Against the backdrop of global climate change and sea-level rise, coastal ecosystems are confronting multifaceted threats, including exacerbated salinization, biodiversity erosion, and ecological functional degradation (Tully et al. 2019). Coastal salinization, as a pervasive environmental issue, not only directly constrains coastal agriculture and ecological security but also profoundly restructures soil physicochemical properties and vegetation distribution patterns. These changes, in turn, significantly alter carbon and nitrogen cycling processes and undermine the stability of coastal wetland ecosystems (Li et al. 2024). Chenier Islands, representative tidal depositional landforms, provide a vivid record of coastline dynamics shaped by land-sea interactions. (Liu et al. 2016). These unique geomorphological features not only chronicle historical sea-level fluctuations and sedimentary dynamics but also serve as a natural laboratory for investigating coastal salinization ecological processes and plant adaptation mechanisms in the context of contemporary global change.
Chenier Islands, a globally rare intertidal depositional landform, are mainly found along the silt-muddy coasts surrounding the Bohai Sea. These islands represent a unique coastal dune formation shaped by the combined effects of tidal dynamics and the accumulation of mollusk shell debris. The soil matrix primarily consists of shell fragments intermixed with yellowish-brown fine sand (Xia et al. 2015). The distinct layered structure of shell and sand provides significant geomorphological insights into coastal ecosystems. As sensitive ecological interfaces within the land-sea transition zone, Chenier ridges serve dual functions: acting as natural barriers against storm surges and supporting specialized saline habitats and soil conditions that promote distinct vegetation patterns (Liu et al. 2005). Perennial plant species include P. australis and Glycine soja Sieb. et Zucc (wild soybean), Limonium bicolor (Bag.) Kuntze (seashore statice), and Tamarix chinensis Lour. (Chinese tamarisk), predominantly occupy key ecological niches in this region (Zhao et al. 2022), while annual halophytes, such as Common seepweed and Suaeda glauca (Bunge) Bunge (Seaside seepweed), are commonly found across the islands; they provide a natural laboratory for studying the adaptation mechanisms of halophytes to saline-stress environments.
Common reed, a perennial grass in the Poaceae family, has a near cosmopolitan distribution. With an extensive root system and a high reproductive capacity, this species propagates through seeds, rhizomes, and stolons to establish new ramets under favorable environmental conditions (Juneau and Tarasoff 2013). Common seepweed is an annual plant in the family Amaranthaceae, genus Suaeda. It is primarily distributed across Asia and Europe, typically found in patches within extreme environments such as saline-alkali soils, wastelands, and tidal flats. The species exhibits a strong capacity for environmental adaptation. Under various environmental stressors, it can mitigate adverse effects by adjusting its physiological and growth characteristics (Cao et al. 2022). It exhibits strong adaptability to salt stress and is commonly found growing in saline-alkali soils, often forming monospecific communities near coastal areas and lakeshores.
The rhizosphere, the narrow region of soil influenced by plant root activity, is a complex and dynamic microecological environment. This zone offers diverse ecological niches that support a wide range of bacterial communities, which play crucial roles in plant growth, nutrient absorption, and stress tolerance (Chapman et al. 2010). The rhizosphere soil environment is central to the interactions between plants and rhizosphere-inhabiting Bacteria. Within this zone, the functioning and regulation of all biological processes are primarily determined by the dynamic interplay between plant roots and associated Bacteria populations (Solomon et al. 2024). Among the microorganisms interacting with plants, bacteria and fungi have the highest proportions, and they play essential roles in promoting plant growth, nutrient uptake, and enhancing plant stress resistance (Trivedi et al. 2020) Autotrophic nitrogen-fixing bacteria promote plant growth by accelerating phosphorus cycling in the soil and the release of heavy metal ions through processes such as nitrogen fixation, phosphate solubilization, production of siderophores, secretion of IAA (indole-3-acetic acid), and organic acids (Huang et al. 2024; Kraepiel 2009) However, not all microorganisms benefit plants; some pathogenic bacteria can cause diseases and inhibit or impair plant growth. Phytopathogens, including Ralstonia pseudosolanacearum, Fusarium oxysporum, Rhizoctonia solani, and viruses such as wheat yellow mosaic virus (WYMV), can infect root, stem, and foliar tissues, causing significant plant damage (Chen et al. 2025). The proliferation of these pathogens disrupts mutualistic plant-microbe interactions, ultimately impairing ecosystem functioning (Singh et al. 2025). Plant rhizospheres harbor abundant bacterial communities inhabiting root surfaces and the surrounding soil microenvironments. Plant species identity, rhizosphere effects, sloughed root tissues, root exudates, and diverse environmental factors collectively shape the composition of these rhizosphere microbiomes (Yetgin 2023; Pantigoso et al. 2022). Employing high-throughput sequencing techniques to characterize rhizosphere soil bacterial community composition and functionality enables the elucidation and harnessing of plant-microbe interactions to promote plant growth, sustain biodiversity, and achieve species coexistence (Yuan et al. 2022). Field investigations revealed that common reed is widely distributed across the Chenier Islands, with well-developed rhizomes interweaving into an underground network. The root network of common reed is near the root systems of common seepweed. Common reed is a perennial herbaceous plant that propagates continuously via underground rhizomes, forming stable, year-to-year persistent communities that provide a long-term, continuous rhizosphere habitat for microorganisms (Čížková and Lukavská 1999; Chen et al. 2025). In contrast, common seepweed is an annual true halophyte that completes its entire life cycle from seed germination to senescence within a single growing season (Song and Wang 2015). Furthermore, their salt tolerance strategies differ fundamentally: common reed adapts to saline environments primarily through salt avoidance and secretion mechanisms (Guan et al. 2017), whereas common seepweed evolves succulent organs for salt inclusion (Song et al. 2022). Providing an opportunity to investigate the composition, characteristics, and functions of the rhizosphere bacterial communities of common reed and seepweed and their interactions. Rhizosphere bacteria regulate plant growth through their secretions, serving as signaling molecules. Plants and their rhizosphere microbiota collectively form a holobiont, whose extended genome significantly enhances the host’s phenotypic plasticity and ecological adaptability under saline stress. (Olanrewaju et al. 2019; Vandenkoornhuyse et al. 2015). Although the rhizosphere microbiota of coastal saline systems has been widely explored, it remains unclear how P. australis and S. salsa, with their interlaced root systems, achieve niche separation and adaptation to the saline environment in the special soil matrix of the Chenier Islands through differentiated rhizosphere microbiota recruitment strategies. This study hypothesized that P. australis and S. salsa employ distinct recruitment strategies for their root-associated microbiota, predicting significant structural differentiation but functional convergence in their rhizosphere microbial communities. By extracting rhizosphere soil DNA and conducting 16S rRNA high-throughput sequencing, we analyzed the composition and predictive functional profiles of the microbial communities associated with both plant species. This approach enabled a systematic comparison of community structure and potential functional roles, aiming to elucidate plant ecological adaptation mechanisms in the unique habitat of Chenier Islands and provide a theoretical foundation for the phytoremediation and management of coastal saline-affected areas.
Soil physical and chemical parameters.
| Sample ID | Sample Type | EC (μs/cm) | pH |
|---|---|---|---|
| PA-1 | PA | 746 | 8.59 |
| PA-2 | PA | 745 | 8.28 |
| PA-3 | PA | 735 | 8.38 |
| PA-4 | PA | 729 | 8.27 |
| PA-5 | PA | 741 | 8.50 |
| PA-6 | PA | 732 | 8.36 |
| SS-1 | SS | 632 | 8.38 |
| SS-2 | SS | 631 | 8.32 |
| SS-3 | SS | 635 | 8.33 |
| SS-4 | SS | 637 | 8.36 |
| SS-5 | SS | 633 | 8.36 |
| SS-6 | SS | 641 | 8.33 |
PA – Phragmites australis rhizosphere soil
SS – Suaeda salsa (L.) Pall. rhizosphere soil
The study area of this experiment is located within the Chenier Islands of the Yellow River Delta (37°54′30″N to 38°19′10″N, 117°45′08″E to 118°05′37″E). This region experiences a warm-temperate semi-humid continental monsoon climate, characterized by cold winters and hot summers, with four distinct seasons. The mean annual temperature ranges from 11.7°C to 12.6°C, and the average annual precipitation ranges from 530 mm to 630 mm (Zhang et al. 2021). Field sampling was conducted on April 6, 2025, in the Binzhou Shell Ridge Island and Wetland National Nature Reserve (37°54′30″ to 38°19′10″N, 117°45′08″ to 118°05′37″E). A rectangular plot of approximately 150 m × 150 m was delineated in an area where common reed (P. australis) and common seepweed (S. salsa) co-occur. Sampling points were arranged along the plot’s diagonal transect, with a minimum spacing of ≥15 m between adjacent points, and avoiding areas with obvious signs of human disturbance. The plant selection criteria were as follows: (i) absence of disease or pest symptoms; (ii) plant height within the top 25% biomass range of the plot (common reed > 30 cm, common seepweed > 15 cm); (iii) only a single individual plant collected per sampling point, with intact root systems ensured; (iv) the spatial distance between the two sampled plant species at each point not exceeding 5 m. Root systems from the 0–30 cm soil layer were carefully excavated. After gently lifting the roots and shaking off loosely adhering soil, the tightly adhering rhizosphere soil was collected by brushing the root surfaces with a sterile brush. The soil samples were then placed into sterile zip-lock bags and stored in a 4°C cooling container. A total of 12 rhizosphere soil samples (about 50 grams per sample) were obtained, comprising six biological replicates per plant species (designated as the PA group for P. australis and the SS group for S. salsa). Subsequently, impurities were removed from the samples by passing them through a sterile 2 mm sieve. Aliquots of 0.5 g to 1 g of the sieved soil were taken for DNA extraction, while the remaining soil was stored at –80°C for subsequent high-throughput sequencing of the 16S rRNA gene.
The pH and electrical conductivity (EC) of the rhizosphere soil samples were measured to assess the edaphic conditions. Air-dried soil was suspended in deionized water (1:5 w/v) and shaken for 30 minutes. The pH was determined using a digital pH meter, and EC was measured using a conductivity meter. Measurements were performed in triplicate for each sample.
