Remnants of horizontal transfers of Wolbachia genes in a Wolbachia-free woodwasp

Wolbachia is a bacterial endosymbiont of many arthropod and nematode species. Due to its capacity to alter host biology, Wolbachia plays an important role in arthropod and nematode ecology and evolution. Sirex noctilio is a woodwasp causing economic loss in pine plantations of the Southern Hemisphere. An investigation into the genome of this wasp revealed the presence of Wolbachia sequences. Due to the potential impact of Wolbachia on the populations of this wasp, as well as its potential use as a biological control agent against invasive insects, this discovery warranted investigation. In this study we first investigated the presence of Wolbachia in S. noctilio and demonstrated that South African populations of the wasp are unlikely to be infected. We then screened the full genome of S. noctilio and found 12 Wolbachia pseudogenes. Most of these genes constitute building blocks of various transposable elements originating from the Wolbachia genome. Finally, we demonstrate that these genes are distributed in all South African populations of the wasp. Our results provide evidence that S. noctilio might be compatible with a Wolbachia infection and that the bacteria could potentially be used in the future to regulate invasive populations of the wasp. Understanding the mechanisms that led to a loss of Wolbachia infection in S. noctilio could indicate which host species or host population should be sampled to find a Wolbachia strain that could be used as a biological control against S. noctilio.


Background
Wolbachia is a symbiont of many arthropod and filarial nematode species. This alphaproteobacteria in the family Anaplasmataceae is estimated to infect over 50% of terrestrial arthropods [1][2][3]. Due to its ubiquity and its effects on host reproduction and physiology, Wolbachia can have significant impacts on arthropod and nematode evolution [4].
Wolbachia uses a variety of mechanisms to modify the reproductive biology of its host and to enhance its chances of maternal transmission [5]. These mechanisms include male killing [6], feminization of genetic males [7], parthenogenesis induction [8] and cytoplasmic incompatibility that prevents embryonic development in crosses between a Wolbachia-positive male and a female that does not carry Wolbachia, or carries a different Wolbachia strain [9]. A Wolbachia infection can also provide advantages including resistance against viruses [10] and facilitating host iron metabolism [11].
A common characteristic of the Wolbachia-host interaction is Horizontal Gene Transfers (HGTs) from the Wolbachia genome to the host genome [12]. Thus far, over 20 species of nematodes, insects and isopods have been shown to carry Wolbachia genes in their genomes [13][14][15][16][17][18][19]. The transferred genetic elements vary in size from single genes to full genomes [15]. It is hypothesized that those HGTs were facilitated by the fact that the bacterial symbiont resides in the germline of the female host [12].
The genetic elements transferred from Wolbachia to their hosts sometimes include genes belonging to bacteriophages such as the Wolbachia-specific WO bacteriophages [19]. These bacteriophages play a crucial part in the Wolbachia-arthropod relationship [20]. It has been hypothesized that the phages can increase Wolbachia virulence and may be responsible for part of the molecular processes behind feminization of genetic males [21] and cytoplasmic incompatibility [22]. In order to be integrated into bacterial genomes, these viruses use specialized proteins that could also be responsible for the horizontal gene transfer of WO phage and Wolbachia genes into the hosts' genomes [23].
The woodwasp, Sirex noctilio Fabricus (Hymenoptera: Siricidae) originates from Europe, Eurasia and Northern Africa [24] and has been introduced in many countries over the last century [25]. Today, it is a very successful invader and a pest in many of the Southern Hemisphere pine forests [26]. Research into control strategies of the wasp has included the sequencing of its genome (Postma et al., unpublished). Analysis of the newly sequenced genome led to the identification of gene sequences apparently originating from Wolbachia. Because of the potential to use Wolbachia as a biological control agent against insect populations [27], this finding warranted further investigation.
In this study we investigated the presence of Wolbachia in South African populations of S. noctilio. We also investigated whether the Wolbachia genes observed in the genome of S. noctilio could have been horizontally transferred into the S. noctilio genome. We screened the entire S. noctilio genome to locate potentially horizontally transferred genes from Wolbachia. Finally, we screened individuals from different South African populations of the woodwasp using specifically designed PCR primers for the presence of the identified genes.

