The transacetylase component (E2) of PDC (pyruvate dehydrogenase complex) plays a critical role in the regulation of PDHK (pyruvate dehydrogenase kinase) activity. The present study was undertaken to investigate further the molecular mechanism by which E2 modulates the activity of PDHK. In agreement with the earlier results, it was found that the inner L2 (lipoyl-bearing domain 2) of E2 expressed with or without the C-terminal hinge region had little, if any, effect on the kinase activity, indicating a lack of direct allosteric effect of L2 on PDHK. In marked contrast, significant activation of PDHK was observed with the construct consisting of L2 and the E1BD (E1-binding domain) of E2 (L2-E1BD didomain) suggesting that co-localization and/or mutual orientation of PDHK and E1, facilitated by E2 binding, largely account for the activation of PDHK by the transacetylase component. Isothermal titration calorimetry and glutathione S-transferase pull-down assays established that binding of adenyl nucleotides to the PDHK molecule facilitated the release of L2 domain. In contrast, binding of the L2 domain caused a significant decrease in the affinity of PDHK for ATP. The cross-talk in binding of adenyl nucleotides and the L2 domain to PDHK may indicate the existence of a highly integrated mechanism whereby the exchange of lipoyl-bearing domains presented to PDHK by E2 is coupled with ADP/ATP exchange.

INTRODUCTION

PDHK (pyruvate dehydrogenase kinase) is an important molecular switch that turns off the activity of mammalian PDC (pyruvate dehydrogenase complex) and thereby down-regulates the oxidation of pyruvate in mitochondria [1]. To date, studies have identified four genetically distinctive forms of PDHK [2,3] differing in tissue expression [2,4] and responsiveness to allosteric regulation [4,5].

A unique feature of this mitochondrial protein kinase is that it functions in a close association with PDC [6]. The mammalian complex consists of multiple copies of three enzymes: E1 (pyruvate dehydrogenase), E2 (dihydrolipoamide transacetylase) and E3 (dihydrolipoamide dehydrogenase). Each complex is built around a core of 60 E2 subunits [7]. Through the E1BD (E1-binding domain) of E2, 30 copies of E1 (a tetramer, α2β2) are directly attached to the core. Through a specialized non-catalytic adapter protein, E3BP (E3-binding protein component of PDC), which is tightly integrated with E2 subunits, 6–12 copies of E3 (a dimer, α2) are attached [7]. The coupling of active sites of E1, E2 and E3 components is performed by the lipoate prosthetic groups, which are covalently attached to the so-called lipoyl-bearing domains. Of these domains, two are located on E2 [the outer and inner lipoyl-bearing domains, L1 (lipoyl-bearing domain 1) and L2 respectively] and a third one L3 on E3BP [7]. On average, each mammalian complex contains 2–3 tightly bound PDHK molecules [8].

The E2 component plays a key role in the anchoring of the kinase molecule to PDC [9]. Depending on the isoenzyme, the addition of E2 brings about a 4–40-fold increase in kinase activity [5,10]. This E2-dependent activation requires the presence of lipoyl-bearing domains [11]. In addition, treatment of fully lipoylated E2 with lipoamidase greatly reduces kinase activity due, in part, to the dissociation of kinase from the complex [12]. Taken together, these results strongly suggest that the lipoyl-bearing domains of E2 are indispensable in the anchoring of PDHK to the complex and, furthermore, that docking to the lipoyl-bearing domains is responsible for setting the overall rate of kinase activity in the complex-bound state.

Despite the importance of the interaction between PDHK and lipoyl-bearing domains, the mechanism whereby docking to the lipoyl-bearing domains causes enhanced kinase functionality remains poorly understood. Initially, it was reported that the L2 domain alone could activate PDHK that suggested the existence of a direct allosteric effect of L2 on kinase activity [5]. Unexpectedly, we found that the addition of a monomeric L2 had no effect on the activities of PDHK1 (isoenzyme 1 of PDHK) and PDHK2 and had a rather small effect on the activities of PDHK3 and PDHK4 when compared with the increase caused by E2 [10]. Therefore, it was concluded that, by itself, binding to the L2 domain is insufficient for the maintenance of the enhanced kinase function observed in the presence of E2. Recently, studies of related protein kinase that regulate the activity of branched chain α-oxo acid dehydrogenase complex, identified an additional structural element of E2 that might be required for the efficient binding to E2. Chuang et al. [13] reported that the hinge region connecting the lipoyl-bearing and E1BDs of E2 along with the lipoyl-bearing domain confer efficient binding [13]. This observation brings about a possibility that E2-dependent activation of PDHK might also require an interaction with the amino acids of the hinge region. This study was undertaken to investigate further the relationship between the PDHK activity and its role in the binding to the lipoyl-bearing domain and to get a further insight into the molecular mechanisms involved in the regulation of kinase activity.

EXPERIMENTAL

Plasmids and strains

Construction of plasmids for the expression of PDHK2, the E1 component and its variants, E2 and its subdomains, lipoyl-protein ligase A (LPL_A), as well as E2–E3BP was described previously [4,1416]. Fragments of E2 cDNA encoding the L2 domain with various lengths of the C-terminal linker (L2, L2176–299 and L2176–322) were made using PCR with Pfu polymerase (Stratagene, La Jolla, CA, U.S.A.) and human E2 cDNA as a template [16]. Amplification primers carried the unique NdeI (upstream primer) and XhoI (downstream primer) restriction sites. Downstream primers also incorporated the translation termination codons. The cDNA encoding L2 corresponded to the bases 544 and 807 of human cDNA. Other fragments were as follows: L2176–299 (bases 544–897) and L2176–322 (bases 544–969). Respective cDNAs were subcloned in pUC19 (New England Biolabs, Beverly, MA, U.S.A.) and sequenced [17]. Sequence-verified plasmids were cut with NdeI and XhoI restrictase. The inserts were purified and subcloned between NdeI and XhoI sites of pET-28a (Novagen, Madison, WI, U.S.A.) producing in-frame fusion with a vector-encoded sequence of six consecutive histidine residues (His6 tag). To construct the L2 domain fused with GST (glutathione S-transferase), pET-28a-based vector for the expression of L2 was cut with NdeI restrictase and blunt-ended with a T4 DNA polymerase. L2 cDNA was cut with XhoI restrictase and ligated into pGEX-4T-1 vector (Amersham Biosciences, Piscataway, NJ, U.S.A.) cut with SmaI and XhoI.

