In most legume-rhizobium symbioses, rhizobial colonization occurs through host-derived intracellular infection threads, which enable rhizobial recruitment while presumably modulating the host immune system to prevent rejection. To investigate post-translational regulation of immune responses during rhizobial infection, we focused on Cyclophilin A (CyPA), a peptidyl-prolyl cis/trans isomerase. The model legume Lotus japonicus encodes three canonical CyPA genes. Using CRISPR/Cas9 mutagenesis, structural modeling, and phylogenomics, we characterized LjCyPA1 as essential for normal intracellular infection by compatible rhizobia. A gain-of-function LjCyPA1 variant in a soybean cultivar promoted symbiosis with both compatible and incompatible rhizobia. Functional association between LjCyPA1 and the immune hub protein LjRIN4 is essential for symbiosis. The putative cis-LjRIN4 promoted intracellular rhizobial infection, while the putative trans-LjRIN4 suppressed it. LjCyPA1 and LjRIN4 acted in concert with the rhizobial type III secretion system (T3SS), highlighting a cooperative role between host and symbiont in facilitating infection. Our results contribute to the understanding of how legumes accept symbiotic partners while balancing immune responses.

Background/Objectives: The dark green leaf color trait in bunching onion (Allium fistulosum L.) is an important agronomic trait closely associated with market value; however, its genetic basis remains poorly understood. This study aimed to identify quantitative trait loci (QTLs) associated with leaf color using SPAD values as a phenotypic indicator. Methods: An F2 population derived from a cross between the dark green line YSG1go and the light green line Asagikei-KUJYO was used. A linkage map was constructed based on RNA-seq-derived SNP markers, and SPAD values were measured for QTL analysis. Results: The linkage map consisted of eight linkage groups with a total length of 2103.0 cM and 765 mapped markers. SPAD values showed significant differences between the parental lines, with high broad-sense heritability (H2 = 0.76), indicating a strong genetic contribution to this trait. Multiple significant QTLs were detected on chromosomes 4 and 5, each explaining 27.4-38.1% of the phenotypic variance. The direction of allelic effects differed among QTLs, suggesting that favorable alleles are distributed between the parental lines. In addition, genes related to chloroplast protein translation were identified within the QTL regions. Conclusions: SPAD values are a suitable indicator for genetic analysis of leaf color in bunching onion, and the QTLs identified in this study provide valuable information for molecular breeding aimed at improving dark green leaf color.

In mutualistic symbiosis between plants and bacteria, the abundance and composition of symbiotic bacterial groups in the soil microbiota can be important for plant growth. Here, we focused on the nitrogen-fixing mutualism between Lotus japonicus and nodule bacteria to investigate whether and how much the abundance of symbiotic rhizobia in the soil microbiota of natural environments contributes to variations in host plant growth. An inoculation experiment of soil microbiota revealed extensive variations in plant growth phenotypes, even between microhabitats. We found that the local presence of L. japonicus and the relative abundance of Mesorhizobium bacteria showed positive correlations with plant growth supported by both 16S amplicon sequencing and shotgun metagenome analyses. Among bacteria investigated, the abundance of Mesorhizobium was most strongly associated with plant growth phenotypes, supporting its role as the primary symbiotic rhizobia in natural environments. Given the specificity and the selectivity of plants for favorable rhizobia, legume–rhizobia interactions could trigger a positive plant–soil feedback that enriches favorable rhizobia into the soil surrounding legume plant habitats.

