Mitochondrial and microbial diversity of the invasive mosquito vector species Culex tritaeniorhynchus across its extensive inter-continental geographic range

Background Culex (Cx.) tritaeniorhynchus is an invasive mosquito species with an extensive and expanding inter-continental distribution, currently reported across Asia, Africa, the Middle East, Europe and now Australia. It is an important vector of medical and veterinary pathogens which cause significant morbidity and mortality in human and animal populations. Across regions endemic for Japanese encephalitis virus (JEV), Cx. tritaeniorhynchus is considered the major vector and has also been shown to contribute to the transmission of several other zoonotic arboviruses including Rift Valley fever virus (RVFV) and West Nile virus (WNV). Methods In this study, we used laboratory vector competence experiments to determine if Cx. tritaeniorhynchus from a Southern European population were competent JEV vectors. We also obtained samples from multiple geographically dispersed Cx. tritaeniorhynchus populations from countries within Europe, Africa, Eurasia and Asia to perform phylogenetic analysis to measure the level of mitochondrial divergence using the cytochrome oxidase subunit 1 ( CO1) gene. We also undertook bacterial 16S rRNA gene amplicon sequencing to determine microbial diversity and used multi-locus sequence typing (MLST) to determine any evidence for the presence of strains of the naturally occurring endosymbiotic bacterium Wolbachia. Results Cx. tritaeniorhynchus from a Greek population were shown be be competent vectors of JEV with high levels of virus present in saliva. We found a signficant level of mitochondrial genetic diversity using the mosquito CO1 gene between geographically dispersed populations. Furthermore, we report diverse microbiomes identified by 16S rRNA gene amplicon sequencing within and between geographical populations. Evidence for the detection of the endosymbiotic bacteria Wolbachia was confirmed using Wolbachia-specific PCR and MLST. Conclusions This study enhances our understanding of the diversity of Cx. tritaeniorhynchus and the associated microbiome across its inter-continental range and highlights the need for greater surveillance of this invasive vector species in Europe.

extensive and expanding inter-continental distribution, currently reported across Asia, Africa, the Middle East, Europe and now Australia.It is an important vector of medical and veterinary pathogens which cause significant morbidity and mortality in human and animal populations.Across regions endemic for Japanese encephalitis virus (JEV), Cx. tritaeniorhynchus is considered the major vector and has also been shown to contribute to the transmission of several other zoonotic arboviruses including Rift Valley fever virus (RVFV) and West Nile virus (WNV).

Methods
In this study, we used laboratory vector competence experiments to determine if Cx. tritaeniorhynchus from a Southern European population were competent JEV vectors.We also obtained samples from multiple geographically dispersed Cx. tritaeniorhynchus populations from countries within Europe, Africa, Eurasia and Asia to perform phylogenetic analysis to measure the level of mitochondrial divergence using the cytochrome oxidase subunit 1 (CO1) gene.We also undertook bacterial 16S rRNA gene amplicon sequencing to determine microbial diversity and used multi-locus sequence typing (MLST) to determine any evidence for the presence of strains of the naturally occurring endosymbiotic bacterium Wolbachia.

Results
Cx. tritaeniorhynchus from a Greek population were shown be be competent vectors of JEV with high levels of virus present in saliva.We found a signficant level of mitochondrial genetic diversity using the mosquito CO1 gene between geographically dispersed populations.Furthermore, we report diverse microbiomes identified by 16S rRNA gene amplicon sequencing within and between geographical populations.Evidence for the detection of the endosymbiotic bacteria

Introduction
The invasive mosquito vector species Culex (Cx.) tritaeniorhynchus has a wide and expansive distribution which includes populations in over 50 countries.Ranging across Asia, the Middle East and Africa (Jeffries & Walker, 2015), in recent decades it has been reported in Europe (Samanidou & Harbach, 2003), Eurasia (Gugushvili, 2002), Cape Verde off western Africa (Alves et al., 2014) and was recently recorded for the first time in Australia (Lessard et al., 2021).Cx. tritaeniorhynchus is the major vector of Japanese encephalitis virus (JEV) in Asia and the Pacific (Jeffries & Walker, 2015) and resulting JEV transmission leads to an estimated 50,000 -175,000 cases of human disease annually (Campbell et al., 2011;Erlanger et al., 2009).Estimates suggest JE clinical disease presentations account for only 1% of total viral infections, indicating overall occurrence of human JEV infections could be in the region of 5 -17.5 million each year, with almost four billion people living in 24 endemic countries at risk (Jeffries & Walker, 2015;Weaver & Reisen, 2010).There have also been recent cases of JEV detection in mosquitoes and birds in Italy (Platonov et al., 2012;Ravanini et al., 2012) and an autochthonous human case in Angola, Africa (Simon-Loriere et al., 2017), in addition to the recent JEV outbreak in Australia, highlighting the possibilities of future viral spread and establishment in novel regions and naïve populations (Gao et al., 2019;Lord, 2021;Weaver & Reisen, 2010).Cx. tritaeniorhynchus is also capable of transmitting several other significant zoonotic arboviruses including RVFV (Jupp et al., 2002) and WNV (Akhter et al., 1982;Fontenille, 1989;Reisen et al., 1982;Turell et al., 2006) with viral geographic distributions extensively overlapping with the current range of Cx. tritaeniorhynchus.
The widespread presence of this invasive species warrants further investigation to determine if outbreaks could occur beyond the current arboviral transmission zones.Cx. tritaeniorhynchus is well established in Southern European countries such as Greece and Albania but the capacity to transmit JEV has not been assessed to date.Although laboratory vector competence studies can measure the capability and efficiency of a mosquito vector population to transmit arboviruses under experimental conditions, the vectorial capacity in wild populations is influenced by multiple factors, including genetic diversity and the composition of the microbiota (Bolling et al., 2015;Cansado-Utrilla et al., 2021;Franz et al., 2015;Jupatanakul et al., 2014;Kilpatrick et al., 2010;Takken & Verhulst, 2013).Although several studies have investigated genetic diversity and vector competency in Cx. tritaeniorhynchus, they were mostly confined to populations within the same region or country -largely within the Asian continent (Ashfaq et al., 2014;Hayes et al., 1984;Luo et al., 2016;Philip Samuel et al., 2010;Sakai & Baker, 1972;Takahashi, 1980;Turell et al., 2006;Xie et al., 2021).
Mosquito vectors have associations with a wide diversity of microbes (Jupatanakul et al., 2014), including bacteria such as Wolbachia, Asaia, Serratia and Pseudomonas, which can affect vectorial capacity (Bolling et al., 2015;Minard et al., 2013).Some studies on the native microbiome or associated bacteria have been carried out in Cx. tritaeniorhynchus (Guo et al., 2016;Kittayapong et al., 2000;Tsai et al., 2004), but cross population microbiome composition studies comparing within and between geographically dispered populations has not been extenstive studied.In this study, we undertook laboratory vector competence experiments using an early generation colony of Cx. tritaeniorhynchus originating from Greece to demonstrate that European populations are competent vectors of JEV.We also obtained geographically dispersed Cx. tritaeniorhynchus from multiple populations (spanning four continents) to examine the mitochondrial and microbial diversity of this invasive and medically important vector species.Our results show evidence for mitochondrial divergence, a high level of microbial diversity and the presence of the endosymbiotic bacterium Wolbachia at only low prevalence and in only some geographic populations.

