SnRK1 kinase-mediated phosphorylation of transcription factor bZIP39 regulates sorbitol metabolism in apple

Abstract

Sorbitol is a major photosynthate produced in leaves and transported through the phloem of apple (Malus domestica) and other tree fruits in Rosaceae. Sorbitol stimulates its own metabolism, but the underlying molecular mechanism remains unknown. Here, we show that sucrose nonfermenting 1 (SNF1)-related protein kinase 1 (SnRK1) is involved in regulating the sorbitol-responsive expression of both SORBITOL DEHYDROGENASE 1 (SDH1) and ALDOSE-6-PHOSPHATE REDUCTASE (A6PR), encoding 2 key enzymes in sorbitol metabolism. SnRK1 expression is increased by feeding of exogenous sorbitol but decreased by sucrose. SnRK1 interacts with and phosphorylates the basic leucine zipper (bZIP) transcription factor bZIP39. bZIP39 binds to the promoters of both SDH1 and A6PR and activates their expression. Overexpression of SnRK1 in ‘Royal Gala’ apple increases its protein level and activity, upregulating transcript levels of both SDH1 and A6PR without altering the expression of bZIP39. Of all the sugars tested, sorbitol is the only 1 that stimulates SDH1 and A6PR expression, and this stimulation is blocked by RNA interference (RNAi)-induced repression of either SnRK1 or bZIP39. These findings reveal that sorbitol acts as a signal regulating its own metabolism via SnRK1-mediated phosphorylation of bZIP39, which integrates sorbitol signaling into the SnRK1-mediated sugar signaling network to modulate plant carbohydrate metabolism.

SnRK1 kinase modulates sorbitol metabolism in apple via phosphorylation of transcription factor bZIP39, which integrates sorbitol signaling into the SnRK1-mediated sugar signaling network.

Introduction

Sorbitol is one of the most common sugar alcohols in plants. While it is present only in small amount in many plants, sorbitol accounts for 60% to 80% of the photosynthates produced in leaves and transported in the phloem of apple (Malus domestica) and all other pome and stone fruits in the Rosaceae family (Bieleski 1982; Loescher 1987; Moing et al. 1992; Cheng et al. 2005). As such, sorbitol plays an important role in carbohydrate metabolism, plant growth and development, and stress tolerance of this important group of tree species. Sorbitol is synthesized from a common hexose phosphate pool shared with sucrose synthesis in the cytosol of source leaves via a 2-step process. Glucose 6-phosphate is first reduced to sorbitol 6-phosphate by aldose-6-phosphate reductase (A6PR) using NADPH, and then, sorbitol 6-phosphate is converted to sorbitol by sorbitol-6-phosphate phosphatase, with the first step being rate limiting (Negm and Loescher 1981; Zhou, Cheng, and Wayne et al. 2003; Zhou, Sicher, et al. 2003). The generated sorbitol enters the sieve element–companion cell complex in the minor vein phloem of leaves via diffusion through plasmodesmata (Reidel et al. 2009; Fu et al. 2011) and then transported to sink organs such as growing shoot tips and fruit. In the sink cells, sorbitol is primarily oxidized by sorbitol dehydrogenase (SDH) to fructose while generating NADH (Negm and Loescher 1979; Yamaguchi et al. 1994). The resulting fructose either enters the sucrose cycle for further metabolism or is transported into vacuole for storage (Li et al. 2012, 2018). When sorbitol synthesis is decreased by antisense repression of A6PR expression in ‘Greensleeves’ apple trees, starch synthesis is upregulated without altering sucrose synthesis or CO₂ assimilation during the day (Cheng et al. 2005). However, sucrose accumulates to a higher level in the leaves of the antisense A6PR lines due to increased sucrose synthesis from enhanced starch breakdown at night, making more sucrose available for export. In response to the decreased sorbitol supply and increased sucrose supply, the expression and activity of SDH are downregulated, whereas the expression and activities of sucrose cleavage enzymes, sucrose synthase (SuSy) in shoot tips, and both SuSy and cytosolic invertase in growing fruit are upregulated in the transgenic apple trees (Zhou et al. 2006; Li et al. 2018). Exogenous sorbitol feeding via transpirational stream restores the expression and activity of SDH in the shoot tips of the transgenic trees (Zhou et al. 2006). However, how sorbitol regulates SDH expression remains unclear.

In addition to being a carbon source, sugars act as signals in regulating plant metabolism, growth and development, and stress tolerance (Wingler 2018; Wurzinger et al. 2018; Baena-González and Lund 2020). Three major sugar signaling pathways have been identified for hexose and/or sucrose in plants. First, hexokinase 1 (HXK1) directly senses cellular glucose level independent of its catalytic activity to relay nutrient, light, and hormone signals for controlling growth and development (Moore et al. 2003). Second, target of rapamycin (TOR) signaling is stimulated by glucose to control metabolic transcriptional network, cell cycle and cell fate, and developmental transitions for many aspects of plant growth and development (Xiong et al. 2013; Ye et al. 2022). Finally, sucrose nonfermenting 1 (SNF1)-related protein kinase 1 (SnRK1) senses cellular energy deficit to adjust metabolism and growth to the prevailing sugar and energy status of the plant (Baena-González et al. 2007; Wurzinger et al. 2018; Baena-González and Lund 2020). Sorbitol has recently been demonstrated to serve as a signal modulating flower bud formation in loquat (Eriobotrya japonica) (Xu et al. 2022) and stamen development, pollen tube growth, and resistance against Alternaria alternata in apple (Meng, He, et al. 2018; Meng, Li, et al. 2018; Li, Meng, et al. 2020). Considering that both HXK1- and TOR-modulated sugar signaling pathways are specific to glucose and sorbitol appears to be a good indicator of apple tree energy status, we postulated that sorbitol signaling on its own metabolism may act through the evolutionally conserved metabolic integrator SnRK1.

SnRK1 functions as a heterotrimeric complex consisting of 1 catalytic subunit (α subunit) and 2 regulatory subunits (β and γ subunits), similar to its counterparts, SNF1 in yeast (Saccharomyces cerevisiae) and AMP-activated kinase in the mammalian system (Broeckx et al. 2016). In Arabidopsis (Arabidopsis thaliana), overexpression (OE) of the catalytic subunit is sufficient to enhance its enzyme activity (Baena-González et al. 2007). Under conditions of adequate energy supply, SnRK1 activity is inhibited by trehalose 6-phosphate (T6P), a proxy for sucrose accumulation (Zhang et al. 2009; Wahl et al. 2013; Fichtner and Lunn 2021), and its gene expression is repressed by sucrose and other sugars (Baena-González et al. 2007; Baena-González and Lund 2020). Upon activation in response to energy deficit, SnRK1 initiates metabolic reprograming by direct phosphorylation of key metabolic enzymes and regulation of a vast transcriptional network to upregulate catabolic processes such as breakdown of sucrose, starch, cell wall, proteins, amino acids, and lipids while downregulating energy-demanding anabolic processes such as synthesis of sugars and starch, cell wall, amino acids and proteins, lipids, and cell cycle. This transcriptional reprograming is achieved by phosphorylation of a number of transcription factors by SnRK1 and subsequent binding of these transcription factors to the promoters of their target genes. In Arabidopsis, basic leucine zipper (bZIP) transcriptional factors characterized as G-box binding factors and closely related S- and C-group bZIP2, bZIP11, bZIP63, and bZIP25 are involved in the process, resulting in activation of nearly 300 genes involved in catabolic processes and repression of over 300 genes responsible for anabolic processes (Baena-González et al. 2007). In rice (Oryza sativa), SnRK1A, the ortholog of yeast SNF1, acts (probably via phosphorylation) upstream of MYBS1, a R1 MYB transcription factor, that binds to the TA box in the promoter of α-AMYLASE 3, activating its expression to degrade starch during germination and seedling growth under sugar starvation (Lu et al. 2002, 2007). In addition, the SnRK1 signaling pathway interacts with the ABA signaling pathway in regulating plant metabolism in response to abiotic stress (Jossier et al. 2009; Carianopol et al. 2020; Van Leene et al. 2022). Of the genes that are upregulated by OE of AtKIN10 (SnRK1.1) in Arabidopsis (Baena-González et al. 2007), we noticed that At5g51970 encodes a functional SDH (Nosarzewski et al. 2012; Aguayo et al. 2013). This gene is also among those downregulated by OE of Escherichia coli ots A gene encoding T6P synthase in Arabidopsis (Schluepmann et al. 2003; Zhang et al. 2009), consistent with the involvement of SnRK1 in regulating SDH expression via inhibition of its kinase activity by the elevated T6P level. Based on these clues, we hypothesized that sorbitol regulates SDH expression via the SnRK1 pathway in apple.

In this work, we show that SnRK1 is involved in regulating the expression of both SORBITOL DEHYDROGENASE 1 (SDH1) and A6PR, encoding 2 key enzymes for sorbitol catabolism and synthesis, respectively, in response to sorbitol. SnRK1 phosphorylates bZIP transcription factor bZIP39, which binds to the promoters of both SDH1 and A6PR in activating their transcription.

