Comparative metagenomics of wild and cultivated Fragaria chiloensis reveals major rhizosphere microbiome shifts linked to stress adaptation

Highlights

• •Rhizosphere microbiomes shift with domestication in F. chiloensis.
• •Cultivated F. chiloensis plants harbor more diverse phyla and stress-adaptive BGCs.
• •Frankia dominates wild rhizospheres with nitrogen-fixing and PKS/NRPS gene clusters.
• •Ectoine and NAPAA BGCs emerge in cultivated microbiomes under soil stress.
• •Metabolic modeling links high-biomass taxa to rhizosphere functional traits.

Abstract

Fragaria chiloensis ssp. chiloensis, endemic to Chile, includes a wild form (f. patagonica) and a cultivated “white strawberry” (f. chiloensis) historically grown in central-south Chile. In this study high throughput metagenomic sequencing was employed to examine the rhizosphere microbial communities of wild and cultivated plants were examined to explore how domestication has shaped microbial structure and function. This enabled binning and functional annotations indicating that wild rhizospheres were dominated by Frankia and Bradyrhizobium, whereas cultivated samples exhibited broader genus‐level diversity yet preserving a core ecological function through functional redundancy. Metabolic reconstructions further spotlight high‐biomass taxa, with Frankia in wild plants and Nocardia in cultivated plants, that harbored extensive biosynthetic gene clusters, linking robust growth to specialized metabolite production and putative osmoprotection. Collectively, these findings illustrated how domestication reshapes the rhizosphere microbiome and contributes to One Health-aligned strategies for sustainable agriculture.

Graphical abstract

1. Introduction

The strawberry, Fragaria chiloensis subspecies (ssp.) chiloensis, is native to Chile and is one of the progenitors of the globally consumed strawberry, Fragaria × ananassa Duchesne (Darby and Islam, 2025). In Chile, two botanical forms (f.) of F. chiloensis ssp. chiloensis occur: the f. patagonica, a “wild strawberry” that produces small red fruits, mostly associated with forests, and the f. chiloensis, known as the “white strawberry,” whose fruit exhibits a distinctive pale red or white color and has been cultivated by Chileans for centuries (Finn et al., 2013).

The rhizosphere corresponding to the soil zone influenced by root exudates is a microcosm in which microbial communities profoundly affect plant health and soil quality. The study of Fragaria species has gained significant attention due to their agricultural and ecological importance (Chamorro et al., 2025). A recent transcriptomic analysis indicates that ripening in F. chiloensis fruit is stimulated by abscisic acid (ABA), involving a complete suite of genes responsible for softening, color development, and aroma formation (Gaete-Eastman et al., 2022). ABA, a sesquiterpene, appears to play a crucial role in mutualistic interactions such as mycorrhizal associations and rhizosphere bacterial partnerships with plants (Lievens et al., 2017). Microorganisms in the rhizosphere produce phytohormones such as indole-3-acetic acid (IAA), which enhance plant stress tolerance. They also facilitate biological nitrogen fixation, secrete organic acids that solubilize phosphate and potassium, and synthesize siderophores to sequester iron (Carrasco-Fernández et al., 2020; Lugtenberg and Kamilova, 2009; Spaepen et al., 2007). Consequently, plant–bacterial interactions are essential for nutrient uptake, disease resistance, and environmental adaptability.

Numerous studies have demonstrated that domestication and modern agricultural practices significantly alter these microbial consortia, often leading to reduced diversity and a microbial ecosystem shaped more by anthropogenic influences than by natural selection (Besset-Manzoni et al., 2018; Goss-Souza et al., 2019; Hannula et al., 2017; Shen et al., 2013). Central to understanding these dynamics is the One Health framework, which highlights the interconnectedness of human, animal, and environmental health, emphasizing the crucial role of agricultural microbiomes in maintaining ecological balance and systemic stability (García et al., 2025; Yousuf et al., 2025). In this context, edaphic drivers, including soil nutrient imbalances, metal contamination, and agricultural practices act as strong ecological filters that determine resource availability and select for microbial taxa with specialized metabolic adaptations (Estrada-Villegas et al., 2020).

Comparative metagenomic analyses between wild and cultivated settings provide a powerful approach to link these edaphic factors with microbiome structure and function, enabling the detection of biosynthetic gene clusters, stress response pathways, and host-microbiome interactions that illustrate how soil and rhizosphere metabolic processes support One Health goals (González et al., 2022; Mathew et al., 2025). Sequencing the rhizosphere of wild and white Chilean strawberry will reveal the composition and function of microbial communities and will capture the genetic material of the host plant. A fundamental step in analyzing metagenomic data is binning, which involves grouping sequences into bins that represent individual microbial genomes or closely related taxonomic units, enabling metagenome-assembled genomes (MAGs) reconstruction from complex microbial communities (Kang et al., 2019; Parks et al., 2017). Since the transfer of mobile genetic elements from microbes to plants has been reported (Ku et al., 2021), the proximity of the rhizosphere to plant roots often results in the acquisition of next-generation sequencing (NGS) reads from the plant, providing an opportunity to examine plant genetics. This integrative perspective allows for a comprehensive analysis of plant and microorganism interactions, plant genetic variability, and plant evolutionary dynamics, all of which are essential to sustaining healthy ecosystems and advancing One Health goals (FAO et al., 2023).

The present study uses comparative metagenomic analysis to explore the abundance and variability of the microbial communities in the rhizospheres of the wild strawberry, herein referred to as “wild”, and the white strawberry, herein referred to as “cultivated”, while capturing plant genomic variability at the mitochondrial and chloroplast levels (Carlson et al., 2005; He et al., 2023; Mulugeta et al., 2023; Pyhajarvi et al., 2020). This study represents the first metagenomic-level exploration of the Chilean strawberry. The connections between the microbial ecosystem and the One Health concept, which may be essential for understanding climate resilience in F. chiloensis.

2. Methods

2.1. Sampling and DNA isolation

Rhizosphere samples (FRA-01 to FRA-10) of F. chiloensis were collected from ^{August 1, 2019} to ^{March 9, 2021} across four Chilean regions (Biobío, La Araucanía, Ñuble, and Los Ríos). Sampling sites spanned geographic coordinates ranging from 36.8183° S, 71.6221° W to 39.4057° S, 72.9434° W. Localities included Los Ángeles, Contulmo, Purén, Yungay, Pinto, and Mariquina (Table 1). In the field, root segments (> 500 mg per plant) were excised with sterile scissors and pooled by code into 50 mL conical tubes containing 35 mL of epiphyte removal solution (Simmons et al., 2018). Tubes were shaken for 10 s to detach rhizosphere particles and then snap-frozen in liquid nitrogen. At the laboratory, tubes were thawed briefly and sonicated at 40 kHz for 10 min in an ultrasonic bath to detach adherent soil particles; roots were removed with sterile forceps, and the suspension was centrifuged at 4000 × g for 10 min at 4 °C to pellet the rhizosphere soil. Genomic DNA was extracted from the rhizosphere soil using the DNeasy PowerSoil Pro Kit (QIAGEN) according to the manufacturer’s instructions. Shotgun metagenomic sequencing (≥10 Gb per sample) was performed by Novogene Corporation Inc. (USA).

