Indigenous microbial communities as catalysts for early marine cements: An in vitro study

Early marine cementation is a fundamental process for many characteristics of carbonates, like the stabilisation of steep slopes. The genesis of early cements is often attributed to physicochemical processes but there is evidence for microbial mediation. To elucidate the role of microbes and associated organic material, in vitro experiments were undertaken in the presence and absence of indigenous microbiota in ooids from Schooner Cays, Bahamas and compared with native grapestones from Joulter Cays, Bahamas. Microscopic examinations by stereomicroscopy, scanning electron microscopy and thin section analysis of in vitro incubations with native flora document rapid grain fusion, resulting in the formation of grapestones within 30–60 days. The initial binding of the grains is primarily facilitated by exudates of extracellular polymeric substances and microbial communities acting as catalysts in the formation of micritic bridges, cements and encrusted aggregates. In vitro grapestones are similar to native grapestones from Joulter Cays with intergranular areas infested with extracellular polymeric substances, microbes, micritic cements, amorphous calcium carbonate nanograins and micritised outer surfaces. These similarities suggest that incubations with native flora follow similar mineralisation mechanisms as in the natural environment. In contrast, sterilised grains remain loose with little crystal formation after 60 days and are devoid of microbes and organic exudates. Owing to the near absence of precipitates, abiotic precipitation is not the driving force promoting early cements. In contrast, grain fusion is microbially mediated via both a passive mechanism, where extracellular polymeric substances and cell surfaces function as templates for crystal nucleation and generation of micritic cements, and through an active mechanism by which biofilm heterotrophs and autotrophs induce chemical alterations of a local environment, facilitating precipitation. This study underscores that microbially mediated cementation can occur at fast rates and that firmground to hardgrounds and slope stabilisation take place shortly after deposition of carbonate grains.

The origins of early carbonate precipitates have largely been attributed to physicochemical precipitation of carbonates in sea water, prompted by evaporation, water mixing, increases in water temperature and CO 2 degassing (Bathurst, 1971;Ginsburg & James, 1973;Tucker & Wright, 1990). Among these factors, the pore exchange of supersaturated sea water-with respect to CaCO 3through the sediments have received the most attention and have been considered one of the main controls in cementation processes in shallow-water carbonates (Moore, 2001;Scholle & Halley, 1985). As a result, fluid models have been developed to predict the volumes of sea water necessary to fill granular pores with abiotic cements (Dunham, 1969;Scholle & Halley, 1985). The theoretical estimates, however, rely on exorbitant volumes of supersaturated water (e.g. 10 000-100 000 pore volumes of sea water/per singular pore grain) in tandem with exceeding precipitation efficiencies, both of which are not attainable in natural systems (Chatalov, 2000;Morse & Mucci, 1984).
Given that abiotic precipitation of carbonates cannot fully explain the extent and origins of many early carbonate precipitates, a growing number of studies are recognising the role of microbes and organic matter in early cement formation (Diaz & Eberli, 2022;Fabricius, 1977;González-Muñoz et al., 2010;Hillgärtner et al., 2001). For instance, endolithic and chasmolithic organisms (e.g. bacteria, fungi, foraminifera and diatoms) have been associated with early cements and are considered important contributors to the formation of grapestones and hardgrounds (Dravis, 1979;Fabricius, 1977;Flügel, 2010;Hillgärtner et al., 2001). Others have theorised that cementation of carbonate sands and marine peloids are biotic carbonate products derived from metabolically active communities able to induce carbonate precipitation by altering solution chemistry (Chafetz, 1986;Dravis, 1979). The microbially induced mechanism challenges the traditional notion of abiotic carbonate precipitation as it has been associated with the genesis of cementation in reefs (Kaźmierczak & Iryu, 1999), intraclasts (Diaz & Eberli, 2022) and microbialite systems Dupraz & Visscher, 2005;Grotzinger & Knoll, 1999), as well as the formation of ooids (Diaz et al., 2014(Diaz et al., , 2015(Diaz et al., , 2017Harris et al., 2019) and stabilisation of slopes (Della Porta et al., 2003;Reolid et al., 2017). In addition, extracellular polymeric substances (EPS) and microbial cell walls have also been implicated in the origins of early cements by acting as templates for mineral precipitation Diaz & Eberli, 2022;Hillgärtner et al., 2001;Reitner, 1993;.
Despite the growing number of studies that provide evidence on the role of microbes in microbial carbonate precipitation, there is little experimental work linking marine cementation to microbial mediation. This study tests the hypothesis that initial sediment stabilisation and agglutination of oolitic grains are mediated by microbial-mineral interactions. Toward this end, an experiment was designed to evaluate-through visual observations using scanning electron microscopy (SEM) and thin section analyses-whether natural occurring microbial communities in the ooid sands intervene in the initial stages of grain aggregation and cementation. The experimental system, which uses freshly collected ooids from Schooner Cays in the Bahamas, involves a series of in-vitro incubations under conditions of no agitation to emulate the quiescent conditions of the flank areas of the shoals where aggregate grains tend to form. For a comparison of microbially mediated versus abiotic cementation, experiments were carried out simultaneously in the presence and absence of indigenous microbiota/organic material (using chemical and sterilisation treatments). Finally, ooids from in-vitro incubations are compared with grapestones from the stabilised sand flats at Joulter Cays in the Bahamas, where carbonate grains are actively undergoing marine cementation. The microbial molecular structure in ooids on Great Bahama Bank has been previously established using functional gene and 16SrRNA analyses (Diaz et al., 2013(Diaz et al., , 2014. This study provides a valuable contribution to the current understanding of the role of microbes in early cementation processes.

