The transcription factor MdBPC2 alters apple growth and promotes dwarfing by regulating auxin biosynthesis

Abstract

Auxin plays important roles throughout plant growth and development. However, the mechanisms of auxin regulation of plant structure are poorly understood. In this study, we identified a transcription factor (TF) of the BARLEY B RECOMBINANT/BASIC PENTACYSTEINE (BBR/BPC) family in apple (Malus × domestica), MdBPC2. It was highly expressed in dwarfing rootstocks, and it negatively regulated auxin biosynthesis. Overexpression of MdBPC2 in apple decreased plant height, altered leaf morphology, and inhibited root system development. These phenotypes were due to reduced auxin levels and were restored reversed after exogenous indole acetic acid (IAA) treatment. Silencing of MdBPC2 alone had no obvious phenotypic effect, while silencing both Class I and Class II BPCs in apple significantly increased auxin content in plants. Biochemical analysis demonstrated that MdBPC2 directly bound to the GAGA-rich element in the promoters of the auxin synthesis genes MdYUC2a and MdYUC6b, inhibiting their transcription and reducing auxin accumulation in MdBPC2 overexpression lines. Further studies established that MdBPC2 interacted with the polycomb group (PcG) protein LIKE HETEROCHROMATIN PROTEIN 1 (LHP1) to inhibit MdYUC2a and MdYUC6b expression via methylation of histone 3 lysine 27 (H3K27me3). Silencing MdLHP1 reversed the negative effect of MdBPC2 on auxin accumulation. Our results reveal a dwarfing mechanism in perennial woody plants involving control of auxin biosynthesis by a BPC transcription factor, suggesting its use for genetic improvement of apple rootstock.

IN A NUTSHELL

Background: Plant height is an agronomic trait that markedly affects crop yield. It is controlled by the environment, plant hormone metabolism, and signaling pathways. Transcription factors play a key role in these processes and have been proposed to be involved in the regulation of plant height. Although little is known about the underlying molecular mechanisms by which they regulate plant dwarfism, a BARLEY B RECOMBINANT/BASIC PENTACYSTEINE (BBR/BPC) transcription factor induces plant dwarfing in Arabidopsis and rice. In apple, the Class I BBR/BPC gene MdBPC2 is highly expressed in dwarfing rootstocks, and MdBPC2-overexpressing plants show a dwarf phenotype. However, the exact mechanism of how BPC transcription factors affect plant structure is not well understood.

Question: What is the role of MdBPC family members in regulating growth and development of apple plants, and how do they function?

Findings: We found that a BBR/BPC transcription factor gene, MdBPC2, was highly expressed in dwarfing rootstocks. When MdBPC2 was overexpressed, auxin content decreased, plant growth was inhibited, plant height decreased, and roots were underdeveloped. These growth defects were reversed by supplementation of exogenous auxin. MdBPC2 bound to the GAGA-rich element of the MdYUC2a and MdYUC6b promoters, inhibiting their expression. We also found that MdBPC2 interacted with MdLHP1 and recruited it to MdYUC2a and MdYUC6b gene loci to mediate the catalysis and maintenance of H3K27me3 methylation, enhancing chromatin condensation. This process led to the inhibition of MdYUC2a and MdYUC6b transcription, which limited auxin synthesis and ultimately stimulated dwarfism of apple plants.

Next steps: In view of the significant decrease of MdGH3.1 in MdBPC2-overexpressing transgenic plants, we will next focus on the mechanism of MdBPC2 involved in maintaining auxin homeostasis, which will deepen our understanding of how plants promote or inhibit their development in response to environmental changes, thereby enhancing their environmental adaptability.

Introduction

The breeding and use of dwarf and semidwarf varieties of rice (Oryza sativa) and wheat (Triticum aestivum) have greatly increased world food production. The primary way to control plant height has been by blocking the biological effect of the plant hormone gibberellic acid (GA; Peng et al. 1999; Sasaki et al. 2002). Recent studies confirmed that other hormones, including auxin, also play important roles in regulating the development of plant architecture (Li et al. 2008; Sazuka et al. 2009; Matthes et al. 2019). The rice auxin-defective mutant tdd1 showed significant auxin-deficient phenotypes such as wider leaf angles, inhibited root growth, and overall plant dwarfing (Sazuka et al. 2009). In contrast, overexpression of the auxin biosynthesis gene YUCCA in Arabidopsis (Arabidopsis thaliana) resulted in elevated auxin content, increased plant height, and decreased branching (Li et al. 2008).

In plants, tryptophan (Trp)-dependent auxin synthesis from indole-3-pyruvic acid (IPA) is the main pathway of auxin synthesis and is involved in major biological processes such as cell division and elongation, vascular differentiation, gravitropism, and phototropism (Mashiguchi et al. 2011). Transformation of the precursor Trp to indole acetic acid (IAA) is a 2-step catalytic reaction: first, a tryptophan aminotransferase (TAA) catalyzes the deamination of Trp to IPA (Stepanova et al. 2008), after which a YUCCA (YUC) flavin monooxygenase converts IPA into active auxin IAA through oxidative decarboxylation (Won et al. 2011). YUC genes encode the rate-limiting enzyme in auxin biosynthesis. Multiple mutants of YUC genes affect auxin accumulation and produce severe developmental defects, including abnormal embryo patterning, reduced plant height, root growth defects, vascular system abnormalities, apical meristem abnormalities, and changes in leaf and flower patterns (Cheng et al. 2006, 2007). The coordinated responses of auxin biosynthesis and inactivation due to specific growth stages and environments are mediated by direct regulation of the expression of auxin metabolic genes by both transcription factors (TFs) and epigenetic mechanisms and have been well studied in Arabidopsis (Casanova-Sáez et al. 2021). During lateral root formation, the B3 domain transcription factor FUSCA3 (FUS3) interacts with LEAFY COTYLEDON2 (LEC2) to coactivate the transcription of the auxin biosynthesis gene YUC4 (Tang et al. 2017). Indeterminate Domain (IDD) transcription factors IDD14, IDD15, and IDD16 promote auxin biosynthesis by directly targeting downstream YUC5 and TAA1, thereby regulating organ morphogenesis and gravitropic responses in plants (Cui et al. 2013). The synergistic feedforward regulation of YUC4 expression by AGAMOUS (AG) and CRABS CLAW (CRC) directs a precise change in chromatin state, inducing the shift from floral stem cell maintenance to gynecium formation (Yamaguchi et al. 2018). The UDP-glycosyltransferase UGT76F1 mediates hypocotyl growth in a light- and temperature-dependent manner by regulating active auxin levels. This process is directly regulated by Phytochrome-interacting Factors such as PIF4, a key integrator of light and temperature signaling pathways (Chen et al. 2020). Therefore, a detailed analysis of the regulation of auxin biosynthesis is essential for understanding the growth and development of plants and how they cope with biological and abiotic stresses.

The small, plant-specific family of BASIC PENTACYSTEINE (BPC) transcription factors includes members involved in regulation of various aspects of plant development, including embryo development, root growth, and flowering (Kooiker et al. 2005; Monfared et al. 2011; Wu et al. 2020). The 7 BPC family members in Arabidopsis can be grouped into 3 classes according to their N-terminal structure: BPC1/2/3 belong to Class I; BPC4/5/6 belongs to Class II (in which BPC5 may be a pseudogene); and BPC7 is the only member in Class III (Meister et al. 2004). Single BPC mutants showed no obvious phenotypes, Class II BPC double mutants showed little change compared with the wild type (WT), while Class I double and triple mutants showed some developmental defects. Severe phenotypes were observed only when multiple members of 2 groups of BPC genes were mutated, suggesting functional redundancy between the BPCs (Monfared et al. 2011; Simonini et al. 2012; Petrella et al. 2020). The BPC TFs have been shown to induce plant dwarfing. In rice, knockdown and knockout lines of OsBPC/GBP1 show enhanced seedling growth and grain length, whereas overexpression of OsGBP1 leads to plants with a dwarf stature and shortened grain length, indicating that OsGBP1 repressed grain length and seedling growth (Gong et al. 2008). Impairing both BPC1 and BPC2 function, in combination with either bpc4 or bpc6, resulted in shorter plants than were found when disrupting BPC4 and BPC6 in combination with either bpc1-1 or bpc2. This suggests that Class I genes are more important for plant height determination than are Class II genes (Monfared et al. 2011). However, the exact mechanism of how BPC TFs affect plant structure is not well understood.

In this study, we found that MdBPC2 expression was higher in dwarf rootstocks than in standard rootstocks. MdBPC2 interacted with polycomb group (PcG) proteins MdLHP1a/b to negatively regulate auxin biosynthesis. In MdBPC2-overexpressing transgenic apple (Malus × domestica) plants, inhibition of MdYUC2a and MdYUC6b genes resulted in decreased auxin accumulation, repressed plant growth, and altered tree structure. Exogenous auxin supplementation restored the height, leaf morphology, and root development of these transgenic apple plants. In WT plants, treatment with the antiauxin p-chlorophenoxyisobutyric acid (PCIB) inhibited auxin biosynthesis, resulting in phenotypes similar to those seen in MdBPC2-overexpressing transgenic apple plants. Our work suggests a mechanism for how MdBPC2 regulates auxin biosynthesis through H3K27me3 modification to coordinate the growth and plant structure of apple.

