Phosphorylation of MdCYTOKININ RESPONSE FACTOR4 suppresses ethylene biosynthesis during apple fruit ripening

Abstract

The plant hormone ethylene plays a central role in the ripening of climacteric fruits, such as apple (Malus domestica). Ethylene biosynthesis in apple fruit can be suppressed by calcium ions (Ca²⁺); however, the underlying mechanism is largely unknown. In this study, we identified an apple APETALA2/ETHYLENE-RESPONSIVE FACTOR (AP2/ERF) transcription factor, MdCYTOKININ RESPONSE FACTOR4 (MdCRF4), which functions as a transcriptional activator of ethylene biosynthesis- and signaling-related genes, including Md1-AMINOCYCLOPROPANE-1-CARBOXYLIC ACID SYNTHASE1 (MdACS1) and MdETHYLENE-RESPONSIVE FACTOR3 (MdERF3), as a partner of the calcium sensor, calmodulin. Ca²⁺ promoted the Ca²⁺/CaM2-mediated phosphorylation of MdCRF4, resulting in MdCRF4 recognition by the E3 ubiquitin ligase MdXB3 ORTHOLOG 1 IN ARABIDOPSIS THALIANA (MdXBAT31), and consequently its ubiquitination and degradation via the 26S proteasome pathway. This in turn resulted in lower expression of MdACS1 and MdERF3 and reduced ethylene biosynthesis. Transiently overexpressing various MdCRF4 proteins with specific mutated phosphorylation sites revealed that the phosphorylation state of MdCRF4 affects the ripening of apple fruit. The results reveal that a Ca²⁺/CaM-MdCRF4–MdXBAT31 module is involved in Ca²⁺-suppressed ethylene biosynthesis, which delays apple fruit ripening. This provides insights into fruit ripening that may result in strategies for extending fruit shelf life.

A calcium module downregulates the ethylene synthesis pathway during apple fruit ripening via regulating the phosphorylation status and degradation of MdCYTOKININ RESPONSE FACTOR4.

Introduction

Apple (Malus domestica) has one of the largest cultivation areas and yields of fruit crops, globally, and is a source of nutritionally valuable components of the human diet (Biedrzycka and Amarowicz, 2008). Apple fruit storability is an important economic trait and improper storage results in major losses in crop value (Tahir et al., 2007). In order to extend the shelf life of apple fruit, they are typically stored under low-temperature conditions or in controlled atmosphere warehouses before entering the supply chain (Hardenburg et al., 1954). In addition, 1-methylcyclopropene (1-MCP), an ethylene perception inhibitor, is a commonly used and very effective agent for preventing ethylene biosynthesis and signal transduction in fruit. Given its effect, 1-MCP is used commercially to extend the fruit shelf life (Watkins, 2006). These storage methods are employed to reduce biosynthesis of the gaseous hormone ethylene during fruit ripening, thereby extending shelf life. Ethylene plays a key role in triggering the ripening of climacteric fruits, such as apple and tomato (Solanum lycopersicum) (Yang and Oetiker, 1994; Barry and Giovannoni, 2007), resulting in cell wall degradation, fruit softening, and loss of quality (Tucker et al., 2017; Shi et al., 2021). Accordingly, regulating ethylene biosynthesis in fruit is a highly effective way to extend the shelf life of climacteric fruit and to improve its economic value.

The ethylene biosynthesis and signal transduction pathways have been widely characterized (De Paepe and Van Der Straeten, 2005). Ethylene biosynthesis begins with the formation of 1-aminocyclopropane-1-carboxylic acid (ACC) from S-adenosyl methionine (SAM) catalyzed by the enzyme ACC synthase (ACS); ACC is then oxidized by ACC oxidase (ACO) to form ethylene. Among the enzymes involved in ethylene synthesis, ACS is considered to be a rate-limiting enzyme (Xu and Zhang, 2015). In apple, MdACS1 is considered to be critical for regulating ethylene synthesis during fruit ripening, and MdACS1 silencing in transgenic apple fruit was found to block ethylene production (Dandekar et al., 2004). During signal transduction, ethylene is detected by its receptors and the signal is transmitted downstream through several components, including CONSTITUTIVE TRIPLE RESPONSE1 and ETHYLENE INSENSITIVE 2 (EIN2). A positive signal is then delivered to the primary transcription factor EIN3/EIN3-like, which induces the secondary transcription factors, ETHYLENE RESPONSE FACTORs (ERFs), which in turn activate the expression of downstream ethylene-responsive genes (Chen et al., 2005). We previously reported that MdERF2 and MdERF3, belonging to the APETALA2/ETHYLENE-RESPONSIVE FACTOR (AP2/ERF) family of transcription factors, directly bind to the MdACS1 promoter to inhibit or promote MdACS1 transcription during apple ripening (Li et al., 2016). Moreover, in Arabidopsis (Arabidopsis thaliana), mitogen-activated protein kinase3 (MPK3)/MPK6, phosphorylated ERF6 to enhance its stability and induce the expression of defense-related genes, including PLANT DEFENSIN1.1 (PDF1.1) and PDF1.2, in response to Botrytis cinerea infection (Meng et al., 2013). Thus, the phosphorylation status of ERF proteins can affect their transcriptional regulation of target genes.

Calcium ions (Ca²⁺) are well-characterized as conserved secondary messengers and mediators of plant responses to a wide range of developmental and environmental stimuli. Variations in intracellular Ca²⁺ concentrations influence the molecular, cellular, and genetic links between Ca²⁺ signatures and multiple downstream signaling events (Boudsocq and Sheen, 2013). Plants have three families of calcium sensors: calmodulin proteins (CaMs), calcineurin B-like proteins (CBLs), and calcium-dependent protein kinases (CDPKs). Among these sensors, the CaM proteins bind Ca²⁺ and undergo conformational changes to regulate diverse cellular processes in response to environmental signals by interacting with downstream target proteins (Boudsocq et al., 2010; Perochon et al., 2011; Qiu et al., 2012). The typical plant CaM protein has four (E)helix-loop-(F)helix (EF)-hand Ca²⁺-binding domains (BDs), located in pairs and embedded within two separate globular regions in the N- and C-termini, separated by a flexible central helix that forms a hallmark dumbbell-shaped structure (Reddy, 2001; Snedden and Fromm, 2001). Generally, CaM-interacting proteins contain the CaM-binding domain (CaMBD), which is responsible for interacting with other CaM proteins and usually has a positive net charge and a propensity to form an amphiphilic α-helix (Perruc et al., 2004; Du and Poovaiah, 2005; Du et al., 2009). It was recently found that upon wounding, the interactions between CaM1, CaM4, or CaM7 and JASMONATE-ASSOCIATED VQ MOTIF PROTEIN1 (JAV1) depend on an elevation in Ca²⁺ levels, which induces the phosphorylation of JAV1. This in turn causes a release of jasmonate (JA) biosynthesis repression, resulting in a rapid burst of JA biosynthesis, thus triggering plant defenses (Yan et al., 2018). This suggests that CaM-mediated protein phosphorylation may represent an important regulatory component of the Ca²⁺response.

Ethylene biosynthesis is affected by many internal and external factors and has also been associated with calcium signaling. For example, treatment of chickpea (Cicer arietinum) hypocotyls with the Ca²⁺ chelator ethylenebis (oxyethylenenitrilo) tetraacetic acid (EGTA) substantially inhibits ACO activity and ethylene production (Gallardo et al., 1999). This result was interpreted to suggest that calcium promotes ACO activity and increases ethylene production during normal hypocotyl growth and development. However, ethylene production was reported to be substantially inhibited in tomato and apple fruits treated with CaCl₂ (Chien Yi et al., 1993; Njoroge et al., 1998), and CaCl₂ treatment of fig (Ficus carica Linn) fruit improved its storage properties (Irfan et al., 2013), suggesting an impairment of ethylene biosynthesis. The above results indicate that calcium can inhibit ethylene biosynthesis in fruits and improve storage properties; however, the molecular mechanisms remain unclear (Gao et al., 2019).

In this study, we show that CaCl₂ treatment promotes Ca²⁺/CaM-mediated phosphorylation of MdCYTOKININ RESPONSE FACTOR4 (MdCRF4) in apple fruit, resulting in its ubiquitination and degradation. MdCRF4 is a positive regulator of MdACS1 and MdERF3 transcription, so reduced accumulation of MdCRF4 results in decreased MdACS1 and MdERF3 transcript abundance and lower levels of ethylene biosynthesis during fruit ripening. The data provide insights into the molecular basis by which Ca²⁺ suppresses fruit ripening.

Results

Ca²⁺ inhibits ethylene biosynthesis in apple fruit during ripening

To study how Ca²⁺ affects ethylene biosynthesis during apple fruit ripening, fruit were treated with 100-mM CaCl₂ and 1-mM LaCl₃ (a plasma-membrane Ca²⁺-channel blocker). Ethylene production was significantly inhibited by the CaCl₂ treatment during fruit ripening, compared with untreated control fruit (Figure 1, A and B). We also investigated the expression of MdACS1 and MdACO1, which control ethylene biosynthesis during apple fruit ripening and found that their transcript levels were significantly lower after the CaCl₂ treatment (Figure 1, C and D). We previously showed that MdERF2 and MdERF3 negatively and positively regulate the transcription of MdACS1 during apple fruit ripening, respectively (Li et al., 2016), so we also tested their expression. MdERF2 expression was not affected by the CaCl₂ treatment, while MdERF3 expression was significantly inhibited (Figure 1, E and F). The simultaneous application of CaCl₂ and LaCl₃ did not result in a reduction in fruit ethylene production or the CaCl₂-mediated effects on MdACS1, MdACO1, and MdERF3 expression (Figure 1), suggesting that Ca²⁺ downregulates ethylene production during apple fruit ripening by inhibiting the expression of ethylene biosynthesis- and ethylene signal transduction-related genes.

