Introduction
The melon (Cucumis melo L.) inodorus landraces are commonly known as ‘winter melons’, as they can survive through the winter. They are traditionally cultivated in the Mediterranean area, and are an important horticultural crop in Italy, with particular relevance for the economy of the southern regions of Italy (Ficcadenti et al., Reference Ficcadenti, Sestili, Luongo, Campanelli, Rosa, Ribeca, Ferrari, Maestrelli, Genna and Belisario2007).
A difficulty in the cultivation of these melon genotypes is that they are often identically named in the same cultivation areas, which results in homonyms and synonyms that produce confusion for the recognition of the specific populations (Sestili et al., Reference Sestili, Daniele, Rosa, Ferrari, Belisario and Ficcadenti2008). Therefore, it is imperative to preserve the genetic variability of these local varieties as part of the conservation of the genetic resources of Italian melons, and for their further use in breeding programme (Ficcadenti et al., Reference Ficcadenti, Sestili, Luongo, Campanelli, Rosa, Ribeca, Ferrari, Maestrelli, Genna and Belisario2007, Reference Ficcadenti, Giardini, Sestili, Lo Scalzo, Palermo, Tumbarello, Monteleone, Bono and Bongiovì2010).
Both molecular markers and agronomic traits are commonly used in studies of the genetic relationships among melon genotypes (Garcia et al., Reference Garcia, Jamilena, Alvarez, Arnedo, Oliver and Lozano1998; Staub et al., Reference Staub, Lopez-Sesè and Fanourakis2004; Dhillon et al., Reference Dhillon, Ranjana, Singh, Eduardo, Monforte, Pitrat, Dhillon and Singh2007). Inter-simple-sequence repeat (ISSR) markers represent a successful tool for the assessment of genetic diversity in melon because these are known to be highly informative and can be used to detect polymorphism in genotypes with both wide and narrow genetic backgrounds (Danin-Poleg et al., Reference Danin-Poleg, Tzuri, Reis and Katzir1998; Stepanski et al., Reference Stepanski, Kovalski and Perl-Treves1999; Sestili et al., Reference Sestili, Minollini, Luciani, Campanelli and Ficcadenti2004, Reference Sestili, Daniele, Rosa, Ferrari, Belisario and Ficcadenti2008).
In the present study, we estimated the genetic diversity among Italian inodorus melon populations using ISSR markers and phenotypic traits.
Materials and methods
Thirteen melon inodorus landraces were used. The dihaploid line Nad-1 and the cultivar Charentais-T were used as reference accessions in the molecular analysis. All of the plants were grown in an open field according to a randomized block design, with two replicates and six plants/replicate.
The phenotypic characterization was carried out on three fruits randomly harvested from each population. Fifteen morphological-agronomic traits were recorded: fruit weight, fruits/plant, total yield, fruit length, fruit width, fruit shape rind thickness, flesh thickness, placenta length, placenta width, skin colour, skin texture, flesh colour, taste and Brix.
The morphological data were subjected to two-way ANOVA (Statistica 7 software), and the mean comparisons were calculated by Duncan's test (P = 0.05).
Genomic DNA was extracted from young melon leaves according to Levi and Thomas (Reference Levi and Thomas1999). One-hundred ISSR primers were used (set no. 9; UBC, Vancouver, Canada). The PCR analysis was performed according to Gupta et al. (Reference Gupta, Chyi, Romero-Severson and Owen1994). PCR products were separated through 1.2% agarose gels and photographed by the Kodak 1D 3.6 documentation system. A 100-bp ladder was used for the molecular weight standards.
All of the statistical analyses were performed using the NTSYS-pc software, version 2.20 (Rohlf, Reference Rohlf2006). The morphological data were first standardized according to Lopez-Sesè et al. (Reference Lopez-Sesé, Staub and Gomez-Guillamon2003), and the resulting binary data matrix was used to calculate the Euclidean distances. Each polymorphic ISSR band was scored as either present (1) or absent (0) for all of the genotypes, and the binary matrix obtained was used to calculate the genetic similarity coefficient for each pair of accessions (Nei and Li, Reference Nei and Li1979). The robustness of the nodes in the molecular tree was tested by bootstrap analysis with 1000 replicates, using the TREECON software (Van de Peer, Reference Van de Peer1994). The levels of correlation between the ISSR and agronomic distance matrices were determined using the Mantel test with 1000 permutations (Mantel, Reference Mantel1967).
Results and discussion
Similarities among the inodorus accessions were established. The binary data matrix that was obtained from the phenotypic data was used to generate the dendrogram (UPGMA), which grouped the genotypes that share similar agronomic features. The cluster analysis revealed two main clades, in which the genotypes are grouped on the basis of the skin colour of their fruits (Fig. 1). In particular, the first clade is distinguished by two groups: one with all of the green genotypes and the other represented by only two accessions with a yellow skin colour. The second main clade grouped all of the remaining yellow genotypes. The Verde Quasi Rotondo accession was included in the green cluster although the skin colour of its fruit is green during growing period and yellow at physiological maturity. Alsia was an unclustered accession, as expected, since the fruit shape is very similar to the watermelon morphology. The cophenetic correlation coefficient of this cluster analysis was r = 0.87 (P = 1).
Fig. 1 UPGMA cluster analysis on the basis of the phenotypic data.
A total of 358 polymorphic bands obtained from 39 of the 100 ISSR primers were used for the genetic-similarity calculations. The cluster analysis (UPGMA) grouped the inodorus melon accessions into two separate main groups on the basis of the skin colour of the fruit, in agreement with the morphological analysis, and with the exception of the yellow genotype Cartucciaru (Fig. 2). Furthermore, the Verde Quasi Rotondo was a distinct group, in contrast to what was observed with the morphological data analysis. Alsia and Alsia Fascista were grouped together in a separate group and were highly similar (bootstrap 100). The dihaploid line Nad-1 and the genotype Charentais-T were an outgroup, as expected (Fig. 2). The cophenetic correlation computed for the molecular cluster analysis was r = 0.91, with P = 1. The distance matrices obtained from the phenotypic and molecular data were further compared with Mantel's test, which revealed a good correlation (r = 0.50; P = 0.99).
Fig. 2 UPGMA cluster analysis on the basis of the molecular data.
The results obtained confirm the efficacy of this approach and open new perspectives to reveal the possible molecular associations with the phenotypic traits analysed. Such analyses of plant diversity using precise morphological and molecular evaluations of regional collections are useful for germplasm curators and plant geneticists, as they help to define accessions according to geographical regions, and they provide solid historical reference data for future genetic studies that are aimed at assessing genetic erosion, exploring genetic potential and site conservation priorities.
This study has confirmed the usefulness of these ISSR molecular markers to distinguish among genotypes that are characterized by a narrow genetic background, and it opens new perspectives towards conservation of the Italian melon genetic resources and their further use in breeding programmes.
