Diverse RNA viruses in the venom-related microenvironment of different animal phyla

Abstract

Venom is known as the source of natural antimicrobial products. Previous studies have largely focused on the expression of venom-related genes and the biochemical components of venom. With the advent of metagenomic sequencing, many more microorganisms, especially viruses, have been identified in highly diverse environments. Herein, we investigated the RNA virome in the venom-related microenvironment through analysis of a large volume of venom-related RNA-sequencing data mined from public databases. From this, we identified viral sequences belonging to thirty-six different viruses, of which twenty-two were classified as ‘novel’ as they exhibited less than 90 per cent amino acid identity to known viruses in the RNA-dependent RNA polymerase. Most of these novel viruses possessed genome structures similar to their closest relatives, with specific alterations in some cases. Phylogenetic analyses revealed that these viruses belonged to at least twenty-two viral families or unclassified groups, some of which were highly divergent from known taxa. Although further analysis failed to find venom-specific viruses, some viruses seemingly had much higher abundance in the venom-related microenvironment than in other tissues. In sum, our study provides insights into the RNA virome of the venom-related microenvironment from diverse animal phyla.

1. Introduction

Venom is a known source of natural antimicrobial products (da Mata et al. 2017). It is generally held that venom and venom-producing tissues are largely sterile environments (Ul-Hasan et al. 2019), and previous studies have largely focused on the biochemical components of venoms and the gene expression of venom-producing tissues (Robinson et al. 2017). However, recent studies have revealed that there may be hidden diverse microbial communities in the microenvironment of venom or venom-producing tissues, such as the venom gland and venom duct (Ul-Hasan et al. 2019; Esmaeilishirazifard et al. 2022). These microorganisms can coexist with the host and even play roles in host reproduction (Zhu et al. 2018). As venom can be used as a defensive weapon in animals, its biochemical components can poison other animals and the microorganisms present in venom may be injected into other animals to cause infection or modulate the host immune reaction (Monteiro et al. 2002; Coffman and Burke 2020).

Metagenomic (DNA) and metatranscriptomic (RNA) sequencing approaches have facilitated the identification of novel and existing microorganisms. In recent years, thousands of novel viruses have been identified in this manner (Shi et al. 2016, 2018; He et al. 2022), with many more to be discovered (Carroll et al. 2018). Viruses are the most abundant biological entities on earth (Paez-Espino et al. 2016), although the vast majority have still to be described (Geoghegan and Holmes 2017; Zhang, Shi, and Holmes 2018).

To date, high-throughput sequencing of venom-producing tissues has resulted in the discovery of many novel venom peptides and accumulation of a large volume of transcriptomic sequencing data (Robinson et al. 2017). However, little is known about the viruses in the venom-related microenvironment (including venom and venom-producing tissues) of animals. Herein, we used these publicly available transcriptome sequencing data to explore the RNA virome in the venom-related microenvironment of diverse animals, representing those in the phyla Chordata, Mollusca, Arthropoda, Annelida, and Cnidaria. In doing so, we identified thirty-six different RNA viruses in thirty-three of the 474 venom-related transcriptome datasets, some of which were phylogenetically distant from known viruses. In addition, we characterized the genomic and phylogenetic features of these viruses and explored their abundance among venom-related tissues and non-venom tissues.

2. Materials and methods

2.1. Dataset retrieving and processing

RNA-sequencing (Seq) data from the venom-related microenvironment of animals was retrieved and downloaded from the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database with the keywords ‘venom’ and ‘RNA-seq’. The final data set contained 474 paired-end SRA files from the venom-related microenvironment of 151 species spanning the phyla Chordata, Arthropoda, Mollusca, Annelida, and Cnidaria (Figure S1 and Table S1), with a total of 2,702.87 Gb.

