plants Article Fruit Morphology and Ripening-Related QTLs in a Newly Developed Introgression Line Collection of the Elite Varieties ‘Védrantais’ and ‘Piel de Sapo’ Miguel Santo Domingo 1,† , Carlos Mayobre 1,† , Lara Pereira 1 , Jason Argyris 1,2 , Laura Valverde 1 , Ana Montserrat Martín-Hernández 1,2 , Jordi Garcia-Mas 1,2, * and Marta Pujol 1,2, * 1 Centre for Research in Agricultural Genomics (CRAG), CSIC-IRTA-UAB-UB, Edifici CRAG, Campus UAB, 08193 Barcelona, Spain 2 Institut de Recerca i Tecnologia Agroalimentàries (IRTA), 08193 Barcelona, Spain * Correspondence: jordi.garcia@irta.cat (J.G.-M.); marta.pujol@irta.cat (M.P.); Tel.: +34-935-636-600 (J.G.-M. & M.P.) † These authors contributed equally to this work. Abstract: Melon is an economically important crop with widely diverse fruit morphology and ripening characteristics. Its diploid sequenced genome and multiple genomic tools make this species suitable to study the genetic architecture of fruit traits. With the development of this introgression line population of the elite varieties ‘Piel de Sapo’ and ‘Védrantais’, we present a powerful tool to study fruit morphology and ripening traits that can also facilitate characterization or pyramidation of QTLs in inodorous melon types. The population consists of 36 lines covering almost 98% of the melon genome, with an average of three introgressions per chromosome and segregating for multiple fruit Citation: Santo Domingo, M.; traits: morphology, ripening and quality. High variability in fruit morphology was found within the Mayobre, C.; Pereira, L.; Argyris, J.; population, with 24 QTLs affecting six different traits, confirming previously reported QTLs and two Valverde, L.; Martín-Hernández, newly detected QTLs, FLQW5.1 and FWQW7.1. We detected 20 QTLs affecting fruit ripening traits, A.M.; Garcia-Mas, J.; Pujol, M. Fruit six of them reported for the first time, two affecting the timing of yellowing of the rind (EYELLQW1.1 Morphology and Ripening-Related QTLs in a Newly Developed and EYELLQW8.1) and four at the end of chromosome 8 affecting aroma, abscission and harvest date Introgression Line Collection of the (EAROQW8.3, EALFQW8.3, ABSQW8.3 and HARQW8.3). We also confirmed the location of several Elite Varieties ‘Védrantais’ and ‘Piel QTLs, such as fruit-quality-related QTLs affecting rind and flesh appearance and flesh firmness. de Sapo’. Plants 2022, 11, 3120. https://doi.org/10.3390/ Keywords: introgression lines; Cucumis melo; fruit quality; climacteric ripening plants11223120 Academic Editor: Sujan Mamidi Received: 16 September 2022 1. Introduction Accepted: 10 November 2022 Cultivated fruits and vegetables are an essential part of the human and animal diet, Published: 15 November 2022 so it is crucial to improve the yield as well as the quality of their edible parts. To help Publisher’s Note: MDPI stays neutral breeders, molecular tools have been developed in many crop species to map and associate with regard to jurisdictional claims in a trait with a genomic region. With these tools, molecular plant breeding has emerged as a published maps and institutional affil- relevant strategy to accelerate plant breeding, and the use of molecular markers has led to iations. the discovery of molecular mechanisms of diverse traits [1–3]. Several mapping populations have been used in modern plant breeding, such as recombinant inbred lines (RILs) [4–6] and F2 populations [7,8]. RILs have been widely used for mapping QTLs associated with traits such as yield in wheat [9] and leaf type in Copyright: © 2022 by the authors. soybean [10], while F2 populations have also been used in the study of yield in rice [11] Licensee MDPI, Basel, Switzerland. and fruit quality traits in pomegranate [12]. In RILs and F2 populations, the genome of This article is an open access article the two founding parents is equally represented, but, in other mapping populations, the distributed under the terms and genomes of the parents are not equally represented, such as introgression lines (ILs) [13,14] conditions of the Creative Commons Attribution (CC BY) license (https:// and chromosome segment substitution lines (CSSLs) [15,16]. creativecommons.org/licenses/by/ IL populations make it possible to mendelize and study different traits without the 4.0/). effect of other interacting loci. This has been used widely for different traits and crops, such Plants 2022, 11, 3120. https://doi.org/10.3390/plants11223120 https://www.mdpi.com/journal/plants Plants 2022, 11, 3120 2 of 18 as fruit weight and soluble solid content in tomato [13], flowering time in maize [17], grain weight in rice [18] or CMV resistance in melon [19,20]. In addition, it allows pyramiding of different QTLs and studying their interactions, as has been conducted for studying resistance to bacterial blight in rice [21] or resistance to Aphanomyces euteiches in pea [22]. ILs are developed by multiple backcrosses