Genomic Survey and Expression Analysis of GLKs in Watermelon (Citrullus lanatus)

Abstract

Golden2-like (GLK) genes positively regulate chloroplast development, increase crop yields, and improve fruit quality. However, there has been no comprehensive identification and characterization of GLKs in watermelon. In this study, a total of 48 ClGLKs were identified in the watermelon genome. Based on phylogenetic analysis, they were divided into five groups. ClGLKs within the same group showed a similar gene structure and conserved motif compositions. Promoter analysis indicated that cis-elements responsive to light were the most abundant, though cis-elements associated with hormones, stress, and developmental regulation were also identified in ClGLKs promoters. Expression analysis indicated significant responses of some ClGLKs to drought and CGMMV stress, suggesting that these genes may participate in responses to biotic and abiotic stresses. Phenotypic analyses revealed enhanced chloroplast development and increased thylakoid density and chlorophyll content in the pericarp of a “dark green” watermelon cultivar. ClGLK8 was identified as the homolog of GLK1-2, the genes that promote chloroplast development and chlorophyll biosynthesis in fruits, and showed significantly increased expression in accordance with chloroplast development and chlorophyll accumulation. Our results provide detailed knowledge of the ClGLKs, which will enhance efforts to further improve the fruit quality of watermelon.

1. Introduction

The chloroplast is an important organelle that is the basic functional unit of plant photosynthesis and the biological factory of organic synthesis in the plant cell. Chlorophyll housed in the chloroplast can use light energy to convert carbon dioxide and water into organic substances. The synthesized organic substances can be used for plant growth and development, and the photosynthates transported into the fruit can improve fruit quality, by the accumulation of sugar and the influence of crop yields. Aside from chlorophyll (Chl), carotenoids are the other main pigments in the chloroplast. The ratios of these pigments in the chloroplasts of different plant organs such as leaves or fruit result in the various observed colors. Thus, enhancing the chloroplast function and manipulating the transformation of its pigments may have great significance in increasing crop yields [1], improving fruit quality [2], and enhancing plant aesthetics [3].

The chloroplast is a semi-autonomous organelle whose development requires coordination between its own genome and the nuclear genome. In fact, the genome of the chloroplast encodes few genes. Most proteins that function in the chloroplast are encoded and transported from the nuclear genome and play important roles in the development, maintenance, and degradation of the chloroplast [4]. Among these, several transcription factors (TFs) have unique contributions to chloroplast development and pigment metabolism. To date, several TF families, such as Golden2-like (GLK) [5], NAM, ATAF, and CUC (NAC) [6], KNOTTED-like homeodomain (KNOX) [7], and BEL1-like homeodomain (BELL) [8], have been found to regulate chloroplast biogenesis and chlorophyll metabolism. For example, members of the GLK and KNOX families have been shown to play positive roles in increasing the number and volume of chloroplasts and promoting Chl biosynthesis [5,7], however, the overexpression of AtNAC046 in Arabidopsis led to an early-senescence phenotype and lower chlorophyll content [6].

GLKs belong to the GARP superfamily of MYB TFs, which encode proteins that usually contain a Myb-DNA binding domain (DBD) and a C-terminal GCT box [9,10]. Since Golden2 was originally identified in maize, homologous genes have subsequently been cloned from many other plant species. To date, GLKs have been identified in Arabidopsis [5], rice [11], tomato [12], and many others. Most genetic and biochemical studies of GLKs have focused on their roles in regulating chloroplast development and functional maintenance. AtGLK1, AtGLK2, SlGLK1, and SlGLK2 have been shown to play positive roles in increasing the number and volume of chloroplasts, increasing the density of thylakoids, and promoting the biosynthesis of chlorophyll [13,14]. In addition, recent research has found that the photosynthetic capacity of C3 crops can be modified by using the GLKs of the C4 plant, significantly increasing the photosynthetic efficiency of rice and increasing the rate of yield up to 70% [15,16]. Overexpression of SlGLK1 and SlGLK2 in tomato resulted in increases in the chloroplast number and chlorophyll content in immature fruits and significantly increased the soluble solid, sugar, and carotenoid content in mature fruits [12,13]. These results demonstrate the significant benefits of exploiting GLKs to breed high-yield crops and improve fruit quality.