2.2. Sequence assembly and virus discovery

For each RNA-seq dataset, sequencing reads were first adapter trimmed and filtered using Fastp v0.19 (Chen et al. 2018), with the ribosomal RNA reads then removed. The remaining reads were de novo assembled with Trinity v2.5.1 (Grabherr et al. 2011) using default parameters. The assembled contigs were searched against the non-redundant nucleotide (nt) and non-redundant protein (nr) databases from NCBI using Basic Local Alignment Search Tool (BLAST) related toolkits, including BLASTn (Camacho et al. 2009) and Diamond BLASTx (Li et al. 2021), respectively. Viral contigs were identified, and by comparing the virus contigs with the known closest references, viral contigs with near complete genomes were subjected to further analyses.

2.3. Virus genome annotation

Open reading frames (ORFs) within the virus contigs were predicted using the ‘Find ORFs’ function of Geneious v2021.0.1. The closest virus references for each virus contig were downloaded from NCBI. Conserved domains (CD) of the virus contigs and their closest relatives were identified using CD search against the NCBI CD Database.

2.4. Phylogenetic analysis

To infer the evolutionary relationships of the newly identified viruses, the closest relatives and representative viral protein sequences of associated genera were retrieved by BLASTx by comparing the new viral contigs against the nr database (NCBI). Sequences of the relatively well-conserved RNA-dependent RNA polymerase (RdRp) were selected for phylogenetic analysis. Multiple sequence alignment was performed using MAFFT v7.407 (Katoh and Standley 2013) employing the L-INS-I algorithm, with ambiguously aligned regions removed using TrimAl v1.4 (Capella-Gutiérrez, Silla-Martínez, and Gabaldón 2009). Phylogenetic trees were inferred using the maximum likelihood method implemented in IQ-TREE v2.2.0 (Minh et al. 2020) with 1,000 bootstrap replicates, and ModelFinder (Kalyaanamoorthy et al. 2017) was used to find the best-fit substitution model in each dataset.

2.5. Abundance calculation of RNA viruses

Based on the data with ribosomal RNA reads removed, the abundance of the newly found RNA viruses was calculated as Reads Per Million (RPM) using the formula, ‘Viral reads/Total non-rRNA reads × one million’. Bowtie2 v.2.4.2 (Langmead and Salzberg 2012) was used to align reads to the virus genomes. SAMtools v1.10 (H. Li et al. 2009) was used to output the number of matched reads from the Bowtie2 SAM result.

3. Results

3.1. Overview of the venom-related RNA-seq data

The RNA-seq data of the venom and venom-related tissues (including venom gland, venom duct, and venom bulb) were retrieved and downloaded from the NCBI SRA database. Only paired-end sequencing data were analysed. The final dataset finally comprised 474 SRA files from 151 species spanning the phyla Chordata (eighty-four species, 353 SRA files), Arthropoda (thirty-nine species, seventy-three SRA files), Mollusca (twenty-five species, forty-four SRA files), Annelida (one species, two SRA files), and Cnidaria (two species, two SRA files), with total sequencing bases of 2,702.87 Gb (Fig. S1 and Table S1). Further analysis showed that the phylum Chordata mainly included samples from Colubroidea, the phylum Arthropoda mainly included samples from Araneae, Scorpiones, Scolopendromorpha, Diptera, and Hymenoptera, while the phylum Mollusca contained samples from Conidae and Turridae (Fig. S1). A total of 160 SRA files had information on collection locations, involving at least nineteen countries or regions (Fig. S2). The samples were mainly collected from North America (n = 68), Asia (47), and South America (21), and the three countries with the most samples were the USA (n = 60), Brazil (n = 20), and China (n = 16).

3.2. Diverse RNA viruses in the venom-related microenvironment

Sequencing reads were de novo assembled into contigs, which were further analysed by BLAST against the NCBI nr/nt databases. A total of forty-five virus strains with complete or partial genome sequences containing at least the RdRp region were recovered from thirty-three RNA-seq data sets. These forty-five virus strains likely represented thirty-six different viruses (Fig. 1 and Table 1), including twenty-eight negative-sense single-stranded (−ssRNA), thirteen positive-sense single-stranded (+ssRNA), and four double-stranded RNA (dsRNA) viruses (Table S3). BLAST analysis revealed that approximately half of virus contigs were divergent from the known viruses: twenty-two viruses shared less than 90 per cent amino acid identity in RdRp with the most closely related viruses, including sixteen that exhibited <70 per cent amino acid identity to known viruses (Table 1).

