Genetic structure and vector competence for DENV-2 of Aedes albopictus (Diptera: Culicidae) populations in China CURRENT

Background: As a vector of over 20 arboviruses, Aedes albopictus has spread throughout the world mostly since the second half of the twentieth century, and it is now found on every continent except Antarctica. Approximately 50-100 million people are infected with dengue virus (DENV) transmitted by Aedes mosquitoes each year, leading to a heavy economic burden on both governments and individuals. Understanding the population genetics and vector competence of vector species is critical to effectively preventing and controlling vector-borne diseases. The aim of this study was to examine the genetic structure and vector competence for dengue virus type 2 (DENV-2) of Ae. albopictus populations across China and their relationship. Methods: From July to September in 2019, Ae. albopictus eggs were collected by using ovitraps in 8 localities across China and reared to adults in laboratory. The mitochondrial gene cytochrome c oxidase subunit 1 ( cox 1) were used to examine the patterns of genetic diversity and population structure among native Ae. albopictus populations. The vector competence was detected by quantitative analysis of DENV-2 virus titer in mosquito tissues (midguts, heads and salivary glands) through qRT-PCR. The correlation between population genetic indices and DENV-2 loads in the mosquitoes’ tissues was also examined. Results: A total of 20 haplotypes of mtDNA cytochrome c oxidase subunit 1 ( cox 1) were identified in the 120 specimens from the eight tested populations. The dominate haplotype H01 was detected in seven geographic populations of mainland China. Genetic parameters such as haplotype diversity ( H d ), nucleotide diversity (π), and fixation index ( F ST ) revealed the population diversity decreasing from south to north, and low population genetic differentiation. STRUCTURE analysis indicated that Ae. albopictus populations in southern China were clustered. In addition, The Mantel test indicated a positive correlation between genetic distance and geographical distance ( R 2 = 0.364, P = 0.003). We observed no correlation between population genetic indices of cox 1 in Ae. albopictus populations and DENV-2 virus loads. However, the southern populations had the low DENV-2 virus loads generally. Conclusion: Conventional genetic markers such as cox 1 may not This study reported not only the genetic structure and from eight locations across China but also the correlation between them. Our results showed the population differentiation among Ae. albopictus was low and positively related with geographic distance, which was similar to our previous study. There was no significant correlation between vector competence and the genetic indices of each population from cox1 analysis. However, the southern populations had the low DENV-2 loads generally, which may be associated with innate immunity and worth deep research. The genetic structure and vector competence of Ae. albopictus populations may have implications for understanding the epidemiology, prevention and control of vector-borne diseases. Further researches are needed to focus on mosquitoes’ innate immune system or immune-related genetic loci to reduce vector competence for prevention and control of vector-borne diseases.