Results

Expression of MdBPCs in dwarf rootstocks and characterization of MdBPC2

To identify genes of the BPC family in apple, Arabidopsis BPC protein sequences were used as queries to search the Malus × domestica genome GDDH13 v 1.1 (Daccord et al. 2017). Six MdBPC candidate genes were identified in apple and were grouped into 2 classes in a phylogenetic tree along with the 7 Arabidopsis proteins (Supplemental Fig. S1 and Files S1 and S2). MD13G1015100 and MD16G1012800 sorted into Class I and were thus named MdBPC1 and MdBPC2, respectively. The sequences of MdBPC1 and MdBPC2 are highly similar, with a similarity of 94.8%. MD05G1054600, MD10G1062100, MD08G1016400, and MD15G1015400 belonged to Class II and were named MdBPC3, MdBPC4, MdBPC5, and MdBPC6. To determine if any of the MdBPCs could regulate plant growth, the expression profiles of the 6 MdBPC genes were compared in 5 dwarfing rootstocks (M9 T337, M9, M26, M9 Pajam 1, and JM7) and 5 standard rootstocks (Malus prunifolia, Malus micromalus, Malus halliana, Malus sieversii, and Malus hupehensis). The expression levels of the Class I BPCs (Supplemental Fig. S2), especially MdBPC2, were significantly higher in the dwarf rootstocks than in the standard rootstocks (Fig. 1A). Therefore, we selected MdBPC2 as a candidate gene for further study.

Figure 1.

Relative expression and characterization of MdBPC2. A) Relative expression levels of MdBPC2 in 5 standard apple rootstocks (M. prunifolia, M. micromalus, M. halliana, M. sieversii, and M. hupehensis) and 5 dwarfing rootstocks (M9T337, M9, M26, M9 Pajam 1, and JM7). MdActin was used for normalization. The expression level of MdBPC2 in M. prunifolia was set as 1. B) Relative expression of MdBPC2 in different tissues, including root, shoot, young leaf, mature leaf, stem tip, and developing fruits at 30, 60, and 90 d after blooming (DAB). The expression level of MdBPC2 in the root was set as 1. The error bars represent Se for 3 independent biological replicates, and each replicate measured 9 plants. C) Subcellular location of MdBPC2 in onion epidermal cells. GFP: green fluorescence; Bright field: visible light; Merge: visible light merged with green fluorescence; Pro35S:GFP: empty pCAMBIA2300-GFP vector; Pro35S:MdBPC2-GFP: recombinant pCAMBIA2300-MdBPC2-GFP vector. Scale bars = 5 µm. D) Transcriptional self-activation activity of MdBPC2 in yeast cells. Fusion vectors pGBKT7-MdBPC2 (MdBPC2-BD), negative control pGBKT7 vector (BD), and positive control pGAL4 vector were separately transformed into yeast strain AH109. Transformants were examined on SD/-Trp and SD/-Trp/-His/-Ade/+X-α-gal plates. E) Transcriptional repression activity of MdBPC2 protein in N. benthamiana leaves. The firefly LUC gene driven by 5 GAL4-binding elements and the minimum CaMV 35S promoter (TATA-box) was used as the reporter gene, the REN gene driven by the 35S promoter was used as the internal reference gene, and the VP16 transcriptional AD was used as the positive control. The GAL4 DNA binding domain (GAL4DBD) was used as the negative control, and the fusion protein of MdBPC2 fused to the N-terminus or C-terminus of VP16 was used as an effector. The 5 vectors carrying the effectors were cotransformed into N. benthamiana leaves along with the vectors carrying the reporter LUC. Three biological replicates were performed, the samples from 3 plants were as 1 replicate, and the error bars represent Se. Statistical tests were performed by one-way ANOVA followed by Tukey's test (P < 0.05). Different letters indicate significant differences.

We investigated the expression profile of MdBPC2 genes in different tissues by quantitative PCR (qPCR; Fig. 1B). The results showed that MdBPC2 was widely expressed throughout the plant: in roots, stems, leaves, stem tips, and developing fruits, with higher expression in stem tips. To determine the subcellular localization of MdBPC2, a MdBPC2-GFP fusion protein was transiently expressed from the cauliflower mosaic virus (CaMV) 35S promoter (Pro35S:MdBPC2-GFP) in onion (Allium cepa) epidermal cells transformed by gene gun-mediated transformation. Compared with the Pro35S:GFP control, Pro35S:MdBPC2-GFP signals were concentrated only in nuclei, indicating that MdBPC2 functions in the nucleus (Fig. 1C). When the transcriptional self-activating activity of MdBPC2 was analyzed in a yeast (Saccharomyces cerevisiae) 2-hybrid system, only the positive control pGAL4 grew well and showed galactosidase activity on SD/-Trp-His-Ade (+X-α-Gal) selection medium (Fig. 1D). Yeast cells carrying pBD-MdBPC2 or the negative control pGBKT7 did not survive on SD/-Trp-His-Ade (+X-α-Gal). These results suggest that MdBPC2 has no transcriptional activation activity, suggesting that MdBPC2 may be a transcriptional repressor. Then, we analyzed its transcriptional activity in plants using the luciferase (LUC) reporter system. The VP16 transcriptional activation domain (AD) was used as a positive control, and fusion proteins of MdBPC2 with VP16 fused to either the N-terminus or C-terminus were made as effectors to examine its repression function (Fig. 1E). Compared with the negative control GAL4BD, the transcriptional AD of VP16 alone significantly increased the LUC/REN ratio. However, the N-terminus or C-terminus of VP16 fused to MdBPC2 significantly inhibited the LUC activity promoted by VP16. These results showed that MdBPC2 had transcriptional inhibitory activity.

Identification of MdBPC2-overexpressing and RNAi transgenic apple lines and their effects on plant growth

To determine the role of MdBPC2 in intact apple trees, we constructed overexpression and RNA interference (RNAi) vectors and transformed them into apple variety ‘GL3'. PCR analysis showed that 3 overexpressing lines (MdBPC2-OE1/3/4) and 3 RNAi lines (MdBPC2-RNAi3/5/6) were obtained (Supplemental Fig. S3, A and B). Reverse transcription quantitative PCR (RT-qPCR) analysis showed that, compared to WT plants, MdBPC2 transcripts were approximately 15 times higher in transgenic MdBPC2-OE1/3/4 and about 60% lower in transgenic MdBPC2-RNAi3/5/6 plantlets (Supplemental Fig. S3, C and D). We also measured the expression levels of the other 5 MdBPC genes in the MdBPC2 transgenic apple plants. In MdBPC2-silenced lines, the transcript levels of MdBPC1 were also inhibited, possibly due to high sequence similarity, and the expression levels of the other 4 genes were not significantly affected (Supplemental Fig. S4).

In the MdBPC2 overexpression lines, plant growth was visibly inhibited (Fig. 2, A, B, and E). The overexpression of MdBPC2 significantly reduced average plant height, which was only 29% that of the WT plants (Fig. 2G). The internode lengths became shorter (Fig. 2H), and the stem diameters increased 1.25-fold (Fig. 2I). The longitudinal sections of the stem showed smaller cell lengths and increased numbers of cells per unit area (Supplemental Fig. S5). The leaf morphology also changed, with a decreased ratio of length/width (aspect ratio) and a more rounded shape (Fig. 2J). In addition, the root system in the overexpression plants was underdeveloped, with shortened root length (Fig. 2K) and reduced number of lateral roots (Fig. 2L). However, in MdBPC2-silenced lines (Fig. 2, C, D, F, and G to K), there were no significant phenotypic changes, possibly due to functional redundancy among the BPC family members, as was shown in Arabidopsis (Monfared et al. 2011).

Figure 2.

MdBPC2 inhibits plant growth. A to F) Phenotypes of MdBPC2-OE#1/3/4 and MdBPC2-RNAi#3/5/6 transgenic apple plants grown in soil for 3 mo after transformation and regeneration. Images show plant height A, C), leaf size, shape, and color B, D), and root structure E, F). Scale bars in A) and C) represent 10 cm; bars in B) and D to F) represent 5 cm. G to L) Analysis of plant height G), internode length H), stem diameter I), leaf aspect ratio J), total length of the root K), and lateral root number L) of the apple plants shown in A to F). The error bars represent Se. Three independent biological replicates were performed, and 10 plants were measured for each replicate. Student's t test was used to determine statistical significance relative to WT (*P < 0.05).

MdBPC2 altered auxin content

Phytohormones regulate plant growth and development, and dwarfism can be caused by changing the content or balance of endogenous hormones. Therefore, we measured the hormone contents, including auxin (IAA), cytokinin (CTK), GA, and brassinolide (BR) in our transgenic overexpression and RNA silencing lines. Compared with WT, the contents of CTK, GA, and BR were unchanged (Supplemental Fig. S6), while the IAA content was significantly reduced in the MdBPC2 overexpression lines (Fig. 3A). The IAA content in the MdBPC2-OE lines was 48.0% to 61.1% that of WT (Fig. 3A).

Figure 3.

Auxin content and IAA application in MdBPC2-OE transgenic apple plants. A) IAA content in stem tips of transgenic and WT apple plants. The error bars represent Se. Three independent biological replicates were performed, and samples from 11 plants were used as 1 replicate. Student's t test was used to determine statistical significance (*P < 0.05). B to D) Relative plant height B), leaf aspect ratio C), and lateral root number D) under the different treatments. Numbers in the x axis represent the concentration of IAA in milligrams per liter, and the values for the plants treated with 0 mg/L IAA were set to 1. The error bars represent Se for 3 independent biological replicates, with 10 plants measured for each replicate. Asterisks indicate significant differences relative to 0 mg/L IAA (Student's t test, **P < 0.01). E to J) The effect of exogenous IAA on the architecture of MdBPC2-OE plants. The transgenic plants and WT plants sprayed with exogenous IAA (0, 0.1, 0.2, 0.5, or 1 mg/L) for 1 mo, using both transgenic plants and WT plants without IAA treatment as controls. Bars, 8 cm. K) Expression levels of BPC genes in plants infected with virus carrying pTRV (set as 1), pTRV-MdBPC1#2, pTRV-MdBPC3#4, pTRV-MdBPC5#6, pTRV-MdBPC1#2#3#4, or pTRV-MdBPC1#2#3#4#5#6. Statistical tests were performed by 1-way ANOVA, followed by Tukey's test (P < 0.05). Different letters indicate significant differences. L) IAA concentrations in pTRV, pTRV-MdBPC1#2, pTRV-MdBPC3#4, pTRV-MdBPC5#6, pTRV-MdBPC1#2#3#4, and pTRV-MdBPC1#2#3#4#5#6 plants. The error bars represent Se. Three independent biological replicates, and 10 plants measured for each replicate. Student's t test was used to determine statistical significance relative to pTRV (*P < 0.05).

