Lithifying and Non-Lithifying Microbial Ecosystems in the Wetlands and Salt Flats of the Central Andes

The wetlands and salt flats of the Central Andes region are unique extreme environments as they are located in high-altitude saline deserts, largely influenced by volcanic activity. Environmental factors, such as ultraviolet (UV) radiation, arsenic content, high salinity, low dissolved oxygen content, extreme daily temperature fluctuation, and oligotrophic conditions, resemble the early Earth and potentially extraterrestrial conditions. The discovery of modern microbialites and microbial mats in the Central Andes during the past decade has increased the interest in this area as an early Earth analog. In this work, we review the current state of knowledge of Central Andes region environments found within lakes, small ponds or puquios, and salt flats of Argentina, Chile, and Bolivia, many of them harboring a diverse range of microbial communities that we have termed Andean Microbial Ecosystems (AMEs). We have integrated the data recovered from all the known AMEs and compared their biogeochemistry and microbial diversity to achieve a better understanding of them and, consequently, facilitate their protection.

Compared to other saline systems around the world, the Central Andean systems have, on the one hand, higher ultraviolet (UV) radiation expositions, higher levels of desiccation, and higher daily temperature fluctuations caused by the high altitudes; and on the other hand, a different chemical composition of the water (e.g., high arsenic concentration) attributable to the volcanic activity [5,6]. These environmental conditions resemble the early Earth and potentially extraterrestrial conditions. Therefore, the study of the AMEs could provide information about the early evolution of life on Earth, as well as knowledge for the search of life on Mars [7].
Biofilms (1) are communities of microbial cells associated with a solid surface and embedded in an exopolymeric substance (EPS) matrix [18]. The microbial colonization of a surface depends on the EPS matrix formation, which can trap inorganic and abiotic compounds as well as immobilize water. Microbial mats (2) are biofilms distributed in layers, defined by light, oxygen, and physicochemical requirements, which colonize both solid and sedimentary surfaces [19,20]. The mat-constructing biota escapes burial by freshly deposited sediments migrating vertically upwards, where it can again colonize newly deposited surfaces. The thickness of living microbial mat communities can reach several centimeters, even decimeters, and they can be formed in a wide variety of shallow aquatic environments, from fresh to thalassic water conditions, and even within halite or gypsum crusts [21].
Some microbial mat communities can influence carbonate precipitation giving rise to structures named microbialites (3). These are organosedimentary deposits accreted by sediment trapping, binding and/or in situ precipitation due to the growth, metabolic activities, and EPS matrix produced by the microorganisms of the microbial mat [22][23][24][25][26][27][28][29]. Geological records indicate that microbialites first appeared more than 3.7 Ga ago and were the main evidence of life on Earth for the next 2 Ga [30][31][32]. As the first communities performing significant oxygenic photosynthesis, they are thought to have played a major role in the oxygenation of the Earth's atmosphere [33,34].
This review aims to examine the current state of knowledge of the AMEs acquired during the past decade of Andean exploration. All the data gathered from the different AME studies have been integrated into a single high-level comparative analysis to assess the microbial and mineral diversity and complexity of these polyextreme environments.

AMEs in Central Andes Region
In the past decade, extensive exploration has been carried out on salt flats, lakes, hot springs, and fumaroles of Bolivia, Argentina, and Chile to study the AMEs [2,3]. Different AMEs were reported in these explorations, including biofilms, mats, microbialites, and endoevaporites (Fig. 4). However, not all types of AMEs were found in every prospected place. Endoevaporites were exclusive of salt flat environments, but not all the salt flats presented endoevaporitic microbial ecosystems. Similarly, not all types of microbialites were found in salt flats, hypersaline lakes, or brackish rivers [2]. Therefore, in an attempt to update all the knowledge generated until now, we have summarized the AMEs found in the Central Andes as well as their chemical conditions, geographical location, main characteristics, and references (Table 1). While some systems have only been reported, others have been described in more detail with respect to the characteristics of their habitat, composition of their mineral phase and microbiota, and certain aspects of microbial activity and physiology.

