Article pubs.acs.org/biochemistry
Utilization of Substrate Intrinsic Binding Energy for Conformational Change and Catalytic Function in Phosphoenolpyruvate Carboxykinase Troy A. Johnson,‡,§ Matthew J. Mcleod,† and Todd Holyoak*,†,‡ †
Department of Biology, University of Waterloo, Waterloo, ON N2L 3G1, Canada Department of Biochemistry and Molecular Biology, The University of Kansas Medical Center, Kansas City, Kansas 66160, United States
‡
ABSTRACT: Phosphoenolpyruvate carboxykinase (PEPCK) is an essential metabolic enzyme operating in the gluconeogenesis and glyceroneogenesis pathways. Previous work has demonstrated that the enzyme cycles between a catalytically inactive open state and a catalytically active closed state. The transition of the enzyme between these states requires the transition of several active site loops to shift from mobile, disordered structural elements to stable ordered states. The mechanism by which these disorder−order transitions are coupled to the ligation state of the active site however is not fully understood. To further investigate the mechanisms by which the mobility of the active site loops is coupled to enzymatic function and the transitioning of the enzyme between the two conformational states, we have conducted structural and functional studies of point mutants of E89. E89 is a proposed key member of the interaction network of mobile elements as it resides in the R-loop region of the enzyme active site. These new data demonstrate the importance of the R-loop in coordinating interactions between substrates at the OAA/PEP binding site and the mobile R- and Ω-loop domains. In turn, the studies more generally demonstrate the mechanisms by which the intrinsic ligand binding energy can be utilized in catalysis to drive unfavorable conformational changes, changes that are subsequently required for both optimal catalytic activity and fidelity.
P
pathways required for cancer cell proliferation has been demonstrated.10,11 In contrast, the role of mPEPCK in metabolism is less well understood, and until recently, its role in metabolism has historically been underappreciated.8,9 Structural studies predominantly focused on cPEPCK have revealed many structural and functional details of the enzyme as it moves through the catalytic cycle.12−17 The structural data illustrate that PEPCK is a monomeric enzyme that is divided into two domains: an N-terminal domain (residues 1−259) and a C-terminal domain (residues 260−622). The active site is located in a cleft formed at the interface formed between the two domains. Inspection of the active site illustrates that it is cationic in nature, composed of the two divalent metal ions, and specific lysine and arginine residues that are positioned to readily conduct the reversible decarboxylation/carboxylation and phosphoryl transfer half-reactions.12 On the basis of a stepwise chemical mechanism that results in the formation of a reactive enolate intermediate, it was postulated that to effectively mediate the reversible chemical conversion of OAA into PEP, the enzyme must both stabilize the enolate intermediate upon formation and protect the enolate from favorable protonation, which would decouple the consumption
hosphoenolpyruvate carboxykinases (PEPCKs) catalyze the decarboxylation of oxaloacetate (OAA), forming an enolate of pyruvate intermediate that is subsequently phosphorylated forming PEP. The isoform of the enzyme found in animals utilizes GTP as the phosphate donor. While in vitro the reaction is freely reversible, the consensus is that in most organisms PEPCK operates primarily in the direction of PEP synthesis.1 PEPCK is a metal-requiring enzyme demonstrating an absolute requirement on divalent cations for activity. One metal ion (M1) binds to the enzyme in the absence of substrates acting as a true metal cofactor, and Mn2+ is typically found to be the most activating M1 cation in the GTPdependent isoforms.2−4 In addition, a second divalent metal ion (M2) is required for the reaction, as the metal−nucleotide complex is the active form of the nucleotide substrate; either Mg2+ or Mn2+ is able to fill this role. In animals, PEPCK is present as both cytosolic (cPEPCK) and mitochondrial (mPEPCK) isoforms, with the relative distribution of the two isoforms being tissue- and species-dependent.5−7 Historically, cPEPCK has been demonstrated to function as a key cataplerotic enzyme, removing TCA cycle anions for usage in downstream metabolic pathways, including the synthesis of glucose, glycerol, triglycerides, and serine (reviewed in refs 8 and 9). It has more recently been suggested that a function of cPEPCK may be to act as a general regulator of TCA cycle flux, and the importance of cPEPCK in the regulation of anabolic © 2015 American Chemical Society
