Introduction

Powdery mildew on tomato, caused by Oidium lycopersicum Cooke & Massee (Noordeloos & Loerakker, 1989), is the only fungal disease which has to be controlled by fungicides in protected tomato cultivation in NW Europe. Complete or high levels of resistance have been found in at least four Lycopersicon species, i.e. L. esculentum var. cerasiforme, L. hirsutum, L. parviflorum and L. peruvianum (Kozik, 1993; Laterrot & Moretti, 1993; Neshev, 1993; Lindhout et al., 1994; Ciccarese et al., 1998). The resistance in Gl.1560 to O. lycopersicum is controlled by an incompletely dominant gene, Ol-1 which was mapped on chromosome 6 between RFLP markers GP79 and TG153, in the vicinity of the resistance genes Mi to Meloidogyne spp. and Cf-2/Cf-5 to Cladosporium fulvum (Van der Beek et al. 1994). More recent data have confirmed the monogenic resistance in G1.1560, and more accurately mapped Ol-1 between SCAF10 and H9A11 (Huang et al., 2000).

Resistance to O. lycopersicum seems to be widely distributed across Lycopersicon species (see above). It is an intriguing question how the corresponding resistance genes have originated during evolution and how related these genes are. Clustering of disease resistance genes is quite common in plants (Michelmore & Meyers, 1998); examples are the Cf-cluster and the Pto cluster in tomato (Hammond-Kosack & Jones, 1997). Each such cluster usually consists of homologues that have different specificities to different races of the corresponding pathogens. So far, only resistance of L. hirsutum G1.1560 has been characterized and the corresponding gene Ol-1 has been mapped on chromosome 6 (Van der Beek et al., 1994; Huang et al., 2000). In order to find out whether the O. lycopersicum resistance genes (Ol-genes) are also organized in one or more clusters, we investigated the inheritance of resistance in L. hirsutum G1.1290 and mapped the resistance gene, designated Ol-3. In addition, we provided more evidence for the map position of Ol-1. Finally, we performed allelism tests for Ol-1 and Ol-3 in order to assess their relationship.

Materials and methods

Plant and fungal materials

The resistant wild accessions L. hirsutum G1.1290 and G1.1560 (Lindhout et al., 1994) were obtained from the Centre of Genetic Resources, Wageningen, The Netherlands. L. esculentum cv. Moneymaker, as susceptible control, was maintained at the Laboratory of Plant Breeding, WAU, Wageningen, the Netherlands. Advanced breeding lines (ABLs), ABL1290.4 and ABL1560.2, derived from either G1.1290 or G1.1560, were obtained from commercial breeding programmes for resistance to O. lycopersicum. All of the populations evaluated are described briefly in Table 1. In addition, two other populations, HOl3 and HOl1, heterozygous for Ol-3 and Ol-1 were generated by crossing Moneymaker with ABLs ABL1290.4 and ABL1560.2, respectively, and used as heterozygous controls.

Table 1 Summary of each experiment of disease tests
Full size table

The stock of O. lycopersicum originated from infected commercial tomato plants (Lindhout et al., 1994), and was maintained on cv. Moneymaker plants in a growth chamber at 20 ± 1°C with 70 ± 3% RH and 16-h day length.

Disease tests

Disease tests were carried out in eight experiments during 1994–98. The experimental was set up according to a completely randomized block design with 2 to 18 blocks (Table 1). Each block always contained susceptible and resistant control genotypes. All plants at the four true leaf stage were inoculated by spraying with a spore suspension at a concentration of 3–4 × 104 conidia ml−1. The inoculum was prepared by washing conidial spores from the freshly sporulating leaves of heavily infected Moneymaker plants in tap water and was used immediately. The inoculated plants were grown in a greenhouse at 20 ± 3°C with 70 ± 10% RH under natural light supplemented with artificial light to provide a photoperiod of 16 h.

Disease evaluation

The susceptibility or resistance was determined macroscopically by the degree of fungal sporulation. Depending on the development of the fungus, evaluations were performed per plant twice to five times from 7 to 29 days post inoculation, by using a disease index (DI) from 0 to 3, based on the degree of sporulation. The DI was defined as: 0, no sporulation; 1, slight sporulation, but less than 5% foliar area affected; 2, moderate sporulation, between 5 and 30% foliar area affected; 3, abundant sporulation, more than 30% foliar area affected. In addition, in order to evaluate each plant more precisely, DI from 0 to 4 was used in the allelism test (also Figs. 5 and 6): 0, no symptom; 1, some mycelia without sporulation; 2, moderate mycelia with poor sporulation; 3, between 2 and 4; 4, abundant mycelia with heavy sporulation.

