Variations in Nodule Microbial Communities and Their Association with Root-Colonizing Arbuscular Mycorrhizal Fungi in Medicago Sativa Authors: Yuxue Zhang, Chunyan Zheng, Shichao Wang, Feng Zhu Published: 2025 ABSTRACT Alfalfa (Medicago sativa L.), known as “Queen of forages”, is valued to its high-nutritional quality and is a key member of Leguminosae family. Its productivity is largely attributed to mutualistic symbioses with arbuscular mycorrhizal fungi (AMF) and rhizobia, which facilitate nutrient exchange and plant growth. However, the coexistence and mutualistic interactions between rhizobia and AMF across alfalfa genotypes with differing yields in native soil remain poorly understood. In this study, we investigated the community composition of rhizobia and AMF colonizing alfalfa roots across different-yield varieties. Our results showed variations in dominant microbial taxa and the structural complexity of root-associated microbial networks among genotypes. Moreover, rhizobia exhibited no significant associations with AMF on genus level, however, negative correlations were observed among genera within the AMF community, and a comparable trend was identified among rhizobial taxa. In summary, our findings offer new insights into how native soil microbiota influence the dual symbiotic relationships of alfalfa, with implications for leveraging native microbial communities to enhance sustainable forage production.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00248-025-02687-x. Introduction Alfalfa (Medicago sativa L.) is a leguminous crop with high protein content and plays an important role in animal husbandry as a key source of livestock feed [1, 2]. With the rapid expansion of the livestock industry and the increasing demand for forage, achieving sustainable high yields and maintaining quality are critical goals for alfalfa cultivation. Recent studies have focused on enhancing alfalfa productivity through utilizing symbiotic relationships between plants and microorganisms [3, 4]. Microbial associations with plant hosts underpin essential ecosystem functions [5], driving nutrient cycling [6, 7], promoting plant growth [8, 9], and enhancing stress resilience [10, 11]. Studies have shown that the plant roots often harbor beneficial microbes that can promote plant development and resistance to abiotic and biotic stressors, such as nutrient deficiency and pathogen infestation [12, 13]. Legumes, including alfalfa, form mutualistic relationships with two major types of soil microorganisms, nitrogen-fixing rhizobia and arbuscular mycorrhiza fungi (AMF). Rhizobia fix atmospheric nitrogen within root nodules via the nitrogenase enzyme complex, while receiving carbon and nutrients from the host plants in return [14]. Similarly, AMF develop highly branched arbuscules within root cortical cells, where they exchange nutrients (e.g. phosphorus) for plant-derived sugars and lipids [15]. Co-inoculation with rhizobia and AMF has been shown to improve nutrient uptake and enhance resistance to various stressors, including insect pests [16], fungal pathogens [17], and salinity [18]. Root nodules are complex habitats that also harbor non-rhizobial bacteria, such as Pseudomonas, which can further contribute to plant health [19]. The synergistic effects of these microbial consortia, especially the interactions among diverse native soil microorganisms, amplify the benefits to the host plants. Thus, microbial diversity and interspecies interactions are critical components of plant productivity. Despite these advances, the specific interactions between AMF and rhizobia in native soils and their dependence on plant genetic background remain poorly understood. Crop breeding and ecological optimization are both essential strategies for improving agricultural sustainability [20]. However, our understanding of plant-soil-microbe interactions remains limited. Soil microbes form intricate, genotype-specific relationships with plant hosts [21], and variations in root endophytic communities are shaped by host identity and microbial preferences [22–24]. Previous studies primarily emphasized host species-level differences, as driven by soil properties and rhizocompartments effects [24, 25]. Increasing evidence now suggests that varietal differences also influence microbial assembly and function [26, 27]. Even under identical environmental conditions, crop varieties can exhibit distinct gene expression profiles due to their unique genetic backgrounds, which in turn shape agronomic traits such as yield potential, stress tolerance, and responsiveness to microbes [28]. Based on this premise, we hypothesize that genotype factor may significantly influence the establishment and functional dynamics of symbiotic interactions between rhizobia and AMF, potentially resulting in distinct symbiotic patterns. In this study, we aimed to unravel the complex tripartite interactions among alfalfa genotypes, nodule microbiota, and root-colonizing AMF. We investigated the