Formation of stromatolite lamina at the interface of oxygenic-anoxygenic photosynthesis.

Texte intégral

(1)1 2. DR. AURÉLIE PACE (Orcid ID : 0000-0001-6295-5118). 3. DR. RAPHAEL BOURILLOT (Orcid ID : 0000-0003-4641-9125). 5 6. Author Manuscript. 4. Article type. 7 8. m. 9. : Research Article. yg. 10 11. m. lam. ic--. ic. the inte he. 12. Aurélie Pace†,§,a, Raphaël Bourillot†, Anthony Bouton‡,‡‡, Emmanuelle Vennin‡, Olivier. 13. Braissant¥, Christophe Dupraz*, Thibault Duteil†, Irina Bundeleva‡, Patricia Patrier#,. 14. Serge Galaup†, Yusuke Yokoyama##, Michel Franceschi†, Aurélien Virgone‡‡ and Pieter. 15. T. Visscher**. 16 17. † Bordeaux. 18. § Université. 19. ‡ Laboratoire. 20. boulevard Gabriel, Dijon 21000, France;. 21. ¥Center for Biomechanics. 22. * Department. 23. Stockholm, 06269, Sweden;. 24. #. 25. 86022. 26. Poitiers, France;. 27. ##. 28. Planetary Sciences, University of Tokyo, 5-- 1- -5 Kashiwanoha, Chiba 277-- 8564, Japan. INP, G&E, EA 4592, F-- 33600, Pessac, France; Bordeaux Montaigne, G&E, EA 4592, F-- 33600, Pessac, France; Biogéosciences UMR 6282 UBFC/CNRS, Univ. Bourgogne Franche-- Comté, 6. and Biocalorimetry, University of Basel, Basel, Switzerland. of Geological Sciences, Stockholm University, Svante Arrhenius väg 8,. Université de Poitiers, UMR 6269 CNRS, HYDRASA, 40 avenue du Recteur Pineau,. Atmosphere and Ocean Research Institute, Department of Earth and. This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/gbi.12281. This article is protected by copyright. All rights reserved.

(2) 1. ‡‡ Total,. 2. **. 3. Road,. 4. Groton, CT 06340, USA. 6 7 8 9 10 11 12 13 14. Department of Marine Sciences, University of Connecticut, 1080 Shennecossett. Author Manuscript. 5. CSTJF, Avenue Larribau, 64018 Pau, France.. a Corresponding. author: aurelie.pace@gmail.com. 15. This study was supported by a grant from the Ministère de l’Enseignement Supérieur. 16. et de la Recherche attributed by the Ecole Doctorale 480 Montaigne Humanités and. 17. complementary funding from Total S.A. The authors are grateful to Emmanuelle. 18. Gérard (IPGP, France) for her help with the Confocal Laser Scanning Microscopy. 19. during this project and the previous ones, Dylan Chaillot (Hydrasa, Poitiers,. 20. France) for the complementary FTIR analyses, Olivier Mathieu and. 21. Philippe Amiotte-- Suchet (Biogéosciences Dijon, Université de Bourgogne) for their. 22. help with water chemistry analyses. Pieter Visscher acknowledges support by NSF. 23. EAR award no. 1561173.. 24 25 26 27 28 29. -. In modern stromatolites, mineralisation results from a complex interplay between This article is protected by copyright. All rights reserved.

(3) microbial metabolisms, the organic matrix and environmental parameters. Here, we. 2. combined biogeochemical, mineralogical and microscopic analyses with measurements. 3. of metabolic activity to characterise the mineralisation processes and products in an. 4. emergent (<18 months) hypersaline microbial mat. While the nucleation of Mg-silicates. 5. is ubiquitous in the mat, the initial formation of a Ca-Mg carbonate lamina depends on. 6. (i) the creation of a high-pH interface combined with a major change in properties of the. 7. exopolymeric substances at the interface of the oxygenic and anoxygenic. 8. photoautotrophic layers and (ii) the synergy between two major players of sulfur cycle,. 9. purple sulfur bacteria and sulfate-reducing bacteria. The repetition of this process over. 10. time combined with upward growth of the mat is a possible pathway leading to the. 11. formation of a stromatolite.. 12 13 14 15 16 17 18. Author Manuscript. 1. 19. Many benthic microorganisms are able to form biofilms through the production. 20. of exopolymeric substances (EPS; Decho, 2000; Flemming and Wingender, 2010). The. 21. organic biofilm matrix formed by EPS fulfills multiple functions for the microbial. 22. community, facilitating adhesion to- and interactions with sediments, providing. 23. protection to unfavourable environmental conditions (e.g., UV, toxic chemicals,. 24. predation), and enabling cell-to-cell communication (Marvasi et al., 2010). EPS is a key. 25. component of the architecture and functioning of microbial mats, which arguably. 26. represent the oldest ecosystem known on Earth (Tice and Lowe, 2006). These complex,. 27. laminated organosedimentary biofilms are characterized by high metabolic rates. 28. resulting in rapid cycling of elements (C, O, S, N, Fe). The combined metabolic reactions. 29. of the microbial community alter the geochemical conditions in the mat, and result in. 30. extreme diel fluctuations of e.g., oxygen, or sulfide concentrations and pH (Dupraz et al.,. 31. 2009). The high rates of carbon cycling are a consequence of tight coupling of CO 2 This article is protected by copyright. All rights reserved.

(4) fixation notably through oxygenic and anoxygenic photosynthesis (using either H 2 O or. 2. reduced S, H 2 , Fe2+ or As3+ as electron donor, respectively) and aerobic and anaerobic. 3. organic carbon oxidation (using e.g., SO 4 2-, Fe3+, or NO 3 - as oxidant; Sforna et al., 2016;. 4. Visscher and Stolz, 2005). The individual metabolic reactions each have a different effect. 5. on the geochemical conditions in the mat. For example, oxygenic and anoxygenic. 6. photosynthesis, sulfate reduction under some conditions and methanogenesis increase. 7. the pH and alkalinity; in contrast, aerobic heterotrophic bacteria, fermenters and. 8. chemolithotrophic sulfide-oxidizing bacteria decrease the pH and alkalinity (Gallagher. 9. et al., 2012; Visscher and Stolz, 2005). The capacity of the entire community metabolism. 10. to either increase or decrease the local alkalinity, referred to as the “alkalinity engine”,. 11. can impact mineral dissolution or precipitation. For example, an increase of the. 12. alkalinity can lead to the lithification of the mat through carbonate mineral precipitation. 13. (Dupraz et al., 2009). The formation of other minerals e.g., silicates, sulfides and oxides. 14. by microbial metabolism has also been documented (Arp et al., 2003; Dupraz et al.,. 15. 2011; Konhauser, 2007; Miot et al., 2009) Poorly crystallized Mg-silicates have been. 16. characterized in several hypersaline, alkaline lake and sea cave microbialites (e.g. Burne. 17. et al., 2014; Léveillé et al., 2002; Zeyen et al., 2015). Their precipitation could be. 18. enhanced by microbial metabolisms increasing pH (e.g. oxygenic photosynthesis, sulfate. 19. reduction, anaerobic Fe respiration; Tosca et al., 2011; Pace et al., 2016).. Author Manuscript. 1. 20. Several members of the microbial mat community, notably autotrophic. 21. organisms, produce copious amounts of exopolymeric substances (Decho et al., 2005).. 22. The specific composition and physicochemical properties of EPS (e.g., the cation binding. 23. capacity) can also inhibit or promote mineral precipitation and influence the mineralogy. 24. and the shape of mineral products (Braissant et al., 2007, 2009). Functional groups in. 25. freshly produced EPS initially bind cations, which can be released upon degradation of. 26. these polymers. The consumption of low-molecular weight organic carbon (LMWOC) as. 27. well as the degradation of complex EPS molecules is a joint venture between many. 28. heterotrophic bacteria allowing the liberation of matrix-bound cations, which can lead. 29. to mineral precipitation. In addition, the EPS can provide a nucleation site for mineral. 30. precipitation, further illustrating a major role of the organic matrix in microbe-mineral. 31. interactions (Dupraz et al., 2009). This article is protected by copyright. All rights reserved.

(5) The formation of microbialites results either from (1) trapping and binding of. 2. sediments, (2) mineral precipitation in the microbial mat or (3) a combination of both. 3. processes (Burne and Moore, 1987). A proper understanding of microbe-mineral. 4. interactions in modern mats is critical for the interpretation of the ancient counterparts.. 5. Repetitive thin lamination is a diagnostic feature of modern and ancient stromatolites,. 6. dating back to the Precambrian (Allwood et al., 2006; Nutman et al., 2016; Riding, 2011).. 7. Lamination in modern open marine stromatolites can form through successive cycles of. 8. grain trapping and binding by filamentous cyanobacteria and mineralization of surface. 9. biofilms (Reid et al., 2000; Riding, 2011; Seong-Joo et al., 2000). Alternatively, the. 10. lithification of alternating horizontally- versus vertically-oriented filamentous. 11. cyanobacteria (Dupraz et al., 2013) can result in laminated structures. The two models. 12. cited above were based on observations of (i) the organization and structure of. 13. laminated mats, (ii) the role of specific metabolisms (e.g., photosynthesis, sulfate. 14. reduction) and/or (iii) the cycling of mat types ultimately resulting in formation of. 15. stromatolite lamination. With the exception of these two models, precipitation. 16. mechanisms are often discussed but the formation of a laminated fabric is not studied.. Author Manuscript. 1. 17. Our study provides for the first time a quantification of the major microbial. 18. metabolisms at play in a lithifying stromatolite along a vertical (i.e., depth) profile. We. 19. selected the hypersaline Cayo Coco Lagoonal Network (CCLN), where lithifying. 20. laminated microbial mats are abundant, actively forming and produce copious amounts. 21. of EPS (Bouton et al., 2016). In order to avoid early diagenetic transformations, we. 22. focused on the surface of actively growing CCLN microbial mats (<18 months).. 23. Metabolic activity measurements are combined with high resolution characterization of. 24. the mineral products and of the physicochemical properties of the exopolymeric. 25. substances. We propose a conceptual model discussing the relative influence of the. 26. alkalinity engine and EPS properties on the mineral composition and distribution during. 27. the initial step of stromatolite formation.. 28 29 30 31. This article is protected by copyright. All rights reserved.

