Virome Analysis of Normal and Growth Retardation Disease-Affected Macrobrachium rosenbergii

ABSTRACT

The giant freshwater prawn, Macrobrachium rosenbergii, is an important aquaculture species in China. Growth retardation disease (GRD) is a common contagious disease in M. rosenbergii, resulting in slow growth and precocious puberty in prawns, and has caused growing economic losses in the M. rosenbergii industry. To investigate the viral diversity of M. rosenbergii and identify potentially high-risk viruses linked to GRD, virome analysis of the GRD-affected and normal M. rosenbergii was carried out using next-generation sequencing (NGS). A total of 327 contigs (>500 bp) were related to viral sequences belonging to 23 families/orders and a group of unclassified viruses. The majority of the viral contigs in M. rosenbergii belonged to the order Picornavirales, with the Solinviviridae family being the most abundant in both the diseased and normal groups. Furthermore, 16 RNA viral sequences with nearly complete genomes were characterized and phylogenetically analyzed, belonging to the families Solinviviridae, Flaviviridae, Polycipiviridae, Marnaviridae, and Dicistroviridae as well as three new clades of the order Picornavirales. Notably, the cross-species transmission of a picorna-like virus was observed between M. rosenbergii and plants. The “core virome” seemed to be present in the diseased and normal prawns. Still, a clear difference in viral abundance was observed between the two groups. These results showed that the broad diversity of viruses is present in M. rosenbergii and that the association between viruses and disease of M. rosenbergii needs to be further investigated.

IMPORTANCE Growth retardation disease (GRD) has seriously affected the development and economic growth of the M. rosenbergii aquaculture industry. Our virome analysis showed that diverse viral sequences were present in M. rosenbergii, significantly expanding our knowledge of viral diversity in M. rosenbergii. Some differences in viral composition were noted between the diseased and normal prawns, indicating that some viruses become more abundant in occurrences or outbreaks of diseases. In the future, more research will be needed to determine which viruses pose a risk for M. rosenbergii. Our study provides important baseline information contributing to disease surveillance and risk assessment in M. rosenbergii aquaculture.

INTRODUCTION

The virome has proven helpful for understanding viral diversity and describing novel viruses in new diseases (1). A viromic method based on metatranscriptomic sequencing not only could detect DNA and RNA viruses with excellent sensitivity but also could show high tolerance to stochastic errors to the relative abundance of viromic taxa (2, 3). This approach has played an important role in discovering novel viruses in various tissue and fecal samples from humans and animals (4–6). Furthermore, more surveys have been conducted to explore the viromes of economically important aquaculture species, including fish (7, 8), crustaceans (9–11), and mollusks (12). Although these studies have greatly expanded the knowledge of viral diversity of the aquatic species, relatively few studies have been conducted on the viral composition of aquatic species in different disease states (13–15). Comprehensive virome analysis of valuable aquaculture species will help to explore potential viral pathogens and further develop diagnostic tools to monitor these viruses, which will be of great help to avoid disease outbreaks resulting in substantial economic losses (9).

The giant freshwater prawn, Macrobrachium rosenbergii, is one of the most economically important aquaculture species and has been widely farmed in China and other Southeast Asian countries (16, 17). China is the largest producer of M. rosenbergii aquaculture, reaching 161,888 tonnes in 2020 (18). Despite the exponential growth of M. rosenbergii aquaculture in China, the incidence of diseases has increased in parallel (15, 19–22). Since 2010, many cultured giant freshwater prawns have been recorded with growth retardation disease (GRD) in China, characterized by growth retardation and precocious puberty (23, 24). The GRD-affected female prawns began to hold eggs when they were up to only 5 cm in body size, and the male prawns engaged in mating behaviors and grew two large blue claws when the body length was about 6 cm, while normal mature prawns are 8 to 10 cm long (25). More than 2,680 ha of farming ponds in Gaoyou City, Jiangsu Province, were affected to various degrees by GRD in 2014 (26). In recent years, GRD has continuously emerged in the main provinces of M. rosenbergii breeding, and the proportion of GRD-affected prawns in farming ponds reached 30 to 40% or even up to 80 to 90% in serious areas (27). Although mass mortalities have not occurred in GRD-affected ponds, GRD in M. rosenbergii has resulted in a production decline of more than 50%, causing substantial economic losses for M. rosenbergii aquaculture (28). Recently, a few studies have been performed to investigate the cause of GRD. The high prevalence of Enterobacter cloacae in slow-growing giant freshwater prawns observed by Gao et al. (29) suggested that the slow growth of M. rosenbergii is probably linked to E. cloacae. Dong et al. (28) detected eight major shrimp pathogens, but none were found in diseased prawns. Furthermore, they have also characterized a novel flavivirus associated with sexual precocity. In addition, some transcriptome analyses have been performed to explore the cause and mechanism of growth retardation of M. rosenbergii (25, 30). Other factors were reported to be associated with slow-growth M. rosenbergii, including antibiotics (31), long-term inbreeding, and deterioration of germ plasm resources (32). To date, the pathogens causing this disease are unknown. Given the powerful functions of the virome, there is a great necessity to explore the viral diversity of M. rosenbergii to screen the high-risk viruses linked to diseases.

In the present study, we investigated the viromes of GRD-affected and normal M. rosenbergii based on metatranscriptomic sequencing and revealed the broad diversity of viruses in M. rosenbergii. The nearly complete RNA viruses were further characterized and phylogenetically analyzed, expanding the virus diversity of the order Picornavirales and the family Flaviviridae in M. rosenbergii. Comparison of viromes between sick and healthy prawns could help to identify potential “high-risk” viruses. The sequences of these novel viruses could further serve as markers, which would be valuable for developing disease surveillance strategies and investigating their potential risks in disease outbreaks.

