Little Rann of Kutch is roughly triangular in shape, located between 22° 55’’ to 24° 35’’ North latitudes and 70° 30’’ to 71° 45’’ East longitudes near the Great Rann of Kutch, Gujarat, with an average annual rainfall of below 400 mm (Ishnava et al., 2011). In summer, temperatures reach highs of 48°C and lows of 10°C during the winter, while the average temperature is 30-35°C (Datta et al., 2020). Geologically, this area was a part of oceanic floor and has emerged in the recent past. The Rann of Kutch was a gulf of the sea with surrounding coastal towns (Frere, 1870). A fairly low rain fall coupled with evaporation led to the receding of water, subsequently leaving behind a crust of halite and gypsum crystals that converted into the clay and sands.

Soil samples were collected from three different study sites of the LRK: Surajbari (N 23° 11’ 16.373, E 070° 43’ 03.929”), Venasar (N 23° 09’ 57.327”, E 070° 55’ 14.575”), and Jogad (N 23° 10’ 38.992”, E 071° 15’ 04.999”) ( Figure 1 ). A total of three subsamples were collected from each site at distances of 100 - 200 m. The collected samples were designated as SB, VS, and JO for Surajbari, Venasar, and Jogad, respectively. One hundred grams of three subsamples were collected from each site as far as possible from plants to avoid rhizosphere effects and at a depth of 10 to 15 cm in order to obtain a sample minimally impacted by aeolian dispersion. Samples were collected into sterile polythene bags, transported to the laboratory, and stored at 4°C until further analysis.

The physicochemical properties of the soil samples were analyzed. The soil pH and salinity were measured in a 1:5 (wt/wt) aqueous solution, while the conductivity was measured in dS/m by digital conductivity meter (CON700, Eutech instruments, Singapore). Estimation of the oxidizable organic carbon was based on the Walkely and Black Method (Walkley and Black, 1934), while phosphorous (P) was estimated by the extraction with sodium bicarbonate (Olsen et al., 1954). Available nitrogen (N) was measured by alkaline permanganate method (Asija and Subbiah, 1956), while potassium and sodium were measured by the flame photometric method and sulphur by turbidity method (Chesnin and Yien, 1951; Jackson, 1974). Exchangeable calcium and magnesium were estimated using versanate EDTA Method. Total dissolved solid (TDS) of soil was measured as described earlier (Jackson, 1974). Chlorine was estimated by titration method (Jackson, 1974). Cu, Fe, Mn, Zn, and B were estimated using MP-AES (Microwave Plasma Atomic Emission Spectrometry) method (Hettipathirana, 2011).

Two approaches, direct plating and Liquid enrichment technique, were used to isolate bacteria from the desert soil of the LRK. In direct plating, five different media (Modified SP medium consisting, (g/l): (NaCl, 49; KCl, 1; MgSO₄.7H₂O, 0.5; CaCl₂.2H₂O, 0.18; NaHCO_3, 0.03; NaBr, 0.115; FeCl₃.6H₂O, 0.5; Bacto trypton, 5; Yeast extract, 10; Glucose, 1; Agar, 30) (Caton et al., 2004), Reasoner’s 2A (R2A) agar (HiMedia Laboratories, India) consisting, (g/l): (Casein acid hydrolysate, 0.5; yeast extract, 0.5; protease peptone, 0.5; dextrose, 0.5; starch, 0.5; KH₂PO₄, 0.3; MgSO₄, 0.024; Sodium pyruvate, 0.30; NaCl, 50; Agar, 30) (Reasoner and Geldreich, 1985), Soil Extract medium (SE) consisting, (g/l): (Glucose, 1; K₂HPO₄, 0.5; Soil extract, 17.75; NaCl, 50; Agar, 15) (Yadav et al., 2015), Complex medium (CM) agar consisting, (g/l): (Glucose, 10; peptone, 5; yeast extract, 5; KH₂PO₄, 5; NaCl, 150; Agar, 30) (Bhatt et al., 2018) and Halophilic medium (HM) agar consisting, (g/l): [NaCl, 220; MgSO_4.7H₂O, 10; KCl, 5; Sodium citrate, 3; KNO_3, 1; CaCl_2.2H₂O, 0.20, Trace minerals, 0.5; Bacto tryptone, 5; Yeast extract, 1; Agar, 30) (Post, 1977)] were used to capture and isolate the diverse microbiota of the desert soil. Moreover, serial dilution technique was used with two different diluents i.e., 1) Distilled water 2) Winogradsky’s salt solution consisting, (g/l): K₂HPO₄, 0.25; MgSO₄, 0.125; NaCl, 0.125; Fe₂(SO₄)₃, 0.0025; MnSO₄, 0.0025). For bacterial isolation, a composite sample was prepared out of the three subsamples collected from each site. The composite soil sample (1g) was thoroughly mixed and suspended in 9 ml of sterile distilled water and Winogradsky’s salt solution using a Vortex mixer. The samples were then diluted from 10^-1 to 10^-7, and 0.1 ml aliquots from each dilution were spread on the five different media as described above and incubated at 37°C for 7 days. After a week of incubation, the bacterial colonies macroscopically differing in morphology, size, and pigmentation were sub-cultured, isolated, and purified using the respective culture medium. Only the plates containing 30-300 colonies were used to calculate total bacterial counts. The results were expressed as mean colony forming units (cfu) per gram weight of soil. In addition, enrichment technique was used to enrich soil samples using three different media: SP medium, CMB medium, and HM medium. In enrichment technique, after inoculation, enrichment media was incubated on the orbitek shaker at 180 rpm and 37°C and the growth was periodically observed. After 72h of incubation, the cultures were serially diluted up to 10^-7 and 0.1 ml of an appropriately diluted culture was spread over the respective agar plate and incubated at 37°C. After 7 days of the incubation, distinct isolated colonies were selected, and pure cultures obtained.

