Ingested soil bacteria breach gut epithelia and prime systemic immunity in an insect

Significance

Immune priming has evolved in memory-deficient invertebrates to combat sudden infections effectively. Although this phenomenon has been reported in diverse insects, understanding of underpinning mechanisms remains limited. We found that a member of the gut microbiota in the bean bug Riptortus pedestris can cross the epithelia and be released into the hemolymph, thereby stimulating systemic immunity and preventing subsequent infection of lethal pathogens. In this process, both humoral and cellular immunity play a pivotal role, and there was no trade-off with insect fitness. This finding not only highlights the utilization of environmental microbes by insects for shaping immunity but also provides a unique, symbiotic perspective on bacterial intestinal barrier breaching, which has been generally discussed only in terms of pathology.

Abstract

Insects lack acquired immunity and were thought to have no immune memory, but recent studies reported a phenomenon called immune priming, wherein sublethal dose of pathogens or nonpathogenic microbes stimulates immunity and prevents subsequential pathogen infection. Although the evidence for insect immune priming is accumulating, the underlying mechanisms are still unclear. The bean bug Riptortus pedestris acquires its gut microbiota from ambient soil and spatially structures them into a multispecies and variable community in the anterior midgut and a specific, monospecies Caballeronia symbiont population in the posterior region. We demonstrate that a particular Burkholderia strain colonizing the anterior midgut stimulates systemic immunity by penetrating gut epithelia and migrating into the hemolymph. The activated immunity, consisting of a humoral and a cellular response, had no negative effect on the host fitness, but on the contrary protected the insect from subsequent infection by pathogenic bacteria. Interruption of contact between the Burkholderia strain and epithelia of the gut weakened the host immunity back to preinfection levels and made the insects more vulnerable to microbial infection, demonstrating that persistent acquisition of environmental bacteria is important to maintain an efficient immunity.

To protect themselves from invasion by environmental pathogens, insects are armed with a well-developed innate immunity (1–4). Unlike vertebrates, insects lack acquired immunity and do not possess immune memory against pathogens. However, recent studies have reported that insects exhibit high defensive capabilities against pathogen infection through a phenomenon called immune priming (5–8). Immune priming is triggered when insects are exposed to sublethal doses of pathogenic microbes or their microbe-associated molecular patterns (MAMPs), such as peptidoglycan, eliciting the immune system and enabling immediate defense upon subsequent infection, even against lethal doses of pathogens. Moreover, immune priming can also be induced by nonpathogenic bacteria (9, 10). Experimental evidence confirmed that immune priming can be achieved through injection of microbes into the hemolymph or ingestion of a diet containing microorganisms (8, 11, 12). However, the mechanisms underlying immune priming in insects are not fully understood.

A major weapon to kill infecting microbes is the systemic humoral immunity, which is activated from the fat bodies (3). Fat bodies produce massive amounts of antimicrobial peptides (AMPs) upon microbial infection. AMPs are secreted in the hemolymph and kill and rapidly eliminate microbes from the insect’s body (3, 13, 14). The humoral immune cascade is generally activated by two different pathways: the Toll pathway and the immune deficiency (IMD) pathways, which are triggered by MAMPs and regulate the expression of the AMPs as well as other effectors (15, 16). In addition to humoral immunity, insects utilize a cellular immune system in the hemolymph mediated by phagocytic immune cells (17). These circulating hemocytes bind infecting microbes via MAMP-recognizing receptors or opsonins, which engage phagocytosis pathways. The activated hemocytes engulf infecting microbes and internalize them into phagosomes where they are lysed (17).

Insect guts are continuously exposed to environmental bacteria including pathogens, which are mostly ingested orally from their food and habitat. Orally ingested pathogens are confronted with a local gut immune response, which is dissimilar to the systemic immunity, wherein microbicidal reactive oxygen species (ROS) are generated and secreted in the gut lumen by the NADPH oxidase enzyme dual oxidase (Duox) that is highly activated by microbial infection in the midgut (18–20). When ROS-resistant pathogens colonize the midgut for a long period, AMPs are produced by the midgut epithelial cells via the IMD pathway (3, 21, 22), as a fail-safe system to eliminate the ROS-tolerant bacteria, thus maintaining gut homeostasis (21). In addition, most insects possess a physical barrier, the peritrophic membrane that lines the luminal side of the midgut epithelia and prevents direct contact by ingested bacteria with the midgut epithelial cells (2, 23). Despite these physical and immunological barriers, some enteropathogenic bacteria such as Serratia and Pseudomonas species can nevertheless penetrate the peritrophic membrane and gut epithelial cells, causing sepsis followed by killing of the insect (24, 25).

The bean bug Riptortus pedestris (Hemiptera: Alydidae) is a notorious pest of legumes and harbors a specific symbiont belonging to the genus Caballeronia, within the Burkholderia sensu lato group (Betaproteobacteria: Burkholderiaceae), in a crypt-bearing, posterior region of its midgut, called M4 (26, 27) (Fig. 1A). Caballeronia symbionts are not essential for the survival of the host insect but significantly enhance their growth and fecundity (28, 29). New-born R. pedestris nymphs are free of microbiota and acquire the Caballeronia gut symbiont together with a large diversity of other bacteria from the soil (28, 30, 31). Thus, the bean bug evolved stringent selection mechanisms including a symbiont sorting organ, the constricted region (CR) located in between the anterior midgut and the posterior symbiotic organ, to winnow out most bacteria derived from the soil (32). Caballeronia symbionts pass through the CR, but other bacteria are mostly filtered out and retained in the M3 midgut region which is located immediately before the CR (32) (Fig. 1A). However, the interactions between the M3-resident bacteria and the host bean bug are poorly understood.

Fig. 1.

Soil-infected and wild bean bugs have more complex gut microbiota than those solely infected with cultured Caballeronia symbiont. (A) Intestinal tract of R. pedestris, highlighting the six distinct midgut sections: M1 (first section), M2 (second section), M3 (third section), CR (constricted region), M4B (bulbous region preceding M4), and M4 (fourth section containing crypts, symbiotic organ). H, hindgut; MG, malpighian tubules. (B) M3 microbiota of Sym^control (n = 5), Sym^Soil (n = 5), and Sym^wild (n = 4 for both nymphs and adults) bean bugs. (C) M4 microbiota of Sym^control, Sym^Soil, and Sym^wild bean bugs. The bacterial community composition in (B) and (C) was determined via 16S rRNA amplicon sequencing.

