|Home | About | Journals | Submit | Contact Us | Français|
Pyruvate carboxylase (PC) is a conserved metabolic enzyme with important cellular functions. We report here crystallographic and cryoEM studies of S. aureus PC (SaPC) in complex with acetyl-CoA, an allosteric activator, as well as mutagenesis, biochemical and structural studies of the biotin binding site of its carboxyltransferase (CT) domain. The disease-causing A610T mutation abolishes catalytic activity by blocking biotin binding to the CT active site, and Thr908 may play a catalytic role in the CT reaction. The crystal structure of SaPC in complex with CoA reveals a symmetrical tetramer, with one CoA molecule bound to each monomer, and cryoEM studies confirm the symmetrical nature of the tetramer. These observations are in sharp contrast to the highly asymmetrical tetramer of R. etli PC in complex with ethyl-CoA. Our structural information suggests that acetyl-CoA promotes a conformation for the dimer of the biotin carboxylase domain of PC that may be catalytically more competent.
Pyruvate carboxylase (PC, EC 188.8.131.52) is a biotin-dependent enzyme and catalyzes the carboxylation of pyruvate to produce oxaloacetate (Attwood, 1995; Wallace et al., 1998; Jitrapakdee and Wallace, 1999; Jitrapakdee et al., 2006; Jitrapakdee et al., 2008). PC is highly conserved and is found in most living organisms. In mammals, PC has crucial roles in gluconeogenesis, lipogenesis, glyceroneogenesis, insulin secretion and other metabolic processes (Jitrapakdee and Wallace, 1999; Jitrapakdee et al., 2006; Jitrapakdee et al., 2008). Inherited PC deficiencies are linked to serious diseases in humans such as lactic acidemia, hypoglycemia, psychomotor retardation and death (Jitrapakdee and Wallace, 1999; Robinson, 2006). Four single-site mutations in PC, V145A, R451C, A610T and M743I, have been associated with these diseases (Carbone et al., 1998; Wexler et al., 1998; Carbone and Robinson, 2003; Robinson, 2006).
In eukaryotes and most bacteria, PC is a single-chain enzyme of approximately 130 kD in molecular weight, and is active only in its tetrameric form (Attwood et al., 1993; Attwood, 1995; Jitrapakdee and Wallace, 1999; Jitrapakdee et al., 2008). In addition, the activity of most PC enzymes is stimulated allosterically by acetyl-CoA and inhibited by aspartate (Jitrapakdee and Wallace, 1999; Sueda et al., 2004; Jitrapakdee et al., 2007). Three domains were identified based on sequence comparisons with other biotin-dependent carboxylases (Cronan Jr. and Waldrop, 2002; Hall et al., 2004; Kondo et al., 2004; Tong, 2005; Studer et al., 2007). The biotin carboxylase (BC) domain at the N-terminus catalyzes the first step of the reaction, the ATP- and Mg2+-dependent carboxylation of biotin, with bicarbonate as the carboxyl donor. The carboxyltransferase (CT) domain follows the BC domain in the primary sequence and catalyzes the transfer of the activated carboxyl group from carboxybiotin to pyruvate to form the oxaloacetate product (Supplementary Fig. 1). The active site of CT contains a tightly-bound Mn2+ or Zn2+ divalent cation. The biotin cofactor is covalently linked to the side chain of a lysine residue in the biotin-carboxyl carrier protein (BCCP) domain, located at the C-terminus of the enzyme.
Crystal structures of human, S. aureus and R. etli PC (HsPC, SaPC and RePC) have been reported recently by us and others (St. Maurice et al., 2007; Xiang and Tong, 2008). They reveal the organization of the tetramer (Fig. 1a) and indicate that the BCCP domain must migrate between the BC active site of its own monomer and the CT active site of another monomer during catalysis, explaining why PC is only active in the tetrameric form. Interactions among the BC, CT, as well as a novel domain, the PC tetramerization (PT) domain, are important for the formation of this tetramer (Fig. 1a). A generally symmetrical tetramer was observed for both HsPC and SaPC, although there are also significant differences in the relative positions of the BC and CT domains in the four monomers of SaPC (the HsPC structure is missing the BC domain) (Xiang and Tong, 2008). In contrast, a highly asymmetrical tetramer was observed for RePC, in complex with ethyl-CoA, an analog of the acetyl-CoA activator (St. Maurice et al., 2007). It was not clear whether this dramatic difference in the tetramer organization is truly due to the presence of ethyl-CoA, as the structures of HsPC and SaPC do not contain an activator. In addition, only the 3′-phospho-ADP portion of ethyl-CoA was ordered in the RePC structure, while the conformation of the rest of the CoA molecule was not observed.
