The ‘Red Zaosu’ and ‘Zaosu’ pear trees were planted in the Pear Germplasm and Innovative Resources Nursery at the Chongzhou Modern Agriculture R&D Base of Sichuan Agricultural University, located at 103°64’52’’ E and 30°55’67’’ N. The trees were of similar age, growth potential, and management level. Tobacco (Nicotiana benthamiana) was cultured under controlled environmental conditions in the Materials Room of the Plant Molecular Genetic Breeding and Biotechnology Laboratory, College of Horticulture, Sichuan Agricultural University, China, for subcellular localisation, dual luciferase experiments, and BiFC tests. The temperature was maintained at 23 ± 2°C, with a relative humidity of 80%–90% and a light-dark cycle of 16/8 h. The light intensity was 220 μmol m^-2s^-1. ‘Red Zaosu’ stems, leaves, buds, flowers, fruits, seeds at various developmental stages, and pear pericarp samples were collected at 25, 45, 65, 85, 105, and 115 Day after anthesis (DAF). The pear fruits were harvested, and the epidermis was removed using a scalpel. All samples were immediately frozen in liquid nitrogen and stored at –80°C. The samples were then sorted by tissue type. The ‘Zaosu’ and ‘Red Zaosu’ pears were bagged 15 d after flowering. The fruit was then collected at 30 and 90 d after bagging (DAB) for the instantaneous injection test.

The epidermis of different regions of the fruit at 65, 85 and 105 DAF was used as the material, total RNA was extracted using the CTAB-based method (Gambino and Gribaudo, 2012) and transcribed into cDNA. The quality of the total RNA was assessed, and BioMarker company (Qingdao, China) was used to construct and sequence the libraries. Each sample included three biological replicates. The sequencing reads were analyzed using our previously described procedures (www.biocloud.net).

The primers listed in Supplementary Table S1 were used to amplify and clone the genes into the pBlunt vector (Yeasen Biotechnology, Shanghai) using the mined gene sequences. The sequence information was determined by comparing and analyzing the cloned sequences in the Genome Database for Rosaceae (GDR) and National Center For Biotechnology Information databases (NCBI). The MYB structural domains were analyzed using the Pfam (pfam.xfam.org) and ESPript 3.0 online websites (Jiang et al., 2023). The ClustalW program in the MEGA 6.06 package was used for multiple sequence comparisons. A phylogenetic tree was constructed using the neighbor-joining method with 1000 bootstrap repeats.

The CDS sequence of PbMYB5 was cloned into the expression vector pCAMBIA-35SN (35S::PbMYB5), and the RNA interference (RNAi) vector pRNAi-35SN (RNAi::PbMYB5). Transient overexpression and silencing were performed in pear fruit per the described protocol (Cong et al., 2021; Ni et al., 2023; Yao et al., 2017; Zhai et al., 2016). Each replicate included at least 10 fruits (30 and 90 DAB), with three biological replicates. The injected fruits were harvested after seven days, and the peel of the injected area was scraped using a scalpel. The harvested peel was stored at –80°C before use. For stable genetic transformation of ‘Red zaosu’ and ‘Zaosu’ healing tissues, the 35S::PbMYB5 and RNAi::PbMYB5 constructs were transformed into Agrobacterium tumefaciens GV3101. Pear healing tissues were infiltrated following the previously described method (Ni et al., 2023). Anthocyanoside synthesis was induced by culturing positive healing tissues in the dark for two weeks, followed by continuous light treatment for 96 h.

SYBR Green (TaKaRa, Dalian, China) was used to detect the PCR products on a CFX96 real-time reaction system (Bio-Rad). The housekeeping gene JN684184 (Actin) was selected and analyzed using the 2^-ΔΔCt method (Bai et al., 2019). The qPCR primers are listed in Supplementary Table S2 .

The total anthocyanin content was determined using the pH difference method (Ni et al., 2023; Yao et al., 2017). PAs were detected by reaction with p-dimethylaminocinnamaldehyde (DMACA) solution (0.1% [w/v] DMACA, 90% [v/v] ethanol, and 10% [w/v] HCl) to produce a blue color, and their concentrations were determined through full-wavelength zymography using the DMACA assay, as previously described (Zhang et al., 2022). Positive healing tissues were stained for PA using DMACA solution (0.2% [w/v] DMACA, methanol: 6M HCL = 1:1 [v/v]), immersed for 48 h, rinsed with distilled water, dried and observed (Zhang et al., 2022).