In this study, we investigated the interactions in the midgut between soil-borne gut bacteria colonizing the M3 region and the bean bug host. We found that a Burkholderia sensu stricto (s.s.) species isolated from the M3 region primed the host systemic immunity by penetrating the gut epithelia and entering into the hemolymph, without any pathogenetic effects on the host insect. The activated immunity protected the host against subsequent pathogen infection, demonstrating that in addition to the M4 crypt-colonizing Caballeronia symbionts, some M3-resident bacteria are also beneficial to the bean bug host.

Results

Diverse Bacteria Colonize M3.

Bacterial communities in the M3 and M4 regions from bean bugs infected with cultivated Caballeronia insecticola symbiont (Sym^control), reared with environmental soil (Sym^soil), or captured in the field (Sym^wild) were analyzed by 16S rRNA gene amplicon sequencing. Bacterial diversity was higher in the M3 than in the M4 symbiotic organ, and Sym^soil and Sym^wild bean bugs possessed a more complex M3 bacterial community than Sym^control insects (SI Appendix, Fig. S1A). The M3 bacterial community of Sym^control bean bugs was relatively simple and low-abundant since the insects were not fed other bacteria besides the native gut symbiont C. insecticola (Fig. 1B and SI Appendix, Fig. S1 A and B). Most M3-colonizing bacteria in Sym^control bugs belonged to the genus Caballeronia (76.6 ± 8.5%), with a second major genus, Lysinibacillus (15.5 ± 4.3%), making up the rest of the community (Fig. 1B). The latter probably originated from the rearing conditions such as the soybeans and water they were fed. By comparison, the proportion of Caballeronia was much lower in the M3 of Sym^soil insects (14.9 ± 19.4%), whereas the genera Serratia (33.4 ± 19.7%), Enterococcus (26.7 ± 18.3%), and Burkholderia (20.3 ± 14.2%) were more abundant (Fig. 1B). The bacterial community of Sym^wild bean bugs was also more diverse than Sym^control insects, with only a small quantity of Caballeronia in the M3 region (6.2 ± 7.7%) (Fig. 1B). Instead, a high proportion of other bacteria, such as Enterococcus, Lactococcus, Citrobacter, Burkholderia, Enterobacter, Elizabethkingia, and Fibrella, were detected from the M3 of Sym^wild bean bugs (Fig. 1B).

In contrast to that of the M3 region, the bacterial community residing in the M4 symbiotic organ was not strongly different across Sym^control, Sym^soil, and Sym^wild bean bugs (Fig. 1C and SI Appendix, Fig. S1A). Both Sym^control and Sym^soil insects, which are maintained under laboratory conditions, predominantly harbored the Caballeronia gut symbiont in the symbiotic organ (Fig. 1C). Sym^wild bugs also possessed a high proportion of Caballeronia symbionts in the M4, but some individuals had in addition a relatively large amount of other bacteria, belonging to genera such as Cutibacterium, Citrobacter, Lactococcus, Enterococcus, and Lysinibacillus (Fig. 1C). The presence of bacteria other than Caballeronia spp. in the M4 could be due to the reopening of the symbiont sorting organ. Indeed, the CR is tightly sealed immediately after colonization of the M4 by the Caballeronia symbiont (33) but reopens at the late developmental stage of the bean bug due to aging (34). However, it should be noted that the core bacterial taxon in the M4 of these Sym^wild insects is still Caballeronia (Fig. 1C).

Soil-Derived Gut Bacteria Enhance Immunity and Protect the Bean Bug against the Pathogen Pseudomonas entomophila.

To investigate the effects of M3-colonizing gut bacteria acquired from the soil on the host bean bug, the survival rate, developmental period, and fitness parameters (body length and dry body weight) of Sym^control and Sym^soil insects were measured (SI Appendix, Fig. S2A). Sym^soil bean bugs showed similar survival rate, growth rate, and body size to Sym^control insects (Fig. 2 A–F), indicating that orally acquired commensal gut microbes did not show harmful effects on the host bean bug. Gut-colonizing bacteria in insects are known to activate local gut immune responses, such as the production of AMPs (14, 35, 36). The enhanced expression level of three known AMPs (riptocin, defensin, and thanatin) (37) in the M3 of Sym^soil bean bugs compared to Sym^control demonstrated that ingested M3-colonizing bacteria similarly induces the local immune response in R. pedestris (Fig. 2G). Interestingly, the expression levels of the AMPs were also up-regulated in the M4 symbiotic organ as well as in the fat bodies although M3-colonizing bacteria generally do not colonize those organs (Fig. 2 H and I), indicating that the orally acquired soil-derived bacteria stimulated not only the local gut immunity but also a systemic immune response in the bean bugs.

Fig. 2.

Soil-borne microbes do not impair bean bug physiology. (A) Host survival rate and (B) time to adulthood of Sym^control and Sym^soil bean bugs (n = 51 and 52, respectively). (C and D) Body length (C) and dry body weight (D) of adult male bean bugs (n = 20). (E and F) Body length (E) and dry body weight (F) of adult female bean bugs (n = 20). (G–I) Expression levels of three AMPs (riptocin, defensin, and thanatin) in M3 (G), M4 (H), and fat body (I) after oral infection with cultured Caballeronia or environmental soil (n = 5, respectively). (J) Survival rate of Sym^control and Sym^soil bean bugs after injection of entomopathogen P. entomophila into the hemocoel (n = 20). The statistical significances were analyzed by the Mann–Whitney U test with Bonferroni correction (A–I) or log-rank test (J). Error bars indicate SDs; ns, not statistically significant; **P < 0.005; ***P < 0.0005; ****P < 0.0001.