populations may be associated with the stable immunity system established in mosquitoes due to the long-term prevalence of dengue disease in these areas. The genetic structure and vector competence of Ae. albopictus populations in this study may have implications for understanding the epidemiology, prevention and control of vector-borne diseases.
Background Aedes (Stegomyia) albopictus (Skuse, 1894) is an anthropophilic mosquito which invaded from Southeast Asia into many countries across all continents except Antarctica [1]. This species is the vector of more than 20 arboviruses, and of which some are strongly pathogenic and transmissible viruses, such as dengue virus (DENV), Chikungunya virus (CHIKV), and yellow fever viruses [2,3].
Global warming, strong adaptability of Ae. albopictus, and human-aided transportation have contributed to the reemergence and wide spread of mosquito-borne diseases with the expansion of mosquito habitats [4][5][6].
Population genetic studies of Ae. albopictus can provide information about the transmission dynamics of specific pathogens, which aids in designing strategies for controlling vector-borne diseases [7,8].
There are approximately 50 million new cases of dengue each year, among an estimated 2.5 to 4 billion people living in over 100 countries where DENV transmission occurs [9,10]. In China, dengue fever and dengue hemorrhagic fever epidemics frequently occurred in the southern regions. Prior to 2000 this mainly included Guangdong, Hainan, Guangxi, and Fujian Provinces [11]. However, dengue fever epidemics have headed northward to Zhejiang province (in 2005 and 2011) in southeastern China and to Henan province (in 2013) in temperate central China, which is far north of any previous epidemics [12]. The recent waves of dengue outbreak in China highlight the need to improve our knowledge of Ae. albopictus population distribution and dynamics. In this study, we performed genetic analysis of Ae. albopictus based on the mitochondrial gene cytochrome c oxidase subunit 1 (cox1), which is also widely used for barcoding, and numerous published sequences from different geographical origins are available [13][14][15].
The vector competence for DENV of Aedes mosquitoes is influenced not only by the ecological structure, climate, virus titer, the serotype and virulence of DENV, but also by population genetic characteristics of mosquitoes [16][17][18][19][20]. The environmental factors and the mosquito density were used to predict the risk of dengue outbreaks. The susceptibility to dengue virus of different mosquito species are significantly different, so are those of different geographic strains of the same mosquito species. Significant variation in susceptibility to four dengue serotypes was observed among 13 geographic strains of Ae. albopictus [21] and among 13 geographic strains of Ae. aegypti [22].
Boromisa et al. [23] reported the eight geographic strains of Ae. albopictus from Asia and North America showed marked variation in susceptibility to oral infection with DENV-1. Paupy et al. [24] found that the populations of Ae. albopictus from the east coast were more susceptible to DENV-2 than those from the west coast on Réunion. The 5-to 7-day-old and non-blood-fed female mosquitoes were starved for 24 h before infection. The females were fed with an infectious blood meal (the fresh DENV-2 virus suspension was diluted in defibrinated sheep blood to 5.77 × 10 5 log 10 copies/µl). After 30 min of exposure to the infectious blood meal, fully engorged mosquitoes were transferred into the 1500-ml plastic bucket covered with nylon mesh. They were provided with 10% glucose and maintained for 14 days in an HP400GS incubator (Ruihua, Wuhai, China) at 28 °C, 80% relative humidity, and a light:dark cycle of 16 h:8 h.

Data analysis
The cox1 gene sequences from 120 mosquitoes were aligned using Clustal W multiple alignment in BioEdit version 7.2.5 [28]. The number of segregating sites (S), haplotype diversity (Hd), average number of nucleotide differences (k), and nucleotide diversity (π) within each population were determined using DnaSP version 5.10.1 [29]. Deviations from selective neutrality were tested using Fu's F s statistic [30] and Tajima's D [31]. Neutrality test was performed for each population to examine population expansion. To determine the genealogical relationships among haplotypes, a haplotype network was constructed using a statistical parsimony algorithm implemented in TCS version 1.21 [32]. The minimum number of mutational steps between sequences was calculated with > 95% confidence. Genetic differentiation among populations was estimated, and analysis of molecular variance (AMOVA) was conducted using Arlequin version 3.5.2.2 [33]. Bayesian clustering analysis was used to infer population structure in STRUCTURE V2.3 [34]. The burn-in was set to 100,000 steps and was followed by 2,000,000 Markov Chain Monte Carlo replications. All runs were repeated 20 times for each number of possible clusters from 1 to 8 (the number of populations). The optimal number of clusters (K) was determined using the Delta K method of Evanno et al. [35] in the online version of Structure Harvester v.0.6.94 [36]. Individuals were partitioned into multiple clusters according to the membership coefficient (Q) that ranges from 0 (lowest affinity to a cluster) to 1 (highest affinity to a cluster) across the K clusters. Finally, the programs CLUMPP [37] and DISTRUCT [38] were used to average replicate runs and to generate bar graphs of structure results, respectively. A total of 20 haplotypes of mtDNA cox1 were detected in the 120 specimens (GenBank: MT188111-MT188130, Additional file 1: Table S1). To determine the relationships among the samples or haplotypes, we constructed a median-joining network using haplotypes based on sequence variation ( Fig. 1). The most samples were identified as H01 (49.2%) in Ae. albopictus populations, but no haplotype H01 in Lingshui population (LS) (Fig. 1). Other haplotypes were either unique to a specific population (such as H03 in HZ; H06, H07, H08, H10 in MS; H11 in GZ; H12, H14, H15 in ZJ; H17,H18,H19,H20 in LS), or had a limited geographical distribution (such as H13 and H16 in southern China (ZJ and LS), H09 in southern-western China (MS, GZ and ZJ)) ( Fig. 1).

