferred between these two groups [ 50 ]. C4H is also uniquely found in land plants and fungi, with distant bacterial homologs (data not shown). The combination of these two genes is therefore likely a late invention/acquisition of land plants and fungi. Land plants added two genes (C4H and PAL) to the 4CL, CCR and CAD already present in their ancestor (Figure 6 ). Whether PAL and C4H displaced an ancestral enzyme(s) synthesizing p -coumaric acid or those enzyme(s) are still present in land plant genomes is currently unknown. The only other enzyme known to synthesize p -coumaric acid is tyrosine ammonia lyase (TAL), which has only been found in a few bacteria [ 50 ]. It can by itself convert the amino acid tyrosine to coumaric acid, suggesting that a single unknown enzyme could carry the same function in eukaryotes lacking PAL and C4H but which have 4CL, CCR and CAD, such as the hapthophyte E. huxleyi and the red alga Calliarthron. It is also possible that enzymes analogous to PAL and/or C4H exist, as PAL activity has been found in the green alga Chlorella pyrenoidosa [ 57 ], but we could not find PAL homologous to plant enzymes in any of the Chlorella genomes screened (Table 1 ). Expansion of the lignin biosynthesis pathway has occurred multiple times independently on land and in the sea Screening of eukaryotic genomes revealed that except for 4CL, CCR and CAD, all other lignin biosynthesis genes found in land plants are missing from Calliarthron (Table 1 ), despite the clear presence of all three lignin types in this red alga [ 20 ]. Assuming the capacity to produce 4CL’s substrate p -coumaric acid, the presence of 4CL, CCR and CAD genes theoretically enables the synthesis of p -coumaryl alcohol as well as its intermediates, which can be used as substrates for the synthesis of G and S lignins (Figure 1 ). This makes convergent evolution of the ability to synthesize G and S lignins in Calliarthron simple, as it would only require the addition of two more enzymes to this core pathway. For example, the lycophyte S. moellendorfii only needed to add caffeic acid O -methyltransferase (COMT) and ferulate 5-hydroxylase ( sm F5H) to the enzymes needed for p -coumaryl alcohol synthesis to be able to make both G and S lignins [ 17 ]. Also, not all land plants can make S lignin, and the presence of both producers and non-producers in various groups of plants suggests that this ability has been gained and lost multiple times in land plants. For example, most gymnosperms do not produce S lignin, but some can, such as Ginkgo biloba (maidenhair tree) [ 15 ]. The fact that modifications of lignin production can easily evolve from a genetic background found in various photosynthetic eukaryotic lineages lends insight into how the red alga Calliarthron likely evolved the ability to produce G and S lignins. Whether this type of convergent evolution has also happened in other lineages is an open question. Provisional biochemical evidence for the presence of p -coumaryl alcohol in brown and green algae, but absence of G and S variants [ 4 ] suggests that numerous eukaryotes could have the capacity to only synthesize p -coumaryl alcohol or H lignin."
} | 9,113 |
32402068 | PMC7348690 | pmc | 9,373 | {
"abstract": "Abstract The origin of plastids (chloroplasts) by endosymbiosis stands as one of the most important events in the history of eukaryotic life. The genetic, biochemical, and cell biological integration of a cyanobacterial endosymbiont into a heterotrophic host eukaryote approximately a billion years ago paved the way for the evolution of diverse algal groups in a wide range of aquatic and, eventually, terrestrial environments. Plastids have on multiple occasions also moved horizontally from eukaryote to eukaryote by secondary and tertiary endosymbiotic events. The overall picture of extant photosynthetic diversity can best be described as “patchy”: Plastid-bearing lineages are spread far and wide across the eukaryotic tree of life, nested within heterotrophic groups. The algae do not constitute a monophyletic entity, and understanding how, and how often, plastids have moved from branch to branch on the eukaryotic tree remains one of the most fundamental unsolved problems in the field of cell evolution. In this review, we provide an overview of recent advances in our understanding of the origin and spread of plastids from the perspective of comparative genomics. Recent years have seen significant improvements in genomic sampling from photosynthetic and nonphotosynthetic lineages, both of which have added important pieces to the puzzle of plastid evolution. Comparative genomics has also allowed us to better understand how endosymbionts become organelles.",
"conclusion": "Conclusion Three photosynthetic lineages—green algae plus land plants, red algae, and glaucophyte algae—harbor plastids that appear to stem directly from a primary endosymbiotic event with a cyanobacterium a billion-plus years ago ( Parfrey et al. 2011 ; Shih and Matzke 2013 ). A wealth of data reveals that the plastids of green and red algae subsequently spread far and wide across the tree of eukaryotes by higher-order endosymbiotic mergers between eukaryotic hosts and endosymbionts. Here, we have focused on insights gleaned from the perspective of genomics, but it is important to note that advances in our understanding of the cell biology, biochemistry, and metabolism of diverse algae and plants have contributed greatly to the broad picture of plastid evolution (see, e.g., Gould et al. 2015 ; Kim and Archibald 2009 and references therein). The challenge is combining all of these data into a single, coherent picture of the birth and spread of plastids. Despite 20-plus years of molecular phylogenetic and genomic investigation, we are still in the dark about how many eukaryote–eukaryote endosymbioses have occurred and, in many cases, who the partner cells even were (although not discussed here, interested readers should refer to Kim and Maruyama [2014] for a provocative discussion of the possibility that the plastid found in green algae and land plants is of secondary endosymbiotic origin). With phylogenomics, the deep structure of the eukaryotic tree of life has become ever more resolved (e.g., Strassert et al. 2019 ; Burki et al. 2020 ; fig. 1 ), thereby providing a framework for mapping plastid gains and losses across the tree. However, the same genomic data that have enabled construction of taxonomically rich phylogenomic trees also reveal that the nuclear genomes of complex algae are mosaics of genes of both red and green algal ancestry. Putative LGTs from bacteria are also increasingly described in the nuclear and plastid genomes of phototrophs representing the full breadth of algal diversity (e.g., Khan et al. 2007 ; Dorrell et al. 2017 ; Ševcíková et al. 2019 ; Novák Vanclová et al. 2020 ). Why (and how) this is so is not yet clear. As we have seen, lineages such as the dinoflagellates and apicomplexans provide a window into the dynamics of plastid gain, loss, and replacement over recent evolutionary timescales. Moving forward, these data will help us to generate and test hypotheses with which to elucidate much older events in plastid evolution. Which complex algal lineage was the initial recipient of the primordial red algal secondary plastid? How many secondary, tertiary, and (possibly) quaternary endosymbioses gave rise to the full breadth of algal biodiversity, and who were the plastid donors and recipients? Answers to these questions will require even greater efforts to improve taxon sampling among both heterotrophic and photosynthetic lineages, as well as creative solutions to hard bioinformatic problems.",
"introduction": "Introduction Algae are diverse and ecologically important organisms found across the eukaryotic tree of life, and they all have at least one thing in common—they are photosynthetic. Although their precise evolutionary paths vary, all canonical plastids are believed to be derived from the same endosymbiotic cyanobacterium. How do we know this? Simply put, it is in their DNA. The genes that remain within modern-day plastid genomes, along with those that have relocated to the nuclear genome, provide important clues as to how plastids evolved across different branches of the eukaryotic tree. The first plant plastid genomes were sequenced in the 1980s (Ohyama et al. 1986 ; Shinozaki et al. 1986 ) and provided preliminary insight into their gene content and structure. Over 4,000 plastid genomes have now been sequenced from a wide diversity of photosynthetic and secondarily nonphotosynthetic eukaryotes. Together with nuclear genome sequences, these data have made it possible to investigate the “who, what, when, where, and how” of eukaryotic photosynthesis. Here, we review recent genomics-based advances in our understanding of how plastids arose and spread. With the immense amount of information now coming from genome sequencing projects, the potential for discovery and insight into these fundamental questions is unparalleled. What is emerging is a more complete and nuanced view of plastid evolution, one that is nevertheless still lacking in important details."
} | 1,483 |
37520357 | PMC10379653 | pmc | 9,376 | {
"abstract": "Microbial electrosynthesis (MES) is an emerging electrochemical technology currently being researched as a CO 2 sequestration method to address climate change. MES can convert CO 2 from pollution or waste materials into various carbon compounds with low energy requirements using electrogenic microbes as biocatalysts. However, the critical component in this technology, the cathode, still needs to perform more effectively than other conventional CO 2 reduction methods because of poor selectivity, complex metabolism pathways of microbes, and high material cost. These characteristics lead to the weak interactions of microbes and cathode electrocatalytic activities. These approaches range from cathode modification using conventional engineering approaches to new fabrication methods. Aside from cathode development, the operating procedure also plays a critical function and strategy to optimize electrosynthesis production in reducing operating costs, such as hybridization and integration of MES. If this technology could be realized, it would offer a new way to utilize excess CO 2 from industries and generate profitable commodities in the future to replace fossil fuel-derived products. In recent years, several potential approaches have been tested and studied to boost the capabilities of CO 2 -reducing bio-cathodes regarding surface morphology, current density, and biocompatibility, which would be further elaborated. This compilation aims to showcase that the achievements of MES have significantly improved and the future direction this is going with some recommendations. Highlights \n – MES approach in carbon sequestration using the biotic component. – The role of microbes as biocatalysts in MES and their metabolic pathways are discussed. – Methods and materials used to modify biocathode for enhancing CO 2 reduction are presented.",
"conclusion": "10. Conclusion The MES is a BES technology proven to be a new study area with unrealized potential in combating climate change and increasing circular bioeconomy. The catalyst development of MES proved that many works in increasing the effectiveness of CO 2 reduction via microbes were done previously. This review also covers the metabolic pathways related to the process, which could help harness it to improve the process, such as hydrogen-mediated pathways. Many approaches are being applied in modifying the biocathode through different ranges, from conventional methods to new methods, materials, and technology strategies, to transform CO 2 into various chemicals. All new research mainly focused on a noble free metal catalyst that could be fabricated economically and work better than previous studies to achieve the final result, the biocompatible electrode. Also, some new key finding, such as 3D printer is an interesting method of biocathode preparation and is a promising technology for making better biocathodes. The factors of the MES process are also discussed to optimize the process better and with the idea of scaling up the MES for deployment on an industrial scale. All the key findings in MES research could help achieve CCU and decarbonise the economy, thus following the global aspiration to pursue sustainable development goals.",
"introduction": "1. Introduction The anthropogenic carbon dioxide (CO 2 ) emission is responsible for global warming and climate. Furthermore, even with the implementation of the Paris Agreement at 33.5 Gt in 2018, climate change has increased to unprecedented levels in recent years. On the positive side, 2019 saw stable CO 2 emissions at 33 Gt due to the implementation of renewable energy in the power sector ( IEA, 2020 ). Then, in 2020, CO 2 emissions in the energy sector dropped by around 5.2% ( IEA, 2020 ) due to the lockdown caused by Covid-19. The slight drop in CO 2 emission around the globe was an excellent short-term effect caused by the outbreak. As the efforts to recover from the pandemic, 2022 saw global CO 2 emissions and energy rebound to reach their highest-ever annual level at a 0.9% increase, translating to 321 Mt. ( IEA, 2023 ). Since 2020, the emissions have increased to 36.3 Gt with an average global CO 2 growth rate of 3.0% annually from the early 2000s before decreasing to 1.2% annually for the last decade (2010–2019) ( Mofijur et al., 2021 ). Thus, an effective and reliable method for capturing and storing CO 2 capture and storage is vital to reduce and limit greenhouse gasses, especially CO 2 , as it is the highest emitted pollutant worldwide. Carbon capture, utilization, and storage (CCUS) are one of the proposed ideas and methods used to address the problem and reach net-zero carbon emissions. CCUS is a holistic approach because it can be used in various sectors, such as manufacturing, construction, and oil and gas. As of last year, around 35 commercial facilities are applying CCUS to industrial processes, fuel transformation, and power, with a total annual capture capacity of almost 45 Mt. CO 2 annually, exist around the globe, with 200 more yet to be completed in 2030, with 200 Mt. sequestration capacity ( IEA, 2022 ). This method revolves around two categories, one of which is carbon capture and storage (CCS), in which the CO 2 is collected, transported, and can be permanently stored in a facility to prevent it from ever reaching the atmosphere. One notable example is the Carbfix project in Iceland, where CO 2 and hydrogen sulfide from the Hellisheidi geothermal power plant is injected into geothermal reservoirs, specifically porous basalt rocks, to form stable minerals for safekeeping ( Sigfússon et al., 2018 ). Another category is called carbon capture and utilization (CCU), where the CO 2 generated by industry or recaptured from the atmosphere is used to produce carbon-based products such as fuel and chemicals. This method is used in enhanced oil recovery (EOR), where CO 2 is injected into oil reservoirs to extract and displace remaining valuable gas where it is difficult to reach. This CCU method is applied worldwide, such as in Jilin Oilfield in China ( Ren et al., 2016 ) and The Gorgon Project in Australia ( Flett et al., 2009 ). This method could recover up to 470 million barrels of oil and 10–35% of the original gas in reservoirs ( Hamza et al., 2021 ). Another application that has caught up with many researchers worldwide, which will be highlighted in this review, is the application of the microbial electrosynthesis system (MES) in CCU and how it will benefit scientists in addressing climate change generally and CO 2 emission, specifically. In an economic sense, the approach is favorable because the value-added product can be used or sold to a wide range of consumers, which could offset the cost of production itself instead of being stored in a facility. In other words, side profit can be generated by integrating these technologies in existing CO 2 -producing facilities, such as metal, power, and cement industries. Thus, the initiatives are incentivized to implement the bio-recycling system widely, contributing to the circular bioeconomy of MES, where the waste generated is recycled and converted to profitable products ( Bian et al., 2020 ). Also, the design and scalability of MES can be easily implemented using existing fuel cell technology as both essentially run on the same principle, design, and component. For example, a small cluster of simple, small-size MES stack in hybrid series and parallel configuration can be used as the scaling-up strategy in implementing this system with a projected surface/volume ratio of 1 cm 2 /mL using flat, multi-chamber reactors ( Jourdin et al., 2020 ). MES is a bio-electrochemical system (BES) that harnesses and reduces CO 2 into other beneficial carbon compounds using biocatalysts. This method combines existing CCU technologies and incorporates biotic components in the system that harness CO 2 as a substrate and transform it into value-added products. However, up to this day, the commercialization of the MES system is far from reality because of some technical barriers that need to be solved to scale up system feasibility. These issues include the energy-intensive requirement of downstream processing and anode material cost, poor selectivity of the valuable product, and the interaction of cathode catalyst and biocatalyst, the most crucial aspect in determining its practicality and efficiency because it is the core part of MES. Thus, this review will concentrate on the recent progress of cathode catalyst fabrication from the past decade by measuring and comparing its performance from various angles in MES."
} | 2,147 |
23100857 | null | s2 | 9,377 | {
"abstract": "Marine mussels utilize a variety of DOPA-rich proteins for purposes of underwater adhesion, as well as for creating hard and flexible surface coatings for their tough and stretchy byssal fibers. In the present study, moderately strong, yet reversible wet adhesion between the protective mussel coating protein, mcfp-1, and amorphous titania was measured with a surface force apparatus (SFA). In parallel, resonance Raman spectroscopy was employed to identify the presence of bidentate DOPA-Ti coordination bonds at the TiO(2)-protein interface, suggesting that catechol-TiO(2) complexation contributes to the observed reversible wet adhesion. These results have important implications for the design of protective coatings on TiO(2)."