RESULTS

Overview of virome.

Fifteen diseased prawns (D) with the gross signs of GRD, including smaller size and precocious puberty, and another pool of 15 healthy prawns (H) without GRD were collected from farms in Jiashan County, Zhejiang Province, China, in 2019. To increase the viral abundance, these prawns were grouped into two pools: the diseased pool (pool D) and the healthy pool (pool H) (see Table S1 in the supplemental material). Overall, 65,370,202 and 33,836,190 trimmed reads were obtained from pools D and H and de novo assembled into 79,010 and 53,427 contigs, respectively (see Table S1 in the supplemental material). A total of 12,297 contigs (>500 bp) from two samples were clustered at 95% nucleotide identity over 80% of the length, resulting in 10,798 representative contigs. A total of 5,939 contigs (55%) were assigned to four categories (Eukaryota, Bacteria, Archaea, and Viruses) against the NCBI nonredundant protein (nr) database using DIAMOND. A total of 4,859 contigs got no significant hits, which were called “unknown” (Fig. 1A). The relative abundance of viruses in the diseased group was much higher than that in the normal group, which possessed 81.15% and 16.79% reads in pool D and pool H, respectively (Fig. 1A and Table S1). In contrast, the Eukaryota and Bacteria proportions (13.77% and 0.39%, respectively) were lower than those in pool H (77.81% and 0.90%, respectively) (Fig. 1A and Table S1).

FIG 1.

Virome comparison between GRD-affected (pool D) and normal M. rosenbergii (pool H). (A) Read proportions of four taxonomic categories (Eukaryota, Bacteria, Archaea, and Viruses) in GRD-affected and normal M. rosenbergii. “unknown” represents contigs that got no significant hits in the NCBI nr database. (B) A bar plot displays the relative abundance of viral orders and families in two M. rosenbergii groups based on a log₁₀ RPM scale.

A total of 327 contigs were invertebrate-associated viruses, 51 (15.6%) viral contigs were shared among D and H, 58 unique contigs were found from pool D, and 218 viral contigs were present only in pool H (Fig. S1). Rarefaction analysis of the viral contigs revealed the identified viral contigs for both GRD-affected and normal prawns were approaching the saturation plateau, suggesting the sampling depths of sequencing data were sufficient, and most viruses were detected in this experiment (Fig. S2). A total of 327 viral contigs were assigned into 23 families/orders under double-stranded RNA (dsRNA) viruses, positive-sense single-stranded RNA (+ssRNA) viruses, negative-sense single-stranded RNA (−ssRNA) viruses, double-stranded DNA (dsDNA) viruses, single-stranded DNA (ssDNA) viruses, and a group of unclassified viruses (Fig. 1B). Sixteen families/orders and some unclassified viruses were shared among pools D and H. A viral sequence of the family Flaviviridae was present in high abundance (91,141 mapped reads per million total reads [RPM]) only in pool D. In contrast, viral sequences from the order Bunuyavirales, including the families Circoviridae, Phenuiviridae, Polycipiviridae, New-Yanvirus, and Picornaviridae, were identified only from pool H (Fig. 1B and Table S3). The order Picornavirales (46.8% of the viral contigs) was the major viral group in M. rosenbergii (Table S2). The family Solinviviridae was most abundant in both the diseased and normal groups, and most of the other families were present at very low abundance (<1,000 RPM) (Fig. 1B and Table S3). Some differences in the abundance of viral families were noted between the diseased and normal groups. Specifically, pool D possessed a higher percentage of viral reads from the families Solinviviridae, Dicistroviridae, and Tombusviridae (883,839, 9,581, and 3,883 RPM, respectively), while pool H possessed such reads at lower abundances (643,510, 532, and 565 RPM, respectively) (Fig. 1B and Table S3 in the supplemental material). Reads of the families Hepeviridae and Marnaviridae and viral sequences belonging to the unclassified Picornavirales and Nodaviridae were abundant in pool H (127,492, 117,820, 43,694, and 21,156 RPM, respectively) (Fig. 1B and Table S3). In comparison, pool D had very low abundance (189, 7,224, 2,890, and 524 RPM, respectively) (Fig. 1B and Table S3).

Here, 16 RNA viral sequences with nearly complete genomes were identified from the GRD-affected prawns and normal ones (Table 1). These viral sequences belong to the order Picornavirales (n = 15) and family Flaviviridae (n = 1). As detailed below, the genome structures of these virus sequences were characterized, and phylogenetic analysis was performed using the amino acid sequences of RNA-dependent RNA polymerase (RdRp) protein.

TABLE 1.

