Subcellular localization and vacuolar targeting of sorbitol dehydrogenase in apple seed

Plant Science 210 (2013) 36–45

Contents lists available at SciVerse ScienceDirect

Plant Science journal homepage: www.elsevier.com/locate/plantsci

Subcellular localization and vacuolar targeting of sorbitol dehydrogenase in apple seed Xiu-Ling Wang a,∗ , Zi-Ying Hu a , Chun-Xiang You b , Xiu-Zhen Kong a , Xiao-Pu Shi a a b

State Key Laboratory of Crop Biology, College of Life Sciences, Shandong Agricultural University, Taian 271018, China State Key Laboratory of Crop Biology, College of Horticulture Science and Engineering, Shandong Agricultural University, Taian 271018, China

a r t i c l e

i n f o

Article history: Received 6 March 2013 Received in revised form 21 April 2013 Accepted 24 April 2013 Available online 10 May 2013 Keywords: Sorbitol dehydrogenase Apple cotyledon Subcellular localization Vacuolar targeting Internal signal peptide

a b s t r a c t Sorbitol is the primary photosynthate and translocated carbohydrate in fruit trees of the Rosaceae family. NAD+ -dependent sorbitol dehydrogenase (NAD-SDH, EC 1.1.1.14), which mainly catalyzes the oxidation of sorbitol to fructose, plays a key role in regulating sink strength in apple. In this study, we found that apple NAD-SDH was ubiquitously distributed in epidermis, parenchyma, and vascular bundle in developing cotyledon. NAD-SDH was localized in the cytosol, the membranes of endoplasmic reticulum and vesicles, and the vacuolar lumen in the cotyledon at the middle stage of seed development. In contrast, NAD-SDH was mainly distributed in the protein storage vacuoles in cotyledon at the late stage of seed development. Sequence analysis revealed there is a putative signal peptide (SP), also being predicated to be a transmembrane domain, in the middle of proteins of apple NAD-SDH isoforms. To investigate whether the putative internal SP functions in the vacuolar targeting of NAD-SDH, we analyzed the localization of the SP-deletion mutants of MdSDH5 and MdSDH6 (two NAD-SDH isoforms in apple) by the transient expression system in Arabidopsis protoplasts. MdSDH5 and MdSDH6 were not localized in the vacuoles after their SPs were deleted, suggesting the internal SP functions in the vacuolar targeting of apple NAD-SDH © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Sorbitol is a major photosynthetic product and phloemtranslocated component in the woody Rosaceae, important crops like apple (Malus domestica Borkh.), pear (Pyrus communis L.), and peach (Prunus persica (L.) Batsch) [1–3]. Sorbitol accounts for about 80% of the total soluble carbohydrate in apple leaves. In Rosaceae plants, sorbitol is synthesized in source organs from glucose-6-phosphate to sorbitol-6-phosphate catalyzed by the NADP-dependent sorbitol-6-phosphate dehydrogenase (S6PDH, EC 1.1.1.200) (also named aldose-6-P-reductase, A6PR) [4,5]. Sorbitol-6-phosphate is further converted to sorbitol by sorbitol6-phosphate phosphatase (SorPP, EC 3.1.3.50) [6]. It has been reported that NADP-dependent sorbitol-6-phosphate dehydrogenase was localized in chloroplasts and the cytosol of protoplasts from apple cotyledons, which means sorbitol was synthesized in the chloroplast and cytosol [7,8]. Sorbitol was detected in the cytosol, chloroplasts, and vacuoles of peach leaves [9]. Sorbitol concentration in cytosol was similar to that in the vacuole, and the highest level of sorbitol was found in the chloroplast. However,

∗ Corresponding author. Tel.: +86 05388247826. E-mail address: [email protected] (X.-L. Wang). 0168-9452/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.plantsci.2013.04.008

most of the sorbitol was in the vacuole because of the larger volume of the vacuole (occupying 68% of the total mesophyll cells) [9]. Sorbitol is converted to fructose mainly catalyzed by NAD+ dependent sorbitol dehydrogenase (NAD-SDH, EC 1.1.1.14) in apple [10–12]. Sorbitol accounts for about 80% of the total soluble carbohydrate in apple leaves, but only 3–8% of that accumulated in the fruit. Sorbitol imported into apple fruit is not stored but metabolized rapidly after unloading by NAD-SDH [13]. More than ten full-length cDNAs of NAD-SDH have been isolated from apple, including MdSDH5 and MdSDH6, and their deduced amino acid sequences have several conserved domains, such as a zinccontaining alcohol dehydrogenase signature, a structural zinc binding site, a NAD-binding pocket, and a catalytic zinc binding site [13–16]. The expression pattern of MdSDH (NAD-SDH in Malus domestica) genes is tissue specific and developmental stage-dependent [14,15]. The expressions of MdSDH2, MdSDH3 and MdSDH4 are restricted in sink tissues, such as young leaves, stems, roots, and fruits, but MdSDH1 is highly expressed in source organs (mature leaves) besides sink tissues [14]. In apple fruit, SDH1 and SDH3 are expressed in both seed and cortex tissues, while the expression of SDH2 is limited to cortex, and SDH6 and SDH9 are only expressed in seed [15]. NAD-SDH protein is ubiquitously distributed in the tissues of apple leaf and fruit [16]. Moreover, the subcellular localizations of NAD-SDH in the fruits and leaves are multiple and different. In apple

