Insect material

The samples of M. gracilicorne used in this study are as listed (see  Supplementary Table S1 (zs200163_TableS1.pdf)). The insects were collected from wild cucurbitaceous plants at 11 localities in Japan (see  Supplementary Figure S1 (zs200163_FigS1.pdf)). Some insects were immediately subjected to experiments, whereas other insects were maintained in the laboratory. The insects were kept at 25°C under constant light in plastic containers (30 cm × 22 cm × 6 cm) in which paper towels, fresh cucumbers as food, and wet cotton pads as water source were provided (Fig. 1B). The insects were transferred to new clean cages every week. Eggs laid on the substrata were collected every week and subjected to rearing and other experiments.

Histology

Adult insects were subjected to dissection of their digestive tract in phosphate buffered saline (PBS: 137 mM NaCl, 8.10 mM Na₂HPO₄, 2.68 mM KCl, 1.47 mM KH₂PO₄ [pH 7.4]), which consisted of morphologically differentiated midgut regions M1, M2, M3, M4B and M4 (Fig. 1C). The whole digestive tracts were fixed in 4% paraformaldehyde at 4°C overnight and subjected to whole-mount fluorescence in situ hybridization (wFISH) as previously described (Koga et al., 2009), in which Alexa555-labelled oligonucleotide probe EUB338 (5′- Alexa555-GCT GCC TCC CGT AGG AGT -3′) visualized bacterial cells and 4′,6-diamidino-2-phenylindole (DAPI) counterstained host cell nuclei. The tissue specimens were observed under a laser scanning microscope (LSM 700, Zeiss), a fluorescence microscope (DM6B, Leica) or a fluorescence stereo-microscope (M165FC, Leica).

For transmission electron microscopy (TEM), the dissected symbiotic M4 regions were prefixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) at 25°C for 1 h, and postfixed with 2% osmium tetroxide in 0.1 M phosphate buffer (pH 7.4) at 4°C for 1 h. Then, the tissues were dehydrated through a water-ethanol series, embedded in Epon812 resin, cut on an ultra-microtome (EM UC7, Leica), mounted on copper meshes, stained with uranyl acetate and lead citrate, and observed under a transmission electron microscope (H-7600, Hitachi).

Cultivation and DNA preparation of bacteria in symbiotic midgut region

The midgut M4 region was dissected from each adult insect in PBS, and half of the tissue was ground in 1 ml of PBS in a plastic tube using a plastic pestle. The suspension was appropriately diluted with PBS and subjected to bacterial cell counting using a hemocytometer under a light microscope. The suspension was also serially diluted with PBS to 1/1, 1/100 and 1/10,000, and 50 µl each was spread on an LB agar plate. The inoculated plates were incubated at 25°C under constant darkness for 24 h, and subjected to bacterial colony counting to obtain colony formation unit (CFU) values per M4 sample. For each M4 sample, eight bacterial colonies were picked at random for colony intact PCR of the bacterial 16S rRNA gene. The other half of the dissected M4 region was directly subjected to DNA extraction using a QIAamp DNA Mini Kit (Qiagen). The DNA samples were used as PCR templates for cloning and sequencing of the bacterial 16S rRNA gene.

Fig. 1.

(A) An adult female of M. gracilicorne on the vine of S. angulatus. (B) Laboratory rearing system of M. gracilicorne with fresh cucumber and water. Arrowheads indicate the insects. (C) Dissected alimentary tract of M. gracilicorne. M1, midgut first section; M2, midgut second section; M3, midgut third section; M4B, M4 bulb section; M4, midgut fourth section with crypts (symbiotic organ). (D) Bright field image of dissected M4 region, on which many crypts are arranged in two rows. (E) Epifluorescence image of dissected M4 region, the same as in (D), in which bacterial 16S rRNA is visualized in red by wFISH. Red bacterial signals are localized to the crypt cavities. (F) Laser confocal microscopic image of M4 region. Red and blue indicate bacterial 16S rRNA and host nuclear DNA, respectively. Red bacterial signals are seen in the crypt cavities. (G) TEM image of an M4 crypt, whose cavity is full of bacterial cells. (H) Magnified image of the bacterial cells.

