Diverse novel Wolbachia bacteria strains and genera-specific co-infections with Asaia bacteria in Culicine mosquitoes from ecologically diverse regions of Cameroon

Background: The endosymbiotic bacterium Wolbachia infects numerous species of insects and Wolbachia transinfection of Aedes mosquito species is now being used for biocontrol programs as Wolbachia strains can both inhibit arboviruses and invade wild mosquito populations. The discovery of novel, resident Wolbachia strains in mosquito species warrants further investigation as potential candidate strains for biocontrol strategies. Methods: We obtained mosquito specimens from diverse Culicine mosquitoes from Cameroon including ecologically diverse locations in the Central and West Regions. Wolbachia prevalence rates were assessed in addition to the environmentally acquired bacterial species Asaia in major Culicine genera. PCR-based methods were also used with phylogenetic analysis to confirm identities of host mosquito species and Wolbachia strains were classified using multi-locus sequence typing (MLST). Results: We report high Wolbachia prevalence rates for Culicine species, including in a large cohort of Aedes africanus collected from west Cameroon in which 100% of mono-specific pools were infected. Furthermore, co-infections with Asaia bacteria were observed across multiple genera, demonstrating that these two bacteria can co-exist in wild mosquito populations. Wolbachia strain MLST and phylogenetic analysis provided evidence for diverse Wolbachia strains in 13 different mosquito species across seven different genera. Full or partial MLST profiles were generated from resident Wolbachia strains in six Culex species ( quinquefasciatus, watti, cinerus, nigripalpus, perexiguus and rima), two Aedes species (africanus and denderensis) and in Mansonia uniformis, Catageiomyia argenteopunctata, Lutzia tigripes, Eretmapodites chrysogaster and Uranotaenia bilineata. Conclusions: Our study provides further evidence that Wolbachia is widespread within wild mosquito populations of diverse Culicine species and provides further candidate strains that could be investigated as future options for Wolbachia-based biocontrol to inhibit arbovirus transmission.


Introduction
Wolbachia are endosymbiotic bacteria which reside within an estimated 40-70% of insect species 1,2 .These bacteria have been detected in numerous mosquitoes that transmit human pathogens including species within the Aedes (Ae.), Culex (Cx.) and Anopheles (An.) genera [3][4][5][6][7][8][9] .Wolbachia is now being used in applied mosquito biocontrol strategies as strains inhibit arboviruses and invade mosquito populations using the reproductive phenotype cytoplasmic incompatibility (CI).Wolbachia-induced CI was first used in mosquito release control trials in the 1960s in attempts to supress Cx. quinquefasciatus populations in Myanmar 10 .More recently, Wolbachia-infected Aedes males have been released to induce CI and the associated sterility from matings with wild-type females, resulting in inviable progeny.This method of suppressing the population, the incompatible insect technique (IIT), has seen field trials for species such as Ae.polynesiensis that contain natural resident strains 11,12 .Embryo microinjection has also resulted in transinfected strains in Ae. aegypti [13][14][15][16] .These transinfected strains, including wMel, have been shown to invade wild mosquito populations and also inhibit major arbovirus transmission, such as dengue virus (DENV) 13,15,[17][18][19] .The wMel Ae. aegypti line, through release programmes, is now present in more than 10 countries and encouragingly, a randomised control trial in Indonesia resulted in a 77% DENV inhibition 20 .The wMel strain is being continually released into additional DENV endemic countries, and based on mathematical modelling, has the capacity to reduce the DENV R0 (basic reproduction number) from 66-75% 21 .The wMel strain also inhibits other medically important arboviruses such as chikungunya virus (CHIKV) 22 , Yellow fever virus (YFV) 23 and Zika virus (ZIKV) 24 .There are other Wolbachia strains also being used in applied strategies with the wAlbB strain 25 now established in Malaysian Ae. aegypti populations and having an impact on dengue incidence 26 .
Wolbachia strains that naturally reside within mosquito populations can provide important comparative data to complement biocontrol strategies.For example, whether Wolbachia strains in natural populations are found at a high prevalence, and whether strains are capable of inducing CI to allow rapid population invasion.This first requires the identification of strains using molecular strain typing from diverse species which would then allow more comprehensive phenotypic characterization.Intra-genera transinfection has been shown to be successful without the need for adapting Wolbachia strains to insect cell lines so identifying strains residing in natural mosquito populations could provide additional candidate strains for biocontrol strategies.The presence of other bacteria, such as the a-Proteobacterium Asaia, has been postulated to compete with Wolbachia to colonise the reproductive tissues of mosquitoes, including the ovaries in females 27,28 .Asaia has also been proposed for biocontrol strategies as this genus of bacteria forms stable associations with numerous insects that sugar feed 29 and can rapidly colonize the midgut and spread to other insect tissues after ingestion from either a sugar or blood meal 30 .Asaia is particularly well studied in Anopheles (vectors of malaria parasites) and can stably associate with multiple species and be the dominant bacteria within some species such as An.stephensi 31 .Unlike Wolbachia, Asaia can be cultured in cell free media and has been genetically transformed 31 .Asaia can also be both horizontally and vertically transmitted providing a mechanism to invade mosquito populations.
Therefore, examining the possibility of co-infections in diverse mosquito populations will investigate if any antagonistic associations between these two common bacteria are present, as has been demonstrated in lab populations 27,28 .Numerous studies which have detected natural Wolbachia strains in Culicines have undertaken analysis of the 16S rRNA gene when looking at the wider microbiome 32,33 .However, a more targeted approach amplifying Wolbachia-specific genes is required to confirm a resident strain is present and phylogenetic analysis allows any newly discovered strains to be compared to existing strains.This then allows a more comprehensive assessment of Wolbachia tissue tropism (microscopy), whole genome sequencing and ultimately phenotypic characterisation.Wolbachia strains can be classified in Supergroups designated from A to H. A and B Supergroup strains are mostly found in arthropods, with only one strain per host, or multiple strains infecting the same host 34 .Superinfection, the infection of more than one strain of Wolbachia in the same host -as can be seen in Ae. albopictus with the wAlbA and wAlbB strains 35 -can comprise strains of differing Supergroups, such as A and B group strain co-infections.
Cameroon is a country in West Africa in which outbreaks of arboviral diseases including DENV, YFV, CHIKV and Rift Valley fever virus (RVFV) 36 have been reported.For example, DENV IgM seroprevalence among febrile children was 14%

Amendments from Version 1
We have taken into consideration all the very relevant reviewer comments and significantly improved all sections.Specifically, we have adjusted our title to reflect the 'genera-specific' associations seen between Wolbachia and Asaia prevalence rates, included more background information on using additional Wolbachia strains for dengue control in Malaysia and significantly increased the rationale for investigating Asaia-Wolbachia interactions by providing more context on previous studies looking at Asaia in mosquitoes.We have provided additional details on qPCR methodology and added an additional section in the methods for statistical analysis.Figure 2 now contains a more universally accessible colour scheme and we have significantly expanded our discussion to address some very valid and important reviewer queries.We have included more discussion around how our Wolbachia and Asaia prevalence rates results are more reflective of a 'genera-specific' pattern that is comparable to other previous studies and have included some limitations based on the challenging nature of mosquito morphological identification.Furthermore, we have expanded our discussion to highlight the limitations of our results based on both using whole bodies (rather than tissue-specific detection) and using monospecific pooling.We have now clearly defined what we would categorise as a 'novel' Wolbachia strain and acknowledged that even MLST allelic profiling may not represent the optimal loci to capture strain variation (despite being widely used to determine Wolbachia strain diversity).
Any further responses from the reviewers can be found at the end of the article from a cross a cross-sectional study performed in 961 children less than 15 years old attending public hospitals of Cameroon 37 .Cases of arboviruses are reported throughout the country and there is concern that rapid urbanisation seen throughout Africa could exacerbate transmission through favouring the breeding conditions of urban-adapted mosquito species 38 .Deforestation has a significant effect on the abundance and diversity of Cameroon mosquito species and could lead to spill-over transmission of additional circulating zoonotic viruses such as Semliki Forest, o'nyong'nyong and Bwamba viruses 39 .Here, we obtained specimens from entomological surveys undertaken in Central and West Cameroon to determine if diverse Culicine species harbour resident Wolbachia strains and if co-infections exist with Asaia, given this bacterium can be environmentally acquired and can compete with Wolbachia.We extracted RNA from preserved samples and undertook qRT-PCR analysis to make any detection of Wolbachia strains more likely to be from actively expressed Wolbachia genes.A combination of phylogenetic analysis and strain typing using multi-locus sequence typing (MLST) revealed a diversity of newly discovered Wolbachia strains.

