Content uploaded by Francesco Ricci
Author content
All content in this area was uploaded by Francesco Ricci on Jan 05, 2025
Content may be subject to copyright.
Chemosynthesis enhances carbon 1
fixation in an active microbialite 2
ecosystem 3
4
Francesco Ricci1,2,3, Pok Man Leung1,2,*, Tess Hutchinson1,2, Thanh Nguyen-Dinh1,2, 5 Ashleigh v.S. Hood3, Vinícius W. Salazar4, Vera Eate5, Wei Wen Wong5, Perran L.M. 6 Cook5, Chris Greening1,2 & Harry McClelland3,6,* 7 8 1. Department of Microbiology, Biomedicine Discovery Institute, Monash 9 University, Melbourne, VIC, Australia 10 2. Securing Antarctica’s Environmental Future, Monash University, Melbourne, 11 VIC, Australia 12 3. School of Geography, Earth and Atmospheric Sciences, University of 13 Melbourne, Parkville, VIC, Australia 14 4. Melbourne Bioinformatics, University of Melbourne, Parkville, VIC, Australia 15 5. Water Studies Centre, School of Chemistry, Monash University, Clayton, VIC, 16 Australia 17 6. Department of Structural and Molecular Biology, University College London, 18 London, WC1E6BT, UK 19 20 *Corresponding authors: francesco.ricci@monash.edu, bob.leung@monash.edu, 21 and h.mcclelland@ucl.ac.uk 22 23 24
Abstract 25
Microbialites—carbonate structures formed under the influence of microbial action—26 are the earliest macroscopic evidence of life. For three billion years, the microbial 27 mat communities responsible for these structures fundamentally shaped Earth’s 28 biogeochemical cycles. In photosynthetic microbial communities, light energy 29 ultimately drives primary production and the ensuing cascade of daisy-chained 30 metabolisms. However, reduced compounds such as atmospheric trace gases and 31 those released as metabolic byproducts in deeper, anaerobic regions of the mat, 32 could also fuel chemosynthetic processes. Here, we investigated the intricate 33 metabolic synergies that sustain microbialite community nutrient webs. We 34 recovered 331 genomes spanning 40 bacterial and archaeal phyla, revealing a 35 staggering diversity fuelled by the biogeochemistry of these ecosystems. While 36 phototrophy is an important metabolism encoded by 17% of the genomes, over half 37
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
encode enzymes to harness energy from reduced compounds and 12% co-encode 38 carbon fixation pathways, using sulfide and hydrogen as major electron donors. 39 Consistent with these genomic predictions, we experimentally demonstrated that 40 microbialite communities oxidise ferrous iron, ammonia, sulfide and gas substrates 41 aerobically and anaerobically. Furthermore, 14C-fixation assays revealed that 42 chemosynthesis contributes significantly to carbon fixation alongside photosynthesis. 43 Chemosynthesis in microbialite communities represents a complex interplay of 44 metabolic synergies and continuous nutrient cycling, which decouples community 45 carbon fixation from the diurnal cycle. As a result, this process mitigates the loss of 46 organic carbon from respiration, thus enhancing the net productivity of these highly 47 efficient ecosystems. 48 49 50 51 52
Introduction 53
Complex microbial communities in aquatic environments can form carbonate 54 structures named microbialites. With the earliest fossils dating to around three and a 55 half billion years1,2, these structures are representative of the most ancient and 56 persistent microbial ecosystems in Earth’s history. Over Earth’s long history, dense, 57 benthic microbial communities played a pivotal role in shaping the composition of 58 Earth's atmosphere3,4. Notably, they contributed to the rise in atmospheric oxygen 59 following the evolution of oxygenic photosynthesis3,5, which amplified global 60 biological productivity by 100-1000 fold6. Much of this new productivity first occured 61 in microbial mats that provided favourable conditions for photosynthetic communities 62 to thrive7–9. Additionally, mat communities mediated large fluxes of reduced gases 63 such as carbon monoxide (CO), hydrogen (H2), and methane (CH4) into the 64 atmosphere6. Today, living microbialites are generally confined to extreme 65 environments such as hypersaline lakes10–14, where grazing metazoans and plants 66 are typically absent3. However, modern stratified microbial communities have a 67 structure resembling those from the Cambrian15 and Archaean5, positioning these 68 communities as natural laboratories for unravelling the functional processes 69 underlying these ancient systems. 70 71 Microbialites are hotspots of microbial diversity16–18, yet the key processes that have 72 sustained the ecosystems associated with these structures remain enigmatic. Light 73 is thought to have had a central role throughout their history, initially sustaining 74 anoxygenic phototrophs and later cyanobacteria and algae, driving organic carbon 75 production, carbonate precipitation, and elemental cycling5,6,19,20. The resultant 76 photosynthetic end-products shape aerobic niches and fuel complex microbial 77 networks3, including by supporting the activity of aerobic heterotrophs21–23. 78 Anaerobes dominate beneath the surface of microbial mats, including fermenters, 79
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
sulfate-reducing bacteria (SRB), and methanogens that catabolise organic matter 80 and gases6,24,25. SRB use organic (e.g. acetate, sugars) and inorganic electron 81 donors (e.g. H2) while releasing H2S. Chemolithoautotrophs can exploit inorganic 82 compounds, encoding genes mediating the oxidation of H2, CO, and various sulfur 83 compounds26–28, though their ecological and biogeochemical roles within these 84 ecosystems remain poorly understood. While photosynthetic carbon fixation is 85 central to biomass production in these ecosystems3,29, this is not the sole process 86 contributing to their primary production. Most notably, the abundant anoxic pockets 87 within microbialites foster ecological niches well-suited for anaerobic autotrophs 88 utilising the Wood–Ljungdahl pathway (WLP)30–32. Although gene- and genome-89 resolved studies have provided insights into the functional potential of microbialite 90 communities, direct activity-based evidence of the complex metabolic interactions 91 that sustain these ecosystems remains limited. 92 93 Here we aim to shed light on the primary production and elementary cycling 94 processes sustaining the biodiversity of calcyfing microbial mats using living 95 microbialites from West Basin Lake, Victoria, Australia. While it is recognised that 96 microbialite communities have the potential for several photosynthetic and 97 chemosynthetic carbon fixation pathways, we still lack a comprehensive 98 understanding of their mediators, the role of various electron donors in driving these 99 processes, and their relative contribution to primary production. The classic view 100 positions photosynthesis as a leading process driving biological productivity3,6,33,34. 101 However, microbialite communities produce substantial amounts of reduced 102 compounds such as H2, H2S, CO and CH46,25, which could fuel diverse 103 chemosynthetic processes. Using living microbialite ecosystems, we integrate 104 genome-resolved metagenomics with high-resolution metabarcoding, 105 biogeochemistry, modelling and phylogenetic analysis to quantify the key 106 metabolisms and their contributions in supporting the nutrient web in microbialite 107 communities. 108 109 110
Results & Discussion 111
112
Taxonomically and metabolically diverse microbes control a 113
complex nutrient web in living microbialites 114
115 We deeply sequenced nine metagenomes across three microbialite communities 116 from West Basin Lake in Victoria, Australia. This effort resulted in the recovery of 117 331 medium- to high-quality metagenome-assembled genomes (MAGs) dereplicated 118
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
at the species level using a 95% ANI threshold. These MAGs depict a remarkable 119 diversity spanning 40 bacterial and archaeal phyla, including the first genomes from 120 a microbialite community for the phyla JAHJDO01 and UBP6 (Fig. 1). Many of these 121 microbes represent elusive and rare lineages, such as Cloacimonadota, 122 Krumholzibacteriota, Hydrogenedentota, Omnitrophota, and Sumerlaeota. 123 Representatives of these phyla have been previously found in extreme environments 124 including an Antarctic lake35, deep-sea trenches36 and geothermal springs37. The 125 most abundant phyla were Proteobacteria and Bacteroidota, represented by 81 and 126 56 MAGs respectively (Fig. 1). Cyanobacteria are considered key members of these 127 ecosystems, though we only retrieved one medium-quality MAG of Halothece (total 128 rel. abundance 3.11%) and 18 MAGs capable of anoxygenic chlorophototrophy (Fig. 129 1). Millimeter-scale 16S rRNA gene community profiling detected 30 cyanobacterial 130 sequences, which were present at considerably different relative abundance 131 between the analysed samples (sample C: 11.2% ± 7.6; sample D: 0.27% ± 0.35) 132 (Supp. Data 1). Archaea are also significant members, represented by 18 MAGs 133 spanning Asgardarchaeota, Halobacteriota, Iainarchaeota, Nanoarchaeota, 134 Thermoplasmatota and Thermoproteota (Fig. 1). Eukaryotes have been found in 135 most modern microbial communities38–40 and accordingly we identified several major 136 members of diatoms and red and green algae (Supp. Data 1). 