Since Wolbachia endosymbionts are difficult to culture, detection and characterization of the bacteria have been principally performed by PCR-based techniques. The most popular gene for bacterial characterization, 16S ribosomal DNA, is almost useless for differentiation of Wolbachia strains due to a paucity of nucleotide substitutions. Therefore, the diversity of Wolbachia endosymbionts has been analyzed by using fast-evolving genes, such as ftsZ and wsp. Based on ftsZ sequences, most of Wolbachia from arthropods are classified into two major clades, A group and B group (41). The A and B groups are further divided into a number of subgroups based on wsp sequences (44).

A single insect individual may be infected with more than one strain of Wolbachia. Previous extensive surveys of natural insect communities detected low levels (1.2 to 5.8%) of multiple Wolbachia infections (28, 36, 39, 40, 42). However, these values cannot be regarded as true multiple-infection frequencies. In these studies, multiple infections were detected by a diagnostic PCR technique using specific primers that can differentiate between A group and B group of Wolbachia. The technique detects only AB double infections, whereas AA and BB double infections are erroneously recognized as single infections. Using the technique, it is impossible to detect triple infections. Therefore, these reported values of multiple Wolbachia infections are no doubt underestimates. In fact, when 89 mosquito species were analyzed by using 12 sets of specific primers for known wsp subgroups of Wolbachia, the diagnostic PCR of much better resolution unveiled a remarkably higher frequency, 15.7% (14 of 89 species), of double infections, consisting of 9.0% (8 of 89) AB, 5.6% (5 of 89) AA, and 1.1% (1 of 89) BB infections (20). It should be noted, however, that this approach tends to fail to detect novel Wolbachia strains because the specific primers were designed for already-known wsp sequences. Therefore, a thorough survey independent of such specific PCR primers would lead to a deeper grasp of Wolbachia diversity in natural insect communities.

The Heteroptera, known as true bugs, is one of the most diverse insect groups with incomplete metamorphosis. Many heteropteran bugs are known to harbor extracellular symbiotic bacteria in their midgut caeca. The symbiotic bacteria are, in general, maternally inherited to the offspring by superficial contamination of the eggshell with excretion containing the bacteria (2, 3, 7, 31). In several heteropteran bugs, deprivation of the symbiont was reported to result in retarded growth and/or mortality of the nymphs, suggesting some important role of the symbiont for the host (1, 5, 11, 22, 32). On the other hand, intracellular symbionts of heteropteran bugs have been very poorly investigated, except for those of blood-sucking reduviid bugs and bedbugs (13, 14, 26).

In this study we conducted an extensive survey of Wolbachia infection in Japanese terrestrial heteropteran bugs. Two hundred twenty-six field-collected heteropteran bugs, representing 134 species from 19 families, were subjected to PCR detection with universal primers for Wolbachia, cloning of the PCR products, restriction fragment length polymorphism (RFLP) analysis of infecting Wolbachia types, and molecular phylogenetic characterization of all the detected Wolbachia strains. In total, 36 single infections, 10 double infections, and 1 triple infection were detected, and 59 Wolbachia strains, some of which belonged to novel lineages, were identified.

Materials. The terrestrial heteropteran bugs used in this study are listed in Table 1. They were collected in 2001 and 2002 in Japan, except for the samples of the Acanthosomatidae that were collected in 1994 and 1995. These bugs were preserved in acetone immediately after collection until molecular analysis (4).

TABLE 1. Distribution of Wolbachia by Japanese terrestrial heteropteran species

DNA extraction. DNA was individually extracted from 1 to 10 insects for each species. Females were preferentially used if available. Small bugs were subjected to whole-body extraction, whereas large bugs were dissected and their abdomens or ovaries were extracted. The tissues were homogenized in 200 μl of lysis buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 0.1 M NaCl, 0.5% SDS, 0.2 mg of Proteinase K/ml) and were incubated at 56°C for 2 h. The lysate was extracted with 200 μl of PCI (phenol:chloroform:isoamyl alcohol [25:24:1]). DNA was recovered from the purified lysate by ethanol precipitation and was dissolved in 200 μl of TE buffer (10 mM Tris-HCl [pH 8.0], 0.1 mM EDTA).

