TRANSGENIC RESISTANCE TO BARLEY YELLOW DWARF VIRUS: REPLICATION, RECOMBINATION, AND RISKS

Gennadiy Koev, B.R. Mohan, Randy Beckett, and W.Allen Miller*

Plant Pathology Department, 351 Bessey Hall, Iowa State University, Ames, IA. 50011.

*For offprint requests: fax:(515)294-9420, email: wamiller@iastate.edu, www: http://www.public.iastate.edu/~wamiller/

SUMMARY

Barley yellow dwarf viruses (BYDVs) are among the most economically important pathogens of cereals. Conventional breeding methods have not provided satisfactory level of protection against BYDV. Thus, genetically engineered virus-derived resistance is a potentially important tool for oat improvement. Along with potential advantages, transgenic resistance to BYDV has possible risks associated with it such as transgene escape to wild relatives, recombination, complementation, and synergistic interaction with the challenging virus. The lab of D. Somers (University of Minnesota), transformed oats with the first two open reading frames (ORFs) including the polymerase gene of BYDV-PAV. Most of the lines have multiple transgene insertions into the plant genome. Transgenic lines are being tested for resistance to the virus, and some of them have shown increased tolerance compared to the untransformed controls. To identify previous recombination events in luteovirus evolution, sequences of divergent isolates of BYDV are being determined. Likely sites of recombination, the subgenomic mRNA promoters, are being mapped to gain an understanding of the recombination mechanism.

Key words: Transgenic-virus-resistance, risk-assessment, transgene-escape, RNA-recombination, luteovirus

INTRODUCTION

Barley yellow dwarf viruses belong to the luteovirus group. Like all luteoviruses, BYDVs are phloem-limited and obligately transmitted in a persistent fashion by specific aphid species. Based on the genome organization, serological and other properties, luteoviruses including BYDV are divided into two subgroups (Miller et al., 1995; Mayo and Ziegler-Graf, 1996). Genes coding for viral RNA-dependent RNA polymerase (RdRp or replicase) of subgroup 1 luteoviruses located in the 5' half of the genome are more similar in sequence to those of diantho- and carmoviruses than to the polymerase genes of subgroup 2 luteoviruses. On the other hand, polymerase genes of the subgroup 2 are more closely related to those of sobemoviruses. The 3' halves of both subgroups' genomes have more homology to each other than to any other viruses (Miller et al., 1995; Mayo and Ziegler-Graf, 1996). This part of the luteovirus genomes contains genes coding for coat protein and proteins involved in aphid transmission and plant systemic infection (Chay et al., 1996a).

Numerous examples exist in which transgenic plants transformed with all or part of the viral RNA-dependent RNA polymerase (RdRp) gene are highly resistant to the virus (Anderson et al., 1992; Baulcombe, 1994; Braun and Hemenway, 1992; Golemboski et al., 1990). Virus-resistant transgenic potatoes expressing the RdRp gene of potato leafroll virus are an example of this strategy applied in luteoviruses (Kaniewski et al., 1994). Thus, in collaboration with David Somers' lab of the University of Minnesota, we developed transgenic oat plants expressing the BYDV-PAV replicase gene. Upon challenge of several lines with viruliferous aphids, one of the transgenic lines was more tolerant to the virus infection than others and untransformed controls. We are continuing testing plants for resistance.

Among potential risks associated with virus-derived resistance, transgene escape, recombination, and complementation are of the most concern. There are examples of the transgene escape from genetically engineered plants to their wild relatives with pollen. Hybrids between transgenic squash expressing zucchini yellow mosaic virus coat protein and wild squash have been shown to be more resistant to the virus infection than wild squash (Fuchs and Gonsalves, 1996). Oats normally self-pollinate, but there is some degree (up to 8.7%) of outcrossing with wild relatives (Shorter et al., 1978). We have begun to test the possibility of cross-pollination between oat, Avena sativa, and a related noxious weed, Avena fatua.

