Turnip yellows virus (TuYV; family Luteoviridae, genus Polerovirus) is a phloem-limited virus that is obligatorily transmitted by aphids (Myzus persicae) and has a wide host range [3, 17]. Since TuYV has been reported from all continents [1, 5, 6, 10, 15, 17, 20], it probably has a worldwide distribution. In the UK, TuYV is the most important viral pathogen of oilseed rape and an emerging threat to its production [17]. In Germany and Australia, yield losses of oilseed rape compared to controls amounted to about 34% and 46%, respectively [4, 11]. There was considerable confusion concerning the taxonomy of beet poleroviruses and TuYV until investigations of host range and molecular properties provided valuable data for distinguishing them [2, 8, 9, 16, 19]. The International Committee on Taxonomy of Viruses (ICTV) has assigned European isolates of beet western yellows virus (BWYV) that do not infect sugar beet to the independent species Turnip yellows virus in the genus Polerovirus [3, 16]. The complete genomic sequence of a French lettuce isolate that was previously referred to as BWYV (BWYV-FL1) is now recognized as the standard TuYV sequence (GenBank acc. no. NC003743) [3, 16, 18]. Although sequences of some open reading frames (ORF0 and ORF3) of crucifer-infecting TuYV isolates are available [8, 14], the complete genome sequences of TuYV isolates from cruciferous crops have not yet been determined.

There was only limited information on the incidence and distribution of brassica-infecting poleroviruses in China until we first reported that TuYV is widely distributed in China, occurring on eight different cruciferous crops and in ten provinces [20]. In this report, the complete genome sequences of two genotypes of a brassica-infecting polerovirus most closely related to TuYV were determined. This revealed that the Chinese brassica-infecting polerovirus was similar to but distinct from TuYV and that the two polerovirus genotypes differed mainly in the 5′-terminal half of the genome.

An investigation of the occurrence of brassica-infecting polerovirus in mainland China was commenced in September 2006. One hundred ninety-two leaf samples showing yellowing or leafroll symptoms (Fig. 1) were collected from nine different cruciferous plant species including cabbage (Brassica oleracea var. capitata), Chinese cabbage (B. pekinensis), cauliflower (B. oleracea var. botrytis), flowering Chinese cabbage (B. chinensis), leaf mustard (B. juncea), oilseed rape (B. napus ssp. napus), rutabaga (B. napobrassica), white glabrous mustard (B. alboglabra) and radish (Raphanus sativus var. oleifera). Total RNA was prepared by SDS-phenol/chloroform extraction and LiCl precipitation as described previously [7]. RT-PCR detection was performed using a universal primer pair (PocoCPR/PoconF) as described previously [13, 20] (Table 1). Sixty-eight of the 192 samples tested positive, providing evidence for polerovirus infections in all of the crops tested and in 11 provinces, including Beijing, Inner Mongolia, Heilongjiang, Liaoning, Henan, Gansu, Shandong, Hubei, Jiangsu, Yunnan and Chongqing. This indicated that brassica-infecting poleroviruses are distributed widely from the northeast to the southwest in mainland China. Furthermore, some RT-PCR products from these crops and provinces were cloned and sequenced. BLAST analysis showed that these sequences shared highest identities with TuYV isolates, suggesting that a TuYV-like virus is the causal agent of crucifer yellowing disease in mainland China.

Fig. 1
figure 1

Yellowing and leafroll symptoms of the brassica-infecting polerovirus. A. Interveinal yellowing in Chinese cabbage collected from Beijing on 3 October 2006; B. Leafroll symptoms on oilseed rape from Jiangsu on 28 March 2007

