Selection for reproductive ability in Globodera pallida populations in relation to quantitative resistance from Solanum vernei and S. tuberosum ssp. andigena CPC2802

Authors


*E-mail: mphill@scri.ac.uk

Abstract

Four field populations of the nematode Globodera pallida were subjected to selection pressure for increased reproductive ability by rearing sub-populations continuously on four partially resistant potato genotypes for 12 generations. The resistance was derived from either Solanum vernei or from S. tuberosum spp. andigena CPC2802. After the 12th generation the original and sub-populations of nematodes were assessed for their reproductive ability on a susceptible genotype and on each of the partially resistant genotypes. Selection pressure was shown to have increased reproductive ability but the increases were specific to the source of resistance used. The average increase on the ex S. vernei clones was from 11% reproduction by the unselected populations to 35·5% reproduction after selection. On the clones derived from CPC2802, which had higher levels of resistance, the increases were larger with an average of 6·6% reproduction for the unselected but 47·4% reproduction after selection. The response to selection differed amongst the initial field populations with some rates of reproduction increased to as much as 79%. A RAPD based analysis of the original and sub-populations after selection indicated small but consistent changes in the genetic structure, which could have been a result of the selection pressure per se and/or the bottlenecks that the populations had gone though.

Introduction

The potato cyst nematodes (PCN: Globodera rostochiensis and G. pallida are damaging pests of potatoes. In the UK the widespread use of cultivars with the H1 gene for resistance to G. rostochiensis has led to the predominance of G. pallida (Minnis et al., 2002), which is now found in 91% of infested ware potato-growing land in England and Wales. No major gene resistance to G. pallida has been identified but there is a limited number of cultivars with partial resistance to this species. These cultivars reduce the rate at which G. pallida can multiply to below that observed on susceptible cultivars, but seldom reduce population levels. Consequently these cultivars have a role to play in control strategies, but in order to gain sustainable control of the nematode populations, have to be used together with other control strategies such as nematicides or long rotations (Trudgill et al., 2003). The use of cultivars partially resistant to G. pallida is limited but their use could provide selection pressure to increase the reproductive rate in G. pallida populations (Turner, 1990).

This study examined the effect of the continuous rearing of populations of G. pallida on partially resistant host genotypes by measuring the populations’ ability to multiply. An earlier detailed study (Turner, 1990) concentrated on resistance derived from Solanum vernei and used the differential host genotypes used in the International Pathotype Scheme (Kort et al., 1977) for potato cyst nematodes. In this study, two sources of resistance were used and the host genotypes were chosen to represent those that have potential to be developed as cultivars. In addition to assessments of reproductive rate this current study included a molecular characterization of the initial and resultant sub-populations of PCN produced, to examine if any biological differences were reflected in general genetic characteristics following selection.

As well as establishing potential biological effects it was of interest to see if there were any molecular markers that might be associated with increased reproduction. Pastrik et al. (1995) reported a 0·43 kb DNA fragment that hybridized with a German population (Kalle) selected for increased reproductive rate on an ex S. vernei cultivar Darwina, but not to the original unselected Kalle population. This fragment was investigated further to establish its sequence and utility as a marker.

Materials and methods

Production of populations

Four field populations (Bedale, Farcet, Halton, Newton) of G. pallida were reared for up to 12 generations on susceptible cultivars and on each of four potato clones, two of which were derived from Solanum vernei and two from S. tuberosum ssp. andigena CPC2802 (Table 1). The PCN populations were chosen to represent a range of reproductive abilities using data from Phillips & Trudgill (1983) (Table 2).

Table 1.  Potato genotypes and sources of potato cyst nematode (Globodera pallida) resistance used for selection
Potato genotypeSource of resistanceNumber of back crosses to the original source
Guardian and MoragS. vernei4
MoragS. vernei5
11415S. tuberosum spp. andigena CPC28022
12674S. tuberosum spp. andigena CPC28023
Table 2.  Populations of the potato cyst nematode (Globodera pallida) used in this study
Field populationsMultiplication rates on differential clone 62-33-3a
  1. aPhillips & Trudgill (1983).