The possibility of recombination between a virus-derived transgenic transcript and an invading virus has been demonstrated under strong (Schoelz and Wintermantel, 1993; Greene and Allison, 1994) and "moderate" (Wintermantel and Schoelz, 1996) selection pressure. A recombination event occurred recently in luteovirus evolution which explains the homology of some of the genome regions to other virus groups (Miller et al., 1995). Recombination sites seem to coincide with the regions the subgenomic RNA transcription start sites. Here, we show our work on mapping BYDV-PAV subgenomic promoter. Avoiding high-likelihood recombination sequences in the transgenic constructs might reduce risk of recombination.

Synergistic interaction has been known to occur between members of the subgroup 1 and 2 of luteoviruses. For example, plants infected with a mixture of BYDV-PAV and BYDV-RPV can show more severe symptoms than the ones infected with either virus alone. Therefore, it's quite possible that transgenic plants expressing a part of the PAV genome might exhibit severe symptoms when infected with RPV due to the synergism. We are testing this possibility.

MATERIALS AND METHODS

BYDV isolates. PAV-Illinois was provided by A. Hewings (formerly USDA/ARS, Univeristy of Illinois). PAV129 is a severe PAV-like isolate from upstate New York (Chay et al., 1996b). This strain and the type strain of RPV (RPV-NY) were gifts from S. Gray (USDA/ARS, Cornell University). The severe Mexican strain of RPV (RPV-Mex1) was provided by Lukas Bertschinger formerly of CIMMYT.

Oat transformation. Oats were transformed by particle bombardment and regenerated in D. Somers' lab, University of Minnesota as described previously (Somers et al., 1992).

DNA extraction and Southern blot analysis. Genomic DNA was isolated from 0.25g freeze-dried tissue of each plant using a modified CTAB method (Saghai-Maroof et al., 1984; Wise and Schnable, 1994). DNA samples (20 ug) were digested with restriction endonuclease BamHI (NEB), 2.5 units/ug, at 37 C over night in the manufacturer's buffer supplied with the enzyme. Restriction fragments were separated in 0.9% agarose gel by electrophoresis, blotted to a Hybond N+ nylon membrane (Amersham), and hybridized with the probe (Wise and Schnable 1994). 32P-labeled probe was made by oligonucleotide random priming method using Promega (Madison, WI) Prime-a-Gene kit. EcoRI fragment of p35S99K was isolated in low melt agarose gel and used as a template for the probe preparation. Phosphorimager cassettes were exposed to the radioactive blots for 7 days.

RNA extraction and northern blot analysis. Total RNA was isolated from plant leaf tissue (0.1g) using QIAGEN RNeasy total RNA isolation kit. Northern blot hybridization was performed as in Seeley et al. (1992). 32P-labeled probe was made by transcription in vitro according to Promega protocol. Phosphorimager cassette was exposed to the radioactive blot for 5 days. For the synergy experiment, the aurintricarboxylic acid method was used for RNA isolation from plants (Seeley et al., 1992).

Crosses. Crosses between transgenic A. sativa and its wild relative A. fatua were performed in a greenhouse as described in Brown (1980).

Screening for tolerance. In the first experiment, 15 transgenic lines, 6 plants per line were inoculated with about 20 viruliferous aphids (Rhopalosiphum padi) per plant at the age of two weeks. In the second experiment, 4 transgenic lines, 10 plants per line were inoculated with about 20 viruliferous aphids per plant at the age of 4.5 weeks. In both experiments, aphids were allowed to feed for two days, then the plants were sprayed with insecticide.

Synergy. In this experiment, two nontransgenic oat lines GAF/Park 1 and 3 (GP-1, GP-3) and two transgenic lines 2811 and 2803 were used. The nontransgenic lines were inoculated as follows: 4 plants of each line with 5 PAV-infected aphids, 4 plants - 5 RPV-infected aphids, and 4 plants - 5 PAV and 5 RPV-infected aphids. Three plants of each line were left as uninfected controls. The transgenic lines were inoculated with RPV-infected aphids only, 5 aphids per plant. Plants were sprayed after two days.

Sequencing and subgenomic RNA promoter analysis. Viral RNAs were cloned via reverse transcription PCR using Superscript II reverse transcriptase (Gibco-BRL, Gaithersburg, MD), and Vent thermostable DNA polymerase (New England BioLabs, Beverly, MA). Dideoxy sequencing was performed on an ABI 377 automated sequencer. Site-directed mutagenesis and infectivity assays for the full-length, infectious clone, pPAV6, of BYDV-PAV were performed as in Mohan et al. (1995).