The nearly complete viral sequences identified in this study

Virus name	Abbreviation	Family or clade	Length (bp)	Closest blastx hit	blastx identity (%)	Viral abundance (RPM) ina:		Novel viral sequences
						Pool D	Pool H
Macrobrachium rosenbergii
Virus 1	MRV1	Solinviviridae	14462	Solenopsis invicta virus 7	21.6	883,839	643,510	Yes
Virus 2	MRV2	Polycipiviridae	11442	Linepithema humile polycipivirus 1	38.1	0	1,025	Yes
Virus 3	MRV3	Marnaviridae	9242	Bivalve RNA virus G3	29.7	285	0	Yes
Virus 4	MRV4	Marnaviridae	8887	Beihai picorna-like virus 35	98.3	101	1,434	No
Virus 5	MRV5	Marnaviridae	8548	Beihai picorna-like virus 6	64.9	57	946	Yes
Virus 6	MRV6	Marnaviridae	9153	Wenzhou picorna-like virus 50	81.4	22	360	Yes
Virus 7	MRV7	Dicistroviridae	11342	Beihai picorna-like virus 87	29.8	32	483	Yes
Virus 8	MRV8	Dicistroviridae	8587	Dicistroviridae sp. squirrel/UK/2011	28.1	22	6	Yes
Virus 9	MRV9	Dicistroviridae	8891	Dicistroviridae sp.	71.5	9,442	17	Yes
Virus 10	MRV10	New clade 1	8867	Changping earthworm virus 1	31.6	137	0	Yes
Virus 11	MRV11	New clade 1	8093	Corey virus	23.1	121	0	Yes
Virus 12	MRV12	New clade 1	9818	Changping earthworm virus 1	28.7	85	6	Yes
Virus 13	MRV13	New clade 1	9883	Changping earthworm virus 1	31.7	33	13	Yes
Virus 14	MRV14	New clade 2	8205	Wenzhou picorna-like virus 39	25.8	77	0	Yes
Virus 15	MRV15	New clade 3	9011	Trichosanthes kirilowii picorna-like virus	98.9	6,722	114,953	No
Infectious precocity virus strain ZJJS2019	IPV strain ZJJS2019	Flaviviridae	12594	Infectious precocity virus	99.62	91,142	0	No

^aRPM, mapped reads per million total reads.

Viruses in the order Picornavirales.

Fifteen picorna-like virus sequences with nearly complete genomes were identified from two pools, and 10 viruses shared a low (21.6% to 38.1%) amino acid identity with the most closely related viral sequences (Table 1). The helicase, protease, and RdRp replication module (Hel-Pro-RdRp) and at least two capsid protein domains were detected in most sequences (Fig. 2). Fifteen viral sequences were assigned to four current families (Solinviviridae, Polycipiviridae, Marnaviridae, and Dicistroviridae) and three new clades in the order Picornavirales (Fig. 3). Among these, 12 novel RNA viral sequences and three known viral sequences were discovered here.

FIG 2.

Genome organization of newly identified viral sequences in the order Picornavirales.

FIG 3.

Midpoint-rooted phylogenetic tree of 15 picorna-like viruses, related viral sequences, and representative and unassigned members of the order Picornavirales. The tree was based on the amino acid sequence of RNA-directed RNA polymerase (RdRp) and inferred using the maximum likelihood approach implemented in IQ-TREE version 1.6.12. Branch support was evaluated with the Shimodaira-Hasegawa approximate-likelihood ratio test (SH-aLRT) and ultrafast bootstrap (UFBoot). The branches were indicated at the nodes when the SH-aLRT value was ≥80% or the UFBoot value was ≥95%. The background colors indicate different families of viruses. Virus names are marked in different colors based on their host taxonomy. The star indicates the current member species, the triangle indicates related unclassified viruses of families, and the check mark indicates novel picorna-like viruses. The scale bar represents the number of amino acid substitutions per site.

(i) Viruses in the family Solinvividae. Solinviviridae, established in 2017, is a relatively new family of picorna/calici-like viruses, with only two classified solinvivirus species in two genera infecting ants. Still, previous studies have obtained related unclassified virus sequences from various other insects and arthropods (33). One solinvi-like viral sequence, tentatively named Macrobrachium rosenbergii virus 1 (MRV1), was detected in two groups. MRV1 was the most abundant virus in pools D and H, reaching 888,839 and 643,510 RPM, respectively (Table 1). The genome size of MRV1 was 14,462 bp, unlike the typical virus genome length (10 to 11 kb), sharing just 21.6% amino acid identity with Solenopsis invicta virus 7 (SINV-7) across the polyprotein (Fig. 2; Table 1). Phylogenetic analysis using conserved RdRp domains revealed that MRV1 was clustered with other related unclassified viruses as described by Shi et al. (5) (Fig. 4), and a long branch length involved that MRV1 may reveal a novel virus in the family Solinvivirdae.

FIG 4.

Midpoint-rooted phylogenetic tree of RdRp protein of MRV1, related viral sequences, and representative and unassigned members of Solinviviridae. Branch support was evaluated with SH-aLRT and UFBoot. The values are indicated at the nodes when the SH-aLRT value was ≥80% or the UFBoot value was ≥95%. The names of the viruses are marked in different colors based on their host taxonomy. The star indicates the current member species, the triangle indicates related unclassified viruses of families, and the check mark indicates novel picorna-like viruses. The scale bar represents the number of amino acid substitutions per site.

