Introduction
Wolbachia, an intracellular facultative symbiotic bacterium of arthropods, is classified under the phylum Proteobacteria, class Alphaproteobacteria, order Rickettsiales, and family Anaplasmataceae. The genus was first identified in the common house mosquito Culex pipiens in 1924 (Hertig and Wolbach, Reference Hertig and Wolbach1924). While primarily transmitted vertically through the maternal cytoplasm, it can also spread horizontally via various routes (Ahmed et al., Reference Ahmed, Breinholt and Kawahara2016). This bacterium significantly influences the physiology of its host by providing essential nutrients, antagonizing pathogens to improve disease resistance, manipulating reproduction via multiple mechanisms, modulating apoptosis, and shortening lifespan (Min and Benzer, Reference Min and Benzer1997). Previous studies have indicated that approximately 40% to 60% of arthropods harbour Wolbachia, with infection rates in insects ranging from 20% to 75% across orders such as Hemiptera, Diptera, Hymenoptera, Lepidoptera, and Orthoptera (Werren et al., Reference Werren, Windsor and Guo1995; Hilgenboecker et al., Reference Hilgenboecker, Hammerstein, Schlattmann, Telschow and Werren2008; Wolfgang et al., Reference Wolfgang, Markus, Dimitrios and Christian2009; Zug and Hammerstein, Reference Zug and Hammerstein2012). Infection rates vary considerably among species and even among individuals within the same species, ranging from 3% to 100% (Wolfgang et al., Reference Wolfgang, Markus, Dimitrios and Christian2009). Furthermore, Wolbachia strains exhibit considerable diversity: a single host individual may harbour multiple strains simultaneously, and geographically distinct populations of the same host species often carry different Wolbachia variants (Kikuchi and Fukatsu, Reference Kikuchi and Fukatsu2003; Riegler et al., Reference Riegler, Sidhu, Miller and O’neill2005).
The insect mitogenome is a circular, double-stranded biological macromolecule, and has been widely used as molecular markers in studies of insect taxonomy, systematics, and evolution (Nass and Nass, Reference Nass and Nass1963; Cameron, Reference Cameron2014). Studies have demonstrated that Wolbachia can influence the evolution and population genetic structure of host mitochondrial DNA (Yang et al., Reference Yang, Ilinsky, Wei, Duan, Zhu and Jiang2026). Due to their shared mode of maternal inheritance, mtDNA and Wolbachia often exhibit strong linkage disequilibrium within host populations (Jiggins, Reference Jiggins2003; Hurst and Jiggins, Reference Hurst and Jiggins2005). Specifically, Wolbachia infection can trigger a selective sweep – a process in which the bacterium rapidly spreads through a host population owing to its adaptive advantage, thereby increasing the frequency of the associated mitochondrial DNA haplotypes. Consequently, this leads to a marked reduction in mtDNA polymorphism within the host population (Jiggins, Reference Jiggins2003). Furthermore, given the frequent hybridization among sympatric, closely-related species, selective sweeps driven by Wolbachia may facilitate mtDNA introgression between species (Narita et al., Reference Narita, Nomura, Kato and Fukatsu2006). Therefore, the impact of endosymbiotic Wolbachia must be rigorously accounted for as a potential confounding variable when employing mtDNA in population genetics research. The family Hesperiidae, a diverse lineage within the order Lepidoptera, encompasses approximately one-fifth of the world’s butterfly species. There are over 560 genera and 4,000 species known worldwide (Warren et al., Reference Warren, Ogawa and Brower2009; Li et al., Reference Li, Cong, Shen, Zhang, Hallwachs, Janzen and Grishin2019). Due to the high species diversity and extensive geographical range, the spread and evolutionary dynamics of symbiotic bacteria associated with skippers have attracted increasing research attention. Research has shown that Wolbachia is also ubiquitous in skippers. Although infection rates vary among different lineages, evidence suggests that horizontal transmission may play a significant role alongside vertical transmission (Zhao et al., Reference Zhao, Zhu, Hoffmann, Cao, Shen, Fang, Ma, Liu, Yu and Tang2021). Wolbachia can influence the genetic architecture of skipper mtDNA, which may reduce intraspecific diversity via selective sweeps, facilitate interspecific mtDNA introgression, and potentially influence sex ratio distortion in host populations (Jiang et al., Reference Jiang, Zhu, Chen, Yang, Du, Chen, Zhang, Yu and Yu2014, Reference Jiang, Zhu, Wu, Li, Li, Ge, Wang, Endersby, Hoffmann and Yu2018). However, these effects are not universal. For instance, studies on the Tagiadini suggest that phylogenetic discordance between mitochondrial and nuclear markers may not necessarily be driven by Wolbachia (Wei et al., Reference Wei, Zhu, Hoffmann, Jia, Xiao, Duan, Zhang, Zhong, Ge, Yu, Zhang and Jiang2024). Collectively, these findings highlight the complex dynamic symbiotic relationship between Wolbachia and the skippers, providing useful insights into the role of symbiotic bacteria in the evolution of this group.