Watermelon is an important horticultural crop due to its sweet taste and rich source of nutrients. The sweetness is a vital characteristic for evaluating the quality of watermelon fruit and is determined by the total sugar content and ratios of the main sugar forms of glucose, sucrose, and fructose [17]. Photosynthesis is the critical mechanism for sugar accumulation in watermelon. Enhanced chloroplast development and chlorophyll content could increase the photosynthetic capacity, leading to a greater accumulation of sugar in fruit [12,18]. Therefore, GLKs could be used in breeding programs to improve the sugar content in watermelon fruit. In this study, GLKs were identified and characterized in watermelon. Then, the expression patterns of GLKs transcripts were analyzed in the fruit to better understand their roles in chloroplast development and Chl biosynthesis. The results of this study could provide candidate genes to improve the quality of watermelon fruit.

2. Materials and Methods

2.1. Identification and Characterization of GLKs in Watermelon

The watermelon (97103) v2 protein file was downloaded from the Cucurbit Genomics Database (http://cucurbitgenomics.org/, accessed on 7 December 2022). To identify GLKs in watermelon, the GLKs of Arabidopsis and tomato were used as query sequences to BLAST against the watermelon protein database. Alignments with E values less than E-5 were considered as candidate GLKs. The candidate proteins were then confirmed to be ClGLKs using SMART (https://smart.embl.de/, accessed on 20 November 2022) and Pfam (http://pfam.xfam.org/, accessed on 20 November 2022) to identify the conserved GLK domains. The detailed characteristics of putative ClGLKs were predicted as follows: the lengths of the amino acids were obtained from the protein data. The molecular weights (MW) and isoelectric points (pI) of the proteins were calculated using the Expasy site (http://web.expasy.org/protparam/, accessed on 21 November 2022). The signal peptides of proteins were predicted using the SignalP 5.0 Server (https://services.healthtech.dtu.dk/service.php?SignalP-5.0, accessed on 21 November 2022). Finally, all proteins were submitted to WoLF PSORT (http://wolfpsort.org/, accessed on 22 November 2022) to predict the subcellular localizations. All of the detailed information about ClGLKs is listed in Table S1.

2.2. Phylogenetic Analysis of ClGLKs

A phylogenetic tree of GLKs in watermelon, Arabidopsis, and tomato was constructed using MEGA-X with the maximum likelihood method. The phylogenetic tree was visualized using the online software tool EvolView (http://www.evolgenius.info/, accessed on 1 December 2022).

2.3. Gene Structure and Motif analysis

The information about the exon-intron organization of ClGLKs was obtained from the genomic database. Protein sequences were submitted to the MEME website (http://meme-suite.org/tools/meme, accessed on 5 December 2022) to identify conserved motifs. Finally, the gene structure and motif composition of ClGLKs were visualized using TBtools software [19].

2.4. Promoter Analysis of ClGLKs

For promoter analysis, sequences 2 kb upstream of the translation start site of ClGLKs were extracted from the genome and analyzed using PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/, accessed on 6 December 2022) to identify putative cis-elements. TBtools was then used to visualize the results of the cis-elements analysis.

2.5. Expression Analysis of Response of ClGLKs to Biotic and Abiotic Stresses

The transcriptome data of uninoculated watermelon fruits and watermelons inoculated with THE cucumber green mottle mosaic virus (CGMMV) obtained from the Cucurbit Genomics Database (http://cucurbitgenomics.org/, accessed on 7 December 2022) (BioProject: PRJNA389184) were used to analyze the expression of ClGLKs in response to biotic stress. The transcriptome data of the watermelons treated with PEG6000 applied to roots for six hours (BioProject: PRJNA326331) were used to analyze the expression of ClGLKs in response to abiotic stress.

2.6. Plant Materials and Sample Collection

Two watermelon cultivars of “W-22-72” and “W-22-301” were used as the plant materials in this study. Thirty plants per cultivar were grown in a greenhouse located in Gangji Country (Hefei City) with about a 30 °C/18 °C day/night cycle. The skin color of the cultivar “W-21-72” changed from green to dark green as it started to mature, while cultivar “W-21-301” kept a light green pericarp during the entire growth process. The pericarp of the fruit was collected at three stages (before, during, and after the change in pericarp color, which corresponded to about 7, 9, and 16 days after pollination of the fruit, respectively) in “W-22-72”, while the “W-22-301” samples were also collected at 7, 9, and 16 days after pollination. Each sample was collected in three biological replicates, and all samples were frozen in liquid nitrogen and immediately stored at −80 °C for further analysis.

