Barley Stripe Mosaic Virus-Mediated Tools for Investigating Gene Function in Cereal Plants and Their Pathogens: Virus-Induced Gene Silencing, Host-Mediated Gene Silencing, and Virus-Mediated Overexpression of Heterologous Protein

Barley stripe mosaic virus (BSMV), the Hordeivirus type member, is the most intensely studied virus that naturally infects two major monocot crops, wheat (Triticum aestivum) and barley (Hordeum vulgare). The tripartite genome of this virus, comprising RNAα, RNAβ, and RNAγ, was sequenced and cloned in the late 1980s. Since then, there have been significant advances in the understanding of the molecular biology of BSMV and the function of the seven main proteins encoded by its genome (Jackson et al., 2009; Fig. 1A).

Figure 1.

The BSMV genome and mechanistic models for BSMV-VIGS, BSMV-VOX, and BSMV-HIGS. A, The BSMV genome is composed of three RNAs that are capped at the 5′ end and form a tRNA-like hairpin secondary structure at the 3′ terminus. RNAα encodes the αa replicase protein containing methyl transferase (MT) and helicase (HEL) domains. RNAβ encodes βa (coat protein) and the βb, βc, and βd movement proteins (also known as TGB1, TGB3, and TGB2, respectively). RNAβ also encodes a minor protein, βd′ (alias TGB2′), which is expressed by translational read through of the βd open reading frame (ORF). RNAγ encodes γa, the polymerase (POL) component of replicase, and the Cys-rich γb protein involved in viral pathogenicity. B, For BSMV-VIGS, BSMV-VOX, and BSMV-HIGS, heterologous sequences of interest are typically inserted directly downstream of the stop codon of the γb ORF. For each experimental scenario, the modified RNAγ is mixed with RNAα and RNAβ (data not shown) and inoculated onto host plants. Bi, In BSMV-VIGS, a fragment from the plant gene of interest is cloned into BSMV RNAγ, usually in the antisense orientation, downstream of the stop codon of the γb ORF. After the virus enters a host plant cell, long dsRNAs formed during viral replication are recognized by host Dicer-like enzymes (DCLs) that cleave it into 21- to 22-nucleotide siRNAs. One strand of each siRNA is then incorporated into host RNA-induced silencing complexes (RISC). RISC mediate endonucleolytic cleavage of single-stranded RNAs with sequence complementary to the incorporated siRNA strand. As some of these siRNAs will have been generated from the plant gene fragment inserted in the BSMV genome, they will also guide RISC to plant mRNAs with sequence complementarity, resulting in silencing of the target endogenous gene(s). Bii, For BSMV-VOX, the coding sequence for the protein of interest is inserted immediately upstream of the in-frame stop codon of the γb ORF. In a new system, a small synthetic 2A gene encoding an autoproteolytic peptide has been inserted between the 3′ terminus of the γb ORF and the gene sequence coding for the heterologous protein. This enables self-processing (data not shown) of the ensuing γb fusion protein during translation of the virally encoded proteins, which occurs very soon after the entry of BSMV to the plant cell, thus releasing the free heterologous protein. Biii, In BSMV-HIGS, a fragment of a fungal gene of interest is inserted in the antisense orientation into RNAγ downstream of the stop codon of the γb ORF. RNA silencing signals generated within the plant cell can then trigger gene silencing in fungal cells in intimate contact with these modified host cells, such as fungal haustorial cells separated from plant cell membranes by only the extrahaustorial matrix (EHM). The mechanisms by which HIGS occurs are not yet known, but fungus-specific siRNAs generated by plant DCL activities are likely to be involved.

In a field situation, BSMV is easily transmitted from infected to healthy crops through plant-to-plant contact, and on most cultivars it causes mild to moderate mosaic symptoms. The virus is also known to be efficiently transmitted via the seed (Jackson et al., 2009). In the laboratory, BSMV can be mechanically transmitted by rub inoculation to maize (Zea mays), oat (Avena sativa), Brachypodium spp., and over 250 other species, mostly in the Poaceae (Jackson and Lane, 1981). However, among these are several dicot species, including Nicotiana benthamiana, in which BSMV causes mild systemic mosaic, and Chenopodium spp., a local lesion host.

