Turnip mosaic virus (TuMV; genus Potyvirus, family Potyviridae) has a c.10 kb plus-sense RNA genome and infects over 318 plant species in 156 genera from 43 families, especially in Brassicaceae; only Cucumber mosaic virus is more important in vegetables worldwide [1–6]. TuMV is transmitted non-persistently by many aphid species [6]. Potyvirus helper component-protease (HC-Pro) functions include aphid-mediated virus transmission; RNA amplification; systemic movement; suppression of post-transcriptional RNA silencing; and C-terminal self-cleaving proteinase [7, 8]. Three major HC-Pro domains correspond approximately to the N-terminal 100, central 200, and C-terminal 150 residues [9]. An N-terminal KITC motif binds aphid stylets [10] and a C-terminal PTK triplet binds to DAG in the coat protein N-terminus [11, 12]. The central domain affecting RNA silencing suppression, genome amplification, synergism with other viruses, and systemic movement includes conserved FRNK, IGN (LAIGN), and CCC motifs [13–15] and other residues associated with non-specific RNA binding [16].

Due to its many hosts worldwide, there is interest in the evolution of TuMV isolates with differential host ranges. Two major groups are known; B-type isolates mainly infect Brassica species, but not radish (Raphanus sativus), while BR-type isolates infect both Brassica and Raphanus [17, 18]. The Asian-BR group includes many Asian isolates, but basal-BR isolates are emerging in Asia [19, 20]. Strains are defined in part by host resistance genes, and by viral pathogenicity determinants [2, 3, 18, 19, 21, 22] which remain poorly defined; HC-Pro variation has been implicated in differences between horseradish strains [23].

We have shown that replication and symptom severity of a potexvirus, Alternanthera mosaic virus, are significantly affected by a single residue change in the Triple Gene Block 1 protein viral suppressor of RNA silencing (VSR) [24]. Here we examine VSR function and subcellular localization of HC-Pro of TuMV isolates from radish, an important crop in Korea, in relation to symptom expression to determine whether the TuMV VSR has similar effects on disease severity.

TuMV isolates R007, R041, and R065 were selected from R. sativus in Korea [25] based on differential symptoms in R. sativus, B. rapa ssp. pekinensis (Chinese cabbage), and Nicotiana benthamiana. Each isolate induced mild mosaic and mottle in R. sativus at 11–20 days post inoculation (dpi) (Suppl. Fig.  1). However, in Chinese cabbage isolates R041 and R065 induced systemic necrosis (SN) by 20 dpi, with plants dead or nearly dead at 30 dpi, while plants inoculated with R007 developed systemic mottle or mosaic without SN by 20 dpi, remaining so at 30 dpi (Fig. 1a). R041 and R065 induced SN in N. benthamiana while R007 caused only mild mosaic [25].

Fig. 1
figure 1

a Turnip mosaic virus (TuMV) pathogenicity: Non-infected plant (Healthy) and TuMV infected (isolates R007, R041 and R065) of Brassica rapa ssp. Pekinensis: pictures taken 20 days post inoculation (d.p.i). b Confocal images of GFP:HC-Pro localization: pGDR:Talin and pGDG:HC-Pro of (i) R007, (ii) R041, and (iii) R065, plus pGD:p19 were agroinfiltrated to N. benthamiana; at 3 days after agroinfiltration, infiltrated leaves were DAPI stained. Blue color (DAPI) indicates nuclei, red color indicates chloroplast autofluorescence, green color indicates GFP:HC-Pro, and white represents DsRed:Talin. Bars indicate 20 µm. GFP:HC-Pro protein expression was confirmed at 77 kDa by Western blot using anti GFP monoclonal antibody (insets). c VSR assay (i) and quantitative RT-PCR (qRT-PCR) results (ii). i. VSR activity of TuMV HC-Pro in N. benthamiana leaves; pGD:TuMV HC-Pro of R007, R041 and R065, empty pGD vector, or pGD:p19 were separately co-agroinfiltrated with pGD:smGFP. R007, R041, and R065. p19 = silencing suppressor of Tomato bushy stunt virus used as a positive control. smGFP: smGFP co-infiltrated with empty pGD vector. ii. RNA extracts from each smGFP expressing region were subjected to qRT-PCR with smGFP specific primers. Relative RNA accumulation of smGFP was normalized to actin from two biological replicates; bars indicate standard error

