• Open Access

Large subclonal variation in Phytophthora infestans populations associated with Ecuadorian potato landraces

Authors

  • R. A. Delgado,

    1. Instituto Nacional Autónomo de Investigaciones Agropecuárias (INIAP), Quito, Ecuador
    2. Wageningen UR Plant Breeding, Wageningen University and Research Center, Wageningen, the Netherlands
    Search for more papers by this author
  • A. R. Monteros-Altamirano,

    1. Instituto Nacional Autónomo de Investigaciones Agropecuárias (INIAP), Quito, Ecuador
    2. Wageningen UR Plant Breeding, Wageningen University and Research Center, Wageningen, the Netherlands
    Search for more papers by this author
  • Y. Li,

    1. Plant Research International, Biointeractions and Plant Health, Wageningen, the Netherlands
    Search for more papers by this author
  • R. G. F. Visser,

    1. Wageningen UR Plant Breeding, Wageningen University and Research Center, Wageningen, the Netherlands
    Search for more papers by this author
  • T. A. J. van der Lee,

    1. Plant Research International, Biointeractions and Plant Health, Wageningen, the Netherlands
    Search for more papers by this author
  • B. Vosman

    Corresponding author
    1. Wageningen UR Plant Breeding, Wageningen University and Research Center, Wageningen, the Netherlands
    • Instituto Nacional Autónomo de Investigaciones Agropecuárias (INIAP), Quito, Ecuador
    Search for more papers by this author

E-mail: ben.vosman@wur.nl

Abstract

The population of Phytophthora infestans on potato landraces in three provinces (Carchi, Chimborazo and Loja) of Ecuador was analysed. All isolates (= 66) were of the A1 mating type. Simple sequence repeats (SSR) were used to assess the genetic diversity of the isolates. The P. infestans isolates from the potato landraces grouped in a single clade together with reference isolates belonging to the clonal lineage EC-1. In the 66 SSR profiles obtained, 31 multilocus genotypes were identified. The 66 isolates constituted 49 different races according to the Solanum demissum differential set ( R1 to R11). The P. infestans population was complex and virulent on 4 to 11 R genes. Analysis showed that the subclonal variation in the Ecuadorian EC-1 clone is increasing over time and is much larger than clonal variation in lineages in the Netherlands and Nicaragua, suggesting high mutation rates and little or no selection in Ecuador.

Introduction

The oomycete Phytophthora infestans is the causal agent of late blight and attacks a range of species in the Solanaceae family. It is the most important disease on potato in Ecuador and worldwide (Oyarzún et al., 2002). Diversity within and among P. infestans populations has been studied using phenotypic and genotypic markers. Among the phenotypic markers used are mating type and virulence. The mating type is based on the fact that P. infestans is heterothallic, so two types, A1 and A2, are necessary for sexual reproduction (Forbes, 1997). For virulence studies, a set of potato genotypes with different R genes is used (Black et al., 1953; Malcolmson & Black, 1966). This evaluation is based on the reaction of P. infestans isolates with Solanum plants carrying specific resistance genes, and grouping those with the same virulence phenotype as races. Among the genotypic markers used are allozymes, restriction fragment length polymorphisms (RFLPs), mitochondrial haplotypes, amplified fragment polymorphisms (AFLPs) and simple sequence repeats (SSRs or microsatellites) (Fry et al., 2009). Studies on Phytophthora associated with Solanaceae in Ecuador showed the presence of four clonal lineages based on the RFLP fingerprinting pattern (Forbes et al., 1997; Adler et al., 2004). Two clonal lineages, EC-1 and US-1, both of the A1 mating type, have been identified already in the previous century (Forbes et al., 1997). The EC-1 lineage is predominant on potatoes and the US-1 on tomatoes (Forbes et al., 1997; Oyarzún et al., 1998). More recently, two additional lineages were found in Ecuador, EC-2 and EC-3, which cause late blight on non-tuber bearing Solanum species such as S. betaceaum, S. quitoense, S. hispidum and S. muricatum (Ordoñez et al., 2000; Adler et al., 2004). Oliva et al. (2010) formally described EC-2 and EC-3 as a distinct species, Phytophthora andina. However, the species status of P. andina is questioned by others (Cárdenas et al., 2012); it may actually be a hybrid between P. infestans and an as yet unknown lineage. Support for this also comes from a study by Blair et al. (2012).