| Sample collection and processing
Sediments from Schooner and Joulter Cays were collected and transferred into sterile 50 and 15 ml Falcon tubes. After collection sea water was decanted and replaced with 80% ethanol or 4.5% formaldehyde for microscopic analysis. These samples were kept at −20°C. Incubation experiments used two sets of freshly collected sediments with no preservative agent added to them and stored at 4°C until further processing. Sediments from Schooner Cays were divided into two cohort groups, the sterilised and the untreated group. While the untreated group represents ooids with their indigenous microbial communities, the sterilised group comprises ooids that underwent heat sterilisation and removal of organics. To remove organic materials from the grain, ooids underwent two successive washes with DI water followed by an overnight incubation with 5% NaOCl. This was followed by five additional washes with DI water and a heat sterilisation (autoclaved) treatment to ensure sanitation and the complete removal of microbes. Ooids used for the incubation experiments were kept at 4°C until further processing.

| Experimental set-up
The incubation chambers-for both experimental groups-consisted of 2 L Erlenmayer flasks filled with 1.8 L of filtered sea water drawn from Bear Cut, Biscayne Bay, Florida, a tidal channel located next to the Rosenstiel School of Marine and Atmospheric Science where the experiment was conducted. The sea water was sterilised using PALL Maxi capsule filters (0.2 μm pore size). The vessels were placed inside a Class 2 biosafety cabinet and maintained at room temperature to allow a containing system free of contaminants. About 30 g wet weight of sterilised and untreated ooids were screened for the potential presence of grapestones and any present were removed with sterilised forceps. The grains were then aseptically transferred and assembled into silicone rubber sleeves ( Figure 1). The packed grains were sealed with two porous disks, permitting the containment of the sediments and the inflow/outflow of sea water through the sleeve. A continuous inflow was applied to the sleeves through a central port (silicone connector) connected to a silicon hose and a dosing pump using a flow rate of 10 ml/min. To avoid contamination, the glass chamber neck was enclosed with parafilm, sanitised with hydrogen peroxide. All components, including sleeves, tubing and stoppers were autoclaved, except for few connectors, which were sanitised with bleach. Regular exchanges of un-supplemented sterilised sea water (sea water not enriched with nutrients) were conducted on weekly bases to ensure a periodic renewal of naturally occurring nutrients. Nutrient levels from the intakes of Bear Cut sea water typically average 0.9 μM NO 3 , 0.3 μM NH 4 and 0.03 μM PO 4 , while dissolved inorganic carbon (DIC) and total alkalinity average 2076 (±45 μmol/kg) and 2432 (±52 μmol/kg) respectively (Langdon et al., 2018). The pH of filter-sterilised sea water ranged from 8.2 to 8.3, while the salinity levels ranged from 34 to 36 parts per thousand (ppt). The incubations were designed to emulate the natural environment in the form of natural sea water as a growth medium to sustain microbial populations. Using this approach, any microbial growth and metabolic reactions that occur within the 'sleeve microenvironment' is sustained at the expense of ambient levels of nutrients and growth factors from environmental sea water. Likewise, sea water was the source of free calcium ions F I G U R E 1 Incubation system used for ooid cementation studies. The system uses glass vessels (Pyrex glass) with a pump to maintain a constant flow of unamended sterilised sea water (free of external amendments of nutrients, cementation reagents or bacteria inoculum) through stacked ooids. The vessels were kept inside a biosafety class two cabinet, maintained at room temperature to allow a containment system free of contaminants. The system uses natural sea water as a growth medium and as a source of free calcium ions for mineralisation. for mineralisation. In addition, the samples were subject to alternating cycles of daylight and dark conditions to sustain both photosynthetic and heterotrophic activity. The incubations were terminated at 30 and 60 days, after which subsamples were aseptically recovered and preserved both in 80% ethanol and 4.5% formaldehyde. Observations of sediment microstructures were performed using a stereomicroscope and SEM to visualise any dimensional topography of biological, organic or mineral origins. Special attention was given to grain contact areas.

| Petrographic investigation
Sediment samples were first examined using a stereoscope and a Dino-lite AM3111T digital microscope. Resin impregnated 1″ x 2″ petrographic thin sections were prepared with Epoxy resin and stained with blue dye. Thin sections were examined with an Olympus BH2 petrographic microscope using plane and crossed polarised lights. The grain size was determined on a subset of samples by sieving and weighing each grain size fraction.

