Soil pH as an external filter shaping stink bug–Burkholderia gut symbiosis Authors: Hideomi Itoh, Hiroyuki Shimoji, Daisuke Nakane, Seonghan Jang, Yoshitomo Kikuchi Published: 2026 ABSTRACT BackgroundMany animals and plants establish intimate symbiotic relationships with specific microorganisms acquired from the environment. Given the immense diversity of environmental microbiomes, selecting appropriate partners from such a vast microbial pool poses a critical challenge for host organisms. To meet this challenge, hosts have evolved sophisticated internal partner-choice mechanisms that ensure stable associations with beneficial microbes. However, because these symbionts primarily inhabit external environments, environmental conditions themselves are also expected to influence the establishment of symbiosis. Despite this expectation, the mechanistic role of external environmental filters in shaping the intended symbiosis remains largely unexplored. Focusing on stink bugs, which acquire their symbiotic bacteria from soil each generation, we investigated how soil properties influence the establishment of gut symbiosis in terrestrial insects.ResultsMicrobiome analyses confirmed that Burkholderia sensu lato overwhelmingly dominates a specific gut organ in six stink bug species from the superfamilies Coreoidea and Lygaeoidea, including serious agricultural pests (relative abundance ranging from 74.5 to 100%). Rearing experiments with isolated Burkholderia revealed that insects were strictly dependent on this symbiont; failure to acquire it from soil severely reduced host growth and reproduction, indicating that the availability of symbionts from soil can represent an ecological constraint. Field surveys identified patches of exceptionally high stink bug density in weedy fields with soil pH < 7.0, whereas such aggregations were absent in fields with pH ≥ 7.0. Laboratory experiments with collected field soils showed that the abundance of Burkholderia in soils was negatively correlated with soil pH, and stink bugs readily acquired their symbionts from soils with pH < 7.0 but rarely from soils with pH ≥ 7.0. Experimental manipulations of soil pH followed by rearing experiments confirmed that increasing soil pH to 7–8 markedly suppressed symbiont acquisition by the host, likely by impairing symbiont growth and motility.ConclusionsWe demonstrate that, beyond host-intrinsic mechanisms, a soil chemical property can act as an externally filter that constrains symbiont acquisition prior to colonization inside the host in a stink bug–Burkholderia symbiosis. This finding highlights how local environmental conditions can shape the assembly of environmentally acquired insect–microbe symbioses. Video AbstractSupplementary InformationThe online version contains supplementary material available at 10.1186/s40168-026-02402-z. Introduction Symbiosis with microorganisms is ubiquitous among animals and plants and has been a major driver of organismal evolution [1]. Given the enormous diversity of microbes in the environment, establishing and maintaining mutualistic relationships with specific partners is a considerable challenge for host organisms. Many insects and some marine invertebrates have evolved vertical transmission systems to ensure partner fidelity [2–4]. By contrast, numerous animals and plants do not vertically transmit their symbionts but instead acquire them from the environment every generation. To secure partner specificity under these conditions, hosts have evolved sophisticated partner choice mechanisms. For example, in the legume–Rhizobium nodule symbiosis and the squid–Vibrio bioluminescent symbiosis, strict partner selection is mediated through chemical crosstalk between host and symbiont [5, 6]. In many animals, the composition of gut microbial communities is shaped by gut physiology and immunity [7]. In environmentally acquired symbioses, the environment has often been regarded simply as a reservoir of symbiotic microorganisms. However, microbial distributions in the environment are far from uniform. Global-scale surveys have revealed pronounced geographic variation in bacterial and fungal communities [8, 9], while more localized studies show substantial heterogeneity even among nearby sites [10]. Among environmental variables, soil pH has been identified as one of the dominant determinants of soil microbial community structure across spatial scales, often exerting stronger effects than other physicochemical factors [8, 11–16]. Accordingly, soil pH represents a plausible ecological factor shaping the environmental context in which symbiotic microorganisms occur. From this perspective, the uneven distribution of symbiotic microorganisms in the environment may directly influence the establishment of symbioses through external, or pre-infection, selection. Despite its potential importance, meanwhile, the role of environmental heterogeneity in shaping host–microbe specificity remains largely unexplored. Phytophagous species of the superfamilies Coreoidea and Lygaeoidea of the order Heteroptera commonly establish gut symbioses with bacteria in Burkholderia sensu lato, including the genera Caballeronia and Paraburkholderia [17–24]. These symbionts are acquired from ambient soil during the early nymphal stage in every host generation and are specifically harbored in a crypt-bearing region of the posterior midgut [17, 25], where they enhance host survival and growth [17, 18] by performing pivotal metabolic roles [26, 27]. The multilayered host-internal mechanisms of symbiont filtering from environment microbiota have been well investigated in the bean bug Riptortus pedestris and its symbiont Caballeronia insecticola. The midgut of R. pedestris is divided into regions associated with food digestion (M1, M2, and M3) and the symbiotic organ (M4) [28]. Between the M3 and M4 regions lies a symbiont-sorting organ, the constricted region (CR), characterized by an extremely narrow lumen of only a few micrometers in diameter [28]. The CR, filled with a mucus-like matrix, functions as a physical barrier, preventing most non-symbiotic bacteria from passing through [28]. Caballeronia symbionts, in contrast, can traverse the CR by utilizing a unique flagellar-wrapping motility, allowing them to successfully colonize the symbiotic organ [28, 29]. In addition, the M3 and M4 regions exhibit high levels of digestive enzymes and antimicrobial peptides, which play crucial roles in selecting the symbiotic bacteria [30]. Furthermore, the CR gate is permanently sealed after symbiont passage and colonization in M4, thereby preventing secondary inoculation of contaminants [31]. Here we show that, beyond these host-internal filters such as physical barriers and antimicrobial peptides [28, 30, 31], an overlooked but critical layer of symbiont selection occurs externally, in the soil. We first identified the reliance on soil-derived Burkholderia in six stink bug species within the superfamilies Coreoidea and Lygaeoidea, including serious rice pests whose symbiotic associations have remained poorly characterized. Then, through characterization, field survey, and controlled rearing experiments, we further showed that soil pH constrains the density and activity of the symbiotic bacteria, and that symbiosis cannot be established in soils with pH ≥ 7. This observed dependence of gut symbiosis on soil fundamentally broadens current understanding of host–microbe interactions by demonstrating that external environmental properties, not only host-internal physiology, can dictate the establishment and evolution of gut symbioses. Results Obligate symbiosis with soil bacteria Wild populations of the stink bug species examined in this study, Cletus punctiger, C. schmidti, R. pedestris, Leptocorisa chinensis, Pachygrontha antennata, and Metochus uniguttatus, possess cryptic or tubular organs that develop at the posterior region of the midgut (symbiotic organ: M4B and M4 in Fig. 1A–C and Fig. S1), where the bacterial symbionts colonize [32]. Microbiome analysis of field-captured adult insects showed that 97.6–100% of individuals were positive for Burkholderia (Table S1), which consistently dominated the symbiotic organ, with relative abundance ranging from 74.5 to 100% (Fig. 1D and Fig. S2). Eggs laid by field-captured females and insects reared in clean Petri dishes were negative for Burkholderia (Fig. 1E and Table S1). Similarly, except for C. schmidti, all reared insects with an adult pair were also negative for Burkholderia (Table S1). Conversely, 93.8–100% of insects reared with the soil have acquired Burkholderia (Fig. 1F and Table S1). In the rearing experiments conducted in the plant pots, all insect individuals reared under soil-accessible conditions were infected with Burkholderia, whereas those isolated from the soil and accessible only to the plant were not (Fig. S3A–C). These results revealed that these stink bug species are entirely dependent on gut symbionts existing in the soil, although C. schmidti has shown partial vertical transmission.Fig. 1Characterization of gut symbiosis in the stink bugs used in this study. A The third instar and adult female of laboratory-maintained C. punctiger. The inset displays a wild individual of C. punctiger invading a rice paddy field (closed