The combined effects of Golden2‐like 2 and Pseudo‐Response Regulator 2 control the fruit pigmentation in pepper

SUMMARY

Fruit pigmentation mainly controlled by chlorophyll, carotenoids, and flavonoids is an important phenomenon for fortifying the nutritional and esthetic value of humans. Our previous study identified Pseudo‐Response Regulator 2 (PRR2) as a regulator of fruit pigmentation through its association with the c1 locus. A single variation of glk2 or prr2 did not affect the pigmentation of immature fruit in Capsicum accessions, while accessions harboring both glk2/prr2 showed pale‐green immature fruits. The immature fruit color in the F₂ population derived from a cross between “Milyang” (GLK2/PRR2, green fruit) and “Goshiki” (glk2/prr2, pale‐green fruit) was highly associated with GLK2 and PRR2. Dual variation in GLK2 and PRR2 caused by natural variation and gene silencing resulted in altered pigmentation in both immature and ripe fruits of “Numex halloween” and “Habanero orange.” Furthermore, co‐silencing of GLK2 and PRR2 in “Royal black” (GLK2/PRR2) resulted in pale‐green immature fruits. Reduced pigmentation was associated with a decrease in plastid size and thylakoid number, and the expression of chlorophyll and carotenoid biosynthesis genes was downregulated in the dual‐variated fruits. Quadruple variation by co‐silencing of GLK2 and CCS in “Numex halloween” (prr2/psy) resulted in white ripe fruits. Additionally, GLK2 was shown to physically interact with PRR2 in yeast two‐hybrid and bimolecular fluorescence complementation assays. Taken together, the combined effects of GLK2 and PRR2 determine c1 and fruit pigmentation, and our findings provide a genetic basis for improving fruit quality traits.

Significance Statement

Understanding the genetic control of fruit coloration is essential for elucidating the mechanisms that shape visual and nutritional traits in pepper. This study identifies the coordinated action of GLK2 and PRR2 as a central determinant of the c1 locus and overall fruit pigmentation, offering a genetic framework for improving fruit quality.

INTRODUCTION

Fleshy fruits are crucial for human nutrition and attract frugivores for seed dispersal, with the shift from unripe to ripe fruit representing a key survival strategy for plants (Karlova et al., 2014). Fruit ripening is a highly coordinated process involving physiological and biochemical changes including color, flavor, taste and firmness (Klee & Giovannoni, 2011). Fruit pigmentation is an important trait for both consumers and breeders as it affects visual appeal and nutritional value (Yuan et al., 2015). Elucidating the genetic regulation of pigment accumulation is essential for improving fruit quality (Taylor & Ramsay, 2005). Carotenoid is one of the major factors controlling fruit pigmentation and functions as precursors for phytohormones and volatile compounds (Song et al., 2023; Sun et al., 2022). In addition to pigment biosynthesis, its sequestration and stable storage in plastids are major factors for enhancing nutritional value (Li et al., 2016).

Tomato (Solanum lycopersicum) has been widely studied as a model for fleshy fruits (Giovannoni et al., 2017), thereby, comparative genomics based on tomato has been exploited for expediting gene mining to improve many fruit traits in pepper (Paran & van der Knaap, 2007). Among the orthologous regulatory genes involved in fruit ripening differentiation, Colorless non‐ripening (SPL‐CNR), Homeobox (HB‐1), and Uniform (Golden2‐like 2, GLK2) showed distinct expression patterns between tomato and pepper (Kim et al., 2014). SPL‐CNR encodes a SQUAMOSA promoter binding protein‐like transcription factor, and its epi‐mutant exhibited a pale‐yellow pericarp with reduced ethylene production (Do et al., 2024; Manning et al., 2006). HB‐1 was found to be involved in fruit ripening and floral organogenesis (Lin et al., 2008). GLK2 acts as a positive regulator of plastid number and pigment accumulation in Arabidopsis, tomato, and pepper (Brand et al., 2014; Fitter et al., 2002; Powell et al., 2012). Gene mining and comparative genomics in Capsicum have revealed species‐specific regulatory patterns in gene expression, focusing on processes such as carotenoid biosynthesis, metabolite accumulation, and key developmental dynamics during fruit ripening (Kim et al., 2014; Osorio et al., 2012).

Overexpression of the Arabidopsis Pseudo‐Response Regulator 2‐like (APRR2‐like) increased plastid development and enhanced pigment levels in tomato fruit (Pan et al., 2013). Overexpression of GLK2 increased the accumulation of transcripts related to the pigment biosynthesis process and plastid organization, resulting in higher levels of pigments, starch, and sugars (Nguyen et al., 2014). The dominant gain‐of‐function Curl mutant of tomato knotted 2 (TKN2) increased chloroplast numbers and chlorophyll levels in tomato fruit by upregulating both SlGLK2 and SlAPRR2‐like (Nadakuduti et al., 2014). Reduced expression of the zinc‐finger homeodomain protein, SlZHD17, resulted in larger chloroplasts and higher chlorophyll levels, which were regulated by increased expression of the chlorophyll biosynthesis gene, SlPORB, and the chloroplast developmental regulator, SlTKN2, respectively (Shi et al., 2021). Overexpression of the UV‐B photoreceptor 8, SlUVR8, promoted plastid development and pigment accumulation through enhancing GLK2 expression (Li et al., 2018). TOMATO AGAMOUS‐LIKE 1 (TAGL1) influences the distribution of chloroplasts, and knockout of TAGL1 resulted in dark green fruit with increased pigment levels (Jeon et al., 2024; Vrebalov et al., 2009). Knockout of the GRAS transcription factor, SlGRAS9, showed dark green fruits with enhanced chloroplast development and chlorophyll accumulation (Shi et al., 2024).

Bulked segregant analysis and genome‐wide association studies identified protein phosphatase 2C, CaPP2C35 as a candidate gene for the LG locus, controlling light‐green immature pepper fruit and chlorophyll accumulation in the exocarp (Wu et al., 2022). Fine mapping using genotyping‐by‐sequencing in C. chinense germplasm revealed that inactive shikimate kinase‐like 2 was a candidate gene regulating immature fruit color (Shu et al., 2023). A zinc finger transcription factor, pepper B‐box 10, activated the expression of magnesium chelatase subunit D and phytoene synthase 1 (PSY1), thereby promoting both chlorophyll and carotenoid accumulation (Wang, Shan, et al., 2025). A splicing mutation in zeaxanthin epoxidase (ZEP) reduced the conversion rate of zeaxanthin to antheraxanthin or violaxanthin, contributing to orange fruit color (Lee et al., 2021). Silencing of ZEP in red pepper resulted in bright red fruit color due to reduced capsanthin and increased zeaxanthin (Lee et al., 2024). An R‐R‐type MYB transcription factor, DIVARIVATA1, directly regulated PSY, capsanthin‐capsorubin synthase (CCS), and β‐carotene hydroxylase, positively controlling capsanthin accumulation in pepper (Song et al., 2023). MADS‐box transcription factor RIPENING INHIBITOR (MADS‐RIN) positively regulates carotenoid biosynthesis by targeting PSY and CCS, and also enhances their expression via interaction with DIVARICATA1, forming a protein complex (Wang et al., 2025).