Specific PCR detection. Diagnostic PCR was performed using specific primers for two Wolbachia genes. The primers ftsF (5′-GTATGCCGATTGCAGAGCTTG-3′) and ftsR (5′-GCCATGAGTATTCACTTGGCT-3′) yielded a 0.8-kb segment of ftsZ (10). The primers wspF (5′-GGGTCCAATAAGTGATGAAGAAAC-3′) and wspR (5′-TTAAAACGCTACTCCAGCTTCTGC-3′) yielded a 0.6-kb segment of wsp (21). To check the quality of DNA samples, a 1.5-kb segment of insect mitochondrial ribosomal DNA was amplified using the primers MtrA1 and MtrB1 (6). PCRs were conducted using Takara TaqDNA polymerase (Takara) and its supplemented buffer system under a temperature profile of 95°C for 4 min followed by 35 cycles of 95°C for 30 s, 50°C (ftsZ) or 55°C (wsp) for 30 s, and 72°C for 30 s. The PCR products were electrophoresed on Tris-acetate-EDTA (TAE)-agarose gels for 15 min, stained with ethidium bromide, and observed on a UV transilluminator. DNA prepared from ovaries of the Wolbachia-infected bruchid beetle Callosobruchus chinensis (21) was used as a positive control sample.

Cloning, RFLP typing, and sequencing of wsp gene segment. PCR products of the wsp gene segment were directly cloned with TA cloning vector pT7Blue (Takara) and Escherichia coli DH5α-competent cells (Takara) by using an ampicillin and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside blue-white selection system. To check the length of the inserted DNA fragment, white colonies expected to contain inserted plasmid were directly subjected to PCR using the primers wspF and wspR. When a PCR product of expected size (0.6 kb) was obtained, the product was double digested with the restriction endonucleases ClaI and DraI and was electrophoresed in TAE-agarose gels for typing of the wsp clone. More than three clones from each of all RFLP types were cultured overnight in 3 ml of Luria-Bertani medium containing ampicillin and were subjected to plasmid extraction by using a QIAprep-Spin Miniprep kit (QIAGEN). The purified plasmids were eluted with 50 μl of distilled water and were used for sequencing. A dye terminator-labeled cycle sequencing reaction was conducted with DNA Sequencing kit FS (Perkin Elmer) and two sequencing primers, Univ19 (5′-GTTTTCCCAGTCACGACGT-3′) and Rev20 (5′-AGCTATGACCATGATTACGC-3′), under a temperature profile of 95°C for 4 min followed by 30 cycles of 95°C for 30 s, 50°C for 1 min, and 60°C for 4 min. The products were analyzed with an ABI PRISM 377 DNA sequencer (Perkin Elmer).

Molecular phylogenetic analysis. The wsp gene sequences determined were subjected to molecular phylogenetic analysis together with wsp gene sequences of Wolbachia representatives retrieved from the DDBJ nucleotide sequence database. A multiple alignment of the wsp sequences was generated by the program package Clustal W (34) and then was manually realigned. Phylogenetic trees were constructed by the neighbor-joining method using Clustal W. Bootstrap tests were performed with 1,000 replications.

Randomization test. A randomization test was performed using pairwise genetic distances between 20 wsp gene sequences from 10 doubly infected bugs. A 20 by 20 distance matrix was constructed using Kimura's two parameter model (19) packaged in Clustal W (34). From 190 pairwise distances in the matrix, 10 distances were randomly sampled and averaged to obtain a mean distance. The sampling was repeated 10,000 times, by which a null distribution of the mean distance was generated. The observed mean distance calculated for the 10 pairs of coinfecting Wolbachia strains was statistically evaluated in comparison with the null distribution.

Nucleotide sequence accession numbers. The wsp gene sequences determined in this study were deposited in the DDBJ nucleotide sequence database, and the accession numbers are shown in boldface in Fig. 1.