Figure 1.

CaCl₂ treatment suppresses apple fruit ripening. A, Apple fruit collected at 140 days after full bloom (commercial harvest day); CaCl₂, apple fruit were treated with CaCl₂; CaCl₂ +LaCl₃, apple fruit were treated with both CaCl₂ and LaCl₃; Untreated, apple fruit not receiving treatment. B, After treatment, ethylene production was measured. C–F, RT–qPCR was used to examine MdACS1, MdACO1, MdERF2, and MdERF3 expression during apple fruit ripening. The X-axes indicate the number of days of storage at room temperature after harvest. For RT–qPCR, fruit sampled at each sampling point were divided into three groups (three fruit per group). The fruit flesh from each group was evenly mixed for RNA extraction. RNA extracted from each group was used as one biological replicate. A total of three biological replicates were analyzed. The values represent means ± se. Statistical significance was determined using a student’s t test (**P < 0.01). n.s., no significant difference.

CaM2 senses Ca²⁺ and interacts with MdCRF4

CaM is ubiquitous in eukaryotic cells, where it binds calcium ions and participates in signal transduction. We identified eight CaM genes in the apple genome (Supplemental Table S1; MdCaM1–MdCaM8), of which MdCaM2, MdCaM3, MdCaM4, MdCaM5, MdCaM7, and MdCaM8 were expressed in fruit (Supplemental Figure S1). Since MdCaM2 had the highest expression (Supplemental Figure S1), we chose this gene for further study. We screened an apple fruit cDNA library corresponding to ripe fruit in storage, with MdCaM2 as a bait and found that an ERF/AP2 family transcription factor, named apple CYTOKININ RESPONSE FACTOR4 (MdCRF4, Sequence ID: NM_001328784.1) (Supplemental Figure S2), showed the strongest interaction. To verify the interaction between MdCaM2 and MdCRF4, an yeast two-hybrid (Y2H) assay was carried out. As shown in Figure 2A, MdCaM2-pGBKT7 and MdCRF4-pGADT7 were co-transformed into yeast cells and the empty vectors were used as controls. The result showed that yeast cells expressing both MdCaM2 and MdCRF4 grew normally in the selective medium and turned blue in the presence of the chromogenic substrate x-α-gal, indicating that MdCaM2 can interact with MdCRF4 in yeast cells (Figure 2A). To investigate whether this interaction is dependent on Ca²⁺, we carried out an in vitro pull-down assay with MdCaM2-GST and MdCRF4-His fusion proteins. MdCRF4-His bound to Ni-NAT Resin was used to capture the MdCaM2-GST protein. We observed that MdCaM2-GST was pulled down by MdCRF4-His, indicating that MdCaM2 interacted with MdCRF4. In addition, this interaction was promoted by Ca²⁺, but severely attenuated by EGTA (a Ca²⁺ chelator) (Figure 2B). Furthermore, an in vivo co-immunoprecipitation (co-IP) assay was performed to determine how Ca²⁺ affects the MdCaM2/MdCRF4 interaction using transgenic apple calli co-overexpressing MdCaM2-green fluorescent protein (GFP)/MdCRF4-Myc and empty-GFP/MdCRF4-Myc. MdCRF4 protein was immunoprecipitated from extracts from MdCaM2-GFP/MdCRF4-Myc transgenic calli, but not from extracts from the empty-GFP/MdCRF4-Myc transgenic calli, using an anti-GFP antibody, confirming the in vivo interaction between MdCaM2 and MdCRF4. In addition, CaCl₂ treatment enhanced MdCRF4 protein abundance co-immunoprecipitated by MdCaM2-GFP (Figure 2C). Finally, we used a luciferase complementation imaging (LCI) assay to confirm this result. Constructs containing MdCaM2 fused with N-terminus of luciferase (MdCaM2-nLuc) and MdCRF4 fused with C-terminus of luciferase (MdCRF4-cLuc) were co-infiltrated into Nicotiana benthamiana leaves, and the infiltrated leaves were treated with CaCl₂ or EGTA. A luminescence signal was detected in the MdCaM2-nLuc/MdCRF4-cLuc co-expressing region, but not in the negative control, consistent with MdCaM2–MdCRF4 protein interaction. Following CaCl₂ treatment, a stronger luminescence signal was observed in the MdCaM2-nLuc/MdCRF4-cLuc co-expressing region, in contrast, EGTA treatment inhibited this signal (Figure 2D). These results indicated that Ca²⁺ enhances the interaction between MdCaM2 and MdCRF4. Notably, the expression of MdCRF4 was not affected by CaCl₂ treatment (Supplemental Figure S3). However, immunoblot analysis revealed that the MdCRF4 protein abundance substantially decreased after CaCl₂ treatment (Figure 2E). This suggested that MdCRF4 is likely posttranslationally regulated.

Figure 2.

MdCaM2 sensing of Ca²⁺ and interaction with MdCRF4 and analysis of MdCRF4 protein expression. A, An Y2H assay showing that MdCaM2 interacts with MdCRF4. The coding regions of MdCaM2 and MdCRF4 were fused with DNA-BD (pGBKT7) and AD (pGADT7) vectors, respectively, as indicated, and co-transformed into the yeast strain Y2H gold. DDO, SD medium lacking Trp and Leu; QDO, SD medium lacking Trp, Leu, His, and Ade; X, QDO medium containing x-a-gal. The SV40 and P53 genes were used as the positive controls, and AD and BD empty vectors were used as the negative controls. The ability of yeast cells to grow on QDO and to turn blue in the presence of the chromogenic substrate x-a-gal was considered as a positive interaction. B, A pull-down assay confirming that MdCaM2 interacts with MdCRF4 in a Ca²⁺-mediated manner. Purified MdCaM2-GST was incubated with immobilized MdCRF4-His in the presence of 1-mM Ca²⁺ or 5-mM EGTA (a Ca²⁺ chelator), and eluted for immunoblot analysis with an anti-His antibody (upper panel) or an anti-GST antibody (lower panel). The band detected by the anti-GST antibody in the pull-down experiment indicates an interaction between MdCaM2 and MdCRF4 (C) A co-IP assay showing that Ca²⁺ treatment enhanced the interaction between MdCaM2 and MdCRF4. MdCaM2-GFP/MdCRF4-Myc and empty-GFP/MdCRF4-Myc were overexpressed in apple calli, respectively, and anti-GFP was used for immunoprecipitation analysis. Anti-GFP and anti-Myc antibodies were used in an immunoblot analysis. The band detected by the anti-Myc antibody in the precipitated protein sample indicates the interaction between MdCaM2 and MdCRF4. CaCl₂ treatment enhanced MdCRF4 protein abundance co-immunoprecipitated by MdCaM2-GFP. D, A firefly Luc complementation imaging assay showing that CaCl₂ treatment enhanced the interaction between MdCaM2 and MdCRF4 in N. benthamiana leaves. Agrobacterium tumefaciens strain EHA105 harboring different constructs was infiltrated into N. benthamiana leaves. CaCl₂, N. benthamiana leaves treated with CaCl₂. Luciferase activities were imaged in these regions 3 days after infiltration. E, Immunoblot analysis of MdCRF4 protein levels during fruit storage using an anti-MdCRF4 antibody. Coomassie brilliant blue (CBB) staining of protein extracts was used as a control to monitor equal sample loading.

A helical wheel projection analysis indicated that 14 amino acid residues, from M1 to 14K, in MdCRF4 may serve as the CaMBD. These domains are responsible for interaction with CaM proteins and typically comprised a stretch of 12–30 contiguous amino acids with a positive net charge and have a propensity to form an amphiphilic α-helix (Perruc et al., 2004; Du and Poovaiah, 2005; Du et al., 2009). This region showed a high degree of sequence conservation with a previously identified CaMBD in JAV1 (also known as CaM-interaction protein) (Figure 3A). To experimentally verify this putative CaMBD, we generated three MdCRF4 variants to alter the putative CaMBD secondary structure (Figure 3B): MdCRF4^H (amino acid substitutions in the hydrophobic surface), MdCRF4^N (amino acid substitutions in the net surface charges), and MdCRF4^△CBD (deletion of the CaMBD). The results of a Y2H assay indicated that MdCRF4^H, MdCRF4^N, and MdCRF4^△CBD all lost their ability to interact with MdCaM2 (Figure 3C), and this conclusion was further supported by an in vitro pull-down assay with MdCRF4^H-, MdCRF4^N- or MdCRF4^△CBD-His, and MdCaM2-GST fusion proteins (Figure 3D). Finally, in an LCI assay, the luminescence signal of the interaction with MdCaM2 was markedly decreased for MdCRF4^N and MdCRF4^△CBD and abolished for MdCRF4^H-His (Figure 3E). These results consistently indicated that the putative MdCRF4 CaMBD indeed mediates an interaction between MdCRF4 and MdCaM2.

Figure 3.

