Introduction

Endosymbionts are frequently detected in a variety of insect species [1, 2]. Some endosymbionts interact with the host in a mutualistic or parasitic manner [2, 3]. As an example, endosymbiotic bacteria [4,5,6,7,8], microsporidia [9], and an RNA virus [10] are known to manipulate host reproduction. The best-studied endosymbiont is the alpha-proteobacterium Wolbachia, which generally is responsible for three types of reproductive manipulation, namely, male-killing, feminization, and cytoplasmic incompatibility (CI) [11]. Recent studies revealed that Wolbachia infects approximately 40% of terrestrial arthropod species [12]. Some Wolbachia strains exert a negative effect on the host, e.g., by distorting the sex ratio, whereas others benefit the host by increasing its fecundity [13]; elongating the host’s lifetime [14]; providing nutrition [15]; and resisting viral infection [16]. Conversely, other Wolbachia strains do not discernibly affect the host fitness [17, 18]. Hence, interactions between Wolbachia and its hosts are complicated.

Wolbachia sometimes coexists with other endosymbiotic bacteria, such as Spiroplasma [19] and Cardinium [20]. Multiple Wolbachia infections have been also reported in many insect hosts, on the order level: Coleoptera [21], Hymenoptera [22, 23], Lepidoptera [24], and Diptera [25]. In some cases, the CI intensity induced by Wolbachia decreases as a result of coexistence with Cardinium [26] or other Wolbachia strains [27]. The phenotype caused by the endosymbionts is appreciably correlated with endosymbiont density [27, 28]. On the other hand, the phenotype and Spiroplasma density are not affected by the presence of Wolbachia because the bacterial microhabitats are different [29]. Hence, phenotypes caused by the endosymbiotic bacteria may exhibit various types when multiple infections occur in an individual.

Homona magnanima (Tortricidae, Lepidoptera) is a serious tea pest in East Asia. Two male-killing agents, a presumed RNA virus [10, 30] and Spiroplasma [31], have been identified in this species. Although Wolbachia has also been detected in the field population of H. magnanima [31], its effect on host reproduction or development had not been investigated in detail.

In the current study, we first evaluated the endosymbiont infection status of H. magnanima collected from a tea plantation in Tokyo, Japan, in 1999. We then established three lines singly infected with Wolbachia and one uninfected line in a laboratory-maintained H. magnanima. Finally, we investigated the effect of each Wolbachia strain on the development and CI modification of H. magnanima.

Materials and Methods

Insects

H. magnanima was originally collected in Akiruno city (Tokyo, Japan) in 1999 and is positive for Wolbachia [31]. Insects were continuously reared under laboratory conditions (16L:8D, 25 °C, and 60% relative humidity). All larvae hatched from an egg mass were reared on an artificial diet, SilkMate 2S (Nosan Co., Ltd., Yokohama, Japan) in a plastic container (23 × 16 × 8 cm). For mating, 15 males and 10 females were placed in a plastic box (30 × 20 × 5 cm) [31]. Those rearing and mating treatments were done in each generation.

Detection of Endosymbiotic Bacteria

To detect endosymbiotic bacteria, total DNA was extracted from the abdomen of newly emerged adult females. Endosymbiotic bacteria showed tissue tropism in its host; however, these bacteria are usually located in the host ovary for maternal transmission [29, 32, 33]. As the abdomen of female H. magnanima bears the ovaries, we chose this organ for DNA extraction and for further experiments. First, the female adult abdomen was separated using forceps and placed into a 1.5-ml plastic tube. Each sample was homogenized using a sterilized pestle in 900 μl of cell lysis solution (10 mM Tris-HCl, 100 mM EDTA, and 1% SDS, pH 8.0) and incubated with 1.1 μg/μl proteinase K (Merck, Darmstadt, Germany) at 50 °C for 5 h. The samples were then incubated with 10 μg/μl RNase (Nippongene, Tokyo, Japan) at 37 °C for 1 h. Two hundred microliters of Protein precipitation solution (Qiagen, Hilden, Germany) was added to collect the debris. After centrifugation, 600 μl of the supernatant was transferred to a new tube, and 600 μl of 100% isopropanol was added. Precipitated DNA was washed with 500 μl of 70% ethanol [v/v], and the supernatant was removed by pipetting. After drying, the DNA was suspended in 30 μl of distilled water and stored at 4 or − 35 °C.