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

leaves where sorbitol is synthesized, NAD-SDH is mainly localized in the cytosol, chloroplasts, and vacuoles. But in the fruit flesh in which sorbitol is utilized, NAD-SDH is localized in the cytosol and chloroplasts, and not in the vacuoles [16]. The different and multiple intracellular localizations of NAD-SDH in different tissues and organs indicate that this enzyme may be very important in regulation sorbitol metabolism to keep the osmotic balance in cells, for sorbitol is an osmotic solute. However, little is currently known about the mechanisms of the targeting and translocation of NADSDH. NAD-SDH activity in the seeds was found to be much higher than that in the cortex tissues of fruits during the early developmental stages of apple [15]. It suggests that this enzyme plays a more important role in sorbitol metabolism in the seed development. Whetter and Taper reported that sorbitol is present in apple seeds at the initial germinating stages. But it is unknown where the functional site of NAD-SDH is in cotyledon cells and whether the sorbitol is accumulated in seed for seed germination [17]. In our current study, we investigated the distribution and subcellular localization of NAD-SDH in apple cotyledon using the antibodies which its high specificity had been previously demonstrated. We found NAD-SDH was ubiquitously distributed with subcellular multi-localization in developing cotyledon. Sorbitol imported into apple seed is not stored but metabolized rapidly after unloading by NAD-SDH. Moreover, our data suggest that NADSDH is imported into the vacuoles mediated by 18 amino-acid residues as an internal SP. The NAD-SDH localized in the seed protein storage vacuoles may play function by being involved in sorbitol metabolism at the early stage of seed germination. 2. Materials and methods 2.1. Plant materials Apple (M. domestica Borkh. cv. Starkrimson) seeds were collected at 8, 10 and 12 weeks after full bloom (AFB). Samples were picked for immediate use or frozen in liquid nitrogen and kept at −80 ◦ C. For immunohistochemistry and subcellular immunogold labeling experiments, samples were fixed immediately after harvest. Preparation, purification and usage of polyclonal antibodies against the NAD-SDH of apple were according to the description by Wang et al. [16]. 2.2. Reverse transcription-polymerase chain reaction analysis Reverse transcription-polymerase chain reaction (RT-PCR) was used to analyze the expression of MdSDH5 (GenBank accession No. AY849315) and MdSDH6 (GenBank accession No. AY849316) we cloned from apple in the transcription level in the seeds. The specific primers are as follows: forward primer for MdSDH5 5 -ACTATTTACTCGCAGCCTGA-3 , reverse primer for MdSDH5 5 -GAATACCAACACTTAAGGGC-3 , forward primer for MdSDH6 5 - CGTGTATTCTGTGTCTTCTGTG-3 , and reverse primer for MdSDH6 5 -CGGAGATCATGGCTTCTTTAAT-3 . EF-1a, an elongation factor of the apple (GenBank accession No. AJ223969), was selected as the control gene. The forward and reverse primers of EF-1a are 5 -ATTGTGGTCATTGGYCAYGT-3 and 5 CCTATCTTGTAVACATCCTG-3 ). 2.3. Immunohistochemistry The anti-NAD-SDH rabbit polyclonal antibodies we used here had been purified by antigen affinity and their high specificity had been demonstrated in our previous work [16]. The cotyledons were cut into small cubes (about 2–10 mm3 ) and were immediately fixed in 4% pre-cooled paraformaldehyde solution (100 mM PBS, pH 7.0) overnight at 4 ◦ C. The fixed samples were dehydrated in a graded

37

ethanol series, replaced with xylene, and embedded in paraffin. Sections made by microtome (6–10 mm thick) were pasted onto slide glasses pretreated with 1% poly-l-lysine solution. The sections were deparaffinized in xylene and rehydrated in serial ethanol rinse (100%, 70%, 50%, and 30% ethanol in PBS). After washing with PBS, the sections were blocked overnight at 4 ◦ C and were incubated with purified anti-NAD-SDH rabbit antiserum solution for 2 h at 37 ◦ C. Followed by another washing in PBS, the samples were incubated in goat anti-rabbit IgG-fluorescein isothiocyanate (FITC) antibodies (diluted 100-fold in PBS solution) for 1 h at 37 ◦ C. The immune signal was detected by Zeiss LSM 510 confocal laser scanning microscope. 2.4. Transmission electron microscopy and immunogold labeling Ultrathin sections were prepared and immunogold labeling was done essentially as described by McCartney et al. [18]. The cotyledons were cut into small cubes immediately after isolation from fruits (about 2 mm3 ), then fixed with a solution of 4% (w/v) paraformaldehyde in 100 mM pre-cooled phosphate buffer (pH 7.2), and incubated for 6 h at 4 ◦ C. After washing with phosphate buffer, the samples were post-fixed overnight in 0.1% (w/v) OsO4 at 4 ◦ C, then rinsed with the pre-cooled phosphate buffer, and dehydrated through a graded ethanol series. After infiltration for 48 h with Spurr epoxy resin at 4 ◦ C, the samples were polymerized at 58 ◦ C for 12 h, and cut into ultrathin sections. It had been proven that heating at 58 ◦ C for 12 h did not destroy the immunogenicity of NAD-SDH in our previously work [16]. The rabbit polyclonal antiserum against apple NAD-SDH we used here had been affinity-purified and evaluated. It had high specificity to NAD-SDH of apple and it had been used to study the subcelluar localization of NAD-SDH in apple fruits [16]. The sections were etched with 560 mM sodium metaperiodate for 50 min, and 0.1 M HCl for 30 min. After washing with Tris-buffered saline Tween-20 buffer (TBST) (10 mM Tris (pH 7.4), 500 mM NaCl and 0.3% Tween 20) for 5 min, the sections were incubated in TBST buffer containing 2% BSA for 1 h. After washing with TBST buffer, the treated sections were incubated in purified anti-NAD-SDH rabbit antibodies (diluted in TBST buffer, containing 0.1% BSA) at room temperature for 3 h. After extensive washing with TBST buffer containing 2% BSA, the sections were incubated in goat antirabbit IgG antibodies conjugated with 10 nm gold (diluted in TBST buffer, containing 0.1% BSA) for 1 h at 37 ◦ C. The sections were rinsed consecutively with TBST containing 2% BSA, TBST buffer, and double-distilled water, and then stained with uranyl acetate. After washed with double-distilled water, the sections were examined under electron microscope. 2.5. Sugar analysis The soluble sugars and sorbitol were detected of apple seeds at the middle stage of seed development (8 weeks after full bloom (AFB)) and the late stage of development (12 weeks AFB). The sugar extracts were obtained from 100 to 200 mg of cotyledons. After drying at 80 ◦ C and grinding, the samples were treated in 20 ml 80% aqueous ethanol at 75 ◦ C for 30 min. The extracts, after condensation and centrifugation, were used for sugar analysis with high-performance liquid chromatography (HPLC). Samples were subjected to HPLC under following conditions: instrument, Waters 966; analytical column, Sugar Pak; column temp, 90 ◦ C; solvent system, aqueous solution of 0.0001 M Ca-EDTA; flow rate, 0.5 ml/min; injection volume, 5 ␮l; detectors, differential refractive index detector. The 0.15% sucrose, and 0.1% fructose, glucose, mannitol and sorbitol were used as standards to evaluate the accuracy.