The MdCRF4 CaMBD responds to its interaction with MdCaM2. A, Helical wheel projection indicating that the putative CaM-binding domain (CaMBD) in MdCRF4 contains a positive net charge and has a propensity to form an amphiphilic α-helix region, which is highly similar to a previously identified CaMBD in AtJAV1. The polar/basic, polar/acid, polar/uncharged and nonpolar amino acids in the MdCRF4 and AtJAV1 CaMBDs are highlighted in green, blue, pink, and cyan, respectively. B, Schematic map of a putative CaMBD covering a region from M1 to K14 in MdCRF4. MdCRF4^H, and MdCRF4^N are highlighted in red; MdCRF4^△CBD, deletion of CaMBD. C, Y2H assay showing that the MdCRF4 CBD is essential for MdCaM2–MdCRF4 interaction. MdCaM2 CDS was fused with DNA-BD(pGBKT7) vector. MdCRF4 and mutated MdCRF4 (MdCRF4^H, MdCRF4^N, and MdCRF4^△CBD) CDS was fused with AD (pGADT7) vector. Vectors with AD and BD were co-transformed into the yeast strain Y2H gold. The details are the same as in Figure 2A. D, Pull-down assay showing that the MdCRF4 CaMBD is essential for the MdCaM2–MdCRF4 interaction. Purified MdCaM2-GST was incubated with immobilized MdCRF4 or mutated MdCRF4 (MdCRF4^H, MdCRF4^N, and MdCRF4^△CBD) in the presence of 1 mM Ca²⁺, and eluted for immunoblot analysis with an anti-His (upper) antibody or an anti-GST antibody (lower). The band detected by the anti-GST antibody in the pull-down experiment indicates an interaction between MdCaM2 and MdCRF4. E, A firefly Luc complementation imaging assay showing that the MdCRF4 CaMBD is essential for MdCaM2–MdCRF4 interaction in N. benthamiana leaves. Agrobacterium tumefaciens strain EHA105 harboring different constructs was infiltrated into N. benthamiana leaves. Luciferase activities were imaged in these regions 3 days after infiltration.

MdCRF4 enhances MdACS1 and MdERF3 transcription by binding to their promoters

Genes in the AP2/ERF family of transcription factors, of which MdCRF4 is a member, are known to be involved in the regulation of apple fruit ripening (Li et al., 2016). We, therefore, hypothesized that MdCRF4 may directly bind to the promoter of ethylene biosynthesis- and ethylene signal transduction-related genes to regulate their transcription, thereby connecting Ca²⁺ signaling and ethylene production. We first silenced MdCRF4 in apple calli using Agrobacterium-mediated transformation (Supplemental Figure S4). Eight transgenic lines were generated, of which lines #1, #3, and #7 showed substantially suppressed MdCRF4 expression (Supplemental Figure S4A). We then evaluated MdACS1, MdACO1, and MdERF3 expression, as well as ethylene production. The expression of all three genes was markedly lower in MdCRF4-suppressed calli than in control callus, and ethylene production showed the same pattern (Supplemental Figure S4, B–E). These results indicated that MdCRF4 is a positive regulator of MdACS1, MdACO1, and MdERF3 transcription and ethylene biosynthesis.

To investigate whether MdCRF4 binds to the promoters of MdACS1, MdACO1, and MdERF3, we performed a yeast one-hybrid (Y1H) assay and observed binding to the MdACS1 and MdERF3 promoters, but not to the MdACO1 promoter (Figure 4A; Supplemental Figure S5A). In subsequent Y1H to various regions of the promoters, MdCRF4 bound to the fragment containing the dehydration-responsive element (DRE) motif in the MdACS1 and MdERF3 promoters (Figure 4A). Furthermore, when MdCRF4 was divided into three fragments MdCRF4-N (N-terminus, 1–88 aa), MdCRF4-M (ERF domain, 88–185 aa), and MdCRF4-C (C-terminus, 185–282 aa), MdCRF4-M containing the ERF domain was found to mediate the binding (Supplemental Figure S6). Next, to confirm that MdCRF4 interacts with the MdACS1 and MdERF3 promoters in planta, a chromatin immunoprecipitation (ChIP)-qPCR assay was carried out with an anti-Myc antibody in transgenic Pro35S:Myc-MdCRF4 and Pro35S:myc apple calli. The results showed that the MdACS1 S1 and MdERF3 S5 promoter regions, which contain DRE-motifs, were highly enriched in the immunoprecipitated chromatin compared to that from the empty vector-expressing callus (Figure 4B), suggesting that MdCRF4 directly binds to MdACS1 and MdERF3 promoters. The effects of MdCRF4 on MdACS1 and MdERF3 expression were further investigated with a β-glucuronidase (GUS) transactivation assay in N. benthamiana leaves. Co-transformation of Pro35S:MdCRF4 with either of the promoter regions of the two target genes fused to the GUS gene (ProMdACS1::GUS or ProMdERF3:GUS) resulted in a significantly increased GUS signal, indicating that MdCRF4 promotes corresponding transcriptional activities (Figure 4C). This result was confirmed using a dual-luciferase reporter assay (Figure 4D). Taken together, the data indicated that MdCRF4 binds to and activates the MdACS1 and MdERF3 promoters.

Figure 4.

MdCRF4 positively regulates MdACS1 and MdERF3 transcription. A, Y1H results showing that MdCRF4 directly binds to the MdACS1 and MdERF3 promoters. The basal concentration of aureobasidin A used was 200 ng mL⁻¹ for MdCRF4 and the MdACS1 promoter, and 150 ng mL⁻¹ for the MdERF3 promoter. The empty vector and the MdACS1 and MdERF3 promoters were used as negative controls. The rec-P53 and P53-promoter were used as the positive controls. B, ChIP-qPCR showing that MdCRF4 binds to the MdACS1 and MdERF3 promoters in vivo. The various cross-linked chromatin samples were extracted from MdCRF4-Myc overexpressing fruit callus and precipitated with an anti-Myc antibody. The eluted DNA was used to amplify the sequences by qPCR. Six different regions (S1–S6) were investigated. Fruit callus overexpressing the Myc sequence was used as a negative control. C, GUS activity analysis showing that MdCRF4 enhances the activity of the MdACS1 and MdERF3 promoters. The MdCRF4 effector vector, together with a reporter vector containing either the MdACS1 or MdERF3 promoter, or a mutated promoter (mProMdACS1 and mProMdERF3), were infiltrated into N. benthamiana leaves to assess the possible regulation of GUS activity. D, Dual-luciferase reporter assay showing that MdCRF4 enhances the activity of the MdACS1 and MdERF3 promoters. The MdCRF4 effector vector, together with a reporter vector containing the MdACS1 or MdERF3 promoters, were infiltrated into N. benthamiana leaves to assess the possible regulation of Luc activity. For the ChIP-qPCR assay, the enriched DNA fragments in each ChIP were used as one biological replicate for qPCR. A total of three biological replicates were analyzed. The values represent means ± se. Statistical significance was determined using a Student’s t test (**P < 0.01). n.s., no significant difference. For GUS activity analysis, one independent transformation experiment was used as one biological replicate. A total of three biological replicates were analyzed. The values represent means ± se. Statistical significance was determined using a Student’s t test (**P < 0.01). n.s., no significant difference.

CaM proteins sense Ca²⁺ to activate the phosphorylation of MdCRF4, leading to its degradation

Ca²⁺/CaM-mediated phosphorylation of target proteins is an important factor in Ca²⁺-mediated signal transduction (Hook and Means, 2001; Zhang and Lu, 2003). Since CaM proteins sense Ca²⁺ and interact with MdCRF4, we hypothesized that this results in MdCRF4 phosphorylation and signaling. Previous studies have shown that ethylene production by apple fruit can be inhibited by CaCl₂ treatment within a few hours (Lieberman and Wang, 1982; Ferguson, 1983). To study the change in the state of the MdCRF4 protein mediated by Ca²⁺, we treated disks excised from apple fruit (10 days after harvest) with 100-mM CaCl₂. This inhibited MdACS1 and MdERF3 expression and ethylene production (Supplemental Figure S7), consistent with the previous study (Ferguson, 1983). We also observed that LaCl₃ blocked the inhibitory effect of Ca²⁺ (Supplemental Figure S7).

Next, an in vivo phosphorylation assay was employed to determine whether MdCRF4 is phosphorylated in apple fruit and how it responds to Ca²⁺. Immunoblotting of protein that was immunoprecipitated by the anti-MdCRF4 antibody from the apple disks treated with CaCl₂, using an anti-phosphoSer/Thr antibody, showed that MdCRF4 was phosphorylated and that this phosphorylation was promoted by CaCl₂ treatment (Figure 5A), indicating that the CaCl₂ treatment promoted MdCRF4 phosphorylation. To investigate whether this Ca²⁺-induced MdCRF4 phosphorylation was dependent on MdCaM, we generated transgenic apple fruit callus expressing 35S:myc-tagged MdCRF4^N or MdCRF4^H for phos-tag mobility shift assays. Ca²⁺-induced MdCRF4 phosphorylation was almost abolished in CaCl₂-treated transgenic callus expressing 35Spro::Myc-MdCRF4^N or 35Spro::Myc-MdCRF4^H (Figure 5B), indicating that disruption of the Ca²⁺-mediated CaM–MdCRF4 interaction impedes Ca²⁺-induced MdCRF4 phosphorylation. In summary, CaM senses Ca²⁺ and interacts with MdCRF4, resulting in changes in the phosphorylation status of MdCRF4. Notably, the Ca²⁺/CaM-mediated MdCRF4 phosphorylation-dependent protein degradation was severely attenuated in the MdCRF4-deficient variants MdCRF4^H-His and MdCRF4^N-His, as determined using an in vitro cell-free degradation assay, with total protein extracted from apple disks treated with CaCl₂ (Figure 5C). These results suggest that Ca²⁺/CaM mediates MdCRF4 phosphorylation and the phosphorylated MdCRF4 undergoes 26S proteasome-mediated protein degradation.

Figure 5.