Presence of Wolbachia in S. noctilio
To test for the presence of Wolbachia in S. noctilio, 14 primers targeting three Wolbachia genes were used (Tables 1 and 2), along with a series of protocols that used three DNA extraction methods, two different Taq polymerases and a total of four cycling protocols with different annealing temperatures (Additional file 11: Table S1).
The general bacterial primers pA (27F) and pH (1492R) consistently produced multiple amplicons across all tested protocols. This prevented the determination of the nucleotide sequence of the amplicons and the  This study   SnW1r  ORF4  TGC CAT CTG GTG AAA TTG AA  This study   SnW2f  ORF5  TCC ATA AGT GGG CTC TCA CC  This study   SnW2r  ORF5  AGA GCC GAA CGC TTA TAT GG  This study   SnW3f  ORF8  CAC ACC TTC TGG AAT GCT GA  This study   SnW3r  ORF8  AAA GTT GCG CTA CCT GAT GG  This study identification of the amplified products through sequencing analysis without fragment separation or cloning. The Anaplasmataceae-specific primers, EHR 16SD and pH (1492R) and 16S 712F and 16S 1401R amplified two bands when tested with the positive control. These combinations of primers were not used further. Primers pA (27F) and EHR 16SR amplified the right target sequence in the positive control (i.e. Wolbachia 16S rRNA gene). However, when tested on S. noctilio, the amplicons obtained had high sequence similarity with Hymenoptera sequences. Primers 16S 567F and 16S 712R and 16S 567F and 16S 1401R amplified the right target sequence in the positive control (i.e., Wolbachia 16S rRNA gene). Amplicons from S. noctilio samples grouped with 16S rRNA gene sequences of bacterial species other than Wolbachia.
Primers Wspecf and Wspecr, Wsp 81 F and Wsp 691 R and ftsZf1 and ftsZr1, respectively, amplified the 16S rRNA, Wsp and FtsZ genes of Wolbachia in the positive controls, but did not amplify anything from S. noctilio samples.

Horizontal gene transfer from Wolbachia to S. noctilio
The first genome-wide searching method used to localize Wolbachia gene sequences used 14 Wolbachia genomes for a BLASTn analysis against the S. noctilio genome.
This search found open reading frames (ORFS) similar to Wolbachia gene sequences in scaffolds 13, 62, 126 and 1255 of the annotated genome of S. noctilio. The second method, that used taxonomic classification of genomic DNA reads from S. noctilio, found ORFs similar to Wolbachia gene sequences in seven scaffolds (scaffolds 13, 15, 62, 79, 106, 126 and 1224). The whole genome alignment using MUMmer identified scaffold 1 as potentially carrying Wolbachia gene sequences.
Using the scaffolds previously identified for a BLASTx against the protein database of NCBI showed that scaffolds 1, 62, 1224 and 1255 did not contain identifiable Wolbachia gene sequences. When restricting the reference database to Wolbachia protein sequences, the BLASTx analysis found similarity between a fragment of scaffold 1224 and two Wolbachia protein sequences. However, the percent identity (maximum value 44.38%) was lower than when the same fragment was compared to arthropod protein sequences (minimum percent identity 56.95%).
Across the scaffolds 13, 15, 79, 106 and 126, the BLASTx analysis found a total of 12 ORFs similar to Wolbachia gene sequences (Fig. 1). Eleven ORFs were either missing the 5' or the 3' end of the gene sequence, contained a premature stop codon or were fragmented across multiple reading frames. Only ORF8 was of the same length as the reference sequences. However, the percent identity was low (maximum percent identity 73.65%).

Ubiquity of horizontally transferred genes in S. noctilio in South Africa
The six primers designed in this study to amplify the horizontally transferred Wolbachia genes found in S. noctilio (i.e., SnW1f and SnW1r, SnW2f and SnW2r and SnW3f and SnW3r) all amplified the target loci. Primers SnW1f and SnW1r were arbitrarily chosen for the rest of the analysis. Out of the 500 samples collected from five South African populations, only 85 did not amplify after the first PCR, but showed amplification after dilution of the DNA samples.

