Interaction of Human 3-Phosphoglycerate Kinase with Its Two Substrates: Is Substrate Antagonism a Kinetic Advantage?

https://doi.org/10.1016/j.jmb.2011.04.048Get rights and content

Abstract

Substrate antagonism has been described for a variety of enzymes with more than one substrate and is characterized by a lowering of the affinity of one substrate in the presence of the other(s). 3-Phosphoglycerate kinase (PGK) catalyzes phosphotransfer from 1,3-bisphosphoglycerate (bPG) to ADP to give 3-phosphoglycerate (PG) and ATP, and is subject to substrate antagonism. Because of the instability of bPG, antagonism has only been described between PG and ATP or ADP. Here, we show that antagonism also occurs between bPG and ADP. Using the stopped-flow method, we show that the dissociation constant for one substrate increases in the presence of the other, and that this decrease in affinity is mainly due to an increase in the dissociation rate constant. As a consequence, there is an increase in the overall interaction kinetics. Interestingly, in the presence of the mirror image of natural d-ADP, l-ADP (a good substrate for PGK), antagonism is absent. Using rapid-quench-flow, we studied the kinetics of ATP formation. The time courses present the following: (1) a lag with l-ADP, but not with d-ADP, the kinetics of which were similar to the interaction kinetics measured by stopped-flow; (2) a burst that is directed by the phosphotransfer; and (3) a steady-state that is rate limited by the release of product kinetics. Structural explanations for these results are proposed by analyzing the crystallographic structure of the fully closed conformation of PGK in complex with l-ADP, PG, and the transition-state analogue AlF4 compared to previously determined structures.

Introduction

A recurrent question regarding two substrate enzymes that bind their substrates randomly is whether or not the binding process for each substrate is independent of the presence of the other. Segel proposed that if the Km/Kd ratio is < 1 for a given substrate, there is substrate synergy by which the binding of one substrate increases the affinity for the second.1 Segel's proposition was based on the assumption that Kd is the dissociation constant for the binary complex E·S1, whereas Km is the dissociation constant for the ternary complex E·S1·S2. On the other hand, if the affinity for one substrate is reduced by the presence of the other, then the ratio is > 1. This effect is called substrate antagonism.2, 3, 4

With the phosphotransferase group of enzymes, both substrate synergy and substrate antagonism have been observed. For example, yeast hexokinase5 and rabbit muscle creatine kinase6 are subject to substrate synergy, whereas phosphofructokinase from Escherichia coli,2, 7 Rho-associated protein kinase (ROCK I; an enzyme that is implicated in cell adhesion and smooth muscle adhesion),8 and, in particular, 3-phosphoglycerate kinase (PGK; EC 2.7.2.3)9, 10 are subject to substrate antagonism.

PGK catalyzes the reversible transfer of a phospho group between 1,3-bisphosphoglycerate (bPG) and ADP:bPG + ADPPG + ATP

PGK catalyzes the first ATP-generating step in glycolysis and has been termed an archetypal phosphotransferase.11 Furthermore, as PGK is relatively nonspecific for its nucleotide substrates, it may be a component of the cascade of kinases activating nucleoside prodrugs to pharmacologically active nucleoside triphosphates.12, 13 l-Nucleoside analogue prodrugs are of particular interest as a new class of antiviral and anticancer agents (Mathé and Gosselin14 and references cited therein). With human PGK (hPGK), l-ADP is almost as competent a substrate as natural d-ADP.15 In order to understand fully the role of PGK in the activation of l-nucleoside prodrugs, one can compare the chemical reaction pathways of PGK with ADP enantiomers, especially the involvement of substrate antagonism.

PGK is composed of two domains, with the 3-phosphoglycerate (PG) or bPG binding site on the N-domain and the nucleotide binding site on the C-domain. In the absence of substrates, the enzyme is in an “open” conformation in which the two sites are separated. For phosphotransfer to be allowed, upon the binding of both substrates, an extensive hinge bending motion occurs that leads to a “closed” conformation, resulting in amalgamation of the sites.16, 17 While extensive structural studies have been carried out on the open form of PGKs from a number of sources (Kovári and Vas18 and references cited therein), the first structure of a closed form was, until recently (see the text below), obtained only for the enzymes from Trypanosoma brucei (the causative agent of sleeping sickness)19 and Thermotoga maritima.20 In these works, it was shown that, in the ternary PGK·PG·ADP complex, there was a dramatic closing of the large cleft that separates the two substrate sites, thereby bringing bound PG and ADP in close proximity. Of course, there is no phosphotransfer in this complex, and it is remarkable that an “abortive” complex mimics the catalytic PGK·bPG·ADP complex. We note that with the enzyme from yeast, White et al. proposed that the conformation of PGK·PG·ADP is less closed than that of PGK·bPG·ADP.21 This study was based on methods involving flash photolysis of caged ATP and infrared spectroscopy.