PCR, cloning, sequencing and molecular phylogenetic analyses of bacterial 16S rRNA gene

The DNA samples of the dissected M4 regions were subjected to PCR amplification of the bacterial 16S rRNA gene with ExTaq (TaKaRa) and the primers 16SA1 (5′-AGA GTT TGA TCM TGG CTC AG-3′) and 16SB1 (5′-TAC GGY TAC CTT GTT ACG ACT T-3′) (Fukatsu and Nikoh, 1998). The PCR products were subjected to TA-cloning with the T-vector pMD20 (TaKaRa) and chemically competent DH5α cells. Three or more transformant Escherichia coli colonies were subjected to intact-colony PCR with AmpliTaq Gold (Thermo Fisher Scientific) and the primers Uni19 (5′-GTT TTC CCA GTC ACG ACG T-3′) and Rev20 (5′-AGC TAT GAC CAT GAT TAC GC-3′). The PCR products were subjected to sequencing with the primers Uni19 and Rev20 by the Sanger method, which was outsourced to Eurofins Genomics. While all the sequences from the same M4 sample were usually identical to each other, minor sequence differences, which were likely due to PCR errors or chimera formation with minor contaminant sequences, were occasionally identified and reconciled by making a consensus.

The bacterial colonies cultured from the dissected M4 regions were individually subjected to intact-colony PCR of the bacterial 16S rRNA gene with AmpliTaq Gold (Thermo Fisher Scientific) and the primers 16SA1 and 16SB1. The PCR products were subjected to nested PCR with ExTaq (TaKaRa) and the primers 16SA2 (5′-GTG CCA GCA GCC GCG GTA ATA C-3′) and 16SB2 (5′-CGA GCT GAC GAC ARC CAT GCA-3′) (Fukatsu and Nikoh, 1998), and direct sequencing with the primer 16SA2.

Multiple alignments of the nucleotide sequences were generated using MUSCLE (Edgar, 2004), from which ambiguously aligned sites and gap-containing sites were removed manually. Maximum-likelihood and neighbor-joining phylogenies were constructed using MEGA X (Kumar et al., 2018) with 500 bootstrap resamplings. Substitution model selection was conducted under the Akaike information criterion using MEGA X. The nucleotide sequences determined in this study were deposited in the DNA Data Bank of Japan (DDBJ) under accession numbers LC594073–LC594427 (see  Supplementary Table S1 (zs200163_TableS1.pdf)).

Vertical transmission test

Seven field-collected adult females (CKSI20-12–CKSI20-18; see  Supplementary Table S1 (zs200163_TableS1.pdf)) were individually reared in plastic cages and allowed to lay eggs. The first egg clutch and the second egg clutch were immediately transferred to new, clean, humidified plastic containers, and kept at room temperature. From each of the egg clutches, two 7-day-old eggs and a 7-day-old newborn nymph were collected. In this way, for each of the adult females, a dissected M4 sample, two 7-day-old egg samples, and two 7-day-old nymph samples were obtained and subjected to DNA extraction and PCR amplification of bacterial 16S rRNA gene with ExTaq (TaKaRa) and the primers 16SA2 and 16SB2. The PCR products were analyzed on 1% agarose gels, cloned as described above, and sequenced with the primer Uni19.