137 138 Screening of key metabolic pathways in the MAGs revealed extensive metabolic 139 diversity within West Basin Lake microbialite communities (Fig. 1), potentially 140 underpinning these diverse microbial assemblages. Many microbes appeared to be 141 facultative anaerobes that can input electrons from organic carbon to aerobic and 142 anaerobic respiratory chains with 63% of genomes encoding at least one terminal 143 oxidase (Supp. Data 2). The widespread capacity to oxidise various inorganic 144 compounds (53.2% of the genomes), including hydrogen, sulfide, and carbon 145 monoxide, for supplemental energy likely allows more efficient use of organic carbon 146 for anabolism than catabolism (Supp. Data 2). Furthermore, our data suggest that at 147 West Basin Lake, most organic carbon is produced in situ. We identified multiple 148 lines of evidence supporting this hypothesis. First, the lake is not connected to any 149 inlet and lacks aquatic vegetation. Second, we measured moderate concentrations 150 of organic carbon in the water, averaging 4.83 mg/L. Notably, organic carbon 151 abundance increases on average from 3.09 mg/L in shallow to 5.60 mg/L in deeper 152 layers within the microbialites (Supp. Data 3). Third, 13.6% of the generated MAGs 153 encode genes for carbon fixation (Fig. 1). Specifically, 19 MAGs encoded the WLP, 154 whereas 15 MAGs co-encoded the CBB cycle with several other energy harvesting 155 genes (Fig. 1; Supp. Data 2). Interestingly, some MAGs possess the genetic 156 machinery to utilise multiple energy sources. For instance, a Thiohalophilus MAG 157 encodes the CBB cycle with uptake [NiFe]-hydrogenases, sulfide quinone 158 oxidoreductase, thiosulfohydrolase, reverse dissimilatory sulfate reductase and iron 159 oxidising cytochrome (Fig. 1). We also found six MAGs encoding the reductive 160 tricarboxylic acid cycle (rTCA), three of which belonged to the enigmatic group 3 161 [NiFe]-hydrogenase-encoding Thermoplasmatota class E2 (Fig. 1). Lastly, 162
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
phototrophy based carbon-fixation via the CBB cycle was mediated by members of 163 the Cyanobacteria, Proteobacteria and Gemmatimonadota (Supp. Data 2) as well as 164 microalgae in the phyla Cercozoa, Chlorophyta, Gyrista and Rhodophyta (Supp. 165 Data 1). 166
To gain a comprehensive, contextual understanding of the metabolic traits 167 supporting energy conservation and primary production in microbialite communities, 168 we conducted a comparative gene-centric analysis of West Basin Lake 169 metagenomes alongside previously sequenced datasets, including thrombolite, 170 oncolite, and stromatolite communities from five global sites (Fig. 1). The capacity for 171 carbon fixation was consistently high across microbialites worldwide. In West Basin 172 Lake, the CBB (rbcL av. 19.6%) and the WLP (acsB av. 9.92%) were predominant. 173 In the global datasets, the CBB was also the major pathway (av. 32.5 ± 17.0%), 174 while the WLP and 3-hydroxypropionate cycle showed occasional dominance in 175 specific samples (Fig. 1). The potential to utilise inorganic substrates exhibited 176 significant parallels across sites. Sulfide and H2 appear to be dominant energy 177 sources, reflected by the widespread oxidation capacity (sqr: West Basin Lake av. 178 60.9%, global 40.1 ± 20.9%; Group 1-3 NiFe hydrogenase: West Basin Lake av. 179 24.4%, global 22.2 ± 16.0%). The capacity for CO and arsenite oxidation–ancient 180 metabolic traits–was also substantial across all microbial communities (cooS & coxL: 181 West Basin Lake av. 17.7%, global 11.2 ± 11.7%; aro: West Basin Lake av. 11.5%, 182 global 10.1 ± 7.3%). Conversely, photosynthesis genes showed some variation. 183 West Basin Lake had moderate capacity for chlorophylls and bacteriochlorophylls 184 mediated phototrophy (psaA, psbA: av. 21.9%) and microbial rhodopsins mediated 185 phototrophy (rho: av. 30.5%), whereas the global datasets exhibited broader ranges 186 (psaA, psbA: 12.6–141.3%; rho: 13.1–113.5%). This variability most likely 187 underscores the influence of local environmental conditions on energy acquisition 188 strategies. Overall, our findings demonstrate that the genetic potential for energy 189 conservation and carbon fixation in West Basin Lake microbialites is broadly 190 representative of other microbial communities forming microbialites that have been 191 studied globally. The high metabolic flexibility observed in these ecosystems enables 192 them to act as efficient engines of biological productivity, supporting diverse and 193 dynamic microbial communities. 194
195 196
Tight microbial interactions support intricate biogeochemical 197
cycles 198
199
Consistent with previous work13,18,28,41,42, our findings reveal that microbialite 200 communities have remarkable metabolic diversity, and we sought to resolve the 201
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
complex molecular interactions underpinning such diversity. Millimeter-scale 202 community analysis revealed high microbial richness throughout the entire 203 microbialite structure (Supp. Fig. 1). While some taxa such as fermenters in the 204 order Clostridiales were confined to deeper anoxic niches (grey layer - Supp. Data 205 1), many other microbes were distributed across the whole microbialite structure 206 (Supp. Data 1), most notably Cyanobacteriales and the sulfur-oxidiser 207 Campylobacterales (Supp. Data 1). Similarly, members of the obligately anaerobic 208 order Spirochaetales were ubiquitous (Supp. Data 1). These data suggest that the 209 spatiotemporal physicochemical shifts such as the expansion of oxic pockets during 210 the day foster tight metabolic interaction that allow diverse functional guilds to 211 coexist and meet their physiological needs throughout the microbialite structure. 212
We developed a model to elucidate the array of molecular exchanges occurring 213 within microbialite communities (Fig. 2). Molecular hydrogen (H2) emerges as a 214 central molecule. In these communities, H2 is primarily produced through 215 fermentation of photosynthetically- and chemosynthetically-derived organic carbon 216 via diverse group 3 [NiFe]-hydrogenases and [FeFe]-hydrogenases, encoded by 90 217 and 55 MAGs respectively (Fig. 2). The genetic potential for other fermentation 218 pathways in the community is high, with many MAGs encoding marker genes 219 associated with acetate (acdA, ack, pta), formate (fdhA, fdoG, pflD), and lactate (ldh; 220 Supp. Data 5) fermentation. Diazotrophic Cyanobacteria and members of six other 221 phyla (16 MAGs) may also contribute to H2 release as an obligate by-product of the 222 nitrogenase reaction. Upon diffusion into aerobic niches, H2 is readily utilised by the 223 abundant community of gas oxidisers spanning 25 phyla (106 MAGs; Fig. 2), which 224 may utilise the electrons derived through this process for carbon fixation (Supp. Data 225 2). Anaerobic processes likely recycle much of the H2, including through coupling to 226 denitrification (109 MAGs) and dissimilatory nitrate reduction to ammonium (55 227 MAGs; Fig. 2). Hydrogenotrophic acetogens were found across each microbialite 228 sample, albeit at low relative abundance (av. 1.2% of mapped reads). H2 also fuels 229 methanogenesis and sulfate reduction to varying degrees, leading to the production 230 of methane and sulfide (Fig. 2). Aerobic and anaerobic CO oxidation are important 231 processes within microbialite communities (34 MAGs; Fig. 2), capable of supporting 232 energy conservation and carbon fixation in eight phyla (Supp. Data 2). A total of 78 233 MAGs, representing chemolithotrophic and photolithotrophic microorganisms across 234 nine phyla, possess the potential to oxidize sulfide under both aerobic and anaerobic 235 conditions (Fig. 2; Supp. Data 2). This capability likely provides a substantial 236 ecological advantage, facilitating their persistence in environments characterised by 237 dynamic redox fluctuations and intense resource competition. 238
239 240
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
Microbialite communities mediate broad aerobic and anaerobic 241
elemental cycling 242
243 To validate our biogeochemical predictions, we performed ex situ microcosm 244 experiments. The potential for oxygenic photosynthesis was assessed using 245 chemical imaging that simultaneously visualises and quantifies O2 concentrations in 246 natural samples43,44. Algal and cyanobacterial populations, abundant in the 247 microbialite communities, exhibited high activity, producing substantial O2 when 248 exposed to light intensities of ~450
μ
mol m
2 s
1, mimicking their natural habitat 249 (Fig. 3; Supp. Fig. 2). The well-illuminated microbialite surfaces demonstrated the 250 highest gross O2 production, though this varied markedly both within individual 251 samples (e.g., Sample 2 ranged from 29.31 to 182.26
μ
mol O2 L
1 min
1) and 252 among samples (Fig. 3; Supp. Fig. 2). Upon transition to darkness, O2 was rapidly 253 depleted, primarily through aerobic respiration and diffusion into the overlying water 254 column (Fig. 3). The rapid O2 consumption likely contributes to niches for the 255 widespread facultative and obligate anaerobes at the millimetre-scale (Supp. Fig. 2). 256 The O2 generated via oxygenic photosynthesis also serves as an electron acceptor 257 for several metabolic pathways, including organotrophy, hydrogenotrophy, 258 carboxydotrophy, and nitrification. Correspondingly, all samples demonstrated 259 oxidation of H2, CO, and NH4
, albeit at varying rates (Fig. 3). 260