(6) Field campaigns in the Cayo Coco lagoonal network (CCLN) were undertaken in January. 2. 2013 and July 2014. The CCLN is located on the south side of the island of Cayo Coco on. 3. the Atlantic coast of Cuba, in the Ciego de Avila province (Supplementary Figure 1a).. 4. CCLN encompasses a number of fully to partially connected shallow (< 1 m) lagoons. 5. linked with the Caribbean Sea through the Perros Bay. The lagoons fill a Pleistocene. 6. paleodune field resulting in an alveolar morphology on satellite images (Supplementary. 7. Figure 1a). This study focuses on the easternmost lagoon, which harbours a wide variety. 8. of lithifying microbial mats and fully lithified stromatolites ( Figure 1; see also Bouton et. 9. al., 2016;). This lagoon is > 1000 m long and 600 m wide, with a maximum water depth. 10. of ca. 75 cm when the water level is at the dynamic equilibrium. It is isolated from the. 11. rest of the network by a 50-m wide bioclastic sand flat, with an ephemeral channel that. 12. is several dm wide and which serves as a connection with the other lagoons (Figure 1a. 13. and Supplementary Figure 1a). The salinity in the lagoon varied seasonally from 54-. 14. 63‰ in July 2014 (wet season; Table 1) to 67-75‰ in January 2013 (dry season).. 15. Abundant microbial mats were observed; fish, insects, gastropods and arthropods were. 16. rather common, with a lower abundance and diversity during the dry season, when the. 17. salinity was higher (Bouton et al., 2016).. 18 19. Author Manuscript. 1. 20. Six surface water samples (see Figure 1 for location) were analysed in the field for pH,. 21. conductivity and alkalinity: pH was measured using a WTW pH 3110 with a Sentix® 41. 22. electrode or a Consort C561 pH-meter with a BioBlock Scientific electrode. Conductivity. 23. was determined with a WTW Cond 3110 and a TetraCon® 325 probe. The total. 24. alkalinity was assessed in the field using the Gran method (Gran, 1950). The alkalinity. 25. samples were filtered using 0.22 µm polyethersulfone syringe filter. Water samples. 26. were stored in glass vials (4 ml vials either under in situ pH conditions or acidified for. 27. analysis of major ions and 10 ml vials in order to determine the organic composition),. 28. kept refrigerated and transported to the laboratory. Major cation (NH 4 +, Na+, K+, Mg2+,. 29. Ca2+) and anion (Cl-, SO 4 2-, NO 3 - and PO 4 3-) concentrations were determined by ion. 30. chromatography (Dionex DX-100 or ICS-1500, with an analytical precision of 0.2 mg/l). 31. and Dissolved Organic Carbon (DOC) content using a Shimadzu TOC-5000A analyser. This article is protected by copyright. All rights reserved.

(7) The concentration of Si(OH) 4 was measured using the silicomolybdate colorimetric test. 2. (method 8185; Hach Lange) with a DR3900 Spectrophotometer (Hach Lange). The. 3. analytical precision was of 0.3 mg/l (Table 1). Salinity values were calculated from. 4. conductivity and temperature values, according to the Aminot and Kérouel method. 5. (Aminot and Kérouel, 2004). The water composition given in Table 1 represents an. 6. average composition of the six water samples.. 7 8. Author Manuscript. 1. 9. Microbiological preparations. Upon completion of field measurements and observations,. 10. lithifying mat samples were stored at 4 °C directly after collection, and then sectioned.. 11. One half of each sample was used for microscopy. The other half was transferred to the. 12. lab, submersed in ca. 5 cm filtered (0.22 µm) lagoon water and incubated in a. 13. greenhouse for two weeks mimicking natural environmental conditions (e.g., light. 14. intensity, T, pH). The mat samples investigated in this study were made of a. 15. superposition of three layers, layer A at the top, layer B and layer C at the bottom. 16. (Figures 1 and 2; for details, see Macroscopic Microbial Mat description section). Layer A. 17. can be subdivided in three laminae, from top to bottom: (i) a green lamina (ca. 2-5 mm. 18. thick); (ii) a white lamina (0.5-1 mm thick) and (iii) a red lamina (extending from 2 to 4. 19. mm). The green lamina, the red lamina and layer B were subsampled with a scalpel.. 20. Each mat fragment was mixed with a drop of deionized water on a glass slide and then. 21. spread under a coverslip. The preparations were then observed with a Nikon Eclipse Ci. 22. Pol microscope.. 23 24. Cryo Scanning Electron Microscopy (Cryo-SEM). SEM analyses were conducted using a. 25. Philips XL 30 field emission environmental scanning electron microscope (FEG-ESEM),. 26. equipped with an electron back-scattering pattern detector as well as an energy-. 27. dispersive X-ray spectrometer (EDS) for chemical analysis in Neuchâtel (Switzerland).. 28. Whole sample slabs were frozen (cryofixation) by immersion in liquid nitrogen at 210 °C. 29. followed by sublimation under vacuum within the SEM chamber using an Oxford high-. 30. resolution cryotransfer system. This cryofixation transforms the water to ice with a. 31. crystalline domain size between 10 and 100 nm, which does not interfere with or This article is protected by copyright. All rights reserved.

(8) modify the three dimensional organization of even highly hydrated samples, which. 2. allows thorough three dimensional observations of microbes, EPS and associated. 3. mineral phases. EDS analyses were used to determine the composition of the main. 4. mineral groups, (e.g., Ca-Mg carbonates, Mg silicates, sulfides). The lower limit of. 5. detection of elements was of 0.1 wt%.. 6. Confocal Laser Scanning Microscopy (CLSM). Sample preparation and analyses were. 7. carried out as described by Gérard (Gérard et al., 2013) without modifications. A thin. 8. section of the soft lithifying mat sample was observed under Confocal Laser Scanning. 9. Microscopy (CLSM), at the Institut de Physique du Globe de Paris (France), using a. 10. FluoView FV1000 CLSM with a spectral resolution of 2 nm and a spatial resolution of 0.2. 11. mm (Olympus, Tokyo, Japan). Fluorescence image stacks were obtained by using 405-. 12. nm laser diode, 488-nm multiline argon and 543-nm helium-neon-green lasers (at 5%,. 13. 10% and 30% of maximum power, respectively).. 14 15. Author Manuscript. 1. 16. X-Ray diffraction (XRD). Five samples were analysed: the green, white and red laminae. 17. of layer A, layer B and layer C. The organic matter was removed in all samples by. 18. reaction with H 2 O 2 (30%) during 24-48h at 40°C. X-ray diffractometry was successively. 19. performed on the bulk and less than 100 µm fraction. The clay fraction (less than 2 µm). 20. was analysed on oriented and unoriented powders. The diffractometer was a Siemens. 21. D500 Bragg-Brentano equipped with a scintillation detector, using 30 kV Cu-K radiation.. 22. Internal standards (Corindon: NIST SRM 1976b, Quartz: NIST SRM 1878b) were used for. 23. calibration. The errors were respectively of 0.02 Å on d-spacing and 0.01° on diffraction. 24. angles. All mineral phases were determined using Bruker AXS software Eva (Diffract+. 25. 14.0).. 26 27. Fourier Transform Infra-Red spectroscopy (FTIR). FTIR was performed on the same. 28. samples as XRD analyses. 1 mg of the powder was mixed with 150 mg of potassium. 29. bromide. The powders were pressed for 5 minutes under 8 tons, then for 1 minute. 30. under 10 tons. Powders were then analysed with a Fourier Nicolet spectrometer. 31. (detector: DTGS CSI, separating: CSI) in the 300–4000 cm-1 wavenumber range (spectral This article is protected by copyright. All rights reserved.

(9) 1. resolution of 0.125 cm-1; intervals of 4 cm-1), at the HYDRASA laboratory (Poitiers,. 2. France). Spectra were interpreted using the OMNIC software (Thermo Fisher Scientific).. 3. Author Manuscript. 4 5. Microelectrode depth profiles. Dissolved oxygen (O 2 ), sulfide (H 2 S/S2-) concentration. 6. and pH profiles with a vertical resolution of 250 µm (Figure 3) were measured in during. 7. peak photosynthesis (between 12:00-14:00) in the mat under natural light (1610-1780. 8. μE.m-2.s-1) and at the end of the dark period (between 4:00-5:30; 0-1 μE.m-2.s-1). The. 9. measurements for O 2 ,, H 2 S/S2 and pH were taken using polarographic and ion-specific. 10. needle microelectrodes in combination with a field microsensor multimeter (Unisense,. 11. Aarhus, Denmark; Pages et al., 2014; Visscher et al., 2000) and measurements for S2-. 12. using a high-impedance millivolt meter (Microscale Measurements, The Hague, The. 13. Netherlands). The [HS-] was calculated by combining H 2 S or S2- readings with pH. 14. measurements at each depth (Visscher et al., 1991). Detection limits for electrode. 15. measurements were better than 1 μM O , and better than 5 μM for H S, and 2 μM for the 2. 2. 16. S electrode. Calibration values determined before and after each deployment were. 17. within 96% for O electrodes, better than 86% for sulfide electrodes and within 0.3 units. 18. for pH microelectrodes. Depth profiles were determined three to five times for each. 19. analyte and representative profiles shown as explained in detail in Pages et al. 2014.. 20. Light measurements were taken with a LiCor LI-250 photometer equipped with a. 21. quantum probe SA-190A. The temperature and conductivity of the overlying water were. 22. determined with an Accument AP-75 handheld meter and the pH using a Mettler Toledo. 23. GFive meter.. 24. 2-. 2. 25. Oxygen production and consumption. The rates of oxygen production and consumption. 26. were determined using the light-dark shift (Pages et al., 2014; Visscher et al., 1998).. 27. Briefly, the O 2 concentration with depth was determined in the light. Following, the. 28. mats were incubated in the dark and O 2 profiles measured every 5-15 min for ca. 1.5-2 h. 29. until profiles resembled dark profiles measured at 04:00 h (i.e., depicting O 2 diffusion. 30. into the mat). Consecutively, mats were exposed to natural sunlight and O 2 profiles. 31. measured as before until the original light profile was approached. Oxygen consumption This article is protected by copyright. All rights reserved.

