Soil-borne wheat mosaic virus (SBWMV), the type member of the genus Furovirus (family Virgaviridae), causes yellow mosaic disease in winter wheat (Triticum aestivum L.) [12]. SBWMV is transmitted by Polymyxa graminis Ledingraham [8]. The two single-stranded positive-sense RNA genomic segments are separately encapsidated in rigid rod-shaped virions. The 7.1-kb RNA1 encodes three proteins: the 150-kilodalton (kDa) and 209-kDa readthrough replicase proteins and the 37-kDa movement protein (MP). The 3.6-kb RNA2 encodes three proteins: the 19-kDa coat protein (CP), the larger 84-kDa CP readthrough protein (CPRT), and the 19-kDa cysteine-rich protein (CyR) (Fig. 1A) [19]. Studies on furovirus molecular biology to date have focused primarily on Japanese soil-borne wheat mosaic virus (JSBWMV; formerly classified as SBWMV-JT) because of the availability of infectious clones [21]. Development of full-length cDNA clones of SBWMV were less successful due to the instability of RNA1 in Escherichia coli, but one was developed using transcripts prepared from two overlapping RNA1 cDNA products [13]. We describe the development of an agro-binary plasmid-based full-length infectious clone of an Ohio SBWMV isolate and evaluation of multiple insertion sites within RNA2 for gene-carrying capacity.

Fig. 1
figure 1

Schematic representation of SBWMV genome and mutant clones. (A) RNA1 and RNA2 genome maps are shown, indicating the 5’ cap, the 3’ tRNA like structure, and the coding regions in boxes with nucleotide coordinates. (B) RNA2-derived cDNA clones with deletions or insertions. Deletions in the wild-type virus sequence are shown with red dashed lines. PDS sequence insertions from N. benthamiana (370 nt) or T. aestivum (409 nt) in the antisense orientation are shown as blue textured boxes (color figure online)

Full size image

Ohio soft red winter wheat samples collected in spring 2012 were used for RNA-Seq to identify virus sequences as described previously [20]. SBWMV 7105-nt RNA1 and 3593-nt RNA2 genomic sequences were amplified and confirmed by Sanger sequencing (MCIC, Wooster, OH; GenBank accession numbers KT736088 and KT736089). This isolate showed 99-100 % nucleotide sequence identity to partially sequenced isolates from New York (AY016007.1, AY016008.1, AF361641.1 and AF361642.1), Illinois (HM133584.1, HM133583.1 and AB002812.1) and Germany (JX468081.1 and JX468082.1) but only 89-90 % identity to the type strain of SBWMV (Nebraska isolate; L07937.1 and L07938.1). Despite their low nucleotide sequence similarity, the Ohio and type SBWMV isolates have 100 % identical predicted CP amino acid sequences.