We observed consumption of H2, CO, Fe2
, and S2
in nearly all anaerobic 261 microcosms (Fig. 3), confirming the high metabolic flexibility of the anaerobic 262 communities. This metabolic versatility aligns with genomic predictions made by 263 previous studies in the literature13,26,41 and reflects the capacity of microbialite 264 communities to exploit alternative electron donors to sustain their nutrient web 265 (Supp. Fig. 1). Such flexibility enables them to optimise energy generation across the 266 dynamic physicochemical gradients of the microbialite environment. Consistent with 267 these findings, physicochemical analyses revealed the presence of diverse electron 268 acceptors in the microbialite samples. Specifically, SO42- was present in high 269 concentrations (av. 67.85 mg/L), whereas NO3- at moderate concentrations (av. 2 270 mg/L; Supp. Data 3). 271
Our genomic predictions further indicate that microbialites host microbial populations 272 with high capacity for reductive metabolisms (Fig. 1). Microcosm experiments under 273 prolonged anoxia confirmed the production of large amounts of reduced inorganic 274 compounds (Fig. 3). Despite methanogens being low in abundance (av. mcrA 0.12% 275 of the community), their elevated metabolic rates enable them to contribute notably 276 to CH4 production (av. 28.7 ppm/h; Fig. 3), a characteristic previously documented in 277 other environments45. Similarly, sulfate reducers were scarce (av. asrA, dsrA: 278 2.44%) but released abundant S2- (av. 7.4 µM/h; Fig. 3). On the other hand, 279 fermenters (av. FeFe-hydrogenase: 34.37%; group 3 NiFe hydrogenase: 48.77%) 280 and nitrogen fixer (av. nifH: 27.49%) abundance was mirrored in the high H2 281
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
production (22.7 ppm/h; Fig. 3). These reduced compounds, diffusing through the 282 microbialite structure, fuel a diverse array of aerobic and anaerobic metabolic 283 pathways, thereby supporting microbial energy conservation and carbon fixation 284 within these dynamic ecosystems. 285
286 287 288
Chemosynthesis and photosynthesis contribute significantly to 289
microbialites' carbon fixation 290
291 While there’s long-standing evidence for photosynthesis in microbial mats19,46,47, 292 recent genomic studies suggest chemosynthetic pathways also facilitate primary 293 production25–27,48, though current evidence is fragmentary. To gain a comprehensive 294 understanding of the diversity of microbialite autotrophic populations, we performed 295 phylogenetic analyses of carbon fixation protein sequences. These phylogenetic 296 inferences were further supported by activity measurements using 14C incorporation 297 assays, conducted with a range of supplemental electron donors (Fig. 4). We 298 analysed the phylogenies of protein sequences encoding the acetyl-CoA synthase 299 (AcsB), ribulose 1,5-bisphosphate (RbcL) and ATP-citrate lyase (AclB) (Fig. 4a-c). 300 This analysis revealed an extraordinary diversity of sequences affiliated with at least 301 16 phyla across the CBB, WLP and rTCA (Fig. 4a-c). These sequences 302 encompassed multiple novel autotrophic clades including two Nanoarchaeota MAGs 303 encoding the CBB, one JAHJDO01 MAG encoding the WLP, and one Polyangia 304 MAG encoding the rTCA (Fig. 4a-c). 305 306 Across the 140 RuBisCO (RbcL) sequences, we recovered four subtypes (Fig. 4a). 307 Notably, RuBisCO subtypes IA-D exhibit intermediate to high specificity for CO2 and 308 are generally found across both aerobic and anaerobic habitats, whereas RuBisCO 309 subtype II has lower CO2 specificity and is more commonly associated with 310 anaerobic or microaerophilic environments49. Consistent with the prevalent 311 anaerobic niches of most microbialites, we recovered 184 acetyl-CoA synthase 312 (AcsB) protein sequences for the WLP (Fig. 4b). However, a large proportion of 313 these sequences most likely were not encoded by autotrophs but rather by members 314 of the phyla Actinomycetota, Desulfobacterota and Chloroflexota that generally 315 utilise the reverse WLP for acetate oxidation coupled with dissimilatory sulfate 316 reduction. Conversely, sequences associated with the Firmicutes, Planctomycetota 317 and Spirochaetota likely represent carbon-fixing acetogens, consistent with prior 318 studies characterizing these phyla50,51. 319
To validate the potential for carbon fixation highlighted through genomic and 320 phylogenetic analysis, we quantified relative chemosynthetic and photosynthetic 321
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
carbon fixation rates using radiolabeled carbon dioxide (14CO2). To capture both 322 aerobic and anaerobic carbon fixation processes, we conducted five-day 323 experiments, during which the microcosms transitioned to anaerobic conditions after 324 approximately 24 hours. This transition was inferred based on O2 measurements 325 taken simultaneously on other microcosms. 14CO2 incorporation was detected in all 326 microcosms, except for H2-stimulated chemosynthesis in microbialite C (Fig. 4d, 327 Supp. Fig. 3), which most likely resulted from an experimental issue. As anticipated, 328 microcosms exposed to light supported higher 14CO2 incorporation through 329 photosynthesis (av. 0.0122 nmol day
¹ g wet
1) compared to those incubated in the 330 dark (av. 0.00279 nmol day
¹ g wet
¹; Fig. 4d). Dark incubations likely captured the 331 combined activity of anaplerotic processes and baseline chemosynthetic carbon 332 fixation. The addition of electron donors significantly boosted chemosynthetic carbon 333 fixation rates (Fig. 4d). Consistent with the high abundance of genes and MAGs 334 encoding sqr and rdsrA genes (Fig. 1), S2--supplemented microcosms exhibited the 335 highest chemosynthetic 14CO2 incorporation rates (av. 0.0084 nmol day
¹ g wet
¹). 336 Similarly, NH4+ and H2 supplementation enhanced chemosynthetic carbon fixation 337 (av. 0.00741 and 0.0062 nmol day
¹ g wet
¹, respectively; Fig. 4d). Contrary to the 338 prevailing view that microbial productivity is predominantly driven by 339 photoautotrophs, our findings reveal that inorganic chemical sources and gas 340 substrates are also critical for the primary productivity of these ecosystems, 341 highlighting the significant role of chemosynthesis in supporting stratified microbial 342 communities. 343
344
345 346
Conclusions 347
Through an integrative approach, our study sheds light on the intricate molecular 348 exchanges that underpin the carbon acquisition, energy conservation, and broader 349 nutrient cycling within a model microbialite community, offering a detailed view into 350 how taxonomic and metabolic diversity intersect in such ecosystems. Microbialite 351 communities in the hypersaline West Basin Lake exhibit a metabolic diversity 352 comparable to that observed in other microbialite ecosystems worldwide (Fig. 1). 353 This diversity likely arises from the presence of organisms with complementary traits, 354 driven by resource facilitation among community members52,53. In microbialite 355 communities, metabolic synergies occur between organisms inhabiting contrasting 356 physicochemical niches26,54, which vary on adiel cycle The resultant mosaic of 357 ecological niches align with the physiological requirements of diverse microbial taxa. 358 Consistent with these observations, our analysis simultaneously reveals millimetre-359 scale overlap of functional guilds throughout the microbialite structure (Supp. Fig. 2) 360 and cycling of key metabolites such as iron, nitrogen and sulfur compounds across 361
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
their major redox states through intricate molecular handoffs (Supp. Data 2). These 362 processes promote minimal energy loss and enhance ecosystem productivity55. In a 363 broader context, the exceptional efficiency of elemental cycling and carbon use 364 efficiency within microbialite communities, combined with the duality of light- and 365 chemically-driven metabolic pathways, suggests that these ecosystems have likely 366 served as hotspots of metabolic innovation throughout Earth's history. 367 368 369
Methods 370
Characterization of microbial mat organosedimentary structures 371
The eastern shoreline of West Basin Lake hosts abundant microbialites and living 372 microbial matgrounds56. Well-developed microbialite buildups and hardgrounds are 373 exposed on subaerially-exposed benches up to several metres above the current 374 lake level. Previous studies have documented extensive sub-aqueous, well-lithified 375 living microbialites with considerable relief above the lake floor, especially in the 376 adjacent East Basin Lake56. The living microbialites documented in this study occur 377 in shallow water of around 0.5 m depth, and are best-developed in benches 378 approximately 1-2 m from the current shoreline in West Basin Lake (Supp. Fig. 4). 379 These living microbial matgrounds appear dark orange-red and have an irregular, 380 undulating surface with a broad-scale relief of several centimetres. In cross section, 381 the mats have a dark orange-red film over a hard to friable, mineralised, cream-white 382 uppermost layer of up to several millimetres in thickness (Supp. Fig. 1). This 383 mineralised layer has an irregular, pustular appearance and is composed of 384 carbonate, likely hydromagnesite56. Underneath this, the mat consists of weakly 385 lithified to unlithified mud with several thin colour zones over a several millimetres 386 depth (dark green, purple and grey), followed by up to several centimetres of light 387 green mat (Supp. Fig. 1; Supp. Fig. 4). The sediment texture is commonly clotted 388 and unlaminated. The substrate for the mat is dark, organic-rich mud which may 389 have weak layering preserved through the presence of occasional coarse sediment 390 laminae. Here we use the term ‘microbialite’ to define these microbial matgrounds, 391 as they do form mineralised structures with cm-scale relief above the lake floor. 392 While these modern and recent microbialites are not as extensively developed as the 393 sub-aerially exposed microbialites, they do appear to form significant accretionary 394 structures over time (for example encrusting anthropogenic debris). 395
