Allelic variation in an expansin, MdEXP-A1, contributes to flesh firmness at harvest in apples

Abstract

Flesh firmness is a core quality trait in apple breeding because of its correlation with ripening and storage. Quantitative trait loci (QTLs) were analyzed through bulked segregant analysis sequence (BSA-seq) and comparative transcriptome analysis (RNA-seq) to explore the genetic basis of firmness formation. In this study, phenotypic data were collected at harvest from 251 F₁ hybrids derived from ‘Ruiyang’ and ‘Scilate’, the phenotype values of flesh firmness at harvest were extensively segregated for two consecutive years. A total of 11 candidate intervals were identified on chromosomes 03, 05, 06, 07, 13, and 16 via BSA-seq analysis. We characterized a major QTL on chromosome 16 and selected a candidate gene encoding expansin MdEXP-A1 by combining RNA-seq analysis. Furthermore, the genotype of Del-1166 (homozygous deletion) in the MdEXP-A1 promoter was closely associated with the super-hard phenotype of F₁ hybrids, which could be used as a functional marker for marker-assisted selection (MAS) in apple. Functional identification revealed that MdEXP-A1 positively expedited fruit softening in both apple fruits and tomatoes that overexpressed MdEXP-A1. Moreover, the promoter sequence of TE-1166 was experimentally validated containing two binding motifs of MdNAC1, and the absence of the MdEXP-A1 promoter fragment reduced its transcription activity. MdNAC1 also promotes the expression of MdEXP-A1, indicating its potential modulatory role in quality breeding. These findings provide novel insight into the genetic control of flesh firmness by MdEXP-A1.

Core

This study demonstrated the role of MdEXP-A1 in the formation of flesh firmness, and the genetic variation of TE-1166 was significantly correlated with flesh firmness at harvest, which could be used as a potential molecular marker for firmness traits at maturity to assist breeding. The deletion of TE-1166 reduced the expression of adjacent gene MdEXP-A1 and inhibited the decline of flesh firmness. The transcriptional regulation of MdEXP-A1 by MdNAC1 further affected the flesh firmness traits.

Gene & accession numbers

The transcriptome date during development were derived from Liu et al. (2021) and are available in the NCBI database: the accession number is PRJNA728501. The sequence data used in this study can be found in the GenBank data libraries under the following accession numbers: MdEXP-A1 (MD16G1070600), MdNAC1 (MD13G1069200), MdACS1 (MD15G1302200), MdACO1 (MD10G1328100), and MdPG1 (MD10G1179100).

Introduction

Fruit ripening and softening are elaborate and highly coordinated programs that accompany dramatic physicochemical changes (Han et al. 2016; Chen et al. 2020). Flesh firmness is an intrinsic quality trait that has an essential influence on consumer preferences and is also an indicator of harvest date and shelf life. A texture that is too soft or overly hard is not desirable (Hu et al. 2020). Excessive softening not only enhances disease susceptibility but also greatly limits transportability and storability (Su et al. 2023). As a quantitative genetic trait, flesh firmness is controlled by multiple genetic effects and external environmental factors (Chagné et al. 2014). Despite a few decades of studies that have explored the physiological or genetic mechanisms underlying fruit softening, their application in marker-assisted selection (MAS) remains limited (Wu et al. 2021b). A clear and systematic understanding of the unique genetic mechanisms of ripening and softening is highly important for providing guidance for genetic improvement.

Apples (Malus × domestica Borkh) are widely appreciated and grown in temperate regions (Duan et al. 2017). Studies on apple segregation population have identified plenty of quantitative trait loci (QTLs) for flesh firmness or softening on chromosomes 01, 03, 06, 08, 10, 15, and 16 (Longhi et al. 2012; Kumar et al. 2013; Bink et al. 2014; Di Guardo et al. 2017; Chagné et al. 2014). However, in the case of wide candidate intervals, the determination of core variations in key traits remains challenging (Guo et al. 2020). To date, several crucial genetic variations, especially in MdACS1, MdACO1, MdPG1, and MdEXP7, have presented an association between flesh softening and genetic control (Costa et al. 2005; Costa et al. 2008; Costa et al. 2010). However, increasing evidence has confirmed that genetic diversity can explain only limited phenotypic variation (Nybom et al. 2013; Wu et al. 2021a). With the abandance of high-quality genomes, many studies have suggested that structural variations (SVs) largely contribute to phenotypic variance in diverse species (Li et al. 2023a; Wang et al. 2023c). Consequently, a novel high-throughput sequencing strategy should be used to identify genetic variants that are more accurately associated with flesh firmness at harvest.

In climacteric fruits such as apples, ethylene plays a pivotal role in fruit maturation. ACC oxidase 1 (ACO1) and ACC synthase 1 (ACS1) are located within a well-known QTL interval for fruit softening and are genetically linked to ethylene synthesis (Harada et al. 2000; Oraguzie et al., 2004). Recently, a sery of allelic variations in ethylene response factors has been shown to be closely associated with firmness loss and ethylene production in apples (Hu et al. 2020; Wu et al. 2021b). However, insight into fruit softening before harvest remains largely unexplored. Because three traits (ripening date, firmness at harvest, and softening after harvest) are closely linked, systematic research is needed to elucidate this relationship.

During the development process of fruit, depolymerization of cell wall components and destruction of structure decrease cell adhesion and accelerate fruit softening (Brummell 2006). Moreover, cell walls often become nonexpandable, which makes expansins crucial since they positively participate in cell enlargement and cell wall creeping (Cosgrove 1997; Rose et al. 1997). Cell wall loosening during the ripening process is caused by synergistic modification of numerous cell wall hydrolases and non-enzymatic proteins known as expansins (EXPs). To date, knocking out only one cell wall-modifying gene other than pectate lyase does not fully explain the loss of firmness. Currently, expansins have been proposed as the main catalysts for breaking the frame construction formed by cellulose microfibrils and xyloglucan, thereby exposing polysaccharides to cell wall hydrolases (McQueen-Mason et al. 1992; Cosgrove, 2016).

EXPs are structural proteins involved in a variety of diverse biological processes, such as fruit ripening and softening (Costa et al. 2008; Perini et al. 2017; Mu et al. 2021) and flesh crispness retention (Costa and Stella, 2005; Trujillo et al. 2012). EXPs decrease the adhesion between adjacent cell wall polymers, as they directly mediate the creep of walls at low pH (Cosgrove 2000). Recently, EXP has been a hot research topic in the fruit softening process (Brummell et al. 1999; Valenzuela-Riffo and Morales-Quintana 2020). In tomatoes (Solanum lycopersicum L.), repression of LeEXP1 resulted in reduced softening and delayed ripening, whereas overexpression of LeEXP1 caused greater softening (Rose et al. 1997; Minoia et al. 2016). In apples, previous data suggested that varieties with a long shelf life have lower levels of MdEXP3 expression during the maturation stage (Wakasa et al. 2003). However, the role of EXP in apples remains uncharacterized.

