A Novel C-Terminal Proteolytic Processing of Cytosolic Pyruvate Kinase, Its Phosphorylation and Degradation by the Proteasome in Developing Soybean Seeds

The Plant Journal (2003) 34, 77±93 A novel C-terminal proteolytic processing of cytosolic pyruvate kinase, its phosphorylation and degradation by the proteasome in developing soybean seeds

Guo-Qing Tang1,2, Shane C.Hardin 1,2, Ralph Dewey2 and Steven C.Huber 1,2,Ã 1US Department of Agriculture, Agricultural Research Service, North Carolina State University, Raleigh, NC 27695-7631, USA, and 2Crop Science Department, North Carolina State University, Raleigh, NC 27695-7631, USA

Received 30 September 2002; revised 17 December 2002; accepted 9 January 2003. ÃFor correspondence (fax 1 919 856 4598; e-mail [email protected]).

Summary

Cytosolic pyruvate kinase (ATP:pyruvate 2-O-phosphotransferase, EC 2.7.1.40) is an important glycolytic enzyme, but the post-translational regulation of this enzyme is poorly understood.Sequence analysis of the soybean seed enzyme suggested the potential for two phosphorylation sites: site-1 (FVRKGS220DLVN) and site-2 (VLTRGGS407TAKL).Sequence- and phosphorylation state-speci®c antipeptide antibodies established

that cytosolic pyruvate kinase (PyrKinc) is phosphorylated at both sites in vivo.However, by SDS±PAGE, the phosphorylated polypeptides were found to be smaller (20±51 kDa) than the full length (55 kDa).Biochem- ical separations of seed proteins by size exclusion chromatography and sucrose-density gradient centrifugation revealed that the phosphorylated polypeptides were associated with 26S proteasomes.The 26S proteasome particle in developing seeds was determined to be of approximately 1900 kDa. In vitro, the 26S

proteasome degraded associated PyrKinc polypeptides, and this was blocked by proteasome-speci®c inhibitors such as MG132 and NLVS.By immunoprecipitation, we found that some part of the phosphorylated

PyrKinc was conjugated to ubiquitin and shifted to high molecular mass forms in vivo.Moreover, recombi-

nant wild-type PyrKinc was ubiquitinated in vitro to a much greater extent than the S220A and S407A mutant proteins, suggesting a link between phosphorylation and ubiquitination. In addition, during seed development, a progressive accumulation of a C-terminally truncated polypeptide of approximately 51 kDa was observed that was in parallel with a loss of the full-length 55 kDa polypeptide.Interestingly, the C- terminal 51 kDa truncation showed not only pyruvate kinase activity but also activation by aspartate.

Collectively, the results suggest that there are two pathways for PyrKinc modi®cation at the post-translational level.One involves partial C-terminal truncation to generate a 51 kDa pyruvate kinase subunit which might have altered regulatory properties and the other involves phosphorylation and ubiquitin conjugation that targets the protein to the 26S proteasome for complete degradation.

Keywords: pyruvate kinase, phosphorylation site, C-terminal truncation, degradation, proteasome, soybean (Glycine max (L.) cv.Ransom).

Introduction Pyruvate kinase (ATP:pyruvate 2-O-phosphotransferase, EC but the effectors of pyruvate kinase in plants appear to be 2.7.1.40) is an enzyme of glycolysis that catalyzes the irre- different from those in animals (Plaxton, 1996). In mamma- versible conversion of phospho(enol)pyruvate (PEP) and lian liver, PyrKinc is activated by fructose-1,6-P2, the product ADP to pyruvate and ATP. Because pyruvate is a key meta- of ATP-linked phosphofructokinase (PFK), whereas plant bolic intermediate of many pathways, such as energy pro- PyrKinc is not. Rather, PEP ± the substrate of PyrKinc ±is duction and the biosynthesis of amino acids, organic acids, an inhibitor of plant PFK (Plaxton, 1996). In addition, amino and fatty acids, there is much interest in elucidating the acids are important allosteric effectors of plant PyrKinc.A regulatory and kinetic properties of pyruvate kinase. Allos- recent study involving rapeseed identi®ed aspartate as an teric regulation of eukaryotic pyruvate kinase is common, activator that could reverse the inhibition by glutamate

ß 2003 Blackwell Publishing Ltd 77 78 Guo-Qing Tang et al.

(Smith et al., 2000) which provided a link between C and N that there are two mechanisms for the post-translational metabolism. modi®cation of PyrKinc. The ®rst is a C-terminal truncation

The phosphorylation of PyrKinc on serine residues by of the 55 kDa full-length PyrKinc to form an approxi- serine/threonine protein kinases in mammals (Ljungstrom mately 51 kDa polypeptide that increased during seed et al., 1976) and Mck1 in yeast (Brazill et al., 1997) have development and might alter regulatory properties. The been established. Phosphorylation of liver (L-type) PyrKinc second mechanism is phosphorylation that could target at the LRRAS site in the N-terminal domain is well docu- the enzyme to the ubiquitin/26S proteasome pathway for mented (Engstrom, 1980; Lone et al., 1986). In plants, this degradation. serine residue is not conserved (Blakeley et al., 1990) and phosphorylation of plant PyrKinc has not been reported to date. Results

We were interested in the role of PyrKinc in developing soybean seeds. Soybean is one of the most important Pyruvate kinase consists of 51 and 55 kDa subunits in economic crop plants in the world, and the regulation of soybean seed carbon partitioning between storage protein and oil production is a central issue for crop improvement. It is cur- An immunological approach was taken to monitor PyrKinc rently thought that sucrose, taken by developing seed, is during seed development. An N-terminal antibody (Ab-NT) converted to PEP via glycolysis in the cytoplasm. The PEP was produced to detect the PyrKinc polypeptide, and can either be transported into plastids to support fatty acid sequence- and phosphorylation-speci®c antibodies were biosynthesis (Ruuska et al., 2002) or converted to pyruvate produced that speci®cally recognized the phosphoserine 220 407 in the cytoplasm by PyrKinc for energy production and À220 (Ab-pSer ) or phosphoserine À407 (Ab-pSer ) organic acids required for amino acid interconversion for sites (Figure 1a). Initially, we probed PyrKinc in protein storage protein biosynthesis. Therefore, PyrKinc is a extracts from seeds (Glycine max (L.) cv. Ransom) at dif- branch-point enzyme and its regulation may be one of ferent stages of development by immunoblot analysis with the factors controlling storage product accumulation. Ear- Ab-NT. To minimize the possible proteolysis and de-phos- lier studies suggested that PyrKinc activity is correlated phorylation during protein extraction, the frozen powdered positively with both storage protein and oil biosynthesis seeds were homogenized directly in SDS sample buffer in developing soybean seeds (Platt and Bassham, 1978; supplemented with a cocktail of protease and phosphatase Smith et al., 1989). While these experimental ®ndings are inhibitors. We found two major polypeptides detected by not consistent with the proposed role of PyrKinc as a Ab-NT: one was a 55 kDa full-length PyrKinc and the other branch-point enzyme, they do support a role for this was approximately of 51 kDa . During seed development, enzyme in storage product biosynthesis and thus it is the 55 kDa full-length PyrKinc decreased, while the 51 kDa important to identify the biological mechanisms controlling polypeptide progressively increased (as shown in activity. While allosteric control of PyrKinc is known (Smith Figure 1b). Similar developmental pro®les were observed et al., 2000), it remains unclear whether control by rever- in three other cultivars (data not shown). sible phosphorylation occurs in vivo as it occurs in mam- Previously, only one cDNA encoding PyrKinc from soy- mals. bean somatic embryos was available in the public data- In the present study, we investigated the phosphorylation base. The full-length deduced sequence (GB# Q42806) of PyrKinc and its regulation in developing soybean seeds. encoded a protein of 511 amino acid. However, no com-