Sample collection, processing and physicochemical parameters 396
Forming microbialites were collected from West Basin Lake (38° 19' 24.6468'' S, 397 143° 26' 51.8928'' E), a hypersaline (salinity 6.5-9.5%) inland crater lake located in 398 Victoria, Australia. Microbialites samples were collected during three field trips in 399 November 2022, March 2024 and May 2024 using pre-sterilized spades and 400
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
transferred into pre-sterilized 5-gallon buckets containing lake water. At first 401 collection in November 2022, the water temperature was 18.9 °C, dissolved oxygen 402 90%, pH 8.47 and oxidative reduction potential 2.72 mV. Irradiance at noon was 403 measured using a Walz Universal Light Meter (ULM-500) equipped with a Mini 404 Quantum Sensor (LS-C) at 30 cm and 50 cm below the water surface was ~650 405
μ
mol photons m–2 s–1 and ~450
μ
mol photons m–2 s–1, respectively. Physicochemical 406 parameters for lakewater and microbialite depth-profile subsamples were measured 407 at the TrACEES Platform, University of Melbourne. The selected physicochemical 408 parameters include total organic carbon, total carbon, inorganic carbon, total 409 nitrogen, sodium, magnesium, potassium, calcium, sulfate and chlorum. Nitrate was 410 measured at Water Studies, Monash University. 411 412
16S rRNA gene sequencing and community analysis 413
Total DNA was extracted from the layers of two different microbialite samples (three 414 technical triplicate per layer) using the DNeasy PowerSoil Kit (Qiagen, Hilden, 415 Germany) on 0.5g of material. Two sample-free negative control were also included. 416 Extracted DNA samples were sent to AGRF (Melbourne, Victoria) for library 417 preparation, PCR-amplification and sequencing of the 16S rRNA gene V1-V3 418 regions on an Illumina MiSeq platform, 2x300bp paired-end reads. Sequences were 419 processed using the QIIME2 pipeline v 2022.257. Primer sequences were trimmed 420 using Cutadapt58, while DADA2 was employed for merging forward and reverse 421 reads, quality filtering, dereplication, and chimera removal59. Taxonomic 422 classification was performed with QIIME2's feature-classifier plugin. The SILVA v132 423 QIIME release was used for 16S rRNA gene taxonomy60. 424 425
Community DNA extraction and sequencing 426
Each microbialite sample was homogenised into a slurry. Total DNA was extracted 427 from three different microbialite slurry samples (technical triplicate per sample for a 428 total of nine samples) using the DNeasy PowerMax Soil Kit (Qiagen, Hilden, 429 Germany) on 10 g of materials as per the manufacturer’s protocol. A sample-free 430 negative control was also included. Extracted DNA samples were sent to AGRF 431 (Melbourne, Victoria) for library preparation and sequencing on two lanes using an 432 Illumina NovaSeq SP Flow-cell, 2 x 150 bp for 500 cycles. 433 434 435
Reads quality control, assembly and binning 436
Across the three technical replicates of each microbialite sample, we obtained an 437 average of over 16 million read pairs for sample 1, over 56 million read pairs for 438 sample 2, and over 40 million read pairs for sample 3. Reads quality control, 439 assembly and binning were implemented within the Metaphor pipeline61. Specifically, 440
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
raw reads derived from the nine metagenome libraries were quality-controlled by 441 trimming primers and adapters, followed by artefacts and low-quality read filtering 442 using fastp62 with parameters length_required 50, cut_mean_quality 30, and extra: --443 detect_adapter_for_pe. The nine quality-controlled metagenomes were coassembled 444 using MEGAHIT v1.2.963 with default parameters. Contigs shorter than 1,000 bp 445 were removed. Assembled contigs were binned using Vamb v4.1.364, MetaBAT 446 v2.12.165 and CONCOCT v1.1.066. The three bin sets were then refined using DAS 447 Tool v1.1.667 and de-replicated using dRep v3.4.268 with 95% ANI integrated with 448 CheckM269. Bins completeness and contamination were estimated using CheckM269. 449 After dereplication, we recovered 331 between medium (completeness >50%, 450 contamination <10%) and high-quality (completeness >90%, contamination <5%) 451 metagenome-assembled genomes (MAGs), according to the MIMAG standard70. 452 MAG taxonomy was assigned according to Genome Taxonomy Database Release 453 R21471 using GTDB-Tk v2.3.272. CoverM v0.6.1 was used to calculate the relative 454 abundance of bins based on the metagenomic reads 455 (https://github.com/wwood/CoverM). 456 457 458
Functional annotation of binned contigs and unbinned contigs 459
The sequences of 51 marker genes representing energy conservation, carbon 460 fixation, trace gas metabolism, sulfur cycle, nitrogen cycle, arsenic cycle, iron cycle, 461 formate oxidation, phototrophy, and aerobic respiration were retrieved from the 331 462 MAGs and unbinned contigs. Open reading frames (ORFs) were predicted using 463 Prodigal v2.6.373, then annotated using DIAMOND blastp74 homology-based 464 searches against a custom database75 of 51 metabolic marker gene sets described 465 below. DIAMOND mapping was performed with a query coverage threshold of 80% 466 for all databases, and a percentage identity threshold of 80% (for psaA) 75% (for 467 mcr, hbsT), 70% (for isoA, psbA, ygfK, aro, atpA), 60% (amoA, pmoA, mmoA, coxL, 468 [FeFe], nxrA, rbc, nuoFL) or 50% (all other databases). MAGs were also annotated 469 using DRAM76. 470
471 472
Metabolic annotation of metagenomic short reads 473
Paired-end reads from the nine samples were stripped of adapter and barcode 474 sequences, then contaminating PhiX and low-quality sequences were removed 475 (minimum quality score 20) using the BBDuk function of BBTools v.36.92 476 (https://sourceforge.net/projects/bbmap/). Resultant quality-filtered forward reads 477 with lengths of at least 100 bp were searched for the presence of the 51 marker 478 genes described above using the DIAMOND blastx algorithm77. Specifically, reads 479 were compared against the custom-made reference databases75 of 51 metabolic 480 marker genes for energy conservation, carbon fixation, phototrophy, and hydrogen, 481
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
carbon monoxide, methane, sulfur, nitrogen, and iron cycling. A query coverage of 482 80% and an identity threshold of 80% (for psaA), 75% (for hbsT), 70% (for atpA, 483 psbA, isoA, ygfK, aro), 60% (for amoA, mmoA, coxL, FeFe, nxrA, rbcL, nuoF) and 484 50% (for all others) was used. The proportion of community members encoding each 485 gene was calculated by normalizing the gene's read count (measured in reads per 486 kilobase million [RPKM]) against the average RPKM of 14 universal single-copy 487 ribosomal marker genes. 488 489 490
Phylogenetic analysis 491
Maximum-likelihood phylogenetic trees for archaeal and bacterial MAGs were built 492 using GTDBtk72 commands identify and align on high quality MAGs (completion 493 >90% and contamination <5%). The archaral and bacterial trees were built using IQ-494 TREE v2.2.2.678,79 with 1,000 ultrafast bootstrap80 using the LG+C10+F+G and 495 WAG+G20 models, respectively. MUSCLE81 was used to align 36 AclB, 184 AcsB 496 and 140 RbcL proteins retrieved between the MAGs and unbinned contigs. Aclb, 497 AcsB and RbcL maximum-likelihood phylogenetic trees were built using IQ-TREE 498 v2.2.2.678,79 with 1,000 ultrafast bootstrap80 and models LG+F+I+R6 for the AcsB 499 tree, LG+I+G4 for the AclB tree and LG+R5 for the RbcL tree. All trees were plotted 500 using iTOL v682. 501 502 503
Chemical imaging 504
The O2 sensitive optode preparation included mixing 100 mg of polystyrene, 1.5 mg 505 of indicator (PT (II) meso-tetra(pentafluorophenyl)porphine), 1.5mg of reference 506 (Macrolex yellow®) and dissolved in 1 g of solvent (Tetrahydrofuran) to form a 507 cocktail. The O2 cocktails were knife-coated on dust-free polyester foils 508 (goodfellow.com) and the final thickness of the coating was <2
μ
m. Once dry, the O2 509 sensitive optode was coated with an antirefractory layer. The antirefractory cocktail 510 preparation included mixing 100 mg hydrogel D4, 100mg carbon black and 1 g of 511 100% ethanol. The antirefractory cocktail were knife-coated on top of the O2 512 sensitive optode and the final thickness of the coating was <3
μ
m. 513
The experimental setup included a modified digital single-lens reflex camera (Canon 514 EOS 1000D) with its near-infrared (NIR) blocking filter removed and equipped with a 515 Sigma 50 mm F2.8 EX DG Macro lens. An emission filter (Schott 530 nm, 516 Uqgoptics.com) was fitted to the lens to detect oxygen fluorescence. Following the 517 protocol described by Larsen et al. (RED), O2 sensitive optode were excited by four 518 high-power blue LEDs (l-peak = 445 nm, LXHL-LR3C, Luxeon, F = 340 mW at IF = 519 700 mA) combined with a 470 nm short-pass filter (blue dichroic color filter, 520 Uqgoptics.com). Microbialite sample cross-sections were illuminated using a Schott 521