FIG.1. Molecular phylogenetic analysis of 59 wsp sequences identified from 43 Japanese terrestrial heteropteran species. A total of 656 aligned nucleotide sites were subjected to the analysis. On the right side are shown the names of groups and subgroups of (more ...)

Detection of Wolbachia infection. By using specific PCR detection of ftsZ and wsp genes, Wolbachia infection was detected in 47 of 134 heteropteran species examined. There was no discrepancy between the ftsZ results and wsp results. These species belonged to 13 of 19 heteropteran families surveyed (Tables 1 and 2).

TABLE 2. Distribution of Wolbachia in Japanese terrestrial heteropteran families

Identification of Wolbachia infection types. The PCR products of wsp gene segments were cloned and subjected to RFLP typing. Of the 47 samples, some showed only one type of RFLP pattern, suggesting either single infection or cryptic double infection. Others exhibited two types of RFLP pattern, indicating double infection. One sample, Charagochilus angusticollis, showed three distinct RFLP patterns, suggesting triple infection. More than three clones for each of the RFLP types were subjected to DNA sequencing. At this stage, some of the single RFLP types turned out to contain two distinct sequences, implying double infection undetected by the RFLP analysis. In total, 59 wsp sequences were identified from the 47 samples. These sequences were assigned to either A group or B group of Wolbachia based on the sequence similarity. Of the 59 sequences, 16 and 43 were placed in A group and B group, respectively. As a result, the 47 species of Wolbachia-infected bugs were classified into 8 species with A infection, 28 species with B infection, 2 species with AA infection, 3 species with AB infection, 5 species with BB infection, and 1 species with ABB infection (Tables 1 and 2).

Molecular phylogenetic analysis. A phylogenetic tree of all the Wolbachia strains identified from the heteropteran bugs was constructed based on the wsp sequences (Fig. 1). Wolbachia strains obtained from the same insect families were not clustered into distinct groups but were scattered throughout the phylogenetic tree. Wolbachia strains obtained from the same insect genera were also scattered on the tree, although several congenic clusters (for example, Megacopta cribraria 1 and Megacopta punctatissima 1, Cletus punctiger and Cletus trigonus, and Elasmucha putoni and Elasmucha amurensis) were identified. Wolbachia strains obtained from the same insect individuals, comprising multiple infections, were not clustered at all (Fig. 2). The phylogenetic analysis revealed several novel lineages of Wolbachia strains. For example, next to the subgroup Con, wsp sequence from Paromius exiguus constituted an isolated lineage, and wsp sequences from E. putoni and E. amurensis formed a distinct clade. The most notable was the identification of “Bugs” subgroup consisting of 11 wsp sequences. Except for two sequences from a dragonfly, Perithemis tenera, and a wasp, Ceroptres cerri, the well-defined but diverse clade was composed of nine wsp sequences from heteropteran bugs (Agonoscelis femoralis, Eurydema dominulus, Nysius plebeius, Orius minutus, Orius strigicollis, Paraeucosmetus pallicornis 2, Piocoris varius 1, Pyrrhocoris sinuaticollis, Spirostethus hospes 2) that represented four families (Anthocoridae, Lygaeidae, Pentatomidae, and Pyrrhocoridae).

FIG. 2. Molecular phylogenetic analysis of 23 wsp sequences identified from 10 doubly infected heteropteran bug and 1 triple-infected heteropteran bug. A and B on the left side indicate the groups of Wolbachia (41). On the right side Wolbachia strains detected (more ...)

Expected and observed frequencies of double infections. From the 134 heteropteran species examined, 16 strains of A-type and 43 strains of B-type Wolbachia were identified. Provided that 11.0% (16 of 146) and 29.5% (43 of 146) are regarded as frequencies of A type and B type in the heteropteran bugs, respectively, frequencies of double infections were expected to be 1.2% for AA, 3.2% for AB, and 8.7% for BB on an assumption of random combination. Observed double infection frequencies were 1.5% (2 out of 134) for AA, 2.2% (3 out of 134) for AB, and 3.7% (5 out of 134) for BB (Table 2). The differences between the expected and observed values were statistically not significant (χ² = 5.29, df = 2, P = 0.071).