(ii) Viruses in the family Polycipiviridae. A novel polycipivirus sequence, namely, Macrobrachium rosenbergii virus 2 (MRV2), was discovered only from pool H. The nearly complete genome of MRV2 was 11,442 bp in length and contained a typical polycipivirus genome organization with four overlapping 5′-proximal ORFs (ORFs 1 to 4) and one long 3′ ORF (ORF5) in the sense direction (Fig. 2). In addition, it had an additional ORF (ORF2b) overlapping ORF2, which is also found in all members of the genus Sopolycivirus, encoding a small protein containing a predicted transmembrane domain (34). Moreover, MRV2 also had a new small ORF (ORF4b) without any significant homology (Fig. 2). This was consistent with a previous report for Linepithema humile polycipivirus 1 (LhuPcV1) found in Linepithema humile (35). MRV2 was closely related to LhuPcV1, sharing 38.1% amino acid identity across ORF5 (RdRp) (Table 1). Phylogenetic analysis of the ORF5 nonstructural polyprotein placed MRV2 within the Sopolycivirus genus in a well-supported clade (100% bootstrap support) (Fig. 5). According to the International Committee on Taxonomy of Viruses (ICTV) guidelines, species demarcation within the Sopolycivirus genus requires >10% amino acid sequence divergence from other species across the ORF5 protein (34). These results support that MRV2 should be considered a novel virus of the genus Sopolycivirus in the family Polycipiviridae. It is interesting to note that nearly all members of the genus Sopolycivirus appear to be detected in ant species (family Formicidae), except for Shuangao insect virus 8 (ShiV8), which is an insect-associated virus from an insect mixed sample in the previous study (5). MRV2 is the first to identify a polycipivirus in the genus Sopolycivirus from the crustacean species M. rosenbergii, expanding the diversity and host range in the family Polycipiviriade.

FIG 5.

Midpoint-rooted phylogenetic tree of full-length ORF5 amino acid sequences of MRV6, related viral sequences, and representative and unassigned members of Polycipiviridae. Branch support was evaluated with SH-aLRT and UFBoot. The values are indicated at the nodes when the SH-aLRT value was ≥80% or the UFBoot value was ≥95%. The background colors show the different genera of Polycipiviridae. The names of the viruses are marked in different colors based on their host taxonomy. The star indicates the current member species, the triangle indicates related unclassified viruses of families, and the check mark indicates novel picorna-like viruses. The scale bar represents the number of amino acid substitutions per site.

(iii) Viruses in the family Marnaviridae. Four sequences of marna-like viruses were identified from two pools that fell within the recently established family Marnaviridae, which we named Macrobrachium rosenbergii virus 3 (MRV3), Macrobrachium rosenbergii virus 4 (MRV4), Macrobrachium rosenbergii virus 5 (MRV5), and Macrobrachium rosenbergii virus 6 (MRV6). The lengths of the four viral sequences were 9,242, 8,887, 8,548, and 9,153 bp, respectively. MRV4, MRV5, and MRV6 contained two nonoverlapping ORFs, while MRV3 contained a single ORF encoding the polyprotein (Fig. 2). Regardless of the fact that the genome structures of these viral sequences were different, the nonstructural proteins and structural proteins were encoded in the 5′ region and 3′ region, respectively (Fig. 2). MRV3 was highly divergent, sharing only 29.7% amino acid identity with bivalve RNA virus G3 found in bivalves (12). MRV4 shared 98.3% amino acid identity with Beihai picorna-like virus 35 strain BHBJDX17224 identified from penaeid shrimp (5) (Table 1). MRV5 and MRV6 shared 64.9% and 81.4% amino acid identities across the polyprotein to Beihai picorna-like virus 6 and Wenzhou picorna-like virus 50, respectively (5) (Table 1). MRV5 shared 33.27% and 31% amino acid similarities to MRV6 across the nonstructural and structural proteins, respectively. According to a recent proposal to the ICTV, species demarcation with the family Marnaviridae requires >10% and >25% amino acid divergence from other species across the RdRp and capsid protein, respectively (36). Phylogenetic analysis placed four viruses into three genera (Labyrnavirus, Kusarnavirus, and Sogarnaviru) of the family Marnaviridae (Fig. 6). MRV3 was a novel virus belonging to the genus Labyrnavirus with 70.3% divergence, whereas MRV4 and Beihai picorna-like virus 35 were the same virus within the genus Kusarnavirus (Fig. 6). MRV5 and MRV6 fell within the genus Sogarnavirus (100% bootstrap support), which suggested that MRV5 and MRV6 were two novel viruses of the genus Sogarnavirus in the family Marnaviridae (Fig. 6).

FIG 6.

Midpoint-rooted phylogenetic tree of RdRp protein of the family Marnaviridae, including identified viruses in this study and related representative viruses. Branch support was evaluated with SH-aLRT and UFBoot. The values are indicated at the nodes when the SH-aLRT value was ≥80% or the UFBoot value was ≥95%. The background colors show the different genera of Marnaviridae. The names of the viruses are marked in different colors based on their host taxonomy. The star indicates the current member species, the triangle indicates related unclassified viruses of families, and the check mark indicates novel picorna-like viruses. The scale bar represents the number of amino acid substitutions per site.

(iv) Viruses in the family Dicistroviridae. Here, we obtained three dicistrovirus sequences from pool D and pool H. Macrobrachium rosenbergii virus 7 (MRV7) was 11,342 bp long, encoding only one large polyprotein, unlike the typical dicistroviruses genome organization containing two nonoverlapping ORFs (Fig. 2). MRV7 shared an amino acid identity of only 29.8% with Beihai picorna-like virus 87 (BPV87) (Table 1), which in turn was predicted to have only one ORF and was assigned to the “Dicistroviridae cluster” within the “Picorna-Calici” clade (5). The genome sizes of Macrobrachium rosenbergii virus 8 (MRV8) and Macrobrachium rosenbergii virus 9 (MRV9) were 8,587 and 8,891 bp, respectively, and the viruses displayed typical dicistrovirus genome structures, with two nonoverlapping ORFs encoding the replication enzyme polyprotein and the capsid protein (Fig. 2). They all showed limited sequence identity with unclassified dicistroviruses: 28.1% and 71.5%, respectively (Table 1). Current ICTV-recognized species in the family Dicistroviridae were separated by greater than 10% amino acid distance in the capsid protein, suggesting that these three viruses found in this study were all novel viruses (37). Phylogenetic analysis showed that these viruses were clustered into three separate clades: MRV7 was clustered with the crustacean-infecting members of genus Aparavirus, whereas MRV8 and MRV9 formed two new sister clades to the genera Triatovirus and Cripavirus within other unclassified dicistro-like viruses, respectively (Fig. 7), and might represent two new genera in the family Dicistroviridae.

