Turnover of LeACS2, a wound-inducible 1-aminocyclopropane-1-carboxylic acid synthase in tomato, is regulated by phosphorylation/dephosphorylation

Authors


(fax +81 52 789 4167; e-mail morihito@agr.nagoya-u.ac.jp).

Summary

1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) is the rate-limiting enzyme of the ethylene biosynthesis pathway. ACS is regulated both transcriptionally and post-translationally. We previously reported that LeACS2, a wound-inducible ACS in tomato (Solanum lycopersicum), is phosphorylated in vivo, and suggested that phosphorylation regulates protein stability rather than enzymatic activity. In this report, we demonstrate that phosphorylation/dephosphorylation of LeACS2 regulates its turnover upstream of the ubiquitin-26S-proteasome degradation pathway. Pulse–chase experiments coupled with treatment with protein kinase/phosphatase inhibitors demonstrated that LeACS2 is stabilized by phosphorylation and degraded after dephosphorylation. The amount of LeACS2 affected by the protein kinase/phosphatase inhibitors significantly influenced cellular ACS activity, ACC content, and ethylene production levels in tomato fruit tissue, suggesting that post-translational regulation by phosphorylation plays an important role in the control of ethylene production. A calcium-dependent protein kinase (CDPK), LeCDPK2, was isolated as one of the protein kinases that are able to phosphorylate LeACS2 at Ser-460. LeACS2 was immediately phosphorylated after translation by CDPK and mitogen-activated protein kinase at different sites in response to wound signaling and almost all functional LeACS2 molecules are phosphorylated in the cell. Phosphorylation at both sites was required for LeACS2 stability.

Introduction

Ethylene influences numerous plant growth and developmental processes, including germination, cell elongation, leaf and flower senescence and abscission, fruit ripening, and the responses to biotic and abiotic stresses (Abeles et al., 1992). Ethylene is biosynthesized from methionine via S-adenosyl-l-methionine (SAM) and 1-aminocyclopropane-1-carboxylic acid (ACC). The conversion of SAM to ACC is the first committed and rate-limiting step in ethylene biosynthesis, which is catalyzed by ACC synthase (ACS; EC 4.4.1.14) (Yang and Hoffman, 1984). ACS is encoded by a multi-gene family and each isozyme is differentially regulated at the transcriptional level by various developmental, environmental, and hormonal factors (Kende, 1993; Zarembinski and Theologis, 1994). Although transcriptional regulation of ACS is a first determinant of the amount of ethylene produced (Kende, 1993) and has been investigated in detail (Van der Straeten et al., 1992; Nakatsuka et al., 1998; Barry et al., 2000; Ge et al., 2000; Llop-Tous et al., 2000; Tsuchisaka and Theologis, 2004), ACS is also regulated at the post-translational level (Chae and Kieber, 2005; Argueso et al., 2007). Previous reports suggested the possibility of post-translational regulation of ACS. In suspension-cultured cells, the increase in ACS activity observed in response to an elicitor is not sensitive to RNA transcription inhibitors (Chappell et al., 1984; Felix et al., 1991). Furthermore, discrepancies between the expression level of ACS mRNA and ACS activity or the rate of ethylene production are observed in tomato suspension cells treated with an elicitor (Oetiker et al., 1997), etiolated seedlings of ethylene-overproducing mutants of Arabidopsis (eto) (Vogel et al., 1998; Woeste et al., 1999), and cytokinin-treated etiolated seedlings of Arabidopsis (Vogel et al., 1998).

Analyses of eto mutants have contributed to our understanding of the post-translational regulation of ACS. The eto2 mutation results from a single base-pair insertion that changes 12 amino acids in the carboxy-terminal (C-terminal) region of AtACS5 (Vogel et al., 1998), thereby increasing AtACS5 stability (Chae et al., 2003). The eto3 mutation is the result of a single amino acid change, Val to Asp, at the C-terminal region of AtACS9 (Chae et al., 2003). Analyses of eto1 revealed that ETO1 is involved in AtACS5 stability in Arabidopsis (Wang et al., 2004). ETO1 and its homologues, EOL1 and EOL2, have a BTB (Broad-complex, Tramtrack, Bric-à-brac) domain in the N-terminal region and six predicted TPR (tetratricopeptide repeat) domains that are involved in diverse protein–protein interactions (Lupas, 1996; Collins et al., 2001; D’Andrea and Regan, 2003). ETO1 recognizes and directly interacts with the C-terminal region of AtACS5 and suppresses ACS activity, allowing CUL3, a component of ubiquitin E3 ligase, to bind AtACS5, which leads to AtACS5 degradation via the ubiquitin-26S-proteasome pathway (Wang et al., 2004). The half-life of myc-tagged AtACS5 is remarkably prolonged in eto1 compared with wild-type (Chae et al., 2003). These findings suggest that the C-terminal region of ACS is involved in ACS stability. A recent report clearly demonstrated that the C-terminal region of ACS is sufficient for regulating ACS stability (Joo et al., 2008).

