Structural and Kinetic Properties of High and Low Molecular Mass Phosphoenolpyruvate Carboxylase Isoforms from the Endosperm of Developing Castor Oilseeds*

Phosphoenolpyruvate carboxylase (PEPC) is believed to play an important role in producing malate as a substrate for fatty acid synthesis by leucoplasts of the developing castor oilseed (COS) endosperm. Two kinetically distinct isoforms of COS PEPC were resolved by gel filtration chromatography and purified. PEPC1 is a typical 410-kDa homotetramer composed of 107-kDa subunits (p107). In contrast, PEPC2 exists as an unusual 681-kDa hetero-octamer composed of the same p107 found in PEPC1 and an associated 64-kDa polypeptide (p64) that is structurally and immunologically unrelated to p107. Relative to PEPC1, PEPC2 demonstrated significantly enhanced thermal stability and a much lower sensitivity to allosteric activators (Glc-6-P, Glc-1-P, Fru-6-P, glycerol-3-P) and inhibitors (Asp, Glu, malate) and pH changes within the physiological range. Nondenaturing PAGE of clarified extracts followed by in-gel PEPC activity staining indicated that the ratio of PEPC1:PEPC2 increases during COS development such that only PEPC1 is detected in mature COS. Dissimilar developmental profiles and kinetic properties support the hypotheses that (i) PEPC1 functions to replenish dicarboxylic acids consumed through transamination reactions required for storage protein synthesis, whereas (ii) PEPC2 facilitates PEP flux to malate in support of fatty acid synthesis. Interestingly, the respective physical and kinetic properties of COS PEPC1 and PEPC2 are remarkably comparable with those of the homotetrameric lowM r Class 1 and heteromeric highM r Class 2 PEPC isoforms of unicellular green algae.

Phosphoenolpyruvate carboxylase (PEPC) is believed to play an important role in producing malate as a substrate for fatty acid synthesis by leucoplasts of the developing castor oilseed (COS) endosperm. Two kinetically distinct isoforms of COS PEPC were resolved by gel filtration chromatography and purified. PEPC1 is a typical 410-kDa homotetramer composed of 107-kDa subunits (p107). In contrast, PEPC2 exists as an unusual 681-kDa hetero-octamer composed of the same p107 found in PEPC1 and an associated 64-kDa polypeptide (p64) that is structurally and immunologically unrelated to p107. Relative to PEPC1, PEPC2 demonstrated significantly enhanced thermal stability and a much lower sensitivity to allosteric activators (Glc-6-P, Glc-1-P, Fru-6-P, glycerol-3-P) and inhibitors (Asp, Glu, malate) and pH changes within the physiological range. Nondenaturing PAGE of clarified extracts followed by in-gel PEPC activity staining indicated that the ratio of PEPC1:PEPC2 increases during COS development such that only PEPC1 is detected in mature COS. Dissimilar developmental profiles and kinetic properties support the hypotheses that (i) PEPC1 functions to replenish dicarboxylic acids consumed through transamination reactions required for storage protein synthesis, whereas (ii) PEPC2 facilitates PEP flux to malate in support of fatty acid synthesis. Interestingly, the respective physical and kinetic properties of COS PEPC1 and PEPC2 are remarkably comparable with those of the homotetrameric low M r Class 1 and heteromeric high M r Class 2 PEPC isoforms of unicellular green algae.
Phosphoenolpyruvate carboxylase (PEPC) 1 is a ubiquitous cytosolic enzyme in vascular plants that is also widely distributed in green algae and bacteria (1). It catalyzes the irreversible ␤-carboxylation of PEP in the presence of Mg 2ϩ and HCO 3 Ϫ to yield oxaloacetate and P i . PEPC is abundant in C 4 and crassulacean acid metabolism (CAM) leaves where it partici-pates in photosynthesis by catalyzing the initial fixation of atmospheric CO 2 . Both allosteric mechanisms and covalent modification are involved in PEPC control in C 4 and CAM leaves (1)(2)(3). Early work established that C 4 and CAM PEPCs are controlled by a diurnal cycle that modulates their sensitivity to L-malate inhibition (1)(2)(3). This cycle is the result of phosphorylation of the PEPC homotetramer by an endogenous Ca 2ϩ -independent PEPC protein kinase and dephosphorylation by a protein phosphatase type 2A at a highly conserved seryl residue localized near the N terminus of the 100 -110-kDa PEPC subunit (1)(2)(3).