In the structures of both HsPC and SaPC, a biotin coenzyme was captured in the active site of the CT domain (Fig. 1b), providing the first molecular insight into how biotin could participate in the carboxyltransfer reaction (Xiang and Tong, 2008). The structures identify a collection of residues in the CT active site that may be important for catalysis, and most of them are highly conserved among the PC enzymes (Fig. 1c). Two of the disease-causing mutations, A610T and M743I, are located in this active site (Hall et al., 2004; Xiang and Tong, 2008), although the functional importance of the other residues in the biotin binding site has not been characterized. The structure of SaPC also reveals an exo site for biotin binding (Xiang and Tong, 2008), which may exist in RePC as well (Jitrapakdee et al., 2008), but the functional role of this site is currently not known.
We report here structural and biochemical characterizations of SaPC carrying single-site mutations in the active site of the CT domain, as well as the structure of SaPC in complex with CoA. The studies demonstrate that the disease-causing A610T mutation abolishes catalytic activity by blocking biotin binding to the CT active site, and indicate that residue Thr908 may play a catalytic role in the CT reaction. Having structural information on SaPC in the absence and presence of CoA allowed us to carry out detailed structural comparisons to define conformational changes in the enzyme upon CoA binding. Surprisingly, our crystal structure showed that the CoA complex of SaPC is a symmetrical tetramer, which we have confirmed by cryoEM studies. This is in sharp contrast to the highly asymmetrical tetramer of R. etli PC in complex with ethyl-CoA (St. Maurice et al., 2007).
Based on a careful examination of the structure of SaPC, we identified six residues in the CT active site that may have important roles in biotin binding and/or catalysis: Ala610, Tyr651, Gln870, Thr908, Ser911, and Lys912 (Fig. 1b). These residues are highly conserved among the PC enzymes (Fig. 1c). (To simplify discussions, the residues in SaPC and RePC are numbered according to their equivalents in HsPC.) Among these, Thr908 is hydrogen-bonded to the N1′ atom and Ala610 is located near the sulfur atom of biotin (Fig. 1b). Ser911 is hydrogen-bonded to the ureido oxygen of biotin. Gln870 is located near this oxygen atom, although it does not have direct interactions with the biotin. The side chains of Tyr651 and Lys912 have van der Waals interactions with biotin, contributing to the formation of its binding site. We have therefore designed the A610T, Y651A, Q870A, T908A, S911A, and K912T mutants. A610T corresponds to the disease-causing mutation, and the K912T mutation was selected as our modeling study suggests that the bulkier Thr side chain could disrupt the biotin binding site.
The other part of the CT active site, involved in the binding of the pyruvate substrate and the divalent cation, has previously been examined by mutagenesis studies in RePC (St. Maurice et al., 2007), the PC enzyme from Bacillus thermodenitrificans (Yong-Biao et al., 2004) and the CT (5S) subunit of the Propionibacterium shermanii transcarboxylase (Hall et al., 2004). The disease-causing M743I mutation is located here and is expected to block binding of the pyruvate substrate (Hall et al., 2004). We have selected one residue in this area of the active site, Arg644, for mutagenesis. This residue is involved in a bi-dentate interaction with the pyruvate substrate in the structures of HsPC and SaPC (Fig. 1b) (Xiang and Tong, 2008), but is pointed away from the substrate in the 5S structure (Hall et al., 2004). This residue is strictly conserved among the PC enzymes (Fig. 1c). To examine the functional role of this residue in catalysis, we have generated the R644K and R644A mutants.
