To construct S. ixodetis sHm genome, we used the laboratory-maintained Spiroplasma-positive MK-inducing line (S+ line) of H. magnanima (Tsugeno et al., 2017). In the present study, we accidentally obtained a Spiroplasma-positive 1:1 sex ratio line (S+M+ line) as a subline of the S+ line. For every generation, the male moths picked up from the 1:1 sex ratio line, which had been confirmed negative for Spiroplasma, Wolbachia, and OGV (NSR line) (Takamatsu et al., 2021), were crossed with the female moths of the S+ and S+M+ lines as described by Arai et al. (2022a). The obtained larvae were reared using artificial diet SilkMate 2S (Nosan Co., Yokohama, Japan) at 25°C under a long photoperiod (16L:8D), i.e., till pupation. To eliminate Spiroplasma from the S+ line, the first instar larvae were reared with SilkMate 2S supplemented with 0.05% tetracycline (w/w) as described by Arai et al. (2019). Adult moths were sexed based on their morphology, and the hatched larvae and the unhatched pharate larvae (mature embryo) were sexed based on the presence or absence of the female-specific sex chromatin body (a condensed W chromosome), which was detected via lactic-acetic orcein staining (Arai et al., 2022a).

Total DNA was extracted from the abdomen of female adults (0-day post eclosion), the whole body of larvae and pupae (0-day post molting), and dissected tissues of H. magnanima larvae (0-day post molting) using cell lysis buffer, as described by Arai et al. (2019). To detect Spiroplasma, a pair of Spiroplasma-specific primers was used to amplify RNA polymerase β subunit gene (RpoB), which is a single copy conserved gene in Spiroplasma spp., from the extracted DNA (adjusted to 50–100 ng/reaction) with EmeraldAmp MAX PCR Master Mix (TaKaRa Bio, Shiga, Japan); the primer sets are listed in Table 1. The PCR conditions were as follows: 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s, followed by 72°C for 7 min. β-Actin gene of H. magnanima was used as the control. To quantify Spiroplasma density, qPCR was performed using the extracted DNA, which was diluted to a concentration of 10 ng/μL with MilliQ water, Spiroplasma RpoB primers (Table 1), and KOD SYBR^® qPCR Mix (Toyobo, Osaka, Japan) in a LightCycler^® 96 system (Roche, Basel, Switzerland). The PCR consisted of 45 cycles of 98°C for 10 s, 60°C for 10 s, and 68°C for 30 s. Relative abundance of the gene was calculated using the expression of elongation factor 1a gene (ef1a) of H. magnanima as the control. Spiroplasma density (RpoB copies) and relative abundance (RpoB/ef1a) were calculated as described in Arai et al. (2019, 2022c).

For genome sequencing of strain sHm, high molecular weight DNA was extracted from the egg masses of S+ line moths using Nanobind Tissue Big DNA Kit (Circulomics Inc., MD, USA) and used for library construction using Ultra-Long DNA Sequencing Kit (Oxford Nanopore Technologies, Oxford, UK) following the manufacture’s protocol. The constructed libraries were sequenced using ONT MinION flow cell (R 9.4.1) (Oxford Nanopore Technologies). The obtained reads were mapped to the H. magnanima reference genome (Jouraku et al., in preparation) with minimap2 (Li, 2018), and the non-mapped reads containing Spiroplasma reads were extracted with SAMtools v.1.9 (Li et al., 2009) and assembled using Canu 1.6 (Koren et al., 2017). The draft Spiroplasma genome (a circular main chromosome and plasmids) was annotated via BLASTn (NCBI nr database). The extracted DNA was also subjected to Illumina paired-end 150 bp sequencing (PE-150) at Novogene (Beijing, China). The Illumina data were used to polish the draft genome using minimap2 (Li, 2018) and Pilon v. 1.23 (Walker et al., 2014). Since no sequence changes were observed after the second polishing, the polished genome was considered as the complete genome of strain sHm. The circularity of the sHm genome was confirmed by BLASTn search, followed by manual deletion of overlapping sequence.

