Background

The WD-repeat (WDR) proteins are commonly referred to as the WD40 repeat proteins or WD40 proteins owing to the highly conserved domain of approximately 40 amino acid stretches [1,2,3,4]. This conserved domain is characterized by a glycine histidine (Gly-His, GH) dipeptide at the N-terminal (start) and a tryptophan-aspartic acid (Trp-Asp, WD) dipeptide at the C-terminal (end) [1, 4]. Structurally, the WD40 proteins typically have multiple tandem repeated WD motifs, forming a series of four-stranded and antiparallel β sheets [2, 5, 6]. Functionally, the WD40 domain has no catalytic activity of its own and is considered an adaptor protein that recruits other factors to form protein-protein or protein-deoxyribonucleic acid (DNA) complexes [7,8,9]. WD40 proteins are involved in histone modification [10], DNA damage repair [11], signal transduction [9], and abiotic stress responses [12].

The MYB and basic helix-loop-helix (bHLH) transcription factors and WD40 repeat protein form an MYB-bHLH-WD40 (MBW) complex, which jointly regulates flavonoid biosynthesis, constituting among the most general functions of the WD40 genes. For example, the TTG1-bHLH-MYB complex promotes flavonoid accumulation and the formation of trichomes in Arabidopsis seedlings [13, 14]. Similarly, TT2 (AtMYB123), TT8 (AtbHLH42), and TTG1 (WD40) synergistically regulate the biosynthesis of the proanthocyanidins pigments in the Arabidopsis seed coat [15]. The expression of several structural genes is regulated by MBW complex, including dihydroflavonol-4-reductase (DFR), anthocyanin dioxygenase/ anthocyanin synthase (LDOX/ANS), BANYULS (BAN) [15, 16]. In pepper, CaTTG1 forms an MBW complex with CaAN1 and CaGL3 to control anthocyanin synthesis by activating the expression of structural genes in the synthesis pathway of anthocyanin [17].

Some studies have demonstrated the participation of the WD40 in various biological processes involved in reproductive development in Arabidopsis. For instance, the XPO1-interacting WD40 protein 1 (XIW1) can interact with ABA insensitive 5 (ABI5) in the nucleus and regulate seed germination and growth [18]. The JGB gene with the WD40 domain is a negative regulator of pollen germination [19]. WDR55 is essential for multiple plant developmental processes, including gametogenesis, seed, and endosperm development [20]. The SWA1 protein is necessary for the normal functioning of the mitotic division cycle by regulating cellular metabolism [21].

The WD40 repeats are widely found in all eukaryotes but are rare in prokaryotic organisms [3, 22,23,24,25,26]. In plants, the WD40 proteins form a superfamily, which are functionally diverse. With the completion of whole genome sequencing of different species, the WD40 protein family has been systematically identified in several [3, 25,26,27,28,29]. The 237 potential WDR proteins from Arabidopsis have been identified. These show conservation and divergence in structure and function [3]. In Triticum aestivum, 743 WD40 proteins have been grouped into 5 clusters and 11 subfamilies and may be involved in reproductive development and mediating responses to multiple stresses [25]. In Oryza sativa, 200 OsWD40 genes have been identified and are suggested to perform diverse functions through a complex network [26]. Li et al. [27] found 191 WDRs in cucumber (Cucumis sativus L.). In Arabidopsis, WD40s show strong conservation during the evolutionary process. Feng et al. [28] identified 220 WD40s in the peach genome, and corresponding proteins localized to different subcellular structures. Chen et al. [29] analyzed 164 barley HvWD40 proteins and revealed that the WD40 gene formed an MBW complex with MYB and bHLH and participated in anthocyanin synthesis. These results suggest that the WD40 gene family has diverse gene numbers, structures, and functions.