Genetic clustering and differentiation
Based on Bayesian clustering analysis and the Delta K method, the optimal partitioning of all samples was obtained for K = 4 with structure analysis by the Delta K method (Fig. 2). The proportional membership coefficient of individuals was showed in the pie charts from the eight Ae. albopictus populations studied (Fig. 3, Additional file 2: Table S2). The largest membership coefficient value (Q), or the proportion of individuals assigned to a cluster, were high for Lingshui (Q = 0.630), suggesting a strong affinity to be included in a single cluster (cluster 3). This cluster was shared with some of the individuals from Zhanjiang and Guangzhou, consistent with the results obtained in the network analysis (Fig. 1). Individuals from the rest populations mainly constituted three genetic clusters (clusters 1, 2, and 4; Fig. 3), suggesting a strong gene flow among these populations. Based on these results and previous study [25], we divided the eight populations into two groups: southern (GZ, ZJ and LS) and other populations (BJ, SJZ, HZ, WH, MS). AMOVA results (    The amount of dengue virus in the mosquito midguts, heads and salivary glands was measured by qRT-PCR (Fig. 4). At 14 days post-infection, the average DENV RNA copies (log 10

Correlation between population genetic indices and vector competence
In order to explore the association between population genetic structure and vector competence, we conducted Pearsonʼs correlation analysis of five genetic indices and the average dengue virus load in each tissue in different populations (Additional file 3: Table S3). The average dengue virus load in midguts and heads showed no statistically significant correlation to the five genetic indices. The average dengue virus load in salivary glands showed a negative correlation with H d (r = -0.772, P = 0.025), k (r = -0.810, P = 0.015), and π (r = -0.810, P = 0.015). However, the significance was lost after the Bonferroni correction for multiple testing (adjusted P-value < 0.00179).