Figure 1.

Viral species identified in the venom-related microenvironment from diverse animals. The number of viral species identified in the venom-related microenvironment is summarized according to the respective host organism. The size of the points indicates the number of viral species. The host genus Crotalus marked with a star includes the species Crotalus cerastes, Crotalus durissus terrificus, Crotalus molossus, Crotalus morulus, Crotalus polystictus, and Crotalus totonacus. The colours of the horizontal bars and animal icons represent host groups; brown, pink, and grey indicate Arthropoda, Chordata, and Mollusca, respectively. The aquamarine, royal blue, and light green of the vertical bars represent negative-sense single-stranded, positive-sense single-stranded, and double-stranded RNA viruses, respectively.

Table 1.

Summary of the viruses identified from the venom-related microenvironment in this study.

No.	Virus name	Host	Baltimore classification	Families or groups	Length (nt)	GenBank accession numbers	Closest relative	RdRp aa identity (%)
1	Crassispira cerithina rhabdovirus b	Crassispira cerithina	(−)ssRNA	Rhabdoviridae	11,539	BK064875	Lyssavirus mokola	48.09
2	Conus episcopatus rhabdovirus b	Conus episcopatus	(−)ssRNA	Rhabdoviridae	5,738	BK064874	Anole lyssa-like virus 1	56.79
3	Urolepis rufipes peropuvirusa	Urolepis rufipes	(−)ssRNA	Artoviridae	12,176	BK064876	Pteromalus puparum negative-strand RNA virus 1	88.35
4	Tapajós virus	Bothrops atrox	(−)ssRNA	Filoviridae	16,181	FAA04063	Tapajós virus	100.00
5	Microplitis mediator mononega-like virus strain CHBb	Microplitis mediator	(−)ssRNA	Xinmoviridae	12,918	BK064877	Gudgenby Calliphora mononega-like virus	53.96
6	Crotalus cerastes sunvirus strain Ccera1a	Crotalus cerastes	(−)ssRNA	Sunviridae	17,220	BK064879	Sunshine Coast virus	86.94
7	Hemiscolopendra marginata myriavirusb	Hemiscolopendra marginata	(−)ssRNA	Myriaviridae	14,106	BK064881	Hubei myriapoda virus 8	63.97
8	Crotalus durissus terrificus chuvirus	Crotalus durissus terrificus	(−)ssRNA	Chuviridae	10,645	BK064882	Herr Frank virus 1	91.17
9	Crotalus chuvirus strain Ctoto1 a	Crotalus totonacus	(−)ssRNA	Chuviridae	11,043	BK064883	Guangdong red–banded snake chuvirus-like virus	74.20
10	Crotalus totonacus bornavirus strain Ctoto1	Crotalus totonacus	(−)ssRNA	Bornaviridae	8,932	BK064870	Caribbean watersnake bornavirus	92.85
11	Agkistrodon piscivorus bornavirus	Agkistrodon piscivorus	(−)ssRNA	Bornaviridae	8,876	BK064872	Caribbean watersnake bornavirus	93.61
12	Crotalus molossus bornavirus	Crotalus molossus	(−)ssRNA	Bornaviridae	8,938	BK064873	Mexican black-tailed rattlesnake bornavirus	100.00
13	Conus consors orthomyxo-like virus b	Conus consors	(−)ssRNA	Orthomyxoviridae	1,431/1,989	BK064896–BK064897	Barns Ness dog whelk orthomyxo-like virus 1	51.30
14	Xibalbanus tulumensis orthomyxo-like virusb	Xibalbanus tulumensis	(−)ssRNA	Orthomyxoviridae	1,469/1,359/1,328/1,446/1,039	BK064898–BK064902	Jos virus	53.17
15	Microplitis mediator orthomyxo-like virus b	Microplitis mediator	(−)ssRNA	Orthomyxoviridae	2,442/2,484/2,284/1,772/1,711	BK064903–BK064907	Hymenopteran orthomyxo-related virus OKIAV173	61.07