For expression experiments, vectors directing synthesis of L2, L2176–299, L2176–322 and GST–L2 were transformed into calcium-competent BL21(DE3) (Novagen) cells along with plasmid directing the expression of bacterial LPL_A [15]. Double-transformants were selected on M9ZB agar supplemented with 50 μg/ml kanamycin or 100 μg/ml ampicillin for GST–L2 and 50 μg/ml chloramphenicol. Several individual colonies from each transformation were tested for their capability to produce significant amounts of soluble recombinant protein. Clones expressing the greatest amount of the soluble enzyme were used for further analysis. The enzymes used in this study to construct the expression vectors were purchased from New England Biolabs. The oligonucleotides were obtained from Life Technologies (Rockville, MD, U.S.A.).

Protein expression and purification

Expression and purification of PDHK2, the E1 component and its variants, E2, E2–E3BP and subdomains of E2 were described previously [4,1416]. L2, L2176–299, L2176–322 and GST–L2 constructs were expressed following a similar procedure. Briefly, expression experiments were performed using 1 litre of M9ZB media supplemented with kanamycin (or ampicillin for GST–L2) and chloramphenicol both added at the final concentration of 50 μg/ml. Cells were allowed to grow at 37 °C with constant shaking (200 rev./min) until A600 reached approx. 0.6. Expression was induced with isopropyl β-D-thiogalactoside at a final concentration of 0.4 mM for 18 h at 16 °C. Cultures also received lipoic acid before the addition of isopropyl β-D-thiogalactoside at the final concentration of 0.2 mM. By the end of the induction period, cells were harvested by centrifugation at 2744 g (JA-10 rotor) for 20 min at 4 °C. When the minimally lipoylated GST–L2 construct was produced, lipoic acid was omitted from the growth media.

Isolation of His6-tagged proteins was performed on TALON™ affinity resin (ClonTech Laboratories, Palo Alto, CA, U.S.A.) following the method previously described for the isolation of the L2 domain. GST–L2 was isolated on glutathione–Sepahrose 4B (Amersham Biosciences) following the manufacturer's instructions. The protein composition of each preparation was evaluated by SDS/PAGE analysis. Gels were stained with Coomassie R250. All preparations used in the present study were more than 90% pure. The extent of lipoylation of L2 constructs was examined following the procedure described by Quinn et al. [18]. ‘Minimally lipoylated’ constructs contained <5% of lipoylated species. The lipoate content of ‘lipoylated’ constructs was >95% (results not shown).

PDHK activity assay

Kinase activity was determined by measuring the initial rate of incorporation of 32P from [γ-32P]ATP into the E1α subunit of recombinant E1 as described elsewhere [4]. Before the assay, E1, the appropriate E2 construct and PDHK2 were reconstituted on ice. Unless specified otherwise, the final protein concentrations of the recombinant proteins in the assay were as follows: E1 and its variants at 5 μM; L1, L2, L2176–299, L2176–322, L1-L2 (didomain of L1 and L2), L2-E1BD (didomain of L2 and EIBD) and L1-L2-E1BD (tridomain of L1, L2 and E1BD) at 40 μM; E2, L2-E1BD-TR (tridomain of L2, E1BD and TR, where TR stands for transacetylase domain) and E1BD-TR (didomain of E1BD and TR) at 20 nM; and PDHK2 at 20 nM. All assays received a negative control (minus PDHK) to determine the non-specific incorporation and were performed in triplicates. Raw kinetic data were fitted and analysed using GraFit software (Erithacus Software Limited, Middlesex, U.K.). The apparent Vmax and Km values for E1 were determined by measuring the initial velocities of the phosphorylation reaction at various concentrations of E1. The activity of PDHK2 was calculated based on the incorporation of [32P]phosphate into E1 during 1 min of the reaction. The reported kinetic parameters represent the means±S.E.M. obtained for one preparation of each protein. Comparable results were obtained with at least two other preparations of each protein.

PDHK2 pull down on GST–L2 construct

Pull-down experiments were perfomed as follows: 50 μl of 50% (v/v) slurry of glutathione–Sepharose beads in PBS was placed in a Spin-X microcentrifuge filter device with pore diameter of 0.22 μm (Corning, Corning, NY, U.S.A.). Excess PBS was removed by centrifugation at 6000 g for 1 min. Beads were washed three times with 0.5 ml of buffer A [25 mM Tris/HCl, pH 8.0, 0.1 mM EDTA, 2.5 mM MgCl2, 0.1 M KCl, 5 mM dithiothreitol and 1% (v/v) glycerol]. Equilibrated beads were incubated with 0.4 ml of the GST–L2 construct (0.5 mg/ml) in buffer A for 10 min at room temperature (23 °C). Decorated beads were washed three times with 0.5 ml of buffer A and mixed with 0.4 ml of PDHK2 (0.25 mg/ml) made in buffer A. Binding was allowed to proceed for 15 min at room temperature. Unbound PDHK2 was removed by centrifugation at 6000 g for 1 min followed by three consecutive washes in buffer A (0.5 ml per wash). To elute bound proteins, 0.4 ml of buffer A containing 10 mM GSH was added to the beads for 10 min at room temperature followed by centrifugation at 6000 g for 1 min. Free and bound PDHK2 were analysed using SDS/PAGE. Gels were stained with Coomassie R250. When the effects of ATP, ADP, GTP, L2 and lipoic acid were studied, their concentrations in the binding buffer were as follows: for ATP, ADP and GTP, 0.1 mM for each of the nucleotides; for L2, 0.5 mg/ml; and for lipoic acid, 40 mM. Appropriate ligands were also present in the solutions used to remove the unbound PDHK2.