By sequence analysis, we found that the enzyme has two plete open reading frame cDNA sequences for PyrKinc were potential phosphorylation sites: FVRKGS220DLVN and reported from developing soybean seeds. By RT-PCR, we 407 VLTRGGS TAKL. Both sites match the minimal consen- cloned two cDNAs for PyrKinc from developing seeds of sus motif (f-X-Basic-X-X-S/T, where f is a hydrophobic Glycine max (L.) cv. Ransom, herein named PyrKinc1 and residue) that is recognized by calcium-dependent protein PyrKinc2. PyrKinc1 (GB# AY128951) was 511 amino acid in kinases (CDPKs) and SnRK1 (SNF1 related) kinases in plants length and PyrKinc2 (GB# AY128952) was 510 amino acid in (Halford and Hardie, 1998; Huang and Huber, 2001). Se- length. The identity of each to the original sequence for quence- and phosphorylation-speci®c antibodies were pro- soybean PyrKinc (GB# Q42806) is 99.4 and 98%, respec- duced for each potential site in PyrKinc in order to monitor tively, and both share very high identity with cDNAs for phopshorylation in vitro and in vivo. Similar af®nity- PyrKinc in other plant species (Figure 1d). The theoretical puri®ed antibodies produced against synthetic phospho- molecular weights of the two PyrKinc isoforms are approxi- peptides have been used to monitor the phosphorylation of mately 55 kDa. The transcriptional pro®les of the two speci®c sites in other plant enzymes, including PEP carboxy- PyrKinc genes were examined by RT-PCR using isoform- lase (Ueno et al., 2000) and sucrose synthase (Hardin et al., speci®c primers under high-stringency PCR conditions. A

2002). The results obtained in the present study suggest dominant expression pattern of PyrKinc1 was observed

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 79

Figure 1. Characterization of PyrKinc from developing soybean (Glycine max (L.) cv. Ransom) seeds. (a) Schematic diagram showing the sequences used to generate antipeptide (Ab-NT) and antiphosphopeptide (Ab-pSer220 and Ab-pSer407; pSer is phosphoserine) antibodies.

(b) Changes in PyrKinc polypeptides during development of soybean seeds (Glycine max (L.) cv. Ransom) as detected by immunoblot analysis with Ab-NT. (c) Nested RNA gel blot analysis of gene expression for PyrKinc in developing soybean seeds. Three types of cDNAs were chosen as probes. The probes 1 and 2 0 were 378 and 719 bp in length, respectively, which covered different 5 regions, and probe 3 is the full length of PyrKinc1. (d) PHYLOGENY PHYLIP program analysis of cloned soybean PyrKinc1 (GB# AY128951) and PyrKinc2 (GB# AY128952). Clustal alignments were performed using the online software of SEQUENCE UTILITIES at BCM Search Launcher at http://searchlauncher.bcm.tmc.edu. The length of the horizontal lines indicates the degree of relatedness and the identity of each sequence to the soybean PyrKinc (Q42806). Accession numbers to the sequences are soybean (Glycine max, Q42806); soybean (G. max (L.) cv. Ransom, AY128951); soybean (G. max (L.) cv. Ransom, AY 128952); potato (Solanum tuberosum, JC1481); Arabidopsis (Arabidopsis thaliana, BAB10461); tobacco (Nicotiana tabacum, CAA82628); Escherichia coli, AAB47952; rat (Rattus norvegicus, P11981); rat (Rattus norvegicus, P11980); rat (Rattus norvegicus, A23612); rat (Rattus norvegicus, P12928); Drosophila melanogaster, NP_524448); yeast (Schizosaccharomyces pombe, T38890); and tobacco (N. tabacum) plastid (S44287).

(e) Differential RT-PCR analysis of gene expression for the two PyrKinc isoforms in developing soybean seeds by using isoform-speci®c primers. b-tubulin was used as an internal control of the RT-PCR analysis.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 80 Guo-Qing Tang et al.

(Figure 1e), and both declined in abundance as seed devel- boxes in Figure 2b) was excised and analyzed by SDS± opment proceeded. PAGE and immunobloting. As shown in Figure 2(c), active In order to determine if the 51 kDa polypeptide resulted bands eluted from the ®rst activity peak consisted mainly of from an independent transcription event or alternative the 51 kDa subunit. Upon SDS±PAGE, the most active splicing, we performed a nested RNA gel blot analysis with bands resolved by native PAGE from RQ peak I (bands 3 two different 50-region cDNAs (378 and 719 bp) and the full- and 4) contained only small amounts of 55 kDa polypep- length sequence as templates for probe labeling. The sub- tide. In contrast, all active bands from RQ peak II had the sequent hybridization was carried out under low-stringency 55 kDa polypeptide as a major component and a lesser conditions. Because the 51 kDa polypeptide was immuno- amount of the 51 kDa polypeptide. All the active bands logically related to the full-length PyrKinc (Figures 1b and resolved by native PAGE migrated slower than the

3e), it was reasonable to anticipate that RNAs for the 55 kDa 200 kDa standard, suggesting that the soybean PyrKinc can full-length and the 51 kDa truncated polypeptide might be exist as a tetramer or higher order oligomer. Indeed, hex- detected in the RNA gel blots, especially in RNA samples americ (Hu and Plaxton, 1996) and decameric (Wu and from the 9- and 10-week-old seeds that had the highest Turpin, 1992) forms of PyrKinc have been observed. Even level of the 51 kDa polypeptide (Figure 1b). However, we though not all forms of PyrKinc might have retained activity could only detect a signal band corresponding to the pre- duringnativePAGE,theresultssuggestthatthe51and55 kDa dicted 55 kDa PyrKinc subunit (Figure 1c). No signal was polypeptides were active subunits of PyrKinc oligomers. observed in RNA gel blots of samples from the 9- and 10- In order to investigate metabolite regulation of soybean week-old seeds, consistent with the results obtained by seed PyrKinc, active forms resolved by native PAGE were RT-PCR (Figure 1e). In addition, because the 51 kDa poly- assayed in the presence and absence of amino acids. For peptide continued to increase from 8 to 10 weeks while the example, fraction no. 18 from RQ peak I contained a major 55 kDa subunit decreased and transcripts were essentially and minor band of activity after native PAGE (band 3 and 2, undetectable, these experiments excluded the possibility respectively). Both bands were unaffected by Gln, Asn, and that the 51 kDa polypeptide was produced from either an Glu, but were activated by Asp (Figure 2d). In contrast, of independent transcriptional event or an alternative spli- the three activity bands resolved by native PAGE from RQ cing. Rather, it seemed that the 51 kDa polypeptide might peak II, the most pronounced effect was inhibition by Glu. be derived from the 55 kDa form at the post-translational Thus, the observed regulatory properties appeared to vary level by C-terminal cleavage. depending on the subunit composition and/or oligomeric structure of the enzyme. Of particular signi®cance is the association of Asp activation with forms containing the The 51 kDa polypeptide might act as a functional subunit 51 kDa polypeptide. of PyrKin in the soybean seed c To further test if the 51 kDa polypeptide had altered For investigation of potential functions of the novel 51 kDa kinetic activity and/or regulatory properties, we compared polypeptide, we partially puri®ed PyrKinc by Resource Q the total PyrKinc activity in preparations from 2-week-old (RQ) anion-exchange Fast Protein Liquid Chromatography versus 7-week-old seeds. Immunoblot analysis of aqueous (FPLC) from extracts of 5-week-old seeds. Enzymatic assays extracts of the seeds indicated that the 55 kDa polypeptide were conducted at pH 6.9 at which the cytoplasmic form is predominated in 2-week-old seeds, whereas the 51 kDa active while the plastid form is inactive (Smith et al., 2000). polypeptide predominated in 7-week-old seeds (Figure 2e). Under these conditions, there were two major peaks of Quanti®cation of immunoreactive bands by densitometry

PyrKinc activity (fraction nos. 18±26 and 27±35, designated indicated that the PyrKinc protein (sum of the 51 and 55 kDa as peak I and II, respectively; Figure 2a). To further separate polypeptides) in 7-week-old seeds was reduced to 70 Æ 2% possible PyrKinc oligomeric forms, we separated proteins of the amount present in 2-week-old seeds (Figure 2f). by non-denaturing PAGE followed by in-gel activity stain- Maximum PyrKinc activity in 7-week-old seeds was reduced ing as described by Rivoal et al. (2002) (Figure 2b). After to roughly the same extent (62% compared to 2-week-old native electrophoresis, each active PyrKinc band (white seeds; Figure 2g). Therefore, these results suggest that the

Figure 2. The 51 kDa PyrKinc polypeptide is enzymatically active. (a) Resolution of two peaks of PyrKinc activity from extracts of 5-week-old seeds by Resource Q (RQ) anion-exchange Fast Protein Liquid Chromatography (FPLC). (b) Non-denaturing PAGE coupled with in-gel activity staining to visualize active PyrKinc forms in the RQ-fractions. The white boxes show the active bands that were excised and subjected to 12.5% SDS±PAGE.