Carbonate Microbial Ecosystems: Mats and Microbialites
Carbonate precipitation can be biologically driven when microbial metabolism is the main process inducing supersaturation of CaCO 3 (by altering alkalinity, pH, and cations availability) or extrinsically driven when physicochemical processes are responsible for the carbonate supersaturation state (evaporation, salinity, CO 2 degassing, etc.) [22]. In HAALs, both processes take place [3,75,76].
Andean carbonate microbial ecosystems (CMEs) are generally located in water-mixed zones where fresh superficial and groundwater inputs reach the lake, promoting microbial development, as well as providing a source of carbonate/ bicarbonate ions for the carbonate precipitation [17,77]. The chemistry of the HAALs is generally thought to result from the weathering of surrounding volcanic rocks by the infiltration and passage of precipitation through them, leading to the formation of dilute inflow waters that subsequently concentrate by evaporation [97], but it is believed that the most important mineral and salinity supply comes from thermal springs that discharge into the superficial and underground waters [98]. Depending on the physicochemical conditions of the water (e.g., pH, Eh, alkalinity, temperature), carbonate, sulfate, and chloride minerals could precipitate. For instance, the abundance of Na + in the water allows the occurrence of gaylussite (Na 2 Ca(CO 3 ) 2 ·5H 2 O) in Laguna Diamante [12] or pure aragonite in Laguna Socompa [10,65].
Compared to low-altitude marine microbialite systems like Cuatro Ciénegas (Mexico), Shark Bay (Australia), or Highborne Cay (Bahamas), the majority of the microbialites present in the Central Andes have a low abundance (ca. <1-4%) of Cyanobacteria, possibly due to their high sensitivity to environmental stresses such as UV, high salt, and arsenic (As) [67,102]. The oxygenic photoautotrophic species of this phylum play a fundamental role in inorganic carbon fixation and biologically induced and biologically influenced mineralization (organomineralization sensu lato) in the low-altitude microbialite systems, but in the Central Andes microbial systems other groups different from Cyanobacteria may largely contribute to these processes [67]. Another characteristic of the Central Andes microbialites, absent in the low-altitude marine microbialites, is the presence of Deinococcus-Thermus species in the upper oxic zones (ca. 0-7 mm) [9][10][11]67]. In Laguna Socompa stromatolites, this phylum even dominates (ca. 35-87%) the first two layers (0-2 mm). The high radioresistance of these species provides them a selective advantage to outcompete other microorganisms in the high-UV irradiance environment of the Central Andes microbialites, while at the same time, protects the rest of the microbial community from UV radiation.
The arrangement of these taxa in a vertically organized structure is mainly determined by steep vertical gradients of light, UV radiation, O 2 , H 2 S, and pH [103]. For instance, in Brava and Tebenquiche soft mats [11,13], radioresistance bacteria (Deinococcus-Thermus), and oxygenic photoautotrophs (Cyanobacteria) are mostly present in the upper layers (ca. 0-1 cm depth), followed by anoxygenic green photoautotrophs (Chloroflexi) mainly located in the intermediate ones (ca. 1-3 cm depth). As Cyanobacteria and Chloroflexi fix inorganic carbon, they share their location with aerobic heterotrophs that consume their organic carbon exudates (Euryarchaeota, Bacteroidetes, Planctomycetes, Verrucomicrobia, and Spirochaetes, among others). Methanogens (Euryarchaeota), sulfate reducers (Crenarchaeota), fermenters (Firmicutes), and anaerobic heterotrophs are generally present in the lower layers (ca. 1-5 cm depth); however, these anaerobes have been reported close to the surface as well (0-1 cm depth). The mechanism(s) that allows them to survive under high oxygen conditions are still unknown [11,13].