In this study, we sequenced the housekeeping genes of Wolbachia from 21 species representing 14 genera and 6 tribes within Hesperiinae. Concurrently, the host mitochondrial COI and Cytb genes, along with the nuclear EF-1α gene, were sequenced. Separate phylogenetic trees were constructed for both the host and the Wolbachia. Integrating these phylogenies with analyses of mitochondrial genetic diversity and neutrality tests, we systematically explored the influence of Wolbachia on the host’s mitochondrial genetic structure and phylogeny. Furthermore, we examined the phylogenetic relationship between the host and Wolbachia, as well as the transmission modes of the bacterium.
Materials and methods
Sample collection and DNA extraction
Adult specimens were mainly collected using the net trapping method. Immediately following collection, specimens were preserved in cryogenic tubes containing anhydrous ethanol. The preservation medium was replaced weekly to ensure sample integrity. After returning to the laboratory, all specimens were stored at −20 °C. Species identification and sex determination were performed based on external morphological characteristics (Supplementary table S1).
Genomic DNA was extracted from adult tissue samples using the Qiagen DNeasy Blood and Tissue Kit. Notably, for samples harbouring Wolbachia, the extracted total DNA contained both host genomic DNA and Wolbachia DNA.
PCR amplification of Wolbachia, mitochondrial, and nuclear genes
In this study, we initially screened for Wolbachia infections in Hesperiinae using the wsp gene. Individuals testing positive were then subjected to PCR amplification of five housekeeping genes (gatB, hcpA, ftsZ, coxA, and fbpA) to confirm infection and determine strain types. Samples that consistently failed to amplify were recorded as negative, provided that positive controls yielded amplicons of the expected size. All primers used for gene amplification were synthesised by Beijing Qingke Biotech Co., Ltd., and the sequences are listed in the Supplementary table S2.
PCRs were performed in a 25 μL final volume under the following conditions: 2 μL template DNA, 1 μL concentrations of all forward and reverse primers (final concentration 0.4 μM), 12.5 μL of 2 × Taq Master Mix, and 8.5 μL of ddH₂O. For Wolbachia genes, the cycling conditions were as follows: initial denaturation at 94 °C for 2 min; 35 cycles of 94 °C for 30 s, 55 °C for 45 s, and 72 °C for 90 s; and a final extension at 72 °C for 7 min. For the host genes (COI, Cytb, and EF-1α), the same reaction mixture was used, with the following cycling conditions: initial denaturation at 95 °C for 5 min; 35 cycles of 95 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min; and a final extension at 72 °C for 10 min.
Negative controls (ddH₂O) and positive controls were included in all PCR runs to monitor contamination and amplification success. The amplification products were subjected to electrophoresis in a 2% (w/v) agarose gel in TAE buffer (0.04 M Tris–acetate, 0.001 M EDTA) with a DL2000 ladder size marker (Sangon Biotech, Shanghai, China) to determine whether the amplification reactions were successful. After electrophoresis in agarose gels, the amplification products were subjected to gel excision for recovery and purification. The purified solution of the target fragment was sent to Beijing Qingke Biotech Co., Ltd. for bidirectional sequencing of the target fragment. The sequencing chromatograms were examined for quality, and the sequences were assembled after verification. The assembled sequences were then compared against the NCBI database using BLAST to confirm their identity as the target genes. Finally, all the sequences obtained were deposited in Figshare (https://doi.org/10.6084/m9.figshare.31994628).
MLST analysis of Wolbachia
The five housekeeping genes were searched in the PubMLST database (https://pubmlst.org/wolbachia/) to identify allele numbers and sequence types (STs). Sequences identical to existing entries were directly assigned the corresponding allele numbers. Novel sequences lacking matches in the database were submitted along with raw sequencing files to obtain new allele numbers. We then compared the five housekeeping genes obtained from sequencing with the existing ST types in the database. If no match was found with the existing ST type, it was determined as a new ST type. In such cases, strain information, supergroup assignment, and MLST allele numbers were submitted to the database to obtain an official ST number. All allele profiles obtained in this study matched known STs in the PubMLST database; no novel STs were identified.