Two major QTLs for pigment content, pc10 and pc1, which modulate plastid development, were identified as GLK2 and LSD one like 1 (LOL1), respectively (Borovsky et al., 2019; Brand et al., 2012; Brand et al., 2014). Null alleles at CaGLK2 were associated with chlorophyll reduction, and GLK2 expression was positively correlated with chlorophyll levels (Brand et al., 2014). Extreme reductions of chlorophyll content in C. chinense germplasm were induced by triple mutations in CaGLK2, CcLOL1, and CcPRR2 (Borovsky et al., 2019). The ripe fruit colors were determined by three loci (c1, c2, and y) (Hurtado‐Hernandez & Smith, 1985; Thorup et al., 2000), and PSY (Huh et al., 2001) and CCS (Lefebvre et al., 1998; Popovsky & Paran, 2000) are determinants of c2 and y, respectively. The c1 co‐segregated with Pseudo‐Response Regulator 2 (PRR2), which regulates the fruit color in both immature and ripe stages (Jeong et al., 2020; Lee et al., 2020). A nonsense mutation in APRR2‐like in pepper led to the white immature fruit color (Pan et al., 2013). PRR2 was identified as positive regulator of pigment accumulation by controlling plastid development and pigment biosynthesis (Lee et al., 2020). Silencing of PRR2 in yellow pepper reduced carotenoid content in the ripe fruit (Jeong et al., 2020).

We previously identified a PRR2 as a candidate of the c1 locus; however, we found that the orange fruit color in “Numex halloween” (psy1/CCS) was not altered by PRR2 silencing. In this study, we identified the interplay between GLK2 and PRR2 in controlling fruit pigmentation and determined the c1 locus in pepper.

RESULTS

Nonsense mutations in GLK2 and PRR2 result in pale‐green fruits with a high reduction of chlorophyll content and c1

In the previous study, we revealed that prr2 in “Habanero peach” caused the peach fruit color due to a reduction of plastid development and pigment biosynthesis (Lee et al., 2020). To further identify PRR2 function, we conducted virus‐induced gene silencing (VIGS) in C. annuum “Numex halloween” (hereafter NH). However, PRR2‐silenced NH showed green immature and orange ripe fruits similar to the control (Figure 1A), implicating that an additional factor may coordinate with PRR2 to control pigmentation. We explored the genetic variation of CCS, PSY1, and PRR2 in NH. No mutation was found in CCS, and a nonsense mutation due to G‐to‐A substitution at the 1431st nt was found in PRR2 (Figure 1B). A 19966 bp deletion of PSY1 was found in NH (Figure 1C). The fruit color model previously suggested by the c1c2Y (prr2/psy1/CCS) (Lee et al., 2020) is inconsistent with NH.

Figure 1.

PRR2 silencing in NH does not affect fruit pigmentation.

(A) No phenotypic difference in PRR2‐silenced NH was observed in immature and ripe fruits. The scale bars represent 1 cm.

(B) Gene structure of PRR2 and a nonsense mutation (G to A) in the 8th exon leading to a premature stop codon in NH.

(C) Gene structure of PSY1 and a null mutation caused by a 19 966 bp deletion at the genic and promoter regions were detected in NH.

Several genes, including GLK2, LOL1, and CaPP2C35, have been reported to control immature fruit color; therefore, these genes were sequenced in NH, “Habanero orange” (hereafter HO), and “Habanero peach” (hereafter HP). Sequence variation in PRR2 and GLK2 was identified in HO and HP (Figure S1). However, sequence variation in LOL1 (Figure S1), CaPP2C35 (Figure S1) was not found in NH, HO, and HP. To investigate the genetic relationship between GLK2 and PRR2 on immature fruit color, we analyzed 30 Capsicum accessions (Figure 2A, Table S1). GLK2/PRR2 conferred green immature fruit in “CM334,” “Milyang,” “Perennial,” “Royal black,” “TF68,” IT250176, IT237362, and IT231169. GLK2/prr2 conferred green immature fruit in NH. glk2/PRR2 conferred light‐green immature fruit in IT235383, IT201635, IT032406, IT203228, IT218849, IT236237, IT242040, IT258995, IT258997, IT199414, IT229690, IT240154, and IT259005, except for HO, which exhibited green. The light‐green accessions with moderate levels of chlorophyll had variations in GLK2 (Table S2). However, although the PRR2 variation in NH and the GLK2 variation in HO were found, their effects on chlorophyll levels were not consistent. glk2/prr2 conferred pale‐green immature fruit in “Goshiki,” IT158808, IT158845, IT158782, “Numex centennial,” HP, and IT236457. Variations in both GLK2 and PRR2 resulted in pale‐green fruits with a ~ 104.6‐fold reduction in chlorophyll levels compared to “TF68” and “Milyang.” These pale‐green accessions are expected to have truncated proteins lacking the conserved GLK/C‐terminal box (GCT box) due to premature stop codons in GLK2 (Figure S2a) and PRR2 (Figure S2b).

Figure 2.

glk2 and prr2 resulted in pale‐green immature fruits.

(A) Diverse immature fruit color mainly determined by combination of GLK2 and PRR2 variations in Capsicum accessions. Cultivar names with variations and fruit photos are shown in Table S1. The scale bars represent 1 cm.

(B) Segregation of immature fruit color in the F₂ population derived from a cross between “Milyang” and “Goshiki.” “Milyang” exhibited a green color (class 1) in immature fruit with GLK2 (GG) and PRR2 (PP) alleles, whereas “Goshiki” exhibited a pale‐green color (class 4) with the glk2 (gg) and prr2 (pp) alleles. The phenotype of 210 F₂ individuals was classified from 1 to 4.

(C) Immature fruit color in the F₂ individuals harboring the indicated allele combinations of GLK2 and PRR2. The Y‐axis represents color intensity in the F₂ population. The X‐axis represents the nine allele combinations at GLK2 (GG, Gg, and gg) and PRR2 (PP, Pp, and pp). Different letters indicate significant differences by one‐way ANOVA followed by Duncan's multiple test (P < 0.05).

To verify the association between pale‐green fruit color and glk2/prr2, we analyzed F₂ individuals derived from a cross between “Milyang” (GLK2/PRR2, green fruit) and “Goshiki” (glk2/prr2, pale‐green fruit) (Figure 2B). The F₂ individuals showed segregation of immature fruit color, which was green to pale‐green. Segregation for each allele at the GLK2 and PRR2 showed a strong correlation with immature fruit color (Figure 2C). GLK2 appeared to have a more pronounced effect than PRR2 in the F₂ population (Figure S3a,b). CaGLK2 was highly expressed during fruit ripening compared to SlGLK2, whereas PRR2 was expressed at low levels in both pepper and tomato (Figure S4a). Distinct expression patterns between CaGLK2 and SlGLK2 are expected to influence intense pigmentation in pepper immature fruits (Figure S4b). To investigate the mismatch of the ripe fruit color model found in NH, variations of GLK2, PRR2, PSY1, and CCS were further examined in Capsicum accessions (Table 1). Nonsense and frame‐shift mutations of GLK2, PRR2, PSY1, and CCS were identified (Figure S5). GLK2/PRR2/PSY1/CCS in “Royal black” (hereafter RB) and glk2/prr2/PSY1/CCS in “Numex centennial” led to rep ripe fruit. Coupled with GLK2/PRR2, psy1/CCS led to orange ripe fruit in IT250176, and PSY1/ccs determined orange‐yellow fruits in IT237362. psy1/ccs resulted in lemon‐yellow in IT231169. Coupled with psy1/CCS, glk2/PRR2 in HO and GLK2/prr2 in NH showed orange ripe fruits. The triple mutation (glk2/prr2/psy1) in HP resulted in peach fruit, and the quadruple mutation (glk2/prr2/psy1/ccs) in IT158782 led to white ripe fruit. These results suggest that GLK2 and PRR2 coordinate fruit pigmentation.

Table 1.