ITC (isothermal titration calorimetry)

Calorimetric measurements of the PDHK2–ATP interaction were performed on a MicroCal VP-ITC microcalorimeter (MicroCal, Northampton, MA, U.S.A.). Before the binding experiment, freshly thawed preparations of PDHK2 and L2 were desalted on a PD-10 column (Amersham Biosciences) equilibrated in buffer B [20 mM potassium phosphate buffer, pH 7.5, 50 mM KCl, 10 mM MgCl2, 5 mM dithiothreitol and 2% (v/v) ethylene glycol]. Desalted proteins were passed through a 0.2 μm filter and degassed by stirring under vacuum before use. Experiments were performed at 15±0.2 °C. The sample cell was filled with PDHK2 solution (20 μM) and while being stirred at 250 rev./min, the system was allowed to equilibrate to an r.m.s noise value of 0.008 J· s−1 before the start of titration. The injection syringe was filled with ATP (1.0 mM) solution made in buffer A and repeated 5 μl injections were made. ATP concentration was checked spectrophotometrically using a molar absorption coefficient ε260 of 15.4× 103 cm2·mol−1. The calorimeter was calibrated using standard electrical pulses generated by the software provided. Calorimetric results were analysed by integration of the resultant peaks (ORIGIN software; OriginLab, Northampton, MA, U.S.A.). The heat change accompanying the addition of buffer to PDHK2 and the heat of dilution of the ligands were subtracted from the raw results after correction for the injection signal of buffer into buffer. The equilibrium association constant of the ligand KA, the molar heat of binding ΔH° and the concentration of binding sites (stoichiometry) were obtained by non-linear regression fitting to the isotherm. A single-site model, a two non-interacting sites' model and a two interacting sites' model were used to analyse the results. To determine the effect of L2 on ATP binding, both cell and injectant solutions received L2 in the final concentration of 40 μM. Three repetitions of each experiment were performed to ensure reproducibility. Representative samples are shown in the Figure 5.

Other procedures

SDS/PAGE was performed according to Laemmli's method [19]. Protein concentrations were determined according to Lowry with BSA as a standard [20].

RESULTS

Regulation of PDHK activity by various constructs derived from the E2 component

Numerous studies had established that the addition of E2–E3BP subcomplex greatly enhances the rate at which PDHK phosphorylates E1 [5,10,21,22]. Initially, this observation was interpreted as an indication that E2–E3BP-dependent activation of PDHK stems from the increase in productive encounters between PDHK and E1 caused by their binding to E2–E3BP [21,22]. Recently, an alternative explanation for this effect was suggested by Baker et al. [5]. The authors proposed that the increase in PDHK activity in the complex-bound state reflects a direct activation of PDHK caused by its binding to the L2 domain of E2, which would require the L2 domain to induce some yet unidentified structural changes in PDHK that accelerate the rate-limiting step in PDHK catalysis [5]. To examine these hypotheses further, we studied the effects of several engineered forms of the E2 component on the activity of PDHK. The following E2 constructs were tested during this investigation (Figure 1A): full-length E2 lacking E3BP (E2); E2 lacking the outer lipoyl-bearing domain (L2-E1BD-TR tridomain); E2 lacking the outer and the inner lipoyl-bearing domains (E1B-TR didomain); the outer and the inner lipoyl-bearing domains expressed individually (L1 and L2 respectively); the didomain comprised of the outer and the inner lipoyl-bearing domains (L1-L2 didomain); the didomain of the inner lipoyl-bearing domain and the E1BD (L2-E1BD didomain); and the tridomain consisting of the outer and the inner lipoyl-bearing domains along with the E1BD (L1-L2-E1BD tridomain). Among these constructs, E2, L2-E1BD-TR and E1B-TR self-assembled into a 60-meric core characteristic of the transacetylase component, whereas L1, L2, L1-L2, L2-E1BD and L1-L2-E1BD existed as monomeric proteins [15].

Regulation of PDHK2 activity by various constructs derived from the E2 subunit

Figure 1
Regulation of PDHK2 activity by various constructs derived from the E2 subunit

(A) Schematic representation of the domain structure of E2 constructs used in this study. (B) Effect of various E2 constructs on PDHK2 activity. Kinase assays were performed as described in the Experimental section. The activities were expressed as a percentage of activity determined in the presence of E2.

Figure 1
Regulation of PDHK2 activity by various constructs derived from the E2 subunit

(A) Schematic representation of the domain structure of E2 constructs used in this study. (B) Effect of various E2 constructs on PDHK2 activity. Kinase assays were performed as described in the Experimental section. The activities were expressed as a percentage of activity determined in the presence of E2.