Mosquito collections
A variety of adult mosquito trapping methods were used in the Central Region of Cameroon in Yaoundé (3°52'22.2"N,11°30'38.0"E)and Olama village (3°24'45.0"N11°17'03.0"E)from June 2019 to July 2019.Yaoundé (the capital of Cameroon) is an urban location characterised by mosquito breeding sites such as tyres, containers, and temporary water pools.In contrast, Olama village is rural, located along the Nyong River (Figure 1) and is characterised by houses built with corrugated metal roofs and mud walls.In Olama village, eight CDC miniature light traps were hung in four houses and two Stealth traps (model 14 which uses ultraviolet light for attraction) were hung in one house (one trap inside and one outside).BG Suna traps containing a carbon dioxide source from fermented yeast as an attractant were hung in trees nearby houses, approximately one metre above the ground.Traps were powered using a 12V battery over a period of 11 days from 17:00-07:00.Human landing catches were carried out in Olama Village as previously described 4 .In Yaoundé, BG Sentinel-2 traps containing a BG Sweetscent lure used as an attractant were used to collect mosquitoes for 15 days from 16:30-10:00.Traps were assembled using manufacturers guidelines and positioned nearby potential Aedes breeding sites under tree coverage.Larval collections using ladles and sieves were also carried out in Yaoundé́ in typical urban breeding sites from the following districts: Etoude, Nkolbissim and Briqueterie.
Mosquitoes were collected in the West Region from the Menoua division as previously described 40 and from two locations in the Dschang sub-division in a rural area; Fonakeukeu (05°24′73′′ N, 010°04′79′′ E) and a peri-urban area; Toutsang (05°25′65′′ N, 010°04′05′′ E).Collections were carried out in both the rainy season (from March to September 2019) and dry season (from November 2019 to February 2020) using sweep nets.Mosquitoes from Central Cameroon were morphologically identified to genera level, and from West Cameroon to species level, using morphological identification keys, then preserved in RNAlater (Invitrogen) and kept at below -20°C, prior to being sent for molecular analysis at LSHTM.

RNA extraction and generation of complimentary DNA (cDNA)
From West Cameroon collections, mosquitoes of the same species, from the same location and season, were pooled prior to RNA extraction (650 mosquitoes, 192 pools, pool size range was 3-5 mosquitoes).From Central Cameroon collections, a subsample (n=576) was selected for molecular analysis based on diversity of genera and collection sites.Mosquito RNA was extracted from either pools or individuals using Qiagen 96 RNeasy Kits and a Qiagen Tissue Lyser II (Hilden, Germany) with 5 mm stainless steel beads (Qiagen) to homogenize mosquitoes.Resulting RNA was eluted in 45 µL of RNase-free water and stored at −70°C.cDNA was generated from RNA using an Applied Biosystems High-Capacity cDNA Reverse Transcription kit.Each reaction (20 µL) was made up of the following: 1 µL reverse transcriptase, 2 µL 10X RT buffer, 2 µL 10X random primers, 0.8 µL 25X dNTP (100 mM), 4.2 µL nuclease-free water and 10 µL RNA.A Bio-Rad T100 Thermal Cycler was used to generate cDNA as follows: 25°C for 10 minutes (min), 37°C for 120 min; 85°C for 5 min with all resulting cDNA stored at -20°C.

Wolbachia and Asaia detection
The detection of Wolbachia strains was initially carried out by amplification of three conserved Wolbachia genes; 16S rRNA gene using primers W-Spec-16S-F: 5'-CATACCTATTCGAAG-GGATA-3' and W-Spec-16s-R: 5'-AGCTTCGAGTGAAAC-CAATTC-3' 41 , Wolbachia surface protein (wsp) gene using primers wsp81F: 5'-TGGTCCAATAAGTGATGAAGAAAC-3' and wsp691R: 5'-AAAAATTAAACGCTACTCCA-3' 42 .A Bio-Rad T100 Thermal Cycler using standard cycling conditions was used to amplify 16S rRNA and wsp gene products which were then separated and visualised using an Invitrogen E-Gel iBase Real-Time Transilluminator with 2% SYBR safe E-Gel EX agarose gels.Real-time PCR reactions for the 16S rRNA gene were carried out with 5 µL of FastStart SYBR Green Master mix (Roche Diagnostics), primers at a final concentration of 1 µM, 1 µL of PCR grade water and 2 µL cDNA (10 µL final volume/reaction) as previously described using no template controls (NTC) and a limit of detection was previously established using a synthetic oligonucleotide standard through ten-fold serial dilutions 4 .A Roche LightCycler 96 System was used to amplify PCR products using the following cycling conditions: 15 min at 95°C, 40 cycles of 95°C for 15 seconds (sec) and 58°C for 30 sec.Asaia detection was also carried out using Real time PCR by amplifying the 16S rRNA gene 30 with the same mastermix, reagent concentrations and cycling conditions as for Wolbachia genes.PCR assays included a dissociation curve (95°C for 10 sec, 65°C for 60 sec and 97°C for 1 sec) to check that the correct amplicon was being amplified.Fluorescence was quantified using LightCycler 96 software (Roche Diagnostics).

Molecular mosquito species identification
For Aedes mosquitoes collected from Central Cameroon sites, a SYBR-green based assay that can distinguish Ae. aegypti from Ae. albopictus based on the internal transcribed spacer 1 (ITS1) 43 was used.PCR cycling conditions for the ITS1 assay were: 95°C for 15 min, 40 cycles of 95°C for 10 sec, 55°C for 30 sec, 72°C for 20 sec and a dissociation curve (see above).For Culex mosquitoes collected from Central Cameroon sites, a multiplex PCR assay targeting the ACE1 gene 44 that can distinguish Cx. pipiens pipiens from Cx. pipiens quinquefasciatus was also undertaken, given these sibling species are morphologically indistinguishable.PCR cycling conditions for the ACE1: 95°C for 10 min, 34 cycles of 95°C for 30 sec, 55°C for 30 sec, 72°C for 1 min and 2°C for 5 min.To determine the species for additional samples that were Wolbachia-positive, Sanger sequencing and phylogenetic analysis of the cytochrome c oxidase subunit 1 (COI) gene 45 was undertaken as this provided the most available sequences for comparison to ensure the optimal method for species identification.

Wolbachia MLST
Five conserved genes (gatB, coxA, hcpA, ftsZ and fbpA) were amplified to determine any newly discovered Wolbachia strains as previously described with the use of M13 adaptors or degenerate primers 46 .MLST PCRs consisted of 10 µL of Phire Hot Start II PCR Master Mix (Thermo Scientific), primers with a final concentration of 1 µM, 1 µL of PCR-grade water and 2 µL template cDNA (20 µL total).PCR cycling was carried out in a Bio-Rad T100 Thermal Cycler using cycling conditions that were optimised for different MLST genes tested with the Phire Hot Start II PCR Master Mix.Three genes (gatB, hcpA and fbpA genes) had the following cycling: 98°C for 30 sec, 34 cycles of 98°C for 5 sec, 65°C for 5 sec, 72°C for 10 sec, 72°C for 1 min.For two genes (coxA and ftsZ) cycling was: 98°C for 30 sec, 34 cycles of 98°C for 5 sec, 55°C for 5 sec and 72°C for 30 sec, 72°C for 1 min.

Sanger sequencing
PCR products were deemed worthy of sequencing when producing a strong, clear band at the correct product size when visualised using an Invitrogen E-Gel iBase Real-Time Transilluminator with 2% SYBR safe E-Gel EX agarose gels run for 10 mins.Products were sent to Source BioScience (Nottingham, UK) for cleanup prior to forward and reverse sanger sequencing.The MLST primers used were gene-specific and in the case of MLST genes just the M13 primers (M13_adaptor_F: 5'-TGTAAAACGACGGCCAGT-3' and M13_ adaptor_R: 5'-CAGGAAACAGCTATGACC-3') were used if these adaptors were included in the initial PCR to generate the product.MEGAX 47 was used for all analysis of sequences with manual checking of both forward and reverse chromatograms.Editing of sequences included trimming and then alignment to produce consensus sequences was undertaken using ClustalW.Nucleotide BLAST (NCBI) database queries and searches against the Wolbachia MLST database were combined to determine if new alleles and strain types were present in our collection.We also submitted our sequences to GenBank and obtained accession numbers.

Phylogenetic analysis
Alignments were constructed in MEGAX and ClustalW was also used to align our sequences alongside additional sequences obtained from NCBI BLAST and Wolbachia MLST database searches.Maximum Likelihood (ML) phylogenetic trees were generated after initially determining the optimal nucleotide substitution model using the "Find Best-Fit Substitution Model (ML)" option within MEGAX.The lowest Bayesian Information Criterion (BIC) score was one of the criteria used and this resulted in two models: the Jukes-Cantor model 48 and the General Time Reversible model 49 .For our phylogenetic analysis, we used the highest log likelihood and included next to the branches the percentage of trees in which the associated taxa clustered together.In all phylogenetic analyses we used a Bootstrap method with 1000 replications and Neighbor-Join and BioNJ algorithms using tMaximum Composite Likelihood (MCL).Our phylogenetic trees were then generated to scale, with branch lengths measured in the number of substitutions per site and we removed all gaps and missing data.

