PbrNAC34a- PbrMYB3/65- PbrACO2 cascade plays a role in citrate difference between the pericarp and cortex tissues of pear (P. bretschneideri Rehd.) fruit

Abstract

See More

Citrate is critical to the flavor of horticultural fruit and governed by ACO. However, the specific ACO and its upstream regulators involved in citrate metabolism during pear (Pyrus spp.) fruit development remained uncharacterized. This study identified and characterized six PbrACOs from the Pyrus bretschneideri Rehd. genome. Comprehensive analyses of citrate levels, cyt/mitACO activities, and PbrACOs expression profiles in the pericarp and cortex tissues of developing 'Yali' and 'Dangshansuli' fruits revealed PbrACO2 as a candidate gene. Subsequently, PbrACO2 was confirmed as a mitochondrial aconitase catalyzing citrate-to-isocitrate conversion in vitro and in vivo. Analysis of differentially expressed transcription factors (TFs) and cis-acting elements in the PbrACO2 promoter identified nuclear PbrMYB3 and PbrMYB65, derived from whole genome duplication/segmental duplication, as candidate upstream regulators. These MYB TFs, without direct relationship, bound, as monomers, to the same two MYB-binding sites in the PbrACO2 promoter to activate its transcription, thereby promoting citrate isomerization in pear and tomato. Further investigation revealed that PbrMYB3 and PbrMYB65 are transcriptionally regulated by PbrNAC34a. Given their tissue-dependent expression profiles, the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade partially accounts for citrate differences between pear fruit pericarp and cortex tissues. These findings enhance understanding of citrate accumulation in Rosaceae fruit and provide genetic resources for pear breeding.

Core

PbrMYB3 and PbrMYB65 could bind to the same two MYB-binding sites in PbrACO2 promoter as monomer to activate its transcription, and thus promote citrate isomerization in pear and tomato. Further study revealed that the expression of these two MYB TFswas under the control of PbrNAC34a. Considering their tissue-dependent expression profiles, the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade was partly responsible for citrate difference between the pericarp and cortex tissues of pear fruit.

Gene and accession numbers

Sequence data in this article could be retrieved from the pear genome database (http://peargenome.njau.edu.cn/) under the accession numbers: PbrNAC34a (Pbr026635.1), PbrMYB3 (Pbr003370.1), PbrMYB65 (Pbr000749.2), PbrACO2 (Pbr039718.1), and PbrWRKY72 (Pbr009294.1).

Introduction

Pear (Pyrus spp.), an economically significant temperate fruit originating from southwestern China, is cultivated globally and valued for its distinctive organoleptic properties (Li et al., 2022b; Niu et al., 2024; Qian et al. 2021). Together with soluble sugars, malate and citrate serve as the primary organic acids contributing to pear fruit flavor formation (Wu et al., 2022). During fruit development of P. bretschneideri Rehd. cv. 'Dangshansuli', malate levels exhibited a general declining trend, while citrate demonstrated an inverse pattern (Zhang et al., 2021). Similar patterns were observed during the maturation of P. pyrifolia cv. 'Yandangxueli' and 'Gengtouqing' fruits (Lu et al. 2011).

Citrate biosynthesis in horticultural fruit begins with the β-carboxylation of phosphoenolpyruvate (PEP) to oxaloacetate (OAA), catalyzed by cytosolic phosphoenolpyruvate carboxylase (PEPC) (Fig. S1) (Etienne et al. 2013; Tahjib-Ul-Arif et al. 2021). Subsequently, mitochondrial citrate synthase (CS), the first enzyme in the tricarboxylic acid (TCA) cycle, catalyzes the addition of an acetyl group from acetyl-CoA to OAA, generating citrate (Fig. S1) (Etienne et al. 2013; Tahjib-Ul-Arif et al. 2021). However, studies have not identified correlations between CS activity (or PEPC mRNA abundance) and citrate content during fruit maturation or among low-/high-citrate cultivars (Hussain et al. 2017; Roongruangsri et al. 2012). Consequently, citrate levels in developing fruit may be regulated by the downstream steps of citrate biosynthesis (Hussain et al. 2017).

Following synthesis, citrate undergoes sequential conversion by aconitase (ACO) and isocitrate dehydrogenase (IDH) into α-ketoglutarate, with isocitrate as an intermediate (Fig. S1) (Etienne et al. 2013; Tahjib-Ul-Arif et al. 2021). Plant ACOs, present in cytosol (cytACOs) and/or mitochondria (mitACOs) (Fig. S1) (Etienne et al. 2013; Tahjib-Ul-Arif et al. 2021; Wang et al. 2016), are classified into five subfamilies (Terol et al. 2010). For IDH, two types of isozymes have been identified in higher plants: the mitochondrial NAD⁺-dependent isozymes (NAD⁺-IDHs) and the NADP⁺-dependent isozymes (NADP⁺-IDHs). These isozymes are distributed across the cytosol, chloroplast, peroxisome, and mitochondrion (Fig. S1) (Gálvez and Gadal 1995; Gálvez et al. 1999; Hussain et al. 2017); with cytosolic NADP⁺-IDHs reportedly accounting for the majority of IDH activity in angiosperm and gymnosperm species (Pascual et al., 2018).

During the early developmental stages of Citrus limon cv. 'Eureka' fruit and C. reticulata cv. 'Sai Num Phueng' and 'See Thong' fruits, citrate accumulation corresponded with reduced mitACO and NAD⁺-IDH activities (Sadka et al. 2000a, b); conversely, increased activities of cytACO and NADP⁺-IDH appeared responsible for decreased citrate levels in later growth stages of these varieties (Roongruangsri et al. 2012; Sadka et al. 2000a). Additional research demonstrated that CitACO3 expression levels were inversely correlated with citrate abundance in late developmental stages of C. reticulata Blanco cv. 'Ponkan' fruit (Li et al. 2017). This pattern was also observed during the maturation of C. sinensis cv. 'Newhall' and 'Skaggs Bonanza' fruits: citrate content progressively decreased, showing negative correlations with CitACO3 and CitIDH1 mRNA levels (Chen et al. 2013). Furthermore, compared to C. sinensis cv. 'Anliu' fruit, lower citrate levels in its bud mutant (cv. 'Hong Anliu') fruit development may be attributed to enhanced cytACO and NAD⁺/NADP⁺-IDH activities and upregulated expression of ACO1-3, NAD⁺-IDH1-3, and NADP⁺-IDH1/3 (Guo et al. 2016). These findings suggest ACO and IDH are involved in citrate metabolism during fruit development. However, minimal citrate level changes were observed in plant tissues following antisense inhibition of a mitochondrial NAD⁺-IDH 1 gene (SlIDH1) or a cytosolic NADP⁺-IDH 1 gene (SlICDH1) from Solanum lycopersicum (Sienkiewicz-Porzucek et al. 2010; Sulpice et al. 2010) or overexpression of a cytosolic NADP⁺-IDH gene from Pinus pinaster (Pascual et al., 2018). Consequently, researchers have paid more attention to ACO's role in citrate metabolism during fruit maturation.

ACO, an iron-sulfur enzyme containing a 4 Fe-4S cluster, catalyzes the reversible isomerization of citrate to isocitrate with cis-aconitate as an intermediate (Wang et al. 2016). Plant ACOs typically exhibit a negative regulatory effect on citrate accumulation (Hussain et al. 2017). Overexpression of CitACO3 reduced citrate accumulation in 'Ponkan' fruit (Li et al. 2017). Similar results were observed in OsACO1-overexpressing Oryza sativa (Senoura et al. 2020). Conversely, antisense suppression of SlACO1 from S. lycopersicum or knockdown of OsACO1 in rice decreased carbon flux through the TCA cycle, thereby enhancing citrate accumulation in these plants (Carrari et al. 2003; Senoura et al. 2020). Beyond regulating carbon status and flux, ACOs contribute to other physiological and developmental processes, including oxidative stress response and cell death (Wang et al. 2016).

