The Importance of the Strictly Conserved, C-terminal Glycine Residue in Phosphoenolpyruvate Carboxylase for Overall Catalysis

Phosphoenolpyruvate carboxylase (PEPC) is a “multifaceted,” allosteric enzyme involved in C4 acid metabolism in green plants/microalgae and prokaryotes. Before the elucidation of the three-dimensional structures of maize C4 leaf and Escherichia coli PEPC, our truncation analysis of the sorghum C4 homologue revealed important roles for the enzyme's C-terminal α-helix and its appended QNTG961 tetrapeptide in polypeptide stability and overall catalysis, respectively. Collectively, these functional and structural observations implicate the importance of the PEPC C-terminal tetrapeptide for both catalysis and negative allosteric regulation. We have now more finely dissected this element of PEPC structure-function by modification of the absolutely conserved C-terminal glycine of the sorghum C4 isoform by site-specific mutagenesis (G961(A/V/D)) and truncation (ΔC1/C4). Although the C4 polypeptide failed to accumulate in a PEPC- strain (XH11) of E. coli transformed with the Asp mutant, the other variants were produced at wild-type levels. Although neither of these four mutants displayed an apparent destabilization of the purified PEPC homotetramer, all were compromised catalytically in vivo and in vitro. Functional complementation of XH11 cells under selective growth conditions was restricted progressively by the Ala, ΔC1 and Val, and ΔC4 modifications. Likewise, steady-state kinetic analysis of the purified mutant enzymes revealed corresponding negative trends in kcat and kcat/K0.5 (phosphoenolpyruvate) but not in K0.5 or the Hill coefficient. Homology modeling of these sorghum C-terminal variants against the structure of the closely related maize C4 isoform predicted perturbations in active-site molecular cavities and/or ion-pairing with essential, invariant Arg-638. These collective observations reveal that even a modest, neutral alteration of the PEPC C-terminal hydrogen atom side chain is detrimental to enzyme function.

cies (1-4) (Fig. 1). It catalyzes the irreversible, biotin-independent carboxylation of phosphoenolpyruvate (PEP) in the presence of HCO 3 Ϫ and Mg 2ϩ (or Mn 2ϩ ) to yield P i and oxaloacetate and, thus, is involved intimately in photosynthetic and/or anaplerotic C4 dicarboxylic acid metabolism in these organisms. Overall catalysis is composed of a threestep reaction sequence in which carboxyphosphate and the enolate anion of pyruvate are deployed as key reaction intermediates (1,2). The former molecular species is subsequently cleaved to P i and free CO 2 within the active site for ␤-carboxylation of the enolate anion to oxaloacetate. In the above PEPC-containing organisms the enzyme is most often found as a tetramer of identical, ϳ95-110-kDa subunits, although notable exceptions to this general pattern exist in the Archaea (3) and green microalgae (4,5).
Given PEPC's cardinal roles in the photosynthetic assimilation of atmospheric CO 2 by certain green plants (i.e. the so-called C4 and Crassulacean acid metabolism (CAM) species), the C4 isoforms and their eubacterial homologues have been studied extensively during the past two decades by a combination of molecular, biochemical, and structural approaches. As a result, an impressive list of advances in PEPC research has been generated and subsequently reviewed (1, 2, 6 -9). Notable among these diverse findings is the realization that (i) only the vascular plant enzyme is co-regulated posttranslationally by both opposing, allosteric metabolite effectors and the reversible phosphorylation of a specific Ser residue near its N terminus by a dedicated, ϳ31-kDa Ser/ Thr kinase and opposing heterotrimeric protein phosphatase 2A (1)(2)(3)(4)8), (ii) several representative green plant and green algal species encode within their nuclear genomes two divergent, PEPC catalytic subunit types, one of which is most closely related phylogenetically to the eubacterial polypeptide that lacks the N-terminal phosphorylation domain typical of plant PEPC and ends in a C-terminal (R/K)NXG tetrapeptide instead of the plant-invariant QNTG motif (2,4,10,11) (Fig. 1), and (iii) both the eubacterial (Escherichia coli) and green plant (Zea mays (maize) C4 isoform) homotetramers are arranged structurally as a "dimer-of-dimers," with each monomer composed of 40 or 41 ␣-helices, respectively, and only 8 ␤-strands. The catalytic domain is situated at the C-terminal side of the 8-stranded ␤-barrel, and the two distinct positive and negative allosteric sites are distanced ϳ15 and 20 Å, respectively, from this domain (2,6,12,13). Notably, two of the mobile elements within the monomer three-dimensional structure contain residues that contribute to both active site and negative allosteric site interactions. Specifically, mobile loop I is composed of the strictly conserved, Gly-rich GXGXXXXRG motif (1-4, 6, 12). Its invariant Arg residue, situated