The viruses

The potato viruses discussed in this paper (and their abbreviations) are listed in Table 1. For a more comprehensive list, including symptoms, see Jeffries (1998). Most of these viruses cause yield losses, which vary between cultivars. Some also cause tuber defects, which can render the tubers unsaleable, including: spraing (necrotic arcs or rings in or on the tubers) resulting from TRV or PMTV; potato tuber necrotic ringspot disease resulting from PVYNTN; net necrosis resulting from PLRV in a few cultivars and deformed tubers resulting from PSTVd. Symptom severity also depends on whether the infection is primary (current season infection) or secondary (tuber-borne).

Table 1 The potato viruses mentioned in this paper
Full size table

PLRV is probably the most damaging and widespread virus of potato and is found wherever potato crops are grown. PVY is next in importance and although it tends to cause less damage than PLRV in some cultivars, it can cause severe damage in others (Beemster & Rozendaal, 1972), and isolates of the PVYNTN subgroup cause severe symptoms. PVY is also widespread, and in some parts of Europe it is more common than PLRV. The viruses PVA and PVV are related to PVY, but are less commonly reported. PVA occurs worldwide (except for the Andes) and causes symptoms that range from very mild to fairly severe. PVV is relatively restricted in its distribution, and tends to cause few symptoms. PVX occurs worldwide and is also important, although its symptoms tend to be mild. However, mixed infections with PVX and other viruses are very damaging. Other viruses that we shall consider, such as PVS, PVM, TRV and PMTV, tend to be locally important. For example, PMTV seems to be restricted to areas with cooler climates such as the Andean region of South America, Canada, China, Japan and Northern Europe. TRV is prevalent in light sandy soils that favour the trichodorid nematode vectors. PSTVd occurs in North and South America, China and parts of Eastern Europe but is considered to be nonindigenous in most of Western Europe.

The need for resistance

Potato is the world’s fourth major food crop. Viruses are a serious problem, not only because of effects caused by primary infection, but also because the crop is vegetatively propagated and they are transmitted through the tubers to subsequent vegetative generations. Crop yield losses resulting from PLRV, PVY and PVX in the UK were estimated, using 1982 prices, at £30–50 million during a year of average infection (Hull, 1984). The potential net gain from using cultivars resistant to these three viruses has been estimated at 22% in Mexico (where only a quarter of ware hectarage uses certified seed) (Quaim, 1998). Yield losses attributable to PLRV worldwide have been estimated at 20 million tonnes per year (Kojima & Lapierre, 1988). Losses can take the form of reduced yield, downgrading of seed crops, and/or tuber blemishes (described above) but, in most countries, the true cost has also to be seen in terms of expensive control measures. These include seed potato production/certification schemes (with the attendant costs of roguing, crop inspections and diagnostics including the increasing use of postharvest tuber testing); the use of pesticides to kill vector aphids (for PLRV control) or nematodes (for TRV control) and other measures such as mineral oil sprays to prevent nonpersistent aphid transmission of PVY. Control of aphids and nematodes by pesticides is expensive, and likely to become more difficult with increasing insecticide resistance in aphid populations, any long-term climatic changes leading to changes in vector populations, concern about environmental effects of pesticides, and consumer concern about pesticide residues in food. The potential benefits of virus resistance are therefore great, because resistant cultivars are the most economic and environmentally acceptable way of controlling virus diseases of potato.

Nomenclature of virus resistance genes

This paper follows the conventions described in Solomon-Blackburn & Barker (2001) for resistance gene nomenclature, indicating the type and origin of resistance and the virus or strain resisted.

Types of resistance

The nomenclature for the responses of plants to virus infection was reviewed by Cooper & Jones (1983), and the types of host resistance in potato by Valkonen (1994) and Valkonen et al. (1996). The following are descriptions of different types of responses that occur in resistant potato plants. Most can be applied to host gene-mediated and transgenic resistance.

Extreme resistance (ER) and hypersensitive resistance (HR) are described in Solomon-Blackburn & Barker (2001). Host gene-mediated ER prevents virus multiplication at an early stage of infection and is not normally associated with cell death (Hämäläinen et al., 1997; Gilbert et al., 1998). HR is well studied and can be considered as a rapid defence response that results in death of a few cells (necrosis) at the site of infection which prevents the infection from spreading further (Dixon et al., 1994).

Resistance to infection can be defined as when the likelihood of infection by natural means (i.e. from vector inoculation) is reduced in resistant plants. Such resistance was called ‘field resistance’ by Cooper & Jones (1983). Infection resistance can also be manifested in plants that are resistant or unattractive to the vector (e.g. aphids) itself (Flanders et al., 1992).

Plants with resistance to virus accumulation are infectible but the virus reaches only a relatively low concentration in the plant. The term ‘moderate resistance’, adopted by Valkonen et al. (1996) for this type of resistance, is not used here, as it could equally apply to several other types of resistance (including resistance to infection).

Resistance to virus movement can be expressed in a number of ways. For example, it has long been recognized with respect to PLRV (Hutton & Brock, 1953; Barker, 1987) that certain potato clones produce a lower percentage of virus-infected daughter tubers from an infected plant, than other clones do. This could be the result of impeded virus movement. HR could also be regarded as resistance to virus movement, because it limits spread within the plant following the initial infection.

Cooper & Jones (1983) defined tolerance as when there is little or no apparent effect on the infected plant, i.e. infected plants are symptomless or have reduced symptom expression in the presence of disease. Tolerance is not synonymous with resistance, but over the years breeders have sometimes used the two terms interchangeably and without definition which can lead to confusion. Tolerance can be better defined as resistance to disease (the effects of the pathogen), but tolerant plants are often susceptible to infection with the pathogen which multiplies extensively in the host.

Sources of resistance

Host resistance

Introduction.

Potato breeding has been reviewed by Bradshaw & Mackay (1994) and potato breeding for virus resistance by Davidson (1980), Ross (1986) and Świeżyński, 1994. Sources of host resistance in S. tuberosum include existing cultivars and improved breeders’ clones. Some clones contain multiple copies of a resistance gene (multiplex state, see below) and in others different virus resistances have been combined (Ross, 1978; Świeżyński, 1983, 1994; Dziewońska, 1986; Mendoza et al., 1996; Solomon-Blackburn, 1998). These represent the efforts of a succession of breeders over many years and provide a large pool of parental material for future breeding. Known resistance genes are also available in wild species that exist in various collections around the world or are in the wild. New resistance genes are still being discovered (Barker, 1996; Hämäläinen et al., 1998) and there are, no doubt, many others as yet undiscovered.

Introduction of potato to Europe and its early decline in vigour.