(10) and production were calculated from the decrease, and increase minus decrease in [O 2 ],. 2. respectively. Oxygenic photosynthesis produces O 2 , and aerobic respiration,. 3. chemolithotrophic sulfide oxidation and some chemical reactions contribute to oxygen. 4. consumption. Dark incubations of slurries amended with sulfide or thiosulfate (see. 5. below) revealed extremely low rates of aerobic sulfide oxidation. The vertical hiatus. 6. (mm’s) between the depth of zero O 2 and the depth of first appearance of sulfide further. 7. corroborated low chemolithotrophic sulfide oxidation rates.. 8. Author Manuscript. 1. 9. Anaerobic respiration – Sulfate reduction (SR). The sulfate reduction activity in intact. 10. mat samples was mapped using 35SO 4 2–coated silver foil (Pages et al., 2014; Visscher et. 11. al., 2000). Silver foil (0.025 mm thickness; Sigma-Aldrich, St. Louis, MO) was washed. 12. with acetone three times, rinsed with distilled water (>18 MΩ) and air dried overnight. 13. in a class 1000 clean bench. Subsequently, the foils were coated with S-SO (specific. 14. activity 1 mCi.ml ; Perkin-Elmer, Boston, MA) by submersion in a bath followed air. 15. drying. This was repeated four times in order to obtain uniform coverage of the isotope. 16. on the foil surface. The silver foils were prepared within a week of application and. 17. stored in the dark at 4 C. Silver foils were prepared in a batch of four. One foil was tested. 18. for homogenous isotope coverage using a BioRad Molecular Imager System GS-525. 19. (Hercules, California) radioactive gel scanner. The foils were deployed ex situ to. 20. estimate the sulfate reduction rate. During a 4-hr incubation of vertically cut mat. 21. sample, sulfate reduction produced sulfide, which precipitated as Ag 2 35S onto the foil. 22. surface. Upon completion of the incubation, the sample was photographed to document. 23. the exact location of the mat sample and the Ag-foil stored in the dark until analyzed.. 24. The two-dimensional distribution of the Ag 2 35S radioactivity was analyzed using a. 25. BioRad Molecular Imager System GS-525 (Hercules, California) radioactive gel scanner.. 26. This method accurately documents the metabolic activity of sulfate reducers as shown. 27. by comparison of Ag-foil maps and confocal scanning laser microscopy using SRB-. 28. specific dsrAB probes and reflects traditional sulfate reduction measurements in which. 29. 35SO 24. 30. reducing activity were imported from the output of the Molecular Imager System into. 31. Adobe Photoshop. The spatial resolution of pixels is ca. 35 μm and the colour is This article is protected by copyright. All rights reserved. 35. 4 2-. -1. o. is injected in sediment cores (Visscher et al., 2002, 2000). Pixel maps of sulfate-.

(11) indicative of the intensity of the reduction rate (i.e., red pixels represent high, orange. 2. intermediate and yellow pixels lower activity; Figure 3a5).. 3. Sulfate reduction rate (SRR) was estimated from previous incubations under identical. 4. conditions in hypersaline microbial mat systems in Puerto Rico (Casillas-Martinez et al.,. 5. 2005) and the Bahamas (Visscher et al., 2000). Briefly, mats were sliced vertically after. 6. which one half incubated on Ag foil, and the other half dissected in individual layers and. 7. micro-injected with. 8. reduction (Fossing and Jørgensen, 1989). Ag-foil maps showing three pixel intensities. 9. were obtained and compared with the rates calculated from 24 individual layers. With. 10. the average weight factors determined for each pixel intensity (15:5:1 for dark (red),. 11. medium (orange) and light (yellow) pixels, respectively), and pixel density (number of. 12. each per mm2), the sulfate reduction rate could be estimated with an error of 10-15%.. 13. Author Manuscript. 1. 35SO 2-, 4. incubated and processed using a single-step chromium. 14. Methanogenesis and sulfide oxidation. Homogenized samples were prepared from. 15. individual mat layers by mixing one volume of mat with an equal volume of Instant. 16. Ocean, prepared at 50 PSU using sterilized deionized water. The potential rate of. 17. methanogenesis was determined by incubating 3 ml of slurry in a 10-ml serum bottle,. 18. crimped sealed with a butyl rubber stopper. The [CH 4 ] in the headspace was determined. 19. for 18 days using a Shimadzu GC-14A gas chromatograph with flame ionization. 20. detection and a Porapak Q column (Buckley et al., 2008; Visscher et al., 1991). The. 21. experiment was carried out in triplicate. The potential rate of sulfide oxidation was. 22. determined in 10 ml slurries that were preincubated for 48 hrs, transferred into glass. 23. vessels (Visscher et al., 1998) and to which known amounts of either Na 2 S or Na 2 S 2 O 3. 24. were added at T 0 . The decrease of sulfide (using sulfide microelectrode) or thiosulfate. 25. (using HPLC in combination with monobromobimane derivatization; (Rethmeier et al.,. 26. 1997) was measured in live slurries in the light and dark and repeated in killed slurries.. 27. The difference between live consumption in the light and live consumption in the dark. 28. plus consumption in killed slurries represents an estimate for photosynthetic sulfide. 29. and/or thiosulfate consumption. Sulfide/thiosulfate consumption experiments were. 30. carried out with two replicates. In order to allow for comparison of the various. 31. metabolic processes, all rate estimates were expressed as μM organic carbon produced This article is protected by copyright. All rights reserved.

(12) 1. or consumed per h using the equations from Visscher and Stolz (Visscher and Stolz,. 2. 2005).. 3. Author Manuscript. 4 5. Chlorophyll a (Chla) and bacteriochlorophyll a (BChla) concentrations were determined. 6. spectrophotometrically according to Stal (Stal et al., 1984). Mat samples were frozen and. 7. individual layers: green lamina, red lamina and layer B (Figure 3) separated.. 8. Photopigments were extracted from the samples (in triplicate each individual layer ;. 9. Supplementary Table 1) in cold methanol in the dark for 3 h. The extinction in. 10. supernatant of the extracts was measured on a Varian Cary 50 Probe spectrophotometer. 11. at 665 nm and 770 nm after centrifugation (5 min at 2000 rpm) and Chla and BChla. 12. concentrations calculated.. 13 14 15. EPS extraction and purification. EPS were extracted from the whole layer A and from. 16. three subsamples: green lamina, red lamina and layer B, which were dissected with a. 17. scalpel under a stereomicroscope. The samples were first homogenized during at least 2. 18. hours in deionized water (ca. three times the sample volume) with addition of 1 mM. 19. EDTA. Samples were then centrifuged (2500 rpm, 10 minutes) and the supernatants. 20. were filtered through 0.5 µm pore size. The filtrate was precipitated in five times its. 21. volume of cold propanol (4°C). The precipitate was recovered by centrifugation (2500. 22. rpm, 20 minutes), placed in dialysis tubing (12 kDa). Samples were dialyzed four times:. 23. twice against 1 mM EDTA and twice against deionized water, each cycle lasting 24 h.. 24. Samples used for Alcian Blue, phenol-sulfuric acid and calcium binding assays were kept. 25. liquid at 4°C and processed in less than 24 hours. The remaining EPS were freeze-dried.. 26 27. EPS depth profiles. The quantity of EPS with depth in the mat was estimated using two. 28. different assays: (i) Dubois’ phenol-sulfuric acid assay (Dubois et al., 1956) and (ii) the. 29. Alcian Blue assay (Passow and Alldredge, 1995). Overall the use of the two assays give. 30. an estimate of (i) the amount of sugar monomers in the sample (i.e., the amount of EPS). 31. and (ii) the amount of cationic dye binding sites (i.e., the amount of acidic reactive sites This article is protected by copyright. All rights reserved.