Representative results showing the effects of E2-derived constructs on kinase activity obtained for the isoenzyme PDHK2 are summarized in Figure 1(B). It was found that the full-length E2 and L2-E1B-TR could support the enhanced PDHK2 activity similar to that of fully assembled E2–E3BP subcomplex, indicating that neither E3BP nor the L1 domain were required for PDHK2 activation. In contrast, the activity of PDHK2 determined in the presence of E1B-TR (E2 lacking the lipoyl-bearing domains) was similar to that determined with free E1. Similarly, L1 and L2 domains alone as well as L1-L2 didomain did not display an appreciable effect on PDHK2 activity. A significant increase in kinase activity was observed on the addition of L2-E1B or L1-L2-E1BD (Figure 1B) with monomeric L2-E1BD didomain being the minimal structure capable of supporting the enhanced kinase activity of PDHK2. Qualitatively similar results were obtained with other isoenzymes of PDHK (results not shown), except that PDHK3 displayed a slight increase in activity in the presence of the L2 domain, which probably reflected a stabilizing effect exerted by L2 on the activity of this particular isoenzyme [10].

A marked increase in PDHK activity caused by L2-E1BD and L1-L2-E1BD along with the lack of effects of L1, L2, L1-L2 and E1BD-TR observed in this study supported the idea of co-localization of PDHK and E1 by E2 as a driving force responsible for the increase in kinase activity in a complex-bound state. However, these results did not completely exclude the possibility that the direct activation of PDHK by L2 could require an additional structural element of E2 that was missing in our L1-L2, L2 and E1BD-TR constructs, namely, the hinge region connecting the L2 and E1BD. To investigate this hypothesis, two additional L2 constructs differing in the length of the C-terminal hinge region were made. The first construct (L2176–299) harboured approximately half of the amino acids of the hinge region, whereas the second construct (L2176–322) carried the entire hinge region (Figure 2A). Respective constructs were produced in a bacterial expression system and purified to near homogeneity using metal affinity chromatography. The resulting recombinant L2 proteins were tested for their capability to stimulate PDHK2 activity in a standard phosphorylation assay. In control experiments, the measurements were conducted with E2–E3BP subcomplex as a source of lipoyl-bearing domains. As shown in Figure 2(B), neither L2176–299 nor L2176–322 construct displayed an appreciable effect on the activity of PDHK2. Qualitatively similar results were obtained for three other isoenzymes of PDHK (results not shown), which indicated that the amino acids of the C-terminal hinge region are not essential for the maintenance of enhanced kinase function. Taken together, the results reported in this paper strongly suggested that the co-localization and mutual orientation of PDHK and E1, facilitated by E2 binding, accounted for the large proportion of E2-dependent increase in PDHK activity.

PDHK2 activity in the presence of various constructs of the L2 domain

Figure 2
PDHK2 activity in the presence of various constructs of the L2 domain

(A) Schematic representation of the amino acid sequence of dihydrolipoamide transacetylase (E2) within the region corresponding to the lipoyl-bearing and E1BD. The elements of the secondary structure of L2 domain [32] determined by NMR are shown above the alignment. The size of C-terminal hinge region present in various constructs of the L2 domain is indicated by open triangles. (B) PDHK2 activity in the presence of L2 constructs containing different lengths of the C-terminal hinge region. Kinase activities were expressed as a percent of activity determined in the presence of E2–E3BP complex.

Figure 2
PDHK2 activity in the presence of various constructs of the L2 domain

(A) Schematic representation of the amino acid sequence of dihydrolipoamide transacetylase (E2) within the region corresponding to the lipoyl-bearing and E1BD. The elements of the secondary structure of L2 domain [32] determined by NMR are shown above the alignment. The size of C-terminal hinge region present in various constructs of the L2 domain is indicated by open triangles. (B) PDHK2 activity in the presence of L2 constructs containing different lengths of the C-terminal hinge region. Kinase activities were expressed as a percent of activity determined in the presence of E2–E3BP complex.

L1-L2-E1BD tridomain mediated activation of PDHK2

Earlier, it was shown that binding to E2 has a 2-fold effect on the kinetics of phosphorylation reaction catalysed by PDHK: it brings an approx. 33-fold decrease in the apparent Km value for E1 and an approx. 2-fold increase in the apparent Vmax value [23]. To evaluate the mechanism of action of monomeric constructs derived from the E2 subunit, we examined the effects of L1-L2-E1BD tridomain on the kinetics of phosphorylation reaction catalysed by PDHK2. It was found that variation of the level of L1-L2-E1BD in the reaction cocktail from 2.5 to 30 μM caused a hyperbolic increase in kinase activity with EC50 of approx. 16±2 μM (Figure 3A). Under similar conditions, variation in the level of L2 over the same range of concentrations caused little, if any, activation of PDHK2. Examination of the kinetics of phosphorylation reaction with variations in the E1 concentration from 1.25 to 40 μM at a single fixed concentration of L1-L2-E1BD of 40 μM revealed a significant approx. 2.5-fold increase in the apparent Vmax value (Figure 3B). The apparent Km value for E1 was largely unchanged (31±4 versus 36±1 μM). This outcome suggested that the activation of PDHK2 by L1-L2-E1BD tridomain was achieved through the enhancement of catalytic efficiency of PDHK2 and, therefore, only partially mimicked the action of fully assembled, 60-meric E2, which exerts both Km and Vmax effects.

Regulation of PDHK2 activity by L1-L2-E1BD tridomain

Figure 3
Regulation of PDHK2 activity by L1-L2-E1BD tridomain

(A) Effect of various concentrations of L1-L2-E1BD tridomain on PDHK2 activity determined at a single fixed concentration of ATP (200 μM) and E1 (20 μM) (●). The activities of PDHK2 in the presence of various concentrations of L2176–268 construct (○). (B) Effect of L1-L2-E1BD tridomain (20 μM) on phosphorylation of various concentrations of E1 by PDHK2 (●). PDHK2 activities in the absence of tridomain are shown by (○). (C) Effects of L1-L2-E1BD tridomain and E2–E3BP complex on phosphorylation of E1 with ablated phosphorylation site 1 (S2,3E1) or site 2 (S1,3E1).