compositional dynamics and network structures of root-associated microbial communities, as well as cross-kingdom associations between nodule bacteria and AMF. The alfalfa genotypes investigated in this study exhibited contrasting high- and low-yield phenotypes. We addressed the following questions: (1) Does alfalfa genotype affect the diversity and structure of root endophytic microbial communities? (2) Does genotype influence the functions and interactions of key symbiotic microorganisms? (3) To what extent do plant productivity phenotypes associate to the variation in microbial communities? Materials and Methods Field Experiment and Sampling This study was conducted as part of a selection experiment to identify high-yield alfalfa genotypes adapted to the local environment in Huanghua, Hebei Province, China (HH; 38°09′N, 117°05′E), a region characterized by a temperate continental monsoon climate. Fifty alfalfa (Medicago sativa L.) genotypes were planted, and their forage yields were measured [29]. Based on yield performance (Table S1), ten genotypes were selected: the top five (Northern Hemisphere, Magnum401, WL903HQ, Magnum2, and Guochan No.1) and the bottom five (Magnum601, SK4030, WL319HQ, Magnum7, and Southern Hemisphere). A comparison of mean annual yield (calculated as the average of the five genotype means within each group) revealed a significant difference between the two groups (Fig. S1). They were subsequently referred to as the high-yield group (30.98 ± 1.25 t ha⁻¹ yr⁻¹) and the low-yield group (25.66 ± 1.63 t ha⁻¹ yr⁻¹). Sampling was conducted in October, 2019. For each genotype, five biological replicates were established by randomly selecting five points in the field. Each replicate consisted of ten adjacent plants surrounding the corresponding point, from which all root nodules were pooled to create a single nodule sample, and all roots were combined to form a root sample. All plants were selected of the approximately same height and the same age. Samples were transported to the laboratory on ice, washed, and surface-sterilized following the protocol of Lundberg [30]. Complete surface sterilization of the samples was confirmed by inoculating the final sterile rinse onto LB agar plates. After incubation at 28 °C for 3 days, the absence of any microbial growth on these plates demonstrated that the surface microbiota had been effectively eliminated. Surface-cleaned root nodules were detached from the root system using sterile scissors and stored at −80 °C separately until further analysis. In total, 50 nodule samples and 50 root samples were obtained. DNA Extraction, PCR Amplification and Illumina Sequencing Genomic DNA was extracted from 0.15 g of each nodule sample or root sample using the DNeasy UltraClean Microbial Kit (QIAGEN, Hilden, Germany; Cat. No.12224-250) according to the manufacturer’s instructions. DNA concentration and purity were determined using a NanoDrop One Microvolume UV-Vis Spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Bacterial and AMF DNA were each amplified in two successive rounds with different primers combinations. For bacteria, the first round used 799 F and 1392R for 27 cycles, followed by a second round using 799 F and 1193R for 13 cycles [30]. For AMF, the first round used AML1F and AML2R for 32 cycles, followed by a second round using AMV4-5NF and AMDGR for 30 cycles [31]. PCR was performed in a 20 µl PCR mixture containing 4 µl 5×FastPfu Buffer, 2 µl 2.5 mM dNTPs, 0.8 µl of each primer (5 µM), 0.4 µl FastPfu Polymerase, 0.2 µl BSA and 10 ng Template DNA. All reactions were carried out in an ABI GeneAmp® 9700 PCR thermocycler (ABI, CA, USA) with the following conditions: initial denaturation at 95 ℃ for 3 min; denaturing at 95 ℃ for 30 s, annealing at 55 ℃ for 30 s and extension at 72 ℃ for 45 s for the indicated number of cycles; final extension at 72 ℃ for 10 min; and end at 4 ℃. Each reaction was performed in triplicates. The PCR products were visualized on 2% agarose gels, purified with the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA), and quantified using Quantus™ Fluorometer (Promega, USA). Purified amplicons were pooled in equimolar and paired-end sequenced on an Illumina MiSeq PE300 platform (Illumina, San Diego, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). Raw reads were demultiplexed, quality-filtered using fastp version 0.20.0 [32] and merged with FLASH version 1.2.7 [33] under the following criteria: (i) the 300 bp reads were truncated at any site receiving an average quality score of < 20 over a 50 bp sliding window, and the truncated reads shorter than 50 bp were discarded, reads containing ambiguous characters were also discarded; (ii) only overlapping sequences longer than 10 bp were assembled according to their overlapped sequence. The maximum mismatch ratio of overlap region is 0.2. Reads that could not be assembled were discarded; (iii) samples were distinguished according to the barcode and primers, and the sequence direction was