38

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

2.6. The prediction of internal signal peptide and transmembrane domain in MDSDH5 and MDSDH6 The bio-informatic tools for the signal peptide prediction of NAD-SDH were iPSORT Prediction (http://ipsort.hgc.jp/predict.cgi) and CBS Prediction Servers (http://www.cbs.dtu.dk/services/ SignalP). Because only the first 30–70 residues are used for prediction by these programs, we deleted the amio-acid residues one by one from the N-terminal regions and the remaining sequences were submitted for predication. Transmembrane domain (TMD) was predicted by using tool “DAS” – Transmembrane Prediction Server (http://www.sbc.su.se/∼miklos/DAS/). The conservation analysis (sequence alignment) of predicted signal peptides and transmembrane segments in all NAD-SDH isoforms was done with software DNAMAN. 2.7. Plasmid constructions and transient expressing The coding sequences of MdSDH5 were cloned as a C-terminal fusion in-frame with the GFP into the pEZS-NL expression vector with EcoRI/SalI and expressed under the control of the 35S promoter. For simplicity, this construct was named MdSDH5-GFP. The mutants of MdSDH5 were created as follows: 1): DNA sequences which code MdSDH5 segments (residues 1–193 and 217–371) were ligated respectively into the expression vector to form these constructs MdSDH51−193 -GFP and MdSDH5217−371 -GFP. 2): the deletion mutant lacking residues “199GAGPIGLVSVLAARAFGA216” (MdSDH5d199−216 ) was preformed by deletion mutagenesis using the mutanBEST kit (TaKaRa). The PCR amplification, digested with EcoRI/SalI and blunting ligation were performed according to the manufacturer’s instructions. The full-length or segments (residues 1–191 and 211–368) of MdSDH6 were ligated into pBI221 with XbaI/BamHI and named MdSDH6-GFP, MdSDH61−191 -GFP and MdSDH5211−368 GFP, respectively. The mutant of MdSDH6 deleting “193GAGPIGLVSVLAARAFGA210” (MdSDH6d193−210 ) was done by the same method described above. Transient expression in the Arabidopsis mesophyll protoplasts was performed essentially with the procedures described by Sheen [19]. Protoplasts were isolated from the leaves of 3–4 weeks old Arabidopsis and transiently transformed using PEG. Fluorescence of GFP was obtained by series of optical sections (Z-series stack at 0.8-␮m intervals) through single protoplasts using confocal laser scanning microscope (LSM 510 meta) after incubation at 22 ◦ C for 24–48 h. Because it is easy to distinguish the chloroplasts and vacuoles by the chlorophyll autofluorescence and the light micrograph, we did not use these fluorescently-tagged markers of chloroplast and tonoplast to co-express with MdSDH-GFP. Because the GFP fluorescence in vacuole and chloroplasts is not discerned clearly in a single optical section, we selected some images in which the vacuole is the observation aim (see from the light micrographs) from the series of optical sections in this study. All transient expression assays were repeated at least three times. 3. Results 3.1. Sorbitol accumulation in apple seeds in the late development stage To elucidate NAD-SDH function during seed growth and development, we analyzed firstly the concentrations of sorbiotl and soluble sugars including sucrose, glucose, and fructose of apple seeds at 8 weeks after full bloom (AFB) (the middle stage of seed development which seed begins to accumulate storage materials) and 12 weeks AFB (the late stage of seed development which is

Table 1 Soluble sugars and sorbitol (micrograms) in apple seeds (fresh weight expressed as grams). Substance

Concentration (mg g−1 ) Stages of seed development

1. Sucrose 2. Glucose 3. Froctose 4. Mannitol 5. Sorbitol a

8 weeks AFB

12 weeks AFB

3.06 1.22 0.06

2.38 2.44 2.81

a

a

a

0.63

Trace < 0.05 mg g−1 ; ABF, after full bloom.

at the end of storage material accumulation and seed approaches maturation) by HPLC. We found that the HPLC elution profiles of these soluble sugars were varied in the cotyledons at these two different developmental stages. The sucrose concentration was decreased, but concentrations of glucose and fructose, especially fructose (from 0.06 mg g−1 FW to 2.81 mg g−1 FW), were increased in the seeds at the late development stage compared with those in seeds at middle stage. (Fig. 1A; Table 1). The fructose concentration was very low (0.06 mg g−1 FW) and sorbitol was undetectable in the seeds at 8 weeks AFB, suggesting that sorbitol imported into apple seed is not stored but metabolized rapidly after unloading in apple seeds at the middle stage of seed development. However, an amount of sorbitol (0.63 mg g−1 FW) was present in the seeds at 12 weeks AFB, which indicate that sorbitol (as fructose, glucose, and sucrose) can be stored in seeds at the late stage of seed development. In our present analysis, sucrose, fructose, glucose, mannitol, and sorbitol were completely separated from the mixture standard, indicating the above data are reliable. 3.2. Gene expression and protein distribution of NAD-SDH in the cotyledon cells of developing apple seeds Sorbitol was undetectable in seeds at the middle stage of development by HPLC. A possible reason for this is that the NAD-SDH in the seeds rapidly catalyzes the conversion of sorbitol to fructose during seed development. Thus, the gene expression and protein distribution pattern of NAD-SDH are required for further understanding the processes in developing seeds. In our previous work, we cloned two NAD-SDH genes from apple fruits, MdSDH5 and MdSDH6, which had high homology to other known NAD-SDH genes in apple [16]. RT-PCR was conducted to analyze their expressions during seed development using total RNA extracts from 8 weeks AFB seeds. The result revealed that MdSDH5 and MdSDH6 were highly expressed in the developing seeds (Fig. 1B). The distribution pattern of NAD-SDH in the developing seeds was assessed by immunohistochemistry with anti-NAD-SDH serum and FITC-coupled goat anti-rabbit IgG secondary antibodies. The polyclonal antiserum we used had been affinity-purified, and its high specificity to NAD-SDH of apple had been proved in our previously experiment [16]. In seeds at 8 weeks AFB, FITC fluorescence was observed in the parenchyma, epidermis, and vascular bundle in the cotyledons (Fig. 1C), indicating that NAD-SDH is ubiquitously distributed in the cotyledons of apple seeds. Moreover, some spherical bodies having FITC fluorescence seemed likely to be distributed in the vacuoles of the parenchyma cells (Fig. 1D). 3.3. Subcellular localization of NAD-SDH protein in the cells of developing cotyledon In order to determine whether the NAD-SDH protein is present simultaneously in the vacuoles and cytoplasm of these cotyledon