Discussion
At present, there have been many studies on the application of cox1 in mosquito population genetics.
But it has not been discussed whether the genetic factors from cox1 analysis could affect mosquito vector competence. As we know, the outbreak of mosquito borne diseases would be affected by climate, mosquito density, vector competence and so on [41][42][43][44]. Would mosquito populations of different geographical strains affect their vector competence due to genetic factors, leading to differences in incidence rate of mosquito borne diseases in different areas? From this aspect, we would carry out experiments and data analysis to explore the genetic characteristics of mosquito populations in different regions and the relationship between the genetic characteristics and mosquito vector competence in these regions.
Population genetic of Ae. albopictus would be helpful for understanding the degree of population spread and establishment, which facilitate dengue prevention and vector control. In this study, the dominate haplotype H01 was detected in 7 geographic populations of mainland China, suggesting it was relatively conservative and primordial. The haplotype diversity decreased from south to north, which was similar to our previous study [25]. The suitable climate in the south may be conducive to the growth, development and reproduction of each genotype of Ae. albopictus, while the cold and dry climate in the north may play a role in mosquito genotype selection [45,46]. Results of Tajima's D and Fu's F s tests ( Table 2) indicated no sign of population bottleneck/expansion but in demographic equilibrium. Thus, these populations could have existed stably rather than recently invaded as a new founding population [47]. Our results showed that F ST -values were low and not significantly different among BJ, SJZ, HZ and WH, and they couldn't be separated by clustering analysis, whereas they were significantly different and separated in our previous study using microsatellites. These results also indicated the strong gene flow among these populations. Through comparing the results of clustering and population differentiation between this study and our previous study [25], we could infer that weaker separation between populations was obtained with compared to microsatellites, which was associated with that microsatellites evolved more rapidly than mitochondrial sequence data [48]. The F ST -value was not significantly different between ZJ and LS, but these two populations had differentiation with any other population, which might be related to the fact that the two populations had gene flow with Ae. albopictus from other countries through ships for they were close to ocean.
Divergence of specific haplotypes in ZJ and LS was observed from network structures, such as H12, H14, H15 and H19. Reasons for this phenomenon needed to explore based on more samples, in order to reveal this differentiation was caused by immigrate samples or local population divergences.
Mantel test suggested that population differentiation of Ae. albopictus had positive relation with geographic distance, which was similar to the previous study [25].
There are several factors influencing vector competence of Aedes mosquitoes, such as mosquitoes' midgut microbiota, innate immune system, climate, the virulence of the virus strain and the genetic background of mosquito populations tested [21,27,[49][50][51][52]. In this study, the eight different geographical strains of Ae. albopictus were fed with a blood meal containing the same DENV-2 strain at the same concentration and kept in the same condition for 14 days. Therefore, the vector's genetic characteristics are probably an important parameter in the observed difference in vector competence.
An advantage of the current study is that we were able to quantify the amount of DENV-2 virus in the midgut and disseminated infections through real-time quantitative PCR while other studies have only compared infection, dissemination or transmission rates [53,54]. Our results showed that there was not significant difference in IR, DR, TR and PTR among the eight population of Ae. albopictus, with the exception of the significant difference in PTR between ZJ and BJ/SJZ population. That indicated ZJ population of Ae. albopictus had lower potential population transmission rate than BJ and SJZ populations. The amount of DENV-2 virus of midguts, heads and salivary glands in GZ population of Ae. albopictus were significantly lower than that in BJ, SJZ and WH populations. The amount of DENV-2 virus of salivary glands in ZJ and LS populations of Ae. albopictus were significantly lower than that in BJ, SJZ and WH populations. We could find that the amount of DENV-2 virus in southern populations (GZ, ZJ and LS) was generally lower than other populations. This seems to be similar to the division of population genetic structure. So we also did the correlation analysis between the genetic indices of each population and dengue loads in the tissues of Ae. albopictus. However, there were no statistically significant correlation between them. One possible reason might be that the mosquito vector competence of the same species is not affected by the genetic factors of different geographical strains, but mainly affected by the climate and the microbial communities in vivo [27,[49][50][51]. Another one might be that the genetic level of the samples we collected is not enough to represent the overall level of the region, which need more collecting sites and more samples. And another one might be that other genetic markers or loci associated with immunity should have been used to analyze the population genetic structure of mosquitoes in different regions in order to establish a link with vector competence [55][56][57][58]. For example, Lambrechts et al. [59] found that Dicer-2 genotype was associated with resistance to dengue virus in Ae. aegypti. The fact of lower DENV-2 loads in southern populations might be associated with the stable immunity system established in mosquitoes due to the long-term prevalence of dengue disease in these areas [11,12].
Viral infection triggers the activation of innate immunity pathways and leads to the transcription of genes responsible for antiviral responses in mosquitoes [60][61][62]. Innate immunity is a functional response against foreign organisms, formed in the long-term evolution process of mosquitoes after frequent viral infection, including RNA interference pathways, JAK-STAT, Toll and Imd pathways [63,64]. We will further focus on studying the relationship between mosquitoes' innate immune system or immune-related genetic loci and vector competence.

Conclusions
This study reported not only the genetic structure and vector competence of Ae. albopictus from eight locations across China but also the correlation between them. Our results showed the population differentiation among Ae. albopictus was low and positively related with geographic distance, which was similar to our previous study. There was no significant correlation between vector competence and the genetic indices of each population from cox1 analysis. However, the southern populations had the low DENV-2 loads generally, which may be associated with innate immunity and worth deep research. The genetic structure and vector competence of Ae. albopictus populations may have implications for understanding the epidemiology, prevention and control of vector-borne diseases.
Further researches are needed to focus on mosquitoes' innate immune system or immune-related genetic loci to reduce vector competence for prevention and control of vector-borne diseases.  Dengue virus RNA copies in infected midguts (a), heads (b), and salivary glands (c) of Ae.

Abbreviations
albopictus in China. The results are expressed as means ± standard error (SE). Different letters with the same color above bars represent significant differences in relative expression levels at the P < 0.05 level

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.