Sediment microbial community structure, enzymatic activities and functional gene abundance in the coastal hypersaline habitats

Salt marsh vegetation, mudflat and salt production are common features in worldwide coastal areas; however, their influence on microbial community composition and structure has been poorly studied and rarely compared. In the present study, microbial community composition (phospholipid fatty acid (PLFA) profiling and 16S rRNA gene sequencing (bacterial and archaeal)) and structure, enzymatic activities and abundance of functional genes in the sediments of salt ponds (crystallizer, condenser and reservoir), mudflat and vegetated mudflat were determined. Enzyme activities (β-glucosidase, urease and alkaline phosphatase) were considerably decreased in saltpan sediments because of elevated salinity while sediment of vegetated mudflat sediments showed the highest enzyme activities. Concentrations of total microbial biomarker PLFAs (total bacterial, Gram-positive, Gram-negative, fungal and actinomycetes) were the highest in vegetated mudflat sediments and the lowest in crystallizer sediments. Nonmetric-multidimensional scaling (NMDS) analysis of PLFA data revealed that the microbial community of crystallizer, mudflat and vegetated mudflat was significantly different from each other as well as different from condenser and reservoir. The most predominant phyla within the classified bacterial fractions were Proteobacteria followed by Firmicutes, Bacteroidetes and Planctomycetes, while Euryarchaeota and Crenarchaeota phyla dominated the classified archaeal fraction. Cyanobacterial genotypes were the most dominant in the condenser. Mudflat and vegetated mudflat supported a greater abundance of Bacteroidetes and Actinobacteria, respectively. The results of the present study suggest that salt ponds had significantly decreased the microbial and enzyme activities in comparison to mudflat and vegetated mudflat sediments due to very high salinity, ionic concentrations and devoid of vegetation. The present study expands our understanding of microbial resource utilization and adaptations of microorganisms in a hypersaline environment.