with the same parental line (the recurrent parent) and a last step of self-pollination to fix the heterozygous regions, providing a set of perpetual lines sharing most of the genome of the recurrent parent but having an introgression in a given chromosome from the other parental line (donor parent). Ideally, with a complete set of ILs, the donor genome should be fully covered. In melon, many mapping populations have been developed to study fruit qual- ity [23–27] and fruit morphology [27–29]. For melon breeding, fruit ripening represents a key complex trait, and several studies have been conducted to identify the genes underlying this process [27,30,31]. Melon presents intra-specific segregation of climacteric ripening, al- lowing the development of populations segregating for ripening behavior [27,30]. Recently, three genes involved in melon ripening (CmCTR1, CmROS1 and CmNAC-NOR) have been validated through CRISPR/Cas9 [32,33], and a collection of lines pyramided with three ripening-related QTLs introgressed in a non-climacteric background has been developed to understand the interaction of these QTLs [34]. Here, we developed a novel IL population with ‘Piel de Sapo’ (PS) as the recurrent parent and ‘Védrantais’ (Ved) as the donor parent. These two elite cultivars were chosen due to their differences in fruit morphology and ripening behavior. This is a new high-value genomic resource for breeding inodorous varieties and offers some insights into the genomic architecture of fruit quality, fruit morphology and fruit ripening behavior in melon. 2. Results and Discussion 2.1. Phenotyping of the Parental Lines PS and Ved were used as parental lines for the population, with PS the recurrent and Ved the donor parent. Both varieties are considered elite cultivars as they are widely consumed for their taste and aroma but with distinctive characteristics. Over the two-year period of analysis, 2020 and 2021, PS produced consistently larger fruits than Ved, even though the PS fruits were smaller in the 2021 season (FW; Figure 1A; Table S1). The soluble solid content was highly dependent on the environmental conditions. In 2020, there were no significant differences between the cultivars, while PS fruits accumu- lated more soluble solids than Ved in 2021 (SSC; Figure 1B; Table S1). This SSC variability has been observed in previous studies: the mean value in PS is generally around 10–11◦ Brix, but, in some environments, it can reach values lower than 9 or higher than 13 [26,27,35,36]. In Ved, the variation has also been found to be highly dependent on the location and the season, from around 7 to 13◦ Brix [27,31,34]. This high dependence on the environment can make it difficult to map QTLs related with this trait. To overcome this problem, the development of ILs that facilitate the analysis of many replicates in different years and locations and without the interaction among loci observed in F2 and RIL populations is essential. In this way, QTLs consistently observed under different environmental conditions will be detected and later used for specific breeding programs. No significant differences were found in firmness of the flesh between cultivars or between years (FIR; Figure 1C; Table S1). This trait is also highly dependent on the environment, especially in PS, where it has been found to vary significantly between years, while Ved tends to be more stable [30,34]. Concerning fruit shape, PS produced more oval melons than Ved in both years, with longer and wider fruits in PS, while Ved produced rounded (FS close to 1) and smaller fruits (FL, FWI, FS; Figure 1D–F; Table S1). Significant differences between the parental lines were observed in traits related to ripening behavior. In climacteric melons, production of ethylene is coupled to certain ripening-related traits, such as production of aroma (EARO) and abscission layer formation (EALF), also affecting harvest date (HAR) [27]. While Ved ripened at around 35 days after Plants 2022, 11, 3120 3 of 18 pollination in both years, PS ripened around 20 days later, generally without producing Plants 2022, 11, x FOR PEER REVIEW 3 of 20 aroma or abscission layer (Figure 1G–I) (Table S1). The ripening behavior observed for both parental lines was in agreement with previous studies [27,30,33]. Figure 1.1. Fruit-quality-, Figure Fruit‐quality‐, fruit-morphology- fruit‐morphology‐ and and fruit-ripening-related fruit‐ripening‐related phenotypes phenotypes of of the the parental parental lines lines PS and and Ved Ved inin2020 2020and and2021. 2021.