To determine whether the systemic immune activation by M3-colonizing gut bacteria protects the bean bugs against pathogen infection, we first tested the pathogenicity of gram-negative and gram-positive bacteria by injecting 10⁶ cells of four bacterial species into the hemolymph of Sym^control insects. This infection route mimics natural infections occurring by wounding or by natural parasites. The entomopathogenic gram-negative bacterium Pseudomonas entomophila exhibited high pathogenicity, but the bean bugs showed resilience against gram-positive bacteria (Staphylococcus aureus and Micrococcus luteus) and nonpathogenic gram-negative bacteria Escherichia coli (SI Appendix, Fig. S3). Since a previous study demonstrated the insecticidal activity of P. entomophila against the bean bug (38), we used P. entomophila as the pathogen in further experiments. A total of 10⁶ cells of P. entomophila were injected into the hemolymph of Sym^control and Sym^soil insects, and the insects’ survival rates were measured. In agreement with previous results showing that P. entomophila is pathogenic for bean bugs (SI Appendix, Fig. S3), Sym^control insects died within 72 h after bacterial injection (Fig. 2J). In contrast, Sym^soil bean bugs showed high survival after P. entomophila systemic infection (65% survivability) (Fig. 2J), suggesting that the systemic immunity up-regulated by M3-colonizing bacteria protects the host bean bugs against an infecting pathogen.

Isolation of M3-Colonizing Bacteria.

To identify bacterial taxa that can enhance host systemic immunity, we isolated M3-colonizing bacteria from the M3 of Sym^soil bean bugs by spreading homogenized M3 lysate on nutrient broth agar plates (SI Appendix, Fig. S1A). A total of four different bacterial species were identified by 16S rRNA sequencing, three of them belonged to the genus Burkholderia s.s. and one to the genus Achromobacter (SI Appendix, Table S1). One of the Burkholderia (Burkholderia sp. HK1; SI Appendix, Fig. S4) and Achromobacter (Achromobacter sp. HK2) isolates were orally administered to aposymbiotic (free of Caballeronia M4 symbionts) second-instar bean bug nymphs, and bacterial titers were measured from each midgut section. Both, Burkholderia HK1 and Achromobacter HK2, were detected in the M1, M2, and M3 midgut regions, and their number increased over time in the M3 (Fig. 3 A and B). Despite their presence in M3, neither bacterial strains were able to colonize the M4B and M4 regions, in concordance with the selective colonization of the symbiotic organ by Caballeronia (29). In addition, fluorescence microscopy showed that red fluorescence protein (RFP)-expressing Burkholderia HK1 and Achromobacter HK2 cells did not pass the CR and were stuck in the M3, confirming that both isolated M3-colonizing bacteria are not M4 crypt symbionts (SI Appendix, Fig. S5).

Fig. 3.

M3-colonizing bacteria elicit systemic immunity, protecting the host from pathogen invasion. (A and B) The number of Burkholderia HK1 (A) and Achromobacter HK2 (B) cells colonizing the gut of aposymbiotic second-instar nymphs was measured in each midgut region at 12, 24, and 48 h after oral administration (n = 3). nd, not detected. (C–E) Relative expression levels of three AMPs (riptocin, defensin, and thanatin) in M3 (C), M4 (D), and fat body (E) in the Burkholderia HK1 or Achromobacter HK2 treatments relative to the control treatment (n = 5, respectively). Burkholderia HK1 elicited the induction of systemic AMPs, while Achromobacter HK2 triggered an immune response only in M3. (F) Total number of hemocytes in the hemolymph of bean bugs after oral infection with either Burkholderia HK1 or Achromobacter HK2 cells (n = 5, respectively). (G) Survival rate of Sym^control, Sym^Bu, and Sym^Ab bean bugs after systemic infection with the pathogen P. entomophila (n = 20, respectively). Different letters indicate statistically significant differences (P < 0.05). The statistical significance was analyzed by the Mann–Whitney U test with Bonferroni correction (A–F) or log-rank test with Bonferroni correction (G). Error bars indicate SDs; ns, not statistically significant.

Burkholderia HK1 Protects the Bean Bug from Pathogens by Priming Host Systemic Immunity.

Since infection with soil microbes activated the immune responses of the bean bugs compared to Sym^control insects (Fig. 2 G–I), we measured the expression levels of AMPs in the M3, M4, and fat body of Sym^control insects colonized by C. insecticola, symbiotic insects colonized by C. insecticola and infected with Burkholderia HK1 (Sym^Bu), or with Achromobacter HK2 (Sym^Ab) to determine whether isolated commensal gut bacteria indeed enhance host immunity. Both Burkholderia HK1 and Achromobacter HK2 stimulated local gut immunity in the M3 they inhabit, although the expression level of thanatin was only up-regulated in Sym^Bu and not in Sym^Ab insects (Fig. 3C). This could be attributed to the inherently low expression of thanatin in the M3 even during bacterial systemic infection (39, 40). Similar to Sym^soil bean bugs, infection with Burkholderia HK1 but not Achromobacter HK2 increased the expression of AMPs in the M4 and the fat body, indicating that commensal Burkholderia HK1 can enhance systemic host immunity after oral ingestion (Fig. 3 D and E). Since Burkholderia HK1 activated systemic immune responses, the cellular immunity was subsequently investigated by measuring the number of hemocytes in the hemolymph. The total number of hemocytes in Sym^Bu insects but not in Sym^Ab was higher than that of Sym^control bean bugs, suggesting that Burkholderia HK1 also activates cellular immune responses (Fig. 3F). The enhanced systemic immunity of Sym^Bu bean bugs was correlated with a protective effect on the host against infection with P. entomophila, while Sym^Ab bean bugs succumbed to the pathogen infection within 72 h similarly to Sym^control bean bugs (Fig. 3G). These results demonstrate that certain M3-resident bacteria such as Burkholderia HK, but not all bacterial species, can prime the host immunity, conferring protection against infection by pathogens.

To distinguish between the possibilities that protection arises from enhanced resistance or alternatively from acquired tolerance to the pathogen, the number of P. entomophila cells in the hemolymph was monitored after their injection in the hemolymph of Sym^Bu and Sym^control insects. This experiment showed that the pathogen was significantly decreased in Sym^Bu insects and dropped more than 100-fold 12 h after P. entomophila injection or more than 1,000-fold 24 h after injection, contrary to Sym^control insects where the pathogen stayed at a high level in the hemolymph (SI Appendix, Fig. S6). This observation suggests that colonization of the M3 by Burkholderia HK1 confers resistance against pathogen infection.