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
Leica KL 2500 LCD Cold Light Source. All components were synchronized via a 522 trigger box (https://imaging.fish-n-chips.de) and controlled using the custom software 523 Look@RGB. Each planar optode was calibrated individually in an aquarium 524 maintained at a constant seawater temperature of 20 ± 1°C in a darkened room. The 525
calibration range for the O2 sensitive optode was 0 – 360
μ
mol L
−1
. 526
All experiments were conducted in a dark room at a constant 20 ± 1 °C to resemble 527 the lake temperature at the time of sampling. Three microbialite samples were cut 528 using a diamond saw exposing their cross-sections. Each of these samples was 529 placed in a 4 L glass aquarium pressing on the O2 sensitive optode, which was 530 attached to the aquarium side. Following overnight acclimation in the aquaria, 531 microbialite sample cross-sections were illuminated from above with approximately 532
450
μ
mol m
−2
s
−1
of light. Irradiance levels in the experimental setup for defined lamp 533
settings were measured using a Walz Universal Light Meter (ULM-500) equipped 534 with a Mini Quantum Sensor (LS-C). Image sequences capturing O2 dynamics 535 across the microbialite cross-sections were taken every 5 minutes. Lake water in the 536 aquarium was aerated with an aeration stone connected to an air pump. 537
Downstream data analysis was conducted using ImageJ v1.53K. Each image was 538 separated into Red, Green, Green2, and Blue RAW TIFF channels. The ImageJ 539 plugin Ratio Plus was used to calculate the ratio of the Red to Green channels (R/G). 540 The resulting ratio images were color-coded using the ‘Fire’ lookup table to visualise 541 O2 dynamics. Calibration was performed using the Curve Fitting function, applying an 542 exponential fit with offset for O2, based on planar optode calibration values. 543 Brightness and contrast settings were adjusted to display minimum and maximum 544
values of 0 – 360
μ
mol L
−1
for O2 images. To minimize influence of water infiltrating 545
in between the O2 optode and microbialite cross-section and to quantify microbial O2 546 production, the first image of each experiment was subtracted from subsequent 547 images using the Image Calculator function. To identify the photosynthetic regions 548 on the microbialite sample cross-sections, we overlaid images with the highest O2 549 production onto microbialite cross-section images. The portions of each microbialite 550 sample that showed O2 production were identified as regions on interes (ROI). 551 Subsequently values were extracted from the ROI. Control sample in Fig. 3 was an 552 image sequence of ROI measuring a deeper, anaerobic microbialite cross-section 553 portion simultaneously to O2 measurements. Proxies for the net photosynthesis (PN) 554 and apparent dark respiration (RD) were calculated by subtracting images taken with 555 a 5 min interval when O2 production and respiration were the highest, respectively. 556 Gross photosynthesis was estimated as PG = PN+|RD|. 557
558
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
Ex-situ biogeochemical measurements 559
We conducted microcosm experiments to evaluate the aerobic and anaerobic 560 metabolism of microbialite communities. Each experiment was performed in 561 triplicate, with three independent microbialite samples per replicate. Heat-killed 562 samples were prepared using gamma radiation followed by one autoclave cycle at 563 121 °C for 30 minutes and served as controls. These controls confirmed that the 564 observed element dynamics were attributable to biotic processes. 565
Aerobic and anaerobic microcosms, including trace gas and S2- and Fe2+ additions, 566 were set up in 120 mL serum vials containing 50 mL of 0.22
μ
m-filtered lake water 567 and approximately 10 g of microbialite to create slurries. The vials were sealed with 568 butyl rubber septa. For aerobic microcosms, the headspace was left with ambient air, 569 while anaerobic microcosms were flushed with helium for 10 minutes to remove O2 570 and then supplemented with 1.5 mM NO3- as an electron acceptor (except for S2- 571 consumption). The weights of microbialite used in each microcosm were recorded 572 and used to normalise downstream calculations. 573
Trace gas microcosms were supplemented with 10 ppm H2, CH4, and CO in the 574 headspace. Sampling of the headspace began immediately after the addition of 575 electron donors and acceptors, with 2 mL of gas extracted at variable time intervals 576 as shown in Fig. 3. In anaerobic vials, the sampled gas volume was replaced with 577 He. Gas concentrations were analysed by gas chromatography using a pulsed 578 discharge helium ionization detector (model TGA-6791-W-4U-2, Valco Instruments 579 Company Inc.), with calibration based on certified standard mixtures of H2, CH4, and 580 CO (0, 10, 100 ppm in N2, BOC Australia). 581
In anaerobic S2- and Fe2+ microcosms, either 100
μ
M Na
S·9H
O (only for 582 consumption) or 6 mM FeCl
was added to helium-purged slurries. At each 583 timepoint, 3 mL of water was sampled and filtered through 0.45
μ
m pore-size filters. 584 For S2- analysis, 2 mL of the filtered sample was preserved with ZnAc, while for Fe2+ 585 analysis, 1 mL was preserved with ferrozine. Both S2- and Fe2+ concentrations were 586 measured using a GBC UV-Visible 918 spectrophotometer, following methods 587 described previously83. 588
Microcosms for nitrification were prepared in uncapped 250 mL Schott bottles 589 containing 100 mL of 0.22
μ
m-filtered lake water, 100
μ
M NH4+, and approximately 590 10 g of microbialite sample. At each sampling timepoint, 10 mL of water was filtered 591 through 0.45
μ
m pore-size filters and stored frozen until further analysis. The filtered 592 samples were analysed for NOx (NO2- + NO3-) concentrations using a Lachat 593 QuikChem 8000 Flow Injection Analyzer (FIA) in accordance with APHA methods84. 594 For oxygen measurements, microcosms were prepared in 120 mL vials containing 595 100 mL of 0.22
μ
m-filtered lake water and approximately 10 g of microbialite slurry. 596 Dissolved O2 concentrations were monitored using a FireSting oxygen probe 597 (PyroScience) until the microcosms approached anoxic conditions. 598
599
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
14C incorporation analysis 600
0.25 g of homogenised microbialite sample with 1 mL of filtered lake water (0.22
μ
m) 601 were prepared in 7 mL scintillation vials with ambient air headspaces. Radiolabeled 602 sodium bicarbonate solution (NaH14CO3, Perkin Elmer, 53.1 mCi nmol-1) was added 603 to a concentration ~0.1 µM. Triplicates of each sample were prepared and subjected 604
to five different conditions, namely light (40
μ
mol m
−2
s
−1
), dark, dark + H2 (100 ppm), 605
dark + S2- (800 µM) and dark + NH4+ (1 mM) and incubated for 5 days. These 606 experiments aimed at capturing both aerobic and anaerobic carbon incorporation 607 metabolisms and according to parallel O2 measurements, microcosms likely 608 transitioned to anaerobic after approximately 24 h. After the incubation period, 609 concentrated HCl was added dropwise to each vial and left for 24h with intermittent 610 shaking to ensure excess unbound dissolved inorganic carbon (DIC) was acidified 611 and released as 14CO2. HCl was added equally to all vials until bubble production 612 ceased and then they were placed at 60 °C under a heat lamp to dry. When dry, 7 613 mL of scintillation liquid (EcoLume™, MP Biomedical) was added and 14C measured 614 on an automated liquid scintillation counter (Tri-Carb 2810 TR, Perkin Elmer). 615 Photosynthetic 14C incorporation values were adjusted to account for the 616 photosynthetic ROI areas present on the 0.25g portion of microbialite used in the 617 light treatment. Assuming that the photosynthetic ROI areas identified through 618 chemical imaging represent the microbialite sample regions performing oxygenic 619 photosynthesis, we normalised by applying the ratio of photosynthetic ROI area of 620 microbialite sample 3 which had the largest photosynthetic surface area 621 (photosynthetic ROI sample 1: 3.11%, sample 2: 2.73%, and sample 3: 6.35%) to 622 total microbialite area. We developed a hierarchical Bayesian model to analyze 14C 623 incorporation rates under different conditions. The model is implemented in Stan and 624 R (see SI for code). The model consists of N total observations, with I samples, J 625 conditions, and K replicates for each sample condition combination. As the Dark+H2 626 condition experiment failed for microbialite C, we ran two models: one excluding 627 microbialite C and included all conditions (where I=2, J=5, K=3 and N=30), and one 628 excluding the Dark+H2 condition and included all samples (where I=3, J=4, K=3 and 629 N=36), so that in each case the model can be fitted to a complete dataset. The 630 likelihood function is given by: 631 632
,
, 633
634 where
ߙ
describes the sample specific effect for the ith sample,
describes the 635
condition specific effect for the jth sample, and
ߪ
2 is the unexplained variance in the 636 data. The Dark condition (j=1) was prescribed to be the control by setting
ߚ
637
0.