EXP, as a structural gene, is controlled by many upstream transcription factors (TFs) and hormone signals (Mu et al. 2021; Chen et al. 2024). To date, numerous TFs that mediate ripening and softening by modifying cell wall-related genes have been identified (Zhang et al. 2022b, 2023a). For example, the ethylene-related TF MaERF11 delays maturation in bananas (Musa × paradisiaca) by inhibiting the transcription of MaACO1 and MaEXP (Han et al. 2016). In maize (Zea mays L.), ZmNAC29 enhances the transcriptional activity of ZmEXPB15, resulting in increases in kernel size and weight (Sun et al. 2022). In addition, it contains an abundant hormone-responsive element in the EXP promoter, which could participate in cell extension and cell wall modification regulated by various plant hormones, such as auxin-induced growth (Spartz et al. 2014; Cosgrove, 2016).

Our current understanding of ripening and softening relies on a few mutants in tomatoes, including non-ripening (NOR), ripening inhibitor (RIN), and colorless non-ripening (CNR) (Vrebalov et al. 2002; Manning et al. 2006; Gao et al. 2018). An increasing number of studies have shown that the NAC TF family plays a key role in regulating the ripening process in climacteric fruit. A number of NAC proteins from various species have been shown to be involved in ripening and quality traits formation by increasing the expression of genes associated with cell elongation, cell wall metabolism, and ethylene accumulation (Gao et al. 2020; Martín-Pizarro et al. 2021; Cao et al. 2023; Zhang et al. 2024).

Currently, data on EXP proteins and the flesh firmness of apples are still limited. Hence, a comprehensive understanding of the molecular regulation and genetic basis of this complex trait would provide potential insights into genetic modifications. Here, novel insights into firmness traits in apple fruit have been presented via the combination of BSA-seq, RNA-seq, and structural variation (SV) analysis. We functionally identified a candidate gene, expansin-A1, in apples, which is closely associated with flesh firmness. Additionally, a 1166 bp deletion in the promoter of MdEXP-A1 decreases its transcription activity, which could be recognized and upregulated by an NAC TF. In addition, MdNAC1 was also found to be capable of activating the expression of MdEXP-A1, which plays a positive regulatory role in ripening and softening.

Results

Segregation of flesh firmness at harvest

The ‘Scilate’ apple is popular worldwide for its hard-crisp texture, whereas the new cultivar ‘Ruiyang’ inherits the soft-crisp taste from its parent, ‘Fuji’. Throughout development, the firmness of ‘Scilate’ fruit was significantly greater than that of ‘Ruiyang’, and both the parents and representative progeny exhibited a dramatic decreasing trend (Fig. 1A-C). To understand the segregation regularity of flesh firmness, the phenotypic values of 251 F₁ hybrids from ‘Ruiyang’ and ‘Scilate’ were collected at harvest for two consecutive years (Tables S1, S2). As a result, the firmness data exhibited broad segregation at harvest, and the major distribution range was 6–10 kg/cm². The frequency distribution of firmness followed a normal distribution, suggesting that firmness at harvest is a multi-genic trait involving complex mechanisms (Fig. 1D, E). The broad-sense heritability (H²) was 84.81% and 91.36% in the two years for flesh firmness, respectively, indicating that the firmness trait at harvest was under strong genetic control in the offspring population (Table S3). The maximum firmness recorded was 14.18 kg/cm², while the minimal value was 3.80 kg/cm² (Table S3), indicating that the flesh firmness at harvest was widely segregated in the F₁ population from ‘Ruiyang’ and ‘Scilate’, which could be utilized for QTL mapping analysis.

Fig. 1

Phenotypic characterization and QTL identification for flesh firmness in the F₁ population derived from ‘Ruiyang’ and ‘Scilate’. A Phenotypes of ‘Ruiyang’ and ‘Scilate’ and their super-hard and super-soft F₁ individuals during diverse developmental periods. Images were digitally extracted for comparison. Scale bar =5 cm. B Flesh firmness changes in ‘Ruiyang’ and ‘Scilate’ apple fruits during various developmental stages. DAFB: days after full bloom. C Flesh firmness values of three super-hard hybrids (H1, H2, and H3) and three super-soft F₁ hybrids (S1, S2, and S3) from F₁ populations of apple fruit during different developmental stages. Different letters represent differences in the fruit development stages. From top to bottom, they are H1, H2, H3, S1, S2, and S3. Significance was defined at P < 0.05 (Student’s t-test). The frequency distribution of flesh firmness at harvest in the F₁ population from a cross of ‘Ruiyang’ and ‘Scilate’ during 2021 (D) and 2022 (E). F The flesh firmness phenotype value of 20 super-hard and 20 super-soft hybrids for bulked segregant analysis (BSA) construction. G Identification of a candidate chromosome region for the flesh firmness trait at harvest via BSA-seq analysis in apples. H Heatmap of candidate cell wall genes and transcription factors within the QTL region for apple flesh firmness. The error bars represent the means ± SDs of three repetitions. Asterisks indicate a significant difference determined via a t test (*P <0.05, **P <0.01)

Identification of MdACS1 and MdACO1 genotypes in F₁ individuals

MdACS1 and MdACO1 are two well-known molecular marker genes associated with ethylene release and postharvest fruit softening. The genotypes of these two markers were identified via published primers (Harada et al. 2000; Oraguzie et al. 2004; Wen et al. 2024). The parental genotypes for ‘Ruiyang’ and ‘Scilate’ was MdACS1-1/2:MdACO1-1/2 and MdACS1-1/2:MdACO1-2/2 (Tables S4, S5). The homozygous MdACS1 genotype was segregated at approximately 1:1 in F₁ super-hard or super-soft hybrids (MdACS1-2/2:MdACS1-1/1), and the segregation ratio of MdACO1 genotype was in a 1:1 ratio of the super-soft hybrids (Tables S4, S5). No significant effects were detected between the MdACS1 or MdACO1 genotype and flesh firmness at harvest in F₁ hybrids from ‘Ruiyang’ and ‘Scilate’. Therefore, we still need to map more QTLs closely linked to flesh firmness at harvest.