RESULTS

Screening for transgene. Putative transgenic plants (R1) were screened for the presence of the insert by using Southern blot hybridization. Most of the plants screened contained multiple copies of the transgene according to the multiple band pattern on the Southern blot. Several lines such as 2880, 2811, 2805, 2606, 2819, 2821, 2822 showed segregation but all plants screened in other lines contained the transgene.

Screening for tolerance. In our first experiment, 15 transgenic lines were inoculated with BYDV-PAV to determine the degree of resistance in comparison to untransformed controls. All plants developed symptoms of the virus infection after 5 to 7 days post inoculation. One month post inoculation most transgenic plants as well as untransformed controls were dead or nearly dead. Plants of two lines, 2880 and 2811 showed much milder symptoms, grew to maturity and produced seed. In the second experiment, these two lines, the most susceptible transgenic line 2803, and untransformed controls were inoculated with the virus at the older age (4.5 weeks). Before inoculations were done, we isolated total RNA from the plants and analyzed for the transgenic transcript accumulation. All lines except for 2880 and untransformed controls showed accumulation of degraded transgene mRNA. As far as tolerance, lines 2880 and 2811 again exhibited milder symptoms than the untransformed plants and line 2803, but the differences were more modest. However, at the later stage, line 2811 performed better than line 2880. Plants of line 2811 showed a recovery phenotype characterized by more green tissue and florets, more vigorous growth, and ultimately more seed production (Figure 1). Southern blot analysis of the plants used in this experiment indicated absence of the transgene in the plants of line 2880 (data not shown). This line had segregation in the first generation, so probably nontransgenic segregants were picked for this experiment. Good performance of the line 2880 at the earlier stage might be explained by somaclonal variation in tissue culture. Although, ultimately the presence of transgene in line 2811 coincided with the increased tolerance of those plants.

Crosses. To determine the possibility for the deliberate transgene transfer from transgenic cultivated oats A. sativa to the wild oat species A. fatua, we performed interspecific crosses where our transgenics (lines 2808 and 2802) were used as male plants, and A. fatua as female plants. Out of about 15 crosses, only two seeds developed. The first attempt to germinate the seeds was undertaken a month after they were collected. The seeds did not germinate. Only three months after collection, the seeds germinated and were planted into pots.

Synergy. To determine the possibility of synergistic interaction between two barley yellow dwarf viruses, PAV and RPV, that belong to the different luteovirus subgroups and also the possibility of synergism between the PAV transgene and the genome of challenging RPV, nontransgenic plants, line GAF/Park were inoculated with both viruses, PAV only, and RPV only. Beside that, transgenic plants of lines 2803 and 2811 expressing PAV polymerase gene were inoculated with RPV. Plants infected with RPV developed only very mild symptoms characteristic of the RPV-NY isolate. PAV only and PAV+RPV-infected plants developed much more severe symptoms: severe stunting, yellowing and reddening of the leaves with subsequent necrosis, no heading. After about a month post inoculation, plants infected with the mixture of the two viruses exhibited slightly more severe symptoms than those infected with PAV only. We analyzed the rate of viruses' replication by Northern blot hybridization of total RNA extracted from the plant leaf tissue. Considering uneven distribution of BYDV in the plant organism (Periera, 1989), we took random tissue samples from up to 10 leaves of infected plants. The Northern blot was first analyzed for the RPV genomic and subgenomic RNA accumulation with the probe complementary to the 3' end of the RPV genomic RNA. The results do not show any difference in RPV replication between the samples. Then the same blot was analyzed for the PAV genomic RNA accumulation with the probe complementary to the 5' end of the PAV genomic RNA. No difference was found in PAV replication either.

Subgenomic mRNA promoter mapping. Homology between the two subgroups of luteoviruses begins and ends near the 5' ends of subgenomic mRNAs. Thus, we proposed that the subgroups arose by recombination at the subgenomic mRNA promoters, where internal initiation of RNA synthesis begins. Thus, we have begun mapping these promoters. The 5' ends of genomic and subgenomic RNAs (sgRNAs) of PAV share the sequence: (A)GUGAAG. Point mutations in this region of the subgenomic mRNA1 leader in the infectious clone of the PAV genome had no effect on PAV RNA replication or subgenomic mRNA accumulation. Another sequence, that is conserved at (subgroup II) or near (subgroup I) the 5' ends of genomic and sgRNAs of luteoviruses in both subgroups, ACAAA (Miller et al., 1995) will be mutated next.