triangle). B Each dissected whole gut of the 4th instar of C. punctiger infected with Burkholderia (Sym) or uninfected (Apo). Abbreviations: M1, midgut first section; M2, midgut second section; M3, midgut third section; CR, constricted region; M4B, midgut fourth section with bulb; M4, midgut fourth section with crypts (symbiotic organ); H, hindgut. C Magnified photo of the M3 to M4 region of Sym and Apo insects. The closed triangles in M4 indicate crypts. Photos of other stink bug species and their whole guts used in this study are shown in Fig. S1. D Community structure of the gut microbiota in wild populations of six stink bug species. Data represent the average composition, with the number of examined individuals shown below each bar. Detailed sample information and individual gut microbiota community structures are provided in Table S2 and Fig. S2, respectively. E and F Infection rates to Burkholderia in insects reared without (E) and with soil (F). Values below each bar indicate the number of examined individuals. G Survival rates (i.e., adult emergence rates) of six stink bug species infected with Burkholderia (Sym) or uninfected (Apo). †Data of R. pedestris are from a previous study [33]. Numbers on the bars represent “number of survivors/total number of examined individuals.” Statistical analysis was performed using Fisher’s exact test (**, P < 0.01; ***, P < 0.001; n.s., not significant (P ≥ 0.05)) Characterization of gut symbiosis in the stink bugs used in this study. A The third instar and adult female of laboratory-maintained C. punctiger. The inset displays a wild individual of C. punctiger invading a rice paddy field (closed triangle). B Each dissected whole gut of the 4th instar of C. punctiger infected with Burkholderia (Sym) or uninfected (Apo). Abbreviations: M1, midgut first section; M2, midgut second section; M3, midgut third section; CR, constricted region; M4B, midgut fourth section with bulb; M4, midgut fourth section with crypts (symbiotic organ); H, hindgut. C Magnified photo of the M3 to M4 region of Sym and Apo insects. The closed triangles in M4 indicate crypts. Photos of other stink bug species and their whole guts used in this study are shown in Fig. S1. D Community structure of the gut microbiota in wild populations of six stink bug species. Data represent the average composition, with the number of examined individuals shown below each bar. Detailed sample information and individual gut microbiota community structures are provided in Table S2 and Fig. S2, respectively. E and F Infection rates to Burkholderia in insects reared without (E) and with soil (F). Values below each bar indicate the number of examined individuals. G Survival rates (i.e., adult emergence rates) of six stink bug species infected with Burkholderia (Sym) or uninfected (Apo). †Data of R. pedestris are from a previous study [33]. Numbers on the bars represent “number of survivors/total number of examined individuals.” Statistical analysis was performed using Fisher’s exact test (**, P < 0.01; ***, P < 0.001; n.s., not significant (P ≥ 0.05)) In the absence of soil-derived Burkholderia symbionts, survival rates were significantly reduced in all species except C. schmidti (Fig. 1G). Notably, M. uniguttatus and P. antennata failed to reach adulthood without acquiring Burkholderia (Fig. 1G). Although aposymbiotic individuals of Cletus species were not entirely incapable of reaching adulthood, they exhibited multiple morphological and reproductive abnormalities. Specifically, aposymbiotic Cletus individuals exhibited a decline in growth rate, dwarfed body size, and forewing atrophy (Fig. 2A–E). Furthermore, aposymbiotic adults suffered reproductive impairments, including immature reproductive organs, reduced mating behavior, lower egg production, and decreased hatching rates (Fig. 2F–J). These results demonstrate that the tested stink bug species are strictly dependent on soil-borne Burkholderia, indicating that the availability of Burkholderia from the ambient soil is a critical factor for development, reproduction, and population maintenance.Fig. 2Effects of Burkholderia symbiosis on host growth and fecundity in Cletus spp. Fitness effects of Burkholderia symbionts on insect growth (A–E) and fecundity (F–J). A Sym and Apo adult females (upper) and male (lower) of C. punctiger. B Developmental period (days) from hatching to adulthood. C Adult body length. D Whole-body dry weight. E Ratio of forewing length to body length in adult insects. F Reproductive organs in Sym and Apo adults: ovary (upper) and testes (lower). G Size of reproductive organs. H Mating rate. I Total number of eggs laid within 2 weeks after first oviposition. J Hatching rate of eggs laid by Sym and Apo females. Numbers in bars indicate “total number of examined insect samples” (B–E, G, I–J) or “number of successful pairs/total number of examined pairs” (H). Error bars represent standard deviation. Statistical analyses were performed using Student’s t-test, Welch’s t-test, Mann–Whitney U test, or Brunner–Munzel test after checking for normality and homogeneity of variance (***, P < 0.001) (B–E, G, I–J), or using Student’s t-test (***, P < 0.001) (H). Detailed results of the statistical analyses are provided in Table S3 Effects of Burkholderia symbiosis on host growth and fecundity in Cletus spp. Fitness effects of Burkholderia symbionts on insect growth (A–E) and fecundity (F–J). A Sym and Apo adult females (upper) and male (lower) of C. punctiger. B Developmental period (days) from hatching to adulthood. C Adult body length. D Whole-body dry weight. E Ratio of forewing length to body length in adult insects. F Reproductive organs in Sym and Apo adults: ovary (upper) and testes (lower). G Size of reproductive organs. H Mating rate. I Total number of eggs laid within 2 weeks after first oviposition. J Hatching rate of eggs laid by Sym and Apo females. Numbers in bars indicate “total number of examined insect samples” (B–E, G, I–J) or “number of successful pairs/total number of examined pairs” (H). Error bars represent standard deviation. Statistical analyses were performed using Student’s t-test, Welch’s t-test, Mann–Whitney U test, or Brunner–Munzel test after checking for normality and homogeneity of variance (***, P < 0.001) (B–E, G, I–J), or using Student’s t-test (***, P < 0.001) (H). Detailed results of the statistical analyses are provided in Table S3 Symbiosis establishment dependent on soil pH A number of studies have investigated the environmental determinants of soil microbiome diversity, consistently identifying soil pH as one of the most influential physicochemical parameters shaping microbial community structure and diversity; it exerts a substantial impact on microbial assemblages, often overriding other edaphic factors such as soil type, moisture, or nutrient content [8, 11–16]. Notably, a continent-wide survey reported that B. sensu lato are more prevalent in mildly acidic soils than in neutral or alkaline conditions [34]. Collectively, these findings suggest that soil pH may affect the local distribution of symbiotic Burkholderia and, by extension, influence the ecological interactions of stink bugs that rely on the environmental acquisition of these symbionts. To investigate patterns of stink bug biomass across fields differing in soil pH, we surveyed stink bug biomass across weedy fields with varying soil pH (Fig. 3A and Table S4). The major Coreoidea and Lygaeoidea species captured in these field surveys were C. punctiger, L. chinensis, and Paromius jejunus (Lygaeoidea) (Fig. S4A). Although the size of the stink bug population was not significantly different among fields, large numbers of nymphs and adults were observed at several sites with soil pH < 7.0 (Fig. 3B). Sensitivity analyses confirmed that these high-abundance data points were outliers (Fig. S4B), suggesting that localized stink bug surges tend to occur in mildly acidic soil environments.Fig. 3Soil pH preferences for symbiosis establishment inferred from field surveys and laboratory experiments. A Schematic overview of the field survey and geographic distribution of the examined weedy fields. Detailed site information and results are provided in Table S4. B Number of stink bugs collected from each weedy field for each insect species. The left and right panels show the number of adults and nymphs, respectively. Data points marked with an asterisk indicate outbreaks identified through statistical analysis (Fig. S4B). Colors represent variations in soil pH. C Schematic diagram of the laboratory infection test conducted on insects using soil samples. Detailed information of the soils is summarized in Table S5. D Bacterial abundance in the examined soil samples, estimated using qPCR analysis. E Relative abundance of Burkholderia in each soil microbiome. F and G Correlation between soil pH and infection rates of Burkholderia in C. punctiger (F) and R. pedestris (G), respectively. Each data point represents the infection rate among 8 individuals. H pH preferences for the growth of symbiotic Burkholderia strains derived from stink bug species. Error bars represent standard deviation (n = 3). Values marked with an asterisk indicate a significant difference compared with those at pH 7.0 (*, P < 0.05), as determined