Fruit color depending on the genotypes of GLK2, PRR2, PSY1, and CCS in Capsicum accessions

Cultivar name	Species	Genotype				Immature fruit color		Ripe fruit color		GLK2/PRR2	PSY1	CCS
		GLK2	PRR2	PSY1	CCS
Royal black*	C. annuum	GLK2‐WT	PRR2‐WT	PSY1‐WT	CCS‐WT	Green		Red		C1	C2	Y
Milyang	C. annuum	GLK2‐WT	PRR2‐WT	PSY1‐WT	CCS‐WT	Green		Red		C1	C2	Y
Numex centennial*	C. annuum	glk2‐B	prr2‐C	PSY1‐WT	CCS‐WT	Pale‐green		Red		c1	C2	Y
Goshiki	C. annuum	glk2‐B	prr2‐C	PSY1‐WT	CCS‐WT	Pale‐green		Red		c1	C2	Y
IT250176	C. annuum	GLK2‐WT	PRR2‐WT	psy1‐C	CCS‐WT	Green		Orange		C1	c2	Y
Habanero orange	C. chinense	glk2‐D	PRR2‐WT	psy1‐B	CCS‐WT	Green		Orange		C1	c2	Y
Numex halloween*	C. annuum	GLK2‐WT	prr2‐C	psy1‐C	CCS‐WT	Green		Orange		C1	c2	Y
Habanero peach	C. chinense	glk2‐D	prr2‐A	psy1‐A	CCS‐WT	Pale‐green		Peach		c1	c2	Y
IT237362	C. annuum	GLK2‐WT	PRR2‐WT	PSY1‐WT	ccs‐A	Green		Orange‐yellow		C1	C2	y
IT231169	C. annuum	GLK2‐WT	PRR2‐WT	psy1‐C	ccs‐B	Green		Lemon‐yellow		C1	c2	y
IT158782	C. annuum	glk2‐B	prr2‐C	psy1‐C	ccs‐B	Pale‐green		White		c1	c2	y

^*Indicate the cultivars accumulating anthocyanin. The immature fruit color was described after AN2 silencing. The scale bars represent 1 cm.

GLK2 silencing in PRR2 ‐mutated NH causes fruit color changes

To identify the effects of both GLK2 and PRR2 on fruit color, VIGS was conducted to induce dual‐variation. GLK2‐silenced NH showed pale‐green immature and peach ripe fruits (Figure 3A). No visible change was observed in GLK2‐silenced leaves, while GLK1‐silenced leaves showed pale‐green color with a significant reduction of chlorophylls and carotenoids (Figure S6). The chloroplast size in GLK2‐silenced NH (8.52 μm²) was 2.4‐fold smaller than the control (3.63 μm²), as analyzed by transmission electron microscopy (TEM). The number of thylakoids per granum was 3.5‐fold lower in GLK2‐silenced NH (1.47) compared to the control (5.07) (Figure 3B,C). In the immature fruit of GLK2‐silenced NH, levels of chlorophyll a and b were 16.1‐fold and 12.0‐fold lower than the control, respectively. Levels of lutein and β‐carotene in GLK2‐silenced NH were 11.4‐fold and 9.1‐fold lower than the control. In the ripe fruits, capsanthin level in GLK2‐silenced NH was 4.7‐fold lower than the control (Figure 3D). In immature fruits, chlorophyll biosynthesis genes including HEMA, CHLH, CRD, DVR, PORB, and CAO were significantly downregulated in GLK2‐silenced NH. In ripe fruits, PSY1 expression in GLK2‐silenced NH was 2.2‐fold lower than the control, and CCS expression in GLK2‐silenced NH was 2.9‐fold higher compared to the control (Figure 3E).

Figure 3.

GLK2 silencing in PRR2‐mutated NH reduced fruit pigmentation.

(A) GLK2 silencing in NH resulted in a pale‐green color in immature fruits and peach color in ripe fruits. The scale bars represent 1 cm.

(B) Ultrastructure of plastids in immature fruit pericarps by TEM.

(C) Reduced plastid structure in GLK2‐silenced NH.

(D) Decreased content of chlorophyll and carotenoid in immature and ripe fruits.

(E) Expression analysis of chlorophyll biosynthesis and carotenoid biosynthesis genes in GLK2‐silenced NH by qRT‐PCR. ND: not detected. Significant differences were determined by Student's t‐test at *P < 0.05 and **P < 0.01.

PRR2 silencing in GLK2 ‐mutated HO causes fruit color changes

Pale‐green immature and peach ripe fruits were observed in PRR2‐silenced HO (Figure 4a). The chloroplast size in PRR2‐silenced HO (7.37 μm²) was 1.9‐fold smaller than the control (13.79 μm²). The number of thylakoids per granum in PRR2‐silenced HO (5.0) was significantly lower than the control (10.6) (Figure 4B,C). In immature fruits, levels of chlorophyll a and b in PRR2‐silenced HO were 12.8‐fold and 48.2‐fold lower than the control, respectively. Levels of lutein and β‐carotene in PRR2‐silenced HO (10.62 μg/g FW and 4.81 μg/g FW) were 6.1‐fold and 5.8‐fold lower than the control. In ripe fruits, capsanthin level was 3.6‐fold lower in PRR2‐silenced HO compared to the control, and zeaxanthin was only detected in the control (Figure 4D). Unlike the expression changes in GLK2‐silenced NH, the expression of CHLH, DVR, and CAO was not altered, and CHLM and CHLG expression were significantly reduced in PRR2‐silenced HO. The chlorophyll a/b ratio was decreased in PRR2‐silenced HO, while no difference was observed in GLK2‐silenced NH. The expression of carotenoid biosynthetic genes, including PSY1, PDS, ZISO, LCYB1, LCYB2, ZEP, and CCS, was significantly reduced in ripe fruit of PRR2‐silenced HO. CCS expression was reduced 3.6‐fold in PRR2‐silenced HO, while its expression was increased 2.9‐fold in GLK2‐silenced NH (Figure 3E). The expression of ZDS and CHY2 was not significantly altered in PRR2‐silenced HO (Figure 4E).

Figure 4.

PRR2 silencing in GLK2‐mutated HO reduced fruit pigmentation.

(A) PRR2 silencing in HO resulted in a pale‐green color in immature fruits and peach color in ripe fruits. The scale bars represent 1 cm.

(B) Ultrastructure of plastids in immature fruit pericarps by TEM.

(C) Reduced plastid structure in PRR2‐silenced HO.

(D) Decreased chlorophyll and carotenoid contents in immature and ripe fruits.

(E) Expression analysis of chlorophyll biosynthesis and carotenoid biosynthesis genes in PRR2‐silenced HO by qRT‐PCR. ND: not detected. Significant differences were determined by Student's t‐test (*P < 0.05 and **P < 0.01).

Dual‐variation of GLK2 and PRR2 affects flavor volatiles in ripe fruits

Volatile compounds in the dual‐variated fruit were analyzed using solid‐phase microextraction‐gas chromatography–mass spectrometry (SPME‐GC–MS). Forty‐seven and 43 volatile compounds were detected in GLK2‐silenced NH (psy1/prr2) and GFP‐silenced NH fruits, respectively (Table S3). Forty and 46 volatile compounds were detected in PRR2‐silenced HO (psy1/glk2) and GFP‐silenced HO fruits, respectively (Table S4). Seventeen flavor volatiles differentially accumulated by the dual variation were divided into three groups (Table 2). Group 1 showed a similar change in dual‐variated fruits of NH and HO, and Group 2 showed significant changes only in GLK2‐silenced NH. Flavor volatiles detected only in PRR2‐silenced HO were classified in Group 3. Among the seven volatiles in Group 1, methyl salicylate was significantly increased, while ocimenol, dill ether, and p‐menth‐1‐en‐9‐al were significantly decreased in both dual‐variated fruits. trans‐2‐heptenal was increased in PRR2‐silenced HO and was only detected in GLK2‐silenced NH. Ocimene quintoxide was detected only in the control fruits (NH and HO), while 1,2‐benzisothiazole was found only in dual‐variated fruits. Among the eight volatiles in Group 2, butyl acetate was increased, and cis‐2‐penten‐1‐ol, hexanal, trans‐2‐hexenal, linaloyl oxide, hotrienol, α‐terpineol, and 2′‐methylacetophenone were significantly decreased in GLK2‐silenced NH compared to the control. In Group 3, 4‐acetyl‐1‐methylcyclohexene and hexyl isopentanoate were found only in PRR2‐silenced HO. These results also suggest that GLK2 and PRR2 cooperate to regulate flavor volatiles in ripening fruits.