Table 1 Primers used for RT-PCR

Since no complete genomic sequence of a brassica-infecting polerovirus had been reported, we decided to analyze the genomic sequence of a Chinese polerovirus isolate from a rutabaga plant showing leaf yellowing in the Haidian district of Beijing in September 2008. Primers for RT-PCR amplification of the genome sequences were derived from the TuYV-FL1 sequence (GenBank acc. no. NC003743) and the sequences that we obtained (Table 1). Two overlapping fragments (Tu5-31F/Tu2750R and Tu1738F/TYR3) of the genome were obtained by RT-PCR and cloned, and at least three independent positive PCR clones were sequenced. Based on genome sequence analysis, we concluded that the rutabaga-infecting polerovirus is a distinct virus, referred to as brassica yellows virus (BrYV). For the fragment (Tu5-31F/Tu2750R), sequence analysis showed there were two distinct types of clones, which shared only 90.6% nucleotide sequence identity. This suggested that there were two genotypes infecting rutabaga in Beijing, named BrYV-ABJ and BrYV-BBJ. Primer TuB1433F was subsequently designed specifically for BrYV-B and a specific fragment (TuB1433F/TYR3) was amplified, cloned and sequenced. For the 5′ and 3′ terminal regions, two fragments (Pocon5F/Tu862R and Tu4877F/Pocon36R) were acquired and sequenced. Thus, the complete genomic sequences of two types of the BrYV isolate from Beijing (BrYV-ABJ and BrYV-BBJ) were determined except for the short region where the primers annealed at the 5′-and 3′-termini (GenBank acc. no. HQ388348 and HQ388349, respectively). Using a similar strategy, polerovirus sequences were also obtained from leafroll-affected oilseed rape samples collected from the Yangzhou district of Jiangsu province in March 2009 and submitted to GenBank (acc. no. HQ388350 and HQ388351). Analysis of the sequences revealed that the oilseed rape samples from Jiangsu also contained two genotypes (BrYV-AJS and BrYV-BJS, respectively) that were similar to those found in the rutabaga sample from Beijing.

Both BrYV-ABJ and BrYV-BBJ were 5666 nt in length, making them slightly larger than the TuYV genome (5641 nt), and had a similar genomic organization. They had six ORFs (ORF0-ORF5), a 5′ UTR of 31 nt, an intergenic NCR of 203 nt (from nt 3269 to 3471) between ORF2 and ORF3, and a 3′ UTR of 185 nt. The first ORF (ORF0) beginning with an AUG at nt 32 to 34 ended with an UGA at nt 779 to 781 and encoded putative proteins of 28.9 kDa (BrYV-ABJ) and 28.8 kDa (BrYV-BBJ). The second ORF (ORF1, 1812 nt) starting with an AUG at nt 174 to 176 was predicted to encode a 65.6-kDa protein (P1) and was 12 nt shorter than that of TuYV. The third ORF (ORF2, from nt 2156 to 3268) was presumed to be translated by a -1 frameshift and produce a putative P1-P2 fusion protein (115.3 kDa for BrYV-ABJ and 115.0 kDa for BrYV-BBJ). The intergenic NCR was 203 nt in length compared to 202 nt for TuYV. The fourth ORF (ORF3, 609 nt) began at nt 3472 and was deduced to encode a major coat protein (P3, 22.5 kDa for BrYV-ABJ and 22.4 kDa for BrYV-BBJ). The fifth ORF (ORF4, from nt 3503 to 4030), which was entirely embedded in ORF3, is proposed to be translated into a protein for cell-to-cell movement (P4, 19.5 kDa for BrYV-ABJ to 19.4 kDa for BrYV-BBJ) by a leaky scanning mechanism. The last ORF (ORF5, from nt 4267 to 5481) directly adjoins ORF3, with ORF3-ORF5 potentially encoding readthrough proteins of 75.1 kDa (BrYV-ABJ) and 75.2 kDa (BrYV-BBJ) past the amber stop codon of ORF3. ORF5 of BrYV-ABJ and -BBJ was 81 nt shorter than that of TuYV, resulting in a 1.8- to 1.9-kDa-smaller protein compared to the 47.3 kDa RTD of TuYV. The 3′ UTRs of both BrYV-ABJ and BrYV-BBJ were 29 nt longer than that of TuYV. The genomic size and structure of both BrYV -AJS and BrYV -BJS were identical to those of BrYV-ABJ and BrYV-BBJ. Taken together, BrYV isolates have a genomic organization similar to that of TuYV, except for the length of ORF1, ORF5, the intergenic NCR and the 3′ UTR.