QTL analysis for flesh firmness in hybrids via BSA-seq

To uncover the key genetic basis and variation locus, two segregant bulks were constructed and re-sequenced from 20 super-hard and 20 super-soft individuals (Fig. 1F). Additionally, their parents were also re-sequenced. Clean reads for each sample (16,218,811,500 for super-hard pools and 16,755,700,200 for super-soft pools) were merged with the apple GDDH genome (Daccord et al. 2017). Candidate regions for flesh firmness at harvest were identified and located on Chr 03 (Chromosomes), 05, 06, 07, 13, and 16 from two segregant bulks (Fig. 1G, Fig. S1; Table 1). Among these QTLs, seven QTLs on Chr 03 overlapped at 33.2–33.3 Mb, and several QTLs were concentrated within an interval from 0‒17 Mb of Chr 16 (Table 1). The main QTL peaks were concentrated on five chromosome intervals (03, 06, 07, 13, and 16), and the G value was greater than the corresponding threshold value. The high enrichment variation of SNPs (Single nucleotide polymorphism) or InDels (insertion-deletion) in the above QTL regions may be positively associated with flesh firmness formation at harvest.

Table 1 QTL identification of apple flesh firmness characteristics

Candidate gene screening for flesh firmness-associated QTLs at harvest

To narrow the research on the basis of flesh firmness genetics, all 922 genes were downloaded from the above QTL signal intervals. By combining previous transcriptome data of extra-hard and extra-soft hybrids (Su et al. 2024), a total of 109 DEGs were screened via a strategy with a |log2 (fold change)|> 1.0 and P values < 0.05 (Fig. S2; Table S6). According to the reference genome annotation, 9 candidate genes involved in cell wall modification and 6 transcription factors were identified and screened (Fig. 1H; Table S7). Furthermore, since fruit softening is well known to involve the disassembly of cell wall matrix polysaccharides, 9 candidate genes involved in cell wall modification were selected for further investigation. These genes included expansins (MD05G1130300 and MD16G1070600), O-fucosyltransferase family proteins (MD07G1003800 and MD16G1026700), galacturonosyltransferase (MD16G1048100), cellulase protein (MD16G1104600), pectate lyase (MD16G1106400 and MD16G1161800), and beta-xylosidase (MD16G1158300). Additionally, on the basis of previously published transcriptome data under fruit development (Liu et al. 2021; PRJNA728501), the transcript levels of these candidate genes during development were analyzed in detail (Fig. S3; Table S8). Finally, the expression level of cell wall-loosening gene (MdEXP-A1, MD16G1070600) and beta-xylosidase gene (MD16G1158300) increased significantly in three accessions during the whole development period, which was closely related to the decline of firmness. Therefore, these two genes were chosen for further study and verification.

Analysis of genetic variation in candidate genes via Sanger sequencing

To explore the genetic variation of key candidate genes, the coding sequence (CDS) and promoter of the candidate genes (MdEXP-A1, MD16G1070600 and Mdβ-xyl, MD16G1158300) were amplified from parents and linked into a 19-T vector for Sanger sequencing. As a result, the MdEXP-A1 CDS in the parent was completely coincident with the ‘Golden Delicious’ reference genome (Fig. S4). However, upon comparison with the reference genome, 20 SNPs and two InDels in the promoter of MdEXP-A1 were identified. Moreover, a novel structural variation (SV) of 1166 bp was absent in the promoter region of MdEXP-A1 (Fig. 2A, Fig. S5 and S6). According to the website of the apple genome and CENSOR software, the SV loci were identified as transposable elements (TEs) that contain one LTR/BEL and one LTR/Gypsy family transposon (Fig. S7C). We used the 1166 bp sequence as a query to perform BLAST analysis on the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and found that similar transposons were highly abundant across distinct chromosomes of the apple genome (Fig. S8). This allelic variation of MdEXP-A1 was subsequently selected as a key genetic variation for further exploration.

Fig. 2

Identification and verification of a novel structural variation in F₁ progeny from two crosses. A Gene structure and upstream genetic variation analysis of the candidate gene MdEXP-A1. B Correlation analysis of the flesh firmness at harvest in 20 extra-hard and 20 extra-soft hybrids of ‘Fuij’ and ‘Cripps pink’ and the SV loci genotypes. C Pie chart analysis of different genotypic ratios of this SV in the MdEXP-A1 gene upstream in the 20 extra-hard individuals from ‘Fuij’ and ‘Cripps pink’. D Pie chart analysis of various genotypic ratios of this SV in the 20 extra-soft individuals from ‘Fuij’ and ‘Cripps pink’. E Correlation analysis of the flesh firmness phenotype in 47 super-hard and super-soft progeny of ‘Ruiyang’ and ‘Scilate’ and the SV locus genotypes. Pie chart analysis of various genotypic ratios of this SV in the 24 super-hard (F) and 23 super-soft individuals (G) from ‘Ruiyang’ and ‘Scilate’

TE-1166 deletion in MdEXP-A1 was associated with super-hard hybrids

To validate whether the SV loci was associated with flesh firmness at harvest, specific primers were designed via Primer 5 software (Table S11). The previous extra-hard and extra-soft F₁ offspring (40 individuals) and their parents (Su et al. 2024), ‘Fuji’ and ‘Cripps Pink’, were subsequently genotyped via PCR (Table S9). Among the parents, the genotypes of ‘Fuji’ and ‘Cripps Pink’ were identified as heterozygous Del:del and homozygous deletion (Del-1166) from our PCR product analysis (Table S9). Interestingly, the Del-1166 genotype was found to be present in almost all extra-hard individuals, accounting for approximately 60% (12/20) (Fig. 2B, C). Conversely, this genotype was observed in only 25% of the extra-soft progeny (5/20) (Fig. 2D). Furthermore, 47 super-hard and super-soft F₁ individuals from the cross of ‘Ruiyang’ and ‘Scilate’ were also genotyped. In the parents, the genotypes of the ‘Ruiyang’ and ‘Scilate’ accessions were identified as heterozygous Del:del (Table S10). Similarly, the Del-1166 genotype was closely linked to the super-hard hybrids, accounting for approximately 54.2% of the population (13/24) (Fig. 2E, F). In contrast, this genotype accounted for only 17.4% of the super-soft progeny (4/23) (Fig. 2G). In summary, the absence of the TE-1166 locus in the MdEXP-A1 promoter was strongly correlated with the ultra-hard phenotype, and it could serve as a potential genotyping marker for assessing harvest firmness in apple MAS. Therefore, we speculated that the adjacent gene MdEXP-A1 plays a crucial role in the firmness formation of apples because of its location within a significant QTL interval for flesh firmness, indicating its potential biological function as an essential candidate gene.