Statistical analysis
Fisher's exact post hoc tests in GraphPad prism version 9 (P<0.05significance threshold) were used to determine any association between prevalence rates of Wolbachia and Asaia for each mosquito genus from the different regions (West and Central).Samples were categorised as Wolbachia-infected, Asaia-infected, co-infected or uninfected.

Ethical approval
We previously obtained permission and ethical approval for mosquito sampling 4,40

Wolbachia and Asaia prevalence rates
We compared the prevalence rates of Wolbachia and Asaia rates using the 16S rRNA gene from the three major Culicine genera collected from both the West and Central Regions, with the caveat that the West Region samples were monospecific pools from the same species at the same location (Figure 2, Table 1).In the Central Region, 97.96% (n=115) of Aedes genera mosquitoes were infected only with Wolbachia, and Asaia was only detected in a single mosquito as a coinfection.In contrast, the majority of Culex genera mosquitoes collected from the Central Region were uninfected (85.83%, n=103), with Asaia detected in 13.45% (n=16) of individuals and only a single individual infected with Wolbachia.A similar infection prevalence was observed in Mansonia collected from the Molecular species identification of selected Wolbachiainfected mosquito samples After using the 16S rRNA and wsp genes to provide a preliminary indication of infection status, COI gene barcoding 45 was done to molecularly identify the species of a sub-sample of mosquitoes, given the difficulties associated with morphological identification of less well-studied species (Table 3-Table 4, Figure 3).Our results showed that within these 13 sub-selected Wolbachia-infected samples, there were seven Culex species, three Aedes species and one species each of five additional genera, confirmed to species level using Sanger sequencing of the COI gene: Cx. quinquefasciatus, Cx. watti, Cx. cinereus, Cx. nigripalpus, Cx. perexiguus, Cx. rima, Cx. cinctellus, Ae. africanus, Ae. denderensis, Ma. uniformis, Ca. argenteopunctata, Lu. tigripes, Er. chrysogaster and Ur.bilineata (Table 2).To differentiate between sibling species within the Cx.pipiens complex, we amplified the ACE1 gene and gel electrophoresis indicated a band size of 274 base pairs, which is diagnostic for Cx.quinquefasciatus.To avoid potentially mis-labelling species without sufficient sequence similarity, samples with species identity below 94% were designated 'cf' as this would be more indicative of a species that is closely related.However, the lack of sequences available for many of these species could result in genetic variation within the same species accounting for lower-than-expected sequence similarities.

Wolbachia genetic diversity and MLST gene allelic profiling
We used 16S rRNA phylogeny to put the strains detected in this study into context with existing strains (Figure 4).Our results showed eight strains are clustering closely together.In addition, there is sequence diversity among strains found infecting Ae. africanus (samples 8-10).An in-depth analysis was undertaken through MLST gene allelic profiling (Table 5) from The remaining samples had sequences of sufficient quality from 2-4 genes.For example, we were only able to obtain MLST gene sequences for two genes for Ma.uniformis (gatB and coxA).All sequences of sufficient quality were submitted to Genbank to obtain accession numbers Table 6.
As expected, the MLST allelic profile for Cx.quinquefasciatus mostly matched with strain type (ST) 9 for the wPip strain which infects Cx. pipiens and Cx.quinquefasciatus, although our sample had a match to ftsZ allele number 241, whereas existing profiles for ST 9 had an ftsZ allele number of 22.However, this represents only four nucleotide differences across the 435 base pairs for ftsZ alleles 241 and 22 resulting in 99.1% sequence identity.In Cx. watti, the Wolbachia strain allelic profile is most similar to Supergroup B Wolbachia strains found in the Coenonympha hero (the scarce heath  butterfly) -(ST 296), the moth Amblyptilia punctidactyla, Coenonympha pamphilus (the small heath butterfly), the Fischer's butterfly Tongeia fischeri (ST 300) and the thrip Pezothrips kellyanus (ST 430).However, only gatB (allele 9) and ftsZ (allele 7) show exact matches to these three strain types with hcpA and fbpA alleles being variable both for the novel strain in Cx. watti and STs 296, 300 and 430.The allelic profile of the Wolbachia strain detected in Cx. cinerus was most similar to 11 different strain types (3,108,151,213,366, 382,387,461,462,472,492) but appears a novel strain as none of these typed strains contain the combination of the four alleles identified in Cx. cinerus (we were unable to sequence the ftsZ gene).For example, ST 108 has the same gatB allele (71) but the remaining three genes for comparison (coxA, hcpA and fbpA) have variable allele numbers which are not the same as those sequenced from the Wolbachia strain in Cx. cinerus.
The Wolbachia strain detected in Cx. cf.nigripalpus produced an exact match to strain type 13 which is found in numerous insect species, including Drosophila recens and Leucophenga maculosa fruit flies and Rhagoletis cerasi (cherry fruit fly).
The widespread occurrence of this Wolbachia strain type across multiple insect genera requires further investigation given 19 isolates are present in the database.We were only able to sequence three genes for the Wolbachia strain detected in Cx. perexiguus (coxA, fbpA and hcpA) resulting in three nearest strain types (108, 187,467) of which none had allele number 11 for coxA.However, a complete MLST profile would help confirm what appears to be a novel strain in Cx. perexiguus given its unique combination of three sequenced MLST genes.Likewise, the Wolbachia strain detected in Cx. rima would appear novel given only two of four allelic loci (gatB and coxA) matched the closest ST (52) previously reported in Anoplolepis gracillipes (the yellow crazy ant).
Given that we collected large numbers of Ae. africanus from the West Region of Cameroon, we included three monospecific pools (A-C) for MLST allelic profiling (Table 5-Table 6).
Our results provide evidence for multiple Wolbachia strains through analysis indicating allelic matches for four of five MLST genes (and divergence seen in the 16S rRNA phylogeny between samples 8-10).Our Sanger sequencing indicated no evidence for the presence of multiple strains within the same pool, but further analysis is required to further determine if superinfections can be present within individual For Lutzia tigripes, we could only produce sequences for three MLST genes, but this strain also appears novel with only three of four loci matching ST 108 -a strain found in the butterfly Brangas felderi.Likewise, novel strains appear to be present in both Eretmapodites chrysogaster and Uranotaenia bilineata as their partial allelic profiles did not match any other strain types in the database.