2.7. Chl Content and Chloroplast Analysis

The Chl content samples was sent to KeMing Bio Co., Ltd. (Suzhou, China) to measure the Chl content. The fruit pericarp at different stages was soaked in FAA fixative solution, and the fixative solution was vacuum infiltrated into the tissue for 30 min. The treated samples were sent to Service Bio Co. Ltd (Wuhan, China) to analyze the chloroplast structure. The sample treatment included dehydration, resin penetration and embedding, polymerization, ultrathin section with 60–80 nm, and staining. Finally, the chloroplast ultrastructure was observed under transmission electron microscopy (TEM) and images taken with a 7000 times lens.

2.8. Expression Patterns Analysis of ClGLKs in the Pericarp of the Fruit

RNA-Seq was performed on the pericarp samples collected from two watermelon cultivars at three stages to analyze the expression profiles of ClGLKs. Genes with fold-changes ≥2 and p-values ≤ 0.05 were identified as differentially expressed genes (DEGs). The log₂-transformed RPKM expression values of the genes were used to generate a heatmap using TBtools.

2.9. RNA Extraction, cDNA Synthesis, and RT-qPCR Analysis

Total RNA was extracted from the samples using the Plant Total RNA Isolation Kit (Sangon, Shanghai, China). cDNA was generated using the MightyScript First Strand cDNA Synthesis Master Mix (Sangon, Shanghai). RT-qPCR was performed using the 2X SG Fast qPCR Master Mix (Sangon, Shanghai) according to the manufacturer’s protocol. The actin gene was used as the reference. The reaction system of RT-qPCR was carried out as follows: 10 uL qPCR Master Mix, 0.5 μL of forward primer and reverse primer, 1 μL of diluted cDNA and 8 uL of ddH₂O. Then, the reaction system was performed on a Bio-Rad CFX96 (Hercules, CA, USA), and the program was used as follows: 95 °C for 60 s, followed by 40 cycles of 95 °C for 10 s, and 60 °C for 30 s. Three biological and three technical replicates were performed and the relative expression of ClGLKs was analyzed using the 2^−ΔΔCT method [20]. The primer sequences used for RT-qPCR are listed in Table S4.

3. Results

3.1. Identification and Characterization of GLKs in Watermelon

By using a bioinformatics approach, a total of 48 GLKs were identified in the watermelon genome. These genes were annotated as ClGLK1 to ClGLK48 according to their chromosomal locations (Table S1). Furthermore, the amino acid lengths, MWs, pIs, signal peptides, and subcellular locations were determined to characterize the detailed physical and chemical properties of ClGLKs (Table S1). ClGLKs encoded proteins ranging from 138 (ClGLK42) to 710 (ClGLK20) amino acids in length with predicted molecular weights ranging from 16.26 kDa (ClGLK42) to 78.78 kDa (ClGLK20). The pI of ClGLK proteins ranged from 5.19 (ClGLK31) to 9.9 (ClGLK34). No ClGLKs were predicted to have signal peptides, indicating that they are all are likely to be non-secreted proteins. In addition, 44 GLKs were predicted to be localized in the nucleus. ClGLK30, ClGLK32, and ClGLK42 were predicted to be localized in the cytoplasm, with ClGLK37 in the chloroplast.

3.2. Phylogenetic Analysis, Gene structure, and Motif Analysis of ClGLKs

A single phylogenetic tree was built using the ClGLK protein sequences. Forty-eight ClGLKs were clustered into five separate groups based on bootstrap values (Figure 1). Group I harbored the largest number of ClGLKs with 13 members. Twelve ClGLKs were contained in groups II and V. Group III had seven ClGLKs, and group IV contained four members.