Immune Priming by M3-Colonizing Burkholderia Has No Obvious Negative Effects on the Bean Bug Host.

We next checked the effects of isolated Burkholderia HK1 and Achromobacter HK2 on the bean bug fitness by measuring the insect survival rate, growth period, and biomass parameters. Burkholderia or Achromobacter cells were orally administered following infection with the gut symbiont C. insecticola in order to compare all parameters with Sym^control bean bugs (SI Appendix, Fig. S2B). Both Burkholderia HK1 and Achromobacter HK2 showed no negative effects on the survival rate, development, and fitness parameters of the bean bugs compared with Sym^control insects (Fig. 4 A–F). Additionally, the population of C. insecticola symbionts in the symbiotic organ was similar between Sym^control, Sym^Bu, and Sym^Ab insects, demonstrating that the infection by both M3-colonizing bacteria did not affect the M4 symbiotic association in the bean bug (Fig. 4 G and H). Taken together, the gut commensal bacteria derived from the soil exhibit no negative impact on the host bean bug. In many insects, there is a physiological trade-off between immune activation and reproduction; an increased immunity due to the presence of bacteria leads to a decrease in fecundity (41, 42). We thus investigated the fecundity of Sym^control and Sym^Bu bean bugs by determining the time at which they start laying eggs and counting the number of eggs produced over a week. In each replicate, two males and one female bean bug were reared together to promote mating. The time before egg laying was not significantly different between Sym^control and Sym^Bu bean bugs (6 to 7 d after molting to the adult stage), and the number of laid eggs was unchanged in Sym^Bu insects when compared to Sym^control insects (Fig. 4 I and J). These results thus demonstrate that immune priming did not affect fecundity in R. pedestris and are in accordance with prior studies in beetles (43, 44).

Fig. 4.

M3-colonizing symbionts do not affect host physiology and M4 gut symbiosis. (A) Survival rate and (B) growth period to adulthood of Sym^control, Sym^Bu, and Sym^Ab bean bugs (n = 26, 25, and 24, respectively). (C and D) Body length (C) and dry body weight (D) of adult male bean bugs (n = 20, respectively). (E and F) Body weight (E) and dry body weight (F) of female bean bugs (n = 20, respectively). (G) Copy number of the dnaA gene of Caballeronia symbiont in the M4 region (n = 5, respectively), as measured by RT-PCR. (H) Fluorescence microscopic images of isolated M4. Green indicates the symbiont-derived GFP signal. (I) Number of eggs laid since the beginning of oviposition (n = 18, respectively). (J) Total number of eggs during a week of oviposition (n = 18, respectively). The statistical significance was analyzed by the Mann–Whitney U test with Bonferroni correction. Error bars indicate SDs; ns, not statistically significant.

Burkholderia HK1 Enhances Systemic Immunity by Breaking through Gut Epithelia and Migrating into the Hemolymph.

To reveal how gut-colonizing Burkholderia HK1 stimulates host systemic immune responses, the bacterial titer in the hemolymph was measured after oral administration of Burkholderia HK1. Two microliters of the hemolymph was collected from insects by removing a leg or antenna and collecting the pooled hemolymph (SI Appendix, Fig. S7). A few colonies of Burkholderia HK1 were detected in the hemolymph 12 h after oral infection and the population increased over time until 48 h (Fig. 5A). The migration of Burkholderia HK1 cells to the hemolymph was confirmed through fluorescence microscopy using the RFP-expressing Burkholderia HK1 (Fig. 5A). On the other hand, Achromobacter HK2, which stimulates only local gut immunity (Fig. 3C), was not detected from the host hemolymph, indicating that migration to the hemolymph from the M3 is required to enhance systemic immunity (Fig. 5A). Confocal microscopy confirmed the penetration of the M3 epithelia by Burkholderia HK1. Twelve hours after oral infection by Burkholderia HK1, bacterial cells entered the M3 region and attached to the gut epithelial cells (SI Appendix, Fig. S8). After 24 h, Burkholderia HK1 was detected in between the M3 gut epithelial cells (Fig. 5B and SI Appendix, Fig. S8). Transmission electron microscope (TEM) images also showed that Burkholderia HK1 cells penetrated the M3 gut epithelia while there was no detectable bacterial penetration in Sym^control bean bugs (Fig. 5 C and D). The C. insecticola gut symbiont, which is temporarily present in the M3 before entering the symbiotic organ, only colonized the luminal side of the M3 (Fig. 5C).

Fig. 5.

Burkholderia HK1 breaches the gut epithelium, migrates into the hemolymph, and stimulates host systemic immunity. (A) Bacterial titers in the hemolymph after oral infection of Burkholderia HK1 or Achromobacter HK2 cells (n = 5, respectively). The Inset shows the RFP signal derived from dsRed-expressing Burkholderia HK1 cells by direct fluorescence microscopy observation of a droplet of hemolymph. (B) Confocal microscopy image of gut-breaching Burkholderia HK1. A few HK1 cells are located between M3 epithelial cells. Red, RFP-labeled Burkholderia HK1 cells; blue, host nuclei; White, host cytoskeleton. (C and D) TEM images of M3 in Sym^control and Sym^Bu bean bugs. In the gut lumen of the Sym^control insect, numerous C. insecticola symbiont cells are present (C). Burkholderia HK1 cells are situated in the M3 epithelial layer, migrating toward the hemolymph side (D). Nu; nuclei of the host gut epithelium, Lu; luminal side of the midgut. (E) Pathogenicity of Burkholderia HK1. A total of 10³ or 10⁶ cells of Burkholderia HK1 were injected into the hemocoel, and the survival rate of bean bugs was measured (n = 15, respectively). (F) Bright field, fluorescence microscopy, and merged images of phagocytosed Burkholderia HK1 cells by the host hemocytes. In the fluorescence image, red represents RFP-labeled Burkholderia HK1 cells, and blue indicates nuclei of hemocytes. (G and H) TEM images of in vitro cultured (G) and in vivo M3-colonizing (H) Burkholderia HK1 cells. White granules in bacterial cells represent polyhydroxyalkanoate (PHA) granules. Yellow arrows point to lysed cells in the midgut. (I) Pathogenicity of in vitro cultured and in vivo M3-colonizing Burkholderia HK1 (n = 15, respectively). The survival rate of bean bugs was measured after injecting each type of bacterial cells into the host hemocoel. (J) Survival rate of Sym^control, Sym^Bu, and immune-suppressed Sym^Bu insects after P. entomophila injection (n = 15, respectively). Relish is the key gene for bean bug’s humoral immunity. Sym^Bu/relish, relish-silenced Sym^Bu insects by RNAi; Sym^Bu/bead, Sym^Bu insects with suppressed cellular immunity; Sym^{Bu/relish/bead}, both cellular and humoral immunity-suppressed Sym^Bu insects. Different letters of figure a indicate statistically significant differences (P < 0.05). The statistical significance was analyzed by the Mann–Whitney U test with Bonferroni correction (A), log-rank test (E and I), or log-rank test with Bonferroni correction (J). Error bars indicate SDs; nd, not detected.