Discussion
In Cameroon, we previously showed that the richness of mosquito species was dependent on both habitat type and seasonality 40 .Therefore, in this current study we analysed mosquitoes from diverse environmental settings to capture as much potential diversity in both mosquito species and corresponding resident Wolbachia strains.We identified what appears to be either novel strains or variants of existing characterised Wolbachia strains in 13 different mosquito species.A natural Wolbachia strain in Cx. quinquefasciatus mostly matching ST 9 is to be expected given the wPip strain is widespread in species of the Cx.pipiens complex 53-55 .Our allelic profiling indicated evidence of some genetic diversity in the ftsZ allele although this was only four of 435 nucleotides (99.1% sequence similarity).This also highlights the requirement of using MLST allelic profiling given the 16S analysis provided little sequence similarity to existing sequences from Wolbachia strains detected in the Cx.pipiens complex (Figure 4).Interestingly, there was no evidence of Wolbachia in Cx. pipiens collected from Madagascar 6 despite the prevailing assumption that the wPip strain widely infects both Cx.pipiens and Cx.quinquefasciatus populations.
Further studies across sub-Saharan Africa are needed to determine variability in both the prevalence rates and genetic strain diversity of the wPip strain in members of the Cx.pipiens complex, given their important role as vectors of multiple human pathogens such as West Nile virus and filarial nematodes such as Wuchereria bancrofti.
Another vector of human pathogens analysed in our study was Ae. africanus -a major vector of YFV.Although Ae. africanus is considered a sylvatic vector in rural areas, recent studies have suggested it has the capacity to colonise peri-domestic and domestic habitats 50 .Our MLST analysis suggests there are multiple Wolbachia strain variants present in Ae. africanus with variation in gene sequences in three of the five MLST genes (Table 5).It could also be possible that resident Wolbachia strain superinfections occur in Ae. africanus as have been seen in Ae. albopictus 56  Although we undertook molecular barcoding by sequencing the mosquito CO1 gene to try and provide as much confidence in species identification, caution must be taken with any results as this is dependent on the availability of sequences for comparison.For example, we identified Wolbachia strains in multiple individuals in which sequence identity was only 94%, suggesting these may be closely related species to the closest match sequence available on GenBank.Sample 7 was identified as Cx.cinctellus but with only 95% identity and 89% coverage, indicating this could also be another closely-related species.The inability to accurately identify mosquito species using molecular barcoding is common for species in which few sequences have previously been made available in databases such as GenBank.However, providing the CO1 sequences will inform future studies looking at closely related species.
Our results comparing Wolbachia and Asaia prevalence rates across major Culicine genera indicated a significant association only in Mansonia mosquitoes.These results for Mansonia are consistent with other mosquito species such as Ae.koreicus in which studies from field collected mosquitoes indicate a mutual exclusion between these two symbionts 58 .In contrast, high levels of co-infections (particularly within Ae. africanus, which dominated our collections from the West Region) add to growing evidence that Wolbachia and Asaia can co-exist in wild mosquitoes 3 despite studies clearly demonstrating an antagonistic association in lab colonies 27,28 .As Asaia can be acquired from the environment throughout the mosquito life cycle, the collection location becomes a significant factor that complicates this tripartite association and therefore our results are limited to both our collection locations and species collected.Another major limitation of our study is that we were unable to provide comparative data to the species level for the Central region due to the high levels of misidentification of Culicline species 59 and missing or damaged morphological features during mosquito collections.Furthermore, as tissue-specific detection was not feasible for the large number of diverse field-collected mosquitoes in our study, it could be possible that Wolbachia and Asaia co-exist within a given individual mosquito but are located in different tissues 60 .Likewise, the detection of both Wolbachia and Asaia in samples from the West region (monospecific pools) needs further investigation given the limitations of using pooled samples.The possibility of morphological misidentification resulting in the addition of an 'incorrect' species to the pool or results reflecting single infections (ie.one Wolbachia-infected individual, one Asaia-infected individual) cannot be ruled out.However, this would seem unlikely given Aedes monospecific pools mosquitoes were either co-infected (75%) or singly infected with Wolbachia (25%) and for Culex 98% were infected with Wolbachia only.Larger cohort collections of mosquitoes from diverse environmental settings will provide further insight into how these two widespread bacteria co-exist (or do not co-exist) in different mosquito species.
Wolbachia MLST gene allelic profiling was undertaken to provide more assurances on detection of genuine endosymbiotic strains found in wild mosquito populations.We defined a 'novel' strain based on MLST to contain either new MLST gene sequences not present in https://pubmlst.org/organisms/wolbachia-sppor a combination of MLST gene sequences that does match an existing strain in the database.Despite being widely used, MLST For Wolbachia strains has limitations and the five genes may not represent the optimal loci to capture strain variation 61 .Furthermore, defining whether a novel strain exists based only on PCR amplification of genes is problematic given the numerous examples of environmental contamination or host genome integration 51,52 .Caution must be taken when extrapolating PCR amplification to indicate the presence of a living endosymbiotic bacterium -particularly so when only a few gene targets such as 16S rRNA are amplified and sequenced.It has been shown that 16S rRNA prokaryotic DNA can be amplified from dead cells 51 and Wolbachia 16S rRNA can be detected just from water that previously contained mosquito larvae 52 .If possible, extraction of mosquito RNA (as carried out in this study) to confirm expression of Wolbachia genes provides further evidence 3 .Once novel strains are detected using MLST profiling, further studies are needed to confirm a genuine stable endosymbiotic association is present with the mosquito host species.This is important when low prevalence rates are detected given this may otherwise suggest that the Wolbachia strain is not inducing CI.Furthermore, there are several additional experiments that can be undertaken to provide further confirmation of resident Wolbachia strains in mosquitoes.Methods that can visualise Wolbachia bacteria in mosquito tissues using microscopy, such as fluorescent in situ hybridization, and Wolbachia genome sequencing, to compare genome depth and coverage of novel strains to those of other known infections, should be carried out to fully characterise novel Wolbachia strains.

Conclusions
Novel Wolbachia strains in Culicine mosquitoes collected from ecologically diverse settings in Cameroon add to the diversity of this highly prevalent endosymbiont in insect populations.Resident Wolbachia strains should be further characterised to determine the tissue tropism and density of newly discovered strains.Our study also suggests that co-infection with environmentally acquired Asaia bacteria is widespread in wild mosquito populations (except the Mansonia genera) and the antagonistic relationship observed in lab colonies may not be present in some wild Culicine populations.Novel Wolbachia strains could be considered as candidate strains for biocontrol strategies given their ability to reside naturally within existing mosquito populations and co-exist with environmentally acquired Asaia bacteria.

Guido Favia
School of Bioscience and Veterinary Medicine, University of Camerino, Camerino, Italy I have read the revised version of the manuscript and in my opinion, the authors improved the its quality well addressing the concerns raised at the first evaluation.Consequently the manuscript is worthy for indexing.
Wolbachia endosymbionts are prevalent in insect populations and can enact dramatic effects on their hosts.Understanding the prevalence and diversity of Wolbachia in natural mosquito populations has important implications for the control of mosquito-borne disease because Wolbachia infections can influence mosquito vectorial capacity and modulate the success of vector control programs.Other microbes such as Asaia can also reside in mosquito tissues where they can influence host phenotypes and interact with Wolbachia.Here, the authors performed field surveys of Culicine mosquitoes in two regions of Cameroon and screen either individuals or pools of mosquitoes for Wolbachia and Asaia using PCR detection and sequencing of 16S rRNA, wsp and MLST markers.They detect Wolbachia and Asaia at variable frequencies which on the mosquito genus and region.While these results provide important data for vector control programs and may lead to candidate strains for Wolbachia release programs, there are a few key limitations with their study design and their conclusions are overstated.
The authors claim to detect co-infections of Wolbachia and Asaia in field-collected mosquitoes, contradicting evidence from laboratory studies that the two microbes have an antagonistic relationship with each other.However, these two observations are not necessarily in conflict.While there may be mutual exclusion of Asaia and Wolbachia in the reproductive tissues of some mosquitoes, this does not preclude the two microbes from residing in the same mosquito but within different tissues.There are several cases where Wolbachia and Asaia have been detected in the same mosquito (e.g. ) but they may not be occupying the same tissues, especially given that Asaia can be environmentally acquired and often resides in the gut.This study also does not provide sufficient evidence for a Wolbachia-Asaia coinfection.Firstly, all samples where both Wolbachia and Asaia were detected are from pools of mosquitoes, so it is possible that these samples contain a mix of singly infected individuals.Secondly, molecular detection is not sufficient evidence of an infection as pointed out by previous studies (Chrostek and Gerth 2019 mBio 5 , Ross et al. 2020 Ecol Evol 6 ).This is acknowledged at the end of the discussion but the rest of the paper is written as if all these detections represent true infections.
There are also some limitations and apparent gaps/inconsistencies in their study design.Mosquitoes from the West region were pooled while mosquitoes from the Central region were tested individually, but mosquitoes were only identified to the species level in the West region.The use of pools is likely to overestimate the prevalence of Wolbachia and Asaia in a sample given that not all individuals need to test positive for the pool to be considered positive.There is also the risk that any errors with morphological species identification could lead to an incorrect assignment of Wolbachia or Asaia status of a pool if one species is positive and the other is negative.The authors also use a mix of 16S rRNA and wsp detection for Wolbachia, but the data for 16S are combined across species within a genus while the wsp data are only presented for individual species.Furthermore, the results for 16S include only Aedes, Mansonia and Culex while there are other genera included for the wsp marker.The justification for this is unclear, and it would be useful to know how concordant the two markers are with each other, especially when the authors note that the use of the 16S rRNA gene has limitations for Wolbachia detection.The authors also mention in the methods that they also used the FtsZ marker for initial Wolbachia detection but these results are not presented.
The authors then select a sample of mosquito species testing positive for Wolbachia for further analysis of the Wolbachia strains with 16S rRNA and MLST markers.The authors place an emphasis on the fact that the Wolbachia strains that they detect are novel and diverse but these terms are not well defined.Has Wolbachia been detected in any of these species besides Cx. quinquefasciatus before, and how different do Wolbachia sequences have to be before they are considered novel and diverse?The authors compared their sequences to the MLST database but this may not represent the full diversity of Wolbachia strains that have been identified given that Wolbachia sequences have been obtained through other approaches (e.g.Scholz et al.Nature Communications).There are also some issues with the use of the MLST system for strain typing and diversity as outlined by Bleidorn and Gerth 2018 (FEMS Microbiol Ecol) 7 and these should be acknowledged.