(c) Immunoblot analysis using Ab-NT of PyrKinc activity bands obtained by non-denaturing PAGE in (b). (d) Effect of amino acids (10 mM) on PyrKinc activity in RQ fraction 18 (top) or 30 (bottom) using the in-gel activity staining method. (e) Immunoblot analysis of PyrKinc polypeptides in three individual replicate samples of 2- and 7-week-old seeds probed by Ab-NT. (f) Comparison of the abundance of total PyrKinc (55 plus 51 kDa polypeptides) in 2- and 7-week-old seeds by densitometry of the immunoblots in (e). (g) Comparison of PyrKinc activity in extracts prepared from 2- and 7-week-old seeds. Assays were conducted in the presence or absence of 20 mM Asp.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 81

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 82 Guo-Qing Tang et al.

51 kDa polypeptide must be an active subunit with equiva- by protein kinase and no phosphorylation of either site was lent catalytic activity compared with the full-length PyrKinc catalyzed in vitro by RQ-SnRK1 (data not shown). subunit. Glutamate has been reported to inhibit PyrKinc We tried to detect the in vivo phosphorylation of PyrKinc (Smith et al., 2000). Consistent with this, 30 mM Glu inhib- using the phospho-speci®c antibodies. There was no sig- ited PyrKinc activity by about 30% in extracts of 2- and ni®cant signal band corresponding to the phosphorylated

7-week-old seeds (data not shown). However, aspartate full-length PyrKinc, but rather abundant phosphorylated at 10 mM activated total PyrKinc activity in 7-week-old and truncated polypeptides in the range between approxi- seeds (16.6 Æ 2% stimulation) but slightly inhibited activity mately 20 and 51 kDa were observed (Figure 3e). This result in extracts from 2-week-old seeds (10 Æ 1% inhibition; implied that phosphorylation and proteolytic cleavage Figure 2g). These results suggest that the 51 kDa polypep- could be closely coupled. tide might have somewhat altered the regulatory properties compared to the 55 kDa polypeptide. Association of phosphorylated and truncated pyruvate kinase polypeptides with the 26S proteasome in soybean The investigation of phosphorylation of pyruvate kinase seeds in soybean seeds Because phosphorylated PyrKinc polypeptides only

The phosphorylation of PyrKinc in plants has not been appeared in truncated polypeptides in vivo, the proteolytic reported so far (Podesta and Plaxton, 1991). However, machinery might be involved. There are two general path- in silico analysis suggested that plant PyrKinc contains ways for protein turnover in eukaryotic cells: one is the two sequences that match the phosphorylation motif tar- ubiquitin/26S proteasome pathway and the other involves geted by plant CDPKs and SnRK1s (Huang and Huber, 2001; soluble proteases present in the vacuole or lysosome. see Figure 3a). Interestingly, the mammalian L- and R-type Selective protein degradation usually involves the 26S

PyrKinc also contain these two potential phosphorylation proteasome, which is a huge complexwith a molecular sites, but phosphorylation at these two sites has not been weight of approximately 2000 kDa in Arabidopsis (Peng reported (Lone et al., 1986). Rather, phosphorylation of the et al., 2001). Therefore, we reasoned that size exclusion mammalian L- and R-type pyruvate kinase occurs at the chromatography of seed extracts could distinguish which serine residue in the sequence LRRAS, located close to pathway was active in degradation of PyrKinc. Cytoplasmic the N-terminus of the protein. pyruvate kinase activity eluted from the size-exclusion We ®rst determined if synthetic peptides based on the chromatography column in a broad peak consistent with two predicted sites could be phosphorylated by endogen- active tetrameric and dimeric forms (Figure 4a). Enzymatic ous seed protein kinases. Initially, we fractionated seed activity was coincident with the 55 kDa polypeptide, detec- proteins by RQ-FPLC. Two major peptide kinase activity ted with Ab-NT (Figure 4b). As shown in the immunoblots peaks were resolved (Figure 3b). The ®rst peak had a two- probed with Ab-pSer220 or Ab-pSer407 (Figure 4b), the heav- 220 fold higher activity with the Ser peptide compared to the ily phosphorylated and truncated PyrKinc polypeptides 407 2 Ser peptide, was Ca -dependent, and had a native Mr of partitioned in the ¯ow-through fractions which were devoid approximately 60 kDa by size exclusion chromatography of PyrKinc activity.

(Figure 3c), and therefore appeared to be a member of Because the phosphorylated and truncated PyrKinc poly- the CDPK family of protein kinases (Harmon et al., 2000). peptides eluted near the column void volume, we specu- The second peak from RQ-FPLC showed a preference for lated that they could be associated with proteasomes. In a the Ser407 peptide, compared to the Ser220 peptide and was subsequent size exclusion chromatography experiment, 2 not Ca -dependent. The native Mr was approximately we detected a chymotrypsin-like activity of a huge proteo- 150 kDa, and immunoblot analysis using antimaize SNF1 lytic complexusing Suc-LLVY-AMC as the substrate catalytic subunit antibodies con®rmed the existence of (Figure 5a). The maximum catalytic activity of the proteo- SnRK1 in those fractions (data not shown). lytic complexwas observed in the presence of MgATP. No

In order to determine if the full-length PyrKinc could be signi®cant 20S-like activity was observed. Moreover, a 26S phosphorylated by CDPK or SnRK1, we performed in vitro S60 ATPase-like component was detected in this proteolytic phosphorylation using recombinant PyrKinc as substrate complexby immunoblot analysis using anti-Hela cell 26S and RQ-CDPK or RQ-SnRK1 as the protein kinases. Phos- S60 ATPase (Figure 5b). Therefore, we believe that the high phorylation at the Ser220 site by RQ-CDPK, as monitored molecular complexis the 26S proteasome. The size of the with sequence- and phosphorylation-speci®c antibodies soybean 26S proteasome particle, determined by size (Ab-pSer220), proceeded in a time-dependent manner, exclusion chromatography, was approximately 1900 kDa, and was completely blocked by the addition of the calcium which is similar to that reported in Arabidopsis (Peng et al., chelator EGTA (Figure 3d). We could not demonstrate the 2001). Indeed, supplementing the extraction buffer with 407 phosphorylation of recombinant PyrKinc at the Ser site the proteasome-speci®c inhibitors, MG132 and NLVS,

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 83

Figure 3. Characterization of protein kinases and phosphorylation of PyrKinc in developing soybean (Glycine max (L.) cv. Ransom) seeds. (a) Alignment showing potential phosphorylation sites in the sequences of cytosolic pyruvate kinase in selected plants. (b) Resolution of peptide kinase activities by Resource Q (RQ) anion-exchange Fast Protein Liquid Chromatography (FPLC) using synthetic peptides based on the Ser220 or Ser407 sites as substrates. (c) Identi®cation of native molecular sizes of protein kinases by size exclusion chromatography and veri®cation of the nature of protein kinases. 220 (d) Immunoblot analysis of in vitro phosphorylation of recombinant PyrKinc by partially puri®ed RQ-CDPK using Ab-pSer . The immunoblot probed with Ab- NT was the control to show the equal addition of PyrKinc in the phosphorylation reactions in vitro. 220 407 (e) Immunoblot detection of in vivo phosphorylated PyrKinc in the developing soybean seeds by Ab-pSer and Ab-pSer .

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 84 Guo-Qing Tang et al.

Figure 4. Size exclusion chromatography separation of PyrKinc and its phosphorylated and truncated polypeptides. (a) Elution of PyrKinc activity during size exclusion chromatography. 220 407 (b) Immunoblot detections of PyrKinc and its polypeptides by Ab-NT, Ab-pSer , and Ab-pSer in the fractions from 4-week-old seeds. increased the recovery of full-length (55 kDa) and 51 kDa et al., 1994; Hu et al., 1999). By this approach, we examined polypeptides (cf. Figure 5c,d) and phosphorylated polypep- 26S proteasomes in developing soybean seeds. After 19 h tides (cf. Figure 5e±h) co-eluting with the 26S proteasomes. of centrifugation at 100 000 g at 48C in a 10±40% sucrose In parallel, we also measured trypsin activity as an inter- gradient, all of the 26S proteasome activity was at the nal control. The elution volume for detectible typsin-like bottom of the gradient when MgATP was present through- activity using the substrate Z-ARR-AMC was located at out (Figure 6a) and there was no detectable 20S-like pro- approximately 24 kDa, similar to the Mr of the trypsin-like teasome activity (data not shown). When MgATP was protease (Shrestha et al., 2002). No trypsin activity was ob- omitted, 26S-like proteasome activity was undetectable served in the ¯ow-through fractions (Figure 5a). Although (data not shown) and only a low level of 20S-like protea- there was a clear 36 kDa polypeptide recognized by Ab-NT some activity was observed (Figure 6b). The results sug- that co-eluted with trypsin-like protease, there were no gest that the maintenance of the 26S proteasome particle is obvious phosphorylated and truncated PyrKinc poly- ATP-dependent. Consistently, as an internal control as in peptides that co-eluted with the trypsin-like activity size exclusion chromatography experiments, we found (Figure 5e±h). These results suggest that the phosphory- typsin-like activity exclusively in the low-molecular weight lated and truncated PyrKinc polypeptides were speci®cally fraction that remained at the top of the gradient (Figure 6c). associated with the 26S proteasome. In the sucrose-gradient centrifugation experiments, inci-