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

39

Fig. 1. Detection of sorbitol, gene expression and protein distribution of NAD-SDH in apple cotyledon. (A) HPLC detection of soluble carbohydrate extracted from apple seeds at 8 weeks AFB (middle stage of seed development) (I) and 12 weeks AFB (late stage of seed development) (II). Sorbiotl (0.63 mg g−1 FW) (arrow) was accumulated in seeds at late stage of development. 0.1% sucrose (1), 0.1% glucose (2), 0.1% fructose (3), 0.1% mannitol (4), and 0.1% sorbitol (5) were used as standards (III). (B) Gene expression of MdSDH5 and MdSDH6 in the 8 weeks AFB seeds. EF-1a was the control gene. (C–E) Immunohistochemical analysis using purified anti-NAD-SDH serum followed by FITC labeled goat anti-rabbit IgG. (C) FITC fluorescence was detected in cells of parenchyma (Pa), and vascular bundle (Vb). (D) Magnified view of parenchyma cells, some spherical bodies having fluorescence (arrows) was seemed to exist in the vacuoles (Va). (E) Control experiment using preimmune serum. I, FITC fluorescence image; II, light micrograph; Scale bars = 200 ␮m (C) and (E); 20 ␮m in (D).

cells, immunogold electron microscopy was used to examine the subcellular localization of this enzyme. The anti-NAD-SDH serum we used had been purified by antigen affinity and its high specificity had been proven in our previous work [16]. We had also confirmed that heating at 58 ◦ C for polymerization of Spurr epoxy resin did not destroy the immunogenicity of NAD-SDH. In cotyledon cells of apple seeds at 8 weeks AFB, the central vacuole was subdividing into smaller ones in which electron-dense proteins began to accumulate (Fig. 2A and B). Some endoplasmic reticulums (ERs) became swollen. A lot of ribosomal particles were attached on the membrane. Therefore, these swollen ERs are rough ERs (Fig. 2C and D). The immunogold labeling results showed that immunogold particles were localized in the cytoplasm, the membranes of the ERs and vesicles, and vacuolar lumen of the cotyledon cells in seeds at 8 weeks AFB (Fig. 3A and B). Immunogold particles were rarely observed in the cell wall, plasmodesmata, plasma membrane, lipid bodies, mitochondria, and tonoplasts of the cotyledon cells (Fig. 3A and B). No gold particles were found in the control experiments labeled with preimmune serum (Fig. 3C), indicating the specificity and reliability of immunogold labeling.

3.4. Transportation of the NAD-SDH to vacuoles and accumulation in the protein storage vacuoles Abundant immunogold particles were localized in the membranes of some vesicles near the vacuoles, and no immunogold particles were observed in the lumen of the vesicles in cotyledon cells of apple seeds at 8 weeks AFB (Fig. 4A). A lot of ribosome particles were attached on the membranes of these vesicles, suggesting that these vesicles may be ER-derived (Fig. 4B). Furthermore, some immunogold particles were found being transported from these

vesicles membrane to the vacuolar lumen (Fig. 4B). There were no immunogold particles on the tonoplasts. In the cotyledon cells of seeds at 10 weeks AFB, the vacuoles were filled with electronopaque materials and were transformed to protein storage vacuoles (PSVs). Abundant immunogold particles were localized predominantly in the opaque contents in the PSVs (Fig. 4C).

3.5. Eighteen amino acid residues are essential for the vacuolar targeting of MdSDH5/6 To determine whether there is a signal peptide (SP) to direct NAD-SDH transportation to the vacuoles, we analyzed their amino acid sequences of using web-based tools (see Section 2). There were no sequences predicted to be a putative N-terminal SP. However, a fragment in the middle region of NAD-SDH such as 194N-216A in MdSDH5 and 193G-210A in MdSDH6 was predicted to be internal SP after around 190 residues were deleted from the N-terminal regions (Fig. 5A and B). Moreover, the residues 197V-213A in MdSDH5 and 192I-210A in MdSDH6 were predicted to be TMD (see Section 2) (Fig. 5C and D). Eighteen amino-acid residues in predicted internal SP and TMD of MdSDH5 (“199G-A216”) or MdSDH6 (“193G-A210”) were highly conserved across all 11 known isoforms of NAD-SDH in apple (Fig. 5E). To determine the role of the predicted internal SP, different deletion mutants of MdSDH5/6 were fused to GFP under 35S promoter and expressed in mesophyll protoplasts of Arabidopsis thaliana. The GFP fluorescence in the cytosol and chloroplasts in the protoplasts expressing MdSDH5-GFP was observed as we shown in our previous work in which the images dominantly exhibited the chloroplasts distribution of MdSDH5 [16]. In this study, the fluorescence was also found in vacuolar lumen in some optical sections

40

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

Fig. 2. Ultrastructure of the cotyledon cells in apple seeds in 8 weeks AFB. (A) The central vacuoles were subdivided into smaller ones. (B) Storage proteins (arrows) began to be accumulated in vacuoles. (C and D) Some ERs were swollen (asterisks) and abundant ribosomal particles (arrows) were attached on the membrane of swollen ERs. Scale bars = 2 ␮m in (A); 1 ␮m in (B), (C), and (D).