by Dunnett’s test Soil pH preferences for symbiosis establishment inferred from field surveys and laboratory experiments. A Schematic overview of the field survey and geographic distribution of the examined weedy fields. Detailed site information and results are provided in Table S4. B Number of stink bugs collected from each weedy field for each insect species. The left and right panels show the number of adults and nymphs, respectively. Data points marked with an asterisk indicate outbreaks identified through statistical analysis (Fig. S4B). Colors represent variations in soil pH. C Schematic diagram of the laboratory infection test conducted on insects using soil samples. Detailed information of the soils is summarized in Table S5. D Bacterial abundance in the examined soil samples, estimated using qPCR analysis. E Relative abundance of Burkholderia in each soil microbiome. F and G Correlation between soil pH and infection rates of Burkholderia in C. punctiger (F) and R. pedestris (G), respectively. Each data point represents the infection rate among 8 individuals. H pH preferences for the growth of symbiotic Burkholderia strains derived from stink bug species. Error bars represent standard deviation (n = 3). Values marked with an asterisk indicate a significant difference compared with those at pH 7.0 (*, P < 0.05), as determined by Dunnett’s test On the other hand, because field surveys inevitably involve several uncontrolled factors—such as insecticide use, differences in weeding schedules, and variation in plant biomass—these local environmental variations may obscure or confound the specific effects of individual soil properties on symbiont acquisition. Mechanistic links between soil pH and symbiont acquisition were therefore examined under controlled laboratory conditions. Using soil samples collected from agricultural and weedy fields across Japan including soils from the 29 insect survey sites (Table S5), we analyzed the relationships among soil pH and Burkholderia availability of stink bugs (Fig. 3C). Soil pH did not correlate with the total bacterial density (generalized linear model (GLM) with Gaussian, LR test, χ2 test, χ2 = 1.580, df = 1, P = 0.205, R2 = 0.075; Fig. 3D and Table S6); however, it significantly negatively correlated with the relative abundance of Burkholderia (GLM with beta-binomial, LR test, χ2 = 23.589, df = 1, P < 0.001, R2 = 0.016, Q < 0.001; Fig. 3E and Table S6). Among 14 chemical variables, soil pH and exchangeable calcium (which is sensitive to soil pH) were identified as the main factors influencing the relative abundance of Burkholderia in the soil microbiota (Table S7), consistent with a previous continental-scale study reporting a negative correlation between soil pH and Burkholderia abundance [34]. Afterward, to assess whether soil pH affects symbiont acquisition, the collected soils were fed with nymphs of C. punctiger and R. pedestris. The infection rates of both stink bug species drastically declined in soils with pH ≥ 7.0, whereas soils with pH < 7.0 supported efficient symbiont acquisition (C. punctiger: GLM with beta-binomial, LR test, χ2 = 43.857, df = 1, P < 0.001, R2 = 0.366; R. pedestris: GLM with beta-binomial, LR test, χ2 = 33.09, df = 1, P < 0.001, R2 = 0.312; Fig. 3F–G and Table S6). This pH-dependence of Burkholderia availability was also observed in four other stink bug species (Fig. S5). These results suggest that soil pH affects symbiont acquisition by stink bugs, with pH 7.0 acting as a tipping point for the success or failure of symbiont acquisition from field-derived soils. The Burkholderia abundance in soils at this tipping point was estimated to be 7.0 × 105 copies g−1 soil, which can be regarded as the threshold required for successful symbiont acquisition (Fig. S6). To confirm how pH affects symbiont growth, we examined the in vitro growth rates of symbiotic Burkholderia strains derived from stink bugs, along with some type strains within the B. sensu lato, under different pH levels. Our results demonstrated that all Burkholderia strains grew well under low pH conditions (< 7.0) compared to pH 7.0 (Fig. 3H and Fig. S7). However, under high pH conditions (> 7.0), the growth rate of Burkholderia significantly decreased (Fig. 3H and Fig. S7). Experimental modification of soil pH inhibits the establishment of symbiosis To further investigate the causal relationship between soil pH and symbiosis establishment, we experimentally increased the pH of mildly acidic soils with pH 5.6 (pH 5.6 fraction) and 5.9 (pH 5.9 fraction) using CaCO3 (Fig. 4A–B). In these pH-adjusted soils (maintained at pH 7–8), the infection rate of C. punctiger significantly decreased as soil pH increased in both mildly acidic soils (pH 5.6 fraction: GLM with beta-binomial, LR test, χ2 = 45.968, df = 1, P < 0.001, R2 = 0.248; pH 5.9 fraction: GLM with beta-binomial, LR test, χ2 = 25.943, df = 1, P < 0.001, R2 = 0.261; Fig. 4C and Table S6). A similar negative effect of elevated pH on symbiosis was observed for R. pedestris (pH 5.6 fraction: GLM with beta-binomial, LR test, χ2 = 46.17, df = 1, P < 0.001, R2 = 0.304; pH 5.9 fraction: GLM with beta-binomial, LR test, χ2 = 34.715, df = 1, P < 0.001, R2 = 0.228; Fig. 4D and Table S6). In parallel, Burkholderia abundance in soil also declined with increasing soil pH (pH 5.6 fraction: GLM with Gaussian, LR test, χ2 = 10.237, df = 1, P < 0.001, R2 = 0.145; pH 5.9 fraction: GLM with Gaussian, LR test, χ2 = 20.736, df = 1, P < 0.001, R2 = 0.385; Fig. 4E and Table S6). Similarly, negative impacts were detected on infection rates when soil pH was increased using steel slag or biochar instead of CaCO3 (Fig. 4F). The overall results from our laboratory experiments demonstrated that soil pH of ≥ 7.0 can inhibit symbiosis establishment via the reduced abundance of Burkholderia in the soil.Fig. 4Effect of soil pH modification on symbiosis establishment. A Schematic representation of the soil pH modification experiment. The mildly acidic soils with pH 5.6 and pH 5.9 correspond to the S32 and S33 soils, respectively (Table S5). B Transition in soil pH following modification to 7–8 via the addition of CaCO3 (n = 3). Error bars represent standard deviation (n = 3). C and D Transition in the infection rate of C. punctiger (C) and R. pedestris (D) after soil pH modifications. Numbers with bars indicate “the number of infected individuals/total number of examined individuals.” E Correlation between soil pH and the abundance of Burkholderia in the soil during the experiments. F Effect of soil pH modification on the infection rate using steel slag and chicken manure biochar. Numbers with bars indicate “the number of infected individuals/total number of examined individuals” Effect of soil pH modification on symbiosis establishment. A Schematic representation of the soil pH modification experiment. The mildly acidic soils with pH 5.6 and pH 5.9 correspond to the S32 and S33 soils, respectively (Table S5). B Transition in soil pH following modification to 7–8 via the addition of CaCO3 (n = 3). Error bars represent standard deviation (n = 3). C and D Transition in the infection rate of C. punctiger (C) and R. pedestris (D) after soil pH modifications. Numbers with bars indicate “the number of infected individuals/total number of examined individuals.” E Correlation between soil pH and the abundance of Burkholderia in the soil during the experiments. F Effect of soil pH modification on the infection rate using steel slag and chicken manure biochar. Numbers with bars indicate “the number of infected individuals/total number of examined individuals” Immediate inhibition of symbiosis via pH-induced loss of symbiont motility Notably, 7 days after soil pH modification, the infection rate of stink bugs with Burkholderia decreased drastically (Fig. 4C–D), although no reduction in Burkholderia abundance was observed in the soil compared with the pre-adjustment levels (Fig. S8). Importantly, in soils adjusted to pH 5.6 and 5.9, Burkholderia densities remained above the estimated threshold for successful colonization (7.0 × 105 copies g−1 soil), even 90 and 30 days after pH modification, respectively (Fig. S8), as inferred from infection experiments using 42 field soils (Fig. S6). This result suggests that the establishment of stink bugs–Burkholderia symbiosis is not solely dependent on the amount of Burkholderia in the soil. A previous study showed that Burkholderia must pass through a narrow constricted region (CR) in the midgut to colonize the symbiotic organ (Fig. 1C and Fig. S1), and this process depends on flagellar motility: Burkholderia uses flagellar motility to pass through the CR, but non-motile mutants cannot enter this CR [28]. As bacterial flagella are driven by a proton gradient across the cell membrane, motility is enhanced under mildly acidic conditions (pH < 7.0) as seen in Escherichia coli and Salmonella enterica [35]. In vitro observations confirmed that Burkholderia cells expressed active motility at pH 5.5–6.5, whereas most cells lost their motility at pH 7.5 or 8.0 (Fig. 5A–B). Direct observations of the flagella showed that most cells retained their flagella at pH 6.0 but were absent at pH 8.0 (Fig. 5C–D). Furthermore, in vivo observations using Burkholderia cultured at various