In recent years, BSMV has become a popular vector for virus-induced gene silencing (VIGS) in barley and wheat (Scofield and Nelson, 2009; Cakir et al., 2010). This arose primarily because of the availability of full-length infectious BSMV clones and an increasingly detailed knowledge of molecular and biological functions of its various genome components (Fig. 1A). VIGS is a powerful functional genomics tool for the rapid targeted down-regulation of host plant genes (Becker and Lange, 2010; Senthil-Kumar and Mysore, 2011). It exploits the fact that infection of plants by viruses activates the posttranscriptional gene silencing defense response (Waterhouse et al., 2001). In VIGS, a short fragment of a transcribed sequence of a plant gene is inserted into a cloned virus genome, and the recombinant virus is then inoculated onto test plants (Fig. 1B). The introduced virus multiplies and spreads from the site of infection into newly developing regions of the plant and triggers posttranscriptional gene silencing. The inserted plant gene sequence also becomes the target for silencing, and so does the corresponding endogenous gene. This leads to a reduction or in some cases the complete abolition of target plant gene function, which in turn results in phenotype changes.

Bread wheat is an allohexaploid, containing three copies of each gene representing ancestrally distinct homeologous genomes A, B, and D. The BSMV-based VIGS system (hereafter referred to as BSMV-VIGS) is effective in simultaneous silencing of all functionally active homeologous gene copies in wheat. It can also target simultaneously at least two unrelated genes (Cakir and Scofield, 2008; Campbell and Huang, 2010). This increases efficiency and is highly cost effective. In these respects, BSMV-VIGS is similar to the stable transformation-based RNA interference (RNAi), a more commonly used technique for obtaining information about the functions of individual genes in cereals (Travella et al., 2006; McGinnis, 2010). However, the major advantage of BSMV-VIGS is that it is very rapid. In addition, it has a moderately high throughput compared with RNAi, which requires genetic transformation. An appropriately replicated single BSMV-VIGS experiment is usually completed in 5 to 6 weeks, and many different plant genotypes can be tested simultaneously. By comparison, stable cereal transformation requires a dedicated facility, well-trained staff, and is also relatively slow. At least 8 months are required to produce the primary transgenic barley or wheat lines (T0 generation), usually in just a single genetic background. Stable transformation is only low throughput, and true breeding RNAi lines are not available until at least the T3 generation.

Due to its speed and moderately high throughput, BSMV-VIGS allows rapid prescreening of candidate genes prior to the use of other, more time-consuming techniques to assess gene function, such as the stable RNAi transformation or targeting induced local lesions in genomes (Comai and Henikoff, 2006).

With food security becoming one of the key strategic research priorities globally, there has been a recent deluge of available sequence information for the transcriptomes and genomes of major cereal crop species, including rice (Oryza sativa), wheat, barley, and maize, and a model grass species, Brachypodium distachyon (Berkman et al., 2012). The importance of BSMV-VIGS as a research tool for all aspects of cereal plant research in conjunction with comparative genomics, computational analyses (bioinformatics), and other currently available reverse genetic tools cannot be overestimated.

BSMV-VIGS already has an excellent track record for unraveling the function of leaf-expressed genes in barley and wheat at the vegetative (seedling) stage of development. This research tool has been used almost exclusively by plant pathologists for determining the function of plant defense signaling genes in multiple host-pathogen interactions. Since the two most recent comprehensive reviews were written (Scofield and Nelson, 2009; Cakir et al., 2010), there have been significant advances in the development and deployment of BSMV-VIGS in several additional monocot species. Furthermore, two new applications of the BSMV vector have emerged. The first, named host-induced gene silencing (HIGS; Nowara et al., 2010), permits transspecies silencing and has been successfully used for analyzing the functions of genes in plant pathogenic fungi that are expressed during plant infection. The second enables heterologous protein overexpression in planta and has been coined virus-mediated overexpression (VOX; R. Wise, personal communication). Moreover, there has recently been a substantial interest in the adoption of BSMV-derived tools by scientists in other plant disciplines. Here, we review all these exciting, very recently published and unpublished (presented and discussed at the International Cereals VIGS Workshop [ICVW], Rothamsted Research, June 22–24, 2011) advances and developments.