Pepper, a typical spicy vegetable, belongs to the genus Capsicum of the Solanaceae family. It originated in South and Central America [30]. Capsicum species have various morphological characteristics, and several pepper species lack anthocyanins [17, 31, 32]. In recent years, several genes controlling anthocyanin synthesis have been identified in pepper. Zhang et al. [32] reported that a splice acceptor site in CabHLH1 resulted in a frameshift mutation in a mutant of deficient anthocyanin and fertility reduction of anther (rpf1). Two genes, CaAN3 and Ca3GT, controlling anthocyanin biosynthesis in pepper fruit were successively reported [33, 34]. CaHY5 regulated anthocyanin accumulation in pepper hypocotyl by directly binding to the promoter of downstream structural genes in the anthocyanin pathway [35]. A structural gene of anthocyanin synthesis, flavonoid 3’,5’-hydroxylase (F3’5’H), controls pigment biosynthesis in stems and anthers in pepper [17].

In this study, 269 CaWD40 members were identified in the Zunla pepper genome. The physicochemical properties, sequence structural characteristics, phylogenetic construction, and tissue expression of the CaWD40 gene family were analyzed by bioinformatics methods. Based on yeast two-hybrid (Y2H) technology, we verified the potential involvement of a novel MBW complex (CaAN1-CaDYT1-CaWD40-91) in anthocyanin synthesis. These results are expected to facilitate the understanding of biological functions and molecular mechanisms of WD40 proteins and provide a basis for studying the anthocyanin regulatory network in pepper.

Results

The CaWD40 gene family shows great variations in sequence length and physicochemical properties

The Hidden Markov Model (HMM) profile of the WD40 repeat domain (PF00400) was used as a query sequence to search WD40 proteins in the Zhangshugang, Zunla, and CM334 genome databases. Preliminarily, 267, 276, and 245 WD40s were retrieved from the Zhangshugang, Zunla, and CM334 reference genomes, respectively. All examined pepper WD40 members were verified, and partial WD40 genes with incomplete WD40 domains, considered pseudogenes, were further removed. Finally, 257, 269, and 237 WD40s were identified in the Zhangshugang, Zunla, and CM334, respectively. Since the knowledge of WD40 members in the Zunla genome is relatively complete, the information on this genome was subject to subsequent analyses. For convenience, pepper WD40 candidates were named CaWD40-1–CaWD40-269 according to their positions on the 12 chromosomes in the Zunla reference genome. Detailed information on gene name, chromosome location, protein length, molecular weight, isoelectric point (pI), and domain is listed in Table S1.

CaWD40 genes varied greatly in terms of protein length and physicochemical properties. The length of these CaWD40 fragments ranged from 100 aa (CaWD40-85, CaWD40-212) to 3,595 aa (CaWD40-87). The molecular weights (MWs) ranged from 10.75 kDa (CaWD40-85) to 399.13 kDa (CaWD40-87). The isoelectric point (pIs) ranged from 4.21 (CaWD40-267) to 9.96 (CaWD40-151) (Table S1). The prediction results for the subcellular localization of the 269 CaWD40s were as follows: 139 in the nucleus, 56 in the chloroplasts, 55 in the cytoplasm, six in the plasma membrane, five in the cytoskeleton, four in the mitochondria, two in the vacuole membrane, one in the endoplasmic reticulum, and one in the peroxisome.

Chromosome localization reveals multiple tandem duplication events

The 245 out of the 269 CaWD40 genes were randomly distributed across 12 chromosomes in the Zunla genome, and the remaining 24 CaWD40s were not anchored to chromosomes (Fig. 1A). The vast majority of CaWD40 members were concentrated on the distal end of each chromosome, with a few genes in the middle of the chromosome. The number of CaWD40s was highest on chromosome 3 (34), followed by chromosome 1 (33); the lowest was on chromosome 5 (8) (Fig. 1B). The distribution of CaWD40s indicated the accumulation of some genes on specific chromosomes.

We analyzed tandem duplications in the CaWD40 gene family in pepper. Nineteen pairs (I to XIX) of genes among 269 CaWD40s were identified as tandem duplication events (Table S2), crucial in gene functional differentiation. Tandem duplication events were unevenly distributed across the 10 chromosomes and were usually concentrated in regions with high gene density. Some tandem repeat genes belonged to the same phylogenetic branch, while others belonged to different branches, hinting at functional differentiation and conservation among the CaWD40 gene family (Fig. 2).

