Mapping of CaPP2C35 involved in the formation of light-green immature pepper (Capsicum annuum L.) fruits via GWAS and BSA

Genome-wide association study, bulked segregant analysis, and genetic analysis delimited the LG locus controlling light-green immature pepper fruits into a 35.07 kbp region on chromosome 10. A strong candidate gene, CaPP2C35, was identified in this region. In pepper (Capsicum annuum L.), the common colors of immature fruits are yellowish white, milky yellow, green, purple, and purplish black. Genes related to dark green, white, and purple immature fruits have been cloned; however, only a few studies have investigated light-green immature fruits. Here, we performed a genetic study using light-green (17C827) and green (17C658) immature fruits. The light-green color of immature fruits was controlled by a single locus-dominant genetic trait compared with the green color of immature fruits. We also performed a genome-wide association study and bulked segregant analysis of immature-fruit color and mapped the LG locus to a 35.07 kbp region on chromosome 10. Only one gene, Capana10g001710, was found in this region. A G-A substitution occurred at the 313th base of the Capana10g001710 coding sequence in 17C827, resulting in the conversion of the α-helix of its encoded PP2C35 protein into a β-fold. The expression of Capana10g001710 (termed CaPP2C35) in 17C827 was significantly higher than in 17C658. Silencing CaPP2C35 in 17C827 resulted in an increase in chlorophyll content in the exocarp and the appearance of green stripes on the surface of the fruit. These results indicate that CaPP2C35 may be involved in the formation of light-green immature fruits by regulating the accumulation of chlorophyll content in the exocarp. Thus, these findings lay the foundation for further studies and genetic improvement of immature-fruit color in pepper.


Introduction
Fruit color is one of the most important traits of pepper (Capsicum annuum L.) appearance, as it is the main reference standard used by consumers to select and purchase peppers ). Thus, research on pepper fruit color has practical significance for the development of economically efficient strategies of pepper cultivation. Pepper fruit color can be divided into two categories depending on the developmental stage of the plant: immature and mature (Wahyuni et al. 2011). Most immature fruits are green because of the accumulation of chlorophyll in their chloroplasts (Paran et al. 2007). Nevertheless, many other pigments, such as lutein, β-carotene, violaxanthin, and anthocyanins, have been found in immature fruits, leading to fruit colors such as milky yellow, yellowish green, purple, and purplish black (Matsufuji et al. 2007;Lightbourn et al. 2008;Liu et al. 2020a, b). As the fruit ripens, the chlorophyll is gradually degraded, and the chloroplasts are converted into chromoplasts in which carotenoids are synthesized and stored; this process results in orange, red, and yellow mature pepper fruits (Bouvier et al. 1994;Hugueney et al. 1996;Sun et al. 2018;Jeong et al. 2020;Lee et al. 2021).
Previous studies have shown that the genetics of white, yellowish-green, and deep-green immature pepper fruit colors are mainly controlled by the SW1, SW2, and SW3 loci, respectively. SW3 is dominant relative to SW1 and SW2, whereas SW2 is dominant relative to SW1 (Lightbourn et al. 2008;Stommel et al. 2014). The dominance of these loci 1 3 indicates that there are genetic differences among different colors of immature pepper fruits. However, there have been only a few studies on immature-fruit color in pepper plants. Brand et al. (2012) identified two major quantitative trait loci (QTLs): pc8.1 (also termed pc1, as subsequent mapping placed the QTL in chromosome 1 [Borovsky et al. 2019]) and pc10.1, which control the degree of dark green color saturation in immature fruits based on their chlorophyll content. In 2013, Pan et al. (2013) cloned the APPR2-Like gene in pepper, which positively regulates chlorophyll synthesis in pepper, and found that APPR2-Like and pc8.1 were in the same region on the genetic map, indicating that APPR2-Like may be a candidate gene of pc8.1. However, Borovsky et al. (2019) used bulked segregant analysis (BSA) to find CcLOL1, a candidate gene of pc1 (pc8.1), and observed that CcLOL1 affected the formation of green-colored immature fruits by regulating the expression of genes related to photosynthesis and redox reactions. Brand et al. (2014) found that a possible candidate gene of pc10.1 is CaGLK2. In addition, two genes (CaMYB A and Ca3GT) have been found to regulate anthocyanin accumulation in purple immature pepper fruits (Borovsky et al. 2004;Liu et al. 2020a, b).
The molecular mechanism underlying the formation of green-colored immature pepper fruits is complex. Sequence analysis of the open reading frame (ORF) of the CaGLK2 in light-green immature pepper fruits has shown that six base insertions, one A-G substitution, and three base deletions occurred at three different positions of its third exon, indicating that there was diversity in the variation of CaGLK2. However, there was no variation in CaGLK2 ORF in "1901," "21-1," "Nayoi," and "1202" pepper accessions, which produce light-green immature fruits, indicating that CaGLK2 may be down-regulated or that other genes regulate immature-fruit color in these accessions (Brand et al. 2014). Borovsky et al. (2019) found that there was no difference in the ORF sequence of CcLOL1 between C. annuum and C. annuum var glabriusculum, which have significant differences in chlorophyll content, indicating that CcLOL1 has a limited ability to regulate immature-fruit color variation in pepper plants.
Although some studies have isolated genes related to the color of immature pepper fruits (Borovsky et al. 2004(Borovsky et al. , 2019Pan et al. 2013;Brand et al. 2014;Liu et al. 2020a, b), only a few have reported the regulatory genes of lightgreen immature fruits. Among all pepper cultivars, there is a large variation in the hue of green-colored immature fruits, ranging from colors characterized by very low chlorophyll content, such as white or milky yellow, to dark green (Mejia et al. 1988;Levy et al. 1995;Hornero-Mendez et al. 2000;Wall et al. 2001). However, it is difficult to distinguish the different depths of green-colored fruits, especially that of yellowish green, milky yellow, light green, green, deep green, and dark green fruits. This difficulty poses a challenge in conducting research on green-colored immature pepper fruits. Some studies have investigated pepper fruit color based on the pigment content (Brand et al. 2012;Borovsky et al. 2019). However, this method is tedious and easily affected by the sampling or determination technique.
Overall, the genetic and regulatory mechanisms of greencolored immature pepper fruits with different depths of saturation are complex, and there have been few studies on the genetics and mapping of genes controlling light-green immature fruits. Considering this, we analyzed the heredity and formation of light-green immature pepper fruits in the present study with two C. annuum accessions, 17C827 (light-green immature fruits; red mature fruits) and 17C658 (green immature fruits; yellow mature fruits). Additionally, we utilized a genome-wide association study (GWAS) and bulked segregant analysis (BSA) to perform gene mapping and functional analysis of the LG locus that controls lightgreen coloration of immature pepper fruits-the LG locus was identified based on visual and colorimetric analyses of the color phenotype. Finally, a molecular marker closely linked to the LG locus was developed. This study provides a valuable reference for future studies on and genetic improvement of the color of immature pepper fruits.

