Development of Cantaloupe (Cucumis melo) Lines Carrying Vat Gene with Favorable Fruit Traits

The most popular Iranian cantaloupe ‘Samsoori’ is highly susceptible to devastating viruses transmitted by Aphis gossypii. A dominant gene (Vat) causing resistance to the aphid and viruses was detected in ‘Ginsen Makuwa’ in spite of its low fruit quality. They were crossed and the segregating offspring were assessed for combining favorable traits with Vat gene. In the F 2 population, moderate to high broad-sense heritability estimates were found for measured traits including, fruit weight (0.78) and soluble solid content (SSC) (0.7). The F 3 families were signicantly different from each other for earliness, fruit shape indices, cavity, esh thickness, SSC, and fruit numbers per plant. Resistant and susceptible plants were determined by genotyping 210 plants in F 4 generation using a dominant DNA marker for the resistant allele of Vat gene. Out of 15 selected F 3 families, four were susceptible, three were homozygote resistant and six showed segregation in their progeny for the Vat gene. Selection assisted by Vat gene marker was a very useful and applied approach for the identication of healthy plants along with phenotypic selection.


Introduction
Melon (Cucumis melo L.; Cucurbitaceae; 2n = 2x = 24), is a cross-pollinated horticultural crop with a wide diversity in fruit shape, avor traits, and climate adaptation. Iran is an important center of variation and also rank in the top ve production countries in the world for melons (FAO 2019). Many high-quality cantaloupe landraces or indigenous cultivars are cultivated in large acreages in Iran due to their unique avor and shape and early concentrated fruiting. Despite consumer preference for local cultivars, the majority of them are susceptible to biotic stresses such as fungal and viral diseases, which cause annual colossal yield loss. Thus, in recent years, a number of farmers prefer to pay for commercial hybrid seeds provided by seed companies to assure healthier plants with higher yields. Viral diseases are the most devastating diseases of melons, which cause a severe damage to the fruit and lower its yield and quality. The most prevalent viruses in melons are cucurbit aphid borne yellow virus (CABYV), cucumber mosaic virus (CMV), cucurbit yellow stunting disorder virus (CYSDV), zucchini yellow mosaic virus (ZYMV), papaya ringspot virus (PRSV), and watermelon mosaic virus (WMV). The most effective approach to control viral diseases is by developing resistant cultivars (Prohens-Tomás and Nuez 2008). There is a line of research describing the virus resistance in melon. A breeding line derived from PI 414723 with resistance to three potyviruses (WMV, ZYMV, PRSV) and powdery mildew. All four resistances displayed dominant monogenic inheritance and a genetic linkage was observed between resistance to WMV and ZYMV (Anagnostou et al. 2000). It was displayed that the resistance to CMV is oligogenic, where different loci confer resistance to different CMV strains, but not necessarily quantitative (Essa et al. 2009). Diaz et al. (2011) constructed an integrated genetic map associated with economically important traits in melon including four virus resistance genes.
Furthermore aphids are also important pests that harm crops either directly through the feeding of the phloem texture or indirectly through the transmission of viral diseases (Ng & Perry, 2004). Most plant viral infections can be partially controlled by elimination of their vectors; however, this is not a recommended method due to environmental concerns by pesticide application. Therefore, the use of resistant cultivars is the best approach for controlling aphids and white ies. The only aphid species that can colonize effectively on melon is Aphis gossypii which is the vector of CMV, ZYMV, PRSV, and WMV (Dogimont et al., 2014). By means of a RIL population developed from a cross between Vedrantais and PI161375, two minor QTLs were found for resistance to biotype B of Bemisia tabaci, and a major dominant gene (Vat) for resistance to Aphis gossypii (Boissot et al., 2010). It was shown that the virus aphid transmission (Vat) gene is located on linkage group 5, and confers double resistance to Aphis gossypii as well as the viruses transmitted by it. The amino acid sequences of this gene for resistance and susceptible alleles were well characterized, and the DNA sequence of the resistance allele was determined (Dogimont et al., 2014).
The availability of suitable molecular markers encourages plant breeders to use them for marker-assisted selection and speed up the breeding process (Sousaraei et al., 2018). This method can overcome the limitations associated with phenotypic selection (Foolad & Panthee, 2012), especially for virus resistance that phenotypic selection is often not straight forward or costly; the use of molecular markers can be a very bene cial tool.
Cultivar 'Samsoori' is the main commercial cantaloupe cultivar in Iran. It is well-known for its early concentrated fruiting, striped and ne netted rind, and desirable juicy avor. However, this cultivar is highly susceptible to viruses and its sugar content is not satisfactory. In a breeding program, 'Samsoori' for improvement was crossed with the Korean melon 'Ginsen Makuwa' which shows high resistance to viruses and has high sugar content. The main objective of this study was to identify the offspring in a F 3 generation of a cross between 'Samsoori' and 'Ginsen Makuwa' carrying Vat resistant allele (Boissot et al., 2016) with desirable traits such as high soluble solid content (SSC), proper shape and fruit avor, early harvest, and high yield. However, initially it was necessary to clarify the e ciency of the molecular marker to distinguish the resistant and susceptible allele of Vat gene in the segregating population.