The formation of horticultural fruit quality is regulated by transcription factors (TFs) through their binding to corresponding cis-acting elements in structural gene promoters (Jia et al. 2023). For example, PuWRKY31 from P. ussuriensis cv. 'Nanguo' binds to the W-box element (core motif, TTGACC/T) in the promoter of Sugars Will Eventually be Exported Transporter 15 gene (PuSWEET15) to induce its expression, resulting in soluble sugar accumulation in fruit (Li et al. 2020). Similarly, PyHY5, a bZIP TF, enhances anthocyanin accumulation in P. pyrifolia cv. 'Yunhongli No. 1' fruit by directly interacting with the G-box element (core motif, CACGTG) in the promoters of anthocyanin-biosynthesis-related genes (PyMYB10 and PyWD40), thereby initiating their transcription (Wang et al., 2020). Recent research has identified several TFs involved in citrate metabolism regulation in horticultural plants. AcNAC1 interacts with the core binding element in aluminum-activated malate transporter 1 gene (AcALMT1) promoter to activate its expression, leading to citrate accumulation during fruit maturation in Actinidia chinensis cv. 'Hongyang' and A. deliciosa cv 'Hayward' (Fu et al. 2023). Conversely, CitNAC62 collaborates with CitWRKY1 in citrate degradation through upregulation of CitACO3 transcription (Li et al. 2017). However, the molecular mechanism of citrate metabolism during pear fruit development remains incompletely understood.

This study examined the function of PbrACO2 and its upstream regulators, including PbrMYB3, PbrMYB65, and PbrNAC34a, in citrate metabolism within the pericarp and cortex tissues during fruit maturation of P. bretschneideri Rehd. The findings revealed that the tissue-specific expression pattern of the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade likely accounts for the differential citrate concentrations between pear fruit pericarp and cortex tissues.

Results

Evolution and characteristics of plant ACOs

A total of 66 ACOs, which emerged in the plant kingdom approximately 1,160 million years ago (MYA) (Fig. S2A), were identified from 18 plant species, and could be categorized into five subgroups (A, B, C, D & E) (Fig. S2B; Table S2). Five horticultural plant species were subsequently selected for detailed analyses.

As shown in Fig. S2B and Table S3, P. bretschneideri, Prunus persica, Musa acuminata, Vitis vinifera, and A. chinensis contained six, three, six, five, and five ACOs, respectively. These genes originated from whole genome duplication (WGD)/segmental and dispersed duplications, and were distributed across five, three, five, four, and four chromosomes in pear, peach, banana, grape, and kiwifruit genomes, respectively, exhibiting diverse physio-biochemical characteristics (Fig. S3A-B; Table S3-S4). 88 cis-acting elements were identified from their promoters, which were classified into eight functional categories (Fig. S3C). Comparative analyses of upstream regions of the paralogous gene pairs revealed promoter divergence (Fig. S4). This variation might lead to distinct regulatory mechanisms, thereby enhancing plant adaptation to various developmental processes and (a)biotic stresses (Qiao et al. 2015). Furthermore, at least nine exons were present in their genomic sequences, except for PbrACO6, which contained only one exon (Fig. S3D). Additionally, five (Motif 1, 3, 7, 8, and 11) out of thirteen motifs were conserved among the 'Aconitase' domains in 25 ACOs from five horticultural plant species (note: at least one motif was detected in each 'Aconitase' domain) (Fig. S3E; Table S5).

Subsequently, we analyzed the expression profiles of PbrACOs in several different tissues and their responses to light and temperature treatments. As shown in Fig. S5 and Table S6, except for PbrACO6, all other members exhibited expression across six tissues of 'Yali' pear, including 15-DAFB fruit, petal, stem, leaf, ovary, and stigma, with tissue-specific expression patterns. Furthermore, PbrACO1-5 demonstrated varying transcriptional responses to light exposure and temperature treatment. As shown in Fig. S6A and Table S7, light exposure enhanced PbrACO2 expression while suppressing PbrACO3 expression (Fig. S6A; Table S7); the other three members showed no significant changes after light treatment (Fig. S6A; Table S7). Similarly, PbrACO5 mRNA levels progressively decreased with temperature elevation (from 0 ℃ to 25 ℃ to 53 ℃), while low temperature (0 ℃) inhibited PbrACO3 transcription (Fig. S6B; Table S8); meanwhile, PbrACO1, PbrACO2, and PbrACO4 transcripts remained stable after high/low temperature treatment (Fig. S6B; Table S8).

Characterization of PbrACO2 as the candidate gene involved in citrate isomerization in the developing pear fruit

To explore the molecular mechanism of citrate metabolism, we initially analyzed the accumulation patterns of organic acids in the pericarp and cortex tissues during the (un)bagged 'Yali' fruit development. As illustrated in Fig. 1A, Fig. S7-S8, and Table S9, several organic acids, including oxalate, tartarate, malate, and shikimate, showed decreasing trends in both tissues concurrent with increasing citrate percentage and citrate-to-malate/oxalate ratio; conversely, citrate demonstrated an increasing accumulation trend (Fig. 1A; Table S9). Bagging treatment had minimal impact on the compositions and accumulation patterns of organic acids throughout fruit maturation (Fig. 1A; Table S9). However, the cortex tissue of the (un)bagged developing fruit contained higher citrate levels than the pericarp tissue, accompanied by an increased citrate-to-malate ratio (Fig. 1A; Fig. S8; Table S9).

Fig. 1

Dynamic change of citrate metabolism during the (un)bagged ‘Yali’ fruit development. A Organic acid content. Color scale represents normalized log2-transformed (mean value of three biological replicates + 1), where red, blue, and white colors indicate high, low, and medium expression levels, respectively. B PbrACOs expression profiles. Data, adapted from transcriptome assay, represent the mean value of three biological replicates; and color scale represents normalized log2-transformed (mean FPKM + 1), where red, blue, and white colors indicate high, low, and medium expression levels, respectively. C Correlations among attributes. Spearman correlations among different attributes are visualized as a heatmap, in which negative correlations are represented in blue color and positive in red color. ‘Yali’ pear were bagged with triple-layer paper bags at 34 DAFB, while the unbagged fruit at the same positions were labelled as well. Pericarp and cortex tissues were sampled at six developmental stages, including 15 DAFB, 34 DAFB, 81 DAFB, 110 DAFB, 145 DAFB, and 160 DAFB

Subsequently, we sought to identify the candidate PbrACO responsible for the aforementioned phenomenon. As shown in Fig. 1B and Table S10, five PbrACOs were expressed during fruit maturation, exhibiting distinct expression patterns. Among these, PbrACO2 expression levels were consistently higher in the pericarp tissue of the (un)bagged developing fruit compared to the cortex tissue (fold change ≥ 2.0 and FRD < 0.01; Fig. 1B; Table S10). Conversely, PbrACO1 and PbrACO5 showed opposite patterns (Fig. 1B; Table S10). Further analyses revealed strong positive correlations between citrate content and mRNA abundances of PbrACO1 (correlation coefficients = 0.75) and PbrACO5 (correlation coefficients = 0.64), but a negative correlation with PbrACO2 expression level (correlation coefficient = -0.76) (Fig. 1C). Similar to the developing 'Yali' fruit, PbrACO2 was the only member whose cortex tissue expression demonstrated positive correlations with cytACO and mitACO activities (coefficient = 0.46 and 0.08) but negative association with citrate content (coefficient = -0.41) during 'Dangshansuli' fruit development (Fig. S9 and Table S9-S10). Real-time quantitative polymerase chain reaction (RT-qPCR) assay results confirmed the accuracy of transcriptome findings regarding PbrACOs expression patterns during P. bretschneideri Rehd. fruit maturation (Fig. S10).

Collectively, these results suggested that PbrACO2, whose CDS and protein sequences showed high identity between the two cultivars (Fig. S11A and S12A), might be the candidate gene involved in citrate isomerization, resulting in differential citrate contents in the pericarp and cortex tissues of developing P. bretschneideri Rehd. fruit.

Functional validation of PbrACO2

As shown in Fig. 2A, PbrACO2-GFP exhibited the identical subcellular localization to the mitochondrial marker MSTP-mcherry (Sun et al. 2022) in Arabidopsis protoplasts, indicating its mitochondrial localization.

Fig. 2

Functional validation of PbrACO2. A Subcellular localization of PbrACO2. MSTP-mcherry was used as the mitochondrial marker (Sun et al. 2022). Bar, 10 μm. B Michaelis–Menten curve for citrate conversion by PbrACO2 in vitro. C Binding model of PbrACO2 with citrate. The green dotted line indicates the hydrogen bond interaction. D Functional validation of PbrACO2 in pear calli. (D-i) Growth status of pear calli. (D-ii) PbrACO2 expression level and mitACO activity. (D-iii) Chromatogram of sample. (D-iv) Organic acid abundance. Calli transformed with the empty vector was used as the control; and PbrACO2 expression level in the control calli is set as 1.0 for RT-qPCR assay. Data represent mean value ± standard deviation (SD) of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05)

To evaluate its catalytic properties, the recombinant His-PbrACO2 protein was obtained through prokaryotic expression. In vitro experiments demonstrated that PbrACO2 catalyzed the conversion of citrate, with K_m and V_max of 0.63 mM and 43.70 mmol s^-1 kg^-1 protein, respectively (Fig. 2B). The Cys³⁰⁷, Arg⁴⁰⁵, Asp⁴⁰⁶, and Trp⁴⁰⁷ residues in PbrACO2 potentially interact with citrate through hydrogen bonds, with a docking score of -5.04 kcal mol^-1, contributing to the catalytic reaction (Fig. 2C; Fig. S13; Table S11).