Phosphorylation is likely to be involved in the regulation of ACS stability. Treatment with a protein phosphatase inhibitor remarkably enhances the induction of ACS activity, whereas treatment with a protein kinase inhibitor blocks ACS induction in tomato suspension cells treated with an elicitor (Spanu et al., 1994) and in tomato leaves treated with ozone (Tuomainen et al., 1997). In a previous report, we demonstrated that LeACS2, a wound-inducible ACS in tomato, is phosphorylated at Ser-460 in the C-terminal region in tomato fruit tissue and the phosphorylation of recombinant LeACS2 by a crude kinase fraction depends on the available calcium (Tatsuki and Mori, 2001). The Ser residue and its flanking sequence, (F/L) RLS (F/L), are highly conserved in ACS isozymes and comprise the motif for phosphorylation by a calcium-dependent protein kinase (CDPK) (Sebastia et al., 2004; Figure S1). On the other hand, genetic and biochemical studies revealed that AtACS6 is phosphorylated by mitogen-activated protein kinase (MAPK) at three Ser residues, which are different from the CDPK phosphorylation site in the C-terminal extended region (Liu and Zhang, 2004). Putative MAPK phosphorylation sites are conserved in a subset of ACS isozymes (Liu and Zhang, 2004; Figure S1). As both CDPK and MAPK are activated in response to extracellular stimuli, they may function cooperatively to regulate ACS and thus ethylene production. ACS phosphorylation is likely involved in protein stability rather than enzymatic activity or substrate specificity (Tatsuki and Mori, 2001; Liu and Zhang, 2004).

Here, we show the post-translational regulatory mechanism of LeACS2 by phosphorylation in wounded tomato fruit tissue. Biochemical analyses revealed that phosphorylation/dephosphorylation of LeACS2 regulates its turnover upstream of the degradation system of the ubiquitin-26S-proteasome. This regulation plays a role in the control of ethylene production in the tissue. Further, we provide evidence that LeACS2 is phosphorylated by CDPK and MAPK at different sites in response to a wounding signal and phosphorylation at both sites is required for LeACS2 stability. In addition, we isolated a CDPK in tomato, named LeCDPK2, as one of the protein kinases that are able to phosphorylate LeACS2 at Ser-460.

Results

We first examined whether ACC levels and ACS activity are affected by treatment with protein kinase/phosphatase inhibitors in the turning-color tomato fruit tissue. ACC content and total ACS activity in the fruit tissue increased in response to wounding, as previously reported (Kende and Boller, 1981). Compared with mock treatment, treatment with the protein kinase inhibitor K252a clearly decreased the ACC content and ACS activity (Figure 1). On the other hand, treatment with a protein phosphatase inhibitor, okadaic acid, remarkably increased both the ACC content and ACS activity (Figure 1). The levels of ethylene production paralleled ACC content after each treatment (Figure S2). We previously reported that LeACS2, a major ACS isozyme in wounded tomato fruit tissue, is phosphorylated in vivo (Tatsuki and Mori, 2001). We therefore expected total ACS activity to be affected by treatments with inhibitors due to changes in the LeACS2 phosphorylation state in the tissue. To examine the effects of protein kinase/phosphatase inhibitors on LeACS2 accumulation, we performed western blot analyses with anti-whole LeACS2 antibody (α-LeACS2). Consistent with total ACS activity, treatment with K252a decreased, whereas treatment with okadaic acid increased the amount of LeACS2 protein (Figure 2a,b). K252a and okadaic acid did not substantially affect LeACS2 transcript expression levels (Figure 2c), indicating that the inhibitors affect the accumulation of LeACS2 protein at the post-translational level. To detect the LeACS2 phosphorylation state, we raised an antibody that recognizes phosphorylated LeACS2 using a phospho-peptide corresponding to the Lys-454 to Tyr-466 sequence as the antigen. The purified antibody specifically recognized the recombinant LeACS2 phosphorylated by a crude kinase fraction of wounded tomato fruit tissue (Figure 2d) (Tatsuki and Mori, 2001). Proteins extracted from fruit tissue under denaturing conditions to prevent proteolysis of the C-terminal region of LeACS2 were subjected to western blot analyses with α-P-LeACS2. The findings confirmed that K252a inhibited LeACS2 phosphorylation and okadaic acid maintained the LeACS2 phosphorylation state (Figure 2b, middle). Thus, the LeACS2 phosphorylation state closely correlated with the amount of LeACS2 accumulation. Therefore, phosphorylation/dephosphorylation inhibitors influence the amount of LeACS2, which in turn affects the ACC content, ACS activity, and ethylene production levels.

Figure 1.

 ACC content and ACS activity in wounded tomato fruit tissue treated with protein kinase/phosphatase inhibitors.
Tomato fruit disks were prepared and treated with 0.5 μm K252a or okadaic acid for the indicated times. ACC content in the crude extract was determined (a) and the same samples were desalted and used to determine ACS activity (b).
Circle, DMSO; triangle, K252a; square, okadaic acid. Values shown are means ± SD (n = 7).

Figure 2.

 Effects of protein kinase/phosphatase inhibitors on the accumulation of LeACS2 in wounded tomato fruit tissue.
(a) Accumulation of LeACS2 in wounded fruit tissue treated with DMSO, 0.5 μm okadaic acid (OA), and 0.5 μm K252a. Fruit disks were treated with inhibitors and incubated for 5 h. Proteins were extracted and analyzed with anti-LeACS2 antibody (top). A representative experiment from three independent experiments is shown. Coomassie brilliant blue staining shows equal loading of total protein (bottom). The relative amount of LeACS2 is indicated. Values shown are means ± SD (n = 3).
(b) Time-course analysis of LeACS2 accumulation. Fruit disks were prepared as described in (a) and incubated for the indicated times (hours), followed by analysis with anti-LeACS2 antibody (top) or anti-phospho-LeACS2 antibody (middle). Coomassie brilliant blue staining shows equal loading of total protein (bottom).
(c) Number of transcripts in wounded pericarp tissue treated with DMSO, okadaic acid, or K252a. LeACS2 expression was quantified with real-time PCR. Accumulation levels of the transcripts are given as the copy number of mRNA per picogram total RNA. Real-time PCR was performed in quadruplicate, and values shown are means ± SD (n = 4).
(d) Characterization of anti-phospho-LeACS2 antibody. Recombinant Trx–LeACS2 was subjected to phosphorylation reactions in the presence (+) or absence (−) of a crude kinase fraction of wounded tomato fruit tissue, which was followed by western blot analysis with anti-phospho-LeACS2 antibody (top). Silver staining shows loading amounts of Trx–LeACS2 (bottom).