Table 1.

Metadata for ten rhizosphere samples (FRA-01 to FRA-10) of F. chiloensis collected from August 2019 to March 2021, across four Chilean regions (Ñuble, Biobío, La Araucanía, and Los Ríos).

Sample ID	Sample type*	Date	Latitude (S)	Longitude (W)	Altitude (m.a.s.l.)	Locality	Region
FRA-01	Wild	2019–08–01	−39.405747	−72.943403	111	Mariquina	Los Ríos
FRA-02	Wild	2020–07–13	−36.8183	−71.622067	697	Pinto	Ñuble
FRA-03	Wild	2020–07–13	−36.818367	−71.622217	697	Pinto	Ñuble
FRA-04	Wild	2020–10–14	−36.885217	−71.46865	1480	Pinto	Ñuble
FRA-05	Wild	2020–10–14	−36.886917	−71.469033	1474	Pinto	Ñuble
FRA-06	Wild	2021–01–26	−37.059767	−71.6441	1335	Yungay	Ñuble
FRA-07	Wild	2021–01–26	−37.05905	−71.64625	1307	Yungay	Ñuble
FRA-08	Cultivated	2021–03–09	−38.051667	−73.189467	539	Purén	La Araucanía
FRA-09	Cultivated	2021–03–09	−38.063333	−73.1937	480	Contulmo	Biobío
FRA-10	Cultivated	2021–03–09	−37.396213	−72.622865	87	Los Ángeles	Biobío

^⁎”wild”, corresponds to wild strawberry, F. chiloensis ssp. chiloensis f. patagonica; “cultivated” corresponds to cultivated strawberry, F. chiloensis ssp. chiloensis f. chiloensis.

2.2. Metagenomic assembly and binning

Illumina raw data libraries, labeled F01RW to F10RW, were trimmed using fastp v0.20.0 (removing adapter sequences and bases with Phred score <20) (Chen et al., 2018). Two metagenome assemblers were employed: Megahit v1.2.9 (Li et al., 2016) optimized for large and complex datasets, executed as megahit −1 forward.fq −2 reverse.fq -o output_dir, and metaSPAdes v3.15.3 (Nurk et al., 2017) which uses multiscale k-mer assembly, executed as metaspades.py −1 FXXRW.R1.fq.gz −2 FXXRW.R2.fq.gz -t 20 -o FXXRW_metaspades. metaSPAdes assemblies were chosen based on superior N50 values. Contigs from wild (F01RW–F07RW) and cultivated (F08RW–F10RW) samples were concatenated into separate catalogs and binned using Vamb v3.0 (Nissen et al., 2021), with a minimum contig length of 2000 bp, leveraging a variational autoencoder to cluster contigs by sequence composition and coverage. The merged assemblies were polished with hypo (https://github.com/kensung-lab/hypo) for consensus error correction, standardized in Anvi’o platform v7.0 (Eren et al., 2015). The Genome Taxonomy Database (GTDB) (Parks et al., 2022) was downloaded through Anvi'o and a supervised binning against the GTDB database using SemiBin tool was performed (Pan et al., 2022) with the command: SemiBin single_easy_bin -i contigs-fixed.fa -b ./S0*.bam –processes 40 –reference-db-data-dir /path/to/GTDB -o SemiBin output. The binned contigs were further manually refined using Anvi'o. Detailed read-mapping, variant-calling and SnpEff annotation procedures for mitochondrial and chloroplast genomes are provided in the Supplementary Material section.

2.3. Cluster analysis

All biosynthetic gene cluster analyses for the wild and cultivated F. chiloensis samples were performed using nextflow (Di Tommaso et al., 2017) by running commands with the nf-core/funcscan pipeline (https://nf-co.re/funcscan) (Ewels et al., 2020), referencing a user-provided CSV that mapped each bin’s FASTA file. The pipeline invoked antiSMASH (Blin et al., 2021) for cluster annotations, storing results in designated directories for each bin. Visualization involved running Clinker (Gilchrist and Chooi, 2021) on the resulting “c_*.gbk” files to produce comparative HTML plots for the wild and cultivated datasets. SVG generated files were then converted to PDF with the Inkscape command-line interface, following installation of Python libraries (rectpack, svgwrite, cairosvg, lxml, and svgutils). This workflow collectively facilitated the identification, annotation, and visualization of specialized metabolite gene clusters within the F. chiloensis rhizosphere metagenomes.

2.4. Pangenomic analyses and pathway reconstruction

Gene calls were obtained in GFF (General Feature Format) from the entire metagenomes using anvi’o’s anvi-get-sequences-for-gene-calls, then parsed bin-specific contigs with a custom BASH script. Separate pangenomes for wild and cultivated Fragaria chiloensis datasets were calculated using Roary (Page et al., 2015) which clusters genes into orthologous groups at ≥ 95 % protein-sequence identity (BLASTP + MCL; –identity 95) and designates core genes as those present in ≥ 99 % of samples (–core_definition 0.99).

To reconstruct metabolic pathways, anvi-run-kegg-kofams was run against the KOfam database (Aramaki et al., 2020) and then anvi-estimate-metabolism was used in bin-mode to calculate module completeness scores—each score defined as (number of required KOfam hits detected / total KEGG Module markers) × 100, with a default threshold of 75 % for calling a module “present” (see https://anvio.org/help/7/programs/anvi-estimate-metabolism/). Enriched pathways were parsed in Python into eight categories (amino acid metabolism; nitrogen and nitrate metabolism; carbohydrate metabolism; lipid metabolism; cofactor and vitamin metabolism; energy metabolism and electron transport chain; nucleotide metabolism; other) and the results were visualized using the seaborn and matplotlib Python libraries.

2.5. Gapseq analysis

The gapseq tool (Zimmermann et al., 2021) was used to reconstruct each bin’s metabolic network and predict potential metabolite exchange (influx and efflux) under defined conditions, thereby revealing functional capabilities and possible cross‐feeding interactions in the rhizosphere. Contigs from wild and cultivated F. chiloensis metagenomic bins were annotated against a curated reaction database combining MetaCyc (Caspi et al., 2016), KEGG (Kanehisa et al., 2021) and ModelSEED (Seaver et al., 2021) databases. Metabolic pathways were predicted and gap-filling per bin was performed with the gapseq_find.sh script using -v 0 -b 200 -p all -t auto. Sequential gap-filling steps were applied to model metabolite influx and efflux: gapseq find-transport, gapseq draft -u 200 -l 100, gapseq medium, and gapseq fill -b 100. Flux balance analysis yielded metabolite fluxes in mmol/(gDW·hr); compounds with near-zero flux in ≥ 90 % of bins were filtered out, and the remaining influx/efflux profiles were plotted for each bin in both wild and cultivated F. chiloensis metagenomes.