(A)(A) Fruit Fruit weight weight (FW)(FW) (g). (g). (B) Soluble (B) Soluble solid solid (°Brix)(◦(SSC). content content Brix) (C) Flesh (SSC). (C) firmness Flesh firmness cm−2(D) −2) (FIR). (kg cm(kg Fruit(D) ) (FIR). length Fruit(cm) (FL). length (E) (FL). (cm) Fruit (E) width (cm) Fruit (FWI). width (F)(FWI). (cm) Fruit shape (F) (FS). Fruit (G) (FS). shape Presence of aromaof (G) Presence (days aromaafter pollination) (days (EARO). (H) after pollination) Abscission (EARO). layer formation (H) Abscission layer (days after(days formation pollination) (EALF). (I) after pollination) Harvest (EALF). (I)date (days Harvest after date pollination) (days (HAR). Replicates after pollination) are rep‐ (HAR). Replicates resented with dark circles. Different letters (a, b) represent significantly different groups (p‐value < are represented with dark circles. Different letters (a, b) represent significantly different groups 0.05). (p-value < 0.05). No significant 2.2. Development differences of the IL Collectionwere found in firmness of the flesh between cultivars or between years (FIR; Figure 1C; Table For the development of the IL population, S1). Thiswetrait is also used highlysets the same dependent of SNPs asoninthe envi‐ Pereira ronment, especially in PS, where it has been found to vary significantly et al. 2021 [27]. Starting the Ved × PS cross in 2014, six or seven generations were needed between years, while to Ved tends complete to be more depending the population, stable [30,34]. on the IL (Figure 2). Concerning fruit shape, BC1 generation carried a mean PS produced of 50.2%more of theoval melons than PS genome, Vedfrom ranging in both years, 10.6% with to 76.6%, longer and and wider an average fruitsintrogressions, of 11.8 in PS, while Ved produced varying between rounded 5 and 17(FSintrogressions close to 1) and persmaller plant. fruits (FL, FWI, FS; Figure 1D–F; Table S1). Twenty BC1 plants with a mean of 8.3 introgressions and 65.6% of the recurrent parent were Significant selected fordifferences the next roundbetween the parentalInlines of backcrossing. BC2were observed generation, in traits related 46 seedlings carryingto ripening behavior. In climacteric melons, production of ethylene is coupled a mean of 5.3 introgressions and 77.3% of the PS genome were selected for the next round to certain rip‐ ening‐related of backcrossing. traits, In BC3 such as production generation, of aroma 55 plants (EARO) carrying and abscission an average layer formation of 3.4 introgressions and (EALF), 82.4% of also affecting the PS genome harvest date (HAR) were selected. Some[27].BC3 While Vedwere plants ripened at around while backcrossed, 35 days after others pollination were in both years, self-pollinated. DuringPS the ripened springaround season20ofdays 2017,later, bothgenerally BC3S1 and without producing BC4 generations aroma were or abscission genotyped, and layer (Figure 92 plants with1G–I) (Table of an average S1). 2.2The ripening behavior introgressions and 89.7%observed for of the PS both parental lines was in agreement with previous studies [27,30,33]. genome were selected. Seven BC3 families were also recovered in order to cover gaps in Plants 2022, 11, 3120 4 of 18 Plants 2022, 11, x FORthe genome. PEER REVIEWAt this point, five lines carried the desired single introgression. To complete 4 the population, in summer 2017, several interesting progenies were genotyped using the complete set of SNPs or a customized number of SNPs, depending on the line. With this new screening, 18 new ILs were obtained. In 2018, 24 additional families were genotyped and 29 ILs were2.2. finalized Development of the IL Collection (considering the ILs obtained in 2017). Moreover, in 2018, three more families were For the development genotyped with theof the IL population, complete set of SNPswetoused coverthe same gaps in sets of SNPs as in Per the genome. Finally, five ILs were added to the population in 2019, and two more in 2020, which werewere nee et al. 2021 [27]. Starting the Ved x PS cross in 2014, six or seven generations derived fromtothe complete the population, lines selected for coveringdepending on the IL (Figure gaps or unfinished lines. 2). Figure 2. Simplified Figurebreeding scheme 2. Simplified for the scheme breeding development for theofdevelopment the IL population, of theshowing the number IL population, showing the n of seedlings analyzed and the average PS genome at each generation. ber of seedlings analyzed and the average PS genome at each generation. In the BC1 generation, BC1 generationthe mean carriedheterozygosity a mean of 50.2% wasofH0 the= PS 0.48, and segregation genome, ranging from 10.6% distortion was observed for 10 SNPs. The highest distortion, χ 2 = 18, was in the SNP 76.6%, and an average of 11.8 introgressions, varying between 5 and 17 introgressions chr7_16723157. The Twenty plant. H0 of this SNP, BC1 located plants at athe with position mean 18,255,891 of 8.3 bp on chromosome introgressions and 65.6% of7,the recur was 0.63, withparent a predominance of the Ved allele. Segregation distortion has previously were selected for the next round of backcrossing. In BC2 generation, been 46 seedl observed in melon carryingin different a mean of chromosomes, including 5.3 introgressions chromosome and 77.3% of