The wild-type SaPC and the designed single-site mutants were over-expressed in E. coli and purified to homogeneity. The catalytic activity of the enzymes was assayed by monitoring the production of the oxaloacetate product at various pyruvate concentrations. The concentrations of the other substrates were kept at saturating levels in the assay, and the activator acetyl-CoA was not included in the reaction. The initial velocity data were fitted to the Michaelis-Menten equation (there was no indication of cooperative behavior) to obtain kinetic parameters for the wild-type enzyme and those mutants with sufficient catalytic activity (Table 1).
The kinetic data showed that two mutants, S911A and Q870A, maintained strong catalytic activity, with only a 1.5- and 2-fold loss in kcat/Km as compared to the wild-type enzyme, respectively (Table 1). It is likely that the S911A mutation did not completely disrupt the interactions between the ureido oxygen atom of biotin and the enzyme, as this atom is also hydrogen-bonded to the main-chain amide of Lys912 (Fig. 1b) (Xiang and Tong, 2008). The kinetic mechanism of PC catalysis suggests that the ureido oxygen may carry a negative charge during the CT reaction (Supplementary Fig. 1) (Attwood and Wallace, 2002; Jitrapakdee et al., 2008), and these interactions may be important for stabilizing this anionic intermediate.
The other mutants that we studied have greater than 30-fold loss in kcat (Table 1), confirming the structural observations and the functional importance of these residues in the catalysis by PC. The data on the Arg644 mutants suggest that the bidendate interaction between this side chain and the pyruvate substrate is important for catalysis, in contrast to the structural observations on the 5S subunit of transcarboxylase (Hall et al., 2004). Our kinetic results on the A610T mutant of SaPC are consistent with those reported earlier on this mutant of HsPC (Carbone et al., 1998; Wexler et al., 1998; Carbone and Robinson, 2003). To provide further information on the molecular basis for the effects of these mutations on the catalytic activity, we have determined the crystal structures of the A610T and T908A mutants.
The crystal structure of the A610T mutant of SaPC has been determined at 2.9 Å resolution (Table 2). The overall organization of the tetramer of this mutant is similar to that of the wild-type enzyme (Fig. 2a), with an rms distance of 0.67 Å among 4057 equivalent Cα atoms of the two tetramers. However, the structure revealed that none of the biotin groups are present in the active site of the CT (or BC) domains. Instead, all of them are located in the exo site (Xiang and Tong, 2008). Detailed inspection of the biotin binding site confirms the expectation that introduction of the bulkier Thr side chain in the mutant causes serious steric clashes with biotin (Fig. 2b), thereby blocking its binding and involvement in catalysis. The distance between residue Ala610 and the sulfur atom of biotin is 4.1 Å. In the mutant, the distance would be reduced to 2.3 Å, much shorter than the allowed contact distance of 3.5 Å.
Coupled with the relocation of BCCP-biotin from the CT active site to the exo site, there is a large conformational change for the C-terminal segment (residues 863-983) of the CT domain in the A610T mutant structure (Fig. 2c), so that it now resembles that of the free CT domain (Xiang and Tong, 2008). This provides further evidence that the C-terminal segment of CT must undergo a conformational change to accommodate BCCP-biotin in its active site (Xiang and Tong, 2008).
The crystal structure of the T908A mutant of SaPC has been determined at 2.7 Å resolution (Table 2). The overall structure of the mutant tetramer is essentially the same as the wild-type tetramer, with an rms distance of 0.55 Å for 4275 equivalent Cα atoms between them. Moreover, one biotin is located in the CT active site while the other three are in the exo site, just like that in the wild-type enzyme. There are essentially no conformational differences in the biotin binding site between the wild-type and mutant (Fig. 3d). The structural information therefore suggests that the hydrogen-bond between Thr908 and the N1′ atom of biotin may not be essential for biotin binding in the CT active site. The fact that the T908A mutant has a >30-fold loss in kcat (Fig. 2b) suggests that this residue may instead be important for catalysis by PC.
