Introduction

The powdery mildew caused by Leveillula taurica (Lèv.) Arn is the most important foliar disease of eggplant (Solanum melongena L.) in Southern Italy and in many other countries in the world. Disease symptoms appear as chlorotic patches on the leaf upper side (Fig. 1) and powdery whitish spots, consisting of conidiophores and conidia of the fungus, occur on the leaf lower side. At later stages of the disease development, affected tissues turn necrotic. The susceptibility of leaves increases with age and therefore symptoms first occur on the basal leaves, but under favourable climatic conditions, apical leaves may also be infected; in this case partial or complete plant shedding may occur.

Fig. 1 
figure 1

Symptoms of powdery mildew caused by Leveillula taurica on the lower side of eggplant leaf (cv Florida Market; a). Conidiophores and conidia of L. taurica (b)

Full size image

The great majority of records on the powdery mildew of eggplant refer to the asexual state Oidiopsis taurica (Lév.) Salmon, while the sexual state L. taurica has been never found on eggplant and is very rare on many other hosts (Palti 1971, 1988). This pathogen differs from most other powdery mildew agents because of its endophytism. In fact, it develops both external and endophytic intercellular mycelia; the latter colonizes the mesophyll and, sometimes, palisade tissue. On eggplant, hyphae from intercellular mycelium emerge through the stomata as conidiophores bearing conidia singly (Fig. 1) or, under high humidity conditions, in short chains.

L. taurica attacks are favored by warm and humid weather, so that epidemics are mainly spread in coastal areas (Palti 1959). In open field in Southern Italy, powdery mildew on eggplant starts from the mid August, when favorable temperature and humidity occur during the night.

Currently the most effective control strategy is sulphur and, more effectively, systemic fungicides such as Triadimefon, Triforina and Fenarimol (Ciccarese and Cirulli 1980), however the timing of fungicide applications limits this strategy especially during the harvest period. The use of resistant cultivars is considered the best control strategy. Up to present time no specific research on the reaction to L. taurica of S. melongena and wild relative germplasm has been done. In this paper, the results of 4-year testing for resistance to L. taurica of a worldwide collection of eggplant and related Solanum species is reported.

Materials and methods

The screening tests were carried out in Bari, Southern Italy, in a field located near the coast, where natural powdery mildew infections occur yearly with a high disease pressure. Starting from 2000 through 2003, a total of 172 accessions (S0) of eggplant and wild related Solanum species from different geographical areas were screened for resistance to L. taurica. Most part of accessions was provided from the United States Department of Agriculture (USDA, Washington, DC, USA) and some more were collected in Southern Italy. The S. melongena collection also included accessions of cvs Badanekai, Badanjan, Badenjan, Badenjan Ghalami, Badenjan I Sian, Badenjan Sefid, Badenjan Sesame, Badinjan, Badinjan I Sosani, Bandenjon, Bengan, Black Long Early, Brinja Busa Purple, Brinja Clustered W, Brinja Murtakeshi, Byderpore, Cluster Purple, Dolg, Faizabad Long, Incesu, Kairyo-Oraga, Karisik, Kemer, Kitta Horyo, Kopek, Makhua Jarn, Makhua Kaitao, Makhua Proh, Makhua Yow, Mofale, Morska Pata, Patlican, Pusa Purple Long, Pusa Purple Round, Rihgna, Ringua, Rosita, Su Muo Chia, Taiwan-Naga, Tarung, Teng Pao Chia, Tong Xuan (1) Hao, Topak, Vehgan, Wantak, White Round and Wyned Giant.

The eggplant-related wild Solanum were S. aculeatissimum Jacq., S. aviculare G. Forst., S. capsicoides All., S. gilo Raddi, S. laciniatum Ait., S. linnaeanum Hep. & Jaeg., S. nigrum L., S. pseudocapsicum L., S. quinquangolare L. (cv Guadalupe 3), S. sisymbriifolium Lam., S. spinosissimum Lodd. and other Solanum spp. The cv Florida Market was included in the tests as susceptible check as proved in previous studies (Ciccarese and Cirulli 1980).

In mid-June, seedlings grown in 60 celled polystyrol trays (45 mm diameter × 60 mm height) were transplanted in the field in single rows, 1.5 m between rows and 0.4 m along rows. 0.1 kg m−2 of a 10:10:10 (=N:P2O5:K2O) commercial fertilizer was applied to the soil before transplanting and 0.1 kg m−2 of a nitrogenous fertilizer (27% total nitrogen) was applied after transplanting three times at 20-day interval. A sprinkler irrigation system was used. Routine insecticide sprays were used but no fungicides were applied.

At the end of each year/trial, seeds from open-pollinated single plants belonging to accessions with a low powdery mildew mean severity (≤2) were secured in order to obtain progenies for further testing. Also, seeds from open-pollinated single plants cv Florida Market were secured. Selected plants were left to open-pollinate because final disease evaluations were protracted beyond fruit ripening time and therefore artificial self-pollination was not possible. Original (S0) S. melongena accessions, first (S1), second (S2) and third (S3) generation of selfings were tested. For the wild Solanum species, only S0 accessions and S1 progenies were tested.