the PS 7 [37]. For the genome other were selected for SNPs that had segregation distortion, the H0 ranged from 0.41 to 0.44, below the mean. next round of backcrossing. In BC3 generation, 55 plants carrying an average of 3.4 in Finally, agressions set of 36 ILs andwas developed, 82.4% of the PScovering genome 97.89% of the genome, were selected. Some BC3 withplants an average were backcros of three unique introgressions per chromosome (Table S2). The information and graphical while others were self‐pollinated. During the spring season of 2017, both BC3S1 and representation of the population are provided in Table 1 and Figure 3. The developed IL generations were genotyped, and 92 plants with an average of 2.2 introgressions population had a mean of three introgressions per chromosome, with a range between two 89.7% of the PS genome were selected. Seven BC3 families were also recovered in o and four, and an average size of 14.6 Mb per chromosome. Both number of introgressions to cover gaps in the genome. At this point, five lines carried the desired single introg and average size are similar to previous IL populations developed in melon [14,27,28]. sion. To complete the population, in summer 2017, several interesting progenies were otyped using the complete set of SNPs or a customized number of SNPs, dependin the line. With this new screening, 18 new ILs were obtained. In 2018, 24 additional fam were genotyped and 29 ILs were finalized (considering the ILs obtained in 2017). Mo ver, in 2018, three more families were genotyped with the complete set of SNPs to c Plants 2022, 11, 3120 5 of 18 Table 1. Summary of the developed introgression lines population, with average introgression size per chromosome in physical and genetic distance. Introgression Size (bp) Introgression Size * (cM) Chr Number of ILs Average Min Max Average Min Max 1 4 15,012,079 7,083,403 23,611,517 49.5 37.7 58.3 2 4 12,741,987 2,293,836 23,903,255 57.1 14.7 100.6 3 2 11,857,082 6,062,824 17,651,340 53.2 36.8 69.7 4 3 17,188,567 11,842,362 24,802,312 69.6 65.5 75.1 5 4 9,353,072 2,799,579 22,919,604 45.5 24.9 63.3 6 4 9,574,343 2,430,563 27,849,153 38.2 21.9 66.7 7 3 19,189,571 8,232,872 25,979,214 81.8 67.5 100.8 8 4 12,919,496 1,895,546 25,002,865 47.8 12.9 92.1 9 2 23,045,254 20,847,233 25,243,276 85.6 61.7 109.5 10 2 15,007,760 5,108,239 24,907,280 70.3 54.0 86.6 11 2 25,057,113 20,321,128 29,793,099 89.6 54.0 99.4 12 2 13,781,830 4,579,847 22,983,813 51.9 47.1 56.8 Total 36 PEER REVIEW14,584,344 Plants 2022, 11, x FOR 1,895,546 29,793,099 58.5 12.9 109.5 6 of 20 * Genetic distance based on the genetic map in Pereira et al. 2018 [23]. Figure 3. (A) Graphical representation of the genotypes of the introgression lines and the two pa‐ Figure (A) Graphical 3. lines. rental representation (B) External of the genotypes and internal appearance of each of of the the introgression linesand introgression lines andthe the two parental parental lines. (B) External and internal appearance of each of the introgression lines and the parental lines. lines. 2.3. Phenotyping of the IL Population and QTL Mapping In 2019, a first phenotypic analysis of a subgroup of 17 ILs was performed (Table S3). In summer 2020, the available collection of 35 ILs was properly characterized and pheno‐ typed (Table S1) and the population was genotyped using the complete set of SNPs de‐ Plants 2022, 11, 3120 6 of 18 2.3. Phenotyping of the IL Population and QTL Mapping In 2019, a first phenotypic analysis of a subgroup of 17 ILs was performed (Table S3). In summer 2020, the available collection of 35 ILs was properly characterized and phenotyped (Table S1) and the population was genotyped using the complete set of SNPs described by Pereira et al. 2021 [27] (Table S2). Finally, in 2021, six unfinished lines in 2020 were phenotyped again, together with a newly developed line covering a gap on chromosome 5 (Table S1). 2.3.1. Fruit Quality Traits During the phenotyping, the following fruit quality traits were analyzed: mottled rind (MOT), immature rind color (ECOL), yellowing of the rind (YELL), presence of su- tures (SUT), flesh color (FC), soluble solid content (SSC) and flesh firmness (FIR). Those related with the external or internal appearance of the fruit were treated as qualitative (Figure 4A–E). The SSC and FIR, as well as morphological and ripening-associated traits, were treated as quantitative. The QTLs for quantitative traits are shown in Table 2. Mottled rind is a characteristic of the parental line PS. Within the IL collection, only VED2.3 did not have mottled rind (Figure 4A). This QTL was mapped at the end of chromosome 2, at interval 20.37–24.77 Mb. A major gene, Mt-2, has been previously described close to this region [23,27,38], and a candidate gene, MELO3C026282.2, encoding a protein essential for chloroplast development, has recently been reported as the causal gene of this QTL [39]. Both VED7.1 and