Halton 2·1
Farcet 3·3
Newton 5·4
Bedale11·2

The potato genotypes used were selected on three criteria. First, they represent the two main sources of resistance used in breeding programmes in the UK; secondly, they have different levels of quantitative resistance; and thirdly, they were agronomically acceptable potential future cultivars at the time of initiation of these experiments.

The combination of four populations each separately reared on four partially resistant host genotypes together with the four unselected populations resulted in 20 sub-populations (Fig. 1). For each generation of any one sub-population, 10 cysts in a ‘Terylene’ net bag were placed in each of four 15 cm square pots filled with sand, together with a sprouted tuber of the appropriate potato genotype. Four pots for every population/genotype combination were used. Plants were left until they became senescent, when the soil was allowed to dry before the cysts were extracted by flotation. The new cysts from the four replicate pots were pooled and four samples of 10 cysts required for the next generation were selected at random. The populations were passed though one generation per year.

Figure 1.

Schematic diagram showing the method of producing lines of Globodera pallida selected for reproductive ability and subsequent testing on host genotypes with different sources of resistance.

At the same time, the ‘original’ field populations were maintained by rearing them each year on a mixture of susceptible cultivars (Désirée, Maris Piper, Pentland Crown and Pentland Dell).

Nematode multiplication tests

The reproductive ability of all populations and sub-populations on all the potato genotypes was assessed at the end of the 12th generation (Fig. 1). The S. vernei-derived differential clone 62·33·3 (Kort et al., 1977) was also included in the test and the cv. Désirée was used as a susceptible control.

The experiment was set up in a glasshouse in pots (10 cm diameter) half-filled with a loam/sand (3:1) mix. Using a calibrated spoon each pot was inoculated with 10 ± 1 cysts, completely filled with more loam/sand mix (total dry weight 450 g) and planted with a tuber piece having a single sprout. There were four replicates for every population × host genotype combination and the experiment was laid out in a completely randomized block design. When all the plants had died, after 15 weeks, the soil was allowed to dry and then cysts were extracted using flotation and counted.

Molecular analysis

DNA extraction

DNA was extracted from cysts that had been stored at 4°C using the procedure described by Pastrik et al. (1995) but with the addition of a phenol: chloroform, chloroform extraction step following the isopropanol precipitation stage. Dilutions of DNA were stored at 4°C and the stock solutions at –20°C.

RAPD PCR reactions

Amplification reactions included 10 ng of DNA in 50 µL of 10 mm Tris-HCl (pH 8·3), 1·5 mm MgCl2, 50 mm KCl, 200 µm dATP, dCTP, dGTP and dTTP, 200 µm primer and 1 unit of Taq polymerase (Roche Applied Science). An initial denaturation was carried out at 94°C for 4 min followed by amplification for 45 cycles of 1 min at 94°C with an annealing temperature of 38°C for 2 min and extension at 72°C for 3 min with a ramp time between the 38–72°C phase of 1°C per 5 sec. There was a final extension at 72°C for 5 min. Each reaction was repeated on at least two separate occasions using a Perkin Elmer 2400 thermocycler (Applied Biosystems). Two sets of products from any one primer were separated by electrophoresis in TBE buffered 1·5% agarose gels (Sambrook et al., 1989). The products were visualized by UV illumination following staining with ethidium bromide, and photographed.

Twelve primers from the Operon kit G (2, 3, 4, 5, 8, 10, 13, 15, 16, 17, 19) (Operon Biotechnologies Inc.) were used, which have been used in previous studies on PCN, together with one primer (E-06) from kit E (Folkertsma et al., 1994; Pastrik et al., 1995; Blok et al., 1997). Amplification products were scored by visual assessment of the duplicate reactions. Only products that were present in both PCR reactions were scored as present. The presence or absence of amplification products (bands) was recorded as 1 or 0, respectively.