Results

Identification of MdbZIP39, a bZIP transcription factor that binds to the promoter of MdSDH1 and interacts with MdSnRK1

In earlier work, we found that antisense repression of A6PR expression in ‘Greensleeves’ apple led to a lower sorbitol concentration, a higher sucrose concentration, and lower expression of SDH in the shoot tips; the SDH transcript level was restored by exogenous sorbitol feeding but further decreased by sucrose feeding via the transpirational stream (Zhou et al. 2006). As SnRK1 has been demonstrated to serve as a metabolic sensor for plant sugar and energy status and its OE led to upregulation of SDH expression in Arabidopsis (Baena-González et al. 2007; Wurzinger et al. 2018), we hypothesized that SnRK1 is involved in sensing sorbitol level as well as sucrose level in plants. To test this hypothesis, we fed the youngest fully expanded leaves from in vitro shoots of wild type (WT) and 2 antisense A6PR lines (A4 and A10) with 50 mM sorbitol or 50 mM sucrose for 3 h and measured the expression of SnRK1 (MD09G1056200), SDH1 (MD01G1110100), and SDH2 (MD01G1195200) via reverse transcription quantitative PCR (RT-qPCR). SDH1 and SDH2 are 2 main SDH-encoding genes, with SDH1 expressed in both source and sink tissues while SDH2 primarily expressed in sinks such as growing shoot tips and fruit (Park et al. 2002; Nosarzewski and Archbold 2007; Li et al. 2012). Sorbitol and sucrose feeding significantly increased concentrations of sorbitol and sucrose in the leaves of WT, A4, and A10, respectively (Fig. 1A). The elevated sorbitol level led to significant increases in the expression levels of SnRK1, SDH1, and SDH2, whereas the elevated sucrose level decreased their expression levels (Fig. 1B).

Figure 1.

Expression of MdSnRK1, MdSDH1, and MdSDH2 in response to sorbitol/sucrose feeding and verification of transcription factors interacting with the promoters of MdSDH1 and MdSDH2 via Y1H assay. A) Concentrations of sorbitol and sucrose in the leaves of WT and 2 antisense A6PR lines (A4 and A10) of ‘Greensleeves’ apple in response to sorbitol (50 mM) or sucrose (50 mM) feeding via the petiole for 3 h. B) Expression levels of MdSnRK1, MdSDH1, and MdSDH2 under sorbitol and sucrose feeding treatments, respectively. In both A) and B), data are the mean ± Sd of 3 biological replicates, with 10 leaves from 5 shoots per replicate. Lsd was used for significant difference at P < 0.05 after ANOVA. Different letters (a to g) represent significant differences between groups. C) Diagram showing the Y1H screening of an apple cDNA library using the promoters of SDH1 and SDH2. D) Verification of the interactions between 4 candidate transcription factors selected from each Y1H screening and the promoters of SDH1 and SDH2 via Y1H, respectively. SD-Trp-Leu, synthetic dextrose medium lacking tryptophan and leucine; SD/-Trp-Leu-His, synthetic dextrose medium lacking tryptophan, leucine, and histidine; 3-AT, 3-amino triazole. Triangles represent the range of yeast concentrations. Each colony was diluted from 1 to 10⁻².

We cloned the SnRK1 gene in apple, which belongs to the SnRK1 subfamily and is an ortholog of AtSnRK1.1 (Supplemental Fig. S1). As the SDH gene expression was increased significantly in Arabidopsis overexpressing KIN10 (SnRK1.1) (Baena-González et al. 2007), we predicted the existence of a transcription factor that binds to the promoter of SDH1 or SDH2 and interacts with SnRK1 to relay the sorbitol signal sensed by SnRK1. We analyzed the promoter regions of MdSDH1 and MdSDH2 and used them as baits in a yeast 1-hybrid (Y1H) system to screen an apple cDNA library (Fig. 1C). The Y1H screen yielded 283 clones (Group 1) and 145 clones (Group 2) for the promoters of SDH1 and SDH2, respectively. By sequencing the clones, we identified 152 genes and 87 genes from the 2 groups, respectively, via NCBI blast and domain search. Based on gene function annotations, 4 transcriptional factors from each group were selected: 6D8, 3D3, 5C8, and 5D8 from Group 1 and 4F3, 51A1, 5D3, and 1A4 from Group 2 (Supplemental Table S1). By transforming all 8 candidate transcription factors into the yeast with the corresponding MdSDH1/2 promoter vector, we confirmed their interactions in Y1Hs (Fig. 1D).

To identify the transcription factor that interacts with SnRK1, we conducted a yeast 2-hybrid (Y2H) experiment where all 8 candidate genes were separately co-transformed with pGBKT7-MdSnRK1 into the yeast. After growth on the SD/-Trp-Leu medium, the yeast were transferred into the selected SD medium SD/-Trp-Leu-His-Ade. Among the 8 proteins, only 6D8 showed a strong interaction with MdSnRK1 (Fig. 2A). 6D8 (MD08G1025800) is a bZIP transcription factor (Li et al. 2016), which we named MdbZIP39. The LacZ activity assays corroborated the yeast growth data (Fig. 2B). We used bimolecular fluorescence complementation (BiFC) analysis to further confirm the interaction between bZIP39 and SnRK1 and verify the localization of MdbZIP39 (Fig. 2C). The YFP signal was detected in the nucleus as demonstrated by 4′,6-diamidino-2-phenylindole (DAPI), indicating that MdbZIP39 interacts with MdSnRK1 in the nucleus.

Figure 2.

Y2H and BiFC assays of the interaction between MdSnRK1 and MdbZIP39. A) Y2H assay of the interaction of MdSnRK1 with MdbZIP39 and other candidate transcription factors. The pair AD-p53 and BD-SV40 was used as a positive control; the pair AD-pGADT7 and BD-pGBKT7 was used as a negative control. AD, activation domain; BD, binding domain; SD-Leu-Trp, synthetic dextrose medium lacking leucine and tryptophan; SD/-Trp-Leu-His-Ade, synthetic dextrose medium lacking leucine, tryptophan, histidine, and adenine. Triangles represent the range of yeast concentrations. Each colony was diluted in 10 µL sterile water and then diluted to 10⁻¹ to 10⁻³. At least 3 colonies were tested for each combination. B) Activity statistics of the reporter enzyme, β-galactosidase (LacZ), in the Y2H experiments. Miller units of LacZ activity were calculated as 1,000 × OD_420nm/(volume × time × OD_600nm). Different letters (a to c) represent significant differences between groups. Data are the mean ± Sd of 3 biological replicates with 1 colony per replicate. Lsd was used for significant difference at P < 0.05 after ANOVA. C) BiFC assay of the interaction between MdSnRK1 and MdbZIP39 in leaves of N. benthamiana. MdSnRK1 and MdbZIP39 were introduced into the pSPYNE and pSPYCE vectors and fused to N-terminal and C-terminal YFP, respectively. YFPc + MdSnRK1-YFPn was used as a negative control. DAPI, 4′,6-diamidino-2-phenylindole. Bars = 25 µm.

MdbZIP39 protein is phosphorylated by MdSnRK1

As SnRK1 is a protein kinase, we predicted that the interaction between SnRK1 and bZIP39 is probably phosphorylation of bZIP39 by SnRK1. To determine if that is the case, we incubated MdSnRK1 with various forms of MdbZIP39 protein in a phosphorylation reaction solution containing fluorescently labeled ATP and assayed the extent of the phosphorylation reaction using ADP-Glo kinase after 1.5 h (Fig. 3). A total of 10 groups were included in this phosphorylation assay. In addition to 3 controls (buffer only, His, or GST), 3 forms of MdbZIP39 were used: His-MdbZIP39, His-MdbZIP39^S41A that has the serine replaced by alanine at position 41 based on the predicted phosphorylation sites and DNA-binding site (Supplemental Fig. S2), and His-MdbZIP39X that has been deactivated by boiling at 100 °C for 10 min. These 3 forms of MdbZIP39 were either alone or combined with GST-MdSnRK1. The consumption of fluorescently labeled ATP was measured as a decrease in the relative fluorescence unit (RFU) over time. We detected significant decreases in RFU in 4 treatments (GST-MdSnRK1, His-MdbZIP39 + GST-MdSnRK1, His-MdbZIP39^S41A + GST-MdSnRK1, and His-MdbZIP39X + GST-MdSnRK1) relative to the 3 controls and 3 treatments involving His-MdbZIP39 alone. Of these 4 treatments, the combination of His-MdbZIP39 and GST-MdSnRK1 consumed significantly more ATP than the other 3; substitution of serine with alanine at position 41 in MdbZIP39 or deactivation of MdbZIP39 abolished its phosphorylation by MdSnRK1. These data indicate that MdSnRK1 is able to phosphorylate not only itself but also MdbZIP39 (Fig. 3A). To verify the phosphorylation of MdbZIP39 by MdSnRK1, we incubated the GST-MdSnRK1 and His-MdbZIP39 with ATP, separated them on a sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) gel, and used a phospho-(Ser/Thr) Phe antibody to detect the phosphorylated proteins. The phospho-specific antibody detected 2 phosphorylated proteins, GST-MdSnRK1 and His-MdbZIP39, only in the combination of His-MdbZIP39 and GST-MdSnRK1, whereas it detected phosphorylated GST-MdSnRK1 in 3 other reactions involving GST-MdSnRK1 (Fig. 3B). These results confirm that MdSnRK1 phosphorylates MdbZIP39 and exhibits autophosphorylation activity.

