Engineering geminivirus resistance in Jatropha curcus

Background Jatropha curcus is a good candidate plant for biodiesel production in tropical and subtropical regions. However, J. curcus is susceptible to the geminivirus Indian cassava mosaic virus (ICMV), and frequent viral disease outbreaks severely limit productivity. Therefore the development of J. curcus to carry on durable virus resistance remains crucial and poses a major biotechnological challenge. Results We generated transgenic J. curcus plants expressing a hairpin, double-stranded (ds) RNA with sequences homologous to five key genes of ICMV-Dha strain DNA-A, which silences sequence-related viral genes thereby conferring ICMV resistance. Two rounds of virus inoculation were conducted via vacuum infiltration of ICMV-Dha. The durability and heritability of resistance conferred by the dsRNA was further tested to ascertain that T1 progeny transgenic plants were resistant to the ICMV-SG strain, which shared 94.5% nucleotides identity with the ICMV-Dha strain. Quantitative PCR analysis showed that resistant transgenic lines had no detectable virus. Conclusions In this study we developed transgenic J. curcus plants to include a resistance to prevailing geminiviruses in Asia. These virus-resistant transgenic J. curcus plants can be used in various Jatropha breeding programs. Electronic supplementary material The online version of this article (doi:10.1186/s13068-014-0149-z) contains supplementary material, which is available to authorized users.


Background
Jatropha curcas, a small woody plant belonging to the Euphorbiaceae family, is a non-food oil seed crop mainly grown in the tropical and subtropical regions. This plant possesses several traits render this woody plant suitable for biodiesel feedstock production. It is easy to propagate and grows rapidly. It has a short gestation period, low seed cost and high oil content. Moreover, the ability of J. curcus to thrive on degraded soil and its wide adaptability to different growth conditions allows the use of marginal or non-arable wasteland for the application of this plant on an industrial scale. However, the productivity of J. curcus in the field is limited by the occurrence of Jatropha curcas mosaic disease (JcMD) [1][2][3]. The disease incidence is particularly significant in the Indian subcontinent; about 25% in northern India [1] and up to 47% in southern India [2].
We have previously reported the first full-length genome sequence of one geminivirus, a strain of Indian cassava mosaic virus (ICMV-Dha), as the causative pathogen of JcMD found in Southern India [3]. Following our report, three other related geminiviruses were isolated from Jatropha plants in Africa and South Asia [4][5][6]. Recently, we reported another highly pathogenic ICMV Singapore strain as the causative agent for JcMD in South-east Asia which shares a 94.5% identity with ICMV-Dha [7]. The recurrent identification of ICMV as an epidemic viral pathogen in various Jatropha plantations prompted us to investigate the biology of the virus in more detail.
Geminiviruses, which are single-stranded DNA viruses infecting a range of economically important crop species (such as cassava, maize, cotton and tomato) in tropical and subtropical regions, have become a major threat to world agriculture in the past decade [8]. Based on genome organization, insect vector and host range, the family Geminiviridae can be classified into four genera: Begomovirus, Mastrevirus, Curtovirus and Topocuvirus. So far, all five Jatropha viral pathogens belong to one genera: Begomovirus. Most of these viruses contain two genomic components termed DNA-A and DNA-B (approximately 2.7 to 3.0 kb) and they are all exclusively transmitted via the whitefly, Bemisia tabaci. The virus DNA-A-positive strand encodes the coat protein (CP/AV1) involved in the encapsidation of viral DNA, virus movement and viral transmission by B. tabaci). Among other encoded proteins, the replication associated-protein (Rep) AC1 is absolutely required for the replication of both genomic components. The transcriptional activator protein (TrAP) AC2 is needed for transcriptional activation of viral gene transcription and plant host gene expression. The replication enhancer protein (Ren) AC3 greatly enhances viral DNA accumulation by interacting with Rep/AC1. Another viral protein, AC4, acts as a gene-silencing suppressor to compromise the host defense system. All these five genes are essential for the virus life cycle and pathogenesis [8].
Because of the capacity of geminiviruses to evolve rapidly by mutation, recombination and pseudo-recombination, the development of plants with durable virus resistance continues to be a major challenge. One strategy involves genetic crossing of resistant and susceptible Jatropha germplasms. This strategy has the advantage that segregation patterns can be clearly observed between resistant and susceptible lines [9]. However, germplasm-mediated resistance via crossbreeding is time-consuming and requires a large number of progeny plants (large-scale field tests) to ascertain segregation patterns in future generations [9]. Therefore, transgenic technology has been considered as the method of choice for improving the virus resistance of J. curcus. Recently, we and other groups have established transformation platforms which facilitate the transfer of foreign genes into the J. curcus genome [10][11][12][13], and we have used this method to produce virus-resistant transgenic J. curcus.
A major strategy to produce transgenic plants with virus resistance is based on the concept of pathogenderived resistance (PDR) in which the transgene is derived from viral sequences. The mechanism of PDR includes protein-mediated resistance and RNA interference (RNAi). Both mechanisms have been shown to confer geminivirus resistance in transgenic tomato, common bean, cassava and others [14][15][16][17][18]. Here, we report the production of several JcMD-resistant J. curcus lines by expressing a hairpin double-stranded (ds) RNA targeting five key geminivirus DNA-A genes. Some of the transgenic lines displayed broad resistance to related geminiviruses, with 94.5% nucleotide identity in the transgene sequences.

