The phosphate moiety of phosphoenolpyruvate does NOT contribute to allosteric regulation of liver pyruvate kinase by fructose-1,6-bisphosphate

Abstract

A linked-function theory for allostery allows for a differentiation between those protein-ligand interactions that contribute the most to ligand binding and those protein-ligand interactions that contribute to the allosteric mechanism. This potential distinction is the basis for analogue studies used to determine which chemical moieties on the allosteric effector contribute to allostery. Although less recognized, the same separation of functions is possible for enzyme-substrate interactions. When evaluating allosteric regulation in human liver pyruvate kinase, the use of a range of monovalent cations (K⁺, NH₄⁺, Rb⁺, Ca⁺, cyclohexylammonium⁺ and Tris⁺) altered substrate (phosphoenolpyruvate; PEP) affinity, but maintained similar allosteric responses to the allosteric activator, fructose-1,6-bisphosphate (Fru-1,6-BP). Because crystal structures indicate that the active site monovalent cation interacts directly with the phosphate moiety of the bound PEP substrate, we questioned if the phosphate moiety might contribute to substrate binding, but not to the allosteric mechanism. Here, we demonstrate that the binding of oxalate, a non-phosphorylated substrate/product analogue, is allosterically enhanced by Fru-1,6-BP. That observation is consistent with the concept that the phosphate moiety of PEP is not required for the allosteric function, even though that moiety likely contributes to determining substrate affinity.

Introduction

In structure/function studies aiming to evaluate allosteric mechanisms, we find it surprising that there are many fewer studies that probe substrate-active site interactions compared to those that target effector binding site-allosteric effector interactions. One outcome of the linked-function theory related to allosteric function is that if binding of an effector alters substrate binding, then binding of the substrate to the enzyme must modify the binding of the effector^1–4. This mutual influence of the two ligands on each other’s binding to the protein is called reciprocity. Reciprocity emphasizes the need to focus on substrate-active site interactions with the same intensity that is used to study allosteric effector-effector binding site interactions.

Linked-function theory of allosteric systems can accommodate a scenario in which there is a separation of function between those protein-ligand interactions that are the primary determinants of ligand (substrate or effector) binding affinity and those protein-ligand interactions that contribute to the allosteric mechanism¹. In our studies of allosteric binding sites, we have identified systems in which there are clear distinctions among interactions that contribute to binding and that contribute to allostery^{5, 6}. We have also identified systems in which the interactions that are primary determinants of ligand affinity are the same interactions that contribute to allostery, i.e., no separation of binding and allosteric functions⁷. The same potential separation of function is possible in the substrate-active site interaction.

In human liver pyruvate kinase (hLPYK), the allosteric activator, fructose-1,6-bisphosphate (Fru-1,6-BP), causes a similar magnitude of allosteric enhancement of substrate (phosphoenolpyruvate, PEP) affinity when assayed in a range of monovalent cation types (K⁺, NH₄⁺, Rb⁺, Ca⁺, cyclohexylammonium⁺ and Tris⁺)⁸. Despite the consistency in the allosteric response, varying the monovalent cation type resulted changes in the observed PEP affinities (in the absence of effector). Furthermore, the monovalent cation that is required for activity is known to bind directly in the active site and interact with the phosphate moiety of PEP⁹. Taken together, we questioned if the phosphate moiety of PEP contributes to the mechanism of allosteric activation by Fru-1,6-BP.

To answer this question, we considered whether a non-phosphorylated substrate/product analogue is allosterically regulated. The analogue evaluated here was oxalate, an analogue and inhibitor that has also been co-crystalized with PYK isozymes¹⁰. Unfortunately, hLPYK is not stable at physiological pH for the longer data collection time used in many biophysical binding assays. It follows that a suitable binding assay has not been identified to-date. As an alternative, in this study, we treated oxalate as a competitive inhibitor in an activity assay. K_i-oxalate values were evaluated over a range of Fru-1,6-BP concentrations to demonstrate that oxalate binding is allosterically enhanced by Fru-1,6-BP, similar to the allosteric enhancement of PEP binding. Therefore, we conclude that the phosphate-moiety of PEP is not required for allosteric activation.

Materials and Methods

Protein preparation and assay design followed that previously reported⁵. The reaction mix included 50 mM HEPES (pH 7.5 at 30ºC), 10 mM MgCl₂, 2 mM (K)ADP, 0.1 mM EDTA, 0.18 mM NADH, and 19.6 U/mL lactate dehydrogenase. PEP, oxalate and Fru-1,6-BP concentrations were varied. The 2mM ADP condition is saturating for this enzyme. The rate of change A₃₄₀ was determined at different PEP concentration to determine the K_app. K_app-PEP determined at various concentrations of oxalate (i.e., the competitive inhibitor) was used to quantify K_i-oxalate. K_i-oxalate was determined at various concentrations of Fru-1,6-BP to evaluate the allosteric coupling between oxalate and Fru-1,6-BP binding. All assays were completed in 96-well plates with the 12 concentrations of PEP and 8 concentrations of oxalate in each plate. Each 96-well plate included one Fru-1,6-BP concentration. K⁺ and Na⁺ contributions from each ligand and buffer component were summed and NaCl and KCl were added for final concentrations of 200 mM Na⁺ and 150 mM K⁺.