Fig. 3. Immunogold localization of NAD-SDH in cotyledon cells of apple seeds at 8 weeks AFB. (A) Immunogold particles were observed in the cytosol (small arrows), the membranes of ER and vesicles (large arrows). (B) Immunogold particles were localized in the swollen ER-membrane (large arrows), the vacuole, and cytosol (small arrows). Ribosome particles were observed clearly on the ER-membrane. (C) Immunogold particle was rarely observed in the cotyledon cells incubated with preimmune serum. Immunoelectron microscopy was carried out with purified anti-NAD-SDH serum and a gold-conjugated anti-rabbit IgG secondary antibody. CW, cell wall; ER, endoplasmic reticulum; L, lipid body; M, mitochondrion; V, vesicle; Va, vacuole. Scale bars = 2 ␮m.

in which vacuole is the observation target with Z-series stack at 0.8-␮m intervals (Fig. 6, panel 1). Moreover, some spherical bodies in the vacuoles were detected with strong fluorescence signal in these cells. However, the GFP fluorescence in the vacuoles

was rarely observed in the protoplasts expressed the mutants, MdSDH51−193 , MdSDH5217−371 , and MdSDH5d199−216 , respectively (Fig. 6, panels 2–4). The similar results were observed for MdSDH6 by series of optical sections through protoplasts. The fluorescence

Fig. 4. NAD-SDH trafficking to the vacuoles and accumulation in protein storage vacuoles in cotyledon cells of apple seeds. (A) Abundant immunogold particles were distributed in the cytosol (boxs) and in the membranes of vesicles (large arrows) near the vacuoles in cotyledon cells of apple seeds at 8 weeks AFB. The immunogold signal was also found in the vacuolar lumen, but not on the tonoplast (small arrow). (B) NAD-SDH with immunogold labeling was transported from the membrane of these vesicles to the vacuolar lumen (large arrow). There were lots of ribosome particles on the membranes of these vesicles (small arrows). (C) Gold particles were mainly localized in the PSV in the cotyledon cells of apple seeds at 10 weeks AFB. CW, cell wall; M, mitochondrion; PSV, protein storage vacuole; V, vesicle; Va, vacuole. Scale bars = 2 ␮m.

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

41

Fig. 5. Analysis of the internal SP and TMD in MdSDH5 and MdSDH6. (A and B) Predicted internal SP regions in MdSDH5 (A) and MdSDH6 (B) using web-based tools (see Material and methods). (C and D) Predicted TMD in MdSDH5 (C) and MdSDH6 (D). (E) The conserved sequences (underlined) in SP and TMD among all identified NAD-SDH isoforms were aligned by DNAMAN.

signals in vacuoles were detected only in the protoplasts expressing full MdSDH6 (Fig. 7, panel 1), not in these cells expressing the MdSDH61−191 , MdSDH6211−368 , and MdSDH5d193−210 , respectively (Fig. 7, panels 2–4). Taken together, these data show that those 18 residues are required for the vacuolar targeting of MdSDH5/6. 4. Discussion In most cases, proteins transported into the endomembrane system (e.g., ER, Golgi, and vacuole, etc.) harbor N-terminal SP and are trafficked in a co-translational manner [20]. However, some proteins targeted to these destinations only contain an internal signal peptide or C-terminal targeting signal and are

transported in a post-translational (SRP-independent) manner [21–23]. It has been previously reported that an internal SP in UDP-glucuronosyltransferase isoform UGT1A6 can mediate the post-translational ER targeting of this protein [21]. In addition, in some single-pass transmembrane proteins, a single internal SP remains in the lipid bilayer as a membrane-spanning alpha helix [24]. Our results presented herein revealed that 18 amino acid residues in MdSDH5/6 proteins may play role in directing MdSDH5/6 translocation to vacuoles as an internal SP and a TMD. These 18 amino acid residues are highly conserved in all known NAD-SDH isoforms (Fig. 5E). Hence, we propose that these amino acids residues as internal signal sequence function in the targeting of all NAD-SDH isoforms to the vacuoles.

42

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

Fig. 6. Residues 199–216 are required for the vacuolar targeting of MdSDH5. (A) Translational fusion constructs of intact MdSDH5 and three deletion mutants to the reporter GFP. Inlet is the amplified portions of the boxed-in area. (B) These constructs were transiently expressed in mesophyll protoplasts of Arabidopsis and GFP fluorescence was observed by series of optical sections (0.8 ␮m) through a single protoplast. These images were optical sections in which the observation target is the vacuoles (see from light micrograph). The presence of GFP fluorescence in cytosol and vacuolar lumen was found in these protoplasts expressing MdSDH5-GFP. Arrows show these spherical bodies in the vacuolar lumen having strong fluorescence signal in the protoplasts expressing MdSDH5-GFP (panel 1). Inlet in panel 1 is the amplified portion of the boxed-in area, showing the MdSDH5-GFP (spherical bodies) locating in the vacuoles. No vacuolar fluorescence signals were observed in those cells expressing MdSDH51−193 -GFP, MdSDH5217−371 -GFP, and MdSDH5d199−216 -GFP (panels 2–4). GFP was used as a control (panel 5). I, GFP fluorescence image; II, chlorophyll autofluorescence image; III, light micrograph; IV, merged image of I and II. Bars = 10 ␮m.

The biosynthesis of vacuolar proteins takes place at the rough ER and the nascent precursors of these proteins are trafficked to the vacuoles via the ER and the Golgi apparatus in plants [25]. However, the ER-derived vesicles can bypass the Golgi apparatus and directly fuse with the vacuoles, referred to as the ERvt (ER to vacuole trafficking) pathway, which involves the transport of both storage proteins and enzyme precursors [26]. Specific TMD and cytoplasmic tail (CT) sequences are proved to be involved in directing proteins to vacuoles by different pathways [27–29]. A reporter

protein containing the TMD and CT of BP-80 can reach the vacuole via the Golgi. However, the CT of ␣-tonoplast intrinsic protein (␣-TIP) can direct the protein from the ER to the PSV, bypassing the Golgi [27,30]. Recently, Rivera-Serrano et al. reported that there were two distinct pathways for trafficking of TIP in Arabidopsis [31]. Golgi-dependent pathway may be used by TIP1;1 (␥TIP), but a Golgi-independent pathway may be used by TIP3;1 (␦TIP) and TIP2;1 (␣-TIP). In our present study, there were a lot of immunogold particles are localized in the membrane of vesicles.