Statistical analysis

Reproduction data

The data for cyst numbers per pot were analysed using standard analysis of variance after square root transformation to normalize the data. The analyses were either of the total data set or of subsets containing each of the four populations and associated sub-populations. Of particular interest were the population × genotype interactions and these are displayed as the interaction effects. Statistical analysis was performed using the GENSTAT software package (NAG).

Percentage susceptibility was calculated from the proportion of cysts produced on the resistant genotypes relative to those produced on the susceptible control cv. Désirée.

Molecular data

The RAPD data were used to derive a similarity matrix based on an estimate of the proportion of shared amplification products (Nei & Li, 1979). The relationships between the sub-populations were examined using neighbour-joining. A bootstrap analysis was also performed. The statistical package GENSTAT and programs from PHYLIP 3·5 provided by J. Felsenstein (Department of Genetics, University of Washington, Seattle, WA, USA) were used for these analyses.

Results

Nematode multiplication test

The results of the test showed that continuous rearing of the sub-populations on partially resistant hosts caused significant changes in reproductive ability in many of the populations. Table 3 shows the proportional cyst production, in relation to cyst production on the susceptible control cv. Désirée, following angular transformation of the data. In all but one instance any sub-population retested on the selecting clone showed a significant increase in proportional reproduction relative to the unselected population. The exception was the Bedale sub-population selected on clone 11415 which did not show significantly higher relative reproduction. However, this sub-population did have a significantly higher proportional reproduction on clone 12674, which is derived from the same source of resistance.

Table 3.  Mean percentage (angular transformation) cyst production of five subpopulations each from four different field populations of potato cyst nematode (Globodera pallida) on five different host genotypes. The percentage cyst production is relative to reproduction on the susceptible control cv. Désirée after 12 generations. Untransformed means in parenthesis
PopulationCloneUnselectedEx MoragEx GuardianEx 11415Ex 12674
  1. Least significant difference (P = 0·05) = 8·49.

  2. a Nematode population selected on host genotypes where the source of resistance was from S. vernei.

  3. b Nematode population selected on host genotypes where the source of resistance was from CPC2802.

  4. c Selected population on selecting clone.

  5. d Significantly greater reproduction than observed on the unselected population.

FarcetMorag16·0 (7·7)33·1 (30·0)acd34·6 (32·8)ad24·0 (17·5)17·2 (8·9)
62·33·313·5 (5·5)22·4 (15·1)ad19·4 (12·1)a30·0 (25·9)d16·7 (8·8)
Guardian19·0 (11)36·8 (36·0)ad37·3 (36·9)acd35·7 (34·0)d22·9 (15·3)d
1141518·6 (10·8)24·2 (17·1)17·6 (9·2)52·4 (62·6)bc61·4 (75·0)bcd
1267411·3 (4·0)12·0 (4·4)9·7 (2·9)34 (31·8)bcd48·7 (68·9)bcd
HaltonMorag15·2 (7·2)29·6 (26·4)acd41·1 (43·3)ad17·6 (9·3)16·8 (8·4)
62·33·310·4 (3·4)17·4 (9·1)a27·4 (21·7)ad13·5 (5·6)15·5 (7·5)
Guardian17·0 (8·7)24·4 (18·8)a40·6 (42·3)acd18·5 (10·2)18·0 (9·6)
1141513·2 (5·5)15·1 (6·9)15·4 (7·5)34·1 (32·1)bcd22·7 (14·9)bd
126748·5 (2·2)12·5 (4·8)9·4 (2·7)15·0 (7·4)bd30·8 (27·7)bcd
BedaleMorag18·2 (9·9)37·8 (38·1)acd32·5 (29·2)ad16·9 (8·6)24·9 (18·2)
62·33·318·7 (10·4)27·3 (21·3)ad32·4 (29·2)ad20·1 (12·2)27·5 (21·5)
Guardian17·5 (9·5)42·7 (46·5)ad32·2 (28·8)acd19·8 (11·6)26·3 (19·7)d
1141519 (10·7)19·2 (10·8)22·5 (14·9)18·1 (9·9)bc45·1 (50·2)b
1267414·4 (6·2)20·2 (12·6)8·8 (2·4)15·7 (7·4)b42·7 (46·0)bcd
NewtonMorag26 (20·4)49·9 (58·3)acd32·5 (29·2)a19·4 (11·0)11·9 (4·3)
62·33·312·9 (5·2)30·3 (25·7)ad19·9 (11·8)a19·1 (11·5)11·9 (4·3)
Guardian19·7 (11·5)37·0 (36·6)ad32·1 (28·2)acd19·5 (11·7)10·3 (3·2)
1141518·6 (10·3)17·0 (9·3)17·2 (8·9)50·1 (79·1)bcd51·3 (60·2)bd
1267410·1 (3·1)14·7 (8·9)11·6 (4·2)26·9 (20·6)bd24·3 (18·7)bcd