European population colonisation and JEV vector competence experiments
Field-collected Cx. tritaeniorhynchus larvae (~500) from Messolonghi, Greece were transported to the London School of Hygiene and Tropical Medicine for initiation of a colony.A range of techniques were employed to optimise insectary conditions for all mosquito life stages including large rearing cages (110 × 55 × 55 cm), light cycles with dusk/dawn simulation and swarm markers.JEV vector competence was assessed on the fourth generation at the Liverpool School of Tropical Medicine.Blood meals (heparinised human blood, NHS transfusion service, Speke) containing JEV (strain CNS138-11), to a final concentration of 6 log 10 plaque-forming units/mL, were provided for three hours, using a Hemotek membrane feeding system and an odourised feeding membrane, to 5-7-day-old adult females from which sugar sources had been withheld for 24 hours.Blood-fed females were incubated at 27°C, 70% humidity, for 14 days prior to collection of saliva using a forced salivation technique (Moreira et al., 2009).The head/thorax and abdomen were separated for surviving females after the 14-day incubation and the dissected body parts were stored for RNA preservation.RNA was extracted using Qiagen RNeasy kits from all saliva and body-part samples and tested by JEV-specific quantitative RT-PCR analysis (Yang et al., 2004) to determine infection rates.qPCR reactions were prepared using 5 μL of FastStart SYBR Green Master mix (Roche Diagnostics) with a final concentration of 1μM of each primer, 1 μL of PCR grade water and 2 μL template DNA, to a final reaction volume of 10 μL.Prepared reactions were run on a Roche LightCycler 96 System and amplification was followed by a dissociation curve (95°C for 10 s, 65°C for 60 s and 97°C for 1 s) to ensure the correct target sequence in addition to the inclusion of positive controls (RNA extracted from JEV strain CNS138-11) and no template controls (NTCs).

Mosquito field collections
Culex tritaeniorhynchus specimens were obtained from fieldcollections in Albania, Greece, Georgia, Ghana, Madagascar and Bangladesh (Figure 1).The locations, year, GPS co-ordinates and number of specimens obtained are shown in Table 1.Specimens were collected in Albania during country-wide  1).Maps were produced using Mapchart licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.(Jeffries et al., 2018b), and specimens were preserved in RNAlater with cold storage to prevent RNA degradation.Adult collections in Bangladesh were carried out in Sept-Nov 2013 from five sites within two districts in the Rajshahi Division in western Bangladesh.Within the district of Rajshahi, mosquitoes were collected from the upazilas (sub-districts) of Paba, Puthia and Bagmara, and within the Naogaon district, specimens were obtained from the upazilas of Manda and Mohadevpur.Samples were stored dry with silica gel desiccant to prevent microbial growth.

Morphological identification
Adult specimens were morphologically identified using keys appropriate for the geographic region (Günay et al., 2017;Lytra & Emmanouel, 2014b;Reuben et al., 1994;Robert et al., 2019;Samanidou & Harbach, 2003;Tantely et al., 2016), observing distinctive morphological features, including the presence of a clear white band on the proboscis, entirely dark wings and ringed tarsi (Samanidou & Harbach, 2003).Adult female mosquito physiological status was recorded and if wild-caught blood-fed, then the stage of digestion and time since blood-feeding was approximated using the Sella score method (Detinova, 1962;Martínez-de la Puente et al., 2013).prior to column-based extraction of the relevant phase using Qiagen DNeasy or RNeasy kits.RNA eluates were converted to cDNA using Applied Biosystems High Capacity cDNA Reverse Transcription kits.To confirm morphological species identification, gDNA or cDNA was used in broad-specificity barcoding PCRs, followed by Sanger sequencing and phylogenetic analysis (as detailed for mitochodrial diversity analysis below) to confirm species identification.
Intra-and inter-population mitochondrial diversity Mitochodrial diversity was assessed through Sanger sequencing of amplified PCR products from multiple assays targeting the cytochrome oxidase subunit 1 (CO1) gene (Bernasconi et al., 2000;Folmer et al., 1994;Kumar et al., 2007;Zittra et al., 2016) depending on the geographic region.Sub-sample testing was used to select a primer set (Kumar et al., 2007) producing a ~700 base pair (bp) product for screening all geographical populations.A primer combination to amplify the full length CO1 gene, binding at the 5' and 3' tRNA respectively, produced a ~1150bp sequence (Bernasconi et al., 2000;Shaikevich & Zakharov, 2010;Simon et al., 1994).