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

43

Fig. 7. Residues 193–210 involve in MdSDH6 targeting vacuoles. (A) The schematic diagram represents GFP-reporter constructs of MdSDH6 and deletion mutants. (B) Fluorescence signals were detected in the cytosol and vacuoles (arrow) in some optical sections in Arabidopsis protoplasts expressing MdSDH6-GFP (panel 1). In protoplasts that express MdSDH61−191 -GFP, MdSDH6211−368 -GFP, and MdSDH6d193−210 -GFP, respectively, GFP fluorescence in vacuolar was not observed (panels 2–4). I, GFP fluorescence image; II, chlorophyll autofluorescence image; III, light micrograph; IV, merged image of I and II. Bars = 10 ␮m.

We presume that these vesicles may be ER-derived, because many ribosome particles are attached on their membranes (Fig. 4B). Therefore, we propose that NAD-SDH may be transported to the vacuoles via ERvt pathway in the apple cotyledons bypassing Golgi. From our present experiment it can be concluded that NAD-SDH is firstly targeted to ERs mediated by the internal SP. After it is transferred into the vesicles, it is then imported into vacuoles, and eventually accumulated in the PSVs in apple cotyledons. NAD-SDH was found in the membranes of ERs and vesicles in the cells of apple cotyledon in our immunogold labeling experiment. However, most puzzling of all, they were in the vacuolar lumens as soluble proteins but not on the tonoplasts as membrane proteins. In plants, similar results had been reported [29]. A recombinant protein which was composed of a SP (from the potato pathogenesis-related protein), yeast invertase (soluble protein, no vacuolar targeting information), and a TMD (from yeast calnexin, acting as a stop-transfer sequence) was found to be transported to vacuoles in a membrane anchored form. The membrane attached

protein is detached from its membrane anchor either just prior to or after delivery to the vacuole [29]. Further investigation is required to understand the mechanism controlling the detachment of NADSDH into vacuoles from its membrane anchor. There are an increasing number of reports that some proteins are transported to more than one cellular compartment mediated by the same or different signal sequences [32–35]. For example, the translocation of rice ␣-amylases to plastids (chloroplasts or amyloplasts) and extracellular compartments is facilitated by the same targeting signal sequence [36,37]. Various forms of the same protein, such as cytochrome b5-A, -B, -C, and -D, a class of membrane proteins, can be targeted posttranslationally to either the ER or mitochondria via a region of hydrophobic amino acids located near the C-terminus [22]. er-cHL, a lyase activity enzyme, had been proved to be localized both in the cytosol and ER [38]. Our current and previous experimental results of transient expression and immunogold labeling experiments reveal that NAD-SDH has three different destination including cytosol, vacuole (PSV), and

44

X.-L. Wang et al. / Plant Science 210 (2013) 36–45

chloroplast in the cells of fruit fresh, mesophyll, and cotyledon [16]. This is consistent with the report that sorbitol is present in the chloroplasts, cytosol, and vacuoles in peach [9]. However the mechanism regulating multiple subcellular localizations of NAD-SDH of apple is still unknown. PSVs are specialized vacuoles devoted to the accumulation of lots of proteins in the storage tissues and organs such as plant seeds. However, PSVs are not limited in the storage tissues. It had been reported that PSVs are present in leaves of common bean, tobacco and Arabidopsis [39]. Phaseolin, one of the major storage proteins, is accumulated in the storage vacuoles of leaves in these plants. Moreover, ␣-TIP (␣-tonoplast intrinsic protein), a tonoplast aquaporin of PSVs, is also found in the tonoplast of the storage vacuoles in common bean, tobacco and Arabidopsis leaves. Therefore, the results of NAD-SDH targeting to vacuoles in protoplasts of Arabidopsis leaves can be used to demonstrate the vacuoles and PSVs targeting of NAD-SDH in apple seeds. Our experimental results of transient expression assay using Arabidopsis mesophyll protoplasts reveal that 18 conserved amino acid residues as internal signal sequence are important for the vacuoles localization of NAD-SDH. Two putative transit peptides for chloroplast localization in NAD-SDH were predicted with the “iPSORT Prediction” (http://ipsort.hgc.jp) (data not shown). Further experiments are needed to identify the roles of putative transit peptides in NAD-SDH for chloroplast targeting. Sorbitol is the major photosynthetic product transported to sink tissues where it is mainly converted to fructose via NAD-SDH in apple [40]. Our current HPLC results show that sorbitol is only marginally detectable in the apple cotyledons at the middle stage of seeds development, but sorbitol is accumulated in the seeds at the late developing stage (Fig. 1A). Consistent with this result is the finding that NAD-SDH is present at high levels in the cytosol in the cells of apple cotyledon at the middle stage of seeds development, but localized mainly in the PSVs in seeds at late stage (Fig. 3C). Moreover, it had been reported that very high activity of NAD-SDH was found in apple seed [15]. These results mean that sorbitol imported into seed is not stored but metabolized rapidly after unloading as in apple fruit, which suggests that NAD-SDH in the cytosol of cotyledon cells plays a vital function in apple seed development. Large amounts of storage proteins are accumulated in PSVs of cotyledon cells during seed development and maturation. These proteins serve as storage materials for seed germination and early seedling growth [16,41]. However, it is clear that NAD-SDH is not a kind of storage protein. Therefore, the NAD-SDH in vacuoles and PSVs may have other functions in apple seeds. It has been reported that the optimum pH for NAD-SDH activity in vitro for sorbitol oxidation is 9.6 (pH range, 7.5–10.5) and for fructose reduction is 7.0 (pH range, 5.0–8.0) [42]. The low pH in vacuoles (about 5.5) is unfavorable for sorbitol oxidation and fructose reduction. Moreover, we detected that sorbitol accumulation only took place in apple seeds at the late stage of development. Whetter and Taper found that sorbitol is present in apple seeds at the initial germinating stages [17]. These studies suggest that the accumulated sorbitol in mature apple seeds may act as storage materials used for seed germination, and NAD-SDH in the vacuoles and PSVs in apple seeds may be inactive. Inactive NAD-SDH in seed will become active to catalyze the conversion of accumulated sorbitol to the fructose during the initial germinating stage. It is similar to the regulation of ␤-amylase, which is synthesized and accumulated as a latent form in the starchy endosperm during seed development and is engaged in the breakdown of starch during seed germination in barley, wheat, and rye [43–45]. The accumulation of NAD-SDH and sorbitol in mature seed is consistent with the idea that NAD-SDH may contribute to sorbitol metabolism during seed germination.