Generally, sub-populations that had been selected on each source of resistance showed increased cyst production on the test clones with resistance from the same source. No sub-population selected on resistance derived from S. vernei showed significantly increased reproduction on clones derived from CPC2802. The Farcet population selected on clone 11415 and the Bedale population selected on 12674 showed some increased reproduction on some of the clones with resistance from S. vernei.

Analysis of the cyst data indicated high levels of significance between the clones, populations and selection treatment and all interactions (Table 4). Of the main effects, most of the variation was accounted for by host genotypes. Differences between the populations, although significant, were relatively small. The largest interaction was between clones and the selection treatment. Partitioning the selection treatment into a contrast between those sub-populations reared on ex S. vernei clones and the ex CPC2802 clones and a contrast between the clones derived from the different sources indicated that this interaction was largely accounted for by those populations selected on a particular source of resistance reproducing better on clones of that source of resistance.

Table 4.  The mean cyst production (square root transformation) of five subpopulations from each of four different field populations of potato cyst nematode (Globodera pallida) reared on six different host genotypes. Untransformed means in parenthesis
PopulationCloneUnselectedEx MoragEx GuardianEx 11415Ex 12674Mean
  • L.S.D. (P = 0·5) 3·876.

  • a

    Selected population on selecting clone.

FarcetDésirée24·6 (608·9)25·3 (658·2)30·6 (941·8)24·0 (585·5)20·7 (444·0)25·1 (652·0)
Morag6·8 (4·8)14·0 (197·8)a17·3 (309·0)9·8 (102·2)6·2 (39·5)10·5 (130·1)
62·33·35·8 (33·5)9·7 (99·2)10·1 (114·2)12·0 (151·5)6·1 (39·2)8·7 (87·5)
Guardian8·0 (67·0)15·4 (237·0)18·6 (347·0)a14·1 (199·2)8·2 (67·7)12·8 (183·6)
114157·9 (65·5)10·5 (112·2)9·3 (86·5)19·1 (366·2)a18·0 (332·7)13·0 (192·7)
126744·8 (24·2)5·4 (29·0)5·2 (27·2)13·4 (186·3)17·3 (305·7)a9·0 (110·7)
HaltonDésirée30·8 (959·2)21·3 (458·7)26·9 (731·9)25·1 (667·8)21·1 (467·4)25·6 (685·5)
Morag8·1 (69·0)10·3 (121·0)a17·7 (317·0)7·8 (62·3)6·2 (39·0)9·6 (110·1)
62·33·35·6 (32·5)6·4 (41·7)12·4 (158·5)6·1 (37·2)5·8 (35·0)6·8 (52·1)
Guardian9·1 (83·0)8·7 (86·0)17·6 (309·5)a8·2 (67·8)6·7 (45·0)9·5 (103·5)
114157·1 (52·7)5·6 (31·5)7·2 (55·0)14·4 (214·0)a8·3 (69·5)8·7 (90·1)
126744·6 (21·2)4·7 (22·0)4·4 (19·5)6·7 (49·3)10·9 (129·5)a5·8 (41·3)
BedaleDésirée25·2 (640·8)14·6 (222·3)34·1 (1187·8)30·9 (958·3)20·3 (419·5)25·6 (715·8)
Morag7·9 (63·2)9·1 (84·7)a18·4 (346·2)9·0 (82·0)8·6 (76·5)11·2 (147·3