Full-length cDNA copies of SBWMV RNAs were cloned downstream of an enhanced cauliflower mosaic virus 35S promoter by replacing the 35S:hygromycin cassette in the original pCAMBIA1380 vector (AF234301.1) [11]. A 49-nt hammerhead ribozyme sequence derived from subterranean clover mottle virus satellite RNA (5’-GTCTGTACTTATATCAGTACACTGACGAGTCCCTAAAGGACGAAACGGG-3’) was inserted at the 3’ end of viral sequence to ensure proper transcriptional termination [5, 11]. Gibson assembly [9] was used for generating the different cDNA clones. E. coli (JM109; Promega) transformations of full-length RNA1-specific clones produced no colonies when maintained at 37 °C, but colonies grew on plates maintained at 28 °C after about 36 hours. Liquid cultures of RNA1 clones were also grown at 28 °C. cDNA clones derived from full-length RNA2 did not show any temperature-related effects. The infectivity of the cDNA clones was tested in N. benthamiana through agro-coinfiltrations by mixing equal volumes of agro-cultures (Agrobacterium tumefaciens strain C58C1; OD600 = 0.8) containing plasmids with SBWMV RNA1, RNA2 and a 35S-driven tomato bushy stunt virus P19 gene construct. P19, a well-characterized suppressor of RNA silencing that sequesters small interfering (si) RNAs, was used in all agro-coinfiltration experiments [16]. The replication and encapsidation competency of SBWMV cDNA clones were verified by Northern blot analysis and immunosorbent electron microscopy assays (data not shown). The upper non-inoculated leaves of agro-infiltrated N. benthamiana plants showed severe mosaic and leaf deformation symptoms approximately 14-20 days post-infiltration (dpi). Both RNAs were essential for producing symptoms in N. benthamiana and wheat, as was reported earlier in wheat and Chenopodium quinoa [13, 18]. Infiltrated leaf tissue extracts of N. benthamiana at 5-7 dpi were used as scaled up and encapsidated inoculum for transmission of the virus to wheat (line OH05-200-74, kindly provided by Dr. Eric Stockinger, Ohio State University) through mechanical rub-inoculations. At approximately 12 days post-rub-inoculation (dpr), newly emerging noninoculated wheat leaves showed virus symptoms. The results demonstrate the infectivity of our agro-based cDNA clone in N. benthamiana and wheat (Fig. 2A and B).

Fig. 2
figure 2

Assessment of SBWMV as a virus-induced gene silencing vector. (A) Systemic leaves of N. benthamiana at 55 dpi, inoculated with RNA1 and the labeled RNA2 cDNA clone. The PDS silencing phenotype observed in the fresh flush of 161 dpi R2∆RT_PDSNB is also shown. (B) Systemic leaf symptoms observed in wheat at 16 dpr with the same set cDNA clones. (C) Systemic leaves of wheat at 15 dpr with BSMV and BSMV_PDSTA400, respectively, showing wild-type disease symptoms and the PDS silencing phenotype. (D) RT-PCR analysis of PDS insertion clones in CPRT for insert stability. R2∆RT_PDSNB samples were collected at 9 dpi (infiltrated leaf) and 27, 35, and 55 dpi (systemic leaves) from N. benthamiana (i) and at 11, 19, 47 and 63 dpr symptomatic systemic leaves of wheat (ii). For R2ΔRT_PDSTA assay samples were collected at 1, 3, 5, 7 dpi (infiltrated leaf) and 14, 25, 55, 71 dpi (systemic leaves) from N. benthamiana and at 13, 18, 31 and 49 dpr from symptomatic systemic leaves of wheat (iii). RT-PCR products were separated on a 1 % agarose gel with a 1 Kb Plus DNA Ladder marker (Invitrogen). Control amplifications from RNA2, R2ΔRT_PDSTA and R2ΔRT cDNA clone templates and mock-inoculated sample RNAs are shown in each panel

Full size image

The bipartite rod-shaped morphology and taxonomic proximity of SBWMV to other viruses commonly used as heterologous gene expression or silencing vectors, such as tobacco rattle virus (TRV) and barley stripe mosaic virus (BSMV), prompted our interest in investigating its efficacy as a viral vector for monocotyledonous plants. Due to many challenges associated with viruses infecting monocotyledonous plants, very few are adapted for virus-induced gene silencing (VIGS). Barley stripe mosaic virus (BSMV) and brome mosaic virus (BMV) are relatively well characterized and successfully adapted for low-throughput virus vector applications, with BSMV useful in wheat, barley, oats and Brachypodium, and BMV can be used in maize, barley, rice and sorghum [6, 7, 15, 17]. We evaluated SBWMV as a viral vector for gene silencing and gene expression by insertion of green fluorescent protein (GFP) or plant phytoene desaturase (PDS) coding sequences at multiple sites in the viral RNA2 (Fig. 1B and 3A)