FIG 7.

Midpoint-rooted phylogenetic tree of the nonstructural polyprotein, containing the RNA helicase, cysteine protease, and RdRp, including the viruses identified in this study and related representative viruses. Branch support was evaluated with SH-aLRT and UFBoot. The values are indicated at the nodes when the SH-aLRT value was ≥80% or the UFBoot value was ≥95%. The background colors show the different genera of Dicistroviridae. The names of the viruses are marked in different colors based on their host taxonomy. The star indicates the current member species, the triangle indicates related unclassified viruses of families, and the check mark indicates novel picorna-like viruses. The scale bar represents the number of amino acid substitutions per site.

(v) Viruses in the unclassified Picornavirales. Six highly divergent viral sequences were found in two groups, namely, Macrobrachium rosenbergii virus 10 (MRV10), Macrobrachium rosenbergii virus 11 (MRV11), Macrobrachium rosenbergii virus 12 (MRV12), Macrobrachium rosenbergii virus 13 (MRV13), Macrobrachium rosenbergii virus 14 (MRV14), and Macrobrachium rosenbergii virus 15 (MRV15), which fell outside any well-defined family in the order Picornavirales (Fig. 3). MRV10, MRV11, MRV12, and MRV13 had a higher abundance in pool D than H (Table 1). They had similar genome organizations with a single ORF containing the structural proteins at the 5′ end and the nonstructural proteins at the 3′ end (Fig. 2). For MRV10 and MRV11, the polyprotein contained six conversed domains, including RNA helicase (Hel), cysteine protease (Pro), RdRP, and three capsid proteins (Fig. 2). In the case of MRV12 and MRV13, the cysteine protease domain was not identified, and two capsid proteins were predicted in MRV12 (Fig. 2). MRV11 shared 23.1% amino acid identity with Corey virus, whereas MRV10, MRV12, and MRV13 were all closely related to Changping earthworm virus 1 (5), with 31.6%, 28.7%, and 31.7% amino acid identity, respectively (Table 1). These four viruses exhibited slow polyprotein amino acid identity (27.49% to 38.8%), suggesting that they were four different viruses. Phylogenetic analysis showed that these four viruses clustered together and formed a high divergent clade, sister to the family Polycipiviridae (Fig. 3 and Fig. S5), indicating that the four novel viral sequences might potentially represent a novel viral family in the order Picornavirales. MRV14 was present only in pool D and shared an amino acid identity of 25.8% with the Wenzhou picorna-like virus 39 (5) (Table 1). Phylogenetic analysis showed that MRV14 was located on a sister branch of the Dicistroviridae and Marnaviridae families (Fig. 3) and formed an outgroup with three relatives inside the order Picornavirales, likely to represent a novel family.

Instead, the abundance of MRV15 was higher in pool H than in D (Table 1). MRV15 exhibited 99% amino acid identity over 100% coverage to Trichosanthes kirilowii picorna-like virus (TKPV) and Forsythia suspensa picorna-like virus (FSPV) identified from the plants Trichosanthes kirilowii and Forsythia suspensa, respectively (38) (Table 1). MRV15, TKPV, and FSPV formed a monophyletic cluster, sister to the Nora viruses in the phylogeny of the order Picornavales (Fig. 3). This indicated that they were the same virus at the level of species. Notably, we confirmed the presence of MRV15 in our samples using reverse transcription-PCR (RT-PCR) followed by Sanger sequencing, and 73.9% (17/23) of GRD-affected prawns collected from Huzhou Province in 2020 were positive in the prevalence investigation of MRV15 (Fig. S3), likely indicating cross-species transmission among M. rosenbergii and plants.

A virus in the family Flaviviridae.

The Flaviviridae are a family of small enveloped viruses with positive-sense RNA genomes of ~9 to 13 kb that encode a large polyprotein precursor. For the maturation of flavivirus particles, the polyprotein precursor is cleaved into structural proteins (capsid protein [C], premembrane protein [prM], and envelope protein [E]) and also nonstructural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, 2K, NS4B, and NS5) (39). Here, one 12,594-bp viral contig had 98.8% nucleotide identity with the previously identified infectious precocity virus (IPV) (28), tentatively named Infectious precocity virus strain ZJJS2019 (IPV strain ZJJS2019). IPV strain ZJJS2019 was obtained only from GRD-affected prawns, with high relative abundance (up to 91,141 RPM) (Table 1). The genome of IPV strain ZJJS2019 presented the same site of programmed −1 ribosomal frameshifting (−1PRF) to produce a single polyprotein (3,707 amino acids) (28), sharing 99.62% amino acid identity across the polyprotein with IPV strain ZJJS2019 (Fig. 8A and Table 1). Based on the sequence alignment of homologous sequences by MUSCLE, potential polyprotein cleavage sites of the IPV strain ZJJS2019 were identified (Fig. 8A and Table S4). The order of proteins encoded in the IPV strain ZJJS2019 polyprotein is 5′-VirC-CTHD-pr-M-E-NS1-NS2A-NS2B-NS3-NS4A-2K-NS2B-NS5-3′, which is typical of flaviviruses (Fig. 8A) (39). Phylogenetic analysis of conserved amino acid sequences in the RdRp (NS5) demonstrated that IPV strain ZJJS2019 has the closest evolutionary relationship to a group of Jingmenvirus and Tamana bat virus (TABV) isolates belonging to the genus Flavivirus (Fig. 8B). TABV is currently listed as a potential member of the genus Flavivirus but sufficiently distinct to potentially merit assignment into a new genus (40). Therefore, we propose that IPV strain ZJJS2019 could be considered a new genus in the family Flaviviridae as described by Dong et al. (28).