Subsequently, we examined the role of PbrACO2 in cellular citrate metabolism. As illustrated in Fig. S14A, transient overexpression of PbrACO2 in pear fruit significantly increased mitACO activity, inhibiting citrate accumulation; conversely, the opposite effect was observed in PbrACO2-silenced fruit (Fig. S14B). To validate these findings, PbrACO2-overexpressing pear calli and tomato fruit with stable inheritance were generated. Overexpression of PbrACO2 in pear calli enhanced mitACO activity, resulting in lower citrate levels compared to control calli; furthermore, quininate and shikimate formation was inhibited in the overexpressing calli, while malate and oxalate content remained unchanged among samples (Fig. 2D). Comparable results were observed in PbrACO2-transgenic tomato fruit, which exhibited increased mitACO activity but decreased citrate content (Fig. S15A), without affecting malate metabolism (data not shown). Additionally, color transition was suppressed in the transgenic tomato fruit (Fig. S15A-i).

Identification of PbrMYB3 and PbrMYB65 as the candidate upstream regulators of PbrACO2

The expression of structural genes is transcriptionally regulated by upstream TFs through interaction with related cis-acting elements in their promoters (Jia et al. 2023). Based on transcriptome analysis, 50 TFs, whose mRNA abundances in the pericarp tissue were consistently higher (46) or lower (4) than those in the cortex tissue of 'Yali' fruit, were identified from the pear genome (fold change ≥ 2.0 and FRD < 0.01; Fig. S16A-B; Table S12). Among these, the expression levels of Pbr003370.1 and Pbr000749.2 showed strong positive correlations with PbrACO2 transcript abundance during P. bretschneideri Rehd. fruit development (correlation coefficient > 0.8; Fig. S16B). Following Cao et al. (2016), Pbr003370.1 and Pbr000749.2 were designated as PbrMYB3 and PbrMYB65, respectively. Using the PlantCARE database, two MYB-binding sites (MYBCORE box motifs, CAACCG) were identified in the PbrACO2 promoter, both predicted to interact with PbrMYB3 and PbrMYB65 according to PlantRegMap database analysis (Fig. S16C). These findings suggest that these two MYB TFs may act as potential upstream regulators of PbrACO2. Further analysis revealed high sequence identity in their CDS and protein sequences between 'Yali' and 'Dangshansuli' fruits (Fig. S11B-C and S12B-C).

PbrMYB65 bound to the PbrACO2 promoter and then activated its expression

Based on its highest correlation coefficient with PbrACO2 as P. bretschneideri Rehd. fruit matured (correlation coefficient = 0.84; Fig. S16B), PbrMYB65 was selected for further investigation. As illustrated in Fig. 3A, a 160% increase in Dual-luciferase/Renilla (LUC/REN) ratio was detected in tobacco leaf co-transformed with PbrMYB65 and reporter containing PbrACO2 promoter compared to the control, with this increase correlating positively to the number of potential PbrMYB65-binding sites; however, the enhanced LUC/REN ratio was eliminated following mutation of the two potential binding sites (CAACCG → CTTCCG). In the yeast one-hybrid (Y1H) assay, the positive control (pGADT7-p53 & p53-pAbAi), negative controls (pGADT7 & PbrACO2pro^S1-pAbAi, pGADT7 & PbrACO2pro^S2-pAbAi), bait-prey co-transformants (pGADT7-PbrMYB65 & PbrACO2pro^S1-pAbAi, pGADT7-PbrMYB65 & PbrACO2pro^S2-pAbAi), and co-transformants containing the mutated elements (pGADT7 & PbrACO2pro^{S1 mut}-pAbAi, pGADT7 & PbrACO2pro^{S2 mut}-pAbAi, pGADT7-PbrMYB65 & PbrACO2pro^{S1 mut}-pAbAi, pGADT7-PbrMYB65 & PbrACO2pro^{S2 mut}-pAbAi) exhibited normal growth on SD/-Leu medium (Fig. 3B). Upon addition of Aureobasidin A (AbA), growth was inhibited in negative controls and transformants containing mutated binding elements, while the positive control and bait-prey co-transformants remained unaffected (Fig. 3B). Chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) analysis revealed more than three-fold enrichment in PbrACO2 promoter fragments containing the PbrMYB65-binding sites compared to controls (Fig. 3C). In vitro electrophoretic mobility shift assay (EMSA) demonstrated formation of protein-DNA complexes when His-PbrMYB65 was incubated with labeled probes, with binding gradually diminishing as unlabeled competitor probe concentrations increased (Fig. 3D); nevertheless, these complexes disappeared following mutation of the PbrMYB65-binding sites (Fig. 3D).

Fig. 3

Confirmation of PbrMYB65 as the upstream regulator of PbrACO2. A Dual-luciferase assay. PbrMYB65 CDS was introduced into the pSAK277 vector, while PbrACO2 promoter fragments of different lengths, containing different numbers of the possible PbrMYB65-binding sites (PbrACO2pro, PbrACO2pro^frag1, and PbrACO2pro^frag2) or the mutated binding sites (PbrACO2pro^mut), into the pGreen 0800-LUC vector. Transformants containing the empty pSAK277 and each reporter were used as the controls. Data represents mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05). B Y1H assay. PbrMYB65 CDS was amplified into the pGADT7 vector, while about 200-bp fragments of PbrACO2 promoter, containing the wide-type (PbrACO2pro^S1 and PbrACO2pro^S2) or the mutated PbrMYB65-binding sites (PbrACO2pro^{S1 mut} and PbrACO2pro^{S2 mut}), were inserted into the bait vector pAbAi. Yeast cell co-transformed with pGADT7-p53 & p53-AbAi was used as a positive control, while yeast cells co-transformed with the empty pGADT7 vector and each bait as the negative controls. C ChIP-qPCR analyses. Calli overexpressing the empty vector was used as a negative control. Data represents mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05). D EMSA assay. FAM luciferase-labelled PbrACO2 promoter fragments, containing the wide-type and the mutated PbrMYB65-binding sites, were named as PbrACO2pro^S1/S2 probes and PbrACO2pro^{S1 mut/S2 mut} probes, respectively, while the unlabeled PbrACO2 promoter fragments were used as competitor probes. The presence and absence of His protein, His-PbrMYB65 protein, labeled probe, or competitor probe are indicated by “ + ” and “ − ”, respectively. Competitor probe concentrations are 50-fold (50 ×) and 100-fold (100 ×) those of the labeled probe

Collectively, these findings indicate that PbrMYB65 interacts with the two predicted PbrMYB65-binding sites in the PbrACO2 promoter to activate its transcription.

Functional validation of PbrMYB65 in vivo

As illustrated in Fig. 4A, PbrMYB65-GFP exhibited the identical subcellular localization to the nuclear marker AtH2B-mcherry (Liu et al. 2007) in Arabidopsis protoplasts, indicating nuclear localization. Additionally, the lack of PbrMYB65 self-interaction in yeast two-hybrid (Y2H) assay suggests that it exists in plant cell as a monomer (Fig. 4B).