Fig. 1
figure 1

Chromosome mapping and gene number of CaWD40 proteins in the Zunla genome. (A) Chromosome mapping and tandem duplication events of CaWD40s. Chromosome numbers are represented on the top, and the scale is shown on the left. Predicted tandem duplicate pairs are indicated in red font, and the categories are marked in parentheses using Roman numerals. (B) Statistical analysis for CaWD40s on 12 chromosomes

Full size image

The combination of phylogenetic and structural analysis reveals the diversity of CaWD40s

In this study, the protein sequences of the 269 predicted CaWD40s were used to generate an unrooted tree to analyze the classification of the CaWD40 gene family in pepper (Fig. 2). Phylogenetic analysis showed that the CaWD40 gene family was categorized into six main branches (branch A to F) with 15, 22, 15, 42, 68, and 107 members, respectively. Among them, branch F was divided from F1 to F4 (Fig. 2). Branches A, B, and C were the earliest clusters in the phylogenetic tree, and genes in the same branch shared similar motif compositions (Fig. 2 and Fig S1). Similarly, most genes in branches D, E, and F followed this rule. Combined with chromosome localization analysis, gene pairs with tandem duplication were mostly distributed in the E and F branches, suggesting that the differentiation of gene function was gradual along with gene duplication (Figs. 1 and 2).

Fig. 2
figure 2

Phylogenetic relationships among CaWD40s in the Zunla genome. Different branches are marked by shading with diverse colors. The gray dots represent the bootstrap, and those with a bootstrap greater than 0.5 are shown. Predicted tandem duplicate pairs are marked with fonts of diverse colors

Full size image

The structural domain of the predicted 269 CaWD40s was counted using SMART (Table 1, Table S1) to analyze the relationship between domain composition and phylogenetic classification. The CaWD40 gene family was categorized into 17 distinct classes based on domain composition (Table 1). Among them, 191 members contained only a single WD40 domain, and the remaining had at least one or more WD40 domains and other domains.

Table 1 Domain composition of the 17 classes of 269 CaWD40s
Full size table

Gene structural characteristics can elucidate the relationship between gene function and evolution. The number of exons and introns in the 269 CaWD40s varied greatly (Fig. S1A and B). Twenty-three CaWD40 genes (8.55%) had only one exon and were intronless, including CaWD40-80 and CaWD40-91 (Fig. S1B). A total of 146 CaWD40s (54.28%) contained 1–10 exons; 83 CaWD40s (30.86%) contained 11–20 exons, and the remaining 17 CaWD40s (6.32%) contained more than 20 exons. Specifically, CaWD40-218 and CaWD40-181 had two exons; the former belonged to branch A and the latter to branch F (Fig. 2, Fig. S1B). All three genes (CaWD40-62, CaWD40-124, CaWD40-160) belonged to branch A. CaWD40-124 had four exons. CaWD40-160 and CaWD40-62 had nine and 29 exons, respectively. These results suggest that the different exon-intron compositions of 269 CaWD40s are randomly distributed across different branches of the phylogenetic tree, implying a high diversity of the CaWD40 gene family in the process of evolution.

Using the Multiple Em for Motif Elicitation (MEME) tool, we identified 10 putative conserved motifs (motif 1 to motif 10) in the CaWD40 gene family (Fig. S1C). No common motifs were observed in 269 CaWD40s. Among the 10 motifs, motif 1 was the most widely distributed, found in 264 CaWD40 genes. Some motifs were present in only a few genes, and motif 9 and motif 10 were present in only 4 (CaWD40-57/117/233/256) and 6 (CaWD40-86/89/98/147/195/222) CaWD40 genes, respectively. Some CaWD40s with the same exon–intron structure exhibited a similar motif composition, such as CaWD40-69 and CaWD40-266 (Fig. S1C).

Synteny analysis of CaWD40s among three genomes

A genome-wide synteny analysis was performed among the CaWD40s in the Zunla, Zhangshugang, and CM334 genomes. Zunla and Zhangshugang had 225 gene pairs co-linked and Zunla and CM334 had 178 gene pairs co-linked (Fig. 3, Table S3). Zunla and Zhangshugang showed a close synteny for the CaWD40 gene, suggesting that these two species may share more similar functions. The relationship of species origin indicates that Capsicum annuum cv. Zhangshugang and Zunla are derived from China [31, 56]; the CM334 pepper is from South Korea [57].