at position 647, 638, and 587 in the maize and Sorghum bicolor (sorghum) C4 isoforms and E. coli enzyme, respectively, composes part of the allosteric inhibitor (Asp, L-malate) binding site in the Asp-bound, E. coli T-state structure (2,6,12). In contrast, in the active, R-state structure of maize C4-PEPC this same basic residue has been translocated ϳ20 Å to near the catalytic domain, marked by an invariant, essential His (at positions 177, 167, and 138, respectively) within the strictly conserved TXHP motif ( Fig. 2A). Both of these basic residues have been shown by mutagenesis to be indispensable for overall catalysis by the E. coli enzyme (2,6,14,15). Thus, loop I is presumably "trapped" away from the active site upon dicarboxylic acid inhibitor binding to the negative allosteric site by virtue of ionic interaction between this essential Arg side chain and one of the effector's two carboxyl groups. Likewise, in this inhibited T-state the C-terminal tetrapeptide that is appended to the final, buried ␣-helix is also re-oriented away from near the catalytic domain by H-bonding of its Asn side chain, situated within the NXG motif, characteristic of all eubacterial-and plant-type PEPCs (Fig. 1), to the allosteric inhibitor (2,6,12). Conversely, in the active, R-state structure of the maize C4 isoform this flexible terminal tetrapeptide is found re-positioned to near the active site where it ion-pairs with loop I's invariant and indispensable Arg side chain by virtue of its strictly conserved C-terminal Gly carboxyl (2, 6, 12) ( Figs. 1 and 2A). Thus, these recent, comparative structural observations implicate the importance of the PEPC C-terminal tetrapeptide for both negative allosteric regulation (Asn) and catalysis (invariant, C-terminal Gly) (2,6,12).
Before the elucidation of the active R-state structure of maize C4-PEPC (12, 13), our functional studies of the closely related, sor-ghum C4 homologue ( Fig. 1) indicated important roles for the enzyme's C-terminal domain in polypeptide stability and overall catalysis (16). Specifically, mutagenic truncation of the terminal 19 residues (⌬C19 mutant in Fig. 1), which deleted the polypeptide's final, hydrophobic ␣-helix encompassing Leu-942-Met-957 (see ␣41 in Fig. 1) and its adjacent, flexible QNTG 961 tetrapeptide, resulted in marked reductions in both C4 enzyme accumulation in transformed, PEPC Ϫ E. coli cells and in vivo and in vitro activity. In contrast, when only the plant-invariant, C-terminal tetrapeptide was "trimmed" from the polypeptide (⌬C4 mutant in Fig. 1), C4-PEPC levels were restored to near control, wild-type (WT) amounts, but yet maximal catalysis remained compromised by Ն99%. Additional studies revealed that the partially purified, ⌬C4 mutant protein was unaltered with respect to its ϳ440-kDa, tetrameric quaternary structure and the phosphorylatability of its N-terminal target Ser-8 residue in vitro (16,17). Clearly, these functional analyses together with the aforementioned structural observations (Refs. 2, 6, and 12; Fig.  2A) collectively underscore the importance of this flexible, C-terminal tetrapeptide in PEPC function (and negative allosteric regulation). In the present study we have more finely dissected this element of PEPC structure-function by specific perturbation of the strictly conserved C-terminal Gly residue at position 961 in the sorghum C4 isoform by site-directed mutagenesis and truncation (⌬C1 mutant in Fig. 1). The results presented herein highlight the importance of this  . Shading indicates the highly conserved C-terminal tetrapeptide residues in various PEPCs, where G is absolutely species-invariant, N is conserved in all PEPCs except archaeal-type, Q is plant-type PEPCs-only, and R/K is conserved in bacterial-type PEPCs-only. The mutants related to this and our previous (16) study were constructed by deleting 19 (⌬C19), 4 (⌬C4), or 1 (⌬C1) residue(s) from the extreme C terminus or by specifically mutating the invariant, C-terminal Gly to Ala, Val, or Asp in the sorghum C4 isoform. Note that (i) the sorghum and maize C4 homologues are 100% identical within these 31 C-terminal residues and 90% identical overall and (ii) the sorghum C4 isoform terminal ␣-helix 41 extends from Leu-942 to Met-957.
invariant, simple residue and its minimal, H-atom side chain for maximal in vivo activity and in vitro catalytic efficiency of this homotetrameric, allosteric enzyme.

EXPERIMENTAL PROCEDURES
Materials-All buffers and biochemical reagents were obtained from Sigma unless noted otherwise. Liquid [ 14 C]NaHCO 3 (55 Ci mol Ϫ1 ) was purchased from MP Biomedicals, Inc. (Irvine, CA), porcine heart L-malate dehydrogenase (ϳ12 units l Ϫ1 ) was from Roche Applied Science, and rabbit anti-(maize leaf PEPC) IgG was from CHEMICON International, Inc. (Temecula, CA). Prestained Protein-Plus mass standards for SDS-PAGE were from Bio-Rad, and the high molecular weight calibration kit for native PAGE was from Amersham Biosciences.
