INTRODUCTION

Evaporitic environments can be dated as early as the Archean (3.8-2.5 Ga) (^{Marty et al., 2018}). These depositional settings are transitional and have varied over time (^{Brigmon et al., 2008}). A lot of ions are present in these systems because more water evaporates than falls as rain, creating a wide range of hypersaline conditions (^{Hollister et al., 2010}). Evaporitic systems occur worldwide in marine, lacustrine, tidal flats, solar salterns, brines, salt mines, and salty ponds (^{Eriksson et al., 2007}). They are inhabited by moderate and extreme halophiles that include a great variety of metabolisms (heterotrophs, phototrophs, lithotrophs, and methanogens) from all domains of life (^{DasSarma & DasSarma, 2017}).

Microbes are ubiquitous in nature and harbor diverse metabolic capabilities that allow them to inhabit a wide range of terrestrial and aquatic environments (^{Paerl & Pinckney, 1996}; ^{Ley et al., 2006}; ^{Brenner & Arnold, 2008}). Microbes are the main components of extremophile environments characterized by high salinity, radiation, extreme temperatures, osmotic stress, or nutrient limitation, including thermal springs, hydrothermal vents, desert soils, and hypersaline evaporitic basins (^{Li et al., 2014}; ^{Cowan et al., 2015}). Halophilic microorganisms have characteristic metabolic strategies, including the capacity to pump salt from the environment and accumulate them in the cytoplasm and inversely, thus controlling intracellular conditions (^{Ning et al., 2019}). In hypersaline environments, only a few groups of Archaea are dominant, while at intermediate saline levels, heterotrophic activity is common (^{Menéndez-Serra et al., 2020}). Hypersaline-inhabited microbes contribute to the biogeochemistry of the system by carrying out metabolisms that include carbon fixation, sulfate reduction, methanogenesis, and nitrogen cycling (^{Kouzuma et al., 2015}; ^{Oren et al., 2016}); hence, there is a growing interest in their study for potential biotechnological applications (^{Oren, 2002}; ^{Eriksson et al., 2007}; ^{Javor, 2012}).

However, little is known about the factors that drive the microbial community structure in these extreme environments. It has been postulated that diversity decreases as salinity increases in aquatic environments (^{Oren, 1999}) and that different adaptive strategies have evolved to deal with osmotic regulation (^{Ning et al., 2019}). Thus, selective pressures must be the main drivers of phylogenetic community structure, leading to a species-sorting scenario where similar habitat selection regimes contain similar microbial assemblages (^{Logares et al., 2018}). Nevertheless, variations in community structure, identified from molecular sequences, could also be explained by stochastic processes such as limited dispersal and ecological drift (^{Zhou & Ning, 2017}).

The Vizcaino biosphere reserve comprises a desertic region in the middle of the Baja California Peninsula in Northwestern Mexico. This geosyncline has been filled with a sequence of continental and marine deposits as a result of eustatic changes. Currently, the climate is dry, with an average rainfall below 175 mm. The area is characterized by the occurrence of several evaporitic systems, including Guerrero Negro, Laguna San Ignacio, and Las Charcas, among others. The microbial mats of the Guerrero Negro saltern have been extensively studied (^{Harris et al., 2013}, ^{García-Maldonado et al., 2015}, ²⁰²³), and recently, the diversity of microbial mats that develop in evaporitic ponds in the Lagoon of San Ignacio (LSI) has been explored, allowing to recover an inverse relationship between diversity and salinity, where each one of the analyzed ponds hosted a specific composition of microbes and also the relevance of methanogenic metabolisms (^{García-Maldonado et al., 2018}). In this study, we aimed to continue the analysis of the microbial composition of evaporitic ponds in the Laguna de San Ignacio, focusing on describing the bacterial and archaeal diversity as well as assessing the factors that determine the community structure in sediments.

MATERIAL AND METHODS

During January 2018, sediment samples were collected in six ponds along an evaporite basin system constituted by a series of independent shallow salty ponds known as “Las Charcas.” These evaporitic salterns are part of an extensive system that adjoins San Ignacio Lagoon and has remained without anthropogenic disturbance (Fig. 1, Table 1). At each pond, three random points were selected, and sediment samples (50 g) were collected from the top 15 cm and pooled to obtain a mixed sample with a sterile plastic spatula. All samples were kept at -20 °C until processed.