Fig. 4

Functional validation of PbrMYB65. A Subcellular localization of PbrMYB65. AtH2B-mcherry was used as the nuclear marker (Liu et al. 2007). Bar, 10 μm. B PbrMYB65 self-interaction determination. (B-i) Y2H assay. Transformants containing AD-T & BD-53, AD-T & BD-Lam, AD & BD, AD & BD-PbrMYB65, and AD-PbrMYB65 & BD were used as the controls. (B-ii) BiFC analyses. Transformants containing YFP^N & YFP^C, YFP^N & PbrMYB65-YFP^C, and PbrMYB65-YFP^N & YFP^C were used as the controls. Bar, 20 μm. C Functional validation of PbrMYB65 in pear calli. (C-i) Growth status of pear calli. (C-ii) PbrACO2 expression level and mitACO activity. (C-iii) Chromatogram of sample. (C-iv) PbrMYB65 expression level and citrate abundance. Calli transformed with the empty vector was used as the control; and the expression levels of PbrACO2 and PbrMYB65 in the control calli are set as 1.0 for RT-qPCR assay. Data represents mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05)

To elucidate the physiological function of PbrMYB65 in vivo, transformations were performed in pear fruit, calli, and tomato. Compared to control fruit, PbrMYB65-overexpressing pear fruit exhibited elevated PbrACO2 mRNA levels but reduced citrate content (Fig. S17A); conversely, silencing of PbrMYB65 in pear fruit produced opposite results (Fig. S17B). Similar results were observed following PbrMYB65 overexpression in pear calli and tomato fruit (Fig. 4C; Fig. S15B). Malate content remained unchanged across samples (data not shown). Furthermore, PbrMYB65-transgenic tomato fruit displayed reduced color development (Fig. S15B-i).

PbrMYB3 demonstrated a similar function in citrate metabolism as PbrMYB65

Subsequently, we investigated the role of PbrMYB3 in citrate metabolism, which was shown to interact with the same cis-acting elements as PbrMYB65 (Fig. S16C). As shown in Fig. 5A, PbrMYB3-GFP exhibited the identical subcellular localization to the nuclear marker AtH2B-mcherry (Liu et al. 2007) in Arabidopsis protoplasts, indicating its nuclear localization. Additionally, lack of PbrMYB3 self-interaction in the Y2H assay implies that it exists in plant cell as a monomer (Fig. S18).

Fig. 5

Determination of PbrMYB3’s role in citrate metabolism in pear. A Subcellular localization of PbrMYB3. AtH2B-mcherry was used as the nuclear marker (Liu et al. 2007). Bar, 10 μm. B Confirmation of PbrMYB3 as the upstream regulator of PbrACO2. (B-i) Dual-luciferase assay. PbrMYB3 CDS was introduced into the pSAK277 vector, while PbrACO2 promoter fragments of different lengths, containing different numbers of the possible PbrMYB3-binding sites (PbrACO2pro, PbrACO2pro^frag1, and PbrACO2pro^frag2) or the mutated binding sites (PbrACO2pro^mut), into the pGreen 0800-LUC vector. Transformants containing the empty pSAK277 and each reporter were used as the controls. (B-ii) Y1H assay. PbrMYB3 CDS was amplified into the pGADT7 vector, while about 200-bp fragments of PbrACO2 promoter, containing the wide-type (PbrACO2pro^S1 and PbrACO2pro^S2) or the mutated PbrMYB3-binding sites (PbrACO2pro^{S1 mut} and PbrACO2pro^{S2 mut}), were inserted into the bait vector pAbAi. Yeast cell co-transformed with pGADT7-p53 & p53-AbAi was used as a positive control, while yeast cells co-transformed with the empty pGADT7 vector and each bait as the negative controls. (B-iii) ChIP-qPCR analyses. Calli overexpressing the empty vector was used as a negative control. C Functional validation of PbrMYB3 in pear calli. (C-i) Growth status of pear calli. (C-ii) PbrACO2 expression level and mitACO activity. (C-iii) Chromatogram of sample. (C-iv) PbrMYB3 expression level and citrate abundance. Calli transformed with the empty vector was used as the control; and the expression levels of PbrACO2 and PbrMYB3 in the control calli are set as 1.0 for RT-qPCR assay. Data represents mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05).

Further experimentation revealed that the LUC/REN ratio in tobacco leaf co-transformed with PbrMYB3 and reporter containing PbrACO2 promoter increased approximately 220% compared to the control, with this increase correlating positively to the number of potential PbrMYB3-binding sites (Fig. 5B-i); however, this increase disappeared following mutation of the two binding sites (Fig. 5B-i). In the Y1H assay, the positive control (pGADT7-p53 & p53-pAbAi), negative controls (pGADT7 & PbrACO2pro^S1-pAbAi, pGADT7 & PbrACO2pro^S2-pAbAi), bait-prey co-transformants (pGADT7-PbrMYB3 & PbrACO2pro^S1-pAbAi, pGADT7-PbrMYB3 & PbrACO2pro^S2-pAbAi), and co-transformants containing the mutated elements (pGADT7 & PbrACO2pro^{S1 mut}-pAbAi, pGADT7 & PbrACO2pro^{S2 mut}-pAbAi, pGADT7-PbrMYB3 & PbrACO2pro^{S1 mut}-pAbAi, pGADT7-PbrMYB3 & PbrACO2pro^{S2 mut}-pAbAi) exhibited normal growth on SD/-Leu medium (Fig. 5B-ii); upon AbA addition, growth was suppressed in negative controls and transformants with mutated binding elements, while the positive control and bait-prey co-transformants remained unaffected (Fig. 5B-ii). ChIP-qPCR analyses confirmed substantial enrichment (> six-fold) of PbrMYB3 in PbrACO2 promoter fragments containing the PbrMYB3-binding sites (Fig. 5B-iii).

Consistent with observations in PbrMYB65-transgenic fruit and calli, PbrMYB3 mRNA levels in transgenic pear tissues showed a positive correlation with PbrACO2 expression and/or mitACO activity, while negatively correlating with citrate content (Fig. 5C; Fig. S19). Malate abundance remained unchanged across samples (data not shown). These findings suggest that PbrMYB3 functions as a monomer in binding the two predicted PbrMYB3-binding sites in the PbrACO2 promoter, transcriptionally activating its expression and consequently reducing citrate accumulation in pear.

Given that PbrMYB3 and PbrMYB65 constitute a syntenic gene pair (Cao et al., 2016; Li et al. 2016), their relationship was examined. Despite high protein sequence identity (Fig. S20A), their promoter sequences differed (2000-bp sequences upstream of the transcriptional start sites (ATG) of PbrMYB3 and PbrMYB65; Fig. S20B). Consistent with the absence of PbrMYB3-/PbrMYB65-binding sites (MYBCORE box motif, CAACCG) in their promoters, no significant increase in LUC/REN ratio was observed in tobacco leaf co-transformed with PbrMYB3 (or PbrMYB65) and the reporter containing PbrMYB65 promoter (or PbrMYB3 promoter) (Fig. S21A). Furthermore, Y2H and bimolecular fluorescence complementation (BiFC) assays showed no interaction between PbrMYB3 and PbrMYB65 (Fig. S21B).

PbrNAC34a was an upstream regulator of PbrMYB3 and PbrMYB65

To identify the upstream regulators of these two MYB TFs, correlation analyses were conducted between PbrMYB3/65 and 48 other differentially expressed TFs. As shown in Fig. S22, the expression levels of 18 members exhibited strong correlations with PbrMYB3, PbrMYB65, and PbrACO2 mRNA levels during the maturation process of P. bretschneideri Rehd. fruit, including Pbr000398.1, Pbr015697.1, Pbr020595.1, Pbr022437.1, Pbr027839.1, Pbr030208.1, Pbr038280.1, Pbr013255.1, Pbr000523.1, Pbr009294.1, Pbr011441.1, Pbr019293.1, Pbr038434.1, Pbr006028.1, Pbr013948.1, Pbr016310.4, Pbr010912.1, and Pbr026635.1 (absolute correlation coefficient > 0.6). Among these, Pbr026635.1, designated as PbrNAC34a by Gong et al. (2019), displayed relatively high coefficients with PbrMYB3, PbrMYB65, and PbrACO2 (Fig. S22). Furthermore, utilizing the PlantRegMap database and previous reports on the common sequences of the NAC-binding sites (Bi et al. 2023; Li et al. 2023), several potential PbrNAC34a-binding sites were identified from PbrMYB3 and PbrMYB65 promoters (2000-bp sequences upstream of the transcriptional start sites (ATG) (Fig. S23). These findings suggested that PbrNAC34a might function as their upstream regulator.