16	Tetrastichus brontispae orthomyxo-like virus 1b	Tetrastichus brontispae	(−)ssRNA	Orthomyxoviridae	2,434/2,503/2,469/1,786/1,690/1,103	BK064908–BK064913	Mason Creek virus	62.08
17	Tetrastichus brontispae orthomyxo-like virus 2 b	Tetrastichus brontispae	(−)ssRNA	Orthomyxoviridae	2,403/2,548/2,420/1,802/1,707	BK064914–BK064918	Longchuan virus	58.73
18	Micrurus arenavirusb	Micrurus surinamensis	(−)ssRNA	Arenaviridae	6,723/1,890/1,388	BK064890–BK064892	Unidentified Reptarenavirus	40.01
19	Tetragnatha versicolor phenuivirus b	Tetragnatha versicolor	(−)ssRNA	Phenuiviridae	6,145/3,334/853	BK064893–BK064895	FinV707 virus	45.95
20	Israeli acute paralysis virus strain SKS	Apis cerana	(+)ssRNA	Dicistroviridae	9,584	BK064921	Israeli acute paralysis virus	100.00
21	Diversinervus elegans virus strain Tbron	Tetrastichus brontispae	(+)ssRNA	Dicistroviridae	8,832	BK064919	Diversinervus elegans virus	100.00
22	Pteromalus puparum dicistrovirus a	Pteromalus puparum	(+)ssRNA	Dicistroviridae	9,029	BK064920	Soybean thrips picorna–like virus 1	83.02
23	Apis picorna–like virus 4 strain SKS	Apis cerana	(+)ssRNA	Solinviviridae	10,152	BK064922	Apis picorna–like virus 4	96.04
24	Sacbrood virus strain SKS	Apis cerana	(+)ssRNA	Iflaviridae	8,792	BK064925	Sacbrood virus	99.68
25	Tetrastichus brontispae RNA virus 3 strain CYN	Tetrastichus brontispae	(+)ssRNA	Iflaviridae	9,971	BK064923	Tetrastichus brontispae RNA virus 3	100.00
26	Nasonia vitripennis virus strain VG	Nasonia vitripennis	(+)ssRNA	Iflaviridae	10,075	BK064924	Nasonia vitripennis virus	99.69
27	Tetramorium bicarinatum iflavirus	Tetramorium bicarinatum	(+)ssRNA	Iflaviridae	10,231	BK064926	Iflaviridae sp.	99.00
28	Phoneutria nigriventer picorna–like virus b	Phoneutria nigriventer	(+)ssRNA	Unclassified Picornavirales	8,942	BK064931	Riboviria sp.	65.59
29	Solenopsis invicta virus 17 strain Ptepi	Parasteatoda tepidariorum	(+)ssRNA	unclassified (Virga-Kita Clade)	10,267	BK064929	Solenopsis invicta virus 17	100.00
30	Tetragnatha versicolor virus a	Tetragnatha versicolor	(+)ssRNA	unclassified (Nido Clade)	19,716	BK064930	Hubei tetragnatha maxillosa virus 7	86.13
31	Pelteobagrus fulvidraco oncotshavirus-1 strain VG	Tachysurus fulvidraco	(+)ssRNA	Tobaniviridae	27,004	BK064927	Pelteobagrus fulvidraco oncotshavirus-1	100.00
32	Scorpaenopsis potexvirusa	Scorpaenopsis cirrosa	(+)ssRNA	Alphaflexiviridae	6,204	BK064928	Hydrangea ringspot virus	79.87
33	Hemiscolopendra marginata reovirus b	Hemiscolopendra marginata	dsRNA	Spinareoviridae	4,333/4,070/3,147/1,892	BK064933–BK064936	Shelly headland virus	27.66
34	Microplitis mediator partiti-like virus b	Microplitis mediator	dsRNA	Partitiviridae	1,469	BK064937	Pennypacker partiti-like virus	64.11
35	Phoneutria nigriventer partiti-like virus b	Phoneutria nigriventer	dsRNA	Partitiviridae	1,308	BK064938	Hubei partiti-like virus 36	34.21
36	Phoneutria nigriventer toti-like virus b	Phoneutria nigriventer	dsRNA	Totiviridae	6,790	BK064932	Hubei toti-like virus 16	53.24

^aTwenty-two viruses shared less than 90 per cent amino acid identity with the most closely related viruses.