Plant materials
In this study, we used Capsicum annuum accessions 17C827 (light-green immature fruits; red mature fruits) and 17C658 (green immature fruits; yellow mature fruits) to generate F 1 and F 2 populations (941 individuals) to map the LG locus, which controls the light-green color in immature pepper fruits. In addition, 287 pepper accessions (Supplementary  Table 1) were used for the genome-wide association study (GWAS) (Wu et al. 2019). All the above-mentioned materials were provided by the Pepper Research Laboratory in the College of Horticulture, China Agricultural University, and were grown in greenhouses at ShangZhuang Experimental Station of China Agricultural University (Beijing, China).

Determination of pigment content
Exocarp and endocarp (including mesocarp) of the immature pepper fruits in each accession were collected and stored at −80 °C. One gram of sample was used for pigment extraction. Twenty-five milliliters of 95% ethanol was added into a 50-mL centrifuge tube containing the sample, and then the tube was sealed with a sealing film and placed under dark conditions for 36 h after ultrasonic (100 W; 26 °C) treatment for 1 h (Zhu et al. 2017;Vendruscolo et al. 2021). When the sample was whitened, the solution was mixed evenly 1 3 and the supernatant was retained. Measurement was taken using a spectrophotometer at 470, 649, and 665 nm, respectively (Amorim-Carrilho et al. 2014). Finally, the content of chlorophyll and carotenoid was calculated using the Aron formula method (Aron et al. 1949).

Chloroplast observation
The exocarp portion of pepper fruits was cut into a sample strip (1 mm × 2 mm × 5 mm) and then immersed in 2.5% glutaraldehyde fixed solution and stored at 4℃ overnight. The next day, ultra-thin sections were prepared according to the slice production process of Jeong et al. (2020). Finally, chloroplasts were observed and photographed using the JEM-1400 Flash (JEOL, Tokyo, Japan) (120 kV) transmission electron microscope.

Color measurement of the immature pepper fruit
The color of immature fruits of the F 2 individuals and the 287 pepper accession was classified using standard colorimetric cards (The Royal Horticultural Society's Colour Chart) and measured using a colorimeter (Konica Minolta Chroma Meter CR-400). Chroma (SD) and shade (SG) were calculated with the formulae described by Yang et al. (2004). For each parameter, the values for each plant were collected from three fruits; for each fruit, the value was the average of three measurements.

