Insights Into an Unexplored Component of the Mosquito Repeatome: Distribution and Variability of Viral Sequences Integrated Into the Genome of the Arboviral Vector Aedes albopictus

Abstract

The Asian tiger mosquito Aedes albopictus is an invasive mosquito and a competent vector for public-health relevant arboviruses such as Chikungunya (Alphavirus), Dengue and Zika (Flavivirus) viruses. Unexpectedly, the sequencing of the genome of this mosquito revealed an unusually high number of integrated sequences with similarities to non-retroviral RNA viruses of the Flavivirus and Rhabdovirus genera. These Non-retroviral Integrated RNA Virus Sequences (NIRVS) are enriched in piRNA clusters and coding sequences and have been proposed to constitute novel mosquito immune factors. However, given the abundance of NIRVS and their variable viral origin, their relative biological roles remain unexplored. Here we used an analytical approach that intersects computational, evolutionary and molecular methods to study the genomic landscape of mosquito NIRVS. We demonstrate that NIRVS are differentially distributed across mosquito genomes, with a core set of seemingly the oldest integrations with similarity to Rhabdoviruses. Additionally, we compare the polymorphisms of NIRVS with respect to that of fast and slow-evolving genes within the Ae. albopictus genome. Overall, NIRVS appear to be less polymorphic than slow-evolving genes, with differences depending on whether they occur in intergenic regions or in piRNA clusters. Finally, two NIRVS that map within the coding sequences of genes annotated as Rhabdovirus RNA-dependent RNA polymerase and the nucleocapsid-encoding gene, respectively, are highly polymorphic and are expressed, suggesting exaptation possibly to enhance the mosquito's antiviral responses. These results greatly advance our understanding of the complexity of the mosquito repeatome and the biology of viral integrations in mosquito genomes.

Keywords: domestication; immunity; mosquitoes; piRNA pathway; repeatome; viral integrations.

FIGURE 1

NIRVS are variably distributed in SSMs. (A) Number of Flavivirus (F-NIRVS) and Rhabdovirus (R-NIRVS) loci mapping within genes, piRNA clusters or intergenic regions, classified on the basis of read-coverage across SSMs. (B) IGV screen shot showing read-coverage at AALF003313 in SSM1 and SSM2. Positions of the three AALF003313 exons, AlbFlavi4 and vepi4730383 are indicated by blue bars. (C) PCR amplification of AALF003313 exon2 in ten Ae. albopictus geographic samples. (D) F-NIRVS and R-NIRVS loci occupancy in the 16 single-sequenced mosquitoes (SSMs) of the Foshan strain is about half of that expected based on the annotated sequences of the reference genome assembly (AaloF1). F-NIRVS are in blue, R-NIRVS are in red.

FIGURE 2

Phylogeographic distribution of NIRVS. Genetic relationships among five mosquito populations shown by a Neighbor-joining trees based on shared-allele distance using data from all 13 genotyped NIRVS (A), only F-NIRVS (B), NIRVS mapping in intergenic regions (C), only R-NIRVS (D) and NIRVS mapping in piRNA clusters (E). Bootstrap values are shown at the corresponding nodes.

FIGURE 3

Prevalence of R-NIRVS (red) and F-NIRVS (blue) in each SSMs. Prevalence was calculated in each sample as the ration between detected NIRVS and annotated NIRVS for both R- and F-NIRVS.

FIGURE 4

NIRVS integration times. Boxplots showing the integration times for the NIRVS whose variability was studied across five geographic populations. Estimates are based on the D. melanogaster mutation rate, i.e., 3.5–8.4 × 10⁻⁹ per site per generation (Haag-Liautard et al., 2007; Keightley et al., 2009), and a range of generations per year between 6 and 16 to include mosquitoes from temperate and tropical regions (Manni et al., 2017). NIRVS of piRNA clusters are statistically older than NIRVS mapping in gene exons (ANOVA, ^∗∗∗P < 0.001).

FIGURE 5

NIRVS polymorphism. Volcano plot showing LoP comparison between SGs and NIRVS, genes encompassing NIRVS (Palatini et al., 2017), genes of the RNAi pathway and FGs. Entities with LoPs statistically different than that of conserved genes are above the red line [adjusted significance of the test (–log10 0.05/99 = 3.32)]. Entities on the left side of the panel (log2FC < 0) have smaller LoP than conserved genes. The opposite for entities on right side of the panel (log2FC > 0).

FIGURE 6

Expression of NIRVS mapping in coding sequences. Heatmap of the expression profiles of genes with NIRVS in the coding sequence (N-Gs) across developmental stages and body tissues (L1-L4: 1st-4th instar larvae; P: pupae; M: male whole body; SF_F/BF_F: sugar/blood fed female whole body; SF_O/BF_O: ovaries from sugar/blood fed females). Color key expresses the log10-fold change relative to larva 1st instar (calibrator). Asterisks indicate significant differences in transcript abundances (Unpaired two-tailed t-tests, ^∗P < 0.05, ^∗∗P < 0.01).

FIGURE 7

Soft clipped reads (SCR) support novel arrangements and longer than annotated viral integrations. (A) AlbFlavi6 and AlbFlavi7 are annotated on the same contig, but 5000 bp apart from each other. Soft-clipped reads (in light blue) and PCR experiments support their contiguity, with a unique ORF with similarity to the Flavivirus NS5. (B) A sequence of 212 bp extending AlbFlavi10 on its left side was identified investigating SCRs and confirmed by further sequencing. Red arrows indicate primer positions.