A randomized block experimental design with three replicates was used; each replicate was composed of ten plants.

One powdery mildew evaluation was made on the leaf lower side when the maximum disease severity (>75% of infected leaf surface) occurred on the susceptible check. For the disease evaluation, basal and middle leaves were considered and the disease rating was calculated according to a scale of six severity classes: 0 = no symptoms (highly resistant), 1 = 1–5% (highly resistant), 2 = 6–20% (resistant), 3 = 21–50% (susceptible), 4 = 51–75% (susceptible), 5 = 76–100% (highly susceptible) of infected leaf surface.

The assumption of normal distribution of data was not acceptable according to the normality test of Shapiro-Wilk and no transformation procedures were able to normalize effectively the data. For the reason, Friedman test (P < 0.05) was used for the analysis of variance (Friedman 1937). For multiple pairwise comparisons, Dunn’s procedure (Dunn 1964) with the correction of the significance level proposed by Bonferroni was applied.

Results

Field natural infections of L. taurica provided a homogenous disease pressure in each year of testing. In fact, powdery mildew mean severity on cv Florida Market ranked 4.47–5 and was not statistically different in the 4 years of testing (Fig. 3). For all trials, lack in detection of statistical difference among replicates indicated homogeneity of the disease pressure within each year throughout the field (data not shown).

S. melongena original accessions (S0) distributed in all severity classes (Fig. 2). Only a small part of the germplasm was ranked as resistant (disease severity = 2) or highly resistant (disease severity ≤ 1) to powdery mildew. All the cultivars were susceptible (disease severity = 3–4) or highly susceptible (disease severity = 5) to the disease, except for cv Mofale (P.I. 232079) originating from South Africa, which was selected for further testing because showing a mean disease severity of 1.84.

Fig. 2 
figure 2

Powdery mildew severity occurring in Solanum melongena germplasm (S0) and three selected open-pollinated progenies (S1, S2 and S3) in 4 years of selfing and field-screening for resistance. Plant populations with same letter are not significantly different according to the Dunn’s procedure (P < 0.05; Bonferroni’s corrected significance level = 0.0083)

Full size image

During the first year of testing, 10 original accessions (S0) were selected for further observations. In the following 2 years of trials, testing of S2 and S3 populations was restricted to most promising progenies originating from 5 and 4 original accessions (S0), respectively.

Statistical analysis showed that single plant selection was effective in eliminating highly susceptible genotypes under the pathogen pressure occurring in the field (Fig. 2). In fact, significant difference in overall disease severity was detected among the S0, S1 and S2 populations. That difference was not observed between the S3 and S2 populations. The S0 population included 42.8% of plants ranking in class 5 (highly susceptible reaction), while no highly susceptible plants were observed in S2 and S3 populations. In addition, resistant (disease severity = 2) and highly resistant (disease severity ≤ 1) plants occurred in 8.8% of all plants of S0 population and in 73.4% of all plants in S3 population.

The S3 population included four selected resistant lines originated from PAVEG 10187, PAVEG 10196 (both from southern Italy), P.I. 232079 (cv Mofale from South Africa) and P.I. 419198 (from China). In S3, a significant improvement in powdery mildew resistance level was obtained only for PAVEG 10187 S3 and PAVEG 10196 S3: powdery mildew mean severity of the S3 progenies was statistically lower than the S2, S1 and S0 originating populations, which were not significantly different in the disease severity (Fig. 3). Powdery mildew mean severity of S0, S1, S2 and S3 populations was not statistically different for both P.I. 232079 and P.I. 419198. PAVEG 10187 S3 and PAVEG 10196 S3 showed mean percentage of infected leaf surface lower than 5% (0.68 and 0.73 disease severity, respectively), while it was 6–20% (1.70 and 1.84 disease severity, respectively) on P.I. 232079 S3 and P.I. 419198 S3.

Fig. 3 
figure 3

Powdery mildew mean severity on accessions (S0) and three selected open-pollinated progenies (S1, S2 and S3) of four Solanum melongena selections in 4 years of selfing and field-screening for resistance. Within each accession/progeny, columns with same letter are not significantly different according to the Dunn’s procedure (P < 0.05; Bonferroni’s corrected significance level = 0.0017)

Full size image

All the accessions of S. laciniatum, S. nigrum and S. quinquangolare showed no symptoms of powdery mildew (Fig. 4). S. linnaeanum, S. aculeatissimum, S. aviculare and S. pseudocapsicum were highly resistant with all or most plants showing percentages of infected leaf surface lower than 5% (disease severity ≤ 1), S. spinosissimum was resistant, S. gilo and S. capsicoides were susceptible or highly susceptible to the disease. A high variability in the powdery mildew severity was detected in S. sisymbriifolium.

Fig. 4 
figure 4

Powdery mildew severity occurring in germplasm (S0) and selected open-pollinated progenies (S1) of wild Solanum species related to eggplant in 2 years of selfing and field-screening for resistance. Within each Solanum species, plant populations with same letter are not significantly different according to the Dunn’s procedure (P < 0.05; Bonferroni’s corrected significance level = 0.05)

Full size image

Reactions to the powdery mildew observed on the accessions (S0) of S. linnaeanum, S. laciniatum, S. aviculare, S. aculeatissimum, S. sisymbriifolium and S. nigrum were statistically confirmed in their S1 progenies. For the other species, no S1 progenies were tested because no viable seeds were obtained from plants S0.