Total genomic DNA was extracted from rhizosphere soil samples of P. australis and S. salsa using the TGuide S96 Magnetic Soil/Stool DNA Kit (Tiangen Biotech, China), following the manufacturer’s protocol. The quality and integrity of the DNA were assessed by 1.8% agarose gel electrophoresis, while concentration and purity were determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). The V3-V4 hypervariable region of the bacterial 16S rRNA gene was amplified using the primer pair 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) (Takahashi et al. 2014). Both forward and reverse primers were tailed with sample-specific Illumina index sequences to enable multiplexed deep sequencing. Libraries were constructed with the Illumina TruSeq Nano DNA LT Library Prep Kit (FC-121-4001). The workflow consisted of: end repair and A-tailing (50°C for 30 min, 72°C for 5 min); ligation of universal adapters (P5: 5′-AATGATACGGCGACCACCGAGAUCTACAC-3′ P7: 5′-CAAGCAGAAGACGGCATACGAGAT-3′) at a final concentration of 15 μM double-size selection with AMPure XP beads (0.8× followed by 0.2× ratios) to enrich for 300–500 bp insert fragments and a 12-cycle PCR enrichment to incorporate the 8 bp dual-indexes. The final libraries showed a primary peak of 460–480 bp and were quantified and normalized to 2 nM. The 20 μl PCR mixture comprised: 5–50 ng of DNA template, 0.3 μl each of forward and reverse primers (10 μM), 5 μl of KOD FX Neo Buffer, 2 μl of dNTPs (2 mM each), 0.2 μl (1.0 U) of KOD FX Neo DNA Polymerase (Toyobo, Japan), and ddH2O up to 20 μl. The thermal cycling conditions were: initial denaturation at 95°C for 5 min; 20 cycles of denaturation at 95°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 40 s; with a final extension at 72°C for 7 min. PCR amplicons were purified using the Omega DNA Clean Kit (Omega, USA), quantified on a Qsep-400 Bio-Fragment Analyzer (Bioptic, Taiwan), and subjected to paired-end (2 × 250 bp) sequencing on an Illumina NovaSeq 6000 platform (Biomarker Technologies, China).
According to quality of single nucleotide, raw data was primarily filtered by Trimmomatic (Bolger et al. 2014) (version 0.33) was used with parameters SLIDINGWINDOW:4:20 MIN-LEN:200 (4-base sliding window with average quality ≥ Q20, reads shorter than 200 bp were discarded) Adapter removal criteria: Cutadapt v1.9.1 was used to identify and remove primer sequences, allowing ≤ 2 mismatches at the 3′ end. After filtering, an average of 69,498 high-quality reads was retained per sample, with a minimum of 53,136 high-quality reads retained per sample. Identification and removal of primer sequences was performed by Cutadapt (Martin 2011) (version 1.9.1). PE reads obtained from previous steps were assembled by USEARCH (Edgar 2013) (version 10) and followed by chimera removal using UCHIME (Edgar et al. 2011) (version 8.1). The high-quality reads generated from the above steps were used for subsequent analyses. Clean reads were then subjected to feature classification via DADA2 trunc-len-f 240, max-ee 2) (Callahan et al. 2016) to generate amplicon sequence variants (ASVs). This method identifies exact sequence variants without relying on sequence similarity clustering. ASVs with counts < 2 across all samples were filtered out. Taxonomic annotation of ASVs was performed using the Naive Bayes classifier in QIIME2 (Bolyen et al. 2019) against the SILVA database (Quast et al. 2012) (release 138.1) with a confidence threshold of 70%.
The DADA2 (Callahan et al. 2016) method in QIIME2 2020.6 (Bolyen et al. 2019) was employed to denoise, merge paired-end sequences, and remove chimeric sequences, yielding the final valid data. Following ASV table generation, Venn diagrams were constructed to visualize ASV overlap among samples, and shared microbial features across environments were identified based on taxonomic annotation. Taxonomic annotation of feature sequences was performed using a Naive Bayes classifier trained on the SILVA reference database. Following taxonomic annotation, ASVs assigned to chloroplasts (taxonomic label: p__Chloroplast) and mitochondria (taxonomic label: p__Mitochondria) were removed using the QIIME2 feature-table filter-taxa command to eliminate plant organelle contamination. Species distribution bar plots across taxonomic levels were generated using QIIME software. Alpha diversity indices (ACE, Chao1, Simpson, Shannon) were calculated in QIIME2. Statistical differences between groups were assessed using the Student’s t-test (independent samples t-test) (R v4.1, stats::t.test) with FDR correction (q < 0.05). All samples were rarified to a uniform sequencing depth of 30,000 reads per sample based on rarefaction curve plateau analysis in mothur (v1.22.2) (Wang et al. 2012). Beta diversity analysis was conducted using Principal Coordinates Analysis (PCoA) (Gower 1966) based on Weighted Unifrac distances in QIIME. Statistical significance of sample differences was assessed via PERMANOVA implemented in the vegan R package, with visualizations created in Python. LEfSe analysis was performed to identify discriminative taxa between groups. The analysis first applied the Kruskal-Wallis test (α = 0.05) to filter features with significant inter-group differences, followed by Linear Discriminant Analysis (LDA) with a logarithmic score threshold of 4.0. The statistical significance of LDA scores was validated using 1,000 permutations. For functional prediction, PI-CRUSt2 (v2.3.0) was employed using the phylogenetic placement approach (non-closed reference), where ASVs were inserted into the reference phylogenetic tree (IMG/GTDB) to infer gene families. FAPROTAX (v1.2.6) was used for ecological function annotation based on taxonomic similarity (≥ 97% identity) without closed-reference clustering. Both predictions were performed on the ASV abundance table without OTU clustering. Study area maps were generated in ArcGIS.