Nucleic acid extraction
Genomic DNA of each individual (or pepper accession) was isolated from six fresh leaves using the cetyltrimethylammonium bromide (CTAB) method (Lee et al. 2017). DNA concentration and quality were determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). DNA sample of concentration ≥ 20 ng/µL and volume ≥ 30 µL was used for sequencing (Wu et al. 2019). DNA of 941 F 2 individuals was genotyped for mapping the LG locus. Furthermore, 30 individuals with light-green immature fruits and 30 individuals with green immature fruit were selected from 941 F 2 individuals. Equal amounts of DNA from each individual were mixed to generate the green pool and light-green pool for the BSA. For each pool, the final concentration was 40 ng/µL.
The total RNA was extracted from various tissues. For 17C827 and 17C658, RNA was extracted from the stems, leaves, flowers, and exocarp at three stages (9, 12, and 20 days after anthesis). For the virus-inoculated 17C827 used in the VIGS analysis, RNA was extracted from lightgreen exocarp and green exocarp of immature fruits. For the control plants used in the VIGS analysis, RNA was also extracted from the exocarp of immature fruits. The total RNA was extracted using the MG RNAzol kit (MGmed, Seoul, South Korea) according to the manufacturer's instructions. Complementary DNA (cDNA) was synthesized from 2 µg of RNA using the EasyScript Reverse Transcriptase kit (TransGen, Beijing, China) with oligo (dT) primers. The resulting cDNA was used for further analyses.

GWAS and BSA methods
The genome of "Zunla-1" (http:// peppe rsequ ence. genom ics. cn/ page/ speci es/ index. jsp, version 2.0) was selected as the reference genome for an electronic digestion prediction experiment. The genome size of "Zunla-1" was ~ 3.36 Gb, and the GC content was 34.97% (Qin et al. 2014). We used the sequencing data of Wu et al. (2019) to perform the GWAS of immature-fruit color (including yellowish white, milky yellow, yellowish green, light green, green, deep green, dark green, purple, and purplish black) in the 287 pepper accessions. The detailed methods of GWAS have been described by Wu et al. (2019). The BSA was conducted according to the method of Xu et al. (2016).

Genetic analysis, linkage map construction, and QTL analysis
To determine the genetics of fruit color, Chi-squared (χ 2 ) analysis was performed to test the phenotypic data for goodness-of-fit to Mendelian segregation ratios. A large number of simple sequence repeat (SSR), insertion-deletion (InDel), cleaved amplified polymorphic sequence (CAPS) and derived cleaved amplified polymorphic sequence (dCAPS) markers within the candidate interval of LG were screened, and seven polymorphic molecular markers were obtained (Supplementary Table 2). Using the phenotypic and genotypic data of the 941 F 2 individuals, the linkage analysis of LG was carried out using JoinMap4 software, and a genetic linkage map was constructed with Mapmaker/EXP 3.0 command (LOD value ≥ 3.0) (Lander et al. 2009;). In addition, to quickly and efficiently validate the linkage mapping analysis results of LG, 351 F 2 individuals were randomly selected from 941 F 2 individuals; L *, a *, b *, SD, and SG values of immature fruits were measured; and the results were used for QTL analysis. Interval-mapping QTL analysis of the 315 F 2 individuals was performed for the L *, a *, b *, SD, and SG values of immature fruits using QTL IciMapping 4.0 software (Meng et al. 2015). Significance threshold level (LOD 3.0) for QTL detection was computed using permutation tests with 1,000 iterations at P < 0.01.
Polymerase chain reaction (PCR) was conducted using a 10 µL system containing 1.0 µL of DNA template, 0.5 µL of forward primer, 0.5 µL of reverse primer, 5.0 µL of Taq polymerase mix (Beijing ComWin Biotech Co., Ltd., Beijing, China), and 3 µL of ddH 2 O. The thermal cycle was as follows: initial denaturation at 94 °C for 5 min; 35 cycles of denaturation at 94 °C for 30 s, annealing at 58 °C for 35 s, and extension at 72 °C for 30 s, before a final extension at 72 °C for 5 min. HinfI was used to digest the PCR product of the dCAP10-1 marker. The digested PCR product was separated by 7% polyacrylamide gel electrophoresis (Wu et al. 2019).

Gene expression and bioinformatic analysis
The expression of CaPP2C35, GSA, HEMA1, CHLD, CHLH, CAO, CHLG, CRD1, DVR, SGR1, PAO, NYC, RCCR , PPH, and SGR2 was measured by real-time PCR; the primers used are listed in Supplementary Table 2. The samples used for gene expression analysis were the exocarp of fruit; the reference gene was UBI (AY486137.1). Relative gene expression levels were obtained using the 2 −ΔΔCT method, with three biological and three technical repeats. Data are means of three biological replicates. RT-PCR, real-time PCR, and PCR product sequencing were performed as described by Wang et al. (2019). DNAMAN (Lynnon Biosoft, USA) and MEGA7 were used to analyze the sequence difference or evolution of related genes or proteins (Kumar et al. 2016). SOPMA (https:// npsa-prabi. ibcp. fr/ cgi-bin/ npsa_ autom at. pl? page= npsa_ sopma. html) and I-TASSER (https:// zhang lab. ccmb. med. umich. edu/I-TASSER/) software were used to analyze the primary and secondary and tertiary structures of CaPP2C35 protein (Ballut et al. 2015;Roy et al. 2010;Yang et al. 2015).