Introduction
Salt marsh vegetation, mudflat and salt production are common features in worldwide coastal areas. Intertidal coastal zones constitute the interface between terrestrial and aquatic ecosystems . The coastal ecosystem is the most productive system which provides services like climate regulation/buffering from natural hazards, nutrient cycling, carbon sequestration, shoreline stabilization, food and fodder, habitat for marine life and biodiversity (Calvão et al. 2013; Thompson and Schlacher 2008). Coastal areas are under the influence of various pressure (natural and anthropogenic), such as mariculture, solar-salt production, coastal mining, erosion, pollution, grazing of riparian vegetation and deforestation of mangroves (Calvão et al. 2013). Microbes play a fundamental role in ecological processes, including organic matter decomposition, nutrient cycling and fixation which are necessary for the functions and services of the coastal ecosystem (Rathore et al. 2017;Wang and Wang 2018). Microbial activities greatly depend on the environmental conditions of sediments, including carbon substrate and nutrient availability, salt concentration, water Communicated by Muthusamy Govarthanan.
1 3 56 Page 2 of 16 content, temperature and pH Rathore et al. 2017;Jeddi et al. 2022). Therefore, it is necessary to investigate the influence of coastal vegetation, mudflats and salt production on the microbial community composition of sediments that generally exist in a coastal ecosystem.
India ranks the third position in global salt production after China and USA (Bhat et al. 2015) and Gujarat state shares the longest coastline (1600 km). Salt ponds are shallow pools with elevated borders designed to store seawater or other brines to produce salts. The salt ponds consist of three distinct interconnected ponds, namely; reservoir, condenser and crystallizer pond (Mani et al. 2012). During the tidal influxes or mechanical pumping, the seawater is collected into the reservoir pond and then seawater is transferred into the condenser ponds, where seawater evaporates by solar radiation. The condenser feeds crystallizer ponds and the salinity of crystallizer ponds is much higher than the rest of the two. The salt concentration is a crucial component of these salt ponds. Salt production ponds invariably increase the concentration of salt (salinity) in the sediments as compared to coastal intertidal sediments (Song et al. 2022;Wei et al. 2022); however, the influence of increased salinity on sediment community composition/structure of microbes particularly in the hypersaline environments is not much studied. Salt pans and marshes are attractive model systems for the investigation of halophilic microorganisms and also the habitat for novel organisms (Berrada et al. 2012;Wei et al. 2022). Higher salt concentrations generate stress for the microorganisms and only tolerant organisms survive under such circumstances (Oren 2009;Wei et al. 2022). Sfax solar salterns in Tunisia, Mediterranean coastal salterns and salterns in South China have previously been studied for microbial communities (Boujelben et al. 2014;Oren 2009;Wei et al. 2022). Apart from salt concentration, high sunlight irradiation (deterministic factor) also affects the halophilic community which establishes the stable and highly changeresistant communities dominated by Haloquadratum sp. and Salinibacter sp. (Viver et al. 2019). The hypersaline sediments of an ephemeral pond from the S'Avall solar salterns, Spain constituted dominance of methanogenic archaea in the upper layers, whereas bacteria with fermentative and anaerobic respiration metabolisms increased with depth (Font-Verdera et al. 2021). Hypersaline environments sediments (salt lake and salt mines) were dominated by the phyla Proteobacteria, Firmicutes and Bacteroidetes, whereas species belonging to the genus Natronomonas were widely distributed (Chen et al. 2020). However, very few studies compared the microbial communities in sediments of coastal vegetation and salt ponds in India.
Salinity depressed the bacterial and actinomycetes biomarker PLFA concentrations but slightly influenced the amount of fungal PLFA in salinized sediments (Wang and Wang 2018). In contrast, elevated salinity increased the bacterial, fungal and total PLFAs concentrations as observed by Li et al. (2017). Bacteria in a hypersaline environment are more susceptible to shifts in salinity than archaea (Leoni et al. 2020). In addition, the archaeal community can withstand wide salinity fluctuations and remain stable (Mani et al. 2012;Leoni et al. 2020). Studies carried out in hypersaline conditions exhibited that archaea followed by bacteria, were dominant halophilic members (Baati et al. 2008(Baati et al. , 2010Trigui et al. 2011). However, many studies demonstrated that the bacterial domain plays a vital role in hypersaline environments (Berrada et al. 2012;Li et al. 2017;Morrissey et al. 2014;Wang et al. 2012). Salicola marasensis belonged to the ɣ-Proteobacteria subdivision and was found predominantly in the non-crystallizer pond (Boujelben et al. 2014). Culturable bacteria in hypersaline environments showed the predominance of Gram-positive (Bacillus species, which can grow up to 25% salinity) and Gram-negative (Vibrio species) bacteria (Berrada et al. 2012). In the salt crystals of salterns, Bacteroidetes were dominant, while ɣ-Proteobacteria and α-Proteobacteria were equally distributed (Baati et al. 2010). Euryarchaeota and Crenarchaeota phyla were previously reported as the dominant phyla in solar salterns (Ahmad et al. 2011;Xie et al. 2017).
Carbon (C) and nitrogen (N) cycling are the key processes carried out by the microorganisms which are involved in biogeochemical cycling in the coastal sediments. The RuBisCO (ribulose-1,5-bisphosphate carboxylase/ oxygenase) is a notably recognized protein for CO 2 assimilation through the Calvin cycle which is encoded by the cbbL gene (Spiridonova et al. 2006). The nitrogen assimilation process is governed by the diazotrophic bacteria and the nifH gene which encodes the dinitrogenase reductase enzyme involved in N-fixation (Poly et al. 2001a). The 16S rRNA, cbbL and nifH were quantified using a quantitative real-time polymerase chain reaction to enumerate the abundance of bacteria, Cand N-fixer, respectively. However, the influence of different sediments from the salt pond, mudflat and coastal vegetation on the abundance of functional genes (cbbL and nifH) in the coastal ecosystem is still indistinct. We hypothesized that the diverse microbial community exists in the hypersaline environment, and microbial community diversity and functional gene abundances would be decreased at higher salinity. Hence, the objective of the present investigation was to elucidate the effect of different coastal sediments (salt ponds, mudflat and vegetated mudflat) on the microbial community structure (phospholipid fatty acid (PLFA) profiling and 16S rRNA gene sequencing), enzyme activities (β glucosidase, urease, phosphatase, and sulfatase) and functional genes (nifH and cbbL) abundances.

Sites and sediment sampling
The present study was carried out at the Experimental Salt Farm of CSIR-CSMCRI, located at the coastal intertidal area (N21° 47.521' to N21° 47.732'; E72°07.417' to E72°07.644') of Bhavnagar district of Gujarat, India (Fig. S1). The study site is located in a hot semi-arid, with minimum and maximum temperatures of 24-44 °C and 15-32 °C in summer and winter, respectively. The annual average rainfall is 593 mm and is unevenly distributed. The texture of sediment ranged from clay loam to clay. Five different types of sediment samples were collected for the present study. Among the five different sediment samples, three sediment samples were collected from different ponds of solar salt pans (reservoir, condenser and crystallizer ponds) and one each from mudflat (without vegetation) and vegetated mudflat (halophyte growing area) (four replicates of each; total samples: 20).
In each sediment type (crystallizer, condenser, reservoir, mudflat and vegetated mudflat), four replicate quadrats (1 × 1 m) were constructed and sampled from 0 to 20 cm depth in January 2018. Sediment samples were immediately brought to the laboratory after collection. Sediment samples were separated into two subsets; one set was used for physicochemical analysis after air-drying and the second subset was kept at − 20 °C for analysis of microbial assay (enzyme activities, PLFA and genomic DNA analysis).