Studies on the race structure of P. infestans populations isolated from cultivated potatoes in Ecuador showed an increase in the complexity from a predominance of avirulent isolates (INIAP, 1974, 1975, 1976) to complex races, which was suggested to have resulted from the replacement of the US-1 lineage on potatoes by the EC-1 lineage (Forbes et al., 1997). Additionally, EC-1 was found to be more aggressive on potato than isolates of the US-1 clonal lineage (Oyarzún et al., 1998). Forbes et al. (1997) found 24 races infecting potatoes in Ecuador, 14 in Carchi, 14 in Chimborazo and eight in Loja. More recently, in commercial varieties and selected clones from INIAP's national potato breeding programme, 27, 17 and 37 races were found in Carchi, Cotopaxi and Pichincha provinces, respectively (Tello, 2008). In Peru and Colombia three clonal lineages were identified that infected potato landraces (Garry et al., 2005; Vargas et al., 2009) and, as in Ecuador, the dominant lineage was EC-1. Some studies on P. infestans populations indicated a low genetic diversity, represented by a single dominant clonal lineage but with high virulence diversity (Forbes et al., 1997; Guo et al., 2009). In regions where sexual reproduction occurs, such as the Nordic countries, populations had a high genetic diversity (Brurberg et al., 2011).

Despite all previous studies on Phytophtora spp., there are no specific studies on populations of P. infestans associated with Ecuadorian potato landraces and it is unknown if there are any other lineages or species present on them.

There are more than 400 potato landraces in Ecuador (Cuesta et al., 2005). These landraces are the result of selection and conservation carried out by small-scale farmers in the highlands (Cuesta et al., 2005). In contrast to conventional potato cultivation, these potatoes are cultivated on small acreages, with low input of pesticides and often several landraces grown together (Cuesta et al., 2005; Monteros & Reinoso, 2010). The landraces belong to three Solanum species: the diploid S. phureja, the triploid S. chaucha and the tetraploid S. andigena (Monteros-Altamirano, 2011). They constitute a potential source of genetic variation for breeding purposes, in characters such as quality, earliness and resistance to biotic and abiotic stresses. Evaluations showed that most landraces in Ecuador were highly susceptible to late blight, but some landraces with field resistance to late blight were also identified (Cañizares & Forbes, 1995; Revelo et al., 1997; Monteros-Altamirano, 2011).

The aims of this study were to: (i) characterize P. infestans populations associated with Ecuadorian potato landraces in three areas of Ecuador, (ii) compare these to populations previously reported on commercial potatoes, and (iii) assess the impact of landraces on the P. infestans population. The genotypic diversity found in the EC-1 lineage present in Ecuador was also compared to the variation found elsewhere.

Materials and methods

Isolate collection

Phytophthora infestans isolates were collected from potato landraces present in three regions of Ecuador: the provinces of Carchi in the north, Chimborazo in the centre and Loja in the south (Monteros et al., 2008). For each potato landrace present on a farm, five to 10 leaves with a single lesion were sampled. Leaves were kept at 4°C until isolation (Forbes et al., 1997). The name and ploidy level of the landrace sampled, farm owner, location, GPS coordinates and altitude were recorded (Table S1). The pathogen was isolated either from infected leaves or from small pieces of necrotic leaves, which were placed between potato slices of the susceptible cv. Superchola and incubated in a humid chamber at 16°C with a 12 h photoperiod. Once mycelium was visible on leaves or slices, it was transferred to Rye B Agar plates (Caten & Jinks, 1968) with antibiotics (Oyarzún et al., 1998). Purified isolates of P. infestans were maintained on Rye A medium (Caten & Jinks, 1968).

Isolate characterization

The mating type was determined for each isolate by pairing it with known A1 (EC3090 or EC3690) and A2 (EC3260) isolates (provided by the International Potato Center (CIP), Lima, Peru) on 10% clarified V8 agar (Forbes, 1997) at 18°C. After 2–3 weeks each plate containing the paired isolates was assessed for the presence of oospores. Isolates that produced oospores in the presence of a known A2 tester were designated as the A1 mating type, and vice versa (Forbes, 1997).