Virus-induced gene silencing of CaPP2C35
To study the function of CaPP2C35, virus-induced gene silencing (VIGS) was carried out in 17C827 according to Cheng et al. (2018) with some modifications. A fragment of CaPP2C35 coding sequence (CDS) was selected as the target, and primers were designed using Primer 5 software (Supplementary Table 2). The constructs consisting of pTRV1, pTRV2, and pTRV2-PDS, pTRV2-CaPP2C35 were transformed into Agrobacterium tumefaciens GV3101, respectively. The transformed Agrobacterium was cultured for 12 h in Luria-Bertani (LB) medium (20 µM acetosyringone, 50 mg L −1 kanamycin, 25 mg L −1 Rif and 10 mM 4-Morpholine Ethane Sulfonic acid [MES]) and then harvested and suspended in MS buffer (200 µM acetosyringone, 10 mM MES, and 10 mM MgCl 2 ; pH = 5.6) to a final OD 600 of approximately 1.0. A mixture of cultures containing 1:1 (v/v) of pTRV1 and pTRV2 was used as tobacco rattle virus (TRV) control; pTRV1 and pTRV2-PDS were used as reporters, and pTRV1 and pTRV2-CaPP2C35 were used to silence CaPP2C35. The effect of virus-induced gene silencing (VIGS) was determined according to Cheng et al. (2018).

Low chlorophyll content in the exocarp of 17C827 affects light-green immature-fruit formation
We examined the cross section of 17C658 (green immature fruits) and 17C827 (light-green immature fruits) immature fruits. As shown in Fig. 1a, both 17C827 and 17C658 immature fruits formed a notable distinction of green coloration at the boundary between the exocarp and mesocarp. The green portion was on the exocarp side, and the light-green portion was on the mesocarp side. Thus, the degree of green in the 17C658 and 17C827 immature fruits is mainly influenced by the exocarp.
Next, the pigment content in the exocarp and endocarp (including the mesocarp) of the two pepper accessions was compared. We found that the chlorophyll content (chlorophyll a [ Fig. 1b] and chlorophyll b [ Fig. 1c]) in the exocarp was significantly different between 17C827 and 17C658 fruits, and the content in 17C827 fruits was lower than in 17C658 fruits from day 12 to M stage. However, there was no significant difference in chlorophyll content in the endocarp (including the mesocarp) between 17C827 and 17C658 fruits at various periods of fruit development (Fig. 1b, c). These results showed that low chlorophyll content in the exocarp of 17C827 fruit was the main factor influencing the formation of light-green immature fruits. Furthermore, the chlorophyll content (chlorophyll a [ Fig. 1b] and chlorophyll b [ Fig. 1c]) in the exocarp and endocarp (including the mesocarp) of the two pepper accessions decreased from day 5 to 9, increased from day 9 to 12, and gradually decreased and eventually stabilized from day 12 to M stage. The change in the carotenoid content was similar to that in the chlorophyll content ( Supplementary Fig. 1). The results showed that the degree of green hue in the immature fruits of 17C658 and 17C827 began to stabilize after day 12. Meanwhile, the total chlorophyll content (0.048-0.382 mg/g·FW) in both 17C827 and 17C658 fruits was considerably higher than the carotenoid content (0.003-0.047 mg/g·FW) during various periods of fruit development ( Supplementary Fig. 1), which indicates that the formation of green-colored immature pepper fruits is mainly affected by chlorophyll.