pH levels showed that, at pH < 7.0, Burkholderia passed through the CR to the midgut’s fourth section with bulb (M4B) in the insect gut, while those cultured at pH ≥ 7.0 remained in the midgut’s third section (M3) without passing through the CR (Fig. 5E and Movies S1–S2). Together, these results suggest that the rapid inhibition of symbiosis observed shortly after soil pH modification to ≥ 7.0 is not attributable to reduced Burkholderia abundance, but can be explained by a direct negative effect of elevated pH on Burkholderia motility, which prevents their passage through the constricted region and delays symbiosis establishment. Feeding stink bugs with soil suspensions at various pH levels revealed that, although the pH of the M3 region responded to the pH of the ingested soil suspension, it remained consistently below 7.0 across all treatments (Fig. S9). The flagellar motility of Burkholderia was restored in vitro when the pH was reverted from ≥ 7.0 to 6.0 (Fig. S10), suggesting that there is a mechanism in the M3 region that prevents motility recovery, independent of pH. It is known that various antimicrobial substances are secreted in the M3 region [30], and the stressful environment inside the insect gut may hinder the recovery of motility in Burkholderia symbionts. Collectively, the temporal decoupling between symbiont abundance and infection success observed in the soil pH modification experiments indicates that reduced bacterial motility provides a plausible explanation for the rapid decline in colonization following pH manipulation, whereas longer-term reductions in symbiont density may contribute to sustained suppression at later time points (Fig. 4C–D and Fig. S8).Fig. 5Rapid impact of pH modification on the motility of Burkholderia observed through in vitro and in vivo experiments. A Movement trajectories (direction: red to blue) of Burkholderia in broth medium with pH 6.0, 7.0, and 8.0 over a 2-s observation period. B Swimming speed distribution of Burkholderia cells (n = 100) recorded 4 h after pH adjustment. C Flagellar observation of Burkholderia cells 4 h after pH adjustment using fluorescence microscopy. D Proportion of flagellated cells calculated using fluorescence microscopy (n = 100). Statistical analysis was performed using Fisher’s exact test: *** (P < 0.001). E In vivo observation: path (CR) to the symbiotic organ in the midgut of R. pedestris fed with Burkholderia treated at pH 5.5–8.0. The green signals represent GFP-labeled Burkholderia. Abbreviations: M3, midgut third section; CR, constricted region; M4B, midgut fourth section with bulb Rapid impact of pH modification on the motility of Burkholderia observed through in vitro and in vivo experiments. A Movement trajectories (direction: red to blue) of Burkholderia in broth medium with pH 6.0, 7.0, and 8.0 over a 2-s observation period. B Swimming speed distribution of Burkholderia cells (n = 100) recorded 4 h after pH adjustment. C Flagellar observation of Burkholderia cells 4 h after pH adjustment using fluorescence microscopy. D Proportion of flagellated cells calculated using fluorescence microscopy (n = 100). Statistical analysis was performed using Fisher’s exact test: *** (P < 0.001). E In vivo observation: path (CR) to the symbiotic organ in the midgut of R. pedestris fed with Burkholderia treated at pH 5.5–8.0. The green signals represent GFP-labeled Burkholderia. Abbreviations: M3, midgut third section; CR, constricted region; M4B, midgut fourth section with bulb Discussion This study demonstrates that soil pH can play a pivotal role in the establishment of symbiosis between stink bugs and their beneficial gut bacteria. Using a multi-scale approach integrating field surveys, soil chemical profiling, laboratory rearing experiments, and bacterial physiology, we found that mildly acidic soils (pH < 7.0) facilitate efficient acquisition of Burkholderia symbionts, whereas neutral to alkaline soils (pH ≥ 7.0) constrain this process under controlled conditions (Figs. 3 and 4). This disruption reflects both a reduction in the environmental abundance of the symbiont and a loss of its flagellar motility—an essential trait for host colonization (Figs. 4E and 5). These results indicate that soil physicochemical properties can function as external environmental filters influencing the availability of symbionts before they colonize the host gut. Although most research has focused on internal partner-choice mechanisms, our results suggest that external environmental selection can represent an additional and previously underappreciated layer shaping the establishment of host–symbiont associations. Although the insects studied here harbor environmentally acquired (i.e., horizontally transmitted) symbioses, many terrestrial invertebrates possess vertically transmitted symbioses, where the symbionts are inherited from the parental individuals via diverse manners such as coprophagy and transovarial transmission [36]. Such vertically transmitted systems are notably rare in marine animals but relatively common in terrestrial invertebrates [37]. Our findings highlight a potential ecological vulnerability of environmentally acquired symbioses in terrestrial habitats: when soil pH exceeds 7.0, stink bugs fail to acquire their beneficial partners. We propose that this ecological uncertainty may have acted as a selective pressure favoring the evolution of vertical transmission in insects. The concept of “partner assurance” has been hypothesized as a major advantage of vertical transmission [38], and our results provide empirical support for this idea in the context of soil-mediated constraints on symbioses. In this context, it is noteworthy that C. schmidti examined in this study (Table S1), squash bug Anasa tristis (Coreoidea), and oriental chinch bug Cavelerius saccharivorus (Lygaeoidea) exhibit partial vertical transmission of Burkholderia symbionts [39, 40], potentially representing intermediate stages in the evolutionary transition from environmental acquisition to vertical inheritance. At the symbiont level, although the evolution of Na+-driven flagellar motility could be considered a potential adaptation to alkaline soils [41], such a transition would likely require extensive physiological reorganization beyond the flagellar motor itself, including cellular pH homeostasis, ion transport systems, and membrane properties. Moreover, given that Burkholderia symbionts must function both as free-living soil bacteria and as insect gut symbionts, extensive adaptation to alkaline environments may be evolutionarily constrained by trade-offs with host-associated fitness, such as reduced colonization efficiency or impaired symbiotic performance within the host. Under such constraints, rapid symbiont evolution or replacement that circumvents soil pH–mediated barriers may be limited, thereby maintaining the sensitivity of environmentally acquired symbioses to abiotic soil conditions. Because many symbiotic microorganisms are essential for their hosts, the distribution of microbes in the environment may ultimately influence insect population dynamics and, potentially, their geographic distribution. While insect population dynamics are conventionally attributed to factors such as food availability, natural enemies, or climatic conditions [42–45], our findings suggest that soil chemical properties—particularly pH—can contribute to the success or failure of symbiosis establishment, which in turn influence host growth and reproduction (Fig. 6). Although previous studies have shown that soil pH is a major determinant of soil microbial communities [8, 11–16], how this abiotic factor shapes interactions between terrestrial animals and environmentally acquired symbionts remains poorly understood. Moreover, despite increasing recognition that some animal–microbe associations depend on soil microorganisms [46–48], the role of abiotic soil factors in regulating these interactions has been largely overlooked in both soil science and entomology. By bridging this gap, our study highlights one plausible cross-kingdom pathway whereby pH-mediated shifts in microbial communities could cascade through symbiotic interactions to influence insect ecology.Fig. 6Conceptual framework illustrating a proposed pathway linking soil pH and insect–microbe symbiosis. The left and right panels illustrate hypothesized relationships among soil pH, soil-derived Burkholderia symbionts, and host stink bugs in weedy fields with mildly acidic soils (pH < 7.0) and those with neutral to mildly alkaline soils (pH ≥ 7.0), respectively. Effects on host growth and reproduction at individual level are supported by laboratory experiments, whereas consequences for population dynamics in the field remain hypothetical and require future studies that incorporate additional ecological factors Conceptual framework illustrating a proposed pathway linking soil pH and insect–microbe symbiosis. The left and right panels illustrate hypothesized relationships among soil pH, soil-derived Burkholderia symbionts, and host stink bugs in weedy fields with mildly acidic soils (pH < 7.0) and those with neutral to mildly alkaline soils (pH ≥ 7.0), respectively. Effects on host growth and reproduction at individual level are supported by laboratory experiments, whereas