Gene structure analysis indicated that the gene structure varied among different groups, but in the same subfamily, most members showed great similarity. The ClGLKs in group I contained introns that varied from four to seven, with more than half of the members containing five introns (Figure 1). The introns in group II varied from three to six, with most genes containing five introns. Most genes in group III contained four introns, except for ClGLK27 and ClGLK47, which contained three introns. No introns were identified in the ClGLKs of group IV. For group V, the intron number varied from four to twelve, with half of the members containing five introns.

Based on MEME analysis, the motif compositions were conserved within each group, though they differed between groups. Motifs 1 and 2 (Table S2), which corresponded to the myb SHAQKYF domain were shared in most ClGLKs except for ClGLK42 (Figure 1). Aside from these, a unique conserved motif composition was observed in each group. For example, motif 3 was only found in the group I proteins. Group II only contained motifs 1, 2, and 15, though 15 was absent in several members. Motifs 6, 9, and 10 were uniquely distributed in the proteins of group III. Motif 8 was present in the ClGLKs of group IV, and the group V ClGLKs contained motifs 4, 5, and 7, except for ClGLK8. The diverse motif compositions might lead to functional diversity among the different groups of ClGLKs. The similar conserved motifs within each group indicated that their ClGLKs may have similar functions.

3.3. Phylogenetic Analysis of GLKs in Watermelon, Arabidopsis, and Tomato

An additional phylogenetic tree was generated using the GLKs of watermelon, Arabidopsis [21], and tomato [22] to infer the potential roles of ClGLKs. This phylogenetic tree was divided into five groups (Figure 2a). The placement of individual ClGLKs within each group was consistent with the results of the single phylogenetic tree of ClGLKs (Figure 1 and Figure 2a). The potential biological functions of ClGLKs could be inferred from the GLKs that have demonstrated functions in Arabidopsis and tomato based on clustering. ClGLK8 clustered together with AtGLK1 (AT5G44190), AtGLK2 (AT2G20570), SlGLK1 (Solyc07g053630), and SlGLK2 (Solyc10g008160). Sequence alignments showed that ClGLK8 had the most similarity to SlGLK1 (59%) in tomato and AtGLK2 (50%) in Arabidopsis (Figure 2b), suggesting that it may have a similar function to these genes.

3.4. Promoter Analysis of ClGLKs

Based on the cis-element analysis, 45 types of cis-elements were identified in the promoter regions of ClGLKs (Figure 3). Among these, cis-elements responsive to light were particularly abundant, accounting for 25 of the 45 cis-element types (Table S3). The promoters of all ClGLKs contained at least one type of light-responsive cis-element, indicating that light may play an important role in regulating the expression of ClGLKs. Aside from light, cis-elements related to stress response (drought, low temperature, and defense), hormone response (MeJA, abscisic acid, auxin, and gibberellin), circadian control, and endosperm expression were also identified in promoter regions (Figure 3, Table S3), indicating that ClGLKs may be involved in a variety of processes of plant development and stress response.

3.5. Expression Analysis of Response of ClGLKs to Biotic and Abiotic stresses

To gain insights into the potential roles of ClGLKs in response to abiotic and biotic stresses, gene expression was assayed via RNA-Seq after subjecting watermelons to the cucumber green mottle mosaic virus and drought stress (Figure 4a). For CGMMV treatment, the expression of 11 ClGLKs (ClGLK23, ClGLK25, ClGLK29, ClGLK30, ClGLK34, ClGLK46, ClGLK1, ClGLK6, ClGLK9, ClGLK16, and ClGLK31) were at low levels and not affected by viral infection. Among the responsive ClGLKs, 30 genes showed increased the expression compared to the control. Among these, 11 genes (ClGLK21, ClGLK41, ClGLK42, ClGLK10, ClGLK27, ClGLK43, ClGLK14, ClGLK22, ClGLK47, ClGLK37, and ClGLK35) increased the expression greater than two-fold. Of ClGLKs that decreased in expression, ClGLK18, ClGLK33, and ClGLK44 were down-regulated more than two-fold. These data suggest that most ClGLKs responded to CGMMV infection, though the variability in responses suggest different mechanisms in response to CGMMV infection.