Overview Map of the Study Area in the Chenier Islands.
Bar chart of rhizosphere bacterial community composition for Phragmites australis and Suaeda salsa. a) Phylum-level bacterial composition; b) Genus-level bacterial composition; c) Family-level bacterial composition: The x-axis denotes sample names. The y-axis represents relative abundance (as a percentage). Colors correspond to distinct taxa, with stacked bars depicting the top 10 taxa by relative abundance at each taxonomic rank.
Core and unique bacterial ASVs and sequencing depth sufficiency between Phragmites australis and Suaeda salsa rhizospheres. a) Venn diagram showing the number of shared and unique ASVs between groups; b) Rarefaction curves of observed ASVs. The plateau of curves indicates adequate sequencing depth for microbial diversity analysis.
Number of ASVs per sample and number of sequences.
| Sample ID | ASVs_Num | Seqs_Num |
|---|---|---|
| PA-1 | 2031 | 48924 |
| PA-2 | 2714 | 52570 |
| PA-3 | 2816 | 48397 |
| PA-4 | 2739 | 51227 |
| PA-5 | 2172 | 49832 |
| PA-6 | 1844 | 53546 |
| SS-1 | 1869 | 60414 |
| SS-2 | 2325 | 57275 |
| SS-3 | 2443 | 57159 |
| SS-4 | 2372 | 58874 |
| SS-5 | 2151 | 58309 |
| SS-6 | 2132 | 58982 |
| Total | 10063 | 655509 |
PA – Phragmites australis rhizosphere soil
SS – Suaeda salsa (L.) Pall. rhizosphere soil
Analysis of the fundamental physicochemical properties of the rhizosphere soil of the two plants revealed that pH values were stable and did not differ significantly between the P. australis and S. salsa rhizospheres. In contrast, the electrical conductivity (EC) was significantly higher in the PA rhizosphere compared to the SS rhizosphere (t-test, p < 0.001).
Alpha diversity of the bacterial microbiota in the rhizosphere of Phragmites australis and Suaeda salsa. Box plots show the (a) Simpson, (b) Shannon, (c) ACE, and (d) Chao1 indices. The center line represents the median, the box limits indicate the upper and lower quartiles, and the whiskers extend to 1.5× the interquartile range. Outliers are shown as points. Significant differences between groups (p < 0.05) are indicated with asterisks and exact p-values.
Based on species annotation and taxonomic analysis, bar charts of species classification were generated for the top 10 most abundant bacterial species in the rhizosphere soil of common reed and common seepweed. At the phylum level, the dominant bacteria in the PA group are Proteobacteria, Bacteroidota, Chloroflexota, Actinomycetota, and Acidobacteriota. In the SS group, the dominant bacteria are Proteobacteria, Bacteroidota, Chloroflexota, Actinomycetota, and Cyanobacteria. At the family level, two dominant bacterial families (relative abundance > 1%) were identified in both groups: unclassified_Coleofasciculaceae: The relative abundance was 1.05% in the PA group and 3.38% in the SS group, showing a significantly higher abundance in the common seepweed rhizosphere. unclassified_Unknown_Family: The relative abundance was 2.48% in the PA group and 1.75% in the SS group, with a higher abundance in the common reed rhizosphere. At the genus level, aside from unclassified species, the dominant genus in both common reed and common seepweed rhizosphere is Zeaxanthinibacter, with relative abundances of 0.60% in the PA group and 2.49% in the SS group. Other major unclassified genera include unclassified_Bacteria: 7.50% (PA group) and 2.72% (SS group), unclassified_Actinomarinales: 4.61% (PA group) and 3.16% (SS group), unclassified_Gammaproteobacteria: 2.44% (PA group) and 2.39% (SS group).
In this study, a total of 10,063 bacterial ASVs were obtained across sample groups. The PA group contained 4,314 unique ASVs (42.8%), while the SS group had 2,739 unique ASVs (27.2%). Both groups shared 3,010 ASVs. Rarefaction curves were generated to validate the adequacy of sequencing depth for capturing Bacterial diversity. Curve plateauing indicated sufficient sequence coverage, confirming the robustness of the data for downstream analyses. When the number of randomly selected sequences reaches 10,000, the curve gradually flattens, suggesting that the data are valid.
Beta diversity analysis of rhizosphere bacterial communities. a) Principal coordinates analysis (PCoA) plot. Samples are colored by group, with ellipses showing 95% confidence intervals. Axes show the percent variation explained by each principal component; b) Boxplots comparing between-group and within-group distances.
LEfSe Analysis Significant microbial differences between the two plants. a) LEfSe analysis phylogenetic tree. The concentric circles radiating from the innermost to the outermost represent taxonomic levels from phylum to species.; b) LDA Value Distribution Bar Chart The vertical axis represents classification units with significant differences between groups. The horizontal axis indicates the degree of difference among classification units, with longer bars signifying greater differences.