adjusted, exact barcode matching, 2 nucleotides mismatched in primer matching. Operational taxonomic units (OTUs) were clustered at 97% similarity using UPARSE version 7.1 [34], and chimeras were removed. Taxonomic assignment was performed with the RDP Classifier v2.2 [35], using the SILVA 138 database for bacteria and the UNITE v7.0 database (https://unite.ut.ee) for AMF, with a confidence threshold of 70%. Statistical Analysis Normality and homogeneity of variance were evaluated using Shapiro-Wilk and Levene’s tests, respectively. Based on the results obtained, parametric (Student t test) or non-parametric (Wilcoxon rank sum test) one-way tests were used to compare microbial variables between high-yield and low-yield groups. The effects were considered to be significantly different if p < 0.05. Alpha diversity was analyzed using Shannon, Simpson and richness indices for bacterial (nodule) and AMF (root) communities. Beta diversity was evaluated by permutational multivariate analysis of variance (PERMANOVA) and community composition was visualized by principal coordinate analysis (PCoA). Differential abundances of the top 10 bacterial and AMF taxa were displayed using post hoc plots. Microbial co-occurrence networks were constructed and visualized in Gephi v0.10.1 [36]. Functional profiles of bacterial communities were predicted with FAPROTAX (http://www.cloud.biomicroclass.com/CloudPlatform). Spearman correlation matrices were generated to explore relationships between bacterial and AMF taxa and visualized as heat map. All statistical analyses were performed in R v4.5.0 (R Core Team, 2025). Post-hoc analyses were conducted with STAMP [37]. Results Bacterial Diversity and Composition After quality control, 988,874 effective sequences were clustered into 295 OTUs. Among them, 234 OTUs were shared between high- and low-yield groups, 30 OTUs were unique to the high-yield group and 31 OTUs were unique to the low-yield group. Taxonomic annotation identified 42 genera and 51 species. Shannon and Simpson indices differed significantly between yield groups, though richness did not (Fig. 1a). PCoA separated the two groups, with the first two axes explaining 61% and 17% of variance (Fig. 1b), and PERMANOVA confirmed significant compositional differences (p < 0.05).Fig. 1The diversity and co-occurrence of nodule microbial communities between high-yield group (H) and low-yield group (L). Three alpha diversity metrics of bacteria in nodule (a). Principal coordinate analysis of nodule endophytic bacterial communities of alfalfa (b). Values on PCoA axes indicate the percentages of total variation explained by each axis. PERMANOVA reflects the significant difference of bacterial communities between two groups. Extended error bar plot identifying significant differences between mean proportions of top 10 genera in high-yield group (H, coral) and low-yield group (L, cyan) samples (c). Corrected p values are shown at right. Co-occurrences network of top 30 genera of nodule bacteria in high-yield groups (d) and in low-yield group (e). The nodes with different colors in the co-occurrences network represent different genera; the node size represents the degree of the genus in the co-occurrences network; the edge between nodes indicates a correlation between genera; the color of the edge represents a positive correlation (red) and a negative correlation (blue). The edge thickness indicates the correlation values; only significant interactions are shown (r > 0.6; p < 0.05) The diversity and co-occurrence of nodule microbial communities between high-yield group (H) and low-yield group (L). Three alpha diversity metrics of bacteria in nodule (a). Principal coordinate analysis of nodule endophytic bacterial communities of alfalfa (b). Values on PCoA axes indicate the percentages of total variation explained by each axis. PERMANOVA reflects the significant difference of bacterial communities between two groups. Extended error bar plot identifying significant differences between mean proportions of top 10 genera in high-yield group (H, coral) and low-yield group (L, cyan) samples (c). Corrected p values are shown at right. Co-occurrences network of top 30 genera of nodule bacteria in high-yield groups (d) and in low-yield group (e). The nodes with different colors in the co-occurrences network represent different genera; the node size represents the degree of the genus in the co-occurrences network; the edge between nodes indicates a correlation between genera; the color of the edge represents a positive correlation (red) and a negative correlation (blue). The edge thickness indicates the correlation values; only significant interactions are shown (r > 0.6; p < 0.05) Over 98% representative OTUs belonged to Proteobacteria. Ensifer dominated both groups (84.9% in high-yield vs. 70.0% in low-yield), followed by Pseudomonas (6.4% vs. 4.0%); both genera were more abundant in the high-yield group, while other dominant genera were enriched in low-yield genotypes (Fig. 