To further determine whether the MdBPC2-OE phenotype was caused by IAA deficiency, we treated MdBPC2-OE transgenic plants with exogenous IAA. Exogenously applied IAA restored plant height, leaf development, and root growth of MdBPC2-OE plants in a dose-dependent manner (Fig. 3, B to E, G, and I). Importantly, IAA promoted the growth of MdBPC2-OE plants, while the phenotype of WT plants was not affected by IAA treatment until growth was inhibited at the high concentration of 1 mg/L (Fig. 3, B to D, F, H, and J). This suggested that MdBPC2-OE plants were highly sensitive to the IAA concentration. Meanwhile, treatment of WT plants with increasing concentrations of PCIB, which inhibits auxin action (Oono et al. 2003), decreased plant height, altered leaf morphology, and dramatically decreased lateral root number in a dose-dependent manner. When the PCIB concentration reached 500 μM, the plant phenotypes were similar to those of MdBPC2-OE transgenic plants, indicating that a decrease in auxin activity was sufficient to cause MdBPC2-OE-like phenotypes (Supplemental Fig. S7). These results indicate that the overexpression of MdBPC2 resulted in a decrease of endogenous auxin content that inhibited the growth of transgenic plants.

Consistent with the phenotype of MdBPC2-RNAi plants, there was no significant change in IAA levels due to decreased expression of MdBPC2 (Fig. 3A). When Class I or Class II BPC genes were silenced by virus-induced gene silencing (VIGS; Fig. 3K), the IAA content of the transgenic lines increased significantly only when all 6 BPC genes were silenced at the same time (Fig. 3L). Thus, in apple, MdBPC2 and the other BPC genes show functional redundancy during auxin accumulation.

Identification of direct targets of MdBPC2

To determine how MdBPC2 overexpression alters endogenous IAA accumulation, we performed RNA sequencing (RNA-seq) using stem tips of WT and MdBPC2-OE1/3/4 plants. To identify differentially expressed genes (DEGs) between WT and MdBPC2-overexpressing apple plants, data were screened using the values of log2 (fold change) greater than 1.5 relative to WT and a threshold of P < 0.05 to identify DEGs. Among the DEGs identified, 433 upregulated genes and 587 downregulated genes (as compared to the WT) were common in all 3 lines overexpressing MdBPC2 (Fig. 4, A to C; Supplemental Data Sets S1 to S3). Genes known to be involved in auxin synthesis and catabolism (Supplemental Fig. S8 and Files S3 and S4; Song et al. 2020) were compared with the 1,020 DEGs of the MdBPC2-OE transgenic lines (Fig. 4D), revealing a total of 3 overlapping genes. These DEGs were MdYUC2a and MdYUC6b, both of which encode rate-limiting auxin biosynthesis enzymes, and MdGH3.1a, which is involved in maintaining auxin feedback regulation (Casanova-Sáez et al. 2021). MdYUC2a, MdYUC6b, and MdGH3.1a were downregulated in the 3 transgenic lines (Fig. 4E). To verify our RNA-seq expression profile data, these 3 genes were analyzed by qRT-PCR, and we found expression trends consistent with the transcriptome data (Fig. 4F).

Figure 4.

MdBPC2 regulates auxin biosynthesis genes. A) The numbers of genes showing differential expression in MdbBPC2-OE lines compared with WT as determined by RNA-seq. Venn diagram analysis of common upregulated B) and downregulated C) genes (| log2(fold change) | ≥ 1; P < 0.05) in MdBPC2 transgenic apple plants compared with WT plants. D) Venn diagram showing the overlap between the DEGs in MdbBPC2-OE and auxin metabolism genes. E) Heatmap showing the transcriptional abundance of the 3 auxin metabolism genes in the overlapping area in D). F) Relative expression levels of MdGH3.1a, MdYUC2a, and MdYUC6b in WT and MdBPC2-OE plants. The error bars represent Se. Three independent biological replicates were performed, and samples from 9 plants were used as 1 replicate. Student's t test was used to determine statistical significance (*P < 0.05). G) Binding motifs of MdBPC2 identified by MEME. The numbers near the elements indicate e value. H) EMSA results showing that MdBPC2 bound to a GAGA-rich motif. The biotin-labeled probe was the GAGA-rich motif, and the competitor was the onlabeled probe. MdBPC2-His represents a purified fusion protein. Labeled probe with His protein as negative control I) The binding peaks (Repeats 1 and 2) and negative control (Input) of MdBPC2 in MdYUC2a (−476 bp) and MdYUC6b (−196 bp) as determined by DAP-seq. The numbers at the left ([0 to 109]/[0 to 54]) refer to the heights of the peaks. The red arrow shows the binding peaks.

DNA affinity purification sequencing (DAP-seq) is a high-throughput method for screening transcription factor binding sites by analyzing the binding of genomic DNA to in vitro-expressed transcription factors. Compared with chromatin immunoprecipitation (ChIP)-seq, it has the advantages of being unrestricted by antibodies and species and is a high-throughput method (Bartlett et al. 2017). Therefore, we used DAP-seq to reveal genome-wide binding sites of MdBPC2. His-tagged MdBPC2 was expressed in vitro and incubated with a genomic DNA library purified from 3-moh-old ‘GL3' plants. All MdBPC2-bound DNA was isolated for high-throughput sequencing analysis. MEME suite analysis revealed that the core sequence of MdBPC2 binding regions was a GAGA-rich element, “GAGAGAGAGAGAGAG” (e value = 4.2e−4392; Fig. 4G). This GAGA-rich nucleotide sequence is more conserved than that of Arabidopsis BPC1 binding sites (C-box: “RGARAGRRA”; Kooiker et al. 2005; Simonini et al. 2012). Five additional potential binding sites of MdBPC2 were also highly enriched. Electrophoretic mobility shift assays (EMSAs) were performed to verify the reliability of the DAP-seq results. A purified MdBPC2-GST fusion protein was incubated with a biotin-labeled GAGA-rich DNA probe. Specific DNA-MdBPC2 protein complexes were detected when labeled GAGA-rich DNA probes were used, and this signaling was attenuated when a competing probe lacking the biotin label was added (Fig. 4H). This specific competitive combination suggests that MdBPC2 can specifically recognize GAGA-rich sequences. Next, we analyzed the binding peaks of the 3 auxin-related DEGs and found that only MdYUC2a and MdYUC6b had significant peaks on their promoters (Fig. 4I), while none were detected for MdGH3.1a (Supplemental Fig. S9). These results suggest that MdBPC2 inhibits auxin biosynthesis by suppressing MdYUC2a and MdYUC6b at the transcriptional level.

MdBPC2 inhibits transcription of MdYUC2a and MdYUC6b

To determine whether transcriptional activation of MdYUC2a and MdYUC6b was directly affected by MdBPC2, we monitored the bioluminescence signals of ProMdYUC2a:LUC and ProMdYUC6b:LUC constructs in a transient expression system in apple callus. When MdBPC2 was expressed in transgenic apple calli, the activity of LUC driven by either the MdYUC2a or the MdYUC6b promoter was significantly decreased (Fig. 5A). These results further suggested that MdBPC2 acts as a transcriptional inhibitor of MdYUC2a and MdYUC6b that negatively regulated auxin biosynthesis.

Figure 5.

MdBPC2 directly binds the promoters of MdYUC2a and MdYUC6b and inhibits their transcription. A) Transient expression assays in apple calli showing that the transcriptional activities of MdYUC2a and MdYUC6b were inhibited by MdBPC2. The upper parts show the effector and reporter constructs. ProMdYUC2a: LUC, ProMdYUC6b: LUC, and empty vector (pGreenII 0800-LUC) were cotransformed with either the MdBPC2 overexpression vector (Pro35S:MdBPC2) or an empty vector control (pGreenII 62-SK), respectively. At least 3 biological replicates were performed, and the samples from 3 transgenic calli lines were as 1 replicate. The error bars represent Se. Student's t test was used to determine statistical significance (***P < 0.001). B, C) Y1H assays showing that MdBPC2 directly bound to the GAGA-rich elements in the MdYUC2a and MdYUC6b promoters. The upper panels show schematic diagrams of the MdYUC2a and MdYUC6b promoters and the promoter fragments (P1 to P6) that were cloned into pAbAi. The numbers indicate the nucleotide positions relative to the ATG start codon. pGADT7 (AD-emp) was used as a negative control. D, E) ChIP-qPCR assay results showing that MdBPC2 bound to the promoters of MdYUC2a and MdYUC6b. The upper panels show the locations of the amplified fragments. The error bars represent Se. Three biological replicates were performed, and samples from 3 transgenic calli lines were as 1 replicate. Student's t test was used for statistical significance (***P < 0.001).