Figure 3
Regulation of PDHK2 activity by L1-L2-E1BD tridomain

(A) Effect of various concentrations of L1-L2-E1BD tridomain on PDHK2 activity determined at a single fixed concentration of ATP (200 μM) and E1 (20 μM) (●). The activities of PDHK2 in the presence of various concentrations of L2176–268 construct (○). (B) Effect of L1-L2-E1BD tridomain (20 μM) on phosphorylation of various concentrations of E1 by PDHK2 (●). PDHK2 activities in the absence of tridomain are shown by (○). (C) Effects of L1-L2-E1BD tridomain and E2–E3BP complex on phosphorylation of E1 with ablated phosphorylation site 1 (S2,3E1) or site 2 (S1,3E1).

As first established by Yeaman et al. [24], the E1 component has three phosphorylation sites (site 1, Ser-264; site 2, Ser-271; and site 3, Ser-203) with site 1 being primarily responsible for the regulation of pyruvate dehydrogenase activity in vivo. Subsequently, it was shown that the complex-bound PDHK2 can phosphorylate two out of three sites (sites 1 and 2), whereas the free PDHK2 phosphorylates the uncomplexed E1 almost exclusively at site 1 [14,25]. In the present study, to examine the site specificity of L1-L2-E1BD-bound PDHK2, we used E1 variants with ablated site 1 or 2 (S2,3E1 and S1,3E1 respectively) [14]. As shown in Figure 3(C), addition of L1-L2-E1BD tridomain caused a profound increase in the rate of phosphorylation of site 1, which was comparable with the effect of the fully assembled transacetylase component (Figure 3C, left panel) suggesting that co-localization of PDHK2 and E1 by L1-L2-E1BD is sufficient to provide effective access for the kinase to the primary phosphorylation site of E1. Somewhat unexpectedly it was found that the effect of L1-L2-E1BD on the rate of phosphorylation of site 2 was significantly lower than that of E2–E3BP (Figure 3C, right panel). The latter observation might indicate that the 60-meric assembly of E2 is still required in order for PDHK2 to gain access to the additional phosphorylation site of E1.

Effect of adenyl nucleotides on the binding of PDHK2 to the L2 domain

Recent structural studies on mitochondrial protein kinases revealed that binding of adenyl nucleotides is associated with a significant interdomain movement [26,27]. This observation brings about an intriguing possibility that there might be a cross-talk between the binding of adenyl nucleotides and the binding of lipoyl-bearing domain to the kinase molecule. To explore this hypothesis, we developed a new pull-down assay in which an L2 domain fused with GST is used as a ‘bait’ to catch PDHK2. As shown in Figure 4(A) (lanes 1 and 2), GST–L2 could efficiently bind recombinant PDHK2. Binding was strictly specific for the lipoyl-bearing domain part of the fusion as evidenced by the competition with free, monomeric L2 (Figure 4A, lanes 3 and 4). As would be expected from the previously published results [10,12], the interaction between L2 and PDHK2 was dependent on the proper lipoylation of L2. The minimally lipoylated GST–L2 did not display an appreciable binding of PDHK2 in this assay system (Figure 4A, lanes 5 and 6). Furthermore, free lipoic acid competed with L2 for binding (Figure 4A, lanes 6 and 7).

Pull-down of wild-type PDHK2 on GST–L2 construct

Figure 4
Pull-down of wild-type PDHK2 on GST–L2 construct

(A) Lanes 1, 3, 5 and 7, free PDHK2; lanes 2, 4, 6 and 8, GST–L2-bound PDHK2; lanes 1 and 2, control; lanes 3 and 4, effect of monomeric L2; lanes 5 and 6, effect of minimally lipoylated GST–L2; lanes 7 and 8, effect of 40 mM lipoic acid; lane 9, position of protein standards. (B) Lanes 1, 3, 5 and 7, free PDHK2; lanes 2, 4, 6 and 8, GST–L2-bound PDHK2; lanes 1 and 2, control; lanes 3 and 4, effect of 100 μM ATP; lanes 5 and 6, effect of 100 μM ADP; lanes 7 and 8, effect of 100 μM GTP; lane 9, position of protein standards.

Figure 4
Pull-down of wild-type PDHK2 on GST–L2 construct

(A) Lanes 1, 3, 5 and 7, free PDHK2; lanes 2, 4, 6 and 8, GST–L2-bound PDHK2; lanes 1 and 2, control; lanes 3 and 4, effect of monomeric L2; lanes 5 and 6, effect of minimally lipoylated GST–L2; lanes 7 and 8, effect of 40 mM lipoic acid; lane 9, position of protein standards. (B) Lanes 1, 3, 5 and 7, free PDHK2; lanes 2, 4, 6 and 8, GST–L2-bound PDHK2; lanes 1 and 2, control; lanes 3 and 4, effect of 100 μM ATP; lanes 5 and 6, effect of 100 μM ADP; lanes 7 and 8, effect of 100 μM GTP; lane 9, position of protein standards.

Having established a new pull-down assay, we investigated the effect of adenyl nucleotides on the interactions between PDHK2 and the L2 domain. As shown in Figure 4(B), both ATP (lanes 3 and 4) and ADP (lanes 5 and 6) tested at a single fixed concentration of 100 μM caused a marked decrease in the amount of PDHK2 recovered in L2-bound form. ATP was found to be more efficient than ADP, probably due to the greater affinity PDHK2 has for this nucleotide [4]. This idea was supported by the observation that the amount of L2-bound PDHK2 was progressively reduced with an increase in concentration of ADP in the binding buffer (results not shown). To explore further the nucleotide specificity, similar pull-down experiments were performed with 100 μM GTP, which does not serve as a substrate for PDHK2 [27] (Figure 4B, lanes 7 and 8). In comparison with the control (Figure 4B, lanes 1 and 2), GTP did not cause any reduction in the amount of L2-bound PDHK2. Taken together, these results strongly suggested that the binding of adenyl nucleotides could modulate the interaction between PDHK2 and L2, presumably by decreasing the affinity of PDHK2 for the L2 domain.