Virulence was determined using a differential set of potato clones containing 11 major P. infestans resistance genes from S. demissum (Malcolmson & Black, 1966). R1 (CIP 801038), R3 (CIP 801041), R4 (CIP 801042), R5 (CIP 801043), R7 (CIP 801045), R8 (CIP 801046) and R9 (CIP 800994) differential cultivars were provided by CIP. R2 (CIP 800987), R6 (CIP 800991), R10 (CIP 800995) and R11 (CIP 800996) differentials, and the susceptible cv. Bintje, which has no known R genes and was used as a control, (Montarry et al., 2010) were obtained from Wageningen University. This set of differentials is commonly used to characterize P. infestans isolates. Each isolate was inoculated on the differentials. From each plant, three leaflets were taken and placed in inverted Petri dishes containing water agar. On the abaxial surface of each leaflet, two 20 μL drops containing 25 × 103 sporangia mL−1 were placed at each side of the mid-vein. The inoculated leaves were placed in a climate chamber at 16°C for 6 days with a 12 h photoperiod, after which the reaction was scored. The reaction was considered compatible if a necrotic lesion and/or sporulation was observed on a leaflet, and as incompatible when a hypersensitive reaction was seen or no lesion was visible.

DNA extraction

All the P. infestans isolates obtained from potato landraces were grown for 10–14 days on Pea broth (Forbes, 1997). The mycelium was harvested and lyophilized. DNA was extracted from each sample with AGOWA sbeadex Maxi Plant kit using a KingFisher96 robot (Thermo Fisher Scientific).

SSR amplification and genetic data analysis

Twelve SSRs were used in this study: Pi04, Pi63, Pi70, G11, D13, Pi4B (Lees et al., 2006); PinfSSR2, PinfSSR3, PinfSSR4, PinfSSR6, PinfSSR8 and PinfSSR11 (Li et al., 2010). These were selected from previously published sets according to their map position, ease of scoring and allelic diversity (van der Lee et al., 2004; Knapova & Gisi, 2002; Lees et al., 2006; Li et al., 2010). Amplifications were run in a PTC200 thermocycler (MJ Research), with an initial denaturation at 95°C for 15 min, followed by 30 cycles of 95°C for 20 s, 58°C for 90 s, and 72°C for 60 s, and a final extension at 72°C for 20 min (Li et al., 2010). A number of isolates from previous surveys carried out in Ecuador were included in the analysis for reference (Table S2). The resulting amplification products were sized by capillary electrophoresis on an ABI 3730 sequencer using the molecular size standards GeneScan-500 ROX, and scored using GeneMapper v. 3.7 (Applied Biosystems). Within the GeneMapper software kits, panels and bin sets were generated, defining the markers and their known allele bins (provided by Plant Research International, PRI, Wageningen, the Netherlands). The known panels and bin sets for the 12 SSRs have been used to size SSR alleles in thousands of isolates from different countries by PRI and the James Hutton Institute, UK. Most P. infestans isolates possessed a maximum of two alleles per locus, as expected for a diploid organism. However, in a number of cases more than two alleles were found, which may result from aneuploidy or polyploidy (see Fig. S1 for examples). This complicates the analysis because the analysis tools assume haploid or diploid data. Therefore, the detected SSR fragments (alleles) were converted to binary presence (1) or absence (0) data. From this a neighbour-joining tree was produced using paup* v. 4.0 Beta software. Robustness of the phylogram branches was inferred from jackknife values after 10 000 replicates (Swofford, 2002).

Diversity analysis

Race and genotypic diversity were estimated using Shannon's index (HS) as H= −∑ (pi × lnpi), where pi is the frequency of the race or genotype. Evenness (E) was estimated by the formula: HS/ln(n), where n is the total number of isolates of the sample and HS is the Shannon index. E varies from 0 to 1, with 1 representing a situation in which all races or genotypes are equally abundant (Magurran, 1988).

The significance of Shannon indices was assessed with Hutcheson's t-test (tH) (Hutcheson, 1970). It was calculated as follows: tH = (H1H2)/[S2H1+S2H2]1/2, where H1H2 is the difference of the Shannon indices between two samples. The variance of each sample (S2) was estimated with the formula: S2 = [∑ pi(lnpi)2 − (∑pilnpi)2/n]/n2, where pi and n are as described above.