^bSixteen viruses shared less than 70 per cent amino acid identity with the most closely related viruses.

The virus with the highest abundance (RPM) was Pteromalus puparum dicistrovirus (Dicistroviridae), found in the venom gland of the arthropod Pteromalus puparum (RPM = 59,818.7), followed by Pelteobagrus fulvidraco oncotshavirus-1 strain VG (Tobaniviridae) from the chordate Tachysurus fulvidraco (RPM = 16,523.3), and Tetragnatha versicolor phenuivirus (Phenuiviridae) from the arthropod Tetragnatha versicolor (RPM = 11,612.0) (Table S3). The abundance of different viruses in the venom-related microenvironment also varied sharply across hosts. For example, the abundance of Crotalus cerastes sunvirus and Crotalus chuvirus was different within the same host or among different hosts (Table S3). Although this is likely due to limited data, it may also reflect the infection status and tissue tropism of the viruses.

The viruses identified here came from twenty-six venom-producing animals, involving the phyla Arthropoda (n = 12), Chordata (n = 11), and Mollusca (n = 3). Viruses in the venom-related microenvironment of Arthropoda species were the most common (n = 24), including eleven +ssRNA, nine −ssRNA viruses, and four dsRNA viruses (Fig. 1). These viruses were identified in fourteen Arthropoda RNA-seq datasets, with a detection rate of 19.18 per cent (14/73). Viruses in the venom-related microenvironment of Chordata (n = 18) largely comprised −ssRNA viruses (n = 16), with only two +ssRNA viruses. Viruses were present in sixteen RNA-seq datasets from Chordata species, with a detection rate of 4.53 per cent (16/353). Three −ssRNA viruses were identified in the venom-related microenvironment of Mollusca, with a detection rate of 6.82 per cent (3/44). As the detection rates documented here are based on the available datasets with potential sampling biases, the true prevalence of these viruses in nature is unclear and requires additional investigation.

3.3. Genome organization of the venom-related RNA viruses

Most of the viruses found in the venom-related microenvironment were similar to their closest relatives in terms of genome structure (Figs. 2 and S3). There were, however, some clear structural variations. Seven Crotalus chuvirus sequences were recovered, with RdRp identities to their closest relative (Guangdong red–banded snake chuvirus-like virus) ranging from 71.86 per cent to 74.20 per cent. ORF prediction showed that these chuviruses had similar genome structures to known viruses, although three strains identified in Crotalus polystictus might possess an additional ORF (Fig. S4). The genome sequences of segmented viruses were difficult to recover from sequencing data alone, and the genomes identified were different from those from known viruses. For example, three genome segments were identified for Micrurus arenavirus (Arenaviridae), which were annotated as RdRp, glycoprotein precursor, and nucleoprotein (Fig. 2). However, the closest viruses had only two segments: one with RdRp and a zinc-binding matrix protein and another with a glycoprotein precursor and a nucleoprotein (Fig. S3). Similarly, Tetragnatha versicolor phenuivirus (Phenuiviridae) was identified with three segments (L, M, and S segments). Although the Tetragnatha versicolor phenuivirus S segment contained one predicted ORF (Fig. 2), while the closest viruses have two ORFs, as the S segment obtained here was only partial, it is possible that the additional ORF existed but was not sequenced (Fig. S3).

Figure 2.

