The economic benefits of pollination are the primary determinants for fruits, vegetables, and seed production, which influence at least 87 leading food crops around the world (Kevan and Viana, 2003; Klein et al., 2007). Pollination is crucial for reproduction in flowering plants whereby the male gamete (pollen grains) from the anther comes into contact with the female gamete (stigma). Likewise, successful reproduction is highly dependent on efficient pollinators (Blacquiere, 2010).

The trichromatic vision of bees is known to effectively handle different photoreceptor classes (blue, green, and UV ranges) (Peitsch et al., 1992). The floral pattern at the center of a flower can guide pollinators in finding the nectar (Lunau et al., 1996; Dafni and Giurfa, 1999; Lunau, 2005; Davies et al., 2012). Therefore, the center of a flower is easily recognizable by bees (Biesmeijer et al., 2005; Davies et al., 2012).

There is an important feature of distinct stigma color in each female flower, which directly provides a great support in plant reproduction. However, stigma color is a neglected trait in many cultivars of different fruit crops but it significantly possesses enormous benefits for desired crop production using different molecular breeding approaches. It has been reported that the level of carotenoid biosynthetic genes is associated with the accumulation of carotenoids and the resultant stigma color in Crocus sativus (Ahrazem et al., 2019). Similarly, during the transition from yellow undeveloped to red fully developed stigmas, the accumulation of zeaxanthin occurred due to the expressions of CsPSY and CsLcyb (Castillo et al., 2005). In a stable inherited yellow stigma tomato mutant (ys) that was obtained using ethyl methane sulfonate, a single recessive gene was found to regulate the yellowing of stigma due to the accumulation of naringenin chalcone in ys (Zhao et al., 2017). In rice, two genes controlling the purple stigma were mapped on chromosomes 1 and 6 using 1,300 F₂ populations that are derived from XQZ (purple stigma and red lemma tip) and Kitaake (white stigma and colorless lemma tip) (Wang, 2016).

Melon (Cucumis melo L., 2n = 24) is an attractive fruit crop due to the extreme divergences in phenotypic diversity with a reported production of more than 42 million tons globally in 2020 (FAOSTAT; http://faostat.fao.org). Melon flowers bear yellow petals and yellow to green stigmas, which are mostly pollinated by bees under the natural field conditions (Rodrigo Gomez et al., 2016). Therefore, the green stigma can be favorable over the yellow stigma for the production of desired fruits and seeds, aimed at different breeding purposes, respectively.

Compared to the traditional breeding approaches, the marker-assisted selection (MAS) system has been proved as a more effective strategy that is used for genetic mapping of different crop traits. Until now, not all the genetic components of melon cultivars have been dissected, and there is a dire need to investigate the genetic patterns and molecular mechanisms associated with desirable traits, which should be an important part of breeding programs, aimed at genetic improvement of crops. Therefore, this study was aimed at identifying the stable quantitative trait loci (QTLs) regulating the stigma color, using the respective mapping population of F₂ and F_2:3 families that are derived from the crossing of MR-1 (green stigma) and M4-7 (yellow stigma) melon lines, respectively. The present novel outcomes would be beneficial to provide the fundamental basis for the genetic understanding of melon stigma color trait.

Two different parent lines of melon MR-1 (P₁, female with green stigma) and M4-7 (P₂, male with yellow stigma) were selected as experimental material (Figure 1) and crossed to produce their F₁ progeny. The field experiments were performed in the plastic greenhouse at Xiang Yang Experiment Agricultural Station, Northeast Agricultural University, Harbin, China (lat. 44°04′N, long. 125°420′E) over 3 years (from 2019 to 2021). The plants were grown using a completely randomized design (CRD), and standard horticultural practices were adopted for successful germination.

In 2019, an F₂ mapping population comprising 133 plants was obtained from the parental lines crossing, and 55 plants with extremely divergent phenotypes were self-crossed to get the respective F_2:3 families. In 2020, all of the 55-F_2:3 families (20 plants for each family and a total of 1,100 individuals) were planted to detect the stable QTLs associated with stigma color. In 2021, an expanded population of 545-F₂ lines was also planted. In this study, fifteen (15) plants, each line for P₁, P₂, and F₁, generations, were grown along with all generations and subsequently utilized for genetic linkage analysis of melon stigma color.

The observation site was located in Harbin, Heilongjiang Province, China. The numbers of flower-visiting insects on the stigmas of both parent lines “MR-1 and M4-7” were visually observed during the peak flowering. The flowers were observed on daily basis from 7:30 up to 9:30 am and photos were taken for 3 consecutive days using a Gopro7 camera in the time-lapse mode.