Sequence analysis. Very few BYDV isolates have been sequenced in their entirety. Knowledge of the natural sequence variation is essential if we are to understand the likelihood of recombination, and safe means of deploying viral transgenes. We are sequencing severe strains of PAV (PAV-129) (Chay et al., 1996b), and RPV (RPV-Mex1). The polymerase gene and a portion of the 3' untranslated region of PAV-129 differs more from the other PAV isolates than they differ from an MAV isolate. The polymerase ORF has 88% amino acid sequence identity to the PAV and MAV isolates, whereas the other PAVs and MAV all have 98-99% identity to each other. The, as yet unsequenced, 5' half of the genome of RPV-Mex1 has no homology to that of RPV-NY based on Northern blot hybridization studies (L. Rasochova, personal communication), whereas the 3' halves of these two genomes share 85% nucleotide sequence identity.

DISCUSSION

Transgenic resistance. Development of cereal cultivars with environmentally safe and economically efficient, genetically-engineered resistance to BYDV would be a very important step towards cereal crops improvement. We have made first attempts to engineer oats with polymerase-mediated tolerance to the virus and study possible risks associated with it. Screening for tolerance showed that one truly transgenic line 2811 performed the best as far as the amount of seeds produced and showed a recovery phenotype. This fact taken together with the degradation of the transgenic mRNA hints that RNA-mediated "sense" strand co-suppression may be the mechanism of tolerance. This phenomenon, in which the transgenic and any invading viral RNA are specifically degraded has been described for other viruses (Mueller et al., 1995; Pang et al., 1996; Smith et al., 1994). Although, we can't explain why other lines (2803, 2806) did not show any resistance to the virus. One explanation might be that the amount of inoculum applied was overwhelming for the plants, especially considering severity of BYDV-PAV-Illinois isolate used in our studies. Relatively good performance of the line 2880 at the earlier stage of infection might be due to the somaclonal variation in tissue culture that is observed often in plant transformation. Transgenic potatoes that expressed the coat protein gene of PLRV showed increased resistance to the virus as well as the vector only-transformed ones (Presting et al., 1995). One of the explanations suggested that the tissue culture process caused somaclonal variation which resulted in increased resistance of the plants. We continue testing our transgenic plants for resistance.

Lack of synergistic interactions. Synergistic interaction between two viruses in mixed infections is characterized by production of more severe symptoms than either virus can produce alone. This phenomenon is best studied on the PVY/PVX model, where presence of PVY genomic RNA enhances the replication rate of PVX (Vance et al., 1995). In luteo- and related viruses, many pairs of a luteovirus subgroupI-like virus and a subgroup II-like virus, have been observed to produce more severe disease than either virus alone (synergy) (Miller et al., 1995). One example of such pairs is the ST9a-associated RNA (subgroup I) and beet western yellows virus (BWYV) genomic RNA (subgroup II) in the ST9 isolate of BWYV. In mixed infections, rates of BWYV replication and symptom production in plants are greatly enhanced owing to the presence of ST9a RNA (Sanger et al., 1994). Mixed infection of two barley yellow dwarf viruses that belong to the two different subgroups, PAV and RPV, can also produce synergistic disease in oats (P.M. Waterhouse, unpublished observation), although the levels of virus and viral RNAs were not. In our experiment, we observed some symptom exacerbation in plants infected with both PAV and RPV. However, results of Northern blot analysis did not indicate any difference in the accumulation of either of the viral genomes. It is possible that different inoculation conditions are necessary to ensure simultaneous infection of individual cells by both viruses. Another possibility is that the PAV/RPV interaction does not result in increase of the replication rate, but is still able to enhance the symptom production by some other mechanism. In this case, the small difference in symptomatology between PAV only-infected plants and PAV+RPV-infected plants can be explained again by the extreme severity of PAV-Illinois which makes any additional symptom exacerbation insignificant. In transgenic plants expressing the PAV polymerase gene infected with RPV we did not observe any synergy either. However this does not rule out the possibility for synergistic interaction between the PAV transgenic transcript and RPV genome because as we reported above, transgenic mRNA is degraded in the plants that were used. Finally, it should be borne in mind that this type of a risk is only to the grower (and seed supplier). If such a crop proved highly susceptible to the synergistic partner virus, the problem could be controlled in the future simply by ceasing to use the particular crop. Although if enough acres are planted without knowledge of such a possibility, the result could be analogous to the worst epiphytotic in U.S. history, the southern corn leaf blight epidemic of 1970. In this case, millions of acres of corn were planted which carried the T-cytoplasm that confers both the desired trait of male sterility, and also extreme susceptibility to the T toxin of the causal agent of southern corn leaf blight: Cochliobolus heterostrophus (Pring and Lonsdale, 1989).