Alternative Carbon Fixation Pathways
Compared to other microbial ecosystems located at low altitudes, the AMEs contain a low abundance of Cyanobacteria (ca. <1-4%) which fix carbon dioxide through the Calvin-Benson cycle. This observation is supported, on the one hand, by the low amounts of Chla (ca. <0.1-3 μg/L) reported in Laguna Socompa stromatolites, Laguna Tebenquiche mats, Laguna Brava mats and microbialites, and Laguna Negra mats [9-11, 13, 100], and on the other hand, by the metagenomic analyses of both 16S rRNA gene amplicons and wholegenome sequencing (WGS) data sets [3]. Therefore, in the AMEs, carbon fixation might occur not only by the Calvin-Benson cycle but also through alternative cycles/pathways. To determine which carbon fixation cycle/pathway dominates each AME, metagenomic analyses of Laguna Socompa stromatolites, Laguna Diamante red biofilms, Laguna Llamará endoevaporites, and Laguna Brava and Tebenquiche mats were carried out. In these analyses, the abundances of key enzyme genes from each carbon fixation cycle/pathway were determined. These key enzymes were the ATP-citrate lyase from the rTCA cycle, the RuBisCO from the Calvin-Benson cycle, the bifunctional enzyme carbon monoxide dehydrogenase/acetyl-CoA synthase from the reductive acetyl-CoA pathway (Wood-Ljungdahl pathway), the 4hydroxybutyryl dehydratase from the dicarboxylatehydroxybutyrate (DC/HB) and hydroxypropionatehydroxybutyrate (HP/HB) cycles, and the 2-methylfumaryl-CoA isomerase from the 3-hydroxypropionate bicycle (3HP) [104]. The metagenomic analyses revealed that, in Laguna Socompa stromatolites, the dominant carbon fixation cycle/ pathway is the rTCA cycle, followed by the Calvin-Benson cycle and the Wood-Ljungdahl pathway [68]. In Laguna Diamante red biofilm, the rTCA cycle was the only one reported [12]. In Laguna Llamará endoevaporites, the oxygenrich upper part is dominated by the Calvin-Benson cycle, whereas the anoxic lower part is dominated by the Wood-Ljungdahl pathway [95]. Finally, in Laguna Brava and Tebenquiche mats, the Calvin-Benson cycle and the Wood-Ljungdahl pathway were reported in similar proportions [105]. All these results support the idea that carbon fixation in the AMEs might not be carried out only by the Calvin-Benson cycle, but also through alternative cycles/pathways such as the rTCA cycle and the Wood-Ljungdahl pathway.

Diatoms: Key Structural Components of AMEs
Optical microscopy and scanning electron microscopy (SEM) images of the AMEs revealed the presence of prokaryotic cells (filamentous cyanobacteria, coccoids and/or bacillus) associated with diatom frustules forming nano-globular carbonate aggregates [101]. Diatoms produce great amounts of EPS that trap and bind detrital sediments and/or provide a matrix for mineral nucleation [76,106]. Therefore, they could be promoting carbonate precipitation in AMEs [101]. As diatoms were reported in most of the studied AMEs, and they are photoautotrophic microorganisms, they might play an important role in the primary production of these ecosystems. Moreover, they are the main source of food for the species of flamingos in the Central Andes [107,108], representing a key component of the trophic network.
The ecological role of microscopic eukaryotic taxa in AMEs, and their participation (directly and/or indirectly) in the carbonate precipitation processes are not well understood. The main limitations to answer these interrogations are the lack of data related to the diversity and abundance of these taxa, as well as the seasonal variation of the physicochemical parameters of the habitat. Therefore, further physicochemical and eukaryotic composition analyses should be performed to have a better comprehension of these processes.