Fig. 5
figure 5

Frequency distribution of resistance to O. lycopersicum infection in L. hirsutum G1.1290, L. hirsutum G1.1560, L. esculentum cv Moneymaker and their ‘BC1wa’ [MM × ‘F1wa’ (G1.1560 × Ν/G1.1290)] progenies of the allelism test. Disease index (DI): 0, no symptom; 1, some mycelia without sporulation; 2, moderate mycelia with poor sporulation; 3, between 2 and 4; 4, abundant mycelia with heavy sporulation. Evaluations were done at 10, 14 and 21 days post inoculation (DPI).

Full size image
Fig. 6
figure 6

Susceptibility of the ‘BC1S1’ lines. (a) Frequency distribution of resistance to O. lycopersicum infection in the ‘BC1waS1’ lines (indicated by Arabic numbers, Lines 2′ & 15′ were repeats of Lines 2 & 15, respectively) at 28 days post inoculation (DPI). These lines were generated by self-pollinating the most susceptible ‘BC1wa’ plants. Here, ‘BC1wa’ was obtained by crossing cv Moneymaker (MM) with the progeny (‘F1wa’) of the cross between L. hirsutum G1.1290 and L. hirsutum G1.1560. HOl3 and HOl1 are heterozygous for Ol-3 and Ol-1, respectively. For each genotype 17–23 plants were tested. Evaluations were done at 10, 14 and 28 dpi. (b) Frequency distribution of resistance to O. lycopersicum infection in the ‘BC1ablS1’ lines (indicated by a–f) at 21 days post inoculation (DPI) (dpi). These lines were generated by self-pollinating the most susceptible ‘BC1abl’ plants. Here, ‘BC1abl’ was obtained by crossing cv Moneymaker with the progeny (‘F1abl’) of the cross between two ABLs ABL1290.4 and ABL1560.2. For each genotype 17–23 plants were tested. Evaluations were carried out at 10, 14 and 21 days post inoculation (DPI). Disease index (DI): 0: no sporulation; 1: slight sporulation, but less than 5% foliar area affected; 2: moderate sporulation, between 5 and 30% foliar area affected; 3: abundant sporulation, more than 30% foliar area affected.

Full size image

Marker analysis

Total DNA was extracted from the frozen young leaves of most plant materials as described by Van der Beek et al. (1992), except for the ABLs in expt 2 from which DNA was isolated from a leaf-disc as described by Wang et al. (1993). DNA (5 μg) was digested with the restriction enzymes EcoRI and HaeIII (Life Technology), based on a pilot test of enzyme-probe combinations. RFLP analysis was performed as described by Van der Beek et al. (1992), and SCAR analysis as by Paran & Michelmore (1993) and Huang et al. (2000). The four TG markers, TG153, TG25, TG164 and TG240, were provided by S. D. Tanksley, Cornell University, New York, USA and H9A11 was supplied by C. Gebhardt, Max Planck Institut für Züchtungsforschung, Cologne, Germany. The SCAR markers were from by P. Zabel, Laboratory of Molecular Biology, WU, Wageningen, The Netherlands.

Linkage analysis

The map** program JOINMAPTM 2.0 (Stam & Van Ooijen, 1995) was used to estimate the map distances. DRAWMAP 2.0 (Van Ooijen, 1994) was employed to produce graphics of the maps.

Results

Inheritance and map** of the resistance to O. lycopersicum in L. hirsutum G1.1290

The inheritance of resistance to O. lycopersicum was studied by applying a disease test on a BC1 population of L. esculentum cv. Moneymaker × L. hirsutum G1.1290. All plants were evaluated for the degree of sporulation expressed as disease index (DI) at scales from 0 to 3. The plants of resistant control G1.1290 either remained uninfected or were scored at most 1, while almost all plants of susceptible control Moneymaker were scored 3 (Fig. 1). All F1 plants were scored 0–1, and hence were considered as resistant (Fig. 1), indicating that the resistance was dominant. The BC1 segregated into 69 resistant and 80 susceptible plants, according to a 1:1 ratio (χ21=0.81, P=0.50–0.25), in accordance with one dominant resistance gene, which we designated as Ol-3. In addition, 29 BC1 plants could not be classified clearly because of their intermediate disease score (2). This might indicate the incomplete dominance of Ol-3 or existence of minor gene effects. It might also be a result of the influence of genetic background of L. esculentum on the expression levels of resistance as sometimes noticed in the ABL (also see Fig. 3g).