In order to evaluate the salt tolerance of the bacteria, the isolates were spot-inoculated onto the surface of the respective culture media with varying salt concentrations in the range of 0-25% at pH 8. The pH tolerance was examined at 37°C in respective growth medium used for the isolation and the pH was adjusted to pH 7.0–11.0 with an interval of 1 pH units by using separately autoclaved Na₂CO₃ (20%, w/v) for pH 6–8, 1 M glycine/NaOH buffer for pH 9.0–10.0, and 1M NaOH or 1M HCl for pH 11. The growth was monitored at 24 hr. intervals and considered positive on appearance after 4 days of incubation for fast-growing bacteria and 7 days of incubation for slow-growing archaea at 37°C.

For bacteria, 16S rRNA genes were amplified using forward and reverse primer pairs 27F (5’-AGAGTTTGATCMTGGCTCAG-3’) and 1492R (5’- TACGGYTACCTTGTTACGACTT-3’), respectively, followed by the sequencing of each amplified product after purification. Whereas for archaea, forward and reverse primer pairs 21F (5-TTCCGGTTGATCCYGCCGGA-3) and 1492R (5’-GGTTACCTTGTTACGACTT-3’) were used for the amplification. The amplified 16S rRNA genes were sequenced on the Applied Biosystems Automatic Sequencer (ABI3730XL). Identification of the phylogenetic neighbors and calculation of pairwise 16S rRNA gene sequence similarity was achieved using the EzTaxon-e server (Kim et al., 2012). The CLUSTAL W algorithm of MEGA 6 software was used for sequence alignments and MEGA 6 software for phylogenetic analysis of the individual sequences (Tamura et al., 2013). Distances were calculated using the Kimura correction in a pairwise deletion manner. Maximum Likelihood (ML) method was used to construct phylogenetic tree. Percentage support values were obtained using a bootstrap procedure based on 1000 replications. The 16S rRNA gene sequences of 87 isolates are deposited in the NCBI GenBank (Accession No. MF321815- MF321853, MK785115 - MK785132, MK779774 - MK779775, MK779747, MK779714, MK779742, MK779796, MK779859 - MK779864, KT008286, and KT008288 - KT008300).

The influence of soil chemical properties on the microbial diversity was determined by the CCA analysis using PAST v3.02 software (Hammer et al., 2001). Before the multivariate analyses, to use the same units for all the environmental variables, the values of all variables were normalized by subtracting the mean of the raw data and dividing by the standard deviation of the raw data to better conform to normality (Pagaling et al., 2009). Ordination triplots were used to represent the effect of environmental variables on the bacterial community structure. The environmental factors were represented as lines. The CCA plots were generated using the PAST v3.02 software (Hammer et al., 2001). Cluster analysis based on genera distribution among sampling sites was carried out using Bray-curtis similarity measure and UPGMA algorithm.

The Good’s non-parametric coverage estimator was calculated according to the equation C = 1 - (n1/N), where n1 is the number of phylotypes for which only one isolate was recovered, and N is the total number of isolates (Good, 1953). The Shannon biodiversity index and Margalef index were also calculated (Magurran, 1988).

In order to assess the geographic distribution of the phylotypes, 16S rRNA gene sequences were compared with the sequences of the cultured strains as well as environmental sequences (from metagenomics and high-throughput sequencing) available in public databases (EMBL and NCBI) using BLAST (Blast Local Alignment Search Tool) (Altschul et al., 1990). In BLAST analysis, only high scoring entries with ≥ 98.7% sequence similarity were considered to obtain geographical data as per the species delineation criteria (Stackebrandt and Ebers, 2006). Data on geographic location of high scoring hits (≥ 98.7%) was obtained, such as isolation source and environment from where strains were isolated, in order to assess any concordance in biome or habitat type. Based on the geographical location data of the high scoring entries (≥ 98.7%), the phylotypes were labeled as: 1) LRK (If no high scoring sequences from non-LRK origin), 2) Marine (Only high scoring sequences from any marine environment – For instance, Sea water/Sea sediments), 3) Saline (At least one high scoring sequence from inland saline habitat/athalassohaline habitat - For instance, saline lake or saline desert), or 4) Cosmopolitan (Phylotypes having similarity with at least one high scoring sequence from non-LRK/non-marine/non-saline environment - For instance, garden soil/fresh water/rhizosphere etc.).