Analysis of promoter sequences of MdYUC2a and MdYUC6b revealed that there were 2 GAGA-rich elements in each promoter region: (GA)₂₄ and (GA)₃₂ in MdYUC2a and (GA)₁₁ and (GA)₂₂ in MdYUC6b (Fig. 5, B and C). To further investigate whether the MdBPC2 protein can bind to the GAGA-rich elements of the MdYUC2a and MdYUC6b promoters, we constructed a series of MdYUC2a and MdYUC6b promoter fragments (P1 to P6) and fused them into the pAbAi vector. The plasmids were transformed into yeast cells to produce the Bait-reporter strain. The MdBPC2 coding sequence was fused to the AD. Yeast 1-hybrid (Y1H) assays were used to detect binding of MdBPC2 to the promoter fragments. The results showed that when promoter regions containing GAGA-rich elements were coexpressed with MdBPC2-AD, transformed yeast cells grew well on SD/-Leu/AbA culture plates, indicating binding of MdBPC2 to these promoter regions. In contrast, both promoter regions lacking GAGA-rich elements and negative control plasmids failed to grow on SD/-Leu/AbA selective media (Fig. 5, B and C). These results demonstrate that all GA-rich elements in the MdYUC2a and MdYUC6b promoter regions can bind to the MdBPC2 protein.

ChIP-qPCR assays in transgenic apple calli showed that the amplicons of the (GA)₂₄ (−534 bp) and (GA)₃₂ (−178 bp) regions of the MdYUC2a promoter, and the (GA)₁₁ (−546 bp) and (GA)₂₂ (−281 bp) regions of the MdYUC6b promoter, accumulated significantly when immunoprecipitated using MdBPC2-GFP but not with GFP alone. DNA without a GAGA-rich element, namely (GA)₀, was also not significantly enriched in MdBPC2-GFP (Fig. 5, D and E). This further shows that MdBPC2 can bind to the MdYUC2a and MdYUC6b promoters via the GAGA-rich elements. These results indicate that MdYUC2a and MdYUC6b, key genes in auxin biosynthesis, are direct transcriptional targets of MdBPC2 and that their transcriptional activities are inhibited by MdBPC2.

MdBPC2 mediates enrichment of repressive histone trimethylation at H3K27

Posttranslational histone modification is associated with the activation or inhibition of gene transcription. Genome-wide occupancy studies of histone 3 lysine 27 trimethylation (H3K27me3) found that epigenetic mechanisms can regulate auxin-related genes (He et al. 2012). To detect whether H3K27me3 inhibitory markers are related to the inhibition of MdYUC2a and MdYUC6b by MdBPC2, we performed ChIP assays to explore potential differences of H3K27me3 inhibitory markers at MdYUC2a and MdYUC6b loci after MdBPC2 overexpression. Nine regions were tested, including the YUC2a/6b promoter region, the UTR region, introns, and exons. An anti-H3K27me3 antibody was used to precipitate the DNA, and qPCR was used for sequence analysis. Our results showed that H3K27me3 modification of MdYUC2a and MdYUC6b was significantly enhanced in the MdBPC2 overexpression lines, especially near the initiation codon (Fig. 6, A and B). As a control, the H3K27me3 levels of MdActin were unchanged in the transgenic plants (Supplemental Fig. S10B). These results suggested that MdBPC2 plays an important role in epigenetic modifications that inhibit the expression of MdYUC2a and MdYUC6b.

Figure 6.

MdBPC2 affects the H3K27me3 level at the MdYUC2a and MdYUC6b loci. A, B) ChIP detection of H3K27me3 levels at MdYUC2a/6b loci in WT, MdBPC2-OE, and MdBPC2-RNAi transgenic lines. The upper panels show schematic diagrams of MdYUC2a and MdYUC6b genome structure, with letters indicating amplicons for ChIP analysis. The genomic DNAs immunoprecipitated using the H3K27me3 antibody were analyzed by qPCR. The error bars represent the Se for 3 biological replicates, and 10 plants measured for each replicate. C) Y2H assays of the interactions of BPC2 with PcG complex subunits. pGBKT7-BPC2/pGADT7 was used as a negative control. 2DO, SD-Trp/-His; 4DO, SD-Trp/-Leu/-His/-Ade. D) BiFC visualization of the MdBPC2 and MdLHP1a/b interactions. Nicotiana benthamiana leaves cotransformed with constructs expressing nYFP-LHP1a or nYFP-LHP1b and cYFP-BPC2 showed YFP fluorescence in the nucleus. Leaves coexpressing either nYFP-LHP1a/b and empty vector cYFP-emp or cYFP-BPC2 and empty vector nYFP-emp were used as negative controls. Bars = 50 μm. E) LUC complementation visualization of the MdBPC2 and MdLHP1a/b interactions. Nicotiana bennetii leaves cotransformed with constructs expressing nLUC-LHP1a or nLUC-LHP1b and cLUC-BPC2 showed LUC chemiluminescence (top right). Leaves coexpressing nLUC-LHP1a/b and cLUC-emp (bottom right), cLUC-BPC2 and nLUC-emp (bottom left), or nLUC-emp and cLUC-emp (top left) were used as negative controls. Bars = 1 cm. F) Pull-down assays analyzing the interaction between MdBPC2 and MdLHP1a/b. Purified GST or MdLHP1a/b-GST were incubated with anti-His-conjugated magnetic beads containing MdBPC2-His. Immunoblotting analyses with anti-GST and anti-His antibodies were performed to detect MBPC2-His (to test pull-down efficiency) and LHP1a/b-GST (to test protein interactions).

BPC2 protein physically interacts with the PcG subunit LHP1a/b

In plants, the inhibitory epigenetic modification H3K27me3 is mainly recognized and catalyzed by PcG proteins (Mozgova and Hennig 2015). Because H3K27me3 inhibitory markers at MdYUC2a and MdYUC6b loci were significantly increased after overexpression of MdBPC2 (Fig. 6, A and B), we used a yeast 2-hybrid (Y2H) experiment to test for interaction of MdBPC2 with inhibitory epigenetic modification-related proteins, including MEDEDA (MEA), CURLY (CLF), SWINGER (SWN), EMBRYONIC FLOWER2 (EMF2), FERTILIZATION INDEPENDENT ENDOSPERM (FIE), and LIKE HETEROCHROMATIN PROTEIN1 (LHP1). Our results showed that MdBPC2 interacted with MdLHP1a/b but not with the other proteins we tested (Fig. 6C). The interaction between MdLHP1a/b and MdBPC2 was further evaluated by bimolecular fluorescence complementation (BiFC). We were able to reconstitute the activity of YFP in the nuclei of tobacco epidermis cells by coexpressing the fusion proteins nYFP-LHP1a/b and cYFP-BPC2 (Fig. 6D), demonstrating interaction of MdLHP1a/b with MdBPC2. We also performed LUC complementation assays in tobacco leaves simultaneously injected with cLUC-BPC2 and nLUC-LHP1a/b (Fig. 6E). Cells expressing both LHP1a or LHP1b and BPC2 showed strong fluorescence during transient transformation. In addition, we performed a pull-down experiment to verify the interaction between MdBPC2 and MdLHP1a/b. Recombinant His-MdBPC2 protein was pulled down by GST-MdLHP1a/b but not by GST alone (Fig. 6F), suggesting that MdBPC2 can also interact directly with MdLHP1a/b proteins in vitro. All of these results suggested that MdBPC2 can recruit the PcG complex to target genes via MdLHP1a/b.

LHP1 is involved in MdBPC2-mediated inhibition of YUC2a/6b expression

In animals, GAGA-motif binding factors (GAFs) recruit PcG proteins by binding to conserved sequence motifs of polycomb response elements (PREs) to influence gene expression via histone modification (Ogiyama et al. 2018). Because MdBPC2 can bind specifically to the GAGA sequence and there is a physical interaction between MdBPC2 and MdLHP1, we hypothesized that MdLHP1 plays a part in MdBPC2-induced inhibition of MdYUC2a and MdYUC6b expression. To verify this hypothesis, we first detected the levels of MdLHP1 enriched at the MdYUC2a and MdYUC6b loci using ChIP-qPCR. DNA was extracted from MdBPC2-OE, MdBPC2-RNAi, and WT plants that were expressing MdLHP1-Flag. In MdBPC2-RNAi and WT plants, enrichment of MdLHP1-Flag was detected in the analyzed region of the MdYUC2a and MdYUC6b loci (Fig. 7, C and D). However, significantly more MdLHP1-Flag protein was detected on MdYUC2a and MdYUC6b promoters in the MdBPC2-OE plants. Silencing of MdLHP1 in the MdBPC2-OE background significantly reduced the H3K27me3 level at MdYUC2a and MdYUC6b loci similar to that of WT plants (Fig. 7, E and F). These 2 experiments suggested that the change in H3K27me3 levels in MdYUC2a and MdYUC6b locus was caused by MdLHP1.

Figure 7.

MdBPC2 and MdLHP1 affect H3K27me3 levels at the MdYUC2a/6b loci. A, B) Schematic diagrams of MdYUC2a and MdYUC6b genome structures, with letters a to i indicating amplicons used for ChIP analysis. C, D) ChIP detection of MdLHP1 enrichment levels at the MdYUC2a and MdYUC6b loci in WT, MdBPC2-OE, and MdBPC2-RNAi seedlings overexpressing MdLHP1-Flag. The genomic DNAs immunoprecipitated using the Flag antibody were analyzed by qPCR. E, F) ChIP assays documented H3K27me3 levels at the MdYUC2a and MdYUC6b loci in WT, MdBPC2-OE, and MdBPC2-OE transgenic plants with VIG-silenced MdLHP1. The genomic DNA immunoprecipitated using H3K27me3 antibody was analyzed by qPCR. The error bars represent Se. Three biological replicates were performed, and samples from 10 plants were used as 1 replicate. G, H) Schematic diagram of the LUC reporter genes. The MdYUC2a and MdYUC6b promoters or their mutants were fused upstream of the LUC gene. Gray box, MdYUC2a/6b promoter. I, J) Relative levels of H3K27me3 at LUC sites driven by the MdYUC2a or MdYUC6b promoters containing intact or mutated GAGA-rich elements were detected by ChIP assays. ChIP assays were performed on 3 independent transgenic callus clumps for each structure. The genomic DNAs immunoprecipitated using the anti-H3K27me3 antibody were analyzed by qPCR. The error bars represent Se for 3 biological replicates, and samples from 10 plants were used as 1 replicate. Student's t test was used to determine statistical significance (***P < 0.001). MdActin was used as a negative control in the above experiments, as shown in Supplemental Fig. S8, C to E.