(13) in the EPS). For a detailed description of the two methods, see Braissant et al. (Braissant. 2. et al., 2009). Three replicates (100 µl of EPS) were used in both assays, except for the. 3. red layer, for which we used two replicates because of lower EPS yields. For the phenol-. 4. sulfuric acid assay, two standards were used: xanthan and glucose, whereas only. 5. xanthan was used for the Alcian Blue assay. The amounts of protein were quantified in. 6. the green and red layers using the micro-bicinchoninic assay (Smith et al., 1985).. 7. Author Manuscript. 1. 8. Calcium and magnesium binding capacity. The calcium binding capacity was determined. 9. on the EPS of the whole layer A according to Braissant et al., 2009 under a N 2 /H 2. 10. atmosphere to avoid ion pairing of Ca2+ with carbonate resulting from atmospheric CO 2. 11. dissolution. Dialyzed EPS was transferred into 4-ml vessel containing a solution of 5mM. 12. KNO 3 and 20mM Tris which was adjusted to pH 9.0. The titration was carried out by. 13. stepwise addition of a 0.5M CaCl 2 solution in increments of either 10 or 20 μL. The. 14. concentration of free calcium ions was determined with a calcium ion-selective. 15. microelectrode (courtesy of Dr. Eric McLamore, Univ Florida) and a calomel reference. 16. microelectrode (Unisense, Aarhus, Denmark) coupled to a high-impedance millivolt. 17. meter (Microscale Measurement, the Netherlands). Aliquots from the EPS prepared for. 18. calcium binding were taken for assessing the magnesium binding capacity. The titration. 19. was carried out under the same conditions with the exception that 0.5M MgCl 2 was. 20. used. Upon addition of 50 μL of the magnesium standard and stirring for 10 min, 50 μL. 21. of the solution were removed and centrifuged for 1 min at 15,000rpm. The magnesium. 22. concentration in the supernatant was determined by ion chromatography (Dionex ICS. 23. 3000; analytical precision better than 0.1 mg/l). All reagents were prepared using. 24. autoclaved deionized water that was allowed to cool to room temperature under. 25. vacuum.. 26 27. ICP-OES analyses. Two to three fragments of freeze-dried EPS from each layer (green. 28. lamina, red lamina, whole layer A and layer B) were weighed using a ultra-high. 29. resolution scale Sartorius MC5 and then rehydrated in 5 mL vials. The elemental. 30. concentration of Ca, Mg, Fe, Si was measured in each sample by inductively coupled. 31. plasma emission spectroscopy (iCAP 6000; Thermo Scientific; Supplementary Table 2). This article is protected by copyright. All rights reserved.

(14) Four standards were used to calibrate the iCAP 6000: Ca (1g/l Scharlau 0.5M/l HNO 3 ),. 2. Fe (1 g/l Fischer Chemical 1.0M/l HNO 3 ), Mg (1g/l Scharlau 0.5M/l HNO 3 ), Si (1g/l CPA. 3. CHEM H 2 O). As EPS were extensively dialysed prior to ICP analyses, these. 4. concentrations are considered to represent elements that are strongly bound to EPS.. 5. Author Manuscript. 1. 6. R. 7. Radiocarbon ages (∆14C) were measured on four samples: layer A, layer B and two lamina. 8. of layer C (Figure 4) at the Atmosphere and Ocean Research Institute, the University of. 9. Tokyo (Japan) using Single Stage Accelerator Mass Spectrometry. Sample preparations. 10. and analytical conditions vary depending on the sample size and are described in detail. 11. in Yokoyama et al., (2010). Ages were calibrated with CalPal online (http://www.calpal-. 12. online.de/), so the values are given in years calBP.. 13 14 15 16 17. In the CCLN most of the microbial mats show a laminated (stromatolitic) mesofabric,. 18. typically composed of three types of layers (Bouton et al., 2016), from bottom to top. 19. (Figures 1, 2 and 4): (i) Layer C is composed of buried ancient microbial mats (between. 20. 511+/- 5 and 751 +/- 20 14C yrs BP; Figure 4d and 4e) and associated sediments, mostly. 21. lime mud and thin shelly beds (Figure 2a); (ii) Layer B is constituted of cohesive. 22. lithifying mats that are precipitating CaCO 3 and are being degraded (16 +/- 52. 23. BP ; Figures 4 and 5). Layer B is referred to as the leathery layer because of its colour. 24. and cohesiveness. Layers B and C can be partially mineralized but poorly lithified (Fig.. 25. 1b, 1d and 1e) or highly mineralized, lithified and forming a laminated crust (Fig. 1f and. 26. g); (iii) The soft Layer A displays a sequence of green and red laminae (Figure 1c) and. 27. has a thickness that can extend for more than 1 cm. Layer A was absent in January 2013,. 28. but well-developed in July 2014, and so developed in less than 18 months (Figure 4).. 29. The maximum spatial extension of Layer A coincided with the limit of July 2014 water This article is protected by copyright. All rights reserved. 14C. yrs.

(15) 1. level (Figure 1a). The object of this study was to investigate the initial step of. 2. mineralization, and so the focus was on the emerging part of the mat (layer A and the. 3. upper part of Layer B).. 5 6. Author Manuscript. 4. 7. Figure 3 shows the microbial activity of Layer A. During the peak of photosynthesis, the. 8. maximum of oxygen concentration (300 µM) was observed between 2 and 3 mm depth,. 9. near the base of the green lamina (Figure 3a) where the oxygen production reached 81. 10. nmol.cm-3.min-1. The O 2 consumption (i.e., predominantly aerobic respiration) peaked. 11. (40-43 nmol.cm-3.min-1) also at this depth, indicating that a large part of the fixed carbon. 12. was available for anaerobic respiration and/or burial (Figure 3a). The green lamina had. 13. a Chla content of 1.66 ± 0.12 mg/cm3 (Supplementary Table 1) and harbored a diverse. 14. microbial community with abundant filamentous (e.g., Schizothrix spp., Figure 2c)- and. 15. coccoid (e.g., Chroococcus sp., Figure 2d; Gloeocapsa sp., Figure 2f and Entophysalis sp.). 16. cyanobacteria, which were likely the main oxygen producers. Pennate diatoms were also. 17. observed in this layer (Supplementary Figure 2a and 2b).. 18. Below 3 mm, the [O 2 ] rapidly decreased (Figure 3a3), and oxygen consumption equalled. 19. the oxygen production, both of which declined progressively to zero in the upper part of. 20. red lamina (Figure 3b2). Aerobic sulfide oxidation was negligible and thus most of the. 21. oxygen was probably consumed by aerobic respiration, with an estimated depth-. 22. integrated 1085 µMC h-1 in the green layer. The depth-integrated oxygen production in. 23. this layer was 1587 µMC h-1. These rate estimates exceeded the approximated rates of. 24. sulfate reduction, anoxygenic photosynthesis and methanogenesis by one to three. 25. orders of magnitude (Figure 3b3).. 26. During the peak of photosynthesis, the pH reached a value of 10.3 at 4 mm depth, in. 27. between the zone of maximum oxygenic photosynthesis and the oxic-anoxic interface (at. 28. 5 mm). A white lamina, which was located between the green and red lamina, was. 29. observed exactly at the depth of the pH maximum (Figures 3a2 and 3a3). The. 30. anoxygenic purple sulfur bacterium (PSB), Thiocapsa sp. (Figure 2e) was abundant in. 31. this layer. Characteristic elemental sulfur granules were observed inside PSB cells This article is protected by copyright. All rights reserved.

(16) (Supplementary Figure 3). The red layer (BChla content 1.38 ± 0.40 mg/cm3;. 2. Supplementary Table 1) still contained coccoid and filamentous cyanobacteria (same. 3. taxa as in the green layer, see above; Chla content 0.30 ± 0.13 mg/cm3), some of which. 4. had lost their green pigments (Figure 5 and Supplementary Figure 3).. 5. Sulfate reduction (SR) was found throughout the mat (Figure 3a5), but appeared. 6. concentrated at 4-5 mm, i.e. at the oxic-anoxic interface, just below the white lamina. A. 7. second sulfate reduction-rich depth horizon was present at 9-10 mm, inside layer B. 8. (Figure 3a5). The estimated sulfate reduction rate was higher in the red layer (ca. 317. 9. µM C.h-1) than in the green lamina (ca. 140 µM C.h-1) and in layer B (ca. 123 µM C.h-1;. 10. Figure 3b3). During the daytime, free sulfide always first appeared well below the oxic-. 11. anoxic interface (Figure 3a3) at the base of the red layer, the concentration increasing. 12. with depth to a maximum value of 307 µM at 12 mm in Layer B.. 13. At night, oxygen production ceased (Figure 3a4), but some O 2 diffused from the. 14. overlying lagoon water to ca. 2 mm, probably consumed by aerobic respiration and. 15. chemical and biological sulfide oxidation. The sulfide concentration increased with. 16. depth and reached 432 µM at ca. 10 mm (Figure 3a3). The diel distribution pattern of. 17. sulfide and potential rate estimates from slurry experiments of 1125 µM C.h-1 fixed. 18. indicates very active anoxygenic photosynthesis in the well-developed red layer (Figure. 19. 3b3). The activity of PSB during the day could be fueled by sulfide produced by SRB,. 20. which activity peaked at the top and bottom of the red layer. At night, the absence of. 21. anoxygenic photosynthetic activity explains the increase of sulfide near the surface. The. 22. estimated methanogenesis was low throughout the mat (<2 µM C.h-1) and had a limited. 23. role in C cycling (and as an alkalinity engine) in the CCLN mats.. 24 25. Author Manuscript. 1. 26. The amount of EPS measured with the phenol-sulfuric acid assay (when expressed. 27. either in xanthan or glucose equivalents) showed large variations among the three. 28. depth horizons that were studied. In the green lamina (0-4 mm deep), the hexose value. 29. was ca. 10 µg EPS.g-1 wet weight (WW) and 36 µg EPS.g-1 WW sample respectively when. 30. using xanthan and glucose as standards. In the red layer (4-6 mm), the amount of EPS. 31. decreased to 0.7 µg EPS.g-1 WW and 11.6 µg EPS.g-1 WW expressed as xanthan and This article is protected by copyright. All rights reserved.