Ca²⁺ triggers calcium-mediated phosphorylation of MdCRF4. A, Ca²⁺ induces phosphorylation of MdCRF4 in vivo. The apple fruit disks (10 days after harvest) were treated with CaCl₂ for 0, 2, 4, 6, and 8 h for protein extraction and immunoprecipitation with anti-MdCRF4 and separated by SDS–PAGE for immunoblotting analysis using an anti-phosphoSer/Thr antibody. B, The CaMBD-mediated CaM2–MdCRF4 interaction is essential for Ca²⁺-induced phosphorylation of MdCRF4 in vivo. Equal amounts of MdCRF4-Myc protein, immunoprecipitated from transgenic apple callus (35Spro::Myc-MdCRF4, 35Spro::Myc-MdCRF4^N, or 35Spro::Myc-MdCRF4^H) after CaCl₂ treatment, were separated in phos-tag mobility shift assays with an anti-Myc antibody (top). The slowly migrating MdCRF4 band in the phos-tag gel represents the phosphorylated form of MdCRF4. As a loading control (lower) equal amounts of the immunoprecipitated MdCRF4 proteins were separated in a normal SDS–PAGE gel followed by immunoblot analysis. C, An in vitro cell-free degradation assay showing that MdCRF4 phosphorylation determines its degradation. The purified MdCRF4-His proteins (MdCRF4, MdCRF4^N, and MdCRF4^H) were added to total protein extracts from apple fruit disks treated with CaCl₂ for 2 h, incubated for the indicated time periods, and subjected to immunoblot analysis with an anti-His antibody. CBB staining of protein extracts was used as a monitor for equal sample loading.

MdXBAT31 interacts with and ubiquitinates phosphorylated MdCRF4 for proteasome degradation

Based on bioinformatic analysis (http://kinasephos.mbc.nctu.edu.tw) (Huang et al., 2005), MdCRF4 was predicted to contain two putative Ca²⁺/CaM-mediated phosphorylation sites: S204 and S259. To test this prediction, we generated MdCRF4 variants with alanine or aspartate mutations at S204 and S259 that represented nonphosphorylated (MdCRF4AA: MdCRF4^S204A/S259A) and continuously phosphorylated (MdCRF4DD: MdCRF4^S204D/S259D) forms of the protein. Next, we generated Pro35S:Myc-MdCRF4 and Pro35S:Myc-MdCRF4AA plasmids and transformed them individually into apple calli. Proteins were extracted from the transgenic calli treated with CaCl₂ and immunoprecipitated with an anti-Myc antibody, and a phos-tag mobility shift assay was performed. A slowly migrating band corresponding to phosphorylated MdCRF4 was observed in phos-tag gel and MdCRF4AA substantially decreased (Figure 6A). This result indicates that the phosphorylation sites in MdCRF4 regulated by Ca²⁺ are S204 and S259.

Figure 6.

MdCRF4 phosphorylation is essential for the MdXBAT31–MdCRF4 interaction. A, The phosphorylation status of MdCRF4 proteins (MdCRF4 and MdCRF4AA) in vivo. MdCRF4AA (S204A/S259A): MdCRF4 variants harboring alanine or aspartate mutations at S204 and S259 to form a nonphosphorylated MdCRF4 form. The proteins were immunoprecipitated with an anti-Myc antibody from transgenic apple callus (Pro35S::Myc-MdCRF4 and Pro35S::Myc-MdCRF4AA) after CaCl₂ treatment and were separated in phos-tag mobility shift assays. B, Y2H assay showing that MdXBAT31 interacts with phosphorylated MdCRF4. The coding regions of MdXBAT31 and MdCRF4DD were fused with DNA-BD (pGBKT7) and AD (pGADT7) vectors, respectively, as indicated, and co-transformed into the yeast strain Y2H gold. The details are the same as in Figure 2A. C, Pull-down assay showing that MdXBAT31 interacts with phosphorylated MdCRF4. Purified MdCaM2-GST was incubated with immobilized MdCRF4DD in reaction solution buffer, and eluted for immunoblot analysis with an anti-GST (upper) antibody or an anti-His antibody (lower). The band detected by the anti-GST antibody in the pull-down experiment indicates an interaction between MdXBAT31 and MdCRF4DD. D, A co-IP assay showing that Ca²⁺ treatment enhanced the in vivo interaction between MdXBAT31 and MdCRF4. MdXBAT31 fused to a GFP tag (MdXBAT31-GFP) and MdCRF4 fused to a Myc tag (MdCRF4-Myc, MdCRF4AA-Myc, and MdCRF4DD-Myc) were overexpressed in apple fruit callus, and an anti-GFP antibody was used for immunoprecipitation analysis. Anti-Myc and anti-GFP antibodies were used in an immunoblot analysis. The band detected by the anti-Myc antibody in the precipitated protein sample indicates the interaction between MdXBAT31 and MdCRF4 (or MdCRF4DD). Intensity, Protein quantification in gel was achieved by using ImageJ to measure the intensity of bands. E, A firefly Luc complementation imaging assay showing that phosphorylated MdCRF4 interacts with MdXBAT31 in N. benthamiana leaves. Agrobacterium tumefaciens strain EHA105 harboring different constructs was infiltrated into N. benthamiana leaves. Luciferase activities were imaged in these regions 3 days after infiltration.

To characterize the mechanistic basis of the degradation of phosphorylated MdCRF4, we performed a Y2H screen for a yeast cDNA library derived from apple fruit cDNA, using MdCRF4DD as a bait molecule. Among the 23 positive clones, a RING finger domain-containing protein, named MdXBAT31, showed the strongest interaction with MdCRF4DD, and was used in a direct Y2H assay. As shown in Figure 6B, the MdCRF4DD and MdXBAT31 proteins physically interacted with each other in the yeast cells. Moreover, an in vitro pull-down assay also showed that MdCRF4DD interacted with MdXBAT31 (Figure 6C), while MdCRF4AA did not interact with MdXBAT31 (Supplemental Figure S8). To investigate how the phosphorylation status of MdCRF4 affects its interaction with MdXBAT31, and how this interaction responds to Ca²⁺ in planta, an in vivo co-IP assay was performed with transgenic callus (MdXBAT31-GFP/MdCRF4-Myc, MdXBAT31-GFP/MdCRF4AA-Myc, or MdXBAT31-GFP/MdCRF4DD-Myc) with or without CaCl₂ treatment. We observed that the interaction between MdCRF4DD and MdXBA31 was strongest, even without CaCl₂ treatment (Figure 6D, Columns 2 and 6), while MdCRF4AA did not interact with MdXBAT31 (Figure 6D, Columns 3 and 7), and the interaction between MdCRF4 and MdXBAT31 was promoted by CaCl₂ treatment (Figure 6D, Columns 1 and 5). Finally, in an LCI assay, the luminescence signal of the interaction with MdXBAT31 was markedly increased for MdCRF4DD and abolished for MdCRF4AA (Figure 6E). Together, these results indicate that MdCRF4 phosphorylation is a prerequisite for its interaction with MdXBAT31.

Given that CaCl₂ treatment promotes MdCRF4 protein degradation during apple fruit ripening (Figure 2E) and the E3 ubiquitin ligase MdXBAT31 interacts with MdCRF4DD, we next investigated whether MdXBAT31 ubiquitinates MdCRF4DD. An in vitro ubiquitination assay was carried out, and in the presence of Ub, E1, E2, and E3 (MdXBAT31-GST), polyubiquitinated MdCRF4DD-His was detected by anti-Ub immunoblot analysis, while lack of MdCRF4DD substrate in the reaction mixture abolished the ubiquitinated bands (Figure 7, A and B). Similarly, polyubiquitinated MdCRF4AA-His protein was not detected by anti-Ub or anti-His immunoblot analysis. In addition, an in vivo ubiquitination detection assay was carried out using single-transgenic Myc-MdCRF4 overexpressing callus (35S:Myc-MdCRF4) and two double-transgenic calli overexpressing Myc-MdCRF4 and silencing MdXBAT31 (35S:Myc-MdCRF4/MdXBAT31-AS). These calli were treated with CaCl₂, and protein gel blot analysis showed that MdXBAT31-AS inhibited the ubiquitination of MdCRF4 (Figure 7C). MdXBAT31 also affected the stability of phosphorylated MdCRF4 (MdCRF4DD) in an in vitro cell-free degradation analysis, such that when proteins extracted from wild-type and MdXBAT31-AS (MdXBAT31-silenced) transgenic apple calli were incubated with the purified MdCRF4DD-His fusion protein, MdXBAT31 silencing slowed the degradation of MdCRF4DD (Figure 7D) which could be substantially blocked by MG132 treatment. These results demonstrated that the degradation of phosphorylated MdCRF4 by the 26S proteasome is mediated by MdXBAT31. To further verify whether Ca²⁺/CaM-mediated phosphorylation of MdCRF4 is important for MdCRF4 ubiquitination by MdXBAT31, we used 35S:Myc-MdCRF4, 35S:Myc-MdCRF4AA and 35S:Myc-MdCRF4^△CBD transgenic apple calli treated with CaCl₂ in an in vivo ubiquitination detection assay, and observed that polyubiquitinated MdCRF4-Myc was detected in an anti-Ub immunoblot analysis, while MdCRF4AA and MdCRF4^△CBD were not (Figure 7E). This result suggested that Ca²⁺/CaM-mediated MdCRF4 phosphorylation is a prerequisite for MdCRF4 ubiquitination by MdXBAT31.

Figure 7.