} | 183 |
31001240 | PMC6454187 | pmc | 9,379 | {
"abstract": "Most microbes can produce surface-associated or suspended aggregates called biofilms, which are encased within a biopolymer-rich matrix. The biofilm matrix provides structural integrity to the aggregates and shields the resident cells against environmental stressors, including antibiotic treatment. Microscopy permits examination of biofilm structure in relation to the spatial localization of important biofilm matrix components. This review highlights microscopic approaches to investigate bacterial biofilm assembly, matrix composition, and localization using Pseudomonas aeruginosa as a model organism. Initial microscopic investigations provided information about the role key matrix components play in elaborating biofilm aggregate structures. Additionally, staining of matrix components using specific labels revealed distinct positioning of matrix components within the aggregates relative to the resident cells. In some cases, it was found that individual matrix components co-localize within aggregates. The methodologies for studying the biofilm matrix are continuing to develop as our studies reveal novel aspects of its composition and function. We additionally describe some outstanding questions and how microscopy might be used to identify the functional aspects of biofilm matrix components.",
"conclusion": "Conclusions and Future Perspectives Understanding of the structural changes that biofilms undergo throughout their lifecycle requires an understanding of the structure and function of the individual biofilm matrix components as well as of how those biomolecules assemble into a three-dimensional architecture. We have used CLSM to investigate that biofilm matrix biomolecules have structural and chemical properties that are specific to their function. For example, matrix components may possess properties that mediate interactions with other biomolecules (i.e., Pel and eDNA interactions) to produce unique properties that support structure and function of the entire biofilm ( Borlee et al., 2010 ; Jennings et al., 2015 ). As such, investigations which use CLSM to complement additional microbiological approaches have identified the structural and functional basis of biofilm matrix stability and function. This improved understanding will provide more nuanced approaches to treating biofilm infections such as strategies to disrupt biofilms so that biofilm bacteria become susceptible to antimicrobial treatment and host immune responses. There are many questions that remain unanswered about the P. aeruginosa biofilm matrix. For example, as of yet, the localization of matrix proteins in P. aeruginosa has not been determined. One could imagine that the correct spatial positioning of a matrix protein might be crucial for reaping the benefits of its extracellular activity. Unfortunately, we currently lack the resolution to resolve the positioning of matrix proteins relative to other components. The use of fluorescently conjugated antibodies to stain proteins in biofilms could be used to determine if Psl-binding proteins (such as CdrA and ecotin) co-localize with Psl during biofilm development. Additionally, advances in super resolution microscopy have permitted the tracking of single matrix components for Vibrio cholerae ( Berk et al., 2012 ), but similar methods have not been applied to P. aeruginosa . Another method that may be useful is electron microscopy, which has been applied to investigate overall biofilm and biofilm matrix morphology ( Hung et al., 2013 ; Serra et al., 2013 ; Hollenbeck et al., 2014 ; Joubert et al., 2017 ). For electron microscopy, immunogold-labeled antibodies can be used to study matrix proteins. Indeed, super resolution fluorescence microscopy might lead to novel insight concerning matrix structure. The one study to date that has been performed led to the discovery that V. cholerae precisely positions its matrix EPS and proteins ( Berk et al., 2012 ). Investigating the functional significance of such positioning, including mechanisms through which it is localized, is an obvious next step in understanding biofilm assembly. For P. aeruginosa , another interesting vista is understanding the advantages and constraints of using different EPS components. We might very well find that P. aeruginosa finely tunes EPS production in response to the environment as a means to maximize the functionality of the matrix.",
"introduction": "Introduction Microbes form multicellular communities called biofilms ( Costerton et al., 1995 ). Within these communities, microbial aggregates are encased in a biopolymer-rich extracellular matrix. Biofilm formation helps microbes to persist in several niches ranging from the natural environment to human infections ( Costerton et al., 1999 ; Hall-Stoodley et al., 2004 ; Flemming et al., 2016 ). In the human host, biofilms cause serious and chronic infectious diseases including recurrent urinary tract infections, biofouling of medical devices, and chronic infections of wounds and burns ( Donlan and Costerton, 2002 ; Parsek and Singh, 2003 ; Metcalf and Bowler, 2013 ). Typically, biofilm microbes exhibit decreased susceptibility to antimicrobial treatments relative to their planktonic counterparts ( Stewart and Costerton, 2001 ). Biofilm matrix composition varies depending upon the microbial species and growth conditions. In general, biofilm matrix contains exopolysaccharides (EPS), proteins, and extracellular DNA (eDNA). These matrix components are assembled into supramolecular structures that aid in shielding microbes from external stresses ( Flemming and Wingender, 2010 ). The matrix fills more than just a structural role. For example, the matrix can retain protective proteins (e.g., ecotin) or serve as a signal (e.g., to increase biofilm matrix production) ( Irie et al., 2012 ; Steinberg and Kolodkin-Gal, 2015 ; Dragoš and Kovács, 2017 ; Tseng et al., 2018 ). Investigations of biofilm matrix composition and structure are challenging due to their inherent complexity. Despite this challenge, several key microscopy and biochemical assays have been developed and successfully applied to annotate biofilm matrix composition and determine the functional roles of the matrix components ( Azeredo et al., 2017 ). Genetic approaches have identified genes that are important for biofilm matrix formation ( O’Toole et al., 1999 ). Putative biofilm-involved genes can be deleted and/or overexpressed, and then the resulting matrix composition and biofilm phenotypes can be assayed. For example, quantitative composition can be profiled using mass spectrometry methods (e.g., glycomics, proteomics, etc.) ( Sauer, 2003 ; Zarnowski et al., 2014 ) or nuclear magnetic resonance (NMR) ( Reichhardt and Cegelski, 2014 ). Changes in levels of known matrix components can be determined using immunoblotting ( Fong et al., 2010 ; Colvin et al., 2012 ). The ability of bacteria to form communities can be measured in several ways including monitoring the ability of bacteria to aggregate in liquid culture ( Borlee et al., 2010 ; Rybtke et al., 2015 ; Cooley et al., 2016 ; Reichhardt et al., 2018 ), form pellicles (biofilms at the air liquid interface) ( Hollenbeck et al., 2014 , 2016 ), and adhere to abiotic and biotic surfaces ( Colvin et al., 2012 ). All these approaches involve generating average values for the entire community, while providing no information regarding heterogeneity in the system. Confocal laser scanning microscopy (CLSM) is a useful tool to study biofilms, and our lab has extensively applied CLSM to study biofilms cultured under continuous flow in flow-cell reactors ( Borlee et al., 2010 ; Colvin et al., 2012 ; Tseng et al., 2013 ; Jennings et al., 2015 ). This method has several advantages including that it is a reproducible biofilm culturing format, live biofilms can be non-destructively imaged at multiple timepoints, and spatial information can be obtained regarding cell and matrix distribution ( Heydorn et al., 2000a ). Additionally, antibody- and lectin-conjugated dyes can be used to identify and spatially resolve biofilm constituents. CLSM in conjunction with image analysis software such as COMSTAT ( Heydorn et al., 2000b ) can be used to quantitatively study biofilm matrix, as well as the amount of adherent biomass, while localization and relative amounts of matrix constituents can also be determined ( Swerhone et al., 2001 ; Berk et al., 2012 ). Additionally, spatiotemporal effects of different nutrient environments or antimicrobial treatments can be monitored ( Tseng et al., 2013 ; Sønderholm et al., 2017 ). This review will discuss ways that our laboratory has implemented CLSM to study matrix composition and function of Pseudomonas aeruginosa biofilms, with our general approach summarized in Figure 1 . Figure 1 CLSM can be used to obtain important structural and functional information about biofilm matrix. (A) CLSM can be used to compare wild-type with matrix mutant strains grown in flow-cells, which provides information about how specific matrix components contribute to the amount of biomass covering the surface of the flow-cell coverslip, biofilm aggregate formation, and the morphology and size of aggregates. (B) The retention and localization of matrix components (e.g., EPS, eDNA) or exogenous molecules (e.g., antibiotics) can be monitored by CLSM. For example, EPS localization can be tracked with fluorescently-conjugated lectins, and the retention of antibiotics can be monitored by using fluorescently-conjugated antibiotics. Fluorescence intensity can be quantitated using image processing software, and correlated to position across the diameter of a biofilm aggregate. In the schematic, red-stained matrix elements localize to the periphery of the aggregate, green-stained elements are present throughout, and yellow-stained matrix elements localize to the aggregate interior. P. aeruginosa is both a model organism for laboratory study of biofilms and an important pathogen that causes chronic infections. Examples of chronic infections caused by P. aeruginosa include respirator-associated pneumonia, infections of burns and wounds, and lung infections in patients with cystic fibrosis (CF) ( Mulcahy et al., 2014 ). P. aeruginosa can use three EPS to assemble its biofilms: alginate, Psl, and Pel. Alginate is the key EPS of mucoid biofilms, and Psl and Pel are predominant in non-mucoid biofilms ( Mann and Wozniak, 2012 ; Moradali et al., 2017 ). Alginate is a negatively charged polymer of guluronic and mannuronic acids ( Schürks et al., 2002 ). Psl is a neutral polysaccharide consisting of pentasaccharide repeats containing D-mannose, D-glucose, and L-rhamnose ( Kocharova et al., 1988 ; Ma et al., 2007 ; Byrd et al., 2009 ). Pel is a cationic polymer of partially deacetylated N-acetylglucosamine and N-acetylgalactosamine ( Jennings et al., 2015 ). These EPS are present in varying amounts depending upon the specific strain and stage of infection ( Ma et al., 2006 ; Colvin et al., 2012 ). For example, alginate, Psl, and Pel are believed to be expressed during different stages of P. aeruginosa CF lung infections ( Martin et al., 1993 ). Proteins also play a key role in the P. aeruginosa biofilm matrix. The structural matrix protein CdrA is important for aggregate assembly and localization and retention of Psl (discussed in detail below) ( Borlee et al., 2010 ; Reichhardt et al., 2018 ). Additional proteins have been found in the matrix, which could impart functions ranging from nutrient acquisition to protection from oxidative stress ( Toyofuku et al., 2012 ). Recently, the serine-protease inhibitor ecotin was identified as a matrix protein that binds to Psl ( Tseng et al., 2018 ). Matrix-bound ecotin was found to protect bacteria from attack by the host immune protease neutrophil elastase. In these ways, the biofilm matrix acts as both a structural scaffold for biofilm assembly and an active functional network. Finally, eDNA has been identified as a component of biofilms formed by several species, including P. aeruginosa ( Whitchurch et al., 2002 ; Moscoso et al., 2006 ; Rice et al., 2007 ; Harmsen et al., 2010 ). The source of eDNA in P. aeruginosa is unclear although it may simply result from cell lysis that occurs during biofilm growth ( Webb et al., 2003 ; Allesen-Holm et al., 2006 ). The addition of DNase to growth medium inhibits biofilm formation at early stages, suggesting that DNA is important for biofilm development ( Whitchurch et al., 2002 ). However, the addition of DNase to established biofilms does not significantly disrupt them due, at least partly, to protective interactions within the biofilm matrix."
} | 3,176 |
38805271 | PMC11161743 | pmc | 9,380 | {
"abstract": "Significance Light distribution within algal cultures is one of the primary limitations to scalable and efficient biomass growth, a pertinent issue given the increasing interest in nonplanktonic growth methods, such as biofilms. Within these, cells experience uneven illumination via either overexposure on the outer surface or underexposure inside the film. We show how light distribution is altered upon cell aggregation, which naturally occurs under confinement, and enhanced through the incorporation of scatterers. Our work provides insights into how future photobioreactors could be engineered to optimize light delivery, allowing efficient cultivation of microalgae at scale. Last, our work also provides a better understanding of light propagation through gel-encapsulated biomass, a key area given the rise of research interest in engineered living materials.",
"discussion": "Discussion By comparing the light management capabilities of biofilms and gel-immobilized cultures, we conclude that gel-immobilized algal cultures have the potential to reach a higher areal biomass density compared to flat, homogeneous algal biofilms. Our results suggest that the formation of cell aggregates upon hydrogel immobilization is crucial, as it reduces the probability of photon-cell interaction, effectively lowering the scattering and absorption coefficient ( 56 – 58 ). As a result, more photons were able to penetrate and reach greater depths in the gel-immobilized algal culture ( Fig. 3 B ). Such an increase in light penetration depth alleviated the self-shading of the algal biomass. Such self-shading is inevitable in dense microalgal biofilms, limiting their thickness to 300 µm or less ( 51 , 53 , 54 ), corresponding to an aerial biomass density of ~30 mm 3 cm −2 . Additionally, gel-immobilized systems were able to achieve significantly higher P net than biofilms at higher biomass densities, particularly when higher incident irradiance was required to counteract self-shading within the growing biomass ( Fig. 3 D and E ). While photoinhibition could be minimized with a lower incident irradiance, the predominance of a light-limiting regime would lower the overall photosynthetic efficiency, especially at high biomass density with significant self-shading ( Fig. 3 D ). Hence, as the biomass density increased, moderate to high levels of illumination could reach the shaded cells in the lower region of the gel-immobilized algal culture. This conclusion is supported by the decrease in overall P net with biomass density and the enhancement of P net with incident scalar irradiance beyond an aerial biomass density of 20 mm 3 cm −2 ( Fig. 3 D ). In comparison to a homogeneous biofilm, a gel-immobilized culture achieved higher P net , with an aerial biomass density exceeding 30 mm 3 cm −2 . A heterogeneous biomass distribution from cell aggregation reduced the proportion of photoinhibited cells among the top layers as not all cells located in the upper layers were exposed to the same irradiance level. Furthermore, self-shading within individual aggregates protected some of the cells against excessive irradiance. In contrast, the algal biomass in biofilms was uniformly exposed and hence equally photoinhibited in the top layer. With further self-shading, part of the shielded culture became light-limited ( Fig. 3 E ). Meanwhile, the shaded biomass beneath the top layer received an optimal light level. In the case of gel immobilization, the percentage of photoinhibited biomass was significantly lower compared to the biofilm, owing to its heterogeneous distribution of cells. Such optimization required finding an optimal trade-off between alleviating the light shading in the lower region with intense illumination irradiance while incurring a smaller degree of photoinhibition in the upper region of biomass. Finally, a gel-immobilized system was able to deliver light more efficiently than a biofilm even when a low incident irradiance is desired, given an optimal scattering matrix. We found that modifying the scattering properties of the hydrogel matrix could enhance the overall photosynthetic performance, both in our simulations ( Fig. 4 C ) and experiments ( Fig. 4 E ). Algal growth in gel-immobilized systems with added scattering particles was promoted through increasing the amount of light available for photosynthesis. This is especially prominent given that these cultures were grown under low photon irradiance (~40 µmol photons m −2 s −1 ). In industrial settings, direct and intense light sources cannot always be guaranteed, and factors like shading or variable sunlight can dilute the light reaching the samples. With the proof of concept shown here, it is now possible to develop a light-sensitive material with dynamic scattering properties upon light exposure of varying intensity, optimizing the light distribution with a better trade-off between the proportion of photolimited and photoinhibited cells within the algal cell population. However, it is important to consider that light management is not the sole factor to consider in a photobioreactor. The availability of gases and nutrients in a hydrogel system depends on the diffusion of molecules within and between cell aggregates. Diffusion-limited growth became evident from our observation of a lower cell density within a thicker bulk hydrogel ( Fig. 4 E ). Potential strategies have been studied to address diffusion limitations in hydrogels, such as 3D bioprinting to increase the surface area-to-volume ratio ( 23 , 59 ) or cocultivation of algae with symbiotic bacteria to enhance gas and nutrient exchange ( 39 ). We also note that our study has used a simplified assumption to capture the main optical properties of a homogeneous biofilm. In reality, biofilms growing on a substrate can exhibit more complex morphology both in terms of uneven surface morphology and bulk porosity ( 51 , 60 ). Notably, there are multiple factors in play concerning the photosynthetic performance of any system. This paper considers the efficiency of light harvesting and utilization for photosynthesis primarily from the perspective of the spatial distribution of algal biomass. With the advancement of cell-matrix composites using encapsulation and immobilization techniques, it is important to understand how these processes implicate optical performance, and our work serves to address the impact of cell aggregation. In conclusion, we showcase the advantages of cell aggregation in gel-encapsulated colonies of C. reinhardtii compared to biofilm growth, when it comes to light management. This aggregation led to improved light transmission and utilization, particularly under optimal incident irradiance. As biomass density increased and self-shading became more prominent, the aggregated system achieved a better balance between photolimited and photoinhibited regimes when exposed to higher incident irradiance. Furthermore, the addition of scattering particles enhanced light harvesting efficiency, resulting in increased growth rates of C. vulgaris under low incident irradiance. By highlighting the collective improvement in light allocation throughout the hydrogel culture, our findings offer insights for optimizing the design and light use efficiency of photobioreactors and microalgae-based photosynthetic living materials."
