In September and October 2018, 5,500 female mosquitoes, including live and newly dead mosquitoes, were gathered in Yunnan Province, China. The obtained mosquitoes covered Armigeres obturbans (Ar. obturbans), Anopheles sinensis (An. sinensis), Culex quinquefasciatus (C. quinquefasciatus), Culex tritaeniorhynchus (C. tritaeniorhynchus), Aedes albopictus (Ae. albopictus), and Aedes aegypti (Ae. aegypti). After identifying the mosquito species, we put every 100 mosquitoes of the same species in a cryopreservation tube ( Table S1 ). Then, based on the collection location, we grouped the mosquito samples ( Table 1 ) with storage at −80°C for later use. This study was approved by the Ethics Committee and the Research Ethics Committee of Wenzhou University. All studies relating to active viruses were performed in bio-safety level-3 laboratory.

The methods used for metagenomic sample preparation and metaviral sequencing were reference to methods of our previous study (Xiao et al., 2018). The barcode DNA used in metagenomic analysis is shown in Table 2 .

The sequences were aligned with non-redundant databases and virus reference databases using BLASTx and BLASTn in GenBank. (https://www.ncbi.nlm.nih.gov/genbank/) The E value ≤10 e^-5 in BLAST hits was considered statistically significant. Sequences of bacterial and eukaryotes were deleted, and virus-like sequences were analyzed.

Viral sequences were aligned with the corresponding viruses in GenBank, the matching positions were found, and then specific primers were designed and synthesized ( Table 3 ) for identifying the virus-like sequences. The viral nucleic acid was extracted using a viral nucleic acid extraction kit from Bioer Technology (Hangzhou, China), followed by amplification with PCR Master Mix (Tiangen, Beijing, China) and designed primers.

The PCR products were sequenced, and the obtained sequences were used to make an alignment with the representative viruses via Clustal W version 2.0. The MEGA 7 was used to build phylogenetic trees based on the tree-building model of maximum-likelihood method with 1,000 bootstrap replicates.

In this study, BHK-21 and Vero cell lines were cultured in DMEM (HyClone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS. HyClone, Logan, UT, USA) and 1% penicillin and streptomycin (Pen Strep. HyClone, Logan, UT, USA); cells were maintained in humidified 5% CO₂ incubator at 37°C.

Mosquito samples verified to be positive by PCR were used for virus isolation. In short, the mosquito sample solution was mixed 1:6 with DMEM containing 2% FBS, followed by inoculation on a monolayer of BHK-21 cells and culture together for 5–6 days. We checked the culture daily for virus-induced cytopathic effect (CPE). Cultures without a CPE were blind passaged 3 times.

After isolation, the obtained virus CHIKV-China/YN2018-1 was first identified by PCR technology. Based on the envelope 1 (E1) protein gene ( Table 3 ), specific paired primers were designed and synthesized. In addition, it was further identified by Western blotting. Anti-E1 monoclonal antibody (Abcam, Cambridge, UK) was used as the primary antibody, and an HRP-conjugated goat anti-mouse antibody (ZSGB-Bio, Beijing, China) was used as the second antibody.

After verification by PCR and Western blotting, the newly isolated CHIKV was inoculated on a monolayer of BHK-21 cells to produce CPE. Then, the cells were collected and repeatedly frozen and thawed three times. After that, the virus suspension was harvested by centrifuging at 10,000 × g for 15 min, followed by centrifugation at 60,000 × g for 4 h. We got the virus precipitate by gently removing and discarding the supernatant, and equal proportion mixture of 6.1% (V/V) glutaraldehyde fixative (pH 7.2) and DMEM was used to resuspend the precipitate, and 25 μl was added to the copper grid, followed by the addition of one drop of 3% phosphotungstic acid for negative staining and observation using a microscope.

The viral titers of the 5th- and 10th-generation CHIKV were detected in BHK-21 cells. The CHIKV was treated with serial tenfold dilution and inoculated in cells, and incubated for 120 h at 37°C. The CPE of CHIKV was assessed by using eight replicates in each dilution. The TCID₅₀ (50% tissue culture infective dose) was acquired based on the Reed-Muench method (Krah, 1991).

CPE was induced in BHK-21 cells with the 5th- and 10th-generation CHIKV. Viral RNA was extracted from the supernatant using an RNA viral kit according to the manufacturer’s protocol (Bioer Technology). After passage in the Vero cell line from the BHK-21 cell line of CHIKV (P5 and P10), the purified RNA was used as template for cDNA synthesis using a SuperScript™ III first-strand synthesis system (Invitrogen) following the manufacturer’s instructions. Subsequently, we amplified the E1 genes of passages 5 and 10 of CHIKV and sequenced them. The four E1 genes were aligned with MEGA 7.0.