(17) glucose equivalents, respectively. When compared to the red layer, the amount of EPS in. 2. Layer B (6-11 mm) increased sixty fold when using xanthan as standard and around. 3. tenfold when using glucose (Figure 3b4). Similar trends were resulted from the Alcian. 4. Blue assay, which is a measure for sugar acidity (e.g., sulfate esters and carboxyl. 5. functional groups; Figure 3b4): in the green lamina, the amount of EPS is 91 µg EPS.g-1. 6. WW, and the value decreased until 15 µg EPS.g-1 WW in the red layer. The protein. 7. concentration was six-fold higher in the red layer (172 µg EPS.g-1 dry weight; DW) than. 8. in the green layer (28 µg EPS.g-1 DW). The maximum binding capacity of the whole layer. 9. A EPS was ca. 386 mg.g-1 EPS dry weight (DW) for calcium and ca. 50 mg.g-1 EPS DW for. 10. magnesium. In comparison, the concentrations of tightly bound calcium and magnesium. 11. were very low in layer A, respectively ca. 8 mg.g-1 EPS DW and ca. 2.3 mg.g-1 EPS DW. 12. (Supplementary Table 2).. 13 14. Author Manuscript. 1. 15. Observations using polarizing microscopy, cryo-SEM combined with EDS and confocal. 16. laser scanning microscopy coupled with spectroscopic analysis (FTIR) and X Ray. 17. Diffraction showed two mineral phases in layer A, a high-magnesium calcite (HMC) and. 18. a Mg-silicate (Figures 2, 6 and 7, Supplementary Figure 4). In the green lamina (0-3. 19. mm), photosynthetically active coccoid and filamentous cyanobacteria were embedded. 20. in the EPS (Figure 2f), in which small 1-5 µm rounded particles sometimes organized in. 21. 10-50 µm globular aggregates could be observed. EDS analyses indicated a strong. 22. enrichment in Mg and Si (Figure 6a, b and h), which was especially abundant around. 23. coccoids, but also locally impregnated the sheaths of filamentous cyanobacteria (Figure. 24. 6b). The FTIR spectrum of the mineral fraction in the green layer showed a broad band. 25. at 3400 cm-1 indicative of strong hydration and a low intensity band at 656 cm-1,. 26. corresponding to the Mg 3 -OH bending vibration of trioctahedral Mg-silicates. 27. (Supplementary Figure 4). No diffraction peaks were found for this Mg-Si phase. 28. indicating poor crystallization at this depth. HMC was detected by XRD (Figure 7) and. 29. could have developed locally in this layer, although this was not observed by SEM. In the. 30. green lamina, pennate diatoms frustules were either pristine (comprising opal) or. 31. transformed into a Mg-silicate phase (Supplementary Figure 2a to 2c). This article is protected by copyright. All rights reserved.

(18) Peloids, 10-20 µm in size (Figure 6c and Supplementary Figures 2d and e), were. 2. abundant within the alveolar EPS matrix of the white lamina (4 mm deep). These peloids. 3. were composed of <1 to 5 µm HMC crystals (Figure 6d).. 4. In the red lamina (4-7 mm deep), both HMC and Mg-silicates were observed. As in the. 5. green lamina, the Mg-Si phase formed globular aggregates (Supplementary Figure 2g),. 6. but also impregnated some of the cyanobacterial cell walls or filled in some of the. 7. coccoid cells (Figure 6f). The top of Layer B is richer in carbonates than the red lamina. 8. (Figure 2a). It shows abundant peloids of HMC locally fused to form patches larger than. 9. 100 µm (Figure 6g) spreading between and filling coccoid cells (Supplementary Figure. Author Manuscript. 1. 10. 2h).. 11. Similar green and red lamina successions were observed in older and deeper layers B. 12. and C, both in moderately (Figures. 4 and 5a to 5e) and highly lithified mats (Figures 4,. 13. 5f and 5g). The interface between the two laminae coincided with a peloidal and micritic. 14. lamina (Figure 5e). Overall, mats were increasingly enriched in HMC with depth, as. 15. shown in the XRD spectra (Figure 7). However, red laminae remained weakly. 16. mineralized (Figure 5d) while green laminae contained more HMC micropeloids that. 17. formed between vertically oriented cyanobacteria (Figure 5d and 5e).. 18 19 20 21 22. We propose a five-step conceptual model explaining the relationships between. 23. microbial activity, EPS properties and the nucleation and precipitation of minerals in the. 24. developing (<18 months) lithifying mats of the CCLN (Figure 8). The main. 25. geomicrobiological reactions involved in the mineralization process are depicted in. 26. Figure 9.. 27 28. (i) Cyanobacteria and possibly diatoms colonize the surface of Layer B to form a biofilm. 29. (Figure 8i). The pH of the water in CCLN is 8.8-9.0, rendering bicarbonate the main. 30. carbonate species. Cosmopolitan cyanobacteria (e.g., Microcoleus spp., Schizothrix spp.). 31. possess extracellular carbonic anhydrase that dissociate HCO 3 - into OH- and CO 2 This article is protected by copyright. All rights reserved.

(19) (Kupriyanova et al., 2007; Merz, 1992). This CO 2 is photosynthetically fixed at a. 2. maximum rate of 1.5 mM C.h-1 producing biomass, LMWOC excretion products and EPS.. 3. The OH- contributes to increase the pH and alkalinity (Figure 9, metabolic reaction 1);. 4. see Figure 8ii). Owing to a large quantity of acidic functional groups in the EPS (Figure. 5. 3b4), this biofilm possesses a high calcium binding capacity and therefore a high. 6. potential to inhibit calcium carbonate precipitation.. 7. Author Manuscript. 1. 8. (ii) The high rate of oxygenic photosynthesis (see Figure 8i) supports proliferation of. 9. aerobic heterotrophic bacteria. These organisms either respire organic carbon with. 10. oxygen or partially degrade large organic molecules into LMWOC during fermentation at. 11. night in the absence of O 2 . Both types of metabolism decrease the alkalinity by. 12. generating CO 2 and/or LMW-organic acids (Figure 9, metabolic reaction 2). The high. 13. rate of aerobic respiration (1.1 mM C.h-1), also evidenced by the steep O 2 gradient,. 14. results in the formation of an anoxic zone below the active cyanobacterial layer (Figure. 15. 8ii). Although the Mg:Ca ratio is ca. 2.9 in the overlying water (Bouton et al., 2016), EPS. 16. from layer A can bind ca. 8 times more Ca than Mg, implying that Mg is likely available. 17. for precipitation. Furthermore, the pH of 10.3 (Figure 3a3) is well above the threshold. 18. value of 8.6-8.7 for nucleation of Mg-silicates (Tosca and Masterson, 2014; Zeyen et al.,. 19. 2015). This pH maximum combined with an elevated Mg2+ activity could lead to the. 20. nucleation of Mg silicates on both EPS and filamentous cyanobacterial sheaths, as. 21. observed in other hypersaline lakes (Burne et al., 2014; Pace et al., 2016). In CCLN, the. 22. Mg-Si phase seems to nucleate preferentially, so potentially first, on the EPS. 23. surrounding coccoid cyanobacteria, or on the actual coccoid cell walls and/or. 24. undetermined globular aggregates (possibly SRB). In hypersaline environments, the. 25. dissolution and/or recrystallization of diatoms frustules is enhanced at elevated pH. 26. (Badaut and Risacher, 1983; Barker et al., 1994; Ryves, 1994). EPS can protect diatoms. 27. against alteration and dissolution (Pike and Kemp, 1999). The opaline-Si to Mg-Si. 28. change observed in the green lamina of CCLN mats (Supplementary Figure 2a to 2c). 29. could occur in the oxygenic photosynthetic zone, where pH is maximum, and be. 30. facilitated by degradation of EPS surrounding diatoms by aerobic heterotrophs.. 31 This article is protected by copyright. All rights reserved.

(20) (iii) Anoxic conditions prevailing at depth support the development of an anaerobic. 2. heterotrophic bloom. Sulfate reduction peaks in the red lamina consuming ca. 0.32 mM. 3. C.h-1 (Figure 3a5), but is also active in the oxic part of the mat (0.14 mM C.h-1; Figure 2b).. 4. SRB produce HS- that accumulates in the red layer (Figure 3a3 and 3a4 and Figure 9,. 5. metabolic reaction 3). SRB can be part of consortia forming small degradation pockets. 6. (Decho, 2010), possibly in anoxic microzones (Villbrandt and Bebout, 1994) in the EPS. 7. matrix, releasing Ca2+ and enhancing carbonate precipitation (Dupraz et al., 2009).. 8. SRB metabolism uses LMWOC and H 2 either decreasing (when using e.g., lactate) or. 9. increasing (when using e.g., H 2 , formate) the pH (Gallagher et al., 2012). The. 10. degradation of certain LMWOC by SRB is probably one of the main contributions to the. 11. alkalinity engine leading to the nucleation of HMC in the CCLN (Figure 8iii).. 12. Furthermore, the removal of sulfate potentially reduces the kinetic inhibition for Mg. 13. carbonates precipitation (Van Lith et al., 2003a, 2003b). HMC peloids identical to those. 14. observed in the CCLN (Figure 6c and 6d) have been linked to the site of maximum SRB. 15. activity in several Bahamian hypersaline lakes (Dupraz et al., 2004; Glunk et al., 2011).. 16. Author Manuscript. 1. 17. (iv) and (v) The HS- produced by SRB act as electron donor for PSB, which thrive at the. 18. oxic-anoxic interface in the red lamina where light and reduced sulfur compounds are. 19. available (Figure 3a3 and 3a4; Figure 8iv). Although PSB are primarily anoxygenic. 20. photoautotrophs, some (e.g., Thiocapsa) can also grow as chemoautotrophs, which. 21. enables survival at this depth horizon with fluctuating concentrations of O 2 and sulfide.. 22. In addition to cyanobacterial photosynthesis in the overlying green lamina, high rates of. 23. anoxygenic photosynthesis (1.12 mM C.h-1) help increase the alkalinity and thus the. 24. carbonate precipitation potential (Figure 9, metabolism 4) especially when sulfide is. 25. fully oxidized (Dupraz and Visscher, 2005; Visscher and Stolz, 2005). The comparable. 26. rates of oxygenic (near the bottom of the green lamina) and anoxygenic (in the upper. 27. part of the red lamina) photosynthesis, possibly supported by sulfate reduction. 28. (depending on the electron donor; Gallagher et al., 2012) result in the observed pH. 29. maximum at the oxic-anoxic interface (Figure 8 iv). The EPS from the CCLN red lamina. 30. show a decrease in the abundance of sugars and acidic sites and a marked increase in. 31. protein fraction compared to the EPS from the green lamina (Figure 3b4). The change in This article is protected by copyright. All rights reserved.