Construction of the Sorghum C4-Ppc Expression Vector-To obtain stable overexpression of sorghum C4-PEPC in a PEPC Ϫ (Ppc Ϫ ) strain (XH11) of E. coli (16,18,19), a novel vector was constructed and named Ppc-P6. This vector is 4702 base pairs in length and harbors a ColE1 origin of replication sequence, an ampicillin-resistance gene for antibiotic selection, and a multiple cloning site for convenient cloning, which is flanked by the highly expressed tac promoter and strong txn transcription terminator to prevent unstable replication (Fig. 3). A His 6 tag sequence was inserted between the promoter and the multiple cloning site for affinity purification of the expressed C4-PEPC, and a thrombin cleavage sequence was introduced into the multiple cloning site for potential use. These latter two sequences resulted in a His 6 -tagged fusion protein containing 17 extraneous amino acids on its N terminus (Fig. 3). However, it should be noted that previous studies have documented that an N-terminal tag domain of 159 extraneous residues has no adverse effects on either the catalytic and allosteric properties or the tetramerization of the maize recombinant C4 isoform (20). Likewise, these observations have been confirmed with the present, much shorter His 6 -tagged WT fusion protein relative to our earlier, untagged sorghum recombinant C4-PEPC (see "Results" and Refs. 19 and 21).
Culturing of Transformed E. coli XH11 Cells-For analysis of in vivo functional complementation of the original Ppc Ϫ strain, each transformed XH11 cell line described above was grown at 30°C on M9 minimal agar medium (16) supplemented with 50 g ml Ϫ1 ampicillin (Amp), 20 g ml Ϫ1 Arg, and 0.5% (w/v) succinate. A single colony of each transformant was selected and streaked onto the above-prepared M9 plates, Ϯ succinate, and incubated at 30°C for up to 2 days.
For cell extraction and subsequent affinity purification of recombinant PEPC, the XH11 cells harboring a specific Ppc plasmid DNA were maintained on LB agar medium (16) containing 100 g ml Ϫ1 Amp. A single colony of each transformant was selected, inoculated into 5 ml of Terrific Broth medium (16) supplemented with 50 g ml Ϫ1 Amp, and incubated overnight with shaking at 37°C. A 2.5-ml aliquot of this starter culture was added to 1 liter of Terrific Broth medium supplemented with 50 g ml Ϫ1 Amp and incubated at 37°C for ϳ8 h until the A 600 reached 1.2. The cells were collected by centrifugation, resuspended, and washed once in buffer containing 0.1 M Mops-KOH, pH 7.3, 10 mM MgCl 2 , 5 mM L-malate (pH adjusted), and the resulting pellets (ϳ6 g) were frozen in liquid N 2 and stored at Ϫ80°C until used.
Extraction and Affinity Purification of Sorghum Recombinant C4-PEPCs-The frozen E. coli cells were resuspended in 25 ml of icecold extraction buffer containing 0.1 M Mops-KOH, pH 7.3, 10 mM MgCl 2 , 5 mM L-malate (pH adjusted), 2 mM phenylmethylsulfonyl fluoride, 10 g ml Ϫ1 chymostatin, and one tablet of Complete EDTA-free protease inhibitor mixture (Roche Applied Science) per 50 ml of buffer. The resulting suspension was passed twice through a chilled French pressure cell at ϳ20,000 p.s.i. The crude lysate was centrifuged at 47,800 ϫ g and 4°C for 30 min, and the supernatant fluid was fractionated at 4°C with 40 -65% saturation ultrapure ammonium sulfate (16). The resulting protein precipitate was dissolved in buffer A (20 mM Tris-HCl, pH 8.0, 5 mM L-malate (pH adjusted), 10% (v/v) glycerol, 10% (v/v) ethylene glycol, 0.5 M ammonium sulfate) containing 2 mM phenylmethylsulfonyl fluoride, 10 g ml Ϫ1 chymostatin, and 1 tablet of the above protease inhibitor mixture per 50 ml of buffer and centrifuged at 16,000 ϫ g for 20 min. The clarified solution was applied to a column at 4°C containing 25 ml of phenyl-Sepharose CL-4B (Sigma) pre-equilibrated with buffer A alone. The column was washed to base line at A 280 with buffer A, and the bound proteins were eluted with a decreasing, 0.5-0 M ammonium sulfate linear gradient in buffer A (16,19). The enriched, PEPC-containing fractions that eluted at ϳ0 M ammonium sulfate were pooled, concentrated, and loaded onto a 2-ml column of His⅐Bind resin (Novagen, Madison, WI) pre-equilibrated with 20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 5 mM imidazole, 10% (v/v) glycerol, and the above protease inhibitors. The affinity purification protocol was performed at 4°C as per the manufacturer's instructions. The fractions eluted by 0.2 M imidazole, pH 7.9, that showed highest PEPC activity were collected, pooled, and concentrated with an iCON TM 20K concentrator (Pierce) to a protein concentration of