Received: November 9, 2015 Revised: December 22, 2015 Published: December 28, 2015 575
DOI: 10.1021/acs.biochem.5b01215 Biochemistry 2016, 55, 575−587
Article
Biochemistry of OAA from the production of PEP through the formation of pyruvate.12,14 Another informative aspect of the structural studies was the demonstration that conformational changes, both global and local to the active site, were occurring during the catalytic cycle.12,13,17−19 The structural data illustrate that three predominant mobile loop elements are positioned at the active site and their position and/or mobility is directly coupled to the ligation state of the enzyme as it proceeds through the catalytic cycle. These loop structures are the R-loop (residues 85−92), the P-loop (residues 284−292), and the Ω-loop (residues 464−474), which functions as a lid and gates access to the active site. The mobile Ω-loop lid in PEPCK is reminiscent of a similar structure that has been well characterized in triosephosphate isomerase.20−23 In addition to the mobile loop elements situated in the vicinity of the active site, a global motion that involves the opening and closing of the N- and C-terminal lobes of the enzyme is also observed. On the basis of the data presented here, it is postulated that the closure of the N- and C-terminal lobes reduces the volume of the active site pocket and thereby results in the phosphoryl donor and acceptor being positioned closer to each other, shortening the phosphoryl transfer distance and therefore facilitating efficient catalysis. The question arises as to how the chemical composition of the ligands at the active site impact the dynamic interplay of the aforementioned structural elements through remodelling the energetic landscape that defines the local and global structure of the enzyme as it moves along the catalytic pathway. The P-loop and R-loop make direct contacts with the ligands occupying the nucleotide and PEP/OAA subsites of the active site, respectively, and thus, the stabilization of these loops would seem to be the result of direct favorable contacts between the enzyme and bound ligands. In contrast, the mechanism by which the Ω-loop lid is stabilized in a closed conformation, an event that is coincident with the closing of the C- and Nterminal lobes, is less obvious. Inspection of the structural data suggests that in neither the Ω-loop’s open nor closed conformation are direct contacts between the mobile lid and substrates present that would influence its stability in either conformation in the presence of ligands versus in their absence. This observation is in contrast to the Ω-loop found in TIM that is stabilized in its closed conformation in the presence of substrate through direct substrate interaction with its phosphate “gripper” element.24 Examination of the structural data suggests the possibility that the Ω-loop lid of PEPCK indirectly senses the occupancy of the PEP/OAA binding pocket via interactions mediated by the R-loop (Figure 1). In this model, the R-loop is stabilized in its closed, ordered state via contacts between the ligand at the OAA/PEP binding pocket and R87, which in turn allows for additional stabilization of the R-loop through contacts between R283 and E89. Importantly, this conformation of the R-loop relieves steric constraints between the R-loop and the Ω-loop lid and, upon lid closure, allows for stabilization of the closed lid through interactions between the E89 carboxylate on the R-loop and H470 on the lid. Additional stabilizing contacts are also established between E469 on the lid and the backbone amides of E89 and S90 on the R-loop. As described, this sensing mechanism provides a plausible mechanism for the lid to sense the occupancy of the OAA binding subsite that in turn can influence the stabilization of the closed form of the lid and subsequently the closed form of the enzyme that is required for optimal catalytic function.
Figure 1. Proposed interaction network communicating OAA/PEP binding site occupancy to the Ω-loop lid domain. The residues comprising the Ω-loop are colored red; the residues comprising the Rloop are colored blue, and the remaining enzyme and associated residues are colored gray. Key residues involved in the proposed interaction network are labeled, and the dashed lines represent potential hydrogen bonds of ≤3.0 Å. Oxalate (green) coordinated to the active site manganese ion (pink sphere) is also illustrated and labeled.