Regulation of LeACS2 turnover was then analyzed by pulse-chase experiments. LeACS2 was immunoprecipitated with α-LeACS2. A competition experiment with recombinant LeACS2 indicated that α-LeACS2 specifically immunoprecipitated LeACS2 (data not shown). After synthesis of 35S-labeled LeACS2, 35S-labeled amino acids were chased using excess amounts of cold amino acids. Sample proteins were extracted and LeACS2 was immunoprecipitated. Under control conditions, the radioactivity of LeACS2 rapidly diminished, with a half-life of 70 min (Figure 3). The half-life of LeACS2, however, was shortened by protein kinase inhibitors in the chase solution (K252a, 35 min; staurosporine, 40 min), whereas it was remarkably prolonged by protein phosphatase inhibitors (okadaic acid, 150 min; calyculin A, 210 min). These results clearly indicate that phosphorylated LeACS2 is stable and non-phosphorylated LeACS2 is unstable, and that LeACS2 turnover, at least through wounding, is regulated by phosphorylation.

Figure 3.

 LeACS2 turnover in the presence of protein kinase/phosphatase inhibitors.
35S-Labeled amino acids (Met and Cys) were applied to wounded pericarp tissue for 2 h. 35S-Labeled amino acids were chased by excess amounts of cold amino acids containing 0.5 μm protein kinase inhibitors [K252a, staurosporine (STS)] or protein phosphatase inhibitors [okadaic acid (OA), calyculin A (CLA)]. DMSO was used as a control. Proteins were extracted and LeACS2 was immunoprecipitated with anti-LeACS2 antibody coupled with Dynabeads protein A. Samples were subjected to SDS-PAGE and radioactivity of the labeled LeACS2 was detected (a).
Chase times (min) are indicated above each figure.
The radioactivity of LeACS2 was quantified (b).
circle, DMSO; diamond, K252a; triangle, staurosporine; square, okadaic acid; cross, calyculin A. Values shown are means ± SD (n = 3). Half-lives of LeACS2 in the cases of DMSO (control), K252a, and okadaic acid are indicated by the dotted line with an arrow.

We next isolated the LeACS2 kinases that phosphorylate LeACS2 at Ser-460 by an expression screening using α-P-LeACS2. Screening of 7.5 × 105 clones led to the isolation of three positive clones, all with the same CDPK sequence. Because two CDPK genes have already been reported: LeCDPK1 (Chico et al., 2002) and LeCPK1 (Rutschmann et al., 2002), we named these CDPKs LeCDPK2. The nucleotide sequence data has been deposited in DDBJ/EMBL/GenBank under accession number AB530160. LeCDPK2 encoded 581 amino acids with a predicted molecular mass of 64 602 Da. LeCDPK2 was classified into Group I of the CDPK family (Cheng et al., 2002; Figure S3). In contrast, LeCDPK1 and LeCPK1 are classified as Group II, and have 59 and 57% identity at the amino acid level, respectively. The closest homologue of LeCDPK2 is NtCDPK2 (93% amino acid identity), which is involved in plant defense signaling (Romeis et al., 2001). To analyze the biochemical characteristics of LeCDPK2, recombinant NusA fused–CDPK was expressed in Escherichia coli. NusA–LeCDPK2 phosphorylated LeACS2 and was inhibited by 2 mm EGTA (7.9% activity remained compared to the activity in the presence of 100 μm Ca2+, which was defined as 100%), Ser/Thr protein kinase inhibitors [100 nm K252a (36.5%) and 100 nm staurosporine (37.2%)], and calmodulin inhibitors [100 μm trifluoperazine (35.2%), 100 μm calmidazolium (CMZ) (1.7%), and 500 μm compound 48/80 (19.2%)] (Chang et al., 1995). NusA–LeCDPK2 also phosphorylated LeACS1A, LeACS3, and LeACS6, which possess putative CDPK phosphorylation sites, but not LeACS4, which does not contain a phosphorylation site (Figure 4a; Figure S1). Furthermore, we confirmed that three CDPKs (LeCDPK1, LeCDPK2, and LeCPK1) could phosphorylate the Ser-460 of LeACS2, based on the following data (Figure 4b); three CDPKs phosphorylated Trx–LeACS2 (top), but not the mutant Trx–LeACS2 (Ser-460 was replaced by Gly) (middle), and three Trx–LeACS2s phosphorylated by three CDPKs were recognized by α-P-LeACS2, respectively (bottom). The apparent Km values of the C-terminal peptide of LeACS2 for each LeCDPK were the same (data not shown).

Figure 4.

In vitro characterization of LeCDPK2.
(a) Substrate specificity of LeCDPK2. Trx–ACS isozymes from tomato were phosphorylated by NusA–LeCDPK2 in the presence of [γ-32P]-ATP and radioactivity of the phosphorylated ACSs was detected (top). Coomassie brilliant blue staining shows each Trx–ACS as a substrate (bottom).
(b) LeACS2 phosphorylation by each tomato CDPK. Trx–LeACS2 (WT, top) or mutated Trx–LeACS2 (S460G, middle) was phosphorylated by NusA–LeCDPK1, NusA–LeCPK1, or NusA–LeCDPK2 and radioactivity of the phosphorylated LeACS2 was detected. Each phosphorylated LeACS2 by each tomato CDPK was also detected with the anti-phospho-LeACS2 antibody (bottom). The amount of Trx–LeACS2 in each reaction is indicated at the bottom of the figure.