Relative to C 4 and CAM PEPCs, the properties of the enzyme from non-green plant tissues are less well understood. Although proposed roles for nonphotosynthetic PEPCs are diverse, a crucial PEPC function is the anaplerotic replenishment of citric acid cycle intermediates consumed during biosynthesis and nitrogen assimilation (1). As with C 4 and CAM PEPCs, the PEPC of C 3 leaves and nonphotosynthetic tissues can be controlled by allosteric effectors and reversible phosphorylation (4 -10). However, despite the probable central role of PEPCs in the metabolism of developing and germinating seeds (11)(12)(13)(14)(15)(16), no seed PEPC has been fully purified and thoroughly characterized.
Storage lipids account for as much as 65% of the weight of mature castor oilseeds (COS). Triacylglyceride accumulation depends on the synthesis of long chain fatty acids, which in developing oilseeds occurs in specialized plastids termed leucoplasts. This process requires the transport of both sucrosederived carbon skeletons and energetic intermediates across the plastid envelope (17). L-Malate supports significant rates of fatty acid synthesis by isolated leucoplasts from developing COS (18). Malate imported from the cytosol into the leucoplast stroma is mediated by a malate/P i translocator within the COS leucoplast envelope (19). Sangwan and co-workers (16) hypothesized that the large increase in PEPC activity and concentration that accompanies COS development facilitates malate production for fatty acid synthesis. The increased PEP to malate flux would also serve as an anaplerotic source of C-skeletons for transamination reactions associated with COS storage protein synthesis.
The aim of this study was to purify and characterize PEPC from developing COS. Here we present unexpected evidence for two PEPC isoforms from developing COS and examine their structural and kinetic properties. Although one isoform is a typical PEPC homotetramer, the other represents a unique high M r PEPC complex unprecedented in vascular plants but remarkably reminiscent of Class 2 PEPC isoforms recently described in unicellular green algae (20 -23). We provide evidence that the association of a common 107-kDa PEPC catalytic subunit with an unrelated but PEPC-like 64-kDa polypeptide is responsible for the dramatic differences in the physical and kinetic properties observed between the PEPC homotetramer and high M r PEPC complex of developing COS.

EXPERIMENTAL PROCEDURES
Plant Material-Castor plants (Ricinus communis L., cv Baker 296) were cultivated in a greenhouse at 24°C and 60% relative humidity under natural light supplemented with 16 h of artificial light. COS were harvested at stages of development previously described (24). Dissected endosperm (free of cotyledon) was frozen in liquid N 2 and stored at Ϫ80°C.
Enzyme and Protein Assays and Kinetic Studies-PEPC activity was assayed at 25°C using a Molecular Devices microplate reader as previously described (8). Standard assay conditions were: 100 mM Hepes-KOH (pH 8), 10% (v/v) glycerol, 2.5 mM PEP, 5 mM KHCO 3 , 5 mM MgCl 2 , 2 mM dithiothreitol, 0.15 mM NADH, and 5 units/ml desalted malate dehydrogenase. All assays were corrected for background NADH oxidation and were linear with respect to time and the concentration of enzyme assayed. One unit of PEPC activity is defined as the amount of enzyme resulting in the production of 1 mol of oxaloacetate min Ϫ1 . Protein concentration was determined by the Coomassie Blue G-250 (25) or bicinchoninic acid (26) colorimetric methods using bovine ␥-globulin as the protein standard.
Apparent V max (V max,app ), K m , and I 50 and K a values (concentrations of inhibitors and activators producing 50% inhibition or activation of PEPC activity, respectively) were calculated using Brooks' computer program (27). All kinetic parameters represent means of at least three separate determinations and are reproducible to within Ϯ 10% (S.E.) of the mean value. Stock solutions of metabolites were made equimolar with MgCl 2 and adjusted to pH 7.5.