VED7.2 had white immature rind compared to the green rind of PS and the rest of the ILs. This QTL was mapped to interval 2.62–20.73 Mb on chromosome 7 (Figure 4B). This region includes the major gene Wi, previously described as controlling this trait [40]. This interval colocalized with the region identified in two other populations of the same parental lines [23,27]. Related to the external color of the fruit as well, the entire population turned yellow during ripening (from around 25 DAP), except for VED10.1 and VED10.2, which remained dark green until harvest. This behavior has been related to flavonoid biosynthesis [41]. This trait was mapped on chromosome 10, in interval 1.76–5.11 Mb (Figure 4C), colocalizing with the previously described causal gene CmKBF [41]. Regarding the external aspect of the rind, VED11.1 and VED11.2 had sutures, while the rest of the population had smooth rind. This trait was mapped to a region on chro- mosome 11, in interval 14.14–29.79 Mb (Figure 4D). This region colocalized with QTLs reported in previous studies [23,27] and contains MELO3C019694.2, a gene that has been proposed as a candidate gene for this trait [42]. As for the internal appearance of the melon, we detected three different flesh colors (Figure 3B). While PS and most of the population had white flesh, other ILs had orange or green flesh. The flesh of VED8.3 was green when mature, mapping this QTL to chromosome 8 at interval 25.00–32.87. This region colocalizes with previously reported QTLs controlling this trait [43,44] and contains two genes that have been proposed as candidate genes for Wf, MELO3C003069.2 [45] and MELO3C003097.2 [42]. The orange flesh of VED9.1 and VED9.2 was similar to the Ved parental line, and this QTL was mapped in interval 1.62–22.56 Mb on chromosome 9 (Figure 4E). Gene CmOr (Gf ), located in our mapping interval on chromosome 9, has been reported as controlling accumulation of carotenoids in melon flesh, provoking the orange coloration [46]. The color of the flesh in melon is controlled by two genes under dominant epistasis, being Gf dominant over Wf. When Gf is not present (gf/gf ), then white is dominant over green [43–45]. In our population, PS has the alleles gf/gf Wf/Wf and Ved is Gf/Gf wf/wf, so VED8.3 had the alleles gf/gf wf/wf, resulting in green flesh, and VED9.1 and VED9.2 had Gf/Gf Wf/Wf, being orange-fleshed. Plants 2022, 11, x FOR PEER REVIEW 7 of 20 Plants 2022, 11, 3120 7 of 18 (Figure 4A–E). The SSC and FIR, as well as morphological and ripening‐associated traits, were treated as quantitative. The QTLs for quantitative traits are shown in Table 2. Figure 4. Figure 4. Fruit‐quality‐related Fruit-quality-relatedphenotypes phenotypesofof thethe population populationin 2020 andand in 2020 2021. (A) Presence 2021. of spots (A) Presence of in the rind (mottled rind), with a representation of chromosome 2 and the reported spots in the rind (mottled rind), with a representation of chromosome 2 and the reported causal causal gene. (B) Immature rind color, with a representation of chromosome 7 and the reported causal gene. (C) Yel‐ gene. (B) Immature rind color, with a representation of chromosome 7 and the reported causal gene. lowing of the rind, with a representation of chromosome 10 and the reported causal gene. (D) Pres‐ (C) Yellowing of the rind, with a representation of chromosome 10 and the reported causal gene. ence of sutures in the rind, with a representation of chromosome 11 and the reported causal gene. (D) (E) Presence Flesh color,of sutures in the rind, withofa chromosomes with a representation representation8ofandchromosome 9 and the 11 and thecausal reported reported causal genes. (F) gene. Soluble solid content (SSC) in °Brix. (G) Flesh firmness (FIR) in kg cm . (F,G) Replicates are genes. (E) Flesh color, with a representation of chromosomes 8 and 9 and −2 the reported causal repre‐ (F) Soluble sented withsolid dark content (SSC) circles; red in ◦ Brix. asterisk (G) Flesh indicates firmness p‐value < 0.05;(FIR) in kg different cm−indicate colors 2 . (F,G) Replicates are different chro‐ mosomes. with dark circles; red asterisk indicates p-value < 0.05; different colors indicate different represented chromosomes. Plants 2022, 11, 3120 8 of 18 Table 2. Summary of the QTLs mapped in the population for the quantitative traits. Genomic Trait QTL IL Year Chr. Interval (Mb) IL Mean PS Mean % Difference a p-Value FWQW7.1 VED7.1 2020 1090.4 −50.0 0.0131 FW VED7.2 2020 7 2.62–20.73 963.6 2175.6 −55.7 0.0148 FPQW4.1 VED4.2 2020 54.17 −13.30 0.0183 VED4.3 2020 4 22.48–27.25 54.58 −12.64 0.0259 FPQW8.1 VED8.1 2020 8 0–6.89 54.65 −12.53 0.0053 FPQW7.1 VED7.1 2020 48.66 62.48 −22.12 0.0004 FP VED7.2 2020 7 2.62–20.73 48.08 −23.05 0.0002 FPQW8.2 VED8.3 2020 8 31.12–34.62 45.2 −27.66 0.0002 FPQW11.1 VED11.2 2020 11 29.79–34.46 54.89 −12.15 0.0071 FPQW4.2 VED4.1 2021 4 0–14.92 71.52 62.2 14.98 0.0216 FAQW4.1 VED4.2 2020 