3. Results

3.1. Characterization of rhizosphere metagenomes from wild and cultivated F. chiloensis

The rhizosphere microbial communities were profiled for the first time in the wild and cultivated Chilean F. chiloensis. After performing metagenomic binning and curation of metagenome-associated genomes, microbial bins from various genera were identified, commonly associated with plant rhizospheres (Wang et al., 2024). In the wild rhizosphere, metagenomes of Methylocella, Mycobacterium, Reyranella, AP-15, Frankia, Fen-1137, Streptomyces, Pseudomonas, UBA7541, Bog_209, and an unclassified Bacteria, among others were identified (Fig. 1A, 1C). The most abundant bins corresponded to Bradyrhizobium (Bin 16) and Alphaproteobacteria (Bin 15), featuring a genome assembly completeness that varied from 0 % (indicating highly fragmented or absent MAGs) to 84.51 % (for Bin 2, Frankia, Supplementary Table 1). At the phylum level, Pseudomonadota was represented by Methylocella, Reyranella, and other bins, while Actinobacteriota included Mycobacterium and Streptomyces bins. Acidobacteriota and Myxococcota were also present, represented by Edaphobacter, UBA7541, Bog_209, and Fen-1137 bins, respectively (Supplementary Table 1).

Fig. 1.

Comparative metagenomic analysis of the rhizosphere of F. chiloensis under wild and cultivated conditions. (A) Phylogenetic tree depicting the diversity of microbial taxa identified from the rhizosphere metagenome of wild F. chiloensis plant roots (n = 7). Each colored branch represents a taxonomically distinct bin, and the taxonomy was determined using the Kaiju and GTDB databases. Adjacent barplots, situated above the graph, show sequencing statistics, including the number of mapped reads, total retained reads, reported insertion-deletions (INDELs), and reported Single Nucleotide Variants (SNVs) for each microbial bin. Each contig was hierarchically positioned on the plot based on its tetranucleotide frequency and GC content. (B) Same as (A) depicting the microbial diversity in the rhizosphere metagenome of cultivated F. chiloensis plant roots (n = 3). The accompanying barplots indicate sequencing statistics comparable to those in panel (A), for the identified microbial taxa across the metagenome. (C) Heatmap displaying the relative abundance of the identified microbial bins from the rhizosphere of seven wild F. chiloensis plant roots. The color gradient represents the percentage of abundance for each bin with respect to the total binned metagenome per sample (FRA-01 to FRA-07). (D) Same as (C) for the relative abundance of microbial bins from the rhizosphere of three cultivated F. chiloensis plant root metagenomes. The color gradient represents the percentage of abundance for each bin with respect to the total binned metagenome per sample (FRA-08 to FRA-10).

Coverage-filtered mapping of non-recruited reads against F. chiloensis and F. × ananassa organelle reference genomes confirmed that both mitochondrial and chloroplast genomes were represented at ≥30 ×, enabling high-confidence SNP/indel detection. Among the sequenced samples, chloroplast DNA of F. chiloensis showed a 4- to 5-fold higher mutation rate, i.e., enrichment of polymorphisms, relative to the mitochondrial genome (Supplementary Fig. 1).

The rhizosphere metagenome from cultivated strawberry included bacterial bins from a wide range of genera such as Rhodanobacter, Humibacter, Edaphobacter, Nocardia, Burkholderiaceae, Streptacidiphilus, Mycobacteriaceae, Rhodoferax, Acidocella, Saccharimonadaceae, UBA4665, 2011-GWC2–44–17, Trinickia, DP-20, AC-32, Rhizobium, Methylovirgula, Arachidicoccus, Amycolatopsis, and Acidocella (Fig. 1B, 1D). In contrast to the wild, the most abundant members in these metagenomes were Rhodanobacter_sp00201105 (Bin 34) and AC-32 (Bin 24). The percentage completion ranged from 0 % to 98.59 % (for Bin 4, Arachidicoccus, Supplementary Table 1). Like the wild, and at the phylum level, Pseudomonadota and Actinobacteriota were present. However, the cultivated strawberry metagenomes also included bins from Acidobacteriota, Patescibacteria, Desulfobacterota, and Bacteroidota phyla, indicating a more diverse range of phyla (Supplementary Table 1).

Both metagenomes from the wild and cultivated plants shared the presence of Pseudomonadota and Actinobacteriota phyla, abundant members at the plant rhizosphere level (Chen et al., 2021). However, the wild had a higher representation of Acidobacteriota and Myxococcota, while the cultivated strawberry showed a more diverse range of phyla, including Patescibacteria, Desulfobacterota, and Bacteroidota. Commonalities, such as the presence of nitrogen-fixing bacteria Bradyrhizobium (Bin 16 in the wild), Streptomyces (Bin 4 in the wild) or Rhizobium (Bin 33 in cultivated plant rhizosphere), indicate key microbial taxa that were robust across among the rhizosphere microbial communities in F. chiloensis. Despite the observed differences in bacterial phyla between the wild and cultivated plant metagenomes, it cannot be discarded that diverse bacterial phyla can perform similar functions in both settings. This is particularly relevant given the complex and dynamic nature of microbial communities, where different species can often fulfill similar ecological roles, a concept known as functional redundancy. Therefore, while the specific bacterial phyla present may vary between the wild and cultivated plants, the overall functionality of the microbial community could remain consistent.

Thus, the analysis of the rhizosphere metagenomes from wild and cultivated F. chiloensis revealed a rich and diverse array of bacterial species. The wild rhizosphere was characterized by the prevalence of nitrogen fixing bacteria like Bradyrhizobium and a broad representation of Alphaproteobacteria, suggesting a microbial community well-adapted to the natural soil environment with a capacity for symbiotic relationships as seen in other species (Leff et al., 2017; Zhang et al., 2019). In contrast, the rhizosphere of cultivated F. chiloensis demonstrated a different microbial profile, with Rhodanobacter and AC-32 being more abundant, reflecting a profound influence of agricultural practices on F. chiloensis rhizosphere composition (Siegieda et al., 2024). Despite these differences, commonalities such as the presence nitrogen-fixing bacteria indicate minimal microbial taxa that are necessary across different environmental conditions for the growth of F. chiloensis.