Relatedness between Wolbachia strains and double infections. To know whether coinfecting Wolbachia strains tend to be closely or distantly genetically related, a randomization test was applied to 20 wsp sequences from 10 doubly infected insects (Fig. 3). The genetic distances between coinfecting Wolbachia strains showed no statistically significant bias (P = 0.61).

FIG. 3. Randomization test to assess the relatedness between Wolbachia strains identified from doubly infected host insects.

Wolbachia infection in the heteropteran bugs. Our extensive survey of Wolbachia infection in 134 species of Japanese terrestrial heteropteran bugs revealed a high infection frequency at 35.1% (47 out of 134). This result clearly indicates that Wolbachia infection is prevailing in the heteropteran species diversity, as has been reported for other insect groups (8, 16, 20, 25, 28, 36, 39, 40, 42). The value of 35.1% is, however, no doubt an underestimate, because quite limited numbers of individuals and populations were examined in this study. To date, no studies have been conducted on the effects of Wolbachia infection in heteropteran insects. To understand the mechanism for maintaining the high infection frequency, phenotypic effects of Wolbachia infection on the heteropteran bugs, such as fitness consequences, vertical transmission rates, and intensity of cytoplasmic incompatibility and other reproductive phenotypes, should be investigated.

Multiple Wolbachia infection in the heteropteran bugs. Our approach, independent of group-specific PCR primers, identified a considerable frequency of multiple infection at 8.2% (11 of 134) in the heteropteran bugs. The frequency was remarkably higher than the multiple infection frequencies at 1.2 to 5.8% reported in previous studies where only AB double infections were detected by diagnostic PCR (28, 36, 39, 40, 42). The frequency of AB infection in this study was 2.2%, which is almost comparable to the frequencies of 1.2 to 5.8%. Therefore, the higher frequency of 8.2% was probably not due to higher multiple-infection rates in the heteropteran bugs but was due to improved detection of non-AB-type multiple infections. In spite of the improved detectability, however, the estimated multiple-infection frequencies are still likely to be underestimates. Although we selected the restriction endonucleases ClaI and DraI for the best detectability of wsp sequence polymorphisms, some multiple infections may fail to be recognized by the RFLP analysis.

Putative horizontal transfers of Wolbachia. Wolbachia endosymbionts are maternally transmitted through host generations by vertical transmission. However, phylogenetic relationships of Wolbachia strains are generally not concordant with those of their host insects (30, 41, 44), which strongly suggests that horizontal transfers of Wolbachia between unrelated host organisms may have taken place at a considerable frequency over evolutionary time. Although the mechanism of horizontal transfer in natural conditions is unknown, a plausible hypothesis is that parasitoids may be involved in the process (9, 12, 23). In the heteropteran bugs, little congruence was found between the Wolbachia phylogeny and the host systematics (Fig. 1 and 2), supporting frequent horizontal transfers of Wolbachia in the evolutionary course of the Heteroptera. In several cases, closely related congenic bugs possessed closely related wsp sequences, which may favor the idea of host-symbiont cospeciation of relatively recent origin. However, the pattern can also be explained by horizontal transfer of Wolbachia between closely related host species that often utilize similar ecological niches and share common parasitoids. In this context, survey of Wolbachia in hymenopteran and strepsipteran parasitoids that exploit heteropteran bugs will be of interest.

Identification of novel lineages of Wolbachia. From the heteropteran bugs we identified 59 wsp sequences, some of which were conceived to represent novel lineages of Wolbachia. Notably, a novel clade of Wolbachia tentatively named Bugs subgroup, whose members are predominantly associated with heteropteran bugs, was identified (Fig. 1). Since the Wolbachia phylogeny in the group did not reflect the host systematics, the association cannot be ascribed to host-symbiont cospeciation. Wolbachia of the group might be adapted to the internal environment of the bugs and/or might be readily vectored between bug lineages by parasitoids. However, the integrity of the group requires further verification. As more wsp sequences are reported from diverse insect taxa, it appears possible that the group would be diluted and disintegrated by the sequences from nonheteropteran insects.