1c). Co-occurrence networks were less complex in high-yield nodules, with 11 significant correlations compared to 27 in low-yield in an analysis of the top 30 bacteria at the genus level (Fig. 1d, e). In high-yield group, Ensifer was negatively correlated with Pseudomonas, Serratia, Rhizobiaceae_unclassified, and Allorhizobium.Neorhizobium.Pararhizobium.Rhizobium. Greater connectivity and more positive correlations characterized low-yield group networks. We observed 15 and 19 nodes with a degree greater than zero in high-yield nodules and in low-yield nodules, respectively. Furthermore, the network in low-yield nodules exhibited a higher maximum degree value. Functional Predictions of Nodule Bacteria FAPROTAX annotated 43 functions linked to carbon, nitrogen, and sulfur cycling, and parasitism. Among the top five functions in terms of relative abundance, nitrogen fixation and aerobic chemoheterotrophy were significantly enriched in high-yield nodules, while chemoheterotrophy and fermentation were more abundant in low-yield genotypes. Nitrate reduction showed no significant difference between the two groups (Fig. 2).Fig. 2Extended error bar plot identifying significant differences between mean proportions of abundance of functions in high-yield group (H, coral) and low-yield group (L, cyan) samples. Corrected p values are shown at right. Only the results with statistical significance (p < 0.05) showed in the extended error bar plot Extended error bar plot identifying significant differences between mean proportions of abundance of functions in high-yield group (H, coral) and low-yield group (L, cyan) samples. Corrected p values are shown at right. Only the results with statistical significance (p < 0.05) showed in the extended error bar plot Composition of Rhizobial Communities Eleven OTUs were associated to Ensifer, with OTU153 (Ensifer meliloti) accounting for more than 79% of all nodule sequences. Other rhizobial genera comprised less than 7% of the community (Fig. 3a). To better visualize minor taxa, relative abundances excluding Ensifer were examined (Fig. 3b). Except for Mesorhizobium, most rhizobia differed significantly between high- and low-yield groups. Ensifer and Bradyrhizobium were enriched in high-yield group, whereas other genera were more abundant in low-yield group (Fig. 3c).Fig. 3Common and different rhizobia between two groups at genus levels (a), rhizobia except Ensifer (b), extended error bar plot identifying significant differences between mean proportions of abundance of functions in high-yield group (H, coral) and low-yield group (L, cyan) samples, only p < 0.05 showed (c) Common and different rhizobia between two groups at genus levels (a), rhizobia except Ensifer (b), extended error bar plot identifying significant differences between mean proportions of abundance of functions in high-yield group (H, coral) and low-yield group (L, cyan) samples, only p < 0.05 showed (c) Root AMF Diversity and Composition After processing, 121 AMF OTUs were identified, of which 100 were shared between yield groups, 18 were unique to high-yield group, and 3 to low-yield group. Richness was higher in high-yield group, but Shannon and Simpson indices were similar (Fig. 4a). PCoA revealed clear compositional separation, with the first axis explaining 44% of variance (Fig. 4b). Fig. 4Common and different species between two groups at different taxonomic levels (a), AMF and beta diversity metrics (b), Top 10 species with relative abundance of species (c). Co-occurrences network of top 30 species of AMF in high-yield groups (d) and in low-yield group (e). The nodes with different colors in the co-occurrences network represent different species; the node size represents the degree of the species in the co-occurrences network; the edge between nodes indicates a correlation between genera; the color of the edge represents a positive correlation (red) and a negative correlation (blue). The edge thickness indicates the correlation values; only significant interactions are shown (r > 0.6; p < 0.05) Common and different species between two groups at different taxonomic levels (a), AMF and beta diversity metrics (b), Top 10 species with relative abundance of species (c). Co-occurrences network of top 30 species of AMF in high-yield groups (d) and in low-yield group (e). The nodes with different colors in the co-occurrences network represent different species; the node size represents the degree of the species in the co-occurrences network; the edge between nodes indicates a correlation between genera; the color of the edge represents a positive correlation (red) and a negative correlation (blue). The edge thickness indicates the correlation values; only significant interactions are shown (r > 0.6; p < 0.05) All AMF OTUs belonged to Glomeromycota, represented by Glomus, Paraglomus, Diversispora, Diversisporaceae_unclassified and Glomeromycetes_unclassified. Glomus dominated (over 99.9%) in both groups. Diversispora was more abundant in