To detect three types of endosymbionts (Wolbachia, Spiroplasma, and Rickettsia), DNA samples were amplified using specific primer sets: wsp81F and wsp691R, for Wolbachia [34]; HA-IN and SP-ITS-N, for Spiroplasma [35]; and RpCS.877P and RpCS.1258, for Rickettsia [36]. PCR was conducted using TaKaRa Ex Taq kit (TaKaRa Bio Inc., Shiga, Japan). The reaction mix (total 10 μl) consisted of 1.0 μl of 10X Ex Taq buffer, 0.8 μl of dNTP mixture (2.5 μM in each), 0.05 μl of TaKaRa Ex Taq (5 U/μl), 6.75 μl of Milli-Q water, 0.2 μl of each forward and reverse primer (10 μM), and 1.0 μl of sample DNA (50–100 ng/μl). The amplification reaction parameters for Wolbachia wsp (wsp81F and wsp691R) were as follows: 3 min at 94 °C; followed by 35 cycles of 1 min at 94 °C, 1 min at 60 °C, and 1 min at 72 °C; with a final extension for 7 min at 72 °C. The conditions for Spiroplasma (HA-IN and SP-ITS-N) and Rickettsia (RpCS.877P and RpCS.1258) were as follows: 3 min at 96 °C; followed by 30 cycles of 30 s at 96 °C, 30 s at 52 °C for Spiroplasma and 55 °C for Rickettsia, and 30 s at 72 °C; with a final extension for 7 min at 72 °C. To verify the success of DNA extraction, a housekeeping gene of lepidopteran insects encoding β-actin was amplified as a control, using the primers actin-S and actin-AS [31]. The temperature profile of the amplification reaction was as follows: 3 min at 94 °C; followed by 30 cycles of 30 s at 94 °C, 30 s at 55 °C, and 30 s at 72 °C; with a final extension for 7 min at 72 °C. DNA, which was extracted, as mentioned above, from laboratory colonies of Laodelphax striatellus and Nephotettix cincticeps, was subjected to PCR as a positive control of each symbiont. Laodelphax striatellus was positive for Wolbachia and Spiroplasma [37, 38], and Nephotettix cincticeps was positive for Rickettsia [39]. PCR products were resolved by electrophoresis on a 1.5% (w/v) agarose gel, stained with ethidium bromide, and the bands observed on a transilluminator.

Establishment of Lines Singly Infected with Wolbachia and an Uninfected Line

Host line infected with three Wolbachia strains (wHm-a, wHm-b, and wHm-c) was designated as “Wabc.” To establish the uninfected line (W−) and the three singly infected lines (Wa, Wb, and Wc), the Wabc line was treated with antibiotics, as follows. The W− line was established from Wabc line reared on SilkMate 2S supplemented with tetracycline [0.05% (w/w)] for one generation. The Wa line was established from the Wabc line reared on SilkMate 2S supplemented with tetracycline [0.0125% (w/w)] for one generation. The Wb and Wc lines were established from the Wabc line reared on SilkMate 2S supplemented with rifampicin [0.06% (w/w)] for two generations. The offspring of each antibiotic-treated line was maintained as mentioned above. To confirm the Wolbachia infection status, specific primer sets for wsp of wHm-a, wHm-b, and wHm-c were designed and used for PCR-based detection, as follows: wHm-a_F173 (5′-CCTATAAGAAAGACAATA-3′) and wHm-a_R565 (5′-TTTGATCATTCACAGCGT-3′), for wHm-a; wHm-b_F176 (5′-GGTGCTAAAAAGAAGACTGCGG-3′) and wHm-b_R667 (5′-CCCCCTTGTCTTTGCTTGC-3′), for wHm-b; and wHm-c_F188 (5′-CATATAAATCAGGTAAGGACAAC-3′) and wHm-c_R603 (5′-CACCAGCTTTTGCTTGATA-3′), for wHm-c. The following amplification reaction conditions were employed: 3 min at 94 °C; followed by 35 cycles of 30 s at 94 °C, 30 s at 50 °C, 55 °C, or 60 °C (for wHm-a, wHm-b, and wHm-c, respectively), and 30 s at 72 °C; with a final extension for 7 min at 72 °C.