Phylogenetic analysis
To contextualise our results, we retrieved sequences from the PubMLST database that matched the strains identified in this experiment and selected representative strains from previous studies based on the literature. The sequences of five housekeeping genes of these reference strains (table 1) were obtained from the PubMLST database for subsequent analysis.
Wolbachia strains
Table 1 Long description
Wolbachia strain names are paired with their host species and assigned supergroup. Sixteen strains are included across insects and one nematode. Supergroup A contains eight strains, including wMel in Drosophila melanogaster, two strains in Muscidifurax uniraptor, and three strains in Polytremis species. Supergroup B contains six strains spanning multiple butterflies and moths, including wBol1 in Hypolimnas bolina and entries for Acraea eponina, Neptis rivularis, Celastrina argiolus, Aphantopus hyperanthus, and Heteropterus morpheus. Supergroup D appears once with wBm in Brugia malayi, and supergroup F appears once with wCle in Cimex lectularius. The table provides categorical assignments only and does not indicate abundance, relatedness strength, or sampling methods.
The phylogenetic tree of Wolbachia was rooted using the wBm strain (supergroup D) and the wCle strain (supergroup F). Individual genes were aligned using MAFFT v7.505 with the L-INS-I strategy (Katoh and Standley, Reference Katoh and Standley2013), and alignment trimming was performed using trimAl v1.4 (Capella-gutiérrez et al., Reference Capella-gutiérrez, Silla-martínez and Gabaldón2009). The five gene sequences were concatenated using the Concatenate Sequence function in PhyloSuite v1.2.3. For model selection and partitioning, the concatenated datasets were analysed using PartitionFinder2 (Lanfear et al., Reference Lanfear, Frandsen, Wright, Senfeld and Calcott2017) under the Bayesian Information Criterion (BIC), with the ‘greedy’ search algorithm to select the optimal partitioning schemes and nucleotide substitution models for each partition. Maximum likelihood (ML) analyses were performed using IQ-TREE v2.2.0 (Nguyen et al., Reference Nguyen, Schmidt, Von Haeseler and Minh2015) with the ultrafast bootstrap (UFB) approximation approach (Hoang et al., Reference Hoang, Chernomor, Von Haeseler, Minh and Vinh2018), employing 10,000 bootstrap replicates to assess node support. UFB values ≥ 95% were considered strongly supported, and values between 70% and 94% as moderately supported.
For phylogenies based on mitochondrial and nuclear genes, Pyrgus malvae (Pyrginae) was designated as the outgroup. All genes were concatenated into two datasets: (1) COI + Cytb, comprising concatenated sequences of mitochondrial COI and Cytb genes; and (2) COI + Cytb + EF-1α, comprising concatenated sequences of mitochondrial genes (COI and Cytb) and the nuclear gene EF-1α. The phylogenetic analyses for these datasets followed the same procedures as described above for the Wolbachia.
Genetic diversity analysis
Based on the COI, Cytb and EF-1α genes, the number of haplotypes (H), haplotype diversity (Hd), nucleotide diversity (Pi), and the average number of nucleotide differences (K) among individuals within the same species were calculated using DnaSP v5.10 (Rozas et al., Reference Rozas, Librado, Sanchez-delbarrio, Messeguer and Rozas2009). According to the neutral theory, values of Tajima’s D statistic and Fu’s F statistic were expected to be zero or close to zero, indicating that the examined sequences conform to neutral evolution. A negative result suggests the presence of numerous low-frequency variants, which may be influenced by factors such as selective sweeps. Positive values suggest an excess of intermediate-frequency alleles, which may be indicative of balancing selection or population contraction. Statistical significance of the test statistics was determined using coalescent simulations with 1,000 replicates; values with P < 0.05 were considered statistically significant. However, for species with small sample sizes (n < 5), the results of these neutrality tests should be interpreted with caution, as small samples may inflate the variance of the statistics and increase the risk of Type I error (Fu and Li, Reference Fu and Li1993; Aris-brosou and Excoffier, Reference Aris-brosou and Excoffier1996).