| SEM analysis
Samples were preserved on site and followed the protocol as described by Diaz et al. (2017). Preserved samples were kept in the dark at 5°C. Once in the laboratory, the preserved specimens were treated with 2% gluteraldehyde and 0.2 μm millepore filtered 0.05 M sodium cacodylate buffered sea water. The latter procedure was undertaken to ensure the integrity of microbial structures. The samples were further treated with three rinses with sea water buffer and a dehydration step consisting of a graded series of ethanol (20%, 50%, 70%, 95% and 100%). The specimens were dried in three changes of hexamethyldisiloxane (HMDS) and coated with palladium (Pd) in a plasma sputter coater for conductivity. Imaging analysis was undertaken with a field emission SEM (Philips XL-30).

| Marine organism identification
Bacterial and fungal identification followed conventional morphological traits such as shape, size and surface structure of cells (Bergmans et al., 2005;Hawksworth, 1988;Vos et al., 2009), whereas other microbial forms such as protists (e.g. diatoms, foraminifera, coccolithopores) were identified based on their morphology and characteristic traits of their cell and/ or body structures (Hasle & Syvertsen, 1996;Young et al., 2003).

| Stereomicroscopy
To establish the initial baseline characteristics of the sedimentary grains, visual observations were undertaken at time zero (starting point of the assay) with a stereomicroscope for each sample (untreated and sterilised groups). Examinations of untreated samples (ooids with their native microbial flora) at time 0 depict well-polished rounded/oval grains, some of which are colonised by biofilm-EPS distributed in a non-homogenous pattern ( Figure 2). As the incubation time progresses from 30 to 60 days, colonisation of biofilm-EPS on grain surfaces and in grain contact areas is more conspicuous and appears as pigmented specks with spotty distribution or mucous films that vary in thickness and density ( Figure 2). The EPS reveal the presence of various pigments of microbial origins, with colour hues ranging from rust (e.g. carotenoids), green (e.g. chlorophylls), blue-green (e.g. phycocyanines) to black-grey (e.g. melanins) ( In contrast, ooids that underwent sterilisation and removal of organics with hypochlorite remain largely clean, polished and devoid of biofilm-EPS and/or microbial pigments at all time intervals ( Figure 2G,H,I). The sterilised ooids-as opposed to untreated ooids-remain unconsolidated at 30 and 60 days but at 60 days a few grains have tiny mineral crystals on their outer cortices ( Figure 2I).

| Petrographic thin sections
Thin section analysis of untreated and sterilised sediments (time 0), collected from a high-energy shoal crest in Schooner Cays, shows moderately well-sorted medium sized grains with no mud or silt and mainly consisting of ooids, followed by peloids and some skeletal fragments and foraminifera that are either broken or abraded ( Figure 3A,B,G). The sand grains typically range from ca 400 to 620 μm on Schooner Cays tidal bars (Rankey & Reeder, 2011). At the sample site, 85% of the ooids are ca 500 μm. Many ooids bear a well-defined nucleus, mainly composed of peloids but sometimes of quartz and shell fragments ( Figure 3A,G), while others show two peloidal or vaguely defined nuclei ( Figure 3A,B). Some cortical laminations are micritised (time 0) ( Figure 3A,B).
Through all the incubation periods, the sterilised grains remain loose with an admixture of well-preserved and micritised grains with no traces of intergranular EPS-mucilage or grain agglutination ( Figure 3G,H,I). In contrast, untreated ooids show initial stages of grain clumping, mainly assisted by EPS-mucilage and/or microbial filaments in intergranular pore areas ( Figure 3C through F). The presence of EPS is more conspicuous at 30 and 60 days, with some grains embedded within an amorphous matrix ( Figure 3F). Some grains are now connected by bridging ( Figure 3D) and meniscus cements ( Figure 3F inset). While clumping of the grains tends to be more abundant at 60 days ( Figure 3E), some ooids also show grain dissolution within the first 30 days (Figure 3C inset).