Archaea: an Abundant Domain in AMEs
Molecular diversity analyses based on the small subunit ribosomal (SSU) rRNA of Archaea have revealed that this domain is well-represented in the AMEs, especially those with high salinity content (ca. 100-200 mS/cm) such as Salar de Uyuni brines (ca. 40-100%) [49,50], Laguna Tebenquiche mats (ca. 50%) and endoevaporites (ca. 97%) [13], Laguna Diamante red biofilms (ca. 94%) [12], and Laguna Brava soft mats (ca. 43%) and microbialites (ca. 16-17%) [11]. The most abundant archaeal phyla in these AMEs are Euryarchaeota (ca. 16-100%) and Crenarchaeota (ca. 14-17%). Within Euryarchaeota, the most representative genera are Although the Archaea domain is also present in low-altitude marine microbialite systems like Shark Bay, Highborne Cay and Cuatro Ciénegas [114], its abundance is lower (ca. <2-10%) than in the AMEs. The high salt concentrations present in the AMEs are absent in the low-altitude marine microbialite systems. These concentrations are required by halophilic archaea (Halobacteria) to grow, which could explain the differences observed in the archaeal composition of these systems.

Arsenic Metabolism
In the HAALs, high As concentrations (ca. 4.10 −4 -354 mg/L) are naturally found in the water (Table 1) [68]. Prokaryotes that colonize the AMEs have specific metabolic pathways that allow them to resist these elevated As concentrations [68,73,96,[115][116][117], and in some cases, obtain energy from the oxidation of arsenite (As 3+ ) and/or reduction of arsenate (As 5+ ) [100]. The general mechanism of As resistance, which is present in Exiguobacterium sp. S17 isolated from Laguna Socompa stromatolites, consists in the cytoplasmic arsenate reduction (ars operon) followed by the extrusion of the resulting arsenite (acr3 or arsB) to the periplasm [96,115]. Depending on the organism, the arsenite in the periplasm could be used later for the generation of metabolic energy through an aerobic arsenite oxidase (aioAB) and/or a respiratory arsenate reductase (arrAB) [68]. This bioenergetic mechanism was proposed to be carried out by Halorubrum sp. AD156 and Halorubrum sp. DM2, isolated from Laguna Diamante red biofilms [12,116].
To determine how extended are these As resistance and bioenergetic mechanisms in the prokaryotes that colonize the AMEs, different studies were performed in Laguna Diamante ([As]: 347 mg/L) [12,121], Laguna Socompa ([As]: 28 mg/L) [68], and Laguna Brava ([As]: 20 mg/L) [100,122,123], all of which present high arsenic concentrations. Metagenomic analysis of Laguna Diamante red biofilm revealed a high abundance of genes used for arsenite oxidation (aioBA) and respiratory arsenate reduction (arrCBA), suggesting that the haloarchaea that dominate the biofilm use arsenic compounds as bioenergetics substrates. Phylogenetic analysis of these sequences even suggested that the origin of arsenic metabolism in haloarchaea is ancient [12]. Metagenomic analysis of Laguna Socompa stromatolites also revealed a surprisingly diverse metabolism comprising all known types of As resistance and energy-generating pathways. Although the ars operon was the main mechanism identified, an important abundance of arsM genes, which encode an arsenite methyltransferase, was observed in Bacteroidetes, Actinobacteria, Firmicutes, Verrucomicrobia, Spirochaetes, Cyanobacteria, and Euryarchaeota phyla [68]. In Laguna Brava mats, no metagenomic analysis was carried out to study the As metabolism, but an arsenic-sulfur (As-S) cycle was suggested [100] based on the following observations: (i) the high As and S concentrations in the water column [11], (ii) the lack of O 2 in the mats, (iii) the reduction of S and As compounds through anoxygenic photosynthesis [124], and (iv) the heterogeneous distribution of arsenate and arsenite [100]. All these findings suggest that an As biogeochemical cycle is present in the AMEs, probably similar to the proposed one in ancient microbial ecosystems [125,126].