Soil samples from all the three study sites have shown considerable variation with respect to their physicochemical properties. However, many properties of Surajbari and Venasar were similar. The detailed physicochemical properties for the three soil samples are shown in Table 1 .

The viable counts in the Surajbari site varied from 1 ± 0.15×10² cfu/g of soil in CMB medium to 2 ± 0.35×10⁷ cfu/g of soil in SP medium. In the Venasar site, viable counts varied from 2.6 ± 0.52×10³ cfu/g of soil in SE medium to 2.5 ± 0.43×10⁶ cfu/g of soil in R2A medium. Whereas, in the Jogad site, viable counts varied from 6 ± 0.51×10² cfu/g of soil in CMB medium to 1.3 ± 0.37×10⁶ cfu/g of soil in SP medium. Overall, the cfu values were high in SP and R2A medium compared to other media used in this study. Moreover, cfu count was higher when Winogradsky’s salt solution was used compared to distilled water as diluents ( Table 2 ).

Cultivation strategies yielded 87 isolates from three sampling sites at the Little Rann of Kutch ( Table 3 ). Out of all isolates, 82.75% (n=72) were obtained by direct plating technique, while the remaining 17.24% (n=15) by enrichment technique. In direct plating approach, maximum isolates were obtained on SP medium (43.05%) followed by R2A medium (30.55%), SE medium (11.11%), Complex medium (8.33%), and Halophilic medium (6.94%). In addition, 65.27% of the isolates were obtained when distilled water was used as diluent, while 34.72% isolates were obtained with Winogradsky’s salt solution as diluent. In enrichment technique, out of 15 isolates, nearly one third were obtained on SP medium, Halophilic medium, and Complex medium.

The bacterial isolates can be diversified based on their growth behavior in different salt concentrations. Overall, the isolates displayed a broad range of salt tolerance. Among the isolates, a majority (∼98%) could grow at 7.5% NaCl, ∼75% at 12% NaCl, and ∼36% at 25% NaCl. In the Surajbari site, all the isolates grew in 3-7.5% NaCl, whereas in Venasar 7.5% NaCl supported growth of all the isolates ( Figure 2 ). While∼ 93% of the Jogad isolates grew in the range of 3-7.5% NaCl. Further, ∼ 73% of the Venasar isolates displayed a broader salt tolerance of 0-25% NaCl ( Figure 2 ). Isolates growing without salt were maximally represented in Jogad site followed by Surajbari and Venasar. Moreover, the number of isolates growing at 15% NaCl were maximum in Venasar (∼73%) followed by Surajbari (∼50%) and Jogad (∼26%). Isolates SB-7, SB-27, SB-28, and SB-29 of Surajbari, VS-7, VS-27, VS-30, VS-34, and VS-36 of Venasar, and JO-30, JO-39, and JO-44 of Jogad could not grow without salt, indicating their true halophilic nature.

In the present study, a majority of the bacterial isolates grew over a wide range of pH 7-10. Out of 87 isolates, a majority (>90%) could grow at pH 8-9 followed by pH 7, 10, and 11. All the isolates of the Surajbari site grew at pH 8-9, while the isolates of Venasar displayed growth at pH 7 and 8 ( Figure 3 ). On the other hand, the majority of the Jogad isolates were able to grow in the pH range of 7-10. Interestingly, ∼93% of the Venasar isolates displayed a broader salt range for growth at pH 7-10 ( Figure 3 ). The isolates growing at neutral pH were largely from Venasar study site followed by Surajbari and Jogad. While the maximum number of the isolates growing at pH 11 belonged to Venasar (∼73%) followed by Jogad (∼45%) and Surajbari (∼42%) ( Figure 3 ). Isolates SB-2 of Surajbari and JO-23, JO-35, and JO-39 of Jogad did not grow at pH 7, indicating their true alkaliphilic nature.