Consensus sequence and alignment assembly
Sequencing analysis was carried out in MEGA11 (  2. Phylogenetic tree construction and analysis Each alignment was examined using the "Find-Best-Fit Maximum Likelihood substitution model" for phylogenetic analysis and tree construction as previously described (Jeffries et al., 2021).Models used were the General Time Reversible model (Nei & Kumar, 2000) (GTR) or the Tamura three-parameter model (Tamura, 1992) (T92).The tree with the highest log likelihood is shown.The percentage of trees in which the associated taxa clustered together is shown next to the branches.Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Joining and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value.The trees are drawn to scale, with branch lengths measured in the number of substitutions per site.Codon positions included were 1st+2nd+3rd+Noncoding.
All positions containing gaps and missing data were eliminated.The phylogeny test was by Bootstrap method with 1000 replications.Evolutionary analyses were conducted in MEGA11 (Tamura et al., 2021).

Mitochondrial diversity and haplotype analyses
Mitochondrial diversity of populations was further assessed through mitochondrial diversity metrics, analysis of haplotypes, generation of haplotype networks, and pairwise comparison of genetic differentiation, including both study-generated and available reference CO1 sequences.Cx. tritaeniorhynchus CO1 alignments were analysed using DnaSP V6.12.03 (Rozas et al., 2017) to assess sequence polymorphisms and determine nucleotide and haplotype diversity.Haplotype networks were constructed within PopART (Leigh & Bryant, 2015) using the TCS inference method (Clement et al., 2000).
Bacterial 16S rRNA gene amplicon sequencing Microbiomes of selected specimens were analysed using barcoded high-throughput amplicon sequencing of the bacterial 16S rRNA gene.To enable analysis of the differences in microbiome between species (Cx.tritaeniorhynchus and concomitant species), physiological status (blood-fed or non-blood-fed) and geographic location (both intra-and inter-country) samples were selected from specific groups for comparison (Table 2).Mosquito specimens were surface sterilised prior to extraction, and negative controls comprising both DNA extraction and RNA extraction-Reverse Transcription blanks were included alongside the samples.Sequencing of each extract was undertaken using universal 16S rRNA V3-V4 region primers (FOR: 5'-CCTACGGGNGGCWGCAG-3', REV: 5'-GGACTACHVGGGTATCTAATCC-3') (Klindworth et al., 2013) in accordance with standard Illumina 16S rRNA metagenomic sequencing library protocols with the Nextera XT Index Kit v2 used to barcode samples for multiplexing.Sequencing was performed on an Illumina MiSeq, with the MiSeq v2 (500 cycle) reagent kit, with libraries sequenced as 250bp paired-end reads (PE).

Alpha and beta diversity analysis
Within the qiime2 phylogeny plugin, the "q2-phylogeny alignto-tree-mafft-fasttree" command was used, incorporating representative sequence artifacts from each of the gDNA and cDNA groups (rep-seqs output from DADA2) to produce rooted phylogenetic trees for diversity analysis.For each comparison set, using the respective filtered feature tables, alpha and beta diversity analysis was conducted through the qiime2 diversity plugin, using "qiime diversity core-metricsphylogenetic" (Faith, 1992).The sampling depth was selected from visualising feature tables ("q2-feature-table summarize") for each comparison, generating alpha rarefaction visualisations ("q2-diversity alpha-rarefaction") with "-p-max-depth" just over the median frequency per sample from the feature table, and then by balancing the number of features, with the number of samples from each group retained (Weiss et al., 2017).The diversity core metrics results were then generated and visualised using the relevant alpha or beta "group-significance" commands (Anderson, 2001;Kruskal & Wallis, 1952).Pairwise PERMANOVA tests with 999 permutations were used for comparisons between groups for the variable of interest and a significance level of P value <0.01 was used as the threshold.The metrics consulted for alpha (within-group) diversity were the Shannon diversity Index, Faith's phylogenetic diversity and the Evenness.For beta (between-group) diversity the metrics consulted were the Bray-Curtis, Unweighted-Unifrac (pairwise) and Weighted-Unifrac.

Differential abundance testing -ANCOM
To test for the presence of any differentially abundant taxa within each sample comparison group the analysis of composition of microbiomes (ANCOM) method was used within the qiime2 composition plugin (Mandal et al., 2015).The "q2-composition add-pseudocount" command was used, followed by "q2-composition ancom" with the relevant variable selected for each comparison, to investigate if any association may be apparent.Results were visualised in volcano plots, and assessed through the test statistic, W, to determine significance.
Wolbachia PCR detection and multi-locus strain typing Amplification of Wolbachia-specific 16S rRNA gene sequences was undertaken targeting the conserved Wolbachia 16S rRNA gene using W-Spec primers (438bp) (Werren & Windsor, 2000), in addition to a primer set designed for quantitiave PCR (target length: 102bp) (Gomes et al., 2017).Wolbachia MLST was undertaken to characterise Wolbachia strains using the sequences of five conserved genes (gatB, coxA, hcpA, ftsZ and fbpA) as molecular markers to genotype each strain (Baldo et al., 2006).In addition, an alternative primer set targeting a 271bp fragment of the ftsZ gene sequence in Wolbachia strains from Supergroups A and B was used on selected samples (de Oliveira et al., 2015).PCR reactions and Sanger sequencing of Wolbachia MLST PCR products were carried out as previously described (Jeffries et al., 2018a).Sequencing analysis was carried out in MEGA11 (Tamura et al., 2021), using methodology previously described, with consensus sequences used to perform nucleotide BLAST (NCBI) database queries, and for Wolbachia gene searches against the Wolbachia MLST database (http://pubmlst.org/wolbachia).Phylogenetic analysis of MLST gene locus sequences was performed as previously described for the mosquito CO1 gene.

Results
JEV vector competence of a Cx.tritaeniorhynchus Greek population After incubation for 14 days following an infectious bloodmeal, high levels of JEV (strain CNS138-11) were detected in both the abdomen and head/thorax of all (28/28) surviving females, indicating JEV was successfully acquired and disseminated (Figure 3).(Figure 4B) and 3) maximising the number of comparative sequences, with sufficient CO1 fragment overlap, 1007 sequences and 414 nucleotides which demonstrates complex phylogenetic relationships (Figure 5).The earliest common ancestors were sequences from India and China and Asian-only clades represent ~70% (673/1007) of sequences.Sequences from Georgia obtained in this study group most closely to sequences from Turkey and Kuwait (Figure 5 -blue inset sub-tree).The next monophyletic group with more geographically diverse sequences included a group from Madagascar (Figure 5, purple inset sub-tree).Another clade, developing from a group of Indian sequences, included sequences from Turkey, Ghana, Bangladesh, Albania and Greece (Figure 5 green inset sub-tree).Lineages with the greatest extent of genetic distance (using Cx. sitiens as an outgroup and Asian ancestral sequences from India and China) were from Australia (Figure 5 -yellow inset sub-tree), followed by Madagascar and Europe.