Plant material
The melon genotypes used in this study were 'Samsoori', 'Ginsen Makuwa', and the derivative progenies of their cross.
The fruit of 'Samsoori' is round to oblate with leathery rind and a clear vein tract and thorough netting. It has full slip from its vine once ripe, and its esh is juicy green. In contrast, the fruit of 'Ginsen Makuwa' is oval shape with a very thin rind and a ridged surface, no netting and yellowish-white color. The esh of fruit is crunchy texture and white color (Fig. 1).
For crossing Samsoori was used as the pistillate and Ginsen Makuwa as the staminate parent, here thereafter referred to as P 1 and P 2 , respectively. The initial cross was made in the spring of 2017, at the greenhouse of Auraihan campus, Pakdasht, Iran. Ten seeds were germinated and grew to maturity and self-pollinated to create the F 2 population (n = 350).
In the spring of 2018, each parent's seed, F 1 s, and F 2 s were sown in potting trays. Within a month after sowing the seed, the plants were hardened off and transplanted to the eld, in a randomized complete block design experiment in three replications. Each block consisted of one row (10 seedlings) of P1, P2, and F 1 and ve rows of F 2 generation. During the owering time, each plant in the F 2 generation was self-pollinated. Thus, on the day of anthesis, male owers of the same plant were used for pollinating androgynous owers, which both were covered in the previous evening.
Phenotypic Evaluation of P1, P2, F1 and F2 generations At the end of the growth season, matured fruits were harvested. Earliness of fruits was recorded as the number of days from the date of transplanting of seedlings to the eld until the harvest date. Fruit yield in each plant was calculated with summing up the weight of all fruits from each plant. Other traits such as fruit length, width, esh thickness, and seed cavity were measured for each fruit after longitudinal cutting. Soluble solid content (SSC) was measured using a handheld Refractometer (Master-20PM, Atago, Japan) from fruit juice extracted by squeezing a piece from the middle of each fruit.

Phenotypic selection of desirable plants in F2
During the growth season, the healthy plants of F 2 population were labeled. Disease free plants with early mature fruit, high SCC (brix index), acceptable fruit shape (round shape with thorough netting and clear vein tracts), desirable esh features (green, juicy and thick) were selected within the F 2 population. Out of 150 plants in the F 2 population, 20 were selected. Since each plant had at least one fruit developed from self-pollinattion, they produced 20 different F 3 families.

Phenotypic evaluation of F3 families
In spring 2019 the seedling of 20 F 3 families as well as parental genotypes (in total 22 treatments) were evaluated for different traits in the experimental station of Aburaihan campus, Pakdasht, Iran. The seedlings were transplanted in randomized complete block design with three replications. Each plot consisted of a row with ten plants. Important traits were measured the same way as the previous generation, and during owering time, all F 3 plants were self-pollinated to produce F 4 generation. Based on the collected data, a total of top rated 70 plants were selected within and between rows of F 3 families. Subsequently, three seeds of each selected F 4 plant were sown in the greenhouse to maturity, and their leaf tissue was used for DNA extraction and genotyping.
Parental and Cultivar Survey for the Vat gene It was necessary to verify the reliability and e ciency of the molecular marker before applying it for selection of resistant plants in F 4 families. Thus, six Iranian native cultivars and 11 other genotypes (Table 1), were clari ed for the presence or absence of the Vat gene. Then F 4 families were tested by extracting DNA from all three plants of 70 families beside 'Ginsen Makuwa', 'Samsoori' and F1 generation as genetic controls.

Data analysis
The analysis of variance of phenotypic data was carried out using SAS software version 9.0 (Cary, NC). Outlier data points were detected by visually inspecting the scattered plots or using any data point with larger two standard deviations from the mean, and were removed from the dataset for further analyses. Comparison of the means of different treatments was carried out using the Duncan's Multiple Range test.
Where H b is broad-sense heritability, V F2 is the phenotypic variance of F 2 population. Vg was calculated with the following formula: (2)

Results And Discussion
Phenotypic evaluation of P1, P2, F1, and F2 Analysis of variance among different generations showed signi cant differences for most traits. Table 2 shows the mean comparison of different generations for measured traits. There are signi cant differences for all traits except yield that was not signi cantly different among different generations. Substantial diversity for different traits was expected as the parents belonged to different groups; 'Samsoori' is t to cantalupensis while 'Ginsen Makuwa' is an oriental melon belonging to conomon cultivar group (Pitrat, 2017). The phenotypic characteristics of F 1 progeny were intermediated ( Fig. 1) and in F 2 progeny were segregating between parents. Moderate to high values for broad-sense heritability for most traits show that this population has the potential for improvement via genetic selection. For most traits in the F 2 population, transgressive segregation was observed, as the range of data in F 2 population is broader than the difference between parents. This result may be explained by the fact that parental genotypes are genetically distantly related from each other.
As an index for sweetness between two parents, a signi cant difference was observed for soluble solid content (SCC%). A wide range in F 2 population for this trait provides an opportunity for the selection of genotypes with higher sugar content.
Low sugar content in Iranian landraces was reported before (Pouyesh et al., 2017); therefore 'Ginsen Makuwa' with high sugar content in addition to its virus resistance appears to be a right complementary parent for improvement of 'Samsoori'.
Although the disease severity index could not be measured in the eld accurately, but the difference of the parents for the level of incidence and segregating of the trait among offsprings was quite obvious.

