Gene expression profile of Aedes aegypti females in courtship and mating

Aedes aegypti is the most synanthropic and anthropophilic mosquito of Culicidae. This species always cohabits with humans and is extremely opportunistic. Vector dispersal is directly related to the ability of the females on successfully finding a mate in a generally patchy urban scenario. In the present work, we investigate transcriptional changes in Ae. aegypti females during the courtship process and after mating. We observe a substantial alteration in gene expression triggered just upon contact with Ae. aegypti males, which in turn was not fully correlated to the changes triggered by the contact. After analysing shared significant differentially regulated genes between conspecific contact and insemination, the major part of the observed transcriptomic change triggered by contact is reversed after mating, indicating an intermediary situation between naive and mating conditions that we hypothesize to be crucial for mating success. Upon contact, several chemosensory related genes are repressed, especially odorant binding proteins. Most of these genes return to higher expression rates after mating. None of these genes are significantly regulated by the encounter of a different species, Aedes albopictus. The results presented here might be applied to an innovative control approach focusing on the semiochemical systems of mosquitoes in an effort to disrupt undesirable host–insect interaction to reduce the risk of pathogen transmission to humans.

sex pheromones, cuticular hydrocarbons and other chemical cues have a major role in courtship and mating behaviour 9 .
In this study, we aim to evaluate the impact of courtship, triggered by the close proximity contact of conspecific and heterospecific (Aedes albopictus) males, on the overall Ae. aegypti female gene expression. We also assess the gene expression profile related to conspecific insemination in order to transcriptionally compare insemination and courtship processes on Ae. aegypti females. We believe that gaining insight into the processes driving mate choice or avoidance during courtship and mating, as well as revealing their underlying mechanisms and chemical cues, might allow the design of alternative control methods for these mosquito vectors.

Material and Methods
Mosquito rearing. Ae. aegypti and Ae. albopictus laboratory strains were recently established from eggs collected in Botucatu city, Sao Paulo State, Brazil (22°53′S 48°26′W) in 2015, using ovitraps. The rearing was conducted in a climate-controlled insectarium maintained at 28 °C and 70% relative humidity, with a 12 h light:12 h dark photoperiod. Firstly, eggs from the field were transferred to a tray (27 cm × 19 cm × 7 cm) containing 1 L of water to hatch. After hatching, larvae were fed with ground fish food until the pupae stage. Species and gender were determined in adults by examination of the scutum and antennae. Adults were placed in screened cages (30 cm × 30 cm × 30 cm) and fed with 10% sterile sucrose solution, which was replaced every 3 days.

Experimental groups.
For the experiments, males and females of both species were kept in different cages from the pupae stage. When the larvae changed to the pupae stage, we started to separate each pupa between males and females. The sexual identification was made by analyzing the last segment of the abdomen from the ventral part of the individual, with males having the last segment elongated and wide, while females have the same last segment thinner and short 10 . Twenty pupae from each sex and species were put separately into small trays of 8 × 5 × 3 cm containing water and then transferred to a screened cage of 30 × 30 × 30 cm. Pilot experiments were then performed to determine the mean time length for 100% insemination of Ae. aegypti and Ae. albopictus females in each conspecific crossing, putting together two males with twenty females. In these conditions, the mean time length for 100% insemination was 90 min for Ae. albopictus and 20 min for Ae. aegypti. These results helped us to define a safe time window for the females entering in contact with males and not being inseminated, therefore we used these times as standards also for heterospecific crosses.
Ae. aegypti females were submitted to four experimental conditions: (i) no contact with males (Naive); (ii) contact with Ae. aegypti males (Aeg); (iii) contact with Ae. albopictus males (Alb); (iv) inseminated by Ae. aegypti males (Ins) (Fig. 1). We used three cages for each one of the four experimental conditions as biological replicates. All mosquitos in the experiments were used 24 h after emergence from pupae. In addition, all females were anesthetised in ice and dissected, separating the head and thorax from abdomen. The insemination status was checked by dissecting female spermathecae on clean microscope slides, according to Gary and collaborators 11 .