} | 1,836 |
21887266 | PMC3162558 | pmc | 9,382 | {
"abstract": "We propose a previously unrecognized kind of informational entity in the brain that is capable of acting as the basis for unlimited hereditary variation in neuronal networks. This unit is a path of activity through a network of neurons, analogous to a path taken through a hidden Markov model. To prove in principle the capabilities of this new kind of informational substrate, we show how a population of paths can be used as the hereditary material for a neuronally implemented genetic algorithm, (the swiss-army knife of black-box optimization techniques) which we have proposed elsewhere could operate at somatic timescales in the brain. We compare this to the same genetic algorithm that uses a standard ‘genetic’ informational substrate, i.e. non-overlapping discrete genotypes, on a range of optimization problems. A path evolution algorithm (PEA) is defined as any algorithm that implements natural selection of paths in a network substrate. A PEA is a previously unrecognized type of natural selection that is well suited for implementation by biological neuronal networks with structural plasticity. The important similarities and differences between a standard genetic algorithm and a PEA are considered. Whilst most experiments are conducted on an abstract network model, at the conclusion of the paper a slightly more realistic neuronal implementation of a PEA is outlined based on Izhikevich spiking neurons. Finally, experimental predictions are made for the identification of such informational paths in the brain.",
"conclusion": "Conclusions There are several points where the spiking model is unrealistic. It would benefit from a simulation with many more neurons and weaker connections between neurons. In this case, it is likely that polychronous groups would be the primitive units (nodes) forming a path [8] , [52] . In other words, the path would be a kind of trajectory in state space, rather than a localized neuronal pathway. Each state in the trajectory would consist of the activation of a polychronous group. This is a further step in abstraction that we hope to consider in later even more realistic models. However, we have used this simplified system to help us think about a preliminary mapping of natural selection onto networks of more general form. We believe this is the genuine novelty of this paper. Also real neuronal networks are recurrent. Preliminary modeling has shown that recurrence produces several problems for the algorithm. If A causes B to fire, and B causes A to fire shortly afterwards, then due to STDP the edibility trace associated with the synapse from A to B is both strengthened and weakened in succession. Therefore, this synapse is not rewarded as much as a chain of synapses would be. Further work is needed to extend PEA to recurrent neuronal networks. Here we have demonstrated that overlapping paths in networks can be a hereditary substrate, yet without being spatially distinct individuals. Paths are capable of evolution by natural selection. Pathway evolution has several features that distinguish it from standard genetic evolution. These all result from the fact that paths overlap. Path overlap may be good or bad, and we have shown that the extent of path overlap can itself be determined by the PEA, see Figure 13 for a clear case of this. The capacity to implement path evolution in the brain with a relatively trivial modification of existing models, lends very strong support to the neuronal replicator hypothesis, that argues that there exist informational replicators in the brain, i.e. autocatalytic entities capable of producing offspring that are correlated with their parent in fitness, and hence capable of accumulation of adaptations by natural selection [24] , [25] , [26] , [27] . Path evolution allows rapid search by activity distributions to modify the frequency of a solution, encoded as synaptic weights. We have not restricted ourselves to a particular cognitive architecture here, but have merely suggested a particular kind of generative variation in neuronal networks that may allow unlimited heredity of information, for a range of possible algorithms. We hope that neuroscientists will be interested in taking the path perspective. For example, in the neurosciences one may attempt to identify paths, and observe their multiplicative and mutational dynamics. One may ask what is the probability of fixation of a novel pathway or edge (synapse) in a real neuronal network as a function of the reward it obtains? Finally, we point out that other implementations are also possible, hence the level of description we have chosen to present the path evolution algorithm. For example, in chemical reaction networks, the internet, or social networks, it is possible that network adaptation takes place by path evolution, a kind of cryptic Darwinism. All that is required is the ability to assign reward to a path, and path growth with bypass mutations. John Maynard Smith's goal applies now as it always has; a major problem for current evolutionary theory is to identify the units of evolution. We claim that the task of identifying the units of neuroevolution is a prescient task for neuroscience, and one that we hope to have defined and contributed to here; showing that units of evolution can overlap thus allowing Darwinian natural selection to operate in a cryptic form in the brain.",
"introduction": "Introduction A unit of evolution as defined by John Maynard Smith is any entity that has multiplication, variation and heredity [1] . If units have differential fitness they can evolve by natural selection. Units of evolution [1] at the same level of selection [2] are generally considered to be discrete non-overlapping individuals, for example, living organisms, B-cells undergoing somatic selection, ribozymes in the RNA world, and binary strings in a genetic algorithm. The mechanism by which the above units multiply with unlimited heredity depends on template replication [3] . The fundamental process of natural selection using explicit multiplication by template replication to copy information is shown in Figure 1 . 10.1371/journal.pone.0023534.g001 Figure 1 One generation of natural selection by template replication. At time t the population consists of 4 individuals with two phenotypes b 1 = 0111 and b 2 = 0101. The frequency of these phenotypes is q 1 = 1 and q 1 = 3. One generation involves template replication (possibly with mutation not shown) and removal of individuals to maintain the same population size. In the above diagram, this results in the same two phenotypes but with different frequencies q 1 = 2 and q 2 = 2 respectively. According to the Price equation, the fact that phenotypic traits covary with fitness causes fitter traits to increase in the population [5] . It was template-replication-based natural selection that inspired John Holland to invent the now famous genetic algorithm [4] . But as we will show, a discrete non-overlapping symbolic sequence, e.g. a ‘genetic’ substrate is not the only kind of unlimited heredity substrate that can be a kind of unit of evolution. This paper proposes an alternative informational substrate (and unit of evolution in a weaker sense) that can accumulate adaptations when in the context of a population of such units, by natural selection, but in the absence of explicit multiplication by template replication of such units. What is more, we propose a realistic physical implementation of these units, which from now on are referred to simply as paths. The notion of a path as a unit of evolution rests on our insight that natural selection need not act between physically independent individuals as shown in Figure 1 . Instead, natural selection can act on paths in a directed graph, e.g. in a neuronal network, if the covariance between the phenotype of that path and the fitness of that path is not outweighed by transmission bias due to mutational exploration, and environmental change [5] . This more general formulation of natural selection was originally discovered by Price, i.e. natural selection takes place when there is covariance between a trait and the probability of transmission of that trait, irrespective of whether that transmission is achieved by explicitly multiplying entities as required by John Maynard Smith or by some other recipe (such as path evolution in which there is no explicit multiplication of paths). Similar ideas have been presented by Steven Frank who uses the generality of Price's formulation to describe Darwinian processes occurring in development and learning [6] , [7] . George Price describes what it is to be a “natural selection cake”. John Maynard Smith describes one way to make the “natural selection cake”. Path evolution is best seen as yet another way to make this class of cake. Paths in a network have some benefits compared with non-overlapping genetic units of evolution. The number of possible paths in a network can be far greater than the number of nodes or edges in the network because each node and edge can be part of many paths [8] . The number of possible paths in a brain-sized network is beyond astronomical; a desirable feature for an informational substrate. The same supra-astronomical property has been described for the more complex organizations known as polychronous groups (stereotyped neuronal spike patterns) observed by Eugene Izhikevich in recent models of spiking neuronal networks with delays [8] . We consider the important relation of neuronal paths to polychronous groups in the Discussion . The majority of this paper examines a special kind of path evolution algorithm, based on a tournament selection genetic algorithm, to show the capacity for paths to act as unlimited heredity informational substrates. Having convinced ourselves that paths in networks (that have some general properties) can indeed exhibit all the crucial behaviours of a unit of evolution, we produce a more realistic neuronal path evolution algorithm based on a spiking neural network with synaptic weights modulated by Dopaminergic reward that preserves these required properties and so also allows natural selection of paths. Using the first abstract model of paths, a standard and extremely unsophisticated genetic algorithm called a microbial genetic algorithm [9] is used to evolve paths in a network, in order to merely demonstrate that paths can in fact act as unlimited hereditary substrates for an evolutionary algorithm. From the population perspective, each path is interpreted as an individual candidate solution, one network consisting of many potentially overlapping paths/candidate solutions. Given appropriate path traversal, weight change and structural plasticity rules (that we will describe in due course) a path may be seen as a unit of evolution in the sense that it exhibits multiplicative growth (although not explicit replication), variation, and heredity. Each path phenotype is associated with a reward that determines whether the edges of that path are strengthened or weakened following traversal. A pair-wise tournament selection genetic algorithm (microbial GA) compares the reward obtained by two paths. The directed edges of the winning path are strengthened, whilst the directed edges of the losing path are weakened. Edges shared by both paths are not changed. Each time a node is activated there is a probability that it will mutate, i.e. produce an alternative route that bypasses that node. This generates the potential for a novel but correlated path with a novel but correlated phenotype. By this process the more frequently traversed paths are responsible for most of the exploration. Nodes that are inactive for some period of time become disconnected. We find that the path-based GA (PEA) compares favourably with the standard gene-based GA on a range of combinatorial optimization problems and continuous parametric optimization problems. However, there are important and interesting differences. For example, the PEA more readily appears to sustain a memory of past selective environments and can store previously discovered characters for reuse in later optimization tasks. Finally, a more realistic neuronal PEA is presented showing for the first time how natural selection can occur in a biologically plausible physical system with unlimited heredity and yet without template replication. What is this paper not? The main aim of this paper is to show that paths can be informational substrates in the brain. It is not to show that a microbial GA acting on paths in the present form is superior to other optimization algorithms in computer science. In fact, we do not believe that this version of a PEA in precisely its present form is implemented in the brain. It is presented here to allow a comparison between standard genetic and novel path based hereditary substrates . The paper is intended to convince the reader that competing, mutating, and crossing over of neuronal paths is a plausible substrate for heredity in the brain, that could potentially be used by a range of possible PEAs. In fact, the more realistic PEA presented at the end of this paper provides a demonstration that more neuronally plausible algorithms could be PEAs, i.e. use paths as hereditary substrates for natural selection. To summarise, methodologically the justification for the comparison between a standard microbial GA and a path based microbial GA is to use a simple (and relatively unsophisticated) genetic algorithm (specifically one that does not require global operations such as explicit sorting of all genotypes) to demonstrate the hereditary capacity of a new kind of informational substrate in the brain. Of-course it may be possible to optimize path phenotypes using other kinds of PEA, even to use path-based information for other algorithms that are not PEAs. At the end of the paper a more realistic PEA is presented that is quite different from the microbial GA, but is still an example of natural selection of paths and hence can be called a PEA. Simple Examples of Paths in Networks \n Figure 2 shows two networks on the left, and all the paths they contain on the right. The top network contains two paths, each of which has a distinct phenotype. The pink path has phenotype 0101 and the green path has phenotype 0111. Unfilled circles represent nodes with node phenotype 0, and filled circles represented nodes with node phenotype 1. The network on the bottom of Figure 2 contains four paths, shown on the right. Three of the paths have the same phenotype (pink, 0101) and one path has phenotype 0111 (green). 10.1371/journal.pone.0023534.g002 Figure 2 Two networks and the paths they contain. Paths with phenotype 0101 are shown in pink. Paths with phenotype 0111 are shown in green. The transition probabilities associated with each edge are marked. Note that here all the outflow transition probabilities from one node sum to one. Path Traversal Note that each directed edge is associated with a weight between zero and one. The sum of weighs out of one node is always normalized to one after any weight change. Weights correspond to transition probabilities (weights) P ij and are used to determine the frequency of a path. The probability of traversal (we will call this the frequency) of that path is the product of the weights P ij along that path. A node can be active or inactive. To generate a path, the start node is activated, and all other nodes are inactivated. In one time-step, the active node will then cause activation of one downstream node, chosen by roulette wheel selection over the outflow weights to all downstream nodes of the active node. The original active node is then inactivated. Therefore, at any one time, only one node is active in the network. This process iterates until the finish node becomes activated, at which point the path has been generated. Given this probabilistic traversal scheme, it is easy to see that both networks at the top and bottom of Figure 2 have the same relative frequency of phenotypes as at time t in the traditional template based natural selection scheme shown in Figure 1 . Each phenotype b , e.g. 0101, we will index with i , giving b i . Each phenotype b i has frequency q i . The frequency q i of a phenotype is defined as the sum of the frequencies of paths with that phenotype b i . The frequency of an individual path is the proportion of times that that particular path is traversed when the start node is stimulated. Note that the fact that two different networks can produce the same frequency of phenotypes (as in the top and bottom networks in Figure 2 ) means there is a redundant (many-to-one) encoding of phenotypes by paths, and this may permit non-trivial neutrality [10] , i.e. the probability distribution of phenotypes reachable by single mutations of paths may differ depending on the underlying configuration of paths that generated them. Later we will see that this allows the network to structure exploration by learning from previous environments. Paths as Units of Evolution Whilst paths exhibit multiplicative growth, but do not explicitly replicate (multiply) in the sense that they do not reconfigure non-overlapping material to take on the same form as a parental entity. The increase or decrease of the frequency of a path occurs because there is strengthening (or weakening) of the transition probabilities P ij along a path. Whether there is strengthening or weakening of these transition probabilities depends on the reward obtained by a path. Paths are units of evolution if multiplicative growth is sufficient, rather than explicit multiplication. Note that because paths are overlapping, the multiplicative growth of one path also is multiplicative growth of parts of other paths. Paths exhibit variation. Variation exists because each path can have a distinct path phenotype constituted by the order of node phenotypes along that path. Paths exhibit heredity by two mechanisms. Firstly, when a path undergoes multiplicative growth by increasing P ij along that path, i.e. when its frequency increases, this results in the increase of the frequency q i of its associated phenotype b i in the population of path phenotypes. Secondly, when a path mutates (to be described later) correlated variability exists because a new path phenotype, whilst not identical to the parental path phenotype, will still resemble the parent's path phenotype because a mutant path is always a short bypass of the parental path and therefore overlaps with much of the parental path, i.e. like begets like. Correlated variability was shown to be a fundamental requirement for evolvability that was lacking in a previous proposal of an alternative to template replication due to compositional inheritance [11] . Hereditary and correlated variation of paths is necessary for them to be units of evolution. Node Mutations The mechanism of pathway mutation is shown in Figure 3A and is based on the idea of quantal synaptic mutation originally developed by Adams [12] and for which evidence has recently been found in the form of activity dependent structural plasticity [13] , [14] . On the top left is seen a mutation of the first node of the network that was previously shown at the top of Figure 2 . Mutants occur with a certain probability, μ , each time a node is activated. A node mutation involves creating a new node at the same layer (drawn in the figures above or below the parent node). The new node has weak initial connection strength from the node that activated the parent node, and a connection of strength 1 to the node that was activated by the parent node. This preserves the original paths, yet creates new alternative paths. Note that ‘creating a new node’ can be equivalent to connecting to and from a previously existing unconnected node, and this will be the neuronal interpretation given in the later more realistic model. The path phenotypes of the alternative pathways will be correlated with the path phenotypes of the paths that contain the node that underwent a mutation. Initially the alternative paths are traversed with low probability, in other words the frequency q i m of a mutant path phenotype b i m in a population of path phenotypes will be low, if that path phenotype did not previously exist in the population. Note that this kind of mutation could not occur in the population shown in Figure 1 . Because a node can be involved in many paths each having different path phenotypes, a single node mutation can change the frequency of many path phenotypes at the same time. This is one of the features that distinguish the path evolution algorithm from a standard genetic algorithm. In some cases this causes interference, but in others this allows constructive guidance of search. We will see in the simulations that the algorithm is capable of controlling the extent of overlap to suit the problem at hand, e.g. in the case where the network is evolved in variable environments, two non-overlapping paths are generated and maintained in memory. 