Table 2.

Volatile compounds detected in dual‐variated fruits. The data represents the mean of the ratio (n = 3)

Group	RT	Volatile compound	Odor description	Chemical class	Ratio
					NH‐GLK2::AN2/NH‐GFP::AN2	HO‐PRR2::AN2/HO‐GFP::AN2
1	19.82	Methyl salicylate	Green, Pepper, Minty	Ester	4.8*	2.5*
	19.16	trans‐Ocimenol	Citrus	Terpene	0.2*	0.7*
	19.7	Dill ether	Herbal, Dill‐like	Terpene	0.4**	0.4*
	20.49	p‐Menth‐1‐en‐9‐al	Fruity, Apple	Terpene	0.4*	0.5*
	12.40	trans‐2‐Heptenal	Green, Vegetable‐like	Aldehyde	+	2.4*
	15.78	Ocimene quintoxide	Citrus, Green, Woody	Terpene	−	−
	20.58	1,2‐Benzisothiazole	−	−	+	+
2	5.98	Butyl acetate	Green, Woody	Aldehyde	1.3*	0.9
	4.67	cis‐2‐Penten‐1‐ol	Sweet, Fruity	Alcohol	0.5*	0.9
	5.51	Hexanal	Green, Banana	Aldehyde	0.7*	0.9
	7.41	trans‐2‐Hexenal	Green, Fruity, Apple	Aldehyde	0.7**	0.9
	13.00	Linaloyl oxide	Floral, Green, Fresh	Terpene	0.5**	0.9
	17.54	Hotrienol	Tropical, Floral	Terpene	0.2**	0.5
	19.91	α‐Terpineol	Lemon, Sweet, Lilac	Terpene	0.3**	0.7
	19.63	2′‐Methylacetophenone	Sweet	Ketone	0.9*	0.7
3	18.27	4‐Acetyl‐1‐methylcyclohexene	Green, Fruity	Terpene	ND	0.5*
	20.98	Hexyl isovalerate	Green, Apple, Sweet	Ester	ND	0.3

Odor description referenced in http://www.thegoodscentscompany.com/; http://www.perflavory.com; https://flavornet.org/. “*” and “**” indicate significant differences at 5% and 1%, respectively, depending on the t‐test. “+” means detected only in dual‐variated fruit. “−” means detected only in control fruit. “ND” means not detected. “RT” means retention time.

Co‐silencing of GLK2 ‐ PRR2 in RB and GLK2 ‐ CCS in NH reduced levels of chlorophyll and carotenoid in immature and ripe fruits

Combining natural variation with gene silencing of GLK2 and PRR2 reduced pigmentation in immature and ripe fruits in NH (Figure 3) and HO (Figure 4). To further strengthen the genetic interplay, GLK2 and PRR2 were co‐silenced in RB (GLK2/PRR2). Silencing of GLK2 or PRR2 in RB showed no visible change in leaves (Figure S6c). Co‐silencing of GLK2 and PRR2 in RB resulted in pale‐green immature fruits (Figure 5A), as shown in Figure 2A. Silencing of PRR2 in RB did not affect the chlorophyll and carotenoid levels. However, silencing of GLK2 resulted in light‐green immature fruits, in contrast to the natural variation of GLK2 in HO (Figure 2A). In contrast, co‐silencing of GLK2 and PRR2 in RB resulted in red ripe fruits. Silencing of GLK2 or PRR2 in RB also resulted in red ripe fruits. The visual assessment of color differences in ripe fruits was difficult in RB.

Figure 5.

GLK2 and PRR2 co‐silencing in RB reduced fruit pigmentation.

(A) GLK2 and PRR2 silencing in RB resulted in pale‐green immature fruits. Single silencing of PRR2 did not induce a visible change, while GLK2 silencing showed pale‐green immature fruits. Downregulation of GLK2 and/or PRR2 in RB showed red ripe fruits. The scale bars represent 1 cm.

(B) Decreased chlorophyll and carotenoid contents in GLK2‐PRR2‐silenced RB.

(C) Expression analysis of chlorophyll biosynthesis in GLK2 and/or PRR2‐silenced RB by qRT‐PCR.

(D) Expression analysis of carotenoid biosynthesis genes in GLK2 and/or PRR2‐silenced RB. ND: not detected. Statistical analysis was conducted by Student's t‐test (*P < 0.05 and **P < 0.01).

In immature fruits, the levels of chlorophyll a and b in GLK2‐PRR2‐silenced RB were 13.4‐fold and 12.8‐fold lower than the control, respectively (Figure 5B). In GLK2‐silenced RB, the levels were 4.3‐fold and 4.1‐fold lower, respectively. The level of lutein and β‐carotene in GLK2‐PRR2‐silenced RB was 6.9‐fold and 5.1‐fold lower than the control, respectively. In GLK2‐silenced RB, the levels were 2.9‐fold and 2.2‐fold lower, respectively. In contrast, no significant differences in chlorophyll and carotenoid levels were observed in PRR2‐silenced RB. In ripe fruits, the levels of zeaxanthin and β‐carotene in GLK2‐PRR2‐silenced RB were 1.5‐fold and 1.7‐fold lower than the control, respectively (Figure 5B). The levels of capsanthin, neoxanthin, and β‐cryptoxanthin were slightly reduced in GLK2‐PRR2‐silenced RB compared to the control (Table S5). No significant difference in carotenoid levels was observed in GLK2‐silenced RB and PRR2‐silenced RB. Phytofluene and lutein were not detected in GLK2 and/or PRR2‐silenced RB.

The expression of chlorophyll biosynthetic genes, including HEMA, CHLM, CRD, DVR, PORB, CAO, and CHLG, was downregulated in GLK2‐PRR2‐silenced RB except for CHLH (Figure 5C). Among the chlorophyll biosynthetic genes, only PORB was reduced in GLK2‐silenced RB, whereas the expression of the other genes was not significantly altered. In PRR2‐silenced RB, no significant difference was observed in the expression of any of the biosynthetic genes. The expression of carotenoid biosynthetic genes, including PSY1, ZISO, CHY2, and ZEP, was significantly downregulated in GLK2‐PRR2‐silenced RB (Figure 5D). CHY2 expression was also reduced in both GLK2‐silenced RB and PRR2‐silenced RB.

The fruit color model, c1c2y, produces white ripe fruit (Hurtado‐Hernandez & Smith, 1985), and IT158782 showing white ripe fruit contained the quadruple mutation (glk2/prr2/psy1/ccs) (Table 1). Co‐silencing of GLK2 and CCS in NH (prr2/psy1) caused the white ripe fruit (Figure 6A,B). Carotenoid was not detected in GLK2‐CCS‐silenced NH and IT158782 (Figure 6C).

Figure 6.

Co‐silencing of CCS and GLK2 in NH and quadruple mutation in IT158782 resulted in white ripe fruits.

(A) Co‐silencing of GLK2 and CCS in NH (psy1/prr2) showed white color in ripe fruit, similar to the white ripe fruit of IT158782 (glk2/prr2/psy1/ccs). The scale bars represent 1 cm.

(B) Reduced expression of GLK2 and CCS in ripe fruit of GLK2‐CCS‐silenced NH. Significant differences were determined by Student’s t‐test at *P < 0.05 and **P < 0.01.