Subgenomic promoter. We were surprised that some of the conserved bases at the 5' end of subgenomic RNA1 of BYDV-PAV RNA were not needed for subgenomic RNA generation or virus replication. This may indicate that the nearby ACAAA sequence which is conserved in both subgroups, but is located precisely at the 5' ends only in subgroup II luteoviruses, may be an important promoter element. This would support our model that a shared promoter is the site at which recombination occurs. Studies are underway to test this possibility.

Sequence variation. As indicated by our preliminary work. The natural sequence variation in BYDV's, especially RPV is broader than initially indicated from the few full-length sequences that had been completed previously. The extreme difference in the 5' end of RPVMex1 suggests that the polymerase gene may have originated by recombination with RNA from a different source than we have seen for other BYDV's. Completion of the sequence will provide the answer.

ACKNOWLEDGEMENTS

We would like to aknowledge Dr. Roger Wise (USDA/ARS, ISU), Arla Bush, and Karin Werner, for their advice in designing Southern blot assay; Dr. Stewart Gray (USDA/ARS, Cornell University) for providing BYDV isolates; Dr. James Colbert for the cDNA clone of the oat actin gene; Lada Rasochova for the information about sequence homology.

REFERENCES

Anderson J, Palukaitis P, Zaitlin M. 1992. A defective replicase gene induces resistance to cucumber mosaic virus in transgenic tobacco plants. Proc Natl Acad Sci USA 89: 8759-8763.

Baulcombe D. 1994. Replicase-mediated resistance: a novel type of virus resistance in transgenic plants? Trends Microbiol 2: 60-63.

Braun C, Hemenway C. 1992. Expression of amino-terminal portions or full-length viral replicase genes in transgenic plants confers resistance to potato virus X infection. Plant Cell 4: 735-744.

Brown CM. 1980. Oat. In: Fehr WR, Hadley HH (eds) Hybridization of crop plants. Am Soc of Agron, Inc, Madison, Wis, pp 427-441.

Chay CA, Gunasinge UB, Dinesh-Kumar SP, Miller WA, Gray SM. 1996a. Aphid transmission and systemic plant infection determinants of barley yellow dwarf luteovirus-PAV are contained in the coat protein readthrough domain and 17-kDa protein, respectively. Virology 219: 57-65.

Chay CA, Smith DM, Vaughan R, Gray SM. 1996b. Diversity among isolates within the PAV serotype of barley yellow dwarf virus. Phytopathology 86: 370-377.

Fuchs M, Gonsalves D. 1996. Is gene flow a serious risk for the release of virus-resistant transgenic crops? APS abstract. Phytopathology. In press.

Golemboski D, Lomonossoff G, Zaitlin M. 1990. Plants transformed with a tobacco mosaic virus nonstructural gene sequence are resistant to the virus. Proc Natl Acad Sci USA 87: 6311-6315.

Greene AE, Allison RF. 1994. Recombination between viral RNA and transgenic plant transcripts. Science 263: 1423-1425.

Kaniewski W, Lawson C, Loveless J, Thomas P, Mowry T, Reed G, Mitsky T, Zalewski J, Muskopf Y. 1994. Expression of potato leafroll virus replicase genes in Russet Burbank potatoes provide field immunity to PLRV. Abstract P1-32 of the Thirteenth Annual Meeting of the American Society for Virology, Madison, WI, July 1994, p. 107.