It has been difficult to obtain the closed structure of mammalian PGK. The artifice of exploiting the abortive PGK·PG·ADP with pig PGK was unsuccessful.22 Recently, Cliff et al. reported a closed structure of hPGK.11 High-resolution structures of the transition-state analogue complexes PGK·PG·MgF3·ADP and PGK·PG·AlF4·ADP, where metal fluorides replace the transferable phosphate of ATP, were obtained. It was suggested that a zero local net charge at the active site is a key feature of the phosphotransfer process.

In early works, it was suggested that yeast PGK binds each substrate independently of the concentration of the other (see, for instance, Scopes23). However, more recently, it was shown that pig PGK is subject to substrate antagonism. Thus, Km/Kd > 1 with PG4, 10, 24 and with both ATP and ADP.3 Merli et al. provided evidence that antagonism is connected to the domain closure induced by the substrates. In a review of the work of Bernstein et al.19 on the closed structure of PGK from T. brucei, Blake25 proposed that the small conformational change caused by PG binding alone “primes” the enzyme to react differently from the ligand-free enzyme to the binding of ATP and ADP, and that the heart of this priming mechanism is helix 14, which belongs to the C-terminal end of hPGK but is located in the N-domain.

In earlier work,9 we gave a transient kinetic explanation for substrate antagonism with hPGK through a study of the kinetics of the formation of the abortive ternary complex PGK·PG·ADP. Because of the pharmacological interest of l-nucleoside analogues in antiviral therapies,14 we also studied the transient kinetics of the interaction of l-ADP with PGK. The two enantiomers of ADP behave differently. Thus, with d-ADP, PG decreases the affinity of PGK for the nucleotide not because kon decreases but because koff increases 10-fold. With l-ADP, PG had little effect on the rate constants of interaction with PGK.

Here, our aim was to characterize both kinetically and structurally the phenomenon of substrate antagonism in the forward reaction of PGK (that is to say, starting from bPG and d-ADP). Although a weakening of the binding of d-ATP by bPG has been observed,10 we are not aware of works that demonstrate antagonism between the substrates d-ADP and bPG. Furthermore, the transient kinetics of the interaction of bPG with PGK do not appear to have been studied. From steady-state and equilibrium studies, the dissociation constant (koff) for bPG is low, probably < 50 s 1.4, 23

The paucity of information on the kinetics of bPG binding is due to difficulties in working with this unstable substrate (see, for example, Jakeman et al.26). Thus, Negelein found that at pH 7.2 and 38 °C, bPG decomposed rapidly (half-life, 27 min).27 Harrigan and Trentham observed that its half-life at pH 7.5 and − 20 °C was 2 weeks.28 Despite this technical difficulty, it seemed important to investigate in detail the kinetics of the interactions of bPG with PGK.

We show that at 4 °C and in a buffer containing 30% methanol, bPG is stable for at least 1 h. This allowed us to exploit fluorescence stopped-flow to obtain the transients of its interaction with hPGK and to compare the effects of d-ADP and l-ADP thereon. By this method, we conclude that the substrate binding kinetics that lead to the formation of the catalytic PGK·bPG·ADP complexes are mutually antagonistic with d-ADP, but much less so with l-ADP—a situation that is very similar to that for the formation of the abortive PGK·PG·ADP complexes.9 We confirmed this conclusion by following the kinetics of ATP formation by a chemical sampling method in which solutions of d-ADP or l-ADP are mixed with hPGK + bPG in a rapid-quench-flow apparatus, the mixtures are allowed to age in the millisecond time range and quenched in acid, and d-ATP or l-ATP is measured.

We also present a crystal structure of hPGK in a fully closed conformation in complex with l-ADP, PG, and the transition-state analogue AlF4. Structural explanations for the substrate antagonism observed with d-ADP, but not with l-ADP, are proposed by comparing this new structure to previously determined structures of PGK in the open or closed conformations. Finally, we address the question: Is substrate antagonism a kinetic advantage or disadvantage for the PGK-catalyzed reaction?

