Genetic characterization of citron watermelon (citrullus lanatus var. citroides [L.H. Bailey] mansf. ex greb.) and development of experimental hybrids.

Abstract

Citron watermelon (Citrullus lanatus var. citroides [L.H. Bailey] Mansf. ex Greb.) is indigenous to sub-Saharan Africa (SSA) with multiple uses, including human food and animal feed. Its succulent leaves are used as leafy vegetables, while the ripened yellow and orange-fleshed fruits are used to prepare various traditional dishes, and the seeds are roasted and consumed as snack. It is an emerging potential rootstock for producing grafted sweet watermelon (Citrulus lanatus var. lanatus) to improve fruit yield and biotic and abiotic stress tolerance. It is also a source of novel genes for breeding in sweet watermelon to improve fruit yield, quality and disease resistance. Citron watermelon in SSA is mainly cultivated using unimproved landrace varieties. Landraces exhibit marked phenotypic variation for fruit shape, size, skin colour patterns, and seed coat colours. Phenotypic and genetic variation among South African citron watermelon landraces is yet to be systematically assessed for diverse use and cultivar design. The overall goal of this study was to initiate a pre-breeding program for citron watermelon through identification and selection of unique and complementary genotypes for production, value-adding and breeding. The specific objectives of this study were: i. To determine the extent of genetic diversity among South African citron watermelon landrace accessions using selected simple sequence repeat (SSR) markers to identify genetically divergent accessions for trait integration and variety development; ii. To assess the phenotypic diversity of citron watermelon landrace accessions of South Africa and to select desirable genotypes with suitable agronomic and horticultural traits for direct production, breeding and conservation; iii. To estimate variance components, heritability and genetic advance of phenotypic traits in citron watermelon to guide the selection of superior genotypes for direct production and breeding; iv. To determine the combining ability and hybrid performance of citron watermelon genotypes for agronomic traits for breeding. In the first study, 48 citron watermelon landrace collections widely grown in the Limpopo Province of South Africa were genotyped using 11 selected SSR markers.

The SSR markers amplified a total of 24 alleles, with a mean expected heterozygosity value of 0.38, indicating moderate genetic diversity among the studied accessions. Analysis of molecular variance attributed 8%, 75%, and 17% of the molecular variation between populations, among accessions and within accessions, respectively. Three distinctive genetic groups were identified based on cluster analysis. The following distantly related genotypes are recommended as breeding parents namely: WWM03, WWM04, WWM15, WWM16, WWM18, WWM22, WWM23, WWM24, WWM25, WWM26, WWM28, WWM33, WWM34, WWM35, WWM38, WWM39, WWM41, WWM66, WWM76, WWM78, WWM81, WWM84, WWM86 and WWM89 (selections from Cluster I), WWM14, WWM37, WWM42, WWM44, WWM46, WWM50, WWM65, WWM79, WWM85 and WWM87 (Cluster II), and WWM38, WWM47 and WWM48 (Cluster III). These are useful parental lines for pre-breeding to develop and release new varieties with multiple uses. In the second study, 36 selected citron watermelon landrace accessions were evaluated under field conditions across two environments using a 6 × 6 lattice design with three replicates. Data on key qualitative and quantitative traits were collected and subjected to non-parametric and parametric statistical analyses. The accessions showed wide phenotypic variation and unique traits for genetic improvement. Positive and significant correlations (p < 0.001) were recorded between total fruit yield per plant with plant height (r = 0.64), number of harvestable fruits (r = 0.70), number of marketable fruits (r = 0.73) and marketable fruit yield (r = 0.96). Seed yield per plant positively and significantly (p < 0.001) correlated with number of male flowers (r =0.68), plant height (r = 0.61) and total fruit yield (r = 0.79). Principal component analysis identified nine components which accounted for 86.38% of total variation amongst accessions for assessed phenotypic traits. The study recommended citron watermelon accessions such as WWM14, WWM16, WWM39, WWM41, WWM67 and WWM79 for use as leafy vegetables owing to their profuse branching ability and longer vine production. Whereas accessions including WWM03, WWM17, WWM35, WWM40, WWM50, WWM67, WWM79 and WWM85 are selected with larger fruit size. Accessions WWM05 and WWM09 are sour-flesh types which are suitable genetic stocks for breeding sweet-and-sour and sweet dessert watermelons. Orange-fleshed accessions such as WWM03, WWM04, WWM46, WWM64, WWM66 and WWM67 are recommended for fresh consumption, cooking, processing or variety design.