ϳ3 mg ml Ϫ1 . The resulting PEPC preparation was stored in a buffered medium (0.1 M Tris-HCl, pH 7.6, 20% (v/v) glycerol, 10% (v/v) ethylene glycol, 2 mM dithiothreitol, one tablet of protease inhibitor mixture per 50 ml of buffer, 5 g ml Ϫ1 chymostatin) at Ϫ80°C until used. The purified proteins were stable for at least 15 days under these conditions. Measurement of PEPC Activity, Steady-state Kinetic Parameters, and Thermostability-Routine, ⌬A 340 -based spectrophotometric assay of PEPC activity during purification was performed at 25°C by coupling through exogenous NADH-malate dehydrogenase (4 units) in 50 mM Hepes-KOH, pH 8.0, 1 mM NaHCO 3 , 10 mM MgCl 2 , 0.2 mM NADH, Ϯ 5 mM mono(cyclohexylammonium) PEP (21). For more detailed analysis of the purified recombinant proteins, a sensitive, fixed-time radiometric assay was modified from Budde and Chollet (22). The standard reaction mixture contained, in a final volume of 1.0 ml, 50 mM Hepes-KOH, pH 7.3/8.0, 10 mM MgCl 2 , 2 mM dithiothreitol, 5 units of NADHmalate dehydrogenase, 1 mM NADH, 5 mM NaH 14 CO 3 (0.2-2.0 Ci mol Ϫ1 ), 0, 2, or 5 mM PEP (pH adjusted), and 0.5 g of PEPC. The reaction at 30°C was initiated by the addition of PEPC and terminated after 60 s by injection of 0.2 ml of 6 M acetic acid. The acidified assay mixtures were taken to dryness at 90°C, and PEP-dependent 14 C-dpm fixed into acid-stable product was determined by liquid scintillation spectroscopy. One unit of PEPC activity was defined as 1 mol of oxaloacetate formed (and reduced to L-malate by enzymatic coupling) per min at 30°C. For steady-state kinetic analysis at pH 7.3 and 30°C, the concentration range of PEP varied from zero PEP to 10 values between 0.4 and 15.5 mM. All kinetic parameters were determined from the Hill equation fitted to the PEP-dependent experimental data by nonlinear regression analysis (21, 23) using SigmaPlot ® 8.0 and Enzyme-Kinetics Module 1.1 (Systat Software Inc., Point Richmond, CA), and k cat values were calculated using a subunit mass of 110,360. Protein concentrations were determined with a sensitive, protein assay kit (Bio-Rad), based on dye binding (24), with crystalline bovine serum albumin as standard.
For comparative analysis of the relative thermostability of the various purified recombinant C4-PEPCs, samples (0.3 mg ml Ϫ1 ) were preincubated in 50 mM Hepps-KOH, pH 8.0, at 35 to 43°C for up to 30 min (25). Thermo-inactivation was terminated by cooling aliquots on ice for 5 min, and then residual activity was measured radiometrically at 30°C, pH 8.0, and 5 mM PEP. Given the extremely low activity of the ⌬C4truncated enzyme at 30°C (see Ref. 16 and "Results"), its heat inactivation was not assessed.
Gel Electrophoresis and Immunoblot Analysis-Non-denaturing PAGE at 4°C and SDS-PAGE were performed on 7 and 10% separating gels according to Dong et al. (16), Budde and Chollet (22), and Laemmli (26), respectively. The gels were either stained with Coomassie Brilliant Blue or electrotransferred onto polyvinylidene difluoride membranes using a Bio-Rad Mini Trans-Blot apparatus. The membrane was incubated with a 1:5000 (v/v) dilution of rabbit polyclonal antibody raised against maize leaf PEPC, and the antigen-antibody complexes were visualized using ECL Western blotting detection reagents and the accompanying protocol from Amersham Biosciences.
Structural Modeling, Mining, and Model Evaluation-The starting point for homology modeling of the sorghum C4 isoform was the active R-state, 3.0-Å resolution structure of the closely related, maize C4-PEPC (6,12,13). These two C4 homologues share 90 and 94% identity and similarity in deduced amino acid sequences, respectively. Homology modeling was performed on the 3D-JIGSAW (27) and SWISS-MODEL (28) web servers. The resulting PDB coordinates were handled using the Change-pdb-chain-id function in the 3D-Dock package (29) for changing chain ID, pdbset in the CCP4 Suite (30) for renumbering atoms, and BBEdit (Bare Bones Software, Bedford, MA) for general text editing. The resulting models were evaluated by "Procheck" in the CCP4 Suite. Structural analysis, including local energy minimization, superimpositions, and distances between atoms, was done using Swiss-Pdb Viewer (31). Graphical representations were prepared using PyMOL (DeLano Scientific, San Carlos, CA). Molecular cavities were searched and evaluated using the volume option of the Putative Active Sites with Spheres software (32). Principal component analysis (33) was performed using the statistical computing environment R and the Rweb server of Jean Thioulouse, Université Claude Bernard-Lyon 1, France.