Sequencing of Wolbachia Genes

To determine the status of Wolbachia infection, DNA of Wabc line was subjected to Wolbachia wsp gene amplification as mentioned above. The PCR product was purified using the Qiaquick PCR purification kit (Qiagen). Purified product was ligated with the pGEM-T easy vector (Promega, Madison, WI, USA) using a ligation mix (TaKaRa). Competent cells (Escherichia coli JM109, TaKaRa) were then transformed with the plasmid. Plasmid DNA was extracted using the Pure Yield Plasmid Miniprep System (Promega, Madison, WI, USA), according to the manufacturer’s protocol, and stored at 4 °C.

DNA extracted from females of each line singly infected with Wolbachia was used as a template for Wolbachia multilocus sequence typing (MLST). MLST genes (gatB, coxA, hcpA, ftsZ, and fbpA) were amplified using specific primer sets, and PCR analysis was conducted as described by Baldo et al. [40]. PCR products were purified using the Qiaquick PCR purification kit and directly sequenced. Sequencing reactions were performed using the BigDye terminator v 3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA, USA). Specific primers for T7 and SP6 promoters were used to amplify the cloned wsp genes, while purified amplimers generated using the MLST-specific primers were sequenced with the primer sets used for PCR. The sequencing was performed in both, forward and reverse direction. Products of each reaction were pelleted with a mixture of 30 μl of 99% ethanol and 2.5 μl of 125 mM EDTA and washed with 70% ethanol. Purified products were suspended in 10 μl of Hi-Di™ formamide (Applied Biosystems) and incubated for 2 min at 95 °C. The sequencing was performed using the 3100 Genetic Analyzer (Applied Biosystems).

Transmission Rate of the Individual Wolbachia Strains

One female from the Wabc line and one male from the W− line were mated in a plastic box (15 × 15 × 7 cm). The crossing was replicated six times; ten larvae were chosen randomly from each egg mass and reared individually until eclosion. Wolbachia infection status was checked in F1 adults by PCR, as described above, using strain-specific primer sets. The transmission rate of each Wolbachia strain was calculated as the number of F1 adults infected with Wolbachia divided by the number of adults in each crossing experiment.

The Effects of Wolbachia on Host Development and Reproduction

To investigate the effects of Wolbachia infection on the host, larvae of each host line were individually reared on an artificial diet, INSECTA LF (Nosan Co., Ltd.). The sex ratio, larval development time, and pupal weight were recorded.

Crossing experiments were performed to evaluate the CI modification and rescue of each Wolbachia strain. Three males (0 ± 1 day post eclosion) from the Wolbachia-infected lines (Wabc, Wa, Wb, and Wc) were mated with three virgin females from the W− line in a plastic box (15 × 15 × 7 cm). Each crossing experiment was replicated four times with different generations. The W− males and W− females were also mated, as a control. Females were allowed to oviposit for 7 days. Since low hatchability was observed in the crosses of Wabc × W− and Wb × W− lines, to characterize the phenotypes, males and females of the Wabc and Wb lines were mated with each other.

The hatchability of F1 generation from each crossing experiment described above was calculated. It was defined as the number of late-stage embryos per number of eggs in an egg mass, because the number of late-stage embryos was almost the same as the number of larvae that hatched successfully. First, to analyze the correlation between the egg mass area and the number of eggs, 50 egg masses from the control crossing (W− × W−), which showed high hatchability, were chosen randomly. Each egg mass area was determined using Image J (https://imagej.nih.gov/ij/) and the eggs were counted under a microscope. Both parameters were analyzed by JMP v 9 (SAS, Cary, NC, USA) using a general linear model (see below) to obtain a regression line. Using this line, the number of eggs in an egg mass from each crossing experiment was estimated from the egg mass area. In addition, late-stage embryos in each egg mass were counted. Finally, the hatchability of each egg mass was calculated as described above.

