Tarim Basin Soil Salinity Reshapes Bacterial Communities While Sparing Core Microbiota

A comprehensive geographic survey spanning more than 3,000 kilometers encircling the Tarim Basin in Xinjiang, China, has shown that soil salinity acts as an environmental filter that reshapes bacterial communities and diminishes overall bacterial diversity. The Tarim Basin is China’s largest inland river basin, covering over one million square kilometers, and forms an extreme saline-arid ecosystem dominated by the Taklamakan Desert, with annual rainfall of only 50 to 80 millimeters. To establish a basin-scale understanding of microbial communities across this hyper-arid landscape, researchers gathered 132 soil samples across 44 sites representing non-saline, slightly saline, moderately saline, highly saline, and extremely saline soils under Food and Agriculture Organization standards. The survey demonstrated that while geography primarily shapes broad community structure, escalating salinity enriches specialized halophilic taxa while preserving a stable core microbiota.

Soil salinization currently affects nearly one-tenth of global drylands, and projections indicate that salinized areas will increase globally by 6.7 percent by 2100. Saline soils contain sufficient soluble salts to impair conventional crop growth, threatening agricultural productivity and land sustainability across arid regions. Soil microorganisms drive essential biogeochemical cycles, where microbial survival and stress adaptations control nutrient turnover. Previous microbiome studies focused primarily on agricultural lands, coastal areas, or experimental salinity manipulations, leaving natural saline soils with minimal human influence largely unexplored. In the Tarim Basin, earlier studies investigated localized habitats such as Tarim River microbial assembly, plant-associated microbiota, or nitrogen and carbon cycling within restricted scopes. Consequently, a basin-wide understanding detailing salt-adapted microbial taxa, community assembly mechanisms, and core microbiota across this extreme region remained lacking.

Soil sampling was conducted between April and May 2023 across 44 sites encircling the core region of the Tarim Basin. Sampling locations covered altitudes from 813 to 1,628 meters, between latitudes 36°49′ to 41°58′ North and longitudes 75°08′ to 89°35′ East, spanning more than 3,000 kilometers around the Taklamakan Desert. At each site, three adjacent soil replicates were collected, providing 132 samples for a nested design. Surface crusts and loose sand measuring two to three centimeters were removed before sampling. Sterile stainless steel shovels and disposable gloves were used to prevent contamination, collecting roughly five kilograms of soil from zero to twenty centimeters depth. Bulk soil was stored at room temperature for physicochemical testing, while 0.25 grams was stored at minus 20 degrees Celsius for DNA extraction.

Twelve soil physicochemical properties were measured per sample. Sodium and potassium were determined using flame photometry, whereas calcium and magnesium were quantified by EDTA titration. Sulfate was analyzed using barium sulfate turbidimetry, chloride through Mohr titration, and bicarbonate via acid-base titration. Electrical conductivity and pH were measured with dedicated meters. Soil organic matter was determined by potassium dichromate oxidation-photometry, salinity gravimetrically, and available phosphorus using molybdenum-antimony colorimetry. Soils were classified into five salinity categories based on salt content: non-saline below 3 grams per kilogram, slightly saline at 3 to 6 grams per kilogram, moderately saline at 6 to 10 grams per kilogram, highly saline at 10 to 20 grams per kilogram, and extremely saline above 20 grams per kilogram.

Genomic DNA was isolated from 0.25 grams of soil using the PowerSoil DNA Isolation Kit. Bacterial 16S ribosomal RNA gene V3-V4 regions were amplified using barcoded primers 341F and 806R, while the fungal ITS2 region was amplified with primers ITS3 and ITS4. Each 30-microliter reaction contained 15 microliters of Phusion High-Fidelity PCR Master Mix, 0.2 micromolar of each primer, and approximately 10 nanograms of template DNA. Cycling conditions included denaturation at 98 degrees Celsius for one minute, 30 cycles of 98 degrees for 10 seconds, 50 degrees for 30 seconds, and 72 degrees for 30 seconds, and final extension at 72 degrees for five minutes. Gel electrophoresis verified amplicons, and negative controls produced no bands, confirming reagent purity before sequencing on an Illumina NovaSeq 6000 platform using 250-base paired-end cycles.

Bioinformatic pipelines processed the sequencing outputs to identify taxonomic variants. Raw paired-end reads were separated by barcodes and trimmed before merging with FLASH version 1.2.11. Quality filtering via fastp version 0.23.1 generated clean tags, and chimeras were identified using UCHIME against SILVA version 138.2 and UNITE version 10.0 databases. DADA2 in QIIME2 version 2022.2 denoised sequences into amplicon sequence variants. Taxonomic assignment utilized SILVA for bacteria and UNITE for fungi, referencing the MicroNT sub-library for unclassified regions. Sequencing depths ranged from 42,630 to 182,795 reads per sample for bacteria and 38,085 to 248,757 reads for fungi. Variants with fewer than 10 total reads were removed from subsequent analyses. Reads were normalized to relative abundance for community evaluations, linear mixed models, and co-occurrence network construction across the sampling transect.

Statistical evaluations revealed distinct responses between bacterial and fungal communities across the basin. PERMANOVA demonstrated that soil salinity explained an R-squared value of 0.05 in both bacterial and fungal community composition. When evaluated using stratified PERMANOVA constrained within sampling sites, the salinity effect was not statistically significant. Instead, geographic sampling location explained 43 percent of bacterial community variation and 39 percent of fungal variation. Dominant bacterial phyla included Pseudomonadota, Actinobacteriota, Firmicutes, Bacteroidota, Halobacterota, Chloroflexi, and Gemmatimonadota. Among these, Halobacterota exhibited the highest relative abundance in extremely saline soils and the lowest in moderately and non-saline soils. Fungal communities were primarily composed of Ascomycota, followed by Basidiomycota.