To test whether the enrichment of MdLHP1 at the MdYUC2a and MdYUC6b promoters requires the binding of MdBPC2, we generated 4 transgenic apple callus lines carrying the LUC reporter driven by MdYUC2a and MdYUC6b promoters containing intact or mutated GAGA-rich elements (ProMdYUC2a, ProMdYUC6b and mProMdYUC2a, mProMdYUC6b; Figure 7, G and H). ChIP was used to detect the relative level of H3K27me3 at the LUC gene locus in the transgenic lines. The results showed that transgenic lines containing mutated GAGA-rich elements had significantly lower levels of H3K27me3 compared to transgenic strains containing intact GAGA-rich elements (Fig. 7, I and J). Consistent with this, LUC expression driven by the mutated MdYUC2a and MdYUC6b promoters was increased (Supplemental Fig. S11). These results suggested that the MdBPC2 binding sites within the target promoter are necessary for MdLHP1-mediated H3K27me3 modification.

MdLHP1 regulates auxin biosynthetic genes

Our data suggested that in apple, MdBPC2 recruits the PcG protein subunit MdLHP1 to the MdYUC2a and MdYUC6b loci through a mechanism similar to the regulation of genes by the homologous animal GAGA-binding protein. It then inhibits the expression of MdYUC2a and MdYUC6b via H3K27me3 modification. To confirm the involvement of MdLHP1 in the regulation of auxin levels by MdBPC2, we overexpressed MdBPC2 in apple calli and obtained 3 transgenic lines, MdBPC2-GFP-OE1/2/3 (Fig. 8, A and B). The expression levels of MdYUC2a and MdYUC6b were significantly reduced in calli from these transgenic lines (Fig. 8C), consistent with our observations in transgenic plants.

Figure 8.

MdLHP1 regulates endogenous auxin content. A) GFP fluorescence in Pro35S:MdBPC2-GFP transgenic calli lines. B) Relative expression of MdBPC2 in overexpression callus lines, with WT ‘Orin’ serving as the control. Bar = 1 cm. C) Relative expression of MdYUC2a and MdYUC6b in transgenic apple callus lines (MdBPC2-GFP-OE1, MdBPC2-GFP-OE2, and MdBPC2-GFP-OE3). D) Relative expression of MdLHP1a/b in the transgenic apple callus lines (Pro35S:MdBPC2-GFP-OE1, Pro35S:MdBPC2-GFP-OE2, and Pro35S:MdBPC2-GFP-OE3 and Pro35S:MdBPC2-GFP#TRV:MdLHP1-OE1#TRV:LHP1, OE2#TRV:LHP1, and OE3#TRV:LHP1). The expression level for WT was set as 1. E) GUS staining and enzyme activity of Pro35S:MdBPC2-GFP and Pro35S:MdBPC2-GFP#pTRV-MdLHP1. The images at the top are representative of the stained calli for each construct. Three independent biological replicates were performed, and samples from 3 transgenic calli lines were used as 1 replicate. The error bars represent Se. Student’s t test was used for statistical significance (*P < 0.05, **P < 0.01).

DR5, a synthetic auxin response element, contains binding sites for numerous ARF transcription factors that respond to auxin signals. Therefore, reporter gene activity expressed under control of the DR5 promoter can reflect the endogenous auxin content of plants (Ulmasov et al. 1997). We analyzed auxin content in MdBPC2-GFP-OE1/2/3 transgenic apple calli using a synthetic ProDR5:GUS construct. As shown in Fig. 8E, GUS staining was significantly reduced in MdBPC2-GFP-OE1/2/3 calli compared with WT calli. When we silenced MdLHP1 (Fig. 8D), GUS activity recovered and was higher than in WT (Fig. 8E). Taken together, these data suggest that MdLHP1 is involved in the regulation of auxin biosynthesis by MdBPC2 during growth and development.

Discussion

Plant height is one of the most important agronomic traits of crops and horticultural crops. It limits the spatial arrangement of the various organs of plants and is closely related to yield and quality (Krost et al. 2013). Plant height is controlled by an interplay among the environment, plant hormone metabolism, and signaling pathways, and transcription factors play an important role in these processes (Jia et al. 2018; Zheng et al. 2018). Although the importance of transcription factors in regulation of plant height has long been proposed, little is known about the underlying molecular mechanisms of how they regulate plant dwarfism. In this study, we found that MdBPC2 regulates the transcription of MdYUC2a and MdYUC6b, inhibiting auxin biosynthesis and thus modulating plant structure.

MdBPC2 acts as a transcriptional suppressor to regulate plant structure

As plant-specific transcription factors, BPC proteins have been identified in a variety of land plants, such as barley (Hordeum vulgare), soybean (Glycine max), rice, potato (Solanum tuberosum), tomato (Solanum lycopersicum), cucumber (Cucumis sativus), peach (Prunus persica), and gladiolus (Gladiolus hybridus; Sangwan and O'Brian 2002; Mu et al. 2017; Lloret et al. 2021; Li et al. 2023). Seven BPC genes were cloned from different organs of Arabidopsis, indicating that BPC genes are widely expressed in the plant and not limited to a single developmental process (Meister et al. 2004). In higher-order mutants of BPC, both vegetative growth and reproductive growth are affected by the complete loss of BPC function (Monfared et al. 2011). In terms of plant height, complete loss of Class I BPC function showed more obvious phenotypes than did Class II BPC mutants, suggesting that Class I BPC genes might have a similar role in regulating plant structure and are more important than Class II BPC genes (Simonini and Kater 2014). In this study, we identified 6 BPC genes from apples with a high degree of sequence conservation with AtBPCs in Arabidopsis (Supplemental Fig. S1). Of these, the expression levels of MdBPC1 and MdBPC2 were increased in dwarf rootstocks (Fig. 1A; Supplemental Fig. S2), further revealing the importance of Class I BPC in regulating plant structure. At the same time, our data showed that the expression levels of MdBPC2 in all dwarf rootstocks tested were significantly higher than were levels in standard rootstocks. Therefore, MdBPC2 seems to play a major role in the regulation of plant height.

Analysis of subcellular localization and transcriptional activity indicated that MdBPC2 is a nucleus-localized transcription inhibitor (Fig. 1, B and C). Overexpression in ‘GL3' apple resulted in shortened internodes and reduced plant height. This observation further demonstrates that MdBPC2 plays an important role in regulating plant height. In addition, transgenic plants had poor root development and underdeveloped root morphology. The ability of this weak root system to absorb and assimilate nutrients may be weakened, which may be another important factor for the dwarf phenotype of the transgenic strains.

MdBPC2 inhibits auxin accumulation by regulating the expression of auxin biosynthesis genes

It is widely believed that IAA, CTK, GA, and BR are directly involved in plant dwarfism, and changes in their local synthesis, transport, and signaling can cause dwarf phenotypes (Clouse 1996; Hedden 2003; Werner et al. 2003; Gan et al. 2018). We did not detect significant changes in CTK, GA, and BR contents in MdBPC2 overexpression lines. However, compared with WT plants, the content of IAA in transgenic lines decreased by about 50% (Fig. 3A). Plant height, leaf morphology, and root development are all regulated by auxin concentration. The IAA-defective rice mutant tdd1 is dwarfed, with narrow leaves and abnormal flower development (Sazuka et al. 2009). In our study, low auxin concentrations reduced transgenic plant height, changed leaf morphology, and significantly reduced root number. Application of exogenous IAA to transgenic apple plants partially restored the normal phenotypes (Fig. 3, A to J). These results supported the notion that the dwarf phenotype of MdBPC2 transgenic apple is caused by auxin deficiency. One remarkable observation is that the silencing of MdBPC2 did not increase plant height or auxin content (Figs. 2 and 3A), as reported in Arabidopsis; single BPC mutants did not have distinct phenotypes (Monfared et al. 2011). It is difficult to obtain higher-order mutants of BPC in apples, but we were able to silence the expression of both Class I and Class II BPC genes by VIGS. The significant increase in auxin levels of pTRV-MdBPC1#2#3#4 and pTRV-MdBPC1#2#3#4#5#6 transgenic lines implies that redundancy between Class I and Class II BPC genes is also present in apples (Fig. 3L).

During plant growth, auxin levels are regulated by biosynthesis, metabolism, and inactivation (Casanova-Saez et al. 2021). The downregulation of MdYUC2a and MdYUC6b in MdBPC2 overexpression lines (Fig. 4E), together with the direct binding of MdBPC2 to MdYUC2a and MdYUC6b promoters, suggests that both MdYUC2a and MdYUC6b are directly repressed by MdBPC2 (Fig. 5, B and C). We also showed that GAGA-rich elements in the promoters of MdYUC2a and MdYUC6b can be specifically recognized by MdBPC2 in apple (Fig. 5, D and E), confirming that MdYUC2a and MdYUC6b are major targets of MdBPC2 and that their inhibition is responsible for the decrease of auxin in MdBPC2-OE plant. Interestingly, we also found auxin local synthesis gene GH3.1a among the potential targets of MdBPC2 (Fig. 4E), suggesting that MdBPC2 function may involve auxin feedback regulation and intracellular auxin homeostasis.