The four groups were obtained as follows: (a) Naive: each experimental cage of this group had five naive females that had never had contact with males of any species. We checked the spermatheca of all females and they were negative for the presence of spermatozoa. Three pools each containing the heads and thoraxes of five www.nature.com/scientificreports www.nature.com/scientificreports/ individuals were put into 1.5 mL microtubes containing 500 µL of TRIzol ® and then stored at −80 °C in a freezer. This pool preparation approach was carried out for every experimental group described below. (b) Aeg: In each experimental cage, we put together two Ae. aegypti males and twenty Ae. aegypti females that had never had contact with adult males. We removed five females from the cage during four consecutive intervals of 5 min. This continued until we obtained a total of five females (one pool) that were found not inseminated after spermatheca checking. Therefore, a maximum of 20 min of contact was allowed. When sperm was found in the spermatheca, these females were discarded and not used to compose the pool of this group. (c) Alb: In each experimental cage, we put together two Ae. albopictus males and twenty Ae. aegypti females that had never had contact with adult males. We removed five females from the cage during six consecutive intervals of 15 min. As we could not find any inseminated females by checking the spermatheca after 90 min of contact, we used five females from the first interval for pooling. (d) Ins: In each experimental cage, we put together two Ae. aegypti males and twenty Ae. aegypti females that had never had contact with adult males. We removed all females after 30 min in the cage and checked the spermatheca to form a pool of five inseminated females.
RNA-Seq library preparation and sequencing. Total RNA was extracted from head and thorax pools of Ae. aegypti females using the TRIzol ® reagent protocol according to the manufacturer's instructions (ThermoFisher Scientific, Waltham, USA). RNA concentrations were quantified in each sample using a Qubit TM 2.0 fluorometer (ThermoFisher Scientific, Waltham, USA). The levels of RNA degradation were assessed by a 1% agarose gel. cDNA libraries were constructed with 200 ng of total RNA using the SureSelect Strand Specific RNA Library Prep Kit (Agilent Technologies, Santa Clara, USA) following the manufacturer's instructions. Library products were then sequenced using an Illumina Nextseq platform (Illumina, San Diego, USA) on a single-end 150 bp run.
RNA-Seq data processing and differential gene expression analysis. A CLC Genomics Workbench 7.01 platform was used to remove the adapter and assess the reads quality from the raw reads. The same platform was used to subject reads to mapping to the Ae. aegypti predicted transcriptome AaegL5.2 available at VectorBase (https://www.vectorbase.org/organisms/aedes-aegypti). Only unique gene reads count were used for posterior analysis (Supplementary Table 1). Differential expression analysis was performed with DESeq2 package version 1.24.0 12 . Genes presenting an adjusted p-value lower than 0.05 were considered as significantly regulated. Principal component analysis (PCA) data visualisation and plotting were performed in R using the pcaExplorer package 13 . Clustering analysis was performed using Euclidean distances and complete linkage method.
Gene ontology and pathway annotation. Gene ontology analysis was carried out for the identified differential expression gene lists using g:Profiler (https://biit.cs.ut.ee/gprofiler/gost). A Fisher exact test was considered as the statistical method (p-value < 0.05) and the Benjamini & Hochberg method 14 was used for false discovery rate correction (<0.05).

Results
General sequencing analysis. To understand the transcriptional changes of Ae. aegypti females due to courtship and insemination by conspecific (Ae. aegypti) and heterospecific (Ae. albopictus) males, we performed RNA sequencing (Illumina) libraries for each condition. Here, Ae. aegypti females without contact with any males, and therefore not inseminated, were considered as the Naive group. There was no evidence of insemination by Ae. albopictus males, consequently three conditions were considered besides Naive females: conspecific courtship (Aeg), heterospecific courtship (Alb) and conspecific insemination (Ins). Three biological replicates per condition were sequenced, resulting in ~19.3 million reads on average per library of which ~10.3 million were mapped to the assembly of the Ae. aegypti transcriptome (AaegL5). A total of 16,721 genes were identified in the libraries prepared after removal of low expressed genes (global counts <12). Raw reads are deposited at the NCBI BioProject database under the accession number PRJNA551490.