10.1371/journal.pone.0023534.g003 Figure 3 Mutation is implemented using bypass routes. (Part A) A single mutation to the network in Figure 2 produces two new paths and two new path phenotypes (Right). (Part B) 2-point crossover between a winning path (green) and a losing path (red). Path Crossover Path crossover occurs with probability χ whenever two distinct paths differ in reward, see Figure 3B . A weak weight is formed from a random layer in the loosing path to the next layer in the winning path. Another weak weight is formed from a random layer (after the first point of crossover) in the winning path to the next layer in the losing path. Thus, this is a two-point crossover that creates a new weak path that consists of part of the loosing path and part of the winning path. The Evolutionary Theory of Neuronal Paths Evolutionary Dynamics of Paths in Fixed Networks Let us consider the evolutionary dynamics over one generation of the simple network at the top of Figure 2 . The frequency of the pink path is also the frequency of the path phenotype b 1 = 0101 , namely q 1 = 0.25, because only one path has that phenotype. The frequency of the path phenotype b 2 = 0111 is q 2 = 0.75, and is the frequency of the green path. For more complex networks the frequency of a path phenotype will be the sum over the probabilities of taking all paths with that phenotype. So now we have the frequencies of phenotypes in the ancestor generation at time t . Let us assume that b 1 has reward r 1 = 2 and b 2 has reward r 2 = 3 . Ignoring mutation for now, let one generation consist of choosing two paths. Each path is generated according to the roulette wheel traversal method described previously. From these two paths the winning path is chosen as the path with the highest reward associated with it. The probability of choosing path 1 twice is P(1,1) = ( 0.75 ) 2 . The probability of choosing path 1 and path 2 is P(1,2)+P(2,1) = 2 (0.75)0.25 . The probability of choosing path 2 twice is P(2,2) = 0.25 2 . Only when different paths (with distinct path characters) are chosen is a winner and looser defined. Therefore, with probability 0.375 per generation, path 2 will be chosen as the winner and path 1 as the looser. The transition probabilities P ij are then modified as follows. The edges along the winning path (not shared by the losing path) will be strengthened according to the following rule… (1) and the edges along the losing path (not shared by the winning path) will be weakened according to the following rule… (2) for the losing path, followed by normalization over each set of outflow edges for which weights were changed. Specifically, if λ = 0.1 then the weight of the edge to path 1 will decrease from 0.75 to 0.75×0.9 and the weight on the edge to path 2 will increase by 0.25×1.1, which after normalization gives values new transition probabilities 0.71 and 0.29 respectively. By this learning rule the path character with the higher reward increases in the population and the path character with the lower reward decreases. Let us consider a more general formulation of the above dynamics. Mathematica File S1 shows a deterministic model constructed with dynamical equations that captures the essence of natural selection in these path-based systems. A path is a genotype. A node on a path is an allele. A locus consists of all nodes on paths a certain number of nodes away from the start node (i.e. in the same layer). The frequency of a path is the probability that activity passes along that path when the start node is stimulated. The frequency of a phenotype is the probability that that phenotype will be produced when the start node is stimulated. An understanding of the system will involve a description of the dynamics and links between these various concepts. The kind of network at the top of Figure 2 can be considered as a system with one locus and two alleles. The two alleles are the two parallel nodes at the same locus (layer) of each path. Let us set the initial weight to one of these nodes as w 1 and the other weight w 2 = 1−w 1 because the total outflow weight from the common preceding node must sum to 1. Weight change only occurs if two different paths are chosen in the two traversals available in each generation. Therefore, weight change occurs with probability 2w 1 (1−w 1 ) . With probability 1− 2w 1 (1−w 1 ) there is no weight change. Let us assume (without loss of generality) that the winning path (i.e. the path with higher reward) is associated with the node with weight w 1 . Then the new weight at time t+1 of w 1 is given by (3) For initial values w 1 = 0.1, w 2 = 0.9, and λ = 0.1, this gives the dynamics shown in Figure 4 . 10.1371/journal.pone.0023534.g004 Figure 4 Selection between two alleles at one locus. The allele associated with higher reward reaches fixation, whilst the other allele goes extinct. The path (and phenotype) associated with higher reward reaches fixation, whilst the one with the lower reward goes extinct. Now let us consider the more complex network in Figure 3A . Here there are four paths and four phenotypes, or two loci with two alleles at each locus. Let the two weights at the first locus be w 1 = x and w 2 = (1−x) and we two weights at the second locus be w 3 = y and w 4 = 1−y. The frequency of each path is then… (4) Again, the weights associated with the winning path are changed as in (1) and the weights associated with the losing path as in (2) followed by normalization. Consider the cases in which w 1 and w 2 will change. This happens only when the path pairs AC, AD and BC are traversed with probabilities P(AC) = 2 P(A) P(C) , P(AD) = 2 P(A) P(D) and P(BC) = 2 P(B) P(C) , respectively. When the other pairs are traversed, either fitness is identical and there is no change in weights, e.g. (B & D), or the paths do not differ at the w 1 and w 2 edge, e.g when paths (D&C) or (A&B) are taken. Assume that in this case we wish to minimize the number of 1's (filled circles) in each path. Looking at each case in turn then, A beats C, A beats D, and B beats C, and so w 1 will always be strengthened or not changed at all in each generation according to the following equation… (5) Note that w 2 is just 1−w 1 . Similarly, w 3 an w 4 will only change when path pairs AB, AC, and CD are traversed in a generation. Figure 5 shows the vector field of the Δw 1 and Δw 3 for the various possible values of w 1 and w 3 , and the dynamics of allele frequencies and phenotype frequencies over time for initial conditions w 1 = 0.2, and w 3 = 0.1, and λ = 0.1. 10.1371/journal.pone.0023534.g005 Figure 5 Selection at two loci, each locus having two alleles. The two fitter alleles (with weights w 1 and w 3 ) reach fixation whilst the other alleles (w 2 and w 4 ) go extinct. The fittest alleles (w 1 and w 3 ) and the fittest path, A, go to fixation, whilst the other alleles and paths go extinct. As the vector field shows this is inevitable from any initial condition of w 1 and w 3 . Effectively the two alleles are in linkage equilibrium. Linkage Disequilibrium of Alleles in Paths The network in Figure 6A is initially fully connected (in the forward direction). It has two loci, each with two alleles. We show that it is possible to establish linkage disequilibrium by weight change alone. Consider the case where the ordering of reward is 10>01>11 = 00. Mathematica File S1 shows a deterministic model of how the weights x,y and z change over time to send the fittest path 10 to fixation. Alternatively, if the fitness function is 10 = 01>11 = 00, both paths 10 and 01 are maintained at non-zero probability, the ratio depending on the initial value of the weight x . The initially more frequent of the 10 and 01 paths reaches a higher steady state value, see Figure 6B . 10.1371/journal.pone.0023534.g006 Figure 6 The network has three parameters x,y and z, and encodes four paths, A,B,C and D. Part A shows the dynamics of weights and path frequences for the fitness function 10>01>11 = 00. Path 10 (B, green) reaches fixation, and all other paths go extinct. Part B shows the dynamics of paths for the fitness function 10 = 01>11 = 00, for different initial weights of x of 0.6 and 0.4. Non-overlapping paths B and C are maintained at different concentrations that depend on the initial value of x. Paths A and D again go extinct. The capacity to maintain non-random assortment of the alleles by i. maintaining 10 and losing 01 (in the selective case) and ii. by maintaining B and C at different frequencies in the neutral case shows the capacity for linkage disequilibrium. The network converges to make one path in the selective case, and two non-overlapping paths in the neutral case. As we saw in Figure 3A , there are some networks that will not permit the maintenance of linkage disequilibrium because it is impossible to establish two non-overlapping paths because of a node bottleneck. In this case, mutations will be required to produce a greater number of nodes at that locus, so that paths can pass without overlapping with each other, thus maintaining multiple linkage disequilibria (pairwise associations) between alleles at loci on either side of the bottleneck. Assignment of Path Phenotypes The reward obtained by a path is a function of its phenotype b i . The assignment of a phenotype to a path is determined by how the path interfaces with the environment. This ‘environment’ may be an effector system, or another region of the neuronal network. The elucidation of realistic genotype-phenotype maps is as difficult in these systems as it is in evo-devo, however, some suggestions are given. Figure 7 shows some idealized examples of paths and their path characters and how these path characters may be associated with reward in various implementations of pathway evolution. See the Discussion for further implementation details in more realistic neuronal settings. Figure 7 part A shows that neurons may (indirectly) innervate distinct effectors such that a particular path comes to represent a sequence of motor actions, for example, at a high level in a motor system, a sequence of left and right turns may be encoded by a path. In this sense, a binary genotype can be encoded. Figure 7 part B shows that neurons may be organized into a topographic map in which adjacent positions have correlated response functions, and this is a way to encode a continuous valued genotype. Such maps are seen in early visual layers for example in which adjacent neurons have similar response characteristics. Figure 7 part C shows that a more complex kind of path phenotype may be a network of condition-action-(next condition) triplets that encodes a feed-forward model of an environment. The possibilities are in fact endless. 10.1371/journal.pone.0023534.g007 Figure 7 Different ways in which a path can have a phenotype. ( A ) Nodes may indirectly encode motor actions, e.g. a pattern of turns in a maze, or any other binary effector system. In this way a binary path phenotype can be encoded. ( B ) Alternatively the position of a node along the x-axis may determine a real-valued character from −1 to 1. Bypass mutants may be more likely to encompass adjacent neurons, thus producing correlated variability in phenotypes ( C ) An even more complex phenotypic interpretation of a path is to think of the network as an (anticipatory) classifier system [53] that can evolve by a modification of PE if nodes are conditions and edges are actions. A condition (t) – action – condition (t+1) triplet is a classifier. The full details of the PEA are given in the Methods section, and the C++ code is available in Code S1 . The Results section compares the performance of PEA with various parameter settings against the equivalent standard gene based microbial genetic algorithm [9] on various combinatorial and real-value optimization problems, and for evolution in variable environments. Finally a more plausible neuronal implementation of a path evolution algorithm is presented.",
"discussion": "Discussion Is Path Evolution Really Natural Selection? We have described the PEA using the language of natural selection: parameter combinations are ‘phenotypes’, graph modifications are ‘mutations’, increasing path probability is ‘multiplicative growth’, node parameter-values are ‘alleles’ and so on. This arises from our interest in the neuronal replicator hypothesis that considers whether evolutionary computation may be possible in the brain [24] , [25] , [26] , [27] , [28] , [29] . In fact, viewing network processes from the evolutionary perspective was crucial in allowing us to see paths as possible hereditary substrates. We also have a longstanding interest in the origin of life and therefore we notice that the algorithm also shows how natural selection can occur in the absence of template replication in a physical system. Template replication was previously thought to be necessary for natural selection with unlimited heredity [3] . It is not. A skeptic may ask, can a path of activity really legitimately be considered to be a unit of evolution? John Maynard Smith said that group selection requires the existence of cohesive, spatially discrete groups, that “reproduce” by sending out propagules, and that can go extinct (1976, p. 282). He defined a population of units of evolution as “any population of entities with the properties of multiplication (one entity can give rise to many), variation (entities are not all alike, and some kinds are more likely to survive and multiply than others), and heredity (like begets like) will evolve. A major problem for current evolutionary theory is to identify the relevant entities” (p. 222, [30] ). We have identified a path as a unit of evolution, however it is not a spatially discrete physical individual in the way John Maynard Smith imagined, it has multiplicative growth rather than explicit replication. A path is capable of multiplicative growth in the population of paths; however, it does not give rise to a distinct spatially separate entity during growth, but strengthens the probability of traversal of its edges. We have demonstrated that path characters can have variation and heredity, by bypass mutations. So the PEA (in this case, a microbial GA acting on paths) implements something that is similar to and different from a conventional natural selection acting on genetic informational substrates as modeled with the microbial GA acting on discrete non-overlapping genotypes. The differences are as follows… Whilst there are a well-defined number of distinct paths in a physical network, e.g. 4 paths in figure 3 , the relative frequency of a path in the virtual population of paths generated by repeated stimulation of the start node is a probability. In standard natural selection the frequency of a genotype is an integer value. In the PEA a single mutation can affect multiple genotypes whereas in standard natural selection a single mutation can affect only one genotype. Thus, individuals are non-distinct on the evolutionary level. In the PEA, multiplicative growth and selection operators will in general have direct side effects on the prevalence of many genotypes besides those that were directly evaluated under selection. Thus, individuals are non-distinct on the ecological level. The PEA has memory for past environments. Paths that were useful in past environments can be more stably preserved than in the population of a standard microbial GA. The PEA exhibits an automatic capacity for non-trivial neutrality because there is a many to one network to phenotype map with some mappings possessing favorable exploration distributions [10] . This was not an automatic feature of the standard microbial GA. The PEA can automatically establish appropriate linkage disequilibrium by controlling the amount of overlap between paths, and is thus able to solve the HIFF problem where the microbial GA is not. A skeptic may claim that rather than demonstrating that a “true Darwinian process” is possible in the absence of distinct units, the paper suggests that the concept of evolution by natural selection is inherently less well defined than previously assumed. If an evolving population can be an implicit one, then this significantly widens the net of processes that could be described as evolutionary. Perhaps even processes as physically simple as annealing could be given an evolutionary slant in this sense? We disagree. Here it is helpful to consider a classification of optimization algorithms, see Table 1 . 10.1371/journal.pone.0023534.t001 Table 1 A classification of search (generate-and-test) algorithms. Solitary Search Parallel Search Parallel Search with Competition (Price) Parallel Search with Competition and Information Transmission (JMS) (Stochastic) hill climbing/Simulated Annealing Independent hill climbers, e.g. with restart Competitive Learning Genetic Natural Selection Markov Chain Monte Carlo Reinforcement Learning Adaptive Immune System Synaptic Selectionism Genetic Algorithms Neural “Darwinism” Neuronal Replicators In solitary search only one candidate solution is maintained. Examples include hill-climbing and stochastic hill-climbing. Next, it is trivially possible to parallelize solitary search. This we call parallel solitary search, and doing so allows a linear speed up. In a denuded sense this is a population of sorts. Increasing in sophistication one may allow parallel solitary search with competition. Here there is competition for a global search resource that can be reassigned between individual candidate solutions, probably to the currently best candidate solutions, where best may have a potentially complex definition. In effect, there is now a simple ecology of competition between candidate solutions. Into this category falls competitive learning [31] , [32] , Hebbian learning [26] , [33] , many reinforcement learning algorithms (in which the competing units are state-action pairs) [34] , and other action [35] , [36] and attention selection [37] models. However, all these models lack information transmission between candidate solutions. This defines a new category of search called parallel solitary search with competition and information transfer between candidate solutions. Natural selection is the archetypical example of this class of algorithm. We call such a population a Full Population. It converts a competitive ecology into a true evolutionary system. Notice that the Price equation is satisfied even by the third class of search, and so in a sense it is a broader definition of natural selection that does not explicitly require information transmission between solutions. Note that in a genetic natural selection, information transfer between candidate solutions occurs in a fixed population size with mutation alone (crossover is not needed). To see this is the case, imagine there are 10 material slots, each configured as a particular candidate solution. When a candidate solution replicates with mutation, a randomly chosen slot is reconfigured with a mutated configuration generated (copied) from the parent candidate solution. By observing the state of the offspring slot that was reconfigured one can reduce ones uncertainty about the parental solution. Thus there is transfer of information between material slots. The PEA contains a kind of information transfer because paths overlap, and bypasses can connect paths together that were previously unconnected. Our algorithm shares with natural selection in organisms, and artificial selection in genetic algorithms, the following properties: a full population with information transmission between individuals and competition between individuals, unlimited heredity (the capacity for long paths/genotypes), and (for the problems considered) covariance in fitness between parent and offspring (i.e. the capacity for micro-mutation by short path bypasses/mutations). In this sense, it follows the spirit and the letter of the law of natural selection, but uses a novel hereditary substrate that adds a rather strange set of previously unnoticed novel properties. Related Approaches in Computer Science There is a related set of algorithms used in computer science, specifically in evolutionary computation. For example, a class of algorithms exists called estimation of distribution algorithms (EDAs) that do not explicitly represent the individuals in a population at all, instead they maintain a probabilistic description for the probability of an allele occurring at each locus, and the novel solutions are obtained by sampling from this distribution. Whilst the path evolution algorithm may therefore be seen as a kind of EDA, it's method of updating the probability distribution of solutions is quite different from the methods traditionally used in EDAs [38] . As far as we are aware one of the goals of EDAs was to remove “arbitrary” operators such as mutation and crossover. This was not our goal in developing the path evolution algorithm in which we stress the importance of generative operators. The path evolution algorithm has no explicit re-construction of a probability distribution on the basis of only the best individuals at each generation. EDAs suffer from the problem that the estimation of such a distribution may be unreliable for a large problem size, therefore EDAs typically make simplifying assumption that alleles at different loci are in linkage equilibrium , in other words that the probabilities of alleles occurring at separate loci are independent variables (e.g. the univariate marginal distribution algorithm UMDA, the population based incremental learning algorithm PBIL, and the compact genetic algorithm CGA). More sophisticated approaches may consider bivariate dependencies (two locus models) or multiple dependencies. In contrast, the path evolution algorithm can automatically explore multiple dependencies between alleles . It does this by adapting the network structure by using simple local structural operators that could be implemented in a biologically plausible neuronal network. EDAs do not fall into the category of full population search with competition and information flow between solutions, because they exhibit no information flow between solutions, as there is in path evolution. In short the PEA provides a much simpler and more elegant framework that (as we will show) has a plausible neuronal implementation. Ant colony optimization (ACO) algorithms were not inspired by the idea that natural selection might occur in the brain, but by the communication between ants about the best paths to food [39] . Unlike EDAs, ACOs do fall into the category of full population search with competition and information flow between solutions. However, interestingly, a recent survey states that it is still an open research question “how and why the method works” [40] . We believe that our explanation here of the function