Endosymbionts in Cotesia
1_ From DNA extracts
Out of the 323 DNA extracts selected for Wolbachia screening, 282 were of good quality based on COI amplification, suggesting most of the specimens had been sufficiently preserved since extraction [44, 45].
The PCR amplifications for Arsenophonus, Spiroplasma or Microsporidium from 56 Cotesia specimens from four countries were negative (Table S1). There was one amplification using the Cardinium 16SrRNA primers in one unique specimen of C. melitaearum cryptic sp. H from Spain. However, our attempts at sequencing this amplificon were not successful, thus we could also not confirm the presence of Cardinium in our Cotesia samples. In contrast, out of the 282 Cotesia samples of good quality, 50 (17.7%) carried the symbiotic bacterium Wolbachia (Table 1, S1), representing at least eight Cotesia species parasitizing Melitaeini butterfly species (Fig. 1). As of the 10th January 2023, there was no record of Wolbachia strain from Cotesia species in the PubMLST database.
2_ From genome projects available in NCBI
By screening the 28 Cotesia genome projects (i.e. SRA projects) available on NCBI, we also identified 14 specimens (50%) containing at least 1000 reads classified as Wolbachia (Table S6). Ten specimens (six specimens from C. glomerata, one from C. sesamiae and three from C. vestalis) included Wolbachia reads distributed throughout the Wolbachia reference genomes (Fig. S2-3), while the last four specimens only included reads with patchy coverage across the Wolbachia reference genomes. These last four projects were considered as potential false positive results for Wolbachia infection, with the Wolbachia reads representing potential contamination, or insertions of Wolbachia sequences in the Cotesia host genomes.
Wolbachia strain diversity
Using the ten Cotesia genome projects found infected with Wolbachia, we partially assembled nine Wolbachia genomes. Three assemblies isolated from C. glomerata, exhibited BUSCO completeness of 86.8% (SRR13990441), 87.7% (SRR13990442), and 41.8% (SAMEA7283786) with corresponding total sizes of 1.10 Mbp, 1.08 Mbp, and 0.52 Mbp, respectively (See Table S7-8), while all other Wolbachia assemblies had a low number of BUSCO genes and were < 0.1 Mbp in size (Table S7-8). We were only able to extract between three and six MLST and wsp markers from the three largest Wolbachia assemblies.
Combining results obtained by direct amplification of the Wolbachia markers by PCRs and by screening the Wolbachia genomic assemblies built from Cotesia genomic sequences available on NCBI for those same markers, we obtained sequences from one to six markers for 38 (out of 61) Wolbachia-infected specimens (Table S1). We identified a total of 14 alleles for the ftsZ gene, nine for the hcpA gene, five for the coxA gene, six for fbpA, and six for gatB (See Table S1 for further details). This resulted in a concatenated alignment of 2559 bp, which allowed us to discriminate ten Wolbachia strains from ten Cotesia species (Table 1). We did not detect multiple infections in any of the individual Cotesia specimens, but two species carried several Wolbachia strains. Specimens of C. koebelei reared from E. editha carried either a supergroup A or a B Wolbachia strain, and Spanish specimens of C. bignellii carried a A-supergroup strain, while French specimens of the same species carried one of two B-supergroup strains (Fig. 1).
Table 1
Metadata for the Cotesia species and cryptic species found to be infected with Wolbachia: their butterfly host species, country of origin, and Wolbachia prevalence. Rows in grey highlight the specimens that were screened for all five symbionts (Table S1), while rows in white include the specimens screened for Wolbachia only.