To investigate the cis-acting elements responsible for the PbrNAC34a-induced activation of downstream gene expression, dual-luciferase reporter assays were conducted using PbrMYB3 and PbrMYB65 promoter fragments of varying lengths, containing different numbers of potential PbrNAC34a-binding sites. As illustrated in Fig. 6A, only fragments containing the 'CTTCGTTT' (site 1 (S1) in PbrMYB3 promoter) and 'AGAAAGAA' (site 4 (S4) in PbrMYB65 promoter) elements interacted with PbrNAC34a to initiate downstream gene transcription. Furthermore, the LUC/REN ratio increase ceased after the mutation of the binding site in PbrMYB3 (CTTCGTTT → TTTTTTTT) or PbrMYB65 (AGAAAGAA → CCCCCCCC) promoter (Fig. 6A). These findings confirmed the critical importance of these two elements for PbrNAC34a to activate PbrMYB3 and PbrMYB65 expression, prompting further investigation. In the Y1H assay, the positive control (pGADT7-p53 & p53-pAbAi), negative controls (pGADT7 & PbrMYB3pro^S1-pAbAi, pGADT7 & PbrMYB65pro^S4-pAbAi), bait-prey co-transformants (pGADT7-PbrNAC34a & PbrMYB3pro^S1-pAbAi, pGADT7-PbrNAC34a & PbrMYB65pro^S4-pAbAi), and co-transformants containing the mutated elements (pGADT7 & PbrMYB3pro^{S1 mut}-pAbAi, pGADT7 & PbrMYB65pro^{S4 mut}-pAbAi, pGADT7-PbrNAC34a & PbrMYB3pro^{S1 mut}-pAbAi, pGADT7-PbrNAC34a & PbrMYB65pro^{S4 mut}-pAbAi) exhibited normal growth on SD/-Leu medium (Fig. 6B). Upon AbA addition, growth was suppressed in negative controls and transformants containing mutated binding elements, while the positive control and bait-prey co-transformants remained unaffected (Fig. 6B). ChIP-qPCR analyses revealed greater than four-fold enrichment in PbrMYB3 and PbrMYB65 promoter fragments containing the PbrNAC34a-binding sites compared to controls (Fig. 6C).

Fig. 6

Identification and confirmation of PbrNAC34a as the upstream regulator of PbrMYB3 and PbrMYB65. A Dual-luciferase assay for the activation of PbrMYB3 (A-i) and PbrMYB65 (A-ii) transcription by PbrNAC34a. PbrNAC34a CDS was introduced into the pSAK277 vector, while PbrMYB3 and PbrMYB65 promoter fragments of different lengths, containing different numbers of the possible PbrNAC34a-binding sites (PbrMYB3 promoter fragments included PbrMYB3pro, PbrMYB3pro^frag1, PbrMYB3pro^frag2, and PbrMYB3pro^frag3; PbrMYB65 promoter fragments included PbrMYB65pro, PbrMYB65pro^frag1, PbrMYB65pro^frag2, PbrMYB65pro^frag3, and PbrMYB65pro^frag4) or the mutated binding sites (PbrMYB3pro^mut and PbrMYB65pro^mut), into the pGreen 0800-LUC vector. Transformants containing the empty pSAK277 vector and each reporter were used as the controls. B Y1H assay of the interactions between PbrNAC34a and PbrMYB3 promoter fragment (B-i) and between PbrNAC34a and PbrMYB65 promoter fragment (B-ii). PbrNAC34a CDS was amplified into the prey vector pGADT7, while about 200-bp fragments of PbrMYB3 and PbrMYB65 promoters, containing the wide-type (PbrMYB3pro^S1 and PbrMYB65pro^S4) or the mutated PbrNAC34a-binding sites (PbrMYB3pro^{S1 mut} and PbrMYB65pro^{S4 mut}), were inserted into the bait vector pAbAi. Yeast cell co-transformed with pGADT7-p53 and p53-AbAi was used as a positive control, while yeast cells co-transformed with the empty pGADT7 vector and each bait as the negative controls. C ChIP-qPCR analyses of the interactions between PbrNAC34a and PbrMYB3 promoter fragment (C-i) and between PbrNAC34a and PbrMYB65 promoter fragment (C-ii). Calli overexpressing the empty vector was used as a negative control. Data represent mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05)

Subsequently, PbrNAC34a function was examined. As demonstrated in Fig. 7A, PbrNAC34a-GFP exhibited the identical subcellular localization to the nuclear marker AtH2B-mcherry (Liu et al. 2007) in Arabidopsis protoplasts, indicating nuclear localization. In vivo analysis revealed that PbrNAC34a overexpression in pear fruit enhanced expression levels of PbrMYB3, PbrMYB65, and PbrACO2, leading to reduced citrate levels compared to control fruit (Fig. 7B-i); conversely, PbrNAC34a-silenced fruit exhibited opposite results (Fig. 7B-ii). Comparable effects were observed in PbrNAC34a-overexpressing pear calli, where PbrMYB3, PbrMYB65, and PbrACO2 transcription increased, resulting in decreased citrate content (Fig. 7C). These results confirmed PbrNAC34a's role as an upstream regulator of PbrMYB3 and PbrMYB65.

Fig. 7

Function validation of PbrNAC34a. A Subcellular localization of PbrNAC34a. AtH2B-mcherry was used as the nuclear marker (Liu et al. 2007). Bar, 10 μm. B Functional validation of PbrNAC34a in pear fruit. (B-i) Transient overexpression of PbrNAC34a. ‘Yali’ fruit transformed with the empty pCAMBIA1300 vector containing a GFP tag was used as the control for the PbrNAC34a-overexpressing fruit. (B-ii) Transient silence of PbrNAC34a. Fruit co-transformed with the empty pTRV2 and pTRV1 vectors was used as the control for the PbrNAC34a-silenced fruit. C Functional validation of PbrNAC34a in pear calli. (C-i) Growth status of pear calli. (C-ii) Gene expression level. (C-iii) Citrate abundance. Calli transformed with the empty vector was used as the control. And the expression level of each gene in the control calli/fruit is set as 1.0 for RT-qPCR assay. Data represents mean value ± SD of three biological replicates, and vertical bars labelled with the same letter are not significantly different between samples (p < 0.05)

Discussion

Pear, belonging to the Pomoideae subfamily of Rosaceae family, is renowned for its distinctive flavor quality (Qian et al. 2021; Wang et al. 2021a, b). Taste quality encompasses the collective sensory attributes resulting from the stimulation of gustatory receptors on the tongue (Baldwin et al. 1998). As essential metabolites contributing to taste, malate and citrate were identified as the predominant organic acids in mature 'Yali' and 'Dangshansuli' fruits (Fig. 1A; Fig. S7 and S9A; Table S9). In agreement with Zhang et al. (2021), malate and citrate in the cortex of both cultivars exhibited inverse accumulation patterns during fruit development (Fig. 1A; Fig. S9A; Table S9).

With the publication of P. bretschneideri Rehd. genome and advancements in plant biotechnology, modification of plant phenotype through genetic engineering has become achievable (Li et al., 2022b). Among > 40,000 protein-coding genes identified from the pear genome (Wu et al., 2013), only a limited number have been functionally validated in organic acid metabolism. A cytosolic NAD⁺-dependent malate dehydrogenase gene (gene ID: Pbr030554.1) and a cytosolic NADP⁺-dependent malate enzyme gene (gene ID: Pbr008772.1) participated in malate biosynthesis and degradation, respectively (Wang et al. 2018). Additionally, an ALMT from white pear (gene ID: Pbr020270.1), situated in the plasma membrane, facilitated malate import into the sink cell (Xu et al. 2018). Nevertheless, understanding of citrate metabolism in developing pear fruit remains limited.

Previous studies suggest that citrate levels in horticultural fruit are regulated by ACOs (Hussain et al. 2017; Pascual et al., 2018; Sienkiewicz-Porzucek et al. 2010; Sulpice et al. 2010), which catalyze the reversible isomerization of citrate to isocitrate (Wang et al. 2016). Consistent with this observation, the mitochondrial PbrACO2, whose expression was positively correlated with cytACO and mitACO activities but negatively with citrate content during pear (P. bretschneideri Rehd.) fruit maturation, catalyzed the isomerization of citrate to isocitrate in vitro and in vivo (Fig. 1C; Fig. 2; Fig. S9D; Fig. S14 and S15A). Similar functionality was observed in homologues from other plants, including CitACO3 from citrus (Li et al. 2017), SlACO1 from tomato (Carrari et al. 2003), and OsACO1 from rice (Senoura et al. 2020), suggesting conservation of ACO function throughout plant evolution. However, other PbrACO isoforms, whose transcript levels were higher during P. bretschneideri Rehd. fruit maturation (Fig. 1B; Fig. S9C and Table S10), may also contribute significantly to citrate metabolism. Beyond transcriptional regulation, ACO activity undergoes posttranslational suppression by H₂O₂ and peroxynitrite (a potent oxidant and nitrating agent formed from superoxide anion and nitric oxide generated by mitochondria) (Bulteau et al. 2003; Han et al., 2005; Verniquet et al. 1991). H₂O₂ variations during pear development (Wang et al. 2021a, b) suggest that posttranslational modification of PbrACOs might also influence fruit citrate levels.