S+ (MK line) and S+M+ H. magnanima lines (non-MK line) were used for DNA extraction as described by Arai et al. (2022a). The DNA extracted from Spiroplasma cells was amplified using whole genome amplification (WGA) by REPLI-g Mini Kit (Qiagen, Hilden, Germany), following the manufacture’s protocol. The WGA products, purified using AMPure XP beads (Beckman Coulter, Inc., CA, USA) and dissolved into TE buffer, were sequenced on Illumina platform (PE-150). The Illumina data assembled with unicycler (Wick et al., 2017) and Illumina raw read data were mapped to the sHm reference genome using minimap2 (Li, 2018) to detect the genomic changes in the genome of sHm in the S + M + line.

The constructed sHm genome was annotated via DFAST (Tanizawa et al., 2018). Effector genes were further annotated using EffectiveDB (Eichinger et al., 2016). Functional analysis of proteins (i.e., domain predictions and Gene ontology annotations) was conducted using InterPro.¹ Phage WO infections were annotated using PHASTER (Arndt et al., 2016). Protein homology between different Spiroplasma strains was analyzed using S. apis B31 (CP006682.1), S. citri strain BLH-MB (CP047437.1–CP047446.1), S. syrphidicola strain EA-1 (NC_021284.1), D. melanogaster endosymbiont S. poulsonii MSRO (CM020866.1–CM020867.1) (sMel, MK strain, Masson et al., 2018), Danaus chrysippus (Nymphalidae) endosymbiont S. ixodetis (NZ_CADDIL010000001.1–NZ_CADDIL010000012.1) (sDa, MK strain, Martin et al., 2020), Lariophagus distinguendus (Pteromalidae) endosymbiont S. ixodetis (NZ_JALMUW010000001.1–NZ_JALMUW010000198.1) [sDis, cytoplasmic incompatibility (CI) strain, Pollmann et al., 2022], and Dactylopius coccus (Dactylopiidae) endosymbiont S. ixodetis (JACSER010000001.1–JACSER010000358.1) (sCoc, non-MK strain, Vera-Ponce León et al., 2021) with OrthoVenn2.² Homology of sHm genes and proteins with spaid from strain sMel (Harumoto and Lemaitre, 2018), ankyrin genes from S. ixodetis (Yeoman et al., 2019; Martin et al., 2020; Vera-Ponce León et al., 2021), and the Wolbachia MK candidate factor responsible for WO-mediated killing (Wmk, presumed helix-turn-helix transcriptional regulator, Perlmutter et al., 2019; Arai et al., 2022b) was evaluated using both BLASTn and BLASTp. Moreover, to verify whether MK microbes of H. magnanima carried conserved genes, the genes on the MK-associated prophage region WOwHm-t76 of MK Wolbachia wHm-t (Arai et al., 2022b) and those of the Partiti-like virus OGVs (Fujita et al., 2021) were compared to the sHm genes using both BLASTn and BLASTp. Unique genomic features of the sHm strain in the MK S+ and non-MK S+M+ H. magnanima lines were analyzed using GView³ and BV-BRC variation analysis service.⁴ Metabolic pathways of S. poulsonii sMel and S. ixodetis sHm were compared by using BV-BRC comparative analysis service (see text footnote 4). Phylogenetic trees of 16S rRNA gene and ankyrin genes of Spiroplasma strains were constructed by maximum likelihood with bootstrap re-sampling of 1,000 replicates using MEGA7 (Kumar et al., 2016). Mycoplasma genitalium G-37 (NR074611.1) was used as an outgroup.