Genome structures of the novel RNA viruses found in the venom-related microenvironment. Genome structures of representative novel RNA viruses (<90 per cent amino acid identity in RdRp with known viruses) are displayed with boxes and coloured blocks. The black boxes represent predicted ORFs, while the coloured blocks indicate predicted functional domains, including viral polymerase (RdRp), glycoprotein (Gly), nucleoprotein (NP), viral methyltransferase (Met), and coat protein (CP), as the legend in bottom right. The host groups of the viruses are also labelled with icons.

3.4. Phylogenetic relationships of the venom-related RNA viruses

Phylogenetic analysis of the forty-five RNA viruses identified from the venom-related microenvironment revealed that they fell within a wide range of taxonomic groups, including at least twenty-two viral families or unclassified groups (Table 1 and Figs 3–5). The −ssRNA viruses were classified into eleven viral families: Rhabdoviridae, Artoviridae, Filoviridae, Xinmoviridae, Sunviridae, Myriaviridae, Chuviridae, Bornaviridae, Orthomyxoviridae, Arenaviridae, and Phenuiviridae (Fig. 3). Thirteen +ssRNA viruses clustered into at least eight viral families and unclassified groups (Fig. 4), while four dsRNA viruses fell into three families: Spinareoviridae, Partitiviridae, and Totiviridae (Fig. 5).

Figure 3.

Phylogenetic trees of negative-sense single-stranded RNA viruses. Maximum likelihood phylogenetic trees were estimated using RdRp sequences. The virus names in red were found from venom gland in this study, with their hosts also labelled. The red circles represent viruses sharing <90 per cent amino acid identity with the most closely related viruses.

Figure 4.

Phylogenetic trees of positive-sense single-stranded RNA viruses. Maximum likelihood phylogenetic trees were estimated using RdRp sequences. The virus names in red were found from venom gland in this study, with their hosts also labelled. The red circles represent viruses sharing <90 per cent amino acid identity with the most closely related viruses.

Figure 5.

Phylogenetic trees of double-stranded RNA viruses. Maximum likelihood phylogenetic trees were estimated using RdRp sequences. The virus names in red were found from venom gland in this study, with their hosts also labelled. The red circles represent viruses sharing <90 per cent amino acid identity with the most closely related viruses.

3.4.1. Negative-sense single-stranded viruses

Nine −ssRNA viruses were discovered in arthropods, sixteen were discovered in Chordata, and three were found in Mollusca. Phylogenetic analyses revealed that the viruses from the families Artoviridae, Filoviridae, Sunviridae, Chuviridae, and Bornaviridae fell within known genera. The others, however, may represent new taxa (Fig. 3). Two rhabdoviruses (Crassispira cerithina rhabdovirus and Conus episcopatus rhabdovirus) were identified from the venom duct of Mollusca (Gastropoda), which fell basal to lyssaviruses in the Rhabdoviridae phylogeny (Fig. 3A). There are few reports of rhabdoviruses in molluscs, with only two rhabdoviruses recently discovered in freshwater mussels (Mollusca: Bivalvia) (Goldberg et al. 2023). Two strains of Microplitis mediator mononega-like virus were recovered from the venom gland of Microplitis mediator, belonging to the family Xinmoviridae, but were divergent from any established genera (Fig. 3D). Five orthomyxoviruses were found in the venom-related microenvironment, although some were divergent to known genera within the family Orthomyxoviridae (Fig. 3H).

3.4.2. Positive-sense single-stranded viruses

Eleven +ssRNA viruses were discovered in venom glands of arthropod animals, with the other two found in Chordata. Most +ssRNA viruses fell within known genera, although some required further classification (Fig. 4). For instance, Apis picorna–like virus 4 strain SKS was also found in the venom gland of Apis cerana collected from South Korea, exhibiting 96.04 per cent amino acid identity to Apis picorna–like virus 4 in RdRp. Phylogenetic analysis showed that Apis picorna–like virus 4 strain SKS fell into the family Solinviviridae, although none genus has yet been assigned (Fig. 4B). In addition, three viruses had no known assigned taxonomy: Phoneutria nigriventer picorna–like virus (Fig. 4C), Solenopsis invicta virus 17 strain Ptepi (Fig. 4F), and Tetragnatha versicolor virus (Fig. 4G).

