Inheritance of Resistance to Powdery Mildew Race 2 in Citrullus lanatus var. lanatus

Information on the mode of inheritance of powdery mildew resistance in watermelon is important for designing a breeding strategy for the development of new cultivars. Resistance in the watermelon accession PI 270545 was investigated by generation means analysis by crossing it with susceptible PI 267677. The analyses showed involvement of two genes, a recessive resistance gene, pmr-1, and a dominant gene for moderate resistance, Pmr-2. Resistance to powdery mildew in the leaf had a large dominance effect and a heritability of 71%. The additive-dominance model was inadequate in explaining variation in leaf resistance as revealed by the joint scaling test. However, nonallelic interactions could not be detected by the nonweighted six-parameter scaling test. For stem resistance, the additive-dominance model was adequate, and inheritance was controlled mainly by additive effects. A high narrow-sense heritability of 79% suggested that selection for stem resistance in early generations would be effective. Powdery mildew in watermelon [Citrullus lanatus (Thumb.) Matsum. and Nakai] caused by the fungus Podosphaera xanthii race 2W has in recent years become a concern among growers as well as plant breeders in United States (Davis et al., 2001; McGrath, 2001), China (Feng, 1996; Zhang et al., 2011), and other parts of the world (McGrath, 2001; Tomason and Gibson, 2006). The disease causes significant yield loss as well as decreased fruit quality (McGrath and Thomas, 1996) through mycelial coverage of the leaves, leaf necrosis, and premature death of the plant (Davis et al., 2001). Powdery mildew of watermelon occurs throughout the southeastern United States, extending north to New York as well as into western states (Davis et al., 2005). Development of genetic resistance is an important objective in watermelon breeding programs. Screening of the U.S. watermelon germplasm collection identified high resistance in several wild accessions of C. lanatus var. citroides (Davis et al., 2007; Tetteh et al., 2010), but none of the accessions in the primary gene pool. However, one accession of C. lanatus var. lanatus, PI 270545, originating from Sudan was found to have intermediate resistance (Tetteh et al., 2010). Resistance in this accession was characterized by few mycelium on leaf and stem, and affected plants survived to fruit production stage (Tetteh et al., 2010). To help breed cultivars resistant to powdery mildew, it is important to understand the inheritance and gene action. Tetteh et al. (2013) evaluated the gene action of leaf and stem resistance in the watermelon accession, PI 189225, and established that mainly additive gene action controlled leaf resistance, whereas for stem resistance, additive, dominance, and epistasic gene actions were significant. There are several reports on inheritance of resistance to P. xanthii in melon (Cucumis melo L.). Most agree that there are several genes controlling resistance (Epinat et al., 1993; Kenigsbuch and Cohen, 1992; McCreight, 2003; McCreight et al., 1987; Pitrat et al., 1998). Perchepied et al. (2005) working with quantitative trait loci (QTL) revealed that powdery mildew resistance in melon was under the control of major gene effects and digenic epistasis. Inheritance of powdery mildew resistance in watermelon was investigated by generation means analysis by Tetteh et al. (2013). Studies have demonstrated a close correspondence between generation mean analysis and QTL mapping (Jung et al., 1994; Perchepied et al., 2005). Estimation of genetic effects in different crosses should inform the breeding strategy for development of resistant cultivars. Although additive and dominance models can be determined with the scaling tests in generation means analysis, the identification of nonallelic interactions requires more powerful tests such as the joint scaling test (Mather, 1949) or QTL analysis (Perchepied et al., 2005). In most cases, the variation unaccounted for by a major gene is provided by digenic epistasis. A major deficiency of generation means is in nondetection of additive effects resulting from dispersion of alleles with similar effects between parents and internal cancellation of dominance effects exhibited in opposite directions at different loci (Crow and Kimura, 1970). Saudhu and Nittal (1988) studied two Gossypium arboreum crosses and reported the absence of nonallelic interaction in the six-parameter model, whereas the joint scaling test predicted the presence of epistasis for yield of seed cotton per plant (Iqbal and Nadeem, 2003). Given the lack of genetic information on powdery mildew resistance in the primary gene pool of watermelon, a generation means analysis was carried out to determine inheritance, gene action, and heritability of resistance in PI 270545. Information on these parameters would be useful for designing an efficient breeding strategy for watermelon powdery mildew resistance. Materials and Methods Plant material. A single population of watermelon segregating for resistance to powdery mildew race 2W-U.S. was derived from a cross between the susceptible P1 (PI 269677) and resistant P2 (PI 270545). The parents were inbred for two generations before crossing to produce plants uniform for powdery mildew resistance. From this, crosses were made to create a total of six generations, F1, F2, BC1P1 (the first backcross to P1), and BC1P2 (the first backcross to P2) for a study of inheritance of resistance. Experimental design. Seeds of inbred powdery mildew-resistant PI 270545 and -susceptible PI 269677, together with their F1, reciprocal F1, F2 (generated by selfpollination of the F1), and backcross generations were produced in 2007 to 2008 in greenhouses at the Department of Horticultural Science, North Carolina State University, Raleigh, NC. Seeds were planted in two sets, each consisting of 10 plants of each parent, 10 plants of F1, 10 plants of F1 reciprocal, 100 plants of F2, and 30 plants of each BC1. Pooled over sets, a total of 20 plants for each parent, 17 plants of F1, 19 of F1 reciprocal, 190 plants of the F2, and 59 plants each of BC1P1 and BC1P2 were evaluated for powdery Received for publication 26 June 2013. Accepted for publication 1 Aug. 2013. This work was supported in part by the graduate school assistantship provided by North Carolina State University, the North Carolina Agricultural Research Service, the Ghana Government Scholarship Fund, and the American Association of University Women. We thank Tammy L. Ellington and Allen Gordon for technical assistance. Previously with U.S. Department of Agriculture, Agriculture Research Service, South Central Agriculture Research Laboratory, Lane, OK 74555. To whom reprint requests should be addressed; e-mail [email protected]. HORTSCIENCE VOL. 48(10) OCTOBER 2013 1227 | BREEDING, CULTIVARS, ROOTSTOCKS, AND GERMPLASM RESOURCES