Results
Sampling coverage and Wolbachia infection prevalence
In this study, 72 samples representing 21 species and 14 genera in Hesperiinae were sequenced, comprising 37 females and 35 males. All infected samples carried a single strain (table 2). A total of 42 samples tested positive for Wolbachia infection, resulting in an overall individual infection rate of approximately 58.33%. Across the 21 examined species, 13 (61.9%) were found to be infected. The infection rate was notably higher in females (70.27%, 26/37) than in males (45.71%, 16/35), and this difference was statistically significant (Pearson χ2 = 4.462, df = 1, P = 0.035). Among the 13 infected species, all sampled individuals tested positive in 9 species, accounting for 69.23% of the infected species. For the remaining 4 infected species, the observed proportions of positive individuals were below 100% (Aeromachus catocyanea, Ochlodes cryptochraceus, Onryza maga, and Notocrypta curvifascia), with O. cryptochraceus showing the lowest at 8% (1/12). It should be noted that for species represented by only one or two individuals, these observed proportions are descriptive of the sampled individuals only and should not be interpreted as robust species-level prevalence estimates.
The infection situation of Wolbachia in the Hesperiinae
Table 2 Long description
Wolbachia infection is summarized by species, listing sample size, infected females, infected males, and the overall infection rate among individuals. Complete infection is reported for A. dioscorides (four sampled), Ampittia virgata (five), P. nascens (seven), Polytremis theca (one), Notocrypta feisthamelii (one), Caltoris cahira (one), O. thibetana (fifteen), Isoteinon lamprospilus (two), and Potanthus confucius (one). Partial infection appears in A. catocyanea (six sampled, one infected female), O. maga (three sampled, one infected male), Notocrypta curvifascia (two sampled, one infected female), and O. cryptochraceus (twelve sampled, one infected female). Several species have no infected individuals recorded and no infection rate reported, including Parnara guttata, Pelopidas agna, Baoris leechi, and Thymelicus leoninus. Some rows include dashes for one sex, indicating missing or not recorded sex-specific counts, and a few infection rates are not provided even when a single individual was sampled. Because many species have very small sample sizes, especially single-individual samples, the reported rates may not represent broader population infection levels.
MLST typing and Wolbachia supergroup assignment
Based on the five housekeeping genes (gatB, coxA, hcpA, ftsZ, and fbpA), all Wolbachia strains identified in this study were assigned to supergroups A and B (fig. 1). The most frequently observed ST was ST374, which was identified in wol_of_Oc_thi_1–15. ST302 was identified in all five A. virgata strains (wol_of_Am_vir_1–5). ST19 was identified in multiple host genera, including wol_of_Po_the, wol_of_Ca_cah, wol_of_No_cur, wol_of_No_fei, and all four A. dioscorides strains (wol_of_Am_dio_1–4), suggesting that this ST may be widely distributed across different host lineages. Five of the six P. nascens strains (wol_of_Po_nas_1–5 and wol_of_Po_nas_7) and wol_of_Po_con exhibited a triple ST combination (ST290 + ST351 + ST502), and the P. nascens strain (wol_of_Po_nas_6) was assigned to ST23. The presence of multiple STs within single individuals strongly suggests either co-infection by multiple Wolbachia strains or sequence heterogeneity, warranting further verification through cloning and sequencing. The wol_of_Om_3 and wol_of_Is_lam_1–2 were assigned to ST432 and ST41, respectively. All allelic profiles obtained in this study matched known sequence types in the PubMLST database; no novel STs were identified.
Maximum likelihood (ML) phylogenetic tree of Wolbachia based on concatenated sequences of five housekeeping genes (gatB, coxA, hcpA, ftsZ, and fbpA) from Hesperiinae. Strains identified in this study are labelled as wol_of_ followed by host species abbreviations. Numbers at nodes indicate ultrafast bootstrap support values (10,000 replicates).
Figure 1 Long description
The diagram is a Maximum Likelihood phylogenetic tree illustrating the relationships among Wolbachia strains based on concatenated sequences of five housekeeping genes. The tree is oriented vertically, with branches extending from top to bottom. Strains are labeled with identifiers such as ′wol_of′ followed by host species abbreviations. The tree is divided into four colored regions representing supergroups: A in green, B in blue, D in purple and F in pink. Each node is marked with ultrafast bootstrap support values, indicating the confidence level of the branching. The topmost branch leads to supergroup F, followed by supergroup D. Supergroup A occupies the central portion, while supergroup B is at the bottom. The branches split and merge, showing evolutionary relationships among the strains. Specific strains like ′wol_of_Oc_thi1–15′ and ′wol_of_Am_vir1–5′ are highlighted, indicating their sequence types. The tree visually represents the distribution and genetic diversity of Wolbachia strains across different host lineages.
Host phylogenetic relationships based on COI + cytb and COI + cytb + ef-1α
In this study, ML phylogenetic trees were constructed based on the COI + Cytb and COI + Cytb + EF-1α datasets, using Pyrgus malvae as the outgroup (fig. 2).