The degrees of freedom (df) for each pairwise comparison were calculated using the formula: df = [S2H1+S2H2]2/[((S2H1)2/n1) + ((S2H2)2/n2)].

To examine the distribution of genetic variation within and among populations, analysis of molecular variance (amova) using Arlequin v. 3.5 (Excoffier et al., 2005) was carried out with the binary data set as input. Population differentiation (FST) was calculated using Arlequin.

To compare the subclonal variation in the Ecuadorian to other populations, the genotyping data of 219 Dutch isolates was used. These isolates were collected during the period 2004–2009 in the Netherlands and genotyped with the same 12 SSR markers. The genotyping data and isolate information can be found in Li et al. (2012a). In addition the Shannon index and the evenness were calculated using the frequency data in Nicaragua reported by Blandón-Díaz (2011). These were based on four multilocus genotypes obtained with six SSRs (D13, G11, Pi04, Pi4B, Pi63 and Pi70) on 165 P. infestans isolates collected from potatoes in Nicaragua.

Results

Isolate collection

A total of 66 P. infestans isolates were collected during the years 2009–2010 from 16 farms (Table 1; details of the isolates are provided in Table S1). The reason for the low number collected in Loja was a severe drought that occurred in that province at the time of collection, making it difficult to obtain samples.

Table 1. Diversity of the Phytophthora infestans population associated with potato landraces in three provinces of Ecuador based on number of races and SSR profiles
 CarchiChimborazoLojaAll sites
  1. a

    HS: Shannon index.

  2. b

    E: evenness.

  3. c

    Values followed by same letter within one line are not significantly different (α = 0.05) according to Hutcheson's t-test for pairwise comparisons.

No. of farms83516
No. of isolates3720966
Virulence phenotype
No. of races2718749
H S 3·15a2·86b1·89c3·78
E 0·880·950·860·91
SSR genotype
No. of genotypes179831
H S 2·44a1·94b2·04b3·03
E 0·680·660·930·72

Isolate characterization using potato R gene differentials

All isolates were of the A1 mating type. There were 49 different races in total (Table 1). The number of races collected and associated virulence factors are detailed in Table S3. Twenty-seven races were identified in Carchi, 18 in Chimborazo and seven in Loja. The P. infestans population observed was complex, and virulent on between four and 11 R-genes. One race (1,3,4,5,6,7,8,9,10,11) was present in all three provinces; the other races were restricted to one province (Table S3).

Isolate characterization using SSR markers

Eight of the 12 SSR markers were polymorphic among the P. infestans isolates collected. Markers Pi04, Pi63, Pi70 and PinfSSR8 were monomorphic (Table S4).

For some markers three or even four alleles were observed in a particular isolate. In the P. infestans isolates from potato landraces, three alleles were observed with Pi4B and PinfSSR4 markers (Table S4). In the reference P. infestans isolates, three alleles were detected with markers Pi4B, Pi63, PinfSSR4 and PinfSSR8 (Table S5). In the case of P. andina, three to four alleles were detected with the Pi4B marker (Table S5). Several private and monomorphic alleles were detected. Allele 279 from locus Pi63 was detected in all P. infestans isolates. Allele 188 from locus Pi70 was present in all US-1 reference isolates and absent from the rest of the P. infestans isolates. In total, 31 different genotypes were observed among the 66 isolates (Table 1) using the SSRs.

Relationship among the materials collected

The phylogenetic tree obtained from the SSR profiles showed that the P. infestans isolates grouped together in a large clade (Fig. 1; structure analysis in Fig. S2). This clade also included the reference isolates known to belong to the clonal lineage EC-1. They were clearly distinct from the reference isolates identified as US-1 and P. andina. There was no relationship between the multilocus genotypes and the race phenotype. Isolates that belonged to the same race could possess different multilocus genotypes and vice versa. Even isolates obtained from the same landrace and sharing the same virulence phenotype could possess different SSR multilocus genotypes (Table S1).

Figure 1.