MdXBAT31-mediated phosphorylation of MdCRF4 causes its degradation via the ubiquitin-proteasome system. A and B, In vitro ubiquitination of MdCRF4DD by MdXBAT31. Recombinant MdCRF4DD-His and MdCRF4AA-His proteins were analyzed for potential MdXBAT31 E3 ubiquitin ligase activity in the presence of ATP, ubiquitin, E1, E2, MdXBAT31. MdCRF4AA-His and MdCRF4DD-His were detected by immunoblot analysis with anti-ubiquitin (A) or anti-His (B) antibodies. C, Ubiquitination of MdCRF4 in Pro35S::MdCRF4-Myc and MdXBAT31-AS + Pro35S::MdCRF4-Myc expressing callus pretreated with 50-µM MG132 and 100-µM CaCl₂. MdCRF4 protein was immunoprecipitated with the anti-Myc antibody. Anti-ubiquitin antibody was used to detect ubiquitinated MdCRF4-Myc. D, In vitro cell-free degradation assay showing that the MdCRF4DD degradation rate was slower in a MdXBAT31-AS transgenic apple callus protein extract than in empty vector transgenic apple callus protein. The purified MdCRF4DD-His protein was added to total protein extracts from MdXBAT31-AS transgenic apple callus, incubated for the indicated times, and subjected to immunoblot analysis with an anti-His antibody. CBB staining of protein extracts was used to monitor equal sample loading. E, Ca²⁺/CaM-mediated phosphorylation of MdCRF4 is a prerequisite for ubiquitination and degradation of MdCRF4 by MdXBAT31. Ubiquitination of the MdCRF4 proteins in Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4AA, and Pro35S::Myc-MdCRF4^△CBD transgenic apple callus pretreated with 50-µM MG132 and 100-µM CaCl₂.

Ca²⁺/CaM-mediated MdCRF4 phosphorylation leads to ubiquitination and degradation by MdXBAT31, suppressing MdACS1 and MdERF3 transcription

Our results indicated that MdCRF4 can directly bind to and activate the MdACS1 and MdERF3 promoters (Figure 4). In addition, phosphorylation of MdCRF4 mediated by Ca²⁺/CaM can lead to its ubiquitination and degradation by MdXBAT31 (Figures 5–7). We, therefore, hypothesized that Ca²⁺/CaM mediates MdCRF4 phosphorylation, which in turn causes changes in the degree of MdCRF4 binding to the MdACS1 and MdERF3 promoters, thereby regulating their transcription. To test this hypothesis, a ChIP-qPCR assay was carried out with an anti-Myc antibody and extracts from apple calli expressing the Pro35S:Myc-MdCRF4, Pro35S:Myc-MdCRF4^N, Pro35S:Myc-MdCRF4^H, or Pro35S:Myc transgenes, with or without CaCl₂ treatment. We observed that Myc-MdCRF4^N and Myc-MdCRF4^H caused a greater enrichment of the MdACS1 and MdERF3 promoter DNA than did Myc-MdCRF4. Moreover, Myc-MdCRF4-enriched MdACS1 and MdERF3 promoter DNA abundance were inhibited by CaCl₂ treatment, while Myc-MdCRF4^H- and Myc-MdCRF4^N-enriched MdACS1 and MdERF3 promoter DNA abundance were not affected by the CaCl₂ treatment (Figure 8, A and B). This suggested that Ca²⁺ regulates MdACS1 and MdERF3 transcript levels through the action of MdCRF4 and that it is dependent on CaM to transmit Ca²⁺ signal. Furthermore, a ChIP-qPCR assay was also carried out with an anti-Myc antibody using transgenic Pro35S:Myc-MdCRF4, Pro35S:Myc-MdCRF4AA, and Pro35S:Myc apple calli, with or without CaCl₂ treatment. Myc-MdCRF4AA caused a greater enrichment of MdACS1 and MdERF3 promoter DNA than did Myc-MdCRF4, and Myc-MdCRF4-enriched MdACS1 and MdERF3 promoter DNA abundance were inhibited by CaCl₂ treatment, while Myc-MdCRF4AA-enriched MdACS1 and MdERF3 promoter DNA abundance were not affected by the CaCl₂ treatment (Figure 8, C and D). This suggested that Ca²⁺/CaM-mediated phosphorylation of MdCRF4 weakens its binding to the MdACS1 and MdERF3 promoters.

Figure 8.

Ca²⁺/CaM- and MdXBAT31-mediated phosphorylation and degradation of MdCRF4, suppress MdACS1, and MdERF3 transcription. A–B, ChIP-qPCR showing that CaCl₂ treatment decreases the binding of MdCRF4 to the MdACS1 and MdERF3 promoters. This treatment did not affect MdCRF4^N and MdCRF4^H binding to the promoter. Cross-linked chromatin samples were extracted from MdCRF4-Myc-, MdCRF4^N-Myc-, and MdCRF4^H-Myc-overexpressing fruit callus treated with or without CaCl₂ and precipitated with an anti-Myc antibody. Eluted DNA was used to amplify the sequences neighboring the DRE motif by qPCR. ProMdACS1-S1 and ProMdERF3-S5 refers to the MdACS1 or MdERF3 promoter region from Figure 4B, respectively. C and D, ChIP-qPCR showing that CaCl₂ treatment decreases the binding of MdCRF4 to the MdACS1 and MdERF3 promoters. This treatment did not affect the binding of MdCRF4AA to the promoter. Cross-linked chromatin samples were extracted from MdCRF4-Myc- and MdCRF4AA-Myc-overexpressing fruit callus treated with or without CaCl₂ and precipitated with an anti-Myc antibody. Eluted DNA was used to amplify the sequences neighboring the DRE motif by qPCR. ProMdACS1-S1 and ProMdERF3-S5 refer to the MdACS1 or MdERF3 promoter region from Figure 4B, respectively. E and F, Ca²⁺ promotes MdXBAT31 destabilization of MdCRF4, thereby suppressing the transcriptional repression activity of MdCRF4 on the MdACS1 and MdERF3 promoters. The reporter and effector vectors used in the transient assay, as indicated at the top, were co-transformed into N. benthamiana leaves. The data indicate the relative GUS activity. G, Apple fruit with transiently silenced MdXBAT31 (MdXBAT31-AS) and an empty vector (control) during storage. MdXBAT31 was silenced in apple fruit by A. tumefaciens-mediated transient transformation. MdXBAT31-AS fruit were harvested 2 weeks after infiltration and stored at room temperature for 20 days. DAI, days after infiltration. DAH, days after harvest. H–J, MdXBAT31, MdACS1, and MdERF3 expression was investigated by RT–qPCR. K, Ethylene production was measured during MdXBAT31-AS apple fruit ripening. For ChIP-qPCR, three biological replicates were analyzed as described in Figure 4. For GUS activity analysis, three independent transformation experiments were performed as described in Figure 4. For RT–qPCR, three biological replicates were used as described in Figure 1. The values represent means ± se. Statistical significance was determined using a Student’s t test (**P < 0.01).

We next carried out a GUS transactivation assay in N. benthamiana leaves that were co-transformed with ProMdACS1::GUS and Pro35S:MdCRF4 and Pro35S:MdXBAT31, or Pro35S:MdCRF4DD and Pro35S:MdXBAT31, or Pro35S:MdCRF4AA and Pro35S:MdXBAT31. The transcriptional activation of the MdACS1 promoter by MdCRF4DD was significantly lower than the activation by MdCRF4, and CaCl₂ treatment inhibited the transcriptional activation of the MdACS1 promoter by MdCRF4 (Figure 8E). The transcriptional activation by MdCRF4AA of the MdACS1 promoter was not affected by the CaCl₂ treatment. Similar results were found with the MdERF3 promoter (Figure 8F). Taken together, these results suggested that Ca²⁺/CaM-mediated phosphorylation of MdCRF4 weakens its ability to activate MdACS1 and MdERF3 transcription.

In addition, to characterize MdXBAT31 function during ripening, we transiently silenced its expression in fruit. Transgenic fruit were harvested 14 days after infiltration with Agrobacterium tumefaciens harboring pTRV-MdXBAT31 and we found that MdXBAT31-AS (MdXBAT31-silenced) fruit ripened more rapidly, compared with control fruit (infiltration empty vector, pTRV) (Figure 8G). When MdXBAT31 transcription in MdXBAT31-AS fruit was tested, we found that expression was significantly lower in MdXBAT31-AS fruit, compared with control fruit (Figure 8H). Since the expression of MdACS1 and MdERF3, as well as ethylene production, was significantly higher in MdXBAT31-AS apple fruit during ripening (Figure 8, I–K), we concluded that MdXBAT31-mediated ubiquitination and degradation of MdCRF4 contributes to ripening-related ethylene biosynthesis.

The phosphorylation status of MdCRF4 affects apple fruit ripening

In order to further characterize Ca²⁺/CaM-mediated phosphorylation of MdCRF4 in ripening apple fruit, three recombinant plasmids (Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4AA, and Pro35S::Myc-MdCRF4DD) were transiently overexpressed in Golden Delicious (GD) apple fruit (Figure 9A), with an empty vector used as a control. Transgenic fruit were harvested 2 weeks after A. tumefaciens infiltration, and MdCRF4AA overexpression (MdCRF4AA-OE) was found to have a stronger effect on promoting fruit ripening, compared with MdCRF4 overexpression (MdCRF4-OE) and MdCRF4DD overexpression (MdCRF4DD-OE) (Figure 9A). The expression of MdCRF4 in ripening fruit was also significantly enhanced in MdCRF4AA-OE, MdCRF4-OE, and MdCRF4DD-OE fruit (Figure 9B). When we measured transcript levels of MdACS1 and MdERF3 in the OE-fruit, we observed higher expression of MdACS1 and MdERF3 in the MdCRF4AA-OE fruit than in MdCRF4-OE and MdCRF4DD-OE fruit, while MdCRF4DD overexpression did not change MdACS1 and MdERF3 expression (Figure 9, C and D). Ethylene production showed the same pattern as MdACS1 and MdERF3 transcription abundance (Figure 9E). These results indicate that the phosphorylation state of MdCRF4 affects fruit ripening.