Acknowledgements This work was supported by Ministry of Agriculture of China (2013ZX08009-003-002). References [1] N. Noiraud, L. Maurousset, R. Lemoine, Transport of polyols in higher plants, Plant Physiol. Biochem. 39 (2001) 717–728. [2] G. Teo, Y. Suzuki, S.L. Uratsu, B. Lampinen, N. Ormonde, W.K. Hu, A.M. Dandekar, Silencing leaf sorbitol synthesis alters long dispartitioning and apple fruit quality, Proc. Natl. Acad. Sci. U.S.A. 103 (2006) 18842–18847. [3] R. Zhou, L. Cheng, A.M. Dandekar, Down-regulation of sorbitol dehydrogenase and up-regulation of sucrose synthase in shoot tips of the transgenic apple with decreased sorbitol synthesis, J. Exp. Bot. 57 (2006) 3647–3657. [4] F.B. Negm, W.H. Loescher, Characterization and partial purification of aldose-6phosphate reductase (alditol-6-phosphate: NADP-oxidoreductase) from apple leaves, Plant Physiol. 67 (1981) 139–142. [5] M. Hirai, Purification and characteristics of sorbitol-6-phosphate dehydrogenase from loquat leaves, Plant Physiol. 67 (1981) 221–224. [6] R. Zhou, L. Cheng, R. Wayne, Purification and characterization of sorbitol-6phosphate phosphatase from apple leaves, Plant Sci. 165 (2003) 227–232. [7] S. Yamaki, Subcellular Localization of NADP-dependent sorbitol-6-phosphate dehydrogenase in protoplast from apple cotyledons, Plant Cell Physiol. 22 (1981) 359–367. [8] D. Liang, M. Cui, S. Wu, F. Ma, Genomic structure, subcellular localization, and promoter analysis of the gene encoding sorbitol-6-phosphate dehydrogenase from apple, Plant Mol. Biol. Rep. 30 (2012) 904–914. [9] J. Nadwodnik, G. Lohaus, Subcellular concentrations of sugar alcohols and sugars in relation to phloem translocation in Plantago major, Plantago maritima, Prunus persica, and Apium graveolens, Planta 227 (2008) 1079–1089. [10] W.H. Loescher, G.C. Marlow, R.A. Kennedy, Sorbitol metabolism and sink–source interconversions in developing apple leaves, Plant Physiol. 70 (1982) 335–339. [11] H. Yamaguchi, Y. Kanayama, S. Yamaki, Purification and properties of NADdependent sorbitol dehydrogenase from apple fruit, Plant Cell Physiol. 35 (1994) 887–892. [12] O. El-Kabbani, C. Darmanin, R.P. Chung, Sorbitol dehydrogenase, structure, function and ligand design, Curr. Med. Chem. 4 (2004) 465–476. [13] K. Yamada, Y. Oura, H. Mori, S. Yamaki, Cloning of NAD-dependent sorbitol dehydrogenase from apple fruit and gene expression, Plant Cell Physiol. 39 (1998) 1375–1379. [14] S.W. Park, K.J. Song, M.Y. Kim, J.H. Hwang, Y.U. Shin, W.C. Kim, W. Chung, Molecular clone and characterization of four cDNAs encoding the isoforms of NAD-dependent sorbitol dehydrogenase from the Fuji apple, Plant Sci. 162 (2002) 513–519. [15] M. Nosarzewski, D.D. Archbold, Tissue-specific expression of sorbitol dehydrogenase in apple fruit during early development, J. Exp. Bot. 58 (2007) 1863–1872. [16] X.L. Wang, Y.H. Xu, C.C. Peng, R.C. Fan, X.Q. Gao, Ubiquitous distribution and different subcellular localization of sorbitol dehydrogenase in fruit and leaf of apple, J. Exp. Bot. 60 (2009) 1025–1034. [17] J.M. Whetter, C.D. Taper, Occurrence of sorbitol (d-glucitol) and certain related sugars in germinating seeds and developing seedlings of Malus, Can. J. Bot. 44 (1966) 51–55. [18] A.W. McCartney, J.S. Greenwood, M.R. Fabian, K.A. White, R.T. Mullen, Localization of the tomato bushy stunt virus replication protein p33 reveals a peroxisome-to-endoplasmic reticulum sorting pathway, Plant Cell 17 (2005) 3513–3531. [19] J. Sheen, Signal transduction in maize and Arabidopsis mesophyll protoplasts, Plant Physiol. 127 (2001) 1466–1475. [20] J.E. Smotrys, M.E. Linder, Palmitoylation of intracellular signaling proteins: regulation and function, Annu. Rev. Biochem. 73 (2004) 559–587. [21] M. Ouzzine, J. Magdalou, B. Burchell, S. Fournel-Gigleux, An internal signal sequence mediates the targeting and retention of the human UDPglucuronosyltransferase 1A6 to the endoplasmic reticulum, J. Biol. Chem. 274 (1999) 31401–31409. [22] Y.T. Hwang, S.M. Pelitire, M.P. Henderson, D.W. Andrews, J.M. Dyer, R.T. Mullen, Novel targeting signals mediate the sorting of different isoforms of the tailanchored membrane protein cytochrome b5 to either endoplasmic reticulum or mitochondria, Plant Cell 16 (2004) 3002–3019. [23] C.H. Chou, R.S. Lee, H.F. Yang-Yen, An internal EELD domain facilitates mitochondrial targeting of Mcl-1 via a Tom70-dependent pathway, Mol. Cell Biol. 17 (2006) 3952–3963. [24] S. High, B. Dobberstein, Mechanisms that determine the transmembrane disposition of proteins, Curr. Opin. Cell Biol. 4 (1992) 581–586. [25] F.C. Reyes, T. Chung, D. Holding, R. Jung, R. Vierstra, M.S. Otegui, Delivery of prolamins to the protein storage vacuole in maize aleurone cells, Plant Cell 23 (2011) 769–784. [26] F. De Marchis, M. Bellucci, A. Pompa, Traffic of human ␣-mannosidase in plant cells suggests the presence of a new endoplasmic reticulum-to-vacuole pathway without involving the Golgi complex, Plant Physiol. 161 (2013) 1769–1782. [27] L. Jiang, J.C. Rogers, Integral membrane protein sorting to vacuoles in plant cells, evidence for two pathways, J. Cell Biol. 143 (1998) 1183–1199.