ADVANCES IN BSMV-INDUCED PLANT GENE SILENCING

Expansion of the Host Range for BSMV-VIGS

The success of BSMV-VIGS as a tool for functional genomic studies in barley and bread wheat has led to its adoption and deployment in a number of other monocots. Silencing of phytoene desaturase (PDS), a gene commonly used as a visual marker in VIGS experiments due to the resulting characteristic photobleached phenotype (Fig. 2A), has been demonstrated in B. distachyon (Demircan and Akkaya, 2010; Pacak et al., 2010; Yuan et al., 2011). In Haynaldia villosa, a wild grass, BSMV-VIGS has already been used for the functional analysis of powdery mildew resistance-related genes (Wang et al., 2010). Other desirable traits of interest in these species, including multiple disease resistance, strong tillering ability, and high protein seed content, can be tested and then exploited by introgression breeding in commercial cereal crops.

Figure 2.

Silencing of PDS in wheat leaves, ears, and in the second plant generation. A, Upper uninoculated leaves taken from plants infected with either BSMV:asGFP control (top) or BSMV:asPDS showing the resulting photobleached phenotype (bottom) at 28 d post inoculation. B to D, Inoculating the flag and/or the penultimate leaf of preheading wheat plants can induce silencing in the ears. The control BSMV:asGFP-infected ears appear to develop normally (B). Silencing of PDS in the ears (C and D) results in visible photobleaching of the glumes, awns, and rachis. Photographs were taken at 28 d post inoculation. E and F, Inoculation of lower leaves of younger wheat ‘Apogee’ plants at the three- to five-leaf stage often induces silencing in the flag leaf and ear. The control BSMV:asGFP-infected plants appear to develop normally (E), whereas the BSMV:asPDS inoculations (F) induce almost complete photobleaching in the flag leaf and partial photobleaching in the ear. The plants were photographed at 28 d post inoculation. G and H, Grain from uninfected plants (G) and the BSMV:asPDS-inoculated plants (H) are comparable in size and appearance. I and J, A proportion (10%–15%) of the progeny from BSMV:asPDS-infected plants also display partial photobleaching of the leaves within 12 d of seed germination.

BSMV-VIGS is currently being optimized for other valuable cereals such as millet (Setaria italic), maize, and several members of the genus Triticum (D. Li, personal communication), although similar experiments with oat have so far met with only limited success (Pacak et al., 2010). There has been little work as yet with noncereal monocots, although a study in culinary ginger (Zingiber officinale) has successfully targeted PDS for silencing, thereby demonstrating the potential for the extension of BSMV-VIGS to a wider range of monocot hosts (Renner et al., 2009).

Application of BSMV-VIGS to Disciplines beyond Plant Pathology

The BSMV-VIGS tool has been used to investigate the involvement of the WRKY53 transcription factor and the phenylpropanoid pathway in resistance to aphids in wheat (Van Eck et al., 2010). This was a natural first step because plant virologists frequently collaborate with entomologists to study insect virus vectors. A number of recent studies have used BSMV-VIGS to explore gene function in other areas of plant science. In one study, silencing of the monocot-specific P23k gene in barley resulted in morphological abnormalities such as strong asymmetry and cracks along the leaf margins, confirming the involvement of this gene in secondary cell wall biosynthesis (Oikawa et al., 2007). BSMV-VIGS was also utilized to investigate the effect of silencing the barley cellulose synthase CesA (Held et al., 2008), another gene predicted to be involved in cell wall biosynthesis. Interestingly, while the gene-silencing construct was designed to specifically target the cellulose synthase gene family, it also triggered the down-regulation of several homologous cellulose synthase-like genes, possibly due to transitive silencing induced by secondary small interfering RNAs (siRNAs). A third study (Pacak et al., 2010) demonstrated BSMV-induced silencing of three barley homologs of the Arabidopsis (Arabidopsis thaliana) genes involved in the regulation of inorganic phosphate uptake and translocation. Silencing of one of these genes, under conditions of abundantly supplied inorganic phosphate, resulted in excessive inorganic phosphate accumulation in the leaf tissues compared with that in control barley plants. This silencing phenotype is reminiscent of the Arabidopsis pho2 phosphate-accumulator mutant (Delhaize and Randall, 1995), indicating that PHO2 is functionally conserved in barley and Arabidopsis.