Figure 9.

Effect of MdCRF4 phosphorylation state on ethylene biosynthesis and apple fruit ripening. A, Apple fruit transiently overexpressing various types of MdCRF4 (MdCRF4, MdCRF4DD, or MdCRF4AA) and an empty vector control during storage. Transient overexpression in apple fruit by A. tumefaciens-mediated transient transformation. Overexpressing fruit (MdCRF4-OE, MdCRF4DD-OE, or MdCRF4AA-OE) were harvested 14 days after infiltration and stored at room temperature for 20 days. B–D, MdCRF4, MdACS1, and MdERF3 expression was investigated by RT–qPCR. E, Ethylene production was measured in ripening apple fruit overexpressing the various MdCRF4 forms. For RT–qPCR, three biological replicates were used as described in Figure 1. The values represent mean ± se. Statistical significance was determined using a Student’s t test (**P < 0.01).

Discussion

Previous studies have shown that exogenous Ca²⁺ is involved in regulating ethylene production in several fruit species. For example, ethylene production in apple and tomato fruit can be inhibited by CaCl₂ treatment (Chien Yi et al., 1993; Njoroge et al., 1998). However, the molecular mechanism by which exogenous Ca²⁺ inhibits ethylene production during fruit ripening has been unclear. In this study, we dissected the regulatory network involving Ca²⁺/CaM-MdCRF4–MdXBAT31 module and its association with the ethylene biosynthesis genes MdACS1 and MdERF3 in Ca²⁺-suppressed ethylene production. By exogenously applying CaCl₂, we observed that Ca²⁺-activated CaM2 interacts with MdCRF4 (Figure 2) and mediates the phosphorylation of MdCRF4 (Figure 5). Phosphorylation of MdCRF4 leads to its interaction with MdXBAT31 (Figure 6), which causes MdCRF4, a positive regulator of MdACS1 and MdERF3 transcription (Figure 4), to be ubiquitinated and degraded (Figure 7), thereby inhibiting the transcription of MdACS1 and MdERF3 (Figure 8) and ethylene biosynthesis in apple fruit.

Ca²⁺ serves as a secondary messenger and plays a critical role in the plant life cycle. It is known to participate in auxin, gibberellin, abscisic acid, and ethylene signaling to regulate fruit set, cell division, and fruit ripening by regulating physiological and biochemical processes (Aghdam et al., 2012; Hocking et al., 2016). Apple fruit Ca²⁺ content is negatively associated with fruit ethylene production during storage and exogenous calcium treatment can reduce ethylene production in apple fruit during ripening (Cheverry et al., 1988). These observations are consistent with the results in this study (Figure 1). Previous studies have described the involvement of Ca²⁺ in regulating fruit ripening by influencing ethylene biosynthesis and ethylene signal transduction (Aghdam et al., 2012). For example, Ca²⁺ treatment inhibited ethylene production by reducing ACC content, ACO, and ethylene-forming enzyme activity (Cheverry et al., 1988). In addition, in A. thaliana, SR1 encodes a Ca²⁺/CaM-binding transcriptional activator, which regulates ethylene-induced senescence by directly binding to the EIN3 promoter (Nie et al., 2012). However, these studies only investigated changes in enzyme activity and transcriptional regulatory mechanisms in Ca²⁺-regulated ethylene production. Here, we found that Ca²⁺/CaM-mediated phosphorylation of MdCRF4 initiated its ubiquitination, which integrated phosphorylation and ubiquitination of transcription factors to regulate the biosynthesis of ethylene in apple fruit. In this pathway, the posttranslational modification of the Ca²⁺/CaM-binding transcriptional activator MdCRF4 is involved in the regulation of ethylene biosynthesis by Ca²⁺.

CaM is a well-documented Ca²⁺ sensor and transmits Ca²⁺ signals to their downstream target proteins to regulate various plant processes (Yang and Poovaiah, 2003; DeFalco et al., 2009). In green ripe tomato fruit, transient overexpression of SlCaM-2 delays fruit ripening, while transient silencing of SlCaM-2 promotes fruit ripening (Yang et al., 2014). This indicates that CaM is involved in the regulation of fruit ripening and may be an important component of Ca²⁺ signal transduction. Based on this, the CaM gene with the highest expression in apple fruit, MdCaM2, was used as a bait molecule to screen an apple fruit cDNA library. Interestingly, a CaM-binding protein, MdCRF4 was identified, leading us to the hypothesis that CaM2–MdCRF4 represents a link between Ca²⁺ signaling and ethylene biosynthesis. Previous studies have shown that the ability of CaM proteins to regulate target protein activity resides in the hydrophobic regions that are exposed when Ca²⁺ binds to CaM and cause a conformational change, thereby enabling the interaction with the CaMBD in the target protein by electrostatic interactions (Hoeflich and Ikura, 2002). Here, we found that the MdCRF4 CaMBD comprised a stretch of 14 contiguous amino acids (from M1 to 14K) that contain a positive net charge and can form an amphiphilic α-helix (Figure 3). When the MdCRF4 CaMBD amino acids are mutated to disrupt the charge distribution in the secondary structure, MdCRF4 greatly loses its interaction capability with MdCaM2 (Figure 3). This indicates that the interaction between MdCRF4 and CaM is an electrical interaction made possible by the amphiphilic α-helix. In addition, we observed that MdCRF4 belonged to the AP2/ERF transcription factor family and that its amino acid sequence also contained a cytokinin response domain (Supplemental Figure S2), so it was named cytokinin response factor 4 (MdCRF4). We therefore suggest that it may also play a role in the cross-linking of cytokinin and ethylene signaling pathways.

Our results showed that Ca²⁺ promotes the phosphorylation of MdCRF4 in apple fruit (Figure 5). In addition, the phosphorylation state of MdCRF4 was not detected when it was mutated to MdCRF4^N and MdCRF4^H (Figure 5), as it could no longer interact with MdCaM2. This indicates that the phosphorylation of MdCRF4 is dependent on Ca²⁺/CaM in response to Ca²⁺. Furthermore, we found that Ca²⁺ promoted the degradation of MdCRF4 via the 26S proteasome pathway and that this degradation did not occur with the mutated proteins MdCRF4^H and MdCRF4^N (Figure 5). This indicated that MdCRF4 degradation was also dependent on Ca²⁺/CaM-mediated MdCRF4 phosphorylation. Next, we showed that MdXBAT31 could only ubiquitinate and degrade MdCRF4 when the latter was phosphorylated (Figure 7). Based on these results, we concluded that Ca²⁺/CaM-mediated phosphorylation of MdCRF4 is a prerequisite for ubiquitination and degradation of MdCRF4 by the E3 ubiquitin ligase, MdXBAT31. We also found that overexpression of MdCRF4AA significantly promoted fruit ethylene production and fruit ripening, while overexpression of MdCRF4DD had almost no effect (Figure 9), because the nonphosphorylated MdCRF4 (MdCRF4AA) maintains a stable state, while phosphorylated MdCRF4 (MdCRF4DD) has low stability. This results in increased expression of MdACS1 and MdERF3, which in turn promotes the production of ethylene and fruit ripening. Taken together, these results show that Ca²⁺/CaM- and MdXBAT31-mediated phosphorylation and ubiquitination of MdCRF4 weakens its stability, resulting in reduced MdACS1 and MdERF3 expression and lower ethylene production in apple fruit. A recent study showed that injury triggers Ca²⁺ to activate Ca²⁺/CaM-mediated phosphorylation of JAV1, leading to the degradation of the JAV1–JAZ8–WRKY51 complex to activate JA biosynthesis (Yan et al., 2018). We conclude that Ca²⁺/CaM-mediated phosphorylation of target proteins to affect their stability is a common element in Ca²⁺ signal transduction, which also suggests that Ca²⁺ can regulate the activity of different phytohormones.

Previous studies have shown that CaM can function as a regulator through interaction with various transcription factors, including members of the NAC and WRKY families to regulate plant growth and development (Park et al., 2005), although no further details or in-depth molecular mechanisms have been elucidated. Here, the role of the Ca²⁺/CaM–MdCRF4 module in regulating ethylene biosynthesis, via the phosphorylation and degradation of transcription factors, has been described, which not only highlights the role of CaM in regulating ethylene biosynthesis, but also highlights an important aspect of fruit ripening. In addition to CaM, calmodulin-like protein, CBL, and CDPK are also Ca²⁺-binding proteins that can transmit Ca²⁺ signals. There is growing evidence that CDPK may be involved in the regulation of fruit ripening. For example, one group of ACS proteins contains the conserved Ser residue, which is a phosphorylation site for CDPK, indicating that CDPK contributes to Ca²⁺-regulated ethylene biosynthesis (Xu and Zhang, 2014). CBLs are involved in all kinds of stress responses, including salt, drought, wound, and cold stress in plant (Luan, 2009). Up until now, few studies have focused on the relationship between CBLs and ethylene biosynthesis. In this study, we did not explore the molecular mechanism of CDPK and CBL regulation of ethylene biosynthesis in detail, which would require further study.