To test our model in which the ligation state of the OAA binding site is communicated to the lid domain via the R-loop, we have generated three variants of PEPCK (E89A, E89D, and E89Q) that differ in the amino acid identity of this central member of the interaction network. We have characterized these enzyme variants structurally and functionally and have concluded that the data are consistent with a model in which the mobile R-loop is stabilized via interactions with ligands at the OAA binding site, which in turn allows for stable interactions between E89 and H470, S90, and E469 that in turn results in the closed form of the lid being stabilized in the presence of the catalytically competent Michalelis and enolate complexes, a process that is essential for the catalytic function of PEPCK.
■
MATERIALS AND METHODS Materials. The nucleotides (GDP, GTP, IDP, ITP, and ADP) and PGA were purchased from Sigma. PEP and NADH were from Chem-Impex. DTT and TCEP were from Gold BioTech. HEPES buffer was from Research Organics. Oxalic acid was from Fluka. 3-Sulfopyruvate (βSP) was synthesized and purified as previously described.25 Ni-NTA resin was purchased from Qiagen. HiQ, P6DG, and Chelex resins were from BioRad. All other materials were of the highest grade commercially available. Enzymes. The coupling enzymes used in the kinetic assays for PEPCK, malate dehydrogenase (Calzyme), lactate dehydrogenase (Calzyme), and pyruvate kinase (Roche) were used as provided by the supplier. His6-SUMO protease was expressed and purified as previously described.15 Generation of the E89A, E89D, and E89Q Variants. The gene for rat cPEPCK encoding the entire 622-residue protein expressed as a fusion construct with SUMO was expressed and purified as previously described.15 This wild-type construct was used as the starting vector to create the E89 enzyme variants. The forward primers (E89A, 5′-GTGGCCAGGATCGCAAGCAAGACGGTC-3′; E89D, 5′-GTGGCCA576
DOI: 10.1021/acs.biochem.5b01215 Biochemistry 2016, 55, 575−587
Article
Biochemistry GGATCGACAGCAAGACGGTC-3′; E89Q, 5′-GTGGCCAGGATCCAAAGCAAGACGGTC-3′) and their respective reverse complement sequences were utilized using the Stratagene Quik Change protocol (altered codons are highlighted). The resultant mutated DNA was isolated with a Hurricane cleanup kit from Gerard Biotech and sequenced in its entirety to confirm the presence of the desired mutation and the absence of any additional mutations introduced via the polymerase chain reaction protocol. The plasmid DNA was transformed into Escherichia coli BL-21(DE3) electro-competent cells for expression. PEPCK Expression and Purification. PEPCK was expressed and purified as previously described.15 The final protein, containing only the 622 residues of the native enzyme, was concentrated using a 30 kDa Centricon concentrator to a final concentration of 10 mg/mL, determined by absorbance 280 nm (ε280 = 1.7 mL mg−1). The protein was then flashfrozen by pipetting 30 μL drops directly into liquid nitrogen, and the protein pellets were stored at −80 °C. Kinetic Experiments. PEPCK was assayed for enzymatic activity in the physiological direction of PEP formation and the reverse direction of OAA formation. In addition, PEPCK was assayed for the spontaneous production of pyruvate during turnover and the decarboxylation of OAA to pyruvate in the presence of GDP. All assays were conducted