191.39 −24.59 0.0336 VED4.3 2020 4 22.48–27.25 193.09 −23.92 0.0053 FAQW5.1 VED5.3 2020 27.63–29.32 200.19 −21.12 0.0203 FAQW8.1 VED8.1 2020 8 0–6.89 202.38 −20.26 0.0163 FA VED7.1 2020 156.19 253.8 −38.46 0.0006 FAQW7.1 7 2.62–20.73 VED7.2 2020 150.7 −40.62 0.0004 FAQW8.2 VED8.3 2020 8 31.12–34.62 136.38 −46.26 0.0002 FAQW11.1 VED11.2 2020 11 29.79–34.46 197.52 −22.17 0.0097 FLQW5.1 VED5.3 2020 5 27.63–29.32 18.3 −15.51 0.0477 FLQW6.1 VED6.3 2020 6 7.47–35.32 18.1 −16.44 0.0429 FLQW8.1 VED8.1 2020 8 0–6.89 18.33 −15.37 0.0461 FL FLQW7.1 VED7.1 2020 16.71 21.66 −22.85 0.0187 VED7.2 2020 7 2.62–20.73 16.76 −22.62 0.0167 FLQW8.2 VED8.3 2020 8 31.12–34.62 15.02 −30.66 0.0131 FLQW11.1 VED11.2 2020 11 29.79–34.46 17.04 −21.33 0.0171 FWIQW4.1 VED4.3 2020 4 30.00–34.31 13.42 −11.48 0.0473 FWIQW7.1 VED7.1 2020 12.07 −20.38 0.0171 FWI VED7.2 2020 7 2.62–20.73 11.41 15.16 −24.73 0.0130 FWIQW8.1 VED8.3 2020 8 31.12–34.62 11.45 −24.47 0.0179 FSQW6.1 VED6.3 2020 6 7.47–35.32 1.19 −16.78 0.0011 FS FSQW11.1 VED11.2 2020 11 29.79–34.46 1.15 1.43 −19.58 <0.0001 FIRQW2.1 VED2.2 2020 4.52 58.04 0.0298 VED2.3 2020 2 3.16–24.77 4.59 60.49 0.0475 FIRQW8.1 VED8.1 2020 8 0–6.89 0.88 −69.23 0.0459 FIR FIRQW8.2 VED8.2 2020 8 6.89–14.98 0.85 2.86 −70.03 0.0334 FIRQW8.3 VED8.3 2020 8 14.96–34.62 0.62 −78.32 0.0130 FIRQW11.1 VED11.2 2020 11 29.79–34.46 4.8 67.83 0.0171 EAROQW6.1 VED6.3 2020 6 7.47–35.32 49.5 −10.59 0.0078 EAROQW8.1 VED8.1 2020 8 0–6.89 40.89 −26.14 <0.0001 EAROQW8.2 VED8.2 2020 8 6.89–14.98 46.1 −16.73 <0.0001 EARO VED7.1 2020 47.17 55.36 −14.79 0.0063 EAROQW8.3 VED7.2 2020 8 31.12–34.62 46.8 −15.46 0.0007 VED8.3 2020 46.9 −15.28 0.0001 EYELLQW1.1 VED1.3 2020 29 0.12 0.0181 VED1.4 2020 1 29.95–34.47 31 0.19 0.0076 EYELLQW8.1 VED8.3 2020 8 25.00–29.53 37 26 0.42 0.0004 EYELL VED10.1 2020 43 0.65 0.0083 EYELLQW10.1 10 1.76–5.11 VED10.2 2020 54 1.08 0.0011 EYELLQW10.1 VED10.1 2021 10 1.76–5.11 56 27 1.07 0.0002 EALFQW6.1 VED6.3 2020 6 7.47–35.32 49.38 −11.52 <0.0001 EALFQW8.1 VED8.1 2020 41.43 −25.76 <0.0001 VED8.3 2020 8 2.63–6.89 44.56 −20.16 <0.0001 EALF EALFQW8.2 VED8.2 2020 8 6.89–14.98 47.33 55.81 −15.19 <0.0001 VED7.1 2020 49.83 −10.71 0.0016 EALFQW8.3 VED7.2 2020 8 31.12–34.62 48.6 −12.92 0.0002 VED8.3 2020 44.56 −20.16 <0.0001 ABSQW6.1 VED6.3 2020 6 7.47–35.32 1 0.89 0.0031 ABSQW8.1 VED8.1 2020 1.11 1.00 0.0008 VED8.3 2020 8 2.63–6.89 2.44 2.33 <0.0001 ABS ABSQW8.2 VED8.2 2020 8 6.89–14.98 1 0.11 0.89 <0.0001 VED7.1 2020 0.83 0.72 0.0369 ABSQW8.3 VED7.2 2020 8 31.12–34.62 1.2 1.09 0.0049 VED8.3 2020 2.44 2.33 <0.0001 HARQW6.1 VED6.3 2020 6 7.47–35.32 52.75 −5.75 <0.0001 HARQW8.1 VED8.1 2020 46.43 −17.05 <0.0001 VED8.3 2020 8 2.63–6.89 47 −16.02 <0.0001 HAR HARQW8.2 VED8.2 2020 8 6.89–14.98 52.63 55.97 −5.97 0.0002 HARQW8.3 VED7.2 2020 52 −7.09 <0.0001 VED8.3 2020 8 31.12–34.62 47 −16.02 <0.0001 aThe background color gradient corresponds to the percentage of reduction (red) or increase (blue) in the IL phenotype when compared to PS. Plants 2022, 11, 3120 9 of 18 Focusing on the other quality traits, we did not detect any significant difference regarding SSC in the population, even though a certain level of variability was observed (Figure 4F) (Table S1). We did, however, map several QTLs related with fruit firmness (Table 2). The flesh in three of the ILs on chromosome 8 (VED8.1, VED8.2 and VED8.3) was less firm compared to PS (FIRQW8.1, FIRQW8.2 and FIRQW8.3), probably caused by their climacteric behavior (Table S1). The effect of FIRQV8.1 and FIRQV8.3 was observed in our first phenotyping in 2019 (Table S3). Three other ILs (VED2.2, VED2.3 and VED11.2) had an increase in firmness (FIRQW2.1 and FIRQW11.1). In VED2.2 and VED2.3, this was significantly higher than in PS, as was the case for VED2.3 firmness in 2021, although this was not significant. FIRQW2.1 was mapped to interval 3.16–24.77 on chromosome 2. This QTL was also detected in the reciprocal IL population containing introgressions of PS in the background of Ved [27], with the overlapping region of both populations being 3.16–16.42 Mb. The VED11.2 fruit flesh was significantly firmer than PS, with a QTL (FIRQW1.1) on chromosome 11 at interval 29.79–34.46 (Figure 4G, Table S1), which colocalized with a previously reported flesh firmness QTL [37]. 