MdBPC2-inhibited expression of MdYUC2a and MdYUC6b is related to histone modification

A variety of covalent modifications on lysine, arginine, serine, and other residues at the N-terminus of histone proteins confer epigenetic regulation. Histone modification directly affects the association of DNA and histone proteins into nucleosomes, affecting the degree of chromatin relaxation and regulating gene expression (Bannister and Kouzarides 2011; Tan et al. 2016). One important histone modification is H3K27me3, which is involved in maintaining gene silencing in a chromatin region. H3K27me3 regulates the expression of many development-related genes and is involved in several developmental processes such as vernalization, flower morphogenesis, and seed formation (Kohler and Villar 2008; Zheng and Chen 2011). In Arabidopsis, genes involved in auxin biosynthesis, metabolism, transport, and signal transduction are tightly regulated by H3K27me3 modification, allowing the plant to complete the developmental transition under ideal growth conditions (Lafos et al. 2011; He et al. 2012). Here, we found that MdBPC2 bound to the MdYUC2a and MdYUC6b promoters and inhibited their expression. The H3K27me3 levels of MdYUC2a and MdYUC6b in MdBPC2-OE plants were significantly higher than in WT plants (Fig. 6, A and B). Therefore, we speculated that H3K27me3 modification markers are involved in the regulation of YUC expression and are related to MdBPC2.

The PcG proteins are transcriptional repressors identified in Drosophila melanogaster that function by forming 2 core polycomb repressive complexes, PRC1 and PRC2. BPC TFs can recruit PRC2 and trigger H3K27me3 modification by binding to GAGA-rich elements in promoters (Mu et al. 2017; Wu et al. 2020; Li et al. 2023). Although MdBPC2 did not physically interact with any PRC2 component, it did bind the plant-specific PcG protein MdLHP1 (Fig. 6C). LHP1 is homologous to animal Heterochromatin Protein 1 (HP1), but its N-terminal chromo domain (CD) can bind to H3K27me3 to inhibit euchromatin genes, making it functionally homologous to PcG (Turck et al. 2007; Zhang et al. 2007). Derkacheva et al. (2013) found that LHP1 interacts with the PRC2 component MSI1 and recruits PRC2 to chromatin carrying H3K27me3, and genome-wide transcription spectroscopy showed that LHP1 and the core subunit CLF of PRC2 have very similar effects. These data confirm that LHP1 is involved in the occurrence and maintenance of H3K27me3 in plants. The interaction between MdBPC2 and MdLHP1 appears to be essential for H3K27me3 modification at the MdYUC2a and MdYUC6b loci based on the following observations: (i) the levels of MdLHP1 enriched at the MdYUC2a and MdYUC6b promoters were increased significantly in overexpressing MdBPC2 transgenic plants; (ii) silencing MdLHP1 expression in MdBPC2-OE plants rapidly decreased H3K27me3 at MdYUC2a and MdYUC6b; and (ii) in the absence of MdBPC2, MdLHP1 could not catalyze H3K27me3 on the MdYUC2a and MdYUC6b loci.

Our study shows that MdBPC2 interacts with MdLHP1 to negatively regulate the expression of 2 YUC genes. This result was further verified in overexpressing MdBPC2 apple calli. In MdBPC2-GFP-OE transgenic apple calli, the auxin content of the calli recovered after the expression of MdLHP1 was reduced (Fig. 8E). Therefore, we propose a possible model in which MdBPC2 regulates apple growth by controlling auxin content (Fig. 9). In our model, the Class I BBR/BPC protein MdBPC2 recruits the PcG component MdLHP1 to the GAGA motif in the promoter of the auxin synthesis genes MdYUC2a and MdYUC6b. MdLHP1 is involved in the catalysis and maintenance of H3K27me3, enhancing the condensed state of chromatin. When MdBPC2 is overexpressed, this process is more extensive, resulting in severe inhibition of MdYUC2a and MdYUC6b transcription, restricted auxin synthesis, and ultimately affecting apple plant height, leaf growth, and root development.

Figure 9.

A model of the protein complex formed to repress the expression of the auxin biosynthesis genes YUC2a/6b during apple growth. During normal expression (top), BPC2 binds to the GAGA-rich element in the YUC2a/6b promoters and recruits LHP1, which promotes methylation at the YUC2a/6b loci. This leads to chromatin agglutination, which causes a slight inhibition of gene transcription. Overexpression of BPC2 (bottom) recruits more LHP1 and causes a higher degree of chromatin agglutination, resulting in more intensive inhibition that restricts the accumulation of auxin, dramatically affecting growth and changing plant structure.

In conclusion, this study showed that MdYUC2a and MdYUC6b are the targets of H3K27me3 modification mediated by MdBPC2-MdLHP1, which regulates auxin biosynthesis during apple growth. Our findings not only provide important genetic evidence for the function of MdBPC2 TF but also help to unravel the regulatory mechanisms of plant dwarfism. Further understanding of the mechanisms that maintain auxin homeostasis during plant growth will deepen our understanding of how plants promote or inhibit their development in response to environmental changes, thereby enhancing their environmental adaptability.

Materials and methods

Plant materials and growth conditions

M. prunifolia, M. micromalus, M. halliana, M. sieversii, M. hupehensis, and dwarfing rootstocks (M9T337, M9, M26, M9 Pajam 1, and JM7) were planted at the Apple Experimental Station of Northwest A&F University in Yangling, Shaanxi Province, China (34°N, 108°E). WT ‘GL3' and MdBPC2 transgenic apple seedlings (Malus domestica Borkh. cv. Royal Gala) were grown on MS medium (M519-100L, PhytoTechnology Laboratories, USA) containing 0.3 mg/L 6-BA (A600743-0025, Sangon Biotech, China) and 0.2 mg/L IAA (A600723-0025, Sangon Biotech, China) for 4 wk and were then transferred to MS rooting medium containing 0.5 mg/L IBA (A600725-0025, Sangon Biotech, China) and 0.5 mg/L IAA. After rooting, genotyped plants were transferred to plastic pots filled with a mixture of soil (JZ-20, Yuanye Biologica, China), vermiculite (gyzs123, Aoqi, China), and organic fertilizer (320L, Biotopped, China; 3:1:1) and grown at 24 °C under a 16-h/8-h light/dark photoperiod supplemented with a nature light intensity of 60 μmol/m²/s for 30 d. After 3 mo of plant growth, physiological indicators were determined, and the stem tips and fully mature leaves were sampled, immediately snap frozen in liquid nitrogen, and stored at −80 °C until further analysis.

Apple calli (Malus × domestica cv. ‘Orin’) were grown on MS medium containing 1.0 mg/L 6-BA and 1.0 mg/L 2,4-D at 26 °C under dark conditions and were subcultured every 2 wk.

Gene cloning

Total RNA was extracted from ‘GL3' leaves by using the cetyltrimethylammonium bromide (CTAB) method (Gasic et al. 2004). A set of specific primers (Supplemental Data Set S4) was used to clone the complementary DNAs (cDNAs) of MdLHP1 and PRC2-related genes. MdBPC2 is highly homologous to MdBPC1, with a sequence identity of 94.8%. These 2 genes do not contain introns. To obtain the full-length coding sequence of MdBPC2, we designed a specific forward primer in the promoter region of the gene and used “gala” DNA as a template for PCR amplification. Then, the full-length primers of the gene were designed, and the PCR product obtained by the above steps was used as the amplification template.

Gene expression analyses

Total RNA was extracted from leaves by a modified CTAB method (Gasic et al. 2004) and treated with DNaseI (B43, Thermo, USA). cDNAs synthesized using PrimeScript RT reagent (TaKaRa, http://www.takarabiomed.com.cn) and 2x Fast qPCR Master Mixture (di-ning, https://di-ning.com.cn/) were used in qRT-PCR reactions that were conducted on an iQ5 multicolor real-time PCR detection system (Bio-Rad). The data were analyzed using iQ5 2.0 standard optical system analysis software and the ΔΔCT method (Livak and Schmittgen 2001).

Vector construction and plant transformation

Gateway cloning technology (Landy 1989) was used to introduce the coding region of MdBPC2 into the pGWB411 (74805, Addgene, USA) binary vector to obtain an overexpression vector. An MdBPC2-specific fragment was cloned into the pK7GWIWG2D(II) (1_29, VIB-UGent Center for Plant Systems Biology, Belgium) vector using Gateway cloning technology to create the silencing vector. Each recombinant plasmid was transformed into Agrobacterium tumefaciens strain EHA105 using chemical transformation according to the manufacturer’s instructions (AC1010, Weidi, China). The plasmids were subsequently introduced into the leaves of ‘GL3' by using the method described by Tian et al. (2023). The coding sequence of MdBPC2 without the stop codon was inserted into the pMDC83 binary vector to form a Pro35S:MdBPC2-GFP vector. The recombinant plasmid was introduced into ‘Orin’ apple calli by Agrobacterium-mediated transformation (Li et al. 2016).

To achieve simultaneous silencing of MdBPCs and MdLHP1, specific fragments were cloned into pTRV2 to generate pTRV2-MdBPCs or pTRV2-MdLHP1, which were independently transformed into Agrobacterium strain GV3101 for VIGS (Zhu et al. 2021). Sequences of all primers used are listed in Supplemental Data Set S4.

Gene gun-mediated transformation

Fresh onion epidermis (1 cm²) was spread on MS medium and cultured at 28 °C in the dark for 24 h. Gold powder (1 μm, 1652263, BIO-RA, USA) was washed with 70% ethanol and sterile ddH2O, respectively, and sterilized ddH2O was added to make gold powder suspension. Under continuous vortex oscillation 5 μL Pro35S:MdBPC2-GFP plasmid (1 μg/μL) 50 μL, 2.5 M CaCl₂, and 20 μL 0.1 M spermidine were sequentially added to 50 μL gold powder suspension to prepare DNA-coated gold powder particles. The onion epidermal cells were bombarded with 10 μL of DNA-coated microparticle suspension using a gene gun (PDS-1000, BIO-RAD, USA). The pressure of the helium tank was 1,300 psi during bombardment. The bombarded onion epidermal cells were cultured at 25 °C in the dark for 24 h and observed using a laser scanning confocal microscope (Leica, Wetzlar, Germany). GFP signals were excited at 488 nm and optimally detected at 498 to 545 nm with a laser value of approximately 8% and a master gain value of around 700.