3.4.3. Double-stranded viruses

Four dsRNA viruses were identified from the venom-related microenvironment, belonging to the families Spinareoviridae, Partitiviridae, and Totiviridae (Fig. 5). Hemiscolopendra marginata reovirus shared 27.66 per cent amino acid identity with Shelly headland virus in the RdRp. Phylogenetic analysis showed that the virus clustered with the viruses of the genera Mycoreovirus and Coltivirus (Fig. 5A). Microplitis mediator partiti-like virus and Phoneutria nigriventer partiti-like virus shared 64.11 per cent and 34.21 per cent amino acid identity with known viruses (RdRp). Phylogenetic analysis showed that the two viruses were grouped with partiti-like viruses found in arthropods, but were divergent to known genera (Fig. 5B). Phoneutria nigriventer toti-like virus exhibited 53.24 per cent amino acid identity in the RdRp region with known viruses, and phylogenetic analysis revealed that the Phoneutria nigriventer toti-like virus clustered with Hubei toti-like virus 16, both of which may belong to an additional genus in the family Totiviridae (Fig. 5C).

A number of observations can be made for the viruses newly identified here as a whole. First, Tetragnatha versicolor virus, together with Hubei tetragnatha maxillosa virus 7, might represent a previously unclassified virus family (Fig. 4G). Second, it is notable that many newly reported viruses were divergent to classified genera, such as Crassispira cerithina rhabdovirus and Conus episcopatus rhabdovirus (Fig. 3A), Tapajós virus (Horie 2021) (Fig. 3C), Microplitis mediator mononega-like virus (Fig. 3D), Conus consors orthomyxo-like virus (Fig. 3H), Tetragnatha versicolor phenuivirus (Fig. 3I), Solenopsis invicta virus 17 strain Ptepi (Fig. 4F), Hemiscolopendra marginata reovirus (Fig. 5A), Phoneutria nigriventer partiti-like virus and Microplitis mediator partiti-like virus (Fig. 5B), and Phoneutria nigriventer toti-like virus (Fig. 5C), indicative of substantial untapped genetic diversity. Finally, prior to this study, the families Sunviridae and Myriaviridae only had one member each. Herein, we report new members from the two families, thereby enriching their genetic diversity.

3.5. Are the newly identified RNA viruses venom-specific?

To assess whether the newly identified RNA viruses were venom-specific, we retrieved the RNA-seq data of other tissues available under the same SRA BioProjects. This comprised forty-three SRA files from different tissues of eight host species (Table S2). The RNA viruses identified from the venom-related microenvironment were then screened in the other tissues (Fig. 6). Eleven viruses identified from the venom-related microenvironment (Conus consors orthomyxo-like virus, Conus episcopatus rhabdovirus, Pteromalus puparum dicistrovirus, Tetramorium bicarinatum iflavirus, Solenopsis invicta virus 17, Nasonia vitripennis virus strain VG, Israeli acute paralysis virus, Sacbrood virus, Apis picorna–like virus 4, Microplitis mediator mononega-like virus, and Microplitis mediator partiti-like virus) were also detected in other tissues, with abundance ranging from 1.1 to 279,236.7, indicating that these viruses might not be venom-specific (Fig. 6). However, some viruses had clearly higher abundance in the venom-related microenvironment. For example, Conus episcopatus rhabdovirus was enriched in the venom duct (RPM = 19.0), compared with those in radular sac (RPM = 1.2) and salivary gland (RPM = 2.4) (Fig. 6A). Similarly, Tetramorium bicarinatum iflavirus was enriched in the venom gland (RPM = 213.2), while the whole body without the abdomen had a low abundance (RPM = 1.1) of this virus (Fig. 6A).

Figure 6.