Fig. 1
figure 1

Frequency distribution of resistance to O. lycopersicum infection in L. hirsutum G1.1290, L. esculentum cv Moneymaker and their F1 and BC1 progenies. The population size is indicated between brackets. Disease index (DI): 0, no sporulation; 1, slight sporulation, but less than 5% foliar area affected; 2, moderate sporulation, between 5 and 30% foliar area affected; 3, abundant sporulation, more than 30% foliar area affected. Evaluations were done at 13, 16 and 20 days post inoculation (DPI).

Full size image
Fig. 3
figure 3

Analysis of putative additional genes (a & b) or suppressor genes (c & d). L. hirsutum G1.1560 (e) and the ABL (g) served as resistant control, with L. esculentum cv. Moneymaker (f) as susceptible control (see text). The experiment consisted of six blocks, each contained nine to 13 ‘F2’ plants, five each of Moneymaker and G1.1560, and four of ABL1560.2. Disease index: 0, no symptom; 1, some mycelia without sporulation; 2, moderate mycelia with poor sporulation; 3, abundant mycelia with heavy sporulation. Evaluations were carried out at 10 and 14 days post inoculation (DPI).

Full size image

As a first step towards map** the Ol-3 gene, a resistant advanced breeding line (ABL1290.4) derived from L. hirsutum G1.1290 was analysed for the markers TG153, TG25, SCAF10 and TG164 on chromosome 6, which are closely linked to Ol-1 (Huang et al., 2000). The resistance in this ABL was found to be associated with L. hirsutum alleles at these markers. Obviously, this ABL contained a fairly large introgression fragment (approximately 10 cM) between markers TG153 and TG164 (Fig. 2). This result at least indicated that Ol-3 was located on chromosome 6, near Ol-1.

Fig. 2
figure 2

Genetic linkage map of part of chromosome 6 of tomato, showing the position of the genes Ol-1 (left) and Ol-3 (right) for resistance to tomato powdery mildew. Their relative position is shown in an integrated genetic map (centre). Ol-1 and Ol-3 were mapped by using an F2 of L. esculentum cv. Moneymaker × L. hirsutum G1.1560 and a BC1 of Moneymaker × L. hirsutum G1.1290, respectively. Markers in normal face are RFLPs, in bold are SCARs.

Full size image

To map the gene Ol-3, RFLP analyses with chromosome 6 markers, were performed on 51 resistant (scored at most 1) and 58 susceptible (always scored 3) BC1 plants. After testing 15 probe–enzyme combinations, 11 combinations generated sufficient polymorphisms between the parents Moneymaker and G1.1290 (data not shown). One combination, TG240 with HaeIII, even showed a difference between the two L. hirsutum accessions G1.1290 and G1.1560, indicating a genetic difference between these two accessions. In the interval where Ol-3 was mapped by using ABL1290.4, five probe–enzyme combinations (TG153, TG164 and TG240 with HaeIII, and TG25 and H9A11 with EcoRI) and an L. hirsutum specific SCAR marker, SCAF10, were chosen for more precise map** (Fig. 2). Based on the marker data and the disease resistance evaluation of the BC1 plants, map distances between Ol-3 and RFLPs as well as SCAR were estimated. This map was very reliable (χ2=0.139) and the order of the markers was the same as reported by Tanksley et al. (1992). The most likely map position of the resistance gene Ol-3 was between TG25/SCAF10 and H9A11 (Fig. 2).

Fine map** of Ol-1

Because Ol-3 had previously been mapped in the same chromosome region as Ol-1, we wanted to investigate further the relationship between these two Ol genes. Although Ol-1 had been mapped between SCAF10 and H9A11 in an F2 (Huang et al., 2000), the accuracy of map** also allows a slightly different position. To improve the accuracy of the map** of Ol-1, the F3 progenies of 17 F2 plants (Table 1) with ambiguous disease resistance evaluation, were genotyped for the SCAR markers and evaluated by a disease test to more accurately evaluate the resistance or susceptibility of these F2 plants. By doing so, most of the F2 plants that had been scored ambiguously could now be classified as either resistant or susceptible, which was confirmed by the SCAR-genotype. For two of the F3 progenies from F2 plants nos 101 and 270, however, the SCAR-genotype and the disease rating were contradictory.