high-yield roots, while Paraglomus and Glomeromycetes_unclassified were exclusive to them. Diversisporaceae_unclassified occurred only in low-yield roots. Among 22 species exceeding the 0.01% abundance threshold, 19 of those were shared between the two groups, while 2 of those were unique to the high-yield group. The ten most abundant taxa all belonged to Glomus. Glomus.sp.VTX00301, Glomus.intraradices.VTX00105, Glomus.viscosum.VTX00063, Glomus.MO.G17.VTX00114 and Glomus.sp.VTX00329 were significantly enriched in high-yield group than that in low-yield group, whereas Glomus.Yamato08.A1.VTX00100 and Glomus.sp.VTX00301 were more abundant in low-yield group (Fig. 4c). Co-occurrence analysis of AMF species revealed nine significant correlations (one positive, eight negative) in high-yield groups and five (three positive, two negative) in low-yield groups in an analysis of the top 30 at the species level (Fig. 4d, e). We observed 15 and 9 nodes with a degree greater than zero in high-yield group and in low-yield group, respectively. Furthermore, the network in high-yield nodules exhibited a higher maximum degree value. Correlations Between Rhizobial and AMF Communities Spearman correlation analysis at the genus level revealed predominantly negative associations within and between rhizobial and AMF taxa (Fig. 5). In high-yield group, correlations ranged from − 0.89 to 0.72. Glomus showed significant negative correlations with Diversispora (−0.71) and Glomeromycetes_unclassified (−0.49). Among rhizobia, Ensifer exhibited strong negative correlations with Allorhozobium.Neorhozobium.Pararhozobium.Rhinzobium (−0.46), Ochrobactrum (−0.51), Pseudomonas (−0.88), Rhzobiaceae_unclassified (−0.61). In the low-yield group, correlations ranged from − 0.82 to 0.80. Glomus was negatively correlated with Diversispora (−0.75) and Diversisporaceae_unclassified (−0.82). Within rhizobia, Ensifer was negatively correlated with Allorhozobium.Neorhozobium.Pararhozobium.Rhinzobium (−0.51), Ochrobactrum, Pseudomonas (−0.66), Shinella (−0.41), but showed a positive association with Pseudomonas (0.40). Overall, AMF genera tended to exhibit negative correlations with each other, a pattern also observed among rhizobial taxa. Fig. 5Spearman correlation heat map of AMF and rhizobia of genus level in different groups. The circle color represents positive (blue) and negative (red) correlations. The circle size indicates the correlation values. Only significant interactions are shown (p < 0.05). The labels in black font represent the Rhizobium genus, while those in blue font denote the AMF genus Spearman correlation heat map of AMF and rhizobia of genus level in different groups. The circle color represents positive (blue) and negative (red) correlations. The circle size indicates the correlation values. Only significant interactions are shown (p < 0.05). The labels in black font represent the Rhizobium genus, while those in blue font denote the AMF genus Discussion Reports on the diversity and composition of nodule bacteria and AMF in alfalfa root under the influence of genotype factors are scarce, but genotype is indeed a factor that affects the mutual selection of plants and microorganisms. Besides, it is necessary to study the interaction between beneficial microorganisms. More adequate analysis and assessment must be given more attention in order to achieve sustainable green agriculture. Microbial Diversity in Response to Genotype Factor Previous studies have extensively explored endophytic microbial communities in leguminous plants at the species level [24]. However, the impact of plant genotype on microbial communities remains unclear. For example, no significant differences were observed in endophyte diversity between the two Medicago truncatula genotypes [38]. In our study, the variation in α-diversity indices was inconsistent between bacteria and AMF. Shannon index and Simpson index reflected significant differences in bacterial diversity, while richness primarily accounted for differences in AMF communities. Both PCoA plots showed clear separation along the first principal coordinate (PC1), indicating that the diversity and composition of bacteria in nodule and AMF in root were influenced by the genotypes with low and high yield capacity (Fig. 4b). Although genotype marginally affected the AMF dominant genus, these findings improve our understanding of endophytic microbial diversity in relation to genotype-driven selection. Differences in Dominant Microorganisms and Their Composition The dominant rhizobia present in legume root nodules can vary by host species [39]. Consistent with prior studies, our results confirmed Ensifer as the dominant bacterial genus in alfalfa nodules, with minimal representation from other rhizobia [40, 41]. Notably, the relative abundance of Ensifer was higher in high-yield group, approaching 90%. Interestingly, Ensifer meliloti was not the most abundant rhizobial species in rhizosphere, suggesting complex selective pressures at the