consequences for population dynamics in the field remain hypothetical and require future studies that incorporate additional ecological factors Beyond direct effects on symbiont acquisition, soil pH can also affect insect performance indirectly through its effects on host plants. Soil pH influences plant nutrient availability, growth, and stress tolerance [49, 50], and these plant-mediated bottom-up effects could independently or jointly shape insect occurrence and performance in the field. Thus, any effects of soil pH on population dynamics of insect are likely to reflect the combined outcomes of multiple pathways, including direct constraints on symbiont acquisition, plant-mediated effects on host nutrition and physiology, and other soil physicochemical or biotic factors operating in natural environments. However, our field survey did not explicitly account for several ecological variables known to influence insect occurrence and performance, including host plant quality, predation pressure, or land management practices such as pesticide or herbicide application history. As a result, the current field data alone cannot provide a direct causal link between soil pH and insect population dynamics. Causal relationships among soil pH, symbiont acquisition, plant-mediated effects, and insect performance therefore remain to be rigorously tested. Future studies using controlled or semi-controlled systems, such as greenhouse or mesocosm experiments in which soil pH and other biotic and abiotic factors can be systematically manipulated, will be essential for disentangling the relative contributions of soil physicochemical properties, plant-mediated effects, and microbial processes in shaping insect–microbe interactions and their ecological consequences. The importance of the gut microbiome is now widely recognized across animal taxa, from vertebrates to invertebrates [51], and there is growing interest in harnessing microbial communities for host control. Most of these strategies focus on manipulating internal microbial communities using prebiotics and probiotics [52]. In contrast, we propose the novel concept of “exobiotics,” referring to environmental interventions that modulate microbial reservoirs outside the host to influence host–microbe interactions. Our findings demonstrate that regulating microbiome dynamics externally can shape the establishment and maintenance of symbiosis. This external dimension has been largely overlooked, except in some marine systems such as corals and bioluminescent squids [6, 53]. By shedding light on this overlooked axis, our work advances a new conceptual framework that emphasizes the triadic relationship among hosts, symbionts, and the environment. This perspective not only enriches symbiosis theory but also provides a scientific foundation for potentially innovative strategies to manage insect populations in agriculture. For example, adjusting soil pH with CaCO3, a low-toxicity compound widely used in farming, may reduce symbiont acquisition in weedy breeding grounds of rice stink bugs [54–56], thereby suppressing their population growth and subsequent migration into rice paddies (Fig. S11). The environment has long been regarded merely as a reservoir of symbiotic microorganisms. This study challenges this view by identifying environmental contexts in which symbionts cannot be acquired, thereby revealing a critical external constraint on symbiosis. These findings highlight the overlooked role of abiotic soil properties in the establishment, maintenance, and evolution of symbiosis, and underscore the importance of symbiont selection processes occurring outside the host. Materials and methods Insects We maintained six stink bug species spanning two superfamilies, four families, and five genera: C. punctiger, L. chinensis, and P. antennata collected from weedy fields (mainly Digitaria ciliaris and Setaria viridis) in Tsukuba (C. punctiger) and Ishioka (L. chinensis and P. antennata), Ibaraki, Japan; Cletus schmidti from Polygonaceae plants (Persicaria lapathifolia) in Sapporo, Hokkaido, Japan; R. pedestris from a soybean field in Tsukuba, Ibaraki, Japan [17]; and M. uniguttatus from a sparse forest in Naha, Okinawa, Japan (Fig. 1A and Fig. S1). All species except M. uniguttatus are agricultural pests [56]. Especially, C. punctiger and L. chinensis are serious rice pests in Eastern Asia, known to cause sterility and pecky rice [56]. Insects were maintained at 25 °C under a long-day regimen (16 h light, 8 h dark), and fed distilled water (DW) and various seeds: C. punctiger and M. uniguttatus with dry wheat (Triticum aestivum) and sunflower (Helianthus annuus); C. schmidti with dry wheat, sunflower, and buckwheat (Fagopyrum esculentum); R. pedestris with dry soybean (Glycine max); L. chinensis with wet foxtail millet (Setaria italica); and P. antennata with dry wheat, sunflower, southern crabgrass (Digitaria ciliaris), green foxtail (Setaria viridis), and Japanese millet (Echinochloa esculenta). To maintain symbiotic rearing lines, soil S32 (Table S5) was routinely added to DW and provided to hatchlings. Insect samples used in each experiment are listed in Tables S2 and S8. Soils Forty-two soil samples were collected from weedy and agricultural fields across Japan (Table S5). Plant residues and stones were removed with a 2-mm sieve before measuring soil pH and water content. Soil pH was measured by mixing sieved soil with DW (soil/water; 1/5, w/v) and using a pH meter (LAQUAtwin pH-22B, Horiba, Kyoto, Japan). Water content was determined by weighing several grams of sieved soil before and after drying at 110 °C for 15 h. The maximum water-holding capacity (MWHC) of soils S32 and S33 was measured using Hilgard’s method [57]. Other soil chemical properties were measured and analyzed according to standard methods [58]: electric conductivity (EC; platinum electrode method); cation exchange capacity (CEC; semi-micro Schollenberger method); exchangeable calcium and magnesium (atomic absorption spectrometry); exchangeable potassium (flame photometry); available phosphate (Truog method); phosphate absorption coefficient (ammonium phosphate solution method); NH4+-N (indophenol-blue method); NO3−-N (UV spectrophotometry); total carbon and nitrogen (dry combustion method); and humus (Kumada method). All measured values were summarized in Table S5. Soil classifications were inferred from collection site coordinates using the Japan Soil Inventory (NARO, https://soil-inventory.rad.naro.go.jp/). Sieved soil was stored at 4 °C for the rearing/infection experiments and DNA extraction. DNA preparation The midgut regions of M4B and M4 (symbiotic organs) were dissected from each insect using fine forceps and tweezers under an optical microscope (S9D, Leica Microsystems, Wetzlar, Germany) in a plastic Petri dish filled with phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1.5 mM KH2PO4, pH 7.5). After digestion in ProK buffer (0.1 M NaCl, 10 mM Tris–HCl (pH 8.0), 1 mM EDTA, 0.2 mg/mL proteinase K) at 56 °C for 2 h, the resulting solution was purified by phenol:chloroform:isoamyl alcohol (25:24:1) extraction and ethanol precipitation, as described previously [32]. The quality of the extracted DNA from the insect samples was checked by PCR amplification of the insect cytochrome oxidase I (COI) gene with a universal primer set, LCO1490 and HCO2198 [59] for C. punctiger, C. schmidti, L. chinensis, R. pedestris, and M. uniguttatus. For P. antennata, however, the COI gene could not be reliably amplified using this primer set, and thus, the 18S rRNA gene was amplified using the primer set, pan18S1552F (5′-CTGTGATGCCCTTAGATGTT-3′) and pan18S1908R (5′-ACCTACGGAAACCTTGTTAC-3′). Soil DNA extraction was conducted by considering 0.25 g (wet weight) of each soil sample using the ISOIL for Beads Beating (Nippon Gene, Tokyo, Japan) according to the manufacturer’s instructions, with the following additional treatments: the addition of 0.02 g skim milk to the lysis buffer to improve DNA extraction efficiency [60] and post-elution purification using ribonuclease A (Nippon Gene). The resulting extracts were further purified using OneStep PCR Inhibitor Removal Kits (Zymo Research, Irvine, CA, USA) to remove the PCR inhibitors as much as possible. The integrity and quantity of the purified DNA were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) and a Qubit 4 Fluorometer (Invitrogen, Carlsbad, CA, USA) with Qubit dsDNA HS Assay Kits (Invitrogen), respectively. Diagnostic PCR for Burkholderia infection To determine Burkholderia infection in insect samples, diagnostic PCR was conducted using specific primer sets, Bf and Br, for the 16S rRNA gene of B. sensu lato [61]. Previous studies have indicated that stink bugs can be infected with Burkholderia sensu stricto, including the B. cepacia complex and B. pseudomallei clade, Paraburkholderia (Burkholderia spp. in plant-associated beneficial and environmental clade), and Caballeronia (B. glathei clade or Burkholderia spp. in stink bug-associated beneficial and environmental clade) [33, 40, 47], as also shown in this study (Fig. S2). Therefore, this study covers B. sensu lato, including these previous Burkholderia clades [62]. PCR amplification was performed using an AmpliTaq Gold 360 Master Mix (Invitrogen) according to the manufacturer’s instructions under the following thermocycling conditions: 95 °C for 10 min, followed by 25 cycles of 95 °C for 30 s, 55 °C for 60 s, and 72 °C for 30 s. The PCR products were stained with Novel Juice (GeneDireX, Las Vegas City, NV, USA) and checked by electrophoresis on a 2% agarose gel. Microbiome analysis of the insect gut and soil Deep sequencing of the PCR amplicon of the variable region (V4) of the bacterial 16S rRNA gene was performed for DNA samples of soils (Table S5) and insects (Table S2). DNA libraries for the Illumina sequencing platform were prepared and sequenced using 515 F and 806R primers [63] on an Illumina iSeq 100 sequencer (Illumina, San Diego, CA, USA), as described previously [24]. Raw sequencing reads were first processed to remove primers, perform quality filtering, and assemble paired-end reads using Trimmomatic ver. 0.39 [64] with the SLIDINGWINDOW:20:30 setting, followed by fastq-join from ea-utils v1.1.2 (https://github.com/ExpressionAnalysis/ea-utils). The resulting assembled sequences were further curated in Mothur ver. 1.42.3 [65]: sequences shorter than 200 bp or longer than 300 bp were discarded using the screen.seqs command, and chimeric sequences were detected and eliminated with the chimera.uchime command using reference-based alignment against the SILVA database (release 128) [66]. Taxonomic classification was conducted with the RDP classifier ver. 2.13 [67], applying a 50% confidence cutoff. Sequences identified as chloroplasts were removed, retaining only those assigned to the kingdom Bacteria for downstream microbiome analyses. All steps were carried out following the procedure described in a previous study [68]. To estimate bacterial abundance, quantitative PCR (qPCR) was performed to amplify the V4 of the bacterial 16S rRNA gene using the LightCycler 96 System (Roche Diagnostics, Basel, Switzerland) with THUNDERBIRD SYBR qPCR Mix (Toyobo, Otsu, Shiga, Japan). The reaction mixture comprised 2 × THUNDERBIRD SYBR qPCR Mix, 0.3 μM primers 515 F and 806R, and soil DNA samples. The PCR conditions were as follows: initial denaturation at 95 °C for 30 s, followed by 45 cycles of 95 °C for 15 s, 55 °C for 30 s, and 72 °C for 30 s. The absolute number of bacterial 16S rRNA gene copies was determined based on a standard curve constructed using tenfold serial dilutions of the target PCR product of Burkholderia sp. SFA1, as described previously [69]. Additionally, Burkholderia abundance was calculated by multiplying the bacterial abundance through this qPCR by the relative abundance of Burkholderia obtained from the above amplicon sequencing. Isolation and inoculation of Burkholderia symbionts Burkholderia symbionts were cultured and orally inoculated into early nymphs following the established methods for R. pedestris [33]. For C. punctiger, R. pedestris, L. chinensis, and P. antennata, we used the strains CPU52, RPE64, LCH90, and PAN135, respectively, which were previously isolated from each insect species [32]. For C. schmidti, a Burkholderia symbiont strain, designated CSC94, was newly isolated from a wild individual collected in Sapporo, Hokkaido, Japan. This strain was obtained by incubating the crypt contents on 1.5% agar plates containing YG medium (0.5% yeast extract, 0.4% glucose, 0.1% NaCl) at 30 °C for 2 days and was identified by sequencing the 16S rRNA gene, as described previously [70]. For M. uniguttatus, we were unable to culture the Burkholderia symbiont from wild individuals and thus used the strain THE68, previously isolated from the gut of Togo hemipterus, within the same family (Rhyparochromidae) as M. uniguttatus [32]. The symbiont strains were precultured overnight in YG medium with shaking at 180 rpm and 30 °C. After an additional 2-h culture in fresh YG medium, symbiont cells were collected by centrifugation, suspended in DW adjusted to 104 CFU/mL, and then fed to hatchlings of each insect by soaking the cell suspension in cotton pads. Once the nymphs molted to the 3rd instar stage, the symbiont-containing water was replaced by symbiont-free DW, and the nymphs were reared to adulthood. Detection of the transmission mode To elucidate the transmission mode of the Burkholderia symbiont in stink bug species, a series of rearing experiments were conducted as follows. The symbiont-infection status of the parents was confirmed by diagnostic PCR for the Burkholderia symbiont as described above.(i)Wild population inspection. To check the prevalence of the Burkholderia symbiont in wild populations of stink bugs, adult insect samples collected from fields in Japan (Table S8) were subjected to diagnostic PCR. The data for R. pedestris and L. chinensis were obtained from a previously published study [71].(ii)Egg inspection. The possibility of vertical symbiont transmission via eggs, such as egg-smearing and transovarial manners, was investigated by collecting eggs from wild individuals and performing diagnostic PCR. The data for R. pedestris was obtained from a previously published study [17].(iii)Rearing of nymphs in clean Petri dishes. Hatchlings from infected parents were reared in a clean Petri dish until the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis) and subjected to diagnostic PCR. Different developmental stages were used for diagnostic PCR due to species-specific biological and technical constraints: M. uniguttatus and P. antennata aposymbiotic nymphs rarely survive beyond the 2nd instar; R. pedestris can be reliably diagnosed at the 3rd instar based on previous studies [25, 33]; and in C. punctiger, C. schmidti, and L. chinensis, 3rd-instar nymphs are relatively small and difficult to dissect reliably, making the 4th instar more suitable for consistent sampling.(iv)Rearing of nymphs in clean Petri dishes with soil. The possibility of environmental symbiont transmission was investigated by rearing nymphs in a clean Petri dish with soil S32 collected from a field in Sapporo, Hokkaido, Japan (Table S5). The soil was suspended in DW and fed to hatchlings during their development, and the resulting insect samples were subjected to diagnostic PCR. Details are described in the “Infection test using soil samples” section of the “Materials and methods” section.(v)Rearing of nymphs in clean Petri dishes with infected parents. To check the possibility of coprophagy, hatchlings were reared in a clean Petri dish with their infected parents with GFP-labeled Burkholderia symbionts that were made using the Tn7 minitransposon system, as described previously [72]. After reaching the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis), these insects were subjected to fluorescence microscopy observation or diagnostic PCR.(vi)Rearing of nymphs in plant pots. To test the possibility of symbiont transmission from plants, hatchlings of R. pedestris were reared in soil pots with soybean plants. The sieved soil S32 was transferred into plastic pots (10 × 8 cm; opening diameter × depth), and a soybean seedling was planted there. After 2 weeks of growing the soybean plant, 15 hatchlings were released into a pot covered with a transparent tube and a paper lid (Fig. S3A), fed with 5 dry soybeans, and maintained at 25 °C under a controlled light cycle of 16 h of light followed by 8 h of darkness. For comparison, to prevent insects from directly contacting the soil, they were also reared in a confined space where the above-ground parts of plants, previously sterilized with 70% ethanol and rinsed with sterilized DW, were covered with non-woven fabric (Fig. S3B). Each experimental pot was prepared with three replicates. After reaching the 3rd instar (10 days of rearing), the collected insects were subjected to diagnostic PCR. (i)Wild population inspection. To check the prevalence of the Burkholderia symbiont in wild populations of stink bugs, adult insect samples collected from fields in Japan (Table S8) were subjected to diagnostic PCR. The data for R. pedestris and L. chinensis were obtained from a previously published study [71]. Wild population inspection. To check the prevalence of the Burkholderia symbiont in wild populations of stink bugs, adult insect samples collected from fields in Japan (Table S8) were subjected to diagnostic PCR. The data for R. pedestris and L. chinensis were obtained from a previously published study [71]. (ii)Egg inspection. The possibility of vertical symbiont transmission via eggs, such as egg-smearing and transovarial manners, was investigated by collecting eggs from wild individuals and performing diagnostic PCR. The data for R. pedestris was obtained from a previously published study [17]. Egg inspection. The possibility of vertical symbiont transmission via eggs, such as egg-smearing and transovarial manners, was investigated by collecting eggs from wild individuals and performing diagnostic PCR. The data for R. pedestris was obtained from a previously published study [17]. (iii)Rearing of nymphs in clean Petri dishes. Hatchlings from infected parents were reared in a clean Petri dish until the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis) and subjected to diagnostic PCR. Different developmental stages were used for diagnostic PCR due to species-specific biological and technical constraints: M. uniguttatus and P. antennata