Evaluation of traits in F3 generations
Analysis of variance (ANOVA) revealed highly signi cant differences among treatments for all measured traits (Table 3). In order to select the best families, means of different traits was compared among 22 treatments. Table 4 illustrates the characteristics of different F 3 families, along with parental genotypes. The number of fruits per plant in 'Samsoori' was on average 2.5 while for 'Ginsen Makuwa' it was 8.1. Fruit weight of 'Samsoori' was higher than that of 'Ginsen Makuwa'.
In F 3 families, a wide range was observed for these two traits. Also, there are two families that are signi cantly had higher SSC than that of 'Samsoori' which means selection for this trait has been led an increase in the sweetness of progenies. The response to selection was more evident for fruit shape (round or oval, netting and vein tract), esh color, and texture. In Table 5 the genotypes were categorized based on fruit-related traits. The esh color, vein tract, netting, and esh texture are differentiated by nominal scale, and the shape index is the ratio of fruit length to width. Considerably the frequency of desirable plant type (round fruit with netting and vein tract, juicy green esh) in most families is higher than unfavorable ones. The length to width ratio for all families is near 1 in contrast to this ratio for 'Ginsen Makuwa', which is 1.37. These ndings con rm the idea that netting skin, vein tract, esh color, and fruit shape are controlled with oligo-genes with high heritability (Dogimont, 2011).  Means with different superscript letters are signi cantly different from each other based on Duncan multiple range test at p ≤ 0.01 Table 5 Classi cation of F3 families considering fruit related traits. Data for esh color, vein tract, netting and esh texture are the frequencies of their nominal scale (percentage of all fruits measured in each treatment). Also, the shape index is the ratio of fruit length to width calculated based on data from

Screening of different genotypes
The speci c primer pair developed by Dogimont et al. (2014) for the resistant allele of Vat produced 121bp fragment. Figure 2 depicts the ampli cation curve of this primer pair with Real time PCR for different genotypes. HRM analysis showed the unique DNA fragment had been ampli ed only in 'Ginsen Makuwa' and "Karno Kiku Makuwa" among 17 tested genotypes (see material and methods). None of other genotypes include Iranian cultivars showed ampli cation curve and hence they had no resistant allele of Vat. These results is in congruity with the ndings of a previous study in which 'Ginsen Makuwa' displayed double resistance toward A. gossypii and CMV (Boissot et al., 2016).
Marker assisted selection of resistant plants in F 4 populations Figure 3 shows the melting peak of ampli ed fragment in control parental genotypes in which the melting peak of resistant allele in 'Ginsen Makuwa' and F 1 generation are clearly different from primer dimer peaks in 'Samsoori'. This experiment was carried out several times to screen the F 4 population. After ampli cation, the melting curve was obtained for all samples in a plate (Fig. 4). For some plants there were only the resistant allele of Vat while for others no ampli cation was observed or a minor primer-dimer peak was observed. Figure 5 shows the electrophoresis of PCR product of the primer for control parental genotypes and some F4 generation plants.
In total 178 plants were genotyped in F 4 generation.

Conclusion
The current research made a major contribution to marker assisted selection for aphid and virus resistance in melon. The marker related to Vat gene was easy, fast and reliable enough to be substituted phenotypic selection. Also, it costs less because evaluating of virus resistance in plants via arti cial inoculation is complicated and time-consuming, though by marker selection large number of plantlets could be screened in a short time.
The study also was an advancement in breeding of Iranian cantaloupe landraces with valued characters for virus resistance and improving sweetness. Further characterization of F 5 families (lines) is necessary for developing and selection of best lines which can be introduced as new cultivars or be included in commercial hybrid production.

Declarations
Funding - The project was funded by Iran National Science Foundation (INSF grant number: 93/S/36998)

Con ict of interest -
The authors declared that they have no con icts of interest that could have appeared to in uence the work reported in this paper.  Melting curves of ampli ed PCR products with the primer pair speci c to resistant allele of Vat in control parental genotypes of 'Ginsen Makuwa' (blue color), 'Samsoori' (red color) and F1 (green color). The electrophoresis of PCR product showed that the peak related to 'Samsoori' is primer dimers.

Figure 4
Melting curve of ampli ed PCR products with the primer pair speci c to resistant allele of Vat in F4 generation plants. In all ampli cation experiments 'Samsoori', 'Ginsen Makuwa' and F1 were included. In all experiments, in each plate (96 well), three wells were assigned for non-DNA template controls.