PICRUSt2 Analysis of Phragmites australis and Suaeda salsa at Different Functional Levels. a) KEGG Level 2 functional hierarchy; b) KEGG Level 3 functional hierarchy. The x-axis represents the sample groups, while the y-axis shows the relative abundance (%) of metabolic pathways.
In bacterial community studies, alpha diversity indices are commonly used to characterize the species features of a community. Among these, the Shannon and Simpson indices primarily measure species diversity, while the Chao1 and ACE indices focus on assessing species richness within the community. The results indicated no significant differences in the Simpson index and Shannon index between the PA and SS groups (p > 0.05). Similarly, no significant differences were observed in the Chao1 and ACE indexes (p > 0.05). The SS groups exhibited a mild upward trend in key alpha diversity indices relative to the PA group: the Shannon index increased by 5.11% (from 8.41 to 8.84), the Simpson index increased by 0.51% (from 0.987 to 0.992), the Chao1 index increased by 10.20% (from 2027.83 to 2234.58), and the ACE index increased by 10.20% (from 2048.67 to 2257.42). The coverage of all samples was significantly greater than 0.99, indicating that all fragment sequences were detected and that the sequencing results accurately reflect the actual conditions.
Principal Coordinates Analysis (PCoA) was employed to visualize similarities and dissimilarities among samples, providing an intuitive representation of divergence and convergence between distinct bacterial communities. Results demonstrated that PC1 and PC2 accounted for 55.01% and 20.01% of the total β-diversity variation, respectively, in the Weighted Unifrac-based PCoA. Robust separation along the PC1 axis between the PA group (P. australis) and SS group (S. salsa) was observed with a clear distribution pattern: the PA group exhibited positive PC1 values (mean: 0.1139, range: -0.0233 to 0.1706) while the SS group showed negative PC1 values (mean: -0.1139, range: -0.1730 to -0.0496), Statistical analysis further confirmed that the inter-group difference along PC1 was extremely significant (p = 0.0001), whereas no significant difference was detected along the PC2 axis (p = 0.2385). PERMANOVA analysis (distance metric: binary_jaccard; permutation method: permanova) further confirmed the statistical significance of inter-group differences. The calculated R2 value of 0.205 (representing the proportion of variance explained by grouping factors) with higher R2 values denoting greater separation significance. The statistically significant p-value (p = 0.008; p < 0.05) validated high analytical reliability.
LEfSe analysis was primarily used to identify differentially expressed biomarkers that were significantly different between sample groups. Analysis of rhizosphere microbes in Common reed and Common seepweed (taxonomic levels: phylum to genus; LDA threshold = 4.0) revealed 21 significantly different taxa between the two plant species. Significant biomarkers, excluding unclassified bacteria, included the phyla Acidobacteriota and Chloroflexota in the reed rhizosphere. Biomarkers for Common seepweed comprised: Bacteroidota (phylum), Cyanobacteria (phylum), Cyanobacteria (class), Bacteroidia (class), Alphaproteobacteria (class), Rhodothermia (class), Cyanobacteriales (order), Flavobacteriales (order), Rhodobacterales (order), Cytophagales (order), Flavobacteriaceae (family), Rhodobacteraceae (family), Nostocaceae (family, corrected spelling), Cyclobacteriaceae (family).
Differential abundance of predicted ecological functions in the rhizosphere bacterial communities of Phragmites australis and Suaeda salsa; x-axis, species; y-axis, relative abundance (%) of ecological functions.
Differential functional predictions between groups. a) PICRUSt2 and b) FAPROTAX analyses. Colors correspond to sample groups. For each function, the left panel displays relative abundance, the middle panel shows the abundance difference with 95% confidence intervals, and the right panel indicates the p-value.
Functional prediction of rhizosphere Bacteria in common reed and common seepweed was conducted using PICRUSt2, focusing on the top 10 functional categories by relative abundance under KEGG Level 2 classification. This analysis revealed that both plant rhizospheres exhibited similar functional profiles with comparable abundances. The predominant metabolic pathway-related functions included global and overview maps (the most dominant function), carbohydrate metabolism, amino acid metabolism, energy metabolism, metabolism of cofactors and vitamins, membrane transport, and nucleotide metabolism; genetic information processing-related functions comprised translation, replication, and repair, while environmental information processing-related functions featured signal transduction.
Further analysis of the top 10 most abundant KEGG Level 3 functional categories via PICRUSt2 prediction revealed functional congruence, with comparable abundances between common reed (P. australis) and common seepweed (S. salsa) rhizospheres. Metabolic pathway-related functions included metabolic pathways (dominant function), biosynthesis of secondary metabolites, biosynthesis of antibiotics, bacteria metabolism in diverse environments, biosynthesis of amino acids, carbon metabolism, and purine metabolism; environmental information processing-related functions comprised two-component system; genetic information processing-related functions featured ribosome; cellular process-related functions involved ABC transporters.