A fifth instar female larva was sterilized in 50% bleach (ca. 3% sodium hypochlorite) for 10 min, in 70% ethanol for 10 min, and dissected in IPL-41 Insect Medium (Gibco, Carlsbad, CA, USA) supplemented with 10% (v/v) fetal bovine serum (FBS). Malpighian tubules of the dissected larva were transferred to flasks containing either the Bombyx mori NIAS-Bm-aff3 (aff3) cell line (Takahashi et al., 2006) or the Aedes albopictus NIAS-AeAl-2 (AeAl2) cell line (Mitsuhashi, 1981) maintained in IPL-41 Insect Medium (Gibco) with 10% (v/v) of FBS. The cells and Malpighian tubules were co-cultured at 23°C. Fresh medium was supplied to the flask every 10 d. Purified cells centrifuged at 1,000 g for 2 min were used to analyze infections and titers of the transinfected strain sHm in the cells. DNA extraction, PCR, and qPCR assays were performed as mentioned in section “Spiroplasma detection and quantification in Homona magnanima.”

Sex ratio bias was assessed using Fisher’s exact test. Spiroplasma densities, male ratio in hatched larvae, and male ratio in unhatched pharate larvae were analyzed using either the Wilcoxon test or the Steel--Dwass test. All analyses were performed using R software v4.0⁵.

The S+ line moths harboring sHm exhibited lower egg-hatching rates than the NSR line (Figure 1A), which is consistent with the results from previous studies that Spiroplasma infection halved the egg hatching rates of H. magnanima (Tsugeno et al., 2017; Takamatsu et al., 2021). Cytogenetic sexing based on the presence or absence of a sex chromatin body (W chromosome) revealed that the sex ratio of hatched larvae was strongly biased toward females in the S+ line moths but not in the NSR line moths (P < 0.01, Figure 1B). In contrast, the sex ratio of unhatched pharate larvae (late-stage embryos) were male-biased in the sHm-infected line (S+) (Fisher’s exact test, P < 0.01), confirming that sHm killed male H. magnanima during embryogenesis. Moreover, the elimination of Spiroplasma by tetracycline treatment resulted in non-biased sex ratios in the subsequent generation (P < 0.01, Figure 1C).

Both Illumina (816.3 Mb, 5,442,459 reads, and 150 bp average length) and Nanopore data (93.5 Mb, 25,820 reads, and 3,624 bp average length) were used to reconstruct a complete genome consisting of a circular main chromosome (2, 102,039 bp in length) and two circular potential plasmids [20,119 bp (pSHM_1) and 16,408 bp (pSHM_2)]. Previously, Tsugeno et al. (2017) reported two 16S rRNA gene variants cloned from Spiroplasma-infected H. magnanima, but they did not elucidate whether the two sequences were interoperonic polymorphs of a single isolate or they were derived from two different strains. The present study confirmed that H. magnanima was infected with the MK S. ixodetis strain sHm that harbored two distinct 16S rRNA gene sequences in its genome (Figure 2). Moreover, Spiroplasma strains often encode multiple ribosomal RNA gene sets in their genome (Chang et al., 2014; Tsai et al., 2018; Vera-Ponce León et al., 2021).

Strain sHm harbored a higher number of coding sequences (CDS; 2,886 CDS) than strains sMel (1.9 Mb in genome size; 2,405 CDS; Masson et al., 2018) and sDa (1.7 Mb in genome size; 1,813 CDS; Martin et al., 2020; Table 2). Although plasmids often contain key accessory genes such as spaid of sMel (Harumoto and Lemaitre, 2018), genes on pSHM_1 (n = 24) and pSHM_2 (n = 20) mostly encoded hypothetical or uncharacterized proteins (Supplementary Table 1). In addition, strain sHm harbored 12 prophage regions (size: 6.6–14.8 kb) in its genome, which is consistent with previous reports that several phage sequences are found in Spiroplasma genomes (Bai and Hogenhout, 2002; Lo et al., 2013; Ramirez et al., 2021). Bacteriophages frequently carry virulence-associated genes that encode toxins (Waldor and Mekalanos, 1996; Brüssow et al., 2004). Recently, the mechanistic bases of Wolbachia-induced cytoplasmic incompatibility (CI) and MK have been attributed to phages (Beckmann et al., 2017; LePage et al., 2017; Perlmutter et al., 2019; Arai et al., 2022b). Besides, phages have also been implicated in the defense phenotype exhibited by bacteria against parasitoids, such as Hamiltonella defensa (Brandt et al., 2017). Therefore, it is possible that the phages of sHm contribute to the manifestation of MK phenotype or confer fitness advantage on hosts by protecting the hosts from natural enemies.