Determination of Wolbachia Density

To evaluate Wolbachia density, five newly emerged males and females from each line were chosen for qPCR determinations. The qPCR primer set was designed based on the wsp sequence, to amplify the wHm-wsp universal region of about 100 bp: wHm-uni_qpcrF, 5′-TGGTGTTGGTGCAGCGTAT-3′; and wHm-uni_qpcrR, 5′-AACTAACACCAGCTTTTGCTTGA-3′. The qPCR reaction was performed as described by Iwata et al. [41]. The reaction mixture contained 10 ng of DNA, 30 μM of each primer, and 5 μl of the FastStart Universal SYBR Green master mix (Roche, Basel, Switzerland). The reactions were performed using the StepOnePlus real-time PCR system (Life Technologies, Carlsbad, CA, USA). The cycle conditions were as follows: 10 min at 95 °C; followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Dissociation curve analysis of the amplified product was performed after the amplification, as follows: 15 s at 95 °C, 1 min at 60 °C, and 15 s at 95 °C. The Ct value of each sample DNA was determined twice as described by Iwata et al. [41]. The quantity of Wolbachia in each sample was calculated based on a standard curve generated using 10−4–10 ng of DNA of a plasmid harboring a cloned partial wsp sequence. The number of wsp copies in 10 ng of sample DNA was estimated from the molecular weight of the wsp sequence.

Data Analysis and Statistics

The obtained sequences were aligned with sequences obtained by Baldo et al. [40] using ClustalW [42], except for Trichogramma deion, which lacks wsp sequences. For phylogenetic tree construction, the maximum likelihood method with bootstrap re-sampling of 1000 replications was performed in MEGA6 [43]. Correlation between the size and number of eggs in an egg mass was determined by regression analysis using a general linear model. χ2 test was used to confirm whether the sex ratio of each host line was biased. The hatchability, larval development time, pupal weights, and Wolbachia densities in each host line were analyzed using the Steel-Dwass test. All statistical analyses were performed using JMP v9. Sequences of Wolbachia were deposited in GenBank under accession numbers LC363921 to LC363938.

Results

Detection of Endosymbionts and the Transmission and Phylogeny of Wolbachia

Laboratory-maintained H. magnanima (Tokyo population; Wabc) was positive for Wolbachia, but negative for Spiroplasma and Rickettsia. The 46 clones of wsp fragments generated from Wabc line could be divided into three sequence types: wsp-a (25 clones), wsp-b (12 clones), and wsp-c (9 clones), indicating a multiple infection of H. magnanima with three Wolbachia strains (wHm-a, wHm-b, and wHm-c). The transmission rate of wHm-a and wHm-c to F1 generation was 100.0% (n = 60/60), while that of wHm-b was 90.0% (n = 54/60, Table 1).

Table 1 Transmission rate of each Wolbachia strain
Full size table

MLST sequence data also supported the presence of three distinct strains (Table 2). Both wHm-a and wHm-b belonged to supergroup A, and wHm-c belonged to supergroup B (Fig. 1).

Table 2 wsp and MLST gene typing of wHm-a, wHm-b, and wHm-c strains
Full size table
Fig. 1
figure 1

Phylogeny of Wolbachia based on wsp and MLST gene sequences. Phylogenetic analysis was conducted using H. magnanima infected with Wolbachia and 36 strains from Baldo et al. [40], after exclusion of T. deion, using maximum likelihood method based on the Tamura-Nei model. Bootstrap values exceeding 50% are shown (1000 replicates)

Full size image

Establishment of the Three Lines Infected with a Single Wolbachia Strain

The Wabc line was firstly treated with tetracycline [0.0125% (w/w)]. The randomly chosen eight individuals of the treated line (Gt1) showed only multiple infection, as revealed by diagnostic PCR experiment: triple infection (6/8 in number of individuals) and double infection with wHm-b and -c (2/8). In Gt2 generation, the two Gt1 triply infected lines were reared on SilkMate without antibiotics. The randomly chosen four individuals in each Gt2 line showed wHm-a single infection (2/4 and 3/4, respectively) and no infection (2/4 and 1/4 in each). The offspring of the Gt2 wHm-a singly infected line showed stable infection after the third generation (Gt3); therefore, we chose one line as a wHm-a singly infected line for further experiments (Fig. 2a).