| SEM analysis
The SEM analysis on freshly preserved sedimentary grains reveals that untreated ooids (time 0) harbour a consortium of microbes, including coccoid cells, rod bacteria, pennate diatoms and filamentous microbes, that is, fungi and cyanobacteria ( Figure 4). As previously observed with stereoscopic examinations, the EPS biofilm is locally pervasive, often appearing as desiccated sheets ( Figure 4D). Micropeloids of aragonite needles with sharp and blunt ends are also present ( Figure 4B). As the incubation time progresses to 30 days, the EPS biofilm is more prevalent at contact areas, sometimes resembling thin flakes with swirled edges and/or smooth films connecting intergranular areas ( Figure 4E,F,H). The grain contact areas are often colonised by diverse bacterial forms, including rod-shaped bacilli, coccoids, spiral-shaped bacteria (spirilla) and single-flagellum biofilm bacteria ( Figure 4E,F,G) Other microbes, include coccolithophores, fungi and cyanobacterial filaments, acontial nematocysts with coiled thread morphology and boat-shape pennate diatoms, that is, Navicula spp., are present and often act as physical bridges connecting the grains ( Figure 4G,H). Micrite cement is documented in intergranular areas ( Figure 4G) as well as clusters of nanograins covering localised areas within the EPS matrix or on the surface of diatom frustules ( Figure 4H). At 60 days, untreated samples show firm grain aggregates and/or encrusted lumps of micritised grains ( Figure 4I,J) and micrite cements ( Figure 4I). Some of the grain contacts and outer cortices are however heavily micritised with areas denoting carbonate dissolution and cementation ( Figure 4J). Conspicuous infestation of biofilm bacteria is apparent with dense biofilm threads acting as anchor points ( Figure 4K). The EPS appear as thin threads, mucous, smooth film structures and dehydrated films. In addition, microbes capable of EPS production were identified, including filamentous cyanobacteria/fungi, rod-shaped bacteria, coccoliths of Umbilicosphaera sibogae and a diverse population of pennate diatoms with morphological traits in common with of Mastogloia sp., Cocconeis sp., Navicula sp. and Amphora coffeaeformis ( Figure 4L). Secretions of mucilage EPS by a Cocconeis sp. and nanograin clusters in association with diatoms, EPS and fibrous aragonite were also evident ( Figure 4L inset).
Conversely and regardless of the incubation time, sterilised ooids remain clean and notably absent of viable biogenic material and EPS. The grains are unconsolidated with no signs of early cements at grain contacts ( Figure 5A,D,E,G). Some ooids, however, are heavily micritised on localised areas of the outer cortex ( Figure 5C,H) and interior of the grain with well-defined bore holes partially infilled with acicular aragonite-nano needles ( Figure 5F). Fingerprints of micritisation are probably the product of endoliths while the sediments were in their natural habitat, prior to sediment collection. Contrary to untreated ooids, which show a progression in grain erosion as the incubation period increases, no major differences in micritisation was documented among sterilised grains at 0, 30 and 60 days (Figure 5A,D,G).

| Joulter Cays
Sediments collected from the sand flat areas of Joulter Cays are largely stabilised by both seagrass beds of Thalassia testudinum and cyanobacterial mats. These sediments consist of medium sorted grains (200-500 μm in diameter), mostly comprising ooids, grapestones and a mixture of skeletal, peloidal sands, carbonate muds and foraminifera ( Figure 3C,D). Most of the ooids are heavily micritised ( Figure 6). The SEM photomicrographs reveal that many of the grains consist of loose single grains and composite grains that are clumped together ( Figure 6D). Micrite cements are commonly seen at grain-contacts and outer layers of the grain. Many of the grains are heavily bored by fungi and/or cyanobacteria. Extracellular polymeric substances are commonly seen at contact areas ( Figure 6E,F) and often associated with diverse  (Chentir et al., 2018;Filali-Mouhim et al., 1993;Hoagland et al., 1993;Underwood & Paterson, 2003) ( Figure 6E). Many of the microbes are embedded within EPS mucilage. The EPS structure is variable, seen as continuous or fragmented sheets.
Amorphous calcium carbonate nanograin precipitates occur along various edges of the fragmented EPS matrix ( Figure 6E). Etching of the grains reveal that many of the bored channels are infilled with both finer and coarse micrite cements and fibrous aragonite needles ( Figure 6G). Some of the bored channels display biological imprints of different morphologies with fine or coarse cements.  Hammes et al., 2003). Less attention, however, has been given to the effect of native microbial communities on carbonate mineralisation in ooid sands. Given the complexity of natural communities and their inherent microbial interactions, which influences net community calcification (driven by metabolic processes that lead to mineral precipitation and dissolution), here the impact of indigenous microbial communities in sediment agglutination is elucidated using in vitro experiments and compared to the effect achieved with naturally occurring grapestones from the microbially stabilised sand flats at Joulter Cays. Contrary to other experiments, which are conducted on artificially prepared solutions or on agar, this experimental approach employs in vitro conditions in the absence of external amendments of nutrients, cementation reagents or inoculum of specific bacterial strains-normally used by others to bolster microbial growth and/or carbonate cementation. The microbial composition of these native communities has been previously documented in similar and surrounding ooid shoals in the Bahamas (Diaz et al., 2013(Diaz et al., , 2014Diaz & Eberli, 2022). These studies have found high diversity at both functional and phylogenetic level, with Proteobacteria (e.g. Alpha-, Beta-Delta-and Gammaproteobacteria) being the most predominant and metabolically diverse group, accounting 50%-61% of the total bacteria diversity. Other important groups include: Cyanobacteria, Bacteroidetes, Verrumicrobia, Actinobacteria and Planctomycetes (Diaz et al., 2014).

