Properties of a newly characterized protein of the bovine kidney pyruvate dehydrogenase complex

The dihydrolipoyl transacetylase component, which serves as the structural core of mammalian pyruvate dehydrogenase complexes, is acetylated when treated with either pyruvate or with acetyl-CoA in the presence of NADH. Besides the dihydrolipoyl transacetylase component, we have found that another protein, referred to as protein X, is rapidly acetylated at thiol residues. Protein X remains fully bound to the transacetylase core under conditions that remove the pyruvate dehydrogenase and dihydrolipoyl dehydrogenase components. Mapping of 125I-tryptic peptides indicated that the transacetylase subunits and protein X are structurally distinct; however, under the same mapping conditions, there is considerable similarity in the positions of acetylated peptides derived from these subunits. Affinity-purified rabbit immunoglobulin G prepared against the dihydrolipoyl transacetylase core reacted exclusively with the transacetylase and with both its tryptic-derived inner domain and outer lipolyl-bearing domain. Those results further indicate that protein X is not derived from the transacetylase subunit Affinity-purified mouse antibody to protein X reacted selectively with large tryptic polypeptides derived from protein X and did not react with the inner domain of the transacetylase. However, the anti-protein X antibody did react with the intact transacetylase subunit, the lipoyl-bearing domain of the transacetylase, and weakly with the transsuccinylase component of the alpha-ketoglutarate dehydrogenase complex. This cross-reactivity reflected specificity of a portion of the polyclonal antibodies for a related structural region in the transacetylase and protein X (possibly a similar lipoyl-bearing region). Furthermore, a major portion of that polyclonal antibody was shown to react exclusively with protein X. Thus, protein X subunits differ substantially from transacetylase subunits but the two components have a region of structural similarity. We estimate that there are about 5 mol of protein X per mol of the kidney pyruvate dehydrogenase complex. Under a variety of conditions that result in a wide range of levels of acetylation of sites in the complex, about 1 acetyl group is incorporated into protein X per 10 acetyl groups incorporated into the transacetylase subunits per mol of complex. That ratio is close to the ratio of protein X subunits of transacetylase subunits in the complex, indicating that there are efficient mechanisms for acylation and deacylation of protein X.     

Properties of the pyruvate dehydrogenase kinase bound to and separated from the dihydrolipoyl transacetylase-protein X subcomplex and evidence for binding of the kinase to protein X

Studies were conducted on four pyruvate dehydrogenase kinase-containing fractions: purified pyruvate dehydrogenase complex, the dihydrolipoyl transacetylase-protein X-kinase subcomplex (E2.X.K), a kinase fraction (K fraction) prepared from the E2.X.K subcomplex, and a kinase fraction generated by limited trypsin-digestion of E2.X.K. We characterized the gel electrophoresis properties of dissociated subunits (one-dimensional and two-dimensional), the catalytic and ATP binding properties of kinase-containing fractions, and the subunit requirements for kinase binding to and being activated by the transacetylase-protein X subcomplex (E2.X). A significant portion of protein X was retained with the transacetylase core following release of virtually all the kinase. The K fraction had four major bands separated by sodium dodecyl sulfate-slab gel electrophoresis which corresponded to the dihydrolipoyl dehydrogenase, protein X, the trypsin-resistant catalytic subunit of the kinase and a chymotrypsin-resistant subunit which had a high pI and comigrated in one-dimensional systems with the chymotrypsin-sensitive alpha-subunit of the pyruvate dehydrogenase component. While purified kidney complex contained only about three molecules of kinase (determined by [14C]ATP binding), one molecule of E2.X subcomplex activated a large number (greater than 15) molecules of kinase associated with the protein X-containing K fraction. Sephadex G-200 chromatography of the K fraction in the presence of dithiothreitol led to coelution of protein X and kinase subunits. Limited trypsin digestion converted the transacetylase into subdomains and cleaved protein X and the high pI subunit of the kinase. Under those conditions, the intact catalytic subunit of the kinase did not bind to the large inner domain of the transacetylase but could be activated by untreated E2.X subcomplex. Thus, binding of the catalytic subunit of the kinase and its activation by E2.X required either protein X or the lipoyl-bearing outer domain of the transacetylase. In combination, our results suggest that protein X serves to anchor the kinase to the core of the complex.     

