Green Sichuan pepper or Chinese pepper (Zanthoxylum schinifolium, Rutaceae) has been cultivated for the production of fruits and leaves in China for a long time. Its essential oil extract is known to have medicinal activities [1]. The extract also contains aromatic substances, so it is widely used as a spice in China and other East Asian countries. Several pathogens, including fungi and phytoplasma, have been described in the crop, causing serious damage [2, 3]. However, no viral disease has been reported in green Sichuan pepper.

Members of the genus Badnavirus in the family Caulimoviridae are important agricultural and horticultural pathogens that cause serious crop losses [4]. These viruses have non-enveloped, bacilliform virions of 120–150 × 30 nm and a single circular double-stranded DNA of 7.2–9.2 kb that contains three open reading frames (ORFs) arranged in a series on the plus strand [4]. In recent years, several new badnaviruses have been identified in various crops [5,6,7,8]. In 2018, virus-like vein clearing and mottling symptoms (Fig. 1a–c) were observed on the leaves of green Sichuan pepper trees in Chongqing Province, China. Total nucleic acids were extracted from leaf tissues of the symptomatic tree using an EASYspin Plus Complex Plant RNA Kit (Aidlab, China). An rRNA-depleted cDNA library was built and subjected to Illumina NextSeq sequencing with a layout of 150-bp paired-end (PE) reads (Biomarker Technologies, Beijing, China). Analysis of total sequenced reads was performed using CLC Genomics Workbench 9.5 (QIAGEN, USA) as described by Shen et al. [9], with sequences of the adaptors and reads of low quality and host sequences removed. The remaining reads were assembled de novo into contigs, and BLAST search programs were used for contig annotation (Supplementary Table S1). To determine the complete viral genome sequence, overlapping primer pairs (Supplementary Table S2) were designed based on the viral contigs. Total nucleic acids obtained from symptomatic leaves using a DNAsecure Plant Kit (Tiangen Biotech, China) were used for PCR amplification. PCR products were purified and cloned into the pEASY-T1 vector (TransGen Biotech, China). At least five clones for each amplicon were sequenced, and the sequences that were obtained were assembled using SeqMan software (DNAStar Lasergene, USA). ORF Finder (https://www.ncbi.nlm.nih.gov/orffinder/), Conserved Domain Search (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi), and DRNApred (http://biomine.cs.vcu.edu/servers/DRNApred/) were used to predict the genomic ORFs, identify conserved protein domains, and predict DNA/RNA binding sites, respectively.

Fig. 1
figure 1

Viral symptoms on a twig (a), a young leaf (b), and mature leaf (c) associated with green Sichuan pepper vein clearing-associated virus (GSPVCaV) in a green Sichuan pepper tree. (d) Schematic genome organization of GSPVCaV showing three open reading frames (ORF1–ORF3) as well as the tRNA primer binding site and conserved functional motifs in viral proteins

Full size image

A total of 55,749,940 clean reads were obtained, and 17,782,807 (31.9%) of them were used for BLASTx annotation after removal of the reads mapped to the referenced Citrus (Rutaceae) genomes, as no genome sequences were available for Zanthoxylum species. De novo assembly of unmapped reads produced 87,834 contigs. One contig of 7,994 nt showed significant nucleotide (nt) sequence identity (56.5%) to grapevine vein-clearing virus (GVCV), suggesting the presence of a new badnavirus in the sample. Mapping of transcriptome reads as well as RT-PCR detection of viral mRNAs (Supplementary Fig. S1 and Supplementary Table S2) suggested that this is an episomal virus rather than fragments integrated into the genome of the host plant. The virus was provisionally named “green Sichuan pepper vein clearing-associated virus” (GSPVCaV, isolate CQ-1), as it was the sole virus found by sequencing the cDNA library from the symptomatic tree.

The complete genome of GSPVCaV is 8,014 nt in size (GenBank accession No. MK371353). Comparisons with other badnaviruses identified a putative tRNAmet-binding site (5′-1TGGTATCAGAGCTTCGGC18), a potential 7,xxxTATA7,xxx+4 box, and a possible polyadenylation signal (7,xxxAATAAA7,xxx+6) within the intergenic region on the plus strand [10]. The genome of the virus is typical of badnaviruses, containing three open reading frames (ORF) (Fig. 1d). ORF1 (nt 359–982) encodes a putative protein (P1, 23.7 kDa) of 207 amino acid (aa) residues that contains a motif (pfam07028, aa 88–189) that is conserved among badnaviruses [8]. The P1 of unknown function has been associated with virions [11]. ORF2 (nt 979–1,368) is predicted to encode a 129-aa protein (P2, 14.4 kDa) with a DNA-binding site at aa 49–129, suggesting non-specific DNA- and RNA-binding activities [12]. ORF3 (nt 1,365–7,538) encodes a large polyprotein of 2,057 aa (P3, 232.5 kDa). Like other badnaviruses, the polyprotein contains the coat protein, movement protein (MP), a highly conserved zinc-finger-like RNA-binding domain (CXCX2CX4HX4C), reverse transcriptase (RT), viral aspartic protease (AP), and ribonuclease H (RNase H) domains [13]. Genome-wide comparisons with badnaviruses confirmed that GSPVCaV is most closely related to GVCV. The sequence identities between the two viruses are 56.5% for the whole genome nt sequence, and 64%, 51.2% and 56.0% for the P1, P2 and P3 aa sequences, respectively. The nt sequence identity of the joint region of the RT and RNase H domains of P3 is 71.9%, which is lower than the species demarcation cutoff (80%) for the genus Badnavirus [4], suggesting that GSPVCaV should be considered a member of a new species in this genus.

A phylogenetic tree was generated by the neighbour-joining method from the alignment of the genome sequences of GSPVCaV and representative badnavirus sequences retrieved from the GenBank database. The phylogenetic analysis placed GSPVCaV and GVCV in the same cluster, confirming their close relationship and supporting that GSPVCaV is a distinct badnavirus (Fig. 2).

Fig. 2
figure 2

Neighbor-joining phylogenetic tree derived from an alignment of the complete genome sequences of GSPVCaV and other extant badnaviruses, with 1000 bootstrap replicates. Bootstrap values are shown at the relevant nodes, and only bootstrap values higher than 50% shown. The scale bar represents substitutions per site

Full size image

Our field surveys showed that vein clearing and similar symptoms were common in green Sichuan pepper in Sichuan Province. Another leaf sample (SC-1) was subsequently collected from trees at different locations and subjected to Illumina NextSeq sequencing. Bioinformatics analysis identified only one viral contigs with high sequence similarity to GSPVCaV-CQ1. The complete genome sequence of 8,070 nt (MK371354) was obtained by Sanger sequencing for this isolate. The SC-1 isolate is 94.0% identical to the CQ1 isolate. Sequence differences mainly concentrated at the NCR (7,525–359 nt), with the exception of a 15-nt insertion at nt 4,669. A virus-specific primer pair (HJV-F/-R; Table S2) was designed based on the RT-RNaseH region for the two isolates and used in PCR to test 25 symptomatic and six asymptomatic trees. The virus was detected in 23 of the 25 symptomatic trees (92.0%) but not in any asymptomatic trees (data no shown), suggesting a tight association of the virus with the vein clearing symptoms. However, the biological characteristics of the isolates CQ-1 and SC-1 and the biological differences between them remain undetermined, as Koch’s postulates are yet to be fulfilled.