Silencing beyond the Leaf

BSMV is readily transmitted through seed (Jackson et al., 2009) and also invades the root system (Lin and Langenberg, 1984). Therefore, due to the biology of this virus, VIGS in flowering tissues, seeds, and roots should be possible. The first report on the successful deployment of BSMV-VIGS in wheat roots has just been published (Bennypaul et al., 2012). In that work, the coronatine insensitive1 gene was targeted, and the silencing (85% decrease in transcript level) resulted in a 26% reduction in root length.

At the ICVW, several teams, including our own, already employing BSMV-VIGS in their research convincingly reported the successful silencing of target genes in floral tissue in a range of different wheat cultivars following inoculation of upper leaves at preheading (Fig. 2, B–D). A number of these teams are interested in identifying host genes involved in susceptibility to Fusarium species that cause Fusarium ear blight (also known as Fusarium head blight or scab), the globally important disease of wheat and barley that leads to deoxynivalenol mycotoxin-contaminated grain at harvest (Dean et al., 2012). In addition, Kulvinder Gill reported the successful deployment of BSMV-VIGS for analyzing the functions of genes expressed in developing wheat grain and meiotic tissues. Silencing of the granule-bound starch synthase (or waxy) gene resulted in substantial reductions in the amylose content in seeds, whereas silencing of disrupted meiosis cDNA1, a gene involved in early meiosis in pollen mother cells, led to the disruption of chromosome pairing at metaphase I. Results of this pioneering study have since been published (Bennypaul et al., 2012). Gene silencing in floral and reproductive tissues and the grain of wheat has been achieved by applying the corresponding BSMV-VIGS constructs to older plants with the fully expanded flag leaf. We and others (P. Schweizer, personal communication) have recently discovered that this can also be achieved in rapid cycling wheat genotypes such as USU-Apogee (Bugbee, 1999) by BSMV vector delivery to young plants (three- to five-leaf stage; Fig. 2, E and F).

Silencing in the Next Plant Generations

BSMV is able to invade the reproductive tissues prior to fertilization, and the spread and survival of this virus in nature is due to its ability to be transmitted through seed in barley and wheat (Jackson et al., 2009). The inheritance of BSMV-induced gene silencing in the progeny of infected plants was initially demonstrated for barley (Bruun-Rasmussen et al., 2007). This same phenomenon was reported for wheat at the ICVW by Kulvinder Gill. This opens up the possibility of using VIGS to target genes that are expressed during seed dormancy, germination, and early seedling development. The presence of BSMV appears not to impair seed development and/or appearance (Fig. 2, G and H). Recent work from both the Kulvinder Gill laboratory (Bennypaul et al., 2012) and our laboratory on the silencing of PDS indicate that approximately 10% to 15% of progeny from the inoculated wheat plants display photobleaching and therefore retain silencing (Fig. 2, I and J). The efficiency of silencing in next generations is highly influenced by the genotype (Bennypaul et al., 2012; W.-S. Lee, unpublished data). All the next-generation plants showing photobleaching contain BSMV with either the original or shortened PDS-specific insert (Bruun-Rasmussen et al., 2007; W.-S. Lee, unpublished data). The silencing phenotype in the next-generation plants, therefore, is likely to be initiated by the recombinant BSMV being carried within the seed rather than involving an epigenetics-based mechanism. Interestingly, the proportion of plants showing PDS gene silencing increases with each subsequent generation of self-pollinated plants, and up to 90% to 100% of the third-generation plants display photobleaching (Bennypaul et al., 2012). Again, striking differences between wheat genotypes in exhibiting photobleaching in later generations were noted.