The history of the potato crop has been reviewed by Bradshaw & Mackay (1994) and Glendinning (1983), and by Davidson (1980) with respect to virus problems. The modern Tuberosum potato (S. tuberosum ssp. tuberosum) was derived from (probably two) introductions of cultivated Andigena (or S. tuberosum ssp. andigena) to Europe from South America in 1570. For many years, there were serious problems with ‘degeneration’ in the crop, which worsened with successive generations. This was finally recognized by Salaman (1921) to be the result of virus infection. This led to targeted breeding for virus resistance, the introduction of virus disease regulations into the UK Statutory Seed Certification Scheme, and then protected regions for seed production. Similar measures were adopted in many other countries.

Early resistance breeding programmes.

The recognition of the virus problem was followed, from the 1930s onwards, by much research on the characterization of numerous virus resistance genes and phenotypes in Solanum species (e.g. Schultz & Raleigh, 1933; Salaman, 1938; Cadman, 1942; Cockerham, 1943a, b; Cockerham, 1945; Stelzner, 1950; Hutton, 1951; Cockerham, 1952; Hutton & Wark, 1952; Ross, 1952; Ross, 1954a, b; Cockerham, 1955; Cockerham, 1958; Easton et al., 1958; Ross, 1958; Ross, 1961; Mills, 1965; Cockerham, 1970). Comprehensive lists of known genes for resistance to potyviruses and PVX, their known relationships and mapped locations, are presented in Solomon-Blackburn & Barker (2001). However, before the virus problem was recognized as such, there had no doubt been selection for virus resistance within the Tuberosum gene pool by selecting cultivars that withstood degeneration better than others. A number of HR genes were already present, including Natbr, Nctbr, Nxtbr and Nbtbr (Cockerham, 1943b). Wild species from South America had been used since 1851 for breeding, and species including the hexaploid S. demissum were used as sources of blight resistance in the UK by Salaman in 1909 (Bradshaw & Mackay, 1994) and by Wilson before 1920 (Davidson, 1980); this probably also introduced some virus resistance. Two of Wilson’s cultivars, Templar and Crusader, were reputed to have resistance to PLRV (Davidson, 1980). Ross (1986) reported that PLRV resistance in some modern German cultivars could be traced back to S. demissum crosses. The HR gene Nytbr in the Australian clone 11–79 bred by Hutton from cvs Katahdin and Snowflake, was a later introduction to the U.K. (Davidson, 1980). As it was already in a Tuberosum background of cultivar quality, its use did not involve introgression from a wild species.

Breeding for HR and ER.

With the identification of HR and ER genes, deliberate attempts have been made to breed for and select such genes in breeding progammes. Greater emphasis has been placed on utilizing sources of HR and ER rather than other forms of resistance to potyviruses and PVX, probably because selection for this type of resistance is relatively straightforward and can be carried out in the glasshouse following simple mechanical inoculation tests, and because HR and ER are effective, quite durable and simply inherited. Where the inheritance of ER or HR to viruses has been investigated in potato (reviewed for potyviruses and PVX in Solomon-Blackburn & Barker (2001)), it is monogenic, with dominance for resistance with one exception, the recessive gene s (or sstbr in the nomenclature used here), which confers resistance to PVS infection (Bagnall & Young, 1972). There are also dominant genes for HR or ER to PVS or PVM Baerecke, 1967; Sacut;wieżyński et al "1993"

ER can be comprehensive (effective against several strains of a virus or even several viruses), e.g. various Rx and Ry genes from S. tuberosum ssp. andigena, S. acaule and S. stoloniferum (reviewed in Solomon-Blackburn & Barker, 2001). Comprehensive ER has obvious advantages over specific HR for breeding purposes, but HR also provides useful protection and is more widely available in cultivars than ER.

Attempts to introgress ER to PVX have been made since about 1952 (Ross, 1954a,; Ross,1986 and to PVY since 1944 Ross, 1978; Davidson, 1980. Relatively few cultivars have ER to PVX and PVY, probably for a variety of reasons. First, in comparison with the ubiquitous HR genes, ER genes are generally much more recent introductions to the S. tuberosum gene pool and were introduced from relatively few sources. S. stoloniferum, the source of genes Rysto and Rystona, does not intercross freely with S. tuberosum as the two species differ in endosperm balance number (EBN) (reviewed by Hawkes, 1994). One of the more frequently used Ry parents, the hybrid MPI 61.303/4, had S. tuberosum ssp. andigena, S. acaule, S. demissum and S. spegazzini as well as S. stoloniferum in its pedigree (Ross, 1986). The Ry cultivars listed by Ross were third to seventh backcross generations from S. stoloniferum. The S. stoloniferum source material was generally highly susceptible to PLRV; PLRV resistance and ER to PVX were introduced to S. stoloniferum breeding lines in the U.K. from other sources (Cockerham, unpubl. report). However, to date, no U.K.-bred cultivars are known to have an Ry gene. This may reflect a higher priority given to cultivars resistant to PVY (including PVYN strains) in the countries where Ry cultivars have been produced (Germany, Holland, Poland and Hungary) and possibly an association of Ry genes with wild characteristics. Furthermore, (Świeżyński, 1994) and Ross (1986) reported that most breeding lines with the gene Rysto were male-sterile. It is also difficult and time-consuming to select ER when HR is also present in the population. Cultivars with Rxtbr, Rxadg or Rxacl have been produced in several countries including USA, Germany, Argentina Ross, 1986,UK and Ireland.

Major gene resistance to viruses in potato has proved quite durable, unlike major-gene resistance to late blight (Phytophthora infestans). Some evolution of resistance-breaking strains is apparent (Jones, 1982, 1985), but it is a relatively slow process and resistance-breaking strains often do not become prevalent (Harrison, 1981): the HR genes Nxtbr (Cadman, 1942; Cockerham, 1970) and Nytbr (Davidson, 1980; Jones, 1990) still confer very useful resistance, though neither offers resistance to all strains.

Resistance to infection.

Resistance to infection is, by its nature, somewhat difficult to select and there is a requirement for plants to be exposed to the viruliferous vector in order to assess this type of resistance. In most cases, this means that expensive and laborious field trials are necessary. Resistance to infection with PLRV appears to be inherited polygenically (Ross, 1958; Baerecke, 1961; Davidson, 1973). It could include several different mechanisms; this could contribute to the polygenic nature of this trait (or traits). Brown et al. (1997) suggested that the most resistant S. tuberosum parents might have major genes for resistance to PLRV infection. Quantitative resistance to PVY infection, from S. phureja, has been reported (Davidson, 1980), but it dissipated in outcrossing and would be difficult to assess accurately enough for selection. It had been thought that this quantitative resistance to PVY might be more durable than major gene resistance, but ER and HR have proved quite durable and offer a higher degree of protection.

Resistance to virus accumulation.