Physico-chemical characteristics of sediments
Moist sediment samples were oven-dried at 105 °C (24 h) to measure moisture content. Sediment samples were passed through a 2-mm sieve after air drying and stored for chemical analysis.
Salinity and pH of sediment samples were determined in a 1:2.5 (w:v; sediment: water) suspension using electrical conductivity and pH meters, respectively (Rathore et al 2017).
Organic carbon (OC) content in sediments was measured by the chromic acid oxidation method (Nelson and Sommers 1982). Water (40 mL) was used to extract dissolved organic carbon (DOC) from sediments (4 g) followed by determination through a TOC analyzer (Liqui TOC, Elementar, Germany) (Chaudhary et al. 2018).
Neutral normal ammonium acetate (1 N NH 4 OAc, pH 7.0) was used to extract potassium (K + ), sodium (Na + ), calcium (Ca 2+ ) and magnesium (Mg 2+ ) (Hanway and Heidel 1952). The concentrations of K + and Na + were quantified using a flame photometer. An inductively coupled plasma spectrometer was used to estimate the Ca 2+ and Mg 2+ .
The chloride from the sediments was extracted with water (Lundmark and Olofsson 2007) and estimated by titration with AgNO 3 .

Basal respiration and enzyme assay
The basal respiration rate of sediments was assessed by determining the evolved CO 2 in the headspace of the incubation bottle for 24 h using the alkali (NaOH) absorption method (Vogeler et al. 2008). The enzyme activity of β-glucosidase (carbon cycling enzyme) was measured using p-nitrophenyl-β-D-glucopyranoside as substrate (Tabatabai 1994). Activity of urease (nitrogen cycling enzyme) was measured using a method of Tabatabai (1994) and the released NH 4 + -N was measured spectrophotometrically (Keeney and Nelson 1982). Alkaline phosphatase (phosphorus cycling enzyme) activity was analysed using p-nitrophenyl phosphate as a substrate (Tabatabai and Bremner 1969). p-nitrophenyl sulfate was used as a substrate to measure the activity of sulfatase (sulfur cycling enzyme) using the method of Tabatabai and Bremner (1970).

Phospholipid fatty acid (PLFA)
PLFAs were extracted from sediments (5 g) using the procedures of Bardgett et al. (1996) and Frostegård et al. (1993). In brief, lipids were extracted from sediments in three steps, using citrate buffer, chloroform and methanol in the ratio of 0.8:1:2 followed by separation of phospho-, neutral-and glycolipids using a silicic acid column and esterified using alkaline methanol and then analysed by the gas chromatograph (GC) system (Agilent 6850, Agilent Technologies, Inc.). The identification and quantification of PLFAs were carried out with MI System (MIDI Inc., Newark, DE) and internal standard (methyl nonadecanoate; 19:0), respectively.
The concentration of PLFAs was expressed as nmol PLFA g −1 sediment, indicating the microbial biomass. Taxonomic groups of microbes such as bacteria, Gram-positive 56 Page 4 of 16 (GM + ve) bacteria, Gram-negative (GM − ve) bacteria, fungi and actinomycetes were indicated by individual PLFA biomarkers (Zelles 1996;Kaur et al. 2005). The ratios of fungal to bacterial PLFAs (F/B) and GM + ve/GM − ve were used to indicate the shift in microbial community structure (Bååth 2003;Li et al. 2017;Wang and Wang 2018). Specific microbial groups (bacterial, fungal and actinomycetes) PLFAs were computed using both absolute concentration (nmole g −1 soil) as well as relative abundances (molar percent of total PLFA).

DNA extraction
Microbial genomic DNA was extracted from 0.5 g of sediment samples using FastDNA™ SPIN Kit for soil (MP Biomedicals, USA) according to instructions of manufacturer's protocol and using FastPrep bead beating instrument (MP Biomedicals, USA) then quantified by Nanodrop (NanoDrop Technologies, USA). The quality of DNA was determined on the agarose gel (0.8% w/v). Extracted genomic DNA was stored at − 20 °C to amplify the functional gene(s) and 16S rRNA gene.