Microbial inoculants are increasingly being used to enhance plant growth and soil environments, yet their ecological effects on native microbial communities remain unclear. We investigated the long-term dynamics of soybean-nodulating Bradyrhizobium strains carrying the nosZ gene encoding N₂O reductase in microcosms comprising three different soils. Analysis of 16S rRNA amplicons revealed that bacterial community compositions were primarily driven by soil type and time, with inoculation having only minor effects. In contrast, nosZ clade I amplicon analysis showed high survivability of the dominant nosZ-sequence groups corresponding to the inoculant strains, which accounted for 20%-50% of the total nosZ community even 249 days after inoculation; despite this dominance, the overall community structures of indigenous nosZ-harboring Bradyrhizobium remained largely unchanged. This observation was further supported by a permutational multivariate analysis of variance, which showed a 90% reduction in R2 when the inoculant sequence groups were excluded. Thus, it is possible to enhance N₂O-reducing function through inoculation without drastically disrupting the indigenous microbial community. To explore the landscapes of inoculation effects obscured in two-dimensional mapping, we applied dimensionality-reduction tools such as principal coordinates analysis beyond their typical use for visualization by extending analyses into higher-dimensional spaces. Varying the number of dimensions revealed that the signal-to-noise ratio and clustering tendency peaked at 4-10 dimensions, with further increases in dimensions leading to homogenization of community patterns. This intermediate dimensional space also revealed differences in community succession by soil type. These findings demonstrate that careful selection of dimensionality can enhance the discovery of ecological patterns.IMPORTANCEDimensionality reduction is widely used to visualize microbiome data in two- or three-dimensional ordination spaces. However, its application in higher-dimensional analysis remains underexplored. We highlighted the use of dimensionality reduction not only as a visualization tool, but as a means of projecting microbiome data into ordination spaces for geometric analyses, which are not limited to two or three dimensions. Communities were projected beyond three dimensions to examine how dimensionality affects evaluation of inoculation effects through geometric analyses. Our findings show that dimension selection strongly influences the ability to detect and resolve ecological signals, which were distorted in low-dimensional spaces or homogenized in extremely high-dimensional spaces. Intermediate dimensionalities better retained the spatial fidelity needed to resolve soil-dependent responses to inoculation. By enhancing the resolution of small but meaningful effects, this approach provides a robust framework for guiding strain selection, application strategies, and risk assessment in microbial inoculation studies.

Nitrous oxide (N2O) is a potent greenhouse gas, and the enzyme Nos catalyzes its reduction to dinitrogen (N2). Bradyrhizobium ottawaense exhibits strong N2O-reducing activity with high nosZ expression. To investigate whether promoter sequences affect nosZ expression, we constructed reciprocal promoter-swapped mutants between B. ottawaense and B. diazoefficiens. The swapping of promoters did not significantly affect expression levels. B. ottawaense mutants maintained approximately 200-fold higher expression levels than B. diazoefficiens, and the introduction of the B. ottawaense promoter into B. diazoefficiens did not increase expression levels. Therefore, the present results indicate that promoter sequence differences are not the primary factor affecting nosZ expression, suggesting regulation by other factors.

Insertion sequences (ISs) are major drivers of genomic plasticity in rhizobia, frequently promoting local recombination events. To quantitatively compare the stability of genomic regions inside and outside of the symbiosis island, we engineered Bradyrhizobium diazoefficiens USDA122 mutants carrying a sacB/aadA counter-selectable cassette at four distinct loci-three on symbiosis island A (SymA) and one in the core genome. During 5 days of saprophytic growth, cassette deletion occurred at frequencies of up to 1.77×10-3 within SymA, whereas the deletion rate in core genomic regions was markedly lower (3.29×10-6). Within SymA, cassettes inserted adjacent to the nif and rhc clusters, where IS copies with the same orientation were enriched, were lost more frequently than those placed in other SymA regions, indicating marked intra-island variability in genomic stability. Similar yet overall lower deletion frequencies were observed in B. diazoefficiens USDA110. These results demonstrate that SymA contains genomic loci with greater susceptibility to IS-mediated rearrangements and also that such recombination events may contribute to the diversification of Bradyrhizobium symbiosis islands. Based on our comparative IS mapping in B. japonicum and B. ottawaense, we discuss the potential for the IS-mediated deletion of genome regions harboring nod genes.

Green manures are widely used to enhance soil health and suppress plant-parasitic nematodes, and their effects on the broader soil food web have been studied. Beyond direct suppression, the role of green manures in supporting and sustaining soil food webs has attracted increasing attention. In this study, we evaluated the use of DNA sequencing to identify various nematode genera and their microbial associates in a field trial using oat (Avena sativa) and hairy vetch (Vicia villosa) as green manures. Nematode index analysis revealed that the oat treatment promoted a structured nematode community. Furthermore, the nematode community structure observed in the oat treatment was linked to specific bacterial and fungal genera. Several beneficial fungi were identified, indicating that oats, used as a green manure, actively enhanced the microbiome. Our results showed that enriching the micro-food web through organic fertilizers can help in the detection of beneficial microorganisms, with the nematode index serving as a potential indicator.