Resistance to virus accumulation is readily assessed by measuring the virus titre using a quantitative ELISA of tissue from glasshouse- or field-grown plants. A high level of resistance to PLRV accumulation in S. brevidens was identified by Jones (1979). Gibson et al. (1990) also found a high level of resistance to accumulation of PVX and PVY, as well as of PLRV in S. brevidens. The results of attempts to transfer this resistance into a S. tuberosum background by somatic hybridization suggested that the gene controlling resistance to PLRV is different from those controlling resistance to PVY and PVX, and that genes conferring resistance to PVY and PVX are linked (Valkonen et al., 1994). The resistance to PVY and PVX in S. brevidens is thought to be associated with slow cell-to-cell spread of the viruses (Valkonen et al., 1991).

Resistance to PLRV accumulation was found in S. tuberosum breeding lines with other species in their pedigrees (Barker & Harrison, 1985), and in a range of potato material by other authors (Gase et al., 1988; Świeżyński et al., 1988; Van Den Heuvel et al., 1993; Wilson & Jones, 1993). The most resistant tetraploid clones had 1–5% of the PLRV concentration found in susceptible clones. Virus is less likely to be acquired (Barker & Harrison, 1986) and transmitted to other plants by aphids from plants with this type of resistance (Barker & Woodford, 1992), which is therefore useful for reducing virus spread within the crop. Barker & Harrison (1985) originally referred to this form of resistance as restricted virus multiplication, but the term ‘resistance to virus accumulation’ allows for the possibility that it may sometimes be a consequence in part, for example, of restricted movement (Wilson & Jones, 1992) rather than restricted virus multiplication per se. In inheritance studies on resistance to PLRV accumulation in S. tuberosum, a dominant major-gene effect has been reported, possibly involving two complementary genes both needed for resistance (Barker & Solomon, 1990; Barker et al., 1994a). This hypothesis is currently under test.

Sources of PLRV resistance have been found in other Solanum species. Very strong resistance to PLRV accumulation found in S. chacoense appeared to be controlled by a single dominant major gene (Brown & Thomas, 1994) although the possibility of a tightly linked group of genes was not ruled out. This resistance may be useful as S. chacoense crosses readily with S. tuberosum, bears tubers and is diploid, which may facilitate the identification of linked molecular markers. Resistance to PLRV accumulation has also been found in S. etuberosum (Chavez et al., 1988a). A similar resistance was found in S. phureja by Franco-Lara & Barker (1999), who suggested that this species may be the most valuable of these three diploid species because it is already cultivated and well recognized as a source of desirable characteristics for potato germplasm improvement.

Resistance to virus movement.

Resistance to virus movement may underlie the mechanism of several forms of resistance. Resistance to accumulation of PVY and PVX in S. brevidens is thought to be the result of slow cell-to-cell spread (Valkonen et al., 1991), and resistance to PLRV in S. tuberosum clones may be the result of impaired movement of virus from sieve elements to companion cells in the external phloem bundles (Derrick & Barker, 1997).

A different form of resistance to virus movement has long been recognized with respect to PLRV (Hutton & Brock, 1953; Barker, 1987) in certain potato clones where infected plants transmit PLRV infection to fewer of the tuber-progeny plants than in susceptible clones. It is associated with phloem necrosis in cv. Bismark (Hutton & Brock, 1953) and cv. Apta (Golinowski et al., 1987), but not in the breeding clone G7445(1) (Barker, 1987). Resistance to phloem transport has been demonstrated in cv. Bismark by Wilson & Jones (1992), who found it to be separate from resistance to accumulation and resistance to infection. Świeżyński et al. (1989) found a high level of PLRV resistance associated with limited virus spread in four diploid potato clones. Resistance to PLRV movement was reviewed in greater detail by Barker & Waterhouse (1999).

Tolerance.

Resistance is obviously preferable to tolerance, where available, but where the choice of resistant cultivars is limited, a tolerant cultivar may be preferred for reasons other than virus resistance (e.g. yield, quality, maturity, other disease resistance). However, there are dangers in selecting tolerant lines, because of the risk of virus spread from infected symptomless stocks grown in proximity to healthy material, or of introducing a soil-borne virus to sites that were previously uncontaminated.

The tolerance problem can be illustrated with the types of reaction of potato clones to TRV. Some clones are susceptible but show no spraing symptoms, and infective virus is found in tubers. Other clones are hypersensitive to TRV (‘susceptible’ to spraing disease) and react by producing spraing symptoms (an HR response) in tubers. Although virus is detectable, infective virus often cannot easily be recovered from tubers of these clones (Xenophontos et al., 1998). When clones with ER are planted in infested soil, tubers do not develop spraing symptoms and no virus can be recovered (Robinson & Dale, 1994). The tolerant clones that become infected but produce no symptoms provide a means by which new virus-free sites could become infected if the vector nematodes are present (Xenophontos et al., 1998).

Summary of available host resistance to each virus.

There is HR and ER to PVY, PVX, PVA, PVV, PVM and PVS in breeding lines, wild species and cultivars. Some of these resistances have been used for many years in cultivars to good effect. There is resistance to PLRV infection, accumulation or movement in a few cultivars and, to a greater extent, in breeding lines and wild species. ER to TRV exists in some cultivars and tolerance exists in others. There are no reports of confirmed host resistance (as opposed to tolerance) to PMTV, and few of host resistance (in wild species) to PSTVd (Bagnall, 1972; Salazar, 1981).

Pathogen-derived transgenic resistance

Introduction.

There are many examples of pathogen-derived resistance to potato viruses, especially coat protein-mediated, which was developed first. Genetic engineering of plants for virus resistance was reviewed by Harrison (1992), Fitchen & Beachy (1993), Wilson (1993), Baulcombe (1994b), Pierpoint (1996) and Fuchs et al. (1997), and specifically to PLRV by Barker & Waterhouse (1999).

Coat protein-mediated resistance.

Transformation with viral coat protein genes was the first approach adopted for pathogen-derived resistance because it was thought it might work like cross-protection (Pierpoint, 1996). Transformation with the coat protein gene of PVX was one of the first attempts to obtain pathogen-derived resistance to a major potato virus (Hemenway et al., 1988). Numerous other examples of coat protein-mediated resistance (CPMR) to potato viruses followed soon after. In many cases, the transgenes were tested in a model species, frequently tobacco, but in many other instances potato was transformed and tested. In some cases, but not all, the resistant transgenic plants produce coat protein. The resistance varies in degree and type but often acts against virus accumulation. Kawchuk et al. (1991) observed PLRV titres as low as 1% of the level found in control plants at the stage of primary infection. Either immunity or extreme resistance has also been reported in potatoes with CPMR, e.g. to PVY (Lawson et al., 1990; Hassairi et al., 1998). CPMR is often strain specific but sometimes not. For example, Okamoto et al. (1996) found that transformation of potato cv. Folva with a coat protein (CP) gene from a PVY strain of the PVYN subgroup, conferred extreme resistance to isolates from the PVYN, PVYO and PVYNTN subgroups. Sometimes it is effective against other viruses, at least within the same virus genus: potato cvs Spunta and Claustar transformed with a CP gene from Lettuce mosaic virus were immune to PVY (Hassairi et al., 1998). However, broad-spectrum CPMR in tobacco (Nicotiana tabacum) was found in two cases to take the form of reduced or delayed symptom development (Anderson et al., 1989; Stark & Beachy, 1989), thus conferring little useful benefit.