Vector Improvements

Several variations have been made to the original BSMV-VIGS system. Each variation has important implications for the throughput and ease of use of this research tool. The originally described system used BSMV complementary DNA clones under the control of the strong bacteriophage T7 promoter. Capped transcripts for the three BSMV genomes, RNAα, RNAβ, and RNAγ, are produced in vitro, subsequently mixed together, and then rub inoculated onto individual host plants. Fragments of target plant genes are introduced into the BSMV RNAγ genome vector via restriction digestion with multiple restriction enzymes and ligation-based cloning (Holzberg et al., 2002).

In vitro transcription is costly and prompted innovation. One variant system from the Roger Wise laboratory avoids this expense by replacing the T7 promoter with the 35S promoter from Cauliflower mosaic virus and introducing a ribozyme sequence downstream of each viral complementary DNA to generate the correct 3′ end after transcription. These modifications allow in planta transcription of the DNA plasmids following their delivery into barley leaf tissues using microprojectile particle bombardment (Meng et al., 2009). Sap extracted from the transfected barley leaves containing reconstituted virus can then be used to infect more plants for VIGS studies.

Another variant of the BSMV-VIGS vector retains the T7 promoter but replaces the original cloning site immediately 3′ to the coding region in the BSMV RNAγ genome with a ligation-independent cloning site in order to facilitate efficient insertion of target gene fragments (Pacak et al., 2010).

The latest report unites the advantages of both systems by coupling a ligation-independent cloning strategy for cloning target gene fragment inserts with an Agrobacterium tumefaciens-mediated delivery system (Yuan et al., 2011). Here, BSMV genomes cloned into a binary Ti vector under the control of the Cauliflower mosaic virus 35S promoter are first delivered via agroinfiltration into the leaves of N. benthamiana, an intermediate host susceptible to both A. tumefaciens and BSMV. Sap is then extracted from the infiltrated leaves and used to inoculate a large number of monocotyledonous plants. Since this system was first described at the ICVW, there has been significant interest in its uptake by members of the cereal VIGS community due to its relative simplicity and low setup costs.

BSMV FOR GENE SILENCING IN PLANT-ASSOCIATED ORGANISMS

An emerging and exciting new technical advancement is HIGS. In HIGS, silencing that is triggered directly within the plant targets genes in plant pathogens or pests (Fig. 1Biii). The triggers, as in VIGS, are short double-stranded RNAs (dsRNAs). However, in the case of HIGS, the dsRNAs target pathogen or pest transcripts rather than plant transcripts. Silencing-inducing dsRNAs are produced in the host plant cells following one of three delivery methods: (1) microprojectile particle bombardment of RNAi constructs into plant leaves, (2) stable plant transformation with RNAi constructs, or (3) using a virus vector such as BSMV (Nunes and Dean, 2012). Plants either transiently or stably expressing the dsRNA triggers are then inoculated with the relevant organism, and the role of a target pathogen or pest gene during the infection can be studied.