Fig. 3
figure 3

Assessment of SBWMV as a gene expression vector. (A) Schematic representation of RNA2-derived cDNA clones with GFP (720-nt) gene insertions. GFP gene insertions are shown as green boxes, and upstream F2A protease, MPce and CWMV CPce sequences are shown, respectively, as red, blue, and yellow boxes. (B) Infiltrated (6 dpi; row 1) and systemic (65 dpi; rows 2 and 3) leaves of N. benthamiana inoculated with RNA1 and the indicated RNA2 cDNA clone. Images were taken under UV (rows 1 and 2) or transmitted light (row 3) to observe GFP expression and virus infection. Inset: single-cell fluorescence observed in infiltrated leaves of R2∆RT_F2A-GFP-inoculated N. benthamiana through confocal microscopic imaging (color figure online)

Full size image

The occurrence of spontaneous deletions in the readthrough region of CPRT during prolonged infection or mechanical virus transmission yielding fully infectious, mechanically transmissible virus was previously reported in several studies on furoviruses [3, 18, 22]. Except for its putative role in viral transmission through soil-borne zoospores, CPRT is not essential for any vital biological functions such as virus replication, encapsidation and systemic infection [14, 21, 22], making the CPRT region an obvious choice for sequence insertions. Our R2∆RT RNA2 deletion mutant lacking 1603 nt of the 1734-nt CPRT (R2∆RT; deletion at nt 865-2467) was able to replicate and systemically infect N. benthamiana (Fig. 2A) and wheat (Fig. 2B). To test the utility of the CPRT region as an insertion site, two different PDS insertion clones were developed by replacing the 1603-nt CPRT fragment with insertion sequences. Partial PDS gene sequences derived from N. benthamiana (370 bp; nt 21-390 of AJ571700) and T. aestivum (409 bp; nt 1025-1433 of FJ517553) were inserted in the antisense orientation. Both of the CPRT site PDS insertion clones, R2∆RT_PDSNB and R2∆RT_PDSTA, caused systemic infection and induced visible photobleaching in the non-inoculated leaves of N. benthamiana (Fig. 2A). Photobleaching could still be observed on the fresh flush of N. benthamiana leaves five months after infiltration (161 dpi) with R2∆RT_PDSNB (Fig. 2A). R2∆RT_PDSNB- and R2∆RT_PDSTA-infiltrated leaf extracts showed systemic infection and symptoms when rub-inoculated to wheat. However, no PDS gene silencing phenotype was observed in wheat from either clone (Fig. 2B). In contrast, a barley stripe mosaic virus vector (pCa-γb:TaPDS400, kindly gifted by Dr. Dawei Li) [23] carrying a similar PDS insert sequence (nt 1029-1428 of FJ517553; 400 nt in the sense orientation) resulted in strong PDS silencing in wheat at 7-10 dpr (Fig. 2C). The 409-nt wheat PDS used in this study (nt 1025-1433 of FJ517553) is the same sequence plus nine additional nucleotides, inserted in the antisense orientation, suggesting that the target sequence selection is likely not limiting VIGS in wheat.

CPRT site PDS insertion clones were evaluated for insert stability using primers flanking the insertion site, SP203 (nt 1-27 of KT736089) and SP188 (nt 3224-3247 of KT736089), 837 and 757 nt apart from the insertion site, respectively. R2∆RT_PDSTA-inoculated N. benthamiana leaf samples were collected at 9 dpi from infiltrated leaves, and at 27, 35, 55, 71 dpi from systemically infected leaves. Systemically infected leaves from the corresponding rub-inoculated wheat plants were analyzed at 11, 19, 27, 47 and 63 dpr. A smear of products was observed from N. benthamiana samples (Fig. 2D i). Samples collected from wheat showed a stable insert up to 47 dpr (Fig. 2D ii). A similar insert stability analysis was conducted with R2∆RT_PDSNB-inoculated plants, with leaf samples collected from N. benthamiana at 1, 3, 5, 7 dpi (infiltrated leaves) and 14, 25, 55 dpi (systemically infected leaves) and from systemically infected wheat leaves at 13, 18, 31, 49 dpr (Fig. 2D iii). The results indicate that the inserted PDS sequence is relatively stable in infected wheat, despite the lack of a silencing phenotype. RT-PCR analysis with the wild-type virus clone yielded similar results (data not shown).

