Nest soil from ants was collected in September 2022 from villages in Jieyang and Yunfu cities, Guangdong Province, China. After collection, the samples were transported to the laboratory within 24 h and stored at −80°C immediately. After being extracted from the nest soil, the ants were washed with 95% ethanol. Ant species were morphologically identified by an experienced technician and confirmed by sequencing the mitochondrial 16S ribosomal RNA (16S rRNA) gene. To better identify the viruses present in the samples, the ants from the two locations were synthesized into one group. After being washed with phosphate-buffered saline (PBS), the ants were homogenized in PBS, followed by centrifugation at 2,500 × g for 5 min at 4°C. The supernatants were collected for RNA extraction with TRIzol LS reagent (Invitrogen, Carlsbad, CA, United States).

Meta-transcriptome sequencing was conducted as previously described (Shi et al., 2016). In brief, after the ribosomal RNA (rRNA) was removed, the RNA was fragmented, reverse-transcribed, and adapted, followed by paired-end (150 bp) sequencing on the Illumina HiSeq 2500 platform. All library preparation and sequencing were performed by Tianjin Novogene Bioinformatics Technology Co., Ltd., Tianjin, China.

The sequencing reads were adaptor-and quality-trimmed via the FASTP program, followed by de novo assembly via the Megahit program, and the resulting contigs were compared against the nr database via the diamond BLASTX program. The confirmed viral contigs with assembly overlaps, or from the same scaffold, were merged via the SeqMan program (version 7.1, DNAstar, Madison, WI, United States). Nested RT–PCR was conducted to fill gaps and verify the obtained sequence. The genomic terminus of the target viruses was determined via 5′/3′ RACE kits (TaKaRa, Dalian, China). The primers used are shown in Supplementary Table S2. The reads were mapped to the target contigs via Bowtie 2, and an integrated genomics viewer (IGV) was used to check for assembly faults to validate the assembly results.

The discovered viruses were classified on the basis of their nucleotide (nt) and amino acid (aa) identities. If the species demarcation criteria remain unclear within a genus, a novel viral species is defined if it holds less than 80% nt identity across the complete genome, or less than 90% aa identity of the RNA-dependent RNA polymerase (RdRp) domain with known viruses. All the novel viruses were named “Guangdong ant,” followed by common viral names according to their taxonomy. Viruses classified into established taxonomies were marked with “Guangdong (GD)” to distinguish them from other viral strains.

Reference virus sequences were acquired from the GenBank database and examined via MAFFT v7.450 to determine the phylogenetic relationships of the identified viruses (Katoh et al., 2019). After testing with ProtTest 3.4, phylogenetic trees were constructed via the maximum-likelihood method in PhyML v3.0 with the Le and Gascuel substitution model for amino acid sequence analysis, which is based on a bootstrap value of 1,000 replicates (Darriba et al., 2011; Guindon et al., 2005).

A total of 432 ants were collected from two distinct locations within Guangdong Province, China Jieyang (n = 324) and Yunfu (n = 108). Species identification was conducted through a comprehensive amalgamation of morphological scrutiny and 16S PCR techniques. Among the collected samples, 411 were unequivocally identified as Solenopsis invicta, and the remaining 21 were characterized as Polyrhachis dives and were excluded from this study (Supplementary Table S1; Supplementary Figure S1).

Ants were pooled into one group for RNA library construction and sequencing. After quality control and adapter trimming, a total of 85,900,504 paired-end clean reads were generated in these libraries, resulting in 79,054 viral reads, which accounted for 0.1% of the total RNA reads, and were assembled into 95 viral contigs. After alignment via BLAST, the viral contigs were ultimately annotated to 17 viruses from the viral families Polycipiviridae, Dicistroviridae, Iflaviridae, Parvoviridae, and unclassified families. The raw data and obtained virus sequences have been uploaded to NCBI with Bioproject Accessions: PRJNA1061181; Biosample Accessions: SAMN39259552; and SRA accessions: SRR27458542.

The genomes of 17 viruses, including 15 previously unknown viruses, were obtained through contig-based PCR and rapid amplification of cDNA ends (RACE). The complete genomes of 14 viruses were obtained and named Guangdong Dicistroviridae ant virus 2–3 (GDAV2-3), Guangdong Iflaviridae ant virus 4–9 (GIAV4-9), Guangdong Parvoviridae ant virus 10 (GPAV10), Guangdong ant virus 12,13,15 (GAV12,13,15), Solenopsis invicta virus 1-GD (SINV1 GD) and Solenopsis invicta virus 4-GD (SINV4 GD). Additionally, three viruses with partial sequences were identified and named Guangdong Polycipiviridae ant virus 1 (GPAV1), Guangdong ant virus 11 (GAV11) and Guangdong ant virus 14 (GAV14) (Table 1).

Metagenomic analysis revealed that approximately 2.2% (1,722/79,054) and 0.2% (195/79,054) of the total reads mapped to the genomes of SINV4 GD and GAV7, respectively (Supplementary Table S3). We obtained two complete viral genomes belonging to Polycipiviridae and designated them SINV4 GD and GAV7.