3.2. Biosynthesis gene cluster (BCG) analysis of metagenomes derived from the rhizosphere of wild and cultivated F. chiloensis

Comparative analysis showed that wild F. chiloensis metagenomes harbor unique biosynthetic gene clusters (BGCs), illustrated by cluster regions from distinct bins, highlighting BGCs for hydrogen cyanide (HCN) synthesis, ribosomally synthesised and post-translationally modified peptides (RiPPs), siderophores, lasso peptides, terpenes, non-ribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs) (Fig. 2A–E). Frankia (Bin 2) (Fig. 2A), depicted multiple PKS and NRPS clusters, along with genes involved in indole, lasso peptide, terpene, and siderophore BGCs, emphasizing the substantial specialized metabolic capacity of Frankia in this environment (Udwary et al., 2011). Notably, additional Frankia bins (not shown) also harbored nitrogen fixation genes alongside analogous clusters, suggesting that Frankia may contribute both symbiotic nitrogen supply and bioactive metabolite production in F. chiloensis rhizospheres in natural settings (Marappa et al., 2020). In Bin 1 Mycobacteriales (Fig. 2B), revealed terpene biosynthesis pathways and PKS clusters that could confer competitive or protective roles in the rhizosphere, already described for this genus in literature (Mayfield et al., 2024; Quadri, 2014). Bin 4 Streptomyces (Fig. 2C) featured a combined PKS–NRPS cluster typical of the genus’s capacity to produce bioactive compounds such as antibiotics, which may modulate nutrient dynamics and suppress pathogens (Khan et al., 2023; Vurukonda et al., 2018). Bog-209 sp003135495 (Bin 7) (Fig. 2D) contains genes for terpene production and redox cofactors, potentially enhancing microbial survivability under oxidative stress and influencing soil nutrient cycling. Finally, Reynarella (Bin 13) (Fig. 2E) included genes for RiPP biosynthesis and HCN production, which can foster antagonistic activity against soilborne pathogens and underscore the defensive potential of these rhizosphere microbes (da Cruz Nizer et al., 2023). These specialized gene clusters build on earlier observations that the F. chiloensis rhizosphere was inhabited by a taxonomically and functionally diverse community, including nitrogen fixers and other beneficial microbes, most notably Frankia, which possesses important biosynthetic potential.

Fig. 2.

Representative biosynthetic gene clusters (BGCs) from metagenomic bins in wild F. chiloensis rhizosphere samples. (A) Genomic region from Bin 2, identified as Frankia. (B) Genomic region from Bin 1, affiliated with Mycobacteriales. (C) Genomic region from Bin 4, classified as Streptomyces. (D) Genomic region from Bin 7, assigned to Bog_209 sp003135495. (E) Genomic region from Bin 13, related to Reynarella.

The wild rhizosphere bins predominantly contain PKS, NRPS, RiPP, siderophore, terpene, and HCN clusters, particularly in Frankia, yet lack many of the more specialized or stress‐adaptive BGCs found in the cultivated F. chiloensis rhizosphere (Fig. 3A–E). In particular, ectoine and NAPAA biosynthetic genes emerge only among the cultivated Fragaria rhizosphere metagenomes, suggesting additional osmotic‐stress tolerance (e.g., in Nocardia and Mycobacteriaceae) (Santos and da Costa, 2002) and potential antimicrobial polymer production based on NAPAA polypeptides (Luz et al., 2018). These findings illustrate how closely related rhizosphere communities can differ in their metabolic pathways, reflecting environmental pressures in the cultivated Fragaria metagenome. Ectoine, for instance, stabilizes proteins, and cell membranes under drought or salinity stress, conferring a competitive edge in managed or fluctuating soils and potentially benefiting the plant host (Eswaran et al., 2024). This function aligns with NAPAA peptide production in terms of antimicrobial defense.

Fig. 3.

Representative biosynthetic gene clusters (BGCs) from metagenomic bins in cultivated F. chiloensis rhizosphere samples. (A) Genomic region from Bin 14, affiliated with Nocardia. (B) Genomic region from Bin 3, classified as Trinickia sp013282575. (C) Genomic region from Bin 15, assigned to Burkholderiaceae. (D) Genomic region from Bin 18, related to Mycobacteriaceae. (E) Genomic region from Bin 23, identified as Streptacidiphilus.

An expanded repertoire of biosynthetic gene clusters (BGCs) among cultivated F. chiloensis rhizosphere taxa, specifically in Nocardia (Bin 14), Trinickia sp013282575 (Bin 2), Burkholderiaceae (Bin 15), Mycobacteriaceae (Bin 18), and Streptacidiphilus (Bin 23) was observed (Fig. 3). Nocardia in Bin 14 possesses ectoine and non‐alpha poly‐amino acid (NAPAA) biosynthetic genes, along with class III lanthipeptides and non‐ribosomal peptide (NRP) metallophore clusters that did not appear in the wild bins. Meanwhile, Trinickia sp013282575 (Bin 3) encodes phosphonate, terpene, aryl polyene, quorum‐sensing homoserine lactone (AHL), and HCN pathways, and the related Bin 15 (Burkholderiaceae) similarly expresses phosphonate and AHL genes absent from the Frankia‐dominated wild metagenomes. Bin 18 (Mycobacteriaceae) features PKS and NRPS homolog clusters. Streptacidiphilus (Bin 23) shows multiple type I PKS, NRPS, and siderophore clusters, indicative of this genus’s prolific capacity for specialized metabolite production (Fig. 5).

Fig. 5.

Metabolic reconstruction of cultivated F. chiloensis rhizosphere metagenomes. (A) Metabolic enrichment analysis terms associated with amino acid metabolism in the cultivated root metagenome. Heatmap of enrichment scores (purple scale, 0–20) for amino-acid biosynthesis and degradation pathways across 32 metagenome-assembled genomes. (B) Same as (A) for carbohydrate metabolism. (C) Same as (A) for energy metabolism. (D) Same as (A) for nitrogen and nitrate metabolism.

3.3. Soil nutrient stress drives microbial metabolic adaptations in cultivated F. chiloensis rhizosphere

Notably, soil analyses revealed that in the cultivated F. chiloensis rhizosphere from the sampling sites, potassium, calcium, iron, and manganese were scarce, while nitrogen phosphorus, aluminum, copper, and boron were in excess compared to wild scenario (Supplementary Fig. 2). The deficiency of Fe often recruit bacteria capable to chelate metals (Molnar et al., 2023), the excess of copper is toxic (Bernard et al., 2009; Bondarczuk and Piotrowska-Seget, 2013; Cervantes and Gutierrez-Corona, 1994) and bacteria often release organic acids and/or produce proteins to solubilize K and P (Thepbandit and Athinuwat, 2024; Wu et al., 2023). These imbalances create a challenging environment that select microbial taxa with specific adaptive strategies, for example, the low availability of nitrogen found in wild strawberry plants, which could explain the recruit of nitrogen fixing bacteria like Bradyrhizobium. As mentioned, in the cultivated metagenome, the expanded BGC repertoire among Nocardia (Bin 14), Mycobacteriaceae (Bin 18), and other taxa reflects such adaptations. Ectoine and NAPAA clusters, for example, produced compounds that stabilize proteins and cell membranes under osmotic stress and help detoxify excess metals, conditions consistent with high levels of Al and Cu in the cultivated soils. Likewise, Trinickia sp013282575 (Bin 3) expresses BGCs for phosphonate, terpene, aryl polyene, and quorum‐sensing homoserine lactone (AHL) production, a likely strategy to manage surplus phosphorus and counter stress induced by toxic metal concentrations. Similarly, Burkholderiaceae (Bin 15) and Streptacidiphilus (Bin 23) contribute additional specialized pathways, phosphonates, AHL quorum‐sensing, and siderophores that are absent in the wild rhizosphere bins, underscoring a distinct metabolic adaptation to these nutrient imbalances.