(C) Carotenoid was not detected in the ripe fruits of GLK2‐CCS‐silenced NH and IT158782. ND: not detected.

GLK2 interacts with PRR2

Physical interaction between GLK2 and PRR2 was analyzed using a yeast two‐hybrid (Y2H) assay (Figure 7A). The combinations of constructs grew on the selective medium (SD‐LWA, SD‐LWH + 3AT). Yeast cells of the positive control (p53_BD/T_AD) could grow under the selective medium, whereas the negative control (BD/AD, GLK2_BD/AD, BD/PRR2_AD) did not survive on the same medium. The co‐transformation of yeast cells with both GLK2_BD and PRR2_AD showed that these yeast cells were able to grow on the selective medium. All the pale‐green fruit accessions were expected to have a deletion in the GCT box of GLK2 and PRR2 (Figure S2; Table 1). To determine whether the GCT box is required for the interaction, we performed Y2H assay with truncated forms of GLK2 (GLK2_∆GCT) and PRR2 (PRR2_∆GCT). Co‐transformation of PRR2_∆GCT with GLK2_BD or GLK2_∆GCT abolished the interaction, whereas GLK2_∆GCT hardly interacted with PRR2_AD. Next, we further verified the interaction by a bimolecular fluorescence complementation (BiFC) (Figure 7B). GLK2 fused to the N‐terminus of yellow fluorescence protein (YFP) (GLK2_nYFP) and PRR2 fused to the C‐terminus of YFP (PRR2_cYFP) were co‐transfected to protoplasts. Additionally, empty cYFP or nYFP vectors were co‐transferred with GLK2_nYFP or PRR2_cYFP for use as a mock control. The YFP signal was observed in the nucleus upon GLK2_nYFP and PRR2_nYFP being co‐expressed. The YFP signal was not observed in cells co‐expressing either GLK2_nYFP with cYFP or PRR2_cYFP with nYFP. These results indicate that GLK2 interacts with PRR2 presumably through the GCT box.

Figure 7.

GLK2 interacts with the PRR2.

(A) Yeast two‐hybrid assay showed the physical interaction between GLK2 and PRR2 in yeast. Yeast clones transformed with each construct were grown on SD‐LW (‐Leu, ‐Trp) and SD‐LWH (‐Leu, ‐Trp, ‐His) with 5 mm 3‐AT, a competitive inhibitor of HIS3. Yeast cells transformed with BD‐p53 and AD‐T were used as positive control, while empty vectors pGBKT7 (BD) and pGADT7 (AD) were used as negative control. Schematic diagrams show the truncated modules of GLK2 and PRR2.

(B) Bimolecular fluorescence complementation assay of GFP verifying interaction of GLK2 (fused with the N‐terminal fragment of Venus) with PRR2 (fused with the C‐terminal fragment of Venus) in rice Oc cell protoplasts. Red fluorescent protein (RFP) was used as a nucleus marker. The scale bars represent 10 μm.

DISCUSSION

Multifarious fruit color is necessary to satisfy the diversity of human preferences. Diverse color differentiation in immature and ripe fruits mainly depends on qualitative and quantitative variations of carotenoids and chlorophylls. We previously identified PRR2 as a positive regulator of fruit pigmentation in both immature and ripe fruits and determined its association with the c1 locus (Lee et al., 2020). However, silencing PRR2 in NH (psy1/CCS) did not alter fruit color, showing a discrepancy with the fruit color model (Figure 1). Therefore, further investigation for in‐depth dissection of the c1 locus was necessary. Several candidate genes including GLK2, LOL1, and CaPP2C35 were reported to control immature fruit color (Borovsky et al., 2019; Brand et al., 2014; Shu et al., 2023; Wu et al., 2022). No variation in LOL1 and CaPP2C35 was observed among NH, HO, and HP (Figure S1). GLK2 is a positive regulator of chlorophyll biosynthesis in tomato (Nguyen et al., 2014; Powell et al., 2012), pepper (Brand et al., 2012; Brand et al., 2014), maize (Rossini et al., 2001), and Arabidopsis (Waters et al., 2009). In tomato, the two GLK homologs regulate pigmentation in different tissue: GLK1 primarily in leaves and GLK2 in fruits (Nguyen et al., 2014). Reduced chlorophyll content was observed in GLK1‐silenced leaves, whereas pigment levels were not significantly changed in GLK2‐silenced leaves (Figure S5). These data suggest that spatial regulation of pigmentation occurs in orthologous GLKs in both pepper and tomato. PRR2 is also a positive regulator of fruit pigmentation in tomato (Pan et al., 2013), cucumber (Jiao et al., 2017), and melon (Oren et al., 2019). CaGLK2 showed a distinct expression pattern compared to SlGLK2, while PRR2 was expressed at low levels in tomato and pepper (Figure S4a). It suggests that the increased expression of CaGLK2 affects the intense pigmentation in pepper (Figure S4b), and GLK2 may have a more significant influence on fruit pigmentation compared to PRR2. Moreover, silencing of GLK2 or PRR2 in RB did not affect pigmentation in vegetative tissues (Figure S6c). These results suggest that both genes spatially regulate fruit pigmentation.

Pepper accessions harboring glk2/prr2 exhibited pale‐green immature fruits compared to single variations (GLK2/prr2 and glk2/PRR2) and wild types (GLK2/PRR2) (Figure 2A). The homozygous genotypes of glk2 and prr2 in F₂ population resulted in pale‐green immature fruits (Figure 2C). Furthermore, the homozygous genotypes of glk2 resulted in light‐green or pale‐green immature fruits, regardless of PRR2 genotypes. The dual variation in GLK2 and PRR2 altered the pigmentation of both immature and ripe fruits in NH (Figure 3) and HO (Figure 4). Co‐silencing of GLK2 and PRR2 in RB led to the pale‐green fruits, while GLK2‐silenced RB conferred light‐green fruits (Figure 5A). These results are consistent in the Capsicum accessions (Figure 2A), showing that glk2/PRR2 accessions (excluding HO) produced light‐green fruits, while glk2/prr2 accessions exhibited pale‐green fruits. These findings suggest that GLK2 functions as the major regulator of immature fruit pigmentation, whereas PRR2 contributes to regulation through an additive interaction. The contribution of CcLOL1 to the coloration of immature fruit in C. chinense is more pronounced than in C. annuum, where the major effect is attributed to CaGLK2 (Borovsky et al., 2019). The exception observed in HO could be attributed to unique characteristics of C. chinense, and further analysis will be necessary to identify other factors contributing to the pigmentation in HO. Knockout of BEL1‐LIKE HOMEODOMAIN 2 (BEL2) in tomato increased the formation of green shoulder by altering the gradient distribution of chlorophyll accumulation (Niu et al., 2022). Knockout of KNOX II gene SlKN5 in tomato enhanced chloroplast number and chlorophyll levels in pericarp and locular (Ezura et al., 2024). C1c2Y resulted in orange fruit color (Huh et al., 2001), while c1c2Y resulted in pale orange (Hurtado‐Hernandez & Smith, 1985) or peach (Thorup et al., 2000). In ripe fruits, psy1 coupled with either glk2 or prr2 produced an orange color, while the triple mutation (glk2/prr2/psy1) resulted in a peach color (Table 1). The quadruple mutation (glk2/prr2/psy1/ccs) in IT158782 and co‐silencing of GLK2 and CCS in NH (prr2/psy1) resulted in white ripe fruits (Figure 6). C. frutescens “AC08‐201” also contained a quadruple mutation, producing white ripe fruit color (Jeong et al., 2020). Taken together, these results indicate that both GLK2 and PRR2 are determinants of the c1 locus, representing a genetic factor regulating immature and ripe fruit color.