Western blot analyses with α-P-LeACS2 revealed that the accumulation patterns of phospho-LeACS2 were similar to that of total LeACS2 (Figure 2b). Therefore, we speculated that LeACS2 is immediately phosphorylated after translation and functions in its phosphorylated form in the cell. To test this hypothesis, we examined the ratio of the phosphorylated/dephosphorylated forms of LeACS2. Precipitated 35S-labeled LeACS2 was dephosphorylated by λ-protein phosphatase and subjected to phosphate affinity SDS-PAGE, in which the polyacrylamide gel contained Phos-tag (Kinoshita et al., 2006), a molecule that interacts with phosphate. LeACS2 dephosphorylated by λ-protein phosphatase migrated faster than the untreated sample (Figure 5, lanes 1, 2). No signal was detected at the position of the dephosphorylated form in the untreated sample (Figure 5, lane 1), indicating that almost all detectable LeACS2 molecules accumulated in its phosphorylated form. This finding indicates that LeACS2 is immediately phosphorylated after translation and immediately degraded after dephosphorylation. Because LeACS2 possesses putative MAPK phosphorylation sites in addition to the CDPK phosphorylation site, we examined whether the wounding signal induces LeACS2 phosphorylation by both MAPK and CDPK. To analyze this, we used three kinds of recombinant protein kinases; LeCDPK2, StMPK1 (salicylic acid-inducible MAPK in potato), and StWIPK (wound-inducible MAPK in potato) (Katou et al., 2005). StMPK1 and StWIPK were previously activated by constitutively active MEK2 (MAPKK in tobacco). Dephosphorylated LeACS2 was then phosphorylated by LeCDPK2, StMPK1, or StWIPK, and each sample was subjected to phosphate affinity SDS-PAGE. Phosphorylation by LeCDPK2, StMPK1, StWIPK, and StMPK1/StWIPK shifted the LeACS2 signal (Figure 5, lanes 3–5, and 8, respectively). Furthermore, LeCDPK2 and StMPK1/StWIPK phosphorylation combinations shifted the LeACS2 signal to the same position as native LeACS2 (Figure 5, lanes 6 and 7), suggesting that LeACS2 is simultaneously phosphorylated by CDPK and MAPK at different sites. Phosphate affinity SDS-PAGE can separate phosphorylated proteins depending on the number of phosphorylated sites (Kinoshita et al., 2006). The migration of LeACS2 phosphorylated by MAPK, however, was the same as that of LeACS2 phosphorylated by CDPK, although the LeACS2 sequence possesses three putative MAPK phosphorylation sites (Figure S1). The migration could be dependent not only on the number of phosphorylated sites, but on the conformation as well.

Figure 5.

 LeACS2 phosphorylation state in wounded tomato fruit tissue.
35S-Labeled LeACS2 conjugated with anti-LeACS2 antibody–Dynabeads protein G was prepared as described in Figure 3(a). LeACS2 was dephosphorylated by λ-protein phosphatase. Dephosphorylated LeACS2 was again phosphorylated by LeCDPK2, activated StMPK1, and activated StWIPK, as indicated. Samples were subjected to phosphate affinity SDS-PAGE. (−) and (+) indicate cathode and anode ends of the gel, respectively.

To determine the contribution of phosphorylation by CDPK or MAPK to LeACS2 stability, we examined the effects of CMZ and U0126 on LeACS2 accumulation. Because there is no specific MAPK inhibitor, we used U0126, a MAPKK inhibitor, as a MAPK inhibitor. Wounded fruit tissues that had been treated with inhibitors were incubated with 35S-labeled amino acids. After 5-h feeding, 35S-labeled LeACS2 was immunoprecipitated. Although inhibitor treatments did not change the expression levels of the LeACS2 transcripts (Figure S4a), the inhibitors clearly decreased the LeACS2 protein accumulation to the same extent as K252a (Figure 6a). The CDPK activity was effectively inhibited by CMZ, but not by U0126 (Figure S4b). The MAPK activity was effectively inhibited by U0126, but not by CMZ (Figure S4c). We then performed similar experiments in which 35S-labeled amino acids were replaced by 32P-labeled phosphoric acid to examine the phosphorylation state of LeACS2 in tissue treated with U0126 or CMZ. Inhibitors decreased the phosphorylation signal derived from 32P-labeled LeACS2 (Figure 6b). Each 32P-labeled LeACS2 was excised from the gel and digested with trypsin. The resultant peptides were fractionated with reversed-phase HPLC. We first confirmed that the phospho-peptide containing the CDPK-site (LSFSK) and another phospho-peptide containing the MAPK sites (MYDESVLSPLSSPIPPSPLVR), which were derived from the phosphorylated recombinant Trx–LeACS2 as substrates for CDPK and MAPK, respectively, were clearly separated (Figure 6c, strips 4, 5, 6) and that no other phospho-signal was detected in any of the fractions. LeACS2 in the tissue treated with DMSO was phosphorylated by CDPK and MAPK (Figure 6c, strip 3). Interestingly, even when treated with U0126 or CMZ, phosphorylation signals were detected from both peptides (Figure 6c, strips 1, 2), suggesting that the LeACS2 shown in Figure 6(b) is phosphorylated at both sites. This finding indicates that LeACS2 phosphorylated at a single site by a single kinase is susceptible to degradation. The LeACS2 signal shown in Figure 6(b) would be due to the remaining molecules on which both inhibitors failed to act completely.

Figure 6.