2.3.2. Fruit Morphology Traits Fruit morphology is an important characteristic for fruit consumption, varying from small, round pocket melons to elongated large-sized varieties. Our parental lines had significantly different morphologic phenotypes in both evaluated years (Figure 1D–F), and the IL population segregated for these traits. On chromosome 7, there was a significant decrease of 50% of fruit weight in VED7.1 and VED7.2 between 2.62 and 20.73 Mb, containing QTL FWQW7.1 (Table S1, Figure 5A). We also detected the effect of this QTL in FP, FA, FL and FWI, with a decrease in all these traits (Figure 5, Table 2). In 2019, VED7.1 also produced smaller melons than PS (Table S3). A consensus QTL in this region has been previously reported, affecting fruit width in a melon RIL population between PI 414723 (subspecies agrestis) and ‘Dulce’ (subspecies melo), named fwi7.1 [47]. This QTL was located at 43 cM in chromosome 7; therefore, it should colocalize with both introgressions in VED7.1 and VED7.2, explaining the similar behavior observed in these two lines. However, our QTL was found to affect other traits related with fruit size. Since fruit morphology traits are shown to be correlated in melon [23,48], fwi7.1 and FWQW7.1 might be different QTLs, fwi7.1 being specific to fruit width, while FWQW7.1 is more general, affecting fruit size. FWQW7.1 can be useful in melon breeding programs for modifying fruit size without affecting fruit shape. A QTL cluster for fruit morphology was detected on chromosome 4, decreasing FP, FA and FWI but not FL, making the fruits thinner (Table S1). It was mapped to interval 22.48–27.25 Mb (Figure 5, Table 2). QTLs affecting fruit morphology have been previously reported in this chromosome [48,49]. This QTL was not detected in 2021 in VED4.2 (Figure 5). This line contained three SNPs in heterozygosis (SNP 30, 31 and 82 (Table S2)) in 2020 that were fixed in 2021. It may be possible that this QTL is unstable, or that, in the fixation process, the allele causing the effect in fruit morphology was lost. During the 2021 season, we detected another QTL for FP in the same chromosome but in a different interval (0–14.92 Mb), FPQW4.2, and with the opposite effect, producing larger melons. FPQW4.2 colocalizes with a QTL detected in a reciprocal IL population with the same parental lines, FPQP4.1 [27]. Both QTLs on chromosome 4, FPQW4.2 and FPQP4.1, consistently affect fruit size, producing larger melons when the Ved allele is present. On chromosome 8, we identified two different QTLs for fruit morphology, one in VED8.1 and the other in VED8.3. The QTL in VED8.1 produced flatter melons (Table S1, Figure 5) and significant changes in FP, FA and FL; it was mapped in interval 0–6.89 Mb at the beginning of the chromosome. This QTL collocated with a previously known QTL for fruit shape in an F2 population with PS as one of the parents [50]. Although not affecting the shape of the fruit, a QTL for fruit weight has also been mapped in the same region, producing smaller melons [23]. The second QTL, affecting VED8.3, was located at 31.12–34.62 Mb (Table S1, Figure 5) and affects fruit size (FA, FP, FL and FWI), producing Plants 2022, 11, 3120 10 of 18 smaller melons. This distal part of chromosome 8 has already been reported as one of the main controllers of melon fruit shape [48]. Close to this region, a gene controlling fruit shape has been characterized, CmOFP13 [29]. Our QTL does not colocalize with this gene, Plants 2022, 11, x FOR PEER REVIEW so the causal gene for the QTL covered by VED8.3 must be a different one. Both QTLs11 of on20 chromosome 8 affecting fruit shape were also detected in 2019 (Table S3). Figure5.5. Fruit‐morphology‐related Figure Fruit-morphology-related phenotypes phenotypesofofthe thepopulation in in population both years both 20202020 years and and 2021.2021. Rep‐ licates are represented with dark circles. Red asterisk indicates p‐value < 0.05 and different colors Replicates are represented with dark circles. Red asterisk indicates p-value < 0.05 and different colors indicate different chromosomes. (A) Fruit weight (FW) in grams. (B) Fruit area (FA) in squared cen‐ indicate different chromosomes. (A) Fruit weight (FW) in grams. (B) Fruit area (FA) in squared timeters (cm2). 2(C) Fruit perimeter (FP) in cm. (D) Fruit length (FL) in cm. (E) Fruit width (FWI) in centimeters (cm ). (C) Fruit perimeter (FP) in cm. (D) Fruit length (FL) in cm. (E) Fruit width (FWI) cm. (F) Fruit shape (FS). in cm. (F) Fruit shape (FS). A QTL cluster for fruit morphology was detected on chromosome 4, decreasing FP, A QTL mapped on chromosome 5 in interval 27.63–29.32 Mb (FLQW5.1) decreased FA and FWI but not FL, making the fruits thinner (Table S1). It was mapped to interval the fruit length and, consequently, the fruit area (Table S1, Figure 5). To our knowledge, 22.48–27.25 Mb (Figure 5, Table 2). QTLs affecting fruit morphology have been previously this is the first time that a QTL controlling fruit morphology has been mapped to the distal reported part in this chromosome of chromosome 5 in melon[48,49]. (Table This 2). QTL was not detected in 2021 in VED4.2 (Figure 5). This line contained three SNPs in heterozygosis We detected two QTLs affecting fruit shape on chromosome (SNP 30, 316 and 82 (Table in interval S2)) in Mb, 7.47–35.32 2020 that were fixed in 2021. It may be possible that this QTL is FSQV6.1 and FLQV6.1. VED6.3 produced more rounded melons than PS, significantly de-unstable, or that, in the fixation process,FL creasing the allele and causing FS (Table S1,the effect