Very recently, HIGS has been successfully applied to the investigation of molecular interactions between obligate biotrophic fungi and cereal crops. In one study (Nowara et al., 2010), both stable and transient methods of delivery of silencing constructs have been explored. First, 76 RNAi constructs, each targeting one of the genes of the barley powdery mildew fungus Blumeria graminis f. sp. hordei known to be expressed in planta during infection, were bombarded into the individual epidermal cells of barley leaves. The ability of the fungus to form a haustorium, a specialized postinfection feeding structure required for absorbing nutrients from the plant, within these epidermal host cells was then analyzed. Remarkably, nearly one-quarter of the tested RNAi constructs induced significant reductions of haustorium formation. The functions of two fungal genes, GTF1 and GTF2, targeted by these constructs were further investigated in transgenic RNAi barley lines (GTF1 only) and also using BSMV-HIGS. Both approaches proved to be successful and revealed different roles for these two genes in fungal development, with GTF1 likely to be involved in initial haustorium formation and GTF2 in the elongation of secondary hyphae, necessary for colony formation. In the second study (Yin et al., 2011), BSMV-HIGS was used to silence 12 genes of another obligate biotroph, the stripe rust fungus Puccinia striiformis f. sp. tritici. The genes selected were known to be expressed during infection of wheat leaves. In these experiments, silencing was only observed for genes that were preferentially highly expressed in haustoria and not in the bulk of the fungal mycelial colony. For both species, fungal haustoria penetrate mesophyll cells in host leaf tissue and represent the interface for signal exchange between the fungus and the invaded plant. Presumably, RNA silencing was most efficient in fungal cells in very close proximity to plant cellular cytoplasm and less so in fungal hyphae, which have an extracellular localization.

The mechanism(s) by which RNA-silencing signals are delivered from the plant cell into fungal cells is currently a mystery. For obligate biotrophic fungal pathogens like barley powdery mildew, it has been hypothesized that these signals are transported from the plant cell into the haustorial cells (Nowara et al., 2010). However, these fungal cells are separated from the host cells by the fungal cell wall and plasma membrane as well as by the extrahaustorial matrix and extrahaustorial membrane derived from the plant plasma membrane (Panstruga and Dodds, 2009; Fig. 1Biii). Therefore, the silencing signal (which most probably is a short dsRNA or single-stranded RNA) needs to cross multiple physical barriers in order to initiate HIGS. One of the suggested transport routes is the exosomal secretory pathway (Nowara et al., 2010). Clearly, further research is required to answer this intriguing question.

HIGS of fungal genes in stably transformed plant RNAi lines in the arbuscular mycorrhizal fungus Glomus intraradices-Medicago truncatula interaction, where the fungal and host cellular membranes are again in close proximity, has also been demonstrated successfully (Helber et al., 2011). Recently, another study demonstrated that expression of a GUS-derived dsRNA hairpin in tobacco (Nicotiana tabacum) could trigger specific silencing of this gene in the mycelia of a transgenic GUS-expressing strain of Fusarium verticillioides during infection (Tinoco et al., 2010). This result is particularly important because F. verticillioides growth is exclusively extracellular, and this fungus is not known to produce an intracellular feeding structure.

Within the past two years, a BSMV-based HIGS system has emerged as a viable alternative, or at least a complementary approach, to that of generating numerous fungal gene knockouts to investigate the role of fungal genes in plant-fungus interactions (Nunes and Dean, 2012). This method is particularly useful for studies of important plant pathogens and symbionts for which stable transformation methods are not yet available or for obligate plant-associated organisms that cannot be cultured in vitro.

BSMV-VOX

VOX using the BSMV vector was first demonstrated in studies involving the GFP reporter and barley (Haupt et al., 2001; Lawrence and Jackson, 2001). Although the overall levels of GFP expression from BSMV was shown to be strong, expression was often reported to be patchy in the systemically infected vegetative tissue of monocotyledonous plant hosts. While Lawrence and Jackson (2001) initially suggested that this may be due to BSMV exiting the vasculature at more than one location in systemically infected leaves, forming multiple infection loci, it is probably also due in part to the relatively large size of GFP (720 nucleotides) and, subsequently, the instability of the GFP coding sequence in the viral genome.

The community consensus is that larger fragments appear to be lost from the BSMV vector more frequently than smaller fragments (Bruun-Rasmussen et al., 2007). Mutants with a reduced insert size stand higher chances of establishing systemic infection because they are generally more fit (i.e. they replicate faster and/or move within the infected plant faster) than their originally sized recombinant counterparts. As reported by different authors, inserts within the range of 140 to 500 bp when integrated into the RNAγ genome of BSMV are relatively more stable than larger inserts (Holzberg et al., 2002; Scofield and Nelson, 2009). Indeed, a flavin-based fluorescent reporter protein, iLOV, that is approximately half the size of GFP (Chapman et al., 2008) appears to be more stably and more uniformly expressed compared with GFP when delivered to wheat and barley leaves using BSMV vector (Fig. 3A; K. Kanyuka, unpublished data).