PbMYB5 was used to the N-terminus of eGFP in the pCAMBIA-35 S-eGFP vector (PbMYB5-eGFP) by homologous recombination. For BiFC, the target genes were cloned into vectors modified with pSAT1-nEYFP-C1 and pSAT1-cEYFP-N1 to produce fused YFP target proteins (Citovsky et al., 2006). The vector was transformed into the Agrobacterium strain GV3101. The cells were incubated overnight at 28°C in a YEP medium supplemented with the appropriate antibiotics. After centrifugation, the collected cells were resuspended in MMA solution (10 mM MES, 10 mM MgCl₂, and 500 μM acetosyringone) and incubated at room temperature for 1 h on a shaker. The bacteria were suspended in fresh buffer and adjusted to a final density of OD 600 = 1. The suspension was injected into the dorsal axial lateral chloroplasts of the tobacco leaves using a syringe. The treated plants were kept in a greenhouse for 2–3 d and then observed under a confocal microscope (Olympus FV1000). The primers used to construct the vectors are listed in Supplementary Table S1 .

Two target genes were cloned into the pGADT7 and pGBKT7 vectors, and positive interactions were detected using the SD/-Trp/-Leu and SD/-Trp/-Leu/-Ade/-His/+ X-α-gal/+3-AT medium. Supplementary Table S1 lists all primers used for vector construction.

Genes were cloned into the pGADT7 vectors, target gene promoter cloning in pHIS vector, and positive interactions were detected using the SD/-Trp, and SD/-His/-Trp + 3AT, and SD/-Trp/-Leu/-His + 3-AT medium. Supplementary Table S1 lists all primers used for vector construction.

To construct the reporter gene, promoter sequences of CHS, CHI, F3H, DFR, LAR, ANR, and UFGT were cloned from the ‘Red Zaosu’ pericarp and inserted into the pGreenII0800-LUC vector. 35S::PbMYB5 was used as an effector. Empty pCAMBIA-35SN was used as a negative control. The vector was transformed into Agrobacterium GV3101 (pSoup-p19). Dual-luciferase transient expression assays were performed as described (Liu et al., 2008). Firefly luciferase (LUC) and Renilla luciferase (REN) were detected using the Dual-Luciferase Reporter Gene Assay Kit (Yeasen Biotechnology, Shanghai) according to the manufacturer’s instructions. Supplementary Table S1 lists all primers used for vector construction.

Unless otherwise noted, all data were analyzed using the IBM SPSS Statistics 23 software. Statistically significant differences between samples were determined using Student’s t-test (P < 0.05).

The transcriptome of ‘Red Zaosu’ pear epidermis was sequenced in the red region (RS) and in the red-yellow-green gradient region (GS) at 65, 85 and 105 d after anthesis, and KEGG and GO enrichment analyses showed that differentially differentiated genes were enriched in the phenylpropanoid pathway, with one of the genes 2459, having a significantly higher abundance in the RS than in the GS region at 65 d after anthesis and no significant difference in its abundance at 85 and 105 d ( Supplementary Figure S1 ). By searching the GDR database, we found that the sequence of gene 2459 was highly similar to that of rna3106-v1.1-pbr and rna18247-v1.1-pbr in white pear, located on chromosome 3 and chromosome 11, with amino acid sequence lengths of 373 aa and 375 aa, respectively. Further sequence comparison showed that the sequence of gene 2459 had the highest homology with rna3106-v1.1-pbr ( Supplementary Figure S2 ). The gene was therefore cloned from the peel of the ‘Red Zaosu’ pear and compared with the NCBI database, and it was found that several genes with high sequence similarity in different species belonged to the MYB family. Subsequent amino acid sequence conservation analyses showed that the ‘Red Zaosu’ pear skin gene 2459, like other MYB genes, has a complete R2R3 conserved structural domain ( Figure 1A ). Furthermore, a phylogenetic tree demonstrated that this gene was closely related to the predicted transcription repressor factor MYB5 in ‘Hongxiangsu’ pear and MdMYB5 in apple, support level of up to 99 ( Figure 1B ). Consequently, it was designated PbMYB5. Bioinformatics analysis revealed that the CDS sequence of PbMYB5 was 1122 bp long and encoded a 373 amino acid protein. The amino acid sequence contained a highly conserved bHLH interaction motif (DLx2Rx3Lx6Lx3R) in the R3 structural domain, and no anthocyanin characterization motifs (ANDV and KPRPRS/TF) or inhibition motifs (LxLxL and TLLLFR) were found (Stracke et al., 2001). PbMYB5 belongs to a remote clade with PbMYB1, PbMYB10, and PbMYB114. Quantitative gene expression analysis revealed that PbMYB5 expression was higher in the leaves, buds, and flowers ( Figure 1D ). In the pericarp at different developmental stages, the highest expression of PbMYB5 was observed at 45 DAF by 105 and 115 DAF ( Figure 1C ). The subcellular localisation analysis revealed that PbMYB5 was present in both the nucleus and cytoplasm ( Figure 1E ).