Plant TFs interact with corresponding cis-acting elements in the promoters of downstream genes to regulate their expression, thereby influencing fruit quality (Jia et al. 2023). Through transcription analyses and experimental validation, this study demonstrated that PbrMYB3 and PbrMYB65 bound, as monomers, to two identical MYB-binding sites in the PbrACO2 promoter to activate its expression, subsequently suppressing citrate accumulation in pear and tomato (Figs. 3, 4, and 5; Fig. S15B and S16-S19). These findings establish these MYB TFs as upstream regulators of PbrACO2. This mechanism parallels the collaborative inhibition of citrate accumulation in 'Ponkan' fruit by CitWRKY1 and CitNAC62 through CitACO3 regulation (Li et al. 2017). Furthermore, consistent with observations in the AcNAC1-mutagenetic kiwifruit showing reduced citrate levels (Fu et al. 2023), malate levels remained unchanged after transformation of pear/tomato fruit and/or calli with PbrACO2 (or PbrMYB3, or PbrMYB65) gene (Fig. 2D; Fig. 4C-5C; Fig. S15), indicating malate metabolism may operate independently of citrate pathways. However, similar to the AcNAC1-mutagenetic kiwifruit, color transition was inhibited in PbrACO2/PbrMYB65-transgenic tomato fruit (Fig. S15), suggesting an association between citrate metabolism and this phenotype.

Previous research established that PbrMYB3 and PbrMYB65 constitute a syntenic gene pair originating from WGD/segmental duplication, with purifying selection serving as the primary evolutionary force (Cao et al., 2016; Li et al. 2016). Despite their highly identical protein sequences (Fig. S20A), no direct interaction was observed between PbrMYB3 and PbrMYB65 (Fig. S21). Gene duplication significantly contributes to gene family expansion, potentially leading to regulatory mechanism divergence that enhances plant adaptation to various developmental processes and (a)biotic stresses (Qiao et al. 2015). Similar to observations in PbrACOs (Fig. S3C and S4), the promoter sequences of PbrMYB3 and PbrMYB65 showed inconsistencies (Fig. S20B). Additional investigation revealed that PbrNAC34a binds to different cis-acting elements at distinct sites in the promoters of these MYB TFs, inducing their expression (Fig. 6; Fig. S22-S23). Analysis of their expression patterns across tissues and responses to abiotic stresses (Fig. S5 and S6; Table S6-S8) suggests distinct regulatory mechanisms for PbrMYB3 and PbrMYB65 expression in pear.

The study revealed no significant difference in organic acid accumulation between bagged and unbagged 'Yali' pear fruits (Fig. 1A; Table S9). However, the cortex tissue exhibited higher citrate levels compared to the pericarp tissue, corresponding with lower expression levels of PbrACO2, PbrMYB3, PbrMYB65, and PbrNAC34a (Fig. 1A-B; Fig. S16b; Table S9-S10 and S12). Previous research on 'Hongyang' kiwifruit demonstrated that elevated citrate content in the outer pericarp tissue relative to the core tissue correlated with upregulated transcription of AcALMT1 and AcNAC1 (Fu et al. 2023). Considering their roles in citrate metabolism, these results suggest that the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade mediates citrate differences between pear fruit pericarp and cortex tissues (Fig. 1–7; Fig. S14-S23; Table S9-10 and S12). Similar regulatory cascades have been documented in banana ripening ethylene biosynthesis (Wei et al., 2023) and grape berry sugar accumulation (Li et al., 2022a). Additionally, experimental validation confirmed that nuclear Pbr009294.1 (designated as PbrWRKY72 by Huang et al. (2015)), despite showing strong positive correlations with PbrMYB3 and PbrMYB65 mRNA levels during white pear maturation (Fig. S22), did not regulate their transcription or citrate metabolism (Fig. S24).

Based on the findings in this study, a schematic model has been proposed (Fig. 8). During P. bretschneideri Rehd. fruit development, nuclear PbrNAC34a binds to the promoters of PbrMYB3 (binding element: CTTCGTTT) and PbrMYB65 (binding element: AGAAAGAA), triggering their transcription. Following translation in the ribosome, these two MYB TFs are transported into the nucleus, where they interact with the same two MYB-binding sites (MYBCORE box motif, CAACCG) in the PbrACO2 promoter as monomers and subsequently activate its expression. After translation and import into mitochondria, PbrACO2 isomerizes citrate into isocitrate, thereby inhibiting citrate accumulation in pear. Due to higher expression levels of PbrACO2, PbrMYB3, PbrMYB65, and PbrNAC34a, the pericarp tissue exhibits lower citrate abundance compared to the cortex tissue.

Fig. 8

Schematic model on the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade regulating citrate difference between the pericarp and cortex tissues of the developing P. bretschneideri Rehd. fruit. During P. bretschneideri Rehd. fruit development, the nuclear PbrNAC34a binds to the promoters of PbrMYB3 (binding element: CTTCGTTT) and PbrMYB65 (binding element: AGAAAGAA), and then triggers their transcription. After translation in ribosome, these two MYB TFs are transported into the nucleus, where they interact with the same two MYB-binding sites (MYBCORE box motif, CAACCG) in PbrACO2 promoter as monomer and subsequently activate its expression. After translation and then import into mitochondria, PbrACO2 isomerizes citrate into isocitrate, inhibiting citrate accumulation in pear. Due to the higher expression levels of PbrACO2, PbrMYB3, PbrMYB65, and PbrNAC34a, the pericarp tissue possessed lower citrate abundance than that in the cortex tissue

Conclusion

This study demonstrates that PbrACO2, localized in the mitochondria, catalyzes citrate isomerization during P. bretschneideri Rehd. fruit development. Two MYB TFs, PbrMYB3 and PbrMYB65, bind, as monomers, to the same two MYB-binding sites in the PbrACO2 promoter and initiate its expression, thereby reducing citrate levels in pear and tomato. Furthermore, PbrNAC34a functions as the upstream activator of the PbrMYB3 and PbrMYB65 genes. Based on their distinct expression profiles in two tissues, this study suggests that the PbrNAC34a-PbrMYB3/65-PbrACO2 cascade regulates citrate differences between the pericarp and cortex tissues of pear.

Materials and methods

Bioinformatics analyses

ACOs from Arabidopsis were used as queries in the BLASTP against plant genome databases (https://phytozome-next.jgi.doe.gov/, https://plantgenie.org/, and http://peargenome.njau.edu.cn/) before confirmation of the conserved domain by the SMART database (http://smart.embl-heidelberg.de/) (Terol et al. 2010; Wang et al. 2016; Wang et al. 2021a, b). Physio-biochemical parameters were calculated using the ProtParam tool (https://web.expasy.org/protparam/) (Wang et al. 2018).

The timescale tree was constructed using TIMETREE (http://www.timetree.org/) (Kumar et al. 2017), and the phylogenetic tree was generated by MEGA7.0 software, employing the Neighbor-Joining (NJ) method with the poisson model (all other settings, including the number of bootstrap iterations (1000), remained default) (Zhang et al. 2019). Gene structure was visualized using the Gene Structure Display Server (http://gsds.cbi.pku.edu.cn/) (Hu et al., 2015); cis-acting elements were identified using the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al. 2002); and motifs were characterized using the MEME Suite tool (https://meme-suite.org/meme/index.html) (Bailey et al. 2015). Divergence between upstream sequences of each paralogous gene pair was measured using the GATA program (Nix and Eisen 2005), with a window size set at seven and a lower cutoff score of 12 bits. The TF-binding sites in the gene promoter were predicted through the PlantRegMap database (http://plantregmap.gao-lab.org/) (Tian et al. 2020).

Chromosomal locations of plant ACOs were determined through genome annotation and visualized using Circos (Krzywinski et al. 2009). A methodology similar to that employed in the Plant Genome Duplication Database (PGDD) was utilized to analyze the syntenic relationship. The duplicated ACOs were classified into several categories: WGD/segmental, tandem, singleton, proximal, and dispersed (Zhang et al. 2019). MCScanX downstream analyses tools were employed to annotate the Ka and Ks substitution rates of the syntenic gene pair, while the KaKs Calculator 2.0 determined Ka and Ks using the Nei-Gojobori (NG) method (Zhang et al. 2019).