Figure 1 Map of the geographic location of the six evaporite ponds analysed in this study in the vicinity of the San Ignacio Lagoon in Baja California Sur, Mexico 

RESULTS

Despite their geographic proximity, the six evaporitic ponds showed different characteristics, specifically in the salinity: ponds 1 and 2 were classified as marine (35.8-45.5 ppt), ponds 3, 5, and 6 were classified as hypersaline (65.7-131.3 ppt), and finally, pond 4 was classified as extreme hypersaline (159.3 ppt). Marine ponds showed a lower amount of organic matter, while the highest amount was found in one hypersaline pond (pond 6) and the extreme hypersaline pond (4). Total nitrogen ranges from 0.05% in marine pond 1 to 0.210% in hypersaline pond 3. All ponds showed low values of nitrites, and the nitrates range widely from very low values of 0.6 mg/kg in marine pond 1 and extreme hypersaline pond 4 up to 176 mg/kg in the hypersaline ponds 3 and 5. Finally, soluble phosphorus showed the lowest value of 0.9 mg/kg in marine pond 1 and the highest in marine pond 2 and hypersaline pond 6 (5.8 mg/kg) (Table 1).

The number of reads recovered for the 16S rRNA amplification of the gene V4 hypervariable region was 631 947, and 230 349 sequences remained after quality filtering. The number of sequences ranged widely from 14 603 (pond 1) to 57 229 (pond 3) (Table 1). The minimum library size was defined at 6, 000 sequences after rarefaction, according to Myrlin. The Shannon index on the rarefaction data set ranged from (H = 5.33, sd = 0.0081) in pond 1 up to (H = 6.23, sd = 0.0104) in pond 6 (Table 2). Similar results were recovered for the non-rarefied dataset (not shown). The Shannon diversity index was positively correlated with electric conductivity (r = 0.8857, P = 0.0333) and organic matter (r = 0.9276, P = 0.0076), hence diversity increased with conductivity.

Bacterial phyla in Las Charcas evaporitic pond sediments were the most abundant, representing over 80% of ASVs in all samples. Desulfobacterota and Myxococcota were the main components in all ponds, ranging from 37.8% in pond 3 up to 57.4% in pond 1, followed by Bacteroidota (Bacteroidetes) (12.6% in pond 1 to 22.7% in pond 5). Other abundant phyla were Firmicutes (Bacillota) (4.9% to 12%), Spirochaetes (Spirochaetota) (2.9% to 6.7%), Mycoplasmatota (Tenericutes) (1.2% to 7%), Cyanobacteria (0.46% to 2.62%), Verrucomicrobiota (1.1% to 7.8%), Chloroflexi (Chloroflexota) (0.5% to 5%), and Woesearcheota (1.5% to 3.9%) (Fig. 2).

Figure 2 Barplot of the most abundant bacterial and archaean phyla present in the six evaporite ponds analysed in this study 

The most abundant genera belong mainly to the sulfate-reducing, fermentative Desulfobacterota, followed by Myxococcota and Campylobacterota. Desulfotignum ranges in abundance from 2.9% to 6.9% and is present in all sites. Other abundant genera varied between ponds: Desulfopila (8.7%) and the Fe-reducing bacteria Desulfuromusa (5.6%), pond 1; Spirochaeta (4.1%) and Marinobacter (3.8%), pond 2; sulfate-reducer Desulfocella (6.4%), pond 4; Spirochaeta_2 (4.9%), Desulfomicrobium (4%), pond 3; and Desulfotignum (6.9% and 5.3%), ponds 5 and 6, respectively (Fig. 3). The proportion of rare genera (abundance lower than 1%) was high, ranging from 84.9% in pond 1 up to 91% in pond 2 (Fig 3), and the proportion of very rare genera (abundance lower than 0.01%) ranged from 49% in pond 1 to 62.5% in pond 2 (Fig. S1).

Figure 3 Barplot of the most abundant bacterial genera present in the six evaporite ponds analysed in this study. 

The percentage of variance explained by the first two axes in the beta diversity analysis ranged from 43% in the unifrac unweighted distance (S2 Fig. B) up to 76.5% for the unifrac weighted distance (Fig. S2 item C), where the beta distance ranges from 0.1217 between pond 4 and pond 6 for the unifrac weighted distance to 0.9100 between ponds 1 and 3 for the Bray Curtis index (Table S1). The first two axes of the CCA analyses explained 43.5% of the variance, where ponds 4 and 6 are related to the amount of organic matter and conductivity, while ponds 2 and 3 are related to the amount of total nitrogen and nitrates (Fig. 4). The Mantel test showed that the differences in community composition among ponds are not related to geochemical parameters or geographic distance (P > 0.05).