The 35S::PbMYB5 overexpression vector and 35S::N empty vector were transiently injected into ‘Zaosu’ pear fruits using Agrobacterium transformation. Injections were administered 30 and 90 d after bagging, respectively. After one week of exposure, significant red accumulations were observed in both injected 35S::N. After one week of treatment, significant red accumulation was observed in all regions injected with 35S::PbMYB 5 , whereas no red color accumulation was observed after 35S::N injection ( Figure 2A ). Examination of the epidermis injected with 35S::PbMYB5 and 35S::N vectors revealed that 35S::PbMYB5 increased the expression of PbMYB5. Expression was higher at 30 d after bagging than at 90 d ( Figure 2B ). Additionally, the expression of lignin synthesis genes PbCAD1 and PbF5H, as well as the PAs synthesis gene PbLAR, was significantly lower than that of 35S::N during both stages. Conversely, after the injection of 35S::PbMYB5, the expression of the PbUFGT was significantly higher than that of 35S::N ( Figure 2C ). The results of the total anthocyanin and PAs content of the injected pericarp revealed that the anthocyanin content was higher than 35S::N and the PAs content was lower than 35S::N after injection of 35S::PbMYB5 at 30 and 90 DAB, respectively ( Figure 2D, E ).

The PbMYB5 silencing vector (RNAi::PbMYB5) and the silencing null vector (RNAi::N) were transiently injected using Agrobacterium transformation into ‘Red Zaosu’ fruits at 30 and 90 d after bagging, respectively. After one week of exposure treatment, it was observed that none of the injected areas showed significant red accumulation after injection of RNAi::PbMYB 5 . There was no significant red accumulation in any of the injected areas and no inhibition of red accumulation in the vicinity of the RNAi::N injection ( Figure 3A ). Examination of the epidermis after RNAi::PbMYB5 and RNAi::N injections revealed that both RNAi::PbMYB5 phases inhibited PbMYB5 expression compared to RNAi::N ( Figure 3B ). Moreover, expression of lignin synthesis genes (PbCAD1, PbF5H) and PAs genes (PbLAR and PbANR) was higher than that of 35S::N, while the expression of anthocyanidin synthesis and transporter gene PbUFGT was significantly lower than that of 35S::N, after injection of RNAi::PbMYB5 ( Figure 3C ). Results of the total anthocyanin and PAs content of the injected pericarp revealed that the anthocyanin content was lower than that of RNAi::N and the PAs content was higher than 35S::N after injection of RNAi::PbMYB5 at 30 and 90 d after bagging, respectively ( Figures 3D, E ).

The PbMYB5 overexpression vector (35S::PbMYB5) and the blank control overexpression vector (35S::N) were transformed into ‘Zaosu’ pear healing tissues. The positive healing tissue analysis revealed that 35S::PbMYB5 promoted PbMYB5 expression comparedto 35S::N ( Figure 4B ). Moreover, DMACA staining revealed that 35S::PbMYB5 suppressed the expression of proanthocyanidins ( Figure 4A ) and related genes PbLAR and PbANR, while promoted the expression of PbANS and PbUFGT ( Figure 4C ). The PbMYB5 silencing vector (RNAi::PbMYB5) and blank control silencing vector (RNAi::N) were transformed into ‘Red Zaosu’ pear healing tissues. Analysis of positive healing tissues showed that RNAi::PbMYB5 was able to repress PbMYB5 expression compared to control. ( Figure 4E ). Moreover, DMACA staining revealed that RNAi::PbMYB5 promoted the expression of PAs ( Figure 4D ) and related genes PbLAR and PbANR, in addition to promoting the expression of lignin synthesis genes PbCCR, PbCAD1, PbCAD2, and PbF5H, and reducing the expression of anthocyanin synthesis gene PbANS (the difference was not statistically significant) ( Figure 4F ).