Separation of protein X from the dihydrolipoyl transacetylase component of the mammalian pyruvate dehydrogenase complex and function of protein X

The dihydrolipoyl transacetylase (E2)-protein X-kinase subcomplex was resolved to produce an oligomeric transacetylase that was free of protein X and kinase subunits. We investigated the properties of this transacetylase E2 oligomer and of a form of the subcomplex from which only the lipoyl-bearing domain of protein X (XL) was removed. While retaining other catalytic and binding properties of the native subcomplex, the oligomeric transacetylase and the subcomplex lacking the XL domain had greatly reduced capacities both to support the overall reaction of the complex (upon reconstitution with other components) and to bind the dihydrolipoyl dehydrogenase component. Our results indicate that protein X, in part through its XL domain, contributes to the binding of the dihydrolipoyl dehydrogenase component and to the overall reaction of the complex.     

Additional binding sites for the pyruvate dehydrogenase kinase but not for protein X in the assembled core of the mammalian pyruvate dehydrogenase complex: binding region for the kinase.

A standard resolution of the bovine kidney pyruvate dehydrogenase complex yields a subcomplex composed of approximately 60 dihydrolipoyl transacetylase (E2) subunits, approximately 6 protein X subunits, and approximately 2 pyruvate dehydrogenase kinase heterodimers (KcKb). Using a preparation of resolved kinase in which Kc much greater than Kb, E2-X-KcKb subcomplex additionally bound at least 15 catalytic subunits of the kinase (Kc) and a much lower level of Kb. The binding of Kc to E2 greatly enhanced kinase activity even at high levels of bound kinase. Free protein X, functional in binding the E3 component, did not bind to E2-X-KcKb subcomplex. This pattern of binding Kc but not protein X was unchanged either with a preparation of E2 oligomer greatly reduced in protein X or with subcomplex from which the lipoyl domain of protein X was selectively removed. The bound inner domain of protein X associated with the latter subcomplex did not exchange with free protein X. These data support the conclusion that E2 subunits bind the Kc subunit of the kinase and suggest that the binding of the inner domain of protein X to the inner domain of the transacetylase occurs during the assembly of the oligomeric core. Selective release of a fragment of E2 subunits that contain the lipoyl domains (E2L fragment) releases the kinase (M. Rahmatullah et al., 1990, J. Biol. Chem. 265, 14,512-14,517). Sucrose gradient centrifugation yielded an E2L-kinase fraction with an increased ratio of the kinase to E2L fragment. A monoclonal antibody specific for E2L was attached to a gel matrix. Binding of E2L fragment also led to specific binding of the kinase. Extensive washing did not reduce the level of bound kinase. Thus, the kinase is tightly bound by the lipoyl domain region of E2.     

Alterations in the pyruvate dehydrogenase complex during adaptation to glucose by Neurospora

A 20-fold induction of the pyruvate dehydrogenase complex, pyruvate dehydrogenase (EC 1.2.4.1) plus dihydrolipoate S-acetyltransferase, (lipoyltransacetylase) (EC 2.3.1.12) plus dihydrolipoyl dehydrogenase, NADH : lipoamide oxidoreductase, (EC 1.6.4.3), from a specific activity of 3.5–65.0 was observed in mitochondrial extracts during adaptation of Neurospora to glucose from acetate media. The extent of ATP-dependent, time-dependent inactivation of the pyruvate dehydrogenase complex was approximately the same in both acetate- and glucose-grown cells, thereby indicating that the low pyruvate dehydrogenerase complex activities in acetate-grown cells did not represent phosphorylated pyruvate dehydrogenase complex molecules. High levels of dihydrolipoyl transacetylase (EC 2.3.1.12) were observed in mitochondrial extracts from acetate-grown cells; this lipoyltransacetylase was analyzed on sucrose density gradients and found to be associated with the pyruvate dehydrogenase complex. Digitonin fractionation of mitochondria revealed that both the pyruvate dehydrogenase complex and lipoyltransacetylase were primarily associated with the mitochondrial outer membrane.     

