Molecular breeding for virus resistance : an applied approach in vegetable crops

Viral diseases cause significant economic losses in most, if not all, crop species throughout the world. Total cost is not only restricted to reduction in crop yield and quality, but also include the development and application of a wide array of disease control measures. Routinely employed culture practices include quarantine measures, eradication of infected plants and weed hosts, crop rotation and the use of certified virusfree seed or planting stock. Additionally, the use of pesticides to control insect vector populations implicated in transmission of the virus, represents an important tool to limit the incidence of viral disease outbreaks. However, none of these non-genetic control measures is likely to provide the long-term answer to combat viral diseases, because of their expense and their sometimes questionable effectiveness and reliability. Moreover, current concern about pollution and food safety is forcing hazardous pesticides of the market.As key control pesticides are progressively abandoned, there is a growing urgency for the development of alternative methods to control viral diseases. Breeding for virus resistance generally provides the best prospects for virus control in the long term. In the past, the introgression of genetic sources for host plant resistance that are naturally present within the gene pool of the crop involved, has been successfully applied to develop virus resistant crop cultivars for a considerable number of agronomically important crops. Although plant breeding for virus resistance still is of great potential, there are limitations to this conventional approach. An appropriate source of resistance may not be available in interfertile relatives, the source may be tightly linked to undesirable traits, or may be multigenic and as such difficult to advance in breeding programs. Consequently, the major barrier inherent to plant breeding for virus resistance is the scarcity of suitable sources of host resistance.The limitations of conventional breeding and routine culture practices urge the need for the development of alternative forms of virus control that can be fully integrated within traditional methods. In this perspective, the concept of pathogen-derived resistance as elaborated by Sanford & Johnston (1985), provides an attractive strategy to produce novel, but genetic forms of virus control, by transforming crop plants with nucleotide sequences derived from the viral genome. Major progress in the molecular characterisation of plant virus genomes and the stable transformation of plant species (Fisk & Dandekar, 1993) opened the avenue for molecular breeding to produce transgenic progenitors carrying novel and 'green' sources of virus resistance, that can be incorporated in routine crop breeding programs.Molecular breeding for virus resistance relies on two basic disciplines. Firstly, the design and construction of pathogen-derived resistance genes (Chapter 3), and secondly, the introduction and expression of such synthetic resistance genes in transgenic plants (Chapter 2). For dicotyledonous plant species, Agrobacterium -mediated transformation generally is the method of choice (Zambryski, 1992; Zupan & Zambryski, 1995), as is illustrated by the successful transformation of elite lines of lettuce, melon and tomato (Chapters 4, 6 and 8 respectively). In spite of the highly similar approaches employed, starting from seedling cotyledons for explant material and using kanamycin resistance as selectable marker, the efficacy of transformation appeared highly variable for the various crops. For tomato, the effective transformation frequency, defined as the percentage of explants yielding independent diploid transformants that express the transgene at measurable levels, was calculated at approximately five percent. The transformation of melon, on the other hand, was severely hampered by the emergence of false positives that escaped from kanamycin selection during regeneration. On average, merely ten percent of the shoots was truly transgenic, thereby decreasing the effective transformation frequency to values below 0.5%, but heavily increasing the resultant effort devoted to the selection of true transformants. General tissue culture conditions, however, comprising parameters such as the concentration and kind of plant hormones and plant vitamins, and the use of feeder layers, are known to have a significant influence on transformation efficiencies and do differ considerably for the transformation protocols of tomato, melon and lettuce. Additionally, there exist large differences in responsiveness of different crop genotypes to transformation and to in vitro tissue culture in general. Consequently, plant transformation remains a highly empirical science for which elaborated protocols do not exist (van Wordragen & Dons, 1992). Only trial and error, based on the knowledge and experience from existing protocols can contribute to the development and optimisation of transformation procedures.For the design and construction of pathogen-derived resistance genes a plethora of strategies can be employed, only limited by the size and content of the viral genome concerned (Chapter 3). The viral coat protein (CP) gene, however, has thus far been more widely applied than any other viral sequence, because coat protein-mediated protection was the first described (Powell Abel et al. , 1986; Beachy et al. , 1990). Accordingly, the transformation of tomato with the CP gene from cucumber mosaic cucumovirus (CMV) was shown to generate high levels of protection to CMV infection in tomato hybrids, not only when challenged by mechanical inoculation, but also when exposed to repeated inoculation by viruliferous aphids in open field (Chapter 5). On the contrary, the transgenic expression of the CP gene from beet western yellows luteovirus (BWYV) in lettuce only yielded marginal levels of protection (Chapter 4). Transgenic lettuce plants never resisted infection with BWYV, but merely showed a delay in systemic symptom development. Apparently, the transgenic expression of viral CP genes to engineer resistance against the homologous virus does not apply with equal success to all virus genera, or to all virus-crop combinations.For potyviruses the protection engendered by the transgenic expression of the potyviral CP gene is known to be mediated at the transcript level, rather than by the accumulation of transgenically expressed coat protein (Lindbo et al., 1993). Likewise, the transformation of melon with the CP gene from zucchini yellow mosaic potyvirus (ZYMV) proved to be established by some mechanism of RNA interference, since a translationally deficient gene cassette of the ZYMV CP gene equally generated high levels of protection (Chapter 6). As such, the general term coat protein-mediated protection (CPMP) suggesting a mechanism mediated by the accumulation of the coat protein, is heavily misleading, and should rather be replaced by coat protein genederived protection, or shorter capsid gene-derived protection (CGDP). A second example of engineered resistance mediated at the transcript level results from the transformation of tobacco with the nucleoprotein (N) gene from tomato spotted wilt tospovirus (TSWV), a negative strand RNA virus. The transgenic expression of a translationally defective N gene cassette similarly afforded high levels of resistance, reaching virtual immunity in transformant lines carrying homozygous copies of the N gene cassette (Chapter 7). Transformation of tomato with the same TSWV N gene likewise engendered high levels of resistance, culminating in the development of experimental hybrids that resisted TSWV infection upon continuous inoculation with the virus by the thrips vector in open field (Chapter 8). Hence, tomato transformant lines carrying the TSWV N gene make suitable progenitors for TSWV resistance that can be incorporated into classical breeding programs, as do tomato transformant lines carrying the CMV CP gene, or melon transformants carrying the ZYMV CP gene for resistance against their cognate viruses.The use of engineered resistance genes has advantages over the use of host genes for virus resistance. A unique feature of genetically engineered virus resistance is their source, being the viral genome. In fact, it is the viral pathogen itself that supplies the basic constituents for engineering virus resistance, which can be cloned and identified fairly easily. This is an essential difference from conventional breeding which, of necessity, is limited to host genes that can be introgressed from interfertile relatives. Moreover, one and the same gene construct can be applied in multiple crops to confer genetically engineered resistance against the virus from which the gene construct was derived. Once engineered resistance genes are incorporated into the crop species of interest, they can be forwarded into breeding programs and manipulated like any other Mendelian trait conferred by a single dominant gene. Their inheritance can easily be traced by means of simple molecular techniques such as Southern blot analysis or PCR analysis. Application of these techniques in backcross programs eliminates the need for repeated and laborious resistance screenings of progeny populations, and can even be adapted to discriminate between homozygous and hemizygous progeny plants. In this manner, the transgene functions as a molecular marker that is one hundred percent correlated to the resistance trait.In contrast to cultivar resistance genes, any recognition events based on highly specific interactions between the protein products from host resistance genes and viral avirulence genes (Keen, 1990; Dawson & Hilf, 1992), are not involved in transgenic virus resistance mechanisms. Consequently, engineered resistance genes are not likely to be easily overcome by mutant virus strains carrying point mutations, and as such may reasonably be expected to provide reliable sources of genetic resistance to viral infections. This reasoning especially applies to engineered resistance mechanisms mediated at the transcript level that do not imply the transgenic expression of any protein product at all. Nevertheless, one consideration that must be taken into account in the deployment of any monogenic resistance source, is the general experience that viral pathogens do vary and may overcome single genes in resistance breeding. The development of resistance by the virus would quickly negate the significant environmental benefits, including the reduced use of pesticides, inherent to the employment of genetic sources of virus resistance and irrespective of their nature. As such, it is important to stack multiple resistance genes operating at different levels, so that if the virus overcomes one level, it will be faced by other levels of protection. The combination of multiple sources of resistance might additionally yield higher levels of protection through complementation. In this perspective, one feasible combination would be the viral coat protein gene and a replicase-derived gene, thereby creating an oligogenic type of 'green' resistance based on different underlying mechanisms preventing the homologous virus from systemic infection and concomitant disease development. Alternatively, the combination of host and engineered resistance genes provides a tempting approach in developing oligogenic, highly durable sources of virus resistance.Although criteria for effective field resistance to viral infections can vary significantly between the crop and the virus concerned, it is of crucial importance to ascertain whether transgene expression effecting protection upon infection under controlled conditions commonly achieved by mechanical inoculation, also holds upon transmission by the natural vector and under field conditions. The practical value of genetically modified progenitors for virus resistance in fact can only be evaluated by extensive field testing, thereby merging the efforts of plant molecular biology and plant breeding. Ultimately, it is the breeder who should decide whether transgenic progenitors can be considered elite material that meets the demands for its incorporation into the crop breeding program. In this respect the genetic stability of the novel resistance trait and the overall field performance are critical factors inherent to the development of transgenic elite lines or a final transgenic cultivar. Unfortunately, relatively few studies of field performance of genetically engineered plants have been published to date, at least when considering the tremendous number of field trials conducted over the past few years (Ahl Goy & Duesing, 1995). Successful field testing of genetically modified crop cultivars not only provides proof of their superiority over existing cultivars, but will also contribute to demonstrate their environmental safety in order to diminish public concern and scepticism (Rogers & Parkes, 1995).In all, the ultimate commercialisation and profit of transgenic sources of virus resistance will depend on an array of factors including field performance, genetic stability, public acceptance and the resolution of environmental concerns and patent related issues. Comprehensive studies on the environmental impact, toxicity and other safety issues must first properly be addressed before releasing engineered virus resistant cultivars. As such, extensive field trials and associated studies are now required to adapt genetically engineered sources of virus resistance for their implementation into practical agriculture.