Under drought stress, a total of 46 ClGLKs showed changes in expression compared to the control. Among these, 26 genes were up-regulated while 20 were down-regulated (Figure 4b). Further analysis found that 20 genes changed more than two-fold compared to the control, suggesting that these genes may be significantly regulated by osmotic stress treatment. Twelve ClGLKs (ClGLK8, ClGLK29, ClGLK30, ClGLK5, ClGLK37, ClGLK16, ClGLK46, ClGLK24, ClGLK26, ClGLK13, ClGLK33, and ClGLK11) were significantly up-regulated, suggesting that these genes may play a positive role in response to osmotic stress. Eight ClGLKs, including ClGLK1, ClGLK32, ClGLK20, ClGLK23, ClGLK42, ClGLK36, ClGLK18, and ClGLK2 were significantly down-regulated. Our data suggest that ClGLKs have varied expression responses to abiotic stress.

3.6. Characterization of Pericarp Phenotype and Chloroplast Development

In order to analyze the color difference between the two cultivars (Figure 5a), the Chl content of the fruit pericarp at three development stages in these two cultivars was determined. Chla, Chlb, and total Chl were found to increase consistently throughout development in cultivar “W-21-72” (Figure 5b). Notably, the contents of all three chlorophyll types were significantly higher at S3, about three-fold higher than at S2, indicating that the dark green pericarp may be caused by significantly increased Chl content. For cultivar “W-21-301”, there was no obvious change in the content of Chla, Chlb, and total Chl between S1 and S2. However, the Chl content at S3 was significantly higher than at S2, about a four-fold increase. In addition, the Chla, Chlb, and total Chl contents in cultivar “W-21-301” were lower than in cultivar “W-21-72” at all three developmental stages. At the S3 stage, Chla, Chlb, and total Chl contents in cultivar “W-21-72” were about two-fold higher than in cultivar “W-21-301”, indicating that the color difference between the two cultivars could be due to higher Chl accumulation.

The chloroplast structure was analyzed to uncover the mechanistic basis for the color difference between the two cultivars. Chloroplast development was found to be accelerated in the pericarp of the cultivar “W-21-72”, in which the chloroplast size and thylakoid density increased following fruit growth (Figure 5c). Compared to the cultivar “W-21-301”, the chloroplast size and thylakoid density were higher in “W-21-72” (Figure 5c). These results suggest that the dark green formation of the pericarp was promoted by chloroplast development and Chl accumulation.

3.7. Organ-Specific Expression Analysis of ClGLKs in the Pericarp of the Fruit

To gain insights into the potential function of ClGLKs in fruit development, their expression patterns were characterized using RNA-Seq data obtained from the pericarp of the two aforementioned cultivars. ClGLKs were found to have both conserved and divergent patterns of expression in these two watermelon cultivars (Figure 6). For example, nine genes (ClGLK6, ClGLK9, ClGLK16, ClGLK23, ClGLK25, ClGLK29, ClGLK30, ClGLK34, and ClGLK46) were either lowly expressed or not detected during fruit development in cultivars “W-21-72” and “W-21-301”. Some genes, such as ClGLK1, ClGLK4, ClGLK8, ClGLK13, and ClGLK15, had high expression levels in both cultivars. In addition, ClGLK8 and ClGLK24 had a similar expression pattern, significantly increasing from S2 to S3 in both cultivars.

Some ClGLKs, however, had different expression patterns between the two cultivars (Figure 6). The expression of ClGLK12, ClGLK32, and ClGLK44 decreased throughout development in “W-21-72”, while no obvious changes were observed in “W-21-301”. ClGLK21 and ClGLK26 increased expression significantly from S2 to S3 in “W-21-301”, while no changes were observed in “W-21-72”.

3.8. Characterization of the Role of ClGLK8 in the Color Change of the Fruit Pericarp

ClGLK8, a homolog of GLK1-2, may also play key roles in chloroplast development and Chl accumulation. From the expression data, ClGLK8 had high expression levels in watermelon fruit (Figure 6). In addition, it was found to significantly increase in expression in the fruit pericarp of “W-21-72” with each growth stage (Figure 6a). In the “W-21-301” cultivar, ClGLK8 significantly increased in expression only from S2 to S3 (Figure 6b), which correspond to the increase in Chl content from S2 to S3 (Figure 5b). These results suggest that ClGLK8 may play important roles in the Chl biosynthesis and chloroplast development of the fruit pericarp.