| Microbial mediation
Based on these experimental observations, the initial binding of ooids is primarily mediated by a consortium of microbes (e.g. biofilm bacteria, diatoms, filamentous cyanobacteria/fungi and coccolithophores) in concert with mucilage secretions of EPS-biofilm producers such as cyanobacteria, diatoms, coccolithopores (Figures 2, 3, 4, 6 and 7). As infestation of the grain becomes noticeable by the growth of diverse microbial forms and EPS colonisation, binding of sedimentary grains progresses into the eventual formation of early cements (i.e. ACC nanograins and micrite cements) and subsequent cementation of micritised ooids. This process, which was most notable in the 30 and 60 days incubations with untreated ooids (ooids with their indigenous microbial flora), is believed to be partially driven by the physiological activities of phototrophs (cyanobacteria, coccolithophores, diatoms) and heterotrophs (biofilm bacteria), both of which can alter solution chemistry to induce carbonate precipitation by increasing the pH and/or level of dissolved inorganic carbon (Diaz et al., 2013(Diaz et al., , 2014(Diaz et al., , 2017Hillgärtner et al., 2001;O'Reilly et al., 2017;Summons et al., 2013). For example, functional gene analysis conducted on ooids from similar sand flats in the Bahamas, where early cements form, identifies an abundant number of genes associated with geochemical processes that foster carbonate precipitation (Diaz et al., 2014). Among these, denitrification appears as the most predominant functional yield followed by sulphate reduction, photosynthesis and ammonification. While a specific physiological activity contributing the most to grain agglutination cannot be identified, as no metabolic profiling studies were undertaken, SEM observations and past molecular and geochemical studies (Diaz et al., 2013(Diaz et al., , 2014(Diaz et al., , 2017, suggest that carbonate precipitation is partially driven by a temporal sequence of metabolic events,  prompted by the synergetic effect of microbes with different physiologies such as: phototrophs, denitrifiers, sulphate reducers, anaerobic sulphide oxidation and ammonifiers. This is in line with earlier studies that have implicated similar microbial metabolisms in early lithification processes (Chou et al., 2011;DeJong et al., 2011;Diaz & Eberli, 2022;Dupraz & Strasser, 1999;Gat et al., 2014;Krajewska, 2018;Morad, 1998;Purdy, 1963aPurdy, , 1963b. Many studies, however, have highlighted the pivotal role of sulphate reduction in the lithification of hypersaline coastal ponds (Baltres, 1975), micritic layers in stromatolites (Visscher et al., 2000) and microcrystalline cements in reef deposits (Kaźmierczak & Iryu, 1999). Similarly, microbial activities of cyanobacteria and heterotrophs contribute to micro-encrusting layers of shallow coral bioherms (Dupraz & Strasser, 1999).

indigenous microbial communities, biofilm-mucilage (EPS), clusters of ACC nanograins (CN) and micropelotoids (Mp) in shallow depressions of ooid cortices at time 0. Rod-shaped bacteria (B), biofilm bacteria (BB), coccoid bacteria (CB) and diatoms (D). (E through H) Photomicrographs illustrating the pervasive nature of EPS and associated microbes in contact areas at 30 days incubation time (E) grain binding by biofilm-EPS and biofilm-bacteria swarming on mucilage (inset). (F) Intergranular areas colonised by EPS and indigenous bacterial communities consisting of coccoid (CB) and rod-shaped bacteria (B). Note the presence of ACC nanograins (N) and clusters of ACC
In addition, microbes could influence early cementation by cell entombing through the adsorption of metal cations (e.g. Ca 2+ or Mg 2+ ) around their negatively charged cellular surfaces comprising carboxyl, phosphate and amine groups (Folk & Leo Lynch, 2001;Hillgärtner et al., 2001). Although this type of passive precipitation is generally assumed to occur on dead and/ or inactive cells, bacterial extracellular precipitation  (Obst et al., 2009;Shiraishi et al., 2020) via the synergetic effect of both metabolic activities and microbial cell surfaces-the latter providing nucleation hot spots for the growth of extracellular precipitates (Hoffmann et al., 2021). However, for active evasion against uncontrolled mineralisation that could lead to cell entombment, some microbes have developed strategies, including for example, shedding of encrusted S-layers (Schultze-Lam et al., 1992); generation of nanoglobules and mineral sheaths (Aloisi et al., 2006;Gilbert et al., 2005); production of nanospheres adjacent to the cell surface (Bundeleva et al., 2014); and shedding of mineralised EPS layers (Douglas & Beveridge, 1998).
Examples of microbial cell walls serving as nucleation centre for micrite formation are documented in Figure 7, where microbial cell surfaces (e.g. rod-shaped bacterium, diatom) are covered by micrite and/or aragonite crystals. Although it is unclear whether these precipitates formed during in vivo or post-mortem events, based on the shape and integrity of the cell walls-with no evidence of fragmentation or cellular collapse-the generation of micrite cements could have occurred, but not exclusively, while the cells were viable. Many of these mineral encasements can eventually lead to biogenic encrustations that ultimately act as binding matrices of the grains (Diaz & Eberli, 2022;Folk & Leo Lynch, 2001;Hillgärtner et al., 2001). For instance, Diaz and Eberli (2022) recently documented a diverse array of biogenic encrustations, for example, sessile sponges, foraminifera, diatoms, coccolithopores, bryozoans in the interstices of composite grains, which contribute to grain agglutination and early cementation processes in intraclasts from the Bahamas and Hamelin pool, Australia.