Comparison of the abundance of the RNA viruses from the venom-related microenvironment and other tissues. The SRA files of other tissues were retrieved from the same Bioproject as the venom data. The abundance of the RNA viruses from different tissues was calculated as RPM based on the non-rRNA reads. For visualization and comparison, the RPM values were normalized with log. (A) Only one RNA virus was identified from each host. (B) Multiple RNA viruses were identified from one host. * Female body carcasses without venom apparatus, ** Whole body without abdomen.

In addition, multiple viruses were identified in individual host species, although with variable abundance among tissues (Fig. 6B). Three viruses were identified from the venom glands of Apis cerana—Israeli acute paralysis virus (Dicistroviridae), Sacbrood virus (Iflaviridae), and Apis picorna–like virus 4 (Solinviviridae). Sacbrood virus was also detected in high abundance in larvae, while Israeli acute paralysis virus and Apis picorna–like virus 4 were absent. In addition, Apis picorna–like virus 4 was not detected in the brain and hypopharyngeal gland. Two viruses were found in the venom glands of Microplitis mediator, including Microplitis mediator mononega-like virus (Xinmoviridae) and Microplitis mediator partiti-like virus (Partitiviridae). Besides the venom gland, these two viruses were also found in the carcass, ovary, and whole body, although they were at the lowest abundance in the venom gland. It should be noted, however, that this heterogeneity might also reflect the pooling strategies employed in the sequencing projects, with each pool including multiple samples of the same host.

4. Discussion

There have been few studies on microorganisms in the venom-related microenvironment (Ul-Hasan et al. 2019), with little attention paid to virome diversity. We explored the RNA virome in the venom-related microenvironment using 474 public RNA sequencing datasets. From this, we identified forty-five viruses representing thirty-six viral species belonging to at least twenty-two viral families or groups. These viruses were identified from twenty-six venom-producing animals, representing the phyla Arthropoda, Chordata, and Mollusca, which were collected from at least seven countries. The diverse host and wide geographic range indicate that RNA viruses are commonly present in the venom-related microenvironment. Considering the limited sample size analysed here, we speculate that there are many more viruses to be discovered in the venom-related microenvironment.

Some of the newly described viruses might represent novel viral species and even genera. In particular, twenty-two viruses exhibited <90 per cent RdRp amino acid identity to their most closely related viruses, representing potentially novel viruses. Among these, sixteen viruses had <70 per cent of known viruses in RdRp, indicative of even more divergent taxa. For example, Crassispira cerithina rhabdovirus and Conus episcopatus rhabdovirus shared low identity with known viruses and formed a separated branch, representing a potential new genus.

The number of viral species in the venom-related microenvironment varied sharply across different animals. Arthropods harboured most viruses, involving thirteen viral families or groups, followed by chordates. In addition, viruses described in the families Dicitroviridae, Iflaviridae, and Partitiviridae were only identified from arthropods, while those from Sunviridae, Chuviridae, and Bornaviridae were mostly found in chordates.

It is notable that the abundance of the same virus species in the venom-related microenvironment of different hosts varied sharply and that most of the viruses described here had not previously been described in the venom-related microenvironment. For example, Crotalus cerastes sunvirus was found in venom gland samples of two Crotalus cerastes individuals, but their abundance varied approximately five-fold, from 287.1 to 1431.4. Furthermore, viruses found in the venom-related microenvironment were also detected in other tissues, but showed different abundance levels. For example, Conus episcopatus rhabdovirus was found in high abundance in the venom duct, but it was in much less abundance in the radular sac and salivary gland. Clearly, further investigation is required to understand the potential heterogeneous tissue tropism of these viruses.

In conclusion, we explored the diversity of RNA viruses in the venom-related microenvironment, identifying twenty-two novel viruses and fourteen known viruses. As such, our study sheds light on the hidden diversity of RNA viruses in the venom-related microenvironment, although we may have only documented the tip of the iceberg of virus diversity in this specific microenvironment.

Data availability

The viral sequences obtained here have been submitted to GenBank with accession numbers BK064870–BK064938.

Supplementary data

Supplementary data is available at Virus Evolution Journal online.

Conflict of interest:

The authors declare no competing interests.