Since bacterial presence in the hemolymph could kill the host bean bug (38, 45), we then tested the pathogenicity of Burkholderia HK1 by injecting bacterial cells directly into the hemocoel. A high dose of bacteria led to high mortality in bean bugs, but a low number of Burkholderia HK1 cells were not lethal, suggesting that a small number of Burkholderia HK1 cells migrating from the M3 into the hemolymph will not kill the host (Fig. 5E). Indeed, host bean bugs rapidly eliminated the migrated Burkholderia HK1 cells in the hemolymph by phagocytosis (Fig. 5F). Moreover, comparison of the morphology of Burkholderia HK1 cells between cultured and M3-colonizing cells showed that the cell membrane of in vivo colonizing Burkholderia HK1 is abnormal (Fig. 5 G and H). This membrane alteration could fragilize the bacteria and we observed indeed that the gut-colonizing Burkholderia HK1 cells were more susceptible to various stresses compared to in vitro cultured Burkholderia HK1 cells (SI Appendix, Fig. S9). This weakening of the bacteria in the M3 could facilitate the rapid elimination of Burkholderia HK1 cells from the hemolymph by the enhanced humoral and cellular immunity. Accordingly, the pathogenicity of gut-colonizing Burkholderia HK1 cells to the bean bugs was attenuated compared to normally cultured HK1 cells (Fig. 5I).

Next, we determined whether bacterial breaching of the gut epithelia is also taking place, and by which bacterial species, in the immune-primed insects reared on soil. These insects carry a more diverse microbiota than gnotobiotic insects inoculated with specific microbes (Fig. 1B). Hemolymph was extracted from third instar Sym^soil nymphs and plated on agar medium. Fifteen colonies were randomly selected, and the bacterial species were identified using 16S rRNA gene sequencing. Different bacteria including Cupriavidus and Paraburkholderia spp. were detected in addition to a majority of Burkholderia s.s. spp., indicating that taxonomically diverse bacteria and not only Burkholderia s.s. can breach the gut epithelial wall (SI Appendix, Table S2). Nevertheless, it is striking that all the identified species are closely related and belong to the Burkholderiaceae family. Since the taxonomic diversity in the M3 microbiota of Sym^soil insects includes members of several other bacterial families (Fig. 1B), our data suggest that the capacity to breach the gut epithelia is mostly restricted to species of the Burkholderiaceae. Thus, the immune priming observed in Sym^soil insects (Fig. 2 G–J) is probably triggered by a similar mechanism as in Sym^Bu insects, involving infection of the hemolymph by gut commensal bacteria belonging to the Burkholderiaceae.

Systemic Immunity Primed by M3-Colonizing Burkholderia Plays a Pivotal Defense Role.

To investigate whether induced host systemic immunity is a main factor protecting host bean bugs against pathogenic infection, we suppressed host humoral and cellular immunity by RNAi of the relish gene to suppress humoral immunity or preinjecting latex beads to induce saturation of phagocytosis of the hemocytes in Sym^Bu bean bugs (SI Appendix, Fig. S2C). Then, P. entomophila was injected into the bean bug hemocoel, and the insect survival rates were measured (Fig. 5J). Cellular immunity–repressed insects (Sym^Bu/bead) succumbed to the pathogen faster than the Sym^Bu insects, at rates similar to Sym^control insects, but humoral immunity-repressed insects (Sym^Bu/relish) died even more rapidly. Accordingly, bean bugs lacking both cellular and humoral immunities (Sym^{Bu/relish/bead}) showed the fastest mortality. Together, these observations demonstrate that both cellular and host humoral immune priming by gut-colonizing Burkholderia HK1 is critical to protect host bean bugs against pathogen infection, although the contribution of humoral immunity seems to be higher than the cellular immunity in our experimental conditions.

Persistent Colonization of Burkholderia HK1 in the M3 Is Important for Continuous Immune Enhancement.

Since M3-colonizing Burkholderia HK1 cells are fragilized, the enhanced immunity in Sym^Bu insects could rapidly eliminate the bacteria from the midgut. To test this, we divided bean bugs into two groups: One group was continuously supplied with Burkholderia HK1 (Sym^Bu), and the other group was provided Burkholderia HK1 only once, after which the bacterial suspension was changed to sterile drinking water (DWA) (Sym^Bu>DWA) (SI Appendix, Fig. S2D). In the M3 of bean bugs continuously infected with Burkholderia HK1, a high number of bacterial cell was continuously detected until the fifth instar (Fig. 6A). However, the bacterial titer gradually decreased from the third until the fifth instar nymph when the supply of Burkholderia HK1 was halted (Fig. 6B). Additionally, the up-regulated local gut immunity as well as the systemic immunity returned to a lower level in Sym^Bu>DWA bean bugs, reaching similar levels as those of Sym^control insects (Fig. 6 C–E). Consequently, Sym^Bu>DWA bean bugs, which lost their up-regulated immunity, also lost the protective effect against infection by pathogen P. entomophila (Fig. 6F). Taken together, these results indicate that continuous oral ingestion of Burkholderia HK1 is required to maintain enhanced systemic immunity in the bean bug R. pedestris and is important to protect the host against pathogens.

Fig. 6.