Specific comments
Title -The title is misleading -the authors have detected Wolbachia and Asaia sequences in a range of field-collected mosquito species, but there is insufficient evidence that these represent true infections.The terms "diverse" and "novel" are not well defined and it is unclear how many of these strains have been detected for the first time.
Introduction -the authors mention that identifying Wolbachia strains can provide important comparative data but they don't elaborate on how this information would be useful for biocontrol strategies.For instance, natural infections could interfere with releases of transinfections because they may change patterns of cytoplasmic incompatibility between the released and resident mosquitoes.The authors also note that novel strains could be considered candidates for biocontrol strategies, but what properties are you looking for in a strain for it to be considered a candidate, and what steps would need to be taken for the strain to be used (e.g.introgression or microinjection into a different mosquito line, virus blocking, cytoplasmic incompatibility and so on)?
Introduction paragraph 1 -There is an emphasis on wMel here specifically, but also consider including the wAlbB strain which was the first transinfected strain in Ae. aegypti (Xi et al. 2005, Science) and which has also been released in wild mosquito populations to inhibit virus transmission (Nazni et al. 2019 Current Biology).
Introduction paragraph 2 -"However, a more targeted approach amplifying Wolbachia-specific genes is required to confirm a resident strain is present" -This really requires additional evidence such as microscopy, removal of the infection, demonstration of maternal transmission and so on.

Introduction paragraph 2 -provide an example of a superinfection comprising strains from the same supergroup
Introduction paragraph 3 -"outbreaks of arboviral diseases including DENV, YFV, CHIKV and Rift Valley fever virus (RVFV) are a possibility." is a bit vague -do they occur and how frequently?
Methods -How many mosquitoes per pool?In the results you mention an average of 5 female mosquitoes, but how much did this vary?Methods -sanger sequencing -which PCR products were considered to be worthy of sequencing and why?
Results paragraph 1 -When presenting West region results, do these n values represent the number of pools?Or were these actually individuals?
Results paragraph 1 -What is the hypothesis being tested with the Fisher's exact tests?Please elaborate on what associations are being tested and include these statistical tests in the methods section

Results
-There appear to be some inconsistencies when presenting frequency data-sometimes samples from the West region are specifically referred to as pools while at other times there is no mention of them being individuals or pools (e.g. in the first paragraph of the results).

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility?Partly

Are the conclusions drawn adequately supported by the results? Partly
Competing Interests: I am currently collaborating on a project with a lab from the Liverpool School of Tropical Medicine of which a coauthor of this paper is currently a member.The project is unrelated to the current manuscript and I have not worked with any of the other authors, so I don't believe this will affect the objectivity of my review.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.
has important implications for the control of mosquito-borne disease because Wolbachia infections can influence mosquito vectorial capacity and modulate the success of vector control programs.Other microbes such as Asaia can also reside in mosquito tissues where they can influence host phenotypes and interact with Wolbachia.Here, the authors performed field surveys of Culicine mosquitoes in two regions of Cameroon and screen either individuals or pools of mosquitoes for Wolbachia and Asaia using PCR detection and sequencing of 16S rRNA, wsp and MLST markers.They detect Wolbachia and Asaia at variable frequencies which depend on the mosquito genus and region.While these results provide important data vector control programs and may lead to candidate strains for Wolbachia release programs, there are a few key limitations with their study design and their conclusions are overstated.
The authors claim to detect co-infections of Wolbachia and Asaia in field-collected mosquitoes, contradicting evidence from laboratory studies that the two microbes have an antagonistic relationship with each other.However, these two observations are not necessarily in conflict.While there may be mutual exclusion of Asaia and Wolbachia in the reproductive tissues of some mosquitoes, this does not preclude the two microbes from residing in the same mosquito but within different tissues.There are several cases where Wolbachia and Asaia have been detected in the same mosquito (e.g. ) but they may not be occupying the same tissues, especially given that Asaia can be environmentally acquired and often resides in the gut.Thank you for these points.We have expanded paragraph 4 of our discussion to provide a more balanced view on the relationship between Wolbachia and Asaia.This study also does not provide sufficient evidence for a Wolbachia-Asaia coinfection.Firstly, all samples where both Wolbachia and Asaia were detected are from pools of mosquitoes, so it is possible that these samples contain a mix of singly infected individuals.We have added a caveat to our discussion on 'pooling' mosquitoes (paragraph 4) but given that you have pointed out that 'there are several cases where Wolbachia and Asaia have been detected in the same mosquito' it would be contradictory to those studies to assume a coinfection is not possible.Unfortunately these collections were undertaken during the covid pandemic so a pooling approach was taken to minimise time in labs under severe travel restrictions.Our data from the West region (Aedes monospecific pools mosquitoes were either co-infected (75%) or singly infected with Wolbachia (25%) and for Culex 98% were infected with Wolbachia only) would also indicate a mix of singly infected individuals is not likely to represent most of our data.Secondly, molecular detection is not sufficient evidence of an infection as pointed out by previous studies (Chrostek and Gerth 2019 mBio 5 , Ross et al. 2020 Ecol Evol 6 ).This is acknowledged at the end of the discussion but the rest of the paper is written as if all these detections represent true infections.We agree there have been many recent papers that have just presented molecular data without providing further evidence but we feel our manuscript provides sufficient caveats to justify our results as more than likely coming from genuine infections (paragraphs 4 and 5 of the discussion).We acknowledge that reliance on just one Wolbachia gene (eg.16S rRNA or wsp) would not be sufficient as has been the case for numerous recent publications.However, having multiple MLST gene sequences from Wolbachia housekeeping genes that match (or are very close in sequence) to existing sequencing in https://pubmlst.org/organisms/wolbachia-sppfrom RNA extracts (so active gene expression) would indicate it's much more likely to represent genuine strains than the alternatives such as environmental contamination.We support the general opinion that further experimental work is needed but methods such as whole genome sequencing and FISH are not feasible on large numbers of field collected mosquitoes (from different species) without prior knowledge of molecular evidence that strain exist (or are likely to exist based on high prevalence rates and consistent strain typing).There are also some limitations and apparent gaps/inconsistencies in their study design.Mosquitoes from the West region were pooled while mosquitoes from the Central region were tested individually, but mosquitoes were only identified to the species level in the West region.Unfortunately we were unable to morphologically identify species in the Central region due to 1) a lack of expertise and 2) missing or damaged morphological features during mosquito collections.Morphological identification of less-well known Culicine species is particularly challenging and often results in misidentification 1 .Therefore, we firstly screened individuals for symbionts using qPCR targeting both Wolbachia and Asaia 16S rRNA genes to identify samples of interest (and limited our analysis to the genera level).MLST sequencing was then carried out on selected individuals for Wolbachia strain typing (our main aim of this work) followed by molecular confirmation of mosquito species using CO1 barcoding.The use of pools is likely to overestimate the prevalence of Wolbachia and Asaia in a sample given that not all individuals need to test positive for the pool to be considered positive.There is also the risk that any errors with morphological species identification could lead to an incorrect assignment of Wolbachia or Asaia status of a pool if one species is positive and the other is negative.We again agree that pooling is not the preferred method for detection but covid restrictions (both in the field for travel and in the lab for analysis) resulted in a pooling strategy to be able to analyse samples.We have added a caveat to our discussion on 'pooling' mosquitoes (paragraph 4) to address this point.The authors also use a mix of 16S rRNA and wsp detection for Wolbachia, but the data for 16S are combined across species within a genus while the wsp data are only presented for individual species.Furthermore, the results for 16S include only Aedes, Mansonia and Culex while there are other genera included for the wsp marker.The justification for this is unclear, and it would be useful to know how concordant the two markers are with each other, especially when the authors note that the use of the 16S rRNA gene has limitations for Wolbachia detection.The authors also mention in the methods that they also used the FtsZ marker for initial Wolbachia detection but these results are not presented.Please see our previous response in terms of the difficulties of morphological ID from the Central region (lack of expertise and damage from collections).We used the wsp gene on our pooled samples from the West to re-enforce what appeared to be a high prevalence rate from the monospecific Aedes pools.In contrast to 16S rRNA, wsp is often not amplified from low density strains (or strains that may not be in genuine endosymbiosis).For example, there are numerous studies that report the presence of Wolbachia in Anopheles species in which wsp failed to amplify (PMID: 32787974).Genbank lists 835 sequences for wsp (search wsp AND "Wolbachia pipientis"[porgn:__txid955]) compared to 2643 for 16S rRNA (search 16S rRNA AND "Wolbachia pipientis"[porgn:__txid955]) suggesting wsp amplification is more indicative of a genuine symbiotic association (and hypervariable regions can be used for strain typing in addition to MLST).The authors then select a sample of mosquito species testing positive for Wolbachia for further analysis of the Wolbachia strains with 16S rRNA and MLST markers.The authors place an emphasis on the fact that the Wolbachia strains that they detect are novel and diverse but these terms are not well defined.We have added a sentence to our discussion (paragraph 5) on what criteria we used based on MLST.Has Wolbachia been detected in any of these species besides Cx. quinquefasciatus before, and how different do Wolbachia sequences have to be before they are considered novel and diverse?This was partly addressed in our discussion for Mansonia but we have now checked the most reliable source for strain typing (https://pubmlst.org/organisms/wolbachia-spp)and can confirm (based on this) that no other Culex species in our study has MLST or wsp sequences.There are also no strain profiles present for Aedes africanus or Aedes denderensis or the remaining samples (Catageiomyia argenteopunctata, Lutzia tigripes, Eretmapodites chrysogaster and Uranotaenia bilineata).We have added a sentence to the end of paragraph 2 of the discussion to address this.The authors compared their sequences to the MLST database but this may not represent the full diversity of Wolbachia strains that have been identified given that Wolbachia sequences have been obtained through other approaches (e.g.Scholz et al.Nature Communications).There are also some issues with the use of the MLST system for strain typing and diversity as outlined by Bleidorn and Gerth 2018 (FEMS Microbiol Ecol) 7 and these should be acknowledged.We have included a sentence in the discussion (paragraph 5) on the limitations of MLST despite this being widely used in the Wolbachia community for strain diversity.