Species | Host species reared from | Country | Infection rate (infected/total tested) | Strains detected |
C. acuminata cryptic sp. B | Melitaea phoebe | Spain | 24.4% (5/17) | Uncharacterized |
C. bignellii | Euphydryas aurinia | France | 100.0% (2/2) | wCbig |
C. bignellii cryptic sp. C | Euphydryas aurinia | Spain | 50.0% (3/6) | wCbigC |
C. koebelei | Euphydryas editha | USA | 100.0% (2/2) | wCkoeA, wCkoeB |
C. melitaearum cryptic sp. D | Euphydryas aurinia | Spain | 10.8% (4/37) | wCmelD |
C. melitaearum cryptic sp. F | Melitaea didyma | Spain | 100% (12/12) | wCmelF |
C. melitaearum cryptic sp. G | Melitaea trivia | Spain | 92.9% (13/14) | wCmelG |
C. melitaearum cryptic sp. H | Melitaea cinxia | Finland | 11.1% (6/54) | wCmelH1 |
C. melitaearum cryptic sp. H | Melitaea cinxia | Russia | 27.3% (3/11) | Uncharacterized |
C. glomerata | Pieris sp. | | 75% (6/8) | wCglo |
C. sesamiae | Stem boring moths | | 100% (1/1) | Uncharacterized |
C. vestalis | Plutella sp. | | 75% (3/4) | Uncharacterized |
Analyses of the Wolbachia genomic assemblies
By comparing the predicted proteomes of our two largest Wolbachia assemblies with an > 50% BUSCO completeness against those of the two Wolbachia reference genomes (wMelPop and wPipPel) using Prokka, we identified 954 protein-coding genes, 30 tRNAs, and one rRNA in the SRR13990441 assembly, and 996 protein-coding genes, 32 tRNAs, and three rRNAs in the SRR13990442 assembly (Table S9). In contrast, the two reference genomes, wMelPop and wPipPel, contained 1304 and 1410 protein-coding genes, 34 tRNAs, and three rRNAs, respectively (Table S9). The comparison using the Orthovenn 3 web server showed a total of 1057 conserved orthologs in all four strains, with 590 of these being single copy. All four strains shared 639 ortholog clusters (Fig. S4). The SRR13990442 assembly contains 876 orthologs, while SRR13990441 has 875, and they both share 71 unique orthologs with the reference B-supergroup Wolbachia wPipPel, but only 21 with the A-supergroup Wolbachia wMelPop reference (Fig. S4). Similarly, the ANI analysis, which calculates the average nucleotide identity among orthologous gene pairs shared between two genomes, revealed a high similarity between wPipPel, SRR13990441, and SRR13990442, with ANI values around 98% in pairwise comparisons (Table S10). In contrast, wMelPop displayed a lower ANI (~ 85%) in pairwise comparisons with wPipPel, SRR13990441, and SRR13990442 (Table S10). Altogether, these results suggest the two Wolbachia assemblies from Cotesia belongs to the B-supergroup Wolbachia.
Finally, we partially extracted the CI-associated genes from our Wolbachia assemblies. With this, we identified one copy of a Type I CifA in the SRR13990441 assembly (Table S11, Fig. S8), and a truncated/partial copy of cifB in both the SRR13990441 assembly (contig 109, position 1492–3201) and the SRR13990442 assembly (contig 221, position 1-1624). The sequences of the cifB gene from our Wolbachia assemblies were highly similar to that previously characterized from the fig wasp Kradibia gibbosae (Hymenoptera: Chalcidoidea) (WP_275944372.1), without any report of the role played by the symbiont in this host species [89].
Phylogenetic analyses
Our phylogenetic tree of the COI mitochondrial gene of 39 Cotesia species shows that the Cotesia wasps parasitizing Melitaeini butterflies belong to three distinct clades (See Fig. 1, S5-6:
-
Clade 1 includes C. melitaearum cryptic species (D, E, F, G, H, I, M, N),
-
Clade 2 includes C. koebelei,
-
Clade 3 includes C. bignellii cryptic species C, and C. acuminata cryptic species (A, B, and K).
This phylogenetic pattern is consistent with previous studies on the same clade [45]. Our phylogeny also preserves the previous grouping of C. xylina with C. yakutabensis [90]. Although the three Cotesia clades are specialists to the Melitaeini butterflies, the phylogeny suggests this host associated clustering is not conserved across the genus Cotesia. Indeed, closely related Cotesia species to each of the three clades have been described as parasitoids of divergent non-Melitaeini butterflies. For instance, Pieris sp. butterflies are host to C. glomerata, Lampides boeticus is host to C. specularis, and diverse moths host other Cotesia species (ie. Chilo sp. for C. flavipes, or Plutella sp. for C. vestalis). Furthermore, each of these butterfly species feeds on a wide diversity of host plants.
The Wolbachia phylogeny confirms that all Wolbachia strains characterized from Cotesia belonged to the A- and B-supergroups, with the majority (49/53, 92.4%) belonging to the B-supergroup (Fig. 1). Despite fewer representative taxa per phylogeny and lower resolution, phylogenies based only on individual gene alignments maintained similar sample groupings, with conserved strain assignment to supergroups A and B (Fig. S5), thus suggesting no recombination has occurred between the strains of the two supergroups in these Cotesia species. A visual comparison supports the lack of congruence and co-phylogeny between the maximum likelihood trees of Cotesia and their symbiotic strains. Phylogenetically close Wolbachia strains were found in phylogenetically distant Cotesia host species (Fig. 1). This was true for both the concatenated alignment as well as for the individual marker alignments inferring the symbiont phylogenies (Fig. 1, S5-6). The wsp gene tree suggests that the Wolbachia strain found in the butterfly host M. didyma is phylogenetically close to the Wolbachia strain from their Cotesia parasitoid, which could have occurred through horizontal transfer between the host and parasitoid. No similar pattern was observed for the Wolbachia strains we were able to extract from the other Lepidoptera hosts including Chilo partellus, Pieris rapae, or Plutella xylostella.