Overall, the discovery of unique clusters, for instance, phosphonates, AHL quorum‐sensing (associated with biofilm formation and virulence gene expression (Li and Nair, 2012)), lanthipeptides with antimicrobial activity (Repka et al., 2017), and metallophores for chelating metals from polluted environments (Gomes et al., 2024), underscores the broader range of biosynthetic capabilities in cultivated rhizosphere samples. Alongside the presence of ectoine and NAPAA, these features highlight a strong emphasis on microbial defense against pathogens, high metal concentrations, and other anthropogenic pressures, reflecting a profound influence of agricultural practices. Thus, the interplay between soil chemistry and microbial genomic diversity reveals how the rhizosphere of cultivated F. chiloensis fosters a community with enhanced stress tolerance and a heightened capacity for specialized metabolite production in contrast to the Frankia‐dominated wild rhizosphere, which displays fewer stress‐adaptive or osmoprotective clusters but still harbors significant potential for nitrogen fixation and biosynthesis.

3.4. Comparative metabolic analysis of microbial rhizosphere populations reveals bacteria with high metabolic capabilities

Pangenomic analysis of microbial rhizosphere populations did not reveal any obvious core genes shared between wild and cultivated communities. However, phylogenetic relationships were observed among certain bins, such as Mycobacteriales (Bin 1) and Frankia (Bin 2) in the wild rhizosphere (Fig. 4A), and Edaphobacter (Bin 13), unclassified Bacteria (Bin 1), Burkholderiaceae (Bin 15), and Rhodoferax (Bin 22) in the cultivated rhizosphere (Fig. 4B). These relationships prompted us to explore whether key metabolic capabilities might be conserved functionally, even in the absence of core genes. Anvi’o platform was used to estimate the metabolic capabilities of each bin in both wild and cultivated metagenomes using the KOFam database. Subsequently, the enriched metabolic terms were calculated for each bin. In the wild rhizosphere, Frankia (Bin 2) showed the most significant enrichment in amino acid metabolism, while Reynarella (Bin 13) and Mycobacteriales (Bin 1) displayed distinct metabolic activities, including serine, leucine, and threonine metabolism not observed in Frankia (Fig. 4C). This Frankia MAG represents a high-quality, putatively novel species-level lineage, distantly related to reference genomes in NCBI, as supported by ANIb/ANIm and pangenomic analysis (Supplementary Fig. 3A–C).

In terms of carbohydrate metabolism, Frankia (Bin 2), Methylocella (Bin 10), Reynarella (Bin 12), and Mycobacterium (Bin 11) displayed pathways that cooperate with Crassulacean Acid Metabolism (CAM) metabolism in plants, glycogen biosynthesis, the Entner–Doudoroff pathway, and the reductive pentose phosphate cycle, respectively (Fig. 4D). Frankia (Bin 2) exhibited the highest enrichment in overall energetics (Fig. 4E) and lipid metabolism (Fig. 4F), but not in nitrogen/nitrate metabolism (Fig. 4G). Notably, dissimilatory nitrate reduction (nitrate to ammonia, with electron flow and energy synthesis/ATP synthesis) and assimilatory nitrate reduction (incorporation of nitrogen in the environment in the form of nitrate) terms were enriched in Alphaproteobacteria (Bin 15) and Reynarella (Bin 13), respectively. These bacteria did not otherwise display significant enrichment in the metabolic functions (Fig. 4 G).

Fig. 4.

Metabolic reconstruction of wild F. chiloensis rhizosphere metagenomes. (A) Pangenome of wild plant root microbe bins represented with a phylogenetic tree (16 microbial bins) alongside with a gene matrix produced with Roary (29,793 gene clusters, dark blue bars). (B) Similar to (A), for the pangenome of cultivated root microbe bins (32 strains, 59,843 gene clusters). (C) Metabolic enrichment analysis terms associated with amino acid metabolism. These enriched terms were obtained using the anvi-estimate-metabolism command from Anvi'o to estimate the metabolic capabilities of the metagenome utilizing the KOfams database as a reference. In the y-axis, the enriched scores for the enriched metabolic terms are plotted across the bin IDs (x-axis). (D) Similar to (C) for carbohydrate metabolic pathways. (E) Similar to (C) for direct energy pathways (F) Similar to (C) for lipid metabolism. (G) Similar to (C) for nitrogen and nitrate metabolism.

In the cultivated metagenome, amino acid metabolism was primarily shared among Trinickia (Bin 2), Arachidicoccus (Bin 4), and Streptacidiphilus (Bin 23), among others (Fig. 5A). The same bins, along with Nocardia (Bin 14), dominated carbohydrate metabolism (Fig. 5B). Among these bins, energy metabolism was enriched in Arachidicoccus (Bin 4), Streptacidiphilus (Bin 23), and Trinickia (Bin 2) (Fig. 5C), while lipid metabolism enrichment was observed in Arachidicoccus (Bin 4), Nocardia (Bin 14), and Burkholderiaceae (Bin 15) (Fig. 5D). As seen in the wild metagenome, dissimilatory nitrate reduction was not enriched in these bins with the most enriched metabolic terms but rather in Humibacter (Bin 12) (Fig. 5E). These patterns suggest functional complementarity among microbial bins in both wild and cultivated rhizospheres, with different organisms contributing distinct, but potentially cooperative, metabolic roles that support the F. chiloensis rhizosphere.

Overall, these results align with the observed distribution of biosynthetic gene clusters (BGCs) (Figs. 3 and 4). While Frankia (Bin 2) in the wild rhizosphere was enriched in nitrogen fixation, PKS, and NRPS clusters, bins in the cultivated rhizosphere such as Nocardia (Bin 14), Trinickia (Bin 2), and Burkholderiaceae (Bin 15) harbored BGCs involved in ectoine, NAPAA, phosphonate biosynthesis, AHL quorum sensing, and other specialized metabolites. This functional alignment suggests that microbial community structure is shaped by ecological pressures, such as cultivation, selecting for taxa with broader metabolic and biosynthetic capacities essential for nutrient cycling, stress tolerance, and pathogen defense in the F. chiloensis rhizosphere.