Enhancing the ultrastructure and biosynthetic capacity of plastid is crucial for increasing pigment accumulation (Sun et al., 2018). Dark‐green fruit in GLK2‐overexpressing tomato is due to increased storage capacity, attributed to an elevated number of chloroplasts and thylakoid grana stacks (Nguyen et al., 2014). Overexpression of APRR2‐like in tomato increased plastid size, indicating enhanced plastid proliferation and enlargement (Pan et al., 2013). Increased chloroplast development in immature fruits leads to ripe fruits with more active chromoplasts, resulting in higher levels of metabolites related to nutritional quality (Cocaliadis et al., 2014). ORANGE, a DnaJ‐like zinc finger domain‐containing protein, is a multifunctional regulator of plastid development and carotenoid biosynthesis that is critical for storage capacity (Sun et al., 2022). The dual variation induced by silencing GLK2 in NH (Figure 3B,C) and PRR2 in HO (Figure 4B,C) reduced pigment levels, which is likely a consequence of a decreased plastid size and thylakoid number.

The chlorophyll biosynthetic genes including HEMA, CRD, and PORB were downregulated in the dual‐variated fruits (Figure 3E; Figure 4E) as well as GLK2‐PRR2‐silenced RB (Figure 5C). The expression of CAO was significantly reduced in GLK2‐silenced NH (Figure 3E) and GLK2‐PRR2‐silenced RB (Figure 5C), while it was not altered in PRR2‐silenced HO (Figure 4E). Overexpression of CAO increased chlorophyll b content and decreased the chlorophyll a/b ratio in tobacco (Biswal et al., 2012). Thus, the decreased chlorophyll a/b ratio observed in PRR2‐silenced HO (Figure 4D) may be a response to maintain CAO expression. Interestingly, the GLK2‐silenced RB exhibited a light‐green color and reduced level of chlorophylls, whereas the PRR2 silencing did not alter pigmentation (Figure 5A,B). The GLK2‐silenced RB resulted only in a significant reduction of PORB expression (Figure 5C). PORB is essential for the conversion of protochlorophyllide to chlorophyllide, ensuring efficient chlorophyll production, and mutation of CqPORB in quinoa leads to reduced chlorophyll and defective thylakoid membranes (Li et al., 2022). Among the chlorophyll biosynthesis genes, only the expression of POR was increased in GLK2‐overexpressing tomato fruits (Nguyen et al., 2014). The light‐green fruit color observed in GLK2‐silenced RB is likely attributable to the downregulation of PORB.

The expression of PSY1 was consistently decreased in dual‐variated fruits of NH (Figure 3E) and HO (Figure 4E), as well as in GLK2‐PRR2‐silenced RB (Figure 5D). The expression of CCS was increased in GLK2‐silenced NH, decreased in PRR2‐silenced HO, and showed no significant difference in GLK2‐PRR2‐silenced RB. PSY1 expression was positively correlated with capsanthin content, while CCS expression showed no correlation with color intensity in pepper fruits (Berry et al., 2019). Reduced level of capsanthin in GLK2‐silenced NH (Figure 3D) and PRR2‐silenced HO (Figure 4D) is likely due to the decreased expression of PSY1. Zeaxanthin and β‐carotene were significantly reduced in the GLK2‐PRR2‐silenced RB, while capsanthin showed a slight decrease (Figure 5B). The expression of PSY1, ZISO, CHY2, and ZEP was downregulated in the GLK2‐PRR2‐silenced RB (Figure 5D). These results suggest that reduced carotenoid level in the GLK2‐PRR2‐silenced RB is primarily due to the downregulation of PSY1, which plays a more dominant role in the fruit carotenogenesis compared to the downstream genes, CHY2 and ZEP.

Flavor is a key determinant of fruit quality and influences consumer preference. Different compositions of flavor volatiles have been identified during pepper fruit ripening (Forero et al., 2009); however, regulatory genes related to flavor have been less studied. Here, we identified that GLK2 and PRR2 can regulate volatile flavor volatiles in pepper fruits (Table 2). A positive correlation between sour taste and four volatile compounds (e.g., dill ether, p‐menth‐1‐en‐9‐al, α‐terpineol, and linalool) was identified in tomato fruit (D'Angelo et al., 2018). In sweet pepper, p‐menth‐1‐en‐9‐al showed a positive correlation with fruit/apple taste and sweetness (Eggink et al., 2012). Levels of dill ether and p‐menth‐1‐en‐9‐al were decreased in GLK2‐silenced NH and PRR2‐silenced HO, suggesting that GLK2 and PRR2 can influence sensory‐related flavors. Methyl salicylate, which decreases during fruit ripening in tomato (Wang et al., 2020) and pepper (Huang et al., 2023), was increased in GLK2‐silenced NH and PRR2‐silenced HO, which negatively affects consumer preference (Frick et al., 2023). trans‐2‐Heptenal, a volatile compound favored by consumers, increased 2.5‐fold during fruit ripening process in tomato (Klee & Giovannoni, 2011; Klee & Tieman, 2013). The level of trans‐2‐heptenal was significantly increased in PRR2‐silenced HO and was only detected in GLK2‐silenced NH. In pepper fruits, trans‐2‐hexenal, a marker volatile for sweet and fruity aroma, increases during fruit ripening (Mazida et al., 2005). The level of trans‐2‐hexenal was reduced in GLK2‐silenced NH, which is a consumer‐like volatile in tomato (Klee & Tieman, 2018; Tieman et al., 2017). Hexyl isopentanoate is one of significant volatile in habanero varieties, which increases during fruit ripening (Pino et al., 2006). The level of hexyl isopentanoate was only detected in HO, and significantly recued in PRR2‐silenced HO. GLK2 and PRR2 do not affect general ripening processes, since regulation of ripening‐dependent flavors was not highly coordinated in the dual‐variated fruits. It seems that their influence on fruit development may be employed through the altered plastid activity and plastid numbers, including the accumulation of carotenoids.