Because of their large size, sucrose-gradient centrifuga- dences of proteasomes and phosphorylated PyrKinc poly- tion is a routine technique to isolate proteasomes (Fujinami peptides were examined by immunoblot analysis.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 85

Figure 5. Proteolytic activity assay after size exclusion chromatography and immunoblot analysis of the association between PyrKinc and 26S proteasome particles using 5-week-old seeds. (a) Proteolytic activity pro®le in the collected fractions of size exclusion chromatography. (b) Immunoblot detection of Hela 26S S06 ATPase homolog in the collected fractions. 220 440 (c±h) Immunoblot detections of PyrKinc and its polypeptides by Ab-NT, Ab-pSer , and Ab-pSer in the collected fractions of size exclusion chromatography experiments, in which the extraction buffer was with or without proteasome-speci®c inhibitor MG132 and NLVS as indicated. ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 86 Guo-Qing Tang et al.

Figure 6. Sucrose density gradient centrifugation separation of 26S proteasome particles and immunoblot detections of PyrKinc. (a) 26S proteasome assay. The extraction buffer and sucrose gradient had MgATP. (b) 20S proteasome assay. The extraction buffer and sucrose gradient did not have MgATP but contained SDS. (c) Trypsin-like activity assay. 220 (d±g) Immunoblot detections of PyrKinc and its polypeptides by Ab-NT and Ab-pSer in separate experiments with/without proteasome-speci®c inhibitors MG132 and NLVS.

Including proteasome inhibitors increased recovery of Pyr- Ubiquitination of pyruvate kinase in vivo and in vitro Kinc polypeptides reactive with Ab-NT (cf. Figure 6d,e), and Ab-pSer407 (cf. Figure 6f,g) in the proteasome-containing Ubiquitination is an essential step for the targeting of most fractions. These results are consistent with those obtained proteins for degradation through the 26S proteasome path- from size-exclusion chromatography experiments, and way. In plants, ubiquitin-conjugation and proteasome- suggest that the phosphorylated and truncated PyrKinc mediated degradations have been con®rmed for a number polypeptides are degradative intermediates associated of important regulatory proteins such as PhyA and AUX/ with 26S proteasomes. Notably, the 55 kDa full-length Pyr- IAA proteins (Cough and Vierstra, 1997; Gray et al., 2001).

Kinc was found associated with the 26S proteasome when Accordingly, we tested if PyrKinc can be ubiquitinated MG132 and NLVS were included in the extraction buffer, either in vivo or in vitro. We carried out in vitro ubiquitina- suggesting that not all subunits of the PyrKinc oligomer are tion using two different approaches. First, we demon- post-translationally modi®ed prior to degradation by the strated in vitro ubiquitination using recombinant PyrKinc proteasome. and a soluble protein fraction prepared from 2-week-old

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 87 seeds that was depleted of proteasomes by ultracentrifugation and supplemented with exogenous ubiquitin. A similar approach was used by O'Hagan et al. (2000) with S100 cell extracts. The ubiquitination of recombinant His6-PyrKinc was detected by the appearance of high-molecular weight

PyrKinc conjugates, immunoprecipitated with anti-His6 antibodies. The ubiquitin conjugation state was veri®ed by immunoblot analysis using antiubiquitin antibodies (data not shown). The second approach was attempted, because the components of the ubiquitin±proteasome pathway have been shown to be pre-assembled at the cytoplasmic surface of the ER in mammals and yeast (Biederer et al., 1997; de Virgilio et al., 1998). We fractionated the membrane preparation from 5-week-old seeds by ultracentrifugation at 150 000 g for 6.5 h, and after re-suspension tested in vitro ubiquitination of recombinant Pyr-

Kinc. Time-dependent ubiquitination of recombinant His6-

PyrKinc, immunoprecipitated with anti-His6 antibodies, was observed. Only the wild-type (WT) recombinant PyrKinc was ubiquitinated in a time-dependent manner (Figure 7a). Among the phosphorylation site mutants, S220A showed less ubiquitination signal and S407A did not show any ubiquitination (Figure 7a). These results suggest that the phosphorylation of serine residues in PyrKinc is essential for ubiquitination. To examine the existence of ubiquitinated conjugates of

PyrKinc in vivo, we immunoprecipitated PyrKinc from seeds at different stages of development using Ab-NT, followed by immunoblot analysis with anti-Ubi (Figure 7b-1), Ab-NT (Figure 7b-2), Ab-pSer407 (Figure 7b-3), or Ab-pSer220 (Fig- ure 7b-4). High Mr ubiquitin conjugates were observed after probing with anti-Ubi antibody (Figure 7b-1) or with Ab-NT (Figure 7b-2). These results con®rm the existence of endo- genous ubiquitin conjugates of PyrKinc in young developing seeds. The abundance of those conjugates was higher in the ®rst 5 weeks of seed development, then decreased gradually until at maturity there were no detectable ubiquitin conjugates. This pattern is in agreement with the ®nding of total protein-conjugates in Lupin (Lupinus albus L.) seeds during formation (Ferreira et al., 1995). In mammals, phosphorylation and ubiquitination are closely coupled processes that target many proteins for Figure 7. In vitro ubiquitination assays. complete degradation (Attaix et al., 2001), in particular, (a) Immunoprecipitation of ubiquitin conjugates by using anti-His6 antibodies and immunoblot detected with Ab-Ubi. The substrate was wild-type those recognized by the SCF E3 ligase (Hershko and Cie- (WT) recombinant PyrKinc, and the ubiquitination reaction was supported chanover, 1998). To determine if the ubiquitin conjugates of by a cell extract from 2-week-old seeds. (b) Immunoprecipitation and immunoblot analysis of in vivo ubiquitination PyrKinc are also phosphorylated, we probed the blots of Ab- 220 407 as above. The substrates were WT recombinant PyrKinc and its mutants NT-immunoprecipitates with Ab-pSer or Ab-pSer .As (S220A or S407A) and ubiquitination reactions were supported by a mem- shown in Figure 7(b-3, b-4), both serine residues were brane preparation from 5-week-old seeds. (b-1) Ab-NT immunoprecipitation phosphorylated in the ubiquitin-conjugates in a develop- and immunoblot detection with Ab-Ubi. (b-2) Ab-NT immunoprecipitation and immunoblot detection with Ab-NT. (b-3) Ab-NT immunoprecipita- mental pro®le. Thus, in agreement with our observation of tion and immunoblot detection with Ab-pSer407. (b-4) Ab-NT immunopre- in vitro ubiquitination (Figure 7b), the ubiquitin conjugates cipitation and immunoblot detection with Ab-pSer220. of phosphorylated PyrKin polypeptides have been con- Abbreviations: PyrKinc-Ubin, ubiquitinated pyruvate kinase; PyrKinc- c 220 220 pSer -Ubin, Ser -phosphorylated and ubiquitinated pyruvate kinase; ®rmed and the results suggest that phosphorylation of 407 407 PyrKinc-pSer -Ubin, Ser -phosphorylated and ubiquitinated pyruvate PyrKinc occurs in concert with ubiquitination. kinase.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 88 Guo-Qing Tang et al.