In some cases, CPMR can be overcome by a high concentration of inoculum, e.g. delayed accumulation of PVA in potato transformed with a CP gene from PVY (Hefferon et al., 1997). CPMR sometimes confers resistance to inoculation with viral RNA, e.g. in cases of CPMR to PVX (Hemenway et al., 1988), PVS (MacKenzie & Tremaine, 1990) and PMTV (Barker et al., 1998b). In some cases CPMR was proportional to the amount of CP produced in transformed tissue (Hemenway et al., 1988) but not in other cases (Angenent et al., 1990). In several cases the resistance is effective against manual inoculation and vector-transmission (Lindbo & Dougherty, 1992; Van Der Vlugt & Goldbach, 1993; Reavy et al., 1995). However, there are examples in which CPMR was effective against manually inoculated virus but not against vector-borne virus, e.g. a case of PVY resistance (Lawson et al., 1990), and one of TRV resistance (Ploeg et al., 1993).

Resistance is conferred in some cases by nontranslatable CP genes (e.g. antisense sequences) or by truncated versions of the viral CP gene (Van Der Vlugt et al., 1992). In many such cases it is presumed that resistance is the result of CP RNA transcript rather than CP itself.

Some clones with CP genes have been tested in field trials, which have confirmed the resistance identified in glasshouse tests, and that some clones retain the intrinsic properties of the cultivar. These include trials for resistance to PLRV in Australia (Graham et al., 1995), the USA (Thomas et al., 1997) and Canada (Kawchuk et al., 1997), to PVY in Switzerland (Malnoe et al., 1994) and to PVX in the Netherlands (Van Den Elzen et al., 1993).

Movement protein-mediated resistance.

Virus-encoded movement proteins, which facilitate spread of virus through the plant, have also been targeted as potential sources of resistance transgenes. Tacke et al. (1996) transformed potato with a sequence from PLRV open reading frame (ORF) 4, which encodes a protein (pr17), believed to be a phloem-specific movement protein. They used mutant versions and wild-type (wt) genes. PLRV accumulation was reduced in transgenic plants with secondary infection. Lines expressing mutant pr17 were also resistant to PVY and PVX accumulation. Transgenic lines that did not express mutant protein but in which its transgene mRNA accumulated were resistant to PLRV but not to PVY or PVX. Transgenic plants expressing wt pr17 were also resistant to PLRV only. Tacke et al. (1996) concluded that PLRV resistance was operating at the RNA level and that PVX and PVY resistance was protein-mediated. They suggested that mutant pr17 proteins might be binding to plasmodesmata and inhibiting cell-to-cell movement of unrelated viruses, possibly by blocking binding sites for their movement protein. Herbers et al. (1997) found plasmodesmatal alterations in phloem but not mesophyll tissue in tobacco plants transformed with the same construct. They also found altered carbohydrate metabolism in these plants. Seppänen et al. (1997) transformed potato with mutant versions of two genes from the triple gene block (TGB) of PVX (the TGB proteins provide the movement function in PVX and some other viruses). In inoculation tests, transgenic plants were resistant to PVX, Potato aucuba mosaic virus, PVS and PVM (which have a TGB), but not to PVY (which does not). The resistance was to virus movement and accumulation, and was dependent on inoculum concentration. If these forms of resistance can be shown to be effective in other cultivars and in the field, they have the potential to provide a valuable type of broad-spectrum resistance.

Polymerase-mediated resistance.

Expression of sequences from a replicase gene was first attempted by Golemboski et al. (1990), with Tobacco mosaic virus. Similar approaches have been tried subsequently with other RNA viruses using both mutant and functional forms of RNA polymerases. In general, resistance generated by this route is characterized by strong inhibition of virus replication and can be very effective against high inoculum levels (Braun & Hemenway, 1992; Mueller et al., 1995). It operates in protoplasts and against viral RNA, but tends to be highly strain-specific. In some cases, e.g. with PVX resistance in tobacco, the most resistant plants showed the lowest levels of replicase (polymerase) protein synthesis (Longstaff et al., 1993), suggesting that resistance may be mediated by mRNA rather than protein, at least in these cases (Baulcombe, 1994a). This was subsequently confirmed by Mueller et al., 1995 (see below).

Resistance to PLRV mediated by a region of the PLRV polymerase gene has been produced in potato by Monsanto, USA (Monsanto, 1994; Shah et al., 1994, 1995) and CSIRO, Australia (Waterhouse, see Barker & Waterhouse, 1999). Monsanto produced a line of cv. Russet Burbank, expressing the whole polymerase (ORFs 2a and 2b), that is highly resistant or immune to PLRV infection, and potato lines (NatureMark’s NewLeaf Plus) that contain the PLRV replicase gene and also express the Bacillus thuringiensis insecticidal protein (Thomas et al., 1995).

Ribozyme-mediated resistance.

Ribozymes are small RNA molecules derived from the satellite RNA of Tobacco ringspot virus, certain viroids and viroid-like satellite RNAs that are capable of cleaving specific RNA molecules in trans. One particular success is the use of a ribozyme designed to target minus-strand PSTVd RNA to transform potato cv. Désirée (Yang et al., 1997). In inoculated plants of most transgenic lines, PSTVd was not detectable; in other lines viroid accumulation was reduced. Vegetative progeny stably inherited the resistance. This was the first example of a ribozyme suppressing a viroid to an undetectable level in planta.

Post-transcriptional gene silencing.

In a number of cases, resistance conferred by various virus-derived transgenes is thought to be mediated by RNA transcript rather than by the encoded protein, as mentioned above under the CPMR heading. Another example is the strain-specific (homology-dependent) PVX resistance conferred in tobacco by the RNA polymerase gene of PVX (Mueller et al., 1995), where the highest resistance is associated with low polymerase accumulation and multiple copies of the transgene, and resistance is also conferred by untranslatable versions of the transgene. The mechanism underlying this form of resistance appears to be a form of post-transcriptional gene silencing (PTGS) (Mueller et al., 1995; Baulcombe, 1996; Van Den Boogart et al., 1998). PTGS is characterized by the highly specific degradation of the transgene mRNA and target RNA. For PTGS to operate, the sequences of these two RNAs must be the same or highly homologous (English et al., 1996). The mechanisms underlying the induction and control of the degradation pathway are not yet understood, although they are much discussed (for reviews see: Bruening, 1998; Grant, 1999).