Figure 3.

MdSnRK1 phosphorylates MdbZIP39. A) Phosphorylation activity of MdSnRK1 measured as decrease in relative fluorescence. Among the 10 groups listed, all proteins involved were obtained after prokaryotic expression and purification. His-MdbZIP39^S41A has the serine mutated to alanine at position 41 based on the predicted phosphorylation sites and DNA-binding site (Supplemental Fig. S2), whereas His-MdbZIP39X denotes the deactivated His-MdbZIP39 by boiling at 100 °C for 10 min. Different letters (a to d) indicate a significant difference between treatment groups and the control (phosphorylation reaction buffer only). Data are means ± Sd (n = 3) with measurements made in duplicate. Lsd was used for significant difference at P < 0.05 after ANOVA. B) Ser/Thr antibody detects phosphorylation of MdbZIP39 as well as MdSnRK1 by MdSnRK1. Proteins labeled with the circled “P” refer to their phosphorylated forms.

MdbZIP39 binds to the promoter of MdSDH1 and enhances its expression

MdbZIP39 has 156 amino acids, which contains a DNA-binding and dimerization domain (cl21462, NCBI conserved domain architecture retrieval tool ACGT https://www.ncbi.nlm.nih.gov/Structure/lexington/lexington.cgi). It is localized to the nucleus by confocal laser microscopy (Fig. 4A), consistent with its predicted function as a transcription factor. By analyzing the MdSDH1 promoter using the PlantPAN2.0 website (http://plantpan2.itps.ncku.edu.tw/), we identified 1 bZIP cis-regulatory element (ACGT) around −384 bp in the MdSDH1 promoter. To confirm the binding of MdbZIP39 to the cis-element, we divided the MdSDH1 promoter into 4 sections: SP1 to SP4, ligated them into the pHis2 vector, and co-transformed them with MdbZIP39 into the yeast. Yeast strains harboring SP1, SP2, or SP4 with the predicted binding motif all grew on the selected SD medium with 60 mM 3-amino-1,2,4-triazole (3-AT), whereas those containing SP3 without the motif did not (Fig. 4B). These results indicate that MdbZIP39 binds to the MdSDH1 promoter via its binding site. To corroborate the Y1H binding assay, we conducted electrophoretic mobility shift assay (EMSA) by using a prokaryote-expressed and purified His-MdbZIP39 fusion protein. When a biotin-labeled oligonucleotide probe containing the binding motif (ACGT) was used, a specific DNA-MdbZIP39 complex was detected, but the formation of this complex was attenuated with increasing amount of the unlabeled probe (cold probe), and the complex was not detected when ACGT in the labeled probe was mutated to GTTC (Fig. 4C). These data confirm that MdbZIP39 directly binds to the bZIP cis-element in the MdSDH1 promoter in vitro. To verify the in vivo binding of MdbZIP39 to the promoter of MdSDH1, we performed chromatin immunoprecipitation (ChIP)-qPCR using ‘Orin’ apple calli expressing bZIP39 fused with RFP or RFP alone. The MdbZIP39-binding site containing region (S2) of the MdSDH1 promoter was significantly enriched in the calli overexpressing MdbZIP39 fused with RFP relative to those only expressing RFP (Fig. 4D).

Figure 4.

Binding of MdbZIP39 to the promoter of MdSDH1 and transcriptional activation of MdSDH1. A) Subcellular localization of MdbZIP39 in apple leaves. Bars = 50 µm. B) Y1H assay on binding of MdbZIP39 protein to the promoter of MdSDH1. Numbers indicate nucleotide positions for different regions (SP1, SP2, SP3, and SP4) upstream of the MdSDH1 coding sequence with the first nucleotide of its start codon designated as 0; the short vertical line denotes the binding motif ACGT at −384. SD-Trp-Leu, synthetic dextrose medium lacking tryptophan and leucine; SD/-Trp-Leu-His, synthetic dextrose medium lacking tryptophan, leucine, and histidine; 3-AT, 3-amino triazole. Triangles represent the range of yeast concentrations. Each colony was diluted from 1 to 10⁻². pGADT7-p53 + pHis2-p53 was used as a positive control and pGADT7 + pHis2-p53 as a negative control. C) EMSA on the specific binding of MdbZIP39 protein to the motif ACGT in the MdSDH1 promoter. The biotin-labeled probe was designed to target the ACGT motif; mutant probe has the ACGT motif mutated to GTTC. The amount of cold probes was 10 and 40 times higher than that of the labeled probe. D) ChIP-qPCR analysis of MdbZIP39 binding to the promoter of MdSDH1. S1–S3 represent different primers targeting the various regions of the MdSDH1 promoter (Supplemental Table S6). ACTIN was used as an internal reference. Data were means ± Sd (n = 3) with calli from 1 petri dish per replicate. Lsd was used for significant difference at P < 0.01 after ANOVA. E) The relative expression levels of MdbZIP39 and MdSDH1 in response to MdbZIP39-OE via agroinfiltration to leaves, with empty vector as control. Data are means ± Sd (n = 3) with 10 leaves from 5 shoots per replicate. Lsd was used for significant difference at P < 0.05 after ANOVA. In both D) and E), different letters (a to b) represent significant differences between groups.

To determine the role of MdbZIP39 in regulating the expression of MdSDH1, we overexpressed MdbZIP39 in apple leaves via infiltration with Agrobacterium tumefaciens GV3101 harboring MdbZIP39-OE. This increased the transcription of MdbZIP39 to a significantly higher level in MdbZIP39-OE leaves relative to the empty vector control, leading to a significantly higher expression of MdSDH1 (Fig. 4E). These data indicate that MdbZIP39 enhances the transcription of MdSDH1.

MdA6PR is highly expressed in the transcriptome of shoot tips overexpressing MdSnRK1 and is transcriptionally activated by MdbZIP39

To better characterize the function of MdSnRK1, we overexpressed MdSnRK1 fused with GFP in ‘Royal Gala’ via agrobacterium-mediated transformation and obtained over 20 lines. Two lines, OE-24 and OE-46, with significantly higher MdSnRK1 expression were propagated. After acclimation, they were transferred to full sun conditions outside. During active shoot growth, shoot tips with expanding leaves were taken from WT, OE-24, and OE-46, with 5 biological replicates each, for RNA sequencing (RNAseq). After cleaning the raw reads by removing adapters, ribosomal RNAs (rRNAs), and low-quality reads, we obtained a total of 297 million high-quality reads, of which more than 94% were mapped to the doubled haploid ‘Golden Delicious’ reference genome (GDDH13) for identifying differentially expressed genes (DEGs) (Supplemental Tables S2 to S4). Principal component analysis (PCA) of the transcriptome data revealed that the 2 MdSnRK1-OE lines differed significantly from WT at the transcriptional level (Fig. 5A), which indirectly indicates the functional importance of MdSnRK1 as an upstream regulatory kinase. Interestingly, we found that the MdA6PR gene (MD10G1062300), encoding the key enzyme for sorbitol synthesis, had higher transcript abundance in MdSnRK1-OE lines (Figs. 5B and S3). MdA6PR showed a trend of co-expression with MdSnRK1, suggesting that MdA6PR may be also regulated by MdbZIP39. By analyzing the MdA6PR promoter, we identified a bZIP cis-regulatory element (ACGTAC) around −254 in the promoter (Fig. 5C). To confirm the binding of MdbZIP39 to the MdA6PR promoter, we divided the promoter into 4 sections: AP1 to AP4, ligated them into the pHis2 vector, and co-transferred them into yeast with MdbZIP39 (Fig. 5C). The AP1, AP3, and AP4 yeast strains grew on selected SD media containing 60 mM 3-AT, which all contained the −254 fragment of the promoter, whereas AP2 yeast strain did not. To verify the in vitro binding of MdbZIP39 to the MdA6PR promoter, we conducted EMSAs using a prokaryote-expressed and purified fusion protein. A DNA–protein complex was detected when a biotin-labeled probe containing the binding motif ACGTAC was used, but formation of the complex was reduced with increasing amount of cold probe, and the complex was no longer detectable when ACGTAC in the labeled probe was mutated to GTTCAC (Fig. 5D). These data show that MdbZIP39 binds to the −254 bp promoter region of MdA6PR. To confirm the in vivo binding of MdbZIP39 to the MdA6PR promoter, we conducted ChIP-qPCR using apple calli overexpressing MdbZIP39 fused with RFP or RFP alone. The A2 region of the MdA6PR promoter, which contains the expected binding motif, was significantly enriched in the MdbZIP39-RFP-OE calli relative to the RFP-OE calli (Fig. 5E).

Figure 5.