GLK2 physically interacts with PRR2 in Y2H (Figure 7A) and BiFC assays (Figure 7B). GLK2 (Ni et al., 2017; Rauf et al., 2013) and PRR2 (Perez et al., 2019; Wang et al., 2024) were localized in the nucleus. GLK2 and PRR2 contained three conserved domains: a DNA‐binding (helix–loop–helix) domain, a proline‐rich region (transactivation domain), and the GCT box (dimerization domain) (Rossini et al., 2001). Additionally, GLKs uniquely contain the AREAEAA motif, which lacks in PRRs (Choi et al., 2024; Fitter et al., 2002). All the variations in GLK2 and PRR2 led to the expected deletion of the GCT box in Capsicum accessions (Figure S2; Table 1). Indeed, the lack of the GCT box in GLK2 and PRR2 abolished their interaction, and their interaction in the nucleus was verified by the BiFC assay (Figure 7). The GCT box plays a crucial role in protein dimerization in GLKs (An et al., 2019; Tachibana et al., 2024). The negative regulators of chloroplast development, PSEUDO‐ETIOLATION IN LIGHT (PEL), interact with the C‐terminus of GLK1 and the zinc‐finger domain of the chloroplast‐localized chaperone PHOTOSYSTEM I ASSEMBLY 2 (PSA2), thereby sequestering both proteins in the cytoplasm and ultimately suppressing chloroplast development and pigment accumulation in rice (Choi et al., 2025). Repressor of photosynthetic genes 1 (RPGE1/PEL) interacts with the GCT box of GLK1 and inhibits its transcriptional activity by disrupting its dimerization (Kim et al., 2023). In Arabidopsis, PRR2 interacted with the calmodulin‐like protein CML9 through its GCT box, suggesting that PRR2 may form a homo‐ or hetero‐complex (Perochon et al., 2010). The green fruit accession NH was predicted to have the deletion of the GCT box in PRR2; however, the interaction between GLK2 and PRR2_ΔGCT was abolished. GLKs are recognized as master regulators of chloroplast biogenesis, enhancing the expression of photosynthesis‐associated nuclear genes across various plant species (Hernández‐Verdeja & Lundgren, 2024). GLK2 coordinates the synthesis of chlorophyll, the development of photosystems, flowering regulation, and the signaling pathways of abscisic acid, interacting with various proteins and target genes involved in these processes (Choi et al., 2024). Even in the absence of protein interaction with PRR2, it is possible that GLK2 interacts with other proteins to regulate fruit pigmentation in NH. The interaction of nuclear factor Y, subunit A genes (NF‐YA9a) with GLK2.1 activates HEMA2 expression and increases the chlorophyll biosynthesis in radish (Ying et al., 2023). GLKs transactivated PSY expression by interacting with GBFs, which bind to the promoter of PSY, and promoted carotenoid biosynthesis in Arabidopsis (Sun, Hazra, et al., 2025). The repressor of photosynthetic genes 1 (RPGE1/PEL) interacts with APRR2 in carrot and suppresses carotenoid biosynthesis by inhibiting the binding activity of APRR2 to the promoter regions of carotenoid biosynthesis genes, such as DcPSY1, DcPSY2, and lycopene ε‐cyclase (DcLCYE) (Wang et al., 2024). Additional factors that interplay with GLK2 or PRR2 could provide valuable insights for dissecting complex regulatory mechanisms of fruit pigmentation. The GLK2_ΔGCT hardly interacted with PRR2, and immature fruits harboring the glk2 variations resulted in a light green color. The contribution of CcAPRR2 in chlorophyll levels is less consistent than CaGLK2 (Borovsky et al., 2019). CaGLK2 is a predominant regulator of immature fruit color, while CaAPRR2 functioned as a secondary factor regulating chlorophyll synthesis (Sun, Zhang, et al., 2025). Considering only the PRR2 genotype in the F₂ population, the PRR2 genotype was weakly correlated with immature fruit color compared to the GLK2 genotype (Figure S3b). GLK2 and PRR2 jointly act and are necessary for fruit pigmentation; however, the light‐green phenotype of glk2 suggests that GLK2 plays a major role. Silencing of GLK2 in RB did not reduce PRR2 expression (Figure 5C,D). Similarly, the silencing of PRR2 in RB did not reduce GLK2 expression. These results suggest that each gene may act in a complementary manner to regulate fruit pigmentation through alterations of plastid activity and gene expression.

In summary, we identified the combined effects of GLK2 and PRR2 associated with immature and ripe fruit colors in pepper. Both GLK2 and PRR2 are determinants of c1 and positively control plastid development and pigment biosynthesis. These findings provide new insights into fruit pigmentation and will be useful for the genetic improvement of fruit quality.

EXPERIMENTAL PROCEDURES

Plant materials

Capsicum accessions were obtained from the National Agro biodiversity Center, Rural Development Administration, Korea (Table 1; Table S1). All plants were grown under greenhouse conditions at Kyungpook National University, Daegu, Korea. Pericarps from immature (30 days post‐anthesis) and ripe fruits (10 days after breaker) were collected, immediately frozen in liquid nitrogen, and stored at –80°C before analysis.

Virus‐induced gene silencing (VIGS)

VIGS was performed according to a previously described method (Kim, Park, et al., 2017). For plasmid construction, fragments of the target gene were amplified from pepper leaf cDNA with primers (Table S6). Agrobacterium tumefaciens strain GV3101 carrying pTRV1 and pTRV2::gene of interest was grown overnight at 28°C in 10 ml LB medium containing rifampicin (50 μg/ml) and kanamycin (50 μg/ml). Agrobacterium cultures were harvested and resuspended to an OD₆₀₀ of 0.7. Cell suspensions were incubated at room temperature with inversion for 2 h. Agrobacterium cultures containing pTRV1 and pTRV2 carrying the gene of interest were mixed at a 1:1 ratio and infiltrated into the abaxial side of both cotyledons. The infiltrated plants were incubated at 16°C in the dark for 1 day and then grown under a 20‐h day/4‐h night cycle at 20°C until harvest.

Sequencing of candidate genes

Genomic DNA was extracted using the CTAB method (Murray & Thompson, 1980) from young leaves. PCRs were performed with e‐Taq DNA polymerase (SolGent Co., Ltd., Daejeon, Korea) containing 50 ng of genomic DNA and gene‐specific primers as previously described (Lee et al., 2020), and listed in Table S6. The PCR program was 95°C for 2 min, followed by 35 cycles of 95°C for 30 sec, annealing temperatures for 30 sec, and 72°C for 1 min. The PCR amplicons were purified with the HiGene™ Gel & PCR Purification System (BIOFACT Co., Ltd., Daejeon, Korea) and sequenced (SolGent Co., Ltd.).

Segregation analysis of GLK2 and PRR2 in the F₂ population

A total of 210 F₂ individuals derived from a cross between C. annuum “Milyang” and C. annuum “Goshiki” was developed. The green immature fruit color in “Milyang” was evaluated as class 1, and the pale green in “Goshiki” was evaluated as class 4. The color of immature fruit in F₂ individuals was classified into four classes based on visual observation of green intensity and independently evaluated by three persons. The dCAPS marker for GLK2 (CA10g02900, chr 10) was designed based on the cytosine (C) to thymine (T) substitution in the third exon, with the following sequence: 5′‐GCTTTATCTAATTAAACCTAGAAAG‐3′ (GLK2 forward) and 5′‐CAGCTTGCTGCCTCAGCTTCCCTTGCTAATAAATGTTATC‐3′ (GLK2 reverse). PCR was conducted with an initial denaturation at 95°C for 3 min, followed by 35 cycles of 95°C for 30 sec, annealing at 60°C for 30 sec, and extension at 72°C for 1 min. The CAPS marker for PRR2 (CA00g25180, chr 1) was designed based on the adenine (A) to guanine (G) substitution in the first exon, with the following sequence: 5′‐GGATGTGCAGATTCAATTGGTA‐3′ (PRR2 forward) and 5′‐ CCTAGTCACATTGCAGCAACT‐3′ (PRR2 reverse). PCR conditions were 95°C for 3 min, followed by 35 cycles of 95°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min. The PCR amplicons from the GLK2 primers were digested with ClaI (Enzynomics, Daejeon, Korea), and PRR2 primers were digested with AluI (New England Biolabs, Ipswich, MA, USA) for 12 h to determine polymorphisms.

Transmission electron microscopy (TEM)

Pericarp tissues from immature and ripe fruits were analyzed following the previously described method (Lee et al., 2020). Polymerized blocks were sectioned at a thickness of 70 nm using a Leica Ultracut UCT ultramicrotome, and ultrathin sections were examined with a transmission electron microscope (Tecnai G2 Spirit Twin; FEI Company, Hillsboro, OR, USA). The chloroplast area and the number of thylakoid per granum were analyzed using ImageJ (National Institutes of Health, Bethesda, MD, USA; http://imagej.nih.gov/ij/).

Pigment extraction and high‐performance liquid chromatography (HPLC) analysis

Approximately 100 mg of frozen and ground powder from pericarps was used for pigment extraction. Extraction and quantification of chlorophylls and carotenoids were performed as previously described (Yoo et al., 2017). For saponification, 20% KOH was added to the extracts and incubated for 1 h at room temperature. Carotenoid analysis was conducted using HPLC with C30 YMC Carotenoid S‐5 (4.6 × 250 mm) and Agilent ChemStation software (Agilent Technologies, Santa Clara, CA, USA) (Kim, Yoo, et al., 2017). The concentrations of carotenoid compounds were determined using calibration curves generated from β‐carotene (Cat. No. C4582; Sigma‐Aldrich, St. Louis, MO, USA) and capsanthin (Cat. No. 19081; Sigma‐Aldrich).