ITC

Considering that ATP causes a decrease in PDHK2 affinity for the L2 domain, we reasoned that L2 could, in turn, affect the interaction between the kinase molecule and ATP. To test this hypothesis, we used ITC. Representative results of an ITC run in which PDHK2 was titrated with ATP are shown in Figure 5(A). The upper panel of the figure shows the raw heat results from the run. The lower panel shows the enthalpy changes as a function of the ratio of ATP to PDHK2. The heat of ATP dilution was measured by titration of ATP into the buffer alone. This value was generally small and was subtracted from each binding titration curve. The interaction was exothermic as evidenced by the negative peaks. The heat released during the titration of ATP into a PDHK2 solution exhibited good agreement with ideal binding, indicating the presence of a single type of binding site and the lack of co-operativity in the interaction. On an average, the fit of the ATP binding data resulted in a value for n of 1.62 per PDHK2 dimer, a binding constant KA of 8.6×105 M−1 and an enthalpy change ΔH° of −20.5 kcal·mol−1 (1 cal=4.184 J), which is consistent with the interpretation that each PDHK2 dimer in solution has two equivalent, non-interacting nucleotide-binding sites.

ITC of ATP binding to PDHK2

Figure 5
ITC of ATP binding to PDHK2

(A) Upper panel: titration of PDHK2 with ATP alone, showing the calorimetric response as successive injections of ligand are added to the reaction cell. Lower panel: binding isotherm of the calorimetric titration shown in the upper panel. The continuous line represents the least-squares fit of the data to a single-site binding model. (B) Upper panel: titration of PDHK2 with ATP in the presence of 40 μM L2 construct, showing the calorimetric response as successive injections of ligand are added to the reaction cell. Lower panel: binding isotherm of the calorimetric titration shown in the upper panel. The continuous line represents the least-squares fit of the results to a single-site binding model.

Figure 5
ITC of ATP binding to PDHK2

(A) Upper panel: titration of PDHK2 with ATP alone, showing the calorimetric response as successive injections of ligand are added to the reaction cell. Lower panel: binding isotherm of the calorimetric titration shown in the upper panel. The continuous line represents the least-squares fit of the data to a single-site binding model. (B) Upper panel: titration of PDHK2 with ATP in the presence of 40 μM L2 construct, showing the calorimetric response as successive injections of ligand are added to the reaction cell. Lower panel: binding isotherm of the calorimetric titration shown in the upper panel. The continuous line represents the least-squares fit of the results to a single-site binding model.

To examine the effect of PDHK2–L2 interaction on the nucleotide-binding sites of PDHK2, we performed titration experiments in which both cell and syringe solutions were supplemented with the L2 domain used at the final concentration of 40 μM. The cell solution also received 20 μM PDHK2, whereas the syringe solution contained 1.0 mM ATP. Figure 5(B) (upper panel) shows typical injection heat of dilution for ATP binding to PDHK2 in the presence of L2. The integrated heats for each injection versus the molar ratio of ATP to PDHK2 after subtraction of the heat of dilution of the ligand are illustrated in Figure 5(B, lower panel). In the presence of L2, the enthalpic binding isotherms for ATP significantly deviated from the isotherms obtained for free PDHK2 showing a delayed saturation. The best fit of the experimental results could be obtained assuming the presence of a single type of binding site and the lack of co-operativity in the interaction. On an average, the binding event was calculated to have an association constant KA of 7.92×104 M−1, an enthalpy change ΔH° of −16.8 kcal·mol−1, and a stoichiometry of 1.3. These results suggested that the binding of L2 to PDHK2 caused a conformational change in the structure of the kinase molecule, which brought an approx. 10-fold decrease in the kinase affinity for ATP.

DISCUSSION

Pioneering studies in Lester Reed's laboratory demonstrated that E2 is the major protein factor responsible for anchoring of PDHK to the complex [9]. Binding to E2 is associated with a several fold increase in PDHK activity [5,10] and also provides means for the allosteric control by a number of metabolites (for a review see [28]). Evidence presented by Roche and co-workers [29,30] suggests that the binding sites for the kinase molecule are provided by the lipoyl-bearing domains of the E2 component with the L2 domains being the primary docking sites. The present study was undertaken to evaluate further the consequences of the interaction between PDHK2 and L2. In good agreement with earlier results, we show that PDHK2 recognizes and binds the L2 domain itself [10]. This interaction requires proper lipoylation of L2, indicating that the lipoate prosthetic group is one of the major determinants recognized by PDHK2. Importantly, binding to L2 has no direct effect on kinase activity, suggesting that the interaction with L2 is required but is not sufficient for the maintenance of enhanced kinase functionality. Furthermore, variation in the length of the linker connecting L2 and E1BDs, which is supposed to be important for the binding of related branched chain α-oxo acid dehydrogenase kinase [13], appears to have little, if any, effect on PDHK2 activity. This is consistent with the interpretation that these two protein kinases interact with their cognate E2 components somewhat differently. A marked increase in PDHK activity is observed during this study with a construct consisting of the L2 and E1BDs connected by a flexible linker (L2-E1BD didomain). This observation provides the first direct evidence that the co-localization and mutual orientation of PDHK2 and E1 caused by their binding to the L2 and E1BDs respectively account for most of the effect of E2. This interpretation is supported by our results showing that the 60-meric E1BD-TR didomain construct does not affect kinase functionality. The latter observation indicates that the interaction between E1 and E1BD does not make E1 a better substrate for PDHK2. Therefore the effect of the L2-E1BD didomain stems strictly from the co-localization and mutual orientation of PDHK2 and E1 caused by their binding to L2-E1BD.