Plant material and treatment

P. bretschneideri Rehd. cv. 'Yali' and 'Dangshansuli' trees with comparable vigor and load, cultivated in an experimental orchard in Jiangsu province, served as research material. 'Yali' pear fruits were enclosed in triple-layer paper bags (red paper exterior, black paper middle layer, and white adhesive-bonded fabric interior) at 34 days after full bloom (34 DAFB), while unbagged 'Yali' and 'Dangshansuli' fruits in corresponding positions were marked. 'Yali' and 'Dangshansuli' represent typical pear cultivars in China; the bagging treatment of 'Yali' fruit is extensively practiced in Hebei province, China's largest pear production region, to enhance fruit appearance quality and satisfy international export market requirements.

The (un)bagged fruits were harvested at six developmental stages: fruit-setting stage (15 DAFB), physiological fruit drop stage (34 DAFB), approximately one month after fruit enlargement stage (81 DAFB), pre-mature stage (110 DAFB), mature stage (145 DAFB), and fruit senescence stage (160 DAFB) (Wang et al. 2021a, b). Following transportation to the laboratory, the pericarp and cortex tissues of 'Yali' fruit and the cortex tissue of 'Dangshansuli' fruit were collected for further analysis. The inclusion of 'Dangshansuli' fruit cortex aims to investigate whether the candidate gene functions in citrate metabolism in another cultivar of P. bretschneideri Rehd. Each treatment and/or cultivar included three biological replicates, with 20 fruits sampled per replicate at each developmental stage.

Quantification of organic acids

Plant tissues, including pear cortex, pericarp, and calli, as well as tomato fruit, underwent homogenization with ultrapure water, followed by filtration through a 0.22-μm Millipore filter for supernatant collection.

Analysis of organic acids in the supernatant, including citrate, oxalate, tartarate, quininate, malate, and shikimate, was conducted using a high-performance liquid chromatography (HPLC) system (Thermo Ultimate 3000, Dionex, Massachusetts, USA) equipped with a Waters Acquity UPLC® HSS T3 column and a photodiode array (PDA) detector (Wang et al. 2018). Individual acids were qualified and quantified according to the method described by Wang et al. (2018). Total acid content was calculated as the sum of five individual acids.

Determination of cytACO and mitACO activities

Cytosol and mitochondria extraction from various plant tissues, including pear cortex and calli as well as tomato fruit, followed the protocol of Matamoros et al. (2013). Subsequently, ACO activity in cytosol (cytACO) and mitochondria (mitACO) was measured using the corresponding assay kit (ACO-1-Z, Suzhou Comin Biotechnology Co., Ltd., Suzhou, China).

Protein concentration in the crude enzyme extract was determined using the bicinchoninic acid protein assay kit (A045-4, Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Transcriptome and RT-qPCR analyses

Transcriptome analysis was conducted following the method of Li et al. (2019). In brief, total RNA was extracted from various pear tissues (including 15-DAFB fruit, stem, leaf, stigma, petal, ovary, pericarp, and cortex) using the EASYspin plant RNA extraction kit (Biomed, China). DNase (Takara Biotechnology Co., Ltd., Dalian, China) was applied to eliminate residual DNA before evaluating RNA integrity, concentration, and purity. The RNA-seq library was then constructed and sequenced on the Illumina HiSeq 2500 platform (Biomarker Technologies Co, Ltd., Beijing, China) (Li et al. 2019). Following quality assessment and data filtering, clean reads were mapped to the P. bretschneideri Rehd. genome (Wu et al., 2013) using HISAT2 software. Fragments per Kilobase Million (FPKM) was utilized to calculate gene expression, and differentially expressed genes (DEGs) were identified using DESeq2_EBSeq software, according to the following criteria: fold change ≥ 2.0 and false discovery rate (FDR) < 0.01 (Jia et al. 2023).

RT-qPCR assay was performed as described by Wang et al. (2018). Gene-specific primers were designed using Primer Premier 6.0 (Table S1). Total RNA from different plant tissues, including pear cortex and calli as well as tomato fruit, was isolated using TRIzol Reagents (Invitrogen, USA) followed by RNase-free DNase treatment (Qiagen, USA). Following first-strand cDNA synthesis, RT-qPCR assay was conducted using TaKaRa One Step SYBR® PrimeScript™ RT-PCR Kit (Perfect Real Time) (Takara Biotechnology Co., Ltd., Dalian, China) (Wang et al. 2018). P. Tubulin (PbrTub) served as the internal reference gene for the gene-overexpressing/silenced pear fruit and calli as well as during pear fruit development (Wang et al. 2018), while S. lycopersicum Actin-51 (SlActin-51) served as the housekeeping gene for the gene-overexpressing tomato fruit (Liu et al. 2020). The relative gene expression level was calculated using the 2^-ΔΔCT method (Wang et al. 2018).

Alignment of the coding sequences (CDSs), protein sequences, and promoter sequences from 'Yali' and/or 'Dangshansuli' fruits

The CDSs of PbrACO2, PbrMYB3, PbrMYB65, and PbrNAC34a were amplified from 'Yali' and 'Dangshansuli' fruits. The corresponding protein sequences were then deduced from their CDSs using the Translate database (https://web.expasy.org/translate/) (Jia et al. 2023). Additionally, the promoter sequences of PbrMYB3 and PbrMYB65 were amplified from 'Yali' fruit. Sequence alignment was conducted using DNAMAN software (Lynnon Biosoft, San Ramon, California, USA).

Subcellular localization analyses

The CDSs of PbrACO2, PbrMYB3, PbrMYB65, PbrNAC34a, and PbrWRKY72 without stop codons were amplified from 'Yali' fruit (Table S1), inserted into the pBI221 vector with a GFP tag, and subsequently co-transformed with the corresponding marker into Arabidopsis protoplasts (Jia et al. 2023). MSTP-mcherry (Sun et al. 2022) and AtH2B-mcherry (Liu et al. 2007) served as the mitochondrial and nuclear markers, respectively. The fluorescence signal was detected using a Leica TCS SP8 confocal laser-scanning microscope (Leica, Wetzlar, Germany).

Functional validation of PbrACO2 in vitro

PbrACO2 CDS was amplified from 'Yali' fruit (Table S1), inserted into the pCold-TF vector, and transformed into E. coli BL21 (DE3) for the expression of the His-tagged recombinant protein (Jia et al. 2023). Following purification through a Ni-NTA His bind resin column (Sangong Bioengineering Co., Ltd., Shanghai, China), the recombinant protein was collected for PbrACO2 activity assay as described above.

Gene function validation in vivo

Transient transformation of pear fruit (a)Transient gene overexpression

PbrACO2, PbrMYB3, PbrMYB65, PbrNAC34a, and PbrWRKY72 CDSs were isolated from'Yali'fruit (Table S1) and subsequently inserted into the pCAMBIA1300 vector containing a GFP tag. These constructs were then transformed into A. tumefaciens strain GV3101 and cultured at 28 °C until OD₆₀₀ = 1.0. The bacterial strain was resuspended in infiltration buffer (10 mM magnesium chloride (MgCl₂), 10 mM 2-morpholinoethanesulphonic acid (MES), and 150 μM acetosyringone), and 20 μL of the solution was carefully injected into the cortex tissue of developing'Yali'fruit before incubation at 25 °C for 3 d (Jia et al. 2023). Fruit infiltrated with the empty pCAMBIA1300 vector containing a GFP tag served as the control. (b)Transient gene silence Approximately 200-bp fragments of PbrACO2, PbrMYB3, PbrMYB65 PbrNAC34a, and PbrWRKY72 CDSs were inserted into the pTRV2 vector (Table S1). The constructed pTRV2 and pTRV1 plasmids were transformed into A. tumefaciens strain GV3101, respectively; and suspended in infiltration buffer until OD₆₀₀ = 1.0. Equal volumes of the recombinant pTRV2 and pTRV1 buffer were combined and carefully injected into cortex tissue of developing 'Yali' fruit prior to incubation at 25 °C for 3 d (Zhang et al. 2019). Fruits co-injected with empty pTRV2 and pTRV1 vectors served as controls.