Discussion
The aim of this study was to characterise Wolbachia in S. noctilio or Wolbachia genes in the genome assembly of S. noctilio. PCR was first used to demonstrate that S. noctilio is unlikely to be infected with Wolbachia, suggesting that the genes were introgressed in the S. noctilio genome. Through a genome wide search and a series of local BLASTx analyses, 13 potentially horizontally transferred Wolbachia genes were then identified. Using individual gene phylogenies, 12 were confirmed to be Wolbachia genes, while one was shown to be an arthropod gene. Finally, we demonstrated that these horizontally transferred Wolbachia genes are present in all populations of S. noctilio in South Africa.  None of the PCR protocols tested in this study lead to the amplification of the Wolbachia genes Wsp, FtsZ or 16S rRNA. This suggests the absence of a free living Wolbachia in S. noctilio in South Africa. The protocols tested included nine different primer pairs, three DNA extraction methods, two Taq polymerases and a total of four cycling protocols. Wolbachia-specific primers are known for their high false negative rates due to a high variability in gene sequences between Wolbachia strains [28]. For this reason, only three Wolbachia-specific primer pairs were tested in this study, namely Wspecf and Wspecr, Wsp 81F and Wsp 691 R and FtsZf1 and FtsZr1. The remaining eight primers, including pA (27F), EHR 16SR, EHR 16SD and pH (1492R) found in the literature [29][30][31][32] and 16S 576F, 16S 712F, 16S 712R and 16S 1401R that were designed in this study, either target all bacterial species or species within the Anaplasmataceae. The broader targeted species range of these primers was tested to account for the high sequence variability among Wolbachia strains and might be useful for future studies on Wolbachia infections. When tested on DNA extracted from the Wolbachia-positive A. pipithiensis, primers pA (27F) and EHR 16SR, 16S 576F and 16S 712R and 16S 576F and 16S 1401R amplified the 16S rRNA gene from Wolbachia. However, when tested on DNA extracted from S. noctilio, the same primers amplified non-target sequences.
A total of 12 Wolbachia genes were found in the genome of S. noctilio (Fig. 1). In total, the 12 confirmed Wolbachia gene sequences are distributed across five different scaffolds within the genome assembly. Out of the 12 genes identified, eleven are pseudogenes as they are spread across different reading frames or contain premature stop codons (Fig. 1). These results confirm that these genes were horizontally transferred from Wolbachia to the S. noctilio genome and that these horizontal transfers are not recent (i.e., due to extensive mutation of the gene sequence). Investigating the presence of these Wolbachia genes in other populations of S. noctilio or in related species could give an indication of the time frame within which these transfers happened.
The phylogenetic analysis gave a first indication of the original function of the horizontally transferred genes in Wolbachia. ORF1, ORF10, ORF11 and ORF12 were similar to protein sequences containing tetratricopeptide and ankyrin repeats. This category includes the phosphocholine transferase AnkX [33]. These repeats enable protein-protein interactions in eukaryotic cells [34]. In Wolbachia, these genes are part of the Wolbachia bacteriophages WO [20], a group of temperate double-stranded DNA phages that use Wolbachia as a host [35]. The genes that contain ankyrin and tetratricopeptide repeats are located in the "eukaryotic association module" of the bacteriophage WO genome [20] and are involved in host biology manipulation [21,22].
The phylogenetic analysis showed that ORF10 shares some sequence similarity with a latrotoxin related protein (Additional file 8: Fig. S8). Latrotoxins are an important component of the venom of the widow spiders in the genus Latrodectus [36]. However, C-terminal domain homologs of the latrotoxin gene are part of the "eukaryotic association module" of the phage WO [20]. Latrotoxin genes might have been acquired by WO bacteriophages through horizontal gene transfer and are now potentially used for eukaryotic host cell disintegration. The horizontal gene transfer of C-terminal domain latrotoxins from a Wolbachia strain to its host was also demonstrated in the genomes of the Wolbachia-positive Halyomorpha halys [18] and Aedes aegypti [20].