of the path evolution algorithm is the best explanation so far for how ACO like mechanisms actually work. That is to say, they work by the natural selection of paths. It is remarkable that this explanation appears nowhere in the ACO literature, however it is not entirely surprising for natural selection is often cryptic as an explanation for adaptation in systems that superficially may appear to lack it [41] . ACOs are slightly more complex than the path evolution algorithm because they determine whether a traversal is ‘feasible’ by referring to the phenotype of a node. In the PEA, phenotype “semantics” never influence genotype “syntax”, i.e. there is no “heuristic information” as in ACOs. We hope that the PEA will be welcomed by the ACO community as a general explanation for the adaptive power of ACOs. Note that particle swarm optimization also falls into the category of full population search with competition and information flow between solutions [42] . The information exchanged (replicated) is the memory of the location in N-dimensional space of local optima between particle (slots). Particles are physical slots between which information is exchanged. Strangely, particle swarm optimization also works by a process of Darwinian natural selection in which the replicator is location information. Confusion arises when people think replication is replication of matter rather than of information. Thus, particle swarm optimization is not made Darwinian by replicating particles themselves! The network we maintain in path evolution is a kind of hidden Markov model but with a rather restricted feed-forward topology [43] . The problems for which HMM learning algorithms are used are not optimization problems but supervised learning problems requiring generalization, in the sense that the final set of desired parameters (outputs) are known, e.g. the desired outputs in the training set may be a string of nucleotide sequences. Our problem is slightly different. We have an unknown set of optimal outputs, and we must use immediate reward information to generate a HMM for them. Viewed in this light, this paper provides an algorithm based on natural selection of paths that is able to produce hidden Markov models for optimization problems. For traditional HMM problems, usually, heuristic algorithms such as Baum-Welch (iterative maximum likelihood estimation) [44] are used to produce a model that can generate this known set of desired outputs. These algorithms may get stuck on local optima because they are solitary (gradient climbing) algorithms. Also, they require assumptions about model size and topology. Previously, evolutionary approaches have been used to evolve HMMs for such problems, e.g. for protein secondary structure prediction by Rene Thomson [45] . However, these algorithms maintain multiple separate HMMs and use operators such as add state, remove state, modify state phenotype, add/delete transition, and crossover between distinct HMMs. They evolve an unlimited number of HMM topologies, including recurrent topologies. Also, they add a component of fitness that is linked to the Bayesian Information Criterion to compress the HMM [46] . In contrast our approach is an evolutionary approach for evolving a single non-recurrent HMM containing multiple paths for optimization problems without supervision, and so far, no capacity (yet) for compression guided by BIC. A Slightly More Realistic Neuronal Implementation of a Path Evolution Algorithm The neuronal networks of the brain provide the most natural implementation of paths as informational substrates with unlimited heredity. We produced a model using Izhikevich spiking neuronal networks as described in [47] but with some modifications that are needed to convert the network to run a PEA that is clearly recognizable to the naked eye. \n Figure 16a shows the initial state of the network of regular spiking neurons that form 10 initial paths, each path being stimulated by a start neuron. The initial weights are set to a random value between 15 mV to 60 mV (maximum weight = 60 mV). This allows a path to be created by single neurons. If weights are made weaker, many neurons are required to sustain a path and the system is considerably more complex. 10 ms into each second, the start neuron is externally depolarized causing it to fire. This results in activation passing downstream activating each neuronal layer. Neurons are connected by delay lines of 1 ms, although variable delays can also be used. Background noise is set so that neurons on average fire at 0.1–1 Hz. Synapses are modified by STDP via eligibility traces modulated by DA reward, as in Izhikevich's paper [47] except that eligibility traces decay 10 times faster and reward decays 4 times faster than in the original paper, thus increasing the specificity of reward. 10.1371/journal.pone.0023534.g016 Figure 16 Izhikevich spiking neuronal network modified to include WTA output competition and activity dependent weight decay and plasticity, solving the 10 bit all 1's problem. Red squares = neurons with phenotype 1, Black squares = neurons with phenotype 0. Thickness (and lightness) of green lines = strength of weights from 15 mV to 60 mV (max). The figure below shows the fitness of the path phenotype in each run, and the moving average used to determine whether to give reward or not, over 5000 trials (with 1 trial per second). Winner-Take-All Competition at Outputs An important modification to Izhikevich's model must be made. To implement a hidden Markov model type network it is necessary to limit the outflow of information from one node to just one possible output. In spiking neurons, this is achieved by winner take all (WTA) output competition between all outflow paths for activation. This introduces variation upon which selection can act, and ensures a single path is generated at a time (rather than a tree of spreading activation). This in turn means that only one path is responsible for behavior and hence credit can be specifically assigned to just that path. Effectively, weight proportionate WTA output competition produces a system with very sparse activation, which helps with specific credit assignment. Many neuronal models assume WTA competition, for example self-organization of spike pattern sensitivity in neurons with winner take all (WTA) lateral inhibition and STDP [31] . This competition results in the frequency distribution of single spike outputs matching the weight distribution of output synapses. Phenotypes in Spiking Neuronal Networks based on Spike Order The phenotype of the network is also interpreted differently from Izhikevich [47] . At 10 ms into each second, the start node is stimulated and a sequence of spikes is produced. This sequence is of varying length, i.e. activity may not propagate all the way to the final layer. Each node is assigned a node phenotype (0, or 1) as in the previous PEA. The identities of the first 10 nodes that spike after stimulation at 10 ms is recorded in an array, and the phenotype is defined as the binary string produced by this ordered list of node phenotypes. In the counting 1 s problem, the fitness of this binary string is the number of 1 s contained in it, and this determines the dopamine reward given at 50 ms. Differential Growth and Selection of Pathways by Dopamine Modulated STDP and Weight Decay Reward is also given in a different way to Izhikevich [47] . At 50 ms into each second, reward is given according to the fitness of the path phenotype compared to a running average fitness window. Running average fitness is defined as fitness average(t+1) = 0.01 x fitness average(t)+0.99 current fitness. If the current fitness (i.e. the reward obtained from the path phenotype of the spikes produced between 10 ms and 50 ms into each second) is greater than the average fitness, then reward is given at 0.5 units of DA per correct bit. The fitness function we use is simply the all 1's task, where we wish a sequence of 10 neurons each with a neuron phenotype of one to fire immediately after the 10 ms stimulation, before any neurons with the 0 phenotype fire. If the current fitness is greater is less than the average fitness then a negative reward is given at 0.5 units of DA per incorrect bit. This simple method of assigning reward is primitive when compared to a full TD learning mechanism that modifies reward up and down on the basis of a difference from the predicted reward, but it approximates this and still works. Reward then modulates weights (up and down) on the basis of their eligibility traces. The forth modification to Izhikevich is that in addition to weight change due to DA modulated eligibility traces based on STDP, there is an activity dependent linear weight decay of 0.00002 mV per ms, if a neuron does not fire at all in one second; this results in a weight reaching the minimum permitted weight of 15 mV within approximately a minute if it does not fire at all. Activity Dependent Structural Plasticity Implements Mutation The fundamental operation of node mutation and pathway crossover that the PE algorithm depends upon is closely related to the synaptic pathway mutations first proposed by Adams [12] in which Hebbian learning is noisy, i.e. when a synapse is strengthened there is also a small probability that synapses will be strengthened from the pre- or the post- synaptic neuron to or from nearby neurons. Adams' insight prefigured the recent discovery of rapid structural plasticity; the formation and breakage of synapses in the order of minutes [14] , [48] , [49] . These operations are eminently suitable for implementation of the bypass mutations required for neuronally plausible path evolution. In real neuronal networks it is possible that path mutations will be able to shortcut several layers, or add layers, producing variable path lengths. Also, recurrent paths may come to exist. However, for purposes of demonstration here we chose to add a simple kind of activity dependent structural plasticity to Izhikevich's model that is constrained in the topology of connectivity that is possible by mutation. Whenever a neuron is active there is a 1% probability that it will produce a new synapse to an adjacent neuron (i.e above, same or row below) in the next layer (column). The neuron to which this new output passes also produces a new output randomly to an adjacent neuron in its next layer, thus there is a 1/3 rd probability that a bypass mutant is produced, and a 2/3 rd probability than a divergent mutation (crossover) is produced that does not return to the original path. If a weight decreases below the minimum level of 15 mV it is removed. No neuron may have more than three output synapses. Whenever a new synapse is formed, if the total weight of synapses out of a neuron exceeds 60 mV, then one synapse is removed from the output synapses of that neuron, in inverse proportion to its weight. These activity dependent structural plasticity rules bias synaptic exploration to those neurons that are currently most active. To overcome the limitation that the random generation of node phenotypes may produce a matrix of neurons that does not contain a single possible path of all 1's from the start node to one of the final layer nodes, we allow random node phenotype bit flipping at a low rate, e.g. once every 1 minute iff that neuron has not spiked once in this time. \n Figure 16 shows an evolutionary experiment conducted with a realistic neuronal implementation of pathway evolution that uses Izhikevich spiking neurons, WTA competition, structural plasticity and dopaminergic reward to evolve pathways. This simple demonstration shows that pathway evolution can be expected to carry over to more realistic neuronal implementations using more realistic reinforcement learning kinds of reward allocation. The network of pathways can be seen as overlapping models to which reinforcement can be given [50] , [51] . What is special here is that the models are evolved in a realistic spiking neuronal network. The software for running the above simulations can be downloaded from Code S1 . Conclusions There are several points where the spiking model is unrealistic. It would benefit from a simulation with many more neurons and weaker connections between neurons. In this case, it is likely that polychronous groups would be the primitive units (nodes) forming a path [8] , [52] . In other words, the path would be a kind of trajectory in state space, rather than a localized neuronal pathway. Each state in the trajectory would consist of the activation of a polychronous group. This is a further step in abstraction that we hope to consider in later even more realistic models. However, we have used this simplified system to help us think about a preliminary mapping of natural selection onto networks of more general form. We believe this is the genuine novelty of this paper. Also real neuronal networks are recurrent. Preliminary modeling has shown that recurrence produces several problems for the algorithm. If A causes B to fire, and B causes A to fire shortly afterwards, then due to STDP the edibility trace associated with the synapse from A to B is both strengthened and weakened in succession. Therefore, this synapse is not rewarded as much as a chain of synapses would be. Further work is needed to extend PEA to recurrent neuronal networks. Here we have demonstrated that overlapping paths in networks can be a hereditary substrate, yet without being spatially distinct individuals. Paths are capable of evolution by natural selection. Pathway evolution has several features that distinguish it from standard genetic evolution. These all result from the fact that paths overlap. Path overlap may be good or bad, and we have shown that the extent of path overlap can itself be determined by the PEA, see Figure 13 for a clear case of this. The capacity to implement path evolution in the brain with a relatively trivial modification of existing models, lends very strong support to the neuronal replicator hypothesis, that argues that there exist informational replicators in the brain, i.e. autocatalytic entities capable of producing offspring that are correlated with their parent in fitness, and hence capable of accumulation of adaptations by natural selection [24] , [25] , [26] , [27] . Path evolution allows rapid search by activity distributions to modify the frequency of a solution, encoded as synaptic weights. We have not restricted ourselves to a particular cognitive architecture here, but have merely suggested a particular kind of generative variation in neuronal networks that may allow unlimited heredity of information, for a range of possible algorithms. We hope that neuroscientists will be interested in taking the path perspective. For example, in the neurosciences one may attempt to identify paths, and observe their multiplicative and mutational dynamics. One may ask what is the probability of fixation of a novel pathway or edge (synapse) in a real neuronal network as a function of the reward it obtains? Finally, we point out that other implementations are also possible, hence the level of description we have chosen to present the path evolution algorithm. For example, in chemical reaction networks, the internet, or social networks, it is possible that network adaptation takes place by path evolution, a kind of cryptic Darwinism. All that is required is the ability to assign reward to a path, and path growth with bypass mutations. John Maynard Smith's goal applies now as it always has; a major problem for current evolutionary theory is to identify the units of evolution. We claim that the task of identifying the units of neuroevolution is a prescient task for neuroscience, and one that we hope to have defined and contributed to here; showing that units of evolution can overlap thus allowing Darwinian natural selection to operate in a cryptic form in the brain."
} | 15,042 |
31942258 | PMC6955109 | pmc | 9,384 | {
"abstract": "Background Interactions among species are a driving force of community structure. The species composition of animal-plant interaction networks can be highly dynamic on a temporal scale, even though the general network structure is usually not altered. However, few studies have examined how interaction networks change over long periods of time, particularly after extreme natural events. We analyzed herein the structure of the hummingbird-plant interaction network in a dry forest of Chamela, Mexico, comparing the structure in 1985–1986 with that in 2016–2017 following the passage of two hurricanes (category 2 Jova in 2011 and category 4 Patricia in 2015). Methods The fieldwork was carried out in the Chamela-Cuixmala Biosphere Reserve in Jalisco, Mexico. In the last 30 years, three severe drought events and two hurricanes have affected this region. Previously, from 1985–1986, hummingbird-plant interactions were recorded monthly for one year in the study area. Then, from 2016–2017, we replicated the sampling in the same localities. We compared the network parameters describing the plant-hummingbird interactions of each period using adjacency matrices. Results We found differences in the number and identity of interacting species, especially plants. The plant species missing in 2016–2017 were either the least connected in the original network (1985–1986) or belonged to groups such as cacti, epiphytes, or trees. The new plant species incorporated in the 2016–2017 network were herbs, vines, and shrubs, or other species barely connected. These changes in the composition are consistent with reports on vegetation damage after strong hurricanes at other study sites. Conversely, all hummingbird species remained in the network, with the exception of Heliomaster constantii , which was primarily connected to a plant species absent in the 2016–2017 network. Migratory and habitat generalist species (i.e., Archilochus spp . ) showed higher abundances following the disturbance events. Conclusions Most of the parameters describing the hummingbird-plant network structure remained unchanged after 30 years, with the exception of an increase in plant robustness and hummingbird niche overlap. However, the network’s generalist core was affected by the loss of some species. Also, core plant species such as Ipomoea bracteata , Combretum farinosum , and Justicia candicans were found to be important for maintaining the hummingbird-plant interaction network. The temporal perspective of this study provides unique insights into the conservation of plant-hummingbird networks across time and extreme natural events.",
"conclusion": "Conclusions The number and identity of interacting species in the hummingbird-plant network of a dry forest in Chamela, Mexico, primarily plant species, changed over a 30-year period. We mostly observed changes in the less connected species and those more prone to be affected (e.g., epiphytes, cacti, and large trees) or benefitted (e.g., herbs, vines, and shrubs) by hurricanes. In the case of hummingbirds, all but one ( Heliomaster constantii ) remained in the network over time. Overall, most network structural parameters remained unchanged, with the exception of plant robustness and hummingbird niche overlap, which both increased. Also, extreme events appear to have impacted the generalist core of species, resulting in the loss of several species. Finally, the most important plant species for conserving the plant-hummingbird interaction network in the Chamela-Cuixmala Biosphere Reserve are Ipomoea bracteata , Combretum farinosum , and Justicia candicans . The temporal perspective of this study provides unique insights into the conservation of plant-hummingbird networks across time and after extreme natural events.",