Transcriptomic analysis of MdSnRK1-OE lines and assays on the binding of MdbZIP39 protein to the promoter of MdA6PR. A) PCA of shoot tip transcriptomes of WT and 2 MdSnRK1-GFP OE lines (OE-24 and OE-46) of ‘Royal Gala’ apple, with 5 biological replicates each. B) Expression levels of MdA6PR homologs in transcriptome of WT, OE-24, and OE-46. Transcript levels are presented as the average of RPKM values of 5 biological replicates in the heat map. C) Y1H assay on binding of MdbZIP39 protein to the promoter of MdA6PR. Numbers indicate nucleotide positions for different regions (AP1, AP2, AP3, and AP4) upstream of the MdA6PR coding sequence with the first nucleotide of its start codon designated as 0; the short vertical line denotes the binding motif ACGTAC at −254. SD-Trp-Leu, synthetic dextrose medium lacking tryptophan and leucine; SD/-Trp-Leu-His, synthetic dextrose medium lacking tryptophan, leucine, and histidine; 3-AT, 3-amino triazole. Triangles represent the range of yeast concentrations. Each colony was diluted from 1 to 10⁻². pGADT7-p53 + pHis2-p53 was used as a positive control and pGADT7 + pHis2-p53 as a negative control. D) EMSA on the specific binding of MdbZIP39 protein to the motif ACGTAC (indicated by the arrow) in the MdA6PR promoter. The biotin-labelled probe was designed to target the ACGTAC motif; mutant probe has the ACGTAC motif mutated to GTTCAC. The amount of cold probes was 10 and 40 times higher than that of the labeled probe. E) ChIP-qPCR analysis of MdbZIP39 binding to the promoter of MdA6PR. A1 to A3 represent different primers targeting the various regions of the MdA6PR promoter (Supplemental Table S6), and the A2 region contains the motif ACGTAC. ACTIN was used as an internal reference. Data were means ± Sd (n = 3) with calli from 1 petri dish per replicate. F) The relative expression levels of MdbZIP39 and MdA6PR in response to MdbZIP39-OE via agroinfiltration to leaves, with empty vector as control. Data were means ± Sd (n = 3) with 10 leaves from 5 shoots per replicate. Lsd was used for significant difference at P < 0.05 after ANOVA. Different letters (a to b) represent significant differences between groups.

To confirm the role of MdbZIP39 in regulating the expression of MdA6PR, we overexpressed MdbZIP39 in apple leaves via infiltration with A. tumefaciens GV3101 harboring MdbZIP39-OE. This significantly increased the transcript levels of both MdbZIP39 and MdA6PR (Fig. 5F), indicating that MdbZIP39 enhances the transcription of MdA6PR.

Overexpression of MdSnRK1 leads to increases in the transcript levels of both MdSDH1 and MdA6PR without altering the expression of MdbZIP39

We used the 2 SnRK1-OE lines, OE-24 and OE-46, to determine if altering MdSnRK1 expression affects the expression of MdbZIP39, SDH1/2, and A6PR and their activities and sorbitol levels. We extracted proteins from shoot tips for immunoblot analysis using antibodies against GFP and AtKIN10, respectively. Overexpression of MdSnRK1 in OE-24 and OE-46 suppressed the expression of native MdSnRK1 protein to some extent but enhanced the overall expression of MdSnRK1 protein to a higher level than WT (Fig. 6A). This led to significantly higher SnRK1 kinase activities (Fig. 6B). Overexpression of MdSnRK1 did not alter the transcript level of MdbZIP39 or MdSDH2 but significantly increased the expression levels of both MdSDH1 and MdA6PR (Fig. 6C). As we have shown that MdSnRK1 phosphorylates MdbZIP39 and MdbZIP39 transcriptionally activates the expression of both MdSDH1 and MdA6PR, the lack of transcriptional change in MdbZIP39 is consistent with the scenario where phosphorylation of MdbZIP39 by MdSnRK1 increases its activity, consequently enhancing the transcription of both MdSDH1 and MdA6PR. Correspondingly, both OE-24 and OE-46 lines had higher activities of A6PR and SDH than the WT control (Fig. 6D). Concentrations of both sorbitol and sucrose were also significantly higher in both OE-24 and OE-46 than the WT control, with fructose and glucose concentrations unaltered (Fig. 6E).

Figure 6.

Upregulation of the transcript levels of MdSDH1 and MdA6PR and their activities by OE of MdSnRK1. A) Immunoblot analysis of MdSnRK1 protein in the shoot tips of WT and 2 MdSnRK1-GFP OE lines (OE-24 and OE-46) of ‘Royal Gala’ apple plants. Sigma anti-GFP and anti-AtKIN10 were used in 1:2,000 dilution, and gel was stained with Coomasie blue. CBB, Coomasie brilliant blue. B) SnRK1 activity in the shoot tips of OE-24 and OE-46 relative to WT. C) Relative expression levels of MdSnRK1, MdbZIP39, MdSDH1, MdSDH2, and MdA6PR in WT, OE-24, and OE-46 lines. D) Enzyme activities of SDH and A6PR in the shoot tips of WT, OE-24, and OE-46 lines. E) Concentrations of sorbitol and sucrose in the shoot tips of WT, OE-24, and OE-46 lines. From B) to E), data were means ± Sd (n = 5) with 5 shoot tips per replicate. Lsd was used for significant difference analysis at P < 0.05 after ANOVA. Different letters (a to b) represent significant differences between groups.

Response of MdSnRK1 expression to sorbitol is signaling in nature, and sorbitol-stimulated expression of both MdSDH1 and MdA6PR is dependent on MdSnRK1 and MdbZIP39

We showed that the expression of MdSnRK1 was increased by sorbitol (Fig. 1B). To confirm that the sorbitol effect is signaling in nature while excluding the possibility of an osmotic effect, we conducted a sugar feeding experiment using 50 mM mannitol, fructose, glucose, and sucrose in addition to sorbitol, with mannitol as an osmotic control and fructose and glucose as equimolar carbon controls. Sorbitol feeding enhanced the expression of MdSnRK1, whereas mannitol or glucose did not; both fructose and sucrose decreased the expression of MdSnRK1 (Fig. 7A). The expression levels of both MdSDH1 and MdA6PR showed similar trends as MdSnRK1 in response to sugar feeding, whereas MdbZIP39 expression was not altered by any of the sugars tested (Fig. 7A). These data indicate that sorbitol acts as a signal in modulating the expression of MdSnRK1, MdSDH1, and MdA6PR.

Figure 7.

Stimulation of MdSDH1 and MdA6PR expression by sorbitol and its dependence on MdSnRK1 and MdbZIP39. A) Relative expression levels of MdSnRK1, MdbZIP39, MdA6PR, and MdSDH1 genes in shoots of ‘Royal Gala’ in response to 50 mM mannitol, fructose, sorbitol, glucose, or sucrose in the medium for 12 h, with no sugar as control. B) Relative expression levels of MdSnRK1, MdbZIP39, MdSDH1, and MdA6PR in shoots of ‘Royal Gala’ in response to OE or suppression (RNAi) of MdSnRK1 and MdbZIP39 alone or in combination via agroinfiltration for 3 d, followed by sorbitol feeding for 12 h. Empty vector agroinfiltration followed by no sugar, 50 mM mannitol, or 50 mM glucose feeding serves as additional controls. In both A) and B), data were means ± Sd (n = 5) with 6 shoots per replicate. Lsd was used for significant difference analysis at P < 0.05 after ANOVA. Different letters (a to d) represent significant differences between groups.

To confirm that the response of MdSDH1 and MdA6PR expression to sorbitol is dependent on the MdSnRK1-MdbZIP39 pathway, we transiently repressed the expression of MdSnRK1 or MdbZIP39 alone or combined RNA interference (RNAi) of either gene with OE of the other (MdSnRK1-RNAi, MdbZIP39-RNAi, MdSnRK1-OE + MdbZIP39-RNAi, and MdSnRK1-RNAi + MdbZIP39-OE) via vacuum infiltration of A. tumefaciens GV3101 harboring the respective vectors into in vitro shoots; 2 empty vector controls (RNAi vector and RNAi vector + OE vector) were included in the sorbitol feeding group to allow for detection of any treatment effect. Three additional controls (H₂O, mannitol, and glucose in combination with empty RNAi vector) without sorbitol feeding were also set up to provide reference points for the sorbitol feeding group to compare with. The shoot cultures of WT ‘Royal Gala’ were incubated under the corresponding sugar treatments for 12 h after 3 d following agroinfiltration. RT-qPCR analysis showed that sorbitol feeding increased the transcript levels of MdSnRK1, MdSDH1, and MdA6PR in both RNAi empty vector and double empty vector controls (Fig. 7B), confirming the stimulation effect of sorbitol on the expression of these genes. However, the sorbitol-stimulated expression of both MdSDH1 and MdA6PR was blocked by RNAi suppression of either MdSnRK1 or MdbZIP39. MdSnRK1-OE + MdbZIP39-RNAi led to similar expression levels of MdSDH1 and MdA6PR as in RNAi of either MdSnRK1 or MdbZIP39 alone, whereas MdSnRK1-RNAi + MdbZIP39-OE partially restored the expression levels of MdSDH1 and MdA6PR. These results indicate that sorbitol-stimulated expression of both MdSDH1 and MdA6PR is dependent on both MdSnRK1 and MdbZIP39.

