Plant microbiomes harbor potential to promote nutrient turnover in impoverished substrates of a Brazilian biodiversity hotspot

Campos rupestres are shallow, acidic, and severely nutrient-impoverished substrates, with phosphorus and nitrogen scarcity limiting plant growth. Despite this, they host exceptionally high plant diversity. Prior research has focused on plant adaptations (e.g., specialized roots, carboxylate exudation), leaving unclear how associated microbiota contribute to nutrient acquisition. This study investigates whether two Velloziaceae species, Vellozia epidendroides (soil patches) and Barbacenia macrantha (exposed rocks), recruit and associate with microbial communities that enhance nutrient turnover. The authors pose four questions: (1) What is the composition and novelty of the communities associated with these plants? (2) What are the differences and similarities between their microbiomes? (3) Are genes linked to microbiota recruitment and nutrient turnover enriched in root-associated microbiomes? (4) What phosphorus and nitrogen turnover mechanisms are encoded, how do they compare between plants, and what role does the shared microbiota play?

Plant adaptations to nutrient scarcity in campos rupestres include durable structures, phosphorus remobilization, and specialized roots secreting carboxylates that enhance nutrient uptake. Many campos rupestres species reportedly show minimal mycorrhizal colonization in severely P-limited substrates, suggesting plant-centric strategies have been emphasized. However, plant-associated microbiomes modulate host responses to stress and can influence speciation, distribution, and diversity, highlighting the plant-microbiome holobiont concept. In other nutrient-limited ecosystems (grasslands, savannas, boreal forests), microbial processes such as nitrogen fixation and mycorrhiza can supply significant fractions of plant N and P, indicating potential microbial contributions in campos rupestres that remain unexplored at high throughput.

Sampling: Substrate (soil for V. epidendroides; rock for B. macrantha) and plant tissues (external and internal compartments of roots, stems, leaves) were collected from six individuals per species in March 2017 within ~200 m² areas. Epi- and endophytic communities were isolated. DNA extraction used DNeasy PowerSoil. Amplicon sequencing: 16S rRNA gene V4 (primers 515FB/806R) and ITS2 (ITS9_Fwd/ITS4_Rev) were PCR-amplified and sequenced on Illumina MiSeq (2×300 bp). ASVs were inferred with DADA2; taxonomy assigned with IDTAXA using GTDB (16S) and UNITE (ITS), filtering organellar sequences (SILVA). Diversity metrics: alpha (richness rarefied to 5,000 reads; Pielou’s evenness), beta (Bray–Curtis, weighted UniFrac). Statistical tests: LMMs for alpha diversity and WACI; PERMANOVA for beta diversity; differential abundance with ALDEx2; family-level enrichment via Kolmogorov–Smirnov tests and hypergeometric tests. Metagenomics: External root and substrate metagenomes from three individuals per condition (soil, rock, rhizospheres of each plant) were sequenced on Illumina HiSeq (2×150 bp). Reads were quality-trimmed (cutadapt) and assembled (MEGAHIT) per sample and as co-assemblies (four total). Protein-level assemblies were generated with PLASS. Contigs were taxonomically assigned (MAGpurify2). MAG recovery used coverage-based binning (MetaBAT2, MaxBin2, CONCOCT, Vamb), aggregated with DAS Tool, dereplicated across co- and individual assemblies (Galah), quality assessed (CheckM), contaminants removed (MAGpurify2), yielding 522 medium/high-quality MAGs. Taxonomy and novelty: MAGs classified with GTDB-Tk; species clusters at ≥95% ANI; phylogenetic novelty quantified as phylogenetic gain (PG) using IQ-TREE and DendroPy. Community-level novelty assessed with the WACI metric (abundance-weighted identity of ASVs vs reference databases via BLAST). Functional annotation and gene abundance: Metagenomes, PLASS proteins, and MAGs annotated for KEGG orthologs (KofamScan/IMG pipeline), Pfam, TIGRFAM; transporters mapped via TCDB with MMseqs2; CAZy with dbCAN. Read mapping to dereplicated genes via Salmon to estimate RPKG (normalized by genome equivalents estimated with MicrobeCensus). Processes evaluated included organic substrate transporters (amino/organic acids), carbohydrate turnover, carbon fixation, phosphorus turnover (transporters, mineralization, solubilization), siderophore BGCs (antiSMASH, BIG-SCAPE), fungal PHO84 transporters (read-level via GraftM), and nitrogen-cycle genes (nifHDK, amoABC/pmoABC, hao, nxrAB, denitrification and assimilatory nitrate reduction genes). LMMs tested enrichment between substrates and rhizospheres. Comparative genomics: Phylogenetic regressions (phylolm) compared copy-number densities of P-turnover genes in campos rupestres MAGs vs GTDB genomes, controlling for phylogeny and genome size. Targeted analyses: nifH phylogenies constructed (MAFFT, trimAl, IQ-TREE); HGT assessment for Isosphaeraceae nif via gene synteny (clinker) and phylogeny; PCR validation of nifH in endophytic root microbiomes (primer IGK3/DVV). Prophage detection via VIBRANT and CheckV; auxiliary metabolic genes annotated; host taxonomy assigned after removing viral regions.