"introduction": "Introduction Drought and extreme winds are the main natural disturbances in Neotropical dry forests ( Durán et al., 2002 ). Natural factors such as these have been found to affect plant–pollinator interactions. It is likely that natural disturbances such as drought and hurricanes will only continue to intensify in the near future as a result of climate change ( Allen et al., 2010 ; Knutson et al., 2015 ). Therefore, it is important to investigate the effects of these natural disturbances on biodiversity and ecosystem functioning, including plant-animal interactions. Interactions among species are the driving force behind the biodiversity of the Earth ( Bascompte, 2009 ) as well as the structure of its communities ( Berlow et al., 2009 ).These interactions can be represented as networks, enabling the topology (or patterns) of the relationships among species to be characterized. Accordingly, these networks provide a simple and powerful framework for understanding species interactions ( Poisot, Stouffer & Gravel, 2015 ). Additionally, network theory has informed several approaches for statistically quantifying and comparing patterns across communities ( Bascompte & Jordano, 2007 ). These approaches are essential for understanding the ecological, evolutionary, and co-evolutionary dynamics of a network that cannot simply be deduced from pairs of interacting species ( Bascompte, 2009 ). At a basic level, network structure reflects the organization of species interactions according to the frequency of interactions per node ( Proulx, Promislow & Phillips, 2005 ; Bascompte, 2009 ). Overall, species interactions tend to be highly heterogeneous: Few species have many interactions, and many species have only a few interactions. In particular, plant–pollinator networks have several common features ( Bascompte et al., 2003 ; Lewinsohn et al., 2006 ; Guimarães et al., 2007 ) as well as specific ecological characteristics and structure ( Bascompte et al., 2003 ; Lewinsohn et al., 2006 ; Guimarães et al., 2007 ; Morales & Vázquez, 2008 ; Vázquez et al., 2009 ). For example, plant–pollinator networks tend to be nested: Generalist species mainly interact with other generalists, forming a core, and specialists also mainly interact with generalists and rarely with other specialists ( Bascompte et al., 2003 ). These patterns of interaction can be explained by ecological processes and evolutionary history ( Poisot, Stouffer & Gravel, 2015 ; Díaz-Castelazo et al., 2010 ). However, there is a scarcity of research on how networks change over long periods of time, especially after extreme natural events ( Dupont et al., 2009 ; Díaz-Castelazo et al., 2010 ). The few studies examining temporal changes in species interaction networks have reported that these networks retain basic topological properties even when plant–pollinator interactions vary over time ( Alarcón, Waser & Ollerton, 2008 ). Notably, species interaction networks that show high species turnover and low variation in their structural parameters ( Petanidou et al., 2008 ; Dupont et al., 2009 ; Luviano et al., 2018 ) may be less sensitive to disturbance effects as a result of climatic changes and hurricanes. Natural events have several wide-ranging effects on natural vegetation that may affect animal-plant interaction networks. In dry forests, most plant species are adapted to drought ( Meir & Pennington, 2011 ). However, extreme winds during hurricanes can result in vegetation damage or fallen trees, which create forest gaps ( Askins & Ewert, 1991 ). Also, vegetation may be severely defoliated, and most flowers and fruits may be dropped ( Lynch, 1991 ; Waide, 1991 ). Such changes to the original canopy structure may further affect the production of flowers, the primary food resource for pollinators ( Askins & Ewert, 1991 ). Following natural events, greater (major) defoliation has been observed in the superior canopy strata compared to the lower canopy strata and understory, where re-growth tends to be faster ( Wunderle Jr, Lodge & Waide, 1992 ; Brokaw & Walker, 1991 ). Aside from vegetation damage, a hurricane can kill birds as a result of wind and rain exposure ( Wiley & Wunderle, 1993 ). Surviving birds can be debilitated and subjected to additional mortality or predation ( Lugo, 2008 ). After Hurricanes Gilbert and Hugo, several studies reported that resident nectarivorous birds were among the most affected ( Askins & Ewert, 1991 ; Lynch, 1991 ; Waide, 1991 ; Wunderle Jr, Lodge & Waide, 1992 ). Nonetheless, the subsequent recovery of their populations to previous numbers suggests that reductions in their populations were due to regional movements as individuals searched for food rather than mortality ( Waide, 1991 ). After Hurricane Gilbert in Jamaica Wunderle Jr, Lodge & Waide (1992) found that nectarivorous species diminished their numbers in the affected areas yet increased their numbers in nearby, less affected areas. With respect to hurricanes, most ecological studies have focused on their impacts on vegetation ( Baldwin et al., 1995 ; Lomascolo & Aide, 2001 ), animal communities ( Covich et al., 1991 ; Lugo, 2008 ), or both ( Ferguson et al., 1995 ; Pascarella, 1998 ; Rathcke, 2001 ; Angulo-Sandoval et al., 2004 ; Horvitz, Tuljapurkar & Pascarella, 2005 ) although, as previously mentioned, few have documented the effects on species interaction networks ( Sánchez-Galván, Díaz-Castelazo & Rico-Gray, 2012 ; Luviano et al., 2018 ). In one case, Sánchez-Galván, Díaz-Castelazo & Rico-Gray (2012) studied the effect of Hurricane Karl on a plant-ant interaction network in Veracruz, Mexico. These authors found that most changes occurred in the presence and proportions of interacting species. Also, Luviano et al. (2018) documented the impact of Hurricane Jova on plant-herbivore interaction networks along a successional chronosequence in a dry forest of the Chamela region in Jalisco, Mexico. These authors found that some network descriptors were altered; however, the overall network topology associated with forest succession remained unaltered. We studied herein the hummingbird-plant interactions in a dry forest of Chamela, Mexico. Our aim was to assess temporal changes in (1) the number and identity of interacting species and (2) the topology of the interaction network over a 30-year period during which several disturbances occurred. Considering the results of previous studies ( Alarcón, Waser & Ollerton, 2008 ; Petanidou et al., 2008 ; Dupont et al., 2009 ; Díaz-Castelazo et al., 2010 ; Dupont et al., 2009 ; Luviano et al., 2018 ), we expected that the hummingbird-plant interaction network in the dry forest of the Chamela Biological Station would maintain most hummingbird species over time and that greater changes would be observed in plant species (including the generalist core), mainly due to changes in vegetation cover caused by the passage of Hurricanes Jova and Patricia. As a consequence, we also expected changes in the number and identity of some interactions (i.e., more connections between hummingbirds and lower-strata plants such as shrubs, vines, and herbs and fewer connections between hummingbirds and higher-strata plants such as large trees),yet we expected the general network structure to remain unaltered (i.e., that most network parameters would remain unchanged).",
"discussion": "Discussion We found that natural disturbance events such as hurricanes may have impacted some of the structural parameters of the studied hummingbird-plant network in Chamela, Mexico. Plant robustness and hummingbird niche overlap increased after the disturbance events, and the generalist species core of the network was reduced in number. However, overall, the number of species in the interaction network after the disturbance events was larger than in the prior survey because of the inclusion of new plant species, even though many of the plant species present in the first survey were lost. Notably, four out of five hummingbird species were retained in the second survey, with the exception of Heliomaster constantii . Temporal changes in the number and identity of species forming interaction networks have been previously reported. For example, a three-year study in California (United States) registered changes in species interactions, although few differences were noted in the number of plants and pollinators ( Alarcón, Waser & Ollerton, 2008 ). Dupont et al. (2009) studied pollination networks across different latitudes and found that species identity and interactions changed from one year to the next, although the number of species and interactions remained constant. Petanidou et al. (2008) reported that most species and interactions changed over a four-year period near Athens (Greece): Only 53% of plant species, 21% pollinator species, and 4.9% of interactions remained constant. Díaz-Castelazo et al. (2010) showed that, over a period of 10 years, ant-plant networks in Veracruz (Mexico) differed in the number of species and interactions, suggesting a high species turnover. These findings suggest that the actual range and strength of interactions in the plant and pollinator community vary over time and are unlikely to be detected over short periods of time ( Alarcón, Waser & Ollerton, 2008 ). Long-term studies on plant–pollinator networks are needed, especially after disturbance events such as droughts and hurricanes, as the composition of interacting species can change following these events. Additionally, the modification of plant abundance, flowering, or pollinator phenology as a result of climatic changes can affect the plant–pollinator network ( Burkle & Alarcón, 2011 ). 10.7717/peerj.8338/table-2 Table 2 Parameters of the plant-hummingbird interaction network in the 1985–1986 and 2016–2017 sampling periods. Numbers in bold indicate significant values ( P ≤ 0.05). Asterisks (*) indicate significant differences ( P ≤ 0.05) in a specific attribute between periods. Period 1985–1986 P 2016–2017 P Plant species 21 27 Hummingbird species 5 4 Connectance 0.43 <0.05 0.42 <0.05 Robustness (hummingbirds) 0.68 0.42 0.62 0.29 Robustness (plants) 0.87 * 0.42 0.93 * 0.29 Specialization H 2 ’ 0.41 0.05 0.05 0.43 Niche overlap (hummingbirds) 0.27 * 0.09 0.41 * 0.41 Niche overlap (plants) 0.42 0.09 0.36 0.41 Nestedness (NODF) 48.61 0.48 41.88 0.66 Two epiphytic plant species were lost in the 2016–2017 interaction network including T. paucifolia , a generalist core species in the 1985–1986 network. Extreme events can cause the direct mortality of epiphytic plants or damage their host trees ( Robertson & Platt, 2001 ). In Chamela’s dry forest, droughts are part of the intra- and interannual dynamics of the region ( Martijena & Bullock, 1994 ; Maass et al., 2017 ); thus, the inhabiting species are likely well adapted to these events. However, the abundance of epiphytic species was impacted, for example, after Hurricanes Andrew, Hugo, and Wilma in the Caribbean ( Migenis & Ackerman, 1993 ; Loope et al., 1994 ; Goode & Allen, 2008 ). As a result of climate change and the corresponding predicted increase in hurricane incidence, epiphytic plants may disproportionally suffer compared to other understory plant species. Cacti adapted to drought conditions such as Nopalea karwinskiana and Opuntia excelsa also disappeared from the 2016–2017 network. The first cactus was restricted to rocky creek edges yet, after unusually heavy rains in 1998, it completely disappeared (MC Arizmendi, pers. comm., 1998). It was a periphery species but was important for H. constantii ( Arizmendi & Ornelas, 1990 ), which was not registered in the second survey. Many O. excelsa individuals (a generalist core plant species, particularly important for H. constantii and A. rutila hummingbirds) fell or lost their terminal cladodes (those that produce flowers) due to wind force and felled trees after Hurricane Patricia (S Díaz Infante, pers. comm., 2016), as reported for other similar events ( Frangi & Lugo, 1991 ). The effects after Hurricane Jova were probably similar but likely occurred at a lower intensity due to the lower wind force of this hurricane. Most of the rest of missing plants in the 2016–2017 network had few interactions and were periphery species loosely connected to hummingbirds, so their probability of interaction with hummingbirds was already low ( Borgatti & Everett, 2000 ). Notably, blooming was not observed in tall trees such as Tabebuia donell-smithii and P. mangense during the 2016–2017 survey. These species are more susceptible to damage and mortality as a result of strong winds during hurricanes ( Wurman & Winslow, 1998 ; Segura et al., 2002 ; Vandecar et al., 2011 ). Trees such as Erythrina lanata and Ceiba aesculifolia , located on slopes and hills, did not bloom either after Hurricane Patricia, as most of their treetops were removed after the hurricane. This probably also occurred with Hurricane Jova. Parker et al. (2017) found that both hurricanes altered the forest structure in Chamela, but Hurricane Patricia had a larger and more devastating effect on fruit and flower abundance and phenology than Hurricane Jova ( Renton et al., 2017 ). Most of the new plant species added to the network are mainly pollinated by insects. They had few interactions with hummingbirds, and none belonged to the generalist core. Thus, the probability of detecting their interactions before was also low ( Van der Pijl, 1961 ; Tripp & Manos, 2008 ). Moreover, most new species were vines or shrubs (8 spp.), which likely benefitted from the newly opened gaps after Hurricane Patricia. For example, even though the upper strata (4 to 11 m) lost up to 83% ± 16% of canopy material, the lower strata (1 to 3 m) increased its cover by 61% ± 68% after the hurricane, mainly due to the growth of vines and shoots ( Parker et al., 2017 ). One unique case is Tabebuia rosea , a tree species damaged by wind but included as a core species in the second survey; most registered individuals were planted after the first survey (JM Verdusco, pers. comm., 1987). Overall, our data concur with reports in other study sites affected by hurricanes. For example, the number of fruits and flowers on woody plants was abnormally low several months after Hurricane Hugo impacted the Virgin Islands ( Askins & Ewert, 1991 ); Waide1991). Likewise, vines, branches, and invasive herbs formed an almost impenetrable mass in the understory of a tropical forest in Quintana Roo, Mexico, several months after the arrival of Hurricane Gilbert ( Lynch, 1991 ). After Hurricane Hugo in Jamaica, the lower strata of foliage recovered faster than the upper strata because of the more rapid growth of herbaceous plants and the sprouting of trees and shrubs. Even after 17 months, treetops had not recovered their original density due to fallen trees and broken trunks and branches ( Wunderle Jr, 1995 ). Similarly, after Hurricane Gilbert in Puerto Rico, understory growth compensated for the loss of treetops ( Wunderle Jr, Lodge & Waide, 1992 ). However, despite the severe damage to vegetation, the species composition of the studied plant communities was speculated to remain unaltered after Hurricanes Hugo and Gilbert ( Brokaw & Walker, 1991 ). In our study, an important fraction of the species absent in the 2016–2017 network were either susceptible to hurricane damage or had low interaction probabilities. Of the newly added species, many had few interactions or were herbs, vines, or shrubs, which appear to have benefitted from the passage of the hurricanes. According to Olesen et al. (2008) , the new species added in a network are usually specialized or often rare. As with plants, the hummingbird species that were originally more connected ( A. rutila , C. latirostris , and C. auriceps ) were maintained in the network, whereas the species with fewer interactions ( H. constantii ) and those associated with disturbance-vulnerable plant species such as O. excelsa , C. aesculifolia , and N. karwinskiana were absent. Amazilia rutila , the only core species in the first network, was closely associated with O. excelsa , a plant affected by hurricane winds. Previous studies have shown that frugivorous and nectarivorous birds are more susceptible to a decline in food resources than other birds, as occurred on the Virgin Islands with the hummingbird Orthorhynchus cristatus after Hurricane Hugo ( Askins & Ewert, 1991 ) and in the Yucatán Peninsula in Mexico with nectarivorous birds in general after Hurricane Gilbert ( Lynch, 1991 ). Also, rare hummingbird species are more susceptible to disappearance from an affected area, as occurred with the Vervain Hummingbird ( Mellisuga minima ) in Jamaica after Hurricane Gilbert. However, even abundant hummingbirds, such as the Streamertail Hummingbird ( Trochilus polytmus ), may suffer considerable mortality from exposure to hurricanes or may wander to nearby less damaged areas( Wunderle Jr, Lodge & Waide, 1992 ), especially considering that tropical nectarivorous birds have a wide range and often migrate seasonally and altitudinally in some tropical regions (e.g., Stiles, 1983 ). On the other hand, the two Archilochus species seemed to benefit from hurricane-borne disturbance in our study. These generalist winter migrant hummingbirds are commonly associated with disturbed areas ( Finch, 1991 ; Udvardy & Farrand, 1994 ; Arizmendi & Berlanga, 2014 ) and mainly associated with the vine Ipomoea bracteata . Previously, according to Lynch (1991) , overwintering Nearctic migrants appeared to be more resilient than year-round residents to hurricane effects. The latter author found that these migrants, who are generally associated with edges, dense scrub, and open zones, invaded a forest damaged by Hurricane Gilbert in Quintana Roo, Mexico. Thus, it seems that migratory generalist species are favored by disturbance events and the growth and blooming of vines and shrubs. With respect to the parameters of the hummingbird-plant interaction network, only plant robustness and hummingbird niche overlap presented significant differences after 30 years. Several studies describing temporal patterns in pollination networks revealed similarities ( Burkle & Alarcón, 2011 ). For example, most studied pollination networks have been shown to be highly nested over time ( Alarcón, Waser & Ollerton, 2008 ; Petanidou et al., 2008 ; Dupont et al., 2009 ; Díaz-Castelazo et al., 2010 ). Connectance values also tend to be conserved over time ( Alarcón, Waser & Ollerton, 2008 ; Olesen et al., 2008 ; Petanidou et al., 2008 ; Dupont et al., 2009 ; Díaz-Castelazo et al., 2010 ). As sampling time increases, specialization tends to diminish ( Petanidou et al., 2008 ) because pollinators may use less preferred floral resources when their favorite plant species are scarce or absent ( Burkle & Alarcón, 2011 ). However, interaction network characteristics appear to remain constant within a framework of annual variation in species identity ( Petanidou et al., 2008 ). In our study, the generalist core of nine plants in 1985–1986 was reduced to five in 2016–2017; only three of the original species were conserved. Also, the only former core hummingbird species ( A. rutila ) lost this status in 2016–2017, probably due to the disappearance of two core plant species ( O. excelsa and C. aesculifolia ) that served as resources for this hummingbird, and H. constantii was also lost after the hurricanes. In southern California (United States), the identity of core species also changed in plant–pollinator networks studied over three years ( Alarcón, Waser & Ollerton, 2008 ), indicating that specialists interact with different groups of generalists in different years given that even generalist species are subject to fluctuations in population size. The fact that mutualistic networks form well-defined and predictable patterns of interdependence supports a community-wide approach for species conservation ( Bascompte & Jordano, 2007 ). Network analyses can contribute to conservation strategies at the community level and can highlight important species for the network that should be conserved and protected ( Lara-Rodríguez et al., 2012 ). In addition, the most relevant species in terms of abundance and interaction frequencies can provide insights into the ecological and evolutionary history of interaction networks ( Bascompte & Jordano, 2007 ). Theoretical evidence concludes that heterogeneous, nested, mutualistic networks confer network robustness to species loss and habitat fragmentation ( Memmott, Waser & Price, 2004 ; Fortuna & Bascompte, 2006 ; Burgos et al., 2007 ). However, overall robustness to random species extinctions may be partly explained by the role of a few highly connected species in a network’s core with which specialists interact ( Bascompte, 2009 ). Thus, the robustness of an entire network depends on these species. In Chamela, I . bracteata , C. farinosum , and J. candicans remained as core species overtime, even after extreme events. Therefore, these species significantly contributed to the connectivity and resilience of the studied plant-hummingbird network ( Sánchez-Galván, Díaz-Castelazo & Rico-Gray, 2012 ). Moreover, an increase in hummingbirds’ niche overlap suggests that generalist pollinators can expand their diet (e.g., Archilochus spp . and C. auriceps ) when resources are scarce ( Waser, 1986 ), thereby increasing plant robustness. In conclusion, it appears that, despite the disturbances caused by hurricanes, most plant species survive ( Burgos et al., 2007 ), especially if the remaining pollinators become more generalist. We also considered other factors that could have caused the observed changes in the interaction networks, even though we did not register plant-hummingbird interactions immediately before the hurricanes. However, to the best of our knowledge, no other natural or human-induced disturbances could be directly related to the observed changes in this well-studied site (besides the ones already discussed). On the other hand, we have previous evidence supporting that hurricanes shape forest structure, influence their species composition and diversity, and regulate their functioning ( Lugo, 2008 ). As stated before, some severe droughts were recorded at the study site, but plants and animals native to the area are adapted to a highly unpredictable and strongly seasonal rainfall pattern ( Maass et al., 2017 ). Drought can affect plants’ flowering, but its effects in dry forests are usually restricted to the seasons/years when it occurs and the following year/season. Thus, if plants are able to resist drought, there is usually not a lasting effect on flowering. Furthermore, in the last 33 years, there have not been any reported vegetation changes in the area due to drought. Thus, we can assume that the reason that many plants or trees did not bloom during the second survey was hurricane damage rather than drought, especially considering that these effects are not normal in the study site. Finally, we were able to directly observe damage to the vegetation cover that was also reported in the literature ( Maass et al., 2017 ; Jimenez-Rodríguez et al., in press ; Parker et al., 2017 ). Even so, distinguishing the effects of a large, infrequent event from the background variation of a seasonal ecosystem is challenging ( Parker et al., 2017 ). Thus, we recognize that only longer-term studies would evidence network fluctuations due to normal inter-annual variation in the abundance and phenology of plant and hummingbird populations. Also, a recovery of vegetation and decrease in new opportunistic species would confirm that natural disturbances had been influencing the interaction network. However, some external factors not considered herein may also contribute to the observed interactions (e.g., external effects on migratory hummingbirds). Either way, if plant–pollinator interactions are dynamic and opportunistic, this should increase the resistance of pollination networks to species loss and phenological changes ( Alarcón, Waser & Ollerton, 2008 ). Thus, even in the face of extreme events, these networks may persist."