3.5. Metabolic pathway prediction and model reconstruction in wild and cultivated rhizosphere metagenomes

Metabolic pathways were predicted, and metabolic models were reconstructed based on genomic information from the bins of both wild and cultivated rhizosphere metagenomes. The gapseq tool was used, which employs curated databases for pathways and reactions such as KEGG and MetaCyc and includes gap-filling in simulations with defined bacterial communities. The simulation resulted of bacterial growth from the wild binned plant root metagenome under ideal nutrient conditions demonstrated that Frankia (Bin 2) reached the highest biomass in the wild metagenome (Fig. 6A). This could be due to Frankia (Bin 2) achieving the highest completeness in the wild metagenome, but it is also known for the high biosynthetic and metabolic capability of the Frankia genus. In the growth simulation for bacterial bins from the cultivated plant root metagenome, Nocardia (Bin 14) and Streptacidiphilus (Bin 23) achieved the most substantial predicted biomass (Fig. 6B).

Metabolic fluxes were also predicted for various substances within different bacterial bins found in the root metagenome from wild plants. Highly abundant substances such as H⁺, H₂O, NH₃, CO₂, and acetate were mainly produced by bacteria, with limited consumption by certain bins (i.e., UBA7541 (Bin 6) and Alphaproteobacteria (Bin 15) consumed water, while others produced it (Fig. 6C, left heatmap). Interestingly, l-cysteine and l-arginine were predicted to be consumed from the medium by Reynarella (Bin 13) and Mycobacteriales (Bin 1), respectively (Fig. 6C, left heatmap). Regarding middle-to-low produced/consumed substances, butyrate was produced by Fen_1137 (Bin 3) and AP_15 (Bin 14). Additionally, AP_15 (Bin 14) and UBA7541 (Bin 6) were predicted to produce succinate, MTTL (methanethiol), propionate, and indole, respectively (Fig. 6C, right heatmap). After treating each bin’s net flux as one observation and applying Wilcoxon signed-rank tests with Benjamini–Hochberg correction (FDR < 0.05), several metabolites whose community-level fluxes differed significantly from zero were identified. While O₂, l-cysteine, and l-arginine were consumed, NO, acetate, H₂, CO₂, and H⁺ were produced. The secondary metabolite indole also showed significant export (Fig. 6E and Supplementary Table 3).

Fig. 6.

Bacterial growth and metabolic flux analysis of wild and cultivated F. chiloensis rhizosphere microbiomes. (A) Simulation results of bacterial growth from the wild plant root metagenome using the gapseq tool under ideal nutrient conditions. The bar chart displays the biomass (in femtograms) produced by each bacterial bin. (B) Same as (A) for the growth simulation for bacterial bins from the cultivated plant root metagenome, also derived using the gapseq tool. (C) heatmaps representing the predicted metabolic fluxes of various substances within different bacterial bins found in the root metagenome from wild plants, obtained with gapseq metabolic reconstruction tool. (left) Heatmap shows highly abundant substances, including gases (like CO₂, H₂), ions (like H⁺), and organic compounds (such as acetate and lactate), with a color gradient representing flux values in mmol/(gDW*hr). Red indicates production while blue indicates consumption. (right) Heatmap depicting same as left but for middle-to-low abundant substances encompassing organic and inorganic compounds. Red indicates production while blue indicates consumption. (D) Same as (C) for the root metagenome from cultivated plants.

For the cultivated root binned metagenome, similar conclusions could be drawn with some differences. Consistently with the predicted flux in the wild metagenome, CO₂, acetate, and H⁺ were mainly produced by bacteria (Figure 6D, left heatmap). Water was predicted to be produced by several bins such as DP_20 (Bin 31), Actinomycetia (Bin 20), Streptacidiphilus (Bin 23), Edaphobacter (Bin 13), and Mycobacteriaceae (Bin 18), among others, while Nocardia (Bin 14) and Actinomycetia (Bin 19) consumed it (Fig. 6D, left heatmap). Interestingly, almost all bins consumed d-fructose, but unclassified Bacteria (Bin 9) was predicted to produce it in high quantities (Figure 6D, left heatmap). As seen in the wild metagenome, l-cysteine is predicted to be consumed from the medium by Nocardia (Bin 19) and Actinomycetia (Bin 30), and l-arginine is also consumed by three bacteria, including Arachidicoccus (Bin 4) (Fig. 6D, left heatmap). Regarding middle-to-low produced/consumed substances, among the substances produced by different bacteria in the wild metagenome, only MTTL and indole were predicted to be produced by Methylovirgula (Bin 5) and Acidocella (Bin 24), respectively, but in much lower quantities than in the wild metagenome (Fig. 6D, right heatmap). Overall, the cultivated community consumed the same core metabolites as in the wild (Fig. 6F vs. 6E, top) and retained significant export of core metabolites (CO₂, acetate, H⁺), including H₂S—a metabolite involved in bacterial defense (Toliver-Kinsky et al., 2019) and in plant stress responses (Li et al., 2022) (Fig. 6F and Supplementary Table 3).

In conclusion, several metabolites were produced and consumed by different microbes in the wild and cultivated metagenomes, with a preference for l-cysteine and l-arginine consumption and production of indole and methanethiol (MTTL), which is predicted to be marginally produced in the cultivated metagenome compared to the wild. Overall, these results complement the earlier findings on specialized biosynthetic gene clusters (BGCs). The high biomass and metabolic fluxes predicted for Frankia (Bin 2) in the wild and Nocardia (Bin 14) and Streptacidiphilus (Bin 23) in the cultivated rhizosphere align with their expanded BGC repertoires, reflecting key ecological roles in nutrient cycling, stress adaptation, and potential antimicrobial defense. Taken together, the integration of metabolic reconstructions and BGC analyses highlights how both common and specialized pathways shape the functional dynamics of the F. chiloensis rhizosphere microbiome, underlining distinctive adaptive strategies in wild versus cultivated environments.

4. Discussion

This study aimed to characterize microbial community dynamics and predict metabolite production in the rhizosphere of the Chilean F. chiloensis, a species of strawberry, in both wild and cultivated environments. The rhizosphere microbiome is pivotal for plant health and productivity (Perez-Jaramillo et al., 2016). Classic studies and reviews have underscored that plants actively recruit beneficial rhizosphere microorganisms through root exudates, releasing sugars, amino acids, and secondary metabolites that shape microbial community composition (Pantigoso et al., 2022; Perez-Jaramillo et al., 2016), in the long-term affecting biogeochemical cycling, plant growth and tolerance to biotic and abiotic stress (Philippot et al., 2013).

Unlike amplicon surveys that only profile community composition, shotgun metagenomics retrieves genetic information on functional genes and pathways present in the rhizosphere, revealing key factors in microbial interactions. This functional insight is central to the One Health framework, which recognizes that soil microbial communities influence plant health and productivity, shaping ecosystem resilience, food safety, and environmental sustainability through complex biochemical interactions (Trivedi et al., 2020).