species level [38]. This dominant endophyte displayed strong host genotype preferences, aligning with earlier findings of genetic variation among strains nodulating different genotypes [42]. High-yield genotypes may selectively exclude low-efficiency nitrogen fixers, partially explaining the differing abundance of Ensifer between yield groups. Legume nodules host not only rhizobia but also a range of non-rhizobial endophytes that may contribute to plant productivity. Co-inoculation with these microbes may increase nodulation efficiency, promote growth, and improve biocontrol potential [43, 44]. In our study, Pseudomonas was the second most abundant bacterial genus in nodules and was particularly enriched in high-yield group, showing a pattern similar to that of Ensifer (Fig. 3). Pseudomonas species are well recognized as plant growth promoting bacteria [45], and co-inoculation with rhizobia has been reported to enhance nodulation, stimulate plant growth, and suppress plant diseases [46, 47]. Some Pseudomonas strains harbor the nodC gene and may possess nodulation-related functions, though their classification as rhizobia remains under debate [48]. The concurrent enrichment of Ensifer and Pseudomonas in high-yield genotypes suggests genotype-mediated microbial selection or superior colonization competence of these taxa within nodules. Overall, non-rhizobial nodule-associated bacteria represent a potentially important component of the nodule microbiome, and future research should further elucidate their physiological roles and the underlying mechanisms governing their interactions with host plants. In our data, Glomus dominated the AMF community, accounting for more than 99.9% of the sequences in both high- and low-yield groups (Fig. 4c). Glomus was also found to be the dominant genus within the AMF community colonizing Camellia oleifera [26]. Plant hosts appear to exert limited influence on AMF composition at the species level, as fungal symbionts are more strongly shaped by dispersal limitations and biogeographical constraints [27, 49]. Despite the taxonomic dominance of Glomus, the AMF community in the high-yield group exhibited greater richness, resulting in a more complex co-occurrence network (Fig. 4d, e). This pattern contrasts with the microbial community structure observed in root nodules. Additionally, negative correlations were detected between Glomus and both Paraglomus and Diversispora in the high-yield group (Fig. 5), as well as between Glomus and Diversispora in the low-yield group. These negative associations likely reflect competitive exclusion by the dominant genus Glomus [50]. Previous studies have shown that Paraglomus and several genera within the same family as Diversispora are sensitive to nitrogen availability, and that their symbiotic efficiency can be suppressed by the substantial nitrogen provision from nodule symbiosis, while Glomus tend to perform better under such conditions [51]. Notably, the increased proportion of positive correlations among AMF taxa in the high-yield group may be linked to greater carbon allocation by high-yield genotypes, potentially alleviating the competitive pressure among Glomus species. This hypothesis, however, requires further experimental validation through direct quantification of carbon allocation and assessment of potential synergistic interactions among these taxa. Functional Roles of Bacteria in Response to Genotype Factor The functional capacity of microbial communities is closely related to the dominance of key taxa. A foundational principle of symbiosis is bidirectional nutrient exchange. This is particularly evident in the nodule endosphere, where resident microbes are physiologically dependent on a substantial supply of carbon and energy from the host plant to sustain their metabolism. Consistent with this host-dependent and carbon-intensive lifestyle, our functional predictions revealed an overwhelming dominance of chemoheterotrophic and aerobic chemoheterotrophic pathways. Rhizobia encode the enzyme nitrogenase which catalyzes the conversion of atmospheric dinitrogen to ammonium, a process that is essential for crop productivity but highly energy demanding [14]. This process is ultimately fueled by aerobic respiration [52], providing a mechanistic explanation for the concurrent enrichment of aerobic chemoheterotrophy and nitrogen fixation functions observed in the high-yield group. The stronger nitrogen fixation capacity in this group was likely driven by the higher abundance of Ensifer and Pseudomonas. Synergistic interactions between these genera have been well documented, with Pseudomonas reported to enhance rhizobial symbiosis and reduce the formation of ineffective nodules [44, 47]. In contrast, the enrichment of chemoheterotrophic functions in the low-yield group was primarily attributable to increased fermentation-related pathways (Table S2). Several fermentation-associated genera identified in this study, including Promicromonospora, Aeromonas, and