Figure 4 Canonical correspondence analysis (CCA) ordination biplot of the six evaporite ponds analysed in this study showing their relationships with the five explanatory biogeochemical variables 

The nearest taxon index (NTI) values were zero in all the analyzed samples, on the other hand, the nearest related index (NRI) yielded negative results in ponds 1 and 5, suggesting the occurrence of phylogenetic overdispersion. In the rest of the ponds, the NRI statistic yielded positive values indicating phylogenetic clustering (Table 3). Further, the Raup-Crick metric suggests that homogeneous selection is the principal factor that determines the differences in community structure among ponds (33.33%), followed by homogeneous dispersal (26.66%), whereas dispersal limitation and ecological drift contribute 20%, with a null contribution of variable selection (Table 4). Finally, the model predicts that most of the OTU abundances in the evaporitic pond communities are explained by a neutral model (93.46%); meanwhile, only a small proportion (3.70%) showed an observed frequency above expected, including bacteria involved in the sulfur cycle such as the phylum Desulfobacterota and archaeans of the phyla Bathyarchaeota and Woesearhcheota that are positively selected in the communities (Fig. S3, Table S2).

DISCUSSION

The research presented here complements previous studies that have characterized the microbial diversity and biogeochemical contributions to evaporitic environments in the LSI Baja California, Mexico. Evaporitic ponds are shown to be similar in their biogeochemical roles yet distinct in their compositions, suggesting a restricted connectivity between ponds that act as independent evaporitic entities. The role of sulfur-reducing bacteria could be suggested as a key component of all sediments analyzed, opening our understanding of evaporitic ecosystems, where microbial mats harbor a different composition, apparently in charge of C-transformations, particularly methane (^{García-Maldonado et al. 2018}), whereas sediments mainly contribute to S-cycling (Fig. S4). We could highlight the dominance and positive selection of both bacteria and archaeans involved in the sulfur cycle (Table S2). For example, the Desulfobacterota (previously Deltaproteobacteria) (^{Liu & Häggblom, 2018}) are involved in sulfate reduction and have been proposed as the main mineralization process in a great diversity of environments, as has been evidenced here in evaporation ponds, including genus such as Desulfotignum, Desulfopila and Desulfocella (Table S3), which are sulfate-reducers that have been characterized in intertidal permeable sediments that serve as sites of accumulation of organic matter, harboring a great microbial diversity (^{Brandt et al., 1999}; ^{Gittel et al., 2010}). On the other hand, the more abundant Archaeans belong to the phyla Woesearcheota (Fig. 2) and are dominant in anoxic environments, sediments, and sulfuric springs (^{Zhou et al., 2018}; ^{Liu et al., 2021b}). It has been documented that several factors influence the bacterial communities in sediments, including salinity, organic matter, phosphorus, and nitrogen content (^{Ikenaga et al. 2010}), as well as geographic distance (^{Finstad et al., 2017}; ^{Violle et al., 2017}); nevertheless, despite all sampled ponds analyzed in this study showed differences in composition and abundance (Fig 2), none of the biogeochemical parameters analyzed or the geographic distance were associated with the observed differences (P > 0.05 according with Mantel test). For the aforementioned, we can postulate that the bacterial community ensembles here analyzed are shaped by a series of deterministic ecological factors such as microbial traits, biological interactions, and environmental conditions (e.g., environmental filtering) that are usually more important in environments with steep gradients, and stochastic processes include colonization, extinction, and speciation that are prevalent in more homogeneous environments (^{Huang et al., 2022}). It is difficult to fully associate the differences in diversity patterns recovered among ponds with a single variable due to the fact that all samples only represent a snapshot of the system, and we were not able to evaluate changes through time and the effect of potential perturbations. However, in other studies performed in evaporite ponds, it has been found that repeated samples through time recovered similar patterns despite the fact that most of the diversity proceeds from the rare biosphere (^{Bowen et al., 2012}), allowing us to suggest selective pressures as an important driver of community ensembles in these systems, as has been suggested by the small proportion of taxa that showed signals of being positively selected according to the Sloan neutral model (Fig. S3, Table S2). On the other hand, the community ensembles analyzed in this study could be mainly determined by stochastic processes such as ecological drift, considering that these evaporitic environments do not share a hydrological flow that would be the main agent for homogenization. Thus, ecological