Transcriptome profiles and novel lncRNA identification of Aedes aegypti cells in response to dengue virus serotype 1

Dengue virus (DENV) is a single-stranded, positive-strand RNA virus that belongs to the family of Flaviviridae, and it is mainly transmitted by the mosquito Aedes aegypti (Ae. aegypti). Understanding the interaction of the virus with mosquito vector is vital for devising new strategies for preventing virus transmission. Although protein-coding genes have been the central focus, many reports indicated that long non-coding RNAs (lncRNAs) were also involved in virus-host interaction. Recently, the latest version of Ae. aegypti genome (AaegL5) was released, and the assembly was up to chromosome level. This prompted us to perform lncRNA identification and characterization using the latest genome release as reference. In this study, we investigated the transcriptome profiles of both protein-coding and lncRNA genes in Aedes aegypti cells upon DENV infection. By combining RNA-seq libraries generated in this study with publicly available datasets, we identified a total of 7,221 novel lncRNA transcripts, of which 3,052 and 3,620 were intronic and intergenic respectively, while 549 were antisense to the reference genes. A total of 2,435 differentially expressed transcripts, of which 956 of them were lncRNAs. Overall, the distribution of lncRNA expression and fold change upon virus infection were lower than that of protein-coding genes. We found that the expression of immune-related genes involved in IMD and MAPK signaling pathways were altered. In addition, the expression of major genes involved in RNA-interference (RNAi) pathway that confers antiviral resistance in mosquitoes were found to be unchanged upon DENV infection. Gene ontology analysis suggests that differentially expressed transcripts, either upregulated or downregulated, generally belong to the same functional categories or working in similar signaling pathways. Taken together, besides providing a new set of lncRNA repertoire, the outcomes of our study offer better understanding of Ae. aegypti responses to DENV infection at gene level. Author Summary Dengue virus (DENV), a single-stranded and positive-strand RNA virus of the family Flaviviridae, is primarily transmitted by Aedes aegypti (Ae. aegypti) mosquitoes. There are four closely related but antigenically different serotypes of dengue virus namely DENV1-4. Our understanding on the interaction of each serotype of DENV with its mosquito vector is still very limited. Since vector-borne viruses pose significant burden to public health, knowledge on the virus-host interaction at the molecular level is essential, especially in developing effective strategies to control virus transmission. In this study, we embarked on investigating the transcriptional response of long non-coding RNAs (lncRNAs) and protein-coding genes upon dengue virus serotype 1 (DENV1) infection. Besides, we also generate a comprehensive list of novel lncRNAs identified from the latest and improved genome version of Ae. aegypti. Similar to protein-coding genes, we discovered that the overall expression of lncRNA was significantly altered, suggesting that lncRNAs were involved in virus-host interaction. The results of this study provide basic understanding on the interaction between DENV1 and Ae. aegypti vector at the transcriptional level.

10 214 analysis [29]. In GO analysis of upregulated transcripts, Blast2GO recognized 38, 31, 29 GO terms in 215 biological process, molecular functions and cellular components respectively. Meanwhile, in 216 downregulated transcripts, Blast2GO identified GO terms of 61 in biological process, 33 in molecular 217 function, and 24 in cellular components. We found that in biological process category, similar GO terms 218 such as signal transduction (GO:0007165), cellular process (GO:0009987), and cellular protein 219 modification process (GO:0006464) were represented in both upregulated and downregulated GO 220 analysis (Fig 5). In fact, transcripts that could be assigned to those GO terms were among the largest in 221 number.

222
Similar observation could be found in other GO categories namely molecular function and

240
The field of lncRNA has become increasingly important in many areas of biology particularly 241 infectious disease, immunity, and pathogenesis [6,11,17,30]. High-throughput sequencing combined 242 with bioinformatics enable scientists to uncover comprehensive repertoire of lncRNA in many species.
243 Here, we present a comprehensive lncRNA annotation using the latest genome reference of Ae. aegypti 244 (AaegL5). Previous study [6] reported that a total of 3,482 intergenic lncRNA (lincRNA) was identified 245 in Ae. aegypti. However, the identification was performed using previous version of Ae. aegypti genome 246 (AaegL3), and only lncRNAs residing in intergenic region were retained.

247
Due to the recent release of Ae. aegypti genome (AaegL5) equipped with dramatically improved 248 gene set annotation [7], we decided to perform lncRNA identification using this latest genome reference.
249 Unlike previous study on Ae. aegypti lincRNA, we also annotated lncRNAs residing within the introns 250 (intronic lncRNA) and antisense to reference genes (antisense lncRNA), resulting in the identification of 251 3,052 and 549 intronic and antisense lncRNA respectively. A total of 7,221 lncRNA transcripts were 252 identified in this study, which was more than previously reported lncRNAs in Ae. aegypti. This 253 observation was presumably due to the high-quality assembly of AaegL5 genome and larger RNA-seq 254 datasets with high sequencing depth being used in lncRNA prediction pipeline.

255
Similar to previous reports, we discovered that lncRNAs identified in our study exhibited typical 256 characteristics of lncRNAs found in other species including vertebrates [10]. Such characteristics are 257 lower GC content, shorter in length, and low sequence conservation even among closely related species.
258 Upon DENV1 infection, we observed that both protein-coding and lncRNAs experienced overall changes 259 in expression level. However, the overall level of fold change displayed by lncRNAs was not as high as 271 IMD pathways [32]. In agreement with previous studies done in Ae. aegypti [33], we observed that the 272 expression level of several genes implicated in immunity was simultaneously up-and down-regulated 273 upon DENV1 infection. For example, core components of IMD signaling pathways such as NF-kB and 274 I kappa B kinase gamma-subunit were transcriptionally repressed.

275
The IMD pathway plays important role in immunity by promoting the production of anti-276 microbial peptides (AMP) such as defensin and cecropin [21]. Due to reduced expression of NF-kB, we 277 discovered that the expression of defensin in DENV1-infected cell was also reduced. On the other hand,