aposymbiotic nymphs rarely survive beyond the 2nd instar; R. pedestris can be reliably diagnosed at the 3rd instar based on previous studies [25, 33]; and in C. punctiger, C. schmidti, and L. chinensis, 3rd-instar nymphs are relatively small and difficult to dissect reliably, making the 4th instar more suitable for consistent sampling. Rearing of nymphs in clean Petri dishes. Hatchlings from infected parents were reared in a clean Petri dish until the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis) and subjected to diagnostic PCR. Different developmental stages were used for diagnostic PCR due to species-specific biological and technical constraints: M. uniguttatus and P. antennata aposymbiotic nymphs rarely survive beyond the 2nd instar; R. pedestris can be reliably diagnosed at the 3rd instar based on previous studies [25, 33]; and in C. punctiger, C. schmidti, and L. chinensis, 3rd-instar nymphs are relatively small and difficult to dissect reliably, making the 4th instar more suitable for consistent sampling. (iv)Rearing of nymphs in clean Petri dishes with soil. The possibility of environmental symbiont transmission was investigated by rearing nymphs in a clean Petri dish with soil S32 collected from a field in Sapporo, Hokkaido, Japan (Table S5). The soil was suspended in DW and fed to hatchlings during their development, and the resulting insect samples were subjected to diagnostic PCR. Details are described in the “Infection test using soil samples” section of the “Materials and methods” section. Rearing of nymphs in clean Petri dishes with soil. The possibility of environmental symbiont transmission was investigated by rearing nymphs in a clean Petri dish with soil S32 collected from a field in Sapporo, Hokkaido, Japan (Table S5). The soil was suspended in DW and fed to hatchlings during their development, and the resulting insect samples were subjected to diagnostic PCR. Details are described in the “Infection test using soil samples” section of the “Materials and methods” section. (v)Rearing of nymphs in clean Petri dishes with infected parents. To check the possibility of coprophagy, hatchlings were reared in a clean Petri dish with their infected parents with GFP-labeled Burkholderia symbionts that were made using the Tn7 minitransposon system, as described previously [72]. After reaching the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis), these insects were subjected to fluorescence microscopy observation or diagnostic PCR. Rearing of nymphs in clean Petri dishes with infected parents. To check the possibility of coprophagy, hatchlings were reared in a clean Petri dish with their infected parents with GFP-labeled Burkholderia symbionts that were made using the Tn7 minitransposon system, as described previously [72]. After reaching the 2nd instar (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis), these insects were subjected to fluorescence microscopy observation or diagnostic PCR. (vi)Rearing of nymphs in plant pots. To test the possibility of symbiont transmission from plants, hatchlings of R. pedestris were reared in soil pots with soybean plants. The sieved soil S32 was transferred into plastic pots (10 × 8 cm; opening diameter × depth), and a soybean seedling was planted there. After 2 weeks of growing the soybean plant, 15 hatchlings were released into a pot covered with a transparent tube and a paper lid (Fig. S3A), fed with 5 dry soybeans, and maintained at 25 °C under a controlled light cycle of 16 h of light followed by 8 h of darkness. For comparison, to prevent insects from directly contacting the soil, they were also reared in a confined space where the above-ground parts of plants, previously sterilized with 70% ethanol and rinsed with sterilized DW, were covered with non-woven fabric (Fig. S3B). Each experimental pot was prepared with three replicates. After reaching the 3rd instar (10 days of rearing), the collected insects were subjected to diagnostic PCR. Rearing of nymphs in plant pots. To test the possibility of symbiont transmission from plants, hatchlings of R. pedestris were reared in soil pots with soybean plants. The sieved soil S32 was transferred into plastic pots (10 × 8 cm; opening diameter × depth), and a soybean seedling was planted there. After 2 weeks of growing the soybean plant, 15 hatchlings were released into a pot covered with a transparent tube and a paper lid (Fig. S3A), fed with 5 dry soybeans, and maintained at 25 °C under a controlled light cycle of 16 h of light followed by 8 h of darkness. For comparison, to prevent insects from directly contacting the soil, they were also reared in a confined space where the above-ground parts of plants, previously sterilized with 70% ethanol and rinsed with sterilized DW, were covered with non-woven fabric (Fig. S3B). Each experimental pot was prepared with three replicates. After reaching the 3rd instar (10 days of rearing), the collected insects were subjected to diagnostic PCR. Fitness measurement To assess the fitness effects of the Burkholderia symbiont in stink bugs, 1 st instar nymphs were prepared for both symbiont-infected (Sym) and uninfected (Apo) groups. The Sym group was inoculated with cultured symbionts at the 1 st to 3rd instar stages while the Apo group remained uninoculated. Nymphs were reared in clean Petri dishes, and the survival rate (adult emergence rate) was recorded for all six stink bug species. For more detailed fitness analyses, we focused on C. punctiger and C. schmidti. C. schmidti showed no significant difference in survival between Sym and Apo individuals, whereas C. punctiger was selected as a representative species showing relatively high survival in Apo insects. The other species showing similar survival patterns in Apo insects, R. pedestris and L. chinensis, were not included because fitness effects of the symbiont in R. pedestris have been well characterized in previous studies [17, 33, 72], and L. chinensis was difficult to maintain for multigenerational laboratory rearing. For C. punctiger and C. schmidti, the nymphal period (time to adulthood), body length, body weight, and forewing length were also investigated, as well as fecundity-related properties. Newly emerged adult females and males were kept separately for 2 weeks, and the size of their reproductive organs (one ovary and one testis per individual) was measured using ImageJ ver. 1.54 g [73]. Each pair of newly emerged female and male from the same group (Sym or Apo) and emergence day was reared together and checked for mating behavior twice daily for 2 weeks. Egg number was determined by counting all eggs laid during the 2 weeks following the first day of laying. Hatching rate was calculated by counting hatchlings born from the eggs laid during the 2 weeks after the first birth. Infection test using soil samples Each soil sample collected from the weedy and agricultural fields (Table S5) was mixed with DW (soil/water; 1/5, w/v). The resulting suspension was soaked into a cotton pad, which was then fed to 10–20 hatchlings in each Petri dish. After the insects reached the 2nd (for M. uniguttatus and P. antennata), 3rd (for R. pedestris), or 4th instar (for C. punctiger, C. schmidti, and L. chinensis), they were subjected to diagnostic PCR to detect Burkholderia infection. The pH-modified soils, as described below, were tested in the same manner. For all 42 field soils and the pH-manipulated soil inoculation experiments, we focused on C. punctiger and R. pedestris. These species were selected as representative and well-studied stink bug pests in rice and soybean agroecosystems, respectively, allowing us to examine soil pH effects on symbiont acquisition in ecologically and agriculturally relevant host species. The other stink bug species were tested using representative subsets of the field soils, as indicated in the corresponding figure (Fig. S5). Field survey A field survey of 29 weedy fields was conducted during the summer season, from July 30 to August 4, 2017, in three adjacent prefectures in Japan to minimize the effects of seasonal and geographical variation (Table S4). All surveyed sites were located in mesic agricultural landscapes, and soil EC ranged from 0.03 to 0.22 mS cm−1 (Table S4), well below levels typically associated with saline stress. Survey sites were selected where both southern crabgrass (D. ciliaris) and green foxtail (S. viridis) coexisted and were dominant, along with a mixture of other Poaceae grasses, as such fields serve as the primary breeding grounds for Coreoidea and Lygaeoidea rice bugs [54–56]. At each site, a 100-m2 plot was established using a measuring wheel (MURATEC-KDS Co., Kyoto, Japan). Surface soil samples (~ 1 cm depth) were collected from the four corners and the center of each plot, suspended in DW (soil/water; 1/5, w/v), and the soil pH was measured on-site using a pH meter (LAQUAtwin pH-22B, Horiba). Plots showing large variations in soil pH (standard deviation > 0.5) were excluded from the survey. Within each 100-m2 plot, insects were collected by sweeping 100 times using a 42-cm diameter silk net (Shiga Konchu Fukyusha, Tokyo, Japan), and the number of Coreoidea and Lygaeoidea insects was recorded. pH preference of the Burkholderia strains The symbiont strains for stink bugs and