Hairpin, double-stranded RNA construct
We have previously identified the causal pathogen for the JcMD in Southern India as the strain of ICMV known as ICMV-Dha [3]. We chose the sequences of this ICMV-Dha strain to engineer virus resistance via RNAi technology. Three viral gene fragments were ligated to generate the sense and antisense arms in the hairpin dsRNA. Fragment 1 (250 bp) targets the gene encoding CP/AV1 and the AC5 gene, fragment 2 (250 bp) targets genes for TrAP/AC2 and Ren/AC3 and fragment 3 (609 bp) targets genes for Rep/AC1 and AC4 ( Figure 1A). The ligated fragment (fragment 1, 2 and 3) with the designated orientation as indicated by an arrow has the potential to generate a hairpin dsRNA structure with an intron ( Figure 1B). The siRNA pool, produced from the hairpin (hp) RNA, should have the potential to silence five key viral genes encoding AC1 (Rep, Rep Replication associated protein), AC2 (TrAP, transcriptional activator protein), AC3 (Ren, (Replication enhancer protein), AV1 (Coat protein, CP) and AC4. We placed this hpRNA-encoding DNA fragment into a chemicalinducible marker excision vector which has been shown to function in J. curcus [13].
We then replaced the original G10-90 promoter with a Cauliflower mosaic virus (CaMV) 35S promoter with double enhancers and named it pX9-hpICMV RNAi. Upon chemical induction with 17β-estradiol, Cre/lox P mediated recombination excised the DNA fragment containing the hygromycin-resistance gene. This recombination event resulted in placing the hpRNA-encoding DNA fragment immediately downstream of the 35S promoter [13]. The PCR product using primer 1 and 2 should be approximately 6 kb before chemical induction and approximately 1.2 kb after induction. However, the approximately 6 kb product could not be amplified using our PCR program. The absence of PCR products using primer 3 and 4 indicated the excision of the hygromycin phosphotransferase (Hpt) gene ( Figure 1C).