Recently, Yeoman et al. (2019) and Vera-Ponce León et al. (2021) reported that D. coccus-infecting S. ixodetis (sCoc) harbored a spaid homolog and Cephus cinctus (Cephidae)-infecting S. ixodetis harbored seven spaid homologs. Our BLAST searches confirmed that sHm did not harbor the spaid gene (Table 3), however, some of the ankyrin genes of sHm were homologous to the alleged gene sequences of sCoc and Cephidae-infecting S. ixodetis (Table 4). It is likely that S. ixodetis do not harbor the spaid gene. The superficial homology could be due to the presence of conserved ankyrin repeats (Table 3). Similarly, the amino acid sequences of ankyrin proteins of sHm (such as SHM_18920) showed partial homology to the ankyrin domain of Spaid from strain sMel (N-terminal 200 amino acids) as per BLASTp search, but the complete amino acid sequences of the proteins of these two strains were not homologous (Table 3). Moreover, we also confirmed the absence of Spaid homologs in a MK S. ixodetis strain sDa by using BLASTp search. Gerth et al. (2021) reported that the Spaid homologs are conserved among S. poulsonii strains regardless of the MK phenotype. Because the spaid gene is not likely to be possessed by S. ixodetis, MK mechanisms may differ between S. poulsonii and S. ixodetis (i.e., having different causative genes).

We then focused on genes conserved among Spiroplasma strains. Distantly related Spiroplasma species such as S. ixodetis (sHm), S. poulsonii (sMel), S. apis B31, S. citri BLH-MB, and S. syrphidicola EA-1 shared 345 protein clusters (Figure 3A). For S. ixodetis strains, two MK strains (sHm and sDa) and two non-MK strains (sCoc and sDis) shared 595 protein clusters (Figure 3A). In addition, MK strains sHm and sDa possessed additional 219 conserved protein clusters. sHm also harbored strain-specific 77 protein clusters (470 genes) associated with metabolism and transposition (Figure 3B and Supplementary Table 1) as well as many putative Type IV secretory system effector genes (n = 144, based on T4SEpre prediction at EffectiveDB, Supplementary Table 1), some of which were located in the prophage regions. In Spiroplasma, RIP toxin irreversibly inactivates eukaryotic cytosolic ribosomes (Hamilton et al., 2016; Ballinger et al., 2019; Garcia-Arraez et al., 2019). Based on our blast searches, RIP-4 encoded by Spiroplasma endosymbiont of Drosophila neotestacea (ASM46790.1) showed low homology to SHM_22560 (79–286 aa, e-value 6.7E-13, bit-score 60.8). Besides, SHM_22560 (hypothetical protein, 788 aa, Supplementary Table 1) was predicted to contain a RIP domain based on Interpro (hit: IPR016138, aligned length: 111–286 aa) and HHpred searches [hit: Sapolin (ID: 3HIQ), aligned length: 105–344 aa, e-value: 2.3E-29]. A homolog of an epsilon-like toxin (WP_252319264.1_36) encoded by CI-inducing sDis was detected in the sHm genome (SHM_25300, Table 4), while AbiEii abortive infection toxin (WP_252320055.1_19) and OTU-like cysteine protease (WP_252320277.1_1) were not detected.