FAPROTAX analysis of the top 10 ecological functions in the rhizospheres of common reed and seepweed revealed shared functional profiles with divergent relative abundances. Functions included: photosynthetic carbon/nitrogen fixation (oxygenic_photoautotrophy: PA 3.53%, SS 11.02%; cyanobacteria: PA 3.53%, SS 11.02%); mixotrophic metabolism (photoautotrophy: PA 6.99%, SS 11.7%; photoheterotrophy: PA 5.25%, SS 2.16%); aerobic organic matter degradation (aerobic_chemoheterotrophy: PA 14.79%, SS 11.58%); anaerobic energy backup (fermentation: PA 6.09%, SS 4.08%); nitrogen cycling (nitrate_reduction: PA 2.18%, SS 2.13%; nitrogen_fixation: PA 0.69%, SS 2.21%). The top-tier functional categories, phototrophy (PA 9%, SS 13.37%) and chemoheterotrophy (PA 18.99%, SS 15.61%), hierarchically encompass the sub-functions above, with phototrophy comprising photoautotrophy and photoheterotrophy. In contrast, chemoheterotrophy includes aerobic chemoheterotrophy and fermentation.
Analysis of the top 30 most abundant KEGG pathways (level 2) revealed limited functional divergence between the two plant species. Microbial functional profiles were largely characterized by conserved core metabolism, with only three pathways showing statistically significant differences (FDR-corrected p < 0.05), all of which were enriched in common seepweed (SS). Two additional pathways approached significance, though all differentially abundant categories exhibited modest effect sizes. At a finer functional resolution (level 3), none of the top 30 KEGG pathways differed significantly between groups. Similarly, FAPROTAX profiling of the top 30 ecological functions identified only one significant group difference: the relative abundance of animal_ parasites_or_symbionts was higher in common reed (PA; 0.53%) than in common seepweed (SS; 0.32%; corrected p = 0.012). No other functions have achieved statistical significance.
Analysis of rhizosphere physicochemical properties revealed that although both plants grew in soil with a consistent pH background (p > 0.05), they created markedly different rhizosphere microenvironments. The common reed rhizosphere showed significantly higher electrical conductivity (p < 0.001), aligning with its salt-exclusion strategy by accumulating salts in the rhizosphere to prevent excessive shoot uptake. In contrast, the lower EC in common seepweed rhizosphere corresponds to its salt-secreting/diluting physiological mechanism. This plant-driven divergence in rhizosphere chemical conditions is likely a key driver of the significant differences in their rhizosphere microbial community structures. These findings demonstrate that, in a shared soil reservoir, host plant species are the primary factor shaping the rhizosphere microenvironment.
Soil bacteria play pivotal roles in material cycling and signal transduction within the rhizosphere (Yang et al. 2025), critically contributing to plant growth promotion, nutrient acquisition, and maintenance of plant health (Mukhtar et al. 2025). Rhizosphere bacteria community composition and functionality are fundamental to these processes. In the context of increasingly severe global coastal salinization, understanding plant-microbe interactions is particularly significant for the conservation and restoration of coastal ecosystems. This study analyzed bacterial communities and their functions in the rhizosphere environments of common reed and common seepweed by extracting rhizosphere soil DNA and using 16S rRNA high-throughput sequencing. Results revealed a distinctive pattern of structural differentiation but functional convergence between the two plant-associated microbiomes, consistent with prior research (Louca et al. 2018). Overall, the rhizosphere microbiomes of the two plant species exhibited a clear pattern of structural differentiation despite functional convergence. Alpha diversity indices showed no significant differences between species, whereas beta diversity analysis based on Weighted Unifrac distance revealed significant compositional divergence (P = 0.008). This structural distinction was further supported by LEfSe analysis, which identified 21 differentially abundant taxa. In contrast, both PICRUSt2 functional predictions and FAPRO-TAX ecological profiling indicated highly similar functional profiles between the two rhizospheres.
The absence of differences in alpha diversity alongside significant PCoA separation may arise because both plants inhabit the same environment, yet differ in salt tolerance mechanisms and root exudates. In addition, rhizosphere samples were collected during the early growth stage (Liu et al. 2023), when recruitment strategies—specifically the enrichment or suppression of specific bacterial taxa—can diverge (Walters and Martiny 2020; Zverev et al. 2021). Alternatively, the unique salinized soils of Chenier Islands may exert a more substantial community-shaping effect than root-mediated selection (Ding et al. 2023; Gao et al. 2020). The lack of alpha-diversity differences indicates that the two rhizospheres possess similar capacities to host bacteria (Wei et al. 2020), providing stable niches for microbes even under salt stress. This finding provides important insights into the mechanisms that maintain stability in fragile coastal wetland ecosystems such as the Chenier Islands. The pronounced PCoA dissimilarity suggests species turnover—replacement of taxa. At the same time, functional roles are retained, allowing plants to eliminate sensitive microbes and recruit functionally equivalent microbes to cope with environmental pressures (Huang et al. 2024).
Functional redundancy enables the rhizosphere microbiomes of the two plants to maintain similar functional potential despite differences in community composition (Louca et al. 2018). This implies that even if the bacterial community shifts, the overall functionality can remain stable provided that the keystone functional guides persist or are compensated by other taxa with equivalent functions. Under the dual pressures of sea-level rise and intensifying salinization in global coastal zones, this functional redundancy mechanism provides a micro-level explanation for the maintenance of coastal vegetation ecosystem resilience.