Abnormal chloroplast thylakoids in the exocarp of 17C827 fruits affects light-green immature-fruit formation
To explore whether light-green immature-fruit formation by 17C827 plants is affected by chloroplasts, we first examined the number of chloroplasts in a single cell and the number of chloroplast grana in a single chloroplast. Irrespective of the evaluation day-day 9 or day 20-there was no significant difference in the number of chloroplasts and chloroplast grana between 17C827 and 17C658 fruits (Fig. 2a). We then measured the perimeter and area of the chloroplast cross section in the exocarp and found that there were significant differences between 17C827 and 17C658 fruits (Fig. 2b, c). On day 9, the chloroplast cross-section perimeter and area of 17C827 fruits were higher than those of 17C658 fruits. However, the contrary was true on day 20. We calculated the increments in the chloroplast section perimeter and area from days 9 to 20 and found that the increments in these values in 17C658 were higher than that in 17C827 (Fig. 2b, c). On day 9, dividing chloroplasts were found in both 17C827 and 17C658 fruits, indicating that both accessions were able to proliferate chloroplasts ( Supplementary Fig. 2a). On day 20, we also found that the chloroplast grana of the mature chloroplasts in light-green immature fruits of 17C827 were thinner and more scattered than those of 17C658, and there were fewer stroma lamellae in 17C827 than in 17C658 ( Supplementary Fig. 2b). The above results indicate that both chloroplast size and their inner structural density in the exocarp of 17C827 affected the formation of light-green immature fruits.
Light-green color of 17C827 immature fruits is regulated by a single dominant locus.
In the present study, F 1 and F 2 individuals were obtained from a cross between 17C827 (light-green immature fruits) and 17C658 (green immature fruits). All F 1 individuals produced light-green immature fruits, similar to the female parent 17C827 (Supplementary Fig. 3). In contrast, the immature-fruit colors of the 941 F 2 individuals were divided into light green (726 individuals) and green (215 individuals), at a ratio of 3:1 (χ 2 = 2.32 < 6.64, df = 1, p = 0.01), which was similar to the separation ratio of red (701 individuals) and yellow (240 individuals) mature fruits. These results indicate that the formation of light-green immature fruits could be attributed to a single dominant locus compared with the formation of green immature fruits. Finally, we found that F 2 could be segregated according to fruit color, light-green immature-fruit and red mature-fruit (534 individuals), lightgreen immature-fruit and yellow mature-fruit (174 individuals), green immature-fruit and red mature-fruit (167 individuals), and green immature-fruit and yellow mature-fruit (66 individuals) color, at the ratio of 9:3:3:1 (χ 2 = 1.45 < 11.35,

LG locus position revealed by GWAS and BSA
To efficiently map the LG locus, a GWAS of immature-fruit color, comprising yellowish white, milky yellow, yellowish green, light green, green, deep green, purple, and purplish black, in 287 pepper accessions was conducted (Supplementary Table 1). As shown in Fig. 3a, there was a candidate region spanning from 157,194,038 to 190,675,776 bp on chromosome 10 that was significantly associated with immature-fruit color. Additionally, a GWAS of immaturefruit colors according to lightness (L*), red and green coordinates (a*), and yellow and blue coordinates (b*) of 287 pepper accessions was conducted, and a candidate region was found on chromosome 10. The candidate region significantly associated with L* spanned from 176,498,851 to 176,669,957 bp on chromosome 10, whereas that associated with a* spanned from 55,687,612 to 62,184,580 bp, and that associated with b* spanned from 175,987,065 to 186,812,911 bp. Therefore, we assumed that the LG locus controlling the color of light-green immature fruits was likely to be located in the region spanning from 157,194,038 to 190,675,776 bp on chromosome 10. In addition, the BSA of the extreme mixed pool based on F 2 individuals showed that there was a candidate region from 156,650,000 to 178,330,000 bp on chromosome 10 that was significantly associated with the light-green color of immature pepper fruits (Fig. 3b). Finally, we selected the intersection of the GWAS candidate region and the BSA candidate region to determine whether the LG locus was located in the region spanning from 157,494,038 to 178,330,000 bp (size: 20.8 Mbp) on chromosome 10. Seven primer pairs from the above region were used for linkage analysis of the LG locus. The results showed that the LG locus was located between the InDel-106 and InDel-109 primers, with a physical distance of 35.07 kbp (from 174,787,932 to 174,823,007 bp) (Supplementary Fig. 4a). The QTL analysis for the colorimetry indices (L*, a*, b*, chroma [SD], and shade [SG]) of the immature-fruit surfaces of 315 F 2 individuals showed similar results ( Supplementary Fig. 4b).

CaPP2C35 is the candidate gene of the LG locus
Using the JBrowser tool of the Sol Genomics Network (https:// solge nomics. net/), we found that there was only one gene, Capana10g001710, located between 174,787,932 and 174,823,007 bp on chromosome 10. The sequence analysis revealed a G-A substitution at the 313th base of the Capana10g001710 CDS in 17C827 ( Supplementary Fig. 5). Based on this base substitution, a dCAPS molecular marker, dCAPS10-1, was designed. As shown in Fig. 4a and b, dCAPS10-1 had a stable polymorphism between the parents and was closely linked to LG (0.2 cM). The QTL analysis of the colorimetry indices are the increments of the chloroplast cross-section perimeter and area from 9 to 20D, respectively. 9D and 20D were 9 days after anthesis and 20 days after anthesis, respectively. '**' indicates significant differences at 0.01 level. Data are means of three independent replicates (L*, a*, b*, SD, and SG) of the immature-fruit surfaces of 315 F 2 individuals showed that the QTLs of L*, a *, b *, SD, and SG were all located near dCAPS10-1 (Fig. 4c). Additionally, the percentage of variation explained (PVE%) by dCAPS10-1 for L*, a *, b *, SD,and SG was 27.35,21.43,20.42,17.76,and 18.14%, respectively. Furthermore, the expression of Capana10g001710 in 17C827 with light-green pools was significantly higher than that of 17C658 with green pools (Fig. 4d, e). Finally, we analyzed the structure of the protein encoded by Capana10g001710 and found that a V-I substitution occurred at the 105th amino acid, which was also the last amino acid in the PP2C35 domain ( Supplementary  Fig. 6a, b). This substitution resulted in the conversion of the α-helix of PP2C35 into a β-fold in 17C827 (Supplementary Fig. 6c, d). Hence, Capana10g001710 was named CaPP2C35. The evolutionary analysis showed that PP2C35 belonged to the C subgroup of the PP2C protein family ( Supplementary Fig. 7).

