Water chestnut (Eleocharis dulcis (Burm. f.) Trin. is a perennial aquatic plant of the genus Eleocharis (Cyperaceae) that occurs naturally in China, Japan, Ryukyu Islands, Nanyang Island and India [1]. Its underground corms can be eaten raw or cooked and are one of the characteristic vegetables in China. There are about 50,000 ha of cultivated water chestnut in China including 20,000 ha in Guangxi province, particularly around Hezhou and Lipu and there is also significant cultivation in Tuanfeng and Shayang counties, Hubei province [2, 3].
Because the plants are propagated vegetatively, viruses are easily transmitted to the next generation when germplasm is exchanged, causing a decline in crop yields and significant economic losses. Several viruses have previously been reported from water chestnuts. Cucumber mosaic virus (CMV) causes chlorosis, dwarfing and deformity and was first identified on water chestnuts in 2014 [4], while a double-stranded DNA virus named water chestnut soymovirus-1 was discovered by high-throughput sequencing in 2019 [2].
Over the period 2016-2020, it was noticed that the quality of corms of a local water chestnut variety in Guangxi had become degenerated; corms were smaller and the leaves had chlorotic spots. Diseased leaf samples were collected in July 2020 and sent to Zhejiang Academy of Agricultural Sciences (Hangzhou, China) for next generation sequencing (NGS). Total RNAs were extracted with TRIzol (Invitrogen, Carlsbad, USA) and RNA integrity was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies). A cDNA library was constructed with the TruSeq RNA Sample Preparation Kit (Illumina) according to the manufacturer’s instructions. An Illumina NovaSeq 6000 platform with PE150bp and CLC Genomic Workbench 11(QIAGEN) was used for sequencing and data analysis with default parameters. A total of 36,026,984 paired-end reads were obtained, and 12,675contigs (903 to 18749bp) were generated de novo and compared with sequences in GenBank using BLASTn or BLASTx. Three of the contigs (3858nt, 2861nt and 2144nt) were clearly related to the three genomic RNAs of viruses in the genus Ilarvirus (Bromoviridae) with the highest nucleotide sequence identities of 57.5%, 74.0% and 59.4% respectively to different members of the genus, suggesting the existence of a distinct ilarvirus. Ilarviruses are a group of isometric and labile viruses that are distributed worldwide and infect many hosts including vegetables, ornamentals and fruit trees [5, 6]. Their genomes encode four or five proteins on three RNA strands, and they have mostly been classified into four phylogenetic sub-groups [7].
To determine the complete genomic RNA sequence of the water chestnut virus, specific primer sets were designed (Supplementary Table 1) to amplify each of the RNAs in two overlapping segments for cloning and splicing. Total RNA was extracted from the infected water chestnut sample using the EASYspin RNA Plant Mini Kit (Aidlab Biotechnologies Co., Ltd, China). Total RNAs were polyadenylated using a Poly(A) Polymerase (TAKARA, Japan), and then reverse transcribed using M4-T primer and the First Strand cDNA Synthesis Kit (Toyobo, Osaka, Japan) [8]. Rapid amplification of cDNA ends (RACE) (Tiosbio, Beijing, China) was used to ensure that the complete sequences were obtained. Each PCR product was inserted into the pEASY-T5 Zero Cloning Vector (TransGen Biotech, Beijing) and more than five clones per reaction were sequenced (ykang, Hangzhou, China). The 5’ and 3’RACE fragments amplified by corresponding primers were then assembled using DNAMAN 8.0 software (Lynnon Biosoft, Canada).
The complete genomic sequences of the three RNA segments were 3578nt (RNA1), 2873nt (RNA2) and 2073nt (RNA3) in length (Fig.1) and were deposited in GenBank (accession nos. MZ170696-MZ170698) with the provisional name of water chestnut virus A (WCVA). The 5’untranslated regions (UTRs) of RNA2 and RNA3 of WCVA (49 and 101 nt in length, respectively) both start with TAAA, while the 3’-UTRs of RNA1 and RNA3 both end with GAUGGC. Sequence analysis suggests that WCVA has four ORFs (Fig.1). The single ORF on RNA1 (P1: nt 43-3387; 125.11 kDa) had two conserved domains predicted by the protein families database (Pfam: http://pfam.xfam.org/ [9]): viral methyltransferase and helicase (Fig.1). In BLASTp the P1 protein of WCVA had the highest amino acid sequence identity (98% query coverage and 57.3% identity) to prune dwarf virus (PDV) (ASJ26572). RNA2 also has a single predicted ORF (P2: nt 50-2710; 99.99 kDa) (Fig. 1) that encodes an RNA-dependent RNA polymerase. The P2 protein had the highest amino acid sequence identity to gungahlin flea-associated ilarvirus (GFAIV) (QIJ70041, 89% query coverage and 58.1% identity). RNA3 is predicted to contain two non-overlapping ORFs, encoding a movement protein (MP: nt 102-912; 33.80 kDa) and a coat protein (CP: nt 1088-1765; 24.76 kDa) (Fig. 1). These had the highest amino acid sequence identities to isolates of potato yellowing virus (PYV) (respectively 85% query coverage and 61.7% identity with AZZ70614 and 98% query coverage and 40.6% identity with QBO24586).
To determine the relationship between WCVA and other known ilarviruses, phylogenetic trees were prepared for each of the ORFs using representative sequences of all previous reported ilarviruses (Supplementary Table 2) and with CMV (genus Cucumovirus) included as an outgroup. Analysis was conducted in MEGA X [10] using the maximum likelihood method with 1000 bootstrap replicates. In all ORFs, WCVA clustered in subgroup 4 with PDV, viola white distortion associated virus (VWDAV), fragaria chiloensis latent virus (FCILV) and PYV, but the exact clustering pattern among these viruses differed slightly (Fig.2). In addition, the P2 protein of WCVA showed the closest relationship with that of the GFAIV. Howerver, the RNA2 was the only sequence of GFAIV reported so far. Inspection of the trees suggests that three unclassified viruses: surrounding legume associated ilarvirus (SLAIV), grapevine associated ilarvirus (GAIV) and solanum nigrum ilarvirus 1 (SNIV-1) are very closely related and are probably members of a single (new) species.
Total protein extracted from leaves of infected water chestnut were tested in DAS-ELISA, with antiserum to the ilarviruses American plum line pattern virus (APLPV), tobacco streak virus (TSV), PDV and prunus necrotic ringspot virus (PNRSV) (Adgia) but there were no positive reactions.
Sequence similarity criteria for demarcation of species within the genus Ilarvirus have not been defined [11] but the low amino acid identities to known ilarviruses in all the predicted proteins (<62%) and the serological results all suggest that WCVA is a distinct novel member of the genus. Further research will focus on the incidence of this virus on different water chestnut varieties and its host range and pathogenicity.