Amplification of 16S rRNA gene
Microbial metagenomics DNA from samples of the same origin was pooled and concentrated in the concentrator and approximately 1000 μg of DNA was shipped to Macrogen Ltd., South Korea, for paired-end sequencing using the Illumina MiSeq platform. For library preparation, the targetspecific primers, Bakt_341F (5'-CCT ACG GGNGGC WGC AG-3') and Bakt_805R (3'-GAC TAC HVGGG TAT CTA ATC C-5') were used to amplify the V3-V4 hypervariable regions of the bacterial 16S rRNA gene (Mizrahi-Man et al. 2013;Sinclair et al. 2015). Sequence files obtained from Macrogen Ltd. were checked for their quality using FastQC software (Galaxy version 19.01). Low-quality sequences with a Phred score lower than 25 were excluded using the Trimmomatic package listed under galaxy server (www. usega laxy. com; Galaxy version 19.01). Further, the sequences were analyzed using the Mothur software package (version 1.39.1), according to the MiSeq Standard operating procedure (SOP) (http:// www. mothur. org/ wiki/ MiSeq_ SOP). Paired-end reads were joined to form contigs using both forward and reverse sequence files (Kumar et al. 2022). Both forward and reverse primers were trimmed and sequences were further screened for the following parameters: minimum length: 403 bp, maximum length: 429 bp, maximum homopolymers: 8, maxambig: 0 and pdiffs: 3. Unique sequences from each file were screened to enhance the speed and avoid computational problems. Chimera detection was performed with the UCHIME algorithm (Edgar et al. 2011) and removed. The resulting sequences were aligned against SILVA bacterial reference database. Archaeal communities were analysed by aligning the processed sequences against SILVA archaeal reference database. The taxonomic assignment to the sequences was done according to SILVA taxonomy (version 132 for bacteria and version 102 for archaea) in Mothur software (version 1.39.1). Sequences flagged as chloroplasts, mitochondria or eukaryotes were also excluded by using the remove lineage command. Further, the remaining sequences were clustered de novo using a distance matrix algorithm at a distance cut-off of 97% similarity. The total number of bacterial and archaeal sequences were 356,048 and 108,179, respectively. Out of these, a total of 12,031 and 9745 OTUs were obtained for bacteria and archaea, respectively. Rarefaction curves for the observed OTUs were generated using the "rarefaction. single" command in Mothur to evaluate the alpha diversity and sampling effort (Fig. S2). Statistical analyses were performed with the web-based microbiome analyst software (Dhariwal et al. 2017). InteractiVenn software (http:// www. inter activ enn. net/) was used to construct the Venn diagram. The sequences were submitted in the NCBI SRA database under bio-project number PRJNA521453.

Statistical analysis
The results of the study are presented as means (n = 4) ± standard errors (SE) and ANOVA (one-way analysis of variance) was carried out to differentiate the significant differences (Tukey's honest significance test; p < 0.05) between treatment means using SPSS version 19.0 (IBM Corp, NY). Nonmetric-multidimensional scaling (NMDS) analysis was carried out using PC-ORD software (McCune and Mefford 2006). Molar percent of PLFAs was used for NMDS analysis using the Sorensen distance measure and a second matrix was prepared by summed values for microbial groups and soil physicochemical properties which was used for creating the joint plot overlaid on the NMDS plot. The NMDS scores were analysed by one-way ANOVA. The relationships between sediment microbial community and physicochemical characteristics were examined using correlation coefficients (p < 0.05).

Microbial respiration and enzyme activities in sediments
Basal respiration and enzyme activities were significantly influenced by different sediment sources (Table 1). Basal respiration was the highest in sediments from vegetated mudflats and the lowest in crystallizer and condenser. β-glucosidase and urease activities were noticeably lower in the sediments of saltpan and mudflat than in vegetated mudflat. Sediments from mudflat and vegetated mudflat had much higher alkaline phosphatase activity than the other sediments. Unlike other enzyme activities, the activities of the sulfatase enzyme were the highest in the reservoir sediment and the lowest in mudflats.
Correlation analysis between sediment characteristics and enzyme activities showed that β-glucosidase was positively correlated with pH (Table S4). The activities of urease and alkaline phosphatase, as well as the rate of basal respiration were positively correlated with NO 3 − and P while it was negatively correlated with sediment moisture content (except basal respiration), EC, OC, K + , Na + , SO 4 2− , Cl − , Ca 2+ and Mg 2+ . Sulfatase activity showed a positive correlation with pH, OC and Ca 2+ while a negative correlation with P.