To test the utility of other sites for sequence insertions, additional clones were developed, replacing the CyR coding region alone (i.e., R2∆CyR) or in combination with CPRT deletion (i.e., R2∆RT&CyR). PDS sequence insertions in these sites and also between the CyR stop codon and the 3’ untranslated region (3’UTR) were assayed (Fig. 1B). None of these clones lacking CyR (R2∆RT&CyR, R2∆RT&CyR_PDSNB, R2∆RT&CyR_PDSTA, R2∆CyR, R2∆CyR_PDSNB, and R2∆CyR_PDSTA) or insertions downstream of CyR (R2-3’UTR_PDSNB, R2-3’UTR_PDSTA) showed systemic infection or photobleaching in either N. benthamiana or T. aestivum. The loss of systemic infection of CyR deletion clones supports the essentiality of the CyR, which has been reported to be a viral suppressor of host RNA silencing [4]. Loss of infectivity of 3’ UTR insertion constructs suggests that the 3’ UTR sequence configuration may also be essential.

In addition to VIGS capacity, the SBWMV clone was tested for gene expression capacity using GFP. CPRT and CyR insertion sites were evaluated for gene expression by inserting a 720-nt cycle 3 GFP [4] coding sequence. Five GFP gene insertion clones were developed: R2∆RT_F2A-GFP, R2∆RT_MPceGFP, R2∆CyR_GFP, R2-3’UTR_MPceGFP and R2-3’UTR_cwCPceGFP (Fig. 3A). In the R2∆RT_F2A-GFP clone, GFP was fused to the C-terminus of CP by replacing the stop codon with the 24-amino-acid F2A protease sequence for expression as a cleaved fusion protein [10]. Constructs with GFP under potential controller elements (ce) of homologous RNA1-MP or the heterologous Chinese wheat mosaic virus RNA2-CP were built using 125 nt upstream sequence of the indicated open reading frame, since the CyR sgRNA controller sequences of JSBWMV are mapped to 125 nt upstream sequence of the CyR start codon [22]. All five GFP clones showed GFP expression in agroinfiltrated leaves of N. benthamiana. The R2∆CyR_GFP clone in which the CyR coding sequence was replaced with GFP showed the strongest fluorescence of the five constructs. The R2∆RT_MPceGFP, R2-3’UTR_MPceGFP and R2-3’UTR_cwCPceGFP clones showed moderate GFP gene expression as foci of multiple cells in inoculated leaves. GFP expression from R2∆RT_F2A-GFP was predominantly restricted to single cells (Fig. 3B). Systemic infection in N. benthamiana was observed only with R2∆RT_MPceGFP, and no GFP expression was observed in the systemic leaves of this or any other construct. Rub inoculations of infiltrated leaf extracts into wheat did not result in symptomatic infection or GFP expression from any of the GFP clones.

The agro-based cDNA clone system developed for the OH isolate of SBWMV is a valuable addition to the currently available reverse genetic tools for this virus. Our results with SBWMV clones with insertions in CPRT showed the ability to induce silencing in the experimental dicotyledonous host N. benthamiana, but not in the natural monocotyledonous host T. aestivum, despite systemic movement and the remarkable stability of 370- to 409-nt PDS inserts. Insert stability is one of the problems associated with the currently available gene silencing vectors such as BSMV and BMV that are commonly used as gene silencing vectors in monocotyledonous plants [1, 2, 23]. This study indicates the importance and necessity of basic science research to understand potential differences in monocotyledonous virus hosts or specific host-virus combinations.