Data fitting was with the nonlinear least-squares analysis of Kaleidagraph (Synergy) software^{8, 11}. Initial rates determined as a function of PEP were fit to:

$$v=\left\{\left(V_{max}{[\text{PEP}]}^{nH}\right)/\left(K_{app-PEP}{}^{nH}+{[\text{PEP}]}^{nH}\right)\right\}+c[\text{PEP}]$$ Equation 1

Where V_max is the maximum velocity, K_app-PEP is the concentration of substrate that yields a rate equal to one-half the V_max, n_H is the Hill coefficient, and c is a linear increase in activity. The c term accounts for a slight but continual increase in activity at high PEP concentrations after the V_max activity has been reached, a property that is well described for hLPYK. K_app-PEP values at various oxalate concentrations were fit for competitive inhibition:

$$K_{app-PEP}=\left\{K_{PEP}\left([\text{Oxalate}]+K_{i-app-oxalate}\right)\right\}/K_{i-app-oxalate}$$ Equation 2

K_{i-app-oxalate} and K_app-PEP were evaluated over a concentration range of Fru-1,6-BP and fit to:

$$K_{app}=K_a\{\left(K_{ix-FBP}+[\text{Fru}-1,6-\text{BP}]\right)/\left(K_{ix-FBP}+Q_{\alpha x}[Fru-1,6-\text{BP}]\right).$$ Equation 3

In equation 3, FBP is used as the parameter subscript in place of Fru-1,6-BP. In Equation 3, when the K_app under consideration is K_{i-app-oxalate}, then the K_a from the fit is K_oxalate that in turn is equal to K_{i-app-oxalate} when [Fru-1,6-BP] = 0. Likewise, when the K_app under consideration in Equation 3 is K_app-PEP, then the K_a from the fit is K_PEP that in turn is equal to K_app-PEP when [Fru-1,6-BP] = 0. K_ix is the dissociation constant for Fru-1,6-BP (X) binding to the protein in the absence of oxalate; and:

$$Q_{\text{ax}}=K_{\text{ia}}/K_{\text{ia}/\text{x}}=K_{\text{ix}}/K_{\text{ix}/\text{a}}.$$ Equation 4

In Equation 4, K_ia/x is the binding of the A ligand (either PEP or oxalate) to the enzyme when the x ligand (Fru-1,6-BP) is at a saturating concentration. K_ix/a is the binding of Fru-1,6-BP when the A ligand is at a saturating concentration.

Results and Discussion

The goal of this study was to evaluate the allosteric coupling between oxalate binding in the active site of hLPYK and Fru-1,6-BP binding in the allosteric site of that enzyme. To reach this goal and due to the lack of a suitable binding assay, we used competitive binding in activity assays. This experimental design required more than 3,000 assays to obtain the final data presented in Figure 2B. Due to the extensive effort required, we chose not to extend our analysis to other substrate analogues.

Figure 2.

Ligand binding in the active site as a concentraiton of allosteric effector Fru-1,6-BP concentraiton. Panel A. Both K_app-PEP values determined at zero oxalate (open circle) and K_PEP values from fits in Figure 1 (open triangle) were compared to K_app-PEP vs. Fru-1,6-BP concentration evaluated in a single plate (solid circle). Panel B. Plot of K_{i-app-oxalate} determined over a dilution series of Fru-1,6-BP. Each data point is derived from one 96-well plate of assays evaluated as in Figure 1. These data were fit to Equation 3. Parameter error estimates from the fits in Figure 2 were used to weight the K_{i-app-oxalate} data during data fitting. Error bars are paramater error estimates obatined from data fits to Equation 2 and when not apparent, they are smaller than data points. Lines are best fits to Equation 3. Fit paramaters are compared in Table 1. Please note the y-axis range differences in the two pannels.

Interestingly, several internal standards were provided by our use of activity assays to determine oxalate binding via competitive binding with PEP. K_app-PEP values at zero oxalate and K_PEP from the K_app-PEP vs. [oxalate] (Figure 1) graph were determined at various Fru-1,6-BP concentrations. Due to the design of varying the concentrations of PEP and oxalate within one 96 well plate and Fru-1,6-BP among 96 well plates, these K_app-PEP and K_PEP values were determined from different plates and at different times. Those data were compared to K_app-PEP vs. [Fru-1,6-BP] in which all values were assayed in the same plate (Figure 2A). Clearly, the data collected across multiple plates show good agreement with data collected simultaneously in one plate, which, in turn, increases confidence in the evaluation of oxalate vs. [Fru-1,6-BP] that requires assays across many 96 well plates.

Figure 1.