Our structural analysis suggests that Thr908 could serve as a general acid/general base during the CT reaction. In fact, a general base is needed to extract a proton from the methyl group of pyruvate and a general acid is needed to protonate the N1′ atom of biotin in the forward direction of the CT reaction (Supplementary Fig. 1) (Attwood and Wallace, 2002; Jitrapakdee et al., 2008). The side-chain hydroxyl group of Thr908 could serve both of these functions, as it is hydrogen-bonded to the N1′ atom of biotin and is about 4 Å away from the methyl group of pyruvate in the structure (Fig. 1b). The pKa value of this hydroxyl group may be in the same range as that for the N1′ atom of biotin and the methyl group of pyruvate (after enolization through binding to the divalent cation).
The structures of both the A610T and the T908A mutants reported here are in the free enzyme state, while that of the wild-type enzyme was in complex with the substrate pyruvate (Xiang and Tong, 2008). As biotin was found in the active site in the T908A mutant, in the same conformation as observed in the wild-type SaPC, it is unlikely that pyruvate binding is needed to facilitate binding of biotin to the CT active site.
Interestingly, we have so far not observed the binding of BCCP-biotin to the BC active site in the structures reported here and in the structure of wild-type SaPC reported earlier (Xiang and Tong, 2008). It may be possible that BCCP-biotin has higher affinity for the CT active site and the exo site under the experimental conditions that we have used. Further studies are needed to identify conditions that will favor the binding of BCCP-biotin to the active site of the BC domain.
To provide further biochemical data on the interactions between acetyl-CoA and PC, we characterized the activating effect of a series of acetyl-CoA analogs on the catalysis by SaPC. The concentration of the pyruvate substrate was kept near its Km (5 mM, Table 1) while the other substrates were present at saturating levels in the kinetic experiments. The data showed that acetyl-CoA is the most potent at activating SaPC, with a Ka of 2.0±0.3 μM, while the Ka for ethyl-CoA is 8.7±1.0 μM (Fig. 3a). In comparison, the Ka values of these two compounds for RePC are 30 and 360 μM (St. Maurice et al., 2007). Removal of the acetyl group led to an 80-fold loss in activation, as the Ka for CoA is 160±60 μM. In the structure of RePC, only the 3′-phospho-ADP portion of ethyl-CoA is ordered (St. Maurice et al., 2007). The closest commercially available analog of this compound is 3′-phospho-AMP (or adenosine 3′,5′-bisphosphate), which has a Ka of 1800±270 μM (Fig. 3a), suggesting that the β-mercaptoethylamine and the pantotheine groups of CoA may also be beneficial for PC activation. The most important group on acetyl-CoA for activating PC appears to be the 3′ phosphate, as 3′-dephospho-CoA had no effect on PC even at 0.5 mM concentration (Fig. 3a).
Further kinetic studies showed that acetyl-CoA gives rise to both an increase in the kcat of SaPC as well as a decrease in its Km for the pyruvate substrate, such that the overall kcat/Km is increased 18-fold in the presence of saturating (100 μM) acetyl-CoA (Table 1). The initial velocity data appear to obey Michaelis-Menten kinetics, with no sign of cooperativity (data not shown). Our kinetic data on acetyl-CoA activation of SaPC are consistent with those reported for other PC enzymes sensitive to this compound (Attwood and Wallace, 1986; Branson et al., 2002; St. Maurice et al., 2007; Jitrapakdee et al., 2008).
To reveal the conformational changes in SaPC upon acetyl-CoA activation, we determined the crystal structure of the enzyme in complex with CoA at 2.9 Å resolution (Table 2). The relatively higher R values for the diffraction data and the atomic model are due to the long c axis of the crystal (373 Å), which caused substantial overlaps among the diffraction spots. This was the best diffraction data set that we collected after screening through many crystals. The crystals were grown in the presence of acetyl-CoA, but only electron density for CoA was observed from the crystallographic analysis (Fig. 3b). The acetyl group was either hydrolyzed during crystallization or disordered in the crystal. In fact, it has been reported that some PC enzymes can hydrolyze acetyl-CoA (Frey II and Utter, 1977; Chapman-Smith et al., 1991). Nonetheless, the binding modes of the pantotheine and the β-mercaptoethylamine groups of CoA are clearly defined by the structure. All four BCCP domains are disordered in the structure, although weak electron density was observed for two biotin groups in the exo site, one in each layer of the structure (Fig. 3c).