Rearing test with or without supplementation of cultured symbiotic bacteria

An isolated and preserved strain of the gut symbiont from an adult female (TSKB13-01; see  Supplementary Table S1 (zs200163_TableS1.pdf)) was cultured in LB medium at 25°C, washed and suspended in sterilized water, and adjusted to 10⁸ CFU/ml. Eggs were collected from mass-reared stocks of M. gracilicorne every week, allocated to several experimental groups (100–400 eggs per group), and kept in clean rearing containers. During the first instar stage, the insects of the symbiont-supplemented groups (19OP + Sym01, 19OP + Sym02, 19OP + Sym03, 20OP + Sym01 and 20OP + Sym02; see  Supplementary Table S1 (zs200163_TableS1.pdf)) were provided with fresh cucumbers and the symbiont-suspended water, whereas the insects of the control groups (19OP–Sym01, 19OP–Sym02, 19OP–Sym03, 20OP–Sym01 and 20OP–Sym02; see  Supplementary Table S1 (zs200163_TableS1.pdf)) were provided with fresh cucumbers and sterilized water. Upon second instar molt and thereafter, the insects of both groups were transferred to new rearing cages and provided with fresh cucumbers and sterilized water. During the rearing period of around three months, the rearing cage, food and water were renewed every week, and the insects' molting and survival were monitored and recorded. Survival rates were statistically compared between the symbiont-supplemented groups and the control groups by generalized Wilcoxon test. When all of the insects either reached adulthood or died, eight adult insects were randomly selected from the symbiont-supplemented groups (20OA+Sym01–20OA+Sym08; see  Supplementary Table S1 (zs200163_TableS1.pdf)) and the control groups (20OA–Sym01–20OA–Sym08; see  Supplementary Table S1 (zs200163_TableS1.pdf)) and subjected to dissection of their midgut M4 region. The dissected symbiotic organs were subjected to DNA extraction, PCR amplification of the bacterial 16S rRNA gene with the primers 16SA2 and 16SB2, and cloning and sequencing with ExTaq (TaKaRa) and the primer Uni19 as described above.

Histology and localization of gut symbiotic bacteria in M. gracilicorne

The dissected digestive tract of M. gracilicorne exhibited morphologically differentiated midgut regions M1, M2, M3, M4B and M4 (Fig. 1C), as generally observed in diverse stinkbugs (Kikuchi et al., 2008, 2011; Ohbayashi et al., 2015; Oishi et al., 2019). In comparison with the other stinkbugs, M. gracilicorne was peculiar in that the M4 region was very thin and long: thinner than any other midgut regions and longer than the total stretch from M1 to M4B (Fig. 1C). Though inconspicuous, close observation of the M4 region revealed small crypts arranged generally in two rows (and partially in three or four rows) (Fig. 1D), as previously described anatomically (Miyamoto, 1961). FISH detected bacterial 16S rRNA signals localized to the inner cavity of the crypts (Fig. 1E and F). TEM visualized the crypt cavity full of rod-shaped bacterial cells (Fig. 1G and H).

Molecular phylogenetic characterization of gut symbiotic bacteria in field populations of M. gracilicorne

The gut symbiotic bacteria of M. gracilicorne were characterized by cloning and sequencing of 16S rRNA gene sequences from the dissected M4 regions. In 2013, 13 adult insects from eight localities (blue in Supplementary Figure S1; see  Supplementary Table S1 (zs200163_TableS1.pdf)) were subjected to the analyses, which consistently yielded identical or very similar 16S rRNA gene sequences belonging to Pantoea-allied γ-proteobacteria from all of the insects (blue sequences in Fig. 2). Of these, nine sequences from six localities (Tsukuba, Miyazaki, Kita-Kyushu, Kumamoto, Kagoshima and Chikusei) were completely or nearly identical to each other, and also to the sequence of an environmental bacterium, designated as type X2, which was shown to be potentially capable of symbiosis with the pentatomid stinkbug Plautia stali (Hosokawa et al., 2016a); one sequence from Fukuyama clustered with the enteric/environmental bacterium Leclercia adecarboxylata; two sequences from Tochigi were almost identical to the sequence of the gut symbiont of the cydnid stinkbug Macroscytus japonensis (Hosokawa et al., 2012a); and one sequence from Kagoshima was identical to the sequence of the cultivable gut symbiont, designated as type E, of the stinkbug P. stali (Hosokawa et al., 2016a) (see Fig. 2). In 2020, we also analyzed 12 insects from four localities (red in  Supplementary Figure S1 (zs200163_FigS1.pdf); see  Supplementary Table S1 (zs200163_TableS1.pdf)), and obtained similar results (red sequences in Fig. 2): six sequences from three localities (Takasaki, Fukuoka and Chikusei) were completely or nearly identical to the type X2 potential symbiont of P. stali; one sequence from Takasaki was allied to L. adecarboxylata; three sequences from Chiba and Fukuoka were allied to Enterobacter ludwigii; one sequence from Fukuoka was related to the sequence of the gut symbiont of the cydnid stinkbug Adrisa magna (Hosokawa et al., 2012a); and one sequence from Chiba was identical to the sequence of the cultivable gut symbiont, designated as type C, of the stinkbug P. stali (Hosokawa et al., 2016a). These results indicated that (i) M. gracilicorne is stably associated with Pantoea-allied γ-proteobacteria within the crypts of the midgut M4 region, (ii) the symbiotic bacteria exhibit a considerable level of diversity across individuals and populations of M. gracilicorne, (iii) the major symbiotic bacteria are the type X2 potential symbiont of P. stali of environmental origin, and (iv) the minor symbiotic bacteria also represent other cultivable symbionts of P. stali and other stinkbugs.