Full size image

HC-Pro genes of each isolate were cloned and sequenced (Macrogen; Daejeon, Korea). Sequences of each were 1374 nucleotides (nt) (genome accession numbers KU140240, R007; KU140241, R041; KU140242, R-065), encoding 458 amino acids (aa; predicted size 50 kDa). Pairwise R041:R065 identities were higher (99.1 % nt, 99.6 % aa) than either R007:R041 or R007:R065 (87.3 % nt, 98.5 % aa each). Phylogenetic analysis by the Maximum-likelihood method (MEGA v.6; [26]) with 33 additional TuMV HC-Pro sequences from GenBank and Potato virus Y (PVY) as outgroup resulted in four TuMV clusters according to pathotype: Basal-B, World-B, Basal-BR, and Asian-BR ([17, 18]; Suppl. Table 1; Suppl. Fig. 2). The three Korean isolates grouped into the Basal-BR clade of mainly European isolates, with 91 % bootstrap support. Apart from identities between Korean isolates, R007 was most closely related to Italian isolate ITA7 (86.5 % nt, 97.8 % aa); R041 and R065 were most closely related to Japanese isolate KWB779 J (97.7 and 97.9 % nt respectively, both 99.3 % aa)(Suppl. Fig. 2, and data not shown). Eight aa residues differentiate the Korean isolates, with six residues distinguishing R007 from both R041 and R065; R041 and R065 each differed by one residue from the consensus sequence (Suppl. Fig. 3).

HC-Pro genes of each isolate were cloned into vectors pGD (for VSR assay) and pGDG (for subcellular localization) [27] and transformed into Agrobacterium tumefaciens EHA105. Agrobacterium transformed with pGDR:Talin (actin marker), pGD:p19 [VSR of Tomato bushy stunt virus (TBSV)], and pGD:smGFP (soluble-modified green fluorescent protein) were also utilized [24, 28]. Equal volumes of pGDG:HC-Pro constructs and pGDR:Talin were mixed, usually with 0.1 volume of pGD:p19; agroinfiltrated leaves [28] of N. benthamiana were examined at 2–3 days post-agroinfiltration (dpa) using a Zeiss 710 confocal microscope [29, 30], with DAPI staining as described [31]. Imaging revealed HC-Pro of all isolates distributed throughout the cytoplasm and at the cell periphery, together with perinuclear and microfilament-like association; R007 and R041 also induced many small punctate aggregates in the cytoplasm, presumed to result from self-interaction (Fig. 1b, Suppl. Fig. 4). In contrast R065 formed punctate aggregates in few cells, and those aggregates were smaller (Fig. 1b, Suppl. Fig. 4). Western blotting showed similar HC-Pro expression levels for each isolate (Fig. 1b). Omission of pGD:p19 caused no apparent differences in HC-Pro aggregation (Suppl. Fig. 4) in contrast to [32].

Agroinfiltrations of pGD:smGFP alone, or mixed with either pGD:HC-Pro, pGD:p19, or empty pGD vector, were examined at 6 dpa under UV light; R065 HC-Pro yielded minimal enhancement of smGFP expression, whereas R007 and R041 HC-Pro significantly enhanced expression. These results were supported by qRT-PCR results (Fig. 1c), indicating that R007 and R041 HC-Pros were more effective VSRs than R007.