Low expression of MdEXP-A1 was closely related to the super-hard phenotype

Phylogenetic tree analysis highlighted the close relationship between MdEXP-A1 and other EXP proteins in the Rosaceae family (Fig. 3A). As indicated by previous transcriptome data under apple fruit development (Liu et al. 2021), this analysis revealed that the MdEXP-A1 gene was present throughout the entire development process. The transcript of MdEXP-A1 exhibited considerable accumulation during the development of the three apple varieties, and it was strongly expressed in both flowers and fruits according to RT‒qPCR (Reverse transcription-quantitative polymerase chain reaction), indicating that the expression of MdEXP-A1 was highly positively correlated with a decrease in firmness during development (Fig. 3B, C). Additionally, the results revealed that the expression of MdEXP-A1 was significantly lower in three super-hard F₁ individuals (4J-11, 4J-63, and 4J-154) with the Del-1166 genotype than in the super-soft population (4J-112, 4J-103, and 4J-130) with the Del/del genotype at harvest (Fig. 3D). Additionally, after the fruits were sprayed with the four hormones, the transcript level of MdEXP-A1 was strongly elevated in the NAA treatment group compared with the control group, indicating that the expression of MdEXP-A1 was potentially regulated by the auxin signaling pathway (Fig. 3E). Similarly, its transcriptional level reached its peak at the fruit expansion stage during the development of the parents, ‘Ruiyang’ and ‘Scilate’, supporting its potential role in regulating cell enlargement and cell wall extensibility during development (Fig. 3F). Overall, we propose that MdEXP-A1 plays an active role in the development stage. In addition, the subcellular localization of the MdEXP-A1 protein was located in the chloroplast, whereas the 35S::GFP was distributed throughout the cell, which may be related to cytoplasmic inheritance (Fig. 3G). As a result, MdEXP-A1 was subjected to further functional validation.

Fig. 3

Identification of the MdEXP-A1 gene correlated with flesh firmness. A Phylogenetic relationship analysis of MdEXP-A1 in apples and orthologs in diverse species. B Expression patterns of MdEXP-A1 in fruits of ‘Fuji’, ‘Cripps Pink’, and ‘Ruixue’ fruits during various developmental stages. C Relative expression analysis of MdEXP-A1 in diverse apple tissues via RT‒qPCR. Different letters represent significant differences between diverse tissues. Significance was defined at P < 0.05 (Student’s t-test). D Transcription level of MdEXP-A1 in super-hard (Del-1166) and super-soft (Del/del-1166) F₁ hybrids at harvest period. E Expression level of MdEXP-A1 in apple fruits after treatment with four hormones. F Relative expression level of MdEXP-A1 in the parents of ‘Ruiyang’ and ‘Scilate’ at various developmental stages. G Subcellular localization of the MdEXP-A1 protein. Scale bars = 20 μm. FPKM: Fragments per kilobase of exon model per million mapped fragments. The values are presented as the means ± SDs. Asterisks indicate a significant difference determined via a t test (*P <0.05, **P <0.01,***P <0.001)

MdEXP-A1 promotes fruit softening in ‘Fuji’ apples

The EXP-A1 protein of nine species was highly conserved and may have similar functions (Fig. S9). To explore the potential function of MdEXP-A1 in firmness formation, we first cloned the CDS of MdEXP-A1 from ‘Golden Delicious’. Next, the MdEXP-A1-pCAMBIA2300 recombinant plasmids were used for transient overexpression in ‘Fuji’ apple fruits. After being placed in the dark for five days, a remarkable increase in MdEXP-A1 expression and the content of water-soluble pectin (WSP) and a dramatic reduction in the content of cell wall materials, hemicellulose, and cellulose were detected in fruits transiently injected with MdEXP-A1-pCAMBIA2300 compared with those injected with the empty vector (Fig. 4A-F). Moreover, the VIGS (virus-induced gene silencing) technique was used to further verify the effects of MdEXP-A1 on flesh firmness. Compared with those in ‘Fuji’ apples infected with a mixture of pTRV2 and pTRV1 vectors, the expression levels of MdEXP-A1 and WSP content in pTRV2-MdEXP-A1-transformed apple fruits were significantly lower after five days of infiltration (Fig. 4B, D). The contents of cell wall material, hemicellulose, and cellulose after silencing by MdEXP-A1 were significantly higher than those after being injected with the empty vector (Fig. 4C, E, and F). These data indicate that MdEXP-A1 can promote apple flesh softening by disrupting pulp cell adhesion and the polysaccharide structure of the cell wall.

Fig. 4

MdEXP-A1 positively facilitates flesh softening in apples and tomatoes. A Transient overexpression and silent expression of MdEXP-A1 in ‘Fuji’ apple fruits. Images were digitally extracted for comparison. Scale bar = 1 cm. B Relative expression levels of MdEXP-A1 in apple fruits after transient injection. C Determination of fruit cell wall material after transient expression of MdEXP-A1. Cell wall component content of WSP (D), hemicellulose (E), and cellulose (F) in instantaneously injected apple fruits. G The relative expression of the MdEXP-A1 gene in break-stage fruits of three independent over-expression lines. H Phenotypic observation of WT and MdEXP-A1-OE lines at diverse stages of fruit development and ripening phases. The fruits from three independent transgenic tomatoes showed an early veraison phenotype. Scale bar = 0.5 cm. I Firmness changes in MdEXP-A1 transgenic tomato lines and wild-type tomatoes during development and post-harvest stages. J Fruit cell structure of MdEXP-A1 transgenic tomato lines and wild-type during the Break+7 stage. Scale bar = 100 μm. The error bar represents the standard error of 3 replicates. Asterisks indicate a significant difference determined via a t-test (*P <0.05, **P <0.01)

MdEXP-A1 promotes early color breakdown in transgenic tomato

Given that tomatoes serve as a classic model for the genetic studies of fruit development and ripening, we overexpressed MdEXP-A1 in ‘Micro Tom’. Several independent transgenic lines of MdEXP-A1 were obtained via the leaf disc method. Furthermore, they were identified via PCR and RT‒qPCR. The amplified target DNA band and increasing transcript level indicated successful overexpression of MdEXP-A1 in tomato (Fig. 4G; Fig. S10). Compared with those of wild-type (WT) fruits, the fruits of three stable T3 generation seedlings whose expression was high presented accelerated ripening (Fig. 4H). Additionally, no significant difference in flesh firmness was detected at the mature green (MG) stage, whereas the overexpression of MdEXP-A1 led to softer tomato fruits, especially at the BR, B7, and B15 stages (Fig. 4I). Our cytological observations revealed that the cell gap of the transgenic fruits was greater than that of the WT fruit at the B7 stage (Fig. 4J). These results demonstrated that the candidate gene MdEXP-A1 was responsible for firmness formation during development.