Recent work by Waterhouse et al. (1998) using the PVY protease gene in sense [s] and antisense [a/s] forms has produced some very intriuging results that may go a long way to explain PTGS and also indicate how the frequency of obtaining PTGS (and strong resistance) in transgenic plants can be improved. Waterhouse et al. (1998) found that the expression of both [s] and [a/s] homologous transcripts leads to much higher frequencies of silencing than in plants transformed with the same gene expressed in only one orientation, thus 44–54% of [s] + [a/s] transgenic lines were resistant or immune, compared with less than 15% of lines expressing only [s] or [a/s]. The immunity conferred was not correlated with multiple loci or multiple transgenes at a single locus. This immunity is stable and inherited in a Mendelian way. Waterhouse et al. (1998) concluded that the use of RNAs with the potential to form duplexes ([s] + [a/s]) may be an important new strategy for virus resistance in transgenic plants.

It is intriguing that PTGS may have parallels with natural general antiviral defence in plants where virus-induced gene silencing is believed to have a role (Dawson, 1996; Carrington & Whitham, 1998; Baulcombe, 1999): it has been suggested that virus-induced gene silencing is part of a defence mechanism mediated by viral RNA in plants, which slows down and stops virus accumulation at a certain stage of infection. It has also been suggested that viruses evolve strategies that overcome this (Carrington & Whitham, 1998), e.g. the helper component proteinase HC-Pro in potyviruses (Kasschau et al., 1997), which mediates synergistic viral disease (Shi et al., 1997).

In research with transgenes other than those for virus resistance, potyviruses including PVY produced strong suppressors of PTGS, whereas PVX was not able to reverse PTGS in transgenic plants (reviewed by Baulcombe, 1999). Béclin et al. (1998) suggested that potyviral infection could reverse RNA-mediated resistance to other viruses in a transgenic variety. Hence, PVY resistance could be important for the effectiveness of RNA-mediated resistance to other viruses, and it would be a wise precaution to ensure that viral genes from potyviruses used for transformation do not include the sequence for HC-Pro, which suppresses gene silencing (Anandalakshmi et al., 1998), in translatable form.

Ecological impact of transgenic potatoes.

It is not our intention to discuss the pros and cons of GM crops, but it is relevant to consider potential risks that may arise from recombination between transgene RNA and RNA of a challenge virus. It is generally acknowledged that this is the most important risk area with the use of viral genes. Thus, it is proposed by some that new viruses or strains could arise by recombination between RNA of infecting viruses and virus-derived transgene RNA. The likelihood of such recombination relative to that of natural recombination between viruses is difficult to determine (Aaziz & Tepfer, 1999), although one report suggested it is common (Malnoe et al., 1997) and likely to result in the survival of natural mutant viruses. However, mutants and recombinant viruses with transgene RNA are likely to be less fit than wild-type viruses (Falk & Bruening, 1994), which have been naturally selected (after natural recombination). Indeed, new variant strains have been detected in plants with host resistance (Dawson & Hilf, 1992), although these are generally less fit than the original virus in susceptible plants. Aaziz & Tepfer (1999) suggested that RNA-mediated resistance based on PTGS could present certain advantages from the biosafety perspective because little or no transgene-derived RNA accumulates, and hence the likelihood of its recombination with the RNA of infecting viruses is expected to be extremely low.

Nonpathogen-derived forms of novel resistance

Nonpathogen-derived transgenic resistance.

A number of alternative means by which transgenic resistance can be induced using nonpathogen-derived sequences have been reported. Several of these are reported to induce broad-spectrum (nonspecific) resistance which may have distinct advantages, particularly if combined with other more specific forms of transgenic resistance.

Pokeweed antiviral protein (PAP) is a ribosome-inhibiting protein found in the cell walls of pokeweed (Phytolacca americana). Plants of potato cv. Russet Burbank transformed with cDNA to the PAP gene showed partial resistance to PVY and PVX infection by mechanical inoculation and to PVY infection by aphids (Lodge et al., 1993).

Another nonpathogen form of transgenic resistance has been obtained using the gene for mammalian 2′-5′ oligo-adenylate synthetase. This gene is activated by dsRNA and it, in turn, activates a latent endoribonuclease which degrades viral RNAs (Truve et al., 1993). Potato cv. Pito was transformed with a copy of a rat gene encoding the 2′-5′ oligoadenylate synthetase, and some transformants showed some resistance to PVX infection and/or accumulation.

PSTVd infection and accumulation was suppressed in transgenic potato cv. Russet Burbank expressing the yeast-derived RNA-specific ribonuclease pac1 (Sano et al., 1997). All the progeny tubers were viroid-free. It was suggested that the pac1 gene product digested double-stranded regions in the PSTVd, formed transiently during replication. This could be a useful form of resistance if it proves to be stable and effective in field conditions.

Plants transformed with an antibody gene can synthesize protein chains of mammalian antibodies to a plant virus (Hiatt et al., 1989) and assemble them into functional complexes. Plant expressed antibodies to virus particles have been shown to induce resistance (Tavladoraki et al., 1993; Zimmermann et al., 1998). However, this approach is in its infancy and it remains to be seen whether plant-produced antibodies can provide useful resistance.

Somaclonal variation.

Somaclonal variants (produced by regenerating plants from isolated protoplasts) were investigated as a means of generating novel phenotypes in potato (Shepard et al., 1980). It has been found that somaclonal variants may be a common by-product of attempts to generate transgenic lines of potato by transformation (Jongedijk et al., 1992). Presting et al. (1995) transformed potato cultivars Russet Burbank and Ranger Russet with the native PLRV CP gene, a modified form to optimize protein expression, and a control consisting of a vector plasmid only. Resistant lines were obtained from transformations with all three constructs, the most resistant resulting from transformation with the control construct. This resistance was dominant and simply inherited and Presting et al. (1995) suggested it could be the result of somaclonal variation. The value of deliberate selection of somaclonal variants to obtain improved clones has been proven by Kawchuk et al. (1997), who obtained a somaclonal variant of Russet Burbank with resistance to PLRV accumulation and to tuber symptoms. It remains to be seen whether such an approach will be beneficial for obtaining resistance to other viruses.

Breeding techniques and strategies

Methods of screening

Introduction.

In order to select for virus resistance genes in a breeding programme, it is necessary to assess resistance in plants. Although linked molecular markers present the possibility of selection on the basis of genotype rather than phenotype in the future (discussed later), traditional forms of screening for resistance are still essential in general where selection is needed.

Field exposure trials.