Discussion

In earlier work, we found that SDH expression is downregulated in both shoot tips and growing fruit of antisense A6PR lines of ‘Greensleeves’ apple with decreased sorbitol synthesis (Cheng et al. 2005; Zhou et al. 2006; Li et al. 2018). Here, we show that SnRK1 expression is enhanced by sorbitol but decreased by sucrose. SnRK1 phosphorylates bZIP39, which binds to the promoters of both SDH1 and A6PR in activating their expression in response to sorbitol. These findings reveal that sorbitol acts as a signal regulating its own metabolism via the evolutionarily conserved SnRK1 sugar signaling pathway.

The initial clue for the involvement of SnRK1 in sorbitol signaling came from the finding that SDH was among the genes upregulated in transgenic Arabidopsis plants overexpressing KIN10, encoding a catalytic subunit of SnRK1 (Baena-González et al. 2007). The work presented here provides several lines of evidence supporting a role of SnRK1-mediated phosphorylation of bZIP39 in modulating the expression of 2 keys genes in sorbitol metabolism, SDH1 and A6PR, in response to sorbitol availability. First, the expression level of SnRK1, SDH1, and A6PR all increases in response to exogenous sorbitol feeding (Figs. 1B and 7A). Second, SnRK1 interacts and phosphorylates bZIP39 (Figs. 2 and 3). Third, bZIP39 binds to the promoters of SDH1 and A6PR as demonstrated by promoter binding assays both in vitro (Y1H and EMSAs) and in vivo (ChIP-qPCR), activating their expression (Figs. 4 and 5). Fourth, OE of SnRK1 in apple enhances the expression of SDH1 and A6PR without altering the expression of bZIP39 (Fig. 6). Finally, sorbitol-stimulated expression of SDH1 and A6PR is dependent on both SnRK1 and bZIP39 (Fig. 7). Based on these findings, we propose a model on how sorbitol modulates its own metabolism (Fig. 8). High availability of sorbitol stimulates the expression of SnRK1 in the nucleus. SnRK1 phosphorylates bZIP39, which binds to the promoters of both SDH1 and A6PR in activating their expression. How sorbitol regulates the expression of SnRK1 is unclear, but sorbitol-elicited SnRK1 expression allows sorbitol metabolism to respond to its own abundance and integrates sorbitol signaling into SnRK1-mediated sugar signaling network that modulates plant carbohydrate and energy metabolism.

Figure 8.

Model of MdSnRK1-mediated phosphorylation of MdbZIP39 as a mechanism for sorbitol modulation of its own metabolism. Sorbitol is synthesized via a 2-step process in leaves: reduction of G6P by A6PR to sorbitol-6-phosphate (Sor6P) and subsequent dephosphorylation of Sor6P by Sor6P phosphatase (SorPP) to sorbitol. Sorbitol is converted to fructose by SDH encoded by MdSDH1 in both source and sink tissues. Sorbitol acts as a signal regulating the expression of MdSnRK1, and subsequent phosphorylation of transcription factor MdbZIP39 by MdSnRK1 enhances its transcriptional activation activity on MdSDH1 and MdA6PR. As a result, both SDH and A6PR enzyme activities are upregulated to allow a higher carbon flux going through the sorbitol metabolic pathway. Circled “P” denotes protein phosphorylation. Solid arrow indicates direct regulation; dotted arrow represents indirect regulation.

The suppression of SnRK1 expression by sucrose observed in apple (Figs. 1B and 7A) is similar to that previously reported in other plants that use sucrose as a main transport sugar and synthesize only a small amount of sorbitol such as Arabidopsis (Baena-González et al. 2007; Nosarzewski et al. 2012; Aguayo et al. 2013). In both apple and Arabidopsis, OE of SnRK1 led to upregulation of SDH expression. These findings suggest that sorbitol catabolism via SDH is enhanced under low sucrose status in apple as well as in Arabidopsis, which is consistent with the SnRK1-mediated upregulation of catabolic processes to provide energy for plants under energy deficit (Baena-González and Sheen 2008; Wurzinger et al. 2018). Considering that sorbitol has more energy than glucose on an equimolar carbon basis due to its more reduced nature, remobilization of the energy stored in sorbitol is expected to be more effective in alleviating energy deficit than catabolism of glucose or sucrose. The inhibition of SnRK1 expression by sucrose (Figs. 1B and 7A) may also allow downregulation of sorbitol catabolism when sucrose accumulates to a higher level under low crop load situations (Klages et al. 2001). SnRK1 expression is also decreased by fructose (Fig. 7A), but the fructose concentration is not altered in the shoot tips of antisense A6PR lines (Zhou et al. 2006), and fruit shows remarkable fructose hemostasis in response to altered sorbitol/sucrose supply in the antisense A6PR lines (Li et al. 2018) and to source/sink manipulations (Klages et al. 2001). The low and unaltered fructose concentration in the shoot tips of SnRK1-OE lines (Fig. 6E) is consistent with that detected in the shoot tips of antisense A6PR lines (Zhou et al. 2006) and high expression levels of FRUCTOKINASE 2 encoding a high-affinity fructokinase in apple shoot tips (Yang et al. 2018).

As the expression of SnRK1 responds to sorbitol in the opposite direction of that to sucrose (Figs. 1B and 7A), it raises an interesting question as to the role of SnRK1 in energy signaling in plant species that synthesize, transport, and utilize sorbitol as a major carbon and energy source (Loescher 1987; Cheng et al. 2005). In these species, sorbitol is ideally positioned to represent the plant carbon and energy status as its concentration in both source and sink organs responds to crop load (Klages et al. 2001; Morandi et al. 2008) in addition to being a more energy-rich carbohydrate than sucrose and hexoses on an equimolar carbon basis. Indeed, the partially aborted stamen development and reduced pollen germination and tube growth observed in transgenic ‘Greensleeves’ apple with decreased sorbitol levels (Meng, He, et al. 2018; Li, Meng, et al. 2020) and the stimulation of sorbitol on flower bud formation in loquat (Xu et al. 2022) are consistent with sorbitol being a signal of tree energy status regulating these developmental processes. Viewed this way, the SnRK1-mediated upregulation of SDH1 expression in response to sorbitol renders an additional role to this highly conserved sugar signaling pathway; i.e. it allows sorbitol metabolism to respond to the energy status represented by the sorbitol level in apple. High sorbitol availability leads to upregulation of A6PR expression as well as SDH1 expression by SnRK1 via phosphorylation of bZIP39 (Figs. 3 to 7). The simultaneous upregulation of A6PR and SDH1 enhances carbon flux through the sorbitol metabolic pathway for plant growth and development under abundant sorbitol supply, which represents a unique feature of SnRK1-mediated sugar/energy signaling in sorbitol-transporting species. It is unclear how the apple SnRK1 evolved to sense sorbitol and sucrose in opposite directions, but it could be related to the inverse relationship between the 2. In antisense A6PR lines of ‘Greensleeves’ apple with decreased sorbitol synthesis, sucrose accumulates to a higher level in both mature leaves and shoot tips, corresponding to lower steady-state sorbitol levels (Cheng et al. 2005). The regulation of A6PR and SDH1 expression by SnRK1-mediated phosphorylation of bZIP39 may also provide a mechanism for downregulation of sorbitol metabolism via the suppression effect of sucrose accumulation on SnRK1 expression (Figs. 1B and 7A) under sink limitation such as low crop load (Klages et al. 2001).

The only situation in plant carbohydrate metabolism that is analogous to sorbitol-modulated own catabolism described here is the regulation of SuSy by sucrose. Sucrose induces the expression of SuSy4 and the gene encoding ADP-glucose pyrophosphorylase (AGPase) B subunit (AGPB) in potato (Solanum tuberosum) tuber (Muller-Rober et al. 1990; Fu and Park 1995). However, the inducibility of SuSy expression by sucrose is abolished in transgenic potato leaves with RNAi repression of SnRK1 expression (Purcell et al. 1998), indicating that SnRK1 plays an essential role in regulating SuSy expression. Overexpression of SnRK1 in potato tubers enhances the expression of both SuSy and AGPB genes (McKibbin et al. 2006). These findings suggest a positive role of SnRK1 in upregulating the expression of SuSy and AGPase for starch synthesis in response to high sucrose availability (Geigenberger 2003; McKibbin et al. 2006). However, sucrose represses SnRK1 expression and lowers SnRK1 activity via the elevated T6P level as shown in Arabidopsis (Baena-González et al. 2007; Zhang et al. 2009). In apple where both sorbitol and sucrose are synthesized in leaves and transported in the phloem, sucrose enhances its own catabolism by upregulating SuSy expression and its activity in shoot tips (Zhou et al. 2006) but represses SnRK1 expression (Figs. 1B and 7A). Similar to what was detected in potato tubers, OE of SnRK1 in apple leads to upregulation of the expression of SuSy genes in the shoot tips (Supplemental Table S5). High sucrose availability triggers upregulation of SuSy expression via SnRK1, but how SnRK1 works in the process remains to be uncovered to reconcile its seemingly conflicting roles in modulating SuSy expression in sucrose depletion and abundance situations.