To further explore the role of ClGLK8 in the color formation of fruit pericarp, its expression level was compared between the two cultivars using both RNA-Seq and RT-qPCR data (Figure 7). The RT-qPCR data for ClGLK8 were consistent with the RNA-Seq data, and the results showed that the relative expression level of ClGLK8 in “W-21-72” was more than two-fold higher than “W-21-301” at the S2 and S3 stages. These results suggest that the differences in ClGLK8 expression may play a role in contributing to the color difference between the two cultivars.

4. Discussion

Watermelon is an important horticultural crop widely cultivated worldwide. Its sweetness, sugar ratios, and skin color, among other traits, are important qualities of the fruit [23], and improving these traits is a principal goal for breeding. GLKs are a class of transcription factors widely found in plants that have been implicated in the regulation of photosynthesis-related gene expression, chloroplast development, differentiation, and functional maintenance. The potential of GLKs to increase crop yields and improve fruit quality has been noted [13,15,16], and they may also be involved in leaf senescence and stress response [24]. Therefore, a comprehensive identification and analysis of GLKs in watermelon is necessary to identify candidate GLKs for the purpose of breeding watermelons to improve fruit quality and stress defense.

In this study, 48 GLKs were identified in the watermelon genome. The number of GLKs in watermelon was less than that in Arabidopsis (55) [21], tomato (66) [22], and maize (59) [25], which might be related to the relatively small genome of watermelon. The protein sequence length, MW, and pI were found to be quite variable among the ClGLKs. Phylogenetic analysis showed that the GLK gene of watermelon can be divided into five groups, while Arabidopsis and tomato were divided into seven and five groups, respectively. The results of the gene structure and motif composition analysis of GLKs in these species showed similar exon-intron organization and conserved motif compositions in each group, suggesting that the phylogenetic grouping of GLKs may have a significant correlation with the gene structure and motif composition. In addition, motifs 1 and 2, which corresponded to the myb SHAQKYF domain were identified in GLKs of Arabidopsis, tomato, and watermelon, indicating that these motifs may play a critical role in maintaining the function of GLKs. Most GLKs in Arabidopsis, tomato, and maize were predicted to localize in the nucleus and some members were predicted as the chloroplast proteins. In watermelon, most ClGLKs were also predicted as the nucleus proteins while ClGLK37 was predicted to localize in the chloroplast, suggesting that it may be related to the growth and development of the chloroplast.

Previous studies have found that different signals such as light, hormones, and stress regulate the expression of GLKs. The expression of GLKs in maize, rice, and Arabidopsis were found to be induced by light [5,10]. In tomato, IAA can directly repress the expression of SlGLK2 [26]. The binding site of ARFs was identified in the promoter region of SlGLK1, and SlARF10 was shown to bind to the promoter to activate the expression of SlGLK1 [27]. In addition, GLKs may also enhance plant defenses against environmental stress [28]. Cis-elements responsive to light, hormones (auxin, SA, ABA), and biotic stress were identified in the promoter regions of ClGLKs, suggesting that these signals may play important roles in regulating the ClGLKs involved in different developmental and growth processes.

Drought is one of the main stressors that limits watermelon growth and crop yields. GLKs may regulate the plant defenses to abiotic stressors such as cold and drought. Overexpression of the cotton GhGLK1 in Arabidopsis significantly improved plant resistance to drought and low temperature [28]. Ectopic expression of the peanut AhGLK1 gene in atglk1/atglk2 double mutants can promote chlorophyll biosynthesis and photosynthesis and promote plant recovery after drought-induced damage [29]. After PEG6000 treatment, the expression of 20 ClGLKs was found to be significantly changed by drought (twelve up regulated, eight down regulated), suggesting that these genes may play a variety of roles in response to drought stress.