Section snippets

Stability of bPG

In view of the reputed lability of bPG, we tested its stability under our experimental conditions in the presence or in the absence of PGK. Thus, Huskins et al. noted that free bPG was hydrolyzed completely when stored at 5 °C for 5 days, whereas a “considerable amount was still observed” when it was bound to PGK.29

As illustrated in Fig. 1, bPG, when kept at 4 °C in a buffer containing 30% methanol, was not hydrolyzed whether or not PGK was present. At 20 °C, in the presence of an equimolar

Conclusions: Causes and Implications of Substrate Antagonism with PGK

Substrate antagonism has been observed with another kinase of the glycolysis pathway, 6-phosphofructokinase. Deville-Bonne et al. provided three possible explanations for the antagonism of the binding of the charged substrates ATP and fructose-6-phosphate to phosphofructokinase: electrostatic repulsion between the substrates, steric hindrance, or a conformational change in the kinase.2 We suggest that the substrate antagonism observed with PGK may be due to a mutual repulsion of the charged

Proteins and reagents

Recombinant hPGK was produced in BL21(DE3)/pDIA17 E. coli cells transformed with the PGK-pET28a plasmid and purified as described previously.9 l-ADP was synthesized according to previous work.9, 43 bPG was prepared as previously described.15 Other chemicals were purchased from Sigma Aldrich.

Experimental conditions

In all kinetic experiments, the buffer contained 20 mM triethanolamine (pH 7.5), 0.1 M potassium acetate, 1 mM free Mg2+ as magnesium acetate, and 30% methanol. In our experiments, ADP and ATP refer to Mg2+

Acknowledgements

We thank Dr. Dominique Deville-Bonne for the generous gift of PGK expressing E. coli strain and Lionel Verdoucq for the use of a cell disruptor. This work was supported by CNRS and ANRS (contract number 2008-090). The authors thank the Partnership for Structural Biology (Grenoble, France) for an integrated structural biology environment.

References (52)

  • RaoB.D.N. et al.

    31P NMR study of bound reactants and products of yeast 3-phosphoglycerate kinase at equilibrium and the effect of sulfate ion

    J. Biol. Chem.

    (1978)
  • SegelI.H.

    Behaviour and analysis of rapid equilibrium and steady-state enzyme systems

  • MerliA. et al.

    Nucleotide binding to pig muscle 3-phosphoglycerate kinase in the crystal and in solution: relationship between substrate antagonism and interdomain communication

    Biochemistry

    (2002)
  • VasM. et al.

    Adenine nucleotides affect the binding of 3-phosphoglycerate to pig muscle 3-phosphoglycerate kinase

    Eur. J. Biochem.

    (1984)
  • ViolaR.E. et al.

    Substrate synergism and the kinetic mechanism of yeast hexokinase

    Biochemistry

    (1982)
  • FurterR. et al.

    Creatine kinase: the reactive cysteine is required for synergism but is nonessential for catalysis

    Biochemistry

    (1993)
  • FuterO. et al.

    Phosphoryl transfer is not rate-limiting for the ROCK I-catalyzed kinase reaction

    Biochemistry

    (2006)
  • GondeauC. et al.

    Differences in the transient kinetics of the binding of d-ADP and its mirror image l-ADP to human 3-phosphoglycerate kinase revealed by the presence of 3-phosphoglycerate

    Biochemistry

    (2008)
  • VasM. et al.

    Antagonistic binding of substrates to 3-phosphoglycerate kinase monitored by the fluorescent analogue 2′(3′)-O-(2,4,6-trinitrophenyl)adenosine 5′-triphosphate

    Biochem. J.

    (1994)
  • CliffM.J. et al.

    Transition state analogue structures of human phosphoglycerate kinase establish the importance of charge balance in catalysis

    J. Am. Chem. Soc.

    (2010)
  • GondeauC. et al.

    Molecular basis for the lack of enantioselectivity of human 3-phosphoglycerate kinase

    Nucleic Acids Res.

    (2008)
  • SchlichtingI. et al.

    Structures of active conformations of UMP kinase from Dictyostelium discoideum suggest phosphoryl transfer is associative

    Biochemistry

    (1997)
  • BanksR.D. et al.

    Sequence, structure and activity of phosphoglycerate kinase: a possible hinge-bending enzyme

    Nature

    (1979)
  • KováriZ. et al.

    Protein conformer selection by sequence-dependent packing contacts in crystals of 3-phosphoglycerate kinase

    Proteins

    (2004)
  • BernsteinB.E. et al.

    Synergistic effects of substrate-induced conformational changes in phosphoglycerate kinase activation

    Nature

    (1997)
  • WhiteE.M. et al.

    Infrared studies reveal unique vibrations associated with the PGK–ATP–3-PG ternary complex

    Biochemistry

    (2008)
  • Cited by (0)

    View full text