Fig. 2
figure 2

Processes for segregation and the infection status of Wolbachia. Wolbachia infection was confirmed by strain-specific PCR. Three alphabets: a, b, and c in the figure indicate wHm-a, -b, and -c, respectively. a The process of the establishment of Wa line. b The process of the establishment of Wb and Wc lines

Full size image

Secondly, the Wabc line was treated with rifampicin [0.06% (w/w)]. The randomly chosen nine individuals of the treated line (Gr1) showed the following infection statuses, as revealed by diagnostic PCR experiments: triple infection (2/9 in number of individuals); double infection with wHm-a and -c (3/9), wHm-a and -b (1/9), and wHm-b and -c (1/9); and single infection with wHm-a (1/9) and wHm-c (1/9). In Gr2 generation, the two lines of Gt1 (doubly infected line with wHm-a and -b and singly infected line with wHm-c) were re-treated with rifampicin [0.06% (w/w)]. The randomly chosen six individuals of the offspring of doubly infected line with wHm-a and -b showed the following three infection statuses: single infection with wHm-b (1/6) and wHm-a (2/6) and no infection (3/6). The randomly chosen six individuals of the offspring of singly infected line with wHm-c showed stable infection only with wHm-c in Gr2 generation (6/6). In Gr3, the singly infected line with wHm-b and -c was reared on SilkMate without antibiotics and showed stable infection only with both Wolbachia strains. Thus, we determined the singly infected line with wHm-b and wHm-c as the Wb and Wc lines, respectively, for further experiments (Fig. 2b).

The Effect of Wolbachia on Host Sex Ratio and Development

The percentages of females in the Wabc, Wa, Wb, Wc, and W− lines were determined, with no significant difference between the lines (χ2 = 6.704, p = 0.152, Table 3).

Table 3 Larval development time, pupal weight, and the percentage of females in host lines singly infected (Wa, Wb, and Wc), multiply infected (Wabc), and uninfected (W−) with Wolbachia
Full size table

The mean development time of female larvae from the Wabc line (19.5 ± 0.17 days, mean ± SD) was significantly shorter than those from the W− line (21.5 ± 0.67 days, Steel-Dwass test, Z = 2.79304, p < 0.05), the Wa line (21.5 ± 0.57 days, Z = − 3.12190, p < 0.05), and the Wb line (22.5 ± 0.62 days, Z = − 4.19888, p < 0.01), but was not significantly different from those from the Wc line (19.4 ± 0.19 days, Z = 0.28314, p = 0.9986). The mean development time of female larvae from the Wc line was also significantly shorter than those from the W− line (Z = 2.76255, p < 0.05), the Wa line (Z = − 3.07285, p < 0.05), and the Wb line (Z = − 4.19888, p < 0.01). No significant difference in the mean development time of male larvae was observed between the host lines (p > 0.05).

The mean female pupal weights in the Wabc line (94.5 ± 1.50 mg) and the Wc line (94.0 ± 1.9 mg) were significantly higher than those in the W− line (86.1 ± 2.10 mg; Steel-Dwass test: Z = − 2.89617, p < 0.05 for the Wabc line comparison; Z = − 3.62427, p < 0.01 for the Wc line comparison) but were not significantly different from those in the Wa line (85.8 ± 2.76 mg, Z = − 0.23769, p = 0.9993) and the Wb line (87.6 ± 3.71 mg, Z = − 0.74415, p = 0.9461). No significant differences in the mean male pupal weights were observed between the host lines (p > 0.05).

The Effect of Wolbachia Infection on Host Compatibility and the CI Strength

A positive significant relationship between egg mass size (x; area, mm2) and the number of eggs (y) per egg mass (general linear model, p < 0.01, r2 = 0.9704, y = 4.7974x + 2.1326) was observed. The mean hatchability for the Wabc male × W− female crossing pair (56.8 ± 3.0%) was significantly lower than that for W− male × W− female (77.5 ± 2.3%, Z = − 4.49572, p < 0.01), Wa male × W− female (76.0 ± 2.0%, Z = − 0.33743, p < 0.01), and Wc male × W− female pairs (76.7 ± 2.5%, Z = 0.17050, p < 0.01). The mean hatchability for the Wb male × W− female crossing pair (31.9 ± 2.7%) was also significantly lower than that for the W− male × W− female (Z = − 9.03885, p < 0.01) and Wabc male × W− female pairs (Z = − 5.58632, p < 0.01, Fig. 3).