The yeast two-hybrid results ( Figure 5 ) revealed that PbMYB5 can bind to PbHLH3/PbHLH33 and does not directly interact with PbWD40. The BiFC assay revealed that PbMYB5 interacts with PbbHLH3/PbbHLH33 in the nuclei of tobacco cells. This suggests that PbMYB5 forms a functional complex with PbbHLH3/PbbHLH33 and forms an MYB-bHLH-WD40 (MBW) complex with the stabilizing protein WD40, which in turn regulates the structural genes involved in the anthocyanin pathway and regulates the metabolic flux and accumulation of anthocyanins and PAs.

The Dual luciferase assay showed that PbMYB5 was not able to effectively activate the promoter activity of structural enzyme genes involved in the anthocyanin synthesis pathway ( Figures 6A, B ). The interaction between PbMYB5 and pro-PbANS was observed in a yeast one-hybrid assay ( Figure 6C ), suggesting that PbMYB5 may lack transcriptional activation abilities. To further support this idea, yeast self-activation assays were performed on PbMYB5, bHLH3/33, and WD40. It was found that, unlike bHLH33, PbMYB5, bHLH3, and WD40 did not exhibit self-activating properties ( Figure 7 ).

In the Pyrus bretschneideri genome (SAMN01797448 and SAMN12751541), PbMYB5 was predicted to be a transcriptional repressor. In this study, PbMYB5 was overexpressed in ‘Zaosu’ pear fruits and healing wounds and was found to promote reddening of the ‘Zaosu’ pear epidermis, promote anthocyanin accumulation, and regulate of PAs synthesis-related gene expression. Silencing the PbMYB5 gene in ‘Red Zaosu’ pear fruits and healing wounds inhibited reddening of the ‘Red Zaosu’ pear epidermis, suppressed anthocyanin accumulation. The PbMYB5 transcription factor promoted the accumulation of anthocyanins and regulated the accumulation of PAs in the pear epidermis, promoting the coloration of the pear epidermis. There are ongoing and well-defined studies on anthocyanin synthesis and metabolic regulation of fruit coloration. However, relatively little research has been done on transcriptional regulators that regulate crosstalk between different secondary metabolic pathways (e.g., anthocyanin, PAs, and lignin pathways). Photoinduction of apple pericarp promotes anthocyanin biosynthesis while reducing lignin production, indicating a negative correlation between anthocyanins and lignin and competition for metabolic flow (Hu et al., 2021). R2R3-MYB6 promotes the accumulation of anthocyanins and PAs and inhibits secondary metabolites like lignin during the growth and development of poplar trees (Wang et al., 2019a). The results of this study revealed that overexpression of PbMYB5 significantly increased the accumulation of red phenotypes in pear pericarp, decreasing the expression of the key enzyme genes for lignin synthesis, CCR, CAD1, and F5H and increasing the expression of the anthocyanin synthesis and transporter gene UFGT. Contrarily, silencing PbMYB5 prevented the accumulation of red phenotype in the pear pericarp. These results are consistent with those of MdMYB1, which promoted anthocyanin accumulation and inhibited lignin synthesis (Zhou et al., 2017).

VmMYBA1 and VmMYBA2 upregulated the expression of genes involved in the anthocyanin and PAs pathways in blueberries. Overexpression of MYB165 and MYB194 inhibited PAs and anthocyanin synthesis in poplar and regulated the synthesis and metabolism of other phenolic substances (Ma et al., 2018). Overexpression of VvMYC1 induced the accumulation of anthocyanins and PAs in grape berries. Furthermore, VvMYB5a overexpression increased the metabolism of anthocyanins, PAs, and lignin; VvMYB5b can induce the accumulation of anthocyanins and PAs derivatives, revealing a spatiotemporal specificity of VvMYB5a and VvMYB5b in regulating the accumulation of lignin, anthocyanins and PAs in grapevine fruits (Deluc et al., 2006, 2008). The R2R3-FaMYB5 TF regulates anthocyanin and PAs metabolism in strawberry fruit and transgenic tissues. The results of this study revealed that PbMYB5 expression was higher in pear fruits at 45 d after anthesis than at 105 and 115 d after anthesis and lower at 65 and 85 d after anthesis than at other times. These results are similar to those observed for grapes. PbMYB5 overexpression increased anthocyanin synthesis in pear pericarp, consistent with the results of MYB5 in strawberries and grapes. Nevertheless, in this study, PbMYB5 overexpression resulted in the suppression of lignin and PAs synthesis gene expression in pear pericarp, and a significant inhibition of PA synthesis in pear pericarp and healing tissues. These results revealed that PbMYB5 regulates the fluctuations of metabolic fluxes of different metabolic pathways and can affect pear skin color formation through metabolic flux competition.