Sepharose-insolubilization of the dihydrolipoyl transacetylase core component of the pyruvate dehydrogenase complex: Preparation and characterization

The dihydrolipoyl transacetylase core component of the bovine kidney and heart pyruvate dehydrogenase complexes were covalently attached through the lipoyl moiety to Sepharose by the thiol-crosslinking reagent, N, N′-p-phenylenedimaleimide.In one approach, the N, N′-p-phenylenedimaleimide was allowed to react with glutathione which was in turn linked by its N-terminal to Sepharose CL-6B. In addition, we found the N, N′-p-phenylenedimaleimide would react directly with Sepharose CL-6B (at undetermined sites) and could be used as the sole bridge in forming a stable linkage of the transacetylase core to Sepharose. With the latter approach the extent of multiple-linkage of the 60-subunit core could more easily be controlled. This should be a generally useful approach for linking proteins with reactive surface thiol residues.Insolubilization of the core of the pyruvate dehydrogenase complex by these methods did not appear to significantly alter the binding of other protein components of the complex, but the catalytic activities of the complex requiring the lipoyl moiety were appreciably altered. Procedures for coupling the transacetylase core to various derivatives of phenylenedimaleimide-Sepharose and techniques described for studying the protein products should be useful in preparation of specialized matrices for both protein purification and the study of protein-protein interactions.     

Isolation and characterisation of the pyruvate dehydrogenase complex of anaerobically grown Enterococcus faecalis NCTC 775.

In this contribution the isolation and some of the structural and kinetic properties of the pyruvate dehydrogenase complex (PDC) of anaerobically grown Enterococcus faecalis are described. The complex closely resembles the PDC of other Gram-positive bacteria and eukaryotes. It consists of four polypeptide chains with apparent molecular masses on SDS/PAGE of 97, 55, 42 and 36 kDa, and these polypeptides could be assigned to dihydrolipoyl transacetylase (E2), lipoamide dehydrogenase (E3) and the two subunits of pyruvate dehydrogenase (E1 alpha and E1 beta), respectively. The E2 core has an icosahedral symmetry. The apparent molecular mass on SDS/PAGE of 97 kDa of the E2 chain is extremely high in comparison with other Gram-positive organisms (and eukaryotes) and probably due to several lipoyl domains associated with the E2 chain. NADH inhibition is mediated via E3. The mechanism of inhibition is discussed in view of the high PDC activities in vivo that are found in E. faecalis, grown under anaerobic conditions.     

Function of the nonidentical subunits of mammalian pyruvate dehydrogenase

The pyruvate dehydrogenase (PDH) component of the bovine kidney pyruvate dehydrogenase complex (PDC) contains two nonidentical subunits. PDH catalyzes the decarboxylation of pyruvate to produce α-hydroxyethylthiamine-PP (HETPP) and the reductive acetylation of the lipoyl moieties of dihydrolipoyl transacetylase with HETPP. Phosphorylation of PDH with PDH kinase and ATP markedly inhibits the first reaction but does not inhibit the second reaction. Since the α-subunit but not the β-subunit of PDH undergoes phosphorylation, these results suggest that the α-subunit catalyzes the first reaction and the β-subunit catalyzes the second reaction. Thiamine-PP reduces the rate of phosphorylation of PDC by PDH kinase and ATP. Phosphorylation of PDC increases the K_D of the PDC-Mg-thiamine-PP complex about 12-fold. It appears that the thiamine-PP binding site and the phosphorylation site on PDH influence each other and that HETPP is bound to PDH in a different orientation or possibly at a different site than is thiamine-PP.     

Enhanced dissociation of pyruvate dehydrogenase from the pyruvate dehydrogenase complex following phosphorylation and regulatory implications

We have shown that the active form of the pyruvate dehydrogenase (PDH_a) component exhibits at least a 9-fold greater affinity for sites on the dihydrolipoyl transacetylase core of the pyruvate dehydrogenase complex than does the inactive (phosphorylated) form of pyruvate dehydrogenase (PDH_b). Consistent with a higher rate of dissociation for PDH_b than for PDH_a, free PDH_a rapidly replaces PDH_b whereas, even at high levels, free PDH_b only slowly replaces PDH_a. Dissociation of newly formed PDH_b, during phosphorylation by the immobile PDH_a kinase, leads to an increased association of free PDH_a as observed by protection against inactivation of the complex, even though PDH_a kinase activity is increased.     