The F3 progeny of no. 101 did carry the Ol-1 resistance allele, as judged by SCAR analysis, but was also susceptible to O. lycopersicum. These plants might contain suppressor gene(s). In order to find a better explanation of such conflicting observations, they were crossed with resistant F3 plants that did carry the Ol-1 resistance allele. Then the ‘F2’ progeny of these crosses were subjected to a disease test. These were as resistant as the resistant parent G1.1560 (Fig. 3a,b), indicating the absence of such putative suppressor gene(s) and the presence of the Ol-1 resistance allele. Therefore, the corresponding F2 plant no. 101 was re-interpreted as resistant. In contrast, the F3 progeny of no. 270 did not contain the Ol-1 allele as judged by SCAR analysis but was evaluated to be resistant. These plants might contain additional gene(s) which also confer resistance to O. lycopersicum. Two of them were crossed with Moneymaker and the two derived ‘F2’ progenies were evaluated in a disease test. Both ‘F2’ populations tested were as susceptible as Moneymaker (Fig. 3c, d), indicating the absence of such putative additional gene(s) as well as the absence of the Ol-1 allele. Therefore, the corresponding F2 plant no. 270 was re-interpreted as susceptible. After re-interpreting the two F2 plants nos 101 and 270, Ol-1 was still mapped between SCAF10 and H9A11, in the same region as Ol-3 (Fig. 2).

In addition, 56 ABLs derived from L. hirsutum G1.1560 were used to determine the introgression fragment(s) around the Ol-1 gene. They were subjected to RFLP and SCAR analyses with five RFLPs and four SCARs, which are tightly linked to Ol-1. The ABLs could be classified into eight classes according to the size of the introgression fragments. Six of them contained a large introgression fragment almost covering the whole region between SCAE16 and TG240 (not shown). The genotypes ABL1 and ABL2 showed smaller introgression fragments, which indicated that Ol-1 was indeed between SCAF10 and H9A11 (Fig. 4). This was in accordance with our earlier map** study of Ol-1 (Fig. 2).

Fig. 4
figure 4

Graphical map of part of chromosome 6, showing the map position of Ol-1 between SCAF10 and H9A11. This map was constructed by using ABLs carrying the Ol-1 resistance. Chromosome fragments in black represent L. hirsutum (Lh) loci, in white L. esculentum (MM) loci, in grey heterozygous loci. The relative map distance of each marker was based on the Ol-1 map (Fig. 2).

Full size image

Allelism test of Ol-3 and Ol-1

As stated above, Ol-1 and Ol-3 were mapped at similar positions on chromosome 6. Their relative distance was about 2 cM in an integrated map (Fig. 2), constructed by using both the Ol-1 map and Ol-3 map, suggesting that Ol-1 and Ol-3 are two different genes. However, because the integrated map was constructed by using linkage data from two different populations — F2 and BC1 — the distance between Ol-1 and Ol-3, and the 4 cM shift of map position of SCAG11, might have been results of experimental error. In order to confirm the relative map position of Ol-1 and Ol-3, allelism tests were conducted by using three-way test-cross progenies derived from wild accessions and ABLs, respectively (Table 1).

Most of plants of the three-way-cross progeny ‘BC1wa’ [Moneymaker × ‘F1wa’(G1.1290 × G1.1560), also see Table 1] were resistant, but some plants might be considered as susceptible (Fig. 5). However, most plants of the selfed progenies (‘BC1waS1’) from the most susceptible ‘BC1wa’ plants were resistant (Fig. 6a), indicating the absence of susceptible recombinants in the test-cross progenies (‘BC1wa’). Because no recombinant was found in the ‘BC1wa’ consisting of 135 plants, the distance between Ol-1 and Ol-3 was estimated to be smaller than 2.2 cM (P=0.05).

In order to test the allelism of Ol-1 and Ol-3 in a more L. esculentum genetic background, ABL1290.4 and ABL1560.2 containing Ol-3 and Ol-1, respectively, were used to make similar three-way-crosses as for the wild accessions. Similarly, most plants of the three-way-cross progeny ‘BC1abl’ [Moneymaker × ‘F1abl’(G1.1290 × G1.1560), also see Table 1] were resistant, but some plants might be considered susceptible (Fig. 7). However, most plants of the selfed progenies (‘BC1ablS1’) from the most susceptible ‘BC1abl’ plants were resistant (Fig. 6b), indicating the absence of susceptible recombinants in the test-cross progenies (‘BC1abl’). Compared with the resistant parental lines and the progeny of the putative susceptible plants from the experiment with wild accessions (Fig. 6a), the higher levels of susceptibility of the selfed plants from the putative susceptible BC1abl plants, might have been a consequence of a higher percentage of L. esculentum genome in the genetic background (e.g. Fig. 3g). Because no recombinant was found in the ‘BC1abl’ consisting of 193 plants, the distance between Ol-1 and Ol-3 was estimated to be smaller than 1.5 cM (P=0.05). This result was fully in agreement with that found using wild accessions.