The stigmas of respective flowers of F₂ and F_2:3 mapping populations were checked, three stigmas from each plant were chosen, and color phenotypes were collected from three flower repetitions using a 3NQ portable colorimeter (NR10QC). The respective color hues of “L,” “A,” and “B” were recorded, where “L” indicated lightness, “A” indicates red-green difference, “B” indicates yellow-blue difference, and “E” denoted the net color difference. The final three-dimensional orthogonal graph was plotted using the color phenotypes, and color difference “E” was calculated for each plant according to the following equation:

The leaf material (0.2 g) was taken from 2-week-old seedlings of P₁, P₂, F₁, and F₂ progeny, chilled in cryogenic liquid nitrogen, and freshly stored at −80°C. Then, high-quality DNA was isolated using the cetyl trimethyl ammonium bromide (CTAB) protocol with slight modifications. The quantified DNA was identified with 1% agarose gels electrophoresis fragmented with Bioruptor (Thermo Fisher Scientific, USA), and 200–300 bp fragments were prepared for the High-throughput Illumina™ X10 sequencing platform. The resultant clean end bases of paired-end sequencing of both parent lines were aligned to de novo assembled reference genome (DHL92, v3.5.1) of melon using the default algorithm of Burrows-Wheeler Aligner (BWA)-Maximal Exact Match (MEM) (https://sourceforge.net/projects/bio-bwa/files/). The Binary Alignment/Map (BAM) files were similarly used for the filtering and alignment of re-sequenced clean end bases. The major single-nucleotide polymorphisms (SNPs) were obtained and annotated using the SnpEff tool (v4.3), then 500 bp flanking sequences were extracted from each random SNP site.

The cleaved amplified polymorphism sequence (CAPS) markers were developed by identifying the suitable SNP-based restriction endonucleases (EcoRI, HindIII, PstI, BamHI, and BclI) using SNP2CAPS v0.6 and primer premier v6, respectively (Amanullah et al., 2020). The PCR products of each CAPS marker were digested with restriction endonucleases to verify the codominant polymorphism of each marker. The PCR reaction mixture was prepared as follows: a pair of primers (8–10 pmol), deoxynucleotide triphosphates (dNTPs) (0.25 mM), Taq buffer (10×), and Taq polymerase (1 unit). A thermocycler PCR was performed by preheating the samples at 94°C for 7 min, followed by 30 cycles of 30 s at 94°C, 30 s at 60°C for the first cycle, and a stepwise down-gradient of 0.5°C per cycle, followed by 60 s at 72°C. Then, 10 cycles of 30 s at 94°C, 30 s at 45°C, and 60 s at 72°C were performed, followed by elongation for 5 min at 72°C. For enzyme digestion, a reaction mixture consisted of 4 μl of PCR product, 4.8 μl of ddH₂O, 1 μl of CutSmart buffer, and 0.2 μl of restriction enzymes and was incubated at 37°C for 4 h. The digested products were subsequently cleaved by 1% agarose gel electrophoresis.

A genetic linkage map was developed and QTLs were mapped by using the default parameter settings of the IciMapping (v4.2) tool as followed (Meng et al., 2015). All the genotyped CAPS markers were evenly scattered across the whole genome chromosomes. The coded genotypic data of respective F₂ mapping populations were grouped and anchored over the whole genome chromosomes according to their exact physical positions and maximum likelihood means. The default chi-squared method and Kosambi's mapping function were selected to determine the segregation ratio of genotypic markers and to estimate the genetic distances, intervals, and positions at p (>0.001). The significant QTLs were defined above the threshold level of the default logarithm of odd (LOD) score (3.00) and the genome-wide type I error at α = 0.05.

Total RNAs were extracted using DiNing DP230-01 Plant Total RNA Purification Kit following the manufacturer's protocol (DiNing Biotech, Beijing, China). The qRT-PCR was performed on RNA extracted from the stigma of MR-1 and M4-7 on the day the female flowers were opened using the Actin gene as MELO03C023264. For qRT-PCR, assays were prepared using 20 ng cDNA and 300 nM of each primer in a 10 μl of reaction mixture with the addition of SYBR Green I Master Mix. qRT-PCR was performed using three biological replicates for each tissue sample and at least three technical replicates of each biological replicate. After normalization of the transcript level of each gene with the most suitable internal control gene for each sample, fold change was calculated by a 2^−ΔΔCT method.

The stigma color phenotype is often considered as a communication signal for natural pollinators of flowers, and its differential diversity is driven by the natural selection process. At the physiological level, pigment content accumulation and chloroplast development are associated with differential coloration in higher plants (Yang et al., 2015). Different concentrations of chlorophyll and carotenoid components were known to cause a variety of colors in cucurbit fruits (Henderson et al., 1998; Burger et al., 2010; Zhang et al., 2016).