Silencing of CaPP2C35 leads to green stripes on the surface of 17C827 immature fruits
To further explore the role of CaPP2C35, we silenced this gene in 17C827 (light-green immature fruits) via TRVmediated VIGS. The immature fruits of 17C827 plants subjected to TRV2-CaPP2C35 silencing continued to have green stripes on their surfaces that were unable to transition into the light-green surface of the fruits (Fig. 5). Additionally, the chlorophyll content in the green and light-green exocarps was determined in fruits treated with TRV2-CaPP2C35. The results showed that the chlorophyll content in the green exocarp was 0.248 mg/g·FW and that in the light-green exocarp was 0.095 mg/g·FW; thus, there was a significant difference between them. In addition, the green and light-green parts of the exocarp of CaPP2C35silenced fruits were sampled for CaPP2C35 expression analysis. The exocarp of plants subjected to no treatment or TRV2 treatment was used as negative controls, and the exocarp of plants treated with TRV2-PDS was used as a positive control. The results revealed that the expression of CaPP2C35 in the light-green exocarp of fruits treated with TRV2-CaPP2C35 was not significantly different from that of the control treatments (Fig. 6a). However, the expression of CaPP2C35 in the green exocarp of fruits treated with TRV2-CaPP2C35 was significantly lower than that in the light-green exocarp of the same fruits. In addition, the tissue-specific expression pattern of CaPP2C35 shows that it was specifically and highly expressed in the fruit tissues of 17C827 ( Supplementary Fig. 8). These results demonstrate that CaPP2C35 plays a role in the formation of light-green immature pepper fruits.

Silencing of CaPP2C35 leads to the induction of biosynthetic chlorophyll genes
To investigate the effect of CaPP2C35 down-regulation on the expression of genes involved in chlorophyll metabolism, we analyzed the expression of key chlorophyll metabolismrelated genes in the green and light-green portions of the same fruits treated with TRV2-CaPP2C35. As shown in Fig. 6b, the expression of the chlorophyll biosynthesis metabolism-related genes-GSA, HEMA1, CHLD, and CAO-in the green portion of the exocarp was significantly higher than that in the light-green portion. However, there was no significant difference in the expression of genes involved in chlorophyll degradation (Fig. 6c). In addition, we analyzed the expression of chlorophyll metabolism-related genes between 17C827 and 17C658 ( Supplementary Fig. 9). The results showed that there was a significant difference in the expression of chlorophyll biosynthesis metabolismrelated genes between 17C827 and 17C658 on day 9 or 12, including the expression of GSA, HEMA1, CHLD, CHLH, CAO, CHLG, and CRD1, which were expressed at lower levels in 17C827 than in 17C658. Moreover, there was a significant difference in the expression of the chlorophyll degradation-related genes-SGR1, PAO, RCCR , and SGR2between 17C827 and 17C658 on day 9 or 12. In addition, the expression of some of them (RCCR and SGR2) was higher in 17C827 than in 17C658 on day 12. Finally, the coexpression analysis of CaPP2C35 and the key chlorophyll metabolism-related genes showed that the expression levels of the genes (HEMA1, CHLH, CAO, CRD1, and CHLD) involved in the chlorophyll biosynthetic pathway were negatively associated with the level of CaPP2C35. In contrast, the expression level of SGR1, involved in chlorophyll degradation, was negatively correlated to that of CaPP2C35, whereas the level of SGR2 was positively correlated to that of CaPP2C35 ( Fig. 7 and Supplementary Fig. 10). These results indicate that CaPP2C35 may participate in the formation of light-green immature fruits in 17C827 by affecting the accumulation of chlorophyll in the exocarp.

dCAPS10-1 can be used in assisted selection for light-green immature pepper fruits
We evaluated the effectiveness of dCAPS10-1 application in the F 2 population (Supplementary Fig. 11a). The results showed that the phenotypic matching rate of the dCAPS10-1 marker in the F 2 population was 95.92%. In addition, dCAPS10-1 was tested using 126 pepper accessions (Supplementary Table 3). The phenotypic matching rate of the 126 pepper accessions was 75.40%, of which the phenotypic matching rate was 79.37% in 63 light-green immature fruit accessions and 71.43% in 63 green immature fruit accessions. Furthermore, we found that there was 1 3 no G-A substitution at the 828th base of the CaPP2C35 sequence (313th base of the CaPP2C35 CDS) in some light-green pepper accessions, such as 17C632 and 17C965 ( Supplementary Fig. 11b). Overall, the results demonstrate that dCAPS10-1 could provide molecular assistance for the breeding of pepper varieties with characteristic greencolored immature fruits within a certain range.