 The effects of CDPK and MAPKK inhibitors on LeACS2 stability.
(a) The effects of calmidazolium (CMZ) and U0126 on LeACS2 accumulation. Fruit disks were treated with 100 μm U0126, 100 μm CMZ, or 1 μm K252a immediately after preparation. After 1-h incubation, 35S-labeled amino acids were applied to disks for 5 h. LeACS2 was immunoprecipitated, followed by SDS-PAGE and detection of radioactivity of LeACS2. A representative experiment from four independent experiments is shown on the left. The relative radioactivity of LeACS2 is indicated on the right. Values shown are means ± SD (n = 4).
(b,c) The effects of CMZ and U0126 on phosphorylation of LeACS2. (b) Similar experiments as (a) were conducted using 32P-labeled phosphoric acid instead of 35S-labeled amino acids. After SDS-PAGE, radioactivity of the phosphorylated LeACS2 was detected. A representative experiment from three independent experiments is shown. Excess amount of Trx–LeACS2 was added to a DMSO-treated sample as a competitor to confirm that the radioactivity derived from LeACS2 (compe.).
(c) The 32P-labeled LeACS2 shown in (b) (in vivo) or recombinant Trx–LeACS2 phosphorylated in the presence of [γ-32P]-ATP by LeCDPK2, activated StWIPK (MAPK), and both kinases (in vitro) were digested with trypsin. Resultant peptides were separated with reversed-phase HPLC and each fraction was spotted on μFocus MALDI plates. The radioactivity of the phospho-peptides was detected. The C-terminal region of LeACS2 sequence (455–485) and a putative fragmentation pattern with trypsin are indicated at the bottom. Asterisk indicates CDPK phosphorylation site and arrowheads indicate putative MAPK phosphorylation sites.

Both CDPK and MAPK are defense-related kinases. To determine the correlations between LeACS2 induction and increases in kinase activity, the time-course changes in CDPK and MAPK activity were examined in wounded fruit tissue (Figure S4b,c). A certain level of kinase activity was detected prior to wounding. In response to wounding, total CDPK and MAPK activity increased and reached a maximum at 60 min and 10 min, respectively, and then gradually decreased to steady-state levels.

An alternative important regulator of ACS is its degradation via the ubiquitin-26S-proteasome pathway. To examine whether LeACS2 is degraded through proteasomes like AtACS5 (Wang et al., 2004) and AtACS6 (Joo et al., 2008), we analyzed the effect of the proteasome inhibitor MG132 on LeACS2 turnover. The amount of LeACS2 in the fruit tissue increased in an MG132 dose-dependent manner and became six-fold higher than that of controls with 500 μm MG132 treatment (Figure 7a). Furthermore, LeACS2 turnover was prolonged by 500 μm MG132 (Figure 7b). These findings suggest that the degradation mechanism via the ubiquitin-26S-proteasome pathway lies downstream of the regulation by phosphorylation.

Figure 7.

 The effect of a proteasome inhibitor on LeACS2 turnover.
(a) Accumulation of LeACS2 in wounded pericarp tissue treated with MG132. Fruit disks were treated with MG132 at the indicated concentrations and incubated for 5 h. Proteins were extracted and analyzed with anti-LeACS2 antibody.
(b) LeACS2 turnover affected by MG132. Pulse–chase experiments were performed using the same procedure as described for Figure 3. MG132 was added to the chase solution at the indicated concentrations. The extract was subjected to immunoprecipitation with anti-LeACS2 antibody. Chase times (min) are indicated at the bottom of the figure.

Discussion

Ethylene has several important roles in plant growth, development, and defense responses. Thus, ACS, a rate-limiting enzyme of ethylene biosynthesis, must be tightly regulated and rapidly degraded after completing its function. Several previous analyses of ACS turnover indicated that ACS is a short-lived protein (Kende and Boller, 1981; Spanu et al., 1990; Yoshii and Imaseki, 1982). Because the content of authentic ACS is very low in plant tissue due to its rapid degradation, post-translational regulatory mechanisms of ACS are usually analyzed under the control of a synthetic/artificial inducible promoter. In the present study, we succeeded in detecting authentic ACS in a dynamic state using radiolabeled molecules. The pulse–chase experiment showed that the half-life of LeACS2 was approximately 70 min in wounded fruit tissue at the turning stage (Figure 3). Kim and Yang (1992) estimated the half-life of LeACS2 to be 48 min and 58 min in two separate pulse–chase experiments. In Arabidopsis, myc-AtACS5 (Chae et al., 2003) and myc-AtACS6 (Joo et al., 2008) are also degraded rapidly in a DEX-inducible system coupled with cycloheximide, with half-lives of 15 min and 60 min, respectively. Although there are some differences in the ACS half-lives among these reports, probably due to differences in experimental conditions or materials used, the results commonly show that ACS is rapidly turned over. A rapid ACS degradation mechanism would be important for controlling the effects of ethylene on plants, in addition to an elaborate regulation of ACS transcription (Tatsuki and Mori, 1999), rapid degradation of the ethylene receptor ETR2 (Chen et al., 2007) and ethylene-regulated transcription factor EIN3 (Guo and Ecker, 2003).