X.-L. Wang et al. / Plant Science 210 (2013) 36–45 [28] L. Jiang, J.C. Rogers, Sorting of membrane proteins to vacuoles in plant cell, Plant Sci. 146 (1999) 55–67. [29] F. Barrieu, M.J. Chrispeels, Delivery of a secreted soluble protein to the vacuole via a membrane anchor, Plant Physiol. 120 (1999) 961–968. [30] L. Jiang, S.S. Sun, Membrane anchors for vacuolar targeting, application in plant bioreactors, Trends Biotechnol. 20 (2002) 99–102. [31] E.E. Rivera-Serrano, M.F. Rodriguez-Welsh, G.R. Hicks, M. Rojas-Pierce, A small molecule inhibitor partitions two distinct pathways for trafficking of tonoplast intrinsic proteins in Arabidopsis, PLoS ONE 7 (2012) e44735. [32] J.C. Mueller, C. Andreoli, H. Prokisch, T. Meitinger, Mechanisms for multiple intracellular localization of human mitochondrial proteins, Mitochondrion 3 (2004) 315–325. [33] A. Villarejo, S. Burén, S. Larsson, A. Déjardin, M. Monné, C. Rudhe, J. Karlsson, S. Jansson, P. Lerouge, N. Rolland, G. von Heijne, M. Grebe, L. Bako, G. Samuelsson, Evidence for a protein transported through the secretory pathway enroute to the higher plant chloroplast, Nat. Cell Biol. 7 (2005) 1224–1231. ˜ [34] Y. Nanjo, H. Oka, N. Ikarashi, K. Kaneko, A. Kitajima, T. Mitsui, F.J. Munoz, M. Rodríguez-López, E. Baroja-Fernández, J. Pozueta-Romero, Rice plastidial N-glycosylated nucleotide pyrophosphatase/phosphodiesterase is transported from the ER-Golgi to the chloroplast through the secretory pathway, Plant Cell 18 (2006) 2582–2592. [35] A. Rödiger, B. Baudisch, U. Langner, R.B. Klösgen, Dual targeting of a mitochondrial protein: the case study of cytochrome c1, Mol. Plant 4 (2011) 679–687. [36] M.H. Chen, L.F. Huang, H.M. Li, Y.R. Chen, S.M. Yu, Signal peptide-dependent targeting of a rice alpha-amylase and cargo proteins to plastids and extracellular compartments of plant cells, Plant Physiol. 135 (2004) 1367–1377.

45

[37] A. Kitajima, S. Asatsuma, H. Okada, Y. Hamada, K. Kaneko, Y. Nanjo, Y. Kawagoe, K. Toyooka, K. Matsuoka, M. Takeuchi, A. Nakano, T. Mitsui, The rice ␣-amylase glycoprotein is targeted from the golgi apparatus through the secretory pathway to the plastids, Plant Cell 21 (2009) 2844–2858. [38] M. Arnedo, S. Menao, B. Puisac, M.E. Teresa-Rodrigo, M.C. Gil-Rodríguez, E. ˜ P. Gómez-Puertas, N. Casals, C.H. Casale, F.G. Hegardt, J. Pié, López-Vinas, Characterization of a novel HMG-CoA Lyase enzyme with a dual location in endoplasmic reticulum and cytosol, J. Lipid Res. 53 (2012) 2046–2056. [39] M. Park, S.J. Kim, A. Vitale, I. Hwang, Identification of the protein storage vacuole and protein targeting to the vacuole in leaf cells of three plant species, Plant Physiol. 134 (2004) 625–639. [40] W.H. Loescher, T. McCamant, J.D. Keller, Carbohydrate reserves, translocation, and storage in woody plant roots, Hort. Sci. 25 (1990) 274–281. [41] H. Zheng, L.A. Staehelin, Protein storage vacuoles are transformed into lytic vacuoles in root meristematic cells of germinating seedlings by multiple, cell type-specific mechanisms, Plant Physiol. 155 (2011) 2023–2035. [42] H. Yamaguchi, Y. Kanayama, Changes in the amounts of the NAD-dependent sorbitol dehydrogenase and its involvement in the development of apple fruit, J. Am. Soc. Hort. Sci. 121 (1996) 848–852. [43] C. Laurière, M. Laurière, J. Daussant, Immunohistochemical localization of ␤amylase in resting barley, Physiol. Plantarum 67 (1986) 383–388. [44] T. Rorat, J. Sadowski, F. Grellet, J. Daussant, M. Delseny, Characterization of cDNA clones for rye endosperm-specific ␤-amylase and analysis of ␤-amylase deficiency in rye mutant lines, Theor. Appl. Genet. 83 (1991) 257–263. [45] I. Hara-Nishimura, T. Kinoshita, N. Hiraiwa, M. Nishimura, Vacuolar processing enzymes in protein-storage vacuoles and lytic vacuoles, J. Plant Physiol. 152 (1998) 668–674.