Specific comments
Title -The title is misleading -the authors have detected Wolbachia and Asaia sequences in a range of field-collected mosquito species, but there is insufficient evidence that these represent true infections.The terms "diverse" and "novel" are not well defined and it is unclear how many of these strains have been detected for the first time.We respectively disagree that the title should be changed to 'sequences' given we are presenting MLST data.There are many published papers reporting Wolbachia strain detection with much less genetic data (sometimes based on sequences obtained from nested 16S rRNA) so this change seems inappropriate.We have changed our title to 'Diverse novel Wolbachia bacteria strains and genera-specific co-infections with Asaia bacteria in Culicine mosquitoes from ecologically diverse regions of Cameroon'' to reflect genera differences for Asaia co-infections.
Introduction -the authors mention that identifying Wolbachia strains can provide important comparative data but they don't elaborate on how this information would be useful for biocontrol strategies.For instance, natural infections could interfere with releases of transinfections because they may change patterns of cytoplasmic incompatibility between the released and resident mosquitoes.The authors also note that novel strains could be considered candidates for biocontrol strategies, but what properties are you looking for in a strain for it to be considered a candidate, and what steps would need to be taken for the strain to be used (e.g.introgression or microinjection into a different mosquito line, virus blocking, cytoplasmic incompatibility and so on)?We have added some sentences in paragraph 2 of the introduction to address this important point.Introduction paragraph 1 -References 11 and 12 refer to Ae. polynesiensis, not Ae.

albopictus. Thank you for pointing this error out -we have changed this
Introduction paragraph 1 -There is an emphasis on wMel here specifically, but also consider including the wAlbB strain which was the first transinfected strain in Ae. aegypti (Xi  et al. 2005, Science) and which has also been released in wild mosquito populations to inhibit virus transmission (Nazni et al. 2019 Current Biology).Thank you for this suggestion and we have added some information at the end of paragraph 1 Introduction paragraph 2 -"However, a more targeted approach amplifying Wolbachiaspecific genes is required to confirm a resident strain is present" -This really requires additional evidence such as microscopy, removal of the infection, demonstration of maternal transmission and so on.We agree and have added a sentence to reflect this Introduction paragraph 2 -provide an example of a superinfection comprising strains from the same supergroup We have not been able to find a clear example (at least within mosquitoes) so have modified this sentence to reflect this.

Introduction paragraph 3 -"outbreaks of arboviral diseases including DENV, YFV, CHIKV and Rift
Valley fever virus (RVFV) are a possibility." is a bit vague -do they occur and how frequently?
We have modified this and added a sentence with an example for DENV in Cameroon.

now been added
Results -There appear to be some inconsistencies when presenting frequency datasometimes samples from the West region are specifically referred to as pools while at other times there is no mention of them being individuals or pools (e.g. in the first paragraph of the results).We have added 'pools' were appropriate to reflect the West region data Discussion -What makes a strain a candidate?Is any newly detected strain a candidate strain, or do they need to display a certain phenotype such as cytoplasmic incompatibility or virus blocking?In my opinion a candidate strain must possess some desirable traits and should be able to be maintained in the laboratory.Thank you for this valid point and opinion.Given that mosquito transinfection has been possible from closely related mosquito species (eg.wAlbB into Ae.aegypti and An.stephensi), strains that natural reside within mosquito species may provide an increased chance of successful transinfection (difficult to test given the technical difficulty of embryo microinjection).Cell line adaptation was seen as key reasons for the success of Drosophila Wolbachia strains wMel and wMelPop into Ae.aegypti so strains that can be firstly identified to be at high prevalence in mosquito field populations will allow further examination in terms of phenotypic effects through eg.cif gene analysis to indicate CI induction.We agree that there are numerous candidate strains in lab colonies but temperature has recently been shown to decrease both density and virus inhibition for some Drosophila strains (eg.wMel).Potential candidate strains that naturally exist in mosquito field populations in countries such as Cameroon with high average temperature may be less likely to see this effect.References 1. Jourdain, F., Picard, M., Sulesco, T., Haddad, N., Harrat, Z., Sawalha, S.S., Gunay, F., Kanani, K., Shaibi, T., Akhramenko, D.

Guido Favia
School of Bioscience and Veterinary Medicine, University of Camerino, Camerino, Italy The work is part of a series of works, now quite numerous, which aim to identify species of insect vectors that are infected by Wolbachia.In this sense, the research is not characterized by particular originality; nevertheless the research has some appreciable aspects as it aims to analyze several species of mosquitoes (some of which are little studied) in some areas of Cameroon, in different eco-ethological contexts.Furthermore, I appreciated that the analysis of Wolbachia distribution was carried out not exclusively through 16S amplification but also through an MLST approach resulting in phylogenetic analysis.
That said, the manuscript has some flaws which, in my opinion, should be corrected before publication.
For experts in the field the decision to verify the co-presence/co-absence of Wolbachia and Asaia is quite clear but for many readers it would be useful to explain why the monitoring concerned Asaia and not other symbionts (it might help in the introduction to point out that these are two of the few symbionts for which insect and vector control approaches have already been proposed and in some cases validated in semifield and field experiments).

1.
Figure 2 compares the positivity between West and East mosquitoes at the genus level.I don't understand why, given that they then move on to the analysis and definition of the species, this datum is not expressed by species; it would definitely be more informative.

2.
Even the logical thread expressed in the materials and methods does not seem acceptable to me.They speak first of the monitoring of Asaia and Wolbachia and then of the definition of the host species.A stringent logic that aims to monitor the distribution of two symbionts in different mosquito species, I think, would mean proceeding first with the identification of the host-species and then with monitoring the distributions of the symbionts.

3.
Another aspect that should be slightly modified concerns the conclusions: the authors state that "Our study also suggests that co-infection with environmentally acquired Asaia bacteria is widespread in wild mosquito populations and the antagonistic relationship observed in lab colonies may not be present in wild Culicine populations."This is not entirely true since there 4.
are studies on some wild-species of culicine that demonstrate an almost total absence of coexistence of the two symbionts, as in the case of some invasive species of aedes.The sentence therefore should be rephrased and referred to and limited to the geographical context analyzed by the authors.publication.

Is
For experts in the field the decision to verify the co-presence/co-absence of Wolbachia and Asaia is quite clear but for many readers it would be useful to explain why the monitoring concerned Asaia and not other symbionts (it might help in the introduction to point out that these are two of the few symbionts for which insect and vector control approaches have already been proposed and in some cases validated in semifield and field experiments).

1.
Thank you for this suggestion.We agree and have extended our introduction (paragraph 2) to incorporate a more expansive introduction to Asaia and its potential role for mosquito control.
Figure 2 compares the positivity between West and East mosquitoes at the genus level.I don't understand why, given that they then move on to the analysis and definition of the species, this datum is not expressed by species; it would definitely be more informative. 1.
Unfortunately we were unable to morphologically identify species in the Central region due to 1) a lack of expertise and 2) missing or damaged morphological features during mosquito collections.Morphological identification of less-well known Culicine species is particularly challenging and often results in misidentification 1 .Therefore, we firstly screened individuals for symbionts using qPCR targeting both Wolbachia and Asaia 16S rRNA genes to identify samples of interest (and limited our analysis to the genera level).MLST sequencing was then carried out on selected individuals for Wolbachia strain typing (our main aim of this work) followed by molecular confirmation of mosquito species using CO1 barcoding.
Even the logical thread expressed in the materials and methods does not seem acceptable to me.They speak first of the monitoring of Asaia and Wolbachia and then of the definition of the host species.A stringent logic that aims to monitor the distribution of two symbionts in different mosquito species, I think, would mean proceeding first with the identification of the host-species and then with monitoring the distributions of the symbionts.