Figure 8. In vitro degradation of proteasome-

associated PyrKinc and its polypeptides detected by immunoblot analysis. (a) Immunoblot detection by Ab-NT of the degradation of the 26S proteasome-associated

PyrKinc and its polypeptides in the ¯ow-through fraction obtained from size exclusion chromatography. (b) Immunoblot detection by Ab-pSer407 of 26S proteasome degradation of its associated Pyr-

Kinc polypeptides in the pellet fraction obtained from sucrose-gradient centrifugation.

the other appears to be a SnRK1. Furthermore, developing The degradation of pyruvate kinase polypeptides soybean seeds are shown to contain typical 26S protea- associated with 26S proteasomes in vitro some particles with a molecular weight of approximately

We next examined the degradation of PyrKinc polypeptides 1900 kDa determined by size exclusion chromatography. associated with proteasomes contained in the ¯ow-through Collectively, we propose that there are two pathways for fraction from size exclusion chromatography (as in the PyrKinc modi®cation at the post-translational level. One experiment shown in Figure 5a). As shown in Figure 8(a), involves conversion of the 55 kDa to the 51 kDa polypep- immunoblot analysis using Ab-NT indicated that the degra- tide, which may produce a novel pyruvate kinase activity. dation was time-dependent and was prevented by MG132. The second involves phosphorylation (at the Ser220 and 407 The full-length PyrKinc and its truncations were only Ser sites) and ubiquitination and might be involved in the degraded during incubation at 308C in the presence of proteasome-mediated degradation of the enzyme. ATP but in the absence of MG132. Similarly, we tested the proteasome-containing fraction The 51 kDa subunit is produced from the 55 kDa pyruvate from sucrose-gradient centrifugation following the removal kinase subunit by a C-terminal truncation of MG132 by dialysis. The time-dependent degradation of proteasome-associated phosphorylated polypeptides at Immunoblot analysis using Ab-NT revealed that develop- 308C was observed on immunoblots probed with Ab- ing seeds accumulate a C-terminally truncated 51 kDa Pyr- 407 220 pSer (Figure 8b) or Ab-pSer (data not shown). These Kinc polypeptide that appears to retain PyrKinc activity. As a results provide additional evidence for PyrKinc degradation part of this study, we attempted to elucidate the basis for through the 26S proteasome pathway. production of the 51 kDa polypeptide. The 51 kDa polypeptide could be produced: (i) from an independent transcription event, (ii) from an alternative splicing event of the full Discussion length transcript, or (iii) from proteolytic processing of the In this study, we have cloned two nearly identical cDNAs 55 kDa polypeptide. We were unable to detect a transcript encoding paralogs of PyrKinc; both encode polypeptides of encoding a 51 kDa polypeptide, which rules the ®rst two approximately 55 kDa and are expressed in a similar devel- possibilities as unlikely. Furthermore, while alternative spli- opmental manner in developing soybean seeds. However, cing to generate a shortened mRNA is common in mam- during development, a 51 kDa polypeptide accumulates, mals (35% of genes), it occurs less often in plants (Miller apparently produced by C-terminal cleavage of the 55 kDa et al., 2001). In rat muscle, alternative splicing produces the polypeptide. Active PyrKinc appears to be a heterooligo- M1- and M2-type PyrKinc and the alternative splicing occurs meric protein composed of 55 and 51 kDa subunits depend- at the 9th and 10th exons (Noguchi et al., 1986). The two M- ing on the developmental stage. We also found that types of PyrKinc have different enzymatic properties and developing seeds contain two kinds of protein kinases that tissue speci®city (Imamura and Tanaka, 1982). To under- 220 can potentially phosphorylate PyrKinc at two sites (Ser stand the possible gene structure of soybean PyrKinc,we and Ser407). Biochemically, one shows CDPK activity and examined the Arabidopsis homolog (GB# NP_196474, TIGR

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 89

accession no. at5g08570, http://www.tigr.org), which has recovery of phosphorylated PyrKinc polypeptides (data three exons and two introns. Therefore, the gene structure not shown). However, the extent of reversibility was not encoding plant PyrKinc is totally different from that of the investigated and is beyond the scope of the present study. mammalian genes. No alternative splicing for the plant More importantly, the recovery of the phosphorylated poly- gene was suggested by computational analysis (http:// peptides with proteasomes was increased by addition of www.tigr.org; search locus At5g08570). This computational proteasome-speci®c inhibitors, MG132 and NLVS, and prediction is consistent with the nested RNA gel blot when proteasomes were incubated in vitro (in the absence results; no detectible shorter transcripts were observed. of inhibitors), PyrKinc polypeptides were degraded

Therefore, proteolytic processing might be responsible. (Figure 8). High Mr Ubi±PyrKinc conjugates could be iden-

This is supported by the observation that the 51 kDa poly- ti®ed in vivo. Ubiquitination of recombinant WT PyrKinc peptide continued to increase from the 9th to 10th week, was demonstrated in vitro, but ubiquitination of phosphor- even though the RNA gel blots and RT-PCR pro®les showed ylation site mutants (S220A or S407A) was signi®cantly almost no transcripts for PyrKinc at that time. Several reduced (Figure 7a). These results suggest that phosphor- critical questions remain. First, which proteolytic pathway ylation is an essential step prior to ubiquitination. Collec- or enzyme (such as proteasome or an unknown protease) is tively, these results suggest that phosphorylation and engaged in this C-terminal truncation and how is it regu- ubiquitination are essential for PyrKinc degradation via lated? Secondly, what is the physiological signi®cance of proteasome pathway. the 51 kDa polypeptide? It is possible that C-terminal clea- In recent years, the ubiquitin/26S proteasome pathway vage of the 55 kDa polypeptide of PyrKinc may provide a has been shown to play pleiotrophic roles in plant biology novel mechanism to complement transcriptional regula- such as hormone regulation (auxin, cytokinins, and abscisic tion in developing soybean seed. We speculated that the acid signaling), photomorphogenesis, senescence, pro- 51 kDa polypeptide may have increased stability, as it grammed cell death, and plant defense responses showed a steady accumulation during development while (reviewed by Callis and Vierstra, 2000; Estelle, 2001; the full-length PyrKinc decreased. Removal of approxi- Schwechheimer and Deng, 2001). Our results suggest that mately 4 kDa (approximately 35 amino acid residues) from degradation of metabolic enzymes such as PyrKinc and the C-terminus could have a signi®cant impact on both sucrose synthase (Hardin et al., 2002) can be added to stability and enzymatic properties. the list as well. For both enzymes, phosphorylation of minor or `cryptic' sites has been implicated as a possible trigger for proteasome-mediated degradation. If this is correct, Phosphorylation and ubiquitination of pyruvate kinase there may be recognition elements contained within the are linked to its degradation via 26S proteasomes 220 407 sequences surrounding Ser and Ser of PyrKinc and Embryogenesis is an active phase of metabolism and re- Ser170 of sucrose synthase (Hardin et al., 2002) that may be quires highly accurate control. The regulation of metabolic recognized by an E3 ubiquitin ligase such as SCF complex. enzymes by turnover contributes to the control extended by Thus, certain phosphorylation sites in metabolic enzymes transcriptional and translational control. Ferreira et al.(1995) may serve as `phospho-degrons' as described for the phos- reported that there were abundant and dramatic changes of phorylation-driven and ubiquitination-mediated destruc- ubiquitin and ubiquitin±protein conjugates during lupin tion of cyclin-dependent kinases (Harper, 2002). seed development. Thus, ubiquitination and protein turnover is signi®cant in developing seeds. Similarly, Doelling Experimental procedures et al. (2001) demonstrated the importance of the ubiquitin/ 26S proteasome pathway to embryo development in Ara- bidopsis. They showed arrest of embryo development at Materials the globular stage in knockouts of ubiquitin-speci®c pro- All peptide-speci®c and phosphopeptide-speci®c antibodies were tease (UBP14). In agreement with these ®ndings, our data produced and af®nity puri®ed by Bethyl Laboratories, Inc. (Mon- tgomery, TX). The peptide sequences were as follows: MANID suggest that metabolic enzymes such as PyrKinc are IEGILKQQQP for Ab-NT; FVRKGpS220DLVN for Ab-pSer220; degraded via the ubiquitin/26S proteasome pathway. 407 407 220 407 VLTRGGpS TAKL for Ab-pSer . The ratio of speci®city of the Phosphorylation of PyrKinc at the Ser and Ser antiphosphopeptide antibodies for the phosphopeptide antigen occurred in vivo primarily on N-terminally truncated poly- compared to the unphosphorylated sequence was greater than peptides that co-fractionated with 26S proteasomes during 99 : 1 by ELISA (as determined by Bethyl Laboratories). [g-32P] ATP size exclusion chromatography (Figures 4 and 5) or (111 TBq mmolÀ1) and [a-32P] dCTP (111 TBq mmolÀ1) were pur- sucrose-gradient centrifugation (Figure 6). We presume chased from Perkin Elmer (Boston, MA); microcystin-LR, proteasome inhibitors (MG132, NLVS) and the ¯uorogenic substrate that the phosphorylation of both sites was reversible by (Suc-LLVY-AMC) for proteasome assay were obtained from Cal- the action of protein phosphatases. This was suggested by Biochem (La Jolla, CA). Pfu DNA polymerase was from Stratagene the observation that phosphatase inhibitors increased (Carlsbad, CA) and Taq DNA polymerase was from Roche (Switzer-