Spontaneous ingestion of commensal bacteria is important for immunopotentiation. (A and B) Bacterial titers in the M3 region of Sym^Bu (A) and Sym^Bu>DWA (B) bean bugs over time. Sym^Bu>DWA; HK1 cells were supplied only once to bean bugs and then changed to DWA (n = 3, respectively). (C–E) The expression levels of three AMPs (riptocin, defensin, and thanatin) in M3 (C), M4 (D), and fat body (E) of Sym^control, Sym^Bu, and Sym^Bu>DWA insects (n = 5, respectively). (F) Survival rate of Sym^control, Sym^Bu, and Sym^Bu>DWA insects after pathogen P. entomophila septic infection (n = 15, respectively). Different letters indicate statically significant differences (P < 0.05). The statistical significance was analyzed by the Mann–Whitney U test with Bonferroni correction (A–E) or log-rank test with Bonferroni correction (F). Error bars indicate SDs.

Discussion

Immune priming is thought of as a mechanism analogous to vertebrate immune memory and plays a crucial role in insects, which only possess an innate immunity, to combat lethal pathogenic infections (6, 7). However, the mechanisms underlying immune priming in natural environments remain poorly understood. In this study, we found that in R. pedestris, the anterior midgut M3-colonizing bacterium Burkholderia HK1, acquired from soil, penetrates the gut epithelia, subsequently migrates to the hemocoel and triggers systemic immunity, which ultimately confers resistance against a lethal dose of pathogens. We previously revealed that the M4-colonizing Caballeronia symbionts strongly contribute to the activation of the host immunity (45). Thus, in addition to the M4 symbionts, also M3-colonizing bacteria boost immunity in an unexpected way, by penetrating the intestinal wall. Cells of the M3-colonizing Burkholderia HK1 exhibited lower stress tolerance compared to cells cultured in vitro in nutrient-rich media (SI Appendix, Fig. S9), suggesting their inactive or less-active state in the M3. Indeed, although a few Burkholderia HK1 cells were detected in the hemolymph of infected insects (Fig. 5A), no pathogenic properties were observed. Notably, immune priming induced by Burkholderia HK1 had no discernible adverse effects on host fitness (Fig. 4). Previous studies have reported a trade-off between immune activation and overall fitness (42, 46, 47). However, the immune priming elicited by commensal gut bacteria in nature, as observed in this study, could provide significant immune benefits with minimal costs, and therefore, Burkholderia HK1 as well as other M3-colonizing and hemolymph-infecting Burkholderiaceae can be considered as symbionts.

The midgut of insects is constantly exposed to a large number of ingested gut microbes, thus, insects have to be armed with a well-developed immunity to maintain gut homeostasis. In Drosophila, many orally infected bacteria, including food-borne pathogens, are eliminated from the midgut by local immunity. ROS, generated by the Duox enzyme upon microbial infection in the midgut, results in the elimination of invading pathogens (18, 19, 48). Genetic blocking of the duox gene actually increases the load of midgut-colonizing bacteria and duox-deficient flies succumb to oral microbial infection (18). In contrast to Drosophila, the Duox is not important in the midgut mucosal immunity of the bean bug. RNAi of duox neither altered the survival rate of the bean bugs upon pathogen infection nor increased bacterial population in the M3 (49). Additionally, the production of AMPs by bacterial infection in gut epithelia of Drosophila complementarily kills ROS-resistant pathogens (21). Besides these immune reactions, Drosophila possesses the acellular chitinous and proteinaceous peritrophic membrane that lines the midgut epithelium. This matrix physically protects midgut cells from direct contact with gut microbes thereby avoiding immune activation (50). It should be noted that while peritrophic membranes are found in most of the insect groups, they are completely absent in many hemipteran insects (51), including the bean bug R. pedestris (49). The M3-colonizing bacteria are thus potentially in more direct contact with M3 epithelia of the host compared to other insects and can more readily activate gut immunity that could immediately affect gut microbes. Indeed, the M3 symbiont Burkholderia HK1 or Achromobacter HK2, highly induced locally the expression of AMPs in the M3 region (Fig. 3).

Even if most gut-dwelling microbes in Drosophila are confined to the gut lumen due to the peritrophic membrane and thus come not in contact with gut epithelial cells, some pathogens (e.g., Serratia marcescens and P. entomophila) nevertheless can cross the peritrophic and epithelial barriers by secreting proteases (24, 25). However, S. marcescens, which had traversed the Drosophila midgut barriers and migrated to the hemolymph, did not induce systemic immunity since most bacterial cells were directly engulfed by hemocytes (24). In contrast, the M3 symbiont Burkholderia HK1 that migrated to the hemolymph in R. pedestris, activated a systemic immune response of the fat body despite the fact that these Burkholderia HK1 cells were also phagocytosed by hemocytes (Fig. 5F). Then, what accounts for this difference in the immune response against gut-breaching bacteria between Drosophila and R. pedestris? One possibility is variations in the velocity and efficiency of phagocytosis by hemocytes between the insect species. Additionally, it is plausible that Burkholderia HK1 possess some mechanisms enabling their survival within the hemocoel, evading complete elimination through phagocytosis and then priming systemic immunity of the bean bug. Another possibility is that Burkholderia HK1 penetrates the gut epithelia with greater frequency and quantity, thereby consistently priming the fat bodies. Indeed, since R. pedestris lacks a peritrophic membrane, gut-colonizing bacteria could more easily penetrate the intestinal epithelium than in other insects with a peritrophic membrane. Alternatively, the difference could be caused by different immune regulations between holometabolous and hemimetabolous insects. The immunity of hemimetabolous insects has been poorly studied, but it is clear that it is distinct from that of holometabolous insects (52, 53). In Drosophila, gram-negative and gram-positive bacteria are recognized by the IMD and Toll pathways, respectively (3, 54). However, many hemimetabolous insects lack key components of the IMD pathway, such as Fadd, Relish, and Imd (52, 53). A recent study of a hemimetabolous insect, the brown-winged green stinkbug Plautia stali (Hemiptera: Pentatomidae), revealed that the IMD and Toll pathways are functionally intertwined to possibly create an immune pathway with new functionalities (55). In this stinkbug species, gram-negative and gram-positive bacteria are recognized by both the IMD and Toll pathway. Therefore, the differently wired immune pathway in the bean bug could account for the systemic immunity activated by phagocytosed Burkholderia HK1 cells. Despite these differences, it is important to note that immune priming is also frequently observed in insects that possess a peritrophic membrane. In these insects, immune priming may not result from bacteria directly entering the hemolymph but could instead be initiated by other immune-stimulating molecules, such as peptidoglycan (56). However, the molecular mechanism of how these molecules can prime insect immunity should be elucidated in future studies.