(21) EPS properties from green to red lamina could imply that cation binding sites decrease,. 2. rendering free Ca available at the interface between these two layers. In addition,. 3. degradation of EPS into smaller molecules by (an)aerobic heterotrophs (Braissant et al.,. 4. 2009) could also explain this change in EPS properties. Furthermore, it has been shown. 5. that xanthan can stimulate SRB activity (Battersby et al., 1984). The elevated SR activity. 6. at the interface of the green and red laminae (Figure 3) supports this notion. EPS. 7. degradation by SRB could liberate Ca2+ ions and increase the alkalinity (see also step iii).. 8. However, an elevated sulfate concentration has a counter effect by inhibiting calcite. 9. precipitation (Busenberg and Niel Plummer, 1985). The C-normalized rates indicate that. 10. PSB are ca. 3.5 times more active than SRB, which is corroborated by depletion of free. 11. sulfide in the upper part of the red lamina. This indicates that PSB outpace SRB and that. 12. sulfate may accumulate, possibly counteracting the microbial increase in alkalinity. The. 13. result is a decrease in the potential for calcite precipitation, which could explain the. 14. relatively low amount of carbonate precipitating in the red lamina (Figure 6e to 6g).. 15. Consequently, the site of HMC peloids precipitation would be concentrated at the. 16. interface between the green and red laminae, i.e. at the oxic-anoxic interface, coinciding. 17. with the pH maximum (Figures 4, 5, 6 and 8 and Supplementary Figure 2).. 18. Author Manuscript. 1. 19. Previous models of stromatolite formation propose either (i) alternating periods of. 20. trapping and binding, mineral precipitation and microboring of grains (e.g., Reid et al.,. 21. 2000) or (ii) changes in orientation of filamentous cyanobacteria combined with EPS. 22. degradation by heterotrophic bacteria (Dupraz et al., 2013). We show here that during. 23. the early development (i.e., the initial 18 months) of a hypersaline microbial mat, the. 24. production a carbonate lamina can also occur at an interface of major change in. 25. metabolisms and subsequent EPS properties. In this particular case at the interface of. 26. oxygenic and anoxygenic photosynthesis.. 27 28 29. EPS are a critical ecophysiological component of the biofilm community, providing an. 30. environment that sustains optimal growth and interactions (Decho, 2000; Dupraz et al.,. 31. 2009; Flemming and Wingender, 2010; Marvasi et al., 2010). The negatively charged This article is protected by copyright. All rights reserved.

(22) acidic groups in EPS bind Ca and Mg that link individual polymers through bidentate. 2. bridges, increasing the mechanical properties. Such acidic functional groups can be. 3. present in mono- and polysaccharides, as well as in certain amino-acids (e.g., aspartic. 4. and glutamic acids), comprising proteins (Gautret et al., 2004; Kawaguchi and Decho,. 5. 2002). As a secondary effect, a high cation binding capacity inhibits carbonate. 6. precipitation, as was observed in the green layer of CCLN mats. The degree of cation. 7. binding depends on the specific EPS properties, which change with depth in microbial. 8. mats (this study, Braissant et al., 2009; Sforna et al., 2016). CCLN mats potentially. 9. sequester two to six times more Ca than previously reported. Furthermore, the capacity. 10. for Ca binding in CCLN EPS exceeds that for Mg adsorption by a factor of ca. 8. A release. 11. of Ca either by EPS degradation or saturation of binding sites, leads to carbonate. 12. precipitation (Decho, 2010; Dupraz and Visscher, 2005; Dupraz et al., 2011; Gautret et. 13. al., 2004; Kawaguchi and Decho, 2002). In CCLN, a vast change in EPS properties at the. 14. interface of green (oxygenic) and red (anoxygenic photosynthetic) layers result in a. 15. locus/zone for geochemical reactions. This has a two-fold/dual effect on the mineralogy:. 16. Mg remains available for incorporation in Mg-Si phases and Ca is transported through. 17. the green layer to the interface with the red layer. A similar Ca transport mechanism. 18. was also proposed for Kiribati microbialites by Ionescu et al. (2015): in their conceptual. 19. model, Ca is bound to EPS in cyanobacteria-dominated layers at the surface of the mat. 20. and released through aerobic heterotrophic degradation of EPS deeper in the mat,. 21. where O 2 is still available The released Ca would then precipitate at the oxic-anoxic. 22. interface as aragonite and gypsum in this case.. 23 24 25 26. Author Manuscript. 1. 27. Mg-Si phase. 28. The selective sequestration of Ca vs. Mg by EPS can facilitate Mg-silicate precipitation in. 29. the green layer of the mat. Dissolved silica from the overlying water and the dissolution. 30. of diatom frustules (pH>9) within the mat likely provide silica to CCLN microbial mats,. 31. similar to what happens in hypersaline Solar Lake (Jorgensen et al., 1983) and Great Salt This article is protected by copyright. All rights reserved.

(23) Lake (Pace et al., 2016). Dissolution of Mg-calcite was also suggested as a source for Mg-. 2. silicates in microbialites of alkaline Lake Satonda (Arp et al., 2003). The overall. 3. community metabolism in CCLN mats results in a pH of ca. 11 at the interface of. 4. oxygenic- anoxygenic photosynthesis. This pH is well above 8.6-8.7 and, if combined. 5. with potentiallyenhanced silica and Mg activities, favorare favourable to Mg-silicate. 6. precipitation (Tosca and Masterson, 2014; Zeyen et al., 2015).. 7. Authigenic Mg silicates are now recognized as major mineral products of hypersaline. 8. and alkaline lake microbialites. However, these minerals show variable morphology and. 9. preferential precipitation loci. In the green lamina of the CCLN mats, the Mg-Si phase. 10. either (i) form globular aggregates in the organic matrix surrounding coccoid. 11. cyanobacteria; (ii) impregnate the coccoid cell walls (Figure 6a, 6b) or (iii) replace. 12. diatom frustules (Supplementary Figure 2b and 2c). In Lake Clifton thrombolites, Burne. 13. et al. (2014) documented poorly crystallized massive to flaky stevensite on both alveolar. 14. EPS and Scytonema sp. sheaths. In another study, kerolite was observed covering EPS,. 15. replacing diatom frustules, impregnating filamentous cyanobacteria cell wall and. 16. sometimes infilling the cells in alkaline crater lakes in Mexican crater lakes microbialites. 17. (Zeyen et al., 2015). In Mono Lake microbialites, both cyanobacteria and the EPS were. 18. coated by Mg-Si phases (Souza-Egipsy et al., 2005). In contrast, the Mg-Si phase. 19. nucleated preferentially on alveolar EPS in Great Salt Lake thrombolites (Pace et al.,. 20. 2016). Due to variable surface charges, bacterial cell walls have been suggested as sites. 21. for the precipitation of authigenic clay precursors (Konhauser and Urrutia, 1999). This. 22. variability in preferential nucleation sites points to important differences in both cell. 23. material and EPS.. 24. It is plausible that Mg silicates could be mineral evidence for metabolisms increasing the. 25. pH (e.g., photosynthesis or sulfate reduction) in fossil microbialites (Arp et al., 2003;. 26. Pace et al., 2016; Tosca et al., 2011). Given their shape and size (<1 to ca. 50 µm), the. 27. CCLN Mg-Si aggregates could develop on the surface of bacteria, on EPS or both. The. 28. patchy Mg-Si distribution indicates possible nucleation around bacteria, particularly. 29. clusters of SRB, which are able to create a rise in alkalinity in the surrounding EPS. 30. (Figure 6b; Decho, 2010). Mg-Si phase were documented in the organic matrix. 31. surrounding SRB in cultures from the hypersaline coastal lagoon Lagoa Vermelha This article is protected by copyright. All rights reserved. Author Manuscript. 1.

(24) (Warthmann et al., 2000).. 2. EPS play a role in nontronite (Fe-rich smectite containing Mg-Si) formation in deep-sea. 3. sediments, presumably through incorporation of Fe and Si in the organic matrix.. 4. Experiments by Ueshima and Tazaki with artificial EPS (Ueshima and Tazaki, 2001; i.e., a. 5. mixture of acidic (pectin) and neutral (dextrin) polymers) confirmed Fe-Mg silicates. 6. precipitation, with similar mineral composition and shape as in the deep-sea samples.. 7. Similarly, Bontognali et al. (Bontognali et al., 2014) demonstrated that the presence of. 8. certain organic acids (e.g., succinic acid) enhanced the precipitation of smectite, but. 9. others (e.g., oxalic and citric acid) had no influence or inhibited nucleation. Succinic acid. 10. is a common product of the degradation of organic macromolecules by aerobic. 11. microorganisms and thus can be found in EPS of natural mats (Krebs, 1970). An. 12. enrichment process of Mg and Si scavenging, followed by site-specific degradation could. 13. be invoked in the CCLN surface layer, in which the EPS contained small amounts of. 14. tightly bound Mg and Si (Supplementary Table 2).. 15. Author Manuscript. 1. 16. Ca-Mg carbonates at the interface between green and red laminae. 17. Although the Ca and Mg binding capacity could not be determined in individual layers,. 18. the acidic sites decreased fivefold in the red lamina, reflecting a significant loss of cation. 19. binding capacity at the interface between green and red laminae. The estimated. 20. cyanobacterial photosynthesis rate in the green lamina was ca. 1.5 times higher than. 21. that of aerobic heterotrophic respiration, indicating depletion of CO 2 . Some HMC. 22. precipitated in the green layer, suggesting that saturation of Ca binding was locally. 23. reached (Dupraz et al., 2009). Combined with a high sulfate-reduction rate (SRR) at the. 24. top of the red lamina (Figure 3a), the formation of a well-developed carbonate layer,. 25. locally exceeding 1 mm, at the pH maximum could be explained. A high [Mg] in the water. 26. column, a relatively low binding capacity of Mg and dissolution of the Mg-Si phase as. 27. discussed above could contribute to Mg incorporation in the HMC mineral.. 28 29. Our results and those of others (Burne et al., 2014; Zeyen et al., 2015) suggest that Ca-. 30. Mg carbonates and Mg silicates have distinct nucleation loci, possibly resulting from. 31. spatial and temporal heterogeneities in EPS that in part could be caused by This article is protected by copyright. All rights reserved.