Weakly informative priors were used. Priors for
ߙ
(
1:
) and
ߚ
(
2:
) 638
were specified as normal distributions with variances an order of magnitude larger 639 than the variance of the total dataset. The prior for
ߪ
was a Cauchy distribution. 640 Posterior probability distributions for values of
ߚ
are presented in figure 4D, with 641
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
Credible intervals defined between the 2.5th and 97.5th, and the 25th and 75th 642 percentiles. 643
644
Statistics and visualization 645
Downstream statistical analyses were performed in RStudio (version 1.2.5033) using 646 several R packages, including decontam85, ggplot286, phyloseq87, and vegan88. 647 Illustrator v24.0.2 was utilised for figure editing. 648 649
Data availability 650
All data supporting the findings of the present study are available. We deposited all 651
the project’s sequences at the Sequence Read Archive. Metagenomic sequences 652
have accession number PRJNA1194634, 16S rRNA gene sequences have 653
accession number PRJNA1194668, and MAGs have been submitted and awaiting 654
release. 655
656
Conflict of interest 657
The authors declare no conflict of interest. 658 659
Acknowledgments 660
We acknowledge Judy and Leon for kindly allowing us to conduct our study on their 661 property at West Basin Lake. F.R. was supported by the Early Career Postdoctoral 662 Fellowship (ECPF24-4273843556) awarded by the Faculty of Medicine, Nursing and 663 Health Science at Monash University, and internal funding awarded to H.M. from the 664 University of Melbourne. P.M.L. was supported by an ARC DECRA Fellowship 665 (DE250101210). C.G. was supported by an NHMRC EL2 Fellowship (APP1178715). 666 A.H. was supported by an ARC DECRA Fellowship (DE190100988). This study used 667 the MASSIVE M3 supercomputing infrastructure. 668 669
Author contribution 670
671
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
F.R., C.G., P.M.L., and H.M. conceptualised the study. Experimental planning and 672 design were conducted by F.R., C.G., T.H., V.E., W.W.W., and P.C. Fieldwork was 673 conducted by F.R., H.M. and A.H. A.H. provided microbialite description. Data 674 analysis was led by F.R. Gas chromatography measurements were carried out by 675 F.R., T.H., and T.N., while V.E. and W.W.W. oversaw nitrification measurements. 676 Chemical imaging, oxygen, sulfide and ferrous oxide measurements were performed 677 by F.R. Carbon fixation incubations were performed by T.H. Metagenome analysis, 678 MAG construction, and annotation were completed by F.R., with extensive 679 bioinformatics support provided by V.W.S. Resources, supervision, and funding were 680 contributed by C.G., P.C., and H.M. The manuscript was written by F.R., with input 681 from all authors. 682 683
References 684
1. Walter, M. R., Buick, R. & Dunlop, J. S. R. Stromatolites 3,400–3,500 Myr old from 685
the North pole area, Western Australia. Nature 284, 443–445 (1980). 686
2. Van Kranendonk, M. J., Philippot, P., Lepot, K., Bodorkos, S. & Pirajno, F. Geological 687
setting of Earth’s oldest fossils in the ca. 3.5 Ga Dresser formation, Pilbara Craton, 688
Western Australia. Precambrian Research 167, 93–124 (2008). 689
3. Dick, G. J., Grim, S. L. & Klatt, J. M. Controls on O 2 Production in Cyanobacterial 690
Mats and Implications for Earth’s Oxygenation. Annu. Rev. Earth Planet. Sci. 46, 123–147 691
(2018). 692
4. Cavalier-Smith, T. Cell evolution and Earth history: stasis and revolution. Phil. Trans. 693
R. Soc. B 361, 969–1006 (2006). 694
5. Nisbet, E. G. & Fowler, C. M. R. Archaean metabolic evolution of microbial mats. 695
Proc. R. Soc. Lond. B 266, 2375–2382 (1999). 696
6. Hoehler, T. M., Bebout, B. M. & Des Marais, D. J. The role of microbial mats in the 697
production of reduced gases on the early Earth. Nature 412, 324–327 (2001). 698
7. Des Marais, D. J. Earth’s early biosphere. Gravitational and space biology bulletin: 699
publication of the American Society for Gravitational and Space Biology 11, 23–30 (1998). 700
8. Hawes, I., Sumner, D. Y., Andersen, D. T., Jungblut, A. D. & Mackey, T. J. 701
Timescales of growth response of microbial mats to environmental change in an ice-702
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
covered Antarctic lake. Biology 2, 151–176 (2013). 703
9. Nold, S. C. et al. Underwater sinkhole sediments sequester Lake Huron’s carbon. 704
Biogeochemistry 115, 235–250 (2013). 705
10. Boidi, F. J., Mlewski, E. C., Gomez, F. J. & Gérard, E. Characterization of 706
Microbialites and Microbial Mats of the Laguna Negra Hypersaline Lake (Puna of 707
Catamarca, Argentina). in Microbial Ecosystems in Central Andes Extreme Environments 708
(ed. Farías, M. E.) 183–203 (Springer International Publishing, Cham, 2020). 709
doi:10.1007/978-3-030-36192-1_13. 710
11. Paul, V. G., Wronkiewicz, D. J., Mormile, M. R. & Foster, J. S. Mineralogy and 711
Microbial Diversity of the Microbialites in the Hypersaline Storr’s Lake, the Bahamas. 712
Astrobiology 16, 282–300 (2016). 713
12. Bischoff, K. et al. Spherulitic microbialites from modern hypersaline lakes, Rottnest 714
Island, Western Australia. Geobiology 18, 725–741 (2020). 715
13. Osman, J. R. et al. Key energy metabolisms in modern living microbialites from 716
hypersaline Andean lagoons of the Salar de Atacama, Chile. Science of The Total 717
Environment 937, 173469 (2024). 718
14. Pace, A. et al. Microbial and diagenetic steps leading to the mineralisation of Great 719
Salt Lake microbialites. Scientific Reports 6, 31495 (2016). 720
15. Laval, B. et al. Modern freshwater microbialite analogues for ancient dendritic reef 721
structures. Nature 407, 626–629 (2000). 722
16. Valdespino-Castillo, P. M. et al. Exploring biogeochemistry and microbial diversity of 723
extant microbialites in Mexico and Cuba. Frontiers in microbiology 9, 510 (2018). 724
17. Centeno, C. M. et al. Microbialite genetic diversity and composition relate to 725
environmental variables. FEMS microbiology ecology 82, 724–735 (2012). 726
18. Myshrall, K. L. et al. Biogeochemical cycling and microbial diversity in the thrombolitic 727
microbialites of Highborne Cay, Bahamas. Geobiology 8, 337–354 (2010). 728
19. Dupraz, C. & Visscher, P. T. Microbial lithification in marine stromatolites and
729
hypersaline mats. Trends in Microbiology 13, 429–438 (2005). 730
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
20. Sabbe, K., Verleyen, E., Hodgson, D. A., Vanhoutte, K. & Vyverman, W. Benthic 731
diatom flora of freshwater and saline lakes in the Larsemann Hills and Rauer Islands, 732
East Antarctica. Antarctic science 15, 227–248 (2003). 733
21. Patterson, D. J. & Simpson, A. G. Heterotrophic flagellates from coastal marine and 734
hypersaline sediments in Western Australia. European Journal of Protistology 32, 423–735
448 (1996). 736
22. Hauer, G. & Rogerson, A. Heterotrophic Protozoa from Hypersaline Environments. in 737
Adaptation to Life at High Salt Concentrations in Archaea, Bacteria, and Eukarya (eds. 738
Gunde-Cimerman, N., Oren, A. & Plemenitaš, A.) vol. 9 519–539 (Springer-Verlag, 739
Berlin/Heidelberg, 2005). 740
23. Post, F. J., Borowitzka, L. J., Borowitzka, M. A., Mackay, B. & Moulton, T. The 741
protozoa of a Western Australian hypersaline lagoon. Hydrobiologia 105, 95–113 (1983). 742
24. Toneatti, D. M., Albarracín, V. H., Flores, M. R., Polerecky, L. & Farías, M. E. 743
Stratified bacterial diversity along physico-chemical gradients in high-altitude modern 744
stromatolites. Frontiers in Microbiology 8, 1–12 (2017). 745
25. Wong, H. L., Smith, D. L., Visscher, P. T. & Burns, B. P. Niche differentiation of 746