drift could be an important stochastic factor that drives changes in the rare biosphere of a microbiome, particularly in the transient rare taxa (TRT), which refers to the taxa that disappeared from one sample to another in opposition to the permanent rare taxa (PRT), which refers to taxa that disappear across biogeographic and seasonal patterns that are driven by deterministic processes (^{Jia et al., 2018}). The role of the rare biosphere has been widely discussed; some authors suggest that members of the rare biosphere represent a genomic reservoir and that these rare taxa could become dominant after a drastic change in environmental conditions (^{Shade et al., 2014}); on the other hand, other studies have revealed their importance in the biogeochemical cycles in terrestrial and marine environments (Martiny et al., 2006). Nevertheless, there is a lack of consensus regarding the threshold abundance value to consider a taxon to be rare. For example, ^{Medina-Chávez and Travisano (2022)} recently split taxa into rare (relative abundance lower than 0.1%) and strictly rare (relative abundance lower than 0.01%); thus, according to their proposal, in our evaporite basins, the proportion of rare taxa was higher than 96% and the proportion of strictly rare taxa was higher than 52%; nevertheless, contrary to what would be expected, the proportion of rare taxa does not increase to their proposal, in our evaporite basins, the proportion of rare taxa was higher than 96% and the proportion of strictly rare taxa was higher than 52%; nevertheless, contrary to what would be expected, the proportion of rare taxa does not increase with the total number of OTUs (^{Locey & Lennon, 2016}), leading to the suggestion that the term rarity must be used beyond an abundance value. Thus, the term rarity must include all the facets that a taxon must match to be considered a real rarity beyond abundance, such as geographic range, habitat specificity, and functional traits (^{Uritskiy et al., 2020}). The diversity levels found in this study (Shannon index) varied widely among the different ponds. One of the possible explanations relies on the use of rarefaction datasets (^{McMurdie & Holmes, 2014}; Willis, 2019). In this study, we used a recently published approach that used repetitive rarefying to avoid the concerns associated with the use of random or omission data (^{Callahan et al., 2016}). However, recently, ^{Schmidt et al. (2022)} warned about the negative effect of repeated rarefaction since this leads to a decrease in the Shannon index of larger libraries and wrongly suggests that the diversity levels are similar among different library sizes. Results from this study showed that the highest range in the Shannon index corresponds to the ponds with the lower number of sequences and that the value decreases as several sequences increase (Table 1 and Table 2), supporting the idea that repeated rarefaction estimators must be used with caution. However, the estimated intervals did not show overlap, and the estimated values were similar to those recovered from analysis without rarefaction (data not shown), so we can conclude that the differences in diversity are real and not an artifact of the procedure. Diversity values found in this study showed a strong correlation with salinity levels (Table 1), a result that contrasts with previous papers that suggest that species richness must decrease with increasing salinity (^{Casamayor et al., 2000}; ^{Benlloch et al., 2002}; ^{Baati et al., 2008}), but is similar to other published studies that found an increase in diversity with salinity levels (^{Yang et al., 2016}, ^{Menéndez-Serra et al. 2021}). For example, ^{Dillon et al. (2013)} found that archaeal diversity is higher than bacterial diversity as salinity increases; however, our results found that the proportion of archaea was low, ranging from 1.65% to 4.53% without a pattern related to salinity levels. Thus, further investigation of the microbial diversity in other biomes within the evaporitic ponds, e.g., water and sedimentary columns, as well as their temporal variation, would be necessary to fully understanding the dynamic of these ecosystems.

CONCLUSIONS

The hypersaline aquatic system analyzed here is made up of contiguous evaporation ponds, although diversity was associated with salinity, other factors remained insignificant in explaining the patterns of pond-specific diversity. Fort the aforementioned, we conclude that the main processes maintaining species coexistence are homogeneous selection and ecological drift due to the high percentage of rare biospheres; thus, low-abundance taxa are more prone to experiencing random fluctuations in their abundance, leading to highly differentiated communities over relatively short distances with no pattern related to the geographic distance. Finally, the study of salty ponds provides unique habitats for organisms, which can provide information on adaptive strategies and a better understanding of the diversity of life, as well as their potential use for biotechnological applications in the future.