Fig. 2.

Molecular phylogenetic placement of the gut symbiotic bacteria of M. gracilicorne collected from field populations. Maximum-likelihood phylogeny inferred from 1325 aligned nucleotide sites of the 16S rRNA gene under the GTR+G+I model is shown, while neighbor-joining phylogeny exhibits substantially the same tree topology. Bootstrap probabilities are indicated at the nodes in the order of maximum-likelihood/neighbor-joining. Accession numbers for the sequences are shown in brackets. As for colors, see the notes on the upper left.

Cultivability of gut symbiotic bacteria of M. gracilicorne

These observations suggested the possibility that the gut symbiotic bacteria of M. gracilicorne may be derived from cultivable environmental bacteria. In 2013, we performed a preliminary cultivation trial of homogenized M4 regions, and observed bacterial colony formation on LB agar plates for 10 of 13 insects examined (see  Supplementary Table S1 (zs200163_TableS1.pdf)). In 2020, we performed more careful and thorough investigation on the cultivability of the symbiotic bacteria using 12 adult females from four localities (Table 1; see  Supplementary Table S1 (zs200163_TableS1.pdf) and see  Supplementary Table S2 (zs200163_TableS2.pdf)). Direct cell counting detected largely constant bacterial titers ranging from 10⁷ to 10⁸ per M4 region (Table 1). Meanwhile, bacterial culturing on LB agar plates detected CFU values mostly ranging from 10⁴ to 10⁵ per M4 region (Table 1), indicating that only a very small fraction (∼1/1000) of the gut symbiont population may be capable of colony formation on agar plates. Plausibly, the majority of the gut symbiont population may be in a non-proliferating status, and the M4 homogenization must be so damaging that the majority of the symbionts are killed in the experimental treatments. Such low efficiencies of symbiont culturing were also observed among cultivable gut symbiotic bacteria of other stinkbugs (T. Hosokawa et al., unpublished data). Notably, three insects from Takasaki (and one insect from Chikusei) exhibited conspicuously low CFU values of 10² to 10³ per M4 region (Table 1). When eight bacterial colonies from each of the insects were subjected to sequencing of the 16S rRNA gene and comparison with the M4-derived symbiont sequence, it turned out that (i) most of the colony-derived sequences exhibited 100% identity to the M4-derived symbiont sequence, indicating that the gut symbiotic bacteria of M. gracilicorne are generally cultivable, (ii) in several cases (e.g., two insects from Chikusei and one insect from Chiba), one or two non-symbiont sequences were detected, indicating that minor non-symbiont bacteria may often coexist with the major symbiotic bacteria in the M4 region of M. gracilicorne, and (iii) as for three insects from Takasaki, all of the eight sequences disagreed with the symbiont sequence and instead represented a variety of other bacterial sequences, suggesting the possibility that the gut symbiotic bacteria of M. gracilicorne of the Takasaki population might be exceptionally resistant to cultivation, although other possibilities, such as accidental sampling of unhealthy insects at Takasaki, should also be taken into account (Table 1; see  Supplementary Table S2 (zs200163_TableS2.pdf)). In this context, it may be notable that a widely used herbicide, glyphosate, suppresses insect gut bacteria (Motta et al., 2018).