• Community composition and novelty:
  • 29,008 16S rRNA gene ASVs and 9,153 ITS ASVs spanning 3 archaeal, 38 bacterial, and 13 fungal phyla were identified. High-rank taxonomic profiles resembled global soil surveys.
  • Bacterial communities exhibited WACI ~91.8–95.8%, indicating dominance of novel genera/families; below-ground fungal communities were more novel (lower WACI) than above-ground.
  • 522 MAGs from 1 archaeal and 17 bacterial phyla were recovered; 331 species-level clusters (≥95% ANI). Novelty: 51.3% MAGs novel genera, 12.6% new families, 3.6% new orders. Substantial PG added to Eremiobacterota (40.4%), Dormibacterota (13.9%), Binatota (11.7%), and Acidobacteriota (8.6%); Elsterales expanded (PG 76.3%).
• Differentiation and shared core:
  • Microbiomes of V. epidendroides and B. macrantha were significantly different (PERMANOVA p<0.001). 74.9–88.0% (16S) and 74.7–95.8% (ITS) ASVs were plant-specific across sample types.
  • Despite differentiation, 12–25% of 16S ASVs and 4.2–25.3% of ITS ASVs were shared and tended to be highly abundant, often exceeding 50% of community abundance. Twenty-one bacterial families were enriched in the shared fraction.
  • Taxonomically structured enrichment: 41 families from 15 phyla differentially enriched between plants; Actinobacteriota families were recurrently enriched in B. macrantha.
• Recruitment potential and carbon cycling:
  • Genes encoding transporters for amino acids and organic acids were systematically more abundant in rhizospheres than substrates (LMM p<0.001; ω²=0.14), consistent with root-exudate-driven recruitment.
  • Carbohydrate turnover genes were more abundant in B. macrantha-associated communities; photosynthetic bacteria and autotrophic MAGs (38 MAGs across 7 phyla) were more common in rock-associated communities.
• Phosphorus turnover:
  • Fourteen phosphorus mobilization processes (transport, mineralization, solubilization) were detected and enriched in rhizospheres (LMM p<0.001; ω²=0.12).
  • Among the top 15 families by total P-turnover gene abundance, 12 were enriched in the shared microbiota; notable contributors included Xanthobacteraceae and Bryobacteraceae (elevated gcd for gluconic acid synthesis), and UBA5184 (exopolyphosphatase and inorganic pyrophosphatase).
  • Campos rupestres MAGs showed significant enrichment (phylogenetic regressions) for multiple P-related gene families, including PIT and ABC phosphate transporters, inorganic pyrophosphatases, and exopolyphosphatases.
  • Siderophore BGCs: 42 siderophore-producing BGCs grouped into 15 GCFs; a large Pseudonocardiaceae clan resembled desferrioxamine-type siderophores but with distinctive organization. Siderophore BGCs were more abundant in B. macrantha roots and rocks; KEGG siderophore pathway significantly enriched (FDR<0.001).
  • Fungal PHO84 transporters (n=312) were detected and enriched in rhizospheres (trend, LMM p=0.18), with 67.6% assigned to taxa containing known endophytes.
• Nitrogen cycle:
  • Full complement of genes for N transformations detected across 14 phyla; slight enrichment in rhizospheres (LMM p=0.06).
  • Nitrogen fixation: nif diversity dominated by Rhizobiales (55/89 nifH) and Isosphaeraceae (22/55). Four Isosphaeraceae MAGs encoded complete nifHDK; phylogenies and synteny indicate acquisition via HGT from Gammaproteobacteria. A Verrucomicrobiota nifH group II was also found.
  • Bradyrhizobium enriched in endophytic compartments by 16S data; nifH confirmed by PCR in endophytic roots of both species; Bradyrhizobium nifH reads present and enriched in rhizospheres (LMM p<0.05; ω²=0.45). Proviruses carrying exoZ detected in Bradyrhizobium contigs from both plants.
  • Nitrification: canonical AOB with coupled amoABC and hao were not evident. Ammonia oxidation potential was found in Nitrososphaeraceae (AOA) and Binataceae (pmoABC functioning as amoABC). Hydroxylamine oxidation (hao) encoded by Isosphaeraceae (with required heme-binding cysteines). No MAG encoded both amo and hao; evidence supports a metabolic handoff where Binataceae oxidize ammonia to hydroxylamine and Isosphaeraceae oxidize hydroxylamine to nitrite.
• Overall, shared core taxa with high abundance possess key P and N turnover genes, suggesting adaptation to nutrient-poor campos rupestres and potential to enhance host nutrient availability.