X-ray small-angle studies of the pyruvate dehydrogenase core complex from Escherichia coli K-12

The pyruvate dehydrogenase core complex from E. coli K-12, defined as the multienzyme complex which can be obtained with a unique polypeptide chain composition, has been investigated in solution with the X-ray small-angle technique. The molecular mass of the core complex of 3.78·10⁶ daltons verifies the ratio of polypeptide chains of 161616 of the three enzyme components, pyruvate dehydrogenase, dihydrolipoamide transacetylase, and dihydrolipoamide dehydrogenase, present in the complex. In connection with the values obtained for the radius of gyration (156.5 å), volume (1.07·10⁷ å³) and amount of solvent associated with the complex (1.03 g/g) a loose packing of subunits in the complex has to be assumed. The maximum diameter of the core complex of 433 å, as determined from the correlation function, corroborates the large extension of the complex. The comparison of experimental and theoretical scattering curves reveals a relatively isometric overall shape of the core complex.Enzymes: Pyruvate dehydrogenase complex = pyruvate dehydrogenase (EC 1.2.4.1) plus dihydrolipoamide transacetylase (EC 2.3.1.12) plus dihydrolipoamide dehydrogenase (EC 1.6.4.3).     

Purification of bovine kidney and heart pyruvate dehydrogenase phosphatase on Sepharose derivatized with the pyruvate dehydrogenase complex

Pyruvate dehydrogenase phosphatase has been purified to apparent homogeneity from mitochondrial extracts of both beef heart and beef kidney. An essential step in this three-step purification is affinity chromatography of a largely purified phosphatase fraction using Sepharose beads to which pyruvate dehydrogenase complex is covalently bound through the lipoic acid residues of the dihydrolipoyl transacetylase component of the complex. The purified phosphatase, which has a native relative molecular mass, Mr, of about 140000, is composed of two nonidentical subunits of Mr 89000 and 49000.     

Regulation of pig heart pyruvate dehydrogenase by phosphorylation. Studies on the subunit and phosphorylation stoicheiometries

1. The molecular weights of the subunits of purified pig heart pyruvate dehydrogenase complex were determined by sodium dodecyl sulphate/polyacrylamide-disc-gel electrophoresis and were: pyruvate decarboxylase, α-subunit 40600, β-subunit 35100; dihydrolipoyl acetyltransferase 76100; dihydrolipoyl dehydrogenase 58200. 2. Inactivation of the pyruvate dehydrogenase complex by its integral kinase corresponded to the incorporation of 0.46nmol of P/unit of complex activity inactivated. 3. Further incorporation of phosphate into the complex occurred to a limit of 1.27nmol of P/unit of complex inactivated (approx. 3 times that required for inactivation). 4. Phosphate was incorporated only into the α-subunit of the decarboxylase. 5. The molar ratio of phosphate to α-subunits of the decarboxylase was estimated by radioamidination of amino groups of pyruvate dehydrogenase [³²P]phosphate complex by using methyl [1-¹⁴C]acetimidate, followed by separation of α-subunits by sodium dodecyl sulphate/polyacrylamide-disc-gel electrophoresis. Inactivation of the complex (0.46nmol of P/unit of complex inactivated) corresponded to a molar ratio of one phosphate group per two α-chains (i.e. one phosphate group/α₂β₂ tetramer). Complete phosphorylation corresponded to three phosphate groups per α₂β₂ tetramer. 6. Subunit molar ratios in the complex were also estimated by the radioamidination technique. Results corresponded most closely to molar ratios of 4 α-subunits:4 β-subunits:2 dihydrolipoyl acetyltransferase subunits:1 dihydrolipoyl dehydrogenase subunit.     

Molecular weight and symmetry of the pyruvate dehydrogenase multienzyme complex of Escherichia coli.