Rapid and stringent regulation of ACS turnover would be accomplished by regulating phosphorylation. The present study demonstrated the importance of phosphorylation in ACS turnover. Inhibition of protein kinases decreased the half-life of LeACS2 and thus decreased the amount of LeACS2 in tomato fruit tissue, whereas inhibition of protein phosphatases had the opposite effect (Figure 3). Phospho-mimicking AtACS6 at the MAPK phosphorylation site, myc-AtACS6DDD, is slowly degraded compared with wild-type, whereas myc-AtACS6AAA is rapidly degraded in Arabidopsis (Joo et al., 2008). These findings indicate that ACS is stabilized by phosphorylation and destabilized by dephosphorylation in vivo. In addition, phosphate affinity SDS-PAGE revealed that almost all detectable LeACS2 molecules that accumulated in the cell were phosphorylated (Figure 5). Taken together, these findings suggest that the regulatory mechanism of LeACS2 is as follows; after translation LeACS2 is immediately phosphorylated and functions in its phosphorylated form. After completing its function, LeACS2 is dephosphorylated and immediately degraded. Thus, phosphorylation/dephosphorylation is a crucial step in the post-translational regulation of LeACS2. Furthermore, inhibitor-induced alterations of LeACS2 turnover rates directly affected ACS activity, ACC content, and ethylene production (Figure 1), demonstrating that post-translational regulation of ACS is important for the control of ethylene production as well as transcriptional regulation.

We examined the contribution of phosphorylation by both CDPK and MAPK to LeACS2 stability. An inhibitor of each kinase decreased the LeACS2 accumulation in wounded fruit tissue (Figure 6a). Nevertheless, we did not detect any LeACS2 molecule that was phosphorylated at only a single site (Figure 6c), suggesting that both phosphorylation sites are required for LeACS2 stability. Analysis with phosphate affinity SDS-PAGE supports this hypothesis; no partially phosphorylated LeACS2 was detected (Figure 5). The contribution of MAPK phosphorylation sites to ACS stability has already been demonstrated (Joo et al., 2008). We extended this finding by revealing the involvement of the CDPK phosphorylation site in ACS stability.

We were unable to specify the actual LeACS2 kinase, although we isolated LeCDPK2 as one of the protein kinases that are able to phosphorylate LeACS2 at Ser-460 by expression screening using α-P–LeACS2 because other CDPKs (LeCDPK1 and LeCPK1), which are classified as a different clade of the CDPK family (Cheng et al., 2002; Figure S3), also phosphorylate LeACS2 as well as LeCDPK2 in vitro. Thus, any of the CDPK isozymes might be able to phosphorylate ACS isozymes.

CDPK and MAPK belong to a family of kinases involved in plant defense responses. LeACS2 would be phosphorylated by wound-inducible kinases. Several CDPKs are transcriptionally induced in response to wounding (Botella et al., 1996; Yoon et al., 1999; Chico et al., 2002). In addition, an increase in the cytosolic calcium concentration in response to wounding appears to be crucial for the activation of CDPK (Knight and Knight, 2001). In tomato leaves, the increase in LeCDPK1 mRNA upon wounding correlates with an increase in the activity of soluble CDPK (Chico et al., 2002). On the other hand, MAPK is also activated by wounding via activation of the MAPK phosphorylation cascade (Zhang and Klessig, 2001). In tomato, LeMPK1 and LeMPK2 are activated post-translationally and LeMPK3 is transcriptionally induced in response to wounding (Mayrose et al., 2004; Higgins et al., 2007). These results suggested that LeACS2 is actually phosphorylated by these kinases in vivo.

ACS isozymes are classified into three groups based on the C-terminal sequence (Argueso et al., 2007; Yoshida et al., 2005; Figure S1). Type-1 ACS possesses an extended C-terminal region containing CDPK and MAPK phosphorylation sites (e.g. LeACS2, AtACS6), whereas type-2 ACS possesses only a CDPK phosphorylation site (e.g. AtACS5, LeACS3). Type-3 lacks a phosphorylation site altogether. In this study, we showed that LeACS2, which is a type-1 ACS, is regulated by phosphorylation and requires phosphorylation at both sites by CDPK and MAPK for its stability. Joo et al. (2008) also reported that multiple phosphorylations in the C-terminal region are required for AtACS6 stabilization. On the contrary, there is no direct evidence indicating that type-2 ACS requires phosphorylation for its stability. The eto3 mutant provides clues to the importance of phosphorylation for type-2 ACS stability (Chae et al., 2003). Because the eto3 mutation, Val to Asp in AtACS9, occurs very close to the CDPK phosphorylation site, the addition of a negative charge to the C-terminal region of AtACS9 would mimic a constitutive phosphorylation state and thus increase AtACS9 stability. Thus, phosphorylation at a single site by CDPK may be sufficient to stabilize type-2 ACS. Although type-1 and -2 ACSs commonly seem to be regulated by phosphorylation, there is a clear difference between them in the region downstream of the regulation by phosphorylation. In Arabidopsis, type-2 ACS degradation is mediated by the ubiquitin-26S-proteasome degradation pathway, using a BTB E3 ligase assembled with ETO1, EOL1, or EOL2 (Wang et al., 2004; Christians et al., 2009). ETO1 and EOL1 interact with type-2 ACSs (AtACS4, AtACS5, and AtACS9), but not with type-1 ACS (Christians et al., 2009). Yoshida et al. (2006) also reported that a homologue of ETO1 in tomato, LeEOL1, interacts with LeACS3, but not LeACS2. LeEOL1 specifically recognizes the WVF motif and R/D/E-rich region, which are conserved in the C-terminal region of type-2 ACS. On the other hand, the turnover of LeACS2 (Figure 7) and AtACS6 (Joo et al., 2008) was delayed by MG132 treatment, similar to AtACS5. Ubiquitination of type-1 ACS might require another unidentified component of E3 ligase. In any case, it is possible that phosphorylation modulates the interaction between the type-1 and -2 ACSs and some E3 ligase complex.