Homology modeling of the sorghum C4 isoform was accomplished successfully based on the maize C4-PEPC structure (PDB code 1jqo (12,13)). The sorghum three-dimensional model began with residue Ile-26 (maize Ile-35) at the start of ␣-helix 1 and ended at C-terminal residue Gly-961 (maize Gly-970). This removed the more variable N-terminal domain that contains the phosphorylatable Ser-8 residue (maize Ser-15). The root mean square of distance between the modeled structures and templates were all 0.6 Å in the backbones. This deviation was due to lack of structure in the template, corresponding to sorghum residues Lys-116 -Ser-132 and Pro-751-Ile-760.

Preparation and in Vivo Analysis of the C-terminal Mutant
Enzymes-Based on earlier insight into the importance of the PEPC C-terminal tetrapeptide provided by complementary functional and structural analyses (see the Introduction and Refs. 2, 6, 12, and 16), the present study was focused specifically on its terminal Gly, this element's only absolutely conserved residue (Fig. 1). The development and use of a novel expression vector for the sorghum C4 isoform (Fig. 3) together with our original XH11/Ppc Ϫ strain of E. coli (16,19) facilitated the production of various recombinant, His 6tagged C4 variants in a PEPC Ϫ eubacterial background. Besides the sorghum WT and ⌬C4 "reference" enzymes (16), the minimal, H-atom side chain of Gly-961 was replaced conservatively by the two progressively larger, aliphatic hydrocarbon side chains of Ala (G961A) and Val (G961V) containing one or two CH 3 groups, respectively. In addition, the "target" Gly residue was deleted specifically (⌬C1), thereby shortening the 961-residue polypeptide chain by a single amino acid and making the new C-terminal side chain longer and more hydrophilic (Thr-960), or rendered more negatively charged by substitution with the two carboxyl groups of a C-terminal Asp (G961D).
These four new C-terminal mutant forms of sorghum C4-PEPC were first evaluated with respect to their ability to functionally complement in vivo the growth phenotype of Ppc Ϫ strain XH11 on minimal agar medium. Relative to the WT transformant, all the C-terminal mutants examined in this study were compromised to varying extents with respect to their growth under selective conditions in the absence of the C4 acid succinate. Specifically, the conservative G961A substitution was least affected as compared with WT followed by the G961V and ⌬C1 variants (Fig. 4). In contrast, neither the ⌬C4 nor G961D mutant C4-PEPC supported detectable growth of strain XH11 in the absence of succinate, whereas all six transformants grew equally well in the presence of exogenous C4 acid (Fig. 4, data not shown, and Ref. 16). These results indicated that the C4-PEPC variants expressed by these mutants are either essentially inactive (⌬C4, G961D) or not sufficiently active in vivo (⌬C1, G961(A/V)) to fully complement the growth phenotype of the XH11/Ppc Ϫ strain and/or are insufficiently stable to be accumulated to WT levels in the transformed cells.
To assess to what extent the C4-PEPC polypeptide actually accumulated in these XH11 cells, soluble crude extracts were prepared from liquid cultures grown in rich Terrific Broth medium, resolved by SDS-PAGE, and analyzed by immunoblotting with a polyclonal antibody against the maize leaf enzyme. Whereas four of the five mutant transformants contained essentially WT levels of soluble PEPC protein (Fig.  5), the non-conservative G961D variant failed to accumulate detectable, steady-state levels in both the soluble fraction and a total cellular lysate (data not shown). Thus, this non-His 6 -tagged, negative charge mutant was not investigated further in this study.
Purification and in Vitro Analysis of the C-Terminal Mutant PEPCs-Given that the four remaining C4-PEPC variants were produced by XH11 cells in WT levels (Fig. 5), their varying degrees of functional incompetence in vivo (Fig. 4) were most likely caused by perturbations in native aggregation state and/or catalytic activity. To explore these possibilities, the soluble His 6 -tagged WT and mutant fusion proteins (Fig. 3) were purified to electrophoretic homogeneity by sequential ammonium sulfate fractionation, hydrophobic interaction chromatography, and immobilized metal (Ni 2ϩ ) affinity chromatography (Fig. 6A). Analysis of their apparent aggregation state was performed by nondenaturing PAGE at 4°C and subsequent immunoblotting with the maize C4-PEPC antibodies. Like the WT enzyme, the four purified C-terminal variants migrated to a position in the separating gel consistent with a homotetrameric PEPC, with no immuno-evidence for dimeric and/or monomeric forms (Fig. 6B). These findings with the affinity-purified fusion proteins are consistent with those generated previously by size-exclusion fast protein liquid chromatography and native-PAGE analysis of the partially purified, untagged sorghum WT and ⌬C4 recombinant enzymes (16). Clearly, neither modification of the C4-PEPC invariant, C-terminal Gly residue nor its flexible "parent" Each transformant was grown at 30°C as a stock culture on an M9 minimalmedium agar plate supplemented with 50 g ml Ϫ1 Amp, 20 g ml Ϫ1 Arg, and 0.5% (w/v) succinate. A single colony of each was subsequently selected and streaked onto a fresh M9 agar plate containing the original levels of Amp and Arg plus (non-selective) or minus (selective) 0.5% (w/v) succinate and incubated for 2 days at 30°C. tetrapeptide causes detectable changes in polypeptide stability, steadystate accumulation, and tetramerization in vivo, except for the nonconservative, negative charge G961D substitution.