} | 7,016 |
30190707 | PMC6115492 | pmc | 9,385 | {
"abstract": "Marine sediments are important sites for global biogeochemical cycling, mediated by macrofauna and microalgae. However, it is the microorganisms that drive these key processes. There is strong evidence that coastal benthic habitats will be affected by changing environmental variables (rising temperature, elevated CO 2 ), and research has generally focused on the impact on macrofaunal biodiversity and ecosystem services. Despite their importance, there is less understanding of how microbial community assemblages will respond to environmental changes. In this study, a manipulative mesocosm experiment was employed, using next-generation sequencing to assess changes in microbial communities under future environmental change scenarios. Illumina sequencing generated over 11 million 16S rRNA gene sequences (using a primer set biased toward bacteria) and revealed Bacteroidetes and Proteobacteria dominated the total bacterial community of sediment samples. In this study, the sequencing coverage and depth revealed clear changes in species abundance within some phyla. Bacterial community composition was correlated with simulated environmental conditions, and species level community composition was significantly influenced by the mean temperature of the environmental regime ( p = 0.002), but not by variation in CO 2 or diurnal temperature variation. Species level changes with increasing mean temperature corresponded with changes in NH 4 concentration, suggesting there is no functional redundancy in microbial communities for nitrogen cycling. Marine coastal biogeochemical cycling under future environmental conditions is likely to be driven by changes in nutrient availability as a direct result of microbial activity.",
"introduction": "Introduction Marine sediments play a vital role in global biogeochemical cycling, particularly in terms of carbon, nitrogen and oxygen dynamics (Glud, 2008 ). The predicted global climate change scenarios (IPCC, 2014 ) will result in marine sediments being subjected to many environmental pressures, e.g., increasing mean temperature, greater temperature fluctuation, and increasing CO 2 levels (ocean acidification: OA) (Doney et al., 2009 ; Dossena et al., 2012 ). As a direct consequence of rising atmospheric carbon emissions, global average temperature is expected to rise by ~4°C by 2100; and ocean pH, as a result of acidification, is predicted to decline to 7.8 in the same time period (0.2 pH units lower than pre-industrial levels) (Caldeira and Wickett, 2003 ; Kroeker et al., 2013 ; IPCC, 2014 ). It is recognized that many of the key ecosystem services (Beaumont et al., 2007 ) provided by marine benthic habitats are driven by microbial activity (Prosser and Head, 2007 ; Bertics and Ziebis, 2009 ; Gilbertson et al., 2012 ), such as the nitrogen fixation carried out by the cyanobacteria genera Trichodesmium and Crocosphaera (Hutchins et al., 2013 ). Biogeochemical cycling within sediments, and at the sediment water interface, varies with sediment type (Aldridge et al., 2017 ; Hicks et al., 2017a ), and this is reflected in the different microbial communities (Currie et al., 2017 ; Kitidis et al., 2017 ). Cohesive coastal sediments, such as those found in estuaries and intertidal mudflats, tend to have a high organic carbon content, and the sediment biogeochemical cycling is heavily influenced by diffusive processes (Hicks et al., 2017a ). Considering the contribution of benthic microbes to ecosystem services (Bell et al., 2005 ), particularly biogeochemical cycling (Dyksma et al., 2016 ), it is vital that we understand how microbial population dynamics are likely to shift under future climate change scenarios, and how this may affect ecosystem service provision. Climate driven changes, such as warming and elevated CO 2 , are known to alter many biogeochemical cycles, such as the nitrogen cycle (nitrification and ammonia oxidation) (Kitidis et al., 2011 , 2017 ) which are mediated by microbial assemblages (Hutchins and Fu, 2017 ). There is substantial evidence that benthic systems will respond to predicted changes in temperature and CO 2 ; both on an ecosystem and individual species level (Bulling et al., 2010 ; Hicks et al., 2011 ; Godbold and Solan, 2013 ; Cartaxana et al., 2015 ). Individual stressor studies have shown how warming elicits varied responses in microbial communities, with some heterotrophic bacteria responding positively with increasing growth (Vázquez-Domínguez et al., 2012 ), and other smaller bacteria decreasing in size (Moran et al., 2015 ), with implications for nutrient cycling in coastal sediments (Alsterberg et al., 2011 ). Changes in pH through ocean acidification (elevated CO 2 ) also show mixed effects on benthic microbial communities, with abundance of ammonia oxidizing bacteria (AOB) and denitrifiers decreasing in Arctic sediments as a response to OA (Tait et al., 2013 ), although ammonia oxidization rates appeared unaffected (Kitidis et al., 2011 ). Anthropogenically-driven environmental changes are likely to occur simultaneously, and integration of multiple stressors into experimental designs is likely to produce differing responses to those measured for single stressor studies (Crain et al., 2008 ; Kenworthy et al., 2016 ; Pendleton et al., 2016 ). This, combined with the natural variability in many intertidal systems (such as changes in temperature, salinity, exposure) (Benedetti-Cecchi et al., 2006 ; Molinos and Donohue, 2010 ; García Molinos and Donohue, 2011 ) adds to the complexity in interpreting and understanding stressor specific responses and potential shifts in microbial community composition (Fu et al., 2007 ). The high diversity typically found within benthic microbial communities may make benthic ecosystems more resistant to environmental change (Kerfahi et al., 2014 ), ensuring the biogeochemical functions of microbial assemblages remain constant. Previous studies examining benthic microbial community composition and diversity have used a range of “fingerprinting” techniques, such as phospholipid fatty acid (PLFA) analysis to estimate biomass and identify key biomarkers (Findlay and Watling, 1998 ; Mayor et al., 2012 ; Sweetman et al., 2014 ; Main et al., 2015 ); terminal restriction fragment length polymorphism (T-RFLP) (Moss et al., 2006 ; Febria et al., 2012 ; Tait et al., 2015 ), and denaturing gradient gel electrophoresis (DGGE) (Bolhuis et al., 2013 ). To-date, few studies have examined the effects of combined environmental stressors on microbial benthic communities (Currie et al., 2017 ), and to our knowledge this is the first to integrate natural variability as an additional stressor. This study uses next generation sequencing (Ilumina MiSeq) to identify changes in microbial community composition from a manipulative mesocosm study with a focus on biodiversity driven changes in biogeochemical function. Experimental environmental change variables included ambient and elevated CO 2 ; elevated temperature; and temporal variability (diurnal temperature fluctuation) which is reflective of in situ changes in intertidal habitats. Predictions of future temperature elevation are often referred to as a mean global rise, and the diurnal variability of temperature in this experimental design represents the change in both mean temperature, but also the extremes experienced particularly in coastal and tidal ecosystems. The 16S rRNA gene was sequenced from environmental DNA extracted from the incubated intertidal cohesive sediment samples at the end of the experiment. This provides insight into microbial responses toward environmental change, and we discuss the implications on marine biogeochemical cycling. This study harnesses advanced sequencing technology to provide essential understanding of the global consequences of climate change on microbial community composition. We hypothesize that shifts in microbial community assemblages will be a response to changing environmental conditions, and this may be synergistic or additive.",
"discussion": "Discussion There is clear evidence of environmental change affecting species distributions and abundances, and this changing biodiversity has been well studied in benthic systems, through a variety of manipulative experiments and observational studies (Ieno et al., 2006 ; Prosser et al., 2007 ; Bulling et al., 2010 ; Hicks et al., 2011 ; Gilbertson et al., 2012 ; Godbold and Solan, 2013 ). However, most of these studies focus on macrofaunal diversity, although it is the microbial assemblages in these habitats that drive biogeochemical cycling (Middelburg, 2011 ; Mayor et al., 2012 ). Studies that examine shifts in microbial communities in relation to environmental changes have tended to focus on only one environmental variable, such as CO 2 gradients (Kerfahi et al., 2014 ; Tait et al., 2015 ) and the impact on relative class or order level abundance (Tait et al., 2013 , 2015 ); or targeting specific genes, for their biogeochemical properties (Kitidis et al., 2017 ). This study generated over 11 million sequences, with taxonomic identification achievable at species level (97% sequence identity). The number of OTUs found through NGS was much higher than numbers found using T-RFLP, ARISA, or DGGE (Massé et al., 2016 ), and provided greater resolution on species level changes that may be masked in studies that sequence to class/order, or only provide information on overall bacterial biomass (Mayor et al., 2013 ; Main et al., 2015 ). The comparison of different resolution (class/order analysis compared to species level analysis) showed the same trends, but a lower species resolution may not only mask species level changes, but also miss interactions between environmental variables. In future, it would be interesting to compare the results observed with direct sequencing of rRNA as it has been shown to eliminate uncertainties associated with primer matching on the rDNA and therefore producing a more robust assessment of bacterial populations (Rosselli et al., 2016 ). Benthic microbes play a vital role in sediment biogeochemistry (Bertics and Ziebis, 2009 ), and their contribution to ecosystem function is determined by community assemblage (Petchey and Gaston, 2006 ; Beveridge et al., 2010 ). This study supports previous research on coastal sediments, which has shown that Proteobacteria (alpha, beta, delta, and gamma), Bacteroidetes, and Planctomycetes dominate relative abundance (Musat et al., 2006 ; Laverock et al., 2010 ; Gobet et al., 2012 ; Tait et al., 2015 ). Overall relative abundance did not change at class or order level in response to changes in CO 2 , as seen in previous manipulative research (Tait et al., 2013 , 2015 ), although microbial community changes have been found along a natural CO 2 gradient in the Mediterranean (Kerfahi et al., 2014 ). This study found that changes in mean temperature, not CO 2 , have a significant effect on shifts in microbial community assemblage, and the contribution of certain taxa to specific ecosystem services (such as nutrient cycling) may be altered with environmental change, particularly with warming temperature (Bertics and Ziebis, 2009 ). Results indicate that this varies between orders and classes, with some remaining constant in relative abundance (e.g., Flavobacteria ), supporting previous work (Musat et al., 2006 ; Laverock et al., 2010 ; Gobet et al., 2012 ), and others such as the Proteobacteria changing in abundance with increased mean temperature. However, this study illustrates the apparent constant abundance may conceal changes in community structure at genus or species taxonomic levels as a result of the level of detail provided by next generation sequencing. Microbial communities play a vital role in benthic carbon cycling and are often the primary degraders of organic matter when it reaches the sediment surface. Bacteroidetes are important for initial organic matter degradation, often breaking down complex polymeric substances (Teeling et al., 2012 ; McKew et al., 2013 ; Taylor et al., 2013 ; Decleyre et al., 2015 ). The microphytobenthic (MPB)-rich sediment used in this study are typical of tidal mudflats, and the extracellular polymeric substances excreted by MPB provide a labile carbon source for heterotrophic microorganisms (McKew et al., 2013 ; Taylor et al., 2013 ; Decleyre et al., 2015 ; Bohorquez et al., 2017 ). Bacteroidetes are the dominant phylum here, in particular Flavobacteria (which make up 80% of the Bacteroidetes abundance), and together with Plactomycetes, they play a vital role in benthic carbon cycling as the initial organic matter degraders (McKew et al., 2013 ; Taylor et al., 2013 ; Bohorquez et al., 2017 ). Despite the changing environmental conditions, their relatively constant abundance suggests the initial degradation of carbon remains unaffected by temperature changes, perhaps unsurprising as many tidal benthic species are facultative anaerobes (McKew et al., 2013 ). Although the relative abundance of the Flavobacteria remains constant, there are changes in the community structure with increasing temperature, such as an increase in Robiginitalea as mean temperature increases (which corresponds to an increase in PO 4 ), and a corresponding decrease in Eudoraea adriatica and Lutibacter species ( L . litoralis is only found at 6°C mean temperature treatment). Species within the Robiginitalea genus are known to have a thermal preference above 10–15°C (Cho and Giovannoni, 2004 ; Manh et al., 2008 ), which may explain why they increase from 7% at 6°C mean temperature to 23% at 18°C, thus maintaining the overall constant relative abundance of the Flavobacteriaceeae family as the Lutibacter and Eudoraea species decline with rising mean temperature. This maintains the functionality of this group as primary carbon degraders, although the species within the family that carry out this process have shifted with increasing temperature, suggesting some redundancy with in the Flavobacteria . In the dominant phylum Bacteroidetes, a decrease in the Saprospiraceae family was observed with increasing mean temperature, which has implications for the carbon cycle, as they are dominant in coastal zones and play an important role in remineralisation of organic matter (Raulf et al., 2015 ). Previous studies have suggested that Saprospiraceae strains are sensitive to environmental changes, although in this study a temperature effect was demonstrated, not a shift due to elevated CO 2 (Raulf et al., 2015 ). It is also possible that these species level changes may cause a shift in the function or capability within a bacterial class or order, although the overall abundance of a class may remain constant, as found for the Flavobacteria (Table 2 ). The change in nutrient concentration for (decreasing) ammonia (NH 4 ) and (increasing) phosphate (PO 4 ) with increasing mean temperature support this concept. Here we demonstrate an increase in sulfate reducing bacteria ( Deltaproteobacteria ) as mean temperature increases, and the presence of thermophilic bacteria ( Deinococcus-Thermus ) at the highest mean temperature treatment (18°C). Sulfate reducing bacteria (SRB) are often found in cohesive sediments (Ravenschlag et al., 2000 ), such as the intertidal muddy sediment used in this study, due to the steep redox gradients determined by the permeability and oxygen penetration depth (Probandt et al., 2017 ). Sulfate reducers are associated with anoxic sediment (Orcutt et al., 2011 ), and the increase in SRB abundance with increasing temperature may also be indicative of lower oxygen concentrations with the warming regimes, driving the redox layer toward the sediment surface and promoting formation of anoxic “hotspots” within the sediment, stimulating sulfate reduction (Mahmoudi et al., 2015 ). There were clear visual differences in the highest mean temperature treatments, with the sediment profile in the mesocosms turning from an oxic brown color to black, suggesting the redox layer has shifted closer to the sediment surface, supporting sulfate reducing conditions. As strict anaerobes, Desulfobacteraceae remineralise organic matter in the absence of oxygen (Probandt et al., 2017 ), and are often found in fine impermeable sediments which promote the development of anoxic niches within the surface sediments, enhanced by the higher mean temperature in this study. A corresponding increase in the abundance of extremophilic species (Deinococcus-Thermus phylum), typically found in harsh environments such as deserts and hot springs (Albuquerque et al., 2005 ; Pikuta et al., 2007 ), was also measured in the highest mean temperature regime. This demonstrates the shifting regime in the benthic microbial community at a genus and species level, and the consequent shift from aerobic processes to favoring anaerobic processes in the sediment surface. Previous work has demonstrated that stable environmental conditions promotes constant and specific microbial communities (Bertics and Ziebis, 2009 ), but it is unclear how quickly these communities may respond to change. The interpretation of change in microbial communities is dependent on the depth of diversity measured (e.g., down to genus or species level or identifying classes or orders). However, while species turnover may be obvious when using the higher taxonomic resolution, lack of turnover does not necessarily result in static functionality. Freshwater microbial communities are often characterized by their metabolic plasticity in response to environmental change, which contributes to their functional redundancy and links their assemblage composition with ecosystem function (Comte et al., 2013 ). In the present study we demonstrate a clear response in the marine benthic microbial community to different mean temperature treatments that would have been overlooked using poorer taxonomic resolution. This changing community reflected a change in nutrient concentrations as mean temperature increased, thus suggesting there is no functional redundancy among the different species which provides resilience to environmental change (Muntadas et al., 2016 ). However, in terms of carbon cycling, there is a shift in the community assemblage within the Flavobacteria , the relative abundance remains constant, suggesting some functional redundancy with organic matter degradation. Much of the nitrogen cycle is driven by archaea (Raulf et al., 2015 ), such as ammonia-oxidizing archaea (AOA), which were not measured in this study due to the bacterial specific primers used. However, ammonia oxidizing bacteria (AOB), predominantly affiliated with Betaproteobacteria (β-AOB) (Bernhard et al., 2005 ), play a significant role in nitrogen cycling (Risgaard-Petersen et al., 2004 ), and can outnumber AOA in coastal sediments (Santoro et al., 2008 ). In this study, increasing mean temperature led to a decrease in Betaproteobacteria abundance, with no Betaproteobacteria present at the highest mean temperature. Although ammonia oxidisers were identified (both Nitrosomonas and Nitrospora ) within the Betaproteobacteria, their abundance was less than 1% across all treatments. The corresponding decrease in NH 4 concentration in the overlying water suggests there may be changes in the nitrogen cycling, possibly influenced by the absence of Betaproteobacteria, and NO x levels remain low across all treatments (Table 2 ). The phosphate increase could be linked to the corresponding decrease in abundance of Gammaproteobacteria, which are instrumental in phosphorous cycling (Sebastian and Gasol, 2013 ) and are usually limited by phosphate availability, and there is a corresponding increase in the abundance of Robiginitalea . The decrease in Gammaproteobacteria means the uptake of phosphate from the overlying water column is reduced, leading to the rising concentrations with rising temperature, directly impacting the phosphorous cycling in this benthic system. In addition, changes in the redox layer in the surface sediment will release iron-bound phosphorous under anoxic conditions (Sinkko et al., 2011 ), enhancing overall phosphorous flux from the sediment into the water. Ammonium and phosphate are typically the preferred nutrients for microbial communities, and the consistently low nitrate (and nitrite) concentrations in this study are typical of coastal oligotrophic systems (Chen et al., 2017 ). The change in NH 4 concentration may be a result of a combination of low abundance of ammonia oxidizer bacteria, reduced microphytobenthos activity or a higher rate of microbial community mineralisation with increasing mean temperature. In conclusion, changes in microbial assemblage were only found between the mean temperature treatment, and not in response to changes in diurnal temperature variability or elevated CO 2 . This supports recent research that has shown seasonal changes mask any response to elevated CO 2 within the environment (Tait et al., 2013 , 2015 ; Currie et al., 2017 ; Hicks et al., 2017b ). However, some of the changes at species level, such as increasing abundance of sulfate reducing bacteria ( Desulfobacteraceae ) and corresponding decrease of Desulfuromonadaceae with increasing mean temperature, suggest that changes to the sulfur cycle may not be noticed at the ecosystem service level, despite a change in species assemblage. Targeted future work should address how changes in some species (e.g., increase of thermophilic species in the Deinococcus-Thermus phylum) may be reflected in a broad range of biogeochemical processes, such as integrating measurements relating to sulfur, nitrogen and carbon cycles. Sediment profiles of oxygen and H 2 S would provide insight into potential shifting from oxic to anoxic (sulfate reduction) conditions, and this linked to corresponding microbial communities would provide direct biogeochemical information on coastal sediment dynamics. This study has focused on intertidal cohesive sediments, but the microbial response will vary with sediment type, driven by changes in oxygen penetration depth (Hicks et al., 2017a ). The depth of taxonomic resolution provided by NGS provides additional information at a genus or species level, allowing identification of species regime shifts that may directly impact biogeochemistry, which may be missed using a lower taxonomic resolution technique. High taxonomic resolution is useful for identifying species shifts and measuring potential functional redundancy for key biogeochemical processes, such as carbon mineralization and nutrient cycling. Since benthic systems provide a variety of ecosystem services (Duffy and Stachowicz, 2006 ; Frid and Caswell, 2016 ) which are often driven by microbial activity, these results suggest some vulnerability (nutrients), and highlights potential functional redundancy (carbon), in benthic microbial communities as a response to climate changes. Importantly, elevated CO 2 does not appear to have any effect on microbial assemblage, based on the results discussed here, although changing mean temperature (and not variability) appears to drive community assemblage change. Whilst there are limitations in realistically interpreting results from artificial mesocosm systems, and caution must be taken in interpreting responses, these experiments are valuable in providing insights on how complex ecosystems may respond to warming or elevated CO 2 (Benton et al., 2007 ; Cartaxana et al., 2015 ). This has implications for environmental change research, particularly in coastal habitats where much of the ecosystem services are generated through microbial interactions that occur in the benthos. Changes to nutrient cycling (such as the availability of nitrogen or phosphate) could promote eutrophication or decrease phytoplankton primary production (Vitousek et al., 1997 ), directly impacting food webs and indirectly affecting benthic carbon mineralization and sequestration. Integrating next generation sequencing with robust biogeochemical parameters is key in understanding the potential consequences of environmental change in coastal habitats."