Plant transformation and virus inoculation
The pX9-hpICMV RNAi vector was transformed into Agrobacterium and a total of 133 T0 transgenic plants were generated. We performed PCR analysis on genomic DNA using two pairs of primers (P1-P2 and P3-1 together with P4-1, Figure 1C). A total of 53 out of 133 plants showed a 1.2 kb product using primers P1 and P2, indicating the occurrence of successful marker excision. However, the presence of PCR products of P3-1 and P4-1 corresponding to the HPT gene suggested that all 53 transgenic plants were chimeric. Figure 2A shows the results of transgenic lines 1 to 54. We purified the PCR products from the P1-P2 primer of line 13,16,19,25,31,38,50, 52 and cloned into T-vector. DNA sequencing of these products confirmed the positive bands were true marker-free products (data not shown). Transgenic plants which were positive for P1-P2 products were used for virus challenge by vacuum infiltration with Agrobacterium-mediated virus infection ( Figure 1D).
Two and a half months after the ICMV-Dha virus challenge, inoculated WT and susceptible transgenic plants showed very obvious curling and a mosaic phenotype on systemic leaves. We extracted genomic DNA and performed quantitative PCR to investigate virus titers in transgenic and WT plants. Uninfected WT plants were used as controls, and the average value of these control plants was set as 1. There was no cycle threshold (Ct) value difference between symptomless resistant transgenic plants and uninfected control plants. In total, 15 independent transgenic lines presented with the symptomless phenotype and no virus could be detected in these resistant transgenic plants. However, we detected very high virus titers in susceptible transgenic plants, which were similar to those found in infected WT plants ( Figure 3). Six months after virus inoculation, the susceptible transgenic plants became obviously stunted with malformations and plant size reduction, whereas the resistant transgenic plants grew normally ( Figure 4). The virus titers correlated very well with the severity of virus symptoms. A much lower virus amount was detected in plants with mild symptoms (for example line 65) compared with plants with strong symptoms (for example line 8) ( Figure 3 and Figure 4D, E). We collected samples from the resistant T0 plants and performed PCR analysis to confirm the occurrence of a marker excision event ( Figure 2B).
Next, we performed Southern blotting to examine the transgene copy number in resistant lines. As there are two EcoRV sites around the two loxP sites, the genomic DNAs were digested with EcoRV ( Figure 5A). Southern blots using probe prepared from the 35S promoter detected the copy number of the transgene irrespective of the marker status. Figure 5B shows that resistant T0 plant lines 7, 20, 25, 50, 82, 113, 131 and 133 contained a single copy of the transgene. Along with quantitative PCR (qPCR) analysis of viral titers and symptomatic observations, we considered these transgenic plants to be candidate-positive resistant plants.

Heritable virus resistance in T1 plants
We chose number 82 as the first candidate for further characterization as two line 82 T0 plants (regenerated from a single transgenic event and confirmed by thermal asymmetric interlaced (TAIL)-PCR sequencing) showed a consistent virus resistance trait ( Figure 4C). T1 seeds derived from T0 plant line 82 were germinated along    Figure 6A). qPCR analysis showed the absence of detectable virus in transgenic plants 82-1 and 82-2, but high virus titers were obtained in line 82-3 and WT plants ( Figure 6B). To analyze the unexpected virus-sensitive phenotype in line 82-3 we performed PCR analysis and Southern blotting using genomic DNAs isolated from leaf samples taken before the virus challenge. We found that no band was amplified from both P1-P2 and P3-P4, indicating line 82-3 was likely a null segregant ( Figure 6C). Southern blotting showed the same bands as T0 transgenic plants in lines 82-1 and 82-2, but no signal was detected for 82-3, confirming it was indeed a null segregant ( Figure 6D). Overall, our results show that transgenic plants based on the multitarget dsRNA approach can confer heritable resistance.
Resistance to a second Indian cassava mosaic virus strain sharing 94.5% nucleotide identity with ICMV-Dha We also tested whether the ICMV resistance conferred by hpRNAi has broad virus resistance. We recently reported a second ICMV isolate from the Singaporean Jatropha tree (named ICMV-SG, accession number [JX518289]), which shares 94.5% nt homology with the ICMV-Dha strain [7]. Before the virus challenge, we carried out PCR-based genotyping with primer pairs P1-P2 and P3-1, together with P4-1. We found a band of the expected size amplified from P1-P2, but not P3-1 and P4-1, from both lines 82-4 and 82-11, indicating these two plants (82-4 and 82-11) might be marker-free plants, whereas line 82-14 was likely a null segregant as was line 82-3 ( Figure 7A). We tested plants of lines 82-4 and 82-11, along with six other plants   (lines 82-5 to 82-10), which were still chimeric with respect to the marker gene, for possible resistance to the second virus (ICMV-SG) using the agroinfection-based infection method. As expected, all of the seven plants in which marker excision had occurred were found to be resistant to ICMV-SG in the T1 generation. Further genotype analysis showed that all seven of these plants were hemizygous (lower panel of Figure 7A). The null segregant line 82-14 and two other plants which still retained the marker gene showed typical virus infection symptoms and accumulated high levels of virus genomic DNA, similar to WT control plants ( Figure 7B,C,D,E).
We reasoned these ICMV-resistant plants should be also resistant to multiple ICMV strains within a similar evolution clade. Multiple sequence alignment of full-length DNA-A was carried out using the ClustalV program Vector NTI® (Life technologies, Carlsbad, United States) with default parameters. Phylogenetic trees were constructed from multiple alignments using the neighbor-joining method in the MEGA4 program [21], and a bootstrap analysis with 1,000 replicates was performed. Only values above 70 were reported on the trees (Figure 8). Based on the evolutionary phylogenetic tree we deduced that the resistant lines should also contain resistance to another two viral isolates recovered from J. curcus.