The findings support the hypothesis that Velloziaceae in campos rupestres form dynamic associations with microbial communities that enhance nutrient turnover. Despite significant taxonomic differentiation driven by host and substrate, both plants share a highly abundant core microbiota enriched in families with phosphorus turnover potential, suggesting convergent selection of beneficial colonizers. Enrichment of organic substrate transporters in rhizospheres indicates potential recruitment via root exudates. Phosphorus turnover processes, including transport, mineralization, solubilization, and siderophore-mediated mobilization, were elevated near roots, implying microbial contributions complementary to plant exudation strategies. Nitrogen cycling involves both endophytic and free-living diazotrophs; notably, Isosphaeraceae appear to have acquired nitrogen fixation via HGT and, together with Binataceae, may participate in nitrification via metabolic handoff, expanding conventional views of nitrification participants. These results highlight the ecological significance of plant-associated microbiota in nutrient-limited ecosystems and their potential roles in sustaining plant nutrition and diversity in campos rupestres.

This work characterizes the composition, novelty, and functional potential of microbiomes associated with two Velloziaceae species inhabiting phosphorus- and nitrogen-poor substrates. It reveals: (i) highly novel microbial communities with substantial phylogenetic expansion; (ii) strong host- and substrate-driven differentiation alongside a shared, abundant core of efficient colonizers; and (iii) enrichment of genes involved in organic compound uptake, phosphorus mobilization, and nitrogen cycling in root-associated communities. The study provides a basis for holistic models of plant fitness that incorporate microbiome functions in nutrient-limited environments. Future research should experimentally validate microbiome recruitment mechanisms, quantify in situ microbial nutrient transformations, culture key taxa (e.g., Isosphaeraceae and Binataceae) to confirm predicted pathways and interactions, and test microbiome manipulations to enhance plant nutrition and resilience.

• Functional inferences are based on metagenomic potential and gene abundances; direct activity or flux measurements were not performed, requiring experimental validation (e.g., stable isotope probing, enzyme assays).
• Endophytic root communities lacked metagenomic assemblies; nif-related conclusions for endophytes relied on PCR and read-level detection rather than MAG reconstruction.
• Transporter enrichment may reflect accumulation of non-exudated organics; the specificity for root exudates and recruitment mechanisms were not directly tested.
• Nitrification handoff between Binataceae and Isosphaeraceae is hypothesized from gene distributions; co-culture or spatial proximity assays are needed to confirm metabolic interactions and hydroxylamine transfer.
• Mycorrhizal contributions were inferred from PHO84 read detection without direct symbiosis characterization; fungal lineages are highly novel and uncharacterized.