Flowering time is an important factor in plant fitness and local adaptation. Genome-wide association (GWA) studies have allowed the identification of candidate genes in certain plant species for various traits, including flowering time. Lotus japonicus is widely found throughout the Japanese archipelago. To obtain flowering time data with more prominent difference as more suitable indicator of environmental adaptation, flowering time data were collected for 132 wild accessions originating from various points across this region under shorter day length conditions than in previous studies. The results showed latitudinal variations in flowering time, with southern accessions flowering earlier. Comparing data from four flowering times with varying conditions revealed greater differences under a shorter day length. It is likely that day length significantly affects flowering time in this species. GWA analyses were conducted on flowering time variation measured in this study and the ratios between flowering time under different conditions. Candidate genes different from previous study were detected, including orthologues of known flowering time genes in each analysis. Correlation tests between flowering time and strongly detected single-nucleotide polymorphisms (SNPs) in the GWA analysis suggested that approximately 60% of flowering time variation can be explained by the two main SNPs. This result suggests that the majority of the variation could be explained by a small number of genetic factors. Considering the strong association with flowering time variation, these candidates may be responsible for these differences and therefore can be related to local adaptation in this species.

Leguminous plants establish root nodule symbiosis, which is initiated by the recognition of rhizobial nodulation factors by plant receptor kinases. However, other factors, such as Type III effector proteins, also affect host specificity. We herein investigated the role of nodulation outer protein M (NopM), a Type III effector of Bradyrhizobium elkanii USDA61, in symbiosis with Lotus japonicus MG-20 and Lotus burttii. NopM, annotated as an E3 ubiquitin ligase, triggers an early senescence-like response, inducing brown nodules that hinder effective symbiosis. NopM shares structural features with E3 ubiquitin ligases derived from both pathogenic and symbiotic bacteria, including a leucine-rich-repeat and E3 ubiquitin ligase domain. The deletion of these domains or substitution of the cysteine residue, predicted to be the active site of the ubiquitin ligase domain, suppressed the formation of brown nodules. These results suggest that NopM interacts with target proteins through its leucine-rich-repeat domain and mediates ubiquitination via its ligase domain, thereby contributing to the induction of brown nodules. A transcriptome ana-lysis further suggested that the early senescence-like response closely resembled the plant hypersensitive response, with the up-regulation of defense-related genes. Therefore, L. japonicus may recognize NopM in infected nodule cells, leading to an immune response that disrupts symbiosis. The present study provides insights into the mole-cular mechanisms by which rhizobial effectors modulate symbiotic interactions in infected nodule cells, highlighting the ability of L. japonicus to activate immune responses even in nodule cells where rhizobia have been accepted.

Symbiosis between Bradyrhizobium strains isolated from Lao People's Democratic Republic (Lao PDR) and intercropped legumes (Arachis hypogaea, Vigna radiata, and V. mungo) was regulated by the type III secretion system (T3SS), which delivers effector proteins (T3Es) into host plant cells to modulate nodulation. To explore this mechanism, we sequenced and analyzed seven Bradyrhizobium genomes, identifying putative T3Es across five T3SS groups (G.1-G.5), which were classified based on the sequence of rhcN, a conserved ATPase gene essential for T3SS function. Phylogenetic analysis of rhcN more closely reflected the evolutionary relationships of nodulation genes than those based on 16S rRNA or whole-genome comparisons, underscoring its symbiotic relevance. Functional assays using rhcN mutants revealed group-specific effects on nodulation; G.1 strains showed neutral effects on A. hypogaea, negative effects on V. radiata, and positive effects on V. mungo. G.2 strains consistently promoted nodulation across all hosts and lacked effectors related to SUMO (small ubiquitin-like modifier) pathways, which have been implicated in host defense regulation. G.3 strains reduced nodulation in A. hypogaea but enhanced it in Vigna species. G.4 strains suppressed nodulation in A. hypogaea, and G.5 strains inhibited nodulation across all tested legumes. These findings highlight the diversity in T3SS organization, effector composition, and symbiotic responses among native Bradyrhizobium strains. The identification of known and uncharacterized effectors suggests roles in host compatibility and specificity. These strains, along with their effector profiles, provide a foundation for future functional studies to better understand T3SS-mediated interactions and support the development of targeted inoculants for legume hosts.IMPORTANCEThis study advances our understanding of legume-Bradyrhizobium symbiosis by examining the genetic organization and evolutionary patterns of T3SS genes. Our findings revealed that T3SS gene evolution does not always align with phylogenies based on 16S rRNA or whole-genome sequences, suggesting that horizontal gene transfer and functional adaptation may shape diversification. The observed variation in T3SS architecture and effector profiles among the five distinct Bradyrhizobium groups was correlated with host-specific nodulation outcomes in A. hypogaea, V. radiata, and V. mungo. We also identified novel candidate genes influencing symbiotic signaling and compatibility. These insights into the diversity and function of T3SS components contribute to a broader understanding of host-microbe communication and may support the development of more targeted and efficient rhizobial inoculants for sustainable legume cultivation and improved biological nitrogen fixation.