GLKs have also shown positive resistance to biotic stress. Overexpression of AtGLK1 can enhance the plant resistance to Fusarium graminearum and Botrytis cinerea while the atglk1/atglk2 double mutant is more susceptible to cucumber mosaic virus and Botrytis cinerea [30,31,32]. Hetero-expression of peanut AhGLK1b in Arabidopsis resulted in strong resistance to Pst DC3000 and Ralstonia solanacearum [33]. GLKs may induce the expression of basic defense genes that significantly up-regulate or bind key factors of viral immunity, thus initiating an immune response [31,32]. CGMMV is an important viral pathogen harmful to Cucurbitaceae crops, causing severe disease symptoms in infected plants. For watermelon, it results in the transformation of the inner pulp to a water-soaked dirty red and even flesh acidulation in the fruit, a phenomenon known as ‘blood flesh’ [34], leading to serious economic losses. After CGMMV infection, the expression of 11 ClGLKs (ClGLK21, ClGLK41, ClGLK42, ClGLK10, ClGLK27, ClGLK43, ClGLK14, ClGLK22, ClGLK47, ClGLK37, and ClGLK35) showed a significant increase while ClGLK18, ClGLK33, and ClGLK44 decreased, indicating that these genes may participate in response to CGMMV.

One pair of homolog GLKs, GLK1 and GLK2, have been shown to promote chloroplast development in plants. For example, both AtGLK1 and AtGLK2 promote increases in the number and volume of chloroplasts and thus the accumulation of Chl in Arabidopsis [5]. In tomato, SlGLK1 and SlGLK2 have similar functions as AtGLK1 and AtGLK2 [13]. In this study, only a single ClGLK, ClGLK8, was clustered together with AtGLK1-2 and SlGLK1-2 in the phylogeny, suggesting that ClGLK8 may have a similar function as these proteins, and another homolog GLK in watermelon may have been lost during the evolution of the ClGLKs.

Though GLK1 and GLK2 have similar functions, they do have different expression patterns. Like Arabidopsis, tomato, and pepper, studies have revealed that GLK1 is often highly expressed in the leaves while GLK2 is highly expressed in the fruit. In Arabidopsis, only AtGLK2 was expressed in the siliques, and an AtGLK2 mutant resulted in a light green silique [5]. In tomato, both SlGLK1 and SlGLK2 are expressed in the leaves, though only SlGLK2 is expressed in fruits and plays a key role in regulating the green shoulder fruit of the U phenotype [13]. In pepper, the expression patterns of CaGLK1 and CaGLK2 were similar to SlGLK1 and SlGLK2, and CaGLK2 is also a key regulator of chloroplast development in immature fruits [35]. From our data, ClGLK8 had high expression in the fruit pericarp of the two watermelon cultivars at different stages, suggesting that ClGLK8 has a similar expression pattern to GLK2. Phenotypic analysis indicated that increased chloroplast volume, thylakoid density, and Chl content were correlated with the color change in the “dark green” watermelon cultivar “W-21-72”. In the “light green” watermelon cultivar “W-21-301”, the Chl content also significantly increased from the S2 to S3 stage. Transcriptome analysis indicated that the expression of ClGLK8 significantly increased in accordance with the changes in chloroplasts, thylakoid development, and Chl content, suggesting that ClGLK8 may play a key role in promoting the color change in watermelon fruit. In addition, chloroplast size, thylakoid density, and Chl content were increased in the “dark green” fruit compared to the “light green” cultivar, and ClGLK8 also had a relatively high expression level in the “dark green” fruit pericarp. This result suggests that ClGLK8 may play an important role in contributing to the color difference between the two cultivars. Overexpression of AtGLK2 and SlGLK2 has been shown to increase the soluble solids, sugar, and carotenoid content in fruits, indicating the potential role of ClGLK8 in the improvement in fruit quality. Further functional analyses are necessary to confirm this.

5. Conclusions

In this study, 48 GLKs were identified in the watermelon genome. Aside from the physical and chemical properties, we also analyzed their phylogenetic relationship, gene structure, and motif composition. Cis-elements and expression analysis suggest that most ClGLKs were induced by abiotic and biotic stress, and some members may play important roles in the modulation of drought stress and CGMMV infection. Furthermore, phenotype analysis of the “dark green” and “green” skin color of the watermelon fruit indicated that the difference in skin color was due to the larger chloroplast and high content of Chl. Analysis of the gene expression profiles suggests that ClGLK8 may play a key role in contributing to this difference in phenotype. Our findings provide a theoretical reference for the further study of the role of the ClGLKs in stress defense, and the suggested genes could be used for breeding to improve the fruit quality of watermelon.