Sorbitol accumulates as a compatible solute in apple and other Rosaceae tree fruits and Plantago species in response to abiotic stress such as drought/osmotic stress (Wang and Stutte 1992; Lo Bianco et al. 2000; Zhang et al. 2011), salinity (Königshofer 1983; Smekens and Tienderen 2001; Al Hassan et al. 2016), and low temperature (Busatto et al. 2018), as mannitol does in celery (Apium graveolens) under salinity (Everard et al. 1994; Williamson et al. 2002). This is accompanied by increases in ABA levels (Robinson and Barritt 1990; Hezema et al. 2021; Lee et al. 2021). To what extent sorbitol regulates its own metabolism and whether sorbitol is perceived differently under abiotic stress are not known. Considering that suppression of sorbitol synthesis leads to downregulation of a large number of stress response genes including ABA synthesis (Wu et al. 2015) and SnRK1 plays a key role in both sugar and ABA signaling (Jossier et al. 2009; Cutler et al. 2010), it is likely that the SnRK1-mediated sorbitol signaling interacts with other players in the SnRK1 pathway and the ABA signaling pathway under abiotic stress. In growing rice seedlings, SnRK1A-interacting negative regulators (SKINs) counteract sugar/energy-deficit activated SnRK1A signaling (Lin et al. 2014). The expression of SKINs increases dramatically in response to drought and moderately to salinity, cold, and hypoxia, and ABA promotes the binding of SKINs to SnRK1A in the cytoplasm to prevent SnRK1A and its phosphorylation target, MYBS1, from entering the nucleus. This reduces transcriptional activation of α-AMYLASE and other genes involved in mobilization of starch and other nutrients stored in the endosperm, leading to inhibition of seedling growth. Homologs of rice SKINs are among the proteins found in the interactome of SnRK1 in Arabidopsis (Van Leene et al. 2022), suggesting that the interaction between SKINs and SnRK1 operates in dicots as well. In addition, Type II T6P synthase-like proteins are found to be an important group of negative regulators of SnRK1 under stress conditions (Van Leene et al. 2022). The SnRK1 signaling pathway and the ABA signaling pathway also share a large number of proteins including those in core ABA signaling and ABA/abiotic stress response (Carianopol et al. 2020). It appears that upregulation of the expression levels and activities of both A6PR and SDH in short-term drought stress (Yang et al. 2019) is still consistent with SnRK1-mediated sorbitol signaling. However, downregulation of SDH expression and activity to enhance sorbitol accumulation during long-term drought/osmotic stress (Lo Bianco et al. 2000; Zhang et al. 2011) is likely related to the inhibition of ABA on SnRK1-mediated sorbitol signaling. In addition, ABA, drought/osmotic stress, and low temperature may enhance the expression of A6PR via the corresponding cis-acting regulatory elements in its promoter (Kanayama et al. 2006; Liang et al. 2012; Zhang et al. 2011; Busatto et al. 2018).

The expression of SDH1 in apple is enhanced by SnRK1-mediated phosphorylation of bZIP39 in response to sorbitol in this study. Upregulation of SDH expression in Arabidopsis by OE of SnRK1 (Baena-González et al. 2007) indicates that its expression is dependent on SnRK1, which is corroborated by downregulation of SDH expression in Arabidopsis via inhibition of SnRK1 activity by the elevated T6P level resulting from OE of E. coli ots A (Schluepmann et al. 2003; Zhang et al. 2009). All 42 angiosperm species including representatives of monocots and eudicots have at least 1 SDH gene (Jia et al. 2015), suggesting the ubiquitous presence of this gene and its functional importance in plants. The SDHs in core eudicots fall into 2 classes, Class I and Class II, with Class I SDHs in apple (SDH2-15) substantially expanded as a result of tandem duplication and whole genome duplication (Velasco et al. 2010; Jia et al. 2015). It is interesting that the SDH gene in Arabidopsis (Class I) and SDH1 gene in apple (Class II) belong to 2 different classes, but both are regulated by SnRK1, suggesting that the SnRK1 signaling pathway plays an important role in regulating SDH expression in all sorbitol-synthesizing plants regardless of its abundance level. The expression of SDH2 gene in apple shoot tips is stimulated by exogenous sorbitol as well (Fig. 1B), but OE of SnRK1 does not alter its expression (Fig. 6B). This indicates that sorbitol-modulated SDH2 expression is not dependent on SnRK1. How the expression of SDH2 and possibly other SDH genes is regulated by sorbitol in apple warrants further research.

Conclusion

In addition to sensing plant sucrose status, SnRK1 modulates sorbitol metabolism via phosphorylating transcription factor MYB73 in response to sorbitol, a major photosynthate and transport carbohydrate in apple and other tree fruits in the Rosaceae family. Apple SnRK1 expression is stimulated by sorbitol, and subsequent phosphorylation of MYB73 by SnRK1 enhances the expression of both SDH1 and A6PR encoding 2 key enzymes in sorbitol metabolism. This allows sorbitol metabolism to respond to the tree energy status represented by sorbitol and integrates sorbitol signaling into the evolutionarily conserved SnRK1-mediated sugar signaling network to modulate carbohydrate metabolism.

Materials and methods

Plant materials and growth conditions

Subcultured in vitro shoots of WT ‘Greensleeves’ apple (M. domestica) and 2 antisense A6PR lines (A4 and A10) with about 10% of the WT transcript levels (Cheng et al. 2005; Wu et al. 2015) were grown on MS (Murashige and Skoog 1962) with 30 g L⁻¹ sucrose, 0.3 mg L⁻¹ 6-benzylaminopurine (6-BA), and 0.03 mg L⁻¹ indole-3-butyric acid (IBA) under a 16 h photoperiod at 23 °C for leaf sugar feeding experiment.

Two-year-old, self-rooted WT, and 2 MdSnRK1-OE lines (OE-24 and OE-46) of ‘Royal Gala’ apple trees were grown in 8 L pots containing medium comprising 1 part sand and 2 parts MetroMix360 (v/v) (Scotts) under natural conditions at Cornell Experimental Orchards in Ithaca, NY. Each genotype was replicated 5 times with 5 trees per replicate in a completely randomized design. The trees received standard horticulture and disease and pest management. During active soot growth in late June, shoot tips with expanding young leaves (top 3 cm) were taken and frozen in liquid nitrogen and stored at −80 °C for extraction of RNA and sugars.

Subcultured WT of ‘Royal Gala’ apple grown on MS medium with 30 g L⁻¹ sucrose, 0.3 mg L⁻¹ 6-BA, and 0.03 mg L⁻¹ IBA at ∼50 µmol photons m⁻² s⁻¹ under a 16 h photoperiod at 23 °C was used for shoot sugar feeding and transient gene expression experiments.

Apple calli derived from ‘Orin’ apple used for ChIP-qPCR were cultured on MS medium with 30 g L⁻¹ sucrose, 1.5 mg L⁻¹ 6-BA, and 0.5 mg L⁻¹ indole-3-acetic acid at 25 °C in the dark.

Nicotiana benthamiana plants for subcellular localization of MdbZIP39 and agroinfiltration were grown in a phytotron under a 16 h photoperiod at 24 °C.

Transformation of ‘Royal Gala’ apple plants

The coding sequence of MdSnRK1 cDNA without stop codon was cloned into the T-vector using XbaI and BamHI restriction sites, and the entire cassette was excised and ligated into pBI121 fused with GFP using XbaI and BamHI restriction sites to make OE vector pBI121-SnRK1-GFP (primers listed in Supplemental Table S6). The construct was transformed into A. tumefaciens strain EHA105 with an additional virulence plasmid pCH32 (Hood et al. 1993). ‘Royal Gala’ was transformed as previously described (Norelli et al. 1994; Borejsza-Wysocka et al. 1999). Briefly, in vitro shoots of ‘Royal Gala’ were subcultured at 4-wk intervals on proliferation medium (Murashige and Skoog 1962) and were then transferred to leaf expansion medium [MS supplemented with 0.1 mg L⁻¹ naphthyleneacetic acid (NAA) and 8.3 mg L⁻¹ 6-(γ,γ-dimethylallylamino) Purine (2iP)] for 4 wk. After the youngest unfolded leaves were excised from in vitro shoots and wounded with a pair of nontraumatic forceps (Norelli et al. 1996), they were co-cultivated with the EHA105 (pCH32) harboring the binary vector pBI121-SnRK1-GFP. Following co-cultivation, the leaves were washed with distilled water containing half-strength MS salt mixture and cefotaxime sodium (500 mg L⁻¹) to remove excess inoculum and placed on regeneration medium containing 100 mg L⁻¹ kanamycin. The regenerated shoots were propagated in vitro, and the expression of SnRK1 was determined via RT-qPCR using RNA extracted from each line.

Exogenous sugar feeding

Leaves from in vitro shoot cultures of WT and 2 antisense A6PR lines (A4 and A10) of ‘Greensleeves’ apple were fed with exogenous sugars via the transpiration stream (Meng, He, et al. 2018; Meng, Li, et al. 2018). Briefly, the youngest fully developed leaves were excised from the shoots with a razor blade, floated on water, and then assigned to 50 mM sorbitol, 50 mM sucrose, or water control. Each treatment was replicated 3 times with 10 leaves from 5 shoots per replicate in a petri dish, where only the petioles were sandwiched between 2 pieces of filter paper (1.5 cm in width) soaked with either sugar or water. The petri dishes were capped and randomly arranged in a fume hood under fluorescent lights at ∼50 µmol photons m⁻² s⁻¹ at 23 °C. Leaf samples were collected at 0, 1, and 3 h after initiation of sugar feeding, frozen in liquid nitrogen, and stored at −80 °C for use.