Bradyrhizobium sp. strain SUTN9-2 demonstrates cell enlargement, increased DNA content, and efficient nitrogen fixation in response to rice (Oryza sativa) extract. This response is attributed to the interaction between the plant’s cationic antimicrobial peptides (CAMPs) and the Bradyrhizobium BacA-like transporter (BclA), similar to bacteroid in legume nodules. The present study reveals that SUTN9-2 can also establish functional endophytic interactions with chili (Capsicum annuum) and tomato (Solanum lycopersicum) plants. When exposed to extracts from chili and tomato, SUTN9-2 exhibits cell elongation, polyploidy, and reduced cell viability, with the effects being less pronounced for tomato extract. Transcriptomic and cytological analyses revealed that genes associated with CAMP resistance, nitrogen metabolism, nitrogen fixation, defense responses, and secretion systems were upregulated, while genes related to the cell cycle and certain CAMP-resistance mechanisms were downregulated, particularly in response to chili extract. This study suggests that SUTN9-2 likely evolves resistance mechanisms against CAMPs found in rice, chili, and tomato plants through mechanisms involving the protease-chaperone DegP, AcrAB-TolC multidrug efflux pumps, and polysaccharides. These mechanisms facilitate efflux, degradation, and the formation of protective barriers to resist CAMPs. Such adaptations enable SUTN9-2 to persist and colonize host plants despite antimicrobial pressures, influencing its viability, cell differentiation, and nitrogen fixation during endophytic interactions with various plant hosts.

Plant trait databases play a crucial role in understanding ecological and evolutionary processes yet remain insufficient due to geographical and data availability limitations. To address this limitation, we developed a novel large-scale text extraction approach using a large language model (LLM) to transform descriptions of Floras into structured trait data, thereby expanding existing databases. We applied this approach to extract flower color information from the Flora of China and integrate it into the TRY Plant Trait Database. After integration, the dataset expanded to 27,252 species, more than doubling the previously available flower color records. Additionally, we linked the dataset with occurrence records from GBIF and environmental data, including climate and soil properties, to disentangle ecological insight and flower color distribution. Our large-scale association analysis of flower colors and environments revealed that white, yellow, and red-type flowers exhibit distinct environments, suggesting that abiotic environment can play a role in flower color evolution. By transforming descriptions of Flora into structured data, our approach organizes traits across more plant species, creating new opportunities for ecological and evolutionary research. The present approach can be extended to other traits, enhancing our understanding of how plants adapt and respond to environmental changes on a global scale. ### Competing Interest Statement The authors have declared no competing interest.

Flooded rice fields are a major source of atmospheric methane, a strong greenhouse gas second only to carbon dioxide. Rice roots are one of the most important hotspots for methane oxidation in rice fields. However, limited information is available on the physiological and genomic characteristics of methane-oxidizing bacteria (MOB) inhabiting rice roots. In the present study, we isolated MOB from rice roots and characterized the strains phenotypically and genomically. We obtained 100 MOB-enriched cultures from the roots of three rice cultivars (Oryza sativa L. subsp. japonica cv. Nipponbare, O. sativa L. subsp. indica cv. Muha, and Tupa 121-3), in which twelve MOB isolates, two Methylomonas sp., three Methylocystis sp., and seven Methylosinus sp., were successfully purified. They showed different morphological features (types of flagellation) and colony formation potentials within the same group in some cases. A genome sequencing ana-lysis revealed variations in the number of genes or the clusters of methane monooxygenase, methanol dehydrogenase, and nitrogenase. The number of plasmid DNAs also differed among the strains. Four strains belonging to the genus Methylomonas or Methylocystis represented putative novel species based on their phenotypic and genotypic characteristics. The present study largely expanded the eco-collection of MOB cultures inhabiting rice fields and rice roots.