Transmission mode of gut symbiotic bacteria of M. gracilicorne

In this study, we developed a laboratory rearing system for M. gracilicorne supplied with fresh cucumbers (Fig. 1B). In our preliminary rearing trial in 2013, field-collected adult insects were kept in a rearing cage, they laid many eggs, and some nymphs from the eggs attained adulthood, of which 23 adult insects were inspected for the bacterial 16S rRNA gene sequence from their dissected M4 region. Considering that stinkbugs of the superfamily Pentatomoidea generally transmit their gut symbiont vertically via egg surface contamination or deposition of symbiont-containing materials with eggs (Salem et al., 2015; Hosokawa and Fukatsu, 2020), we expected that the laboratory-reared adult insects of M. gracilicorne should inherit the gut symbiont maternally. Unexpectedly, however, most of the insects did not harbor the maternal gut symbiont: only one insect hosted the type X2 gut symbiont, and the remaining 22 insects were associated with a variety of Pantoea-allied γ-proteobacteria, the majority of which were allied to either the type X1, X3 or X5 potential symbiont of P. stali (Hosokawa et al., 2016a) (Fig. 3). These observations suggest the possibility that M. gracilicorne exhibits neither egg surface contamination nor vertical symbiont transmission. In 2020, in order to confirm this, we individually reared seven field-collected adult females, collected their eggs and nymphs serially, and PCR-amplified, cloned and sequenced partial bacterial 16S rRNA gene sequences from these samples (Table 2; see  Supplementary Table S3 (zs200163_TableS3.pdf)). In the majority of the egg samples (8/14), none of eight clones per sample represented the symbiont sequence (Table 2), indicating the absence of stable smearing of the symbiotic bacteria onto eggs. Concordantly, in the majority of the nymphal samples (11/14), none or a few (0, 1 or 2) of eight clones per sample represented the symbiont sequence (Table 2), indicating the absence of stable vertical transmission of the symbiotic bacteria. Meanwhile, some eggs and nymphs exhibited the symbiont sequence at low frequencies (mostly one, two or three, and at most five, per eight clones) (Table 2), suggesting that the eggs and the nymphs may be occasionally contaminated by the symbiotic bacteria of either parental or environmental origins. The nymphs exhibited relatively higher symbiont infection frequencies than the eggs (Table 2), which is probably due to active symbiont exploitation by the nymphs. The non-symbiont 16S rRNA gene sequences obtained from the eggs and the nymphs were phylogenetically diverse (see  Supplementary Table S3 (zs200163_TableS3.pdf)), which plausibly reflected the environmental microbiota. Notably, an Ehrlichia-like 16S rRNA gene sequence was frequently detected from the eggs and the nymphs (see  Supplementary Table S3 (zs200163_TableS3.pdf)), which may represent a rickettsial facultative endosymbiont of M. gracilicorne as in Nysius seed bugs (Matsuura et al., 2012), and its microbial identity should be investigated in future studies.

Table 1.

Direct bacterial counts, colony formation units (CFUs), and symbiont detection rates among cultured colonies from dissected midgut M4 region of M. gracilicorne. For details, see  Supplementary Table S2 (zs200163_TableS2.pdf).

Fig. 3.

Molecular phylogenetic placement of the gut bacteria of laboratory-reared adult insects of M. gracilicorne. Maximum-likelihood phylogeny inferred from 1332 aligned nucleotide sites of the 16S rRNA gene under the GTR+G+I model is shown, while neighbor-joining phylogeny exhibits substantially the same tree topology. Bootstrap probabilities are indicated at the nodes in the order of maximum-likelihood/neighbor-joining. Accession numbers for the sequences are shown in brackets. As for colors, see the notes on the upper left.