Microbial community composition/structure of sediments (PLFA)
The origin of different coastal sediment samples significantly affected the amount of total PLFAs, Gram-positive, Gram-negative, total bacterial and actinomycetes biomarker PLFAs ( Table 2). The highest concentrations of PLFAs microbial biomarkers were exhibited in vegetated mudflat sediments, whereas the lowest was in crystallizer sediments. There was no significant influence of different sediments on the fungi/bacterial ratio of PLFAs. The ratios of GM + ve / GM − ve PLFAs showed significant differences among the sediments and the ratio was the highest in crystallizer and mudflat while the lowest in vegetated mudflat sediments.
Further, attempts were made to correlate the PLFA concentrations of biomarkers of microbial groups with sediment  (Table S5). Amounts of total PLFAs, Grampositive and actinomycetes biomarkers were negatively correlated with EC, K + , Na + SO 4 2− , Cl − and Mg 2+ while positively correlated with NO 3 − . The EC, K + , Na + SO 4 2− , Cl − , Ca 2+ and Mg 2+ were negatively correlated with the amount of Gram-negative, total bacterial and fungal PLFAs while the same microbial groups were positively correlated with NO 3 − . Furthermore, the biomarker PLFA of fungi had a positive relationship with the P concentration of sediments. Amounts of total PLFAs and biomarkers of microbial groups were positively correlated with the activities of all enzymes (except sulfatase) (Table S6).
Molar percent of PLFAs was subjected to NMDS (Nonmetric-multidimensional scaling) ordination analysis and resulted in a two-dimensional elucidation (Fig. 1). The ordination of both axes (2 and 3) explained 93.2% of the total variability (70.5 and 22.7% by axis two and three, respectively; Fig. 2). Gram-positive, Gram-negative, total bacteria, fungi and actinomycetes had a negative relation with axis 2. Among sediment characteristics, EC, OC, NH 4 + , K + , Na + , SO 4 2− , Cl − , Ca 2+ and Mg 2+ were positively correlated while NO 3 − and P were negatively correlated with axis 2. Axis 3 was positively associated with the F/B ratio while negatively associated with Gram-negative and total bacteria. The sediment microbial communities of crystallizer, mudflat and vegetated mudflat were significantly different from each other and distinct from the condenser and reservoir. However, the microbial community of the condenser and reservoir was similar to each other.
The molar percent (relative abundance) of different microbial groups is presented in Fig. 2 and was significantly influenced by the sediment sources. The molar percent of Gram-positive was similar in the vegetated mudflat, mudflat and condenser while significantly lower in the crystallizer. Vegetated mudflat sediments had the highest molar percent of Gram-negative, total bacteria and actinomycetes while the lowest was in sediments from crystallizer. A similar abundance of fungi was observed in a crystallizer, reservoir, mudflat and vegetated mudflat.

Bacterial community structure (16S rRNA sequencing)
After aligning paired-end raw sequence reads, sequences were aligned into 492,804 contigs (ranging from 161,398 to 172,103 per sample). A total of 12,031 OTUs at a 97% similarity cutoff were observed. The observed OTUs were assigned to 22 Phyla, 52 classes, 98 orders and 154 families. The rarefaction curves of the observed OTUs reached saturation, indicating sufficient sampling depth and large library size (Fig. S2). The highest OTUs observed were in vegetated mudflat (4521) followed by mudflat (4372), reservoir (4315), condenser (3529) and crystallizer (2250).
The reads obtained from different samples were compared for alpha diversity measures such as richness (Chao1), abundance-based coverage estimator (ACE) and species diversity (Shannon index). The values of Chao1 and ACE were the highest in mudflat and vegetated mudflat, whereas the lowest was in the crystallizer ( Table 3). The highest Shannon index was observed in the reservoir and the lowest was in the condenser and crystallizer. Simpson index (calculates a measure of diversity) was highest in condenser followed by mudflat, crystallizer, vegetated mudflat and reservoir.

Archaeal community structure (16S rRNA sequencing)
After aligning the sequences with the archaeal silva reference database, taxonomic affiliation with a 97% similarity threshold yielded 9745 OTUs. From the observed OTUs, archaeal community composition was calculated. More than 97% of archaeal OTUs were unclassified and the remaining OTUs were classified into two different phyla (Fig. 5). In classified archaeal phyla, Euryarchaeota was the dominant followed by Crenarchaeota in all the samples. The highest percentage of Euryarchaeota phyla was observed in the condenser followed by crystallizer and mudflat, while the lowest percentage was observed in the reservoir. For the Crenarchaeota phyla, the highest percentage was observed in the reservoir followed by crystallizer and condenser while the lowest percentage was observed in mudflat samples. Further on the class level, the phyla Euryarchaeota was classified into Halobacteria and Methanomicrobia while a large fraction of the same phyla was unclassified (Fig. 5). This showed that the saltpan archaeal community has been poorly studied using the high-throughput NGS technique and needs to be studied. The OTUs from the Crenarchaeota phyla were further classified into Thermoprotei class (18%) and the rest OTUs remain unclassified (Fig. 5). Thermoprotei class was most abundant in crystallizer and condenser while it was absent in mudflat samples. The highest percentage of Halobacteria was observed in the condenser followed by mudflat and crystallizer while the same class was absent in the reservoir and vegetated mudflat. Methanomicrobia class was detected in mudflat, vegetated mudflat and crystallizer while it was absent in condenser and reservoir samples.
The community richness and diversity index values for archaeal communities were calculated (Table S7). Chao1 and ACE values were observed highest in the reservoir followed by condenser and mudflat. The highest Shannon index value was observed in the reservoir, mudflat and vegetated mudflat, while the lowest was in the crystallizer. Simpson index values were highest in the condenser and crystallizer.