Fig. 3
figure 3

Synteny analysis for WD40s among different pepper genomes. The identified collinear genes are shown with blue lines

Full size image

CaWD40s exhibit distinct expression profiles in various tissues and stages

By RNA-seq analysis, expression profiles of the putative CaWD40 genes at various tissues and developmental stages of Capsicum line 6421 were analyzed (Fig. 4). CaWD40s showed varied expression patterns, which could be divided into six types (type i–vi) based on a hierarchical clustering of their expressional characteristics (Fig. 4, Table S4). Specifically, type i contained 48 genes, most relatively highly expressed in different tissues and periods. Type ii comprised 101 genes with medium or low expression in various tissues or stages. Type iii included 21 CaWD40s, highly expressed in almost all tissues. Type iv included 49 genes showing low-to-no expression in most tissues. Type v had 16 members with tissue-specific high expression in specific tissues or stages. For example, four genes (CaWD40-51, CaWD40-135, CaWD40-199, and CaWD40-223) were highly expressed in flower and stamen tissues and showed almost no expression in other tissues. CaWD40-93 and CaWD40-125 showed high expression levels in fruit or placenta. Group vi contained 34 genes with low expression levels in all tissues.

Fig. 4
figure 4

Expression analysis of CaWD40 genes in various tissues and stages based on RNA-seq data. Cluster hierarchy based on expression levels. Heat map; blue/yellow/red colors represent low/medium/high expressions, respectively

Full size image

Verification of expression in the eight CaWD40s

We randomly selected eight CaWD40s from each hierarchical cluster in the heatmap to determine their expression in different tissues by RT-qPCR to verify the expression of the CaWD40 gene family. The expression trend identified by RT-qPCR was similar to that of RNA-seq (Figs. 4 and 5). For example, the expression of CaWD40-26 was not detected in diverse tissues; CaWD40-69 and CaWD40-266 were highly expressed in almost all examined tissues and stages (Fig. 5), and CaWD40-135 and CaWD40-223 tended to be expressed later during flower development.

Fig. 5
figure 5

Relative expressions of CaWD40-26, CaWD40-57, CaWD40-69, CaWD40-91, CaWD40-127, CaWD40-135, CaWD40-223, and CaWD40-266. F1, F4, F5, F7, F8, and F9 represent different stages of buds in Capsicum line 6421

Full size image

A novel MBW complex may be involved in anthocyanin biosynthesis and genic male sterility

Previous reports have shown that WD40 typically acts as part of the MBW complex to regulate anthocyanin biosynthesis [13,14,15, 17]. A recent report has demonstrated that the MBW complex of CaAN1-CaGL3-CaTTG1 affects pigment accumulation and male gametophyte development in the Cha1 mutant of pepper [17]. In this study, CaWD40-91 (Capana04g000080) protein showed high phylogenetic and structural similarity with CaTTG1 (CaWD40-80/Capana03g001813) (Figs. 4 and 5, Fig S2, 3). Subcellular localization results also demonstrated CaWD40-91 protein localization in the nucleus and cytoplasm (Fig. 6A), consistent with the localization results for CaTTG1 [17]. Therefore, we reasonably speculate that CaWD40-91 might function similarly to the CaTTG1 gene.

To confirm whether CaWD40-91 can bind to the CaGL3 and CaAN1 proteins, a Y2H experiment was performed. CaWD40-91 could interact with CaAN1 but not with CaGL3 (Fig. 6B). Considering that CaWD40-91 could indeed bind to CaAN1, we replaced other bHLH proteins to verify the protein interaction. A recent study from our laboratory showed that a candidate bHLH (CaDYT1) activated CaDFR expression to control anthocyanin synthesis and male sterility in pepper (unpublished data). Therefore, we verified the interaction between CaWD40-91 and CaDYT1 proteins. The results indicated that CaWD40-91 interacted with the CaDYT1 and CaAN1 proteins, resulting in the formation of a novel MBW complex (CaAN1-CaDYT1-CaWD40-91) (Fig. 6B and C), which may regulate synergistically anthocyanin accumulation and genic male sterility of pepper.