The deletion of TE-1166 in the MdEXP-A1 promoter reduces its transcription activity

Owing to the SV loci occurring in the promoter region of MdEXP-A1, the specific 1166 bp sequence was analyzed in detail according to the PlantCARE website (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/). As a result, a series of cis-acting elements that can be recognized by hormones, light signals and multiple TFs, such as MYC, ERF, and NAC, were identified (Fig. S11). In addition, the SV loci contain multiple CAAT boxes, which play a strong regulatory role in enhancing promoter activity and increasing gene transcription levels. Moreover, many cis-acting elements of the full-length promoter (2207 bp) were also identified (Fig. S12). To further investigate whether this SV affects its transcriptional activity, the promoter of MdEXP-A1 was amplified via segmentation and independently linked into the pCAMBIA1301 vector with 35S removed. Tobacco leaves were subjected to transient transformation via Agrobacterium injection. As shown in Fig. 5A, transformants of pro1829-MdEXP-A1::GUS (containing TE-1166) presented greater GUS staining intensity than did those of pro643-MdEXP-A1::GUS (lacing TE-1166). These data suggested that the TE-1166 deletion of the MdEXP-A1 promoter could reduce its transcription activity (Fig. 5A, B).

Fig. 5

Transcriptional activity analysis of the structural variation and identification of a regulatory transcription factor, MdNAC1. A B-glucuronidase (GUS) staining of the truncated fragment of the MdEXP-A1 promoter in transgenic tobacco leaves. B The relative activity of GUS in the truncated promoter of MdEXP-A1. C Transcript accumulation of MdNAC1 in ‘Fuji’, ‘Cripps Pink’, and ‘Ruixue’ fruits during development. D Expression patterns of MdNAC1 in parents of ‘Ruiyang’ and ‘Scilate’ fruits throughout the entire developmental stage. E Subcellular localization signal of the MdNAC1 protein in tobacco leaves. Scale bars = 20 μm. F Transcriptional activation activity analysis of the MdNAC1 protein in the yeast system. The error bars represent the standard error of 3 replicates. Asterisks indicate a significant difference determined via a t test (**P <0.01)

MdNAC1 binds the MdEXP-A1 promoter and promotes its transcriptional activity

Owing to the absence of a large fragment on the MdEXP-A1 promoter, it was speculated that transcriptional regulation is one of the major causes of the different expression patterns of MdEXP-A1 in super-hard and super-soft hybrids. Additionally, several cis-regulatory elements of the SV have been explored, and two NAC-specific binding motifs (AC GTA) have been discovered (Zhong et al. 2010; Nieuwenhuizen et al. 2015). Numerous studies have shown that the NAC protein is a strong predictor of maturity and softening (Yeats et al. 2019; Cao et al. 2021). Next, we identified an NAC TF called MdNAC1, which is involved in the regulation of pigment accumulation (Liu et al. 2023). Furthermore, the expression of MdNAC1 was assessed in both parents and three cultivars during development (Fig. 5C, D). These results indicate a significant increase in its transcriptional level in ‘Ruiyang’, ‘Scilate’ and other varieties throughout development, supporting its potential role in ripening and softening. The MdNAC1 protein subsequently localized to the cell nucleus with transcriptional activation activity in the yeast system (Fig. 5E, F), suggesting its typical TF characteristics. To validate whether MdNAC1 can bind to the deleted promoter of MdEXP-A1, the fragment (P‒2) of the promoter was inserted into the pAbiA vector for yeast one-hybrid analysis. The result indicated that MdNAC1 can interact with the deletion promoter fragment of MdEXP-A1 (Fig. 6A, B). Furthermore, two specific probe primers were designed for the EMSA (electrophoretic mobility shift assay) experiment, and the results again revealed that MdNAC1 can bind these two sites (Fig. 6C, D). To explore the effect of binding on MdEXP-A1 transcription, a luminescence (LUC) assay involving genetic transformation of tobacco was performed. The LUC intensity results indicated that MdNAC1 can increase the expression of MdEXP-A1 (Fig. 6E–G). In conclusion, our findings revealed that MdNAC1 could bind to the promoter fragment and activate the transcription of MdEXP-A1.

Fig. 6

MdNAC1 binds the MdEXP-A1 promoter and promotes its transcriptional activity. A Location analysis of NAC-binding cis-acting elements in the MdEXP-A1 promoter sequence. B Yeast one-hybrid assay results showing showed that MdNAC1 can interact with the P-2 fragment of the MdEXP-A1 promoter, and the empty vector (AD) was used as a negative control. C The probe in the MdEXP-A1 promoter sequence, the binding probe, is a biotin-labeled fragment containing an [ACG] motif. The cold probe is an unlabeled competitive probe (50 times the concentration of the binding probe). The mutation probe contains seven or six nucleotide substitutions. D Electrophoretic mobility shift assays revealed that MdNAC1 binds to the specific motifs P1 and P2 of the MdEXP-A1 promoter. The ‘ + ’ and ‘-’ represent the presence and absence of the probe or protein, respectively. E Constructs used to test the role of MdNAC1 in MdEXP-A1 expression in transient expression assays. F and G LUC assay indicating that MdNAC1 positively promotes the activity of the MdEXP-A1 promoter. The values are the means ± SDs of three independent biological replicates; Asterisks indicate significant differences determined via a t test (**P < 0.01)

MdNAC1 expedites fruit ripening and softening

MdNAC1 can increase the transcription level of MdEXP-A1, which may play a positive regulatory role in the ripening and softening of fruit. Three independent MdNAC1-overexpressing lines were subsequently obtained from ‘Orin’ apple calli. The positive transgenic calli lines were subsequently verified via PCR and RT‒qPCR (Fig. 7A, Fig. S13). Furthermore, by using |log2FC|> = 1 and q < 0.05 as the criteria, a total of 1899 DEGs were identified via RNA-seq analysis of MdNAC1 transgenic lines and WT calli (Fig. 7C, Fig. S14A). GO analysis revealed that those DEGs were enriched in cell wall metabolism, plant hormone signaling, and the defense response pathway (Fig. S14D). Compared with those in the control lines, 34 cell wall-related DEGs, including MdEXP-A1 gene, were upregulated and 7 were downregulated (Fig. 7D, highlighted in red box). The expression of the firmness-related marker genes MdACS1, MdACO1, and MdPG1 was obviously elevated, supporting its regulatory role in ethylene synthesis and cell wall metabolism. As expected, the relative expression of the MdEXP-A1 gene was significantly greater in the MdNAC1-overexpressing lines than in the control calli (Fig. 7E). Combined in vitro and in vivo experiments provided novel insight into the regulation of flesh firmness formation by demonstrating the involvement of MdNAC1 in the modulation of the deleted fragment of MdEXP-A1. In summary, the above studies demonstrated that the presence of TE-1166 can be recognized by upstream MdNAC1 transcription factor and promote the expression of the downstream gene MdEXP-A1, which is involved in the formation of flesh firmness at harvest (Fig. 8).