FIG 8.

Genome structure and phylogenetic analysis of a novel flavivirus discovered. (A) Genome architecture of IPV strain ZJJS2019. P, H, and R indicate the localization of the NS3 protease, the NS3 RNA helicase, and the NS5 RdRP domains, respectively. (B) Phylogenetic tree of NS5 protein (~420 amino acids) of the family Flaviviridae. The triangle indicates a node SH-aLRT value of ≥80% or UFBoot value of ≥95%. The virus described in this study is marked with a solid red asterisk. The scale bar represents 0.5 amino acid substitution per site.

DISCUSSION

Our findings support previous observations that crustaceans’ and other invertebrates’ viromes have harbored a broad diversity of viruses of the order Picornavirales. Shi et al. (5) found that numerous and divergent picorna-like viruses were ubiquitous in invertebrate species. In addition, previous virome analyses of honeybees and other insect pollinators also observed that Picornavirales was the dominant group (41–43). Furthermore, 46 picorna-like virus sequences were recently identified in multiple marine invertebrate species from three different seas (the South China Sea, the East China Sea, and the Yellow Sea) in China (44). Although the pathogenicity of picorna-like viruses that were mostly discovered from invertebrate species without any obvious clinical symptoms is still unclear, several picornaviruses are known to be highly pathogenic to beneficial arthropods. For instance, a variety of important honeybee pathogenic viruses, such as deformed wing virus (DWV) (45, 46), acute bee paralysis virus (ABPV) (47), and black queen cell virus (BQCV) (48), caused widespread disappearance and death of honeybee populations (49). Also, three dicistroviruses were implicated as the causative pathogens of disease outbreaks among some crustaceans. Taura syndrome virus (TSV), infecting several penaeid shrimp species, has ranged from 40% to >90% in cultured populations of postlarval (PL), juvenile, and subadult P. vannamei (50, 51). Mud crab dicistrovirus (MCDV) is known to cause sudden and massive mortalities in crabs (52). Furthermore, Macrobrachium rosenbergii Taihu virus (MrTV) was confirmed as the causative agent of the larval mortality syndrome in M. rosenbergii (22). A shred of growing evidence has indicated that viruses of the order Picornavirales might pose high risks to many hosts. The potential risks of these picorna-like viruses in M. rosenbergii should be investigated further.

Virome comparison between diseased and healthy prawns could help to find potential pathogens for hosts. Fifty-one viral contigs were shared among two groups from different ponds, likely indicating the core virome of M. rosenbergii, as was recently observed for brown shrimp (9), rodents (6), and mosquitoes (53–55). However, the differences in the viral relative abundance were noted between diseased and normal prawns, indicating that some viruses become more abundant in occurrences or outbreaks of diseases, and they might be potential high-risk viruses for M. rosenbergii. MRV1, a novel Solinvi-like virus, was the most abundant virus in both diseased and normal prawns. Some members of Solinviviridae are pathogenic viruses for ants and some arthropods, like Solenopsis invicta virus 3 (SINV-3)-infected Solenopsis invicta queens, resulting in significant reductions in fecundity and body weight and slower growth of ant colonies (56). Acyrthosiphon pisum virus (APV) has been reported to reduce the growth of pea aphids and increase the time to reach maturity (57). So, this indicated that MRV1 might pose a potential risk to M. rosenbergii due to the high concentrations. It would be necessary to further investigate the pathogenicity of MRV1 in M. rosenbergii.

The second potentially dangerous virus could be the IPV strain ZJJS2019, belonging to the family Flaviviridae, which was the next most abundant virus and was present only in the diseased pool. Many flaviviruses are known to be important human and veterinary pathogens, including yellow fever virus, dengue virus, hepatitis C virus, Japanese encephalitis virus, and classical swine fever virus, causing fatal hemorrhagic fever, chronic liver or neurological disease in humans, and economically important diseases in domestic or wild animals (39). IPV strain ZJJS2019 is the first flavivirus associated with M. rosenbergii, expanding the host range infecting the invertebrate species, especially the crustaceans, over what is currently reported (58). A previous study indicated that IPV strain ZJJS2019 was associated with sexual precocity in M. rosenbergii (28). Furthermore, Parry and Asgari demonstrated that the horizontal transmission of dual host invertebrate-vertebrate flaviviruses was present between crabs and sharks and offered the potential insights to explore the origin or emergence of terrestrial vector-borne flaviviruses (58). Thus, future studies need to investigate the pathogenicity and host range of IPV strain ZJJS2019 from invertebrates to vertebrates to reveal the epidemiology of IPV strain ZJJS2019 and elucidate potential cross-species transmission threats.