Global mitochondrial diversity metrics
Genetic diversity metrics were generated from 21 countries and 6 regions (Table 3).When all sequences were compared individually across the 414 nucleotide positions in 1007 sequences, 139 variable sites (S) and 444 haplotypes (h) were identified.The overall haplotype diversity (Hd) was 0.97864, the average number of nucleotide differences (K) was 9.16425, and the nucleotide diversity per site (Pi) was 0.02214.The    highest within-country nucleotide diversity per site was seen in South Korea (Pi = 0.03543) and lowest was in Greece (Pi = 0.00129).Asian sequences produced the highest nucleotide diversity per site (Pi = 0.02203), and European sequences had the lowest within-region diversity (Pi = 0.00184).
Haplotype networks and pairwise comparison analysis CO1 haplotype networks using full-length CO1 gene sequences (20 sequences, 1500 positions) (Figure 6A) suggested a linear pattern with sequences from Asia clustering separately to sequences from Eurasia, Europe and Africa (the latter being most divergent from the Asia haplogroup).Partial CO1 gene sequences (1007 sequences, 414 positions) (Figure 6B) produced a more complex haplotype network with one group from South Asia and two groups from East Asia.Haplotypes found in Australia and Timor-Leste appear to be branching from the large South Asian cluster.The haplotypes from Georgia appeared to branch from haplotypes present in Turkey and the Kuwait, sitting between the two large South, and East Asian foundational haplogroups.In a separate cluster, haplotypes from Greece and Albania branched off from the main South Asian haplogroup, linked alongside some haplotypes present in Turkey and Ghana.The haplotype present in Saudi Arabia also originated from this branch, and the haplotypes present in Madagascar then extended and diverged further from the end of this branch.Pairwise comparison analysis heatmaps for differences in genetic diversity within and between populations (Figure 7) revealed sequences from Australia as having the greatest difference to other countries/regions, and sequences from East Asia having the greatest intra-group differences.The matrix of pairwise fixation index (F ST ) indicates high genetic differentiation between populations in different countries and regions, particularly for the Australian, as well as the African and European groups.The divergence time between populations is also relatively lower for these populations.

Bacteria identified using microbiome analysis
We compared the composition and diversity of microbes present between collection locations or countries, physiological states (blood-fed or non-blood-fed), or between Cx. tritaeniorhynchus and concomitant mosquito species (Table 2).Species of potential relevance to mosquito biocontrol were detected including Wolbachia, Asaia, Serratia, Pseudomonas and Apibacter but presence and abundance levels were variable across individuals and populations.From 57 Madagascar samples, 34 had Apibacter (0.05-99.87% relative abundance), seven had Pseudomonas (1.0-29.92%),three had Asaia (1.30-34.06%)and one had Serratia (20.38%).Pathogenic bacteria were also detected including Escherichia shigella, Vibrio cholerae and species within the Bartonella, Anaplasma, Rickettsia, Mycoplasma, Enterobacter, Helicobacter and Providencia genera.Bartonella was found in an Albania sample at a relative abundance of 27.45%.Escherichia shigella was found in two Albanian samples (5.81% and 0.34% relative abundance) with this second sample also containing Vibrio cholerae at 6.28%.In Madagascar, Escherichia shigella was identified in five specimens with abundance ranging from 0.41-30.00%,Bartonella in one specimen with 45.32% abundance and this specimen also had Pseudomonas at 8.63% abundance.Anaplasma was also found in Madagascar samples with relative abundance ranging from 5.02-73.71%.Division down to taxonomic level 7 showed these ASVs were identified as Anaplasma marginale, Anaplasma platys and the rest classified within the Anaplasma genus.Mycoplasma was also identified in seven individuals (0.33-21.61%) and Escherichia shigella in six individuals from Madagascar (0.03-1.91%).

Microbiome diversity analysis
Concomitant species comparisons were possible for mosquitoes collected from Fier in Albania and Tsaramandroso in Madagascar.For Fier, gDNA samples from whole, non-blood-fed Cx. tritaeniorhynchus (n=16), Cx. pipiens (n=16) and Oc.caspius (n=16) specimens demonstrated variation in microbial composition.Alpha diversity within each species group showed no significant differences, however, beta diversity analysis showed clear differences between the species (Bray-Curtis p=0.001 and Weighted-Unifrac p=0.005).ANCOM identified Wolbachia as the only significant differentially abundant feature between the groups (W=265) due to high abundance in Cx. pipiens.From Tsaramandroso, cDNA extracted from abdomens of nonblood-fed female Cx.tritaeniorhynchus (n=12) and Cx.antennatus (n=14) demonstrated no significant difference in alpha or beta diversity and no significant differentially abundant taxa in ANCOM analysis.Comparisons were also made from abdomens of non-blood-fed and blood-fed female specimens collected from Fier (non-blood-fed n=15, blood-fed n=12) and Tsaramandroso (non-blood-fed n=12, blood-fed n=15).
For both countries, alpha-and beta-diversity and ANCOM highlighted no differences which were statistically significant.

Microbiome variation of intra-country Cx. tritaeniorhynchus populations
For intra-country comparisons specimens from two sites 35km apart in Bangladesh were compared: Paba (n=10) and Bagmara (n=7).No statistically significant differences were found through alpha-or beta-diversity, or ANCOM analysis.Within Albania, cDNA from non-blood-fed female abdomens also collected 35 km apart from Fier (n=15) and Vlore (n=10) were compared.The differences between individuals within these groups were found to be significant (Faith's phylogenetic diversity metric, p=0.0087), as well as between the groups (Unweighted-Unifrac, p=0.016) and Enterobacteriaceae was found to be significantly differentially abundant (ANCOM, W=156), with a higher abundance in Vlore than in Fier.