Fig. 3
figure 3

Mating experiments with H. magnanima lines, uninfected or infected with Wolbachia. The center line within the box (a border between dark gray and white) represents the median. The upper and lower boundaries of the box indicate upper quartile and lower quartile, respectively. Sample size in each data point is indicated in parentheses below the plots. Different letters indicate significant differences between groups (Steel-Dwass test, p < 0.05)

Full size image

These results indicated that wHm-b was involved in CI and hence, the hatchability for mating between the Wabc male × Wabc female and Wb male × Wb female pairs was examined to confirm whether CI would be rescued. The hatchability for the Wabc male × Wabc female pair (72.7 ± 1.6%) was significantly higher than that for the Wabc male × W− female pair (Z = − 4.49572, p < 0.01); that for the Wb male × Wb female pair (70.5 ± 1.6%) was also significantly higher than that for the Wb male × W− female pair (Z = − 8.62484, p < 0.01). These results indicated that the CI-inducing factor was wHm-b.

Wolbachia Density in Each Host Line

No significant difference in wsp copy number per 10 ng of DNA was apparent between males and females from each line (p > 0.05, data not shown). The wsp copy number in the Wa line was significantly higher than that in the Wb (Z = − 4.79807, p < 0.01), Wc (Z = − 5.48650, p < 0.01), and Wabc lines (Z = − 5.54392, p < 0.01). Wolbachia density in the Wb line was also significantly higher than that in the Wc (Z = − 4.72656, p < 0.01) and Wabc (Z = − 4.60516, p < 0.01) lines, whereas no significant difference in the wsp copy number was observed between the Wc and Wabc lines (Z = 1.79579, p = 0.2753, Fig. 4).

Fig. 4
figure 4

Wolbachia density in each Wolbachia-infected host line. Different letters indicate significant differences between groups (Steel-Dwass test, p < 0.05)

Full size image

Discussion

In the current study, we demonstrated multiple Wolbachia infections in H. magnanima, their phenotypes, and high transmissibility to the host’s offspring. From the three investigated Wolbachia strains, wHm-b differently induced CI in the host, depending on the Wolbachia infection status, while wHm-c increased host pupal weight and shortened the larval development time. No distinct phenotype was observed for the wHm-a infection.

Wolbachia is frequently found in various arthropod species, affecting the host in various ways [12]. CI is one of the most representative reproductive manipulations of Wolbachia [11]. Since CI-inducing Wolbachia kills the host’s offspring when an infected male mates with uninfected females, the phenomenon is directly correlated with a rapid increase in the prevalence of CI-inducing Wolbachia [44]. Although the reported intensities of CI caused by Wolbachia or other endosymbionts are different in different studies, some authors have shown that Wolbachia density in the host is positively correlated with the intensity of CI [27, 28]. In the current study, the intensity of CI and Wolbachia density in the Wb line were significantly higher than in the Wabc line, suggesting that the density of wHm-b indeed determined the intensity of CI.