62·33·38·1 (66·7)6·8 (47·3)18·3 (346·7)10·6 (115·3)9·5 (90·0)11·1 (149·1)
Guardian7·6 (60·7)9·9 (103·3)18·2 (342·0)a10·5 (111·5)9·1 (82·5)11·5 (153·9)
114158·2 (68·7)4·9 (24·0)13·2 (177·2)9·6 (95·0)a13·7 (210·5)10·1 (117·4)
126746·3 (40·0)5·1 (28·0)5·3 (28·5)8·4 (70·5)13·8 (193·0)a7·1 (59·0)
NewtonDésirée25·1 (696·0)23·4 (563·5)34·9 (1220·0)19·7 (396·3)23·4 (547·9)25·3 (686·4)
Morag11·4 (142·2)18·1 (328·3)a18·7 (356·0)6·6 (43·8)4·8 (23·5)12·7 (196·0)
62·33·35·9 (36·0)11·9 (145·0)11·9 (144·2)6·5 (45·5)4·8 (23·5)8·6 (85·0)
Guardian8·9 (80·0)14·2 (206·2)18·5 (344·5)a6·6 (46·2)4·2 (17·5)11·2 (152·4)
114158·4 (71·5)6·9 (52·5)10·4 (108·3)17·4 (313·3)a18·0 (330·0)11·2 (148·8)
126744·6 (21·7)5·9 (50·2)7·0 (50·8)9·0 (81·7)9·5 (102·5)a6·9 (56·8)

This specificity of response was found with all the populations and is illustrated in Fig. 2(a–d), which shows the interaction effects for each population. Bars with positive values indicate that the relative reproduction of sub-population × clone interaction was greater than would be expected from the grand means and bars with negative values indicate lower than average reproduction. The relative cyst production on clones derived from S. vernei was highest with populations selected on ex S. vernei resistant genotypes and relatively lower with populations selected on ex CPC2802 resistance. Conversely, the latter populations showed relatively increased reproduction on the ex CPC2802 derived material.

Figure 2.

Estimated interaction effects of clone × PCN (Globodera pallida) sub-population interaction effects for Halton, Farcet, Newton and Bedale populations and selected subpopulations. The Y axis indicates the deviation of the observed mean from that which would be predicted from the average clone and population effects. ⋆ Indicates selected population on selecting source of resistance.

Overall the levels of increased reproduction were highest in the sub-populations reared on the ex CPC2802 resistant clones (Table 5). However, this was largely due to this occurring most markedly in the Farcet population. Within the other three original populations the levels of increased reproduction were similar on both sources of resistance.

Table 5.  Summary of the mean percentage cyst production (angular transformation) by Globodera pallida classified by the original population, the mean of clones with a common source of resistance and the mean of the selection treatment. The percentage cyst production is relative to reproduction on the susceptible control cv. Désirée after 12 generations
Test clonesPopulationSelection treatment
UnselectedEx S. verneiEx CPC2802
  1. S.E.D. (P = 0·05) = 6·73.

  2. a Selected population on selecting source of resistance.