Fig. 7

Overexpression of MdNAC1 in stable transgenic apple calli. A Phenotype of MdNAC1 stable overexpressing ‘Orin’ calli lines. B The relative expression levels of the MdNAC1 gene in transgenic apple calli were determined via RT‒qPCR. C The number of differentially expression genes (DEGs) between WT and MdNAC1 transgenic calli. up: up-regulated; down: down-regulated. D Heatmap of all DEGs involved in cell wall metabolism from WT and stable MdNAC1 transgenic calli lines. E Transcription levels of firmness-related genes in WT and 35S::MdNAC1 calli

Fig. 8

A proposed model for illustrating the effects of the TE-1166 and MdNAC1 on the expression of MdEXP-A1 in regulating apple flesh firmness. The thin and thick lines with arrows indicate low and high expression of MdEXP-A1, respectively. The wavy line indicates the abundance of MdEXP-A1 expression protein. TE-1166 can be recognized and upregulated by the upstream regulatory factor MdNAC1, resulting in super-soft-fleshed hybrids. The dotted arrows indicate that the mechanism of ERF on MdEXP-A1 is unknown

Discussion

Flesh firmness at harvest is a major objective breeding trait worldwide. In the practical production of apples, flesh firmness is commonly used as an indicator of ripeness. In general, early-maturing cultivars tend to have shorter shelf life (Kenis et al. 2008). Additionally, for one cultivar, higher firmness at harvest leads to a longer shelf-life and highlights the tight genetic linkage in those traits. Nevertheless, the underlying mechanisms remain elusive. Owing to the high heterozygosity of apple, the key challenge in determining the genetic mechanism of flesh firmness is to identify major genes and functional variants.

A total of 251 hybrids from ‘Scilate’ and ‘Ruiyang’ were investigated for flesh firmness at harvest over two years. These populations exhibited extensive segregation (Fig. 1D, E), suggesting that flesh firmness is a typical quantitative trait controlled by multiple genes. Next, we combined BSA-seq and QTL analysis to identify major genetic loci mapped, especially to Chr03, Chr06, Chr07, and Chr16. In previous studies, numerous QTLs and candidate genes associated with apple fruit softening have been identified. For example, NAC18.1 was characterized within a major QTL on Chr03 for apple maturity date, and flesh firmness at harvest variation on its CDS led to various harvest dates (Migicovsky et al. 2016; Yeats et al. 2019). In addition, a series of natural variations in apple ERF TFs have been revealed on Chr03 and Chr16, which are responsible for ethylene production and firmness loss (Hu et al. 2020; Wu et al. 2021b). Those studies indicated that Chr03 and Chr16 were hotspot intervals of flesh firmness and postharvest softening.

To date, abundant genetic variations have been progressively elucidated and converted into molecular markers for firmness and shelf-life prediction, especially in MdACS1 and MdACO1 (Costa et al. 2005; Oraguzie et al. 2007; Costa et al. 2010). In our study, there was no significant correlation between the MdACS1 and MdACO1 genotypes and firmness diversity in the populations of ‘Ruiyang’ and ‘Scilate’. It is possible that MdACS1 or MdACO1 may function only in certain combinations of parental genotypes. Additionally, a growing body of research has reported that the predictive power of these functional markers is limited in terms of phenotype diversity (Nybom et al. 2013; Wu et al. 2021a), suggesting that many other genes and variants that may contribute to firmness change. These findings further indicate that the mechanisms of flesh firmness at harvest and postharvest softening vary, which is in agreement with previous research (Wu et al. 2021b).

Through multi-omics analysis of BSA and comparative transcriptome sequencing in extra-hard and extra-soft hybrids, nine cell wall-associated candidate genes within the QTL region were screened. In addition, combined with the transcriptome data under fruit development previously published by Liu et al. (2021), the EXP and beta-xylosidase candidate genes with significantly increased expression at a major QTL interval (Chr16) were selected for further study. With the development of high-throughput re-sequencing technologies, many studies have revealed that large SVs play essential roles in the adaptive evolution of species and key trait formation (Zhang et al. 2019; Li et al. 2023a). Next, a comparison of the results of genome analysis by next-generation re-sequencing and Sanger sequencing helped identify a novel SV in the promoter of MdEXP-A1 (Fig. S5, S6). Furthermore, the genotype of this TE-1166 was validated in F₁ super-hard and super-soft populations, and the results revealed that homozygous deletion of this SV was strongly associated with the super-hard phenotype at harvest, which could be a potential diagnostic marker for the firmness trait of apple MAS application. Together, the evidence again suggested that the adjacent gene MdEXP-A1 of this variant was closely linked to hardness formation at harvest.

As mentioned above, the MdEXP-A1 proteins are highly conserved in diverse species, indicating that they may perform similar biological functions. EXP is a non-enzymatic cell wall loosening protein that participates in cell enlargement and cell wall creep, which is positively associated with fruit ripening and softening (Cosgrove 1997; 2016; Rose et al. 1997). For example, MdEXP7 was mapped on Chr01 in a major QTL interval, and its allelic variation in the promoter was associated with fruit softening (Costa et al., 2008). Owing to the extension and expansion of the cell wall during development, the cell turgor and fruit firmness decreased sharply, and the intercellular space increased, so the EXP protein was proposed to be associated with changes in flesh texture. In this study, a candidate EXP gene was selected from a multi-omics conjoint analysis. As the fruit ripens, the transcript level of MdEXP-A1 in the parents peaks during the expansion stage, and its transcript level decrease after plung into ripening (Fig. 3F). These results are consistent with previous reports on apples (Wakasa et al. 2003) and pears (Hiwasa et al. 2003), supporting the potential role of fleshy fruits during development.

EXP crude protein was first extracted from cucumber seedlings and is well known to induce the extension and relaxation of cell walls via the ‘acid growth’ mechanism (McQueen-Mason et al. 1992; Cosgrove, 2016). According to our earlier study, flesh malic acid sharply decreases throughout development (Su et al. 2022). Therefore, we speculated that its expression activity was inhibited at harvest. Additionally, preharvest NAA treatment promotes the expression of MdEXP-A1, which is consistent with their role in the response of cell elongation to auxin-induced acidification (Cosgrove, 2016), but the in-depth mechanism of their interaction requires further exploration. In our study, transient overexpression of MdEXP-A1 clearly disrupted the cell wall polysaccharide matrix structure of apple, and ectopic expression of MdEXP-A1 facilitates the ripening and softening of tomato fruits (Fig. 4), which have functions similar to those of SlEXP1 in tomatoes (Brummell et al. 1999). Cell wall disassembly during the softening process is caused by the coordinated action of cell wall hydrolases and expansin (Brummell and Harpster 2001; Brummell 2006). A two-step model system for fruit softening in which expansins act in the expansion stage and cell wall polysaccharide are subsequently degraded by cell wall enzymes has been proposed. Inhibition of EXP1 gene alone had no significantly effect on fruit firmness in tomatoes. A recent study demonstrated that SlExp1 and SlCel2 coordinate to regulate the decline in tomato firmness (Su et al. 2023). Overall, flesh firmness is a multigenic trait associated with cell wall metabolism.