Fig. 6
figure 6

Subcellular localization and interaction. (A) Subcellular localization of CaWD40-91. Scale bars represent 50 μm. (B) Interactions of CaWD40-91 with CaAN1 and CaDYT1 via a yeast two-hybrid assay. (C) The dual-luciferase reporter gene assay.

Full size image

Mutation in CaDYT1 causes male sterility in pepper

We previously cloned the CaDYT1 gene by crossing the dyt1 mutant with wild type (WT) to obtain the F2 genetic population (unpublished data). The sequencing results are shown in Fig. 7A. The second exon of the CaDYT1 gene had a 7 bp deletion in the dyt1 mutant, resulting in a frameshift mutation. Phenotypic analysis showed that the anthers of the dyt1 mutant were thinner and yellower compared with WT (Fig. 7B). Scanning electron microscopy (SEM) showed a round and full WT anther, with a closely arranged convex surface structure. Several pollen grains are produced following anther dehiscence. In contrast, dyt1 anthers were small and folded, no pollen grains were observed, and only impurities and starch grains were found (Fig. 7B).

Fig. 7
figure 7

Phenotypic analysis of the wild type (WT) and dyt1 mutant. (A) Mutation site detection between WT and dyt1. (B) The phenotype of anther and pollen grain in WT and dyt1. Anther phenotype is shown on the left, and SEM analysis of anther and pollen grain is shown on the right

Full size image

Discussion

Diversity and conservation of CaWD40 proteins in pepper

In plants, WD40 is not directly involved in recognizing target gene promoters, but it plays a role in the MBW complexes with MYB and bHLH proteins [9, 15]. The MBW complex is important in the anthocyanin biosynthesis pathway [17]. In the MBW complex, the WD40 protein is usually located at the center of the triadic structure. It may protect the MBW complex by preventing other transcriptional regulators from binding to MYB or bHLH [15]. The functions of WD40 are diverse. In addition to anthocyanin synthesis, WD40 proteins are critical for pollen germination [36], drought stress [37], RNA processing [38], regulation of plant biological clock [39], skin and fur development [40], and other processes.

In this study, 269 CaWD40s were identified in the Zunla pepper genome. Interestingly, 237 and 257 CaWD40s were found in CM334 and Zhangshugang genome, respectively. Different pepper cultivars may lead to the difference in CaWD40 numbers in the three genomes [31, 56, 57]. Finally, the 269 CaWD40 members from the Zunla genome were analyzed. These genes were divided into six main branches based on their gene structure and motif compositions. The CaWD40 members within the branch showed high similarity in domain compositions and sequences, suggesting a conserved gene function.

The WD40 gene family has been reported in many species. The number of members varies greatly between species, such as 207 in tomato [41], 237 in Arabidopsis [3], 200 in rice [26], 191 in cucumber [27], 743 in Triticum aestivum [25], 164 in barley [29], and 220 in Peach [28]. Based on published genomic databases, the genome sizes of pepper, tomato, and Arabidopsis are predicted to be 3.5 Gb, 0.95 Gb, and 0.125 Gb, respectively. The above results indicate that the number of WD40s is not related to the genome size of a species. We found a high polarization in physical and chemical properties: protein length ranges from 100 aa (CaWD40-85, CaWD40-212) to 3,595 aa (CaWD40-87), and molecular weights ranged from 10.75 kDa (CaWD40-85) to 399.13 kDa (CaWD40-87). These results are consistent with previously reported studies [25, 26, 28, 29].