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 90 Guo-Qing Tang et al.

land). All other chemicals were from Sigma±Aldrich (St. Louis, KOH (pH 6.9), 25 mM KCl, 10 mM MgCl2,2mM PEP, 1 mM ADP, MO), unless stated otherwise. 1mM DTT, 5% (w/v) PEG 8000, 0.15 mM NADH, and 4 units mlÀ1 Soybean [Glycine max (L.) cv. Ransom] plants were grown in lactate dehydrogenase in a total volume of 1 ml. Assays were controlled growth rooms with 9 h photoperiod and 268C day/228C initiated by the addition of ADP and PEP and incubated for 5 min at night temperature regime in the South-eastern Plant Environment 258C. The decrease in absorbance at 340 nm as a result of the Laboratory. A 3 h light interruption was included during the dark oxidation of NADH was followed with time. period for the ®rst 5 weeks. Pods were harvested at the size depending on the designed experiments and seeds were frozen Cloning of cDNAs encoding pyruvate kinase from directly in the liquid N2. developing seeds

Sample preparation for SDS±PAGE and immunoblot Total RNA was isolated from 0.3 g tissue of 4±5-week-old seeds analysis following the routine RNA isolation protocol (Kiefer et al., 2000). Total RNA (1 mg) was used for the reverse transcription reaction

Frozen seeds were powdered in liquid N2 and mixed with four following instruction from the manufacturer (InVitrogen, Carlsbad, volumes of 1.5Â stock of SDS sample buffer (no dye), which was CA) except that incubation was at 458C for 1 h. The forward primer, supplemented with a cocktail of protease inhibitors (1 mM 4-(2- GAGAACATATGGCGAACATAGACATCG, and the reverse primer, aminoethyl)-benzene sulfonyl ¯uoride hydrochloride (AEBSF- GAGCCCGGGTTATTTGACTATGCAGATC, were chosen from the HCl), 1 mM phenylmethylsulfonyl ¯uoride (PMSF), 1 mM benza- cDNA sequence in the NCBI database (GB#: L08632). The PCR was midine HCl, 5 mM trans-epoxysuccinyl-l-leucylamido-(4-guani catalyzed by Pfu DNA polymerase (Stratagene, La Jolla, CA) under dinobutane (E64), 50 mM leupeptin, 20 mgmlÀ1 soybean trypsin the conditions of 948C for 2 min followed by 36 cycles of 948C, inhibitor and 5 mM ethylenediamine tetraacetic acid Na2-salt 50 sec denaturation; 478C, 45 sec annealing; 728C, 2 min exten- (EDTA)), and phosphatase inhibitors (50 mM sodium ¯uoride sion. The PCR product was checked by electrophoresis in 0.8% and 1 mM microcystin-LR). Proteins were denatured by boiling at agarose in 1Â TAE buffer. For TA cloning, a dA tail was added to the 1008C for 4 min, and 20 mg total protein was loaded into each lane PCR product by Taq DNA polymerase, and the desired band was and separated by 12.5% SDS±PAGE, which was followed by a excised and then eluted by the Qiaquick gel extraction Kit (Qiagen, standard immunoblot analysis (ECL; Amersham Biosciences, UK). Germany). The eluted cDNA was ligated into pGEM-T-easy vector (Promega, Madison, WI). Positive clones were identi®ed by Partial purification of protein kinases from soybean seeds sequencing, and sequence analysis was carried out by using the online software of SEQUENCE UTILITIES at BCM Search Launcher Partial puri®cation of protein kinases was performed as detailed by (http://searchlauncher.bcm.tmc.edu) and PHYLOGENY PHYLIP pro- Bachmann et al., (1996). Brie¯y, frozen seeds (6 g) were homo- gram (http://bioweb.pasteur.fr/seganalphylogeny/phylip-uk.html). genized in 50 ml of extraction buffer containing 50 mM MOPS- The two cDNAs cloned for soybean PyrKinc (PyrKinc1 and NaOH, pH 7.5, 10 mM MgCl2,5mM DTT, 1 mM EDTA, 0.5 mM PyrKinc2) were submitted to the NCBI database and their accession phenylmethylsulfonyl ¯uoride, and 0.1% (v/v) Triton X-100 in a numbers are AY128951 and AY128952. cold mortar. The homogenate was ®ltered through two layers of Miracloth (CalBiochem, LaJolla, CA). Soluble proteins that preci- RT-PCR analysis and RNA gel blot analysis pitated from 5 to 15% (w/v) polyethylene glycol 8000 (PEG 8000) were collected by centrifugation at 34 540 g for 20 min at 48C. The The expression of PyrKinc genes was analyzed by RT-PCR. The protein pellet was re-suspended in 10 ml buffer A (50 mM MOPS- isoform-speci®c primers for each PyrKinc were as follows: the NaOH, pH 7.5, 10 mM MgCl2, 2.5 mM DTT). After clari®cation by forward primer for PyrKinc1 was TTGAAGCAGCAGCAGCCTT centrifugation at 34 540 g, 20 min, 48C, the sample was applied to and the reverse primer was TGTTGGGTACACCCCATCC; the a 10 ml Resource Q anion exchange column (Amersham Bios- forward primer for PyrKinc 2 was ATCTTGAAGCAGCAGCTG ciences, UK). After washing with buffer A, the bound protein was and the reverse primer was TGTTGGGTACACCCCATTG. The eluted with an 80 ml linear gradient from 0 to 500 mM NaCl in PCR was driven by Taq DNA polymerase (Roche, Switzerland) buffer A at a ¯ow rate of 1 ml minÀ1. Active fractions were pooled under the conditions of 948C for 2 min followed by 28 cycles of separately and dialyzed against buffer A without NaCl. 948C, 50 sec denaturation; 588C, 45 sec annealing; 728C, 1 min extension. Peptide kinase assay For nested RNA gel blots, the DNA for probe 1 was 378 bp in length covering 50 region, which was released by PCR. The forward The peptide kinase assay was carried out in a 40 ml volume of primer was GAGAACATATGGCGAACATAGACATCG and reverse 32 50 mM MOPS-NaOH, pH 7.5, 10 mM MgCl2, 0.1 mM ATP ([g- P] primer was CTCCTGATTCTCAACC. The DNA for probe 2 was À1 À1 0 ATP 200 CPM pmol ), 0.1 mg ml peptide, 0.1 mM CaCl2 or 719 bp in length covering 5 region and produced by PCR using 1mM EGTA (as speci®ed in the text) and containing 4 mlRQ the forward primer GAGAACATATGGCGAACATAGACATCG and column fraction. Reactions were initiated by addition of ATP the reverse primer GGACAGCAGATGATGC. The probe labelings and incubated for 10 min at room temperature. After incubation, were using the Random Labeling Kit following the instruction from 30 ml of the reaction mixture was spotted onto P81 phosphocellu- the manufacturer (Roche, Switzerland). lose paper squares (4 cm2). After three 10 min washes with 75 mM O-phosphoric acid to remove unincorporated g-32P ATP, the radio- activity on the paper square was detected by liquid scintillation Expression of recombinant pyruvate kinase protein spectrometry. The vector plasmid DNA pET15b (Novagen, Germany) was Pyruvate kinase assay sequentially digested by BamH1, ®lled with Klenow, then digested with NdeI, and fractionated by 0.8% agarose gel electrophoresis.