Wild populations of the bean bug necessarily acquire a large number of nonsymbiotic bacteria from the environment when the insect obtains its Caballeronia M4 symbiont from the soil. Indeed, the proportion of bacteria belonging to the genus Caballeronia in field-collected soil was less than 0.1% (57). In other words, more than 99.9% of soil-living bacteria are not M4 symbionts. Despite the high proportion in the soil of bacteria that are not M4 symbionts, R. pedestris winnows out bacteria through the highly efficient CR that functions as a symbiont sorting organ (32). Since most bacteria can’t pass the CR, they are stuck in the M3 region located anteriorly to the CR, making up the microbiome of the M3. Insect gut microbiota generally have diverse beneficial effects on the host such as assisting food digestion, providing nutrients, detoxification, enhancing immunity, and many others (58–63). The composition of insect gut microbiota is immensely affected by conditions such as host habitats and food source (64). The M3-microbiota of R. pedestris is also different across Sym^control, Sym^soil, and Sym^wild insects (Fig. 1B). Although the composition of bacteria in the M3 at the genus level differed between the three groups of bean bugs, in each case a few specific genera seemed to dominate the population. Even though the specific microbiota of the soil in this study has not been characterized, the M3-microbiota associated with the bean bug, irrespective of the source of infection, is relatively simple when compared to the known complexity and diversity of general soil microbiota (65) (Fig. 1B), because of multiple selection layers inside the gut, including nutrients, pH, oxygen level, and antimicrobial peptides (49, 66–68). Another notable point is that, in Sym^soil bean bugs, diverse bacteria including Burkholderia s.s., Paraburkholderia, and Cupriavidus were isolated from hemolymph in soil-reared nymphs (SI Appendix, Table S2), strongly suggesting that these bacteria breach gut epithelia and thereby prime immunity. Although further inspections are still needed, diverse bacterial species may play a pivotal role in immunity in field bean bug populations as well. Bean bugs are constantly exposed to bacteria that could potentially prime their immunity throughout their lives. For example, bean bugs routinely interact with environments teeming with microbes, including soil and plant surfaces, and they consume microbial-contaminated foods.

Insect guts are generally colonized with more or less complex microbiota, which provide numerous and diverse services to their host, such as synthesizing nutritional supplements, facilitating the degradation of food, detoxifying environmental toxins, and providing colonization resistance (69). Although the penetration of gut epithelia has been reported for some enteropathogenic bacteria (24, 25), our study provides an example of gut-breaching by a member of the soil-derived gut microbiota, whereby the bacterial species primes host immunity without obvious harmful effects to the host insect. This type of gut symbiotic bacteria has not yet been well studied, and further molecular, immunological, and histological studies remain needed in the future. Our study not only highlights how insects employ soil bacteria to cope with pathogens but also provides a unique, symbiotic perspective on bacterial intestinal barrier breaching, which has been generally discussed only in terms of pathology.

Materials and Methods

Insect Rearing and Symbiont Infection.

The bean bug R. pedestris was originally collected from a soybean field of Tsukuba, Ibaraki, Japan, and reared in laboratory conditions for over 10 y. The bean bugs were maintained under a long-day regimen (16 h light and 8 h dark) at 25 °C in Petri dishes (90 mm diameter and 20 mm height). They were supplied soybean seeds and distilled water containing 0.05% ascorbic acid (DWA) which were periodically renewed. The bean bugs were infected with gut symbiont C. insecticola through diluted bacterial solution in DWA (10⁷ cells/mL) at the second-instar nymphal stage. To produce soil-infected Sym^soil insects, second-instar nymphs of the bean bug were infected by a suspension of soil collected from Sapporo, Japan. A total of 1 g of soil was placed on the cotton pads, which was moistened by adding distilled water (DWA) to the soil-covered pads.

Measurement of Host Survival Rate, Development, and Fitness Parameters.

The survival rate of insects was assessed by daily counts of carcasses. The growth rate was determined by the time taken to reach adulthood. Fitness parameters, specifically dry body length and weight, were measured in adult insects. Detailed procedures for the measurement of these fitness parameters as well as for determining the insect survival rate following bacterial injection can be found in SI Appendix, Materials and Methods.

Construction of RFP-Labeled Bacteria.

Burkholderia HK1 and Achromobacter HK2 were genetically engineered to express RFP by introducing the pIN29 plasmid, which carries the dsRed gene. The construction process is detailed in SI Appendix, Materials and Methods.

16S rRNA Amplicon Sequencing.

The gut microbiota of the M3 and M4 regions of Sym^control, Sym^soil, and Sym^wild bean bugs were analyzed through 16S rRNA sequencing, specifically focusing on the variable region 4 (V4), using an Illumina iSeq 100 sequencer. Detailed information for sequencing and microbiota analyses are available in SI Appendix, Materials and Methods.

CFU Assays from the Host Midgut and Hemolymph.

The quantity of M3-colonizing Burkholderia HK1 or Achromobacter HK2 cells was determined by spreading the M3 lysate of insects infected with RFP-expressing bacteria onto yeast-glucose (YG) agar plates supplemented with 15 μg/mL of chloramphenicol and counting the colonies after 2 d incubation at 30 °C. To measure the number of Burkholderia HK1 and Achromobacter HK2 cells in the hemolymph, 10⁷ cells/mL of RFP-expressing bacterial cells were orally administered to the fifth instar nymph of Sym^control bean bugs. Following 12 h, 24 h, or 48 h of oral infection, 2 μL of the hemolymph was extracted by cutting the antenna and forelegs and collecting leaking hemolymph with a micropipette. The obtained hemolymph was immediately mixed with 100 μL of phosphate-buffered saline (PBS), and this suspension was spread on YG agar plates with 15 µg/mL of chloramphenicol. After 2 d of incubation at 30 °C, the number of colonies was counted.