Strains Isolated from AMEs Provide Insight into Arsenic and UV-B Radiation Resistance
Isolation and characterization of microbes from AMEs have provided knowledge on how these extremophiles deal with the diverse environmental conditions present in the Central Andes [127,128], such as high arsenic concentration and UV-B (280-320 nm) radiation.
To study the arsenic resistance mechanisms, Exiguobacterium sp. S17 was isolated from Laguna Socompa stromatolites [96,115]. This strain tolerates an arsenic concentration (10 mM As 3+ ) 20 times higher than the maximum concentration (0.5 mM As 3+ ) tolerated by other Exiguobacterium strains isolated from an estuarine system in Goa, India, with high arsenic content. The analysis of its genome revealed that the S17 strain presents copies of both arsB and acr3 genes, which encode arsenite efflux pumps that export As 3+ to the periplasm. The combined work of both efflux pumps might explain the enhanced tolerance of this strain to As 3+ [115]. A proteomic analysis, carried out on the same strain, revealed that proteins involved in cellular stress responses were overexpressed under the presence of As (e.g., superoxide dismutase, heat-shock proteins, prolyl-tRNA synthetase, elongation factor TS, among others) [96]. Therefore, such proteins have also been proposed to be involved in the arsenic resistance mechanism of the S17 strain.
To study the molecular mechanisms involved in the resistance to high-UV-B radiation, Salinivibrio sp. S10B, Salinivibrio sp. S34, and Salinivibrio socompiensis S35 were isolated from Laguna Socompa stromatolites [129,130]. These strains have shown to be extremely resistant to UV-B radiation (19 KJ/m 2 , corresponding to 240 min of exposure). The analyses of their genomes revealed several genes with a potential role in DNA repair. These include the complete set of genes for RecBCD helicase/nuclease and UvrABC endonuclease holoenzymes, homologs for recA and recX genes, and gene homologs of CPD photolyases, proteorhodopsins, and xantorhodopsins. Photolyases are enzymes that repair cyclobutyl pyrimidine dimers, which are formed between adjacent bases on the same DNA strand upon exposure to UV radiation. Proteorhodopsins and xantorhodopsins enhance the harvest of solar energy, allowing photosynthesis to occur even in periods where there is a lack of solar energy [127,131,132]. Halorubrum sp. BOL3-1, isolated from Salar de Uyuni, also contains gene homologs for photolyases and bacteriorhodopsins in its genome [133], suggesting that these genes belong to a UV-B radiation resistance mechanism shared among the Andean extremophiles.

Final Considerations
During the past decade, several explorations have been carried out to identify and study the AMEs located in the Central Andes region, which thrive under extreme environmental conditions resembling those of the early Earth. The analyses performed to these AMEs have revealed that (1) they display diverse mesostructures and mineralogical compositions; (2) they contain low abundance of Cyanobacteria (ca. <1-4%), and therefore, carbon fixation might occur not only by the Calvin-Benson cycle but also through alternative cycles/ pathways (e.g., rTCA cycle and Wood-Ljungdahl pathway); (3) most contain high abundance of Archaea (Halobacteria) probably due to their high salinity content; and (4) they harbor prokaryotes which present high arsenic and UV-B radiation resistance mechanisms that allow them to grow under the extreme conditions of the environment.
Nevertheless, to have a complete understanding of the AMEs, further analyses need to be done to determine the abundance and diversity of viruses; their impact in the communities of the AMEs; the ecological role of microscopic eukaryotes like fungi, protists, and microinvertebrates and their participation (directly and/or indirectly) in the mineral precipitation processes. Although some analyses have been performed related to these interrogations [2,51,[134][135][136][137], there are no clear answers to them.
The several studies performed during the last years have been of great importance to declare as protected areas Laguna Socompa, Ojos de Mar de Tolar Grande, and Laguna Tebenquiche, which demonstrate that scientific research and exploration are essential for the protection and conservation of these unique ecosystems.