Interaction between coinfecting Wolbachia endosymbionts. When different endosymbionts share the common habitat of the host body, it is expected that various types of interactions may arise between the symbionts. The frequent occurrence of multiple Wolbachia infections in the heteropteran bugs provides an opportunity to test the ideas concerning microbiological and ecological interactions between coexisting endosymbionts.

Competitive exclusion between Wolbachia strains? Since available resources and space are strictly limited in the host body, intersymbiont competition may result in exclusion of either of the symbionts. If Wolbachia coinfections are generally unstable due to such competitive exclusion, it is expected that observed frequencies of multiple infections in host populations may be lower than expected. In the heteropteran bugs, expected double infection frequencies were calculated to be 1.4% for AA infection, 3.8% for AB infection, and 10.3% for BB infection. The values were not significantly different from the observed frequencies of 1.5% for AA infection, 2.2% for AB infection, and 3.7% for BB infection. Therefore, competitive exclusion between coinfecting Wolbachia strains was not detected in this study, although the possibility that weak antagonistic interactions do exist cannot be ruled out. It may be notable that the observed frequency of BB infection (3.7%) was remarkably lower than the expected one (10.3%).

Coinfecting Wolbachia strains genetically closely or distantly related? If intersymbiont competition leads to intrahost niche division, differential tissue tropism of the symbionts may be observed, and different types of symbionts may be favored to coexist. If Wolbachia coinfections tend to require such functional divergence, it is expected that genetic divergence between the coexisting symbionts may be greater than expected. On the contrary, similar types of symbionts may tend to coexist on the grounds that the symbionts have to adapt to the same environment in the host body. If so, it is expected that genetic divergence between the coinfecting Wolbachia strains may be smaller than expected. Contrary to either of the expectations, the randomization test indicated that genetic divergence of coinfecting Wolbachia strains exhibited no significant difference from the expected value (Fig. 3). In the heteropteran bugs, therefore, relatedness between the coinfecting Wolbachia strains was neither close nor distant but was almost at the background level. It should be noted, however, that the frequencies of coinfection are inherently underestimated and therefore may be affected by the bias.

Tentative conclusion. These results suggest that, in the heteropteran bugs, interactions between coinfecting Wolbachia strains are generally not intense and that Wolbachia coinfections have been established through a stochastic process probably depending on occasional horizontal transfers. Recent studies have suggested the occurrences of putative genetic recombination between Wolbachia strains (17, 27, 38), which might also contribute to such coinfections and diversity of Wolbachia endosymbionts. It should be noted, however, that those arguments above are based merely on the global patterns of combination and phylogenetic relationship of coinfecting Wolbachia strains. Future detailed investigation of specific heteropteran bugs infected with multiple Wolbachia strains would reveal various types of ecological, physiological, and microbiological interactions, such as differential tissue tropism reported in a bruchid beetle (15).

Perspective. Owing to this study, the Heteroptera is now rated among the insect taxa whose Wolbachia infection and diversity are best documented. A number of notorious agricultural pests are known from the Heteroptera, and some of them can be conveniently maintained in the laboratory by using grains, seeds, potted plants, and/or artificial diets (18, 29, 35, 43), which will provide excellent model systems to investigate host-symbiont and symbiont-symbiont interactions. Although many heteropteran bugs are known to harbor symbiotic bacteria in the gut, microbiological nature and biological function of the symbionts are, in general, poorly understood (2, 3, 5). Not only interactions between coinfecting Wolbachia strains but also interactions between Wolbachia and gut symbiont will be of interest.

We thank H. Higuchi, N. Ijichi, M. Ishizaki, K. Itoh, T. Kashima, A. Kikuchi, K. Kohno, S. Kudo, S. Moriya, M. Takai, T. Takemoto, and T. Tsuchida for providing bug samples, A. Sugimura, H. Ouchi, S. Tatsuno, S. Suo, and K. Sato for technical and secretarial assistance, and M. Yokoyama for programming and J. Kojima for encouragement.

This research was supported by the Program for Promotion of Basic Research Activities for Innovation Biosciences (ProBRAIN) of the Bio-Oriented Technology Research Advancement Institution.