Neighbour-joining tree of the Ecuadorian Phytophthora infestans populations associated with potato landraces, based on 12 SSRs. Numbers at the nodes are jackknife values supporting the branches based on 10 000 replicates. (a) European P. infestans isolates; (b) P. infestans isolates, clonal lineage US-1; (c) P. andina isolates; (d) P. infestans isolates, clonal lineage EC-1.

Diversity among regions and development of diversity over time

In total, 31 different multilocus genotypes were identified. Only one genotype (EC-1_001) occurred in all three provinces, but it predominated in Carchi; another one (EC-1_003) was found in two provinces (Carchi and Chimborazo). The others were restricted to one province (Table S4). To examine the distribution of genetic variation among and within subpopulations, as defined by the three geographical regions, amova was performed. This showed that 82% of the variance was present within subpopulations whereas the remaining 18% was among the three populations. There was a clear relationship between genotype and geographic origin (FST = 0·18; < 0·0001).

Carchi had the highest Shannon diversity index for race phenotype (3·15) and it was significantly more diverse than Chimborazo (t= 23·28, df = 46·29, < 0·0001) and Loja (t= 7·31, df = 15·32, < 0·001; Table 1). Chimborazo, with a Shannon index of 2·86, was more diverse than Loja (t= 17, df = 17·13, < 0·0001), which showed the lowest diversity value (1·89). The evenness among provinces showed slight, non-significant, differences.

Based on the SSR markers, the P. infestans population in Carchi (HS = 2·44) was genetically more diverse than the population in Chimborazo (t= 18·56, df = 30·39, < 0·001) and Loja (t= 5·48, df = 9·43, < 0·001). The populations in Chimborazo and Loja were not significantly different from each other (t= −0·80, df = 10·73, = 0·372; Table 1). The total race diversity (H= 3·78; Table 1) was significantly higher than the genotypic diversity (HS = 3·03; t= 12·66, df = 66·28, = 0·007; Table 1). Analysis of the genotype changes over time in the Ecuadorian population of P. infestans indicated there was an increase in diversity, evenness and number of genotypes (Table 2).

Table 2. Changes in number of races, Shannon index (HS) and evenness (E) of Phytophthora infestans populations over the years in Ecuador
 YearProvinces
CarchiChimborazoLoja
  1. a

    Calculated from data of Forbes et al. (1997).

  2. b

    Calculated from data of Tello (2008).

  3. c

    Data from this research.

  4. d

    Not evaluated.

H S 1990–1993a2·332·121·55
2007b2·81n.e.dn.e.
2009–2010c3·152·861·89
E 1990–1993a0·640·580·45
2007b0·67n.e.n.e.
2009–2010c0·880·950·86
Races1990–1993a14 (= 39)14 (= 38)8 (= 31)
2007b27 (= 68)n.e.n.e.
2009–2010c27 (= 36)18 (= 20)7 (= 9)

Comparison of the variation within EC-1 in Ecuador to other clonal lineages

Genetic variation was also compared within the Ecuadorian population (the three regions taken together) with the variation in a clonal lineage that is dominant in Europe. In total, 219 Dutch isolates from the clonal lineage Blue_13 were studied with the same set of SSR markers (Li et al., 2012a). The markers split the Blue_13 clonal lineage into 32 different subclones. The genotypic diversity (Hs = 1·98, = 0·37) in the Dutch clonal lineage was significantly (tH = −5·00, df = 177·88, = 1·36 × 10–6) lower than the diversity (Hs = 3·03, E = 0·72) in the Ecuadorian population. The diversity present in the Nicaraguan population, based on the data provided by Blandón-Diaz (2011) (165 isolates, 4 SSR multilocus genotypes, Hs = 0·19, = 0·04), was also much lower than in the Ecuadorian population measured with the same set of SSRs (Hs = 1·89, = 0·45, t= 46·38, df = 94·86, < 0·0001).