Stage VII developing COS endosperm (100 g) was homogenized (1:2; w/v) in buffer A using a Polytron. After centrifugation, the supernatant was filtered through two layers of Miracloth and recentrifuged. PEG 8000 (50% (w/v) in 50 mM Hepes-KOH, pH 7.5) was added to the supernatant to a final concentration of 20% (w/v) and stirred for 30 min. After centrifugation, PEG pellets were resuspended in buffer B (lacking (NH 4 ) 2 SO 4 ) to a final protein concentration of about 10 mg/ml. After centrifugation, solid (NH 4 ) 2 SO 4 was added to the supernatant to 20% (saturation). The solution was stirred for 20 min, centrifuged, and loaded at 4 ml/min onto a column (3.2 ϫ 5.7 cm) of butyl-Sepharose 4 Fast Flow equilibrated with buffer B. The column was connected to an Ä KTA FPLC system and washed with buffer B until the A 280 approached base line. PEPC activity was eluted by 50% buffer C (50% buffer B) (9 ml/fraction). Pooled peak fractions were diluted with an equal volume of 50% (w/v) PEG 8000, stirred for 30 min, and centrifuged. PEG pellets were dissolved in buffer D to a protein concentration of approximately 10 mg/ml, centrifuged as described above, and loaded at 0.6 ml/min onto a column (1.1 ϫ 12 cm) of Fractogel EMD DEAE-650 (S) that had been connected to the FPLC and equilibrated with buffer D. The column was washed with buffer D, and PEPC activity was eluted with a 0 -400 mM KCl gradient (96 ml) in buffer D (3 ml/fraction). Pooled peak fractions were concentrated to 1 ml using an Amicon YM-30 ultrafilter and applied at 0.3 ml/min onto a Superdex 200 HR 16/50 column equilibrated with buffer E (1 ml/fraction). Two PEPC activity peaks were resolved (see "Results"). Each PEPC pool was concentrated to approximately 0.2 ml and applied separately at 0.2 ml/min onto a Superose-6 HR 10/30 column equilibrated with buffer E. Pooled peak fractions were concentrated to Ͻ1 ml, divided into 25-l aliquots, frozen in liquid N 2 , and stored at Ϫ80°C. PEPC activity was stable for at least 3 months when stored frozen. In some instances, the final PEPC2 preparation was subjected to anion-exchange FPLC on a Mono-Q HR 5/5 column equilibrated in buffer D.
Electrophoresis and Immunoblotting-SDS and nondenaturing PAGE, subunit M r estimates via SDS-PAGE, and immunoblotting using affinity-purified rabbit anti-(Brassica napus PEPC) IgG were performed as described (8,28). Gels were stained for protein with Coomassie Blue R-250 or Sypro Red or for PEPC activity using the Fast Violet B method (28). Sypro Red-stained gels were visualized using a Typhoon 8600 fluorescence imager, and the relative band densities were quantified using software provided by the manufacturer. For second dimension PAGE, Coomassie Blue-stained PEPC was excised from a nondenaturing gel and subjected to SDS-PAGE as described (9).
N-terminal Sequencing and Mass Spectrometry-Sequencing was performed by automated Edman degradation at the Biotechnology Research Institute (Montreal, Quebec, Canada). Similarity searches were performed using the BLAST program available on the National Center for Biotechnology Information web site (www.ncbi.nlm.nih. gov/). For MS analyses, Coomassie Blue-stained polypeptides were excised from SDS gels and digested with trypsin using standard protocols. Samples destined for MALDI-TOF MS were mixed 1:1 (v/v) with 5 mg/ml ␣-cyanohydroxycinnamic acid matrix in 50% (v/v) acetonitrile and 0.1% (v/v) trifluoroacetic acid before analysis on a Micromass MALDI (positive ion mode). Samples destined for ESI/Q-TOF MS/MS were applied automatically by capillary liquid chromatography to a Micromass Q-TOF Ultima GLOBAL using a C18 column (75 m ϫ 150 mm) running at 200 nl/min. Mass data from both machines were used to search the NCBInr data base using MASCOT (www.matrixscience.com).