Quantification of auxin

The content of IAA was measured by high performance liquid chromatography (Müller and Munné-Bosch 2011). For IAA quantification, the stem tips of 90-d-old WT and transgenic apple plants were quick-frozen and ground in liquid nitrogen. Each sample (0.1 g) was oscillated in 1 mL of ethyl acetate at 2,000 rpm for 10 min, and the supernatant was removed into a new centrifuge tube after centrifugation at 4 °C and 12,000 × g. The organic solvent was removed by evaporation using a Termovap (NDK200-2N, MIULAB, Hangzhou, China) sample concentrator. The IAA content was analyzed using HPLC-MS (QTRAP 5500, AB SCIEX, USA); 150 mm × 4.6 mm, 5 μm packed chromatographic column (InsertSustainTM AQ-C18, Shimadzu, Japan) was used. The flow rate was 0.3 mL/min, and the column temperature is set at 25 °C. Mobile phase A consisted of 0.1% formic acid (85178, Thermo, USA), and mobile phase B consisted of methyl alcohol (67561, Thermo, USA).

IAA and PCIB treatment

After rooting, the ‘GL3' and MdBPC2-OE transgenic apple plants were transplanted into the substrate (soil:vermiculite:perlite 4:1:1) and cultured in the greenhouse for 3 mo before being treated with auxin. For IAA treatment, ‘GL3' and MdBPC2-OE transgenic plants were divided into 5 groups, and each group was sprayed with IAA at different concentrations (0, 0.1, 0.2, 0.5, and 1 mg/L) once every 3 d. For PCIB treatment, ‘GL3' seedlings were divided into 5 groups, and each group was sprayed with PCIB at different concentrations (0, 50, 100, 200, and 500 μM) once every 3 d. One month after treatment, the plants and organs were photographed, and relevant physiological indicators were measured, including plant height, leaf size, and number of lateral roots.

RNA-seq and data analysis

The polysaccharide polyphenol plant total RNA extraction kit (Tiangen, Beijing, China) was used to extract total RNA from the leaves of WT plants and MdBPC2-OE transgenic plants. Oligo dT magnetic beads (61005, Thermo, USA) were used to specifically bind to the poly(A) tail of messenger RNA (mRNA) to remove other RNA. The mRNA was fragmented in NEB Fragmentation Buffer (E6186AVIAL, New England Biolabs, USA) according to the manufacturer’s instructions. A library was constructed according to the method of Parkhomchuk et al. (2009). After the quality of each library was inspected by Agilent 2100 Bioanalyzer (G293BA, Agilent, USA), 1 μL purified PCR products was analyzed by Agilent DNA 1000 Kit (5067-1504, Agilent, USA). The different libraries were pooled according to the requirements of the machine used for Illumina sequencing (Illumina, San Diego, CA). Fastp software (Chen et al. 2018) was used to process raw reads in fastq format, and clean reads were obtained after removing low-quality reads for subsequent data analysis. The clean reads were mapped to the Malus Genome GDDH13 Version 1.1 (Daccord et al. 2017) using HISAT2 (Kim et al. 2015). FPKM (Roberts et al. 2011) of each gene was calculated, and the read counts of each gene were obtained by HTSeq-count (Anders et al. 2015). Differential expression analysis was performed using the DESeq2 (Love et al. 2014). P < 0.05 and log2 (fold change) >1.5 was set as the threshold for significantly DEGs. For sequencing, 3 sequencing biological replicates were used for the MdBPC2 overexpression lines and WT plants, and samples from 5 plants were used as 1 replicate.

DAP-seq sampling and data analysis

DAP-seq analysis was conducted according to a described method (O'Malley et al. 2016) and performed at Bluescape Hebei Biotechnology Company. Leaf genomic DNA (gDNA) was extracted from 90-d-old ‘GL3' apple seedlings using the CTAB method. The gDNA was segmented by Covaris M220 (Covaris, USA) into fragments of approximately 100 to 400 bp and then constructed into a library using a NEXTFLEX Rapid DNA Seq Kit (PerkinElmer, Inc., Austin, TX, USA). The coding sequence of MdBPC2 was cloned into the pFN19K HaloTag T7 SP6 flexxi vector, expressed using the TNT SP6-coupled wheat germ extraction system (Promega, Madison, WI, USA), and captured using Magne Halo Tag Beads (Promega). The purified protein was incubated with the gDNA library. DNA was eluted from 2 technical replicates as well as from the library incubated with beads but without added protein as a negative control. The captured DNA was sequenced on an Illumina NavoSeq6000. Bowtie2 (Langmead and Salzberg 2012) was used to map the reads to the Malus Genome GDDH13 Version 1.1 (Daccord et al. 2017). MACS2 callpeak (Zhang et al. 2023 ) and IDR software (Li et al. 2011) were used to merge the peaks of the 2 biological duplicates with P < 0.05 and to score the reliability of these repeated peaks. HOMER (Heinz et al. 2010) was used to annotate the bound peaks. Conserved motifs in the peak regions were identified using MEME software (Machanick and Bailey 2011).

Y1H and dual-LUC reporter assays

Yeast (S. cerevisiae) one-hybrid assays were performed using the Matchmaker Gold Y1H system according to the manual (TaKaRa, https://www.takarabiomed.com.cn). In brief, different fragments of the MdYUC2a and MdYUC6b promoters were cloned into the pAbAi vector and integrated into the YIHGold strain by homologous recombination to generate bait-specific reporter strains. The minimum inhibitory concentrations for each bait strain were determined on SD/-Ura plates containing different concentrations of aureomycin A (AbA). The full-length coding sequence of MdBPC2 was inserted into the pGADT7-Rec vector as the prey vector. MdBPC2-pGADT7-Rec was transformed into bait-specific reporter strains according to the method described in the manufacturer’s protocol. The transformed yeast cells were spread onto SD/-Leu/AbA plates and cultured at 30 °C for 3 d.

For the dual-LUC assays, the transcriptional regulatory elements of MdYUC2a and MdYUC6b were cloned upstream of the firefly LUC gene in the pGreenII 0800-LUC vector (V010545, NovoPro, China) to generate the reporter gene construct. The MdBPC2 coding region was cloned into the pGreenII 62-SK vector to generate the effector construct. The reporter vector and effector vectors were separately transformed into A. tumefaciens GV3101 (Psoup-p19; AC1003, Weidi, China), selected positive clone, then coinfected the reporter and effector strain into apple calli. The apple calli were cultured under dark conditions for 24 to 48 h and sampled using a Promega Chemiluminescence detector (Promega, GM2010, USA) to detect the activities of firefly LUC and Renilla luciferase (REN) for subsequent calculation of the LUC/REN ratio. Chemiluminescence imaging was performed using a live plant imaging system (PlantView100, Biolight, China). Three biological replicates were established for each dual-LUC assay, and samples from 3 transgenic calli lines were used as one replicate.

EMSA and ChIP-qPCR

For the EMSAs, we first obtained purified MdBPC2-His recombinant protein according to the instructions for Ni-NTA Bind Resin (7sea Biotech, Shanghai, China). The probe sequence with or without 5′-terminal biotin labeling was 5′-CTCTCTCTCTCTCTCTCTCTC-3′, and the mutant probe sequence was 5′-CACACACACACACACACACAC-3′. EMSA was performed with MdBPC2-His recombinant protein and biotin-labeled probe according to instructions provided in the LightShift Chemiluminescence EMSA Kit (Thermo, MA, USA). In brief, 2 μg of MdBPC2-His recombinant protein was incubated in EMSA binding buffer for 5 min in an ice bath followed by addition of 1 μL of labeled probe or mutant probe and incubation for 30 min at room temperature. The incubated samples were electrophoresed on a 4% nondenaturing polyacrylamide gel and then transferred to an Amersham Hybond-N+ membrane. Membranes were analyzed using an Alliance Q9 Advanced imaging system (UVITEC, Cambridge, UK).

The ChIP-qPCR assay was performed according to the method of Ma et al. (2022), with slight modification. In brief, MdBPC2-GFP transgenic apple calli were cross-linked in 1% formaldehyde under vacuum for 20 min, and then the reaction was terminated with 0.25 M glycine. The ChIP assay was performed using the BeyoChIP Enzymatic ChIP Assay Kit (Protein A/G Magnetic Beads; Beyotime, Shanghai, China) using an anti-GFP antibody (1:2,000 dilution, 50430-2-AP, Proteintech, USA). The amount of immunoprecipitated chromatin was quantified by qPCR using the appropriate corresponding primers (Supplemental Data Set S4). Three biological replicates were established for each ChIP-qPCR assay, and samples from 3 transgenic calli lines were as 1 replicate.

Luciferase complementation assay and BiFC

The luciferase complementation assay (LCA) was performed according to a published protocol (Chen et al. 2008). The CDS of MdBPC2 and MdLHP1a/b without stop codons was cloned into LAC plasmids containing Cluc and Nluc coding sequences: Pcambia1300-Cluc (PVT14650, lifescience-market, China) and Pcambia1300-Nluc (PVT14649, lifescience-market, China), and then the recombinant plasmids were transferred into A. tumefaciens strain GV3101. Agrobacterium cells containing the 2 plasmids were injected into 4- to 6-wk-old tobacco (Nicotiana benthamiana) leaves. Leaf tissues expressing test proteins were collected, and the intensity of bioluminescence was detected using a live plant imaging system (PlantView100, Biolight, China). Three biological replicates were established for each LCA assay, and samples from 3 plants were used as 1 replicate.