(25) heterotrophic degradation. This is of critical importance when interpreting the. 2. biogeochemical conditions during formation of fossil microbialites. The minerals. 3. comprising these fossils are typically the transformation product of primary (microbial). 4. phases, indicating the need to study diagenetic pathways in extant microbialites (Pace et. 5. al., 2016; Sforna et al., 2016).. 6 7. Author Manuscript. 1. 8. Anoxygenic photosynthesis by PSB contributes to carbonate precipitation (Ionescu et al.,. 9. 2015; Kremer et al., 2008; Visscher and Stolz, 2005; Visscher et al., 1998). In laboratory. 10. manipulations of mats from Niva Bay, the green cyanobacterial layer produced mostly. 11. <1 µm HMC crystals and the red PSB layer formed >1 µm HMC crystals and aragonite,. 12. presumably due to differences in EPS composition between the two layers (Kaźmierczak. 13. et al., 2015).. 14. The well-developed red layer in CCLN comprising PSB (e.g., Thiocapsa sp.) supported. 15. high potential rates of anoxygenic photosynthesis (APS; Figures 3a and 8), resulting in a. 16. pH maximum at the interface of green and red laminae. In contrast, lithifying mats with. 17. a much lower APS rates than CCLN develop in e.g., Big Pond, a seawater-fed hypersaline. 18. lake (Glunk et al., 2011) and in the continental hypersaline Great Salt Lake (Pace et al.,. 19. 2016). In these systems, the pH peak which is largely superposed to the O 2 maximum,. 20. points to a dominant cyanobacterial contribution in the alkalinity engine.. 21. In order to assess the potential contribution of PSB to the alkalinity engine, an. 22. evaluation of their metabolic versatility is pertinent. PSB and SRB metabolisms, both of. 23. which are coupled to sulfur cycling, peak in the red layer of CCLN mats. This suggest a. 24. close interaction between both groups, as has been documented for other microbial. 25. mats (Caumette, 1993; van Gemerden, 1993). PSB and SRB both have a positive effect on. 26. the alkalinity in the mat (Visscher and Stolz, 2005; Figure 9), increasing the potential for. 27. carbonate precipitation. In concert, these two groups interact as follows: (1) HS-. 28. produced by SRB act as electron donor for PSB (e.g., Thiocapsa sp.); (2) photoautrophic. 29. growth of Thiocapsa removes HS-, high concentrations of which are toxic to SRB (Decho. 30. et al., 2010; Reis et al., 1992) and cyanobacteria; (3) EPS and LMWOC produced by PSB. 31. stimulate the growth and metabolism of SRB (Gallagher et al., 2012). This article is protected by copyright. All rights reserved.

(26) In microbial mats, the depth profiles of O 2 and sulfide fluctuate throughout the diel cycle. 2. (Dupraz et al., 2009). During the middle of the day, oxygen penetrates into the upper. 3. part of the red lamina in CCLN but Thiocapsa is capable of photosynthesis under oxic. 4. conditions. Prolonged (> 12 h; Visscher et al., 1992) oxic conditions prevent BChla. 5. synthesis, forcing Thiocapsa to shift to chemolithoautotrophic sulfide oxidation,. 6. resulting in a decrease of the alkalinity . Some SRB are insensitive to O 2 exposure but the. 7. effect of their metabolism on the alkalinity engine depends on the type of electron. 8. donor. Consequently, a careful assessment on short temporal and small spatial scales of. 9. sulfur cycling processes is needed to unravel the relative contribution of PSB and SRB. 10. metabolisms to the formation of the Ca-Mg carbonate lamina (Visscher and Stolz, 2005).. 11 12. Author Manuscript. 1. 13. The characteristic lamination of microbial mats is often confused to represent the. 14. lamination in stromatolites. Whereas the lamination results from combined processes. 15. that precipitate a mineral layer in repetitive fashion often by a specific microbial mat. 16. community (Dupraz et al., 2013; Reid et al., 2000), the lamination in mats is the effect of. 17. the light conditions (van Gemerden, 1993). Surface layers may comprise EPS rich in. 18. photopigments that quench light. A green layer of cyanobacteria at the subsurface uses. 19. short wavelength (the blue-green part of the spectrum), and underneath the green layer,. 20. purple sulfur bacteria form a red lamina where they harvest photons with a longer. 21. wavelength that penetrate deeper into the sediments. The deepest anoxic layers of mats. 22. typically contain sulfides and are dark colored (grey to black) from iron sulfide. 23. precipitation (Des Marais, 2003; Sforna et al., 2016).. 24. The peak of oxygenic photosynthesis in CCLN is near the bottom of the green layer. 25. (Figure 3b), and the maximum of anoxygenic photosynthesis presumably near the top of. 26. the red layer, where the light intensity is most favourable to support this (Figure 6 –. 27. darker red appearance near the top of the red layer indicates high BChla. 28. concentrations). Potential metabolic rate estimates at the interface between green and. 29. red laminae during peak photosynthesis suggest that cyanobacteria and PSB are major. 30. contributors and SRB minor players in increasing the alkalinity. Aerobic heterotrophs. 31. are present in low abundance and aerobic sulfide oxidizers are absent, minimizing This article is protected by copyright. All rights reserved.

(27) microbial processes that lower the alkalinity at this interface. The overall effect of. 2. microbial metabolisms on the alkalinity and the changing properties of the. 3. cyanobacterial and PSB EPS at the interface of the oxygenic and anoxygenic. 4. phototrophic zones creates ideal conditions for Ca-Mg carbonate precipitation.. 5. A challenging question is whether the observed mineralization process repeats itself on. 6. longer timescales, and how the mineral products evolve during diagenesis. The 14C ages. 7. show a discontinuous preservation of CCLN mats, probably due to storm-related erosion. 8. and/or prolonged subaerial exposure (Bouton et al., 2016). In deeper, hence older layers. 9. of CCLN mats (layers B and C), horizontal to slightly undulated carbonate laminae. 10. alternate with porous patchy carbonate laminae rich in organic matter, resulting in a. 11. laminated (stromatolitic) mesofabric (Figures 1d to 1g, 4d and 4e). Microscopically,. 12. these laminae show a repetitive succession similar to the young mat: (i) a green lamina. 13. dominated by filamentous cyanobacteria overlying (ii) a weekly mineralized lamina. 14. dominated by PSB coccoid clusters (Figure 5). The interface between the two layers. 15. often coincides with a dense micropeloidal and micritic lamina. The main difference. 16. with layer A is that the green lamina is highly mineralized and richer in HMC peloids.. 17. This observation is confirmed by XRD, layers B and C being characterized by increasing. 18. carbonate (Mg-Calcite) content (Figure 7). This increasing lithification could result from. 19. progressive replacement of the EPS by carbonates through prolonged degradation of. 20. EPS by heterotrophic bacteria, e.g., sulphate-reducing bacteria. This phenomenon is. 21. relatively common in hypersaline microbialites (Dupraz et al., 2004, 2013; Pace et al.,. 22. 2016).. 23. Similar successions are observed in partially lithified mats (Figures 1d, 1e, 5a to 5e) to. 24. fully laminated crusts (Figures 1f, 1g, 5f and 5g), which can be classified as. 25. stromatolites.. 26. These. 27. Mineralization and biogeochemical models) has probably repeated over time and that. 28. the preserved carbonate-rich laminae could record the successive positions of previous. 29. oxygenic-anoxygenic photosynthetic interfaces as the mat grew upward (Figures 4d, 4e. 30. and 5a). A similar superposition of carbonate lamina, which could have precipitated at. 31. the oxygenic-anoxygenic (green-red) interface, was observed in the hypersaline mats of This article is protected by copyright. All rights reserved. Author Manuscript. 1. results. indicate. that. the. documented. precipitation. mechanism. (see.

(28) 1. Lagoa Vermelha, Brasil (Vasconcelos et al., 2006). This suggests that the vertical. 2. organization of oxygenic and anoxygenic photosynthesis could have contributed to the. 3. formation of lamination in numerous fossil stromatolites.. 5 6 7. Author Manuscript. 4. 8. This study documents the processes and products of mineralization during the early. 9. development of a hypersaline microbial mat. Oxygenic photosynthesis predominates. 10. over aerobic respiration in the surface green lamina of the mat, where Mg-silicate. 11. globules precipitate in the EPS matrix, often in close association with coccoid. 12. cyanobacterial clusters. In this zone, the cyanobacterial EPS bind calcium, thus inhibiting. 13. carbonate precipitation. Anoxygenic photosynthesis performed by PSB is prevalent in. 14. the red lamina of the mat. The interface between these two major photosynthetic zones,. 15. coinciding with the daytime pH maximum, forms the locus for the precipitation of a Mg-. 16. carbonate lamina. A major decrease of cation binding sites in EPS between the green and. 17. red laminae, makes Ca available at their interface. Fuelled by sulfate produced by PSB,. 18. SR activity peaks just below this interface. SRB have a dual effect leading to carbonate. 19. precipitation. First, by degrading the LMWOC, SRB can locally increase the alkalinity.. 20. Second, by consuming sulfates, inhibition of carbonate precipitation is removed. Our. 21. study shows that the formation of the initial carbonate lamina depends on (i) the. 22. creation of a high-pH interface between oxygenic and anoxygenic photoautotrophs and. 23. (ii) the synergy between two major players of the sulfur cycle, PSB and SRB. The. 24. repetition of this process over time combined with upward growth of the mat leads to. 25. the formation of stromatolites in the studied lagoon. These results could have major. 26. implications for the interpretation of fossil stromatolites.. 27 28 29 30. This study was supported by a grant from the Ministère de l’Enseignement Supérieur et. 31. de la Recherche attributed by the Ecole Doctorale 480 Montaigne Humanités and This article is protected by copyright. All rights reserved.