Purification and Physical, Immunological, and Structural Characterization
PEPC Purification-Initial purification trials resulted in a PEPC preparation (specific activity, 28 units/mg) having a native molecular mass of approximately 406 kDa as estimated via analytical gel filtration (Fig. 1). SDS-PAGE and immunoblot analysis of this preparation revealed an approximate 1:1 ratio of 107-and 98-kDa protein-staining polypeptides (p107 and p98, respectively) that cross-reacted with anti-(B. napus PEPC) IgG with similar intensities (Fig. 2, A and B, lane 1). BLAST analysis of the N-terminal amino acid sequence of each polypeptide indicated that p98 probably arose via the action of a COS endopeptidase that hydrolyzed an approximate 120amino acid polypeptide from the N terminus of p107 (Fig. 3). Thus, we modified the purification protocol by adding 2 mM 2,2Ј-dipyridyl disulfide to the extraction buffer and 5 mM malate to all purification buffers. 2,2Ј-Dipyridyl disulfide is an active site-directed covalent affinity label of papain (29) that also suppresses the activity of COS cysteinyl endopeptidase(s) (30). Malate helps to preserve the integrity of the N-terminal phosphorylation domain of vascular plant PEPCs during extraction and subsequent purification (1). With these buffer additions, partial degradation of p107 was prevented (Fig. 2, A  and B). Moreover, two distinct peaks of PEPC activity were resolved during Superdex 200 FPLC (Fig. 1B). Additional gel filtration via Superose 6 FPLC resulted in an approximately 200-fold purification of PEPC1 and PEPC2 to a final specific activity of approximately 10 units/mg (Table I). Although Superose 6 FPLC did not increase the specific activity of PEPC1 or PEPC2 beyond that achieved at the Superdex 200 step, it was included to ensure a clean separation of PEPC1 from PEPC2. With the bicinchoninic acid protein assay (26) the specific activity of the final PEPC1 and PEPC2 preparations was increased to 24.2 and 29.2 units/mg, respectively. Calibration of the Superdex 200 column with molecular mass standards yielded respective native molecular masses of 410 Ϯ 5 kDa for PEPC1 and 681 Ϯ 9 kDa for PEPC2 (n ϭ 3).
PAGE and Immunoblot Analysis-A Coomassie Blue-and PEPC activity-staining polypeptide that cross-reacted with anti-(B. napus PEPC) IgG was observed after the nondenaturing PAGE of PEPC1 and PEPC2 (Fig. 4, A-C, lanes 1 and 2). This analysis was consistent with the respective native M r estima-tions by gel filtration, because the smaller PEPC1 migrated significantly further than the larger PEPC2. The additional faster migrating protein-staining polypeptide observed during nondenaturing PAGE of PEPC2 (Fig. 4A, lane 2) is a probable contaminant. It did not stain for PEPC activity or cross-react with the anti-PEPC IgG (Fig. 4, B and C, lane 2) and was eliminated when the final PEPC2 preparation was subjected to Mono-Q FPLC (Fig. 1C).
SDS-PAGE of the final PEPC1 and PEPC2 preparations resolved a protein-staining p107 that cross-reacted with anti-(B. napus PEPC) IgG (Fig. 2, A and B, lanes 2 and 3). However, PEPC2 contained two additional protein-staining polypeptides of approximately 64-kDa (p64) and 57-kDa ( Fig. 2A, lane 3). When the protein-and PEPC activity-staining band obtained after nondenaturing PAGE of PEPC2 was excised, equilibrated with SDS, and subjected to SDS-PAGE, p107 and p64 were resolved (Fig. 2A, lane 4), indicating that the p64 was complexed with p107 in the native PEPC2. This result was corrob-orated by SDS-PAGE of fractions collected during analytical Mono-Q FPLC of PEPC2 in which both the p107 and p64 co-eluted with PEPC activity and a symmetrical A 280 peak (Fig.  1C). The p64 was not recognized by the anti-(B. napus PEPC) IgG (Fig. 3A). Densitometric analysis of Sypro Red-stained SDS gels of Mono-Q-purified PEPC2 allowed us to estimate a p107:p64 molar ratio of 1:1. The native PEPC2 therefore appears to exist as an unusual hetero-octomeric complex composed of four p107 and four p64 subunits with a combined theoretical molecular mass of 684 kDa. This value closely agrees with the molecular mass of 681 kDa estimated for native PEPC2 during Superdex 200 FPLC. PEPC1, by contrast, is a typical PEPC homotetramer of p107 subunits, which likely correspond to the same p107 found in PEPC2.
Nondenaturing PAGE of clarified extracts (prepared in buffer A) followed by in-gel PEPC activity staining indicated that the ratio of PEPC1:PEPC2 progressively increases during COS development such that only PEPC1 is detected in stage IX (maturation phase) and dry (mature) COS (Fig. 4B). Overall PEPC activity was relatively abundant, peaking at approximately 2.6 units⅐(g fresh weight) Ϫ1 in stage VII developing COS endosperm (Fig. 4B, inset). Immunoblots of the same clarified extracts were probed with the anti-(B. napus PEPC) IgG and demonstrated that p107 was present throughout COS development but declined after stage VII (Fig. 2B). Immunoreactive polypeptides of 98 -102 kDa were detected in the later developmental stages. These polypeptides may represent in vivo proteolytic degradation products of p107, because the same antigenic polypeptides were observed on an immunoblot of stage IX COS extract prepared under denaturing conditions in 10% (w/v) trichloroacetic acid (31) (results not shown).