1.
Please see our response to query 2. We firstly identified samples of interest from the Wolbachia/Asaia qPCR results and then proceeded to undertake Wolbachia MLST strain typing and confirm species using CO1 barcoding given the difficulties of Culicine morphological identification which in a study that used an external quality assessment resulted in only 64% accuracy to the species level across 19 participant laboratories 1 .
Another aspect that should be slightly modified concerns the conclusions: the authors state that "Our study also suggests that co-infection with environmentally acquired Asaia bacteria is widespread in wild mosquito populations and the antagonistic relationship observed in lab colonies may not be present in wild Culicine populations."This is not entirely true since there are studies on some wild-species of culicine that demonstrate an almost total absence of coexistence of the two symbionts, as in the case of some invasive species of aedes.The sentence therefore should be rephrased and referred to and limited to the geographical context analyzed by the authors.

1.
Thank you for this suggestion and we agree.We have modified our discussion in paragraph 4 to address these very valuable points.References 1. Jourdain, F., Picard, M., Sulesco, T., Haddad, N., Harrat, Z., Sawalha, S.S., Gunay, F., Kanani, K., Shaibi, T., Akhramenko, D. In their manuscript 'Diverse novel Wolbachia bacteria strains and widespread co-infections with Asaia bacteria in Culicine mosquitoes from ecologically diverse regions of Cameroon', the authors use RT-qPCR screening of mosquito specimens collected from two areas in Cameroon to look for evidence of novel Wolbachia infections, as well as examining patterns of Wolbachia-Asaia cooccurrence/co-exclusion. They observe geographic and mosquito clade-specific patterns of Wolbachia and Asaia occurrence.They then use MLST sequencing data and phylogenetics to examine the similarity of their putative, new Wolbachia strains to each other and those previously characterized.Such strains, if proven to be viable and heritable, may be useful as targets to generate novel Wolbachia transinfections, but may also provide important information on the ecology, physiology, and vector competence of several mosquito species in the region.
The manuscript is well written and does a great job of mentioning caveats.The authors have been particularly conscious of the issue of false positive detection of Wolbachia in their choice of methodology, and in their acknowledgements that all of these putative strains must be further validated, for instance, by using imaging techniques.I have a few suggestions and clarifications for the revised version of the text: The issue of false positive detection of Wolbachia in mosquitoes is intrinsic to this work and field of Wolbachia research and it could be briefly mentioned in the introduction.This will give your readers improved context for your choice of RNA-based quantification of Wolbachia. 1.
Details on Wolbachia detection criteria (calling positives vs negatives) as well as the positive and negative controls used in the RT-qPCR assay have not been provided in the methods section.These details are vital to demonstrate efforts have been made to reduce false positive detection.

2.
Please clarify how Fisher's tests were used to analyze your data.These tests were not mentioned in the methods section and are applied in pairwise comparisons, which might 3.
explain why some tests in Table 1 have such high P values.
At the beginning of the results section it would be useful to include a brief summary of which mosquito species were collected, and where and when they were collected.The when is important given samples were collected across many years.

4.
I do not see an equivalent of table 2 for the central region data.It looks as though Table 2 just deals with the western region.

5.
Queries about phylogenetic trees: -In Fig. 3, is there an explanation for why Cx pipiens and Cx quinquefasciatus don't cluster together.
-None of the phylogenetic appear to have outgroups -Fig.4 -while it is useful to have a tree based on the Wolbachia 16s gene.It could be valuable to include an unrooted consensus tree based on the MLST data.

6.
Is there a reason why Figure 2 and Table 2 collapse the prevalence of Wolbachia and Asaia to the genus level?This reduces the biological relevance of your findings as mosquitoes within a genus can fill distinct biological niches.
The discussion describes the collection areas in Cameroon as containing a high degree of environment-driven variation.Reflecting on that, the decision to bin samples into two homogeneous regions (West vs Central) does not make sense.Are there any site-specific patterns that could be mentioned that better reflect site-to-site diversity in mosquito species/Wolbachia prevalence? 9.
The discussion described Wolbachia/Asaia co-infection as being "widespread" but this does not appear to reflect the data given that Asaia was only highly prevalent in one mosquito genus.

10.
If the introduction were to briefly discuss the impact of Asaia on mosquito vector competence and immunity it could add important context outlining the rationale for studying Asaia infections in mosquitoes.

If applicable, is the statistical analysis and its interpretation appropriate? Partly
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.
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, however I have significant reservations, as outlined above.
Details on Wolbachia detection criteria (calling positives vs negatives) as well as the positive and negative controls used in the RT-qPCR assay have not been provided in the methods section.These details are vital to demonstrate efforts have been made to reduce false positive detection.

1.
Apologies we have corrected this error and included details on the inclusion of NTCs and referenced the previously generated 16S rRNA standard curve providing a limit of detection 1 .The raw qPCR data can also be found at https://doi.org/10.17605/OSF.IO/V75DU.
Please clarify how Fisher's tests were used to analyze your data.These tests were not mentioned in the methods section and are applied in pairwise comparisons, which might explain why some tests in Table 1 have such high P values. 1.

Apologies we erroneously missed a 'statistics' section in the methods and have now addressed this
At the beginning of the results section it would be useful to include a brief summary of which mosquito species were collected, and where and when they were collected.The when is important given samples were collected across many years.

1.
Unfortunately there was a type for the collection 'years' for the Central region (this was only in 2019) and we have modified this in the methods.This reduces any confusing in terms of 'when' they were collected and all location data is present in the methods so we don't feel a summary is now warranted.I do not see an equivalent of table 2 for the central region data.It looks as though Table 2 just deals with the western region.

1.
That is correct.Unfortunately we were unable to morphologically identify species in the Central region due to 1) a lack of expertise and 2) missing or damaged morphological features during mosquito collections.Morphological identification of less-well known Culicine species is particularly challenging and often results in misidentification 2 .Therefore, we firstly screened individuals for symbionts using qPCR targeting both Wolbachia and Asaia 16S rRNA genes to identify samples of interest (and limited our analysis to the genera level).MLST sequencing was then carried out on selected individuals for Wolbachia strain typing (our main aim of this work) followed by molecular confirmation of mosquito species using CO1 barcoding.Queries about phylogenetic trees: -In Fig. 3, is there an explanation for why Cx pipiens and Cx quinquefasciatus don't cluster together.There are multiple 16S sequences available on GenBank for Wolbachia strains detected in species within the Culex pipienscomplex indicating genetic diversity in the strain group loosely labelled as 'wPip'.This would be reflective of numerous studies showing that there are multiple Wolbachia strains that co-existing within Culex pipiens populations 3-5 .We have added a sentence to our discussion to highlight this point and the value of undertaking MLST to provide greater depth of sequencing data.-None of the phylogenetic appear to have outgroups Out grouping for both mosquito barcoding and Wolbachia phylogenetic trees, to our knowledge, is not commonly undertaken (at least based on the vast majority of similar publications).Instead, we (and previously published studies) include mosquito barcoding gene sequences present on GenBank that closely match sequences generated in this study.For Wolbachia 16S, we (and previously published studies) include 16S sequences from well characterised Wolbachia strains such as wMel and wRi for context.-Fig.4 -while it is useful to have a tree based on the Wolbachia 16s gene.It could be valuable to include an unrooted consensus tree based on the MLST data.Although we agree we were not able to generate complete MLST profiles for all mosquito species (please see table 5) which is consistent with numerous other published studies.Therefore, we feel a phylogenetic tree based on MLST data would be incomplete and provide little beneficial data not present in table 5.
Is there a reason why Figure 2 and Table 2 collapse the prevalence of Wolbachia and Asaia to the genus level?This reduces the biological relevance of your findings as mosquitoes within a genus can fill distinct biological niches. 1.
I think this is referring to Figure 2 and yes we agree that it does limit our conclusions.Please see our response to point 5 but we have also included a few sentences in the discussion (paragraph 4) to highlight this. Figure 2 is not visually accessible (red/green colors).1.
Thank you for pointing this out -we have changed the colour scheme.
The discussion describes the collection areas in Cameroon as containing a high degree of environment-driven variation.Reflecting on that, the decision to bin samples into two homogeneous regions (West vs Central) does not make sense.Are there any site-specific patterns that could be mentioned that better reflect site-to-site diversity in mosquito species/Wolbachia prevalence? 1.
We agree but unfortunately the collections were affected by the covid pandemic and we had to treat samples differently (individuals vs pools, morphological identification) between West and Central collections due to restricted travel and time working in labs.In our methods we do include details on collection locations but do not feel we have sufficient 'site-specific environmental data' to justify further sub-divisions within the two Regions.
The discussion described Wolbachia/Asaia co-infection as being "widespread" but this does not appear to reflect the data given that Asaia was only highly prevalent in one mosquito genus.