On the basis of our variant validation, the absence of TE-1166 in the MdEXP-A1 promoter was obviously correlated with super-hard hybrids at harvest. In addition, the expression of MdEXP-A1 in the super-hard offspring with Del-1166 was noteworthy lower than that in the super-soft populations of the Del:del genotype (Fig. 3D). Furthermore, the presence of Del-1166 increased its transcriptional activity (Fig. 5A, B). We speculated that the existence of this SV may facilitate the activity of the MdEXP-A1 protein. In addition, this variation location could be recognized by several TFs, such as MYC, ERF, and NAC (Fig. S11). In particular, NAC has been proposed as a strong determinant of fruit harvest date and firmness-related traits (Yeats et al., 2019). These findings suggested that the variant site of MdEXP-A1 may be subject to specific transcriptional regulation. These results again indicate that the MdEXP-A1 is a major functional protein affecting flesh firmness and that its allelic variation may play a key role in the evolution and domestication of apple fruits.

The highly conserved function of the NAC orthologs in diverse species has been proven to play an active regulatory role in ripening and softening (Zhang et al. 2024). What’s more, the expression level of MdNAC1 trended to increase during development in multiple varieties and exhibited transcriptional activation activity (Fig. 5C, D, and F). These results may support its critical effect on the ripening pathway and downstream structural gene regulation. In our study, the upregulated DEGs from the stable overexpression lines suggested that MdNAC1 has multiple positive effects on fruit ripening and softening (Fig. 7C, D). However, it cannot be excluded that other apple NAC members and TFs play similar roles in apple maturity. Our study integrated EMSA, Y1H (yeast one-hybrid), Dual-Luciferase, and stable overexpression assays and revealed that MdNAC1 upregulates the MdEXP-A1 mechanism for flesh firmness. This finding is in agreement with a previous study in which the expression of EXP was inhibited in NOR mutants (Gao et al. 2018). Furthermore, the expression of MdACS1 and MdACO1 was significantly elevated in the MdNAC1 transgenic calli. These results demonstrated that MdNAC1 may promote the ripening and softening of apples by controlling ethylene production and cell wall creeping. Recently, reports have revealed strong similarities in the modification of NAC homologs in tomato (Gao et al. 2020), strawberry (Li et al. 2023b), banana (Shan et al. 2012), and peach (Zhou et al. 2023; Zhang et al. 2024). Additionally, the NAC protein participates in the regulation of quality formation by modifying specific structural genes; therefore, investigating distinct genetic networks to balance ripening and quality is crucial for breeding programs (Cao et al. 2023).

Materials and methods

Plant materials and growth conditions

A total of 251 F₁ individuals from a cross of ‘Ruiyang’ and ‘Scilate’ were planted on M26 rootstock at the experimental station of Northwest A&F University, Baishui County, Shaanxi Province, China. Both parents and super-hard or super-soft progeny were sampled individually from 80 days after full bloom (DAFB) until ripening, with a 20-day interval. The root, leaf, flower, pericarp, pulp, and branch tissues were collected from the ‘Fuji’ cultivar. The management level of the orchard was coincidental for all apple offspring. Apple fruits were chosen according to taste, base color, days after full bloom, and degree of starch clearance (starch value of 7). Three biological replicates of 4–6 fruits per replicate were selected for detection.

Exogenous hormone treatment

One hundred fruits of ‘Ruixue’ apple (each twenty) were sprayed with 0.5 mM jasmonic acid methylester (MeJA), 0.5 g/L ethylene (ETH), 4 mM naphthylacetic acid (NAA), or 0.4 g/L abscisic acid (ABA) two weeks before harvest, and water spray was used as a control. The surface of the apple fruits was sprayed with the above hormone mixture until it had dripped down. The fruits were sprayed again after 7 days and finally picked at harvest.

Determination of flesh firmness

The firmness of all the apple samples was assessed at opposite sides of the equatorial portion of the peeled fruit with a texture analyzer (GS-15, Germany). The flat probe has a diameter of 10 mm and a test depth of 8 mm. Alternatively, the hardness of the tomatoes was determined by another texture analyzer (TA-XT plus, UK) with a 2 mm diameter probe (1.5 mm/sec, depth: 8 mm). Three biological replicates were performed, with six fruits as one replicate.

BSA-seq and QTL identification

On the basis of firmness data from the hybrids for two consecutive years, a total of 40 individuals with super-hard or super-soft phenotypes were selected to establish two pools. Young leaves of representative hybrids or parents of ‘Ruiyang’ and ‘Scilate’ were used to extract total genomic DNA. Next, two bulked libraries were generated on the basis of the DNA equivalence principle. These libraries were subsequently sequenced via the Illumina sequencing platform, and 150 bp paired-end reads were obtained. Burrows-Wheeler Aligner (BWA) was used to align the processed reads of each sample against the reference genome. SNP/InDel detection and annotation were performed via GATK and ANNOVAR software. The Δ (SNP index) and G values analysis between the super-hard pool and the super-soft pool were subsequently calculated to obtain a candidate region correlated with flesh firmness at harvest.

Candidate gene prediction on the basis of BSA-seq and RNA-seq data

QTL identification of firmness at harvest was narrowed down to special regions by BSA analysis. Using combined transcriptome data from previous individuals with an extreme firmness phenotype or during development, several key candidate genes were subsequently screened to explore genetic variation and function. Briefly, all genes were downloaded from the QTL regions for flesh firmness at harvest, and DEGs were further screened from previous RNA-seq data. According to the annotation of the apple genome, genes associated with cell wall metabolism were chosen as candidate genes for flesh firmness trait. The transcript levels of candidate genes during development were subsequently analyzed via previously published data (Liu et al. 2021).

Extraction of genomic DNA and RNA, and real-time qPCR analysis

The genomic DNA of representative F₁ hybrids and parents, tomatoes, and apple calli lines was generated with a plant DNA kit, and the total RNA of apple, tomato, and calli samples was extracted using a plant quick RNA kit (Tiangen, Beijing, China). Then, first-strand cDNA and RT‒qPCR were conducted as previously described (Su et al. 2022). Mdactin (MD01G1001600) was used as the internal reference. The primers used for RT‒qPCR are listed in Table S11.