The PyrKinc assay followed the procedure of Turner and Plaxton The insert DNA for PyrKinc was prepared by sequential double (2000). Brie¯y, the assay buffer system consisted of 50 mM Hepes/ digestion by Smal and NdeI, ligated with the linearized vector DNA,

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 91 and transformed into E. coli DH5a. Positive transformants were Immunoprecipitation and protein immunoblot analysis identi®ed and the plasmid DNA was used to transform E. coli BL21. The 378C overnight culture of transformed BL21 cell was diluted Immunoprecipitation was performed in 100 ml volume consisting into 100 volumes in LB medium. After growing at room tempera- of 50 mM Tris±HCl, pH 7.5, 150 mM NaCl, 0.1% (v/v) Triton X-100, 5mM EDTA, 1 mM PMSF, and 10 mM MG132. The antibody was at ture for 12 h, the induction of recombinant PyrKinc was initiated by À1 addition of isopropyl b-D-thiogalacto-pyranoside (IPTG) to a ®nal a concentration of 10 mgml . After binding for 2 h at 48C with concentration of 0.5 mM and continued for 10 h at room tempera- gentle shaking, 10 ml washed protein-G beads (Sigma±Aldrich, St. Louis, MO) were added and binding was continued for another 1 h ture. Soluble (His)6-PyrKinc and its mutants (S220A and S407A) were af®nity-puri®ed following the manufacturer's instruction at 48C. After extensive washing with the buffer above, the protein (Qiagen, Germany). bound to the beads was eluted by addition of SDS sample buffer and boiled at 1008C for 10 min. After electrophoresis, the protein was transferred to nitrocellulose membrane (Hybond ECL) and In vitro phosphorylation of pyruvate kinase probing followed the protocol detailed by the manufacturer (Amer- The procedure for in vitro phosphorylation was the same as for sham Bioscience, UK). peptide kinase assays, except that only non-radioactive ATP was applied. Speci®c phosphorylation of the Ser220 and Ser407 sites was In vitro ubiquitination determined by immunoblot analysis using sequence- and phosphorylation-speci®c antibodies for both sites. Soybean seed extracts were prepared following the procedure described by O'Hagan et al. (2000) with some modi®cations. Brie¯y, 5 g frozen seeds were homogenized in 10 ml extraction Size exclusion chromatography separation of 26S buffer containing 40 mM Hepes-KOH, pH 7.2, 5 mM MgCl2,8mM proteasome particle KCl, 0.5 mM PMSF, and 10 mgmlÀ1 leupeptin. The homogenate Sixgrams of soybean seeds were homogenized in 50 ml of extrac- was clari®ed by centrifugation at 34 540 g for 10 min at 48C. The resulting supernatant was then centrifuged at 100 000 g for 1 h at tion buffer containing 50 mM MOPS-NaOH, pH 7.5, 100 mM NaCl, 48C in a Beckman L8-80 Ultracentrifuge. For the membrane pre- 10 mM MgCl2, 1.5 mM ATP, 2 mM DTT, 10% glycerol, and 0.5 mM phenylmethylsulfonyl ¯uoride (PMSF) in a cold mortar. The homo- paration, after 100 000 g centrifugation for 1 h at 48C, the pellet genate was ®ltered through two layers of Miracloth (CalBiochem, was discarded and the supernatant was centrifuged again La Jolla, CA) and fractionated with PEG 8000. Soluble proteins that (150 000 g,48C, 6.5 h). The in vitro ubiquitination assay was per- precipitated between 5 and 15% (w/v) PEG 8000 were collected by formed according to the protocol detailed by Podust et al. (2000) centrifugation at 34 540 g for 20 min at 48C. The protein pellet was with some modi®cations. Brie¯y, the 40 ml reaction mixture con- re-suspended in 10 ml of extraction buffer. After an additional tained 10 mM Hepes-KOH, pH 7.2, 6 mM MgCl2,2mM ATP, 1 mM centrifugation at 100 000 g for 1 h at 48C, 6 ml of the supernatant okadaic acid, 30 mM ubiquitin, 10 nM MG132, 3 mg recombinant was applied to a Fractogel TSK HW-55 column (bed volume protein, and 10 ml S100-like supernatant or membrane preparation, 126 ml) (MCB manufacturing chemists. Inc., Darmstadt, Germany) and incubated at 308C for 0±60 min. The conjugates were immu- linked to FPLC (Amersham Biosciences, UK). The equilibration noprecipitated by anti-His6 and probed with antiubiquitin antibodies (Sigma±Aldrich, St. Louis, MO) on immunoblots. buffer and elution buffer were 50 mM MOPS-NaOH, pH 7.5, containing 10% glycerol, 100 mM NaCl, 3 mM MgCl2, 1.6 mM ATP, and 2 mM DTT. The ¯ow speed was at 30 ml hÀ1 and the separa- Pyruvate kinase native gel activity assay tion was performed at 48C. The 8% native gel system was the same as described by Doucet et al. (1990) except that the SDS was omitted. Activity staining of Partial purification of proteasome by sucrose-gradient pyruvate kinase was performed by the protocol detailed by Rivoal centrifugation et al. (2002) with some modi®cations. Brie¯y, the electrophoresis was run for 2 h at 120 V and 48C. After that, the gel was immedi- Powdered soybean seeds (2.5 g) were homogenized in 10 ml of ately equilibrated in 25 ml of 50 mM Hepes-KOH, pH 6.9, 25 mM extraction buffer containing 25 mM Tris±HCl, pH 7.5, 10 mM KCl, 10 mM MgCl2,1 mM DTT, and 5% (w/v) PEG 8000 for 15 min at MgCl2, 4mM ATP, and 1 mM DTT. The homogenate was ®ltered room temperature. The gel was developed in the same buffer through Miracloth (CalBiochem, La Jolla, CA) and centrifuged for containing 2 mM PEP, 1 mM ADP, 0.2 mM NADH, and 4 units mlÀ1 15 min at 34 540 g and 48C. Supernatants were collected and 1 ml lactate dehydrogenase. After incubation for 10 min at room tem- was applied to 4 ml 10±40% sucrose gradients in extraction buffer. perature, the stained gel was photographed on the UV transillu- The gradient was then centrifuged at 100 000 g and 48C for 19 h in minator. The active PyrKinc appeared in dark bands against the a Beckman L8-80 Ultracentrifuge. Fractions (200 ml) were collected ¯uorescent background. manually after ultracentrifugation. Acknowledgements Proteasome activity assays This research represents cooperative investigations of the US Proteasome assays were performed as described by Kisselev Department of Agriculture (USDA), Agricultural Research Service, et al. (1999) with minor modi®cations. For 26S proteasome acti- and the North Carolina Agricultural Research Services. Mention of vity, 10 ml enzyme samples were added to 0.99 ml of 20 mM Suc- a trademark or proprietary product does not constitute a guarantee

LLVY-AMC in 20 mM Tris±HCl, pH 7.5, 1 mM DTT, 5 mM MgCl2, or warranty of the product by the USDA or the North Carolina and 1 mM ATP. After a 30 min incubation at 308C, the released Agricultural Services and does not imply its approval to the exclu- AMC was detected by ¯uorescence with 380 nm excitation sion of other products that might also be suitable. The research 460 nmÀ1 emission. For 20S proteasome activity, 0.02% SDS was supported in part by fund from the United Soybean Board was substituted for 1 mM ATP in the same buffer and incubation (project number 1240) and the US Department of Energy (Grant conditions. DE-AI05±91ER20031 to SCH).

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 92 Guo-Qing Tang et al.