Inter-country microbiome comparisons
Samples from Bangladesh and Albania (gDNA, whole-body, non-blood-fed) were compared and alpha-diversity demonstrated no significant difference between individuals within each group, whereas beta-diversity highlighted differences between each country, with Bray-Curtis (p=0.001) and Unweighted-Unifrac (p=0.001).ANCOM analysis showed there were several differentially abundant taxa between the groups from Bangladesh and Albania.The three taxa which were most statistically significant were two Erwinia species (W=235 and W=224) and Asaia (W=219), with their abundance in Albania much greater than in Bangladesh.
For comparisons between Albania and Madagascar, one comparison was made between non-blood-fed and the other between blood-fed Cx. tritaeniorhynchus from the two countries (cDNA, abdomen samples).For non-blood fed, none of the alpha-or beta-diversity indexes showed significant results but ANCOM highlighted Anaerobacillus as a significant differentially abundant taxa (W=260), with higher abundance in Madagascar than Albania.This genus was present in 9/12 samples from Tsaramandroso, Madagascar (relative abundance range of <1% to 6.08%), present in 4/10 samples from Vlore, Albania, (<1% to 1.99%) and in 0/15 samples from Fier (Sop), Albania.For the comparison between blood-fed females across the two countries, alpha-diversity showed no significant difference between the individuals in each group; from Fier (Sop), Albania (n=12) and Tsaramandroso, Madagascar (n=15).The Bray-Curtis beta-diversity metric showed a significant difference between the groups from each country (p=0.003) and ANCOM highlighted that reads classified in the Bacillus genus were significantly differentially abundant (W=266), with a relatively high abundance in samples from Madagascar, and absent from the blood-fed mosquitoes from Albania.
Detection and phylogeny of Wolbachia bacteria strains Taxonomic abundance analysis from microbiome analysis did show evidence for the presence of Wolbachia in Cx. tritaeniorhynchus samples from Bangladesh, Albania and Madagascar.One sample from Bagmara, Bangladesh exhibited a relative abundance of Wolbachia comprising 39.77% of the total microbial composition, and two further specimens from the same location had 5.94% and 4.72% relative abundance.A blood-fed specimen collected in Fier (Sop), Albania, had a Wolbachia relative abundance of 22.50%.Three non-blood-fed samples from Albania, collected in Fier (Sop) and Vlore, also showed the presence of Wolbachia with relative abundances of 5.88%, 1.11% (Sop) and 1.32% (Vlore).Concomitant Albanian Cx. pipiens mosquitoes (a species known to be naturally infected with the wPip strain of Wolbachia) were shown to have variable Wolbachia relative abundances, ranging from 0-38.24%.Confirmation of Wolbachia-specific 16S rRNA amplification was possible for some samples from Albania and Bangladesh (Table 4).Further analysis using Wolbachia MLST showed a variable pattern of amplification and sequencing but partial MLST profiles were obtained for samples from Albania and Bangladesh (Table 5).Partial MLST allelic profiles and phylogenetic analysis (Figure 8) indicated that these  Wolbachia strains were different from one another but both placed within Supergroup B. Using the Wolbachia fbpA locus, the two strain do appear closely related (Figure 8).