Flavonoid pathways are generally influenced by the spatiotemporal expression specificity of their structural genes and, to a lesser extent, by regulation of the MBW complex (Gonzalez et al., 2008, 2010; Pu et al., 2021). In Arabidopsis, MYB113/114 forms a complex with TTG1 and bHLH to regulate the anthocyanin synthesis pathway in seedlings, whereas TT2-TT8-TTG1 regulates PAs in seeds (Gonzalez et al., 2008; Xu et al., 2013). R2R3-FaMYB5, FaEGL3, and FaLWD1/FaLWD1-like may form a new MBW complex that regulates flavonoid metabolism in strawberries, FaMYB9/FaMYB11-FabHLH3-FaTTG1 regulates the PAs pathway in fruits, and FaMYB10-FabHLH3/FabHLH33-FaTTG1 may regulate anthocyanin metabolism (Jiang et al., 2023). In apples, MdHB1 regulates anthocyanin synthesis by restricting MdMYB10, MdbHLH3, and MdTTG1 to the cytoplasm, indirectly repressing the transcription of MdDFR and MdUFGT (Jiang et al., 2017). A ternary complex formed by PyMYB10, PybHLH, and PyWD40 transcription factors regulates anthocyanin biosynthesis and accumulation in the ‘Yunnan Hong’ Pear (Cui et al., 2021). PpMYB114 and PpbHLH3 promote PpMYB114-induced PpUFGT expression and red color accumulation; anthocyanin accumulation is transcriptionally regulated by the MYB-bHLH-WD40 complex (Ni et al., 2019). In this study, we discovered that PbMYB5 could interact with bHLH3/bHLH33 but not with WD40. This result is consistent with previous reports that MYB transcription factors interact with bHLH3 to form MYB-bHLH and that WD40 acts as a stabilizing protein that interacts with bHLH, forming the MYB-bHLH-WD complex (Alabd et al., 2022; Cheng et al., 2014; Ni et al., 2021; Wang et al., 2019b; Ye et al., 2016). The spatiotemporal specificity of PbMYB5 contributes to the dynamics of PbMYB5-bHLH3/bHLH33-WD40 complex, which in turn inhibits the expression of lignogenic anthocyanin genes, reduces the flux of related metabolites, promotes the expression of anthocyanin-related genes and accumulation of anthocyanin glycosides, and influences the coloration of pear skin.

Changes in the expression of anthocyanin structural genes are controlled by transcriptional regulators. In strawberries, R2R3-FaMYB5 and FaMYB10 both promote the expression of most of the structural genes involved in the flavonoid biosynthetic pathway (PAL, C4H, 4CL-2, and F3’H), and R2R3-FaMYB5 can specifically regulate the PAs-associated gene LAR (Jiang et al., 2023). PpMYB17 positively regulates flavonoid biosynthesis in pear fruit by activating PpCHS, PpCHI, PpF3H, and PpFLS in the flavonoid biosynthetic pathway independently of the bHLH or WD40 cofactors in the MBW complex (Premathilake et al., 2020). Our findings suggest that PbMYB5 does not have inherent activation ability, but can specifically bind to the PbANS promoter to enhance anthocyanin production. It appears that bHLH33 may form complexes that activate PbMYB5 and promote PbANS gene expression. The slight discrepancy in results compared to previous studies may be due to the direct regulation of PbUFGT by the PbMYB5-bHLH-WD40 complex, as well as the transactivation of PbANS by PbMYB5 with the assistance of bHLH33, resulting in increased anthocyanidin synthesis and disruption of PbANS Regulation ( Figure 8 ).