Maximum likelihood (ML) phylogenetic trees of Hesperiinae reconstructed from two concatenated datasets: mitochondrial genes (COI + Cytb; right) and mitochondrial + nuclear genes (COI + Cytb + EF-1α; left). Numbers at nodes indicate ultrafast bootstrap support values (10,000 replicates).
Figure 2 Long description
Two Maximum Likelihood phylogenetic trees of Hesperiinae are presented side by side. The left tree is based on the COI plus Cytb plus EF-1α alpha dataset, while the right tree uses the COI plus Cytb dataset. Both trees use Pyrgus malvae as the outgroup. The trees are annotated with ultrafast bootstrap support values at the nodes, derived from 10,000 replicates. The trees are color-coded to indicate different tribes: Aeromachini in pink, Ancistroidini in red, Baorini in green, Hesperiini in blue, Iseotini in yellow and Taractrocerini in purple. The outgroup is marked in gray. The left tree shows a more complex branching pattern compared to the right, reflecting the additional nuclear gene data. Each branch represents a lineage, with labels indicating specific taxa. The trees illustrate the evolutionary relationships among the taxa, with branches diverging from common nodes, representing speciation events.
In the phylogenetic tree constructed based on the dataset COI + Cytb, the monophyly of Hesperiini was moderately supported (BS = 84), while Ancistroidini and Taractrocerini formed a single branch as sister groups to each other (BS = 79). Baorini and Aeromacini were not recovered as monophyletic groups. Specifically, A. virgata (tribe Aeromacini) and the Caltoris cahira (tribe Baorini) clustered together and Polytremis lubricans, P. nascens, and P. theca were placed in distinct branches while Parnara guttata formed a separate branch. Isoteinon lamprospilus (tribe Isoteinonini) and O. maga (tribe Aeromachini) were grouped together, with A. catocyanea and A. dioscorides forming a basal branch. Notably, A. virgata was nested within the tribe Baorini.
In the phylogenetic tree constructed using the COI + Cytb + Ef-1α dataset, the monophyly of four tribes was supported: Hesperiini (BS = 64), Ancistroidini (BS = 100), Baorini (BS = 99), and Taractrocerini (BS = 66). However, the genus Polytremis within Baorini was not monophyletic; P. lubricans grouped with C. cahira, B. leechi, and Pelopidas agna. Additionally, Isoteinonini was nested within Aeromachini and clustered with O. maga (BS = 51) (fig. 2).
Comparison between host and Wolbachia phylogenies
We compared the phylogenetic trees established based on the datasets (COI + Cytb and COI + Cytb + EF-1α) with the phylogenetic tree constructed based on the five housekeeping genes of Wolbachia (gatB + coxA + hcpA + ftsZ + fbpA). Within Aeromacini, all samples of A. dioscorides and A. virgata were infected with Wolbachia. In Baorini, P. nascens, P. theca, and C. cahira were infected, whereas species in other branches were not. In Hesperiini, all samples of O. thibetana were infected, and these strains belonged to supergroup B. In Taractrocerini and Ancistroidini, infections were present, and the infecting Wolbachia strains all belonged to supergroup A. I. lamprospilus, which was nested within Aeromachini, was also infected with Wolbachia; notably, the Wolbachia strain infecting I. lamprospilus showed a relatively close genetic relationship to that infecting O. maga (figs 3 and 4). By comparing the phylogenetic trees constructed using Wolbachia and the two datasets, it can be observed that the topological structures of the two were not consistent overall. The phylogenetic relationship of Wolbachia did not accurately reflect the genetic relationship of the hosts, suggesting that Wolbachia may not persist stably in the hosts nor co-evolve strictly with them. However, these topological incongruences are only suggestive of horizontal transmission and do not constitute direct evidence; alternative explanations, such as incomplete lineage sorting, cannot be excluded.
Comparison of the ML phylogenetic tree of Hesperiinae based on mitochondrial genes (COI + Cytb; left) and the ML tree of Wolbachia based on concatenated ST genes (five housekeeping genes; right).
Figure 3 Long description
Compares two ML phylogenetic trees side by side. On the left, the tree is based on mitochondrial genes COI plus Cytb, showing various groups such as Outgroup, Aeromacini, Ancistroidini, Baorini, Hesperiini, Isoteinonini and Taractrocerini, each represented by different colored backgrounds. On the right, the tree is based on Wolbachia genes, divided into supergroups A, B, D and F, also distinguished by colored backgrounds. Both trees display branching patterns with labeled nodes and taxa. The left tree includes taxa like P. nascens and C. cahira, while the right tree shows Wolbachia strains such as wCI and wBa. Dashed lines connect corresponding taxa between the two trees, indicating relationships and infections. The diagram highlights the incongruence between the phylogenetic relationships of the hosts and the Wolbachia strains.