Unsupervised data clustering and general expression profile. The full filtered dataset was subjected to PCA and hierarchical clustering (Fig. 2a,b). Both showed consistent results for the three biological replicates. Overall, the first axis (PC1), which accounts for 30% of the observed gene expression variation in Ae. aegypti females, separated the libraries into two groups: with male contact (Aeg and Alb) and Naive/inseminated libraries. The second PCA axis (PC2-20%) then separated the heterospecific male contact (Alb) and Naive libraries from the two other conditions. Gene expression profile after male contact is depicted through a comparative heatmap of four treatment groups (Fig. 2b). Together, these results suggest that Naive Ae. aegypti females, upon contact with males, alter the expression of a determined set of genes. This expression status, after insemination, changes back to a pattern very similar to the basal Naive condition.
Differential expression genes in Ae. aegypti females after male contact. The identification of differential expression genes between conspecific/hetero-specific male contact and Naive groups was performed. Significance was defined using an adjusted p-value of <0.05. Ae. aegypti females in contact with conspecific males (Aeg vs. Naive) presented greater gene regulation compared to Ae. albopictus male contact (Alb vs. Naive). A total of 1,048 genes were shown to be up-regulated in Ae. aegypti male contact, whereas there were only 732 genes in (2019) 9:15492 | https://doi.org/10.1038/s41598-019-52268-5 www.nature.com/scientificreports www.nature.com/scientificreports/ Ae. Albopictus. Similarly, 1,140 genes were found to be down-regulated for the conspecific male contact while heterospecific accounted for 667 genes. Fewer genes were significantly differentially expressed for Aeg vs. Alb, a total of 466 genes were up-regulated and 200 were down-regulated (Fig. 3). Interestingly, when conspecific contact is compared to heterospecific contact (Alb vs. Aeg), we can observe a positive correlation between significant differentially expressed genes with 661 shared significant regulated genes. A total of 1,527 genes was found exclusively in Aeg and 738 in Alb (Fig. 4b).
Differential expression genes in Ae. aegypti females after insemination. In general, the gene expression profile of inseminated Ae. aegypti females was similar to the Naive group (Ins vs. Naive). However, the inseminated group showed a higher number of regulated genes than Naive when using Ae. aegypti male contact as the baseline (Fig. 3). Surprisingly, direct comparison between the Naive and inseminated libraries showed very few regulated genes, with 65 up-regulated and 275 down-regulated. When conspecific contact is compared to  www.nature.com/scientificreports www.nature.com/scientificreports/ inseminated females (Aeg vs. Ins), we found a strong negative correlation with 1,278 shared significant regulated genes. A total of 976 genes was found exclusively in AEG and 910 in INS (Fig. 4a).

Functional analysis of significantly regulated genes. Significantly regulated genes from Aeg vs.
Naive, Alb vs. Naive and Aeg vs. Ins comparisons were subjected to analysis using the g:Profiler web server 15 for functional analysis. Genes were ordered by log2FC and significance was defined using p-values of <0.05 after Bonferroni correction for the three classes of gene ontology (GO), which are molecular function, the biological process and the cellular component. The GO analysis of regulated genes for Aeg vs. Naive revealed two significant terms: Odorant Binding and Arginine Metabolic Process. For Alb vs. Naive, ten GO terms were found enriched: Structural Constituent of Cuticle, Arginine Metabolic Process, Alpha-amino acid Metabolic Process, Respirasome and six others related to cellular respiration. For Aeg vs. Ins, 25 GO terms were found enriched: Respirasome, Catalytic activity, acting on a protein, another 14 related to general cellular respiration and another nine related to endo and exopeptidase activity evidencing general proteolysis (Table 1).

Expression profile of chemosensory gene families.