For many purposes a field exposure trial to determine virus resistance is the only way one can be sure of obtaining results that are relevant to the crop situation. For some forms of resistance, glasshouse or screenhouse testing might be used in the initial stages, but the final confirmation and proof that the resistance is operating satisfactorily, is likely to come from a field trial. Some forms of resistance can only be assessed in the field. Thus, resistance to PLRV infection is usually assessed by field exposure trials (Davidson, 1973), where the virus is spread from infector plants to exposed plants by aphids (predominantly Myzus persicae), and the frequency of secondary infection in tuber progeny plants is assessed by symptoms, ELISA or a combination of the two, in comparison with standard cultivars. For selection purposes, it is convenient to expose the plants to PVY at the same time (Davidson, 1973), because it is also aphid-transmitted. In general, field trials are grown in areas of high natural aphid infestation, but are more successful in some years than others, because the levels of aphid infestation vary (though natural infestation can be supplemented by applying cultured aphids to infector plants). Thus, estimates of quantitative resistance to PLRV infection can only be rough. However, all exposed plants of the most resistant clones regularly remain PLRV-free, at least in U.K. conditions (Solomon-Blackburn & Barker, 1993).

Field exposure is necessary if quantitative field resistance to PVY is being assessed. It is also useful for confirming the resistance of clones already selected for major-gene resistance (or the unscreened progeny of multiplex parents, see below); most clones with ER or HR genes to PVY remain healthy in field exposure conditions, or with a low frequency of infection, e.g. 4% in clones with the HR gene Nytbr, but some develop systemic necrosis.

Screening in the glasshouse.

Field trials for virus resistance are expensive, and therefore substantial cost savings can be made if satisfactory or preliminary tests can be performed on glasshouse-grown plants. Thus, for major gene resistance to manually transmissible viruses, such as PVX and PVY, glasshouse–grown plants can be inoculated with infective leaf extracts and the response observed over the following weeks. Local or systemic necrosis (indicating HR) or mosaic symptoms are recorded, and noninoculated leaves tested for infection using serological or biological methods, e.g. S. demissum detached leaflet tests for PVY and PVA (Cockerham, 1970). However, a null response to sap-inoculation could be attributable either to ER or to failure of inoculation. Alternatively, plants can be graft-inoculated using infected scions, but although this is a more reliable method, it is more laborious.

Large numbers of young seedlings at the cotyledon stage in seed pans can be inoculated with PVX or PVY, using infective plant extracts mixed with an abrasive in a spray gun (Wiersema, 1972). Susceptible seedlings (showing mosaic symptoms) can be discarded from about two weeks after inoculation. This rapid inoculation method can be used for early selection or for progeny tests (see below). However, to minimize the possibility of ‘escapes’ from inoculation, an additional (manual) inoculation two weeks later is advisable.

Resistance to PLRV accumulation can be assessed by first graft-inoculating glasshouse-grown plants and then determining the virus titre in the leaves of glasshouse-grown daughter plants the following year, using a quantitative ELISA technique, and comparing with resistant and susceptible controls (Barker & Harrison, 1985).

Resistance and tolerance to PMTV and TRV can be assessed by field trials on infective land (with viruliferous Spongospora subterranea or trichodorid nematodes), and scoring the tubers for severity and frequency of spraing symptoms, but problems of uneven distribution of the viruliferous vector in the soil and variable weather conditions make such trials difficult (Dale & Solomon, 1988). A glasshouse test was developed to overcome some of the problems with TRV field trials (Dale & Solomon, 1988). Tubers are grown in pots of tested field soil containing viruliferous nematodes, and daughter tubers scored for the severity of spraing symptoms (but not frequency). In order to distinguish between resistance and tolerance, it is necessary to test symptomless tubers for infection. Transgenic resistance to PMTV was evaluated in a screen house trial by Barker et al. (1998a) using infested field soil and irrigation. Such glasshouse tests can overcome the problems of weather and patchy distribution of viruliferous vectors, and hence need less replication.

Progeny testing.

Progeny tests are used to determine the relative breeding value of parents or crosses, the gene dosage in a resistant parent (see multiplex parents below), or to select the best progenies. This type of screening is carried out on glasshouse-grown plants.

To determine the number of copies of a major gene for resistance to PVY or PVX in a potato clone, it is test-crossed with a susceptible parent. A sample of each progeny is sown, then spray-inoculated as described above. The gene dosage in the resistant parent is deduced from the segregation ratio of resistant to susceptible seedlings in the progeny (Solomon-Blackburn & Mackay, 1993).

Another application of progeny tests was developed by Chuquillanqui & Jones (1980), where families of true seedlings were assessed by aphid inoculation in a screen house in Peru, to identify progenies with resistance to PLRV infection. This method was adapted for experimental use in Scotland, using caged aphids in a glasshouse, and compared with a field trial (described above). The seedling progeny test did distinguish between resistant and susceptible progenies, but was not effective for selecting the more resistant individuals within progenies (Solomon-Blackburn et al., 1994).

Breeding strategies

Development of parents with multiplex resistance genes.

One of the difficulties with potato breeding is that so many different characteristics are desirable in a cultivar. High yield and good quality characteristics are now a prerequisite in many countries, and breeders have the task of combining these with resistance not only to virus diseases but also to a range of other diseases and pests. Added to this is the complication of tetrasomic inheritance, because S. tuberosum is autotetraploid.

One approach to these problems has been the production of multiplex resistant parents (Wastie et al., 1992; Bradshaw & Mackay, 1994). The term multiplex is used here to mean triplex or quadruplex, i.e. with three or four copies of the (dominant) resistance gene. When a multiplex resistant parent is crossed, all of the progeny, or (in the triplex case) nearly all, allowing for some double reduction (Mendoza et al., 1996), will be resistant, even when the other parent is susceptible (Cadman, 1942). By using these multiplex parents, screening and selection of the progeny for resistance is avoided, resources are saved by not having to raise seedlings that will be discarded as susceptible, and the chances of combining the virus resistance with the many other desirable attributes for potato cultivars are increased, because these parents can be crossed with parents that have other attributes but not virus resistance, and still produce all resistant progeny.

By cycles of crossing, test-crossing and progeny tests, parent clones with multiplex PVY resistance and with multiplex PVX resistance have been produced (Mendoza et al., 1996; Solomon-Blackburn, 1998).

Combining host resistances.

There would be great value in obtaining a potato clone with resistance to all the major potato viruses. However, we know of no cultivars with such multiple virus resistance, probably because the cost, time and effort required to achieve this is great and many other features have higher priority. Nevertheless, there are some notable examples of combining different types of host resistance. For example, resistance to PLRV accumulation is not necessarily associated with resistance to PLRV infection, but clones containing both types of resistance have been identified (Solomon-Blackburn & Barker, 1993). These include cv. Pentland Crown and some breeding clones that have greater resistance of both types. These should have great benefit in diminishing PLRV spread. They also have genes for resistance to PVY, and also PVX in some cases. Dziewońska & Waś (1994) have reported that a diploid clone has been selected that has excellent resistance to PLRV infection and accumulation, and in addition has extreme resistance to PVX and high resistance to PVM. Crosses have been made to combine multiplex PVY resistance with both sorts of PLRV resistance (Solomon-Blackburn, 1998). Potato cultivars with ER to PVY, PVA and PVV are known, though this combination is the consequence either of a comprehensive gene or possibly a tight linkage group (Barker, 1997).