Discussion
Despite the presence of Cx. tritaeniorhynchus first being reported in Albania in 1960 (Danielovi & Adhami, 1960), further published European occurrence reports were scarce until the 2000s when this species was recorded in Greece (Lytra & Emmanouel, 2014a, Lytra & Emmanouel, 2014b;Patsoula et al., 2017;Samanidou & Harbach, 2003).Recent extensive entomological surveys carried out in Albania (including this study) have identified Cx. tritaeniorhynchus within multiple areas for the first time (Alves et al., 2014;Gugushvili, 2002;Lessard et al., 2021;Samanidou & Harbach, 2003), highlighting a trend towards expansion of the known geographical range.Entomological surveillance in Europe has also identified other invasive mosquito vector species such as Aedes albopictus, highlighting the risk of exotic vector species becoming established (Medlock et al., 2012).Concurrently, there has been an increase in outbreaks and circulation of mosquitoborne arboviruses such as WNV in Europe (Bakonyi et al., 2013;Calzolari, 2016;Engler et al., 2013).To our knowledge, this study is the first to assess JEV vector competence in a European population of Cx. tritaeniorhynchus and our results emphasise the possibility of future introductions and JEV epidemics outside of Asia.Previous detection of JEV RNA in mosquitoes and birds in Italy further reinforces this point (Platonov et al., 2012;Ravanini et al., 2012).
In our study, a lower number of saliva samples (compared to mosquito body parts) had detectable virus after the 14-day incubation which might be expected, as the excretion of virus in the saliva is the final process in the infection pathway, following after viral acquisition and dissemination.Caution must be taken in over-extrapolating these results given 1) we were unable to generate comparative vector competence data between geographically dispersed populations due to colonisation difficulties and 2) direct extrapolation of laboratory vector competence to wild populations is likely to be imprecise given the complexity of transmission dynamics in wild populations.Comparisons to previous studies are also problematic due to variation in infection and detection methods.For example, a study in Korea resulted in 33-67% JEV transmission (via capillary tube saliva collection, or onward infection of chickens) (Turell et al., 2006) and in India, using ELISA in whole bodies, variable infection rates were reported, from 0-48% (Philip Samuel et al., 2010).The relatively lower infection rates seen in the latter study may be, at least in part, a result of reduced sensitivity of ELISA for virus detection or differences in JEV infectious doses.
Our study should enable a more accurate taxonomic classification of Cx. tritaeniorhynchus -particularly important as hybridisation within species complexes (e.g.Cx. pipiens) can influence arbovirus transmission (Shaikevich et al., 2016).The mitochondrial CO1 gene has been most frequently used (Ashfaq et al., 2014;Karthika et al., 2018;Kumar et al., 2007;Li et al., 2020;Rajavel et al., 2015;Xie et al., 2021) allowing species discrimination and investigation of maternal inheritance patterns (Cywinska et al., 2006;Hebert et al., 2003;Karthika et al., 2018).Our genetic diversity metrics quantified the genetic distances and divergence within population groups, identifying 444 haplotypes and 139 variable sites, with a haplotype diversity of 0.97864 and a nucleotide diversity per site of 0.02214.
To our knowledge this is the first published study examining such a geographically diverse mitochondrial dataset -previous studies have identified 28 (Karthika et al., 2018) & Emmanouel, 2014b).Our samples from the same location identified two haplotypes which were also present in Albania, and a further two haplotypes were found in Albania only.
The haplotype network and pairwise comparison analysis also indicated that (as expected) geographical location influences mitochondrial diversity.For example, there was a distinct grouping of 14 haplotypes in Madagascar, none of which were found in any other countries or regions.The greater genetic distances of some groups, such as Australia, Madagascar, and Europe, is logical but are genetic bottlenecks or selection pressures occurring during adaptation to new environments?Maternal lineages can provide insight into possible movement patterns and although Cx. tritaeniorhynchus are estimated to have an average flight distance of just under 70 metres, some studies have found that during long-distance wind-assisted dispersal, they can migrate 200-500 kilometres (Verdonschot & Besse-Lototskaya, 2014).As adults overwinter, this species may use a combination of long-distance migration and hibernation in situ, as strategies to survive unfavourable conditions in temperate regions (Min & Xue, 1996).The ability to disperse over such long distances and adapt to variable conditions is likely to provide more opportunities for range expansion and to increase gene exchange among different populations (Xie et al., 2021).
Our microbiome analysis revealed some bacteria such as Asaia that was present in all populations and differentially abundant in Albania when compared to Bangladesh.However, as Asaia can be environmentally acquired (Favia et al., 2007), it may depend on differing exposures in local habitats, rather than a country-wide distinction.In Albania, Cx. pipiens had a greater abundance of Wolbachia (ANCOM, W=265) but a lower abundance of Asaia than the other two species.
Although not statistically significant and some individuals had both Wolbachia and Asaia, further studies should be undertaken to determine any antagonistic associations as seen in mosquito lab colonies (Hughes et al., 2014;Rossi et al., 2015).
Apibacter was detected in Madagascar samples (both at high prevalence and for some indiviudals high relative abundance) and in individuals from Bangladesh and Albania.
Apibacter have been relatively recently isolated and classified in 2016 from various bee species (Kwong & Moran, 2016;Praet et al., 2016), as well as being reported in house flies (Park et al., 2019) and Cx.fuscocephala from Thailand (Thongsripong et al., 2021).These bacteria are thought to be beneficial endosymbionts with characteristics of adaptation to the gut environment and a degree of host-specificity (Kwong et al., 2018).Apibacter may also confer a degree of protection against pathogens with a recent study finding an association between Apibacter in bees and decreased infection by a trypanosomatid gut parasite Crithidia bombi (Kwong et al., 2018;Mockler et al., 2018).
Several genera which contain pathogenic species including Anaplasma, Rickettsia, Bartonella, Vibrio, Helicobacter, Providencia, Mycoplasma and Escherichia (Christou, 2011) were identified but it was not possible to classify the ASVs beyond genus level to determine whether they were pathogenic species.Escherichia shigella and Vibrio cholerae were identified and can cause dysentery and severe cholera respectively.These pathogenic bacteria may have been present in local aquatic environments and environmentally acquired.However, detection does not imply Cx. tritaeniorhynchus has the capacity for onward transmission, although it may theoretically be possible to mechanically disperse bacteria between water sources.Anaplasma marginale and Anaplasma platys, which can cause anaplasmosis in cattle and dogs respectively, are vector-borne pathogens, although they are mainly thought to be transmitted by ticks.A recent study in China found a wide range of Rickettsiales, including Anaplasma spp., in mosquito species, including Cx. tritaeniorhynchus (Guo et al., 2016).Phylogenetic analysis suggested a potential role for mosquitoes in vector-borne transmission of Anaplasma marginale, with other Anaplasma species suggested to be vertically transmitted (Guo et al., 2016).Finding these bacteria at relatively high abundance in blood-fed mosquitoes in our study, and not in the matched non-blood-fed mosquitoes may suggest these bacteria were present in the blood meals and not disseminated infections.Even without vectorial capacity, mechanical transmission during blood-feeding may be possible and the high abundance in blood-fed-females would suggest a relatively frequent exposure to Anaplasma, particularly Anaplasma marginale, from feeding on cattle.Bacteria from the Bartonella genus are also vector-borne, with ticks, fleas, lice and sandflies implicated as vectors (Billeter et al., 2008;Jacomo et al., 2002).Bartonella species can infect humans and animals causing bartonellosis.Although Bartonella was only identified in a few Cx.tritaeniorhynchus individuals from each country, when present it was highly abundant.Bartonella was found dominating the microbiome in 4/13 concomitant Cx. antennatus specimens in Madagascar despite none found in the matched group of Cx. tritaeniorhynchus from the same location.
Our microbiome analysis identified the presence of Wolbachia in populations with variable levels of relative abundance (0-40%) suggesting the likely presence of multiple strains at low prevalence spread across several continents.Phylogenetic analysis of the Wolbachia 16S and MLST gene loci confirms the strains from individuals in Albania and Bangladesh are placed within Supergroup B. Although the strains do differ from one another where comparison was possible on the Wolbachia fbpA locus, they appear to be closely related.Some previous studies have not identified Wolbachia in Cx. tritaeniorhynchus (Kittayapong et al., 2000;Nugapola et al., 2017;Ravikumar et al., 2010;Tsai et al., 2004), however, a study from Thailand (Wiwatanaratanabutr, 2013) and recently from Singapore (Ding et al., 2020), reported Wolbachia in small numbers of individual mosquitoes.Cx. tritaeniorhynchus is implicated as a vector of Dirofilaria immitis, a filarial nematode with an obligatory symbiotic relationship with Wolbachia, requiring its presence for survival.However, phylogenetic analysis carried out in our study indicates the Wolbachia strains are not likely to result from filarial infections given the placement within Supergroup B (Dyab et al., 2016).It remains to be determined whether these Wolbachia strains influence reproductive success through the cytoplasmic incompatibility phenotype, are vertically transmitted with high rates of maternal transmission, or impact vectorial capacity.Low infection rates may suggest they do not possess the beneficial phenotypic characteristics most useful for mosquito biocontrol and may not inhibit arboviruses as seen in some other studies on native strains (Tsai et al., 2006).However, low-level natural Wolbachia strains in Cx. tritaeniorhynchus populations are unlikely to be prohibitive to the development of Wolbachia-based biocontrol strategies, such as through transinfection of non-native strains.Further analysis of larger sample numbers from diverse geographical areas is needed including non-PCR based methods such as microscopy to visualise bacteria in mosquito tissues (Walker et al., 2021).