Discussion
Based on the results of previous genetic studies on white, yellowish-green, and deep-green immature fruits of pepper (Lightbourn et al. 2008;Borovsky et al. 2004;Liu et al. 2020a, b), it can be inferred that more saturated immaturefruit colors are likely to dominate those that are less saturated. However, in our study, the light-green immature-fruit color was associated with a single dominant locus compared with the green immature-fruit color; this result enriched the genetic theory of immature-fruit color in pepper. To further elucidate the underlying mechanism, we performed genetic mapping and preliminary functional analysis of a candidate gene involved in the formation of light-green immature-fruit color in pepper.
Considering the complexity of immature-fruit color genetics in green-colored immature fruits, Brand et al. (2012) studied the genetics between dark green and lightgreen immature fruits based on chlorophyll content and found two related major QTLs (pc8.1 and pc10.1). Here, immature-fruit color was identified using both visual observation and colorimetry. We verified that the QTL positions of L*, a*, b*, SD, and SG were consistent with the linkage analysis results of light-green immature fruits (LG) that were identified visually, which showed the reliability of the results. There was only one gene, Capana10g001710 (also termed CaPP2C35 because it encodes a PP2C35 protein), in the candidate region, and its physical distance from the genes controlling dark green coloration of immature fruits (CaGLK2) and purple coloration (CaMYB A ) was 166.45 and 8.16 Mbp, respectively (Brand et al. 2014;Borovsky et al. 2004). The expression of CaPP2C35 in 17C827 with lightgreen pools was significantly higher than that of 17C658 with green pools, which confirmed that CaPP2C35 is the candidate gene for the LG locus. Interestingly, we also found that there is a G-A substitution at the 313th base of the Cap-ana10g001710 CDS in 17C827, which potentially changes the structure of PP2C35. This raised the possibility that the light-green phenotype results from the mutation or expression or both, which will be studied in the future.
Protein phosphatase 2C (PP2C) is attributed to serine/ threonine-protein phosphatases (STPs). In plants, PP2C is widely involved in various signaling pathways, such as those involving abscisic acid (ABA), jasmonic acid, and salicylic acid (Umezawa et al. 2009;Schweighofer et al. 2007;Manohar et al. 2017), which further regulates plant growth and development indirectly, including root growth (Song et al. 2008), cell amplification (Ren et al. 2018), stem cell identity (Yu et al. 2003), organ development (Stone et al.1998), and pigment metabolism (Yusuf et al. 2008). As a regulator of most signaling pathways, PP2C binds directly to kinases (Schweighofer et al.2004), regulatory proteins (Lee et al. 2007), or even directly binds to DNA (Luis Carrassco et al. 2014) to regulate gene expression. Thus, PP2C is variably involved in plant growth and development and stress responses. For example, in tomato, SlPP2C3 has been identified as a regulator of the ABA signaling pathway, thus indirectly affecting fruit color-change and maturation (Liang et al. 2021). FaABI1 (PP2C) functions similar to SlPP2C3 (Jia et al. 2013), during strawberry fruit maturation. Here, we silenced CaPP2C35 in the light-green immature fruits of 17C827 to further analyze its function (Fig. 5). Unlike control immature-fruits, there were green stripes on the surface of the TRV2-CaPP2C35-inoculated immature-fruits. Based on the results of the gene expression analysis, we speculate that the up-regulation of CaPP2C35 in 17C827 promoted the degradation and inhibited the synthesis of chlorophyll, thus reducing the accumulation of chlorophyll in 17C827 ( Supplementary Fig. 12).
We determined the chlorophyll content in light-green and green immature fruits, which revealed that CaPP2C35 may be involved in the formation of light-green immature fruits in 17C827 by affecting the accumulation of chlorophyll in the exocarp (Fig. 1). This is similar to the mechanisms of APPR2-Like, CaGLK2, and CcLOL1 (Pan et al. 2013;Brand et al. 2014;Borovsky et al. 2019). APPR2-Like indirectly regulated the accumulation of chlorophyll in pepper fruit through the ABA signal pathway (Pan et al. 2013). CaGLK2 affected the accumulation of chlorophyll by adjusting the size of the chloroplast chamber (volume) (Brand et al. 2014). CcLOL1 affected the green color of immature fruits by regulating chloroplast size and chlorophyll content (Borovsky et al. 2019). In contrast, CaPP2C35 may regulate chloroplast grana lamella stacking and stroma lamella formation, resulting in differences in the thylakoid area in the chloroplasts and affecting the metabolism and accumulation of chlorophyll ( Supplementary Fig. 2), similar to the mechanisms underlying the formation of white immature cucumber fruits (Liu et al. 2016). Therefore, our findings shed light on the formation of light-green