LEfSe analysis identified Acidobacteriota as a differential biomarker in the common reed rhizosphere, primarily associated with carbohydrate metabolism and Global and overview maps pathways, consistent with prior research (Gonçalves et al. 2024). This phylum facilitates plant adaptation to hypersaline conditions on the Chenier Islands by decomposing complex soil carbohydrates, releasing sugars, small-molecule compounds, and antioxidants (Wang et al. 2025). Concurrently, it coordinates diverse metabolic activities to drive carbon, nitrogen, and sulfur cycling, thereby enhancing soil quality in the rhizosphere (Douglas et al. 2020). Meanwhile, Chloroflexota predominantly performs Energy metabolism and amino acid metabolism (Bovio-Winkler et al. 2023), participating in glycolytic and TCA cycle pathways to supply energy, carbon, and nitrogen to plant roots. It further regulates intracellular redox homeostasis (Liu et al. 2022) and osmoregulation, strengthening salt-stress resilience (Ingrisano et al. 2023). In common seepweed, differential biomarkers predominantly comprised Cyanobacteria, Proteobacteria, and Bacteroidota. Their core functional pathways encompassed Global and overview maps, Energy metabolism, and Amino acid metabolism. Research indicates these bacteria facilitate plant growth and enhance salt-stress resilience through photosynthesis, carbon/nitrogen fixation, phytohormone secretion, organic matter decomposition, and nutrient cycling (Gao et al. 2019; Kollmen and Strieth 2022; Ravanbakhsh et al. 2019).
The results indicate that both common reed and common seepweed prioritize functional rather than taxonomic composition in their rhizosphere microbiomes. By secreting specific root exudates, the plants recruit microbes that possess the required functional traits (Sasse et al. 2018) to meet the metabolic demands of coping with salt stress, irrespective of the exact species that provide these functions. This confirms that plants are the primary architects of their rhizosphere bacterial communities, a conclusion consistent with previous findings (Yang et al. 2025). In the typical intertidal ecosystem of the Chenier Islands, plants effectively maintain survival and reproduction in highly saline environments through function-oriented microbial recruitment strategies, providing theoretical references for vegetation restoration in similar coastal ecosystems worldwide.
PICRUSt2 and FAPROTAX—functional prediction tools based on distinct principles and databases—revealed convergent metabolic potentials in the rhizosphere microbiomes of common reed and seepweed. While PICRUSt2 predictions showed identical functional categories with highly similar relative abundances at both KEGG Levels 2 and 3, FAPROTAX ecological profiling, despite confirming functional congruence, revealed quantitative disparities in relative abundances. This outcome significantly reinforces the hypothesis that plant-driven selection of rhizosphere microbes prioritizes functional convergence over taxonomic identity. FAPROTAX predicts bacterial community functions based on taxonomic information and known ecological classifications (Sansupa et al. 2021). When the composition and structure of the rhizosphere microbiota differ between the two plants, this may lead to differences in the relative abundance of functions (Guo et al. 2025; Li et al. 2025), even if they can perform similar functions. Thus, differences in the relative abundance of functions in the FAPROTAX analysis do not undermine the validity of the conclusion above. This also confirms that functional redundancy is a vital foundation for ecosystem stability (Biggs et al. 2020). When the core functional microbes of common reed and common seepweed are reduced due to environmental factors, other species with the same functions can compensate and maintain functional output (Louca et al. 2018). This is crucial for their adaptation to the saline stress of the Chenier Islands and various environmental changes. With increasing environmental uncertainties in coastal zones due to global climate change, revealing this functional redundancy mechanism has profound implications for predicting and regulating functions of salinized ecosystems, while also offering new perspectives for bioremediation strategies in coastal saline-alkali lands.
The widespread functional redundancy observed in this study, in which distinct microbial taxa support overlapping metabolic functions, carries profound ecological implications for plant adaptation in dynamic saline environments. This mechanism provides halophytes with functional resilience, ensuring the stability of critical processes, such as nutrient mineralization and stress hormone regulation, even when microbial community composition fluctuates due to environmental disturbances. For the host plants, this means that their adaptation does not rely on the presence of specific, potentially sensitive microbial species, but rather on a buffered, functionally versatile microbiome. Consequently, common seepweed can depend on a consistent supply of fixed nitrogen from various photoautotrophic taxa, while common reed can secure its carbon and energy needs through multiple heterotrophic pathways. This ecological strategy effectively decouples plant fitness from taxonomic volatility, allowing both species to thrive under the saline-alkaline stresses of the Chenier Islands by maintaining core physiological functions through diverse yet functionally equivalent microbial partnerships.
The observed functional convergence despite taxonomic divergence suggests the operation of profound ecological and evolutionary mechanisms that shape the rhizosphere microbiome in this saline environment (Li et al. 2024). First, environmental filtering imposed by the pervasive salt stress on Chenier Islands serves as a primary selective force, constraining the total repertoire of metabolically viable functions. The high salinity acts as a filter, permitting only microorganisms with specific genetic traits for osmolyte synthesis, ion homeostasis, and salt-tolerant enzymes to establish and function, thereby limiting the range of possible functional profiles. Second, functional redundancy, the existence of multiple phylogenetically distinct taxa capable of performing the same biochemical process, provides the necessary ‘raw material’ for this convergence (Louca et al. 2018). This redundancy allows each plant host to recruit from a common regional species pool but arrive at distinct taxonomic assemblages that are functionally equivalent. Most critically, the convergence is likely driven by host-mediated selection for key functions. Rather than selecting for specific microbial taxa per se, the plants appear to exert selective pressure on microorganisms through their root exudates that can provide host-beneficial services (Pantigoso et al. 2022; Yetgin 2023). For common reed, this may manifest as a strong selection for heterotrophic degraders to mineralize organic matter (Guan et al. 2017), whereas for common seepweed, the selection likely favors phototrophic and nitrogen-fixing partners (Song et al. 2022). In essence, the plants address similar challenges of nutrient acquisition and stress tolerance in a saline habitat by cultivating microbial communities that provide these essential services, regardless of their taxonomic identity. This represents a sophisticated form of ecosystem-level optimization, prioritizing functional stability over taxonomic composition to ensure resilience in a dynamic intertidal environment.