Discussion
During the last decade geminiviruses have emerged as one of the major causative pathogens of economically important crops such as cassava, cotton and tomato. As resistance genes are not always available in the relevant crop germplasm, biotechnological approaches have been used to generate transgenic plants to confer virus resistance. Several different strategies have been reported to confer virus resistance to transgenic model plants by expression of: (1) trunked viral proteins, (2) artificial zinc finger nuclease, (3) peptide aptamers capable of binding to viral proteins, (4) non-pathogen-derived antiviral agents and so on [18,[22][23][24][25]. However, successful applications of these strategies to generate virus-resistant crop plants are rare. In the case of RNA viruses, expression of virus-derived hpRNA has been shown to provide resistance in transgenic plants; more recently, this strategy has been shown to work for DNA viruses such as geminiviruses as well [26]. Successes have been reported with RNAi technology using either single or multiple viral gene sequences, or by artificial miRNA [27][28][29]. Single target gene approaches have first been used. For example, sequences targeting the AC1 gene (which encodes the multifunctional Rep protein), the coat protein gene and AC4 gene have been shown to confer geminivirus resistance in tomato, bean and tobacco plants [14][15][16][30][31][32]. In the case of the Chickpea chlorotic dwarf Pakistan virus (CpCDPKV), a member of Mastrevirus (also a member of the Geminiviridae family) a multi-target approach was adopted. The RNAi trigger sequences include those encoding the N-terminus of the Rep gene, the large intergenic (non-coding region; LIR) and the N-terminus of the movement protein gene [32]. Notably, the first field proven transgenic geminivirus-resistant plant-common bean (Phaseolus vulgaris) showed that the begomovirus Bean golden mosaic virus (BGMV) can also be suppressed by the expression of a hpRNA transgene derived from AC1 [33].
Here, we employed an RNAi strategy using hairpin dsRNA to silence five key viral genes of the ICMV. Five viral genes were chosen as targets: genes coding for AC1, AC2, AC3, AV1and AC4. This multi-target strategy conferred transgenic J. curcus plants with high resistance to multiple virus infection. After two rounds of virus challenge, resistant plants were free of virus symptoms and no virus could be detected (Figures 4 and 6). Using qPCR to measure and quantify viral titers, we found strong virus resistance in the resistant transgenic lines as compared to WT and susceptible lines (Figures 3, 6 and 7). Our results, together with those of others [30][31][32][33], confirm that the hairpin dsRNA-mediated RNAi approach is indeed a promising technology in generating virus-resistant plants and/or conferring stable and effective resistance in plant crops.
An interesting feature of the work presented here is that we were able to challenge the transgenic plants with two distinct viruses (94.5% nt identity for the whole DNA-A), thereby allowing us to ascertain resistance to related viruses. There is 95.3% nt identity between the ICMV-Dha and ICMV-SG strains if we include sequences that we used for the hpRNAi construction. Two other J. curcus-isolated ICMV strains, ICMV-JC3 and ICMV-Jalgon, share at least 95% nt identity with the Dha strain sequences that we used for hpRNAi construction ( Figure 8 and Additional file 1). In view of this high sequence homology we expect that the transgenic lines generated in this study may also be resistant to these two additional strains. Further field trials are necessary to ascertain whether the multi-target hpRNAi approach can indeed confer durable and heritable resistance to geminiviruses. Nevertheless, here we provide evidences that the multi-target hpRNAi strategy can confer resistance to the ICMV. There were several transgenic T0 lines showing virus resistance to the first ICMV-Dha infection, especially for the line 82. A total of 10 T1 progeny transgenic plants derived from line 82 were free of virus symptoms with no detectable virus, whereas the null segregant 82-3 was as sensitive as WT plants, when challenged with either ICMV-Dha or ICMV-SG strains (Figures 6 and 7). Lines 82-4 and 82-11 are good candidate plants for further commercial development since they are marker-free and contain only a single T-DNA insert. Taken together, our results support and reinforce the notion that the transgenic resistance trait is not only durable but also heritable.