utilizing coupled, continuous assays as previously described.15 Nucleotide Binding. PEPCK’s affinity for nucleotide was determined using intrinsic fluorescence quenching experiments as previously described15 in a mixture containing 50 mM HEPES (pH 7.5), 10 mM DTT, 1.4 mM MgCl2, and PEPCK (1−5 μM). Crystallization. Crystals of the E89 variants of PEPCK were grown by vapor diffusion using the hanging drop method. The well solution consisted of 22−26% PEG 3350, 100 mM HEPES (pH 7.4), and water to produce a final volume of 700 μL. All crystals were grown on a siliconized cover slide in a liquid drop that contained 4 μL of 10 mg/mL protein, 2 μL of the well solution, 0.5 μL of 30 mM GDP or GTP, and 0.5 μL of 75 mM MnCl2. The crystals were cryoprotected, and the corresponding complex was formed by soaking the crystal for 30−60 min in a solution containing 25% PEG 3350, 10% PEG 400, 100 mM HEPES (pH 7.4), 2 mM MnCl2, 5 mM GDP or GTP, and, where required, 10 mM PGA, βSP, or oxalate to generate the corresponding PEPCK-Mn2+-Mn2+GTP, PEPCKMn2+-βSP-Mn 2+GTP, PEPCK-Mn2+-PGA-Mn2+GDP, and PEPCK-Mn2+-oxalate-Mn2+GTP complexes. All crystals were cryocooled by being immersed in liquid nitrogen. Collection of Diffraction Data. Data on the cryocooled crystals of the various enzyme complexes maintained at 100 K were collected on beamline 7-1 at the Stanford Synchrotron Radiation Laboratory (Menlo Park, CA) and CMCF beamline 08ID-1 at the Canadian Light Source (Saskatoon, SK). A summary of the data statistics is presented in Tables 1 and 2. Structure Determination and Refinement. The structures of the E89 variants were determined by the molecular replacement method using MOLREP26 in the CCP4 package27 and previously determined WT rat cPEPCK structures of the same complexes [Protein Data Bank (PDB) entries 2QEY, 3DT4, 3DT7, and 3DTB].16,17 For each complex, the molecular replacement solution was refined using Refmac5 followed by manual model adjustment and rebuilding using COOT.28 Ligand, metal, water addition and validation were also performed in COOT. A final round of TLS refinement was
Table 1. Crystallographic Data and Model Statistics for the E89A, E89D, and E89Q Variants of PEPCK in Complex with GTP E89A-Mn2+GTP
E89DMn2+GTP
E89QMn2+GTP 5FH2 0.97949 P21 a = 44.7 Å b = 118.8 Å c = 60.5 Å α = γ = 90.0° β = 109.5° 59.4−1.5 92558
PDB entry wavelength (Å) space group unit cell dimensions
5FH0 0.97949 P212121 a = 60.4 Å b = 85.7 Å c = 118.4 Å α = β = γ = 90.0°
resolution limits (Å) no. of unique reflections completenessa (%) (all data) redundancya I/σ(I)a Rmergea no. of ASU molecules Rfreea Rworka average B-factor (Å2) protein metals GTP waters estimated coordinate error based on maximum likelihood (Å) bond length rmsd (Å) bond angle rmsd (deg) Molprobity statistics (score, percentile, no. of Ramachandran outliers)
69.4−1.6 81130
5FH1 0.97949 P21 a = 45.5 Å b = 119.1 Å c = 60.9 Å α = γ = 90.0° β = 107.7° 59.6−1.55 84925
99.2 (92.9)
94.7 (61.9)
96.4 (71.8)
13.9 37.1 0.06 1 19.8 17.6
7.4 (5.6) 26.8 (4.4) 0.05 (0.29) 1 20.3 (32.4) 17.2 (30.8)
7.0 (4.7) 18.4 (5.0) 0.07 (0.31) 1 17.2 (26.1) 14.8 (23.2)
32.0 28.7 25.6 36.7 0.06
36.7 28.0 29.2 44.5 0.07
26.0 18.7 17.8 37.1 0.04
0.01 1.6
0.02 1.8
0.02 1.9
0.81, 99th, 0
0.83, 100th, 0
1.00, 99th, 0
(10.8) (7.7) (0.36) (23.5) (21.3)
a
Values in parentheses represent statistics for data in the highestresolution shells.