Nineteen pairs (40 CaWD40s) were identified as tandem duplication events, and these genes were unevenly distributed in 10 chromosomes (Fig. 1). Tandem repeat genes from a pair also sometimes clustered on the different phylogenetic branches (Fig. 2), suggesting functional divergence in the CaWD40 gene family. Sequence characteristics revealed that the number of exons and introns in the 269 CaWD40 gene family varied greatly, that is, from intronless to 30 introns (Fig. S1B and C). However, some CaWD40s possessed similar exon–intron and motif compositions (CaWD40-69 and CaWD40-266) (Fig. S1B and C). Overall, CaWD40 proteins were structurally dynamic, which could determine the functional diversity of the WD40 gene. In humans, WD40s are involved in multiple cellular networks, many of which have been implicated in diseases [42]. Two proteins, WDR5 and EED, are involved in chromatin complexes [42]. In Arabidopsis, several WD40 proteins are related to gametogenesis, seed, and endosperm development [19, 20]. The WD40 gene family is a superfamily of plant species. Some supergene families, such as WRKY, bHLH, and MYB, are usually accompanied by structural and sequence similarities and differences in the evolution process [19, 43,44,45,46].

CaWD40-91 is a potential candidate involved in anthocyanin biosynthesis and pollen fertility

In plants, flavonoids are synthesized through a conserved metabolic pathway, producing several secondary metabolites (SMs), including flavones, anthocyanins, and proanthocyanidins [47]. These SMs have pivotal roles in growth and development as they provide plants with different pigments that scavenge reactive oxide species (ROS) [32], defend against UV damage [48], or mediate plant-microbe interactions [49]. The SM biosynthetic pathway is among the most extensively studied in plants. Several structural genes related to anthocyanin biosynthesis, including CHS, CHI, F3H, F3’5’H, DFR, ANS, and UFGT, are conserved in most plants [47, 49, 50]. In recent years, several reports suggest a close association between flavonoids and pollen fertility [32, 48, 51,52,53].

Wang et al. [17] showed that CaTTG1, CaAN1, and CaGL3 formed an MBW complex, which regulated anthocyanin synthesis. In Arabidopsis, TTG1 can bind to AtMYB123 and AtbHLH42 transcription factors to regulate the biosynthesis of the proanthocyanidins pigments in the seed coat [13, 14]. DYSFUNCTIONAL TAPETUM 1 (DYT1), acting as an upstream regulator of tapetum development, is required for pollen formation in Arabidopsis and pepper [54, 55]. Our recent study suggests that the cha1 mutant (CaTTG1) exhibits partially aborted pollen grains (unpublished). In this study, CaWD40-91, a paralog protein of CaTTG1 (named CaWD40-80 in this paper), was confirmed to interact with CaAN1 and CaDYT1. In general, genes that are highly homologous in sequence and structure tend to have similar functions. More interestingly, we recently found that DYT1 could bind to DFR synergistically and regulate anthocyanins and male sterility in pepper (unpublished). Based on the above results, CaWD40-91, CaAN1, and CaDYT1 may form an MBW complex, which can regulate the formation of anthocyanins and pollen fertility.

Conclusions

From bioinformatics analysis, we identified 269, 237, and 257 CaWD40 genes in the Zunla, CM334, and Zhangshugang genomes, respectively. A total of 269 candidates from the Zunla genome were analyzed, and they were named CaWD40-1 to CaWD40-269 according to their arrangement on the chromosomes. The protein length and physico-chemical properties varied greatly. Protein lengths ranged from 100 to 3,595 aa, and molecular weights ranged from 10.75 kDa to 399.13 kDa. The analysis revealed 245 CaWD40 proteins clustered unevenly along chromosomes, with 24 CaWD40s remaining unlinked to chromosomes. Further analysis revealed 19 tandem duplicate pairs that clustered in different branches of the phylogenetic tree. Sequence analysis indicated that the numbers of exon (1–31) and intron (0–30) varied greatly in the CaWD40 gene family, and the composition of conserved motifs was diverse. However, several CaWD40s showed highly conserved sequence characteristics, expression patterns, and biological function. Finally, we found that CaWD40-91 showed high homology with CaTTG1 based on phylogenetic relationship, gene structure, and domain composition, with further speculations that they may share similar functions. A recent report revealed that CaTTG1 could interact with CaGL3 and CaAN1 to form an MBW complex, which could regulate anthocyanin synthesis in pepper [17]. In this study, we found that CaWD40-91 interacted with CaAN1 but not with CaGL3. Interestingly, we observed that CaWD40-91 could interact with CaDYT1 from the Y2H assay. Moreover, CaDYT1 activated CaDFR to synergistically regulate anthocyanin synthesis and male sterility. Therefore, we speculate that CaWD40-91 can interact with CaAN1 and CaDYT1 to control pigment accumulation and sterility in pepper.