The isolates identified and analyzed on the basis of 16S rRNA gene sequences were classified in the kingdom bacteria and archaea. The EZ taxon server analysis suggested homology of the isolates in the range of 96.6-100% with the reference strains of the NCBI database. The isolates were distributed in four phyla, namely Firmicutes, Proteobacteria, Actinobacteria, and Euryarchaeota, belonging to 19 different genera: Bacillus, Virgibacillus, Halomonas, Gracilibacillus, Thalassobacillus, Piscibacillus, Halobacillus, Oceanobacillus, Salimicrobium, Alkalibacillus, Bhargavaea, Proteus, Marinobacter, Kocuria, Corynebacterium, Planococcus, Pseudomonas, Micrococcus, and Natronococcus. Overall, 90.81% isolates were Gram positive while the rest displayed a Gram-negative character. The relatively abundant genera were Bacillus (33 isolates: 37.9%), Virgibacillus (21 isolates: 24.1%), Halomonas (5 isolate: 5.74%), Bhargavaea (3 isolate: 3.44%), Gracillibacillus (3 isolate: 3.44%), Kocuria (2 isolate: 2.29%), Corynebacterium (2 isolate: 2.29%), Piscibacillus (2 isolate: 2.29%), Oceanobacillus (2 isolate: 2.29%), Thalassobacillus (2 isolate: 2.29%), Halobacillus (2 isolate: 2.29%), Alkalibacillus (2 isolate: 2.29%), Natronococcus (2 isolate: 2.29%), Micrococcus (1 isolate: 1.14%), Salimicrobium (1 isolate: 1.14%), Pseudomonas (1 isolate: 1.14%), Marinobacter (1 isolate: 1.14%), Proteus (1 isolate: 1.14%), and Planococcus (1 isolate: 1.14%).

Based on the phylogenetic analysis, a majority of the isolates belong to the phylum Firmicutes, which included Low G+C Gram-positive bacteria of different families. Nine genera (Bacillus, Virgibacillus, Gracillibacillus, Thalassobacillus, Piscibacillus, Alkalibacillus, Halobacillus, Oceanobacillus, and Salimicrobium) belonged to the family Bacillaceae, while Planococcus and Bhargavaea were associated with Planococcaceae ( Figure 4 ). The second most abundant phylum was Proteobacteria, which included gram negative bacteria of various genera; Proteus, Marinobacter, Halomonas, and Pseudomonas belonged to the families Enterobacteriaceae, Alteromonadaceae, Halomonadaceae, and Pseudomonadaceae, respectively ( Figure 5 ).

Five strains, namely JO-4, JO-18, JO-22, JO-29, and SB-5, were affiliated to phylum Actinobacteria, comprising high G+C Gram-positive bacteria of two different families, Corynebacteriaceae and Micrococcaceae, and three different genera, Corynebacterium, Kocuria, and Micrococcus ( Figure 6 ). Strains VS-36 and JO-44 were identified as archaea affiliated to phylum Euryarchaeota, family Halobacteriaceae, and genus Natronococcus.

The isolates of the Surajbari were classified in the kingdom bacteria and phylum Firmicutes and Actinobacteria, which included Gram-positive bacteria related to the families of Bacillaceae, Planococcaceae, and Micrococcaceae. The isolates were distributed among nine different genera with 20 species. The relative abundance of different genera in three study sites is as depicted in Figure 7 .

The isolates of the Venasar study site were classified in the two kingdoms, bacteria and archaea, with phyla from Firmicutes, Proteobacteria, and Euryarchaeota, which overall included Gram-positive bacteria belonging to the families of Bacillaceae and Planococcaceae and Gram-negative bacteria related to families Halomonadaceae, Enterobacteriaceae, and Alteromonadaceae as well as members belonging to the family of Halobacteriaceae. The isolates were distributed among 11 different genera, displaying 17 different species ( Figure 7 ).

Similarly, the isolates of the Jogad study site were classified in two kingdoms, bacteria and archaea, with phylums from Firmicutes, Proteobacteria, Actinobacteria, and Euryarchaeota which overall included Gram-positive bacteria related to the families of Bacillaceae, Planococcaceae, Micrococcaceae, and Corynebacteriaceae and Gram-negative bacteria related to families Halomonadaceae, Pseudomonadaceae, and members of the family Halobacteriaceae. The isolates were represented by 11 different genera associated with 23 different species ( Figure 7 ).

Out of 19 genera isolated from all the three study sites, 5.26% (1 genus: Alkalibacillus) was common among Surajbari and Venasar, while 15.78% (3 genera: Halomonas, Piscibacillus, and Natronococcus) were common among Venasar and Jogad. Similarly, 5.26% were represented by 1 genus, Oceanobacillus, which was common between Surajbari and Jogad. While 21% (4 genera: Bacillus, Virgibacillus, Gracilibacillus, and Bhargavaea) were common in all study sites ( Figure 8 ).

Further, some genera were exclusively confined to a specific study site. Halobacillus, Salimicrobium, and Micrococcus were exclusively identified in Surajbari, while Proteus, Planococcus, and Marinobacter were present only in Venasar. On the other hand, Corynebacterium, Kocuria, Thalassobacillus, and Pseudomonas were exclusively present in Jogad ( Figure 8 ).