performed for the E89D-Mn2+βSP-Mn2+GTP and E89QMn2+PGA-Mn2+GDP models in Refmac5. A total of 5 and 10 groups per chain were used for the E89D-Mn2+βSP-Mn2+GTP and E89Q-Mn2+PGA-Mn2+GDP structures, respectively, as refinements using more groups per chain did not significantly improve R or Rfree. The optimal TLS groups were determined by submission of the PDB files to the TLSMD server (http:// skuld.bmsc.washington.edu/_tlsmd/index.html).29−31 Model validation, including Ramachandran statistics, was performed using the Molprobity web server (http://molprobity.biochem. duke.edu).32 A summary of the model statistics for all structures is presented in Tables 1and 2. B-Factor Comparison. A quantitative measure of the relative order of the R-loops in the various complexes was determined by calculating the ratio of the average B-factor for the R-loop (amino acids 84−90) to the overall average B-factor for all protein atoms in the structure. This ratio was termed B84−90/Bav and is presented in Table 4. Average B-factor values were determined using Moleman2 [Uppsala Software Factory (http://xray.bmc.uu.se/usf/)].33 577
DOI: 10.1021/acs.biochem.5b01215 Biochemistry 2016, 55, 575−587
Article
Biochemistry
forms. In concert with the increase in the KM value for OAA, there is a general reduction in the kcat for the conversion of OAA to PEP for all mutants by a factor of 5−15. The most significant catalytic defects are observed in E89A, but the differences in kcat among the three mutants (at most a factor of ∼3) are not appreciably significant. When combined, the decrease in the kcat value and the corresponding increase in the KM values for OAA result in an overall decrease in the kcat/KM value for OAA by factors of 102, 151, and 270 in E89D, E89Q, and E89A, respectively. In contrast to the effect of the E89 substitutions on the KM value for OAA, the KM values for GTP are relatively unchanged, exhibiting only a slight reduction in the parameter. This minor effect on the KM value is consistent with the minimal effects on the binding of GTP as determined from the fluorescence quenching experiments using the inosine form of the nucleotide (Table 3). The partially compensating reduction in the KM values for GTP results in the E89 substitutions having a minimal impact on the kcat/KM values for GTP, with the largest reduction being a factor of approximately 8 in E89A. PEP → OAA Direction. When the reversible reaction is examined in the direction of OAA synthesis, a more significant deficit is seen in the kcat for the catalyzed reaction when compared to that in the OAA → PEP direction (Table 3). The E89A and E89Q mutations exhibit a decrease in the parameter by a factor of 68, while the E89D substitution results in a reduction in the parameter by a factor of 29. In contrast to the effect observed on the KM for OAA, the KM values for all substrates are little changed in the PEP → OAA direction. Together with the change in the kcat values, this results in a modest decrease in the kcat/KM values for all substrates (Table 3). More specifically, the kcat/KM value for PEP is moderately affected with a reduction in the parameter by a factor of 19 in the E89A and E89Q mutations, and a factor of 28 in E89D. The most significant effect on the kcat/Km values is seen for GDP, with E89D demonstrating a decrease in the parameter by a factor of 38 and E89Q and E89A demonstrating decreases by factors of 100 and 129, respectively. As described above, the direct binding data support the conclusion that the changes in the kinetic data for GDP are not a result of a perturbation in the binding of the nucleotide to the enzyme as the Kd values for IDP are essentially unchanged (Table 3). The kcat/KM values for CO2 behave in a fashion similar to that of the same parameter for GDP with 29- and 36-fold reductions in E89D and E89Q, respectively, while in E89A, the value decreases by a factor of 84. Decarboxylation Half-Reaction. In contrast to the full kinetic reaction, utilizing OAA and GDP to stimulate the decarboxylation half-reaction, a reaction that results in the formation of pyruvate due to the protonation of the enolate intermediate in the absence of a phosphoryl donor, suggests that all three enzymes can perform the decarboxylation of OAA as effectively as WT. As illustrated in Table 3, the kcat for this process decreases by a factor of