Introgression of host resistance genes from wild species

Introgression usually involves hybridization of the donor species and S. tuberosum, followed by repeated backcrossing to Tuberosum and selection for the desired genes. However, not all potato species intercross freely (if at all): species differ in ploidy and endosperm balance number (Bradshaw & Mackay, 1994; Hawkes, 1994; Hermsen, 1994). This has posed problems for the introgression of resistance genes from host species.

One solution has been through bridging crosses and ploidy manipulation using intermediate species. This method has been used to introgress resistance to PLRV, PVY and PVX from the nontuberous S. brevidens into a S. tuberosum background (Johnston & Hanneman, 1982; Ehlenfeldt & Hanneman, 1984), and PLRV resistance from the nontuberous S. etuberosum into a tuber-bearing Solanum gene pool (Hermsen et al., 1981; Chavez et al., 1988a, b). The diploid S. sparsipilum has been used as a link between diploid and tetraploid species for introducing virus resistance (Cockerham, 1970). However, this approach is time-consuming, demanding of resources and limited in genetic efficiency (Hermsen, 1987).

Dihaploids have also been used for introgressing genes for virus resistance from diploid species (Cockerham, 1970). Dihaploids are diploids produced from tetraploids (e.g. S. tuberosum) by pseudogamy. Their use for potato breeding was reviewed by Ross (1986) and Bradshaw & Mackay (1994). Most of the more evolutionarily advanced tuber-bearing wild species are diploid and cross readily with dihaploids of S. tuberosum (Bradshaw & Mackay, 1994). The resulting hybrids are usually long-day adapted and produce tubers. Improved diploid material produced from these can then be hybridized with tetraploid Tuberosum. Because genetics is simpler with diploids than with tetraploids, rapid progress can be made in breeding at the diploid level. Furthermore, the capacity of diploids to produce unreduced gametes means that tetraploid offspring with all the genes of the diploid parent can be produced from 4x × 2x crosses (De Maine, 1982). The use of dihaploids for breeding purposes is limited by the difficulty of producing enough fully fertile true dihaploids of Tuberosum.

An alternative approach is somatic hybridization via protoplast fusion (Barsby et al., 1984). This has been used to transfer resistance to PLRV, PVY and PVX from S. brevidens into S. tuberosum hybrids (Austin et al., 1985; Gibson et al., 1988; Valkonen et al., 1994). This method can produce many hybrids, but their identification and genetic instability have been a problem (Helgeson et al., 1988; Fehér et al., 1992).

Applications of molecular marker technology

Introduction.

Molecular markers linked to resistance genes can be used for marker-assisted selection or for mapping, and thence gene cloning. Prospects for applying these techniques to a tetraploid and heterozygous outbreeding species such as potato have been increased by improvements in marker technology including the use of bulked segregant analysis (BSA) to identify markers (Michelmore et al., 1991), and use of simple sequence repeats (SSRs) together with amplified fragment length polymorphisms (AFLPs) (Milbourne et al., 1997, 1998). Those authors published a map of a large (and expanding) set of SSRs in potato, which should speed up mapping and provide common points of reference between mapping studies in different laboratories.

In the longer term, large-scale sequencing may lead to the development of a complete gene map of potato, provide a short cut to cloning genes, and provide improved markers, single nucleotide polymorphisms (SNPs), which are likely to supersede SSRs (W. De Jong, pers. comm.).

Marker-assisted selection.

Screening by molecular markers (linked to resistance genes) is quick and accurate (Watanabe, 1994). Marker-assisted selection may be useful where phenotypic selection is difficult, or where it is not possible or convenient to use the virus for direct screening. It can also be useful for backcross breeding, for the introgression of resistance genes from wild species whilst selecting against the undesirable characteristics of the wild parent (Young & Tanksley, 1989). Small leaf samples can be used (Deragon & Landry, 1992), so marker-assisted selection could save some years in a potato breeding programme by selecting true seedlings for several unlinked traits at the same time. It may be used on quantitative trait loci (QTLs), with the advantage of selecting pairs of parents with genes at different QTLs for the same trait, provided that genes with large enough effects can be found (Bradshaw et al., 1998) and a large enough population (150–250) is used to map markers (Hackett et al., 1998). The success of marker-assisted selection will depend on close linkage between the resistance genes and (preferably flanking) markers.

Marker-assisted selection for virus resistance in potato has been applied to the gene Ryadg (Hämäläinen et al., 1997) but, so far, no research on its effectiveness in a breeding programme has been reported. The applicability of marker-assisted selection may increase with the identification of SSR markers (for mapping) because they may overcome the problem of population specificity found with AFLPs (Milbourne et al., 1998). The most likely practical use for marker-assisted selection for virus resistance in potato seems to be the introgression of resistance genes from wild species or otherwise poor parents, or possibly for selecting seedlings for several traits simultaneously. In the short term, marker-assisted selection does not seem a cheap or easy option for most practical purposes (except where suitable markers have been identified for other purposes, e.g. mapping), but is likely to become a more viable proposition with further improvements in the technology in the long term.

Molecular mapping and host gene isolation.

To date, eight virus-resistance genes have been mapped in potato (summarized in Solomon-Blackburn & Barker, 2001). Disease resistance gene clusters (covering a diverse range of pathogens) have been found in many plant species, including potato (see Solomon-Blackburn & Barker, 2001). These linkage groups would afford the advantage of efficient selection for several disease resistances at once without screening for them all, whether by marker-assisted or by phenotypic selection. Throughout the history of potato breeding and natural selection for disease resistance, it is probable that some advantage has already accrued from the fact that these resistances are linked.

Following mapping, several virus resistance genes in potato have been or are being cloned (reviewed in Solomon-Blackburn & Barker, 2001). The Rxadg transgene was inserted into the susceptible potato cv. Maris Bard and found to be effective, conferring the same phenotype as in the source cv. Cara, both in protoplasts and whole plants in seven out of eight transgenic lines (Bendahmane et al., 1999). It also conferred ER when inserted into Nicotiana species. The Rysto gene is potentially particularly useful, as this gene (or set of closely linked genes) confers extreme resistance to all known PVY strains including PVYNTN (Barker, 1996; ">Le Romancer & Nedellec, 1997) and also to PVA (Ross, 1961; Cockerham, 1970) and PVV (Barker, 1997). Ryadg would also be a useful transgene as it also confers extreme resistance to isolates from all PVY subgroups, including PVYNTN (Le Romancer & Nedellec, 1997).