Hierarchical clustering using UPGMA method and Bray-Curtis similarity index indicated more than 68% similarity between the three study sites based on the spatial distribution of genera ( Supplementary Figure S1 ). Venasar and Jogad sites displayed less variation in genera distribution across the spatial profile compared to the Surajbari study site.

Six putative novel taxa were identified in the phylum Firmicutes, Actinobacteria, and Proteobacteria ( Figure 9 ). They had 16S rRNA gene sequence identities below conservative threshold (98.7% identity) with their closest relatives for species level identification (Stackebrandt and Ebers, 2006). Four novel taxa (JO-4 and JO-22, JO-18, JO-21, and JO-39) from Jogad, one from Venasar (VS-10), and one from Surajbari (SB-26) each were obtained.

In the phylum Firmicutes, two putative novel genera and two novel species were identified, related to the family Bacillaceae. JO-21 had a low 16S rRNA gene sequence similarity of 96.60% with their closest phylogenetic relative Bacillus hwajinpoensis SW-72^T followed by Bacillus algicola KMM 3737^T (95.92%) and Bacillus hemicentroti JSM 076093^T (95.77%) ( Figure 9A ). Moreover, the strain has a sequence similarity of 94.12% with the type species of genus Bacillus, Bacillus subtilis subsp. subtilis DSM10^T. Strain JO-21 had equal low sequence similarity with the species of neighboring genera Ornithibacillus contaminans CCUG 53201^T (95.53%), Anaerobacillus isosaccharinicus NB2006^T (95.38%), Anaerobacillus alkaliphilus B16-10^T (94.98%), Anaerobacillus alkalidiazotrophicus MS6^T (94.97%), Pradoshia eiseniae EAG3^T (94.84%), Fermentibacillus polygoni IEB3^T (94.83%), and Desertibacillus haloalkaliphilus KJ1-10-99^T (94.77%). As the sequence identities with the most closely related genera were equally low, it may be a new genus and thus further studies are required. Similarly, strain JO-39 had the highest 16S rRNA gene sequence similarity and clustered together with the species of two different genera, Oceanobacillus limi H9B^T (98.22%) and Ornithinibacillus salinisoli LCB256^T (98.11%), and thus could not be assigned to an existing genus ( Figure 9B ). They either belong to one of these genera or alternatively may represent new genera.

Strain VS-10 showed low 16S rRNA sequence similarity with their closest phylogenetic relatives, Bacillus halosaccharovorans E33^T (96.98%), Bacillus endolithicus JC267^T (96.92%), and Bacillus crassostreae JSM 100118^T (96.53%) and thus represents novel species ( Figure 9E ). Strain SB-26 shared equally low 16S rRNA gene sequence similarity with two different species, Halobacillus trueperi (98.7%) and Halobacillus dabanensis (98.7%), thus representing novel species.

In the phylum Actinobacteria, two putative novel species were identified related to genera Corynebacterium and Kocuria. Strains JO-4 and JO-22 have shown 97.7% sequence similarity with their closest phylogenetic relative Corynebacterium auris DSM 328^T (97.4%). However, in the phylogenetic tree, the two novel strains were not only placed far apart from their phylogenetic relatives but also formed a separate branch ( Figure 9D ). Owing to a low 16S rRNA gene sequence similarity and the tree topology, these two strains represent a single novel species. 16S rRNA gene sequence similarity among these two strains was 99.93%, meaning these two strains belong to a single species. Similarly, strain JO-18 showed low sequence similarity of 98.3% with Kocuria sediminis FCS-11^T and 97.93% with Kocuria turfanensis HO-9042^T and therefore represents a potential new species ( Figure 9C ). All these isolates are therefore indicative of new genera/species and should be studied further to establish their classification and taxonomic positions.

Multivariate analyses are used to elucidate the associations between the abundance of certain genera and environmental parameters. In this study, the Canonical Correspondence Analysis (CCA) (Ter Braak and Verdonschot, 1995) was used to analyze the relationships between the microbial communities of LRK as a function of the physicochemical variables of the habitat ( Figure 10 ). Among the physicochemical parameters, pH, Electrical Conductivity (EC), TDS, organic carbon (OC), nitrogen (N), phosphorous (P), potassium (K), sulphur (S), copper (Cu), ferrous (Fe), Mg, Ca, Na, Cl, B, Zn, and Mn were considered.

Overall, organic carbon and nitrogen content were found to be the main environmental variables affecting the microbial diversity ( Figure 10 ). Chu et al. (2016) found that soil organic carbon is the best predictor of microbial community distribution. In this study, organic carbon was reported highest in the Venasar site followed by Surajbari and Jogad. Interestingly, a negative correlation between soil microbial diversity and organic carbon was observed. The results are in line with some previous reports (Yao et al., 2020; Cui et al., 2021). On the contrary, a positive correlation was observed between nitrogen content and microbial diversity indices. Nitrogen content was highest in jogad site followed by surajbari and venasar. The results clearly match with the trend of Shannon-Weaver Index (H’) among three sites.