Discussion

Sixty-six isolates of P. infestans were obtained from 16 farms in three provinces of Ecuador. As a first step in the characterization, their virulence was analysed on a set of differential genotypes containing the P. infestans resistance genes R1 to R11 (Black et al., 1953; Malcolmson & Black, 1966), as these genes and functional homologues are widely distributed in Solanum spp. (Vleeshouwers et al., 2011). Using this differential set of genotypes, 49 (75%) of the isolates were found to be unique races. Simple races capable of overcoming one to three R genes were not observed. A similar high frequency of different races among isolates has been reported from Costa Rica (37 out of 40; Barquero et al., 2005). However, lower frequencies were observed in other studies, e.g. Nepal 30 out of 251, Estonia 86/432, Finland 66/269, Norway 38/105, Nordic countries (Denmark, Finland, Norway and Sweden) 31/177 and China 61/125 races (Hermansen et al., 2000; Ghimire et al., 2001; Lehtinen et al., 2008; Li et al., 2009; Runno-Paurson et al., 2010). It should be noted that in Denmark, Estonia, Finland, Norway and Sweden sexual populations are dominant, and the populations seems to be clonal in China and Nepal (although both mating types can be found), whereas the population in Ecuador is clonal.

The variation in the Ecuadorian population may be maintained due to a lack of selection for a particular race. Most of the Ecuadorian potato landraces are highly susceptible to late blight (Cañizares & Forbes, 1995; Revelo et al., 1997; Monteros-Altamirano, 2011). In addition, fungicides are rarely used in the small-scale farming system used for these potato landraces. Grünwald et al. (2006) reported that populations of P. infestans not exposed to the fungicide metalaxyl showed more genetic diversity than exposed ones.

The current study has confirmed that P. infestans attacking potato in the Ecuadorian highlands is of the EC-1 clonal lineage. Notwithstanding the fact that all isolates from potato landraces belonged to the same clonal lineage, genetic variation was detected using the SSRs and by phenotypic characterization. Thirty-one multilocus SSR genotypes were detected among the 66 isolates. This variability is large for a clonal lineage. Clonal lineages have been identified in Europe (Cooke et al., 2006, 2012) and Nicaragua (Blandón-Díaz et al., 2012) using microsatellites. Guo et al. (2009) identified a single clonal lineage in northern China using two SSRs; Li et al. (2012b) identified three clonal lineages in western China using ten SSRs. In the Nordic countries (Denmark, Finland, Norway and Sweden), 169 SSR multilocus genotypes were identified using seven microsatellite markers from a sample of 191 P. infestans isolates. This high genetic diversity observed in Nordic countries is attributed to the sexual reproduction of the pathogen (Brurberg et al., 2011). Comparing the results of the current study to published data is not easy as different sampling strategies and different numbers of SSR markers have been used: the best comparison is with the European Blue_13 lineage as this study was carried out with the same set of SSR markers. The variation in the Ecuadorian EC-1 lineage was significantly greater than the diversity in Blue_13 as measured by the Shannon index and evenness. When the results of the current study are compared to the population present in Nicaragua (only considering markers that were present in both studies), the variation in the Ecuadorian population was greater than that in the Nicaraguan population. The high diversity in Ecuador might be related to a high mutation frequency, e.g. due to increased UV radiation as potatoes are grown in Ecuador at high altitudes (>2400 m). It is also possible that the greater diversity in the EC-1 population than in the clonal lineage Blue_13 is due to the age of the population: Blue_13 was detected for the first time in samples from 2004 in the Netherlands (Li et al., 2012a), whereas EC-1 was described in the 1990s (Forbes et al., 1997).

Mechanistically, the subclonal variation may partly be explained by loss of chromosome regions or mitotic recombination. Nevertheless, new alleles were detected that can only be explained by changes in the number of repeat units during mitosis. The presence of more than two peaks in some of the SSR profiles supports the hypothesis that polyploidization or gene duplication followed by mutations caused the genetic diversity observed in the P. infestans population of Ecuador (Tooley & Therrien, 1991; Cooke et al., 2012). The occurrence of more than two alleles at a specific locus has been reported before (Knapova & Gisi, 2002; Lees et al., 2006; Chacón Acosta, 2007; Akino et al., 2009; Oliva Pérez, 2009; Cooke et al., 2012; Li et al., 2012a,b). The same mechanism seems to be active in P. andina also, where four alleles were observed (Fig. S1). Changes in virulence spectrum have been attributed to (partial) chromosomal deletions (van der Lee et al., 2001). Also, mutations in avirulence genes have been found to cause changes in the virulence phenotype (Armstrong et al., 2005). Copy number variations, amino acid replacements, and gene gains and losses have been suggested as sources of the variability within the clonal lineage Blue_13 in the UK (Cooke et al., 2012). The high race and SSR diversity observed in the Ecuadorian population might be related to the cultivation of potatoes all year round, so several generations of the pathogen can occur, increasing the chance of the appearance of new genotypes.