Assuming that chemosensation is an essential sensory factor in mosquitoes to locate mates and hosts, we investigated the expression profiles of five chemosensory-related proteins families, namely, CSPs, IRs, OBPs, ORs and SNMPs (Fig. 5). A total of 65 OBPs were identified out of the 16,721 expressed genes here characterized. Apart from two genes (OBP22 and putative OBP56a), all genes suggested significant down-regulation between conspecific male contact and Naive females. Among them, 24 genes showed significant log2FC after multiple test correction. However, the Ae. albopictus male's courtship did not alter this set of genes in the same way. For the SNMPs, nine genes were identified; one was found significantly down-regulated in conspecific male contact with Naive (AAEL027927); one in the inseminated females (SCRBQ1) and one in hetereospecific contact (SCRB8). We identified 38 genes in the CSP family, among them, two were significantly up-regulated with conspecific contact and down-regulated after insemination; (AAEL005694 and AAEL002011). AAEL005694 is also significantly up-regulated upon heterospecific contact. Two other genes (AAEL005687 and AAEL001985), which are putative serine/threonine kinase, presented significant down-regulation for Aeg vs. Naive comparison. AAEL005687 is found to be significantly up-regulated in the inseminated females. Two other putative serine/threonine kinase genes (AAEL12383 and AAEL002022) were significantly regulated only after heterospecific contact. Twenty-two IRs were identified, among them, three (Ir8a, glur5(AAEL002922) and AAEL02757) were significantly regulated by conspecific contact but not by heterospecific contact, with Ir8a also being regulated by insemination. Another glur5(AAEL002538) is regulated only by heterospecific contact. We could identify 63 ORs but none of them were significantly regulated in any comparison (Fig. 5).

Discussion
In the present work, we investigated transcriptional changes in Ae. aegypti females during the courtship process with both Ae. aegypti and Ae. albopictus males and after conspecific mating. Following our experimental approach for heterospecific crosses, we could not observe heterospecific mating (satyrisation) after 90 min of contact. It is well documented that in both laboratory and field conditions, Ae. albopictus males are capable of inseminating Ae. aegypti females; however, this process only seems to occur after long periods of heterospecific contact (~3 www.nature.com/scientificreports www.nature.com/scientificreports/ weeks) in laboratory conditions. Moreover, the geographic origin of Ae. albopictus males might determine the satyrisation potential, which might be the case for our field population 16 .
We observed a substantial alteration in the gene expression triggered just upon contact with Ae. aegypti males, which in turn was distinct from changes triggered by contact with Ae. albopictus males (Fig. 2a,b). Interestingly, a significant number of regulated genes triggered by conspecific contact are reversed after mating in a switch-like manner: when a gene is found up-regulated in one condition, in the other condition it is present down-regulated and vice-versa. From 1,278 shared significant regulated genes between conspecific contact (Aeg) and insemination (Ins), only six do not follow this pattern (Fig. 4a). Conversely, when we compare conspecific contact against heterospecific contact (Alb vs. Aeg), we observe only 661 shared significant regulated genes, with 1,527 genes found exclusively in Aeg and 738 in Alb (Fig. 4b). It is noteworthy that the expression pattern shows a positive correlation for these two datasets; with the exception of 7 out 661, shared genes are regulated in a similar fashion: when a gene is found up-regulated in one condition, in the other condition it is also found up-regulated, with the same being true for down-regulated genes. These results suggest an overlap of the gene expression profile between these two conditions, reflecting core genes that are regulated upon contact regardless of male species. www.nature.com/scientificreports www.nature.com/scientificreports/ To further investigate the overall differential gene expression profile, we performed a GO analysis for the heterospecific courtship (Naive vs. Alb), insemination (Aeg vs. Ins) and conspecific courtship (Naive vs. Aeg) processes (Table 1). GO analysis of significantly regulated genes in the heterospecific courtship dataset revealed ten enriched terms, with seven of them related to general cellular respiration, two related to the bioprocessing of amino acids and one term named structural constituent of cuticle. Interestingly, cuticular hydrocarbons play a central role in the chemical communication of insects. As they are usually non-volatile, they are mainly detected by contact 9 . Although this term is related to the structural features of the cuticle, they might have a role in the efficient detection of specific cuticular hydrocarbons by the heterospecific male. In the GO analysis of the insemination process, we could find 25 enriched terms, 15 of them are related to general cellular respiration and ten of them are related to endo and exopeptidase activity evidencing general proteolysis. This result is in agreement with a previous study assessing mating-induced transcriptome changes in the reproductive tract of female Ae. aegypti 17 . Even though we did not analyze reproductive tissues in our study, the proteolysis profile was found to be very similar. In the GO analysis of the conspecific courtship, we could find two enriched terms: Odorant Binding and Arginine Metabolic process. The Odorant Binding term is comprised of genes coding for proteins related to olfaction that are usually involved in foraging, courtship, mating and the choice of oviposition. As a major representative group of these molecules, OBPs have been described as important components in the recognition of odours 18 . The Arginine Metabolic process term is comprised of genes related to the general bioprocessing of arginine with a major role on the urea cycle and was also found in the heterospecific courtship.