Several genera, such as Thalassobacillus, Corynebacterium, Kocuria, and Pseudomonas, were most positively correlated with nitrogen followed by manganese and phosphorous, while Pseudomonas and Thalassobacillus showed a negative correlation with the conductivity, sodium, chlorine, calcium, and magnesium. A positive correlation was apparent between the carbon content and the occurrence of the Proteobacteria, i.e., Marinobacter, Proteus, and Halomonas. Similarly, Halomonas, Piscibacillus, and Natronococcus have shown a positive correlation with soil pH. A negative correlation existed with sulphur.

Alkalibacillus shows a positive correlation with the electrical conductivity (EC), TDS, calcium, sodium, and chlorine, while it was negatively correlated with nitrogen and phosphorous. Sulphur content had positive correlation with Oceanobacillus, while this genus was negatively correlated with carbon and pH. None of the physicochemical parameters were significantly correlated with the occurrence of Micrococcus, Halobacillus, or Salimicrobium. However, a marginal positive correlation with EC and negative correlation with pH was evident. The other bacterial genera, such as Bacillus, Virgibacillus, Gracillibacillus, and Bhargavaea, were closely aligned to the center of triplot, reflecting an insignificant correlation between their abundance and soil properties.

On the basis of the Good’s Estimator, 77%, 73.4%, and 77.5% of the cultivable diversity at the genus level was retrieved from the Surajbari, Venasar, and Jogad study sites, respectively. Overall, taking the entire three sites together, 93% of the cultivable diversity was retrieved. Shannon-Weaver Index (H’) was highest in Jogad (2.86) followed by Surajbari (2.79) and Venasar (2.46), while Margalef Index was found to be highest in Jogad (6.115) followed by Surajbari (5.832) and Venasar (4.704).

The phylotypes were clustered into four categories, namely LRK, Marine, Saline, and Cosmopolitan, as per the origin and habitat type ( Tables 3 , 4 ). Overall, among the three study sites, maximum phylotypes (45%) were mostly of a cosmopolitan nature in their distribution followed by Saline (33.3%), LRK (13.3%), and Marine (8.3%) ( Figure 11 ).

Nearly 58% of the Surajbari phylotypes were associated with the cosmopolitan distribution followed by Saline (21%), Marine (16%), and LRK (5%). Similarly, a majority of the Jogad phylotypes (41.6%) had cosmopolitan distribution followed by Saline (29.1%), LRK (25%), and Marine (3.8%). However, among the Venasar phylotypes, 52.9% were of saline type followed by cosmopolitan (35.2%), Marine (5.8%), and LRK (5.8%).

In phylum Firmicutes, a majority of the phylotypes belonging to the genus Bacillus had cosmopolitan distribution ( Table 3 ). However, Bacillus salecetis (JO-7 and VS-19) and Bacillus halosachharovorans (SB-18) had saline distribution while Bacillus paranthracis (SB-6) had marine distribution. Other genera, such as Oceanobacillus and Bhargavaea, were of the cosmopolitan type in their distribution. Further, the phylotypes belonging to the genera Virgibacillus, Planococcus, Gracilibacillus, Piscibacillus, Alkalibacillus, and Thalassobacillus had saline distribution, while Halobacillus and Salimicrobium had marine type distribution. On the other hand, the phylotypes VS-10, JO-21, JO-39, and SB-26 had no significant similarity with any sequences in the public databases and thus are at present known from LRK only.

Within the phylum Actinobacteria, Corynebacterium strains (JO-4 and JO-22) and Kocuria strain (JO-18) showed no significant similarity with any sequences in the public databases and thus are at present known from LRK only. The Phylotypes Micrococcus aloeverae (SB-5) and Kocuria palustris (JO-29) had cosmopolitan and marine distribution patterns, respectively ( Table 3 ). Among the five Proteobacteria phylotypes, Halomonas ventosae (VS-11 and VS-13) and Marinobacter adhaerens (VS-18), Halomonas denitrificans (VS-7), Proteus mirabilis (VS-26), and Pseudomonas veronii (JO-46) had saline, Marine, and cosmopolitan distribution, respectively. Phylotype JO-19 belonging to Halomonas genus showed no significant similarity with any sequences in the public databases and, therefore, is currently associated with the LRK type only ( Table 3 ). In phylum Euryarchaeota, the phylotype Natronococcus jeotgali (VS-36 and JO-44) was associated with the saline distribution.

The phylogenetic analysis revealed four phyla, Firmicutes, Proteobacteria, Actinobacteria, and Euryarchaeota, with 19 different genera. Archaea dominates the microbial community of the hypersaline soil over bacteria (Hacěne et al., 2004). However, in this study, bacteria were dominant over archaea in LRK, a trend also reflected in a previous metagenomic study (Patel et al., 2015). Saline and hypersaline environments are known to possess the dominance of the gram-negative bacteria (Baati et al., 2010; Makhalanyane et al., 2015; Ronca et al., 2015). However, in the present investigation, the gram-positive bacteria were dominant.