Serratia, have previously been described as opportunistic nodule endophytes [53]. It has been shown that nitrogen fixation rates during the reproductive stage of pea are closely linked to carbon availability within nodules [54], and that higher biological nitrogen fixation efficiency is associated with greater host biomass accumulation [55]. Together, these observations suggest that high-yield genotypes may preferentially allocate limited carbon resources to efficient rhizobia rather than to opportunistic microbes, thereby optimizing nitrogen fixation. Co-occurrence network analysis reveals a simpler microbial network in the high-yield group compared with the low-yield group. This difference may partly reflect the higher Shannon diversity observed in the low-yield group, which could give rise to more complex microbial interactions. However, the dominant genus Ensifer showed negative correlations with other nodule-associated microbes in both yield groups (Figs. 1d and 5), a pattern that contrasts with some previous reports [56]. Hence, these findings indicate that the general principles governing microbial interactions within nodules remain incompletely understood and warrant further investigation. Genotype-dependent differences in microbial functional potential, such as ligninolysis, have also been reported in soybean [57], consistent with our observations. Although ligninolysis accounted for only a small proportion of predicted functions and was associated with a single OTU in our dataset (Table S2), it further underscores the influence of host genotype on microbial function. Despite considerable efforts to isolate and culture these bacteria, successful re-inoculation into their original host plants has remained challenging [58], highlighting gaps in our understanding of the mechanisms underlying these functional differences. Our findings support the development of microbial inoculants comprising highly efficient Ensifer strains, either alone or in combination with Pseudomonas, to improve alfalfa growth and yield. Our work provides valuable reference data for the future microbial bio-fertilizer formulations. The Correlation Between Rhizobia and AMF Contrary to expectations, no significant positive correlation was observed between the dominant taxa Ensifer and Glomus. Although the effectiveness of co-inoculation with rhizobia and AMF has been widely investigated, their interactions remain unpredictable [59]. Some studies suggest AMF promote rhizobial nodulation or vice versa [60], while others report no significant effects [61]. Notably, AMF may serve as vectors for rhizobia through their hyphal networks or colonize senescent nodules post-fixation [62]. In our study, Ensifer negatively correlated with other rhizobial genera in high-yield group. Similarly, Glomus showed negative correlations with three other AMF genera. Negative correlations among dominant taxa may arise from niche competition for limited nutrients and space within the root environment [63–65]. Although genotype factor did not exert consistent effects on microbial interactions in our genus-level correlation analysis, an extensive body of prior research consistently demonstrates that the dual inoculation of rhizobia and arbuscular mycorrhizal fungi can substantially promote plant growth. Microbial symbionts play crucial roles in enhancing plant resilience and productivity [66]. Inherited microbial communities have even been linked to traits such as drought tolerance and disease resistance [67]. Harnessing this microbial potential is essential for improving soil health, crop performance, and ecosystem sustainability. Given its considerable potential for supporting environmentally sustainable agricultural development, this approach is expected to remain a high-priority focus in future research. Conclusion This study reveals that plant genotypes with varying yields harbor endophytic microbiomes disproportionately, which play crucial roles in plant growth and productivity. The dominant endophytic bacterium is identified as Ensifer within the root nodules of alfalfa and the dominant fungus colonizing the roots is Glomus. We have showed that the root-associated microbial community of alfalfa was significantly influenced by the genotype factor, as demonstrated by the distinct relative abundance of dominant bacterial taxa and the differential distribution of microbial functional traits. We argue that the significant negative correlations between dominant microbial taxa and other bacteria are primarily mediated through plant-microbe mutual selection, the provision of beneficial functional traits, and competitive interactions. In aggregate, these findings emphasized the importance of host genotype-mediated microbial recruitment. Given that the AMF and rhizobial symbioses are critical for crop growth, our findings might inform strategies to improve agricultural management. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (DOCX 48.1 KB) Supplementary Material 1 (DOCX 48.1 KB)