Subcultured in vitro shoots of ‘Royal Gala’ were placed in MS medium containing 50 mM mannitol, 50 mM sorbitol, 50 mM fructose, 50 mM glucose, 50 mM sucrose, or no sugar control and cultured in an artificial climate room under fluorescent lights at ∼50 µmol photons m⁻² s⁻¹ at 23 °C. Each treatment was replicated 5 times with 6 shoots per replicate in a completely randomized design. After 12 h, the shoots were collected and frozen in liquid nitrogen and stored at −80 °C for use.

Extraction and analysis of leaf sugars

Soluble sugars were extracted from ∼100 mg of tissue in 1.4 mL 75% (v/v) methanol, with ribitol added as internal standard. After fractionation of the nonpolar metabolites into chloroform, 5 µL polar phase of each sample was dried under vacuum without heat and then derivatized sequentially with methoxamine hydrochloride and N-methyl-N-trimethylsilyltrifluoroacetamide. The metabolites were then analyzed using Agilent 7890 GC/5975C MS (Agilent Technology, Palo Alto, CA, USA) (Wang et al. 2010).

Extraction and assay of SDH and A6PR enzyme activities

Enzymes were extracted from shoot tips following Lo Bianco et al. (1998) as modified by Zhou et al. (2006). Briefly, each fresh sample was rapidly chopped to small pieces and ground with a precooled mortar and pestle in 4- mL extraction buffer containing 100 mM Tris–HCl (pH 9.0), 8% (v/v) glycerol, 1% (w/v) insoluble polyvinylpolypyrrolidone (PVPP), and acid-washed sand. The homogenate was centrifuged at 16,000 × g for 10 min at 2 °C, and 1 mL of the supernatant was desalted with a PD-10 column (Amersham BioSciences, Piscataway, NJ, USA). The SDH enzyme activity was measured spectrophotometrically as a rate of increase in absorbance at 340 nm as a result of NADH production (Lo Bianco et al. 1998). The activity of A6PR was assayed as a rate of decrease in absorbance at 340 nm due to consumption of NADPH (Negm and Loescher 1981; Cheng et al. 2005).

RNA extraction and RT-qPCR

Total RNA was extracted using the CTAB method (Gasic et al. 2004). After treatment with RQ1 DNase (Promega), RNA was quantified using a NanoDrop spectrophotometer, and RNA integrity was confirmed by agarose gel electrophoresis. One microgram of total RNA was reverse transcribed to cDNA using an iScript cDNA Synthesis kit (Bio-Rad, USA). Detection of gene transcript levels was performed in triplicate with gene-specific primers (Supplemental Table S6) on an ABI QuantStudio 6 Flex (Thermo Fisher, USA) with MagicSYBR Mixture (Kangwei Shiji, China). The relative expression of genes was analyzed using 2^−ΔΔCT method, with ACTIN as an internal reference gene.

Gene cloning and construction of expression vectors

Coding sequences of MdSnRK1 and MdbZIP39 without stop codon were separately inserted into vector pROK2, with GFP fused to the 3′ end of each gene. A 200 bp cDNA fragment specific to MdSnRK1 and a 225 bp fragment specific to MdbZIP39 were inserted into the MCS1 region of the NcoI/SwaI restriction site and the MCS2 region of the Xbal/BamHI restriction site of pFGC5941, respectively, to generate RNAi vectors. Plant OE and RNAi vectors were constructed with ClonExpress Ultra One Step Cloning Kit (Vazyme Biotech Co., Ltd, Nanjing, China). Primers are listed in Supplemental Table S6.

Transient transformation of in vitro ‘Royal Gala’ shoots via vacuum infiltration

Overexpression and RNAi vectors of MdSnRK1 and MdbZIP39 were separately introduced into A. tumefaciens GV3101 and cultured in liquid medium containing 25 mg L⁻¹ rifampicin and 50 mg L⁻¹ kanamycin for 12 to 15 h. The liquid culture was spun at 5,000 rpm for 10 min, and the precipitate was resuspended in MS liquid medium (pH 5.8). Actively growing in vitro shoots of ‘Royal Gala’ were immersed in the agrobacterium solution and placed in a vacuum suction vessel at a pressure of 50 kPa for 8 min. After vacuum release, the in vitro shoots were washed with water and transferred to 1/5 Hoagland nutrient solution for 3 d for further analysis.

Subcellular localization of MdbZIP39

The coding sequence of MdbZIP39 was cloned into a vector with a C-terminal RFP tag (primers listed in Supplemental Table S6), and the recombined plasmid was transformed into ‘Royal Gala’ apple leaves according to Hu et al. (2016) with modifications. Fluorescence images were obtained using a confocal laser scanning microscope (Leica TCS SP8, Germany). The RFP signal was excited by 532-nm laser line and was detected at 590 nm (bandwidth 20 nm). The GFP signal was detected at 500 to 530 nm after excitation at 488 nm.

Protein purification

MdSnRK1 and MdbZIP39 were inserted into the pGEX-4 T and pRSFDuet⁻¹ vectors, respectively. The resulting constructs were introduced into the BL21 (DE3) E. coli (TransGen Biotech, China, https://www.transgenbiotech.com). The recombinant proteins, GST-MdSnRK1 and His-MdbZIP39, were purified using glutathione agarose and Ni-NTA agarose, respectively.

Assays of SnRK1 kinase activity

SnRK1 kinase activity was determined using the Promega Kinase-Glo Luminescent Kinase Assay Platform (product number: part of #TB372). We tested 3 forms of bZIP39 protein as substrates for SnRK1: His-MdbZIP39, His-MdbZIP39^S41A that has the serine at position 41 substituted with alanine based on the predicted phosphorylation sites using NetPhos-3.1 (https://services.healthtech.dtu.dk/service.php? NetPhos-3.1) and the DNA-binding site in bZIP39 according to the protein annotation result on the NCBI website (https://www.ncbi.nlm.nih.gov/protein/XP_008377201.2) (Supplemental Fig. S2), and His-MdbZIP39X that has been deactivated by boiling at 100 °C for 10 min. There are a total of 10 groups: in addition to 3 controls (buffer only, His, or GST), 3 forms of MdbZIP39 (His-MdbZIP39, His-MdbZIP39^S41A, and His-MdbZIP39X) were either alone or combined with GST-MdSnRK1. Each group was replicated 3 times in a randomized design. The phosphorylation reaction (0.1 mM fluorescently labeled ATP, 10 mM MgCl₂, 50 mM Tris–HCl pH 7.5, 1 mM DTT, and 1 mM PMSF with 30 µg substrate and 10 µg kinase) was measured in duplicate in a 60 µL reaction system at 30 °C as a decrease in the RFU for 1.5 h.

SnRK1 kinase activity was also detected by using phospho-Ser/Thr antibody on the 3 forms of bZIP39 protein (His-MdbZIP39, His-MdbZIP39^S41A, and His-MdbZIP39X). They were individually mixed with GST-MdSnRK1 in a 60 µL reaction system (1 mM ATP, 10 mM MgCl₂, 50 mM Tris–HCl pH 7.5, 1 mM DTT, and 1 mM PMSF with 30 µg of the substrate and 10 µg kinase) and incubated at 30 °C for 2 h. The protein mixture was then separated using SDS–PAGE. Phosphorylation activity was detected with the phospho-(Ser/Thr) Phe antibody (product number: 9631, Cell Signaling Technology, USA).

SnRK1 kinase activity from the shoot tips of ‘Royal Gala’ plants was measured as previously described (Zhang et al. 2009; Yu et al. 2021). Briefly, 1 g plant tissue was ground in liquid nitrogen for protein extraction, and the homogenate was centrifuged at 13,000 × g at 4 °C, and the supernatant was spin desalted (2.5 mL Sephadex G-25 column, GE Healthcare). SnRK1 activity was measured with AMARA peptides as substrates by using Promega Kinase-Glo Luminescent Kinase Assay Platform (product number: part of #TB372).

Y1H library screening and Y1H assay

The promoter fragments of MdSDH1 (proMdSDH1) and MdSDH2 (proMdSDH2) were inserted into the pHis2 vector (primers listed in Supplemental Table S6). Total RNA was extracted from shoot tips for cDNA synthesis. The pHis2-proMdSDH1 and pHis2-proMdSDH2 plasmids were transformed into Y187 yeast (S. cerevisiae) with cDNA and pDADT7-Rec vectors, respectively, for intracellular homologous recombination to screen the cDNA library. Yeast culture and screening of positive clones are the same as described below for Y1H assay, and the obtained positive clones were identified by sequencing.