Conclusions
We generated transgenic J. curcus plants with resistance to ICMV via expression of a hairpin dsRNA with sequences homologous to five key ICMV DNA-A genes. Virus resistance was confirmed through two rounds of inoculation with J. curcus-isolated strains of ICMV, ICMV-Dha and ICMV-SG. The T1 progeny transgenic plants were virus resistant, confirming that the resistance phenotype was heritable. Transgenic plants generated from this study can be used in various Jatropha breeding programs.

Plasmids construction
We replaced the G10-90 promoter in pX7-GFP [20] with a synthetic 35S promoter harboring a double enhancer (detailed sequence in Additional file 2). The resulting vector was named as pX9-GFP. Three fragments of ICMV-Dha DNA-A (Genbank accession number [GQ924760]) were chosen as the target region. The ligated DNA fragment (fragment 1: 521-770 nt, fragment 2: 1210-1459 nt and fragment 3: 1671-2279 nt; see Figure 1B) was generated by two rounds of overlapping extension PCR using the following primers: Sequences for fragments 1, 2 and 3 were compared with those from related ICMV strains isolated from J. curcus (Additional file 1).
The PCR fragment was inserted in the sense orientation into the XhoI/HindIII sites of a pSKint vector [20] to generate pSK-int-sense ICMV. Another fragment, amplified with forward primer 5'-AGCGCGAATTCTAGCTGGAA TTGGGCCCTGGATTGCAGA-3' and reverse primer 5'-AGCGCACTAGTCGTTTGAATCTAGACAC-GATGTGCTCCA-3' , was subsequently placed in the antisense orientation into the EcoRI/SpeI sites of pSK-intsense ICMV to give pSK-int-ICMV RNAi. Finally, the entire RNAi cassette comprising the sense and antisense fragments joined by the Arabidopsis thaliana Actin II intron was excised from pSK-int using the flanking XhoI/ SpeI sites and inserted into the XhoI/SpeI site of pX9-GFP, yielding the construct pX9-hpICMV RNAi (pX9-vRNAi).

Virus challenge assay
Infectious clones of ICMV-Dha and ICMV-SG carrying tandem repeats of DNA-A and DNA-B were described [3,7]. Virus challenge assays were performed by Agrobacterium-mediated vacuum infiltration [35]. Plants were vacuum-infiltrated two times within a 10-day interval.

Real-time PCR to detect viral titers in challenged plants
Equal amounts of genomic DNA were used for analysis. Real-time PCR was performed with Power SYBR™ Green PCR Master mix (Applied Biosystems, Foster City, California, United States) and ran in ABI7900HT. Forward primer 5'-CTGCACAATGTGGGACCCTTTG-3' and reverse primer 5'-CTTCGCCCTGATGACAGA-GATC-3' were used for the amplification of viral DNA-A of either ICMV-Dha (128-290 nt) or ICMV-SG (127-289 nt). All samples were run in triplicate and the data was analyzed with RQ manager at a preset threshold cycle value (Applied Biosystems, Foster City, California, United States). The Jatropha curcas rbcL DNA served as an internal control using forward primer 5'-GGAGTTCCGCCTGAGGAAG-3' and reverse primer 5'-CTTCTCCAGCAACGGGCTC-3'. As described by Prisco et al. [36], the relative quantification of virus titers was determined based on the value of the Ct, using the comparative C T method and the formula 2 −ΔΔCt .

Southern blotting
Genomic DNA was digested with restriction enzymes and separated on 0.8% agarose gels. The gels were processed and resolved DNA bands transferred to a nylon Hybond-N + membrane (GE Biosciences, Buckinghamshire, United Kingdom) following standard procedures [13]. Membranes were hybridized with a CaMV 2 × 35S promoter probe which included the double enhancers with 3' ends at −76 (A) of 35S promoter. The probes were DIG-dUTP-labeled by PCR using a PCR DIG probe synthesis kit (Roche Shanghai, China) according to the manufacturer's instructions and signals were detected by autoradiography.