In sharp contrast to the asymmetrical RePC structure, the structure of SaPC in complex with CoA is symmetrical, with a CoA molecule bound to each monomer of the tetramer (Figs. 3c, 3d). In fact, the structure in complex with CoA is even more symmetrical than that in the absence of CoA. The four monomers of SaPC in the absence of CoA have large differences in the relative positions of their BC and CT domains (Xiang and Tong, 2008). When the CT domains of the four monomers are superimposed, a 6-18° difference is seen in the relative orientations of the four BC domains. In contrast, for the structure in complex with CoA, the difference in the relative orientations of the four BC domains is only 6° (Fig. 3e). This shows that the domain organization of the four monomers has become more similar to each other and the tetramer has become more symmetrical in the CoA complex. The rms distance between equivalent Cα atoms for any pairs of the domains in the CoA complex is 0.6 Å.
The asymmetrical RePC structure possesses a single two-fold axis of symmetry, relating the two monomers in the same layer of the tetramer. In comparison, the structure of the CoA complex of SaPC possesses nearly 222 symmetry, which also relates monomers in the different layers of the tetramer. The presence of only two copies of biotin, in the exo site on different layers, represents a deviation from this 222 symmetry. The significance of this deviation remains to be determined, as all four BCCP domains are disordered in the current crystal. In addition, only very weak, discontinuous density is observed for the four B domains of BC, and they are not included in the current atomic model (Fig. 3c). As a result, there are no adenine nucleotides in the BC active sites in this structure.
To obtain direct, independent evidence for the symmetrical tetramer for the CoA complex observed in our crystal structure, we next carried out cryoEM studies on SaPC. Consistent with earlier observations on chicken liver PC (Attwood et al., 1993), the SaPC tetramer is not very stable at low concentrations in the absence of acetyl-CoA, which hindered the three-dimensional averaging for this state of the enzyme. In contrast, the PC tetramer is much more stable in the presence of acetyl-CoA, allowing the production of a cryoEM model at roughly 13 Å resolution, which clearly resembles the overall features of the PC tetramer observed in the crystal structure (Figs. 4a-c).
Although only two-fold symmetry was enforced during the three-dimensional averaging, the cryoEM map is remarkably symmetrical, consistent with overall 222 symmetry for the tetramer (Figs. 4a-c). In fact, our crystal structure of the CoA complex of SaPC could be readily fit into the cryoEM map (Fig. 4d), with all the structural features located within the boundaries of the cryoEM envelope (Fig. 4e). In contrast, the asymmetrical tetramer of RePC could not be completely accommodated within the cryoEM model (Fig. 4f), and the discrepancy is especially obvious for the PT domains (Fig. 4g). We did not observe any evidence for an asymmetrical tetramer for SaPC from the cryoEM studies. Overall, the cryoEM data provide clear evidence that SaPC in solution forms a symmetrical tetramer in the presence of acetyl-CoA, confirming our observations in the crystal structure.
A symmetrical PC tetramer is consistent with biochemical data. Studies of chicken liver PC showed that four acetyl-CoA molecules can bind to each tetramer, with apparent positive cooperativity (Hill coefficient of 1.9) (Frey II and Utter, 1977). Studies of acetyl-CoA activation of yeast PC also demonstrated positive cooperativity, with a Hill coefficient of 2, but only in the presence of the L-aspartate inhibitor (Cazzulo and Stoppani, 1968; Jitrapakdee et al., 2007). SaPC behaves similarly to the yeast enzyme, with the cooperative behavior (Hill coefficient of 2) being manifested only in the presence of aspartate (Fig. 5a). While further studies are needed to define the exact binding site for aspartate, the kinetic data showed that the inhibitory effect of this compound could be overcome at high concentrations of acetyl-CoA (Fig. 5a). This suggests that the conformation of PC in complex with acetyl-CoA may not allow aspartate binding.