Comparison of the ML phylogenetic tree of Hesperiinae based on concatenated mitochondrial and nuclear genes (COI + Cytb + EF-1α; left) and the ML tree of Wolbachia based on concatenated ST genes (five housekeeping genes; right).
Figure 4 Long description
Compares two maximum likelihood phylogenetic trees. On the left, the tree is based on concatenated mitochondrial and nuclear genes labeled COI plus Cytb plus EF-1α alpha. This tree includes various groups such as Outgroup, Aeromacini, Ancistroidini, Baorini, Hesperiini, Isoteinonini and Taractrocerini, each represented by different colored backgrounds. On the right, the tree is based on concatenated ST genes of Wolbachia, including gatB, coxA, hcpA, ftsZ and fbpA. This tree is divided into supergroups A, B, D and F, each with distinct colored backgrounds. Both trees display branching patterns with nodes and branches indicating evolutionary relationships. The left tree shows species names and their respective groupings, while the right tree shows Wolbachia strains and their supergroup classifications. Dashed lines connect corresponding species and strains across the two trees, indicating relationships between host species and Wolbachia strains.
mtDNA diversity and neutrality tests in relation to infection status
Genetic diversity analysis of the mitochondrial genes (COI + Cytb) was performed for species represented by three or more individuals. Among these, P. nascens exhibited the lowest haplotype diversity, average number of nucleotide differences, and nucleotide diversity. All the tested samples of P. nascens were infected with Wolbachia. Conversely, the highest levels of haplotype diversity, average number of nucleotide differences, and nucleotide diversity were found in O. maga (table 3).
Population genetic diversity based on tandem genes (COI + Cytb) of Hesperiinae
Table 3 Long description
Population genetic diversity and neutrality statistics are reported for eight Hesperiinae species using combined COI and Cytb sequences, alongside sample size and infection counts by sex. Infection is complete in five species: A. dioscorides, A. virgata, P. nascens, and both O. thibetana and O. cryptochraceus, while A. catocyanea is low at sixteen percent and B. leechi is zero. O. maga has three samples with one infected male, giving thirty three percent infection. Genetic diversity is highest in O. maga, with three haplotypes, maximum haplotype diversity, and the largest average nucleotide differences, polymorphic sites, and nucleotide diversity. A. virgata also shows high diversity with five haplotypes and relatively high nucleotide diversity, whereas P. nascens has low haplotype diversity and very low nucleotide diversity despite full infection. A. catocyanea shows a significant negative Tajima’s D and a significant negative Fu’s F, suggesting a departure from neutrality, while several species have non-significant values and some tests are not reported for O. maga and B. leechi. Comparisons should be interpreted cautiously because sample sizes are small for several species, which can limit the reliability of diversity and neutrality estimates.
H, number of haplotypes; Hd, haplotype diversity; K, average number of nucleotide differences; Pi, nucleotide diversity; S, polymorphic site; D, the value of Tajima’s D; Fs, the value of Fu’s F; *P < 0.05, **P < 0.01.
In this study, Tajima’s D and Fu’s FS values for the mitochondrial genes of A. catocyanea, A. dioscorides, and P. nascens were all negative, indicating an excess of low-frequency variants. Notably, the Tajima’s D of A. catocyanea was significantly negative (P < 0.05), and the value of Fu’s FS was also significantly negative (P < 0.01) (table 3). However, this significant negative value cannot be simply interpreted as definitive evidence of adaptive evolution. On the one hand, the relatively small sample size for A. catocyanea (n = 6) may inflate the sampling variance of the statistical tests, possibly leading to an overestimation of significance (Type I error inflation). On the other hand, demographic processes such as population expansion can also generate similar signals of low-frequency variant excess. Furthermore, for O. maga and B. leechi, the extremely small sample sizes (n ≤ 4) precluded the reliable estimation of the test statistics (indicated by ‘–’), further highlighting the limitations of this dataset for population genetic inference.
Taken together, we interpret the significant negative values in A. catocyanea as suggestive evidence that its mitochondrial genes may deviate from strict neutrality. However, we are currently unable to distinguish between the relative contributions of natural selection and demographic history, and we caution against overinterpreting the statistical uncertainty arising from small sample sizes. Future studies with larger sample sizes and nuclear markers are needed for more robust inferences.