There was no correlation between the phenotypic and genotypic diversity. Isolates of a particular race showed different multilocus genotypes. These occurred even in isolates obtained from the same landrace or farm (Table S1). Thus, the observed variation appears to be produced randomly with no selective pressure. The high race and SSR diversity observed in this study contrasts with populations from USA, France and China, where clonal lineages were shown to possess a wide diversity in races but a low genetic diversity (Abu-El Samen et al., 2003; Montarry et al., 2006; Guo et al., 2009).

The genetic diversity was highest in the Carchi province and significantly different from Chimborazo and Loja (Table 1). The FST analysis showed clear differentiation of the populations in the three regions and most of the variation was present within populations. This difference in diversity observed in Carchi could be due to the higher number of isolates and farms sampled. The highest evenness value was measured in Loja (0·93), although the number of samples was small (Table 1) and the value probably reflects under-sampling. The one genotype that occurred in all three provinces might have migrated through the Ecuadorian highlands, perhaps as a result of the exchange of landrace seeds among farmers (Monteros-Altamirano, 2011). Alternatively, this genotype may have arisen multiple times independently. The microsatellite analysis also showed some P. infestans lineage- and species-specific alleles. These may be used to distinguish clonal lineages (Akino et al., 2009). The SSR markers used in this study reflected the genetic diversity within the current P. infestans population in Ecuador: in future they may be used to monitor changes in the population and displacement of pathogen genotypes across the country.

Comparisons with previous studies have to be treated with caution because of the different sampling strategies used. However, an increase in the Shannon index and evenness over time was apparent, as was an increase in the number of races. The differences in Carchi in 2007 and 2009–2010 may be due to the fact that the present study survey included eight farms, whereas Tello (2008) collected all the samples from just one farm in the province and some races were much more frequent than others. The evenness values are close to 1 for all provinces, which was not the situation in 1990–1993, where in Carchi two races represented 40% of the isolates. For Chimborazo and Loja, the situation was similar, two races at each location representing 55 and 70% of the isolates sampled (Forbes et al., 1997).

All P. infestans isolates from potato landraces grouped in a clade together with known EC-1 isolates, separate from US-1 and P. andina isolates (Fig. 1). There is currently some discussion as to whether P. andina should be considered a separate species or a hybrid. Although SSR markers are not particularly suitable for species identification, some of the alleles identified in P. andina are rare and have never been reported in P. infestans isolates. In addition, in P. andina there were four alleles at locus Pi4B and null alleles for locus Pi70 which have not been found for other P. infestans isolates (Y. Li & T. van der Lee, Plant Research International, Wageningen, The Netherlands, personal communication). This is not a typical pattern for a simple hybridization, but clearly distinguished P. andina isolates from P. infestans isolates.

This study has identified genetic variation within the P. infestans clonal population in Ecuador. The high number of races and their complexity constitute a challenge for late blight management in the country. It is necessary to incorporate new R genes not belonging to the S. demissum group into the National Breeding Programme. Other sources of resistance (Wang et al., 2008; Jacobs et al., 2010; Lokossou et al., 2010) may be useful for potato breeding against P. infestans in Ecuador. However, one promising broad-spectrum resistance gene, Vnt1 (Pel, 2010) cannot be used because the avirulence gene triggering the defence response is not expressed in EC-1 genotypes.

Acknowledgements

The authors thank the Netherlands Organization for International Cooperation in Higher Education (NUFFIC), the Secretaría Nacional de Educación Superior, Ciencia, Tecnología e Innovación (SENESCYT), Ecuador and the Instituto Nacional Autónomo de Investigaciones Agropecuárias (INIAP), Ecuador for financial support, and the International Potato Centre (CIP), Ecuador for sharing lab facilities and technical assistance. Dr Ronald van den Berg is also thanked for valuable discussions and producing the dendrogram. Two anonymous reviewers are thanked for their critical comments on an earlier version of the manuscript.

Ancillary