The use of transgenes in breeding programmes

One use of transgenes (host- or pathogen-derived) is to insert genes for attributes that are lacking in an existing cultivar. This has already been applied in the U.S.A.: NatureMark’s ‘NewLeaf’ series of cultivars are transformed versions of established cultivars such as Russet Burbank and Shepody. They have transgenic resistance to insects (Colorado beetle) conferred by the Bt gene, and pathogen-derived transgenic resistance to PLRV and PVY (Monsanto, 1997). This resistance to PLRV gives protection against the net necrosis symptoms seen in tubers of the original cv. Russet Burbank when PLRV-infected (Thomas et al., 1995). In the U.K., a change in regulations might be a prerequisite to the use of such cultivars differing in only few genes from existing cultivars: at present, new cultivars must be visibly distinct from existing cultivars (or never grown on the same seed farm).

Transforming clones at an early stage in a breeding programme would be inefficient, as large numbers of clones would have to be transformed and then most of them rejected. Another disadvantage would be that the transgenes might be superseded by better ones before ever reaching cultivar release. Transformation of parent clones would have the same disadvantage, and also only a proportion of the progeny would inherit the transgenes. The most efficient way to use transgenes in a breeding programme might be to insert a series of genes for different disease resistances/attributes in selected advanced clones nearing cultivar selection, allowing time for trialling after transformation. Resistance to PVY (transgenic or natural) may be important for the effectiveness of PTGS-based transgenic resistance to other viruses, as discussed above.

Conclusions

Viruses are relatively easily controlled in temperate potato crops and there is much input into virus control measures in temperate regions, notably in seed certification schemes and application of pesticides to control aphid vectors. However, such measures can have high financial and environmental costs. Other diseases (e.g. late blight and PCN) are more important in temperate regions. For these and other reasons, potato breeding in such areas puts less emphasis on selection of virus resistance than other characteristics. Furthermore, there has been a traditional view that virus resistance is not in the best interest of seed producers, as they would lose trade through the use of ‘home-saved’ seed. However, although virus problems are the main force behind the seed industry, seed producers have also benefited from virus-resistant cultivars by reduced rejection or downgrading of seed crops (Howell, 1977). The situation is different in tropical areas with a higher inoculum potential and where production and use of certified seed is less practical: the need for resistance is greater in these regions. However, there is an increasing need for resistance wherever potatoes are grown as pesticide use comes under close scrutiny and restrictions may be imposed, as has already happened in Denmark (Jørgensen & Secher, 1996). In future, changes in regulations governing pesticide use, or the imposition of an environmental tax on agrochemical use, could quickly change growers’ attitudes to the desirability of virus-resistant cultivars.

Although there is very good host–gene-mediated resistance to several viruses, traditional breeding has produced cultivars with resistance to only two or three viruses, and many cultivars have little or no virus resistance. Furthermore, in many cases the resistance is not fully effective. For example, it may be strain-specific, or confer only partial resistance that breaks down under high inoculum pressure or only reduces the incidence of infection. Partial resistance to PLRV has been enhanced by combining different types of host resistance (Solomon-Blackburn & Barker, 1993). Combining resistances to several viruses and other diseases with desirable agronomic features in a single cultivar is difficult and time-consuming by conventional means. However, progress has been made, for example by the production of multiplex resistant parents (in UK and Peru) and by combining resistance to different viruses in some parent clones (Ross, 1978; Świeżyński, 1983, 1994; Dziewońska, 1986; Solomon-Blackburn, 1998).

New molecular technologies seem likely to benefit resistance breeding in future, as they promise to shorten the breeding process through the more rapid and cost-effective introgression of desirable characteristics, including virus resistance. As discussed above, marker-assisted selection may, in future, prove to be particularly useful for the introgression of resistance genes from wild species or parents with poor agronomic characteristics. The ability to select seedlings for several traits simultaneously will also be valuable, as more and better markers are developed. Marker-assisted selection may also prove useful for selecting ER genes in populations containing genes for HR to the same virus. However, even with the advent of marker-assisted selection, it is unrealistic to expect that conventional breeding could succeed in producing potato cultivars with resistance to the 10 most common viruses, as well as all the other attributes required in a cultivar. This is where transgenic resistance offers great potential.

As research on mapping progresses, more will become known about the locations and relationships of host genes, and more will probably be identified which might be exploited. The cloning of host genes is at present slow and expensive, but likely to become easier as knowledge expands. In future, use of the sequences of cloned host genes to make transgenes, possibly with improved characteristics, for reinsertion into breeding clones is an attractive possibility. However, transgenes produced in this manner seem likely to raise concerns in the minds of the public in some of the same ways that transgenes using pathogen-derived sequences have done, and this could inhibit their exploitation.

A frequently occurring problem with pathogen-derived resistance is that many of the transgenic lines in a transformation do not confer adequate resistance. Screening lines for resistance is expensive and time-consuming, so measures to increase the success rate by deliberate targeting of PTGS through transformation with sense and antisense sequences may be a useful approach. Cases where transgenic resistance has been developed to viruses for which there are no host resistance genes, or where they are only partially effective or improvable by combination (e.g. for PMTV and PLRV, respectively) (Barker et al., 1994b, 1998a) are very promising, as are the two cases of transgenic PSTVd resistance. If such forms of resistance are stable and effective in the field, and if there are no side effects, they should have great potential.

Future trends in virus resistance breeding are very difficult to predict, and will increasingly depend on politics, the consumer and the manner of science funding. Considerable scope remains for conventional breeding and the use of host genes, but new opportunities are emerging with molecular techniques for gene identification and selection. Thus far, the main interest in pathogen-derived resistance at a commercial level has been for the transformation of special-purpose cultivars that lack resistance to a particular virus (already taken up in the U.S.A.). In future, however, pathogen-derived resistance is likely to be much more useful for constructing transgenes that confer resistance to a range of viruses (for example, the 6–10 most widespread viruses in potato). Such resistance could be developed quite soon and could provide a practical means of obtaining multiple virus resistance much faster and more economically than by any other route, including use of cloned host gene sequences in transgenes and marker assisted selection in conventional breeding programmes. Several broad-spectrum (nonspecific) approaches to obtaining multiple virus resistance using nonvirus sequences in transgenes have been tested. However, questions remain over the validity of these approaches in terms of effectiveness and their public acceptability.

Outstanding problems of how multiple virus resistance transgenes based on virus sequences should be made, and the determination of their stability and durability in a range of field conditions, need to be examined. Nevertheless, for the short-term goal of producing broad-spectrum virus resistance, effort in this area is likely to be more fruitful than the pursuit of alternative approaches. However, uptake of many of these possibilities will depend upon the willingness of consumers to eat GM potatoes. They will need to be convinced that it is not dangerous but instead beneficial to both consumer and environment.