Mayo MA, Ziegler-Graf V. 1996. Molecular biology of luteoviruses. Adv Virus Res 46: 413-460.

Miller WA, Dinesh-Kumar SP, Paul CP. 1995. Luteovirus gene expression. Critic Rev Plant Sci 14: 179-211.

Mohan, BR, Dinesh-Kumar, SP, and Miller, WA 1995. Genes and cis-acting signals involved in replication of barley yellow dwarf virus-PAV RNA. Virology 212: 186-195.

Mueller E, Gilbert J, Davenport G, Brigneti G, Baulcombe DC. 1995. Homology-dependent resistance: transgenic virus resistance in plants related to homology-dependent gene silencing. Plant Journal 7: 1001-1013.

Pang S-Z, Jan F-J, Carney K, Stout J, Tricoli DM, Quemada HD, Gonsalves D. 1996. Post-transcriptional transgene silencing and consequent tospovirus resistance in transgenic lettuce are affected by transgene dosage and plant development. Plant Journal 9: 899-909.

Periera A-MN, Lister RM, Barbara DJ, Shaner GE. 1989. Relative transmissibility of barley yellow dwarf virus from sources with differing virus contents. Phytopathology 79: 1353-1358.

Presting GG, Smith OP, Brown CR. 1995. Resistance to potato leafroll virus in potato plants transformed with the coat protein gene or with vector control constructs. Phytopathology 85: 436-442.

Pring, DR, Londsdale, DM (1989) Cytoplasmic male sterility and maternal inheritance of disease susceptibility in maize. Annu. Rev. Phytopathol. 27: 483-502.

Saghai-Maroof MA, Soliman KM, Jorgensen RA, Allard RW. 1984. Ribosomal DNA spacer-length polymorphisms in barley: Mendelian inheritance, chromosomal location, and population dynamics. Proc Natl Acad Sci USA 81: 8014-8018.

Sanger M, Passmore B, Falk BW, Bruening G, Ding B, Lucas WJ. 1994. Symptoms severity of beet western yellows virus strain ST9 is conferred by the ST9-associated RNA and is not associated with virus release from the phloem. Virology 200: 48-55.

Schoelz JE, Wintermantel WM. 1993. Expansion of viral host range through complementation and recombination in transgenic plants. Plant Cell 5: 1669-1679.

Seeley KA, Byrne DH, Colbert JT. 1992. Red light-dependent instability of oat phytochrome mRNA in vivo. Plant Cell 4: 29-38.

Shorter R, Gibson P, Frey KJ. 1978. Outcrossing rates in oat species crosses (Avena sativa L. X A. sterilis L.). Crop Science 18: 877-878.

Smith HA, Swaney SL, Parks TD, Wernsman EA, Dougherty WG. 1994. Transgenic plant virus resistance mediated by untranslatable sense RNAs: expression, regulation and fate of nonessential RNAs. Plant Cell 6: 1441-1453

Somers DA, Rines HW, Gu W, Kaeppler HF, Bushnell WR. 1992. Fertile, transgenic oat plants. Bio/Technology 10: 1589-1594.

Vance VB, Berger PH, Carrington JC, Hunt AG, Shi XM. 1995. 5' proximal potyviral sequences mediate potato virus X/potyviral synergistic disease in transgenic tobacco. Virology 206: 583-590.

Wintermantel WM, Schoelz JE. 1996. Isolation of recombinant viruses between cauliflower mosaic virus and a viral gene in transgenic plants under conditions of moderate selection pressure. Virology 223: 156-164

Wise RP, Schnable PS. 1994. Mapping complementary genes in maize: positioning the rf1 and rf2 nuclear fertility restorer loci in Texas (T) cytoplasm relative to RFLP and visible markers. Theor Appl Genet 88: 785-795.

Figure 1. Average amount of seeds produced by mature oat plants infected with BYDV-PAV. Plants were infected with BYDV-PAV by viruliferous aphids (see 'Materials and methods' for details); seeds were harvested and quantified after 3 months post inoculation. Lines used in the experiment: 2811 - transgenic, most tolerant line; 2880 - nontransgenic segregants of the line 2880; 2803 - transgenic, most susceptible line; GP-8 - nontransgenic control; non-inf. - non-infected plants of line GP-8.