ORF2 and ORF4 were similar to IS4-family transposases. Insertion elements, such as the ones belonging to the IS4 family, are a type of transposable element widely distributed among bacterial genomes [37,38]. Their capacity to move to other loci in the genome is mediated by a transposase [39].
ORF3 and ORF5 both clustered with proteins of the recombinase family. Recombinases are proteins essential for genome replication in bacteria and are also crucial components of mobile genetic elements such as integrons, plasmids, transposons and bacteriophages [40]. Recombinases can lead to the integration of new DNA sequences in the host genome through strand exchange between the mobile genetic element and the target sequence in the host genome. ORF5 also clustered with a DNA invertase, a type of recombinase protein [41].
ORF9 showed sequence similarity to group II intron reverse transcriptases/maturases and RNA-directed DNA polymerases. These proteins indicate that ORF9 might be a specific type of reverse transcriptase found in bacteria, called retrointrons [42]. These types of retroelements can integrate into a DNA strand by binding to the host DNA as retrointron RNA and by being reverse transcribed into the target DNA strand [43].
ORF8 clustered with proteins with PQQ and YWTD domains. These domains are present in β-propeller proteins, a group of homologous proteins with a characteristic central "barrel" surrounded by a varying number of twisted β-sheets that form "blades" [44]. These proteins are found in viruses, bacteria, archaea and eukaryotes and assume a wide variety of functions [45]. The fact that ORF8 also clustered with a dehydrogenase indicated that, in Wolbachia, ORF8 could have taken part in the oxidation of methanol or ethanol, functions sometimes executed by proteins with a PQQ domain [44].
ORF6 and ORF7 both clustered with phage related proteins. While the function of ORF7 cannot be determined, ORF6 clustered with phage tail proteins. These proteins are the building blocks of the phage tail involved in adsorption to and infection of the bacterial host [46].
While further functional studies would be necessary to determine the exact functions of the 12 Wolbachia protein coding genes found in S. noctilio, the phylogenetic analysis gave a first indication of how these horizontal gene transfers occurred. ORF2, ORF3, ORF4, ORF5 and ORF9 seem to be genes directly involved in transposition of various types of mobile genetic elements, such as retrointrons, transposons and bacteriophages. These genes have the capacity to introgress themselves into new host genomes. On the other hand, ORF1, ORF6, ORF7, ORF8, ORF10, ORF11 and ORF12 do not have this capacity. ORF1, ORF6, ORF7, ORF10, ORF11 and ORF12 seem to be part of the Wolbachia bacteriophage WO while ORF8 does not seem to be part of any transposable element, but part of the core Wolbachia genome. In scaffold 13 ORF1, ORF6, ORF7 and ORF8 were found in the flanking regions of ORF5 (Fig. 1) indicating that these genes might have hitch-hiked with ORF5 from the Wolbachia genome to the genome of S. noctilio [22,46].
Horizontal gene transfers from Wolbachia to arthropod hosts putatively resulting in host genome evolution and expansion [47][48][49] and gene acquisition [12] events, have been observed in a number of studies. In S. noctilio, the fragments transferred from Wolbachia to the genome of the wasp are unlikely to have such impact. The fragments are relatively small, spanning a total of 8957 bp and have gone through substantial sequence variation.
Observing horizontally transferred Wolbachia genes in a Wolbachia-free insect species is interesting. These results demonstrate that the source population from which S. noctilio was introduced in South Africa carried Wolbachia at some point in its evolutionary history. This population could have lost the infection either prior to introduction in South Africa or after introduction and during the invasion process. An investigation into the presence of Wolbachia in native populations of S. noctilio would shed light onto the mechanisms that led to South African populations of S. noctilio to be Wolbachia-free.