Since nitrogenase is intrinsically sensitive to oxygen (O2), diverse aerobic diazotrophs need strategies to cope with nitrogenase damage by O2. In the present study, we investigated the mechanisms by which aerobic methane-oxidizing bacteria (methanotrophs) enable the concurrent activities of methane monooxygenase, which uses O2, and nitrogenase in the cytoplasm of the same cell. By using 15N labeling, we confirmed the capacity of alphaproteobacterial methanotroph Methylosinus sp. 3S-1 for nitrogen fixation and diazotrophic growth across a wide range of O2 concentrations <20%. When the initial O2 concentration was increased from 2 to 20% in a diazotrophic culture, similar decreases were observed in fixed nitrogen and NifH protein levels. In contrast, the mRNA levels of nitrogen fixation genes (nif genes) markedly increased and remained elevated for the duration of slow growth at high O2 concentrations. This pattern of nif expression in response to O2 may be attributed to the properties of the nif-specific transcriptional regulator NifA. The present results suggest that the increase in nif transcription is one of the strategies by which this methanotroph maintains nitrogen fixation on the background of aerobic methane oxidation.

Japanese morning glory (Ipomoea nil), a short day plant, has been used for studying flowering times. Here, quantitative trait loci (QTL) analysis for days from sowing to flowering (DTF) of F2 between I. nil var. Tokyo Kokei Standard (TKS) and I. hederacea line var. Q65, an early flowering variety, revealed seven QTLs: QTL Ipomoea Flowering 1-7 (qIF1-7). The position of qIF3, which had the most significant effect among the seven QTLs, corresponds with that of I. nil (or I. hederacea) CONSTANS (InCO/IhCO) in the linkage map. There is a single-base InDel in the coding sequence of InCO/IhCO. The single-base deletion (SBD) causes a frame-shift mutation and loss of function in TKS allele (inco-1). I. nil accessions bearing inco-1 tend to flower early, similarly to rice varieties bearing the loss of function allele of CO ortholog, hd1. The TKS allele of qIF3 reduces DTF and corresponds with the inferred effect of inco-1. Based on the distribution of inco-1, a hypothesis was proposed that the SBD in inco-1 might have played an important role in the expansion of Japanese morning glories, originally native to the tropical regions of the Americas, into temperate Asia.

Bradyrhizobium vignae ORS3257 is an efficient symbiotic strain for Vigna unguiculata and V. mungo but fails with V. radiata due to an effector-triggered immunity response mediated by the nodulation outer protein P2 (NopP2). To understand this incompatibility, we identified NopP2 interacting proteins in V. radiata cv. KPS1, including enolase, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), monodehydroascorbate reductase (MDHAR), and serine hydroxymethyltransferase (SHMT) as targets. Protein-protein interaction assays confirmed that NopP2 binds to these enzymes, and further analysis revealed their co-localization on the plasma membrane. Comparative transcriptomic analysis revealed NopP2 stimulates genes related to plant defense response (PR1, PR5, MYB13, and TAO1), hydrogen peroxide (SOD, POX10, and POX16), and cell wall lignification (LAC). NopP2 did not alter the expression of genes encoding the target enzymes but interfered with MDHAR activity, leading to high H2O2 accumulation in roots. These findings suggest that NopP2 contributes to symbiotic incompatibility in V. radiata by inducing a multifaceted defense response and initiating cell wall lignification early in infection.

Symbiotic nitrogen fixation is a highly coordinated process involving legume plants and nitrogen-fixing bacteria known as rhizobia. In this study, we investigated a novel Fix- mutant of the model legume Lotus japonicus that develops root nodules with endosymbiotic rhizobia but fails in nitrogen fixation. Map-based cloning identified the causal gene encoding the filamentation temperature-sensitive H (FtsH) protein, designated as LjFtsH4. The LjFtsH4 gene was expressed in all plant organs without increased levels during nodulation. Subcellular localization revealed that LjFtsH4, fused with a fluorescent protein, localized in mitochondria. The Ljftsh4 mutant nodules showed signs of premature senescence, including symbiosome membrane collapse and bacteroid disintegration. Additionally, nodule cells of Ljftsh4 mutant displayed mitochondria with indistinct crista structures. Grafting and complementation tests confirmed that the Fix- phenotype was determined by the root genotype, and that protease activity of LjFtsH4 was essential for nodule nitrogen fixation. Furthermore, the ATP content in Ljftsh4 mutant roots and nodules was lower than in the wild-type, suggesting reduced mitochondrial function. These findings underscore the critical role of LjFtsH4 in effective symbiotic nitrogen fixation in root nodules.