All data were input in SPSS 17.0 for statistical analysis. t-test was applied in the comparison between groups. Each experiment was set in triplicate, and each independent experiment was repeated three times. It was considered statistically significant when the value of p < 0.05.

We have deposited the metagenomic sequencing data in the GenBank Sequence Reads Archive. The obtained accession numbers were SRR15058239 to SRR15058241. We deposited the amplified sequences in the GenBank, whose accession numbers were as follows: CHIKV-China/YN2018E1-1/2/3/4 (MZ494489–MZ494492), CHIKV-China/YN2018NS3-1/2/3/4/5 (MZ494493–MZ494497), GETV-China/YN2018E2-1/2/3 (MZ494498–MZ494500), GETV-China/YN2018NS3-1/2 (MZ494501–MZ494502), and RRV-China/YN2018C-1/2 (MZ494503–MZ494504), respectively.

In order to acquire clean data of mosquito viruses, we eliminated host sequences and barcode DNA. A total of 25,768,166 reads were obtained by Illumina sequencing, and the mean read length was 184.13 nt ( Table 1 ). The content of viral sequences in samples I, II, and III accounted for 17.79%, 7.27%, and 18.83%, respectively, of which the relative abundance of viral families is shown ( Figure 2A ), and their intersection of three samples is illustrated ( Figure 2B ). Viral families in samples I, II, and III were 9, 16, and 13 respectively. Interestingly, all three samples contained representative sequences from 2 viral families: Togaviridae and Bunyaviridae.

SOAPdenovo was performed to assemble virus sequences into contigs, and 14,127 viral contigs were obtained, of which many virus-like contigs were remarkably similar, such as 63 CHIKV-like contigs with a read coverage of 37× (292–906 nt), 126 GETV-like contigs with a read coverage of 54× (226–872 nt), and 35 RRV-like contigs with a read coverage of 12× (235–843 nt). The CHIKV-like contigs have 92.6%–95.1% nt homology with the known CHIKV sequences. The GETV-like contigs have 83.7%–91.3% nt homology with the known GETV sequences. The RRV-like contigs have 81.6%–90.5% nt homology with the known RRV sequences. Next, we designed and synthesized specific primers to amplify the identified viruses, to further confirm the results of metavirome sequencing ( Table 3 ). The results of PCR verification of mosquito virome were as follows.

Considering the successful amplification of CHIKV in sample II, we tried to validate all mosquito samples from sample II and then detected and isolated CHIKV from Aedes albopictus samples. The obtained CHIKV strain was first detected by PCR experiment. Then, Western blot was performed to further detect the levels of intracellular virus (E1 protein). As predicted, E1 gene/protein was positively expressed in CHIKV-infected cells rather than mock-infected cells. In order to further verify the isolation of CHIKV, we observed the virus particles through negative-stain electron microscopy, which were shown rounded with approximately 70 nm diameter and tiny protrusions on the surface, which presented CHIKV-like particles ( Figure 6 ).

After inoculation in BHK-21 cells and calculation, the viral titers of CHIKV-China/YN2018-1 of the 5th and 10th generation were 2.65 × 10⁴ TCID₅₀/0.1 ml and 5.14 × 10⁴ TCID₅₀/0.1 ml, which suggested that multiple passages might lead to the change of viral titers.

After sequencing, MegAlign was utilized to align the CHIKV-China/YN2018-1 E genes of P5 and P10. Nucleotide identity analysis showed that E1 gene of P5 shared 99.2% nucleotide identity with that of P10. Then, follow-up infection was performed on Vero cell lines, and the 5th and 10th generations of CHIKV-China/YN2018-1 E genes were sequenced. Nucleotide comparison analysis showed that 9 nucleotide sites of P5 and 14 nucleotide sites of P10 of E genes presented polymorphism. Alignment of the amino acid sequence showed that one vital amino acid site (position: 226 aa) of E1 protein in P5 and P10 was mutated, with amino acid A (alanine) changed to V (valine). Moreover, after detecting the viral titers in Vero cells, the P5 and P10 of CHIKV-China/YN2018-1 were 1.67 × 10⁵ TCID₅₀/0.1 ml and 2.14 × 10⁵ TCID₅₀/0.1 ml, which were markedly elevated (p < 0.05) in comparison with detection in BHK-21 cells.