Figure 3.

BSMV-mediated overexpression of small heterologous proteins. A, A fragment of wheat leaf infected with recombinant BSMV expressing a fluorescent reporter protein, iLOV. The image was obtained using confocal laser microscopy. Plant cell walls are colored purple and chloroplasts are in blue, whereas cells containing iLOV fluoresce green. B, Recombinant BSMV expressing full-length Nip1 protein of R. commune induces cell death on leaves of barley ‘Atlas 46’ expressing the cognate resistance gene Rrs1. C, Recombinant BSMV expressing full-length Nip1 protein of R. commune induces cell death on leaves of barley ‘Atlas 46’ containing the cognate resistance gene Rrs1 but not on a nearly isogenic Atlas line devoid of Rrs1 (top). Truncated Nip1, with the N-terminal signal peptide removed (no SP), is unable to induce cell death on barley (bottom). All photographs were taken at 7 d post inoculation.

Despite this significant size constraint on BSMV-VOX, it is already a viable strategy for investigating the function of small proteins, such as those predicted to be fungal effectors. The majority of known fungal effectors are small (i.e. comparable in size with iLOV), usually Cys-rich secreted proteins containing no conserved functional domains (Stergiopoulos and de Wit, 2009). Confirmed fungal effector proteins are of particular interest, because for some their production in planta is known to be essential for successful pathogen invasion and colonization, while others are the cues detected by plants that lead to the triggering of resistance (R) gene-mediated defense (Jones and Dangl, 2006). BSMV-VOX has already been used to express the ToxA effector protein from Pyrenophora tritici-repentis, the causal agent of tan spot disease in wheat (Tai et al., 2007; Manning et al., 2010). Previously, it was known that ToxA produced in vitro and purified from the culture filtrate induces necrosis on wheat cultivars susceptible to the disease when infiltrated into the extracellular spaces of the leaf (Ballance et al., 1989).

Some fungal effectors may be expected to either suppress plant innate immunity during early invasion or stimulate cell death or the induction of necrosis at later stages of disease development. BSMV-VOX has a major advantage over microprojectile particle bombardment, another methodology for the delivery of constructs for transient protein expression in plant leaf cells, in that it avoids tissue damage caused by bombardment that could be confused with effector-induced cell death. The ease of inserting protein-coding sequences into the BSMV vector also means that this approach could potentially be used in a forward genetic screen to characterize the large numbers of putative effectors now predicted from interrogation of the recently sequenced fungal genomes, such as Blumeria graminis (Godfrey et al., 2010) and Fusarium graminearum (Brown et al., 2012).

In all BSMV-VOX studies to date involving cereal-infecting fungal species, heterologous proteins were expressed as direct C-terminal fusions to the viral γb protein. This approach, at least in some cases, may be expected to compromise functionality or to affect the cellular localization pattern of the expressed protein and/or the viral γb protein. An alternative and more preferable approach involves the use of self-processing peptide bridges, such as the 18-amino acid-long catalytic autoproteolytic 2A peptide sequence of picornaviruses (El Amrani et al., 2004), between the fused proteins. This approach has been successfully used in the past for the expression of GFP as an N-terminal fusion to the BSMV γb protein (Torrance et al., 2006). We modified the BSMV RNAγ vector by inserting a synthetic 2A gene at the 3′ terminus of the viral γb for the production of C-terminal protein fusions (Fig. 1Bii). This, in theory, should allow cotranslational self-processing of the fusion protein, resulting in some free heterologous protein, and also permit the expression of heterologous proteins with their native N termini (e.g. proteins such as fungal effectors containing N-terminal signal peptides) for secretion into the plant apoplast. Our preliminary experiments on BSMV-VOX of NIP1, the necrosis-inducing small secreted effector protein from Rhynchosporium commune (previously known as Rhynchosporium secalis) that operates in the plant apoplast (Wevelsiep et al., 1991), confirm the above theory, albeit indirectly. R. commune NIP1 is known to elicit defense reactions specifically in barley genotypes carrying the cognate Rrs1 resistance gene (Rohe et al., 1995). Indeed, expression of the full-length NIP1 using a modified BSMV as explained above resulted in genotype-specific Rrs1-dependent cell death. By contrast, a truncated version of NIP1 (with the signal peptide removed) incapable of being secreted failed to induce cell death on barley genotypes with or without Rrs1 (Fig. 3, B and C; K. Kanyuka, unpublished data).