The molecular weight and polypeptide chain stoichiometry of the native pyruvate dehydrogenase multienzyme complex from Escherichia coli were determined by independent techniques. The translational diffusion coefficient (D^o_20,w) of the complex was measured by laser light intensity fluctuation spectroscopy and found to be 0.90 (±0.02) × 10^?11m²/s. When this was combined in the Svedberg equation with the measured sedimentation coefficient (s^o_20,w = 60.2 (±0.4) S) and partial specific volume ($\text{v}\text{?}\text{= 0.735 (±0.01)}\text{ml/g}$), the molecular weight of the intact native complex was calculated to be 6.1 (±0.3) × 10⁶. The polypeptide chain stoichiometry (pyruvate decarboxylase: lipoate acetyltransferase: lipoamide dehydrogenase) of the same sample of pyruvate dehydrogenase complex was measured by the radioamidination technique of Bates et al. (1975) and found to be 1.56:1.0:0.78.From this stoichiometry and the published polypeptide chain molecular weights estimated by sodium dodecyl sulphate/polyacrylamide gel electrophoresis, a minimum chemical molecular weight of 283,000 was calculated. This structure must therefore be repeated approximately 22 times to make up the native complex, a number which is in good agreement with the expected repeat of 24 times if the lipoate acetyltransferase core component has octahedral symmetry. It is consistent with what appears in the electron microscope to be trimer-clustering of the lipoate acetyltransferase chains at the corners of a cube. It rules out any structure based on 16 lipoate acetyltransferase chains comprising the enzyme core.The preparation of pyruvate dehydrogenase complex was polydisperse: in addition to the major component, two minor components with sedimentation coefficients (s^o_20,w) of 90.3 (±0.9) S and 19.8 (±0.3) S were observed. Together they comprised about 17% of the total protein in the enzyme sample. Both were in slowly reversible equilibrium with the major 60.2 S component but appeared to be enzymically active in the whole complex reaction. The faster-sedimenting species is probably a dimer of the complex, whereas the slower-sedimenting species has the properties of an incomplete aggregate of the component enzymes of the complex based on a trimer of the lipoate acetyltransferase chain.     

Dihydrolipoamide transacetylase from Escherichia Coli: evidence for internal gene duplication.

The dihydrolipoamide transacetylase component of the pyruvate dehydrogenase complex from $\text{Escherichia}$ coli consists of identical subunits of 89.000 daltons. During the purification procedure it is partially cleaved into active fragments of 82.000 daltons, 37.000 daltons and 35.000 daltons. Fingerprint analysis shows at least 80% similarity of the fragments compared with the native component. This provides evidence for two large homologous domains within same polypeptide chain.     

Escherichia coli pyruvate dehydrogenase complex: particle masses of the complex and component enzymes measured by scanning transmission electron microscopy

Particle masses of the Escherichia coli pyruvate dehydrogenase (PDH) complex and its component enzymes have been measured by scanning transmission electron microscopy (STEM). The particle mass of PDH complex measured by STEM is 5.28 X 10(6) with a standard deviation of 0.40 X 10(6). The masses of the component enzymes together with their standard deviations are (2.06 +/- 0.26) X 10(5) for the dimeric pyruvate dehydrogenase (E1), (1.15 +/- 0.17) X 10(5) for dimeric dihydrolipoyl dehydrogenase (E3), and (2.20 +/- 0.17) X 10(6) for dihydrolipoyl transacetylase (E2), the 24-subunit core enzyme. The latter value corresponds to a subunit molecular weight of (9.17 +/- 0.71) X 10(4) for E2. The subunit molecular weight measured by polyacrylamide gel electrophoresis in sodium dodecyl sulfate is 8.6 X 10(4). STEM measurements on PDH complex incubated with excess E3 or E1 failed to detect any additional binding of E3 but showed that the complex would bind additional E1 under forcing conditions (high concentrations with glutaraldehyde). The additional E1 subunits were bound too weakly to represent binding sites in an isolated or isolable complex. The mass measurements by STEM are consistent with the subunit composition 24:24:12 when interpreted in the light of the flavin content of the complex and assuming 24 subunits in the core enzyme (E2).     

Synthesis of the E. coli pyruvate dehydrogenase complex: non-dependence on 3'-5'-cyclic AMP

A determination of the level of the pyruvate dehydrogenase complex in a 3′–5′-c-AMP deficient mutant of $\text{E.}$$\text{coli}$ K12 has been carried out. The deficiency has no effect on specific activities for derivatives carrying either the inducible genes for two components of the complex or constitutive mutants. We conclude that synthesis of the complex is not sensitive to catabolite repression.     

Component requirements for NADH inhibition and spermine stimulation of pyruvate dehydrogenaseb phosphatase activity