Our structure showed that the 3′-phospho-ADP portion of CoA is bound at the interface between the BC-PT domains of one monomer and the BC domain of another monomer (Fig. 5b). The adenine base is buried in a small pocket on the surface of the enzyme (Fig. 5b), flanked by the side chain of Lys1056 in the PT domain and the main chain of Tyr78′ in the BC domain (with the prime indicating the other monomer). The N6 atom of adenine is hydrogen-bonded to the main-chain carbonyl oxygen of Ala80′, and the N1 atom is located near the side chain of Ser83′, although this residue is not conserved. The 3′-phosphate group is surrounded by a cluster of four Arg side chains, Arg398, Arg451 and Arg453 from the BC domain, and Arg1085 from the PT domain, explaining its importance for binding (Fig. 3a). R451C is one of the disease-causing mutations, and our earlier kinetic studies have shown that this mutant is much less sensitive to acetyl-CoA activation (Xiang and Tong, 2008). The α- and β-phosphates of CoA interact with the side chain of Arg453 as well as the main chain amides of residues 495-497, at the hinge region between the BC and PT domains (Fig. 5b).
The rest of the CoA molecule follows the interface of the BC dimer (Fig. 5c), having mostly van der Waals interactions with the enzyme. The thiol group of CoA is located in a small depression in the surface of the dimer, surrounded by the side chains of Arg54′, Ala57′, Lys79′, Arg445 and Glu449 in the two BC domains (Fig. 5b). Residues Arg54, Lys79 and Arg445 are strictly conserved among the single-chain PC enzymes. Moreover, Arg54 is also conserved in the BC subunit of E. coli acetyl-CoA carboxylase (ACC), which shares a similar mode of dimerization. Mutation of this equivalent residue in E. coli BC, Arg19, to Glu can disrupt the dimer of that enzyme, and the monomeric form of this mutant has a 3-fold loss in catalytic activity (Shen et al., 2006). This suggests that one function of acetyl-CoA binding may be to stabilize the dimer of BC domains, consistent with our observation in the cryoEM studies that the SaPC tetramer is much less stable in the absence of acetyl-CoA and earlier studies with the chicken liver PC (Attwood et al., 1993). Half-of-the-sites reactivity has been proposed for the E. coli BC dimer (Janiyani et al., 2001; de Queiroz and Waldrop, 2007; Mochalkin et al., 2008), although the substrate complex of BC is fully symmetrical (Chou et al., 2009) and monomeric mutants of the enzyme are catalytically active (Shen et al., 2006).
Based on the asymmetrical tetramer of RePC, it was suggested that acetyl-CoA stimulates PC activity by reducing the distances between the BC and CT active sites (St. Maurice et al., 2007). However, in the symmetrical tetramer of the CoA complex of SaPC (Fig. 4d), the distances between the BC and CT active sites are essentially the same as those in the free enzyme, roughly 75 Å (Xiang and Tong, 2008). Therefore, this mechanism of activation is unlikely to apply to SaPC.
Detailed comparisons between the structures of SaPC in the absence and presence of CoA show that the organization of the BC dimers has large differences between the two tetramers (Fig. 6a), while the organization of the CT dimers is not affected by CoA binding (data not shown). With one monomer of the BC dimer in overlay, the other monomers of the dimer show a difference of 14° in their orientations. A re-organization of the BC dimer is consistent with the binding of CoA to its interface as well as earlier biochemical data showing that acetyl-CoA primarily affects the BC reaction (Jitrapakdee et al., 2008). Therefore, the structural and biochemical data suggest that acetyl-CoA may promote and stabilize an organization of the BC domain dimer that is catalytically more competent. Studies with E. coli BC showed that changes in the dimer interface could affect catalysis in the active site (Shen et al., 2006), suggesting long-range communication between the two regions (Janiyani et al., 2001; de Queiroz and Waldrop, 2007; Mochalkin et al., 2008). The activation of PC by acetyl-CoA may involve a similar mechanism, although further studies are needed to characterize the molecular details of this communication.