Functional gene abundances in sediments
The genomic DNA concentration varied from 7.4 to 20.4 ng µl −1 and the highest concentration was observed in the condenser followed by vegetated mudflat, reservoir, mudflat and crystallizer sediments. The copies of the bacterial 16S rRNA gene varied from 1.8 × 10 7 (crystallizer) to 153.0 × 10 7 (vegetated mudflat) per g of sediment. The abundance of cbbL gene was markedly different in sediments and the highest copy number was exhibited in a vegetated mudflat, whereas the lowest was in crystallizer sediments. Sediments significantly affected the abundance of nifH gene and the highest abundance was observed in the condenser followed by vegetated mudflat, mudflat, reservoir and crystallizer sediments (Fig. 6). The copies of the bacterial 16S rRNA and cbbL genes showed a positive correlation with all enzyme activities and microbial biomarker PLFA concentrations (Table S8).

Microbial respiration and enzyme activities
Basal respiration is extensively used to assess the microbial activity in the sediments/soils and measures the overall decomposition rate of the organic carbon (Vogeler et al. 2008). The basal respiration rate was higher in vegetated mudflat sediments than any other sediments, even though other sediments (crystallizer and condenser) had higher organic carbon which is burial and preserved carbon in the precipitated salts (Table 1). This indicates that the higher sediment salinity significantly reduced the decomposition rate of organic matter which is confirmed by the negative relationship between basal respiration rate and salinity (Table S4). Previous studies have also shown the increased salinity reduced basal respiration considerably salinity (Mahajan et al. 2015). Enzyme activities are sensitive indicators of the quality of sediments and were closely related to physico-chemical characteristics of sediments  (Mahajan et al. 2015;Yang et al. 2017). The enzyme activities (β-glucosidase, urease and alkaline phosphatase) considerably decreased in sediments of salt ponds while vegetated mudflat sediment had the highest enzyme activities. The reduced enzyme activities in salt ponds (with higher salinity) might be due to increased osmotic stress on microbes (Frankenberger and Bingham 1982;Wei et al. 2022) which shows the reduced capability of sediments to mineralize nutrients (C, N and P) as evident in Tables (S2 and S3) as well as a negative correlation with salinity. Enzyme activities under vegetated mudflats could be stimulated by the higher microbial activities sustained by root exudation and litter decomposition (Yang et al. 2017). The salinity did not affect the sulfatase activity as observed by Oshrain and Wiebe (1979).