The regulatory mechanism of type-3 ACS remains to be elucidated. Because type-3 ACS lacks a C-terminal region involved in the post-translational regulation, its turnover would not be regulated stringently. Interestingly, type-3 ACSs include LeACS4 that is expressed only during fruit ripening in tomato and sex determination-related ACSs in Cucurbitaceae, CmACS-7 (melon) (Boualem et al., 2008) and CsACS2 (cucumber) (Boualem et al., 2009; Li et al., 2009), suggesting that type-3 ACSs related to plant programmed development do not require a rapid protein inactivation system to function. On the other hand, almost all type-1 and -2 ACSs are induced by biotic and abiotic stresses irregularly. The stringent regulation of theses ACSs by phosphorylation would be an important system to stop ethylene production immediately when the stresses are removed. Thus, the C-terminal region of ACS might differentiate into appropriate forms for the specific role of each isozyme.

Experimental procedures

Materials and treatment

Tomato fruit (Solanum lycopersicum) in the turning stage was obtained from a local market or harvested at the farm of Nagoya University in Aichi. Fruit pericarp disks (1 cm diameter × 5 mm thick) were treated with protein kinase/phosphatase inhibitors or MG132 by vacuum infiltration. Treated disks were incubated in high humidity at 25°C and frozen at −80°C until use.

ACS activity assay and determination of ACC content

ACS activity was assayed as described previously (Yoshii and Imaseki, 1982). ACC content was determined using the method of Lizada and Yang (1979). One unit of enzyme converts SAM to ACC at 1 nmol h−1 at 30°C.

Antibody preparation

The antibody for the detection of LeACS2 was prepared as described previously (Tatsuki and Mori, 2001). Phospho-peptide (NH2-CKNNLRLpSFSKRMY-OH) was synthesized based on the C-terminal sequence of LeACS2 flanking the phosphorylation site (Ser-460), corresponding to the Lys-454 to Tyr-466 sequence. A Cys-residue was added to the N-terminus of the peptide to conjugate with bovine serum albumin via m-maleimidobenzoyl-N-hydroxysuccinimide ester. A rabbit was immunized with phospho-peptide conjugated bovine serum albumin by multiple intradermal injections. The rabbit’s serum was applied to a non-phospho-peptide conjugated column and the flow-through fraction was applied to the phospho-peptide column. Bound IgG was eluted with 0.1 m glycine–HCl (pH 2.5) and immediately adjusted to pH 8.0.

Protein extraction under denaturing conditions

Fruit disks were homogenized in an equal volume of extraction buffer [200 mm Tris–HCl (pH 9.0), 0.2 m LiCl, 5 mm EDTA, 1% SDS] and a two-fold volume of Tris-saturated phenol (pH 9.0). The homogenate was centrifuged at 6000 g for 15 min. The phenol fraction was recovered and added to a four-fold volume of methanol containing 0.1 m ammonium acetate and put on ice for 30 min. Precipitates were centrifuged (6000 g for 15 min at 4°C), and washed twice with methanol containing 0.1 m ammonium acetate and once with acetone. Proteins were dissolved in 2× Laemmli-sample buffer [125 mm Tris–HCl (pH 6.8), 2% SDS, 2% 2-mercaptoethanol]. Protein concentration was determined using the RC DC protein assay (Bio-Rad, http://www.bio-rad.com/) with γ-globulin as the standard.

Western blot analysis

Proteins were separated using SDS-PAGE (10% acrylamide gels) and blotted onto nitrocellulose membranes (BA-S 85; Whatman, http://www.whatman.com/). The membrane was blocked with 5% dried skimmed milk and 0.05% Tween 20 in Tris-buffered saline [50 mm Tris–HCl (pH 8.0), 150 mm NaCl]. Purified anti-LeACS2 antibody or anti-phosphorylated-LeACS2 antibody was used at a concentration of 50 μg ml−1. The membrane was washed with 0.05% Tween 20 in Tris-buffered saline and then reacted with horseradish peroxidase-conjugated goat anti-rabbit IgG (Pierce Chemical, http://www.piercenet.com/) at a dilution of 1:1000. Signals were detected using Super Signal West Femto (Pierce Chemical).

RNA extraction and real-time PCR

Fruit disks were homogenized in a 2.5-fold volume of Plant RNA Purification Reagent (Invitrogen, http://www.invitrogen.com/). Extraction procedures were performed according to the manufacturer’s instructions. Total RNA (50 ng) was used for real-time PCR quantification. Real-time PCR was performed with One Step SYBR ExScript RT-PCR Kit (TaKaRa, http://www.takara-bio.com/). The following set of primers was used to amplify DNA fragments of LeACS2 (sense, 5′-GAGGTTCGTAGGTGTTGAGAAAAGT-3′; antisense, 5′-GGAATAGGTGACGAAAGTGGTGACA-3′).

Pulse–chase experiment

Fruit disks were pre-incubated for 2 h and then 1.85 MBq of the diluted Pro-mix L-[35S] in vitro cell-labeling mix (GE Healthcare Bioscience, http://www.gehealthcare.com) was dropped onto each disk. After a 2-h incubation, disks were dipped in chase solution [150 mm MES (pH 6.8), 50 mm methionine, 50 mm cysteine, and 0.5 μm inhibitor], followed by vacuum infiltration for 15 sec. After incubation for 0–150 min, samples were stored at −80°C until use. Three disks were used for each sample.

In vivo radiolabeling

Fruit disks were treated with kinase inhibitors immediately after preparation. After a 1-h incubation, each disk was incubated with 1.85 MBq of Pro-mix L-[35S] or 12.3 MBq of [32P]Pi (PerkinElmer, http://www.perkinelmer.com/) for 5 h. [32P]Pi was buffered with 10 mm MES (pH 6.8). Three disks were used for each sample.