To summarize, we have identified MdCRF4 as a direct interactor with the MdACS1 promoter and a positive regulator of its transcription to promote ethylene synthesis in apple fruits at normal basal Ca²⁺ levels. Apple fruit treated with CaCl₂ showed increased Ca²⁺/CaM-mediated phosphorylation of MdCRF4, which resulted in ubiquitination and degradation of MdCRF4 by the E3 ligase MdXBAT31, thereby weakening MdACS1 transcription and ultimately inhibiting ethylene biosynthesis during ripening. Our data provide insights into the molecular basis by which Ca²⁺ suppresses fruit ripening to artificially extend fruit shelf life (Figure 10).

Figure 10.

A model showing the molecular mechanism of Ca²⁺ suppression of ethylene biosynthesis in apple fruit. MdCRF4 directly binds to the MdACS1 promoter and positively regulates its transcription to promote ethylene synthesis in apple fruits at normal base level Ca²⁺. CaCl₂ treatment can promote Ca²⁺/CaM-mediated phosphorylation of MdCRF4, which result in ubiquitination and degradation of MdCRF4 by MdXBAT31 E3 ligase, thereby weakening MdACS1 transcription and ultimately inhibiting the ethylene biosynthesis during fruit ripening at high Ca²⁺ levels.

Materials and methods

Plant materials and treatments

Apple (M. domestica “Golden Delicious”; GD) fruit were collected from the Liaoning Pomology Institute orchard (Xiongyue, Chian). The commercial harvest date of GD fruit in the region is 140 days after full bloom when the content of total soluble solids reached 12% (Li et al., 2015), and apples harvested at this time were stored. Fruit were divided into three different groups and kept at room temperature (24°C) for 20 days with sampling every 5 days. The first group was the control without any treatment; the second group was treated by spraying with 100-mM CaCl₂ every 5 days; the third group was treated with a solution of 100-mM CaCl₂ and 1-mM LaCl₃ every 5 days. Disks (1-mm thick and 10-mm diameter) were taken from the fruit cortex and stored for 10 days after harvest at room temperature. CaCl₂ treatment of disks was performed as previously described (Lieberman and Wang, 1982). Approximately 10-g disks were incubated in 600-mM sorbitol and 10-mM 2-(N-morpholino) ethanesulfonic acid (MES) (Sangon Biotech, China) buffer (pH 6.0) to stabilize the pH, and then transferred to 100-mL conical flasks. Fifty microliters of 600-mM sorbitol, 10-mM MES (pH 6.0), and CaCl₂ at final concentrations of 10, 100, or 250 mM, was added to each flask. The flasks were placed at 24°C with shaking for 0, 2, 4, 6, and 8 h. For disks treated with both CaCl₂ and LaCl₃, 1-mM LaCl₃ was added to the above solution. After this, the fruit and disks were cut into smaller pieces, frozen in liquid nitrogen, and stored at −80°C. Apple (“Orin”) calli and N. benthamiana plants were grown in our laboratory as previously described (Li et al., 2016). For CaCl₂ treatments, callus were immersed in a liquid Murashige and Skoog (MS) medium (Phyto Technologies Laboratories, USA) containing 100-mM CaCl₂ and shaken at 24°C for the corresponding time. For MG132 treatment, the apple callus was immersed in a liquid MS medium containing 50-μM MG132 (Sangon Biotech, Shanghai, China) and shaken at 24°C for 12 h.

RNA extraction, cDNA synthesis, RT–qPCR, and ethylene production rate

Total RNA isolation, cDNA synthesis, and reverse transcription-quantitative PCR (RT–qPCR) were carried out as previously described (Li et al., 2016). RT–qPCR was performed using an iCycler iQ5 system (Bio-Rad, Hercules, CA, USA) with three biological replicates. The apple fruit and callus ethylene production was measured according to previous methods (Li et al., 2016). All primers used in this study are listed in Supplemental Table S2.

Protein expression and purification

The coding sequence (CDS) of MdCaM2 and MdXBAT31 were separately cloned downstream from the GST encoding sequence in the pGEX4T-1 vector to generate GST fusion proteins. The MdCRF4, MdCRF4^H, MdCRF4^N, MdCRF4^△CBD, MdCRF4AA, or MdCRF4DD CDS were cloned into the pEASY-E1 vector (Transgen Biotech) to express His fusion proteins. The transformation of the resulting plasmids into Escherichia coli BL21 cells, and the induction of the target protein were performed as previously described (Li et al., 2016). The final isopropyl β-d-1-thiogalactopyranoside concentration for inducing protein expression was 0.5 mM, and the purification of GST- or His-target proteins was performed as previously described (Li et al., 2016).

Y2H assay

A cDNA library was constructed with mRNA from apple fruit, which was harvested at commercial maturity and stored for 0, 5, 10, 15, and 20 days, using the Make Your Own Mate & Plate Library System (Clontech, Mountain View, CA, USA). The MdCaM2 or MdCRF4DD CDS was introduced into the pGBKT7 vector. The recombinant plasmid was used as bait to screen the cDNA library using the Matchmaker Gold Y2H Library Screening System Kit (Clontech, USA). For protein interaction assays, the MdCaM2 CDS was cloned into the pGBKT7 vector containing a BD. The MdCRF4 or MdCRF4DD CDS were separately cloned into the pGADT7 vector containing an activation domain (AD). AD and BD plasmids were co-transformed into the Y2H-gold yeast strain. The detection of protein interactions between two proteins was conducted using the Matchmaker Gold Y2H Library Screening System Kit (Clontech, USA).

Pull-down assay

Pull-down assays were carried out as previously described (Yan et al., 2018). Briefly, 30-μg bait protein (MdCRF4-His, MdCRF4^H, MdCRF4^N, MdCRF4^△CBD, MdCRF4DD-His, or MdCRF4AA-His) was captured with Ni-NAT resin (Transgene, Beijing China; https://www.transgen.com.cn), washed 5 times, then incubated with 10-μg prey protein (MdCaM2-GST, MdXBAT31-GST, or empty GST) at 4°C for 1 h in RB buffer (50-mM Tris–HCl, pH 7.8, 100-mM NaCl, 0.1% [v/v] glycerol, and 20-mM β-mercaptoethanol). After washing 5 times with the RB buffer to remove nonspecifically bound proteins, the precipitated prey proteins were released by boiling in sodium dodecyl sulfate (SDS) sample buffer at 100°C for 5 min before detection by immunoblot analysis using an anti-His antibody (Cat. no. HT501, Transgen, China). The relative input of the prey proteins was detected by immunoblot analysis using an anti-GST antibody (Cat. no. HT601, Transgen, China). For Ca²⁺-mediated interactions between MdCRF4 and MdCaM2 assays, 1-mM CaCl₂ or 5-mM EGTA (Sangon Biotech, China) was added to the RB buffer.

Co-IP assays

For co-IP assays of apple callus, the MdCaM2 or MdXBAT31 CDS was separately inserted into the pRI101 (TaKaRa, Japan) vector containing a GFP sequence and the CaMV35S promoter to construct the Pro35S::GFP-MdCaM2 and Pro35S::GFP-MdXBAT31 plasmids. The MdCRF4, MdCRF4AA, or MdCRF4DD CDS were cloned into the pRI101 vector containing a Myc sequence and the CaMV35S promoter to construct the Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4AA, or Pro35S::Myc-MdCRF4DD plasmids. The recombinant Pro35S::GFP-MdCaM2 and Pro35S::Myc-MdCRF4, Pro35S::GFP-MdXBAT31, and Pro35S::Myc-MdCRF4, Pro35S::GFP-MdXBAT31, and Pro35S::Myc-MdCRF4AA, or Pro35S::GFP-MdXBAT31 and Pro35S::Myc-MdCRF4DD plasmids were co-transfected into apple callus as previously described (Xie et al., 2012). The transgenic callus was treated with 1-mM CaCl₂ and was used for co-IP analysis. Transgenic apple callus was treated with 50-μM MG132 (Sigma-Aldrich, St. Louis, MO, USA) for 14 h to maintain MdCRF4, MdCRF4AA, or MdCRF4DD protein stability. Callus co-transformed with the Pro35S::GFP-MdCaM2 and Pro35S::Myc, or Pro35S::GFP-MdXBAT31 and Pro35S::Myc plasmids were used as the negative controls. The transgenic callus was used for co-IP analysis. co-IP was performed as previously described (Li et al., 2016). Briefly, a Pierce classic IP Kit (Cat. no. 26149, Thermo Scientific) was used to immunoprecipitate MdCaM2-GFP or MdXBAT31-GFP using an anti-GFP antibody (Cat. no. HT801, Transgene, China). The precipitate was analyzed by immunoblot analysis with the anti-Myc antibody (Cat. no. HT101, Transgene, China).

Firefly LCI assay

The MdCaM2 or MdXBAT31 CDS was inserted into the pCAMBIA1300-nLuc (Chen et al., 2008) vector and MdCRF4, MdCRF4AA, MdCRF4DD, MdCRF4^H, MdCRF4^N, or MdCRF4^△CBD were inserted into the pCAMBIA1300-cLuc (Chen et al., 2008) vector. The corresponding plasmids were co-transformed into A. tumefaciens strain EHA105 using electroporation and cultures were infiltrated into N. benthamiana leaves as previously described (Li et al., 2016). Luciferase activity was detected as previously described (Li et al., 2017). CaCl₂ treatment (1 mM) was applied to Agrobacterium-infiltrated N. benthamiana leaves 2 h before detection.

Y1H assay

Each fragment (MdCRF4, MdCRF4-N, MdCRF4-M, and MdCRF4-C) was ligated into the pGADT7 vector. The fragment from the MdACS1 promoter was cloned into the pAbAi vector. The Y1H-gold assay was performed using the Match-maker Gold Yeast One-Hybrid Library Screening System kit (Clontech, USA).