Out of 19 genera reported in this study, Bacillus, Halobacillus, Oceanobacillus, and Virgibacillus were earlier reported (Bhatt et al., 2018), while the rest are being reported for the first time from the LRK. The Firmicutes were most dominant followed by the Proteobacteria, Actinobacteria, and Euryarchaeota, a trend also reflected in some earlier reports of the deserts (El Hidri et al., 2013; Li et al., 2017). However, Proteobacteria were reported as dominant in metagenomic studies of the LRK (Patel et al., 2015), the Great Rann of Kutch (Pandit et al., 2015), and Thar desert of India (Sivakala et al., 2018).

In Surajbari, the Firmicutes dominated followed by Actinobacteria, while Proteobacteria and Euryarchaeota were absent. The Firmicutes were also dominant in Venasar followed by Proteobacteria and Euryarchaeota with the absence of Actinobacteria. In Jogad, Firmicutes still dominated followed by Actinobacteria, Proteobacteria, and Euryarchaeota. The Firmicutes and Actinobacteria were reported as dominant phyla in two geographically distinct saline pan sediments of the Kalahari Desert of Southern Africa (Genderjahn et al., 2018). It is well established that spore forming bacteria can withstand harsh environmental conditions. Once the environmental conditions become favorable and water availability enhances, the cells begin to divide (Jones and Lennon, 2010; Crits-Christoph et al., 2013).

Phylogenetic analysis revealed that the identification of certain isolates was not straightforward, primarily due to the equally high sequence similarities with two different species. For instance, strains SB-28, SB-6, VS-5, and JO-46 display equally high sequence similarities with two different species, Halobacillus trueperi (98.3%) and Halobacillus dabanensis (98.3%), Bacillus paranthracis (99.5%) and Bacillus nitratireducens (99.5%), Bacillus vallismortis (99.8%) and Bacillus subtilis subsp. spizizenii (99.8%), and Pseudomonas veronii (99.4%) and Pseudomonas simiae (99.4%), in that order. Similarly, other isolates such as SB-2, JO-11, SB-23, VS-24, JO-17, VS-22, JO-32, and JO-36 shared equally high sequence similarities with two different species. Therefore, either they belong to one of the species or may represent a new species. However, in order to probe further and ascertain their taxonomic status, the polyphasic characterization is required. The bacterial diversity was highest in Jogad followed by Surajbari and Venasar. This was a fact also revealed by the Shannon-Weaver Index (H’) and Margalef Index.

As revealed by the CCA analysis, many physicochemical parameters were positively correlated with the bacterial community. Organic carbon, pH, nitrogen, TDS, EC, and Mn were found to be the main environmental factors significantly affecting the microbial diversity. Organic carbon and nitrogen content were significantly high in all three sites.

In saline soils, the SOC content is influenced by two opposing factors: reduced plant inputs which may decrease SOC, and reduced rates of decomposition (and associated mineralization of organic C to CO2) (Datta et al., 2020). Despite the absence of any plant C input to soil of the salt desert, a considerably high amount of organic carbon might be due to the deposition of early Holocene sediments under estuarine deltaic environment (Datta et al., 2020). Furthermore, higher salt concentration improved the soil structure with good aggregation, which can retain organic carbon for a longer time in absence of decomposition due to lower microbial activity (Datta et al., 2020).

On the other hand, Fe, Zn, P, Cu, and B were not associated with the bacterial diversity ( Figure 10 ). Bacillus, Virgibacillus, Gracilibacillus, and Bhargavaea were commonly present in all three study sites. However, Alkalibacillus was exclusively present in Surajbari and Venasar. CCA analysis suggests that Na⁺, Cl^-, Ca²⁺, Mg²⁺, and TDS displayed positive correlation with Alkalibacillus. High concentrations of these ions in Surajbari and Venasar compared to Jogad further supports this correlation. Similarly, a positive correlation of Alkalibacillus with Na⁺ content in saline and alkaline soil of Cuatro Cienegas, Sayula, and San Marcos lakes was identified (Delgado-García et al., 2018).

Halomonas, Natronococcus, and Piscibacillus were exclusively shared between Venasar and Jogad. A positive correlation with pH might be one of the reasons for the exclusive presence of these two genera in these sites. pH in Venasar and Jogad was more alkaline compared to Surajbari. A negative correlation with EC would have played a role in the absence of Halomonas in Surajbari. Earlier, a positive correlation of pH with Halomonas was reported (Oueriaghli et al., 2014). The genus Oceanobacillus was exclusively present in Surajbari and Jogad due to a positive correlation between this genus and sulphur content ( Figure 10 ).