1.
We agree and have addressed this in our discussion (paragraph 4) and have changed the title to reflect this with a more balanced conclusion.
If the introduction were to briefly discuss the impact of Asaia on mosquito vector competence and immunity it could add important context outlining the rationale for studying Asaia infections in mosquitoes.

Figure 2 .
Figure 2. Wolbachia, Asaia and co-infection prevalence rates from major Culicine mosquito genera collected from the West and Central regions of Cameroon.Mosquitoes from the Central region were individually extracted and analysed.Mosquitoes from the West region were extracted from monospecific pools (same species from same collection location) and prevalence analysis represents pooled samples.No Mansonia mosquitoes were collected from the West region.

Figure 3 .
Figure 3. CO1 gene phylogenetic analysis of mosquito species collected from Cameroon.Maximum Likelihood phylogenetic analysis using the General Time Reversible model.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.6443)).The tree comprises 30 nucleotide sequences with 725 positions in the dataset.Drawn to a 0.05 scale.

Figure 4 .
Figure 4. 16S rRNA gene phylogenetic analysis of Wolbachia strains.The tree was produced using the Maximum Likelihood method and Jukes-Cantor model.The tree contains 31 nucleotide sequences and 1610 positions in the dataset.Wolbachia strains detected in this study are highlighted in red, sequences from additional strains obtained from Genbank with accession numbers are shown in black.Drawn to a 0.05 scale.

-
How many mosquitoes per pool?In the results you mention an average of 5 female mosquitoes, but how much did this vary?We have added this (3-5 mosquitoes/pool) in the methods Methods -sanger sequencing -which PCR products were considered to be worthy of sequencing and why?We have modified the beginning of the this section to provide further details Results paragraph 1 -When presenting West region results, do these n values represent the number of pools?Or were these actually individuals?We have edited this paragraph for clarity Results paragraph 1 -What is the hypothesis being tested with the Fisher's exact tests?Please elaborate on what associations are being tested and include these statistical tests in the methods section We erroneously left out a statistics section in the methods which has , et al. (2018).Identification of mosquitoes (Diptera: Culicidae): an external quality assessment of medical entomology laboratories in the MediLabSecure Network.Parasit Vectors 11, 553.10.1186/s13071-018-3127-7. Competing Interests: No competing interests were disclosed.Reviewer Report 13 July 2023 https://doi.org/10.21956/wellcomeopenres.20603.r61511© 2023 Favia G.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.
, et al. (2018).Identification of mosquitoes (Diptera: Culicidae): an external quality assessment of medical entomology laboratories in the MediLabSecure Network.Parasit Vectors 11, 553.10.1186/s13071-018-3127-7. Competing Interests: No competing interests were disclosed.Reviewer Report 13 July 2023 https://doi.org/10.21956/wellcomeopenres.20603.r60867© 2023 Caragata E. 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.Eric P. Caragata Florida Medical Entomology Laboratory, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida, USA

Table 1 . Prevalence rates of Wolbachia (16S rRNA gene), Asaia and coinfection from mosquito genera collected from the West and Central regions of Cameroon. Mosquitoes from
the Central region were individually extracted and analysed.Mosquitoes from the West region were extracted from monospecific pools (same species from same collection location) and prevalence analysis represents pooled samples.

Table 2 . Wolbachia infection prevalence using the wsp gene. Based on
morphological identification to genera/species and wsp gene amplificaiton in mosquitoes collected from the West Region of Cameroon.
representatives of each mosquito species from wsp-positive individuals (Central Region) or monospecific pools (West Region) after species identification was confirmed.Complete MLST sequences are present for Cx.quinquefasciatus, Cx. watti, Ae. africanus (Sample 8) and Ca.argenteopunctata.

Table 3 . CO1 gene sanger sequencing for molecular confirmation of mosquito species.
The NCBI BLAST percentage (%) identity and coverage are shown alongside the closest NBCI accession number (no.) and associated species.For identity 94% and under 'cf' has been added given the uncertainty of species identification.

Table 4 . CO1 and Wolbachia 16S rRNA GenBank accession numbers
. Location, species and sample code are shown alongside Genbank accession numbers.Sample sequences without accession numbers were of insufficient quality to obtain GenBank accession numbers.

Table 6 . GenBank accession numbers for MLST gene sequences
Ae. africanus.A novel strain was detected in Ae. denderensis as only three loci (gatB, hcpA and fbpA) matched ST 467 of a Wolbachia strain found in Cabera pusaria (Common white wave moth).A complete MLST profile was generated for the Wolbachia strain in Ca. argenteopunctata which appears novel given only two loci in combination match existing strains.
. Sample sequences without accession numbers were of insufficient quality to obtain GenBank accession numbers.
Kenya 8but not fbpA, suggesting the possibility of a different strain variant present in populations from Cameroon.The remaining novel Wolbachia strains that we have identified were in mosquito species that are considered either minor vectors of human disease or implicated in transmission of WNV, such as Cx.perexiguus, Cx. watti, and Cx.rima.With the exception of Cx. quinquefasciatus and Ma.uniformis, no Wolbachia MLST sequences are available (pubmlst.org/organisms/wolbachia-spp)for the remaining mosquito species.
. Mansonia uniformis has been shown to transmit numerous arboviruses, such as Murray Valley encephalitis and Ross River virus, and has been shown to be a vector of Bancroftian lymphatic filariasis in Ghana 57 .Interestingly, although Wolbachia has been previously identified in this species, no complete allelic profile is present.Our results match gatB and coxA allelic loci from Ma. uniformis collected in Schrieke et al. 2022 Computational and Structural Biotechnology 1 , Hegde et al. 2018 Frontiers in Microbiology 2 , da Silva et al. 2022 genes 3 , Chen et al. 2020 Frontiers in Microbiology 4

Table 1
-results for Wolbachia 16S are reported, but what about other genes?Were they concordant with the 16S results?

Table 1 -
What is the Fisher test comparing?
Although mentioned as a caveat in the text, it is misleading to put "co-infection" for the West samples when these consist of pools of mosquitoes -it is plausible that pools could consist of singly infected Wolbachia and Asaia individuals Table 2 -Only the last three rows indicate that pools were tested, but aren't all samples in this table pools?Figure 1 -Please provide the 16S data by species to allow for a comparison with the wsp dataResults-"Wolbachia 16S rRNA sequences has previously been detected in water containers that contained larvae of mosquitoes" -Is this not also possible for the other genes?Discussion -What makes a strain a candidate?Is any newly detected strain a candidate strain, or do they need to display a certain phenotype such as cytoplasmic incompatibility or virus blocking?In my opinion a candidate strain must possess some desirable traits and should be able to be maintained in the laboratory.

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? Partly
Schrieke et al. 2022 Computational and Structural Biotechnology 1 , Hegde et al. 2018 Frontiers in Microbiology 2 , da Silva et al. 2022 genes 3 , Chen et al. 2020 Frontiers in Microbiology 4

Table 1
-results for Wolbachia 16S are reported, but what about other genes?Were they concordant with the 16S results?

We used qPCR targeting the 16S rRNA gene for both Wolbachia and Asaia to generate CTs to provide a consistent comparison and did not compare to endpoint PCR results. All raw data is available at
https://doi.org/10.17605/OSF.IO/V75DU.

Table 1 -
What is the Fisher test comparing?

Please see our new 'statistical analysis' section in the methods. "Fisher's exact post hoc tests in GraphPad prism version 9 (P<0.05 significance threshold) were used to determine any association between prevalence rates of Wolbachia and Asaia for each mosquito genus from the different regions (West and Central). Samples were categorised as Wolbachia-infected, Asaia- infected, co-infected or uninfected.''Table 1 -
Although mentioned as a caveat in the text, it is misleading to put "co-infection" for the West samples when these consist of pools of mosquitoes -it is plausible that pools could consist of singly infected Wolbachia and Asaia individuals We

agree that we need to include more caveats in our discussion (see new paragraph 4) but disrespectfully disagree for table 1 given we have the following in the legend 'Mosquitoes from the West region were extracted from monospecific pools (same species from same collection location) and prevalence analysis represents pooled samples' to clearly show these are pooled samples.Table 2 -
Only the last three rows indicate that pools were tested, but aren't all samples in this table pools?

Thank you for spotting this error -we have amended this table Figure1
-Please provide the 16S data by species to allow for a comparison with the wsp data This

is a map showing the co-ordinates of the collections and has no species data Results
-"Wolbachia 16S rRNA sequences has previously been detected in water containers that contained larvae of mosquitoes" -Is this not also possible for the other genes?

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? Partly 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? Partly
Competing Interests: I am working in the same field.Nevertheless this hasn't in anyway affected my review Reviewer Expertise: Molecular Entomology, Molecular Parasitology, Insect symbiosis I confirm that I