coloration in immature pepper Fig. 3 The position of LG locus revealed via GWAS and BSA. a GWAS of immature-fruit color (Imfc), L*, a*, and b* in 287 pepper accessions; b BSA-seq analysis of light-green color of immature fruits in F 2 constructed from a cross between 17C658 and 17C827. The dashed boxes in Manhattan plot indicate the associated peaks associated with each trait. The color bar is the position on the chromosome 10 of association candidate regions for each trait ◂ Fig. 4 Linkage analysis of LG locus and expression analysis of Cap-ana10g001710. a Polymorphism detection of dCAPS10-1 marker between 17C658 and 17C827; b linkage analysis of LG locus controlling light-green immature fruits in F 2 constructed from a cross between 17C658 and 17C827; c QTL analysis of the value of L*, a*, b*, chromaticity (SD) and shade (SG) measured by the colorimeter for immature-fruit color in F 2 (315 F 2 individuals) constructed from a cross between 17C658 and 17C827; d: Expression analysis of Capa-na10g001710 between 17C658 and 17C827; e: expression analysis of Capana10g001710 between green mixed pool and light-green mixed pool. Green mixed pool was constructed from six F 2 individuals with green immature fruits. Light-green mixed pool was constructed from six F 2 individuals with light-green immature fruits. The samples used for gene expression analysis were the exocarp of fruits, and the reference gene was UBI (AY486137.1). Gene relative expression data were calculated by 2 −ΔΔCT method, with three biological and three technical repeats. '**'Indicates significant differences at the 0.01 level. Data are means of three biological replicates In addition, we tested the dCAPS10-1 marker using the F 2 population and 126 pepper accessions (Supplementary Fig. 11a and Supplementary Table 3). Interestingly, there was a phenomenon related to the low phenotypic matching rates for a gene marker. Combined with the PVE% of dCAPS10-1 for L*, a*, b*, SD, and SG, a possible explanation is that CaPP2C35 has a limited ability to regulate the formation of light-green immature pepper fruits, similar to the ability of CcLOL1 (Borovsky et al. 2019). Meanwhile, we found that there was no G-A substitution at the 828 th base of the CaPP2C35 sequence (313th base of the CaPP2C35 CDS) in some light-green pepper accessions ( Supplementary Fig. 11b), indicating that CaPP2C35 may be up-regulated or that other genes may participate in the formation of light-green immature fruits in these accessions. Similar results have been reported by studies on CCS, CaGLK2, and CcLOL1 (Lefebvre et al. 1998;Brand et al. 2014;Borovsky et al. 2019). Therefore, dCAPS10-1 can be used to assist the selection of light-green immature pepper fruits at the seeding stage within a certain range.
In the current study, the GWAS, BSA, and linkage analysis revealed that the LG locus controlling lightgreen immature-fruit coloration in pepper was located in the region between 174,787,932 and 174,823,007 bp on chromosome 10. Analysis of the CaPP2C35 sequence and its expression confirmed that CaPP2C35 is the candidate gene for LG. Silencing of CaPP2C35 in 17C827 resulted in an increase in the chlorophyll content in the exocarp and the appearance of green stripes on the surface of the fruits. These findings shed light on the formation of lightgreen immature pepper fruits. In future research, we will further study the specific regulatory molecular mechanism of CaPP2C35 to determine whether it indirectly regulates the development of chloroplast thylakoids and the accumulation of chlorophyll through the ABA signal pathway or related transcription factors. Such research will enrich our understanding of the mechanisms involved in immaturefruit color in pepper.

Fig. 6
Expression analysis of CaPP2C35 and chlorophyll metabolism genes in silenced fruits. a Expression analysis of CaPP2C35 in silenced fruits; b expression analysis of key genes of chlorophyll synthesis in silenced fruits; c expression analysis of key genes of chlorophyll degradation in silenced fruits. Control is the blank control; A is the exocarp of fruits from TRV2-treated plants (mock); B is the light-green exocarp of fruits from CaPP2C35-silenced plants; C is the green exocarp of fruits from CaPP2C35-silenced plants. The samples used for gene expression analysis were the exocarp of fruits, and the reference gene was UBI (AY486137.1). Gene relative expression data were calculated by 2 −ΔΔCT method, with three biological and three technical repeats. Data are means of three biological replicates Fig. 7 Expression patterns of key genes of chlorophyll metabolism and CaPP2C35 in 17C658 and 17C827. 9D, 12D, and 20D are 9 days, 12 days, 20 days after anthesis, respectively. The samples used for gene expression analysis were the exocarp of fruits and the reference gene was UBI (AY486137.1). Gene relative expression data were calculated by 2 −ΔΔCT method, with three biological and three technical repeats. Data are means of three biological replicates