Ex S. verneiFarcet16·1630·44a24·42
Ex CPC2802Farcet14·9415·8950·20a
Ex S. verneiHalton14·1227·97a16·60
Ex CPC2802Halton10·8213·3326·23a
Ex S. verneiBedale18·1333·87a22·81
Ex CPC2802Bedale16·6917·2430·36a
Ex S. verneiNewton19·5133·61a16·66
Ex CPC2802Newton14·3215·1335·87a

Molecular analysis

The RAPD study resulted in 181 scorable bands of which 111 were polymorphic. An unrooted tree (Fig. 3) derived from the RAPD data showed that the sub-populations tended to cluster depending on their original source, with the exception of the Farcet population reared on 12674 which clustered, with weak support from the bootstrap analysis, with the Halton derived sub-populations. There was relatively strong support for the clustering of the Farcet derived sub-populations (81%) and less for the Bedale (60%) and Newton (72%) sub-populations. Within each population group, those that had been selected on CPC2802 resistance were generally the most distant from the original unselected population.

Figure 3.

Unrooted dendrogram of the genetic similarity (%) of four populations and selected sub-populations of Globodera pallida based on molecular characterization analysis with RAPD markers. The dendrogram was based on a similarity matrix derived from an estimate of the proportion of shared amplification products (Nei & Li, 1979) and the neighbour-joining clustering method. Numbers at the nodes represent bootstrap values supporting the branches, generated from 1000 random permutations. Population codes: F = Farcet, H = Halton, N = Newton, B = Bedale. Selection codes: M = Morag, G = Guardian, 1 = 11415, 2 = 12674, U = unselected.

No clear associations were observed between the presence or absence of specific RAPD fragments and reproductive ability. Within each of the four original populations some markers appeared or disappeared in the selected lines but there was no consistency across all populations to suggest that these markers were associated with reproductive rate.

Discussion

The genetics of the two sources of resistance used in this study have often been described as polygenic (Dale & Phillips, 1982). However, recent studies mapping quantitative trait loci (QTLs) have shown that with S. vernei derived resistance there are at least two, but more likely three, QTLs involved on linkage groups V and IX (Rouppe van der Voort et al., 2000; Bryan et al., 2002). With resistance from CPC2802, two QTLs have been identified on linkage group IV and XI (Bradshaw et al., 1998; Bryan et al., 2004), suggesting that the latter is under different genetic control.

Selection on the pathotype differentials (Kort et al., 1977) derived from S. vernei was first reported by Turner et al. (1983). The authors reported that the rate of reproduction relative to that on a susceptible host increased from 10 up to 88% on clone 62·33·3 and up to 100% on clone 65·346/19 after eight generations. In the study reported here, no year-to-year assessment was made, but after 12 generations the maximum increase in reproductive rate was to 58% (Newton). The S. vernei clones used in this study were potential cultivars and were derived after several more rounds of backcrossing to susceptible S. tuberosum ssp. tuberosum parents than the differential clones. Consequently it is possible that there are relatively fewer genes of small effect from S. vernei in these genotypes than in the differential clones and that this resulted in a lower selection pressure.

Beniers et al. (1995) demonstrated increased reproductive ability in G. pallida populations after 8 years of continuous cropping in field experiments of cv. Darwina (derived from S. vernei). At two sites, reproductive rate rose from an initial 10 and 13% to 32 and 35%, respectively. This is comparable to the results shown here where the average increase in reproduction due to selection on cv. Morag was from 11 to 38% (a 3·3-fold increase) and on cv. Guardian from 10 to 34%; a maximum increase of 58% was obtained for the Newton population selected on Morag. Schouten & Beniers (1997) reported that increase in reproductive ability was due to a greater proportion of individual nematodes becoming female rather than an increase in fecundity.

In contrast to S. vernei, the effects of resistance derived from CPC2802 has received less attention; in this study it showed a more marked increase in reproductive ability from 9·3 to 45·9% (a 4·9-fold increase) on 11415 and from 4% to 40% on 12674, with a maximum of 79% for the Newton population selected on 12674.