(29) complementary funding from Total S.A. The authors are grateful to Emmanuelle Gérard. 2. (IPGP, France) for her help with the Confocal Laser Scanning Microscopy during this. 3. project and the previous ones, Dylan Chaillot (Hydrasa, Poitiers, France) for the. 4. complementary FTIR analyses, Olivier Mathieu and Philippe Amiotte-Suchet. 5. (Biogéosciences Dijon, Université de Bourgogne) for their help with water chemistry. 6. analyses. Pieter Visscher acknowledges support by NSF EAR award no. 1561173.. 7 8 9. Author Manuscript. 1. 10. Allwood, A.C., Walter, M.R., Kamber, B.S., Marshall, C.P., and Burch, I.W. (2006).. 11. Stromatolite reef from the Early Archaean era of Australia. Nature 441, 714-718.. 12 13. Aminot, A., and Kérouel, R. (2004). Dissolved organic carbon, nitrogen and phosphorus. 14. in the N-E Atlantic and the N-W Mediterranean with particular reference to non-. 15. refractory fractions and degradation. Deep Sea Res. Part Oceanogr. Res. Pap. 51, 1975–. 16. 1999.. 17 18. Arp, G., Reimer, A., and Reitner, J. (2003). Microbialite Formation in Seawater of. 19. Increased Alkalinity, Satonda Crater Lake, Indonesia. J. Sediment. Res. 73, 105–127.. 20 21. Badaut, D., and Risacher, F. (1983). Authigenic smectite on diatom frustules in Bolivian. 22. saline lakes. Geochim. Cosmochim. Acta 47, 363–375.. 23 24. Barker, P., Fontes, J.-C., Gasse, F., and Druart, J.-C. (1994). Experimental dissolution of. 25. diatom silica in concentrated salt solutions and implications for paleoenvironmental. 26. reconstruction. Limnol. Oceanogr. 39, 99–110.. 27 28. Battersby, A.R., Fookes, C.J.R., and Snow, R.J. (1984). Synthetic studies relevant to. 29. biosynthetic research on vitamin B12. Part 1. Syntheses of C-methylated chlorins based. 30. on 1-pyrrolines (3,4-dihydropyrroles). J. Chem. Soc. [Perkin 1] 0, 2725–2732.. 31 This article is protected by copyright. All rights reserved.

(30) 1. Bontognali, T.R.R., Martinez-Ruiz, F., McKenzie, J.A., Bahniuk, A., Anjos, S., and. 2. Vasconcelos, C. (2014). Smectite synthesis at low temperature and neutral pH in the. 3. presence of succinic acid. Appl. Clay Sci. 101, 553–557.. Author Manuscript. 4 5. Bouton, A., Vennin, E., Pace, A., Bourillot, R., Dupraz, C., Thomazo, C., Brayard, A.,. 6. Désaubliaux, G., and Visscher, P.T. (2016). External controls on the distribution, fabrics. 7. and mineralization of modern microbial mats in a coastal hypersaline lagoon, Cayo Coco. 8. (Cuba). Sedimentology 63, 972–1016.. 9 10. Braissant, O., Decho, A.W., Dupraz, C., Glunk, C., Przekop, K.M., and Visscher, P.T. (2007).. 11. Exopolymeric substances of sulfate-reducing bacteria: Interactions with calcium at. 12. alkaline pH and implication for formation of carbonate minerals. Geobiology 5, 401–411.. 13 14. Braissant, O., Decho, A.W., Przekop, K.M., Gallagher, K.L., Glunk, C., Dupraz, C., and. 15. Visscher, P.T. (2009). Characteristics and turnover of exopolymeric substances in a. 16. hypersaline microbial mat. FEMS Microbiol. Ecol. 67, 293–307.. 17 18. Buckley, D.H., Baumgartner, L.K., and Visscher, P.T. (2008). Vertical distribution of. 19. methane metabolism in microbial mats of the Great Sippewissett Salt Marsh. Environ.. 20. Microbiol. 10, 967–977.. 21 22. Burne, R.V., and Moore, L.S. (1987). Microbialites: Organosedimentary Deposits of. 23. Benthic Microbial Communities. PALAIOS 2, 241–254.. 24 25. Burne, R.V., Moore, L.S., Christy, A.G., Troitzsch, U., King, P.L., Carnerup, A.M., and. 26. Hamilton, P.J. (2014). Stevensite in the modern thrombolites of Lake Clifton, Western. 27. Australia: A missing link in microbialite mineralization? Geology 42, 575–578.. 28 29. Busenberg, E., and Niel Plummer, L. (1985). Kinetic and thermodynamic factors. 30. controlling the distribution of SO32− and Na+ in calcites and selected aragonites.. 31. Geochim. Cosmochim. Acta 49, 713–725. This article is protected by copyright. All rights reserved.

(31) 1 Casillas-Martinez, L., Gonzalez, M.L., Fuentes-Figueroa, Z., Castro, C.M., Nieves-Mendez,. 3. D., Hernandez, C., Ramirez, W., Sytsma, R.E., Perez-Jimenez, J., and Visscher, P.T. (2005).. 4. Community Structure, Geochemical Characteristics and Mineralogy of a Hypersaline. 5. Microbial Mat, Cabo Rojo, PR. Geomicrobiol. J. 22, 269–281.. 6. Author Manuscript. 2. 7. Caumette, P. (1993). Ecology and physiology of phototrophic bacteria and sulfate-. 8. reducing bacteria in marine salterns. Experientia 49, 473–481.. 9 10. Decho, A.W. (2000). Microbial biofilms in intertidal systems: an overview. Cont. Shelf. 11. Res. 20, 1257–1273.. 12 13. Decho, A.W. (2010). Overview of biopolymer-induced mineralization: What goes on in. 14. biofilms? Ecol. Eng. 36, 137–144.. 15 16. Decho, A.W., Visscher, P.T., and Reid, R.P. (2005). Production and cycling of natural. 17. microbial exopolymers (EPS) within a marine stromatolite. Palaeogeogr. Palaeoclimatol.. 18. Palaeoecol. 219, 71–86.. 19 20. Decho, A.W., Norman, R.S., and Visscher, P.T. (2010). Quorum sensing in natural. 21. environments: emerging views from microbial mats. Trends Microbiol. 18, 73–80.. 22 23. Des Marais, D.J. (2003). Biogeochemistry of Hypersaline Microbial Mats Illustrates the. 24. Dynamics of Modern Microbial Ecosystems and the Early Evolution of the Biosphere.. 25. Biol. Bull. 204, 160–167.. 26 27. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A.T., and Smith, F. (1956). Colorimetric. 28. method for determination of sugars and related substances. Anal. Chem. 28, 350–356.. 29 30. Dupraz, C., and Visscher, P.T. (2005). Microbial lithification in marine stromatolites and. 31. hypersaline mats. Trends Microbiol. 13, 429–438. This article is protected by copyright. All rights reserved.

(32) 1 Dupraz, C., Visscher, P.T., Baumgartner, L.K., and Reid, R.P. (2004). Microbe–mineral. 3. interactions: early carbonate precipitation in a hypersaline lake (Eleuthera Island,. 4. Bahamas). Sedimentology 51, 745–765.. 5. Author Manuscript. 2. 6. Dupraz, C., Reid, R.P., Braissant, O., Decho, A.W., Norman, R.S., and Visscher, P.T. (2009).. 7. Processes of carbonate precipitation in modern microbial mats. Earth-Sci. Rev. 96, 141–. 8. 162.. 9 10. Dupraz, C., Reid, R.P., and Visscher, P.T. (2011). Microbialites, Modern. In Encyclopedia. 11. of Geobiology, J. Reitner, and V. Thiel, eds. (Springer Netherlands), pp. 617–635.. 12 13. Dupraz, C., Fowler, A., Tobias, C., and Visscher, P.T. (2013). Stromatolitic knobs in Storr’s. 14. Lake (San Salvador, Bahamas): a model system for formation and alteration of laminae.. 15. Geobiology 11, 527–548.. 16 17. Flemming, H.-C., and Wingender, J. (2010). The biofilm matrix. Nature Reviews. 18. Microbiology 8, 623–633.. 19 20. Fossing, H., and Jørgensen, B.B. (1989). Measurement of bacterial sulfate reduction in. 21. sediments: Evaluation of a single-step chromium reduction method. Biogeochemistry 8,. 22. 205–222.. 23 24. Gallagher, K.L., Kading, T.J., Braissant, O., Dupraz, C., and Visscher, P.T. (2012). Inside the. 25. alkalinity engine: the role of electron donors in the organomineralization potential of. 26. sulfate-reducing bacteria. Geobiology 10, 518–530.. 27 28. Gautret, P., Camoin, G., Golubic, S., and Sprachta, S. (2004). Biochemical Control of. 29. Calcium Carbonate Precipitation in Modern Lagoonal Microbialites, Tikehau Atoll,. 30. French Polynesia. J. Sediment. Res. 74, 462–478.. 31 This article is protected by copyright. All rights reserved.