Distributions of dominant genera illustrated structural changes across salinity levels. In extremely saline soils, dominant bacterial genera were Alcaligenes, Stenotrophomonas, and Achromobacter. Highly saline soils were dominated by Palleronia-Pseudomaribius, Achromobacter, and Gramella. Moderately saline soils showed highest abundances of Ralstonia, Escherichia-Shigella, and Achromobacter, while slightly saline soils were enriched in Lactobacillus, Nitrosarchaeum, and Escherichia-Shigella. Non-saline soils were dominated by Nitrosarchaeum, Candidatus Nitroscosmicus, and Domibacillus. Among fungal genera, extremely saline soils exhibited highest abundances of Lulwoana, Leptospora, and Stolonocarpus. Highly saline soils were dominated by Herpotrichia, Trichophaeopsis, and Berkeleyomyces. Moderately saline soils showed higher abundances of Collariella and Subramaniula. Slightly saline soils were enriched in Subramaniula, Gloeoporus, and Dimorphosporicola, whereas non-saline soils showed highest abundances of Immersiella, Marcelleina, and Dimorphosporicola, confirming distinct generic turnover along the basin gradient.

LEfSe analysis identified biomarker bacterial genera across salinity tiers. In extremely saline soils, biomarkers included Natronomonas, Salinimicrobium, Stenotrophomonas, Bacteroides, Halorussus, Halomicrobium, and Salinibacter. Highly saline biomarkers were Achromobacter, Halomonas, Marinobacter, Alifodinibius, Planococcus, and unidentified Actinomarinales. Moderately saline soils were marked by Rubrobacter, Tunicatimonas, Truepera, Pseudomonas, and Sphingobium. Slightly saline soils were characterized by Pontibacter, Lactobacillus, Sphingomonas, Microvirga, Arthrobacter, Massilia, and Rubellimicrobium, while non-saline soils featured Tumebacillus, Adhaeribacter, Domibacillus, and MND1. For fungi, LEfSe identified Cladosporium, Paraphoma, Malassezia, and Lachancea as biomarkers for extremely saline soils, and Alternaria, Trichophaeopsis, Rhizophlyctis, Tausonia, Bathanamyces, Ascobolus, and Coprinopsis for highly saline soils. ANCOM-BC showed Bacillus and Marinobacter as top differentially abundant bacteria in extremely saline soils compared to non-saline soils, with fungal orders Pleosporales and Sordariales enriched under extreme salinity.

Geochemical analyses indicated that major cations (calcium, potassium, magnesium, sodium), major anions (chloride, sulfate), and electrical conductivity decreased significantly from extremely saline to non-saline soils. In extremely saline conditions, these parameters were at least three times higher than in other salinity conditions. Soil pH decreased significantly with increasing salinity, correlating negatively with cation and anion concentrations. Correlation analyses showed that ASV47 (Halomicrobiaceae), ASV31 (Natronomonas), and ASV39 (Natronomonas) correlated negatively with magnesium, sulfate, chloride, total salt, electrical conductivity, sodium, and organic matter, but positively with pH. Conversely, Arthrobacter, Micrococcaceae, Stachybotrys, Rubrobacter, Rubellimicrobium, and Fusarium declined under higher salinity.

The study identified core microbiota based on three criteria: mean relative abundance exceeding 0.01, occurrence in more than 80 samples, and no significant variation in relative abundance across salinity levels. In the co-occurrence network, 34.21 percent of core variants belonged to Pseudomonadota, including Halomonas, Palleronia-Pseudomaribius, Pseudomonas, Stenotrophomonas, and Ralstonia. Bacteroidota accounted for 23.68 percent, including six Salinimicrobium variants and Balneolaceae. Actinobacteriota represented 21.05 percent, including Rubrobacter, Kocuria, and Nitriliruptoraceae. Halobacterota comprised 10.53 percent, containing Natronomonas and Halomicrobiaceae. Firmicutes constituted 5.26 percent through Bacillaceae, alongside Gemmatimonadota and Truepera within Deinococcota.

The authors noted methodological limitations in their analysis. Core microbiota were defined based on relative abundance, which may not fully represent in situ ecological functioning or absolute biomass. Furthermore, the core definition relied on sample occurrence rather than site-level prevalence, meaning the nested design of three replicates per site could have influenced taxa identification. While negative controls showed no detectable bands, trace reagent contamination cannot be entirely excluded in low-biomass desert samples, although consistent detection across high-biomass soils supports data authenticity. Amplicon sequencing also provided genus-level resolution, which cannot distinguish species with divergent salinity tolerances, such as Natronomonas pharaonis requiring high salt versus Natronomonas halophila thriving at lower salt levels.

Subject of Research: Soil salinity gradients and microbial community assembly in the Tarim Basin

Article Title: Salinity-adapted and core microbiota along the soil salinity gradients encircling the Tarim Basin

Article References: Liu, X., Jia, Q., Tang, Q., Ding, Y., Zhang, Z., Gu, M., Huang, W., Wang, N., Zhang, L., & Zhu, J. (2026). Salinity-adapted and core microbiota along the soil salinity gradients encircling the Tarim Basin. Environmental Microbiome, 21(1), Article 118. https://doi.org/10.1186/s40793-026-00918-3

Image Credits: AI Generated

DOI: 10.1186/s40793-026-00918-3

Keywords: Tarim Basin, Taklamakan Desert, soil salinization, microbial community assembly, core microbiota, halophilic bacteria, Natronomonas, Salinimicrobium, Stenotrophomonas, Salinibacter, Halomonas, soil electrical conductivity, amplicon sequencing, 16S rRNA gene, ITS2 region, arid soil ecology