Recently, there has been a preliminary (conference) report of a new four-component BSMV vector system that potentially enables much larger proteins to be stably expressed in planta (R. Brueggeman, personal communication). This system is outlined in Figure 4. If this system is indeed proven to work, it has major implications for the use of BSMV-VOX in gene-function studies. For example, R gene candidates identified in mapping studies could be expressed directly in susceptible genotypes, and genes from major resistance quantitative trait loci could be screened and identified. This type of vector would also be highly applicable to gene-function investigations through overexpression for a wide range of predicted cereal proteins.

Figure 4.

A new, four-component BSMV vector system for the expression of larger sized heterologous proteins. The system comprises the wild-type BSMV RNAα and RNAβ and two differently modified RNAγ (RNAγ1 and RNAγ2). RNAγ1 has retained the functional γa replicase gene, but a large deletion in the γb coding region renders this gene nonfunctional. The almost complete deletion of the γb gene is required to ensure the retention of, and heterologous protein expression from, the recombinant RNAγ2. The latter has a large deletion in the γa coding region, which allows sequences of up to 2 kb to be inserted into this region, while sequences downstream of γa including the γb gene have been left intact. During plant infection, it is anticipated that the intact γa coding region and γb coding region expressed from the two differently modified RNAγ will be able to function in this “trans”-configuration and thereby generate infectious BSMV expressing larger sized, potentially up to approximately 65 kD, heterologous proteins of interest.

FINAL REMARKS AND NEW OPPORTUNITIES

The advent of the VIGS technology has revolutionized gene-function discovery in many dicots and has now become a powerful tool for plant science research. For monocots, VIGS is not yet center stage. However, recent significant improvements and developments in BSMV-VIGS, which include (1) marked reduction in cost, (2) expanded host range, (3) ability to silence not only leaf genes but also those expressed in root, floral tissues, and the grain, and (4) maintenance of silencing in the next generation of plants, will certainly make this research tool more attractive to plant scientists and present an exciting opportunity for new researchers entering the field. For example, in the near future, we expect to see BSMV-VIGS commonly used in plant research for rapid prescreening/selecting candidate genes from various gene cloning projects. BSMV-VIGS will also be very useful in a rapidly growing number of genomics projects aiming to translate knowledge of gene function from model plants to crops, such as wheat and barley. In coming years, we also foresee expansion of the usage of BSMV as a vector for host-induced silencing of genes in various cereal-infecting plant pathogenic species, especially for obligate biotrophs that cannot be grown in culture and other species that are not yet amenable to stable transformation. Effector biology of plant-associated organisms is a dynamic, new, and exciting research area. No doubt, BSMV-VOX will play an important role in the identification and functional characterization of small effector proteins from plant-associated organisms. Next-generation derivatives of BSMV vectors allowing stable expression of larger proteins are rumored to become available in the near future. When this happens, it will open an enormous number of new possibilities for functional protein analyses through overexpression in planta.

Glossary

BSMV

Barley stripe mosaic virus

VIGS

virus-induced gene silencing

RNAi

RNA interference

HIGS

host-induced gene silencing

VOX

virus-mediated overexpression

ICVW

International Cereals VIGS Workshop

dsRNA

double-stranded RNA

ORF

open reading frame

siRNA

small interfering RNA

RISC

RNA-induced silencing complexes