N-terminal Sequencing and Mass Spectrometry-The 20 Nterminal amino acids of the p107 and p98 of proteolyzed PEPC were sequenced by Edman degradation (Fig. 3). BLAST analysis revealed significant matches with the corresponding region of various plant PEPCs and included the conserved regulatory seryl phosphorylation site found in all plant PEPCs Lanes 4 and 5 contain 2 g of PEPC1 and PEPC2, respectively, that had been excised from a nondenaturing gel (PEPC active bands; see Fig. 4), equilibrated with SDS, and subjected to SDS-PAGE. Staining was performed using Coomassie Blue R-250. B, immunoblot analysis was performed using affinity-purified rabbit anti-(B. napus PEPC) IgG (8). Lanes 1-3 contain 50 ng each of purified proteolyzed PEPC, PEPC1, and PEPC2, respectively. The remaining lanes contain clarified extracts (corresponding to 1.5 mg of fresh tissue) from various stages of COS development. Lanes (developmental stage) III, V, VII, and IX correspond to the heart-shaped embryo, midcotyledon, full cotyledon, and maturation stages of development, respectively (24). Lane "Dry" designates a fully mature COS. Asterisk denotes the stage at which PEPC purification was conducted. examined to date (Fig. 3). The sequences of 12 amino acid residues of the N termini of the p107 of PEPC1 and PEPC2 were determined and found to be identical to that of the p107 of proteolyzed PEPC. This result indicates that PEPC1, PEPC2, and proteolyzed PEPC may share a common p107. This idea was corroborated by MALDI-TOF MS analysis of PEPC1 and PEPC2 p107 tryptic peptides. Their mass fingerprints were very similar, with 21 identical peptides (results not shown). Both mass fingerprints best matched the same tomato (Lycopersicon esculentum) PEPC (GenBank TM GI number 6688531).
The p64 of PEPC2 was also subjected to MALDI-TOF MS. However, the p64 mass fingerprint did not produce any significant matches in the NCBI data base. In addition, none of the 19 tryptic peptides of p64 occurred in those of PEPC1 or PEPC2 p107 (results not shown). The p64 was further analyzed by Q-TOF MS/MS. The Q-TOF data best matched two putative PEPCs from rice (Oryza sativa) and Arabidopsis genome sequence databases (Fig. 5A). Neither of these PEPCs was identified by MALDI-TOF MS of p107, further suggesting that p107 and p64 are structurally dissimilar. The most significant hit matched 7 individual peptides of a putative rice PEPC (Fig.  5B). BLAST analyses of this 102-kDa rice PEPC indicated that its first 70 N-terminal amino acids are unique compared with most plant PEPCs, although the full-length sequence contains all three conserved domains believed to be required for PEPC activity (Fig. 5B) (1).

Kinetic Properties
Effect of pH and PEP Saturation Kinetics-Similar to other plant PEPCs (1), PEPC1 and PEPC2 activity increased with pH in the range of 6.5-8.0. However, PEPC1 exhibited optimal activity at pH 8.5, whereas PEPC2 displayed optimal activity at pH 8.0 (Fig. 6). Moreover, PEPC1 displayed a significantly greater sensitivity to pH changes within the physiological range. Between pH 8 and 6.5, PEPC1 activity decreased by more than 30-fold, whereas PEPC2 activity decreased by less than 3-fold (Fig. 6).
PEP saturation kinetics and response to metabolite effectors

FIG. 3. Comparison of the N-terminal sequences of COS p107 (A) and p98 (B) with those of other plant
PEPCs. The COS (R. communis) PEPC sequences were determined by Edman sequencing of p107 and p98 of the "proteolyzed PEPC" preparation (see Fig. 2, A  and B, lane 1). Other PEPC sequences were derived by translation of the corresponding genes. Sequence numbering represents amino acid position relative to the N terminus. Hyphens denote amino acid residues that are identical to those of the respective COS PEPC sequences. An asterisk indicates the conserved regulatory seryl phosphorylation site, and underlined letters indicate the consensus target sequence for plant PEPC protein kinase (1).
were determined at pH 8 and 7.3. Hyperbolic PEP saturation kinetics was always observed. At pH 8.0 and 7.3 the respective V max,app values of PEPC1 were 24.5 and 20.7 units/mg, whereas those of PEPC2 were 29.7 and 28.8 units/mg. No difference was noted in the K m (PEP) values for PEPC1 and PEPC2, which were both approximately 0.06 and 0.12 mM at pH 8 and 7.3, respectively.