ChIP-qPCR analysis

The MdCRF4, MdCRF4^N, MdCRF4^H, or MdCRF4AA CDS was cloned into the pRI101 vector containing a Myc sequence and the CaMV35S promoter to construct the Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4^N, Pro35S::Myc-MdCRF4^H, or Pro35S::Myc-MdCRF4AA plasmids. The recombinant constructs were transformed into apple callus as described above. The transgenic calli were treated with or without CaCl₂ (2 h), and the ChIP assays were performed as previously described (Li et al., 2016) with an anti-Myc antibody (Transgen, China). The amount of immunoprecipitated chromatin was determined by qPCR as previously described (Li et al., 2017).

GUS analysis

A reporter construct containing the MdACS1 promoter sequence (1,191-bp upstream of the start ATG) was prepared as previously described (Li et al., 2016). A mutation was introduced into the DRE motif of the MdACS1 promoter using a Fast Mutagenesis System Kit (Transgen Biotech, Beijing, China). The MdCRF4, MdCRF4DD, MdCRF4AA, or MdXBAT31 CDS were separately introduced into the pRI101 vector to generate the effector constructs. The transformation of reporter and effector constructs into N. benthamiana leaves and measurements of GUS activity were performed as previously described (Li et al., 2016). CaCl₂ treatment (1 mM) was applied to Agrobacterium-infiltrated N. benthamiana leaves 2 h before detection.

Dual-luciferase reporter assay

The MdACS1 promoter sequence was inserted into the pGreenII0800-Luc (Hellens et al., 2005) vector to generate the reporter construct. The MdCRF4 CDS was inserted into the pGreenII62-SK (Hellens et al., 2005) vector to generate the effector construct. The transformation of reporter and effector constructs into N. benthamiana leaves and measurements of luciferase activity were performed as previously described (Wang et al., 2020).

In vivo phosphorylation assay

A phos-tag mobility shift assay was used to detect in vivo phosphorylated MdCRF4 protein as previously described (Yan et al., 2018). Briefly, apple fruit (10 days after harvest) disks and transgenic callus (Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4^N, Pro35S::Myc-MdCRF4^H, or Pro35S::Myc-MdCRF4AA) were collected at 4 h after CaCl₂ treatment and powdered in liquid nitrogen. Proteins of transgenic callus were extracted in RB buffer (50-mM Tris–HCl, pH 7.4, 150-mM NaCl, 0.1% [v/v] Nonidet P-40, 1% [v/v] Triton X-100). The anti-Myc antibody was used to immunoprecipitate corresponding proteins according to the method of a Pierce classic IP Kit (Cat. no. 26149, Thermo Scientific). After washing 5 times with RB buffer, the precipitated proteins were released by boiling in SDS sample buffer for 10 min. The immunoprecipitated proteins were separated in a 10% SDS–PAGE gel containing 50-μM phos-tag and 100-μM MnCl₂. After proteins were transferred to a polyvinylidene fluoride (PVDF) membrane, Myc-tagged MdCRF4 was detected using the anti-Myc antibody. As a loading control, the precipitated proteins were separated in a 10% SDS–PAGE gel and detected by immunoblotting using the same procedures. In a second method, proteins from apple fruit disks were immunoprecipitated with anti-MdCRF4 as above. The immunoprecipitated proteins were separated by 10% SDS–PAGE for immunoblotting using an anti-PhosphoSer/Thr antibody (Cat. no. ab117253, Abcam, USA).

In vitro protein degradation assay

For the in vitro phosphorylated MdCRF4 protein degradation assay, apple fruit disks were treated with 100-mM CaCl₂ and then powdered in liquid nitrogen prior to extraction in protein extraction buffer (25-mM Tris–HCl, pH 7.5, 10-mM NaCl, 10-mM MgCl₂, 5-mM 1,4-dithio-dl-threitol (DTT), and 5-mM ATP). The supernatant was collected, and the concentration of the extracts determined using the Bradford assay (TransGen, China). Then, 1 μg of purified His-tagged protein (MdCRF4-His, MdCRF4^H, and MdCRF4^N) was added to 1-mL total protein extract (0.5 μg μL⁻¹) solution supplemented with MG132 (50 μM) or dimethyl sulfoxide and incubated 0, 2, 4, and 6 h at 25°C. To determine whether MdXBAT31 mediated the degradation of phosphorylated MdCRF4, MdCRF4DD-His, and MdCRF4AA-His proteins were obtained as described above. A partial MdXBAT31 CDS (300–702 bp) was cloned into the pRI101 vector in the reverse direction to generate the pRI101-MdXBAT31-AS antisense construct to silence MdXBAT31 in apple callus according to previously described (Li et al., 2017). Protein extracts from pRI101-MdXBAT31-AS transgenic callus, as well as purified MdCRF4DD-His and MdCRF4AA-His were used for the in vitro protein degradation assay according to the same procedures described. Transgenic pRI101 empty vector callus was used as a control.

In vitro and in vivo ubiquitination assay

The in vitro ubiquitination assay was performed as previously described (Li et al., 2012). Briefly, 50 ng of human E1 (Boston Biochem), 50 ng of human E2 (Boston Biochem), 200 ng of E3 (MdXBAT31), 10 μg of His-6-ubiquitin (Boston Biochem), and 100 ng of MdCRF4AA (or MdCRF4DD) were incubated in 30 μL of ubiquitination reaction buffer (50-mM Tris–HCl pH 7.5, 10-mM MgCl₂, 10-mM ATP, and 1-mM DTT) at 30°C for 4 h. The proteins were separated in an SDS–PAGE gel, and the ubiquitinated MdCRF4AA or MdCRF4DD was detected with an anti-ubiquitin antibody (Cat. no. A-100, UBBiotech, Changchun, China) or anti-His antibody (Transgen Biotech, Beijing, China). In vivo ubiquitination assays were performed as previously described (An et al., 2017). Briefly, Pro35S::Myc-MdCRF4 transgenic callus and Pro35S::Myc-MdCRF4/pRI101-MdXBAT31-AS (silenced MdXBAT31) co-transgenic callus was pretreated with 50-μM MG132 proteasome inhibitor for 16 h. A Pierce classic IP kit (Thremo Scientific) was used to immunoprecipitate MdCRF4 with an anti-Myc antibody, and the eluted proteins were analyzed by immunoblot analysis with anti-ubiquitin antibody. We made Pro35S::Myc, Pro35S::Myc-MdCRF4, Pro35S::Myc-MdCRF4AA, and Pro35S::Myc-MdCRF4^△CBD transgenic callus to investigate how the phosphorylation status of MdCRF4 affects its ubiquitination. The above-mentioned transgenic callus was pretreated with MG132 for 16 h and CaCl₂ for 2 h. The ubiquitination analysis method was the same as above.

Agrobacterium infiltration

To silence MdXBAT31 expression in apple fruit, a partial MdXBAT31 (300–702 bp) sequence was inserted into the pTRV vector. The recombinant plasmid was transformed into A. tumefaciens strain EHA105 by electroporation. The preparation of the infection suspension and silencing of MdXBAT31 expression in apple fruit were as previously described (Li et al., 2016). To overexpress MdCRF4, MdCRF4DD, or MdCRF4AA in apple fruit, the MdCRF4, MdCRF4DD, or MdCRF4AA CDS was separately ligated into the pRI101 plasmid. The recombinant plasmids were transformed into A. tumefaciens strain EHA105. The preparation of the infection suspension and overexpression of MdCRF4, MdCRF4DD, or MdCRF4AA in apple fruit were performed as previously described (Li et al., 2016).

Accession numbers

Sequences data in this study can be found in the Genome Database for Rosaceae (https://www.rosaceae.org) and apple genome at the NCBI (National Center for Biotechnology Information). Sequence accession numbers are: MdCaM1 (MDP0000316087), MdCaM2 (MD14G1092700), MdCaM3 (MDP0000865414), MdCaM4 (MDP0000183898), MdCaM5 (MD03G1163500), MdCaM6 (MDP0000234624), MdCaM7 (MDP0000277474), MdCaM8 (MDP0000208420), MdCRF4 (NM_001328784), MdERF2 (AB288348), MdERF3 (XM_008339725), MdACS1 (U89156), MdACO1 (AF030859), and MdXBAT31 (XM_008359188).

Supplemental data

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

Supplemental Figure S1. MdCaM gene expression in apple fruit.

Supplemental Figure S2. Sequence alignment of MdCRF4 with homologs from A. thaliana.

Supplemental Figure S3. MdCRF4 expression during apple fruit ripening.

Supplemental Figure S4. MdCRF4 silencing causes downregulation of MdACS1 and MdERF3 expression, thereby suppressing ethylene biosynthesis.

Supplemental Figure S5. MdCRF4 does not bind to the MdACO1 promoter.

Supplemental Figure S6. The MdCRF4 AP2/ERF domain binds to the MdACS1 and MdERF3 promoters.

Supplemental Figure S7. Apple fruit disks treated with various CaCl₂ concentrations.

Supplemental Figure S8. MdXBAT31 does not interact with nonphosphorylated MdCRF4AA.

Supplemental Table S1. Eight CaMs names and ID in apple draft genomic sequence V1.0.

Supplemental Table S2. List of primers used in this study.

 

A.W. and T.L. conceived and designed this research. T.L. performed most of the experiments. Y.X. and Y.W. generated the constructs for protein purification. W.L. and Y.J.W. performed the RT-qPCR assay. L.L. performed protein purification. Z.L. and T.X.L. provided apple fruit. T.L. wrote the article. All authors analyzed the data and discussed the article.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Aide Wang (awang@syau.edu.cn).