Sediment microbial community composition/ structure (PLFA)
The biomass of microbes and their activities show the size and magnitude of the microbial population associated with nutrient cycling and biogeochemical processes occurring in the sediments. In the present study, culture-independent (PLFA analysis) approach was used for the estimation of microbial biomass and microbial community structure which is extensively used for sediment-microbial interaction studies Rathore et al. 2017;Wang and Wang 2018). The values of total PLFAs, Gram-positive, Gram-negative, total bacterial and actinomycetes biomarker PLFAs were the lowest in crystallizer and the highest in vegetated mudflat sediments ( Table 2). The increased content of PLFA in vegetated mudflat sediments was mainly attributed to higher contents of NO 3 − and lower salinity (salinity-associated ions) as revealed by the correlation study (Table S6). These results are supported by earlier studies (Wang and Wang 2018) which described that the rise in salinity would depress microbial activity and biomass. The crystallizer and condenser had 4 times higher salinity while e reservoir and mudflat had two times higher salinity than the sediments from the vegetated mudflats (Table S2). Furthermore, crystallizer and condenser had a significantly higher amount of organic carbon than vegetated mudflat sediments; however, there was no relationship between organic carbons and microbial biomass, as observed by Li et al. (2017) in coastal sediments. It shows a more dominant effect of salt content than organic carbon on microbial biomass. Further, fungal biomass was positively correlated to the P concentration of sediments which implies that adequate availability of P is necessary to sustain fungal biomass (Teste et al. 2016). Different sediments did not influence the ratio of F/B which might be due to the adaptation of fungus to the high salt concentration of the coastal wetlands Wang and Wang 2018). In all studied sediments, the biomass of Gram-negative bacteria was higher than Gram-positive bacteria and the ratio of Gram-positive /Gram-negative was significantly affected by different sediments (Table 2). Berrada et al. (2012) also observed that Gram-positive bacteria were widely represented in higher salinity which agreed with the highest Gram-positive/Gram-negative ratio in crystallizer sediments. Sediment microbial communities of crystallizer, mudflat and vegetated mudflat have varied from each other and also differed from condenser and reservoir (Fig. 1) due to the changes in the abundance of bacterial, fungal and actinomycetes PLFA biomarkers (Fig. 2). The increased salinity in sediments from mudflat and salt ponds had shifted the bacterial abundance (dominance of Gram-positive) (Morrissey et al. 2014). There was also a marginal reduction in the abundance of fungus at higher salinity. The Gram-positive bacteria are slow-growing as compared to Gram-negative bacteria and adopt k-strategists (low growth rate with high resource use efficiency) which is related to the resistance of the bacterial community to salinity and increased the ratio of Gram-positive/Gram-negative with a rise in salinity (de Vries and Shade 2013). In earlier studies, it was observed that the genera and species numbers decreased from marsh to salterns at Lower Loukkos (Morocco) (Berrada et al. 2012) and Wendeng salterns of China (Song et al. 2022). Vijayakumar et al. (2007) observed higher abundances of actinomycetes in mangrove sediments followed by mudflat and saltpan sediments (Vijayakumar et al. 2007). Bacterial community structure (16S rRNA sequencing) The bacterial 16S rRNA gene OTUs rarefaction curve reached saturation depicting sufficient sampling depth and large library size to capture a vast majority of the diversity in all samples (Fig. S2). Proteobacteria were detected as the most dominant phyla in crystallizer and reservoir after unclassified bacteria (Fig. 3). Core microbiome analysis of the selected phyla showed that most of the bacterial OTUs belonged to unclassified phyla, indicating that the diversity of the hypersaline ecosystem is poorly studied (Najjari et al. 2015;Zhong et al. 2016). Further, Hu et al. (2014) also reported that the abundance of Proteobacteria is influenced by a change in salinity and is directly proportional to salinity. Apart from Proteobacteria phyla, Bacteroidetes, Firmicutes, Chloroflexi, Actinobacteria, Cyanobacteria and Planctomycetes were also observed in all the samples. Our results corroborate with previous studies, bacterial communities were dominated in sediments of the estuary of the Jiyun River, northern China by Proteobacteria, Firmicutes and Bacteroidetes phyla (Wang and Wang 2018) while salt production stimulated the relative abundances of Actinobacteria, Spirochaetes, Tenericutes, and Chlamydiae . Firmicutes produce spores under extreme conditions for their survival (Yu et al. 2012). In mudflat and vegetated mudflat samples, Bacteroidetes (in mudflat) and Actinobacteria (in vegetated mudflat) were the most abundant phyla. Actinobacteria can withstand harsh environmental conditions in the dormant stage or sporulation, or in an inactive but viable form. When the condition becomes favourable the Actinobacterial cells start dividing again (Jones and Lennon 2010;Crits-Christoph et al. 2013). Similarly, Actinobacteria, Clostridia and Flavobacteria were dominant classes of coastal sediments of the Sfax coastal area (Tunisia) with higher organic matter (Jeddi et al. 2022). Gammaproteobacteria, the most predominant Proteobacterial class observed in this study (Fig. S4) which is phylogenetically and physiologically diverse and involved in nutrient cycling (Evans et al. 2008). The absence of Acidobacteriaceae and Opitutaceae family in the crystallizer and condenser showed that these bacterial families could not tolerate higher salinity (Fig. S5). Desulfovibrionaceae and Desulfobulbaceae represent incomplete oxidizers while Desulfobacteriaceae represents members of complete oxidizers in a hypersaline environment (Foti et al. 2007) which were detected abundantly in crystallizer, condenser and reservoir. Sorokin et al. (2004) demonstrated the sulfur-reducing activity in the Siberian soda lakes with saturated salinity. Many sulfur-reducing bacteria have previously been reported in hypersaline environments (Foti et al. 2007;Song et al. 2022;Wei et al. 2022). Mudflat and vegetated mudflat shared the highest OTUs between them while the crystallizer and condenser shared the second-highest OTUs (Fig. 4). This indicated that the bacterial communities in the mudflat and vegetated mudflat were similar while the bacterial communities in the crystallizer showed similarity with the condenser.
In the present study, bacterial diversity was decreased with an increase in salinity (Table 3) which is also supported by previous studies (Baldwin et al. 2006;Song et al. 2022). Based on the OTU analysis and the different diversity indices (Shannon), the highest diversity was observed in the reservoir and vegetated mudflat samples while the lowest diversity was observed in the crystallizer and condenser which is due to the higher concentration of salt (Baldwin et al. 2006;Song et al. 2022).

