Swab samples were collected from 34 healthy pangolins in Guangxi Province, China. The pangolins were raised in cages and fed with the same diet in the same feeding environment. The obtained anal and throat swab samples were collected from pangolins of three genetic generations—generation I (GI), generation II (GII), and generation III (GIII)—which were offspring from different parents. The pangolin samples were grouped by generation ( Table 1 ) and preserved at −80°C. Our study was approved by the Research Ethics Committee of Wenzhou University. Experiments related to viral operation were conducted in a P-3 biosafety laboratory.

All experiments employed in metavirome sample preparation and in metaviral sequencing were referenced to the approaches in our previous research (Pengpeng et al., 2018; Guanrong et al., 2022). The barcode DNA used in metavirome analysis is displayed in Table 2 .

We aligned the sequences from Illumina sequencing with non-redundant databases and virus reference databases through BLASTx and BLASTn in GenBank (https://www.ncbi.nlm.nih.gov/genbank/). Statistical significance was reached when E ≤ 10e⁻⁵ of BLAST hits. Sequences belonging to bacteria and eukaryotes were removed and virus-like sequences were analyzed.

The viral sequences with known sequences listed in the GenBank database were compared. Subsequently, specific primers were designed and synthesized according to the matching positions ( Table 3 ) to detect virus-like sequences. Abstraction of viral nucleic acid was conducted through a viral nucleic acid extraction kit (BIOER, Hangzhou, China), and amplification of viral nucleic acid was performed using PCR Master Mix (Trans, Beijing, China) and specific primers.

After amplification, the PCR products were sequenced, which were aligned with the representative viral sequences through ClustalW 2.0. Thereafter, phylogenetic trees using a maximum-likelihood algorithm with 1,000 bootstrap replicates were built with MEGA software, version 7.0.

BHK-21 and Vero cells were used in this study. The cell lines were cultured using Dulbecco’s modified Eagle’s medium (DMEM; HyClone, Logan, UT, USA), in which 8% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) and 1.2% penicillin and streptomycin (Pen Strep.; Gibco, Grand Island, NY, USA) were supplemented. Cells were cultivated at 37°C, in 5% CO₂ conditions in the thermostat. PCR-positive pangolin samples were selected for virus isolation. Briefly, the grinding liquid supernatant of the pangolin samples was mixed 1:2 with DMEM, which was supplemented with 2% FBS. In addition, the liquid mixture was inoculated on monolayer BHK-21 cells with a growth density of 70%–80%, followed by culture and observation for 5–7 days. The cytopathic effect (CPE) caused by viral infection was monitored daily. Cultures were blind passaged three times when no CPE was observed.

A PCR assay was performed for identification of the isolated CHIKV-China/P2020-1. According to the specific sequence of the CHIKV envelope 1 (E1) gene, the primers were synthesized as designed ( Table 3 ). Western blot analysis further verified the reliability with the primary antibody of the anti-E1 monoclonal antibody (Abcam, Cambridge, UK) and the second antibody of the horseradish peroxidase (HRP)-conjugated rabbit anti-mouse antibody (Trans, Beijing, China).

The newly isolated CHIKV was inoculated on BHK-21 cells and induced CPE. The cells were then harvested, subjected to three freeze–thaw cycles, and subsequently centrifuged at 10,000 × g for 12 min. Viral suspension was then obtained with centrifugation at 60,000 × g for 4 h. After gently removing the supernatant, the virus precipitate was obtained. A 6% glutaraldehyde fixative (pH 7.2) was mixed with an equal amount of DMEM, and then the precipitate was resuspended using the liquid mixture, from which 30 μl was drawn and added to a copper mesh. Finally, for negative staining, one drop of 3% phosphotungstic acid was added. Observation was performed using an electron microscope.

CHIKV was cultured and passaged in BHK-21 cells. The viral titers were tested at the 5th and the 10th generation. The viral fluid was serially diluted 10 times, followed by inoculation on monolayer BHK-21 cells with a growth density of 70%–80%, with incubation at 37°C for 120 h. Each dilution was established with eight holes and the CPE of CHIKV was observed. The TCID₅₀ (50% tissue culture infective dose) was determined according to the Reed–Muench approach (Krah et al., 1991).