The Y1H assay was performed using the Y187-pHis2 Yeast One-Hybrid Media Kit (China, www.bjbalb.com) following the manufacturer's instructions. The MdbZIP39 gene was fused in-frame with the GAL4 AD in a pGADT7 vector (prey plasmid); the promoter fragments of MdA6PR (proMdA6PR) and proMdSDH1 were ligated to the pHis2 vector (bait vector) using primers listed in Supplemental Table S6. The pHis2-proMdA6PR and pHis2-proMdSDH1 were transformed into Y187 yeast strain and selected on an SD/-Trp plate with 3-AT. The pGADT7-MdbZIP39 constructs were transformed into strain Y187 harboring pHis2-bait and selected on an SD/-His-Leu/3-AT plate (60 mM 3-AT). To confirm the results, positive clones (co-transformed) were spotted in a series of dilutions (1:1, 1:10, and 1:100) and cultured on the SD/-His-Leu/3-AT medium at 30 °C for 4 d. pGADT7-p53 + pHis2-p53 was used as a positive control and pGADT7 + pHis2-p53 as a negative control.

Y2H assay

For the Y2H assay, the cDNAs of protein fragments 6D8, 3D3, 5C8, 5D8, 4F3, 51A1, 5D3, and 1A4 were cloned into the pGADT7 vector as a prey, and MdSnRK1 was cloned into the pGBKT7 vector as a bait (primers listed in Supplemental Table S6). The pair, pGADT7-p53 and pGBKT7-SV40, was used as a positive control, whereas the pGADT7 and pGBKT7 empty vectors were used as negative controls. Various combinations of activation domain (AD) and binding domain (BD) vectors were transformed into the yeast strain AH109. After growth on SD/-Leu-Trp medium for 4 to 6 d at 30 °C, the clones were transferred into the selective medium (SD/-Leu-Trp-His-Ade) at 30 °C for 3 to 4 d. The positive clones (co-transformed) were spotted in a series of dilutions (1:10, 1:100, and 1:1000). At least 3 independent experiments were performed, and the results from a single representative experiment were presented.

Immunoblotting

Proteins were extracted from shoot tips of ‘Royal Gala’ in the RIPA Lysis Buffer (Beyotime, product number: P0013B), and immunoblotting was performed using Sigma anti-GFP and anti-AtKIN10. Briefly, 0.1 g shoot tip tissue was ground in liquid nitrogen and extracted for total proteins in 500 µL RIPA Lysis Buffer. The proteins were separated by SDS–PAGE and then transferred to polyvinylidene fluoride (PVDF) membrane. The membrane was sealed with 5% (w/v) semi-skim milk powder at room temperature and incubated with the corresponding primary and secondary antibodies. Immunoblot mages were obtained with a ChemiDoc Touch Chemiluminescence Imaging System (Bio-Rad, Hercules, USA).

Electrophoretic mobility shift assays

The MdbZIP39 transcription factor was inserted into the pRSFDuet⁻¹ vector and transformed into BL21 (DE3) E. coli, and the His-MdbZIP39 fusion protein was obtained by prokaryotic expression and purification. The purified fusion protein was concentrated in ultrafiltration tubes and replaced with 25 mM Tris–HCl (pH 8.0) buffer containing 1× Complete Protease Inhibitor Cocktail (Sigma). MdSDH1 and MdA6PR biotin-labeled probes, cold competition probes, and mutant probes (Supplemental Table S2) were synthesized by Rui Biotechnology Company (Beijing, China), and the probes were mixed in equal amounts, incubated at 95 °C, and gradually cooled to 25 °C for annealing. EMSA experiments were performed using Light Shift Chemiluminescent EMSA Kit (Cat: 20148, Thermo, Waltham, USA).

BiFC assay

Coding sequences of MdSnRK1 and MdbZIP39 were inserted into the 35S-pSPYNE and 35S-pSPYCE vectors, respectively, to obtain the fusion constructs MdSnRK1-YFPn and MdbZIP39-YFPc (primers listed in Supplemental Table S6). The plasmids were transformed into A. tumefaciens GV3101, and infiltration of N. benthamiana leaves was conducted as previously described (Meng, He, et al. 2018; Meng, Li, et al. 2018). After 48 h incubation in the dark, the YFP signal was detected at 520 to 550 nm after excitation at 505 nm using a confocal laser scanning microscope (Leica TCS SP8). Each BiFC assay was repeated at least 3 times.

Apple callus transformation and ChIP-qPCR assay

Both pROK2-MdbZIP39-RFP and pROK2-RFP were introduced into A. tumefaciens GV3101. Transformation of ‘Orin’ apple calli was performed as previously described (Li, Dougherty, et al. 2020). Briefly, 3-wk-old calli were incubated with the agrobacterium harboring either plasmid in liquid medium with shaking at 150 rpm for 15 min at 25 °C and then transferred to solid MS medium and co-cultured for 2 to 4 d. After washed with sterile water for 3 to 4 times, the transformed calli were transferred to a selection medium containing 30 mg L⁻¹ kanamycin and 250 mg L⁻¹ carbenicillin, followed by another round of selection after 3 to 4 wk. The transformed calli were then confirmed by PCR. The selected transgenic calli were cultured for 3 wk, and samples were taken for ChIP-qPCR.

ChIP-qPCR assay was performed as previously described (Meng, He, et al. 2018; Meng, Li, et al. 2018). The ‘Orin’ transgenic calli expressing MdbZIP39-RFP or RFP (control) were ground into fine powder in liquid nitrogen and suspended in 10 mL nucleus separation buffer [10 mM HEPES pH 8.0, 1 M sucrose, 5 mM KCl, and 5 mM EDTA, 0.6% (v/v) Triton X-100, 0.4 mM PMSF, and fresh protease inhibitor cocktail], with 1% (v/v) formaldehyde added for cross linking. The nuclei were separated by centrifugation and then resuspended in a nucleus lysis buffer. Chromatin was fragmented to an average size of 100 to 500 bp with Bioruptor (JY92-IIN, Scientz, China) and immunoprecipitated with RFP antibody (RFP Antibody MA5-15257, Thermo Fisher, USA, https://www.thermofisher.cn/). After elution, the protein–DNA complex was reverse cross-linked, and the enriched DNA was purified for qPCR analysis using the primers listed in Supplemental Table S6.

Transcriptome sequencing

mRNA was isolated from 5 µg of total RNA extracted from shoot tips using NEBNext Poly(A) mRNA Magnetic Isolation Module kit (New England Biolabs, Ipswich, MA, USA). A total of 15 cDNA libraries were constructed for RNAseq, with 5 biological replicates each for WT, OE-24, and OE-S46, as previously described (Meng, He, et al. 2018; Meng, Li, et al. 2018). The libraries were sequenced in 50 bp single-end sequencing on an Illumina HiSeq 4000 sequencer (Illumina Inc, San Diego, CA, USA) at the CLC Genomics and Epigenomics Core Facility at Cornell University. The total RNAseq reads were processed, and the cleaned reads were aligned to the Golden Delicious double-haploid genome as described previously (Meng, He, et al. 2018; Meng, Li, et al. 2018). After alignments, raw counts of mapped reads for each apple gene model were derived and then normalized to the reads per kilobase of exon model per million mapped reads (RPKM). DEGs between the MdSnRK1-OE lines and WT were identified via an R package DEseq (Anders and Huber 2010) using cutoff criteria with OE/WT ratio > 1, WT RPKM value > 1, P < 0.05, for upregulated genes (Supplemental Table S3), and OE/WT ratio < 1, WT RPKM value > 1, P < 0.05, for downregulated genes (Supplemental Table S4).

Statistical analysis

Statistical analysis was performed via software SPSS Statistics 26.0 (IBM). Least significant difference (Lsd) was used for significant difference at P < 0.05 after analysis of variance (ANOVA).

Accession numbers

For apple genes (https://www.rosaceae.org/) in M. × domestica genome (GDDH13 V1.1): MdSnRK1 (MD09G1056200), MdbZIP39 (MD08G1025800), MdSDH1 (MD01G1110100), MdSDH2 (MD01G1195200), MdA6PR (MD10G1062300), and MdACTIN (MD13G1153100).

Author contributions

L.C. and D.M. designed the research; D.M., H.C., Q.Y., M.Z., E.B.-W., H.W., and L.C. performed the experiments. H.C., D.M., and Z.F. analyzed the data. A.D. contributed to plant material. L.C., D.M., and H.C. wrote the manuscript with inputs from all other authors.

Supplemental data

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

Supplemental Figure S1. Phylogenetic analysis and protein domain analysis of MdSnRK1.

Supplemental Figure S2. Prediction of phosphorylation sites in MdbZIP39.

Supplemental Figure S3. The relative expression levels of 4 MdA6PR homologs in shoot tips of WT, OE-24, and OE-46, detected via RT-qPCR.

Supplemental Table S1. List of transcription factors obtained in Y1H library screening.

Supplemental Table S2. List of DEGs in MdSnRK1-OE lines (OE-24 and OE-46) versus WT.

Supplemental Table S3. List of genes upregulated by MdSnRK1-OE.

Supplemental Table S4. List of genes downregulated by MdSnRK1-OE.

Supplemental Table S5. List of SuSy genes upregulated by MdSnRK1-OE.

Supplemental Table S6. List of primers used in this study

Funding

The work was support in part by USDA National Institute of Food and Agriculture – Specialty Crop Research Initiative project 2016-51181-25406 and National Natural Science Foundation of China No. 31922058.

Data availability

Data supporting the findings of this study are available from the corresponding author upon reasonable request.