Metabolite Effects-A wide variety of compounds were tested as possible effectors of PEPC1 and PEPC2 at pH 7.3 and pH 8 with subsaturating PEP (0.2 mM). The following compounds exerted little or no influence on the activity of either isoform (Ϯ 20% of the control rate): 2-P-glycerate, dihydroxyacetone-P, Fru-1,6-P 2 , NAD ϩ , Gly, Glu, Arg, Ala, Leu, Asn, Phe, pyruvate, and AMP (2 mM each); CoA, malonyl-CoA, acetyl-CoA, and oleyl-CoA (50 M each). Table II lists those compounds that significantly influenced PEPC activity. Similar to other plant PEPCs (1), PEPC1 and PEPC2 displayed pH-dependent modulation by several metabolites that were more effective at pH 7.3 than at pH 8. PEPC1, however, was much more sensitive to the various metabolite effectors than PEPC2 (Table II). PEPC1 was potently activated at pH 7.3 by the hexose-mono-Ps and by glycerol-3-P, whereas PEPC2 was only weakly activated by these compounds. Similarly, PEPC1 was far more sensitive to inhibition by malate, Asp, Glu, and ATP, relative to PEPC2. Fig. 7 demonstrates the marked differential response of PEPC1 and PEPC2 activity to increasing concentrations of the most widely recognized allosteric effectors of plant PEPC (1), namely malate and Glc-6-P. DISCUSSION When partial in vitro proteolysis of p107 was prevented, two COS PEPC isoforms that significantly differed in their physical and kinetic properties were resolved by Superdex 200 FPLC and highly purified. Tissue-and/or developmentally specific PEPC isozymes are known to occur in vascular plants (1)(2)(3), and genetic evidence indicates that developing Glycine max (soybean) and Vicia faba seeds express more than one PEPC gene (15,32,33). To our knowledge, this is the first report of the isolation of two PEPC isoforms from the same plant tissue.
COS PEPC1 is a p107 homotetramer, typical of most other plant PEPCs studied to date. By contrast, PEPC2 appears to exist as an unusual 681-kDa hetero-octamer composed of the same p107 found in PEPC1 and an associated p64 that is structurally and immunologically unrelated to p107. Nevertheless, Q-TOF MS/MS analysis of tryptic peptides revealed that p64 is highly similar to two putative PEPCs identified by annotation of the rice and Arabidopsis genomes (Fig. 5). Although they contain conserved regions required for PEPC activity, these putative PEPCs exhibit a unique N-terminal region that lacks the regulatory seryl phosphorylation site thought to be conserved among all plant PEPCs (1) and have predicted molecular masses of 102 (rice) and 110 kDa (Arabidopsis). Although it is feasible that p64 represents an in vivo or in vitro degradation product of a larger polypeptide, the data indicate that p64 is a PEPC-like polypeptide that interacts with p107 to give rise to the COS PEPC2 heteromeric complex. This association may either physically block the allosteric sites FIG. 4. Nondenaturing PAGE (5% gel) analysis of COS PEPC isoforms. A, staining was performed with Coomassie Blue R-250. Lanes 1 and 2, respectively, contain 1 g each of purified PEPC1 and PEPC2. B, immunoblot analysis was performed using affinity-purified rabbit anti-(B. napus PEPC) IgG (8). Lanes 1 and 2, respectively, contain 50 ng of purified PEPC1 and PEPC2. C, in-gel PEPC activity staining was performed as described (9). Lanes 1 and 2, respectively, contain 1 g each of purified PEPC1 and PEPC2. The remaining lanes contain clarified extracts from various stages of COS development, as described in the legend for Fig. 2. Inset, corresponding PEPC activity in clarified extracts of COS endosperm (means Ϯ S.E., n ϭ 3). Ϫ using the standard assay as described under "Experimental Procedures," except that the pH was varied and a mixture of 50 mM MES and 50 mM bis-Tris-propane was used as the buffer. The pH of kinetic assays was determined immediately after the completion of each set of assays. All values represent the means of three separate determinations and are reproducible to within Ϯ10% (S.E.) of the mean value.
of p107 or promote an allosteric transition in p107 such that effectors have limited access to their respective sites.