Anderson (1986) examined the effect of growing two clones with resistance from CPC2802 in field trials over 3 years and found no evidence of increased multiplication. Whitehead (1991) used two other breeding clones derived from CPC2802 to examine five PCN populations in microplots over four generations. The initial population densities were relatively high (40–100 eggs per g soil) and make the interpretation difficult as estimates of nematode reproduction will have been affected by damage to the roots. There was little evidence for selection on the basis of the relative susceptibility of the resistant clones, although pot tests indicated some increase in three of the five populations examined. Other studies have suggested that more than four generations are required to detect increases and that rates of selection will vary depending on the frequency of genes for pathogenicity in the different populations (Schouten & Beniers, 1997). Turner & Fleming (2002) included a host genotype with resistance from CPC2802 in their study of different sources of resistance. They examined one population of G. pallida and observed no selection on CPC2802 for the first five generations, after which reproductive ability increased.

Attempts to use biochemical or molecular markers to examine genetic changes or to find markers related to increased reproductive ability have been previously reported (Turner, 1990; Pastrik et al., 1995). The latter used isozymes and observed differences between unselected and selected lines and between the numbers of bands present in unselected and selected lines. There was no consistency, however, in whether there were more or less bands associated with increased reproduction. In this study using RAPDs a similar situation was observed in that there were no constant patterns with regard to the numbers of polymorphic bands found in each sub-population; however, the differences were small, in the range of 122 to 134 bands per population. All the populations selected from Newton and Bedale showed fewer bands than the unselected, whilst with Halton population the number of polymorphic bands in the selected lines was usually higher than the unselected populations. In the absence of any concordance between RAPD markers and reproductive rate it is most probable that these effects are a result of founder effects or the bottlenecks experienced by these populations in the early stages of selection.

Pastrik et al. (1995) identified two RAPD bands that differentiated an unselected and selected line from a G. pallida population in Germany. Selection had been on the S. vernei derived clone Darwina. They isolated one of these amplification products and found that it hybridized to RAPD amplification products from medium or highly reproductive field populations but not to those with low multiplication rates. The populations in the current study were examined to see if similar hybridization patterns were observed. There was hybridization of the 0·43 kb DNA fragment reported by Pastrik et al. (1995) to RAPD amplification products from the unselected and selected lines but with no consistent pattern to differentiate unselected or selected populations (data not shown), suggesting that the probe is not very tightly linked to reproductive ability and does not have universal applicability. The sequence of the 0·43 kb fragment (GenBank Acc. No. AM292864), while showing no homology to a coding region, appears to be the 3′UTR that is contiguous with ubiquitin sequences found in expressed sequence tags of both PCN species (e.g. Acc. Nos. CV579085 and BM355124). Among eukaryotes, ubiquitins form a large highly conserved gene family and play an important role in regulating the cell cycle, DNA repair, embryogenesis, the regulation of transcription and apoptosis.

The main conclusion that can be drawn from this study is that selection for increased reproductive ability occurs with both sources of resistance but that the selection is specific to the source of resistance used. The degree of selection found is not as high as has been reported by other workers and this may in part be explained by the use of more advanced breeding lines further removed from the original wild species from which the resistance was derived. More marked selection was found when the populations were selected on material derived from CPC2802. There could be at least two explanations for this. First, that this resistance is genetically different, being explained by two QTLs and lacking resistance to G. rostochiensis, whereas S. vernei resistance may be explained by three QTLs on different linkage groups and which also confers resistance to G. rostochiensis. Secondly, the resistance of the two exCPC2802 clones used was slightly greater than the ex S. vernei clones and, therefore, would have exerted a greater selection pressure.

In common with other studies it would seem sensible in developing control strategies to deploy, if possible, cultivars with different sources of resistance rather than use any one repeatedly.

Acknowledgements

The authors acknowledge the technical skills of Anne Holt, the statistical advice of James McNicol (Biomathematics and Statistics Scotland) and funding by the Scottish Office Agriculture, Environment and Fisheries Department and EU Project No. AIR3 CT 92·0062. We also thank Wolfgang Burgermeister for supplying the E-06 Probe.

Ancillary