Although spaid gene is the only ankyrin-coding gene identified in the genome of strain sMel (Harumoto and Lemaitre, 2018), sHm carried 10 ankyrin genes (Figure 3C). Some ankyrin genes harbored additional domains, such as those encoding for DnaJ (SHM_12110), cyclin-dependent kinase inhibitor (SHM_12110 and SHM_28080), and the protein ubiquitination associated FEM1A_DROME (SHM_21210); however, they did not encode for signal peptides, ovarian tumor-like deubiquitinase (OTU), or helix domains found in the spaid gene of strain sMe1 (Paredes et al., 2015; Harumoto and Lemaitre, 2018; Masson et al., 2018; Gerth et al., 2021). Wolbachia induces CI by the CI-inducing factors (Cif) harboring ankyrin repeats in insects (LePage et al., 2017). Pollmann et al. (2022) reported that CI-inducing sDis strain did not harbor the cif gene. Similarly, sHm-encoding 10 ankyrin genes has low homologies to those of other bacteria such as Wolbachia and Rickettsia and were not homologous to the cif as well as spaid genes. Intriguingly, MK sDa and CI sDis strains had 11 and 12 ankyrin genes, respectively. These findings suggest that S. ixodetis has similar characteristics to Wolbachia endosymbionts (Duplouy et al., 2013; Arai et al., 2022b) in terms of phenotypes (i.e., CI and MK) and genetic compositions (i.e., multiple ankyrin genes). Some ankyrin genes encoded by S. poulsonii and Wolbachia have been implicated in reproductive manipulation (LePage et al., 2017; Harumoto and Lemaitre, 2018), and the ankyrin genes found in the sHm genome may also be involved in MK mechanisms.

Homona magnanima harbors three different types of MK endosymbionts (i.e., Spiroplasma sHm, Partiti-like virus OGVs, and Wolbachia wHm-t strain), some of which can coinfect the same host (Arai et al., 2020; Takamatsu et al., 2021). Moreover, microbes sharing the same niche frequently exchange virulence-associated genes (Kent and Bordenstein, 2010; Wiedenbeck and Cohan, 2011). However, we found that strain sHm did not harbor any gene homologous to those of MK Partiti-like virus OGVs (Fujita et al., 2021). Moreover, strain sHm did not harbor wmk or effector genes (e.g., CifB-like) that are present on the MK-associated prophage WOwHm-t76 region of strain wHm-t (Arai et al., 2020, 2022b). The wmk gene, a candidate gene for Wolbachia-induced MK (Perlmutter et al., 2019, 2021; Arai et al., 2022b), possesses a helix-turn-helix (HTH) domain containing putative transcriptional regulator. Although no wmk homologs were identified, strain sHm harbored 87 HTH domain-encoding genes, namely putative transposase (classified into IS-30, IS-3, and IS-5 type transposase, n = 83), a type II toxin-antitoxin system antitoxin HipB (SHM_ 03650), an AAA family ATPase (SHM_24830), an XRE family transcriptional regulator (SHM_17560), and a helix-turn-helix transcriptional regulator (SHM_05440). Notably, a putative transposase SHM_03660, encoded by a gene adjacent to sHm-specific HipB-like SHM_3650, was homologous to the Wolbachia transcriptional regulator. Recently, Arai et al. (2022c) demonstrated that strains sHm, wHm-t, and OGVs affect H. magnanima males in different manners. Specifically, both strains sHm and wHm-t trigger abnormal apoptosis and interfere with sex determination in male embryos (manifested by the alteration of doublesex gene splicing), but only strain wHm-t impairs the dosage-compensation system of the host (manifested by the alteration of the global gene expression on sex chromosomes). In contrast, the OGVs do not affect sex-determination cascades or dosage-compensation systems. These findings and our current results support the view that phylogenetically distinct microbes have independently developed different MK machinery even for the same host, i.e., H. magnanima. Therefore, an unknown factor in the sHm genome may be responsible for the embryonic male death of H. magnanima.