The genome of SINV4 GD includes five ORFs (Figure 1) encoding the RdRp and RNA helicase. The length of the SINV4 GD is 11,965 bp, and it shares high identity (96.1% of the RdRp aa sequence, 90.8% of the complete nt sequence) with Solenopsis invicta virus 4, which was identified in the United States (Olendraite et al., 2017). The GPAV1 genome is 10,504 bp in length, and its RdRp gene shares 83.9 and 85.8% nt and deduced aa sequence identity with the closest viral strain, Polycipiviridae sp., which was identified in China (Chen et al., 2022). Phylogenetically, both SINV4 GD and GPAV1 identified in this study were grouped into the Polycipiviridae family.

In total, 28,419 reads (35.9%; 28,419/79,054) were mapped to viruses in the Dicistroviridae family. A total of 0.3% of the reads (209/79,054) in the pools were matched with SINV1 GD, 0.2% of the reads (261/79,054) were mapped with GDAV2, and 35.4% of the reads (27,994/79,054) were matched with GDAV3 (Supplementary Table S4).

The length of the SINV1 GD was 9,944 bp, with high identity (98.2% of the RdRp aa sequence and 94.2% of the complete nt sequence) to Solenopsis invicta virus 1, which was identified in the United States (Valles et al., 2013). The length of GDAV3 is 10,144 bp, and its RdRp gene shares 45.9 and 66.9% nt and deduced aa sequence identity, respectively, with the Wuhan arthropod virus 2 strain WHSFII47608 (Shi et al., 2016). The length of GDAV2 was 7,418 bp, and its RdRp gene shares 62.0 and 61.7% nt and deduced aa sequence identity with the strain Beihai picorna-like virus 76 (Supplementary Table S4).

Phylogenetically, the SINV1 GD, GDAV2, and GDAV3 genes identified in this study were grouped within the Dicistroviridae clade (Figure 2).

Metagenomic analysis revealed that a total of 6.8% (5,408/79,054) of the viral reads mapped to the genomes of the iflavirus, which included 1.4% (1,114/79,054) GIAV4, 0.4% (345/79,054) GIAV5, 0.2% (172/79,054) GIAV6, 0.6% (506/79,054) GIAV7, 2.2% (1,743/79,054) GIAV8, and 1.9% GIAV9 (1,528/79,054) reads.

The length of GIAV4 is 9,727 bp, and its RdRp gene shares 61 and 59% nt and deduced aa sequence identity with the infectious flounder virus strain (Isawa et al., 1998). The length of GIAV5 was 8,459 bp, and its RdRp gene shares 58.2 and 44% nt and deduced aa sequence identity with the strain King virus. The length of GIAV6 is 9,827 bp, and its RdRp gene shares 47 and 51.2% nt and deduced aa sequence identity with the strain Opsiphanes invirae iflavirus 1 (Silva et al., 2015). The length of GIAV7 is 7,799 bp, and the RdRp gene of GIAV7 is 46.1% nucleotide and has amino acid sequence identities with the Lampyris noctiluca iflavirus 1 strain, with amino acid identities ranging from 33.9 to 34.9%. Compared with those of the Moku virus isolate Big Island, the lengths of GIAV8 and GIAV9 are 10,320 bp and 10,336 bp, indicating nucleotide identities of 36 and 36.3%, respectively, and amino acid identities of 56.1 and 54.8%, respectively, in the RdRp region (Supplementary Table S5) (Mordecai et al., 2016).

Phylogenetically, the six viruses identified in this study were grouped into the Iflaviridae clade (Figure 3).

Approximately 44.4% of the reads (35,116/79,054) were mapped to the genome of Guangdong Parvoviridae ant virus 8 (GPAV10), which is located in the genus Pefuambidensovirus in the phylogenetic tree. We designated this virus GPAV10, whose genome shares 65.7% and 26.7 ~ 67.9% nt and deduced aa sequence identity with Periplaneta fuliginosa densovirus (Guo et al., 2000) (Figure 4).

The length of GAV11 is 5,249 bp, and its RdRp gene shares 49 and 91% nt and deduced aa sequence identity with the strain Dicistroviridae sp. 1. The length of GAV14 is 7,345 bp, and its genome shares 45.3% nt identity with that of the Lycopersicon esculentum picorna-like virus. The length of GAV12 is 9,640 bp, and its RdRp gene shares 44% nt and 32% nt and deduced aa sequence identity with the strain Beihai picorna-like virus 70 (Shi et al., 2016). For GAV11, GAV14 and GAV12 may only have partial RdRp information or high diversity with their closest viruses; thus, these viruses were deemed unclassified viruses in this study.

Vespa velutina-associated acypi-like virus, first discovered in the Asian yellow-legged hornet Vespa velutina nigrithorax, is an unclassified virus (Dalmon et al., 2019). In this study, a total of 344 (0.4%, 344/79,054) reads were mapped to the Vespa velutina-associated acypi-like virus. We named this virus Guangdong ant virus 13 (GAV13), which has 9,860 nt and shares 46.3% and 98 ~ 98.9% nt and deduced aa sequence similarity with the Vespa velutina-associated acypi-like virus.

Changjiang crawfish virus 6 is an unclassified RNA virus (Shi et al., 2016). In this study, a total of 3,214 (4.1%, 3,214/79,054) reads were mapped with the Changjiang crawfish virus 6. We named this virus Guangdong ant virus 15 (GAV15), whose genome is 9,901 nt long and shares 44.9 and 36.5% nt and whose aa sequence similarity is with that of the Changjiang crawfish virus 6.