| Extracellular polymeric substance
Thin sections and SEM images also confirms the occurrence of EPS on outer cortices and in the interstices between the grains that appear more pronounced as the incubation time is lengthened in the non-treated samples (Figures 3 and 4) as well as in the grapestones from Joulters Cays ( Figure 6). This observation is further supported by stereomicroscopy examinations documenting EPS and associated pigments that range from rust colour-herein interpreted as carotenoids and likely the product of photosynthetic bacteria, such as purple non-sulphur bacteria (de Carvalho, 2018)-to blue-green and black-grey hues ( Figure 2). While the blue-green biofilm is typically associated with phototrophic groups-for example, cyanobacteria, green sulphur (e.g. Chlorobium) and green non-sulphur bacteria-the black-grey biofilm has been associated with photoheterotrophs, such as Chloroflexi (Kusumi et al., 2013). Although, phylogenetic determinations based on pigment colouration are rather speculative, some of these microbial groups have been previously identified using DNA-based analyses and as such implicated as potential contributors to carbonate precipitation in the ooid sands of the Bahamas Diaz et al., 2013Diaz et al., , 2014Summons et al., 2013). The EPS, which is often seen embedding and binding the grains, testify to the putative role of EPS in sediment agglutination and formation of early cements. Aside from multiple ecological functions-for example, trapping of nutrients, cell aggregation, protection against environmental stressors and source of organic carbon for heterotrophs (Chenu, 1995;Costa et al., 2018)-the EPS is known for its cohesive bonding properties (Chenu, 1995;Costa et al., 2018) and passive precipitation processes in microbialite systems Dupraz et al., 2009;Kaźmierczak et al., 2015;Reitner, 1993), ooid sands and early cements (Diaz & Eberli, 2022;Hillgärtner et al., 2001). Although there are diverging views regarding the mechanism by which EPS facilitates carbonate precipitation, it is generally accepted that as the EPS degrades-through physical or biological processes-metal cations (e.g. Ca 2+ or Mg 2+ ) are released into the surrounding environment, creating high concentrations of free divalent cations, thereby increasing calcium carbonate supersaturation and enhancing carbonate precipitation through crystal nucleation (Défarge et al., 1996;Dupraz & Visscher, 2005;Dupraz et al., 2009;Trichet et al., 2001). While there are instances when precipitation in non-degraded EPS matrices can occur by a constant supply of cations and alkaline conditions, generally un-degraded EPS can inhibit precipitation given its high binding capacity for divalent cations that results in depletion of free metals from solution (Arp et al., 1999;Dupraz et al., 2009).
Untreated ooids and grapestones from Joulter Cays are also colonised by a myriad of organisms with capabilities for EPS production, which is not limited to photoautotrophs such as cyanobacteria (Chentir et al., 2018;De Philippis et al., 2001), diatoms (Hoagland et al., 1993;Underwood & Paterson, 2003) or coccolithopores (De Jong et al., 1979) but include heterotrophs like biofilm bacteria (Braissant et al., 2007) and filamentous fungi (Mahapatra & Banerjee, 2013;Selbmann et al., 2003). Some of these microbial communities, are herein reported (e.g. pennate diatoms and coccolithopores) exuding copious amount of EPS ( Figures 4L and 7D), swarming on mucilage (e.g. biofilm bacteria) ( Figure 4E inset), entrenched in EPS ( Figure 6F), or attached to the grains through a dense mesh of EPS with fimbriae-like structures ( Figure 4K). Similar microbial communities and putative occurrence of EPS have been reported on preserved intraclasts from various depositional environments in the Bahamas and Hamelin Pool, Australia (Diaz & Eberli, 2022).