Fig. 7
figure 7

Frequency distribution of resistance to O. lycopersicum infection in L. hirsutum G1.1290, L. hirsutum G1.1560, advanced breeding lines ABL1290.4 containing Ol-3 and ABL1560.2 containing Ol-1, ‘F1abl’ of ABL1290.4 × ABL1560.2, L. esculentum cv Moneymaker (MM) and the ‘BC1abl’ [MM × ‘F1abl’ (ABL1290.4 × ABL1560.2)] populations. The DI scores were as defined in Fig. 5. Evaluations were carried out at 7 and 14 days post inoculation (DPI).

Full size image

Discussion

In the present study, we concluded that the resistance in L. hirsutum G1.1290 was controlled by one major gene designated Ol-3, that mapped between markers TG25/SCAF10 and H9A11 on chromosome 6 of tomato. The existence of some BC1 plants that were ambiguous during disease evaluation, suggested that Ol-3 was incompletely dominant. To map Ol-3, a BC1 population was preferred over an F2 because the larger L. esculentum background in BC1 prevents negative effects of wild accession genes on the disease test, thus increasing the accuracy of disease evaluation. In the BC1, there are only two genotypes, Ol-3/ol-3 and ol-3/ ol-3. The dominant L. hirsutum-specific SCAR marker SCAF10 is informative in discriminating ol-3 heterozygotes and homozygotes. However, the resistance allele(s) cannot be detected with the dominant L. esculentum-specific SCAR markers such as SCAE16, SCAK16 and SCAG11, as they cannot discriminate heterozygote resistant from homozygote susceptible individuals.

Notwithstanding the possible ambiguity in disease evaluations, especially for segregating populations, the disease test of the F3 progenies and SCAR-genoty** showed that only 2 out of 17 F2 plants, which were ambiguously evaluated (Huang et al., 2000), had to be re-interpreted in the present study. Therefore, all the results of disease tests in the present or previous (e.g. Huang et al., 2000) studies were reliable enough for interpreting inheritance of disease resistance. In addition, our investigation on the putative additional gene or suppressor gene strongly demonstrated the necessity of testing more than one generation, and the power of marker-genoty** in genetic studies.

The map positions of Ol-1 and Ol-3 were in the same region between SCAF10 and H9A11. By allelism test, Ol-1 and Ol-3 were demonstrated to be allelic or tightly linked. SCAF10 is the most closely linked marker and can be used, as a simple PCR marker, for indirect selection for either Ol-1 or Ol-3 resistances. However, pyramiding Ol-1 and Ol-3 in one tomato cultivar is still not feasible.

The two tomato powdery mildew resistance genes identified so far, Ol-1 and Ol-3 may be in a cluster on chromosome 6. Disease resistance genes are commonly organized in complex loci or clusters (e.g. Hammond-Kosack & Jones, 1997; Michelmore & Meyers, 1998). The members of each cluster may have originated from genetically similar accessions within a species, such as Cf-2 and Cf-5 from L. pimpinellifolium PI270254 and PI187002, respectively (Stevens & Rick, 1988). However, they may also have originated from very diverse species, like Cf-1, Cf-4 and Cf-9 (Kerr & Bailey, 1964; Jones et al., 1993) from L. esculentum, L. hirsutum and L. pimpinellifolium, respectively (Stevens & Rick, 1988).

Genetically, no difference between Ol-1 and Ol-3 has yet been found. Although Ol-1 and Ol-3 may be identical, they may also be homologues of a gene cluster. Morphologically, L. hirsutum G1.1290 differs from L. hirsutum G1.1560 by its broader leaves and by its higher level of necrosis under greenhouse conditions, especially at high humidity. Molecular analysis also demonstrated variation between the two accessions. For example, a polymorphism between L. hirsutum G1.1290 and G1.1560 was found for TG240. Large genetic variation between and within L. hirsutum accessions has also been reported by Miller & Tanksley (1990). Microscopically, while G1.1560 showed only single cell HR (necrosis confined to the haustorium-invaded cell), G1.1290 plants showed a spreading hypersensitivity response (HR) (necrosis spread to the neighbouring cells of the haustorium-invaded cell) after inoculation with O. lycopersicum (although spreading HR was not always manifest) (Huang et al., 1998). In addition, a lower level of resistance in ABLs carrying Ol-3 than in ABLs containing Ol-1, was often observed (pers. obs.) especially at the seedling stage. However, it is still not known whether such differences between G1.1560 and G1.1290 reflect the difference between the two powdery mildew resistance genes Ol-1 and Ol-3. Therefore, precise comparison of Ol-1 and Ol-3 awaits sequence information that might be obtained after cloning the two genes.