Methods

Retrieval and identification of WD40 genes in pepper

In this study, the candidate WD40 proteins were retrieved as follows: first, the protein, nucleotide, and genome sequences of the Zunla, CM334, and Zhangshugang genomes were downloaded from the Sol Genomics Network (https://solgenomics.net/organism/Capsicum_annuum/genome) [56, 57] and Pepper Genomics Database (http://ted.bti.cornell.edu/cgi-bin/pepper/index) [31], respectively. Second, the HMM of WD40 protein (PF00400) was downloaded from the Pfam database (http://pfam-legacy.xfam.org/) [58]. Furthermore, the HMMER (v3.0) software [59] was used to search the predicted WD40 proteins using the cut-off value of the default parameter. Finally, the WD40 domain of all potential candidates was examined in InterPro (https://www.ebi.ac.uk/) [60] and SMART database (http://smart.embl-heidelberg.de/smart/batch.pl) [61], and those without the complete WD40 domain were manually deleted.

Sequence analysis and structural characteristics

Protein length, molecular weight, and theoretical pIs of CaWD40 proteins were analyzed using the Expasy tool (https://web.expasy.org/compute_pi/) [62]. The supposed subcellular localization of CaWD40 proteins was predicted using the online tool WOLF PSORT (https://wolfpsort.hgc.jp) [63]. The protein sequences were submitted to the MEME program (https://meme-suite.org/meme/tools/meme) [64] to assess conserved motifs.

Chromosome localization, tandem duplication, and synteny analysis

Chromosome locations and gene position in pepper were obtained by searching the Sol Genomics Network. Chromosome mapping of the CaWD40 gene family was visualized using MG2C (http://mg2c.iask.in/mg2c_v2.1) [65]. Tandem duplication events were further confirmed using the following criteria: (1) the alignment length had a coverage rate of more than 70% of the full length of the CaWD40 genes; (2) the identity of the aligned region was over 70%, (3) and an array of two or more genes was less than 100 kb distance. The genome sequence was downloaded and annotation files of three different pepper species from NCBI (https://www.ncbi.nlm.nih.gov/assembly/?term=pepper) were obtained. TBtools-II [66] was used to analyze the synteny of CaWD40 genes among the three pepper genomes.

Phylogenetic analysis

The phylogenetic tree was generated in the following three steps: first, the CaWD40 protein sequences were imported into Clustal X to produce a multiple sequence alignment file. Second, the alignment result was used to build an unrooted tree using MEGA11 with a bootstrap of 1000 replicates and neighbor-joining (NJ) methods [67]. Third, the newly produced phylogenetic tree was visualized using the Interactive Tree of Life online website (https://itol.embl.de/) [68].

RNA-Seq analysis of CaWD40 genes

Transcriptome sequencing (RNA-seq) data of development was used to explore the distribution of gene expression in pepper (the elite Capsicum line 6421) [69] to gain insight into the expression profiles of the CaWD40 gene family in different tissues across periods. Expression levels were determined in the following tissues and stages: leaf tissues were sampled at 2, 5, 10, 15, 20, 25, 30, 40, 50 and 60 days after emergence, and marked correspondingly as L1, L2, L3, L4, and L5, L6, and L7, L8, L9, and AL; floral buds were sampled at 0.25, 0.35, 0.5, 0.8, 1.0, 1.2, and 1.7 cm, and marked correspondingly as F1, F2, F3, F4, and F5, F6, and F7, F8, and F9; petals, stamens, and ovaries with stigmas were sampled in fully blossomed flowers, and marked correspondingly as P10, STA10 and O10; fruits were collected at 3, 7, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, and 60 days after flowering (DAF), and marked correspondingly as FST0, FST1, G1, G2, G3, G4, G5, G6, G7, G8, G9, G10, G11; seed samples were collected at 10, 15, 20, 25, 30, 35, 40, 45, 50, 55 and 60 DAF, and marked correspondingly as ST1, ST2, S3, S4, S5, S6, S7, S8, S9, S10, S11; placenta samples were collected at 20, 25, 30, 35, 40, 45, 50, 55 and 60 DAF, and marked correspondingly as T3, T4, T5, T6, T7, T8, T9, T10, and T11. Stems and roots were marked AS and AR, respectively. The treatment methods of all samples were based on those published by Liu et al. [69]. All data of CaWD40 genes were normalized (log2(FPKM + 1)), and a heatmap was drawn using TBtools-II (v2.019) [66].