In BiFC assays, the CDS of MdBPC2 and MdLHP1a/b without stop codons was cloned into BiFC vectors containing YFP N-terminal or C-terminal fragments: Pspyne-35S/Puc-SPYNE (KL-35S2-EA, KeLei Biological, China) and Pspyce-35S/Puc-SPYCE (KL-35S1-EA, KeLei Biological, China). The recombinant plasmids were then transferred into A. tumefaciens strain GV3101. Agrobacterium containing these 2 plasmids were injected into tobacco leaves and expressed for 3 d so that the recombinant DNA could be delivered to plant cells and expressed stably. The YFP signals of the recombinant protein were detected using a laser scanning confocal microscope (Leica, Wetzlar, Germany). YFP signals were excited at 514 nm and optimally detected at 520 to 570 nm with a laser value of approximately 8% and a master gain value of around 700.

Y2H and His pull-down assays

Y2H assays were performed as described by Su et al. (2023). The MdBPC2 coding region was cloned into the PGBKT7 vector (P2135, NovoPro, China) to generate the bait protein. The recombinant plasmid was transferred into yeast strain AH109 to detect self-activation of the bait protein on SD-Trp/-His/-Ade culture plate. Then the coding sequences of MdLHP1a/b without the stop codons were cloned into Y2H vector Pgadt7 to generate prey proteins. Bait protein and prey protein constructs were cotransferred into AH109 yeast, and cells were spread onto SD-Trp/-Leu culture plates and grown for 2 to 3 d. Four single colonies were selected from each selection plate and spotted onto SD-Trp/-Leu, SD-Trp/-Leu/-His, and SD-Trp/-Leu/-His/-Ade synthetic dropout plates. The growth of colonies was observed after 3 d of inverted culture in an incubator at 30 °C.

For the pull-down experiment, we first obtained purified recombinant proteins of MdBPC2-His and MdLHP1a/b-GST using Ni-NTA Bind Resin. Purified recombinant His-tagged protein was added into the prepared Anti-His magnetic beads (P2135, Beyotime, China) and incubated overnight at 4 °C. The magnetic beads were held in the tubes using a magnetic rack and then washed with 1 × TBS (10 mM Tris-HCl, pH 7.6, 150 mM NaCl) buffer solution. The magnetic beads were then resuspended in binding buffer (50 Mm Tris-HCl, pH 7.6, 200 mM NaCl, 1 mM EDTA-Na₂·2H₂O, 1% NP-40, 1 mM DTT, and 10 mM MgCl₂·6H₂O) after which purified GST-tagged recombinant protein was added and incubated at room temperature for 2 h. After centrifugation at low temperature (4 °C, 1,000 × g, 2 min), the magnetic beads were washed 5 times with binding buffer. Finally, the magnetic beads were used for SDS-PAGE electrophoretic analysis.

H3K27me3 histone modification analysis

The anti-H3K27Me3 rabbit polyclonal antibodies (1:2,000 dilution, 39535, Proteintech, USA) were used for immunoprecipitation. Leaves (3 g) were collected from WT ‘GL3' apple, MdBPC2-OE, and MdBPC2-RNAi transgenic lines. ChIP-qPCR was performed according to the method described above. The input and immunoprecipitated samples were analyzed by real-time fluorescence qPCR (Bustin 2005), and the input percentage was calculated. The relative enrichment of MdActin7 was used as a negative control for H3K27me3 markers.

GUS staining

The GUS activity analysis was performed according to the method described by Zhu et al. (2023). Different transgenic apple calli lines were incubated in GUS staining solution (0.1 M Tris-HCl [pH 7.0], 0.05 M NaCl, 0.01% [v/v] Triton X-100, 0.5 mM K₃[Fe(CN)₆], 0.5 mM K₄[Fe(CN)₆], 0.01 M EDTA, 0.1% (v/v) glutaraldehyde, and 1 mg/mL X-Gluc] at 37 °C for 16 h. The calli were photographed with a digital camera (EOS 750D, Canon, Japan).

Statistical analysis

All data were analyzed using IBM SPSS Statistics 27 and graphed with Sigma Plot 12.5 software. Data were analyzed using independent t tests or 1-way ANOVA tests, with a significance level accepted at P < 0.05. Results of detailed statistical analyses are presented in Supplemental Data Set S5.

Accession numbers

Sequence data from this article can be found in the Arabidopsis (At) Genome Initiative, the Malus × domestica (Md) genome (GDDH13 v1.1) database, and GenBank databases under the following accession numbers: AtBPC1 (AT2G01930), AtBPC2 (AT1G14685), AtBPC3 (AT1G68120), AtBPC4 (AT2G21240), AtBPC5 (AT4G38910), AtBPC6 (AT5G42520), AtBPC7 (AT2G35550), MdBPC1 (MD13G1015100), MdBPC2 (MD16G1012800), MdBPC3 (MD05G1054600), MdBPC4 (MD10G1062100), MdBPC5 (MD08G1016400), MdBPC6 (MD15G1015400), MdYUC2a (MD15G1184800), MdYUC6b (MD08G1119300), MdGH3.1a (MD02G1180300), MdCLFa (MD02G117500), MdCLFb (MD15G1285900), MdSWNa (MD00G1179500), MdSWNb (D031294100), MdEMF2a (MD03G1252000), MdEMF2b (MD11G1274700), MdEMF2c (MD11G1274400), MdVNR2a (MD06G1050700), MdVNR2b (MD04G1057300), MdFIEa (MD11G1227100), MdFIEb (MD03G1211700), MdLHP1a (MD02G1231000), MdLHP1b (MD07G1081800), and MdActin7 (MD01G10016000).

Acknowledgments

We are grateful to Dr. Jing Zhang and Miss Wenjing Cao (Horticulture Science Research Center, Northwest A&F University, Yangling, China) for providing professional technical assistance.

Author contributions

H.Z. and M.L. designed this research. H.Z., S.W., Y.H., X.L., T.J., Z.Z., and B.M. performed the experiments. H.Z., L.Z., and M.L. analyzed the data. H.Z. and M.L. wrote the manuscript. M.L. and F.M. supervised the study. All authors read and approved the final manuscript.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Phylogenetic analysis of BPC proteins from apple (Malus × domestica) and A. thaliana.

Supplemental Figure S2. Relative expression levels of MdBPCs in 5 standard rootstocks (M. prunifolia, M. micromalus, M. halliana, M. sieversii, and M. hupehensis) and 5 dwarf rootstocks (M9T337, M9, M26, M9 Pajam 1, and JM7).

Supplemental Figure S3. Identification of MdBPC2 transgenic apples.

Supplemental Figure S4. Expression levels of other BPC family members in MdBPC2 overexpression, MdBPC2-silenced transgenic, and WT ‘GL3' apple plants.

Supplemental Figure S5. Longitudinal sections of stems stained with safranin. The second internodes from ‘GL3' and transgenic apple plants were used to prepare sections.

Supplemental Figure S6. Contents of CTK, BR, and GA in transgenic and WT apple plants.

Supplemental Figure S7. The phenotype of ‘GL3' plants treated with the auxin inhibitor PCIB was close to that of MdBPC2-OE transgenic plants.

Supplemental Figure S8. Phylogenetic analysis of the main genes for auxin metabolism from apple (Malus × domestica) and A. thaliana.

Supplemental Figure S9. The binding peaks (Repeats 1 and 2) of MdBPC2 and negative control (mock) in MdGH3.1a by DAP-seq.

Supplemental Figure S10. Control experiments establishing ChIP specificity in this study.

Supplemental Figure S11. Expression and activity of LUC gene driven by intact and mutated promoters of YUC2a/6b.

Supplemental Data Set S1. Annotation of DEGs in WT plants and the MdBPC2-OE1 transgenic line.

Supplemental Data Set S2. Annotation of DEGs in WT plants and the MdBPC2-OE3 transgenic line.

Supplemental Data Set S3. Annotation of DEGs in WT plants and the MdBPC2-OE4 transgenic line.

Supplemental Data Set S4. List of primers used in this study.

Supplemental Data Set S5. Detailed statistical analysis in this study.

Supplemental File S1. BPC protein sequences used to generate the phylogenetic tree shown in Supplemental Fig. S1.

Supplemental File S2. Phylogenetic tree of BPC proteins in Newick format.

Supplemental File S3. Protein sequences used to generate the phylogenetic tree of the main genes for auxin metabolism shown in Supplemental Fig. S8.

Supplemental File S4. Phylogenetic tree of the main genes for auxin metabolism in Newick format.

Funding

This work was supported by the Program for the National Natural Science Foundation of China (31872043) to M.L., the Shaanxi Science and Technology Innovation Team Project (2022TD-18) to M.L., the China Postdoctoral Science Foundation (2023T160536) to L.Z., and the Earmarked Fund for the China Agriculture Research System (CARS-27) to F.M.

Data availability

All data are incorporated into the article and its online supplementary material.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• BPC5 Gramene: AT4G38910

• BPC5 Araport: AT4G38910

• BPC7 Gramene: AT2G35550

• BPC7 Araport: AT2G35550

• BPC6 Gramene: AT5G42520

• BPC6 Araport: AT5G42520

• YUC5 Gramene: AT5G43890

• YUC5 Araport: AT5G43890

• MSI1 Gramene: AT5G58230

• MSI1 Araport: AT5G58230

• YUC4 Gramene: AT5G11320

• YUC4 Araport: AT5G11320

• BPC4 Gramene: AT2G21240

• BPC4 Araport: AT2G21240

• TAA1 Gramene: At1g70560

• TAA1 Araport: At1g70560

• BPC2 Gramene: AT1G14685

• BPC2 Araport: AT1G14685

• h3k27me3 CHEBI: CHEBI:85045

Author notes

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plcell) is: Mingjun Li ([email protected]).

Conflict of interest statement. None declared.