} | 6,070 |
26824062 | PMC4730847 | pmc | 9,386 | {
"abstract": "A bidirectional freezing method to assemble ceramic particles into porous nacre-like scaffolds aligned over centimeter-scales.",
"conclusion": "CONCLUSIONS We have developed a bidirectional freeze-casting technique to fabricate large-scale (centimeter-scale) lamellar structures. Using HA powder, we investigated the mechanism of gradient nucleation and propagation of ice crystals under dual temperature gradients. This was achieved by placing a PDMS wedge with different slopes in between the cold finger and the slurry. In contrast to conventional unidirectional freeze casting, the bidirectional freeze-casting technique developed in this study can manipulate the alignment of ice crystals in two directions and can be applied to generate larger porous scaffolds with highly controlled, ordered structures. Although HA particles were used in this paper as a proof of concept, the technique can also be applied to multiple materials systems, for example, ceramic particles or platelets. Apart from the slope angle and the cooling rate, many other parameters, such as solid loading or various additive compounds, are currently being investigated in our laboratory for their effect on the bidirectional freezing technique. Our newly developed technique could thus provide an effective way of designing and manufacturing larger-scale novel, bioinspired, structural materials, especially advanced materials such as composites, where a higher level of control over the structure is required.",
"introduction": "INTRODUCTION Natural materials, such as bone, teeth, shells, and wood, show outstanding properties despite being porous and made of weak constituents ( 1 ). The secret usually lies in their sophisticated hierarchical architecture ranging from nano/microscopic to macroscopic levels ( 2 – 5 ). Such architectures have been perfected over the past billions of years, resulting in wonderful materials that are very often strong, tough, and lightweight and serve as a source of inspiration for every materials designer. Porous ceramic structures, in particular, are desirable for a wide range of applications in areas such as supported catalysis ( 6 ), scaffolds for tissue engineering ( 7 , 8 ), foams ( 9 ), fuel cell electrodes ( 10 ), filters for water purification ( 6 ), and many others ( 11 ). Multiple techniques ( 12 , 13 ), such as replica, direct foaming, or sacrificial templating, have been developed to manufacture such scaffolds. Most recently, three-dimensional (3D) printing ( 14 ) has also been used as an alternative technique. However, these techniques have several limitations because they are often time-consuming or size-limiting processes, not environmentally friendly, too costly, or just do not allow precise control over the final structure. An ideal strategy for engineering pores into materials in a more controllable way and at a larger scale has yet to be developed. Freeze casting can overcome many of these previous limitations ( 15 ). This very promising technique enables assembly of ceramic particles into scaffolds that have a highly aligned 3D porous network. The technique uses lamellar ice crystals as a template to assemble building blocks for making biomimetic scaffolds or composites ( 16 – 18 ). It offers the advantage of being applicable to a wide spectrum of materials [such as ceramics ( 16 – 18 ) and/or polymers ( 19 )], having various shapes [that is, particles ( 16 , 17 , 19 ), nanowires ( 20 ), and ceramic platelets ( 18 , 21 ) or graphene sheets ( 22 )]. In addition, the technique is environmentally friendly, with water usually being used as the solvent. Finally, easy control of the structural features at multiple length scales is achievable by modifying ice crystal morphology with additives and/or the cooling rate ( 23 – 26 ). Nevertheless, in the case of conventional freeze casting (also referred to as “ice templating” or “unidirectional freezing”), the slurry starts freezing under a single temperature gradient, causing the nucleation of ice to occur randomly on the cold finger surface. As a result, multiple small-size (submillimeter scale) domains, that is, various ice crystal orientations in the plane perpendicular to the freezing direction, are observed ( 27 ). Despite a pressing demand for the development of new processing techniques that can build large-scale porous aligned lamellar structures, this limitation severely hinders the scale-up fabrication of layered structures aimed for larger applications. Here, we report on a new bidirectional freezing technique which can assemble small building blocks (ceramic particles, platelets, and/or polymer) into a large-size single-domain (centimeter-scale) porous lamellar structure comparable to natural nacre, albeit without the “mortar.” This was achieved through a proper control of nucleation of ice crystals and growth under dual temperature gradients generated by covering the cold finger with a polydimethylsiloxane (PDMS) wedge having different slopes. (To illustrate the breakup of the gradient into vertical and horizontal components during freezing, we use in this paper the expression “dual temperature gradients,” although the resulting combined temperature gradient is actually singular. This allows the ice to grow both vertically and horizontally, which is critical for obtaining large-scale aligned lamellar structures.) Although hydroxyapatite (HA) was used as a proof of concept to study the bidirectional freezing mechanism in detail, this technique can also be applied to any other ceramic or polymeric materials of any shape. Our approach could provide an effective way of designing and manufacturing novel, bioinspired, structural materials, in particular advanced materials, such as composites, where a higher level of control over the structure is required."
} | 1,444 |
25706146 | PMC4338226 | pmc | 9,390 | {
"abstract": "Methanosarcina acetivorans , considered a strict anaerobic archaeon, was cultured in the presence of 0.4–1% O 2 (atmospheric) for at least 6 months to generate air-adapted cells; further, the biochemical mechanisms developed to deal with O 2 were characterized. Methane production and protein content, as indicators of cell growth, did not change in air-adapted cells respect to cells cultured under anoxia (control cells). In contrast, growth and methane production significantly decreased in control cells exposed for the first time to O 2 . Production of reactive oxygen species was 50 times lower in air-adapted cells versus control cells, suggesting enhanced anti-oxidant mechanisms that attenuated the O 2 toxicity. In this regard, (i) the transcripts and activities of superoxide dismutase, catalase and peroxidase significantly increased; and (ii) the thiol-molecules (cysteine + coenzyme M-SH + sulfide) and polyphosphate contents were respectively 2 and 5 times higher in air-adapted cells versus anaerobic-control cells. Long-term cultures (18 days) of air-adapted cells exposed to 2% O 2 exhibited the ability to form biofilms. These data indicate that M. acetivorans develops multiple mechanisms to contend with O 2 and the associated oxidative stress, as also suggested by genome analyses for some methanogens.",
"introduction": "Introduction The reactive oxygen species (ROS) are toxic for most cells because they induce (i) oxidation of polysaccharides and polyunsaturated fatty acids, as well as amino acid residues, particularly of sulfhydryl groups in proteins; (ii) loss of metals in metalloproteins; and (iii) DNA mutations, among many others [ 1 ]. Aerobic microorganisms have developed multiple strategies to handle ROS stress including: (i) enzymes that scavenge ROS such as superoxide dismutase (SOD), catalase (CAT) and peroxidases (PXs); (ii) protein repair mechanisms such as the thioredoxin system; (iii) DNA damage repair enzymes such as RecA; and (iv) anti-oxidant metabolites such as glutathione, α-tocopherol, carotenes, ascorbate, and trypanothione, which are able to directly inactive ROS [ 1 – 4 ]. The organisms belonging to the Archaea domain generally live under extreme conditions [ 5 ]. Indeed, many live under complete anaerobic conditions; therefore, it has been frequently assumed that most anaerobic archaea do not interact with O 2 and therefore they lack mechanisms able to cope with oxidative stress. Methanogens, the main Archaea group, grow in anoxic environments such as the rumen, sewage digesters, landfills, freshwater sediments of lakes and rivers, rice paddies, hydrothermal vents and coastal marine sediments [ 6 ]. Therefore, most of the methanogens are cultivated in the presence of high Na 2 S (1–3 mM) to yield an anoxic and reducing medium (-300 mV). Biochemical and genetic (genome and transcriptome) analyses have suggested that methanogens have the ability to develop mechanisms to cope with oxidative stress [ 7 ]. Methanogens such as Methanosarcina spp and Methanocella spp have been isolated from soil crusts of arid regions where aerobic conditions are predominant [ 8 ]. In these places, methane production by these methanogens is detected, but methanogenic rates are much lower when O 2 is present. Increased transcription of the peroxide-detoxifying kat gene (catalase) was found in these methanogens, but the enzyme activity was not determined [ 9 ]. \n Methanobrevibacter arboriphilicus SA, Methanobacterium fomicicum and Methanosarcina mazei TMA isolated from paddy soils are able to deal with periods of aeration and water stress for up to 30 days [ 10 ]. Analyses of the genomes of these methanogens show the presence of genes encoding antioxidant enzymes, which may be the main reason of the different abilities to resist aerobic conditions, rather than differences in the habitats that may act as shelters for methanogens during the long-term stress period. In Methanobrevibacter cuticularis and Methanobrevibacter curvatus isolated from microaerofilic regions of the hindgut of termites, CAT and SOD activities are detected [ 11 ]; however, these organisms immediately cease growth and methane production when the cultures are initiated in the presence of 0.16–1.6% O 2 in the head space [ 12 ]. In Methanosarcina barkeri , pulses of H 2 O 2 , but not of O 2 , induce the activity of both CAT and Fe-dependent SOD [ 13 – 15 ]. Methanosarcina mazei contains a methanoferrodoxin with superoxide reductase activity which contributes to the protection of cells from ROS formed by flavoproteins during periodic exposure to oxygen in natural environments [ 16 ]. The marine archaeon Methanosarcina acetivorans WWM73 strain can tolerate high H 2 O 2 concentrations without a complete loss of viability [ 17 ]. Also, a functional thioredoxin reductase system has been reported for this methanogen [ 18 ]. \n Methanosarcina spp and Methanosaeta spp are the only methanogens able to consume acetate for methane production [ 19 ], which may account for 75% of the biological methane on earth. Despite this crucial role in the carbon cycle, knowledge regarding the mechanisms present in Methanosarcina spp to contend against oxidative stress is still incomplete. To assess the mechanisms of resistance against oxidative stress in methanogens, M . acetivorans was adapted to grow in the presence of permanent low O 2 (0.4–1% O 2 atmospheric ). These air adapted cells showed increased transcripts of sod , kat and NADH-dependent peroxidase genes and activities of SOD, CAT and NAD(P)H-, cytochrome c - and CoM-SH-dependent peroxidases (PXs). An increase in the contents of thiol molecules and polyP was also observed. Moreover, long exposures (up to 18 days) to higher O 2 concentrations led to formation of biofilms constituted by DNA, CHOs and proteins. The physiological relevance of these mechanisms in methanogens to cope with O 2 in a marine environment is discussed.",
"discussion": "Discussion 3.1 Analysis of genes related to protection against oxidative stress in methanogenic genomes The genomes analyzed here, indicated that rubrerythrin (a non-haem iron protein) is wide-spread among methanogens, and together with rubredoxin and the SOR activity, is necessary for a complete ROS detoxification system [ 16 , 25 ] ( Table 1 ). F 420 H 2 oxidase catalyzes the reduction of O 2 to water and may play an important role against oxidative stress in methanogens [ 26 ]. Genes encoding thioredoxins (Trx), well known proteins involved in oxidative stress handling, were present in all 27 genera analyzed. Methanosarcina spp possesses up to 8 different genes encoding Trx suggesting multiple metabolic roles for this protein [ 18 ]. On the other hand, M . acetivorans and M . barkeri are the only methanogens with genes annotated for quinol: cyt bd oxidase, which suggests that this enzyme may not have an important role against oxidative stress ( Table 1 ). Except for the NADH-peroxidase gene (which was only found in M . acetivorans among methanogens), genes for oxidative stress management found in bacteria ( E . coli ) showed high identity with those identified in M . acetivorans : cyt c peroxidase (30%), Cu-Zn SOD (34%), Fe-Mn SOD (40%) and catalase/peroxidase (60%). Archaea such as Archaeoglobus sp and Methanobacterium sp showed 61 and 80% identity (respect to M . acetivorans gene) for catalase/peroxidase, respectively. Fe-Mn SOD showed high identity (57–82%) among archaeal genomes. In contrast, Pyrococcus spp does not contain any of these enzymes coded in their genome. 3.2 Role of thiol-molecules and polyP as anti-oxidant metabolites In air-adapted cells, the content of thiol-molecules increased 2 times versus anaerobic control cells ( Table 2 ). If an intracellular volume of 0.7 μL (mg protein) -1 is assumed for M . acetivorans [ 27 , 28 ], high concentrations of 70–114 mM and 7–14 mM may be reach up for Cys and CoM-SH, respectively. Hence, it is possible that this high Cys level directly reacts with O 2 and ROS, and may induce the expression of antioxidant genes in response to oxidative stress such as Trx [ 29 ] or Rbr, Rbx, Prx and glutaredoxin-like proteins ( Table 1 ). In turn, CoM-SH was an electron donor for PX activity ( Table 3 ). We previously reported that the contents of Cys, CoM-SH and sulfide also increase (>7 times) in M . acetivorans exposed to Cd 2+ respect to control cells without Cd 2+ [ 22 ]. Therefore, an essential role for these two metabolites (Cys and CoM-SH) in the anti-oxidant machinery of this archaeon is proposed. Increased synthesis of polyP is another mechanism in air-tolerant organisms involved in coping with different types of stresses such as heavy metals [ 30 ] and oxidative stress [ 31 ]. In the present work, it was described that air exposure triggered increased synthesis of polyP ( Table 2 ; S1 Fig. ), although acetate-grown cells accumulated more polyP than methanol-grown cells. Clearly, more work is needed to determine whether (i) there is an O 2 threshold that triggers the polyP synthesis in M . acetivorans and whether (ii) Pi and polyP directly may react with ROS. In bacteria, involvement of polyP in the resistance to oxidative stress has been shown. PolyP is essential for biofilm development, quorum sensing and virulence in bacteria [ 32 , 33 ]. It has been suggested that polyP or polyP kinase regulate the transcription of genes involved in the stress oxidative response such as CAT and SOD in E . coli [ 34 , 35 ]. Since M . acetivorans is a marine methanogen where acetate and phosphorus are present at low levels, it might be an evolutionary advantage for its survival to possess highly efficient mechanisms for uptake and storage of Pi. 3.3 ROS production and the effect of O 2 on antioxidant enzyme transcripts and activities \n M . acetivorans was able to consume O 2 , being higher in methanol than in acetate grown cells ( Fig. 1 ), because higher O 2 levels were used in methanol grown cells. Enzyme activities directly involved in O 2 consumption were not determined in the present study, but M . acetivorans contains F 420 H 2 oxidase and several ferredoxins, flavodoxins and iron-sulfur proteins which may react with oxygen [ 36 ]. ROS production was significantly higher in methanol ( versus acetate) grown cells and anaerobic control cells ( versus air-adapted cells; S2 Fig. ) after the short-term (2 h) exposure to 2% O 2 . Lower SOD and CAT activities may be the reason for the higher ROS levels in these cells. Basal activities of SOD and CAT as well as increased activities induced by oxidant stressors have been also found in Methanosarcina barkeri [ 13 ] and other methanogens [ 12 , 37 ]. M . acetivorans has been identified as an archaeon with resistance to O 2 [ 38 ] and hydrogen peroxide in short-term exposures [ 17 ], whereas other methanogens are extremely sensitive to O 2 [ 10 ]. The absence of genes encoding antioxidant enzymes in the latter group may be the reason for their extreme sensitivity. Air-adapted cells showed significantly increased CAT and SOD activities. In gel enzyme activities indicated that Zn 2+ and Fe 2+ increase SOD activity, as reported for the M . arboriphilus enzyme [ 37 ]. There are genes annotated for Zn/Cu–dependent SOD (MA2422) and Fe/Mn-dependent SOD (MA1574) in the genome of M . acetivorans . On the other hand, the increase in the SOD transcript induced by O 2 in air-adapted cells (2.4 times) correlated with the increased SOD activity (3.8 times), which may be further stimulated by heavy metal divalent cations. On the other hand, there are 4 annotated genes encoding for PXs with putative specific electron donors: catalase/peroxidase (MA0972), chloride peroxidase (MA0993), NADH peroxidase (MA1426) and cytochrome c peroxidase (MA2908). Hence, PX activity determined here with ascorbate, a non-physiological electron donor in M . acetivorans , may have underestimated total PX activity. Hence, other electron donors were examined. Because chloride PX is involved in the detoxification of polychlorinated biphenyls pollutants rather than in oxidative stress [ 39 ], this activity was not determined. Instead, CoM-SH was tested as PX substrate because of its high physiological levels and to be a potential electron donor like glutathione. PX activities were found for NADH, NADPH, cyt c and CoMSH. Transcription of the NADH Px gene increased 5 times in air-adapted cells grown on acetate. However, all PX activities were high in both control and air adapted cells, suggesting that these enzyme activities may be constitutive and hence required for protecting the cell against basal levels of oxidative stress. CAT and SOD were in turn over-expressed in air adapted cells, indicating that these enzymes are involved in contending against acute oxidative stress generated by external stressors such as O 2 , as proposed by Pedone et al [ 20 ]. The presence of these antioxidant enzymes in M . acetivorans suggests that episodes of oxidative stress in the marine environment in which this archaeon grows may be recurrent. Indeed, changes in the O 2 concentration occur during disturbances of the deep sea by earthquakes and other meteorological events [ 40 ]. 3.4 Biofilm formation induced by O 2 stress Composition analysis of the cell agglomerates, in control and air adapted cultures grown on high or low salt, demonstrated that in all conditions non-soluble carbohydrates were present, with 10-fold higher levels in the low salt cultures, in which secretion of methanocondrioitin and S-layer is involved [ 24 ]. However, air adapted cells exhibited higher content of non-soluble CHOs and extracellular DNA, an essential component of biofilms ( Fig. 5D ). Our results showed that long exposure of M . acetivorans cultures to O 2 (at least 18 days) led to the formation of biofilms, apparently as a strategy to gain resistance against higher O 2 concentrations. The significant increase in the content of non-soluble carbohydrates in air adapted cells was, however, only 50% higher than that of control cells. By comparison, M . acetivorans cells exposed to 1.4 mM CdCl 2 exhibit 800% increased content of non-soluble carbohydrates respect to control cells [ 22 ], suggesting that in M . acetivorans O 2 is a weak biofilm-inducer. In this regard, it has been documented that cells within biofilms show increased tolerance to stressful environmental conditions. For instance, the biofilm made by the archaeon A . fulgidus in metal-depleted medium is induced by non-physiological drastic changes of pH and temperature, high concentrations of metals or by addition of xenobiotics or O 2 . Essential metals sequestered within the biofilm stimulate the growth, suggesting that cells may produce biofilm as a mechanism for concentrating cells and attaching to surfaces, as a protective barrier and as a nutrient reservoir [ 41 ]. Due to the fact that similar biofilms are formed by other archaea, biofilm formation might be a common stress response mechanism within the Archaea domain [ 23 ]. Viable methanogens have been detected in dry, aerobic environments such as dry reservoir sediment, dry rice paddies and aerobic desert soils, suggesting that methanogens have mechanisms for long-term survival under various environmental stresses [ 42 – 44 ]. In Methanosarcina barkeri , desiccation and the synthesis of extracellular polysaccharide are indeed survival mechanisms against oxygen, probable because minimize oxygen diffusion into the cell [ 45 ]. Then, it is clear that to elucidate (i) the specific mechanisms that archaeal cells have developed to cope with O 2 ; and (ii) the specific interactions between biofilm and cells, further studies are required. In conclusion, the generation of stable cultures of air-adapted cells of M . acetivorans allowed to clearly determining (i) variation in the expression and activity of the anti-oxidant enzymes SOD, CAT and PXs; (ii) changes in thiol-molecule and polyP contents; and (iii) the development of biofilm. These cellular mechanisms are required to maintain the cell viability, which might become molecular targets for enhancing biogas production under oxidative stress conditions."
} | 4,075 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.