We observed that one of the sublines of the MK S+ line ceased to induce MK (Figure 4A). This subline, referred to as the S+M+ line, exhibited stable sHm infections for at least four generations. We simultaneously re-sequenced the genome of strain sHm from S+M+ and S+ lines at the second-generation stage since their divergence. We previously demonstrated from a genomic comparison of MK Wolbachia (wHm-t) and non-MK Wolbachia (wHm-c) that an MK-associated 76 kb prophage region was inserted only in wHm-t (Arai et al., 2022b). Similarly, we mapped the MK and non-MK sHm re-sequenced Illumina reads to the complete sHm genome (main chromosome and two plasmids) but did not detect any large-scale structural variation (insertions or deletions) as observed in wHm-t (Figure 4B). On the other hand, we found mutations specific to the non-MK sHm mutant (i.e., frameshifts or insertion of stop codons) in 21 genes encoding hypothetical proteins (n = 4), tyrosine-tRNA ligase (n = 1), and transposase (n = 16) (Supplementary Table 2). The 21 genes were found on the main chromosome, not in plasmids. Moreover, the sHm density in the S+M+ line was lower than that in the S+ line (Steel–Dwass test, P < 0.05, Figure 4C).

Rapid genetic evolution leading to resistance against MK Spiroplasma has been reported in various hosts, such as the lacewing (Hayashi et al., 2018) and the planthopper (Yoshida et al., 2021). However, we can exclude the host genetic changes from the possible cause of the observed loss of MK phenotype because females of the S+ and S+M+ lines were parallelly mated with the males of the genetically homogeneous NSR line that had been maintained via inbreeding in the laboratory for over 10 years (>120 generations).

Spiroplasma-induced phenotypic changes have been repeatedly observed in previous studies. For example, spontaneous loss of MK was found in S. poulsonii strains of Drosophila flies, wherein substitutions and deletions occurred in the MK gene spaid (Harumoto and Lemaitre, 2018). Moreover, the MK strain S. poulsonii NSRO and its non-MK variant NSRO-A exhibit difference in bacterial densities in D. melanogaster (Anbutsu and Fukatsu, 2003). Indeed, bacterial density is one of the crucial factors for Spiroplasma- and Wolbachia-induced MK phenotype (Hurst and Jiggins, 2000; Kageyama et al., 2007; Arai et al., 2020). Based on these results, we speculate that the loss of MK phenotype of sHm-infected H. magnanima was due to (i) reduced sHm density and/or (ii) mutations in sHm MK gene(s) or factors regulating MK gene expression levels. However, we still do not know how the small genomic rearrangements (i.e., inversions and insertions) detected in this study are involved in the phenotypic changes of sHm. Future de novo genome construction of sHm from S+M+ lines and gene function analysis would help in elucidating MK mechanisms.

Strain sHm was abundant at the late-developmental stages of H. magnanima (Figure 5A), and sHm densities drastically increased from pupal to adult stages of the insect (Figure 5B). In D. melanogaster, S. poulsonii copy numbers gradually increase as the host larval development proceeds and are generally higher in pupae than in larvae (Anbutsu and Fukatsu, 2003). In contrast to S. poulsonii, which is reported to be the most abundant in hemolymph (Anbutsu and Fukatsu, 2003, 2006), strain sHm exhibited low density in the hemolymph and high density in Malpighian tubules in the fifth instar larva stage (Figure 5C). High titers in Malpighian tubules are also a characteristic of Wolbachia; Wolbachia present in Malpighian tubules protects the host from RNA-virus infections and may constitute a secondary pool of vertically infected bacteria (Faria and Sucena, 2013; Pietri et al., 2016). The localization of strain sHm in somatic tissues may have contributed to the fitness of H. magnanima Although there have been no reports of S. ixodetis localization patterns in insects, our findings suggest that S. poulsonii and S. ixodetis have distinct proliferation strategies. The hemolymph is a nutrient-rich environment but is likely an extreme habitat for microorganisms because it is well-defended by the immune system of the host (Blow and Douglas, 2019). Indeed, only a few microbial taxa are known to persist in the hemolymph of insects for extended periods without causing insect morbidity and death (Blow and Douglas, 2019). Intriguingly, S. poulsonii sMel encoded more metabolic genes in its genome than S. ixodetis sHm (Supplementary Table 3). Hemolymph-inhabiting S. poulsonii may have developed specific adaptations for its habitat, which are distinct from those of S. ixodetis. Further characterization of genomic features and localization patterns of Spiroplasma strains will clarify the distinct proliferation strategies of the two species (e.g., nutrient requirements).