Based on these observations, the PEPC activity of the four purified variants was next evaluated relative to the control, WT sorghum C4 isoform. For these initial in vitro analyses a sensitive 14 C-based assay was employed, and PEP-dependent activity was determined at suboptimal but near physiological levels of pH (7.3) and PEP (2 mM) (21). Consistent with the above findings from the in vivo functional complementation study (Fig. 4), the four C-terminal mutant enzymes also displayed varying degrees of PEPC activity in vitro ranging from ϳ23% of WT with the modest G961A substitution to only ϳ0.2% for the ⌬C4-truncated form (Fig. 6C).
Kinetic Properties of the C-terminal Mutant Enzymes-Given that the above collective findings indicated that all four C-terminal variants were compromised significantly with respect to both in vivo and in vitro PEPC activity, steady-state kinetic analysis of the purified enzymes became imperative. Because various recombinant C4-PEPC homotetramers have been reported to display varying degrees of cooperative behavior depending on the assay pH and/or specific amino acid modification(s) (13,21,23), the 14 C-based experimental data were fitted to the Hill equation for calculation of the kinetic parameters K 0.5 (PEP), V max , and the Hill coefficient (21,23).
Initial determination of the kinetic properties of the purified WT, His 6 -tagged fusion protein at pH 7.3 revealed that its respective K 0.5 and V max values of 1.6 mM and ϳ36 units mg Ϫ1 (Table 1) were similar or superior to those reported previously for the untagged or "trimmed" sorghum and maize recombinant C4 isoforms, respectively (13,21). This documents that neither this fusion protein's in vivo activity (Fig. 4) nor in vitro kinetic constants (Table 1) were affected adversely by its modest 17 amino acid N-terminal extension. These observations agree with earlier related findings with the maize recombinant C4 isoform tagged with 159 extraneous residues on its N terminus (20).
The kinetic properties of the four compromised C-terminal variants were determined next under identical radiometric assay conditions at pH 7.3. Whereas the V max and related k cat values of the three Gly-961specific modifications ranged from ϳ25% (G961A) to ϳ10% (G961V) of WT, those of the ⌬C4-truncated form were only ϳ0.3% (Table 1). These relative kinetic parameters of the purified enzymes essentially mirrored the trends observed in the in vivo functional complementation study (Fig. 4). In contrast, the K 0.5 (PEP) values did not vary markedly from that of the WT enzyme, with the exception of the modest, Ͻ2-fold increase in the ⌬C4-truncated form K 0.5 (Table 1). Consistent with the above kinetic findings, the corresponding decreases in apparent catalytic efficiency or specificity constant (k cat /K 0.5 ) of the four C-terminal variants largely reflected those in k cat alone (Table 1) and likewise in in vivo functionality (Fig. 4). Finally, although there were significant, but Յ2-fold decreases in the Hill coefficient (h) of the four C-terminal mutants, there was no consistent trend in these changes. For example, the h values of the ⌬C1/C4 and G961V variants were all ϳ50% of the WT coefficient (2.4), whereas the respective changes in apparent catalytic efficiency and in vivo functional complementation varied markedly between the ⌬C4-truncated form and these two other mutants (Table 1, Fig. 4).
Thermal Stability of the Purified Gly-961 Mutant Enzymes-One significant experimental variable between the above in vivo and in vitro activity analyses (Figs. 4 and 6C, Table 1) and the related aggregationstate studies (Fig. 6B) was temperature, namely 30 versus 4°C, respectively. Given this difference and the adverse effects on activity measured at 30°C, it was important to further confirm that the various C-terminal modifications did not grossly perturb the apparent structural stability of the active C4-PEPC homotetramer. To this end, in vitro thermal stability experiments were performed with the purified WT and G961 variants (cf. Refs. 25 and 34). Inactivation of the ⌬C4-truncated form by elevated temperature was not assessed due to its already marginal activity at 30°C (Fig. 6C, Table 1, and Ref. 16). Neither of the four recombinant enzymes was inactivated by preincubation for 30 min at 35°C (data not shown). When the preincubation temperature was increased to 40 or 43°C, they were all progressively inactivated as a function of time, but there were no consistent, major differences between the mutants' ther-  mal inactivation profiles and that of WT C4-PEPC. For example, after a 10-min preincubation at 40°C, the residual activities measured at 30°C ranged from 71 (G961V) to 44% (⌬C1), whereas after 2 min at 43°C they varied from 62 (G961V) to 39% (WT) (data not shown). Likewise, all four recombinant C4-PEPCs were inactivated Ն99% after a 10-min preincubation at 43°C. Thus, deleting the invariant, C-terminal Gly residue (⌬C1) or increasing the size of its side chain (G961(A/V)) has no gross effect on apparent homotetramer stability in vitro as compared with WT PEPC (Fig. 6B, data not shown).