It is possible that the source population from which S. noctilio was introduced into South Africa had lost Wolbachia before introduction. Werren and Windsor [32] and Bailly-Bechet et al. [50] have investigated the global equilibrium in Wolbachia incidence in arthropod species. They concluded that the loss of a Wolbachia infection is part of the Wolbachia-host interaction, and that arthropod species lose their Wolbachia infection more often than they acquire a new one. The mechanisms by which Wolbachia is lost still require investigation. There is evidence that once a Wolbachia strain is fixed into an arthropod population, the mechanisms by which it spread, such as cytoplasmic incompatibility, are relieved of their selective pressures and eventually erode [51]. Hornett et al. [52] have also shown that Hypolimnas bolina (Lepidoptera: Nymphalidae) evolved resistance against male-killing by a Wolbachia strain. Without a mechanism to efficiently spread through a population, Wolbachia could then slowly be removed from the host population.
It is possible that S. noctilio lost its Wolbachia infection over the course of the invasion process in South Africa or elsewhere. This phenomenon has been observed in the Argentine ant Linepithema humile after its introduction in Australia, Spain and France [53]. This loss could have happened through a founder effect. In South Africa, populations of S. noctilio were founded by a small number of individuals [54]. It is possible that none of the founding females carried Wolbachia. If the founding individuals carried Wolbachia, in such a small, introduced population, drift could have also led to a loss of infection through stochastic events. Finally, the Wolbachia infection could have been selected against during establishment and invasion. Environmental conditions such as temperature and nutrition affect Wolbachia titers in hosts, decreasing the capacity of the bacteria to get transferred from mother to offspring [55,56]. Because the population of S. noctilio was introduced with a very low genetic diversity, a Wolbachia strain causing cytoplasmic incompatibility could have also been selected against as it would prevent cross fertilization.
The mechanisms by which S. noctilio lost its Wolbachia-infection has implications for the potential use of Wolbachia as a biological control agent against S. noctilio. If S. noctilio lost its Wolbachia infection because the Wolbachia strain it used to carry was no longer able to induce reproductive parasitism, closely related species of wood wasps might carry Wolbachia strains which may still have this ability. These strains could be good candidates for a biological control program. However, if S. noctilio lost Wolbachia because the wasp evolved a resistance mechanism against the bacteria, reintroducing Wolbachia in S. noctilio would be more challenging. Thankfully, Wolbachia strains have very different effects on hosts. For example, ten strains of Wolbachia have already been artificially introduced in A. aegypti, a mosquito species that rarely carries Wolbachia in the wild [27,57]. Those strains have various effects on the reproductive biology, ecology and physiology of A. aegypti. As such, S. noctilio might be resistant to some Wolbachia strains but could be susceptible to others.
If S. noctilio lost its Wolbachia infection during invasion due to stochastic events related to the specific population dynamics of small populations, it might be possible to artificially introduce the Wolbachia strain from the population of origin into South Africa. Due to the distribution of pine trees in South Africa, the distribution of S. noctilio is patchy. This, along with the fact that S. noctilio is a haplodiploid species would slow down the spread of Wolbachia between populations [58,59]. However, this could be remedied through multiple releases of infected individuals. Finally, if S. noctilio lost its Wolbachia infection due to unfavourable environmental conditions, Wolbachia strains potentially present in other pine pests in South Africa could be of interest.