Surajbari, having close proximity with the Gulf of Kutch, has a regular flow of sea water, subsequently causing higher salinity as compared to other sites located at farther distances. Halobacillus, Salimicrobium, and Micrococcus have a marginal positive correlation with the electrical conductivity (EC) and thus were exclusively identified in Surajbari. Proteus, Planococcus, and Marinobacter were exclusively present in Venasar. A positive correlation between the occurrences of these genera with the organic carbon explains their exclusive presence in Venasar. Earlier, a significant correlation between the organic carbon and Gammaproteobacteria were reported from the freshwater Poyang Lake in China (Ding et al., 2015). Further, the exclusive presence of Corynebacterium, Kocuria, Thalassobacillus, and Pseudomonas in Jogad is explained by a positive correlation between these genera, nitrogen (N), and manganese (Mn) contents of the habitat. Previous reports suggests that proteobacteria have a vital role in degrading sedimentary organic nitrogen (Zhou et al., 2013; Zhang et al., 2015). Similarly, Garcia-Lopez et al. (2019) reported a correlation between Pseudomonas and high levels of nitrogen content. Similarly, previous reports suggest the important role of manganese in sporulation of Bacillus sp. (Charney et al., 1951; Weinberg, 1964; Sinnelä et al., 2019). Therefore, spore forming Thalassobacillus might have shown a positive correlation with manganese.

The hierarchical clustering based on the genera distribution revealed considerable variation in the microbial diversity and abundance in different study sites. However, Venasar and Jogad clustered together while Surajbari was placed apart ( Supplementary Figure 1 ). Since Surajbari is in close proximity with the Gulf of Kutch with a regular flow of sea water, its microbial diversity appears different than the other two sites. EC, TDS, Na⁺, Cl^-, and Ca²⁺ decreased with the increasing distances from the sea, affecting the microbial community in the three study sites.

Based on an earlier metagenomic study, some novel taxa were reported in the LRK (Patel et al., 2015). Two novel taxa, Desertibacillus haloalkaliphilus KJ1-10-99^T (Bhatt et al., 2017) and Chryseobacterium salipaludis JC490^T (Divyasree et al., 2018), were reported from the LRK. A 97.0% of the 16S rRNA gene sequence similarity is generally accepted as the threshold to regard bacterial species as different (Stackebrandt and Goebel, 1994). A sequence similarity in the range of 98.7–99.0% could be regarded as a threshold range above which DNA–DNA hybridization is required for species identification (Stackebrandt and Ebers, 2006). In the present study, nearly 15% of the phylotypes have shown <98.7% 16S rRNA gene sequence similarity with their phylogenetic neighbors. Based on 16S rRNA gene based phylogenetic analysis with the corresponding phylogenetic neighbors, six strains were identified as the putative novel species ( Figure 9 ). Among these, a majority were from Jogad followed by Venasar and Surajbari These phylotypes are affiliated with the Firmicutes for Surajbari and Venasar while with the Firmicutes and Actinobacteria for Jogad.

The 16S rRNA gene sequences of the isolates were compared with the public databases to assess the geographic distribution patterns of different phylotypes. A majority of the phylotypes had cosmopolitan distribution, followed by the saline habitats. Meanwhile, ∼13% of the phylotypes had no significant sequence similarity with their phylogenetic neighbors and thus were considered as the exclusive native of the LRK ( Figure 11 and Table 4 ). Further, ∼8% phylotypes had their distribution confined to marine environment, despite being obtained from the desert, possibly due to the close proximity of the LRK with the Gulf of Kutch (Gupta and Ansari, 2014). In a similar study from Antarctica, 72% of the phylotypes were associated with the cosmopolitan distribution with only 23% being exclusive to Antarctica (Peeters et al., 2011).

A majority of the Firmicutes and Proteobacteria reflected a cosmopolitan and saline distribution ( Tables 3 and 4 ), a trend which might be due to their spore forming ability enabling them to adapt in saline and non-saline environments. The phylotypes within the phylum Actinobacteria had equal distribution to four categories.

The marine distribution of phylotypes decreased with the increasing distance from the Gulf of Kutch, suggesting their endemic nature to the marine environment. It is important to mention that the distribution pattern highlighted in this study is based on the current knowledge of the bacterial diversity and ecology bound by limitations. Therefore, the taxa which are exclusively identified from LRK may turn out as cosmopolitan with the increasing data of the microbial diversity from other habitats.

In the era of metagenomics, there is still a need to improve the conventional cultivation methods to recover the yet unexplored majority of the microorganisms from unusual and unexplored habitats. Metagenomics will not be enough to understand the roles and structures of the microbial communities (Borsodi et al., 2005; Joshi et al., 2008; González-Rocha et al., 2017). In this regard, besides the discovery of potentially new taxa and genera, the availability of the cultures will facilitate detailed investigations including biotechnological avenues. The study also represents the cultivability and geographical distribution of the phylotypes of a saline desert.