Immunoprecipitation of LeACS2

Tomato pericarp disks were homogenized in a two-fold volume of extraction buffer [200 mm Tris–HCl (pH 8.0), 150 mm NaCl, 0.1% SDS, Complete™ (Roche Diagnostics, http://www.roche.com)]. After centrifugation (15 000 g for 30 min at 4°C), NP 40 was added to the supernatant (final concentration 1%). The samples were mixed with Dynabeads Protein A or G (Invitrogen) preincubated with anti-LeACS2 antibody. After incubation for 1 h at room temperature, the beads were washed three times with RIPA buffer [50 mm Tris–HCl (pH 8.0), 150 mm NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS]. Bound proteins were eluted with 2× Laemmli-sample buffer and separated with SDS-PAGE. Radio-images were detected using BAS2000 (FUJIFILM, http://www.fujifilm.com).

Protein digestion and peptide fractionation

After SDS-PAGE, a gel band containing LeACS2 was excised, followed by trypsin digestion in the presence of 0.01% Protease MAX surfactant (Promega, http://www.promega.com). Resultant peptides were fractionated using a DiNa nanoHPLC system with reverse phase chromatography (KYA Technologies, http://www.kya-tech.co.jp/) and each 0.3-μl fraction was spotted onto a μFocus MALDI plate (Hudson Surface Technology, http://www.maldiplate.com/) by a DiNa Map system (KYA Technologies). A MALDI plate was then exposed to a BAS2000 imaging plate (FUJIFILM).

Expression cloning of LeACS2 kinase

Screening of kinases that phosphorylate LeACS2 at Ser-460 was performed according to Matsuo et al. (2001). The LeACS2 cDNA fragment that corresponded to the full coding region was cloned into pET-32a (Novagen, http://www.emdchemicals.com/). The plasmid was transformed into E. coli strain BL21 (DE3) pLysS (Novagen). Overnight culture was harvested and infected with the λZAP II cDNA library from wounded ripening tomato fruit. Immuno-screening was performed according to a standard procedure using the anti-phosphorylated-LeACS2 antibody.

In vitro phosphorylation assay by a crude kinase fraction of tomato fruit or the recombinant CDPK

A kinase fraction was prepared using polyethylene glycol, and phosphorylation assays were performed as described previously (Tatsuki and Mori, 2001). Full-length cDNA encoding each CDPK was cloned into pET44a (Novagen). The plasmids were transformed into E. coli BL21 (DE3) pLys (Novagen). NusA–CDPKs with a His-6 tag were affinity-purified with TALON metal affinity resin (Clontech, http://www.clontech.com/). Each 30-μl reaction mixture contained 50 mm HEPES–KOH (pH 8.0), 0.1 mm ATP [or 0.1 mm [γ-32P]-ATP (1.85 Bq pmol−1)], 0.1 mm CaCl2, 10 mm MgCl2, 1 mm dithiothreitol, 5–40 pmol Trx–ACSs, Complete™-EDTA-free (Roche Diagnostics), and 0.05 μg partially purified CDPK fraction or 0.5 μg recombinant NusA–CDPKs. Each reaction was incubated for 30 min at 30°C. EGTA, K252a, staurosporine, trifluperazine, calmidazolium (CMZ), and compound 48/80 were added to analyze the effects of the inhibitors.

In-gel kinase assay

Fruit disks were homogenized in a three-fold volume of extraction buffer [300 mm Tris–HCl (pH 7.9), 10 mm MgCl2, 10 mm 2-mercaptoethanol, Complete™]. The proteins (20 μg) were separated by SDS-PAGE in 10% gels polymerized in the presence of 0.25 mg ml−1 myelin basic protein. An in-gel kinase assay was performed according to Katou et al. (2005).

Dephosphorylation by λ-protein phosphatase and phosphorylation by CDPK or MAPK

Immunoprecipitated 35S-labeled LeACS2 coupled with Dynabeads Protein A, prepared as described above, was incubated in reaction mixture [50 mm Tris–HCl (pH 7.5), 0.1 mm EDTA, 5 mm dithiothreitol, 0.01% Brij 35, 2 mm MnCl2, 10 μm pyridoxal-5′-phosphate, CompleteTM, 50 units of recombinant λ-protein phosphatase (New England BioLabs)] for 15 min at 30°C. After the beads were washed three times with 20 mm Tris–HCl (pH 8.0) to remove the λ-protein phosphatase, the samples were phosphorylated by NusA–LeCDPK2, Trx–StMPK1, or Trx–StWIPK (Katou et al., 2005) in kinase buffer [50 mm HEPES–NaOH (pH 7.5), 10 mm MgCl2, 0.1 mm CaCl2, 0.1 mm ATP, 1 mm dithiothreitol, CompleteTM].

Phosphate affinity SDS-PAGE

Phosphate affinity SDS-PAGE was performed with 7.5% polyacrylamide gels containing 50 μm MnCl2 and 25 μm Phos-Tag (NARD, http://www.nard.co.jp/) according to the manufacturer’s protocol.

Accession numbers

Accession number for LeCDPK2: AB530160.

Acknowledgements

This work was supported in part by Grant-in-Aid for Scientific Research (B) (grants nos. 19380018 and 21380026) to H.M., by Grant-in-Aid for JSPS Fellows (grant no. 19010368) to Y. K., and by Grant-in-Aid for Young Scientists (B) (grant no. 19780031) to M.T. from the Japan Society for the Promotion of Science. We thank H. Yoshioka (Nagoya University) for providing the active recombinant Trx–StMPK1 and Trx–StWIPK proteins.

Ancillary