In the wild rhizosphere of F. chiloensis, bacterial species from genera such as Methylocella, Mycobacterium, Reyranella, and others are present, with Frankia being the most abundant. It has been shown that Frankia grows well in natural rhizospheres thanks to host-derived compounds and interactions that are absent in axenic cultures. For instance, in Casuarina cunninghamiana, it’s been shown that root exudates to a propionate‐based minimal medium significantly enhanced the biomass yield of Frankia sp. CcI3 in vitro (Beauchemin et al., 2012; Ghodhbane-Gtari et al., 2014). Moreover, genomic analysis of the uncultured cluster‐2 strain ‘Candidatus Frankia datiscae’ Dg1 revealed substantial genome reduction, including loss of stress‐response genes, implying a dependency on symbiotic factors not supplied by standard laboratory media (Persson et al., 2015). Therefore, to achieve robust ex planta growth, Frankia must be supplemented with the complete suite of host-derived exudates and signaling molecules characteristic of the F. chiloensis rhizosphere. Moreover, the wild soil microbiome, enriched in Acidobacteriota and Myxococcota, provides complementary metabolic functions and community interactions that further foster Frankia proliferation. Frankia, an actinorhizal microsymbiont, is known for its significant biosynthetic potential (Udwary et al., 2011), also shown here as it dominates enriched metabolism in aminoacids, carbohydrates, lipids, among other terms. Nevertheless, the results suggested that Frankia members found in the wild rhizosphere of F. chiloensis did not fix nitrogen and conversely, Alphaproteobacteria and Reynarella (also Alphaproteobacteria) members did, a function previously described for bacterium belonging to this class (Tsoy et al., 2016). In the case of the metagenomes from the rhizosphere from cultivated F. chiloensis, the biosynthetic function that Frankia displayed on the wild metagenome, can be attributed to bacteria belonging to Arachidicoccus, Methylovirgula and Streptacidiphilus. The nitrogen fixation in the cultivated environment may be attributed to the environmental Burkholderia sensu lato member Trinickia (Estrada-de Los Santos et al., 2018).

A metabolic flux analysis of the wild F. chiloensis rhizosphere metagenome (Fig. 6C) revealed a detailed picture of which microbes produce or consume key metabolites. Notably, CO₂ production is highest in AP_15 (Bin 14, Stellaceae Family, order Rhodospirillales), while H⁺ was predominantly generated by Alphaproteobacteria (Bin 15). Acetate was produced by Streptomyces (Bin 4), Frankia (Bin 2), and Methylocella (Bin 10), underscoring the overlapping yet complementary roles these taxa play in carbon cycling. Meanwhile, L‐cysteine was consumed solely by Reynarella (Bin 13), suggesting a specialized function in amino acid turnover within this community. These microbial interactions illustrate how rhizosphere communities contribute to key ecosystem processes such as nutrient cycling, soil pH regulation, and organic acid turnover, factors that directly influence plant health and indirectly affect environmental stability and food production. This highlights the relevance of adopting a One Health perspective, in which the microbiome is recognized as a central mediator of the interconnected health of soils, crops, and ecosystems (Trivedi et al., 2020).

A particularly noteworthy finding is the production of both indole and MTTL by AP_15 (Bin 14) and, in the case of indole alone, by UBA7541 (Bin 6, an uncultivated Acidobacteriia). Positive flux for indole in these two bins indicated that both may synthesize this key signaling compound, IAA, one of the most abundant phytohormones in plants, which is known to influence plant growth, development, and stress responses (Sun et al., 2022). MTTL flux is additionally detected in AP_15 (Bin 14), emphasizing that a single bacterium can generate multiple ecologically important metabolites, supporting nutrient cycling, intercellular signaling, and possibly even defense. Intriguingly, AP_15 (Bin 14) also showed enriched terms for dissimilatory nitrate reduction (nitrate to ammonia), highlighting a role in nitrogen cycling. As for UBA7541 (Bin 6), it has been proposed to rename this taxon ‘Candidatus Acidoferrum’, reflecting potential iron‐cycling capabilities within the Acidoferrales lineage (Epihov et al., 2021). The simultaneous production of indole by this bin hints at a broader ecological strategy: coordinating metal homeostasis with biochemical signaling in the rhizosphere. Also, indole is produced from tryptophan via bacterial tryptophanase, highlighting the tight coupling between primary amino acid metabolism and secondary signaling functions that underpin rhizosphere adaptation (Roager and Licht, 2018; Tennoune et al., 2022).

The cultivated F. chiloensis rhizosphere exhibited a rich diversity of bacterial bins, including Rhodanobacter, Humibacter, Edaphobacter, and Nocardia, with Burkholderia sensu lato being the most abundant genus. Taxonomically, these samples also encompass a broader range of phyla, such as Patescibacteria, Desulfobacterota, and Bacteroidota, than the wild rhizosphere. Metabolic flux analyses (Fig. 6D) indicate that CO₂ is produced most strongly i.e.: by Humibacter (Bin 12) and Actinomycetia (Bin 20), while acetate is generated predominantly by Bog_1198 (Bin 25), Nocardia (Bins 14 and 19), and certain Actinomycetia (Bin 20). H⁺ appears to be contributed mainly by Methylovirgula (Bin 5), Actinomycetia (Bin 30), and a few additional bins. Several microbes, including Nocardia (Bin 14) and Actinomycetia (Bin 19), also show positive flux for water production, reflecting shared or overlapping metabolic roles in the community.

The cultivated F. chiloensis rhizosphere exhibited a rich diversity of bacterial bins, including Rhodanobacter, Humibacter, Edaphobacter, and Nocardia, with Burkholderia sensu lato being the most abundant genus. Taxonomically, these samples also encompass a broader range of phyla, such as Patescibacteria, Desulfobacterota, and Bacteroidota, than the wild rhizosphere (Figs. 1B and 6D, left). In the cultivated F. chiloensis rhizosphere, several bins stood out for their strong contributions to carbon and nitrogen fluxes. Nocardia (Bins 14, 19) and Burkholderiaceae (Bin 15) produce substantial amounts of CO₂, likely reflecting a high respiratory activity stimulated by agriculture practices as tillage (Sainju et al., 2008). These same bins also generate notable levels of acetate, indicating robust fermentation or incomplete oxidation pathways (Fig. 6D, left). Meanwhile, negative or positive H₂O flux values differ greatly among microbes; for instance, Nocardia (Bin 14) shows significant water consumption, whereas Burkholderiaceae (Bin 15) produces H₂O (Fig. 6D, left). Regarding ammonia (NH₃), Nocardia (Bins 14 and 19) appeared prominent, suggesting catabolic release of ammonium in the rhizosphere; this result is supported by empirical ammonia production by Norcardia species (Alotaibi et al., 2023). In contrast, D‐fructose was heavily produced by Bin 9 Bacteria, potentially supplying a carbon and energy sources to stimulate organic P mineralization other community members (Wang et al., 2024). Amino acids such as L‐cysteine and L‐arginine generally showed negative flux, pointing to net consumption by organisms like Nocardia (Bin 19) or Arachidicoccus (Bin 4). Lastly, specialized metabolites including indole and MTTL appeared at lower flux in the cultivated environment, with Methylovirgula (Bin 5) and Acidocella (Bin 24) as minor producers, presumably because exogenous inputs reduce microbial reliance on these signaling or defensive compounds (Fig. 6D, right). Ecologically, these flux patterns suggest that intensive cultivation fosters microbial taxa with high respiratory and fermentative capacities, thereby accelerating carbon turnover and CO₂ release under nutrient-enriched conditions. The net release of ammonium and acetate by dominant genera such as Nocardia and Burkholderiaceae can increase nitrogen and carbon availability for plants. Overall, these shifts illustrate how cultivation reshapes rhizosphere microbial metabolism toward faster nutrient cycling.