Archaeal community structure (16S rRNA sequencing)
Archaeal 16S rRNA gene sequencing results depicted a higher percentage of unclassified archaeal sequences at the phylum level which shows that the diversity in the salt ponds is poorly studied using the NGS technique (Fig. 5). We found that the Euryarchaeota was the most dominant phyla followed by Crenararchaeta in all the samples (Fig. S5). The abundance of Euryarchaeota has been reported by many researchers in saline soils (Xie et al. 2017;Walsh et al. 2005;Jeddi et al. 2022). Jeddi et al. (2022) observed more dominant Euryarchaeota phylum while less dominant Crenarchaeota and Parvarchaeota phyla in the coastal sediments of Sfax (Tunisia). The presence of halophilic archaea such as Halobacteria (Fig. 5) in this study is in accordance with other reported studies in hypersaline environments (Maturrano et al. 2006;Youssef et al. 2012;Weigold et al. 2016). The halophilic archaea (Halobacteria) adopt a salt-in strategy for osmoregulation which requires less metabolic energy compared with the synthesis of compatible solutes (Kulp et al. 2007;Genderjahn et al. 2018), therefore, detected only in condenser, crystallizer and mudflat. The Crenarchaeota phyla observed in the present study were very diverse, ranging from chemolithoautotrophs to chemoorganotrophs which also vary from aerobes to facultative anaerobes to anaerobes (Ahmad et al. 2011;Jeddi et al. 2022). The majority of the Crenarchaeota phyla were unclassified at the class level while classified sequences showed the abundance of Thermoprotei class in all the samples except for mudflat which showed that the class Thermoprotei could also withstand a wide salinity range (Yan et al. 2018). Venn diagram analysis of the archaeal community showed that the mudflat shared maximum OTUs with the vegetated mudflat and reservoir (Fig. S6). This may be because of the lower salinity in both vegetated mudflat and reservoir. Archaeal diversity was the highest in the reservoir followed by mudflat, vegetated mudflat and condenser while the lowest diversity was observed in the crystallizer (Table. S6). The diversity pattern indicated that the archaeal communities did not follow the same pattern as bacteria and did not show a reduced diversity pattern with respect to salinity.
Taxonomic assignment and community composition depend on the choice of variable regions. In the present study, we have targeted V3-V4 regions for the amplification of microbial 16S rRNA gene analysis, while studies suggest that targeting V4-V5 regions gives superior recognition of Archaea (Willis et al. 2019;Parada et al. 2016;Satari et al. 2021). A large number of unclassified archaeal communities on the phylum level may be overcome by targeting the archaeal-specific V4-V5 regions.

Functional gene abundances in sediments
Quantification of key functional genes offers an exceptional tool to study sediment microbial communities insitu without cultivation biases for environmental samples (Spring et al. 2000). Abundances of bacterial 16S rRNA gene and two functional biomarker genes (cbbL and nifH) involved in C and N cycling were quantified (Fig. 6). The relative abundance of gene copy numbers (per gram sediment) occurring in different coastal sediments (crystallizer, condenser, mudflat and vegetated mudflat) was significantly affected by the sediment types. Relatively low copy numbers of the 16S rRNA gene in the crystalline sediment indicates hostile habitat condition due to the high salt concentration and nutrient-deficient environment in comparison to other sediments which imposed additional stress conditions on microbes so that the consumption of C substrate will not be efficient (Marinari et al. 2012;Keshri et al. 2015). Sediments from vegetated mudflats appeared to have a higher abundance of bacterial flora which can be helpful in improving stressed environmental conditions. The carbon-fixing bacterial communities (abundance of cbbL gene) were most abundant in vegetated mudflat sediments and their copy number was inversely proportional to the sediment salinity (lowest abundance in crystalline sediments). The abundance of nifH gene was the lowest in crystallizer sediments while the remaining sediments had a similar abundance. The negative correlations were observed between the abundance of functional genes (cbbL and nifH) and salinity in saline sediments (Sorokin et al. 2008;Keshri et al. 2013Keshri et al. , 2015.

Conclusion
The present study found notable differences in the chemical and microbial characteristics of sediments collected from salt ponds (crystallizer, condenser and reservoir), mudflat and vegetated mudflat. These changes advocate that salt production processes strongly affect the biogeochemical processes and nutrient cycling in the coastal ecosystem. It was established from the present study that the key controller of the microbial community structure and enzyme activities in sediments are sediment salinity and ionic concentration. Vegetation (halophyte) created the most conducive environment for microbial activities in the sediments. The majority of phyla belonged to the unclassified bacteria and archaea which shows that the diversity of hypersaline ecosystems is poorly studied. The bacterial population was dominated by Proteobacteria, Bacteroidetes, Firmicutes and Chloroflexi while Euryarchaeota and Crenarchaeota dominated archaeal communities. The abundance of genes in sediments enhances our information and understanding of the role of microbes in nutrient biogeochemical cycling.

Acknowledgements
The authors gratefully acknowledge the financial assistance GAP2012 and GAP2125/CRG/2020/000542 rendered by the Ministry of Earth Sciences (MoES) and Science and Engineering Research Board (SERB), New Delhi, respectively. The authors are thankful to Mr. S. C. Upadhyay and Aditya P. Rathore for the help received during sample collection and analysis. CSIR-CSMCRI communication No.: 115/2018.
Author contributions MK and VK did analytical work, and MK and DRC wrote the manuscript, prepared the figures and tables. Conception and study designed by DRC. Acquisition of data and statistical analysis carried out by the MK, VK and DRC. All authors reviewed the manuscript.

Funding
The funding was received from the Ministry of Earth Sciences (MoES) (GAP2012) and the Science and Engineering Research Board (SERB), (GAP2125/CRG/2020/000542), New Delhi to carry out the study.

Data availability
The datasets generated during and/or analysed in the current study are available from the corresponding author upon reasonable request.

Conflict of interest
The authors declare that there is no conflict of interest (financial or non-financial).
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.