References Hershko, A.and Ciechanover, A. (1998) The ubiquitin system. Annu. Rev. Biochem. 67, 425±479. Attaix, D., Combaret, L., Pouch, M.N. and Taillandier, D. (2001) Hu, Z.and Plaxton, W.C. (1996) Puri®cation and characterization of Regulation of proteolysis. Curr. Opin. Clin. Nutr. Metab. Care, 4, cytosolic pyruvate kinase from leaves of the castor oil plant. 45±54. Arch. Biochem. Biophys. 333, 298±307. Bachmann, M., Shiraishi, N., Campbell, W.H., Yoo, B.-C., Harmon, Hu, Z., Zhang, Z.S., Doo, E., Coux, O., Goldberg, A.L. and Liang, T.J. A.C. and Huber, S.C. (1996) Identi®cation of Ser-543 as the major (1999) Hepatitis B vitrus X protein is both a substrate and a regulatory phosphorylation site in spinach leaf nitrate reduc- potential inhibitor of the proteasome complex. J. Virol. 73, 7231± tase. Plant Cell, 8, 505±517. 7240. Biederer, T., Volkwein, C. and Sommer, T. (1997) Role of Cue1p in Huang, Z.J. and Huber, S.C. (2001) Phosphorylation of synthetic ubiquitination and degradation at the ER surface. Science, 278 peptides by a CDPK and plant SNF1-related protein kinase. (5344), 1806±1809. In¯uence of proline and basic amino acid residues at selected Blakeley, S.D., Plaxton, W.C. and Dennis, D.T. (1990) Cloning and positions. Plant Cell Physiol. 42, 1079±1087. characterization of a cDNA for the cytosolic isozyme of plant Imamura, K.and Tanaka, T. (1982) Pyruvate kinase isozymes from pyruvate kinase: the relationship between the plant and non- rat. Meth. Enzymol. 90, 150±165. plant enzyme. Plant Mol. Biol. 15, 665±669. Kiefer, E., Heller, W. and Ernst, D. (2000) A simple and ef®cient Brazill, D.T., Thorner, J. and Marin, G.S. (1997) Mck1, a member of protocol for isolation of functional RNA from plant tissue rich in the glycogen synthase kinase 3 family of protein kinases, is a secondary metabolites. Plant Mol. Biol. Rep. 18, 33±39. negative regulator of pyruvate kinase in the yeast Saccharo- Kisselev, A.F., Akopian, T.N., Woo, K.M. and Goldberg, A.L. (1999) myces cerevisiae. J. Bacteriol. 179, 4415±4423. The size of peptide generated from protein by mammalian 26 Callis.J.and Vierstra, R.D. (2000) Protein degradation in signaling. and 20S proteasomes. J. Biol. Chem. 274, 3363±3371. Curr. Opin. Plant Biol. 3, 381±386. Ljungstrom, O., Berglund, L. and Engstrom, L. (1976) Studies on Cough, R.C. and Vierstra, R.D. (1997) Phytochrome degradation. the kinetic effects of adenosine-30:50-monophosphate-depen- Plant Cell Environ. 20, 713±721. dent phosphorylation of puri®ed pig-liver pyruvate kinase type Doelling, J.H., Yan, N., Kurepa, J., Walker, J. and Viestra, R.D. L. Eur. J. Biochem. 68, 497±506. (2001) The ubiquitin-speci®c protease UBP14 is essential for Lone, Y.-C., Simon, M.-P., Kahn, A. and Marie, J. (1986) Complete early embryo development in Arabidopsis thaliana. Plant J. 27, nucleotide and deduced amino acid sequence of rat L-type 393±405. pyruvate kinase. FEBS Lett. 195, 97±100. Doucet, J.P., Murphy, B.J. and Tuana, B.S. (1990) Modi®cation of a Miller, W.A., Waterhouse, P.M., Brown, J.W.S. and Browning, K.S. discontinuous and highly porous sodium dodecyl sulfate-poly- (2001) The RNA world in plants: post-transcriptional control III. acrylamide gel system for mini gel electrophoresis. Anal. Bio- Plant Cell, 13, 1710±1717. chem. 190, 209±211. Noguchi, T., Inoue, H. and Tanaka, T. (1986) The M1- and M2-tryp Engstrom, L. (1980) Regulation of liver pyruvate kinase by phos- isozymes of rat pyruvate kinase are produced from the same phorylation±dephosphorylation. In Molecular Aspects of Cellu- gene by alternative RNA splicing. J. Biol. Chem. 261, 13807± lar Regulation. Recently Discovered Systems of Enzyme 13812. Regulation by Reversible Phosphorylation, Vol.1 (Cohen, P., O'Hagan, R.C., Ohh, M., David, G., Alboran, I., Moreno, D., Alt, ed.). New York: Elsevier/North-Holland Biomedical Press, pp. F.W., Kaelin, J.W.G. and DePinho, R.A. (2000) Myc-enhanced 11±31. expression of cull promotes ubquitin-dependent proteolysis and Estelle, M. (2001) Proteases and cellular regulation in plants. Curr. cell cycle progression. Genes Dev. 14, 2185±2191. Opin. Plant Biol. 4, 254±260. Peng, Z., Staub, J.M., Serino, G., Kwok, S.F., Kurepa, J., Bruce, Ferreira, R.M.B., Ramos, P.C.R., Franco, E., Ricardo, C.P.P. and B.D., Vierstra, R.D., Wei, N. and Deng, X.W. (2001) The cellular Teixeira, A.R.N. (1995) Changes in ubiquitin and ubiquitin±pro- level of PR500, a protein complexrelated to the 19S regulatory tein conjugates during seed formation and germination. J. Exp. particle of the proteasome, is regulated in response to stresses Bot. 46, 211±219. in plants. Mol. Biol. Cell, 12, 383±392. Fujinami, K., Tanahashi, N., Tanaka, K., Ichihara, A., Cejka, Z., Platt, S.G. and Bassham, J.A. (1978) Photosynthesis and increased Baumeister, W., Miyawaki, M., Sato, T. and Nakagawa, H. production of protein. Adv. Expl. Med. Biol. 105, 195±247. (1994) Puri®cation and characterization of the 26S proteasome Plaxton, W.C. (1996) The organization and regulation of plant from spinach leaves. J. Biol. Chem. 269, 25905±25910. glycolysis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 47, 185±214. Gray, W.M., Kepinski, S., Rouse, D., Leyser, O. and Estelle, M. Podesta, F.E. and Plaxton, W.C. (1991) Kinetic and regulatory (2001) Auxin regulates SCF (TIR1)-dependent degradation of properties of cytosolic pyruvate kinase from germinating castor AUX/IAA proteins. Nature, 414, 271±276. oil seeds. Biochem. J. 279, 495±501. Halford, N.G. and Hardie, D.G. (1998) SNF1-related protein kinases: Podust, V.N., Brownell, J.E., Gladysheva, T.B., Luo, R.S., Wang, C., global regulators of carbon metabolism in plants? Plant Mol. Coggins, M.B., Pierce, J.W., Lightcap, E.S. and Chau, V.A. (2000) Biol. 37, 735±748. Nedd8 conjugation pathway is essential for proteolytic targeting Hardin, S., Winter, H. and Huber, S.C. (2002) Sucrose synthase is of p27Kip1 by ubiquitination. Proc. Natl. Acad. Sci. USA, 97, phosphorylated on a novel secondary site (serine 170) that may 4579±4584. promote proteasome-dependent degradation. In Proceedings of Rivoal, J., Smith, C.R., Moraes, T.F., Turpin, D.H. and Plaxton, W.C. Plant Biology 2002, 3±7 August, Deveur, Colorado. Rockville, (2002) A method for activity staining after native polyacrylamide MD: American Society of Plant Biologists, (http://www.aspb.org/ gel electrophoresis using a coupled enzyme assay and ¯uores- annual-meeting/pb-2002/2002.html) (Abstract #752). cence detection: application to the analysis of several glycolytic Harmon, A.C., Gribskov, M. and Harper, J.F. (2000) CDPKs ± a enzymes. Anal. Biochem. 300, 94±99. kinase for every Ca2 signal? Trends Plant Sci. 5, 154±159. Ruuska, S.A., Girke, T., Benning, C. and Ohlrogge, J.B. (2002) Harper, J.W. (2002) A phosphorylation-driven ubiquitination Contrapuntal networks of gene expression during Arabidopsis switch for cell-cycle control. Trends Cell Biol. 12, 104±107. seed ®lling. Plant Cell, 14, 1191±1206.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93 The processing and degradation of pyruvate kinase 93

Schwechheimer, C.and Deng, X.W. (2001) COP9 signalosome Turner, W.L. and Plaxton, W.C. (2000) Puri®cation and character- revisited: a novel mediator of protein degradation. Trends Cell ization of cytosolic pyruvate kinase from banana fruit. Biochem. Biol. 11, 420±426. J. 352, 875±882. Shrestha, M.K., Peri, I., Smirnoff, P., Birk, Y. and Golan-Goldhirsh, Ueno, Y., Imanari, E., Emura, J., Yoshizawa-Kumagaye, K., Naka- A. (2002) Jojoba seed meal proteins associated with proteo- jima, K., Inami, K., Shiba, T., Sakakibara, H., Sugiyama, T. and lytic and protease inhibitory activities. J. Agric. Food Chem. 50, Izui, K. (2000) Immunological analysis of the phosphorylation

5670±5675. state of maize C4-form phosphoenolpyruvate carboxylase with Smith, C.R., Knowles, V.L. and Plaxton, W.C. (2000) Puri®cation speci®c antibodies raised against a synthetic phosphorylated and characterization of cytosolic pyruvate kinase from Brassica peptide. Plant J. 21, 17±26. napus (rapeseed) suspension cell cultures. Eur. J. Bichem. 267, de Virgilio, M., Weninger, H. and Ivessa, N.E. (1998) Ubiquitination 4477±4485. is required for the retro-translocation of a short-lived luminal Smith, A.J., Rinne, R.W. and Seif, R.D. (1989) Phosphoenolpyru- endoplasmic reticulum glycoprotein to the cytosol for degrada- vate carboxylase and pyruvate kinase involvement in protein tion by the proteasome. J. Biol. Chem. 273, 9734±9743. and oil biosynthesis during soybean seed development. Crop Wu, H.B. and Turpin, D.H. (1992) Puri®cation and characterization Sci. 29, 349±353. of pyruvate kinase from the green algae Chlamydomonas rein- hardtii. J. Phycol. 28, 472±581.

ß Blackwell Publishing Ltd, The Plant Journal, (2003), 34, 77±93