BHK-21 cells produced CPE caused by the newly isolated CHIKV in the 5th and 10th generations. The total RNA of the virus was isolated from the lysate with an RNA extraction kit following the manufacturer’s instructions (Bioer Technology, Hangzhou, China). When passaged in Vero from BHK-21 cells of CHIKV (5th and 10th), which viral RNA was employed as amplification template with SuperScript™ III (Invitrogen) in accordance with the directions for use. The E1 sequences of the 5th and 10th generations of CHIKV were amplified, sequenced, and then aligned using MEGA 7.0.

SPSS 17.0 was applied for statistical analysis, using t-test for comparisons between groups. Each experiment was repeated three times. P < 0.05 was considered a significant difference. Metagenomic sequencing data have been registered in GenBank (accession nos. SRR15264668, SRR15264669, and SRR15264670). The amplified sequences were also accepted by GenBank, with accession numbers: JEV-China/P2020E-1/2/3(OM416153–OM416151), JEV-China/P2020NS4a-1/2/3/4 (OM416150–OM416147), CHIKV-China/P2020E1-12/3/4 (OM416163–OM416160), CHIKV-China/P2020NS1-1/2/3 (OM416159–OM416157), and GETV-China/P2020NS3-1/2/3 (OM416156–OM416154).

Host sequences along with the barcode DNA were excluded to acquire clean data from pangolin viromes. Using Illumina sequencing, 20,538,379 reads were obtained, with an average read length of 182.21 nt ( Table 1 ). The viral sequences from GI, GII, and GIII occupied 17.30%, 12.27%, and 15.45%, respectively. The relative abundance of the virus families is displayed in Figure 2A , with the intersection of three samples shown in Figure 2B . There were 18 virus families observed in GI, 20 in GII, and 24 in GIII. Interestingly, representative viral sequences from two families—Flaviviridae and Togaviridae—were obtained in all three groups.

After contig assembly using SOAPdenovo, 15,922 viral contigs were generated. Some virus-like contigs showed high sequence similarity; for instance, 103 JEV-like contigs had a read coverage of 42× (275–861 nt), 136 CHIKV-like contigs had a read coverage of 39× (236–917 nt), and 67 GETV-like contigs had a read coverage of 23× (219–875 nt).

The JEV-like contigs shared 80.7%–97.3% similarity with the nucleotides of the known JEV sequences, the CHIKV-like contigs had 85.6%–92.8% similarity with the nucleotides of the known CHIKV sequences, while the GETV-like contigs had 84.1%–95.2% similarity with the nucleotides of the known GETV sequences. To further verify the results of the metavirome sequencing, the identified viruses were amplified by specific primers. The verification results are detailed in the following sections.

After the amplification of CHIKV in sample GIII, we examined all anal swab samples from GIII in order to isolate CHIKV. Firstly, the isolated virus was tested using PCR, and Western blot was conducted to determine the viral expression level (E1 protein). As shown in Figure 6 , the E1 sequence and the protein were both positively expressed in viral-infected cells, but not in mock-infected cells. CHIKV-like particles can be observed through negative-stain electron microscopy, displaying diameter of approximately 70 nm, with tiny protrusions on the surface, which further confirmed the above analysis.

CHIKV-China/P2020-1 was passaged on BHK-21 cells and then the viral titer tested. Viral titers of 4.17 × 10⁴ and 2.65 × 10⁴ TCID₅₀/0.1 ml were detected in the 5th and 10th generations, respectively, indicating that multiple passages might induce viral titer changes.

The CHIKV-China/P2020-1 E1 genes of the 5th and 10th generations were sequenced and aligned. Sequence analysis revealed that the E1 sequences of the 5th generation shared 99.2% homology with the 10th generation in nucleotide sequence. CHIKV-China/P2020-1 was subsequently used to infect Vero cells, with the E1 genes of the 5th and 10th generations sequenced and their homology analyzed. Nucleotide comparison analysis revealed that 8 nucleotide sites in the E1 genes of the 5th generation and 17 sites in the 10th generation showed polymorphisms. Comparison of the amino acid sequences exhibited E1 gene mutations identified in both 5th and 10th generations: mutation from A (alanine) to V (valine) at 226 aa, which displayed the same type and site. In addition, the viral titers in the 5th and 10th generations in Vero cells were 1.08 × 10⁵ and 2.63 × 10⁵ TCID₅₀/0.1 ml, respectively, which were markedly higher than those in BHK-21 cells.