Transformation of pear calli

PbrACO2, PbrMYB3, PbrMYB65, and PbrNAC34a CDSs were amplified and inserted into the pCAMBIA1300 vector containing a GFP tag as described above, then transformed into fruit calli derived from P. communis cv. 'Clapp's Favorite' pear fruitlet (Jia et al. 2023). Following selection on hygromycin B-containing MS medium (carbon source: sucrose), positive lines were verified at both the DNA level by PCR and the RNA level by RT-qPCR. The confirmed transgenic lines were transferred to hygromycin B-containing MS medium with mixed sugar (sucrose/sorbitol (1:1), 15 g L⁻¹) as the carbon source to simulate fruit growing conditions (Jia et al. 2023). Calli transformed with the empty vector served as the control.

Transformation of tomato fruit

Previously constructed PbrACO2-/PbrMYB65-overexpressing vectors were transformed into S. lycopersicum cv. 'MicroTom' following the protocol of Cheng et al. (2018). Positive transgenic lines were selected on 100 mg L⁻¹ kanamycin-containing medium and verified at both DNA level by PCR and RNA level by RT-qPCR. Plants were maintained in greenhouse conditions (18 h light at 25 ℃ and 6 h dark at 18 ℃, 60% relative humidity). Tomato fruits at 45 DAFB from control (wild-type) and transgenic homozygous lines (T2 generation) were collected for subsequent analysis.

Protein and DNA interaction analyses

LUC assay The CDSs of PbrMYB3, PbrMYB65, PbrNAC34a, and PbrWRKY72 were isolated from'Yali'fruit and subsequently inserted into the pSAK277 vector. Additionally, promoter fragments of PbrACO2, PbrMYB3, and PbrMYB65 of varying lengths, containing different quantities of potential PbrMYB3-/PbrMYB65-binding sites (MYBCORE box motif, CAACCG; PbrACO2pro, PbrACO2pro^frag1, and PbrACO2pro^frag2) or potential PbrNAC34a-binding sites (PbrMYB3 promoter fragments included PbrMYB3pro, PbrMYB3pro^frag1, PbrMYB3pro^frag2, and PbrMYB3pro^frag3; PbrMYB65 promoter fragments included PbrMYB65pro, PbrMYB65pro^frag1, PbrMYB65pro^frag2, PbrMYB65pro^frag3, and PbrMYB65pro^frag4) as well as the mutated binding sites (CAACCG → CTTCCG in PbrACO2 promoter, PbrACO2pro^mut; CTTCGTTT → TTTTTTTT in PbrMYB3 promoter, PbrMYB3pro^mut; AGAAAGAA → CCCCCCCC in PbrMYB65 promoter, PbrMYB65pro^mut), were inserted into the pGreen 0800-LUC vector to generate various reporter constructs (Table S1). Subsequently, a combination of A. tumefaciens containing pSAK277-PbrMYB3 (or pSAK277-PbrMYB65, or pSAK277-PbrNAC34a, or pSAK277-PbrWRKY72) and each reporter was infiltrated into N. benthamiana leaf (Liu et al., 2022). LUC and REN activities were measured using a dual-LUC reporter assay system (Promega Corporation, Madison, Wisconsin, USA). Transformants containing the empty pSAK277 vector and each reporter served as the controls.

Y1H determination The CDSs of PbrMYB3, PbrMYB65, and PbrNAC34a were amplified from'Yali'fruit and inserted into the prey vector pGADT7 (Table S1). The promoter fragments of PbrACO2, PbrMYB3, and PbrMYB65, approximately 200-bp in length, containing the PbrMYB3/65-binding sites (MYBCORE box motif, CAACCG; PbrACO2pro^S1 and PbrACO2pro^S2) or the PbrNAC34a-binding sites (CTTCGTTT in PbrMYB3 promoter, PbrMYB3pro^S1; AGAAAGAA in PbrMYB65 promoter, PbrMYB65pro^S4) along with their mutated binding sites (CAACCG → CTTCCG in PbrACO2 promoter fragments, PbrACO2pro^{S1 mut} and PbrACO2pro^{S2 mut}; CTTCGTTT → TTTTTTTT in PbrMYB3 promoter fragment, PbrMYB3pro^{S1 mut}; AGAAAGAA → CCCCCCCC in PbrMYB65 promoter fragment, PbrMYB65pro^{S4 mut}), were cloned into the bait vector pAbAi (Table S1). The Y1H assay was performed using the Matchmaker Gold Yeast One-Hybrid Library Screening System (Shanghai Weidi Industrial Co., Ltd., Shanghai, China) (Liu et al. 2019). Self-activation was assessed using SD/-Ura medium supplemented with varying concentrations of AbA for PbrACO2pro^S1/S2, PbrACO2pro^{S1 mut/S2 mut}, PbrMYB3pro^S1, PbrMYB3pro^{S1 mut}, PbrMYB65pro^S4, and PbrMYB65pro^{S4 mut} to determine optimal AbA concentration. The positive control comprised yeast cells co-transformed with pGADT7-p53 and p53-pAbAi, while negative controls consisted of yeast cells co-transformed with empty pGADT7 vector and each bait.

ChIP-qPCR analysesPbrMYB3-/PbrMYB65-/PbrNAC34a-overexpressing calli and control (empty vector), generated according to previously described methods, underwent DNA–protein cross-linking in 1% (v/v) formaldehyde. Following homogenization and cell lysis, the extracted chromatin was sonicated to obtain soluble sheared chromatin with DNA fragments ranging from 200–500 bp. A portion of the sheared chromatin served as input DNA, while the remainder underwent immunoprecipitation using anti-GFP antibody (Ab290, Abcam) (Li et al. 2018). The enrichment of PbrACO2, PbrMYB3, and PbrMYB65 promoter fragments was quantified via qPCR assay (Table S1).

EMSA assay The His-tagged recombinant PbrMYB65 protein was produced following previously described methods (Table S1). FAM luciferase-labelled DNA probes approximately 30-bp in length, containing either wild-type (MYBCORE box motif, CAACCG) or mutated (CAACCG → CTTCCG) PbrMYB65-binding sites, and unlabeled competitor probes were synthesized by Sangong Bioengineering (Shanghai) Co., Ltd. The EMSA was conducted according to the protocol provided in the LightShift™ Chemiluminescent EMSA Kit (Thermo Fisher Scientific, Inc. Sunnyvale, California, USA) (Zhang et al., 2024).

PbrMYB3 and PbrMYB65 interaction determination

Y2H assay The CDSs of PbrMYB3 and PbrMYB65 were isolated from'Yali'fruit and subsequently inserted into the pGADT7 (AD) and pGBKT7 (BD) vectors (Clontech) (Table S1). The AD-PbrMYB3/65 and BD-PbrMYB3/65 constructs were co-transformed into S. cerevisiae AH109 and cultured on synthetic dropout nutrient media, including SD/-Leu/-Trp and SD/-Leu/-Trp/-His/-Ade (Ma et al. 2020). Control transformants included AD-T & BD-53, AD-T & BD-Lam, AD & BD, AD & BD-PbrMYB3/65, and AD-PbrMYB3/65 & BD.

BiFC analyses The mature protein-encoding CDSs of PbrMYB3 and PbrMYB65 were amplified from'Yali'fruit (Table S1) and integrated into 35S-pSPYNE-YFP^N and 35S-pSPYCE-YFP^C vectors. The resulting constructs, PbrMYB3/65-YFP^N and PbrMYB3/65-YFP^C, were introduced into A. tumefaciens strain GV3101 prior to infiltration into N. benthamiana leaf epidermal cells (Ma et al. 2020). Control transformants comprised YFP^N & YFP^C, YFP^N & PbrMYB3/65-YFP^C, and PbrMYB3/65-YFP^N & YFP^C.

Homology modeling and molecular docking

AlphaFold2 was utilized to predict the protein structure of PbrACO2, while citrate's structure was transformed into a 3-D configuration using Gaussview through geometry optimization (B3LYP with the def2-TZVPP) (Jia et al. 2023). Molecular docking of PbrACO2 and citrate was performed using MOE-Dock (Vilar et al. 2008). The configuration exhibiting the highest binding score was selected and visualized using PyMOL software (Zhang et al., 2021).

Data analyses

The data presented represent the mean value of three biological replicates, with the exception of gene expression profiles during'Dangshansuli'fruit development (one replicate). Statistical analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC), particularly the analyses of variance (PROC ANOVA) with multi-comparison correction. Duncan's multiple range test at the 0.05 level was employed for mean separation. The Spearman correlation coefficient between attributes was calculated using R package, where correlations of 0.8–1.0 (absolute value) were considered extremely strong, and 0.6–0.8 (absolute value) were considered strong (Long et al. 2014).