To isolate gut-breached bacteria, the hemolymph of Sym^soil third instar nymphs was collected. The collected hemolymph was suspended in 100 µL of PBS, spread on YG agar, and incubated at 30 °C for 2 d. Bacterial colonies were then subjected to colony PCR using universal primers for the 16S rRNA gene, and approximately 1,400 bp were sequenced for bacterial identification. As a control, Sym^Control insects fed only sterilized DWA were examined in the same manner to confirm that no bacterial colonies were detected in their hemolymph.

Stress-Sensitivity Tests.

Different concentrations of chemical reagents (EtOH, NaCl, SDS, polymyxin B, proteinase K, and Tween 20) were diluted in PBS and mixed with approximately 10⁵ cells of either in vitro cultured or in vivo M3-colonizing Burkholderia HK1 cells. After 1 h incubation, the number of viable bacteria was counted by the CFU assay. Detailed information is provided in SI Appendix, Materials and Methods.

qPCR.

The number of M4-colonizing C. insecticola symbionts in Sym^control, Sym^Bu, and Sym^Ab insects were quantified through qPCR. This was performed using the BSdnaA-F and BSdnaA-R primer set (SI Appendix, Table S3). In addition, the expression levels of three AMPs (riptocin, defensin, and thanatin) were measured by qPCR. Detailed information is available in SI Appendix, Materials and Methods.

Immune Suppression.

To suppress the humoral and cellular immunity in bean bugs, two methods were employed: RNAi of the relish gene was used to inhibit humoral immunity, and the saturation of hemocytes with injected microbeads was implemented to suppress cellular immunity. Detailed information is available in SI Appendix, Materials and Methods.

Infection of Burkholderia HK1 and Achromobacter HK2 cells.

To make Burkholderia HK1-infected (Sym^Bu) and Achromobacter HK2-infected (Sym^Ab) insects, second-instar nymph R. pedestris were initially infected with cultured C. insecticola gut symbiont. Two days post symbiont infection, these insects were subsequently infected with 10⁷ cells of either cultured Burkholderia HK1 or Achromobacter HK1 cells. The bacterial cells were supplied to the insects through their water source by suspending them in DWA. To make Sym^Bu>DWA bean bugs, the bacterial solution was replaced with DWA, which did not contain any bacteria, 2 d after the infection with Burkholderia HK1 or Achromobacter HK2 cells.

Microscopy Analyses.

Fluorescence microscopy.

The midgut colonized by bacterial cells was isolated in PBS, promptly transferred to a glass-bottom dish (Matsunami), and subsequently covered with a cover glass. The GFP or RFP signal, originating from the gut-colonizing bacteria, was observed with an epifluorescence microscope (DMI4000B; Leica). To observe the FITC-derived fluorescence of beads engulfed by host hemocytes, the hemolymph was collected as described above and mixed with 100 µL of PBS. Hemocytes were then harvested by centrifuging the solution at 300 rpm for 10 min at 4 °C. The supernatant was carefully discarded, and the pellet was resuspended in 20 µL of PBS. The prepared samples were deposited onto a glass slide, covered with a coverslip, and the phagocytosed beads were then observed with an epifluorescence microscope.

Confocal microscopy.

The M3 region of R. pedestris was dissected in PBS 12 h or 24 h postinfection with Burkholderia HK1 and transferred into a 2-mL microcentrifuge tube. The midgut samples were fixed by 4% paraformaldehyde in PBS for 15 min at room temperature (RT) and then permeabilized using PBS containing 0.5% Triton X-100 (PBST) for 10 min at RT. Subsequently, the samples were stained with DAPI (Thermo Fisher Scientific) for host nuclei and Alexa Fluor™ 647 Phalloidin (Thermo Fisher Scientific) for the host cytoskeleton, for 20 min at RT. To observe hemocytes and phagocytosed Burkholderia HK1 cells, hemolymph was extracted as described above from the HK1-infected insects and immediately fixed and stained with DAPI for 10 min. After staining, the samples were rinsed three times with fresh PBS. The prepared samples were then placed on a glass-bottom dish, mounted by ProLong Gold Antifade Mountant (Thermo Fisher Scientific), and covered with a coverslip. Observations were made using a confocal microscope with a 40× magnification oil objective (×40/1.3 HC PL Apo CS oil).

Transmission electron microscopy.

To observe bacterial morphology, in vitro Burkholderia HK1 cells were prepared by inoculating precultured bacteria, which had been cultivated overnight, into fresh YG liquid media and then cultivating it for an additional 8 h. In vivo Burkholderia HK1 cells were prepared by dissecting the M3 of fifth instar Sym^Bu bean bugs in PBS, homogenizing the sample with a pestle, and subsequently filtering it through a 5 µm pore size filter to eliminate gut debris. To observe gut-colonizing bacteria as well as epithelia-breaching bacteria, M3 samples of Sym^control and Sym^Bu bean bugs were prepared by isolating the midgut region in a fixative solution (0.1 M sodium phosphate buffer containing 2.5% glutaraldehyde, pH 7.4). Each of these prepared samples was prefixed in the fixative solution at 4 °C overnight and then postfixed in 2% osmium tetroxide at 4 °C for 1 h. After undergoing a series of dehydration steps with ethanol, each sample was embedded in Epon812 resin (TAAB). Ultrathin sections were made using an ultramicrotome (EM UC7; Leica), placed onto a copper mesh, stained with uranyl acetate and lead citrate, and observed under a transmission electron microscope (H-7600; Hitachi).

Statistical Analyses.

Statistical significance for host natural survival, growth period, fitness parmeters, gene expression levels of AMPs, the count of bacteria and host hemocytes, and the number of eggs laid was determined using the Mann–Whitney U test with Bonferroni correction. For the host survival rate after bacterial injection into the hemolymph, the log-rank test was used for two samples, and the log-rank test with Bonferroni correction was applied for more than three samples.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The 16S rRNA nucleotide sequence of gut commensal bacterium, Burkholderia HK1, has been deposited in the GenBank nucleotide database with accession number OR451220 (70). Raw Illumina sequencing data of M3 and M4 microbiome from R. pedestris have been deposited in the DDBJ Sequence Read Archive (DRA) with accession codes SAMD00637910 to SAMD00637945 (71).