The subunit and subdomain requirements for NADH inhibition as well as Ca+ and spermine activation of the pyruvate dehydrogenaseb phosphatase were analyzed. The transacetylase-protein X subcomplex (E2-X) was required for all three effects. The oligomeric inner domain of the transacetylase did not support any of these regulatory effects. The presence of at least a portion of the outer (lipoyl-bearing) domains of the transacetylase but not the lipoyl-bearing portion of protein X was essential for expression of these regulatory effects on phosphatase activity. The inner domain of protein X may contribute to some effects. The E2-X subcomplex, alone, had no effect on phosphatase activity in the absence of Ca2+, but the subcomplex did support both NADH inhibition and spermine activation in the absence of Ca2+. Studies with peptide substrates established that spermine is directly bound by a phosphatase subunit. With the resolved pyruvate dehydrogenase component (E1b) used as the substrate, the E2-X subcomplex transformed the effect of spermine from inhibiting to stimulating the rate of dephosphorylation by the phosphatase. The above observations suggest that binding of E1b to the E2-X subcomplex alters its presentation to the phosphatase. We also present several observations that are consistent with NADH inhibition of the phosphatase being mediated through a dihydrolipoyl dehydrogenase-dependent reduction of lipoyl moieties in the E2-X subcomplex. Overall, our data establish that the outer, lipoyl-bearing domains of the oligomeric transacetylase core have an essential role in the function and regulation of the pyruvate dehydrogenase phosphatase.     

Fluoropyruvate: an unusal substrate for Escherichia coli pyruvate dehydrogenase.

The pyruvate dehydrogenase component of the $\text{E. coli}$ pyruvate dehydrogenase complex catalyzes the decomposition of 3-fluoropyruvate to acetate and fluoride ions in equimolar amounts and at about one-tenth the rate at which it catalyzes the conversion of pyruvate and ferricyanide to acetate and ferrocyanide. When the reaction is carried out in [³H]H₂O the product is [³H]acetate. The reaction is strictly dependent upon added thiamin pyrophosphate, and a mechanistic role is proposed for this coenzyme.     

Escherichia coli pyruvate dehydrogenase complex. Thiamin pyrophosphate-dependent inactivation by 3-bromopyruvate

Inactivation of the pyruvate dehydrogenase complex by 3-bromopyruvate is thiamin pyrophosphate (TPP)-dependent. Inactivation with 2-14C- or 3-14C-labeled 3-bromopyruvate results in TPP-dependent covalent labeling of more than 60 sites in the complex, all of which are associated with the dihydrolipoyl transacetylase component. Inactivation by 3-bromo[1-14C]pyruvate labels up to 20 sites associated with dihydrolipoyl transacetylase, also with TPP dependence. Systemic chemical degradation of the complex inactivated by 3-bromo[2-14C]pyruvate under conditions that would convert lipoyl groups to S,S,-biscarboxymethyl dihydrolipoic acid produces S,S,-bis[14C]carboxymethyl dihydrolipoic acid. It is concluded that 3-bromopyruvate inactivates this complex by initially undergoing the first two steps of the usual catalytic pathway, TPP-dependent decarboxylation followed by reductive bromoacetylation of lipoyl moieties. The sulfhydryl groups of S-bromoacetyl dihydrolipoyl moieties generated by reductive bromoacetylation are then alkylated by 3-bromopyruvate as well as by bromoacetyl thioester groups associated with the complex.     

Configuration of interdomain linkers in pyruvate dehydrogenase complex of Escherichia coli as determined by cryoelectron microscopy.

The dihydrolipoyl transacetylase (E2p) component of the pyruvate dehydrogenase complex (PDC) of Escherichia coli is a multidomain polypeptide comprising a catalytic domain, a domain that binds dihydrolipoyl dehydrogenase (E3-binding domain), and three domains containing lipoic acid (lipoyl domains). In PDC 24 subunits of E2p associate by means of interactions involving the catalytic domains to form the structural core of PDC. From cryoelectron microscopy and computer image analysis of frozen-hydrated isolated E2p cores it appears that the lipoyl domains are located peripherally about the core complex and do not assume fixed positions. To further test this interpretation the visibility of the lipoyl domains in electron micrographs was enhanced by specifically biotinylating the lipoic acids and labeling them with streptavidin. In agreement with the studies of native, unlabeled E2p cores, cryoelectron microscopy of the streptavidin-labeled E2p cores showed that the lipoic acid moieties are capable of extending approximately 13 nm from the surface of the core. Localization of the E3-binding domains was accomplished by cryoelectron microscopy of E2p-E3 subcomplexes prepared by reconstitution in vitro. Frequently an apparent gap of several nanometers separated the bound E3 from the surface of the core. The third component of PDC, pyruvate dehydrogenase (E1p), appeared to bind to the E2p core in a manner similar to that observed for E3. These results support a structural model of the E2p core in which the catalytic, E3-binding, and three lipoyl domains are interconnected by linker sequences that assume extended and flexible conformations.