Our structure is for the CoA complex of SaPC, although acetyl-CoA is 80-fold more potent at stimulating this enzyme than CoA (Fig. 3a). It is possible that the acetyl group could enhance the interactions with the binding site in the BC dimer interface, and the negative charge on the sulfur atom of CoA may also be detrimental for the stimulatory effect. In addition, the current structure does not contain any substrates in the active sites (Fig. 5b), which could also affect the stimulation by acetyl-CoA.
The structural comparisons show that CoA binding and the re-organization of the BC domain dimer also caused a change in the PT dimer of SaPC (Fig. 6b), although the PT domain remains in the tetramer interface and residue Tyr1077 is still located in the center of the PT dimer interface in this complex (Fig. 3c). Our earlier kinetic data showed that the Y1077A mutant of SaPC is catalytically inactive in the absence of acetyl-CoA, but could be rescued by the presence of 0.5 mM acetyl-CoA (Xiang and Tong, 2008). These data suggest that acetyl-CoA binding stabilizes the BC domain dimer such that a mutation in the PT domain can be tolerated. The Tyr1077 residue contributes roughly 70 Å2 of the 350 Å2 surface area buried at the PT dimer interface. A much higher concentration of acetyl-CoA is needed to rescue the Y1077A mutant (Ka of 120 μM) than to activate the wild-type enzyme (Ka of 2 μM), demonstrating that the PT domain still has an important contribution to the stability of the SaPC tetramer in complex with CoA.
The largest structural difference between the RePC tetramer and the CoA complex of SaPC is the organization of the PT domains (Fig. 6b). While a dimeric association is observed in SaPC, the two PT domains (called the allosteric domains) in RePC show essentially no interactions with each other (Fig. 6b) and are exposed to the solvent (St. Maurice et al., 2007). On the other hand, the organizations of the BC and CT dimers of RePC are generally similar to those of SaPC. Therefore, the asymmetry of the RePC tetramer is due primarily to the large change in the PT domains.
The bound position of 3′-phospho-ADP in RePC relative to the PT domain is similar to that seen in the CoA complex of SaPC (Fig. 6b). However, the adenine base is located farther away from the BC dimer interface in the RePC structure (Supplementary Fig. 2). In the two monomers of RePC that do not have this compound, the PT domain occupies the binding site for the adenine base, due to the large conformational change in this domain (Supplementary Fig. 2), suggesting that the conformation observed for RePC is incompatible with the binding of four acetyl-CoA molecules.
In summary, our structural studies have revealed a symmetrical tetramer for SaPC in complex with CoA, which is supported by biochemical data. Acetyl-CoA may activate PC by stabilizing the BC dimer and promoting a conformation of the tetramer that is more catalytically competent. Mutagenesis, kinetic and structural studies have shown that Thr908 may play a catalytic role in the CT reaction of PC.
Staphylococcus aureus PC (SaPC, residues 20-1178) was subcloned into vector pET28a (Novagen) and then over-expressed with a compatible plasmid that carries the bacterial biotin ligase (birA) gene in E. coli BL21Star cells, as reported earlier (Xiang and Tong, 2008). SaPC was purified by nickel-agarose affinity (Qiagen) and gel-filtration (S-300, GE Healthcare) chromatography, concentrated to 10 mg/ml, flash-frozen in liquid nitrogen and stored at −80 °C in a buffer containing 20 mM Tris (pH 7.5), 200 mM NaCl, 2 mM DTT, and 5% (v/v) glycerol. An avidin binding gel-shift assay confirmed that the protein was fully biotinylated (data not shown).
All mutants were made with the QuikChange kit (Stratagene) and verified by sequencing. The mutant proteins were expressed and purified following the same protocol as that for the wild-type protein.
The catalytic activity of wild-type and mutant SaPC was determined at room temperature following the appearance of oxaloacetate, which was coupled to NADH oxidation through malate dehydrogenase (Modak and Kelly, 1995). The reaction mixture contained 100 mM Tris (pH 7.5), 200 mM NaCl, 5 mM MgCl2, 2 mM ATP, 50 mM sodium bicarbonate, varying concentrations of pyruvate, 0.2 mM NADH, 0.1 μM SaPC, and 10 units of malate dehydrogenase (Sigma).