Conclusions
Our study provides evidence that a European population of Cx. tritaeniorhynchus is a competent vector of JEV and shows that a high degree of mitochondrial and microbial diversity is present across geographically dispersed populations of this species.Future vector competence experiments should incorporate factors, such as temperature, to more closely mimic the current and potential future environmental niches inhabited by this vector.Establishing whether geographically diverse populations of Cx. tritaeniorhynchus are competent vectors for other medically important arboviruses such as WNV and RVFV is also a priority as exemplified by it's involvement in the first incursion of RVFV outside of Africa.This study also provides evidence for the presence of diverse bacteria, including those considered candidates for biocontrol, and pathogenic bacteria.Further studies should be undertaken to determine if Wolbachia strains naturally reside within Cx. tritaeniorhynchus populations.Increased surveillance and novel control strategies for Cx.tritaeniorhynchus will be important with climate change increasing the range of this invasive species which can transmit multiple arboviruses of concern to public health.

Yasmina Martínez Barciela
Universidade de Vigo, Vigo, Spain The work is focused on the study of three aspects related to Culex tritaeniorhynchus: the vector competence of Greek populations to transmit Japanese encephalitis virus (JEV) under laboratory conditions (i), the phylogenetic relationship between populations from different continents by mitochondrial analysis using the cytochrome oxidase subunit 1 (CO1) gene (ii), and the microbial diversity (such as Wolbachia) naturally present in the different populations by bacterial 16S rRNA gene amplicon sequencing and multi-locus sequence typing (MLST).
The results obtained are novel and of interest not only in phylogenetic terms, but also regarding vector surveillance and environmental health.The manuscript is well developed in easily understandable English.
Regarding the structure, the different sections are discussed: Abstract.It is well structured.A spelling mistake is detected in the form of the verb "to be" in the first line of the results.Introduction.Appropriate length and content.It is recommended not to abbreviate Rift Valley fever virus (RVFV) and West Nile virus (WNV) the first time they are used.When the species Culex tritaeniorhynchus is mentioned at the beginning of a sentence, the criteria should be unified: either with or without the abbreviation for the genus (Cx.) (it is recommended not to abbreviate it in these cases).Methods.Appropriate structure and explanations apparently sufficient to be able to replicate the method of molecular analysis.However, the information on the capture of adult specimens in the field is not sufficiently well specified (what type of traps are used in each case?).Likewise, it would be useful to add the sampling period including the months of the year and not only the years of sampling.
Results: Well structured and adequately explained for their understanding.Discussion.Well supported by the bibliography and easy to read.When mentioning the pathogenic species found in Culex tritaeniorhynchus, it would be interesting to clarify in which populations (countries) each pathogens were found (this information is not clarified in all cases).Conclusions.Appropriate length and content.References.Adequate length and updated content.No errors have been detected.
Regarding the quality of this review, it is important to point out that it may be necessary to have experts in phylogenetics, molecular analysis and statistics to make a more appropriate assessment of the study.

Evangelina Muttis
Investigadora Asistente en CEPAVE, Buenos Aires, Argentina The study comprises three distinct analyses of Culex tritaeniorhynchus, the primary vector of Japanese Encephalitis Virus (JEV), across six countries and four continents: genetic divergence of COI sequences, bacterial diversity determined by amplification of the bacterial 16S rRNA gene, and studies of vector competence within the mosquito population for JEV transmission.
Firstly, significant genetic divergence of the COI gene among populations was demonstrated.The provided information serves as a dataset for future research endeavors in vector biology or taxonomic studies.
Secondly, the investigation into microorganisms associated with this mosquito vector holds relevance due to potential interactions with arboviruses.Therefore, data on certain populations naturally infected with Wolbachia represent an initial stride towards considering vector control strategies.
Lastly, studies on vector competence revealed that Greek Cx. tritaeniorhynchus mosquitoes are proficient in transmitting JEV, at least under laboratory conditions, exhibiting a high infection rate.
While the content of this article remains pertinent, several inquiries emerge from it.For instance, are there correlations among the factors under study?Is microbial diversity associated with mitochondrial diversity?Can JEV outbreak zones be compared with the factors studied?In my perspective, the various topics addressed should be interconnected in some manner.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate?I cannot comment.A qualified statistician is required.
Are all the source data underlying the results available to ensure full reproducibility?Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Medical entomology I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Reviewer Report 15 April 2024 https://doi.org/10.21956/wellcomeopenres.22976.r76789 analyses underlying conclusions as supplementary tables.While they posted their raw amplicon data in OSF.IO database, I could not see metadata for these fastq.gzfiles -there is no information on which file corresponds to which sample.I strongly feel that these data and complete metadata need to be deposited in a dedicated repository such as NCBI SRA archive.
Further, I have an issue with the use of controls.The authors do not explain how the controls were used in mosquito experiments.What is the background signal in qPCR assessment of viral infections?What is the likelihood that some of the virus-negative results for mosquito saliva samples were caused by the failure of saliva collection in the first place?Likewise, the authors mention negative controls in their microbiome study, but do not explain how they were used.
Negative controls can be critically important for controlling the reagent and laboratory contamination, but the authors need to explain their use clearly, and ideally illustrate what they did.
Specific comments: Table 1 is not helpful if you do not provide information on how the samples were used.For example, 613 larvae collected in Albania in 2017 sound impressive enough -but were any of them used for any experiments/analyses at all? Please provide details of PCR reactions: reaction mix composition, cycling details Figure 2: I do not think that providing the full nucleotide sequence of mosquito COI gene is useful, and definitely not as the main-text figure.On a related note, despite this figure, I struggled to understand how specimens were selected for COI amplification, and which primers, and for which region, were ultimately used for the amplification.Please try to clarify these explanations.How was surface sterilization done?Table 3 and Figure 7: I have not found them informative or helpful.Move them to Supplement? Discussion of the microbiota: you focus on mammalian-pathogenic microbes, but please consider that many of those listed are known to live in a much broader spectrum of environments.For example, some strains of Bartonella are members of the healthy gut microbiota of honey bees, and about many other strains, we do not know!