PDC is a large enzyme comparable in size with the prokaryotic ribosome [7]. On an average, each complex contains two or three PDHK molecules tightly bound to the lipoyl-bearing domains of the transacetylase component [8]. This arrangement creates a mechanistic problem because, to phosphorylate the numerous copies of the E1 component spread around the E2 core, the kinase must be capable of physically moving around the core. However, PDHK has been shown to bind to the L2 domain tightly [10,31], which, theoretically, should immobilise the kinase molecule to a particular surface of the core, thereby preventing its movement. In this context, our results showing the existence of a cross-talk between the binding of adenyl nucleotides and the L2 domain to PDHK2 molecule appears to be particularly intriguing. According to our results, binding of the nucleotide substrate is associated with a decrease in the affinity of PDHK2 for the lipoyl-bearing domain, whereby binding of the lipoyl-bearing domain is associated with a decrease in the affinity of PDHK2 for nucleotide. These observations may indicate the existence of a tightly integrated mechanism whereas the binding of ATP to one PDHK protomer promotes the release of the lipoyl-bearing domain, thereby facilitating the exchange of lipoyl-domains presented to PDHK molecule by the E2 component. In turn, binding of the second PDHK protomer to the neighbouring lipoyl-bearing domain may promote the dissociation of ADP, thereby facilitating nucleotide exchange. Studies to test this hypothesis are currently underway.

This work was supported by grants GM 51262 and DK 56898 from the U.S. Public Health Services.