Figure 5).inAn fruit morphology association was this between lost.region During ofthe 2021 season, chromosome 6 we detected another QTL for FP in the same chromosome but and fruit shape has been reported previously in two different RIL populations [23,24] in a different interval (0– and an14.92 Mb), FPQW4.2, F2 population [50]. and with the FSQV6.1 wasopposite effect,inproducing also detected 2019 (Table larger S3). melons. FPQW4.2 col‐ ocalizes with a QTL detected in a reciprocal IL The last fruit-morphology-related QTL (FSQV11.1) was detected population with the same parental lines, on chromosome 11 FPQP4.1 [27]. Both QTLs on chromosome 4, FPQW4.2 and (29.79–34.46 Mb). VED11.2 produced rounder and also smaller melons, affecting FS, FL, FPQP4.1, consistently affect FA fruitFP. and size, This producing QTL has beenlargerpreviously melons when the Ved reported allele in an RILispopulation present. funded by the same On parents [23]. chromosome 8, we identified two different QTLs for fruit morphology, one in VED8.1 and the other The presence of QTLsin VED8.3. affecting TheallQTL traitsinrelated VED8.1toproduced flatter melons fruit morphology, (Table together S1, with Figure other QTLs5) and significant affecting changes only some in FP, FAbrings dimensions, and FL; outitthe wascomplex mappedgenetic in interval 0–6.89 Mb architecture of at the fruit beginning shape of the chromosome. and morphology in melon. Some This QTL collocated unstable QTLs werewithdetected, a previouslysuchknown QTL as the QTL for fruit shape in an F2 population with PS as one of the parents affecting size in VED4.2 (FPQW4.1 and FAQW4.1). This dependance on the environment [50]. Although not affect‐ ining QTLthemapping shape of thefor fruit, a QTL for traits morphological fruit weight has been hasreported also been mapped before in the when in melon same region, using producing different smaller melons populations, such [23]. as RILsThe[23]second and QTL, affecting ILs [14], and alsoVED8.3, was located in tomato at 31.12– [51]. Moreover, in34.62 Mb (Table tomato, genetic S1,interactions Figure 5) and affects among fruitSUN OFP, size (FA, FP, FL and and TRM5 genes FWI), haveproducing smaller been described, melons. This distal part of chromosome 8 has already been reported as one of the main controllers of melon fruit shape [48]. Close to this region, a gene controlling fruit shape has been characterized, CmOFP13 [29]. Our QTL does not colocalize with this gene, so the causal gene for the QTL covered by VED8.3 must be a different one. Both QTLs on chro‐ mosome 8 affecting fruit shape were also detected in 2019 (Table S3). Plants 2022, 11, 3120 11 of 18 affecting cellular elongation, division and, thus, fruit shape [52]. This variable behavior, together with genetic interactions between QTLs and the large number of genes found within these families in melon, explains the difficulty in controlling those traits in the field. Although many QTLs affecting fruit morphology have already been mapped, the discovery of new QTLs (FLQW5.1 and FWQW7.1) proved the importance of continuing generating genetically diverse material. Unravelling the complete and complex genetic architecture of quantitative traits such as fruit morphology will allow us to better predict the phenotypic performance of melon fruits. 2.3.3. Ripening Traits The analysis of ripening was centered around four related phenotypes: earliness of aroma production (EARO), earliness of abscission layer formation (EALF), harvest date (HAR) and the level of abscission (ABS). As PS is a non-climacteric melon, most of the fruits did not produce aroma or an abscission layer and were harvested at 55 DAP, around 20 days later than Ved, when they were fully ripe (Figure 1G–I). There were high levels of abscission only in climacteric line Ved, and we detected segregation for these four traits in the IL population, obtaining climacteric ILs. A QTL on chromosome 6 in the region 7.47–35.32 Mb affected all ripening-related traits. VED6.3 produced climacteric melons, with aroma and an abscission layer around 50 DAP (Table S1, Figure 6). A QTL related to climacteric ripening has been previously reported in this region in two IL populations. One was with the same parental lines, where the allele of PS delayed ripening in a Ved background [27], and another in an IL population with the same recurrent parent PS, where the introgression of the alleles of PI 161375 induced a climacteric response [53,54]. The causal gene for this QTL has been characterized Plants 2022, 11, x FOR PEER REVIEW 13 of 20 and identified as transcription factor CmNAC-NOR [33,55]. Figure Figure 6. Ripening‐related 6. Ripening-related phenotypesofofthe phenotypes thepopulation population in in 2020 2020 and and 2021. 2021. Replicates Replicates are are repre‐ represented withsented dark with dark circles. circles. Red Red asterisk asterisk indicatesindicates p-valuep‐value < 0.05