Development and usage of bi-parental F₂ population are known as a conventional and rapid method to identify the significant putative genes regulating the important traits. The size of F₂ mapping population can reduce the influence of environmental factors and reduce the biasness of accurate findings (Amanullah et al., 2021). In this study, we used fairly good mapping populations of F₂ and F_2:3 families and effectively identified two stable QTLs in chromosomes 2 and 8, harboring the potential genes associated with stigma color. Further, genetic mapping in the expanded F₂ population also signified the delimited target regions of SC2.1 and SC8.1, respectively.

The QTL SC8.1 was consistent with the genetic position of major QTL GS8.1 that is reported in the previous report of our lab (Qiao et al., 2021), which similarly suggested the conservation of genetic mechanisms governing stigma color across distinct germplasms in melon. The gene MELO03C003165 was the only gene with nsSNP mutation locus in both candidate QTLs (SC8.1 and GS8.1) (Table 6). However, this gene exhibited to encode the uracil phosphoribosyl transferase (UPRT), triggered the plastid levels (Mainguet et al., 2009). The UK/UPRT has a dual role in coding both uridine kinase and uracil phosphoribosyltransferase that form uridine 5′-monophosphate (UMP) through the pyrimidine salvage pathway in Arabidopsis and regulates chlorophyll content in plants (Islam et al., 2007; Mainguet et al., 2009).

Further, our RT-qPCR analysis revealed 6 candidate genes harboring putative involvement in stigma color and similarly have explicit description, e.g., MELO03C017116, MELO03C017120, MELO03C017121, MELO03C017130, MELO03C017134, and MELO03C017147 (Table 5). The gene MELO03C017116 encodes kinesin-like protein (KAC), which is required for chloroplast dispersion within the cell under standard culture conditions (Suetsugu et al., 2012). In addition, KACs mediate the chloroplast light evasion response in an actin dependent (Shen et al., 2015). The gene MELO03C017120 encodes peroxidase, which mediates chlorophyll degradation in the chloroplast or vacuole in the presence of phenolic compounds, such as apigenin (Yamauchi et al., 2004). The gene MELO03C017121 encodes ring-type E3 ubiquitin transferase, which is homologous to the rice GW2 gene encoding ring-type E3 ubiquitin transferase. The earlier studies have confirmed that OsGW2 controlled the chlorophyll content with positive regulation of leaf senescence through genetic analysis of a knockout mutant (Shim et al., 2020). The gene MELO03C017130 encodes enhanced disease susceptibility 1 (EDS1), which has been shown to affect many biological processes, such as chlorophyll content and reactive oxygen species (ROS) metabolism in annual plants (Arabidopsis thaliana) and woody plants (Populus tremula L. × P. tremuloides) (Su et al., 2014; Bernacki et al., 2018). The gene MELO03C017134 encodes transcription factor MYBR1, and loss-of-function in Arabidopsis plants showed more rapid chlorophyll loss and senescence (Jaradat et al., 2013). The gene MELO03C017147 encodes zinc finger protein, and it has been reported that overexpression of zinc-finger protein gene (such as RHL41, AtZFP1) resulted in an increase in chlorophyll content of Arabidopsis (Kazuoka et al., 2010; Han et al., 2014). PSA2 is a member of the DnaJ-like zinc finger domain protein family that affects the light acclimation and chloroplast development (Wang et al., 2016). In conclusion, further screening of our candidate genes using more rigorous methodologies will provide more light in this direction.

At present, we infer that the difference in the stigma color of melon might be caused by the difference in total chlorophyll content in accordance with our results. There are other studies with similar results, chlorophyll deficiency in mutant tomato fruits triggered the yellowish skin color due to abnormal chloroplast development (Liu et al., 2021). At higher latitudes, some cucumbers are naturally subjected to a longer time for photosynthesis and chlorophyll synthesis, which results in green flesh color (Bo et al., 2019). However, the involvement of candidate genes in the synthesis and degradation of chlorophyll similarly indicate the specific biological pathway responsible for the differences in color of melon stigma, which need to be further investigated.

In addition, our field observations confirmed this favored behavior in bees feeding on melon flowers and showed a preference to visit the flowers with green stigma than flowers with yellow stigma. This confirms beyond the doubt that green stigmas are of greater value to melons both for farming and breeding programs. Thus, novel outcomes would be beneficial to provide the fundamental basis for the in-depth genetic understanding of the melon stigma color traits.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

YL performed the experiment, data curation, formal analysis, manuscript draft, and reviewed and edited the manuscript. PG and SL guided for the theoretical and practical experiments. XF, TZ, TL, and XW helped in formal analysis. SA reviewed and edited the manuscript language. FL supervised the research project and reviewed and edited the manuscript. All authors approved the final version of manuscript and disclosed no conflicts of interest.

This work was supported by the National Nature Science Foundation of China (31772331 and 32030094), the China Agriculture Research System of MOF and MARA (CARS-25), and Taishan Industrial Leading Talents Project [grant no. LJNY202112].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.