Discussion
Wolbachia prevalence in Hesperiinae compared with other Lepidoptera
This study investigated the skippers from 11 provinces within China, covering six tribes, 14 genera, 21 species, and 72 individuals. Wolbachia infection was detected in 13 species, corresponding to a species infection rate of 61.9%. This proportion is lower than that reported in a previous large-scale global survey (which covered 10 superfamilies, 17 families, 300 species, and 7,604 individuals from all continents except Antarctica, with a species infection rate of 80%) (Ahmed et al., Reference Ahmed, Araujo-jnr, Welch and Kawahara2015). Recent studies suggest that environmental temperature may exert a stronger influence than host factors on microbial transmission; certain symbiotic bacteria are sensitive to thermal adaptation, and their activity is closely tied to temperature fluctuations (Wernegreen, Reference Wernegreen2012; Kikuchi et al., Reference Kikuchi, Tada, Musolin, Hari, Hosokawa, Fujisaki and Fukatsu2016). Given that the majority of the samples in this study were collected in the southwestern region of China, differences in geographical distribution may partly account for the lower infection rate observed in this research compared with previous studies.
Evidence and limitations regarding horizontal transmission
It was observed that all Wolbachia strains infecting the samples belong to the A and B supergroups. Previous research indicated that the Wolbachia infecting lepidopteran insects mostly belong to the A supergroup and B supergroup (Ahmed et al., Reference Ahmed, Araujo-jnr, Welch and Kawahara2015).
The wol_of_Po_the (of P. theca) is grouped together with the EkueA strain (of E. kuehniella) and the ST19 type. Previous research found that EkueA can induce the cytoplasmic incompatibility in E. kuehniella (Lewis et al., Reference Lewis, Champion de Crespigny, Sait, Tregenza and Wedell2011), and the ST19 type may have a horizontal transmission pathway mediated by the host plant and spread among Lepidoptera, Hymenoptera, and Coleoptera (Masui et al., Reference Masui, Kamoda, Sasaki and Ishikawa2000; Bordenstein and Wernegreen, Reference Bordenstein and Wernegreen2004). In this study, despite the hosts belonging to different genera and being collected from distinct locations, the wol_of_Am_dio_1-2 (of A. dioscorides) and wol_of_No_fei (of N. feisthamelii) are closely related on the phylogenetic tree, most likely caused by a pattern of horizontal transmission. Similarly, the Wolbachia strains from P. nascens located in different phylogenetic lineages are probably explained by horizontal transmission via parasitoid wasps. In parasitic insects like M. uniraptor, horizontal transfer of Wolbachia between parasitoid and host is common, with the mouthparts of infected parasitoid wasps likely facilitating transmission (Ahmed et al., Reference Ahmed, Araujo-jnr, Welch and Kawahara2015; Sanaei et al., Reference Sanaei, Charlat and Engelstädter2021). The wol_of_Oc_thi_1, 3-4, 6, 8-11, 14-15 (of O. thibetana), and Nriv1 strain were clustered together. Hosts of wol_of_Oc_thi_7 and 15 were collected from Yunnan province, while the remaining samples originated from Sichuan province. However, the phylogenetic relationships among their Wolbachia strains are highly complex, suggesting either that infection occurred prior to geographic isolation or that horizontal transfer events may have taken place. Additionally, wol_of_Is_lam_1–2 grouped with Carg_B, wBol1, and ST476. The wBol1 strain is known to induce male killing and has been associated with reduced host haplotype diversity and facilitated mtDNA introgression (Gross, Reference Gross2006). In our study, both wol_of_Is_lam_1–2 strains were obtained from female hosts, which may be related to Wolbachia infection, although small sample sizes and the absence of male-killing assays preclude confirmation.
Nevertheless, these phylogenetic incongruences are only suggestive of horizontal transmission and do not constitute direct evidence. Alternative explanations, such as incomplete lineage sorting, unsampled genetic diversity, or weak phylogenetic signal from the five MLST loci, cannot be ruled out. Therefore, the evidence for horizontal transmission presented here should be interpreted as indirect and treated with caution. Formal co-phylogenetic analyses would be required to statistically test the congruence between host and symbiont phylogenies.
Potential influence of Wolbachia on mitochondrial diversity
The impact of Wolbachia on host mitochondrial diversity is largely mediated by its vertical transmission mode, which links the bacterium’s spread to changes in mtDNA haplotype frequencies. A sex-biased infection pattern was observed in our study: females showed a significantly higher infection rate (70.27%) than males (45.71%). This difference was statistically significant (Pearson χ2 = 4.462, df = 1, P = 0.035). This pattern is consistent with the primarily maternal mode of vertical transmission of Wolbachia (Werren, Reference Werren1997). In Lepidoptera, Wolbachia can alter offspring sex ratios through cytoplasmic incompatibility, feminization, or male killing (Giordano et al., Reference Giordano, Jackson and Robertson1997). The female-biased infection observed here may reflect such reproductive manipulation, although sex-specific sampling bias or differential mortality cannot be entirely excluded.