RNA extraction and RT-qPCR analysis

Capsicum line 6421 was grown at 27 ℃/22 ℃ day/night cycles under a 16/8 h (light/dark) photoperiod. Total RNA was extracted using the TransZol Up (TransGen, China) and reverse-transcribed using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, China). RT-qPCR sample contained 2 × ChamQ Universal SYBR qPCR Master Mix (Vazyme, China) and was performed on a LightCycle® 96 Real-Time PCR System (Roche, Switzerland).

The UBI-3 gene was used as a reference gene [70]. Gene-specific primers were designed. These are listed in Table S5. Three biological repeats and three technical repeats of RT-qPCR were set, and the relative expression of each gene was calculated using the 2−ΔΔCt method.

Y2H assays

Using specific primer pairs, the coding sequences of CaWD40-91, CaAN1, CaDYT1, and CaGL3 were amplified and inserted into pGBKT7 and pGADT7 vectors, respectively (Table S5). Four plasmids, pGBKT7-CaWD40-91, pGADT7-CaAN1, pGADT7-CaDYT1, and pGADT7-CaGL3, were extracted and co-transformed into Y2HGold receptor cells (WEIDI, China). All transformation products were plated and grown on SD/-Trp/-Leu medium for three days. Protein-protein interactions were detected on the SD/-Trp/-Leu/-His/-Ade selective medium following the instructions in the Matchmaker® Gold Yeast Two-Hybrid System User Manual.

Dual-luciferase reporter assays

The primers for Nluc-CaWD40-91, Cluc-CaAN1, and Cluc-CaDYT1 were designed (Table S5). CaWD40-91 was cloned and inserted into the pCambia1300-Nluc. CaAN1, and CaDYT1 were cloned and inserted into pCambia1300-Cluc linearized by KpnI and SalI. These recombinant plasmids were transferred into Agrobacterium tumefaciens GV3101. The plasmids were delivered into tobacco leaves, and a fluorescence signal was captured three days later using the Vilber Fusion FX7 Spectra device system (Vilber Bio Imaging, Paris, France). Nluc-EV and Cluc-EV were used as controls.

Subcellular localization

The full-length ORF sequences of CaWD40-91 without the termination codon were cloned into the pCAMBIA1300-GFP vector and transformed into Agrobacterium tumefaciens GV3101. Arabidopsis histone H2B-mCherry was used as a nuclear marker. CaWD40-91 fusion constructs and H2B marker were co-transformed into tobacco (Nicotiana benthamiana) leaves in a ratio of 1:1. After three days, the fluorescence signals were observed and captured using a confocal laser scanning microscope (Zeiss LSM510 META, Germany).

SEM analysis

Floral buds were sampled and fixed with a fixative solution. Samples were washed with 0.1 M PB (pH 7.4) thrice for 15 min each. The samples were transferred into 1% OsO4 in 0.1 M PB (pH 7.4) for 1–2 h at room temperature. Subsequently, samples were washed in 0.1 M PB (pH 7.4) thrice for 15 min each. Then, they were transferred to 30%, 50%, 70%, 80%, 90%, 95%, 100%, and 100% ethanol for 15 min each. Finally, the samples were incubated in isoamyl acetate for 15 min and dried using a Critical Point Dryer. The dried samples were attached to metallic stubs using carbon stickers and sputter-coated with gold for 30 s. Samples were observed and photographs were captured with a scanning electron microscope (SU8100, Hitachi, Japan).