DISCUSSION
This laboratory has had a longstanding interest in the regulatory phosphorylation and structure-function relationships of eukaryotic PEP carboxylase (for review, see Refs. 1, 2, and 8) (4). This cytoplasmic, allosteric enzyme plays a diverse array of important photosynthetic and/or metabolic roles in the varied physiological contexts of green leaves, developing seeds, legume root nodules, and unicellular green algae, to name a few. With respect to its structure-function properties, PEPC has been analyzed in detail since the early 1990s by mutagenesis, x-ray crystallography, and reciprocal C4/C3 enzyme chimeras (for review, see Refs. 1, 2, 6, and 7) (13). These studies have revealed key insights into the homotetramer's "dimer-of-dimers" quaternary structure and the monomer's largely ␣-helical secondary structure, catalytic domain, and distinct phosphorylation and allosteric effector sites. Although the N-terminal region of the vascular plant polypeptide uniquely harbors the single target Ser residue subject to reversible phosphorylation in vivo (1,2,8,19), the more widely conserved C-terminal domain ( Fig. 1) houses elements important for protein stability, negative allosteric regulation, and maximal overall catalysis (2,6,12,16,35). Specifically, the C-terminal tetrapeptide of plant-and bacterial-type PEPC contains two conserved residues implicated by x-ray crystallography in allosteric inhibitor binding (Asn) and active site ionic interactions with an essential, invariant Arg side chain (C-terminal Gly) (see the Introduction, Figs. 1 and 2A, and Refs. 2, 6, 12, and 35). Prompted by these recent structural observations together with our earlier findings on the importance of the C-terminal, QNTG tetrapeptide for overall catalysis by the sorghum C4 isoform (16), we set out to investigate this strictly conserved Gly residue at the extreme C terminus of sorghum C4-PEPC in the present study.
Perturbation of Gly-961 by either conservative, neutral substitution with Ala (G961A) or Val (G961V) or specific deletion (⌬C1) resulted in neither adverse effects on polypeptide accumulation in a transformed, Ppc Ϫ /XH11 strain of E. coli nor apparent destabilization of the purified homotetramer at below or above ambient temperatures (Figs. 5 and 6B and data not shown). Although similar observations were recorded with the more drastic truncation of the entire parent QNTG tetrapeptide (⌬C4), modification of this invariant Gly residue by a negative charge substitution (G961D) resulted in the complete absence of immunologically detectable C4-PEPC in the transformed bacterial cells (data not shown). Evidently, PEPC appears to not tolerate additional negative charge at its extreme C terminus beyond that of the main chain free CO 2 Ϫ group.
Despite the aforementioned lack of effect of the G961(A/V) and ⌬C1/C4 modifications, these four individual changes led to significant negative perturbation of PEPC activity both in vivo and in vitro. Functional complementation of XH11 cells under selective, minus succinate growth conditions was restricted progressively by the Ala, ⌬C1 and Val, and ⌬C4 modifications (Fig. 4). Likewise, the effects on in vitro activity of the corresponding purified enzymes, assessed at physiological levels of PEP and pH, showed a similarly progressive decline, ranging from ϳ23% to only ϳ0.2% of WT C4-PEPC in the G961A and ⌬C4 mutants, respectively (Fig. 6C). Evidently, even the simplest neutral alteration of the PEPC invariant, C-terminal Gly hydrogen atom side chain to a single CH 3 group (i.e. G961A) is detrimental to overall catalysis by the enzyme in vivo and in vitro. Steady-state kinetic analysis of the purified WT and C-terminal mutant enzyme forms revealed a progressive, significant decrease in the mutant V max , k cat , and specificity (k cat /K 0.5 ) constants relative to the WT C4 isoform, with little or no major change in the enzyme K 0.5 value (Table 1). Thus, specific perturbation of C4-PEPC's strictly conserved C-terminal Gly residue or its parent QNTG tetrapeptide by conservative neutral substitution (G961(A/V)) or truncation (⌬C1/C4) causes up to ϳ10 -1000-fold decreases in overall catalytic efficiency due primarily to changes in k cat alone. These new functional findings provide novel insight into the presumed importance of this enzyme's invariant Gly residue at its extreme C terminus recently derived from the three-dimensional structure of the active R-state form of maize C4-PEPC (see the Introduction and Refs. 2, 6, and 12).