Abbreviations

     
  • E1BD

    E1-binding domain

  •  
  • E3BP

    E3-binding protein

  •  
  • GST

    glutathione S-transferase

  •  
  • ITC

    isothermal titration calorimetry

  •  
  • L1

    lipoyl-bearing domain 1

  •  
  • PDC

    pyruvate dehydrogenase complex

  •  
  • PDHK

    pyruvate dehydrogenase kinase

  •  
  • TR

    transacetylase domain

References

References
1
Holness
M. J.
Sugden
M. C.
Regulation of pyruvate dehydrogenase complex activity by reversible phosphorylation
Biochem. Soc. Trans.
2003
, vol. 
31
 (pg. 
1143
-
1151
)
2
Gudi
R.
Bowker-Kinley
M. M.
Kedishvili
N. Y.
Zhao
Y.
Popov
K. M.
Diversity of the pyruvate dehydrogenase kinase gene family in humans
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
28989
-
28994
)
3
Rowles
J.
Scherer
S. W.
Xi
T.
Majer
M.
Nickle
D. C.
Rommens
J. M.
Popov
K. M.
Harris
R. A.
Riebow
N. L.
Xia
J.
, et al. 
Cloning and characterization of PDHK4 on 7q21.3 encoding a fourth pyruvate dehydrogenase kinase isoenzyme in human
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
22376
-
22382
)
4
Bowker-Kinley
M. M.
Davis
W. I.
Wu
P.
Harris
R. A.
Popov
K. M.
Evidence for existence of tissue-specific regulation of the mammalian pyruvate dehydrogenase complex
Biochem. J.
1998
, vol. 
329
 (pg. 
191
-
196
)
5
Baker
J. C.
Yan
X.
Peng
T.
Kasten
S.
Roche
T. E.
Marked differences between two isoforms of human pyruvate dehydrogenase kinase
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
15773
-
15781
)
6
Linn
T. C.
Pettit
F. H.
Reed
L. J.
α-Keto acid dehydrogenase complexes, X. Regulation of the activity of the pyruvate dehydrogenase complex from beef kidney mitochondria by phosphorylation and dephosphorylation
Biochemistry
1969
, vol. 
62
 (pg. 
234
-
241
)
7
Zhou
Z. H.
McCarthy
D. B.
O'Connor
C. M.
Reed
L. J.
Stoops
J. K.
The remarkable structural and functional organization of the eukaryotic pyruvate dehydrogenase complexes
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
14802
-
14807
)
8
Patel
M. S.
Roche
T. E.
Molecular biology and biochemistry of pyruvate dehydrogenase complexes
FASEB J.
1990
, vol. 
4
 (pg. 
3224
-
3233
)
9
Linn
T. C.
Pelley
J. W.
Pettit
F. H.
Hucho
F.
Randall
D. D.
Reed
L. J.
α-Keto acid dehydrogenase complexes, XV. Purification and properties of the component enzymes of the pyruvate dehydrogenase complexes from bovine kidney and heart
Arch. Biochem. Biophys.
1972
, vol. 
148
 (pg. 
327
-
342
)
10
Tuganova
A.
Boulatnikov
I.
Popov
K. M.
Interaction between the individual isoenzymes of pyruvate dehydrogenase kinase and the inner lipoyl-bearing domain of transacetylase component of pyruvate dehydrogenase complex
Biochem. J.
2002
, vol. 
366
 (pg. 
129
-
136
)
11
Ono
K.
Radke
G. A.
Roche
T. E.
Partial activation of the pyruvate dehydrogenase kinase by lipoyl domain region of E2 and interchange of the kinase between lipoyl domain regions
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
26135
-
26143
)
12
Radke
G. A.
Ono
K.
Ravindran
S.
Roche
T. E.
Critical role of a lipoyl cofactor in the binding and enhanced function of the pyruvate dehydrogenase kinase
Biochem. Biophys. Res. Commun.
1993
, vol. 
190
 (pg. 
982
-
991
)
13
Chuang
J. L.
Wynn
R. M.
Chuang
D. T.
The C-terminal hinge region of lipoic acid-bearing domain of E2b is essential for domain interaction with branched-chain alpha-keto acid dehydrogenase kinase
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
36905
-
36908
)
14
Kolobova
E.
Tuganova
A.
Boulatnikov
I.
Popov
K. M.
Regulation of pyruvate dehydrogenase activity through phosphorylation at multiple sites
Biochem. J.
2001
, vol. 
358
 (pg. 
69
-
77
)
15
Karpova
T.
Danchuk
S.
Kolobova
E.
Popov
K. M.
Characterization of the isozymes of pyruvate dehydrogenase phosphatase: implications for the regulation of pyruvate dehydrogenase activity
Biochim. Biophys. Acta
2003
, vol. 
1652
 (pg. 
126
-
135
)
16
Harris
R. A.
Bowker-Kinley
M. M.
Wu
P.
Jeng
J.
Popov
K. M.
Dihydrolipoamide dehydrogenase-binding protein of the human pyruvate dehydrogenase complex. DNA-derived amino acid sequence, expression, and reconstitution of the pyruvate dehydrogenase complex
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
19746
-
19751
)
17
Sanger
F.
Nicklen
S.
Coulson
A. R.
DNA sequencing with chain-terminating inhibitors
Proc. Natl. Acad. Sci. U.S.A.
1977
, vol. 
74
 (pg. 
5463
-
5467
)
18
Quinn
J.
Diamond
A. G.
Masters
A. K.
Brookfield
D. E.
Wallis
N. G.
Yeaman
S. J.
Expression and lipoylation in Escherichia coli of the inner lipoyl domain of the E2 component of the human pyruvate dehydrogenase complex
Biochem. J.
1993
, vol. 
289
 (pg. 
81
-
85
)
19
Laemmli
U. K.
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
Nature (London)
1970
, vol. 
227
 (pg. 
680
-
685
)
20
Lowry
O. H.
Rosebrough
N. J.
Farr
A. L.
Randall
R. J.
Protein measurement with the folin phenol reagent
J. Biol. Chem.
1951
, vol. 
193
 (pg. 
265
-
275
)
21
Linn
T. C.
Pettit
F. H.
Hucho
F.
Reed
L. J.
α-Keto acid dehydrogenase complexes, XI. Comparative studies of regulatory properties of the pyruvate dehydrogenase complexes from kidney, heart, and liver mitochondria
Proc. Natl. Acad. Sci. U.S.A.
1969
, vol. 
64
 (pg. 
227
-
234
)
22
Davis
P. F.
Pettit
F. H.
Reed
L. J.
Peptides derived from pyruvate dehydrogenase as substrates for pyruvate dehydrogenase kinase and phosphatase
Biochem. Biophys. Res. Commun.
1977
, vol. 
75
 (pg. 
541
-
549
)
23
Hucho
F.
Randall
D. D.
Roche
T. E.
Burgett
M. W.
Pelley
J. W.
Reed
L. J.
α-Keto acid dehydrogenase complexes, XVII. Kinetic and regulatory properties of pyruvate dehydrogenase kinase and pyruvate dehydrogenase phosphatase from bovine kidney and heart
Arch. Biochem. Biophys.
1972
, vol. 
151
 (pg. 
328
-
340
)
24
Yeaman
S. J.
Hutcheson
E. T.
Roche
T. E.
Pettit
F. H.
Brown
J. R.
Reed
L. J.
Watson
D. C.
Dixon
G. H.
Sites of phosphorylation on pyruvate dehydrogenase from bovine kidney and heart
Biochemistry
1978
, vol. 
17
 (pg. 
2364
-
2370
)
25
Korotchkina
L. G.
Patel
M. S.
Site specificity of four pyruvate dehydrogenase kinase isoenzymes toward the three phosphorylation sites of human pyruvate dehydrogenase
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
37223
-
37229
)
26
Machius
M.
Chuang
J. L.
Wynn
R. M.
Tomchick
D. R.
Chuang
D. T.
Structure of rat BCKD kinase: nucleotide-induced domain communication in a mitochondrial protein kinase
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
11218
-
11223
)
27
Steussy
C. N.
Popov
K. M.
Bowker-Kinley
M. M.
Sloan
R. B.
Jr
Harris
R. A.
Hamilton
J. A.
Structure of pyruvate dehydrogenase kinase. Novel folding pattern for a serine protein kinase
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
37443
-
37450
)
28
Roche
T. E.
Baker
J. C.
Yan
X.
Hiromasa
Y.
Gong
X.
Peng
T.
Dong
J.
Turkan
A.
Kasten
S. A.
Distinct regulatory properties of pyruvate dehydrogenase kinase and phosphatase isoforms
Prog. Nucleic Acid Res. Mol. Biol.
2001
, vol. 
70
 (pg. 
33
-
75
)
29
Liu
S.
Baker
J. C.
Roche
T. E.
Binding of the pyruvate dehydrogenase kinase to recombinant constructs containing the inner lipoyl domain of the dihydrolipoyl acetyltransferase component
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
793
-
800
)
30
Yang
D.
Gong
X.
Yakhnin
A.
Roche
T. E.
Requirements for the adaptor protein role of dihydrolipoyl acetyltransferase in the up-regulated function of the pyruvate dehydrogenase kinase and pyruvate dehydrogenase phosphatase
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
14130
-
14137
)
31
Hiromasa
Y.
Roche
T. E.
Facilitated interaction between the pyruvate dehydrogenase kinase isoform 2 and the dihydrolipoyl acetyltransferase
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
33681
-
33693
)
32
Howard
M. J.
Fuller
C.
Broadhurst
R. W.
Perham
R. N.
Tang
J. G.
Quinn
J.
Diamond
A. G.
Yeaman
S. J.
Three-dimensional structure of the major autoantigen in primary biliary cirrhosis
Gastroenterology
1998
, vol. 
115
 (pg. 
139
-
146
)