Discovery and validation of genetic variation

The CDSs and promoters of the candidate genes were amplified and sequenced via Sanger sequencing. The design of specific primers for the MdEXP-A1 promoter was based on the reference genome (Table S11). The genotypes of the F₁ hybrids were identified by corresponding primers for MdACS1 and MdACO1 as previously described (Harada et al. 2000; Costa et al. 2005; Wen et al. 2024). The PCR system for this genetic structure of MdEXP-A1 was performed as previously reported by Ma et al. (2021).

Subcellular localization analysis

The recombinant proteins of MdEXP-A1 and MdNAC1 with a GFP label were ligated into the pCAMBIA2300 vector, and the corresponding primers used are shown in Table S11. The empty vector and each recombinant plasmid were individually transformed into Agrobacteriun strains (GV3101), after which they were separately infected into tobacco leaves (one month old). After three days of co-expression, GFP images were acquired via a confocal laser scanning microscope at an excitation wavelength of 488 nm (Leica TCS-SP8 SR).

Transcriptional activity assays

The transcriptional activity assay was performed as described by Wang et al. (2023a). The MdNAC1 CDS, lacking a termination codon, was constructed into the pGBKT7 carrier with a BD domain. The primers used above are listed in Table S11. The MdNAC1-BD fusion protein, pGBKT7 vector (negative control), and positive control were individually transformed into the yeast strain AH109. After being cultured on Trp-free media, the yeast cells were transferred to the SD-Trp-His condition and finally added to the X-α-Gal solution. The growth and staining results of yeast cells can be used as a reference for self-activated activity.

Transient overexpression and VIGS

The MdEXP-A1 CDS was amplified from ‘Golden Delicious’ fruits and inserted into the pCAMBIA2300 vector with a GFP label using double restriction enzyme site (KpnI and BamhI) to generate the recombinant constructs. The primers used are shown in Table S11. The specific fragment (400 bp) of the MdEXP-A1 CDS was subsequently cloned and inserted into the pTRV2 carrier for gene silencing. The control and fusion plasmids were subsequently individually transferred into the GV3101 Agrobacterium strain. The methods used for the transformation of apple fruits were previously reported (Wang et al. 2023b). The relative expression of the samples subjected to transient transformations was determined via RT‒qPCR.

Determination of cell wall components

The total cell wall material was extracted according to Su et al. (2022). In brief, frozen samples (5.0 g) were subjected to crude CWM extraction via dissolution in acetone and dimethyl sulfoxide. Then, distinct types of pectin polysaccharides were extracted via sodium acetate, (1, 2- cyclohexylenedinitrilo)-tetra acetic acid, and Na₂CO₃ (0.05 M) and determined via the carbazole ethanol method. Hemicellulose was subsequently extracted with KOH (4 M), which was measured via anthrone-sulfuric acid method. The resulting cell wall residue was dominated by cellulose.

Genetic transformation of apple calli and Micro-Tom

Apple calli was isolated from the ‘Orin’ embryo, which was subcultured every two weeks. The CDSs of MdNAC1 and MdEXP-A1 were connected to the pCAMBIA2300 vector for stable overexpression. Then, the MdNAC1 and MdEXP-A1 constructs were separately transferred into EHA105 Agrobacterium, and apple calli transformation of MdNAC1 was conducted on the basis of Zhang et al. (2022a). The tomato transformation of MdEXP-A1 was performed using the same method described previously (Fillatti et al. 1987). The positive seedlings were confirmed via PCR with detection primers (Table S11).

Yeast one hybrid assay

The Y1H assay was accomplished as previously described (Zhang et al. 2023b). The CDS of MdNAC1 was cloned and inserted into the pGADT7 effector carrier. The MdEXP-A1 promoter fragment (P-2) was linked to the pAbiA vector (reporter). The primers used for Y1H are listed in Table S11. Y1H yeast cells harboring the MdEXP-A1-P-2-pAbiA plasmid were cultured on SD-Ura medium, supplemented with aureobasidin A (AbA) at different concentrations to inhibit self-activating activity. The pGADT7-MdNAC1 construct was subsequently co-transformed into a Y1H cells containing MdEXP-A1-P-2-pAbiA. Next, the transformation products were transferred to screening medium lacking Leu supplemented with the corresponding AbA. The interaction results were evaluated by observing the growth of the strain carrying the pGADT7 vector as a control.

Electrophoretic mobility shift assay

The EMSA experiment was carried out as previously outlined (Wang et al. 2023b). The MdNAC1 CDS was linked into the pGEX-6p-1 vector and transformed into BL21 (DE3) Escherichia coli for GST-tag protein expression. Specific binding, competition, and mutation probes were obtained on the basis of the sequence of the MdEXP-A1 deletion fragment. The above probes were synthesized and processed at Sangon Biotech. EMSA was performed with a Chemiluminescent EMSA Kit (GS009, Beyotime). The primers used are shown in Table S11.

Transient dual-luciferase assay

For the dual-luciferase experiment, the MdNAC1 CDS was merged into the pGreenII 62-SK vector driven by CaMV35S (the effector). The full-length promoter of MdEXP-A1 (2179 bp) was integrated into pGreenII 0800-LUC (the reporter). All the constructs and empty vectors were introduced into the GV3101 strain harboring the pSoup vector. Then, they were transiently infected with one-month-old tobacco leaves. After three days of transformation, the LUC signal and activity were measured as previously described (Fu et al. 2021).

GUS staining analysis

To investigate whether the 1166 bp variation in the MdEXP-A1 promoter affects its transcriptional activity, truncated promoter fragments (Pro1809, Pro1049, and Pro643) were amplified from the DNA of ‘Golden Delicious’ through PCR. As previously described, a pCAMBIA1301 reporter vector without 35S was constructed with different segments of the above MdEXP-A1 promoter via restriction enzyme sites (HindIII and EcoRV). Three recombinant plasmids and the pCAMBIA1301 empty vector were transformed into to Agrobacterium GV3101 which was subsequently injected into tobacco leaves (one-month-old). Three days after injection, the leaves were clipped and stained with a GUS stain kit (SL7160, Coolaber) for 12h. Each experiment was repeated more than three times. All primers used are listed in Table S11.

Statistical analysis

The SPSS software (IBM) was used for statistical analysis. The data were analyzed according to a one-way ANOVA followed by Student’s t test analysis. Significant differences are indicated with asterisks (*P < 0.05, **P < 0.01).