Furthermore, some novel viral sequences related to already known crustacean viruses in the phylogenetic trees might pose potential risks to M. rosenbergii aquaculture. For instance, MRV9 was assigned to the same family, Dicistroviridae, as TSV, MCDV, and MrTV, which were already known to cause massive mortalities in crustacean species. MRV9 was the third most abundant virus in pool D and at much higher abundance than in pool H, suggesting the potential threat for M. rosenbergii. Although Dong et al. (28) showed that the healthy M. rosenbergii prawns infected by a viral preparation extracted from “iron prawn syndrome” (IPS)-affected prawns could show clinical signs of growth cessation and sexual precocity, and all infected prawns were positive for IPV, molecular diagnostics were not performed to detect other potentially high-risk viruses in the infected prawns. Except for IPV, MRV1 and MRV9 also might be associated with GRD and were more abundant in diseased prawns than normal ones, and relatedness to the known pathogenic viruses. The potential roles of the high-risk viruses in different disease states need to be confirmed by more molecular techniques and further monitoring of their prevalence in emerging infectious diseases in M. rosenbergii.

Cross-species transmission (i.e., horizontal virus transfer) in animal and plant populations may lead to emerging infections in the recipient host (38, 59). Here, the nearly complete genome of MRV15 shared >99% amino acid identity with two viruses found in the plants Trichosanthes kirilowii and Forsythia suspensa (38). The two plants are important medicinal plants, widely distributed in China (60). Forsythia suspensa plant extracts have proved to be potential substitutes for antibiotics due to their anti-inflammatory antioxidant and intestinal barrier function in aquatic animals, which are as feed additives used in aquaculture (61, 62). We detected MRV15 in M. rosenbergii with high prevalence, which was consistent with a high viral abundance in metatranscriptome sequencing. In addition, in the Picornavales phylogenetic tree, MRV15 formed a sister clade with unclassified picorna-like viruses infecting arthropods in a previous study (5). These results indicated that M. rosenbergii was the host of MRV15, and cross-species transmission might have occurred between crustaceans and plants. Previous studies also reported examples of viruses with host ranges spanning both the plant and animal kingdoms. Li et al. (63) proved that a plant-pathogenic virus, tobacco ringspot virus (TRSV) of the Secoviridae family, could replicate and produce virions in honeybees, (Apis mellifera), resulting in systemic infections throughout the body. Furthermore, some viruses known to infect arthropods can also infect plants. For instance, Flock House virus (FHV), an insect virus belonging to the Nodaviridae family, has been shown to replicate in plants and mammalian and yeast cells (64). Emerging infectious diseases resulting from cross-species transmission often threaten host health. These viruses should be of concern due to their invasiveness and ability to spread among different species (65). Thus, the pathogenicity and transmissibility of MRV15 in M. rosenbergii deserve further investigation.

Different environments might affect the viral composition of hosts. Zhang et al. found that multiple marine invertebrate species from three seas in China seemingly contained sea-specific virus groups (44). In addition, some differences in the viral composition and abundance in sediments were noted with the culture species added to different mariculture systems (66). A large-scale virome project involving more extensive ranges and more organisms would more significantly capture the viral diversity of hosts and provide more detailed information about the virome heterogeneity and connectivity of different environments. The GRD-affected prawns contained a higher viral abundance than the normal groups, while the normal group exhibited higher viral diversity. This might be explained by the replications of some high-risk viruses being enhanced in the diseased state of M. rosenbergii, resulting in a decrease in viral diversity. However, our study remains preliminary due to the limitation of sample size, and the topic still needs to be further explored with a greater diversity of viromes in M. rosenbergii within a larger geographic range and a longer time.

In conclusion, this study has significantly expanded our understanding of the viral diversity of M. rosenbergii, especially numerous viruses from the order Picornavirales. The characterization of viruses in M. rosenbergii will facilitate our knowledge of virus evolution and cross-species transmission and reduce the potential threat of cross-species infections. Comparison of the viral composition and abundance in the GRD-affected prawns with those in normal prawns will help to identify potentially high-risk viruses and provide valuable information for future disease surveillance and prevention of M. rosenbergii.

MATERIALS AND METHODS

Sample collection and preparation.

In August 2019, giant freshwater prawns (Macrobrachium rosenbergii) stocked for 70 to 80 days were collected from farming ponds in Jiashan County, Zhejiang Province, China. The diseased and normal prawns collected from different ponds in the study originated from the same hatchery with the same genetic background. Fifteen GRD-affected giant freshwater prawns, with body lengths ranging from 6.0 to 6.8 cm, were collected from diseased farm ponds with typical GRD. The clinical signs of prawns with GRD included smaller body size, hard shell, and precocious puberty. Fifteen healthy ones were acquired from the farm ponds without GRD, with similar seedling release times, appropriate feeding, and management methods. The body lengths of healthy ones were 8.2 to 9.0 cm. The gill, hepatopancreases, and muscle tissues of all prawns were immediately dissected and stored separately in RNAlater before transferring them to a freezer set at −80°C.

Virus enrichment, RNA library construction, and sequencing.

To increase the likelihood of virus discovery, the gill, hepatopancreases, and muscle tissues from 15 prawns with the same health condition were pooled in equal weight (~50 mg), resulting in two pools—the diseased (D) pool and the healthy (H) pool, respectively. The sample pools were homogenized with phosphate-buffered saline (PBS) (10% mass/vol) in a Lu kα sample freezing grinder (LUKYM-I) for 3 min at 60 Hz and 4°C, frozen, and thawed three times on dry ice. The suspension of each pool was subjected to centrifugation (10 min at 12,000 × g). The supernatant was filtered through a 0.45-μm-pore filter (Millipore) to remove eukaryotic and bacterial cell-sized particles. All filtrates were concentrated using ultracentrifugation to enrich the viral particles. RNA was extracted using the QiaAmp mini-viral RNA kit (Qiagen, Germany) and quantified using Qubit (Thermo Fisher, USA). To eliminate possible reagent contaminants, we included sterile water as a control (67). The rRNA was removed for library construction by the NEBNext rRNA depletion kit (NEB, USA). Sequencing libraries were generated using RNA library prep kit for Illumina (NEB, USA) following the manufacturer’s recommendations. Then, the 150-bp paired-end sequencing of the RNA libraries was performed using the Illumina NovaSeq platform (Illumina, USA) at Novogene (Beijing, China).

