Host Plants, Not Soils, Shape the Phylogenetic Assembly of Culturable Legume-Associated Bacteria

Technical Summary: Host Plants, Not Soils, Shape the Phylogenetic Assembly of Culturable Legume-Associated Bacteria

Problem Statement
A central unresolved question in plant-microbe ecology concerns the relative contributions of soil origin versus host plant identity in determining the phylogenetic structure of bacterial communities associated with legumes. While soils act as reservoirs of microbial diversity influenced by edaphic factors (pH, organic matter, mineralogy), it remains unclear whether legume hosts exert strong selective pressures that override soil-derived community composition. Conflicting literature suggests that soil properties may overwhelmingly shape community composition, while other studies indicate strong host-driven filtering where plants grown in different soils converge toward similar bacterial assemblages. This study addresses this gap by evaluating whether culturable legume-associated bacteria show detectable tendencies toward phylogenetic filtering relative to the available edaphic reservoir, and whether different legume species exhibit distinct, reproducible recruitment patterns across contrasting soils.

Methodology
The study employed a controlled greenhouse experiment involving three legume species—cowpea (Vigna unguiculata), groundnut (Arachis hypogaea), and mungbean (Vigna radiata)—grown in five contrasting, nutrient-poor, sandy soils from the KwaZulu-Natal region of South Africa. The experimental design included:

1. Sample Collection: Soil samples were collected prior to planting. Post-harvest, root nodules were excised from the legumes.
2. Isolation: Bacterial isolates were obtained from both the pre-planting soils (targeting phosphate-solubilizing, nitrogen-cycling, and nitrogen-fixing bacteria) and the root nodules. A total of 77 pure bacterial isolates were characterized.
3. Molecular Analysis: Isolates were identified via 16S rRNA gene sequencing (primers 63F/1387R) and taxonomic assignment using BLASTn against the NCBI GenBank database.
4. Phylogenetic and Diversity Analysis: A Maximum Likelihood phylogenetic tree was constructed. Diversity was assessed using richness, Shannon diversity, and rarefaction analyses. Community composition and phylogenetic structure were evaluated using Bray–Curtis dissimilarity, Principal Coordinates Analysis (PCoA), and unweighted UniFrac distances. Phylogenetic clustering was quantified using Mean Pairwise Distance (MPD) and Mean Nearest Taxon Distance (MNTD) with standardized effect sizes (SES).

Key Results

• Soil-Derived Communities: Soil isolates displayed broad phylogenetic dispersion across major bacterial lineages (including Burkholderia sensu lato, Pseudomonas, Paenibacillus, Bacillus, and Sphingomonadaceae). There was no evidence of soil-specific phylogenetic clustering; isolates from all five sites were distributed across multiple clades. Rarefaction curves indicated substantial variability in richness among soils, reflecting the heterogeneous nature of the edaphic reservoir.
• Host-Associated Communities: In contrast to the dispersed soil communities, nodule-associated bacteria exhibited host-specific clustering. While isolates from each host species spanned multiple clades, they showed a tendency to group within specific regions of the phylogeny.
• Host Specificity: Groundnut-associated communities were distinct from those of cowpea and mungbean. Cowpea and mungbean shared more closely related bacterial assemblages (lower phylogenetic distance) compared to the more divergent groundnut assemblage.
• Diversity Metrics: Mungbean-associated isolates showed significant phylogenetic clustering (SES-MPD and SES-MNTD values below -1.96). Rarefaction curves for host plants showed a narrower range and convergence in richness compared to soils, suggesting nodules harbor a filtered subset of available soil diversity.
• Compositional Shifts: PCoA based on Bray–Curtis dissimilarity showed soils distributed broadly across ordination space, whereas plant-associated communities formed clearer clusters, particularly for groundnut.

Key Contributions
The study provides an integrated evaluation of microbial assembly by simultaneously considering soil origin and host identity within the culturable fraction of the microbiome. It demonstrates that:

1. Soils function as diverse reservoirs of phylogenetically widespread taxa without forming distinct soil-specific clusters in the culturable fraction.
2. Legume hosts act as selective filters that reorganize this diversity, resulting in communities that are phylogenetically and compositionally more constrained than their soil counterparts.
3. Host identity drives the phylogenetic structure of culturable symbionts, with different legume species recruiting distinct subsets of the available soil diversity.

Significance and Claims
The authors conclude that plant hosts play a central role in structuring nodule-associated bacterial communities not by generating novel assemblages, but by filtering and reorganizing the phylogenetic diversity present in the soil. The findings support the "reservoir–filter" model, where soils define the available diversity, but plant identity determines which clades are enriched. The study highlights that host-driven filtering is a dominant force in the assembly of legume microbiomes, even when the surrounding soil community is taxonomically broad and unstructured. The authors note that these results are specific to the culturable fraction of the microbiota, acknowledging that this represents a selective but ecologically meaningful subset of the total soil microbiome. The work underscores the importance of considering multiple dimensions of diversity (taxonomic, compositional, and phylogenetic) when studying microbiome assembly in legumes.