| Amorphous calcium carbonate-ACC
Nanograin precipitates with similar size and morphological characteristics as ACC (Addadi et al., 2003;Diaz & Eberli, 2022;Diaz et al., 2015Diaz et al., , 2017Ihli et al., 2014;Jones & Peng, 2012;Mann, 1988;Meldrum & Cölfen, 2008) are prevalent features in the non-treated group and grapestones from Joulter Cays. These precipitates are often seen in the periphery of decaying EPS sheets as isolated nanoparticles (Figures 4 and 6), embedded within mucilaginous matrices or forming clusters of nanoparticles in grain contact areas colonised by thick and dense layers of EPS (Figures 4 and 7D). Based on previous studies, these amorphous precipitates are the least stable polymorphs of CaCO 3 (Ihli et al., 2014), and as such considered a transient precursor of crystalline carbonates (Addadi et al., 2003;Dupraz et al., 2004;Ihli et al., 2014). Despite being very unstable, ACC is far more prevalent than previously thought and its relevance in the generation of biominerals has been well documented for a wide variety of taxa, including diatoms (Diaz & Eberli, 2022;Stanton, 2019), coccolithophores (Durak et al., 2016), crescent-shaped bacillus (Diaz et al., 2017), cyanobacteria (Blondeau et al., 2018;Chung et al., 2010;Kamennaya et al., 2012;Riding, 2006) and invertebrates with skeletal tissues, including bivalve molluscs and gastropods (Addadi et al., 2003;Ihli et al., 2014;Marxen et al., 2003;Weiss et al., 2002). Most recently, it has been suggested that ACC of biogenic origin acts as seeds for cortex accretion in ooids Diaz et al., 2017) and formation of micritic cements (Diaz & Eberli, 2022). According to Diaz and Eberli (2022), the initial stages of micrite cements begin with the prenucleation of ACC nanograins on a degraded EPS matrix that as it progressively decays, forms individual nanograins that coalesce into clusters. Subsequent reactions of ACC nanograins with organic matter (i.e. carboxyl, phosphate, amino and sulphydryl groups) and metal cations (i.e. Ca 2+ , Mg 2+ ), lead to the formation of a biogenically stable ACC/organo-complex that functions as secondary nucleation sites for the conversion of ACC into stabilised crystalline phases, for example, micrite, calcite crystals and aragonite.

| Micritisation
While untreated samples show evidence of agglutination and early cement formation that appear to progress as the incubation time increases-as opposed to sterilised samples-fingerprints of micritisation and dissolution are also evident, indicating that microbes are actively contributing to precipitation and carbonate dissolution through microbial processes taking place contemporaneously ( Figures 3C inset and 4J). On the other hand, given the absence of microbes in the sterilised group, any micritisation of the grains is attributed to diagenetic events that occurred prior to sample collection, while the sediments were in their natural habitat.
Based on observations reported on here the micritisation of the grains, which in some extreme cases can lead to the destruction of grains by repetitive boring cycles, is primarily driven by microbial endoliths, some of which are capable of dissolving or weakening the grain largely through the release of acidic metabolites-mainly generated by aerobic heterotrophic activities-or proteolytic enzymes (Baumgartner et al., 2009;Dupraz & Visscher, 2005). In addition, the extrusion of Ca 2+ ions via active calcium pumps has been identified as another mechanism contributing to the dissolution of carbonates by cyanobacteria (García-Pichel, 2006;Ramírez-Reinat & Garcia-Pichel, 2012). However, the prevalence of oxygenic phototrophs (e.g. cyanobacteria, diatoms and coccolithopores)-as herein documented-can counterbalance some of the bioerosion activities and as such render them as plausible players in the formation of micrite. For instance, the potential for pennate-diatoms to precipitate carbonates has been recently documented in blooming events (Gomez et al., 2018). Likewise, microboring activities of fungal endoliths can both be offset by fungal activities that control calcium concentrations (Bindschedler et al., 2016) and fungal metabolisms that induce increases in alkalinity (e.g. physicochemical degassing of respired CO 2 ) (Bindschedler et al., 2016), organic acid oxidation (Guggiari et al., 2011), urea mineralisation (Burbank et al., 2011) and nitrate assimilation (Hou et al., 2011). In addition, the chitin content of fungal cell walls-known for its ability to bind divalent cations-together with their

A B C
intertwined mycelial network can both serve as preferential sites for cation adsorption and deposition of carbonate minerals (Burford et al., 2006;Manoli et al., 1997). Owing to the fungal ability for CaCO 3 precipitation some have explored their potential utility as biological 'healers' of concrete cracks (Menon et al., 2019). While the concurrent activities of bioerosion and calcification have been controversial, some theories centre either on the temporal and spatial separation of photosynthetic and boring activities or the expulsion of Ca 2+ ions via active calcium pumps-allowing dissolution of carbonates around the apical cells while enabling high interstitial pH, resulting in carbonate precipitation (García-Pichel, 2006;García-Pichel et al., 2010;Macintyre et al., 2000;Ramírez-Reinat & Garcia-Pichel, 2012).

