Rhizosphere and Soil Depth Differentially Shape Microbial Community Composition and Assembly in Phragmites australis Salt-Marsh Soils

Common reed (Phragmites australis) is a native dominant plant in many coastal wetlands. To determine how rhizosphere effects and soil depth shape microbial communities, we sampled the rhizosphere and three bulk-soil layers (0–15, 15–30, and 30–45 cm) in a monodominant common-reed stand in a coastal salt marsh. Soil physicochemical properties and bacterial and fungal α-diversity, community composition, and assembly processes were evaluated using one-way ANOVA, principal coordinates analysis (PCoA), permutational multivariate analysis of variance (PERMANOVA), neutral community models, and phylogenetic null models. Rhizosphere pH was lower than that of 0–15 cm bulk soil (mean 8.434 vs. 8.712) but remained alkaline; soil organic matter, total nitrogen, hydrolyzable nitrogen, and total phosphorus were greatest in the rhizosphere. Neither bacterial nor fungal richness or Shannon diversity differed significantly among compartments (p > 0.05); fungal Shannon means ranged from 2.083 to 3.077, with relatively higher Bacteroidota and lower Acidobacteriota abundance in the rhizosphere. Fungal composition did not differ significantly (pseudo-F = 0.609, R2 = 0.102, p = 0.9112), although Mucoromycota and Rozellomycota were relatively more abundant in the rhizosphere. Phylogenetic null models indicated predominantly deterministic bacterial assembly, with 10% dispersal limitation in the 15–30 cm layer. Fungal assembly was predominantly stochastic in bulk soils, whereas the rhizosphere was an exception: heterogeneous selection accounted for 60% of pairwise comparisons and median βNTI exceeded +2. The Mantel test identified only the association between total phosphorus and bacterial diversity as significant (0.01 < p < 0.05). These results show that rhizosphere filtering strongly structured bacterial composition and imposed deterministic selection on rhizosphere fungi.