type strains of B. sensu lato were precultured overnight in YG medium with shaking at 180 rpm and 30 °C. After an additional 2 h of culture in fresh YG medium, the cultured cells were washed and suspended in DW. The cell suspension was adjusted to a final concentration of optical density at 600 nm (OD600) = 0.01 and incubated at 25 °C for 36 h in YG medium adjusted to each pH (ranging from pH 5.5–8.0 in increments of 0.5) using 50 mM potassium phosphate buffer. The OD600 after incubation was measured using a microplate reader (Infinite F200 PRO; Tecan Japan Co., Kanagawa, Japan). Soil pH modification experiment Soils S32 (pH 5.6) and S33 (pH 5.9) (Table S5) were used in this experiment. Each 15 g of the sieved soils was transferred into plastic tubes (4.5 × 10.4 cm; opening diameter × depth), and CaCO3 (Fujifilm Wako Pure Chemical Co., Osaka, Japan) was added to adjust the soil pH to approximately 7, 7.5, and 8. These tubes were covered with aluminum foil and incubated at 25 °C for 360 days, with water added twice a week to maintain the water content at 50% MWHC. Soil samples were collected at 7, 30, 60, 90, 120, 240, and 360 days after CaCO3 addition. These samples were used for DNA extraction to check the density of Burkholderia in soils and were also suspended in DW (soil/water; 1/5, w/v) for both pH measurements and infection experiments with C. punctiger and R. pedestris, as described above. For soils S32 and S33, triplicate tubes were prepared and analyzed individually. For the infection experiments, soil suspensions from the triplicates were pooled and fed to hatchling insects. Control systems were maintained at 50% MWHC without CaCO3 addition. Furthermore, soil pH modification experiments were conducted using agricultural materials derived from industrial waste, steel slag [74] and biochar [75], as alternative pH adjusters instead of CaCO3. Steel slag (The Sangyo Shinko Co., Tokyo, Japan) and chicken manure biochar (Sowa Recycle Co., Tokyo, Japan) were added to soil S32 to adjust the soil pH to approximately 7.5. Soil samples were collected at 7 and 30 days after incubation under the same conditions as the CaCO3-treated soils and were fed to hatchlings of C. punctiger for infection experiments. In vitro and in vivo observations of Burkholderia symbiont motility For in vitro observation, we used a previously established assay [29] with B. insecticola RPE64T (currently classified as Caballeronia insecticola), a symbiotic strain for R. pedestris, to assess flagellar motility. Strain RPE64T was cultured overnight at 30 °C in 1/4 strength YG medium, transferred to fresh medium, and incubated for an additional 5 h. Then, 50 µL of the culture was inoculated into 1 mL of fresh 1/4 strength YG medium adjusted to pH 5.5–8.0 (in 0.5 increments) using 50 mM potassium phosphate buffer. After 4 h of incubation, flagellar motility was observed using a phase-contrast microscope (IX73, Olympus) equipped with a × 20 objective lens (UCPLFLN20X, NA 0.70, Olympus), a CMOS camera (DMK 33UX174, Imaging Source), and an optical table (HAX-0806, JVI). Coverslip were glow-discharged by a hydrophilic treatment device (PIB-10, Vacuum Device), and sample chambers were assembled with coverslips and double-sided tapes. Cell suspensions were poured into the chamber, and both ends of the chambers were sealed with nail polish to prevent drying. Flagellar formation at pH 6.0 and 8.0 was also observed under a fluorescence microscope (IX73, Olympus) equipped with a × 100 objective lens (UPLSAPO 100 × OPH, NA1.4, Olympus), an optical filter set (Cy3-4040C, Semrock), a mercury lamp (U-HGLGPS, Olympus), and a CMOS camera (Zyla 4.2; Andor). Greyscale images were recorded and saved as uncompressed sequential TIF files. Fluorescent labeling followed a previous protocol [29]. Furthermore, cells incubated for 4 h in pH 6.0, 7.0, and 8.0 were transferred to fresh pH 6.0 medium, incubated for another 4 h, and re-evaluated for motility. For in vivo observation, we followed a protocol using R. pedestris and GFP-labeled symbionts as described previously [33]. The GFP-labeled strain RPE64T was cultured overnight in YG medium, transferred to fresh medium, and incubated for an additional 4 h. The resulting culture was mixed 1:1 with 100 mM potassium phosphate buffer (pH 5.5–8.0) and incubated at 25 °C for 1 h. This mixture was then provided to 2nd instar nymphs via cotton pads. Six hours after administration, the midgut was dissected, placed on a microscope slide, and covered with a coverslip. Samples were observed using a laser scanning confocal microscope (TCS SP8, Leica Microsystems). pH measurement of the insect gut Third instar nymphs in the Apo group of C. punctiger and R. pedestris were provided with either DW or each soil suspension derived from soils S32 (pH 5.6), S33 (pH5.9), S34 (pH 6.7), S51 (pH7.0), S35 (pH 7.5), S36 (pH 8.0), and S37 (pH 8.6) (Table S5) as drinking water, as described above. After 1 week of rearing, the midgut of the nymphs was dissected in a plastic Petri dish filled with DW and divided into distinct parts: M1, M2, M3, M4B, and M4. Each part, collected from five insects, was homogenized in 100 µL of DW by pipetting, and this procedure was performed in triplicate. The pH of the resulting homogenates was measured using a pH meter (LAQUAtwin pH-22B, Horiba). Statistics Categorical data comparing symbiotic and aposymbiotic insects, as well as flagellated and nonflagellated Burkholderia cells, were analyzed using Fisher’s exact test. Continuous data comparing symbiotic and aposymbiotic insects were first assessed for normality using the Shapiro–Wilk test and for variance homogeneity using the F-test. Based on these assumptions, statistical tests were selected as follows. Student’s t-test was used for normally distributed data with homogeneous variances, and Welch’s t-test for normally distributed data with unequal variances. When normality was violated, nonparametric tests were applied. The Mann–Whitney U test was used when variance heterogeneity was not detected, whereas the Brunner–Munzel test was applied when variances were heterogeneous, as this test does not assume equal variances or identical distributional shapes. To evaluate the outbreak of stink bugs, we conducted a sensitivity analysis using the following method. First, a generalized linear model (GLM) with a negative binomial distribution was constructed, using pH as the explanatory variable and the number of adult or nymph stink bugs as the dependent variable. Submodels were then created by sequentially removing one data point from the explanatory variable, generating as many submodels as there were data points. Finally, the deviance of all these submodels was plotted to identify, which data point removal significantly improved the goodness of fit. GLM with Gaussian distribution was used to analyze the effect of the soil’s chemical properties on total bacterial abundance. In this analysis, abundance as a responsive variable was log-transformed. We used GLM with a beta-binomial distribution to analyze the same effect on the relative abundance of Burkholderia in the soil microbiome and the infection rate for stink bugs. We used the likelihood ratio of the constructed model to the null model to obtain the P value (LR test). All parameters of the constructed model, the test of over-dispersion, and the coefficient of determination (R2) are summarized in Tables S6 and S7. We obtained the above values using the glmmTMB function in the glmmTMB package [76], the testDispersion function in the DHARMa package version 0.4.7 [77], and manual calculations. Because we conducted an analysis of the effect of all chemical properties on the relative abundance of Burkholderia in the soil microbiome, we calculated the Q value using the Storey procedure as a method of false discovery rate (FDR) adjustment [78]. Using the qvalue function in the qvalue package version 2.38.0 [79], we obtained the proportion of true null P values (π0 = 0.2597403) using the bootstrap method and then calculated the Q values. All statistical analyses were performed using R version 4.4.1. pH preferences for the growth of Burkholderia strains were analyzed using Dunnett’s test, comparing values at different pH levels to those at pH 7. To examine the relationship between soil pH and Burkholderia abundance, we estimated the absolute abundance of Burkholderia in soils by combining total bacterial density with the relative abundance of Burkholderia. Data from 42 soil samples were log10-transformed to stabilize variance. Linear regression models were fitted with log10-transformed Burkholderia abundance as the response variable and soil pH as the predictor. Quadratic models including a pH2 term were also evaluated. Model comparison using ANOVA indicated that inclusion of the quadratic term did not significantly improve model fit (F = 0.586, P = 0.449); therefore, the linear model was retained. Predicted Burkholderia abundance at pH 7 was estimated from the linear model and back-transformed to the original scale. Supplementary Information Additional file 1.Additional file 2.Additional file 3.Additional file 4.Additional file 5.Additional file 6. Additional file 1. Additional file 2. Additional file 3. Additional file 4. Additional file 5. Additional file 6.