Conclusions
The presence of Wolbachia genes in the genome of S. noctilio suggests that S. noctilio is a potential host for Wolbachia. This could be determined by investigating the presence of Wolbachia in other populations of S. noctilio, either in the native range or in the introduced range. Because of its capacity to cause cytoplasmic incompatibility, Wolbachia has been investigated as a way to control mosquito populations [27] and might also help to control other insect pests in the future [60,61]. As such, Wolbachia could offer new solutions for the regulation of S. noctilio in the Southern Hemisphere.

DNA extraction
Three DNA extraction kits were tested on eggs and testes using the manufacturer's instructions. The prepGEM Insect DNA extraction kit (ZyGEM Corporation Ltd, Hamilton, New Zealand) was used on 14 male samples and two female samples, the Zymo Quick DNA Fecal/ Soil Microbe kit (Zymo Research, California, USA) was used on three male samples and the NucleoSpin DNA purification kit (Macherey-Nagel, Düren, Germany) was used on 13 female samples.

PCR
Wolbachia-specific primers previously designed in the literature have low success rates due to Wolbachia gene sequences being highly variable among Wolbachia strains [28]. For this reason, 14 different primers targeting the wsp, the FtsZ and the 16S rRNA genes were tested (Table 1 and associated references and Table 2). Ten primers were found in the literature [29-32, 63, 64]. Primers Wspecf, Wspecr, Wsp 81F, Wsp 691 R, ftsZf1 and ftsZr1 are Wolbachia-specific. Primers pA (27F) and pH (1492 R) are general bacterial primers and EHR 16SD and EHR 16SR are specific to the Anaplasmataceae.
Two Taq polymerases were used; KAPA Taq polymerase (KAPA Biosystems, Cape Town, South Africa), using the manufacturers instruction and MyTaq Taq polymerase (Meridian Bioscience, Cincinnati, USA). The total reaction volume of 25.5 µL contained 18.25 µL of Sabax water, 5 µL of MyTaq reaction Buffer, 0.5 µL of each primer diluted to 10 µM, 0.25 µL of MyTaq Taq polymerase and 1 µL of DNA (≈ 100 ng). The MyTaq Taq polymerase has a higher specificity than the KAPA Taq polymerase. The KAPA Taq polymerase would often amplify products when MyTaq Taq polymerase did not. However, the KAPA Taq polymerase also led to multiple product amplifications. A total of four different cycling protocols (Additional file 11: Table S1) were tested. From the amplified products 2 µL were mixed with 1 µL of 30X Gelred (BIOTIUL, Hayward, California, USA) and visualized using agarose gel electrophoresis on a 2% agarose gel using BioRad Gel Doc ™ Ez Imager and the software Image Lab 4.0.

DNA sanger sequencing
Amplicons were characterised through DNA Sanger sequencing. The PCR amplicons were purified using 6% Sephadex G-50 gel filtration (Merck KGaA, Darmstadt, Germany). The purified products were visualized on an agarose gel using the protocol described above. For sequencing, we used a 10 µL sequencing reaction volume containing 5.5 µL of PCR grade water, 1 µL of BigDye ™ (Applied BioSystems, Foster City, USA), 1 µL of sequencing buffer, 0.5 µL of primer diluted to 10 µM and 2 µL of purified PCR product. The cycling conditions included one cycle at 96 °C for 2 min, followed by 30 cycles of 30 s at 96 °C, 15 s at 50 °C and 4 min at 60 °C. Cycle sequencing products were purified using Sephadex G-50 gel filtration. Sequencing was performed on the ABI Prism ™ 3500xl automated DNA sequencer (Applied Biosystems USA, Foster City, California, USA) at the University of Pretoria sequencing facility. The reverse and forward sequences obtained were aligned on CLC Main Workbench 8 (Qiagen, Hilden, Germany) and the consensus sequence was used for a BLASTn analysis [68] against the NCBI nucleotide database [69].

Horizontal gene transfer from Wolbachia to S. noctilio
The S. noctilio genome assembly used in this study has been sequenced and assembled by Postma et al. (unpublished). Briefly, the S. noctilio genome was assembled and scaffolded into 6250 scaffolds using VelvetOptimiser [70] and SSPACE [71]. The genome assembly is estimated to be 185 Mb in size, with a N50 of 825 kb. The completeness of this genome assembly was estimated at 96.6% using BUSCO [72].

Local BLAST using Wolbachia genomes against the S. noctilio genome
The first approach used to locate putative Wolbachia sequences in the genome of S. noctilio was series of local BLAST [68] searches, using complete Wolbachia genomes as queries against the genome of S. noctilio. The complete genomes of 14 Wolbachia strains were downloaded from NCBI [69] (Table 4). BLASTn analyses were performed using the 14 Wolbachia genomes as query and the S. noctilio genome as a reference sequence (0.001 e-value cutoff ). The first BLASTn analysis only included eleven Wolbachia strains chosen either for the quality of their annotation or because their hosts belonged to the Hymenoptera family (i.e. wPip, wInc_Cu, wMel, wNo, GBW, wUni, wWitB, wNfla, wTpre) ( Table 4).
Subsequently, the genomic sequences from S. noctilio which exhibited significant similarity to Wolbachia were subjected to BLASTx analyses [68] against the NCBI protein database [69]. These sequences helped to identify four additional Wolbachia strains (i.e., wCauA, wCfeJ, wDi, wAna) ( Table 4) with a higher percent identity than the previously identified eleven strains. We then added the complete genomes of these four strains to that of the previous eleven and executed a second BLASTn analysis.

Taxonomic classification of S. noctilio sequence data
The second approach used to identify Wolbachia sequences in the S. noctilio genome was a taxonomic classification of genomic DNA reads from S. noctilio using Kraken 2 [73]. The DNA reads were compared to the standard Kraken2 database.