Sequence assembly and virus discovery.

For each library, raw reads were filtered and quality-trimmed with fastp (version 0.21.0) (68), and clean reads were de novo assembled by using Trinity (version 2.12.0) (69) with default settings. Contigs longer than 500 bp from two pools were clustered for redundancy at 95% nucleotide identity over 80% of the length using CD-HIT (version 4.8.1) (54, 70). Representative contigs were annotated against the NCBI nonredundant protein (nr) database using DIAMOND (blastx, version 2.0.11) (71), setting an E value of 1E−5 to remove false-positive results. We filtered the blastx results with the keyword “Viruses.” To remove false positives (e.g., endogenous viral elements and nonviral host sequences), all putative viral contigs were compared against the host reference genome of M. rosenbergii (an in-house M. rosenbergii genome) and nonredundant nucleotide (nt) database with an E value threshold of 1E−5 using the blastn algorithm. The taxonomic lineage information of viral contigs was collected based on the blastx results; viral contigs that are related to virus families well known to infect plants were excluded (72).

To confirm the relative abundance of viral contigs, clean reads were mapped back to the nonredundant contigs using Bowtie2 (version 2.4.2) (73) and inspected using Integrative Genomics Viewer (IGV [version 2.5.0]) (74). To identify and eliminate possible contamination, we used the following steps to exclude false positives of these viral contigs suggested by previous study (75). The relative abundance of each viral contig was quantified as the number of mapped reads per million total reads (RPM), and we set RPM at ≥1 and the number of identified reads at ≥10 for a positive virus hit. To reduce the cross-contamination due to index hopping, the reads of viral contigs in the low-abundance library were removed if the abundance of a viral contig in one library was <0.1% of the higher abundance library. Rarefaction analyses were conducted to visualize the difference in the viral contigs composition (76).

PCR confirmation and prevalence investigation.

To confirm the presence of a subset of the newly identified viruses (see Table S2 in the supplemental material), nested reverse transcription-PCR (RT-PCR) was performed (Fig. S3) using specific primer pairs designed according to assembled viral contigs (Table S4). The target PCR products were visualized on 1.5% agarose gel and were validated by Sanger sequencing (Fig. S4). In order to investigate the prevalence of MRV15, 23 prawns with typical GRD status collected from Huzhou, Zhejiang Province, in 2020 were tested for MRV15 by nested RT-PCR assay, and RNase-free water was used as a negative control. PCR products were confirmed using Sanger sequencing.

Virus genome annotation.

The potential open reading frames of the obtained viral contigs were predicted with the NCBI Open Reading Frame Finder (https://www.ncbi.nlm.nih.gov/orffinder/), based on two criteria: (i) the minimal ORF length was longer than 100 amino acids, and (ii) a nested ORF was ignored unless it had a homolog in the closely related virus. To characterize the functional domains, predicted viral ORFs were analyzed using the NCBI Conserved Domain Search (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi/) (77) together with InterPro (http://www.ebi.ac.uk/interpro/) (78) with default settings. Prediction of putative polyprotein cleavage sites of the flaviviruses was conducted as previously described (58). Briefly, the prediction of transmembrane helices of the polyprotein was performed using TMHMM version 2.0 (https://services.healthtech.dtu.dk/service.php?TMHMM-2.0/). To identify signal peptides, a 70-amino-acid sliding window of the polyprotein was assessed with the SignalP version 6.0 webserver (https://services.healthtech.dtu.dk/service.php?SignalP-6.0/). In addition, the cleavage sites were predicted using the NetChop 3.1 server (https://services.healthtech.dtu.dk/service.php?NetChop-3.1/). The furin cleavage sites were identified using the ProP 1.0 webserver (https://services.healthtech.dtu.dk/service.php?ProP-1.0/). Putative N-glycosylation sites were predicted using the NetNGlyc 1.0 server (https://services.healthtech.dtu.dk/service.php?NetNGlyc-1.0/).

Phylogenetic analysis.

We selected representative viral sequences proposed by the International Committee on Taxonomy of Viruses (ICTV [https://ictv.global/report/]) to determine the evolutionary history of 16 nearly complete RNA viruses. Their closest viral relatives are based on the search in blastp (https://blast.ncbi.nlm.nih.gov/Blast.cgi/). The viral amino acid sequences were then aligned using MEGA-X version 10.2.6, employing the MUSCLE algorithm (79). All alignments were trimmed with TBtools version 1.09852 (80) utilizing the TrimAL program (81). The trimmed alignments were then checked manually. Maximum likelihood phylogenetic trees were reconstructed with the best-fit model of amino acid substitution determined using IQ-TREE version 1.6.12 (82). Branch support was accessed using the Shimodaira-Hasegawa-like approximate likelihood ratio test (SH-aLRT) and ultrafast bootstrap (UFBoot). The trees were visualized with iTOL (https://itol.embl.de/) (83) and refined by Adobe Photoshop CC version 19.1.6.

Data availability.

The raw sequence reads we analyzed in this study have been deposited in the SRA databases under the project accession no. PRJNA832012. All viral sequences with nearly complete genomes were deposited in the GenBank database under accession no. ON382564 to ON382579.