Given the availability of this 3.0-Å resolution structure of the maize C4 isoform, we attempted to evaluate our functional observations in relation to the modeled structure of sorghum C4-PEPC. Notably, the deduced primary structures of these two C4 isoforms are very closely related, bearing a 90 and 94% identity or similarity, respectively, throughout the entire, 961-or 970-residue polypeptide chain and 100% identity within the final 54 C-terminal residues, including all of terminal ␣-helix 41 (sorghum Leu-942-Met-957) and its appended flexible QNTG tetrapeptide ( Fig. 1 and data not shown). Homology modeling of the WT sorghum C4-PEPC was in excellent agreement with the maize structure (12,13) in regard to main-chain folding and, specifically, the active-site domain marked by the C-terminal side of the 8-stranded ␤-barrel and the invariant catalytically essential His and Arg residues (14,15) at positions 167 (177) and 638 (647), respectively (sorghum (maize) numbering) (see Fig. 2A). Likewise, the strictly conserved C-terminal Gly-961 (-970) carboxyl group was shown to be ion-paired with the guanidino side chain of Arg-638 (-647) situated within 3.1 Å. This active-site Arg, which is embedded in the invariant Gly-rich motif of mobile loop I, is presumed to partially dissipate the negative charge of the PEP phosphate moiety upon substrate binding to the active R-state enzyme (2,6,12). As in WT sorghum C4-PEPC, the distance between this essential Arg side chain and the main-chain CO 2 Ϫ group at the extreme C terminus of the G961(A/V) variants is maintained at 3.1 Å. In contrast, this ion pair is compromised in the modeled truncated enzyme forms, with the C-terminal carboxyl of ⌬C1 Thr-960 and ⌬C4 Met-957 situated 6.4 and 11.5 Å away, respectively, from the guanidino side chain of Arg-638 (Fig. 2B). Finally, we sought to gain additional structural insight into the progressively compromised k cat and specificity constants of the four C-terminal mutant enzyme forms ( Table 1) by comparison of putative molecular cavities in modeled structures of the sorghum C4 isoform using the volume option in the Putative Active Sites with Spheres software program (32). From this analysis we found suggestive evidence for alterations in three predicted cavities within the catalytic domain, (i) "behind" essential residue His-167, (ii) by the "gate" to the active site, and (iii) in the "hole" of the active site that houses the essential Mg 2ϩ (or Mn 2ϩ ) cofactor (Fig. 2B, Table 2). For the cavity behind His-167 (situated between Ala-172 and Thr-226), all four variants showed a decreased pocket volume. However, only the G961A mutant also displayed an altered shape of this cavity. This concomitant perturbation in cavity shape and volume behind His-167 resulted in a change in the conformation of the residues near this pocket. With respect to the active-site hole cavity between Arg-638 and the extreme C terminus, the G961V mutant enzyme possessed an enlarged pocket, whereas the ⌬C4-truncated variant displayed a large positional shift of this cavity toward C-terminal Thr-960 relative to the WT enzyme. The G961A and WT enzyme forms shared a similar shape and size of the gate pocket between Arg-447 and Arg-638, whereas the three other molecular forms had an enlarged gate cavity. To clarify these modeled effects of the mutations, principal component analysis (see "Experimental Procedures") was performed. These three cavity volumes and locations were projected on a principal component axis that explained 93% of the entire variance. The trends in the projected values (Table 2) correlated reasonably well with the experimentally determined perturbations in k cat and catalytic efficiency (k cat /K 0.5 ) relative to the WT sorghum enzyme except for the severely compromised ⌬C4 variant, which has a large spatial translation of the hole cavity's center due to deletion of the flexible C-terminal QNTG tetrapeptide. In the case of the simplest and most conservative G961A substitution, modest (ϳ4-fold) perturbation of its specificity constant was also enhanced by a concomitant change in the microenvironment around essential His-167. Thus, formation of the requisite reaction intermediates during the 3-step PEPC catalytic sequence (see the Introduction and Refs. 1, 2, and 6) possibly requires more time in these four C-terminal mutant enzymes due to the increased volume of the predicted gate and/or hole cavities, thereby effecting adverse changes in k cat and catalytic efficiency.
In summary, this study provides new functional and structural insight into the invariant nature of the PEPC C-terminal Gly residue and its importance for maximal overall catalysis by the enzyme. As a next step we are presently investigating the conserved proximal Asn at position 959 in sorghum C4-PEPC (see Fig. 1) given its presumed role in negative allosteric regulation (2,6,12,35) and the marked perturbation in spec-ificity constant of the "feeble" C-terminal ⌬C4 variant relative to the three Gly-961-specific mutants described herein.