Whole genome alignment using MUMmer
The third approach used to identify Wolbachia sequences in the S. noctilio genome was a series of whole genome alignments using MUMmer [74]. The genome of S. noctilio was aligned to the genomes of four Wolbachia strains (i.e., wCauA, wCfeJ, wDi, wInc_Cu) ( Table 4).

BLAST of scaffolds from the S. noctilio genome against NCBI
The BLASTn [57] analysis and the taxonomic classification methods both identified scaffolds within the S. noctilio genome assembly that potentially contained Wolbachia sequences. To determine the position and length of these sequences as well as identify possible Wolbachia genes on the identified scaffolds, we used the full scaffolds for a BLASTx analysis [68] against the NCBI protein database [69]. This also allowed us to extract the DNA sequences of the horizontally transferred genes and to annotate them.

Phylogenetic relationships of candidate horizontally transferred Wolbachia genes
To confirm that the genes identified were transferred from Wolbachia and were not of eukaryotic origin, we constructed individual gene phylogenies. A BLASTx analysis [68] was performed against the protein database of NCBI [69]. The output of the BLASTx analysis was filtered by selecting sequences extracted from fully sequenced Wolbachia genomes. Whenever possible, the protein sequences used as outgroups were selected from bacterial species belonging to taxa outside of the alphaproteobacteria. However, for ORF1, ORF10, ORF11 and ORF12, similar sequences could only be found in other Wolbachia strains or in other Rickettsiales.
Each dataset was aligned in MEGA X: Molecular Evolutionary Genetics Analysis [65] using the Clustal W alignment tool and the default parameters. The sequences were then trimmed manually and the reference sequences that did not overlap with the sequences from the S. noctilio genome were taken out. A maximum likelihood analysis was performed in IQ-TREE 2 [75] using 1000 bootstrap replicates. The best substitution models were selected using ModelFinder [76]. The phylogenetic trees were edited in iTOL [77].

Ubiquity of horizontally transferred genes in S. noctilio in South Africa
Once the sequences of the horizontally transferred genes were identified, we used these sequences to design six primers using Primer3 4.0.1 [66,67] (Tables 1 and 2). These primers allowed us to screen for the presence of the horizontally transferred Wolbachia genes in various populations of S. noctilio in South Africa, and to confirm that those genes are ubiquitous in these populations. We sampled 100 individuals from five populations that correspond to five pine growing regions in South Africa; Western Cape, Southern Cape, Eastern Cape, KwaZulu-Natal and Mpumalanga. The sampling process was similar to previously described except for the fact that only males were sampled for this experiment. After dissection, the DNA was extracted using the prepGEM Insect DNA extraction kit (ZyGEM Corporation Ltd, Hamilton, New Zealand) and the PCR amplification was done using the KAPA Taq PCR kit (KAPA Biosystems, Cape Town South Africa) as previously described. The DNA purification process, visualization of the PCR amplicons and sequencing protocol are as described above.
To confirm that primers SnW1f and SnW1r, SnW2f and SnW2r and SnW3f and SnW3r were amplifying the desired Wolbachia sequences, the PCR amplicons from one female and from one male sample for each of the six different primers were sequenced. The sequences obtained were used for a BLASTn analysis [68] against the S. noctilio genome in CLC Main Workbench 8 (Qiagen, Hilden, Germany). Those samples were used as positive controls for the remaining PCRs. When visualizing the PCR amplicons using agarose gel electrophoresis, the presence of a band at the same height as the positive control indicated the presence of the horizontally transferred Wolbachia gene in the sampled individual. The quantity of DNA in the samples showing no bands was measured using a nanodrop and the DNA was then diluted to obtain a DNA concentration around 100 ng/nL.

Abbreviations
HGT: Horizontal gene transfer; ORF: Open reading frame.
Additional file 2: Figure S2. Maximum likelihood tree. It was constructed with the protein sequence of ORF2 compared to similar protein sequences of 12 Wolbachia strains and one protein sequence from Herpetosiphon llansteffanense (Terrabacteria: Herpetosiphonales) (out group). The branch indicated in red represents the position of ORF2 among other Wolbachia protein sequences. All Wolbachia strains are named after