Presumable benefit of gut symbiotic bacteria for survival of M. gracilicorne

Finally, we evaluated the biological consequences of the gut symbiotic bacteria in M. gracilicorne. We generated several experimental groups consisting of 100–400 eggs laid by field-collected adult females. In the symbiont-supplemented groups, first instar nymphs were provided with fresh cucumbers and symbiont-supplemented water, whereas in the control groups, first instar nymphs were supplied with fresh cucumbers and sterilized water. From the second instar and on, the insects of both groups were provided with fresh cucumbers and sterilized water only. In 2019, we generated three symbiont-supplemented groups and three control groups. In all of the groups, while first instar nymphs suffered high mortality, second instar and older nymphs exhibited relatively good survivorship. The adult emergence rate, 25.6% (161/629), in the symbiont-supplemented groups was significantly higher than the adult emergence rate, 9.5% (52/548), in the control groups (Fig. 4A). In 2020, we repeated the rearing experiments with two symbiont-supplemented groups and two control groups, which yielded similar results: high mortality at the first instar and stable survivorship at the later stages, and significantly higher adult emergence rate in the symbiont-supplemented groups, 13.1% (62/474), than in the control groups, 4.7% (25/532) (Fig. 4B). These results suggested that the gut symbiont of M. gracilicorne may affect host's survival positively.

Table 2.

Vertical transmission rate of gut symbiotic bacteria from mother to offspring of M. gracilicorne. For details, see  Supplementary Table S3 (zs200163_TableS3.pdf).

Origin of gut symbiotic bacteria in non-inoculated control insects of M. gracilicorne?

However, when we inspected the gut bacteria of the symbiont-supplemented adult insects and the control adult insects obtained in 2020, perplexing patterns of bacterial infections emerged. As expected, all of the symbiont-supplemented insects were infected with the gut symbiotic bacteria of the inoculated type (Table 3). On the other hand, unexpectedly, some of the control insects (five of eight) were also infected with the gut symbiotic bacteria of the inoculated type, whereas the other control insects (three of eight) were associated with bacteria allied to cultivable gut symbionts (e.g., type F, type X5, etc.) of P. stali and other stinkbug species (Table 3; see  Supplementary Figure S2 (zs200163_FigS2.pdf)). How could the gut symbiotic bacteria be acquired by the control insects despite the absence of symbiont inoculation? There are two possible routes: (i) experimental contamination from the symbiont-supplemented insect populations maintained in parallel, and (ii) environmental contamination from fresh cucumbers supplied as food. The rearing experiments of M. gracilicorne from egg to adult continued for as long as three months, during which accidental contamination of the symbiotic bacteria from the symbiont-supplemented cages to the control cages might have occurred despite our careful experimental handling of them. Alternatively, considering that surface sterilization of cucumbers may facilitate rotting and cause negative effects on the insect rearing, we used fresh cucumbers washed with tap water as laboratory food for M. gracilicorne, which can be the source of environmental Pantoea-allied bacteria capable of symbiosis with stinkbugs (Hosokawa et al., 2016a). Notably, some 16S rRNA gene sequences detected in the control insects were nearly, but not completely, identical to the sequence of the inoculated symbiont (ex. 20OA–Sym03, 20OA–Sym04 and 20OA–Sym07 in Table 3), which favors the idea of food-derived bacterial infections. It is also notable that all of the non-inoculated control adult insects we inspected were not aposymbiotic but infected with Pantoea-allied bacteria (Table 3; see  Supplementary Figure S2 (zs200163_FigS2.pdf)). In this context, it seems relevant that, even when nymphs of the stinkbug P. stali were maintained under aseptic conditions, accidental bacterial infections occasionally occurred from food peanuts, and the insects infected with the non-symbiont bacteria exhibited improved growth and survival in comparison with the bacteria-free insects (Nishide et al., 2017). Considering all of these results, observations and circumstances taken together, our current hypothesis to account for the survival patterns of the symbiont-inoculated insects and the control insects (Fig. 4) is that (i) M. gracilicorne requires Pantoea-allied gut symbiotic bacteria for normal growth and survival, (ii) the symbiont-supplemented nymphs acquired the symbiotic bacteria immediately, (iii) by contrast, the control nymphs acquired the food-derived Pantoea-allied bacteria only gradually and sporadically, and (iv) consequently, the symbiont-supplemented insects exhibited better survival than the control insects. Needless to say, we recognize that there are a number of uncertainties regarding this scenario, including infection dynamics, phenotypic effects, and presence on cucumbers of the Pantoea-allied gut bacteria, which will be investigated in future studies.