Gene expression analysis

Total RNA was extracted from pericarps using Ribospin™ Seed/Fruit (GeneAll Biotechnology Co., Ltd., Seoul, Korea) and reverse transcribed to synthesize cDNA using a DiaStar™ RT Kit (SolGent Co., Ltd.). Quantitative RT‐PCR was performed using Power SYBR® Green PCR master mix (Applied Biosystems™, Foster City, CA, USA) containing 100 ng of cDNA. qRT‐PCR was conducted with an internal denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min on a CFX Connect Real‐Time PCR Detection System (Bio‐Rad, Hercules, CA, USA). At least three biological replicates and two technical replicates were performed per primer, and CaUBIQUITIN (DQ975458.1) was used for normalization. The primers used for qRT‐PCR were previously described (Lee et al., 2020) and listed in Table S6.

Volatile analysis using SPME‐GC–MS

Volatile analysis by SPME‐GC–MS was performed according to Yoo et al. (2019) and Yoo et al. (2023) with modifications. Approximately 100 mg of frozen and ground powder from pericarps was introduced into a 10 ml headspace glass vial and incubated in a water bath at 37°C for 10 min. Subsequently, 500 μl of EDTA (pH 7.5) and 1.1 g of calcium chloride dihydrate (CaCl₂·2H₂O) were added and sonicated for 5 min. All samples were analyzed within 12 h using a gas chromatograph‐mass spectrometer (7890B‐5977B GC/MSD; Agilent, Santa Clara, CA, USA) at the Kyungpook National University Instrumental Analysis Center (Daegu, Korea). Sample‐wise peaks were identified and matched to the library using the MSD ChemStation Data Analysis software (Agilent). The experimental MS spectra of the volatiles were tentatively compared with the mass spectral library of the Wiley Registry (11th Edition/NIST 2017) to identify the volatiles.

Yeast two‐hybrid assay

The full‐length cDNA of GLK2 was inserted into the pGBKT7 vector, and the full‐length cDNA of PRR2 was inserted into the pGADT7 vector (Clontech, Mountain View, CA, USA) using specific primers (Table S6). The truncated sequence of GLK2 (1–241 a.a.) and PRR2 (1–477 a.a.) were inserted into pGBKT7 and pGADT7 vectors using specific primers (Table S6). The recombinant constructs were transformed into yeast strain AH109 using lithium chloride‐polyethylene glycol (PEG) as described. The transformed yeast cells were grown for 3 days at 30°C on synthetic defined (SD) medium lacking leucine and tryptophan (SD‐LW) and then replicated on to SD medium lacking leucine, tryptophan, and histidine (SD‐LWH) containing 5 mm 3‐amino‐1,2,3‐trizole (3‐AT). The co‐transformation of pGBKT7 and pGADT7 was used as a negative control, while that of pGBKT7‐p53 (murine p53) and pGADT7‐T (T‐antigen) was used as positive control.

Bimolecular fluorescence complementation assay

The full‐length of GLK2 sequence was cloned into pGA3816 vector containing the N‐terminal of Venus (1–172 a.a.) to generate GLK2_nYFP, whereas PRR2 sequence was cloned into pGA3815 vector containing the C‐terminal of Venus (155–238 a.a.) to generate PRR2_cYFP. The BiFC was conducted by transforming protoplasts isolated from rice root‐derived callus suspension cells as previously described (Cho et al., 2024). Fluorescence was observed using a confocal laser scanning microscope (LSM780; Carl Zeiss, Jena, Germany) under a bright field or green fluorescent protein (GFP) channel. Primers used for BiFC were listed in Table S6.

AUTHOR CONTRIBUTIONS

JML designed the research. SL, GL, HJY, and LHC performed the experiments. SL and JML analyzed the data and wrote the manuscript. All the authors read and approved the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Supporting information

Figure S1. Sequence alignments of PRR2 (a), GLK2 (b), LOL1 (c), CaPP2C35 (d) in NH, HO, and HP.

Figure S2. Schematic diagrams of the domain structure of GLK2 and PRR2. The phospho‐accepting receiver‐like domain (REC, red box), the GARP DNA‐binding domain (DBD, orange box), the AREAEAA motif (AREAEAA, yellow box), the proline‐rich region (PRD, green box), and the GLK/C‐terminal box (GCT box, blue box) are indicated. The amino acid changes resulting from each mutation in GLK2 (a) and PRR2 (b) are depicted. The numbers next to each schematic represent the corresponding amino acid size.

Figure S3. Relationships between immature fruit colors and the genotypes of GLK2 and PRR2 in the F₂ individuals. (a) Immature fruit colors harboring the allele combinations of GLK2. (b) Immature fruit colors harboring the allele combinations of PRR2. The Y‐axis represents color intensity in the F₂ population. The X‐axis represents allele at GLK2 (GLK2 and glk2) and PRR2 (PRR2 and prr2). Different letters indicate significant differences by one‐way ANOVA followed by Duncan's multiple test (P < 0.05).

Figure S4. Spatial regulation of pigmentation in GLK2 and PRR2. (a) Comparison of transcripts in leaf, root, and fruit (six stages) of pepper (C. annuum “CM334”) (Kim et al., 2014) and tomato (S. lycopersicum “Heinz 1706”) (Sato et al., 2012). MG; mature green, B; breaker. (b) Immature fruits of pepper and tomato.

Figure S5. Gene structures and variations of GLK2, PRR2, PSY1, and CCS in Capsicum accessions. The gray box indicates the expected amino acid size resulting from a premature stop codon. (a) Nonsense mutations resulting from the C‐to‐T substitution at the 436th nt and 445th nt from the start codon are designated as glk2‐A and glk2‐B, respectively. A nonsense mutation resulting from the G‐to‐A substitution at the 486th nt is designated as glk2‐C. A frame‐shift mutation resulting from the 115 bp deletion at the 568th nt is designated as glk2‐D. A frame‐shift mutation resulting from the G‐to‐C substitution at the splice acceptor site in the 4th intron is designated as glk2‐E. (b) A frame‐shift mutation resulting from the 41 bp deletion in the junction area of the 3rd exon and intron is designated as prr2‐A. A nonsense mutation resulting from the C‐to‐T substitution at the 1171st nt is designated as prr2‐B. A nonsense mutation resulting from the G‐to‐A substitution at the 1428th nt is designated as prr2‐C. (c) A frame‐shift mutation resulting from the 1 bp deletion at the 120th nt is designated as psy1‐A. A frame‐shift mutation resulting from A‐to‐T substitution at the splice acceptor site in the 6th exon is designated as psy1‐B. A large deletion including both genic and promoter regions is designated as psy1‐C. (d) A frame‐shift mutation resulting from the 1 bp insertion at the 1024th nt is designated as ccs‐A. A large deletion including promoter to the coding region is designated as ccs‐B.

Figure S6. Gene silencing of GLK1 and GLK2 in leaves. (a) GLK1 silencing causes leaf color changes. (b) Pigment contents were reduced only in GLK1‐silenced leaves. Significant differences were determined by Student's t‐test (**P < 0.01). (c) Silencing of GLK2 or PRR2 shows no color changes in the leaf.

Table S1. Diverse fruit color depending on the variations of GLK2 and PRR2 in Capsicum accessions.

Table S2. Chlorophyll contents of immature fruits in the Capsicum accessions.

Table S3. Volatile compounds detected in GLK2‐silenced NH and the control by SPME‐GC–MS.

Table S4. Volatile compounds detected in PRR2‐silenced HO and the control by SPME‐GC–MS.

Table S5. Carotenoid contents in the ripe fruits of GLK2 and/or PRR2‐silenced RB.

Table S6. Primers used in this study.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the supplementary material of this article.