Further sequence comparisons were done using DNAMAN software (Version 6.0, Lynnon Biosoft, Quebec, Canada) and Bioedit software (Version 7.0.4.1). The nucleotide sequences of the two genotypes of BrYV were compared to those of TuYV and other poleroviruses. The entire genome of BrYV-ABJ shared 94.6% nucleotide sequence identity with BrYV-BBJ, where the 5′ terminal ORFs displayed more divergence (88.8–92.5%) and the 3′ terminal ORFs showed higher nucleotide sequence identities of 99.2–99.8%. The nucleotide sequence identity between BrYV-AJS and BrYV-BJS was similar to that between BrYV-ABJ and BrYV-BBJ. Nucleotide sequence identities for the full-length sequence and each gene of BrYV-ABJ and BrYV-AJS/BrYV-BBJ and BrYV-BJS were in the range of 96.0% to 99.6%, and their gene products shared amino acid sequence identities ranging from 94.8% to 99.4%. These results indicated that BrYV-ABJ and BrYV-AJS/BrYV-BBJ and BrYV-BJS are different isolates of the same species. However, the full-length sequence of the two BrYV genotypes shared only 80.6-80.7% nucleotide sequence identity with that of TuYV. Sequence identities between the two BrYV genotypes and TuYV were 87.6–88.7% for ORF0, 87.0–87.3% for ORF1, 89.0–89.1% for ORF1-2, 90.6–91.1% for NCR, 94.1–94.6% for ORF3, 94.3–94.7% for ORF4, 67.0–67.3% for ORF3-5, 50.5–50.8% for ORF5 and 56.8–57.3% for the 3′ UTR. At the amino acid level, the sequence identities were 81.1–83.5%, 84.4–84.8%, 89.1–90.1%, 94.1–94.6%, 89.1%, 59.9–60.8% and 39.2–40.0% for P0, P1, P1-2, P3, P4, P3-5 and P5, respectively. Therefore, the gene products of the two BrYV genotypes shared less than 90% amino acid sequence identity with those of TuYV, with exception of the highly conserved P3 (coat) protein. In order to determine the taxonomic status of BrYV, detailed sequence comparisons between BrYV and other reported poleroviruses were also conducted. The results showed that the sequence divergence of all of the gene products of BrYV from those of other poleroviruses was greater than 10%, except for the coat proteins of BrYV and beet poleroviruses, which differed by 7.9–9.4%. Based on the differences in genome length and sequence comparisons, we propose that BrYV represents a distinct species whose members infect cruciferous crops.

To better understand the relationships of BrYV to other poleroviruses, including TuYV, phylogenetic trees were generated by the Clustal W method and visualized with MEGA (Version 4.1) (Fig. 2) [12]. Phylogenetic analysis showed that the 5′-terminal ORFs (ORF0, ORF1, ORF1-2) and their gene products were closely related to, but distinct from, those of TuYV (Fig. 2a and b). However, they differed in their 3′-proximal ORFs (ORF3, ORF4 and ORF5) and their gene products (P3, P4, P3-P5 and P5), which was consistent with the sequence comparison results (Fig. 2c and d). Surprisingly, the BrYV P5 showed less sequence identity to that of TuYV but was more similar to those of cereal poleroviruses (cereal yellow dwarf virus-RPS and -RPV), suggesting that the difference in P5 sequence between the BrYV and the TuYV isolates could be linked to different aphid transmission specificity or efficiency [16]. Taken together, the phylogenetic analysis results strongly indicated that BrYV was distinct from TuYV and other members of the genus Polerovirus.

Fig. 2
figure 2

Phylogenetic trees generated from aligned amino acid sequences of poleroviruses. The phylogeny reconstruction was carried out via the N-J method and visualized by using the MEGA4.1 program, where the bootstrap value was for 500 replicates. A. P0; B. P1-P2 fusion protein; C. P3, coat protein; D. P5 readthrough domain. The following virus sequences (virus abbreviations and accession numbers in parentheses) were obtained from the GenBank database: beet chlorosis virus (BChV, NC002766), beet mild yellowing virus (BMYV, NC003491), beet western yellows virus (BWYV, NC004756), cucurbit aphid-borne yellows virus (CABYV, NC003688), chickpea chlorotic stunt virus (CpCSV, NC008249), carrot red leaf virus (CtRLV, NC006265), melon aphid-borne yellows virus (MABYV, NC010809), potato leafroll virus (PLRV, NC001747), turnip yellows virus (TuYV, NC003743), cereal yellow dwarf virus-RPS (CYDV-RPS, NC002198), CYDV-RPV (NC004751), sugarcane yellow leaf virus (SCYLV, NC000874), BrYV-ABJ (HQ388348), BrYV-B BJ(HQ388349), BrYV-AJS (HQ388350) and BrYV-BJS (HQ388351)

In conclusion, this article describes the first BrYV genome sequences isolated from Brassica crops and provides evidence for the presence of two distinct genotypes infecting rutabaga in Beijing and oilseed rape in Jiangsu. With the exception of the highly conserved P3 protein, the amino acid sequence identities of all of the gene products were lower than 90%, thus meeting the molecular criteria for species demarcation in the genus Polerovirus [3]. However, since there is no information on the pathogenicity and aphid transmission specificity and efficiency of the BrYV genotypes, our future investigations will emphasize these aspects. Moreover, the determination of complete genomic sequences of virus isolates from different crucifer crops and geographical regions is essential for elucidating the genetic diversity and epidemiology of crucifer poleroviruses in China and elsewhere.