K_app-PEP values (from fits of the primary data to the equation 1) as a function of oxalate concentration. The representative triplicate (solid squares, open circles and open triangles) data sets included here were collected at 8.3 μM Fru-1,6-BP. Similar data were collected at a range of Fru-1,6-BP concentrations. Data were fit to Equation 2 for competitive binding. Parameter error estimates obtained from the data fits used to obtain K_app-PEP were used to weight the K_app-PEP data when fitting this data to Equation 2. Both the K_app-PEP determined at zero oxalate (a measured value) and the K_PEP determined from the data fit to a competitive equation (a fit value) are marked. These two values were used in Figure 2A. K_{i-app-oxalate} values from fits like the one presented here were used in Figure 2B. Error bars are paramater error estimates obatined from data fits to Equation 1 and when not apparent, they are smaller than data points. Lines are best fits to Equation 2.

Once K_i-oxalate values were obtained, they were assessed across a range of Fru-1,6-BP concentrations (Figure 2B). Fits of this data to Equation 3 allow for a quantitative determination of Q_{ax-oxalate/FBP}, the allosteric coupling parameter between oxalate binding in the active site and Fru-1,6-BP binding in the allosteric site. As another internal control, the K_ix-FBP determined from the evaluation of oxalate binding are similar to the K_ix-FBP value determined from the evaluation of PEP affinity (Table 1). The obtained Q_{ax-oxalate/FBP} value is very similar to the Q_ax-PEP/FBP value. At the same time, oxalate binds approximately 7 times tighter to the active site than PEP (Table 1: compare K_PEP vs. K_oxalate). Therefore, consistent with previous observations for changing the active site monovalent cation type, the allosteric coupling does not appear to rely on the phosphate/monovalent cation interaction.

Table 1:

Fit values from Figure 2

	From panel A	From panel B
aKPEP [mM]	0.66±0.01
Koxalate [mM]		0.098±0.003
Kix-FBP [μM]	0.62±0.03	0.34±0.04
Qax-PEP/FBP	16.2±0.5
Qax-oxalate/FBP		14.6±0.5

^aValues taken from the response of K_PEP to Fru-1,6-BP conentration in Figure 2A.

Our conclusion that the phosphate-moiety of PEP is not required for the allosteric activation of hLPYK by Fru-1,6-BP is particularly noteworthy given the interactions of bound PEP with the PYK protein. Primarily, 1) the phosphate moiety of PEP interacts with the monovalent cation; 2) the phosphate moiety interacts with the protein bound divalent cation; 3) a carboxyl oxygen and the oxygen that bridges carbon and phosphate both interact with the protein bound divalent cation, 4) the phosphate moiety interacts with the side chain of Arg85, and 5) the sidechain of Thr340 may interact with the carboxyl group of PEP (depending on which structure of PYK is evaluated). This complete list of interactions of the substrate with the protein is striking in the very limited number of direct interactions between the bound PEP and the protein. Instead, most interactions are indirect through bound cations. However, many of the entries in this list of PEP interactions involve the phosphate moiety of PEP. Thus, we find it surprising that the allosteric mechanism does not appear to depend on the phosphate moiety of PEP.

Given the current conclusion that the phosphate moiety is not required for allosteric function and the small size of PEP, we find it convenient to speculate that it is the carboxyl group of PEP that contributes to allosteric regulation by Fru-1,6-BP. However, there are only a limited number of observations that can support that speculation. There is little evidence that rabbit muscle PYK is regulated by Fru-1,6-BP, especially if it is not first inhibited by the addition of Phe^{12, 13}; nonetheless, an infrared spectroscopy study of the influence of Fru-1,6-BP on the vibrational spectra of PEP bound to that enzyme identified a band shift that might represent a change in the carboxylate group of PEP¹⁴. Interestingly, a study of a range of substrate analogues in the allosteric inhibition of rabbit muscle PYK by Phe also resulted in the speculation that the carboxyl group of PEP was the primary contributor to the allosteric mechanism in that system¹⁵. Clearly, data to support the speculated role of the carboxyl group of PEP in either activating or inhibiting allosteric mechanisms is scarce. There are even fewer observations that can provide information about what structural/dynamic change in the protein near the carboxyl group of the bound substrate could be the cause of increased or decreased substrate binding in response to an allosteric regulation. Nonetheless, existing data indicate that the phosphate moiety of PEP does not contribute to either allosteric activation or allosteric inhibition in mammalian PYK isozymes.

Figure 3.

A diagram of the active site of hLPYK with phosphoenolpyruvate (PEP) bound. This diagram is largely based on the cocrystalized structure of M₁PYK with phospholactate⁹.

Note that PEP primarily coordinates to enzyme bound K⁺ and Mg²⁺, rather than forming many interactions directly with amino acid residues from the protein. Structures of pyruvate and oxalate are included in a similarly orientation for comparison.

Highlights.

1. Oxalate binds to human liver pyruvate kinase competitively with the substrate, PEP.

2. Like PEP, the binding affinity of oxalate to human liver pyruvate kinase is allosterically enhanced by the activator, fructose-1,6-bisphosphate.

3. Based on the response of the non-phosphorylated analogue, the phosphate of PEP does not appear to be required for allosteric regulation by fructose-1,6-bisphosphate.

Abbreviations:

PYK

pyruvate kinase

hLPYK

the pyruvate kinase from human liver

PEP

phospho(enol)pyruvate