Conclusion and perspective

In conclusion, the gut bacterial symbiosis in M. gracilicorne is unique among pentatomoid stinkbugs: the microbial nature of the symbionts as Pantoea-allied γ-proteobacteria and the symbiont localization within the midgut crypts are typical of pentatomoid stinkbugs, whereas the lack of vertical symbiont transmission and the environmental symbiont acquisition are exceptional among pentatomoid stinkbugs, being rather close to the β-proteobacterial gut symbiosis found among coreoid and lygaeoid stinkbugs. The unique gut symbiosis of M. gracilicorne may provide unique opportunities to better understand the ecology and evolution of symbiosis, given that whether symbiont acquisition occurs vertically or environmentally pivotally affects the evolutionary fate of host-symbiont relationships such as virulence, infection cost, infection stability, fitness effects, symbiont genome reduction, host-symbiont interdependence, etc. (Ewald, 1987; Kikuchi et al., 2007; Bright and Bulgheresi, 2010; McCutcheon and Moran, 2012; Russell, 2019). Although this study is the first to unequivocally report the cultivable and environmentally acquired gut symbiotic γ-proteobacteria in the Pentatomoidea, we suspect that many other cases are to be found and characterized. For example, see Fig. 2, in which the gut symbiotic bacteria of the soil-dwelling cydnid stinkbugs Macroscytus japonensis and Adrisa magna (see Hosokawa et al., 2012a) are genetically almost identical to the cultivable gut symbiotic bacteria of M. gracilicorne; the cultivable gut symbiotic bacteria types C, D, E and X2 of the pentatomid stinkbug P. stali (see Hosokawa et al., 2016a), are phylogenetically intermingled with the cultivable gut symbiotic bacteria of M. gracilicorne; and the cultivable gut symbiotic bacteria types C, D and E are also found in subtropical stinkbug species such as Axiagastus rosmarus (Pentatomidae), Lampromicra miyakona (Scutelleridae) and Solenosthedium chinense (Scutelleridae) (see Hosokawa et al., 2016a). Therefore, by comparatively investigating the vertically transmitted and environmentally acquired gut symbiotic associations among diverse pentatomoid stinkbugs, we will be able to empirically test a variety of evolutionary hypotheses regarding symbiosis, which we believe should be a promising direction of future studies.

Fig. 4.

Survival curves of laboratory-reared M. gracilicorne with or without supplementation of cultured symbiotic bacteria. (A) The result of the experiment in 2019. (B) The result of the experiment in 2020. Red symbols and blue symbols show survival rates of the symbiont-supplemented groups and the control groups, respectively. Red lines and blue lines are drawn based on the mean survival rates of the symbiont-supplemented groups and the control groups, respectively, with numbers indicating the total numbers of surviving insects for each of the treatment groups. The initial number of first instar nymphs and the final number of adult insects for each treatment group are shown on the upper right. The survival rates were significantly different between the symbiont-supplemented groups and the control groups in both the 2019 and 2020 experiments (P < 0.0001, generalized Wilcoxon test). The high mortality in the first instar is probably because the cucumber peel is too thick for tiny first instar nymphs to penetrate with the mouthpart.

Table 3.

Bacterial infections in laboratory-reared M. gracilicorne with and without supplementation of cultured symbiotic bacteria.