The effect of acetyl-CoA and various analogs on the catalytic activity of wild-type SaPC was determined by running the reactions in the presence of 5 mM pyruvate and varying concentrations of the compounds.
Wild-type and mutant SaPC proteins were crystallized under similar conditions using the sitting-drop vapor diffusion method. The reservoir solution consisted of 20% (w/v) PEG3350 and 200 mM ammonium tartrate. For the mutants, 5 mM ATP was added to the protein solution. For the acetyl-CoA complex, 5 mM acetyl-CoA and 5 mM ATP were added to wild-type protein solution. All crystals grew within 1-3 days at room temperature. Cryo-protection was achieved by transferring the crystals to a 5 μl drop consisting of the reservoir solution supplemented with 20% (v/v) ethylene glycol.
Crystals of the A610T and T908A mutants are isomorphous to those of wild-type SaPC (Xiang and Tong, 2008) and there is a tetramer in the asymmetric unit. Crystals of wild-type SaPC grown in the presence of acetyl-CoA are in a new crystal form. There is a tetramer in the asymmetric unit.
X-ray diffraction data were collected at the National Synchrotron Light Source (NSLS) beamlines X4A and X4C. The diffraction images were processed with the program HKL (Otwinowski and Minor, 1997), and the data processing statistics are summarized in Table 2. The crystal grown in the presence of acetyl-CoA had a long c axis (373 Å) and a relatively long b axis (164 Å), which produced substantial lunar overlaps among the diffraction spots. A lower mosaicity value had to be used during data processing to alleviate this overlap problem, which reduced the quality of the data set and the refinement statistics (Table 2). This was the best diffraction data set that was collected after screening through many crystals.
The structures were solved by molecular replacement using the wild-type SaPC structure as the search model with the program COMO (Jogl et al., 2001). Due to large conformational differences, the structure of the CoA complex was determined only when the individual domains of the monomer were used as the search models in the molecular replacement calculations. Manual rebuilding of the atomic models was carried out using O (Jones et al., 1991) and Coot (Emsley and Cowtan, 2004), and the structure refinement was accomplished with CNS (Brunger et al., 1998) and Refmac (Murshudov et al., 1997), with TLS refinement. The refinement statistics are summarized in Table 2.
SaPC at a concentration of 0.1 mg/ml in a buffer containing 20 mM Tris (pH 7.5), 2 mM NaCl, and 2 mM DTT was incubated with 2 mM acetyl-CoA for 20 minutes. CryoEM grids were prepared following standard procedures and vitrified samples were examined on a JEM-2200FS/CR transmission electron microscope (JEOL Europe, Croissy-sur-Seine, France) at an acceleration voltage of 200 kV. Micrographs were taken on KODAK films under low-dose conditions at a magnification of 50,000, and were digitized on a Z/I PHOTOSCAN (ZEIS) scanner obtaining a final pixel size of 2.82 Å/pixel.
Particles were manually selected in the digitized micrographs and matched to particular reference-based projections. A three-dimensional reconstruction was performed using the Spire-spider package (Frank et al., 1996; Baxter et al., 2007), imposing 2-fold (C2) symmetry. In the calculation of the final 3D density map, a data set of 22,258 individual images was used. Resolution of the cryoEM density map was estimated using a cutoff of 0.15 in the Fourier Shell Correlation (Rosenthal and Henderson, 2003). The atomic coordinates for PC were rigidly fitted into the reconstructed cryoEM map using Chimera (Pettersen et al., 2004), which was also used to produce figures rendering cryoEM maps.
We thank Randy Abramowitz and John Schwanof for setting up the X4A and X4C beamlines at the NSLS; Melisa Lázaro for assistance during cryoEM image processing; C. Huang for careful reading of the manuscript; W.W. Cleland for helpful discussions. This research is supported in part by a grant from the NIH (DK067238) to LT, the Etortek Research Programmes 2007/2009 (Department of Industry, Tourism and Trade of the Government of the Autonomous Community of the Basque Country) and by the Innovation Technology Department of the Bizkaia County to MV.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.