Accessions WWM02, WWM03, WWM08, WWM14, WWM16, WWM23, WWM38, WWM40, WWM41 and WWM67 have red and white seed coat colour which are superior selections to prepare roasted citron watermelon seed snack. In the third study, variance components, heritability and genetic gains of phenotypic traits were estimated involving 36 accessions of citron watermelon grown under field conditions across two test environments using a 6 × 6 lattice design with three replicates. High broad-sense heritability and genetic advance as percent of the mean were recorded for fruit length at 83.86 and 4730.45%, seed length (77.73 and 1731.27%), hundred seed weight (73.73 and 4027.36%), fruit diameter (70.44 and 2949.64%) and fruit weight (70.39 and 8490.05%), respectively. Step-wise regression analysis revealed marketable fruit yield and total number of fruits per plant explaining 89% (R2 = 0.89) of total variation for total fruit yield per plant, whereas number of seed per fruit and hundred seed weight explained 92 (R2 = 0.92) of total variation for seed yield per fruit. Citron watermelon landrace accessions WWM03, WWM14, WWM16, WWM39, WWM65, WWM67 and WWM79 with high total fruit yield and seed yield per fruit were selected for production or breeding programme. In the fourth study, five selected parental genotypes were crossed in a 5 × 5 half-diallel mating design to develop 10 hybrids. The 15 families (five parents and 10 F1 hybrids) were evaluated across two environments using a randomized complete block design (RCBD) with three replications. General combining ability (GCA) and specific combining ability (SCA) effects were significant (p < 0.001) for most traits. Environment × GCA was non-significant, whereas Environment × SCA effects were significant (p < 0.001) for most traits. The ratios of GCA/SCA variances were less than unity for most traits, indicating non-additive gene action of the traits. Broad-sense heritability varied from low to moderate, implying variable selection response of the assessed traits among the F1 hybrids. The parental genotypes WWM16 with positive GCA effects for fruit and seed yield and WWM66, with positive GCA effects for the number of seeds per fruit and seed yield, were identified for hybrid breeding. The following F1 hybrids, namely: WWM04 × WWM16, WWM03 × WWM66 and WWM16 × WWM50 with positive SCA effects on total fruit yield per plant and marketable fruit yield per plant, and WWM04 × WWM50, WWM03 × WWM16 and WWM03 × WWM66 with positive SCA effects for number of seeds per fruit and total seed yield were identified. The study identified novel and best-performing F1 hybrids of citron watermelon for economic traits and are recommended for multi-environmental evaluations, variety registration and commercialization. Overall, the study revealed genetic and phenotypic variation in citron watermelon to select and recommend suitable genotypes for production and for breeding new generation varieties based on market needs and consumer preferences. The study recommends accessions such as WWM14, WWM16, WWM39, WWM64, WWM67, WWM76 and WWM79 with high fruit yield, and WWM03, WWM04, WWM14, WWM15, WWM16, WWM24, WWM28, WWM37, WWM46, WWM66 and WWM68 exhibiting high fruit and seed yield for breeding or direct production. The parents WWM04, WWM03 and WWM16 were identified as good combiners for fruit or seed yield and related-component traits for future breeding. The F1 hybrids derived from these parents, including WWM04 × WWM16, WWM03 × WWM16, WWM03 × WWM66, WWM16 × WWM50, and WWM04 × WWM50 were best performing for economic traits and new breeding population development.