DNA marker-based discrimination between a wild-type specimen and mutants obtained using heavy-ion beams (because of small genomic and genetic mutations) is challenging in mutation breeding. Protecting the plant breeders' right to mutate cultivars is challenging because tissue culture techniques that use the protocorm-like body (PLB) can quickly propagate. Thus, we present two methods for generating DNA markers in Cymbidium using genomic mutations induced by heavy-ion beams. In the first method, we selected 8-16 plants from three cultivars after irradiation that had no mutations in morphology or traits, followed by genome scanning using 240 sets of 15-mer random amplified polymorphic DNA primers and arbitrarily primed PCR primers based on partial retrotransposon sequences. Three polymorphic patterns were generated in each cultivar. In the second method, cultivar candidates were generated by crossbreeding, and a strain was selected. Genome scanning was performed on 43 plants irradiated by applying a carbon- or neon-ion beam to the PLB and using five random primer sets that allowed many scorable bands. Polymorphic patterns were detected in two strains at all micropropagation steps via the PLB and in second-flowering plants. These data demonstrate that strains with the same appearance after heavy-ion beam irradiation can be distinguished using polymorphic DNA patterns alone. These patterns were defined as "DNA marks" for intracultivar identification. In cases where multiple strains of DNA marks are obtained in one cultivar, DNA marks can realize the feasibility of DNA-level traceability from production to market.

Although microbial inoculation may be effective for sustainable crop production, detrimental aspects have been argued because of the potential of inoculated microorganisms to behave as invaders and negatively affect the microbial ecosystem. We herein compared the impact of rhizobial inoculation on the soil bacterial community with that of agricultural land-use changes using a 16S rRNA amplicon ana-lysis. Soybean plants were cultivated with and without five types of bradyrhizobial inoculants (Bradyrhizobium diazoefficiens or Bradyrhizobium ottawaense) in experimental fields of Andosol, and the high nodule occupancy (35-72%) of bradyrhizobial inoculants was confirmed by nosZ PCR. However, bradyrhizobial inoculants did not significantly affect Shannon's diversity index (α-diversity) or shifts (β-diversity) in the bacterial community in the soils. Moreover, the soil bacterial community was significantly affected by land-use types (conventional cropping, organic cropping, and original forest), where β-diversity correlated with soil chemical properties (pH, carbon, and nitrogen contents). Therefore, the effects of bradyrhizobial inoculation on bacterial communities in bulk soil were minor, regardless of high nodule occupancy. We also observed a correlation between the relative abundance of bacterial classes (Alphaproteobacteria, Gammaproteobacteria, and Gemmatimonadetes) and land-use types or soil chemical properties. The impact of microbial inoculation on soil microbial ecosystems has been exami-ned to a limited extent, such as rhizosphere communities and viability. In the present study, we found that bacterial community shifts in soil were more strongly affected by land usage than by rhizobial inoculation. Therefore, the results obtained herein highlight the importance of assessing microbial inoculants in consideration of the entire land management system.

The genetic makeup of natural plant populations often allows them to persist for many years without succumbing to disease. Transferring such properties to crops could increase resilience and reduce reliance on chemical pesticides​. However, the population genomics of plant-microbiome interactions remain poorly understood. Here, we use Lotus japonicus, which has persisted as a natural population in Japan for ~20’000 years​, to identify genomic selection signatures caused by soil microbes. We found strong genetic associations with root microbiome structure within a gene we name ROOT MICROBIOME ESTABLISHMENT 1 (ROOMIE1) located in a region with high haplotype richness. This led us to quantify haplotype richness genome-wide using a new metric, HAPk. We found haplotype-rich regions strongly enriched for microbiome genetic associations, suggesting that soil microbes were imposing negative frequency-dependent selection. We validated this hypothesis by showing increased migration rates of microbiome associated alleles and by determining that ROOMIE1 shows standing variation in local populations and impacts rhizosphere colonization in native Japanese soil. Our results indicate that interactions with soil microbes represent a major selective force in plant genome evolution and suggest that haplotype-rich regions constitute a genetic resource for improvement of crop resilience.