bacterial communities at a millimeter scale in Shark Bay microbial mats. Scientific Reports 747
5, 1–17 (2015). 748
26. Shoemaker, A. et al. Wood–Ljungdahl pathway encoding anaerobes facilitate low-749
cost primary production in hypersaline sediments at Great Salt Lake, Utah. FEMS 750
Microbiology Ecology 100, fiae105 (2024). 751
27. Wong, H. L., MacLeod, F. I., White, R. A., Visscher, P. T. & Burns, B. P. Microbial 752
dark matter filling the niche in hypersaline microbial mats. Microbiome 8, 135 (2020). 753
28. Saghaï, A. et al. Comparative metagenomics unveils functions and genome features 754
of microbialite
associated communities along a depth gradient. Environmental 755
Microbiology 18, 4990–5004 (2016). 756
29. Gutiérrez-Preciado, A. et al. Functional shifts in microbial mats recapitulate early 757
Earth metabolic transitions. Nat Ecol Evol 2, 1700–1708 (2018). 758
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
30. Ragsdale, S. W. & Pierce, E. Acetogenesis and the Wood–Ljungdahl pathway of 759
CO2 fixation. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics 1784, 1873–760
1898 (2008). 761
31. Fuchs, G. Alternative Pathways of Carbon Dioxide Fixation: Insights into the Early 762
Evolution of Life? Annu. Rev. Microbiol. 65, 631–658 (2011). 763
32. Campbell, M. A. et al. Functional gene expression in Shark Bay hypersaline microbial 764
mats: adaptive responses. Frontiers in Microbiology 11, 560336 (2020). 765
33. Bernstein, H. C. et al. Trade-offs between microbiome diversity and productivity in a 766
stratified microbial mat. The ISME journal 11, 405–414 (2017). 767
34. Moorhead, D. L., Wolf, C. F. & Wharton, R. A. Impact of light regimes on productivity 768
patterns of benthic microbial mats in an antarctic lake: A modeling study. Limnology & 769
Oceanography 42, 1561–1569 (1997). 770
35. Williams, T. J., Allen, M. A., Berengut, J. F. & Cavicchioli, R. Shedding light on 771
microbial “dark matter”: insights into novel Cloacimonadota and Omnitrophota from an 772
Antarctic Lake. Frontiers in Microbiology 12, 741077 (2021). 773
36. Li, Y., Liu, H., Xiao, Y. & Jing, H. Metagenome sequencing and 982 microbial 774
genomes from Kermadec and Diamantina Trenches sediments. Scientific Data 11, 1067 775
(2024). 776
37. Fang, Y. et al. Casting Light on the Adaptation Mechanisms and Evolutionary History 777
of the Widespread Sumerlaeota. mBio 12, e00350-21 (2021). 778
38. Edgcomb, V. P. et al. Active eukaryotes in microbialites from Highborne Cay, 779
Bahamas, and Hamelin Pool (Shark Bay), Australia. The ISME journal 8, 418–429 (2014). 780
39. Couradeau, E. et al. Prokaryotic and eukaryotic community structure in field and 781
cultured microbialites from the alkaline Lake Alchichica (Mexico). PloS one 6, e28767 782
(2011). 783
40. Bonacolta, A. M., Visscher, P. T., Del Campo, J. & White III, R. A. The eukaryome of 784
modern microbialites reveals distinct colonization across aquatic ecosystems. npj Biofilms 785
and Microbiomes 10, 78 (2024). 786
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
41. Breitbart, M. et al. Metagenomic and stable isotopic analyses of modern freshwater 787
microbialites in Cuatro Ciénegas, Mexico. Environmental Microbiology 11, 16–34 (2009). 788
42. Kanik, M. et al. Unexpected Abundance and Diversity of Phototrophs in Mats from 789
Morphologically Variable Microbialites in Great Salt Lake, Utah. Appl Environ Microbiol 790
86, e00165-20 (2020). 791
43. Ricci, F. et al. Fine
scale mapping of physicochemical and microbial landscapes of 792
the coral skeleton. Environmental Microbiology 1462-2920.16369 (2023) 793
doi:10.1111/1462-2920.16369. 794
44. Larsen, M., Borisov, S. M., Gunwald, B., Klimant, I. & Glud, R. N. A simple and 795
inexpensive, high-resolution color-ratiometric planar optode imaging approach: 796
Application to oxygen and pH sensing. Limnology and Oceanography: Methods 2, 348–797
360 (2011). 798
45. Hall, N. et al. Aerotolerant methanogens use seaweed and seagrass metabolites to 799
drive marine methane emissions. bioRxiv 2024–10 (2024). 800
46. Martínez-Alonso, M. et al. Distribution of phototrophic populations and primary 801
production in a microbial mat from the Ebro Delta, Spain. International Microbiology 7, 802
19–25 (2004). 803
47. Canfield, D. E. & Des Marais, D. J. Biogeochemical cycles of carbon, sulfur, and free 804
oxygen in a microbial mat. Geochimica et Cosmochimica acta 57, 3971–3984 (1993). 805
48. Kurth, D. et al. Arsenic metabolism in high altitude modern stromatolites revealed by 806
metagenomic analysis. Scientific reports 7, 1024 (2017). 807
49. Prywes, N., Phillips, N. R., Tuck, O. T., Valentin-Alvarado, L. E. & Savage, D. F. 808
Rubisco Function, Evolution, and Engineering. Annu. Rev. Biochem. 92, 385–410 (2023). 809
50. Garritano, A. N., Song, W. & Thomas, T. Carbon fixation pathways across the 810
bacterial and archaeal tree of life. PNAS nexus 1, pgac226 (2022). 811
51. Fuchs, G. CO2 fixation in acetogenic bacteria: variations on a theme. FEMS 812
Microbiology Reviews 2, 181–213 (1986). 813
52. Loreau, M. & Hector, A. Partitioning selection and complementarity in biodiversity 814
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
experiments. Nature 412, 72–76 (2001). 815
53. Cardinale, B. J. et al. Effects of biodiversity on the functioning of trophic groups and 816
ecosystems. Nature 443, 989–992 (2006). 817
54. Dunham, E. C. et al. An ecological perspective on dolomite formation in Great Salt 818
Lake, Utah. Frontiers in Earth Science 8, 24 (2020). 819
55. Seto, M. & Iwasa, Y. Microbial material cycling, energetic constraints and ecosystem 820
expansion in subsurface ecosystems. Proc. R. Soc. B. 287, 20200610 (2020). 821
56. Last, W. M. & De Deckker, P. Modern and Holocene carbonate sedimentology of two 822
saline volcanic maar lakes, southern Australia. Sedimentology 37, 967–981 (1990). 823
57. Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data 824
science using QIIME 2. Nature biotechnology 37, 852–857 (2019). 825
58. Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing 826
reads. EMBnet. journal 17, 10–12 (2011). 827
59. Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina 828
amplicon data. Nature Methods 13, 581–583 (2016). 829
60. Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data 830
processing and web-based tools. Nucleic acids research 41, D590–D596 (2012). 831
61. Salazar, V. W. et al. Metaphor—a workflow for streamlined assembly and binning of 832
metagenomes. GigaScience 12, giad055 (2023). 833
62. Chen, S. Ultrafast one
pass FASTQ data preprocessing, quality control, and 834
deduplication using fastp. Imeta 2, e107 (2023). 835
63. Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-836
node solution for large and complex metagenomics assembly via succinct de Bruijn 837
graph. Bioinformatics 31, 1674–1676 (2015). 838
64. Nissen, J. N. et al. Improved metagenome binning and assembly using deep 839
variational autoencoders. Nat Biotechnol 39, 555–560 (2021). 840
65. Kang, D. D. et al. MetaBAT 2: an adaptive binning algorithm for robust and efficient 841
genome reconstruction from metagenome assemblies. PeerJ 7, e7359 (2019). 842
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
66. Alneberg, J. et al. Binning metagenomic contigs by coverage and composition. 843
Nature methods 11, 1144–1146 (2014). 844
67. Sieber, C. M. K. et al. Recovery of genomes from metagenomes via a dereplication, 845
aggregation and scoring strategy. Nature Microbiology 3, 836–843 (2018). 846
68. Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and 847
accurate genomic comparisons that enables improved genome recovery from 848