Taken together, these findings point towards an intermediary state between naive and mating conditions. To our knowledge, this is the first time that such an intermediary condition has been disclosed and we can hypothesize that it is crucial for mating success. Once we found the Odorant Binding term enriched throughout this intermediary state, which in turn reflects conspecific contact, we decided to investigate the expression profiles of five chemosensory related proteins families for the above-mentioned processes (Fig. 5).
From the overall analysis, we can draw the assumption that upon contact, several genes that may be related to host seeking are repressed, especially OBPs. Most of these genes return to higher expression rates after mating. Out of twenty-five differentially expressed OBP genes only two were found to be significantly regulated by the encounter of a different species, namely, Ae. albopictus. Interestingly, the inseminated females presented an overall gene expression profile very similar to the naive ones, which is also true for the OBP related genes. Conversely, www.nature.com/scientificreports www.nature.com/scientificreports/ we find an IR co-receptor (IR8a) significantly up-regulated by conspecific contact and down-regulated by insemination but not significantly regulated by heterospecific contact. These types of co-receptors form an odor-responsive ion channel with other odor-tuned IRs and are reported to detect volatile chemicals in the environment. IR8a appears to be strongly associated to host seeking behavior since it was recently shown that this co-receptor is required for mosquito attraction to lactic acid, a major component of human sweat 19 . In this case the up-regulation upon conspecific contact seems to contribute to an early priming of females for a future host seeking, reflecting a more complex interplay between genes related to courtship and mating in Ae. aegypti.
In general, insects can use a myriad of chemical cues, such as animal odours and pheromones, to encounter each other and also to detect plant nectar or vertebrate hosts 20 . These compounds are generally small hydrophobic molecules, which enter the sensory organs passing across the sensillum lymph surrounding the olfactory neurons. This passage is thought to be facilitated by OBPs, which are capable of transporting these compounds to the insect ORs 21,22 . Interestingly, Cabrera and Jaffe 23 showed that in laboratory conditions, swarming Ae. aegypti males, settled upwind in an olfactometer evoked female flight activity. More importantly, it was observed that test females reacted with flight to the odours produced by male mosquitoes, but not to the presence of odours from a rat. When the females detected the odours from a rat, they flew upwind and clustered at the area of the olfactometer where the air stream hit the cage and attempted to probe to find a blood meal. If both odours were present at the same time, the females presented the characteristic flight, suggesting that the male's odour elicited behaviour that overrode that produced by the host. This clearly indicates that when females can sense the presence of a male, they shift from their host seeking behaviour to mating behaviour, regardless of the perception of host cues. Taking the present results into account, we can draw the hypothesis that conspecific contact with males, or at least male perception by Ae. aegypti females, is an important cue for triggering a mating response and preparing females for insemination. From our point of view, this may be achieved by down-regulation of OBPs capable of switching off female host seeking behaviour.
By extrapolating these results to field conditions, we put forward a scenario to illustrate the importance of courtship in the mating behaviour of Ae. aegypti females. Once Ae. aegypti possess the so-called skip oviposition behaviour, this can shape the distribution pattern of potential mosquito breeding sites, leading to highly fragmented "breeding units". As a direct consequence of this pattern, encounters between males and females tend to be rarer, thus compromising the mating process. Thus, females need to be able to respond to male stimuli in a mandatory fashion, ignoring any other cues that are not related to male perception.
Mosquito chemical ecology is generally accepted as a major theme of investigation on which upcoming vector-borne disease control strategies may rely on. The results presented here might be applied to an innovative control approach based on mosquito chemical communications where we could develop molecular strategies (i.e. target mutagenesis in OBP related genes) to lock females in "mating mode", rendering them less effective in finding a host for blood meal and consequently reducing the transmission of vector-borne diseases.
Finally, we hope that this dataset will serve as an important resource to guide the selection of candidate genes involved in mosquito courtship and mating behaviour to be precisely targeted mutated with now available CRISPR-Cas9 tools, allowing the development of stable engineered mosquito lines to test the function of candidate genes.