In this study, some samples of the same species were collected from the same location and time period. Skipper populations in a given area may have been historically infected with Wolbachia. Through long-term coexistence and strict vertical transmission, the strain became stably inherited within the local population. For species with relatively large sample sizes (n ≥ 5), the observation that all sampled individuals tested positive – such as in P. nascens from Tangjiahe, Sichuan (n = 7) and O. thibetana (n = 15) – may be related to the long-term maintenance of the strain through vertical transmission. Furthermore, among the 13 species in which Wolbachia was detected, the observed proportion of infected individuals reached 100% in 9 species, accounting for 69.23% of the total infected species. This data distribution further supports the aforementioned viewpoint. Additionally, a limitation of this study is the small sample size for several species, precluding reliable species-level infection estimates. For these taxa, observed proportions reflect only the infection status of sampled individuals, not true population prevalence, and should not be overinterpreted. Future surveys with larger sample sizes are needed to accurately characterise infection frequencies across species.
Tajima’s D and Fu’s Fs neutrality tests were negative for A. catocyanea, A. dioscorides, and P. nascens, indicating an excess of low-frequency variants in these species. However, only A. catocyanea showed statistically significant deviation. This pattern is consistent with either a selective sweep or a recent population expansion. Importantly, for A. catocyanea, the sample size is relatively small (n = 6). In such cases, the tests may suffer from Type I error inflation, where significance is overestimated due to sampling variance. Therefore, we interpret this result as suggestive evidence of departure from neutrality, rather than definitive proof of selection. Future studies with larger sample sizes and additional nuclear markers are needed to distinguish between selection and demographic history. However, it should be noted that the low mitochondrial diversity observed in P. nascens may also be explained by local population history, such as founder effects or recent range expansion, given that all samples were collected from the same locality (Tangjiahe, Sichuan). Thus, while the observed pattern is consistent with a Wolbachia-mediated selective sweep, alternative demographic interpretations cannot be ruled out.
Limitations and future directions
A limitation of this study is the small and uneven sample sizes for several species, which constrains the statistical power of genetic diversity estimation and neutrality tests. Species with less than three individuals were excluded from relevant analyses; for those with moderately small samples, significant test results should be interpreted with caution, as small samples may inflate sampling variance and increase the risk of Type I error. In addition, the topological incongruence between host and Wolbachia trees provides only indirect evidence of horizontal transmission and requires formal co-phylogenetic analyses for confirmation. Thus, our inferences regarding horizontal transmission should be regarded as suggestive rather than conclusive. Future studies with expanded sampling and reproductive manipulation assays are needed to clarify the effects of Wolbachia on host reproduction and to disentangle its relative contributions from host demographic history.
Conclusion
The results of this study are consistent with the occurrence of horizontal transfer events among the genera and tribes of Hesperiinae, although the evidence is indirect and requires further validation. The variation in Wolbachia infection across Hesperiinae species suggests asynchronous infection events and ongoing selective pressure, which may have contributed to non-neutral mitochondrial evolution in affected populations. Among the intraspecific populations surveyed, those exhibiting 100% Wolbachia infection prevalence likely experienced earlier infection events, with the symbiotic bacteria subsequently maintained through multiple generations of host reproduction.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0007485326101217.
Acknowledgements
We thank Xuan Zhang (Northwest A&F University) for assistance with the software. We are thankful for the specimens collected by Jintian Xiao (Northwest A&F University) and Mengfei Yi (Northwest A&F University). We thank Xiangyu Hao (Northwest A&F University) and Zhanghui Liu (Northwest A&F University) for their guidance.
Author contributions
H.B.: Data curation; H.B.: Formal Bioinformatic Analysis; Y.S.: Conceptualization; W.C.: Software; H.B.: Writing – original draft; F.M.: Writing – review & editing; X.Y.: Conceptualization; X.Y.: Funding acquisition; X.Y.: Writing – review & editing.
Data availability statement
The gene sequences of hosts and Wolbachia are available through Figshare (https://doi.org/10.6084/m9.figshare.31994628).
Financial support
This study was supported by the National Natural Science Foundation of China (No. 31970448).
Conflicts of interest
The authors declare no conflict of interest.