metagenomes through de-replication. The ISME Journal 11, 2864–2868 (2017). 849
69. Chklovski, A., Parks, D. H., Woodcroft, B. J. & Tyson, G. W. CheckM2: a rapid, 850
scalable and accurate tool for assessing microbial genome quality using machine 851
learning. Nature Methods 20, 1203–1212 (2023). 852
70. Bowers, R. M. et al. Minimum information about a single amplified genome (MISAG) 853
and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nature 854
biotechnology 35, 725–731 (2017). 855
71. Parks, D. H. et al. GTDB: an ongoing census of bacterial and archaeal diversity 856
through a phylogenetically consistent, rank normalized and complete genome-based 857
taxonomy. Nucleic acids research 50, D785–D794 (2022). 858
72. Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk v2: memory 859
friendly classification with the genome taxonomy database. Bioinformatics 38, 5315–5316 860
(2022). 861
73. Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site 862
identification. BMC Bioinformatics 11, 119 (2010). 863
74. Altschul, S. F. et al. Gapped BLAST and PSI-BLAST: a new generation of protein 864
database search programs. Nucleic Acids Research 25, 3389–3402 (1997). 865
75. Greening, C. Greening lab metabolic marker gene databases. (2021). 866
76. Shaffer, M. et al. DRAM for distilling microbial metabolism to automate the curation of 867
microbiome function. Nucleic acids research 48, 8883–8900 (2020). 868
77. Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using 869
DIAMOND. Nature methods 12, 59–60 (2015). 870
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
78. Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K., Von Haeseler, A. & Jermiin, L. S. 871
ModelFinder: fast model selection for accurate phylogenetic estimates. Nature methods 872
14, 587–589 (2017). 873
79. Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic 874
inference in the genomic era. Molecular biology and evolution 37, 1530–1534 (2020). 875
80. Hoang, D. T., Chernomor, O., Von Haeseler, A., Minh, B. Q. & Vinh, L. S. UFBoot2: 876
improving the ultrafast bootstrap approximation. Molecular biology and evolution 35, 518–877
522 (2018). 878
81. Edgar, R. C. Muscle5: High-accuracy alignment ensembles enable unbiased 879
assessments of sequence homology and phylogeny. Nature Communications 13, 6968 880
(2022). 881
82. Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the 882
phylogenetic tree display and annotation tool. Nucleic Acids Research gkae268 (2024). 883
83. Islam, Z. F. et al. Two Chloroflexi classes independently evolved the ability to persist 884
on atmospheric hydrogen and carbon monoxide. ISME J 13, 1801–1813 (2019). 885
84. Rice, E. W., Bridgewater, L. & Association, A. P. H. Standard Methods for the 886
Examination of Water and Wastewater. vol. 10 (American public health association 887
Washington, DC, 2012). 888
85. Davis, N. M., Proctor, D. M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple 889
statistical identification and removal of contaminant sequences in marker-gene and 890
metagenomics data. Microbiome 6, 1–14 (2018). 891
86. Wickham, H. ggplot2. WIREs Computational Stats 3, 180–185 (2011). 892
87. McMurdie, P. J. & Holmes, S. phyloseq: an R package for reproducible interactive 893
analysis and graphics of microbiome census data. PloS one 8, e61217 (2013). 894
88. Dixon, P. VEGAN, a package of R functions for community ecology. Journal of 895
vegetation science 14, 927–930 (2003). 896
897
898
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
Figures legend 899
900 901 Figure 1 | Maximum-likelihood genome trees depicting the taxonomic diversity of 902 331 archaeal and bacterial metagenome-assembled genomes (MAGs), built with 903 1000 ultrafast bootstrap replicates using the LG+C10+F+G and WAG+G20 models, 904 respectively. The top heatmap highlights the metabolic potential of microbes co-905 encoding energy acquisition enzymes with carbon fixation pathways. The side 906 bargraph shows the number of MAGs in each microbial family (GTDB taxonomy). 907 The bottom heatmap shows the abundance of each gene in the metagenomic short 908 reads in West Basin Lake samples (bold) and across 17 publicly available 909 microbialite community metagenomes from five global sites. Homology-based 910 searches were used to identify signature genes encoding enzymes associated with 911 metabolic pathways. To infer abundance, read counts were normalised to gene 912 length and the abundance of single-copy marker genes. 913 914
Figure 2 | Metabolic model predicted based on genome- and gene-resolved data of 915
the dominant microbial guilds within the microbialite communities of West Basin 916 Lake. This model represents the inferred metabolic pathways and interactions based 917 on the genomic content of these guilds. Blue and brown lines indicate the direction of 918 electron acceptors and donors, respectively. Note that the graphic is an artistic 919 generalization of the data and should not be interpreted as an exact depiction of the 920 microbial community structure. Asterisk (*) denotes that specific metabolic marker 921 genes were exclusively recovered from metagenomic short-read data. 922 923
Figure 3 | Biogeochemical assays illustrating the metabolic activities of microbialite 924 communities under aerobic (a) and anaerobic (b, c) conditions in microcosms. 925
Oxygen dynamics were assessed using chemical imaging on independent 926 microbialite samples incubated in 4 L glass aquaria, with data presented as the 927 mean ± standard deviation for defined photosynthetic regions of interest. Trace 928 gases, sulfide and ferrous ion measurements were taken in 120 ml sealed serum vial 929 containing 10 g of microbialite slurry and 50 mL of 0.22
μ
m-filtered lake water. Trace 930 gas microcosms were supplemented with 10 ppm H2, CH4, and CO in the 931 headspace. S2- and Fe2+ microcosms were supplied with either 100
μ
M Na
₂
S·9H
₂
O 932
(only for consumption) or 6 mM FeCl
₂
. All anaerobic microcosms except for S2-
933 production were supplemented with 1.5 mM NO3- as an electron acceptor. 934 Nitrification (NOx = NO2- + NO3-) measurements were taken in uncapped 250 mL 935 Schott bottles containing approximately 10 g of microbialite slurry, 100 mL of 0.22 936
μ
m-filtered lake water and 100
μ
M NH4+. In the oxygen plot, asterisks (*) indicate the 937 time when light was switched off, mimicking the onset of darkness. All microcosm 938 experiments were performed in triplicate, with results expressed as the mean ± 939
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint
standard deviation across three replicates. All measurements except for oxygen 940 dynamics were normalised to microbialite slurry wet weight. 941 942
Figure 4 | Maximum-likelihood phylogenetic trees of 140 RbcL (a), 184 AcsB (b), 943 and 36 AclB (c) amino acid sequences obtained from the three microbialite samples, 944 constructed using 1000 ultrafast bootstrap replicates. The LG+R5 (a), LG+F+I+R6 945
(b), and LG+I+G4 (c
) substitution models were applied. Sequences derived from
946
binned contigs are highlighted in bold, while those from unbinned contigs are shown
947
without emphasis. Bootstrap support values ≥ 90 are denoted by white circles (
a–c). 948
Panel (d) Credible intervals of effect (electron donors) relative to dark condition 949 denoting 2.5th and 97.5th (thin bar) and 25th and 75th (thick bar) percentiles of 950 posterior probability of the effect of each assay condition relative to the dark 951 condition. 952 953
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted January 4, 2025. ; https://doi.org/10.1101/2024.12.18.629060doi: bioRxiv preprint