Tsugeno et al. (2017) reported that sHm is horizontally transmitted by inoculating non-infected H. magnanima with concentrated hemolymph collected from sHm-infected H. magnanima. Moreover, we revealed that S. ixodetis sHm exhibited Wolbachia-like genetic characteristics (i.e., multiple ankyrin genes) and localization patterns in somatic tissues. Wolbachia can infect and be maintained stably in insect cell lines derived from insect taxa that are distantly related to their native hosts (Fallon, 2021). To examine whether sHm can infect insect cells, we transinfected sHm to the cell lines of A. albopictus (AeAl2) and B. mori (aff3), which are known to be susceptible to Wolbachia. sHm proliferated successfully by placing a piece of fat bodies or Malpighian tubules derived from an S+ female larva into a flask containing the AeAl2 or aff3 cells (Figure 6). sHm was stably maintained in the cell lines for 12 weeks (Figure 6A) but not in cell-free medium IPL-41. qPCR revealed that sHm titers in AeAl2 cells were significantly higher at 12 weeks than at 4 weeks after the introduction of sHm (Figure 6B). This implies the potential of sHm to survive in a wide host range besides Homona (Tsugeno et al., 2017), such as other lepidopteran and dipteran insects. S. ixodetis strains isolated from Japanese ticks were also shown to be culturable in the A. albopictus cell line C6/36 (Thu et al., 2019). We hypothesize that S. ixodetis strains have a broad host range like that of Wolbachia. It is not clear whether S. poulsonii has a broad host range because there is no but one report by Hackett et al. (1986) that showed the infectivity of strain WSRO (derived from D. willistoni) in the Trichoplusia ni cell line IPLB-TN-R². Several attempts to transinfect S. poulsonii (strain NSRO; derived from D. nebulosa) and MK Spiroplasma (derived from the lacewing Mallada desjardinsi) into AeAl2 and aff3 cells failed (personal observation by DK). It has been shown by hemolymph injection that S. poulsonii can infect drosophilid flies but not houseflies, suggesting its narrow host range (Williamson and Poulson, 1979).

In this study, we sequenced and analyzed the genome of an MK S. ixodetis strain sHm. S. poulsonii possesses the Spaid toxin as the MK factor, whereas our study revealed that MK S. ixodetis did not harbor spaid homologs. We speculate that MK S. ixodetis strains found in a diverse range of insects (Hurst et al., 1999; Jiggins et al., 2000; Simon et al., 2011; Tabata et al., 2011; Sanada-Morimura et al., 2013) harbor yet-unknown MK gene(s), other than spaid; thus, future studies should focus on the identification of these MK genes. Besides, high infection efficiencies of strain sHm in other insect cells led us to speculate that MK S. ixodetis has been horizontally transmitted among insect species, like Wolbachia, which has expanded its host range (Zhou et al., 1998; Baldo et al., 2006). Further studies would be required to understand whether closely related MK Spiroplasma strains (i.e., the S. ixodetis group) share common or different MK mechanisms, which will answer evolutionary questions such as how frequent novel MK genes arose, how MK genes moved between different Spiroplasma strains (if it did), and whether MK genes are associated with host sex determining systems.