Overall, a commonality both the wild and cultivated F. chiloensis rhizospheres was the overlap of high biomass producers with expanded biosynthetic repertoires. In the wild environment, Frankia (Bin 2) stood out, while in cultivated samples, Nocardia (Bin 14) dominated, each showing high predicted biomass in growth simulations (Figs. 6A and 6B) and harboring multiple BGCs (Figs. 2 and 3). This duality suggested that these microbes possess the genetic machinery for producing stress‐adaptive or bioactive metabolites, a confluence of active central metabolism and specialized metabolic potential. To support the latter, a global survey of >4500 plant-associated bacterial genomes found that the capacity to produce specialized metabolites is widespread and phylogenetically conserved, with certain BGC classes (e.g. terpenes, aryl polyenes) especially common in phytobiomes (Mukherjee et al., 2023).

Such capacity for adaptation was further corroborated by stress‐responsive gene clusters detected in cultivated bins like Nocardia (Bin 14) and Mycobacteriaceae (Bin 18), which encoded ectoine or NAPAA polypeptides used for osmoprotection (Czech et al., 2018) (Fig. 3), in concordance with soil analysis demonstrating deficient in Fe and P and contamination with metals (Supplementary Fig. 2). Although flux models do not currently show major osmolyte production, perhaps due to relatively mild greenhouse conditions, these clusters clearly harbor a reserve capacity for withstanding more severe stresses (e.g., drought, salinity, lack of potassium and acidification, among other stress). In line with the latter, a recent genome analysis of a halotolerant Brevibacterium found gene clusters for ectoine, NAPAA, siderophore, and terpene biosynthesis in the same species (Huang et al., 2022).

There was a significant intersection between nitrogen metabolism and specialized metabolite pathways. In the wild, Frankia (Bin 2) and Reynarella (Bin 13) displayed both nitrogen‐fixing or nitrate‐reducing genes and biosynthetic clusters for antibiotics or siderophores, indicating roles in nutrient acquisition and microbial competition, an observation already described in literature (Udwary et al., 2011) (Fig. 4, Fig. 5). Meanwhile, in the cultivated set, Trinickia (Bin 2) and Burkholderiaceae (Bin 15) encoded phosphonate and quorum‐sensing systems that tie into nutrient acquisition (phosphorus) and community regulation (AHL‐mediated signals), also already described for the genera (DeAngelis et al., 2008; Paulo et al., 2024). Flux observations of NH₃ production or consumption by these same taxa confirmed that fundamental nitrogen metabolism could interface directly with defense or mutualistic strategies, reinforcing how core metabolic processes and specialized metabolic functions combine to shape the rhizosphere microbial community. Further work could evaluate ectoine or NAPAA expression under drought, high salinity and metal contamination, to confirm osmoprotective functions of these bacteria.

Aside from microbial descriptions, organelle-wide population scans of the same metagenomic reads uncovered signatures of balancing selection in atp9, atpA/H, and rpoC2 (Supplementary Fig. S1 and Supplementary Material). These loci may modulate root exudation or immunity and thus help shape the microbiome described here.

5. Conclusions

In summary, this study presents the first metagenomic characterization of F. chiloensis rhizosphere communities in Chile using MAGs, revealing distinct taxonomic and functional patterns between wild and cultivated environments. The identification of metabolically versatile and biosynthetically rich taxa such as Frankia (bin 2) in the wild rhizosphere, and Nocardia (bin 14) or Burkholderiaceae (bin 15) in cultivated settings, shows how environmental conditions and agricultural practices shape microbial assembly, metabolic flux, and ecological function. Importantly, these findings reinforced a One Health perspective, emphasizing that rhizosphere microbiomes are not isolated entities, but integrative components of ecosystem and crop health. Through processes such as carbon turnover, nitrogen cycling, and the production of signaling molecules and osmoprotectants, these microbes contribute directly to plant resilience, soil function, and environmental stability. The detection of strain-specific BGCs and metabolic interactions provides a foundation for future work on microbe-mediated stress mitigation and sustainable cultivation of native crops like F. chiloensis.

Funding

CF was supported by the ANID Subvención a la Instalación en la Academia (Convocatoria 2021), grant No. SA77210106. JFC was funded by Fondecyt Iniciación (ANID) 11191074.

CRediT authorship contribution statement

Carlos Farkas: Methodology, Software, Formal analysis, Data curation, Visualization, Writing – original draft. Matías Guerra: Investigation, Data curation, Writing – original draft, Formal analysis. Adan Andreu Heredia: Visualization, Methodology. Jean Franco Castro: Formal analysis, Conceptualization, Funding acquisition, Supervision, Project administration, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The sequencing data obtained in this work were deposited in the NCBI Sequence Read Archive under BioProject PRJNA1232186. Ten libraries represent rhizosphere communities of Fragaria chiloensis ssp. chiloensis: cultivated botanic form chiloensis (F08RW–F10RW) and wild botanic form patagonica (F01RW–F07RW), with SRA run accessions SRR32598462 (F10RW), SRR32598463 (F09RW), SRR32598464 (F08RW), SRR32598465 (F07RW), SRR32598466 (F06RW), SRR32598467 (F05RW), SRR32598468 (F04RW), SRR32598469 (F03RW), SRR32598470 (F02RW), and SRR32598471 (F01RW). Three high-quality (>80 % completeness) metagenome-assembled genomes have been deposited in GenBank: Bin 28 UBA4665 (JBNQGE000000000, cultivated), Bin 4 Arachidicoccus sp.003600625 (JBNQGD000000000, cultivated), and Bin 2 Frankia (JBNQGC000000000, wild). Draft assemblies capturing the cultivated (F_chiloensis_chiloensis.fasta; 14 778 contigs, 69.0 Mb) and wild (F_chiloensis_patagonica.fasta; 16 748 contigs, 34.9 Mb) metagenomes, including the recovered MAGs, were archived on Zenodo (DOI 10.5281/zenodo.15343608).