Discussion

This is the first research on the screening of a large worldwide collection of S. melongena and wild species related to eggplant for resistance to powdery mildew caused by L. taurica. A wide range of reactions to powdery mildew occurred in eggplant germplasm, which included highly resistant (disease severity = 0, no symptoms) to highly susceptible (disease severity = 5, 76–100% of infected leaf surface) accessions. However, most S. melongena accessions were susceptible or highly susceptible showing 21–75% or 76–100% of infected leaf surface, respectively, and only very few of them were resistant or highly resistant (6–20% or 0–5% of infected leaf surface, respectively).

By selecting single plants belonging to accessions/families with a low disease severity (≤2), the overall resistance level of S. melongena germplasm shifted towards resistance, and highly susceptible phenotypes (disease severity = 5) did not occur in the populations S2 and S3.

Since eggplant is mainly self-pollinating, it would be realistic to assume that the eggplant collection consisted of a mixture of resistant and susceptible individuals with a high degree of homozygosity, and genes for resistance or susceptibility were probably fixed in the population. Therefore, a great effect of the selection in eliminating susceptible individuals would be expected in the first selection cycle but a little effect in the subsequent ones. Actually, because of environmental influence on the expression of resistance, or varying rates of outcrossing with susceptible genotypes, the complete elimination of highly susceptible individuals needed two selection cycles, while that of susceptible ones would need more than three.

Four promising resistant lines were obtained: PAVEG 10187 S3, PAVEG 10196 S3, P.I. 232079 S3 and P.I. 419198 S3. A significant improvement in powdery mildew resistance level was obtained only in S3 progenies of PAVEG 10187 and PAVEG 10196 (disease severity < 1), compared with the originating populations S2, S1 and S0. Such improvement was not observed for P.I. 232079 S3 and P.I. 419198 S3 (disease severity < 2), but further selections and selfings might increase the resistance level. Lack in increase of resistance level for P.I. 232079 S3 and P.I. 419198 S3 could be due either to field environmental effect on the resistance or to variable rate of out-crossing with susceptible plants. Eggplant is a self-pollinated species because of the cone-like shape of its anthers that envelops short- or medium-styled pistils. Nevertheless, the stigma may projects beyond the anthers (long-styled pistil), therefore cross-pollination may occur (Kowalska 2006) at a rate from 2 to 48%, depending on genotype, location and insect activity (Lal 1993).

Among the wild Solanum species tested, all of the tested plants of S. laciniatum and S. nigrum accessions were not infected by the pathogen in the S0 and S1 generations as well as the plants of the sole S0 accession of S. quinquangolare. This suggests that probably these species are non-host of L. taurica, however closer studies would need to ascertain this type of resistance. Actually, the ‘non-host resistance’ term is used for entire plant species completely resistant to entire pathogen species (Heath 2000).

S. linnaeanum, S. aculeatissimum, S. aviculare and S. pseudocapsicum were highly resistant and S. spinosissimum was resistant. Plants of S. gilo and S. capsicoides were susceptible or highly susceptible. S. sisymbriifolium showed a wide variability in the disease reaction ranging from 0 to 75% of leaf surface infected by powdery mildew in both S0 and S1 populations.

Up to present time, very few studies have been made on eggplant resistance to L. taurica. In Israel, variety trials did not provide adequate information (Palti 1971). In a field-screening in India, difference in susceptibility has been observed on 26 eggplant varieties: 11 and 7 varieties showed 6–10% and 11–20% of leaf infected by powdery mildew, respectively, while other eight varieties showed higher disease severity (Datar 1976). Cv Pusa Purple Long, reported as moderately resistant (6–10% of leaf infected) by Datar (1976), was susceptible (51–75% of leaf infected) in our tests. This discordance could be due very likely to different pathogen pressures in the field. Also, the existence of host-specific races of L. taurica could be probable, but studies have not agreed on this and further experimental evidences are needed (Palti 1971).

In the genus Capsicum, an oligogenic resistance has been found (Shifriss et al. 1992; Daubeze et al. 1995) and in Lycopersicon chilense the gene Lv has been identified (Stamova and Yordanov 1990).

In this research, the four S. melongena resistant lines and some wild Solanum species showed a low powdery mildew severity and thus considered valid sources of resistance to L. taurica. Such resistance could be transferred to susceptible cultivars either by classical breeding or by recombinant DNA techniques. Experiments of crossability of eggplant with wild related species have been attempted and, sometimes, fertile F1 progenies have been obtained. In many other cases, partially fertile, sterile or no F1 progenies were obtained. In order to overcome incompatibility barriers, somatic hybridization has been successfully used, while genetic transformation of eggplant and its wild relatives is still in its early stages (Collonnier et al. 2001).

Resistant lines of S. melongena and wild relative species could be also used for studying the genetic of resistance to L. taurica.