Is the work clearly and accurately presented and does it cite the current literature? Partly
Is the study design appropriate and is the work technically sound?Yes

Are sufficient details of methods and analysis provided to allow replication by others? Partly
If applicable, is the statistical analysis and its interpretation appropriate?Yes Are all the source data underlying the results available to ensure full reproducibility?

Figure 1 .
Figure 1.Locations of collection sites for Cx.tritaeniorhynchus samples analysed in this study.Yellow dots represent estimated collection locations within the six countries in our study (GPS co-ordinates, collection methods and number of specimens are shown in Table1).Maps were produced using Mapchart licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Figure 2 .
Figure 2. CO1 primer sets and sequence alignment positions.Excerpt of nucleotide sequence from the complete mitochondrial reference genome of Cx. tritaeniorhynchus (NC_028616).Red, underlined text -primer binding sites o background -nucleotides comprising the full CO1 gene; Blue, underlined -Kumar et al. primer binding sites; Blue background -Position of alignment (a) of sequence set binding regions; Text in bold -Location of alignment (b) maximizing the length and with almost full CO1 gene coverage; Green text -Location of alignment (c) maxi included.

Figure 3 .
Figure 3. JEV vector competence experiment on European Cx. tritaeniorhynchus colonised from Greece.Scatter dot plot of quantitative PCR data.Horizontal bars represent mean JEV copies/µl per group.Boxed numbers show the number of JEV positive samples / total number of samples tested per group.

Figure 6 .
Figure 6.CO1 haplotype networks for Cx.tritaeniorhynchus.(A) Full CO1 gene haplotype network for Cx.tritaeniorhynchus (maximising the length of sequences).(B) Global partial CO1 haplotype network for Cx.tritaeniorhynchus (maximising number of reference sequences).Haplotype networks were constructed using the TCS network method in PopArt (Leigh & Bryant, 2015) with nodes coloured according to country-of-origin.

Figure 7 .
Figure 7. Global genetic diversity Country and Regional populations pairwise comparison heatmaps.Cx. tritaeniorhynchus partial CO1 maximum reference sequences alignment used for haplotype and pairwise analyses within Arlequin software and visualised in R. F ST = pairwise fixation index: 0=two populations genetically identical, 1=two populations are genetically different; maximum genetic diversity between two populations.

Figs 4- 5 :
Two full-page figures showing multiple trees based on different-length regions of the COI gene are excessive -especially if similar results are also shown in a (more informative, in my view) network plot!Please decide on what way of showing these results would be the most useful for conveying your message.The current set of low-resolution trees, including many duplicates for site/sequence, and with multiple colored inserts, is not really accessible.

Table 1 . Cx. tritaeniorhynchus collection locations and samples collected. Latitude and
(Orsborne et al., 2019)re in decimal degrees.NBFF = non-blood-fed females, BFF = blood-fed females, *denotes from field-collected larvae.entomologicalsurveys in 2015-2016, in addition to focused field-work collections in 2017, including in the rural village of Sop in the Fier district, south-western Albania.A variety of adult trapping techniques and larval dipping were used.Specimens were preserved in Invitrogen RNAlater at -20°C followed by -70°C long-term storage.Field-collected fourth instar larvae were obtained from three sites in south-eastern Georgia in September 2015; using larval dipping in semi-permanent water bodies with vegetation, then stored in 70% ethanol.In Ghana, specimens were collected as adults from Dogo village in the Greater Accra region in June 2017(Orsborne et al., 2019).

Table 3 .
Cx. tritaeniorhynchus population genetic diversity metrics.Total number of sites: 414.n: Number of samples; S: Number of variable sites; h: Number of haplotypes; Hd: Haplotype diversity; K: Average number of nucleotide differences; Pi: Nucleotide diversity (per site); PiJC: Nucleotide diversity (Jukes-Cantor).

Table 5 . Wolbachia partial MLST gene allelic profiles for resident strains in Cx. tritaeniorhynchus populations
. "CM" = Allele number of the closest allelic match, with the number of nucleotide differences in brackets."EM" = Exact match on that locus to the allele number provided."SG" = Super group to which isolates with that allele at that locus belong.*denotes where the query sequence was truncated, therefore the full locus wasn't available for comparison."-" denotes where sequencing was attempted from PCR products but the sequence data quality wasn't sufficient for analysis."NS" denotes where no clear PCR amplified product was obtained and therefore sequencing was not attempted.
and 303(Xie et al., 2021)haplotypes, with the latter finding a Hd of 0.97 and Pi of 0.02434 which is comparable to our study.Analysis of regional population groups identified 412 haplotypes in Asia, four in Australia, 19 in Africa, four in the Middle East, eight in Eurasia and four in Europe.The only previous study for European Cx. tritaeniorhynchus found two haplotypes within the same population, collected in a single rice field in western Greece (Lytra

the work clearly and accurately presented and does it cite the current literature? Yes Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Partly If applicable, is the statistical analysis and its interpretation appropriate?
I cannot comment.A qualified statistician is required.

all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.Reviewer Expertise: Entomology, vectorial ecology, environmental health I

confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.