The observed pattern of functional convergence amidst taxonomic divergence finds strong parallels in other saline ecosystems yet highlights the unique rhizosphere mediation by coastal halophytes. In inland salinized grasslands, halophytes create “saline fertile islands” that enrich functional genes essential for nutrient cycling, demonstrating how plant activity directly enhances microbial metabolic capabilities (Liang et al. 2025; Zhao et al. 2024). Similarly, the temporal dynamics of microbial functional genes in reclaimed salinized farmland reveal that community function can stabilize despite compositional changes (Yin et al. 2024), echoing the functional redundancy we observed in the Chenier Islands. However, our study reveals a more specialized mechanism: host-specific functional guild selection. While studies of saline lake sediments show that bacterial communities adapt to environmental stress through functional gene shifts (Yang et al. 2022), our findings demonstrate that common reed and common seepweed actively recruit distinct taxonomic groups to maintain convergent functions—a strategy more targeted than the broader environmental filtering observed in bulk soils. This represents a refined ecological strategy in which coastal halophytes in shell-sand matrices optimize functional acquisition through taxonomic flexibility, thereby ensuring ecosystem resilience amid increasing salinization pressures.
Although rhizosphere microbes in intertidal saline systems have been extensively investigated, existing studies have predominantly focused on either “community structure description” or “single functional genes,” leaving the “structure-function” coupling mechanism and the differentiation of plant energy strategies unresolved. This study is the first in Chenier Island sedimentary environments to integrate ASV-level structural differences with PICRUSt2/FAPROTAX functional profiling and statistical testing, revealing how P. australis and S. salsa achieve niche separation through contrasting “photoautotrophy-nitrogen fixation” versus “chemoheterotrophy-nitrate reduction” pathways. We propose a rhizosphere model of “structural differentiation but functional convergence”. These findings provide new perspectives for plant-microbe synergistic adaptation theory in saline lands and offer insights applicable to other coastal wetlands for optimizing vegetation restoration and carbon sequestration management.
This study has several limitations that should be considered. First, we acknowledge the inherent constraints of the 16S rRNA amplicon sequencing approach, which, unlike shotgun metagenomics, provides limited phylogenetic resolution and cannot directly access the full functional gene repertoire of microbial communities (Zhu et al. 2022). The functional profiles derived from PICRUSt2 and FAPROTAX are probabilistic predictions based on taxonomic data and existing databases, reflecting metabolic potential rather than actual microbial activity. These predictions are further constrained by the completeness and annotation accuracy of the reference databases themselves, which may not fully capture the unique genetic potential of environmental microbes. Experimental validation is needed to confirm their ecological relevance. Second, the rhizosphere samples were collected during the early plant growth stage and thus may not fully represent stabilized microbial communities in later developmental phases. Third, although the stark contrast in microbial communities between the two plant species—grown in immediate proximity and thus sharing the same soil pool—strongly suggests host-specific selection, the absence of bulk-soil controls prevents definitive partitioning of the rhizosphere microbiota into “soil-derived” and “plant-recruited” components. Future studies incorporating paired bulk soil sampling will be essential to precisely quantify plant enrichment effects in this ecosystem. Finally, as bacterial assembly is influenced by multiple factors, this study focused primarily on plant-specific effects, and other potential influences warrant further investigation.
In summary, this study demonstrates that although common reed and common seepweed recruit phylogenetically distinct rhizosphere bacterial communities, these assemblages exhibit functional convergence under saline conditions. The observed niche partitioning—photoautotrophic dominance in seepweed versus chemoheterotrophic specialization in reed—reflects divergent microbial recruitment strategies, yet both pathways ultimately support plant adaptation through complementary mechanisms. This pattern of “structural differentiation but functional convergence” highlights the importance of functional redundancy in maintaining ecosystem stability under environmental stress, indicating that different bacterial community compositions can achieve analogous ecological functions in helping plants adapt to saline environments. However, it should be noted that the functional roles attributed to bacterial taxa are based on predictive tools (PICRUSt2 and FAPROTAX) rather than experimental validation. Future work incorporating metatranscriptomics or enzyme activity assays will be essential to empirically verify these putative functions and further elucidate the mechanistic basis of plant-microbe interactions in saline ecosystems. Additionally, future research should explore how these plant-specific microbial recruitment strategies vary across seasonal dynamics and environmental gradients. From an applied perspective, our findings provide a microbial ecological basis for optimizing vegetation restoration strategies in coastal saline lands by leveraging the complementary functions of different halophyte species.