CEMENTATION RATES
The striking similarity of the fused grains generated in the intertidal flats of Joulter Cays with those derived from in vitro incubations that bear indigenous microbial communities can be taken as an indication that this study's in vitro cements follow a similar mechanism of formation as those in natural systems (Figure 8). This is further supported by similarities in microbial groups and ubiquitous occurrence of EPS and nanograins, both of which are prominent features in grapestones and intraclasts from natural habitats as well as in the non-treated carbonate grains (ooids with indigenous flora) of this study. The latter observations provide evidence on the prominent role of microbially mediated precipitation and suggests that physicochemical factors alone are not the main driving force promoting early marine cements in carbonate sands.
In addition, the in vitro experiments provide a time frame for the formation of grapestones, intraclasts and marine cementation. An experiment in the natural environment by Grammer et al. (1999) has so far been the only study that assessed the rate of early marine cementation. In their experiment, which was carried out with samples of ooids suspended above the sea floor at various depth across the platform margin in the Tongue of the Ocean, it was shown that partial lithification of carbonate sands-as determined by the formation of fibrous aragonite cementoccurred within a minimum of 8 months in water depths of up to 60 m (Grammer et al., 1999). The in vitro experiments indicate that microbially mediated cementation is even faster, producing grapestones in 30 days. Recently, Reolid et al. (2017) documented in the matrix and encrusting bioclasts of Holocene and Miocene carbonate-platform slopes (ca 35°), distinctive microfabrics indicative of microbial activity during deposition. Based on in vitro experiments, the stabilisation and eventually the submarine cementation of platform-derived sediments can occur in just a few months, documenting that accretion and cementation of the steep slopes along platform margins is a syndepositional process. Likewise, the formation of firmgrounds to hardgrounds on the platform top, where microbes (e.g. bacteria, fungi, diatoms, foraminifera) have been implicated as key contributors to hardground formation (Dravis, 1979;Fabricius, 1977), might form rapidly after deposition of the carbonate grains. Observations made during this study have further implications for how we understand the physical attributes of sediments and soil behaviour as microbial-biomediation can alter the mechanical properties of the soil, by improving the shear strength of the sediments while reducing permeability and porosity (Soon et al., 2013;Umar et al., 2016).

| CONCLUSIONS
The importance of organic matter and microbial activity in the cementation of carbonate sands is well-established (Diaz & Eberli, 2022;Hillgärtner et al., 2001). In fact, many of the early diagenetic characteristics of early cements-known for their filamentous, meniscus and micritic fabrics-have been related to microbes (Diaz & Eberli, 2022). Incubation studies on ooid sands reported on here suggest that under quiescent conditions, when carbonate sediments are not stirred up by waves, tides or currents, microbes and associated products, that is, EPS can trigger the initial stages of sediment stabilisation and subsequent calcification.
Owing to the pervasive occurrence of diverse microbial communities, many of which are intrinsically associated with mucilaginous EPS, this study indicates that early cementation of sand grains is governed by microbial mediation mechanisms. One of these mechanisms follows a passive precipitation process, upon which the EPS and microbial cell walls act as nucleation surfaces for sequential precipitations that lead to an ACC metastable phase, prior to mineral crystallisation, for example, micrite, cements and aragonite. Another follows an active process, whereas precipitates are obtained by chemical modification of the local environment (e.g. carbonate alkalinity, DIC and pH) induced by metabolic activities. Through this mechanism, the net accumulation of micrite cements within inter/intragranular domain areas would largely depend on the balance between metabolic activities that are conducive to carbonate precipitation and dissolution as well as temporal-spatial fluctuations. Given the functional complexity of ooid microbial communities (Diaz et al., 2014), determining with precision the leading metabolic pathway responsible for cement precipitation is difficult to establish as metabolic activities from one group are intertwined with the metabolic processes of other microbes. However, based on the in situ incubations a temporal succession of metabolic events at interplay is foreseen. During the day, as oxygen becomes available-through the photosynthetic activities of cyanobacteria, diatoms and coccolithophoresan increase in oxygen levels and organic carbon, induce an increase in pH and alkalinity, leading to carbonate precipitation. However, as oxygen is produced and used as a terminal electron acceptor during the remineralisation of autotrophic organic matter by a consortium of microbes (e.g. aerobic heterotrophs, sulphide oxidisers, nitrifiers), a decline in pH is foreseen, leading to carbonate dissolution. At night, as microaerophilic or anoxic conditions develop, terminal electron acceptors, such as sulphate and nitrate are preferentially used for the remineralisation of autotrophic biomass by anaerobic heterotrophy (e.g. sulphate reduction, denitrification, nitrate-driven sulphide oxidation by sulphide oxidisers), resulting in supersaturation conditions that could pave the way for the formation of cements. This sequence of metabolic events can take place at grain contact areas and within spatially structured microenvironments (e.g. grain depression, pits and bored tunnels), where redox conditions develop by microbial syntrophic interactions.
Although our understanding of microbioerosion entails further studies, it is clear that the biogenic erosion of the carbonate grains (e.g. mechanical destruction, aerobic heterotrophy, acid secretions), although pervasive in some grains, can be countered by the biogenic formation of micrite, which occurs in both lightly or heavily bored grains. These precipitation events develop as endolithic organisms move from one grain to another, allowing fusing of the grain with fine micrite or thick crust precipitates. These early cements could form in a span of weeks to months, suggesting the rapid nature of this process.