Wolbachia may also be beneficial for its host [13,14,15,16]. For example, one Wolbachia strain reportedly contributes to iron metabolism in Drosophila [14, 15]. In the current study, wHm-c infection correlated with the host development in terms of host pupal weight and development time. Previous studies of Lepidoptera indicated that the weight of female pupae is positively correlated with the oviposition period, the number of eggs laid, and the longevity of the adult [45,46,47]. Thus, the increase of pupal weight associated with the wHm-c infection was beneficial for the survival and reproduction of the host. In terms of the host development, host death is a crucial problem for Wolbachia. Variety of parasitoids and insect pathogenic viruses has been isolated from H. magnanima, characterized, and shown to act as natural enemies [48,49,50]. Since virus infection is mainly restricted to the host larval stages, short larval development time associated with a wHm-c infection may contribute to a lower risk of viral infection. Unlike wHm-b, no differences in the phenotype caused by wHm-c were apparent when singly and multiply infected lines were compared. The density of Wolbachia in the Wabc line was significantly lower than in the Wb line but did not differ from the Wc line, indicating that the phenotype of wHm-c was less dependent on bacterial density than that of wHm-b. In Drosophila, Wolbachia and Spiroplasma exhibit their own specific localization patterns or host organ specificity [29, 51], and Spiroplasma also restricts Wolbachia density in a co-infected tissue [29]. It is likely that the microhabitats of Wolbachia in H. magnanima or Wolbachia-Wolbachia interactions affect bacterial density and the associated phenotypes.

Infections with some Wolbachia strains were shown to be associated with no cost or benefit to the host [17, 18], despite successfully invading the host population [52]. Although wHm-a infection was not associated with any apparent cost or benefit to H. magnanima in the current study, this strain was highly transmissible to the host’s offspring. Another curious observation for wHm-a was that in the Wa line, Wolbachia density was significantly higher than in any other host line. Previous studies demonstrated that Wolbachia density depends on the host’s genetic background [28, 53, 54]. In the current study, all host lines used in the experiments were established from the Wabc line, and their genetic background may be similar to the Wabc line. Thus, it is reasonable to assume that the difference in strain densities depends on the characteristics of each Wolbachia strain. Although wHm-a did not show any distinct phenotype in the current study, the existence of wHm-a reduced the density of wHm-b during a multiple infection, which could remedy the death of offspring caused by wHm-b. From that perspective, wHm-a may indirectly contribute to the host’s fitness. In the current study, we examined the Wolbachia density and its phenotype using only laboratory-maintained colonies. As the density of Wolbachia can be altered depending on field conditions or seasons [55], it is also worth studying the prevalence, phenotype, and density of Wolbachia in field populations of H. magnanima for understanding the dynamics of Wolbachia-host interactions.

Previous studies revealed that multiple Wolbachia infections are highly prevalent in several insects in nature [21,22,23,24,25]. Interestingly, these multiply infected hosts often harbor one or more CI-inducing Wolbachia strains [21, 24, 27]. For example, Eurema hecabe harbors CI-inducing Wolbachia (wCI) and feminization-inducing Wolbachia (wFem), and wFem is only detected together with wCI [24]. Since it is well-known that CI-inducing Wolbachia strains are maintained and spread rapidly in wild host populations [44], CI-inducing Wolbachia is thought to a play crucial role in the establishment of a multiple infection. Considering the induction of CI by wHm-b and data from studies mentioned above, it is reasonable to propose that the co-infection of wHm-a and wHm-c with wHm-b is highly prevalent. For the host, such multiple infections might be beneficial not only in terms of a decreased offspring mortality compared with a single wHm-b infection, but also in terms of reproductive and developmental benefits associated with wHm-c infection. The three strains of Wolbachia cannot co-infect their host with such high density as during a single infection, although they continue to invade the host population on account of their high transmissibility. Hence, infections with multiple endosymbionts are thought to constitute a compromise between selfish microorganisms and their host.

In the current study, we characterized three Wolbachia strains isolated from H. magnanima found in Tokyo. We also demonstrated that none of the three strains caused sex ratio distortion in H. magnanima. Interestingly, a presumed RNA virus and Spiroplasma have both been identified as the male-killing agents in Ibaraki and Shizuoka H. magnanima populations, respectively [10, 30, 31]. Although Wolbachia is also well-known as a sex ratio-distorting factor [56], wHm-a, wHm-b, and wHm-c do not appear to be involved in host sex abnormalities in Tokyo populations of H. magnanima. Previous studies indicated that the phenotypes caused by Wolbachia strains depend on the genetic background of the host [57, 58]. Hence, further studies of the phenotypes and prevalence of Wolbachia in various field populations may contribute to the understanding of Wolbachia strain dynamics in fields. In addition, the presented H. magnanima-symbiont system will likely provide useful insights to facilitate the understanding of the host-endosymbiont-endosymbiont interactions.