Because one residue (F395 in R007, R041; L395 in R065; Suppl. Fig. 3) correlated with formation of punctate HC-Pro aggregates and VSR efficiency (Fig. 1b,c), we mutated residue 395 Phe > Leu in R007 and R041 (F395L), and Leu > Phe in R065 HC-Pro (L395F) by overlap extension PCR [33] using primers shown in Table 1. Mutants were cloned into pGD and pGDG for comparison to the respective WT constructs to assay effects on HC-Pro aggregation and VSR activity. Mutants pGD:HC-Pro R007F395L and R041F395L had obviously reduced VSR efficiency, but R065L395F gained VSR efficiency relative to the respective WT (Suppl. Fig. 5). Whereas WT pGDG:HC-Pro R007F395 produced many punctate aggregates (Fig. 1b, Suppl. Figs. 4,5), R007F395L produced few punctate aggregates (Suppl. Fig. 5). WT R065L395 produced few punctate aggregates with most signal along filaments (Fig. 1b, Suppl. Figs. 4,5), while mutant R065L395F yielded obvious punctate aggregates without clear filament association (Suppl. Fig. 5). Both WT R041F395 and mutant R041F395L produced obvious punctate aggregates scattered throughout the cytoplasm (Fig. 1b, Suppl Figs. 4, 5).

Table 1 Primers used in this study
Full size table

HC-Pro is a well-known VSR affecting viral RNA stability and replication levels, with multiple sequence differences between isolates (Suppl. Fig. 3). VSR activities of R007 and R-041 were similar, and more than twice that of R065, not corresponding with differences in SN (Fig. 1a,c). R041 and R065 HC-Pro differ by only two residues, of which only R041 Leu207 is in the central domain associated with VSR activity. As R065 (weak VSR) and R007 (similar to R041) have Phe207, this cannot explain relative VSR efficiency unless other changes in R007 compensate for ‘negative’ effects of Phe207. The other difference is Phe395 (R041) or Leu395 (R065) in the C-terminal protease domain. Two mutants in the central, and one in the C-terminal domain, of Tobacco etch virus (TEV) HC-Pro reduced both VSR activity and symptom severity; other mutants throughout TEV HC-Pro either ablated VSR activity and systemic infection or were neutral or increased both VSR and symptom expressions [34]. A mutation of the Zucchini yellow mosaic virus FRNK motif to FINK caused dramatic symptom amelioration in squash and ablation of symptoms in other cucurbits despite accumulation levels similar to WT [35]. Although none of the differences between R007, R041, and R065 affect any previously identified HC-Pro motifs ([8, 10–13, 15, 36–39], Suppl. Fig. 3), R065 Leu395 may affect HC-Pro folding, local structure, and/or self-interactions affecting VSR activity, restored in R065L395F (Suppl. Fig. 5).

Subcellular aggregation of GFP:HC-Pro corresponded closely with VSR activity, suggesting a relationship between HC-Pro self-interaction and VSR activity. Lack of R065 HC-Pro aggregation appears related to differences in self-interaction, but aggregation is not directly related to SN induction. The single C-terminal variation between R007/R041 and R065 HC-Pro (F395L) may influence self-interaction and inhibit aggregation (Suppl. Fig. 5), as this domain contributes to TuMV HC-Pro self-interaction [40]. However, R041F395L reduced VSR activity but did not ablate aggregation, suggesting that L207 may modify the effect of F395 on aggregation (Suppl. Fig. 5).

Differences in GFP:HC-Pro localization correlated with differences in VSR efficiency (R007 and R041, higher VSR activity, HC-Pro aggregation; R065, weaker VSR, few punctate aggregates). However, these HC-Pro characters do not correlate with symptom severity (R041, R065 induction of SN in Chinese cabbage and N. benthamiana, R007 mosaic in both; this study, [25]). TuMV symptom determinants appear to be host specific, even within the Brassicaceae.

Different potyviral proteins affect symptoms in a host-specific manner [e.g., 41–44], and the lack of correlation between VSR efficiency and SN induction suggests that differences in viral accumulation due to HC-Pro are not responsible. We will further examine differences in HC-Pro identified here through construction of infectious clones, and substitution of different HC-Pro sequences and single residues to form chimeras in a common TuMV backbone genome. Through such studies, we hope to further advance research into TuMV biology and identify determinants of pathogenicity.