COS PEPC1 and PEPC2 Emulate Green Algal "Class 1" and "Class 2" PEPC Isoforms-Interestingly, the respective properties of COS PEPC1 and PEPC2 are remarkably comparable with those of the homotetrameric low M r Class 1 and heteromeric high M r Class 2 PEPC isoforms of the green algae Selenastrum minutum and Chlamydomonas reinhardtii (20 -22). All algal PEPC isoforms share the same 102-kDa catalytic subunit (p102). Similar to COS PEPC2, the algal Class 2 PEPCs also contain associated polypeptides that are immunologically unrelated to p102. MALDI-TOF MS and microsequencing revealed that like the p64 of COS PEPC2, the 130-kDa subunit (p130) of S. minutum Class 2 PEPC represents a distinct PEPC-like polypeptide that is only distantly related to p102 (22). Moreover, relative to COS PEPC1 and algal Class 1 PEPCs, the COS PEPC2 and algal Class 2 PEPCs demonstrate significantly enhanced thermal stability and a much lower sensitivity to allosteric effectors and pH changes within the physiological range (20,21). Taken together, the data imply that high and low M r PEPC isoforms arose in green algae before the evolution of vascular plants, with this feature being conserved as a key structure-function aspect of at least some plant PEPCs.
Possible Functions and Interconversion of COS PEPC1 and PEPC2-Similar to PEPCs from other non-green plant tissues (6,8,9,11,15), COS PEPC1 was activated by hexose-mono-Ps and potently inhibited by malate, Asp, and Glu at pH 7.3 (Table  II). This result indicates that PEPC1 may fulfill a key anaplerotic role to replenish dicarboxylic acids consumed through transamination reactions required to support storage protein synthesis. The inhibition of PEPC1 by Asp and Glu provides a tight feedback control that could closely balance PEPC1 activity with the production of C-skeletons (i.e. oxaloacetate, 2-oxoglutarate) required for NH 4 ϩ assimilation or transamination reactions. The "effector-insensitive" PEPC2, by contrast, may facilitate PEP flux to malate in support of leucoplast fatty acid synthesis despite the significant malate levels present in developing COS (18). Nondenaturing PAGE of clarified COS extracts followed by in-gel PEPC activity staining revealed that PEPC2 increases during COS development, peaking at stage VII, and then rapidly disappears during COS maturation (Fig.  4C). This pattern parallels triglyceride accumulation in this tissue, which also peaks at stage VII (34). The developmental profile for PEPC1 (Fig. 4C), by contrast, parallels that of storage protein accumulation (24), with both becoming maximal during the maturation phase of COS development. Further studies using transgenic plants and/or pharmacological inhibitors will help to fully evaluate the metabolic functions of COS PEPC1 and PEPC2.
It remains to be determined whether and how COS PEPC1 and PEPC2 interconvert. However, protein-kinase-mediated phosphorylation of p102 appears to be involved in the control and structural organization of green algal (S. minutum) Class 2 PEPCs (23). COS p107 contains the N-terminal regulatory seryl phosphorylation site characteristic of most plant PEPCs (Fig. 3A). It will be of interest to determine whether COS PEPC1 and PEPC2 are interconverted via a phosphorylationdephosphorylation mechanism involving p107 and/or p64.

TABLE II
Influence of various metabolites on the activity of COS PEPC1 and PEPC2 PEPC activity was determined at pH 7.3 or 8.0 with subsaturating PEP (0.2 mM) in the presence and absence of each effector at 2 mM. Activities are expressed relative to the respective control determined in the absence of any additions and set at 100% (PEPC1 control activity ϭ 18.9 and 12.2 units/mg at pH 8.0 and 7.3, respectively; PEPC2 control activity ϭ 22.8 and 16.8 units/mg at pH 8.0 and 7.3, respectively). Shown in parentheses are the K a (Glc-6-P, Fru-6-P, and glycerol-3-P) or I 50 (malate, Asp, Glu) values (expressed as mM) for several effectors (determined at pH 7.3 with 0.2 mM PEP). All values represent the mean of three separate determinations and are reproducible within Ϯ 10% (S.E.) of the mean value.