Impaired Assembly of E1 Decarboxylase of the Branched-chain α-Ketoacid Dehydrogenase Complex in Type IA Maple Syrup Urine Disease

The E1 decarboxylase component of the human branched-chain ketoacid dehydrogenase complex comprises two E1α (45.5 kDa) and two E1β (37.5 kDa) subunits forming an α2β2 tetramer. In patients with type IA maple syrup urine disease, the E1α subunit is affected, resulting in the loss of E1 and branched-chain ketoacid dehydrogenase catalytic activities. To study the effect of human E1α missense mutations on E1 subunit assembly, we have developed a pulse-chase labeling protocol based on efficient expression and assembly of human (His)6-E1α and untagged E1β subunits inEscherichia coli in the presence of overexpressed chaperonins GroEL and GroES. Assembly of the two35S-labeled E1 subunits was indicated by their co-extraction with Ni2+-nitrilotriacetic acid resin. The nine E1α maple syrup urine disease mutants studied showed aberrant kinetics of assembly with normal E1β in the 2-h chase compared with the wild type and can be classified into four categories of normal (N222S-α and R220W-α), moderately slow (G245R-α), slow (G204S-α, A240P-α, F364C-α, Y368C-α, and Y393N-α), and no (T265R-α) assembly. Prolonged induction in E. coli grown in the YTGK medium or lowering of induction temperature from 37 to 28 °C (in the case of T265R-α), however, resulted in the production of mutant E1 proteins. Separation of purified E1 proteins by sucrose density gradient centrifugation showed that the wild-type E1 existed entirely as α2β2 tetramers. In contrast, a subset of E1α missense mutations caused the occurrence of exclusive αβ dimers (Y393N-α and F364C-α) or of both α2β2 tetramers and lower molecular weight species (Y368C-α and T265R-α). Thermal denaturation at 50 °C indicated that mutant E1 proteins aggregated more rapidly than wild type (rate constant, 0.19 min−1), with the T265R-α mutant E1 most severely affected (rate constant, 4.45 min−1). The results establish that the human E1α mutations in the putative thiamine pyrophosphate-binding pocket that are studied, with the exception of G204S-α, have no effect on E1 subunit assembly. The T265R-α mutation adversely impacts both E1α folding and subunit interactions. The mutations involving the C-terminal aromatic residues impede both the kinetics of subunit assembly and the formation of the native α2β2 structure. The E1 decarboxylase component of the human branched-chain ketoacid dehydrogenase complex comprises two E1α (45.5 kDa) and two E1β (37.5 kDa) subunits forming an α2β2 tetramer. In patients with type IA maple syrup urine disease, the E1α subunit is affected, resulting in the loss of E1 and branched-chain ketoacid dehydrogenase catalytic activities. To study the effect of human E1α missense mutations on E1 subunit assembly, we have developed a pulse-chase labeling protocol based on efficient expression and assembly of human (His)6-E1α and untagged E1β subunits inEscherichia coli in the presence of overexpressed chaperonins GroEL and GroES. Assembly of the two35S-labeled E1 subunits was indicated by their co-extraction with Ni2+-nitrilotriacetic acid resin. The nine E1α maple syrup urine disease mutants studied showed aberrant kinetics of assembly with normal E1β in the 2-h chase compared with the wild type and can be classified into four categories of normal (N222S-α and R220W-α), moderately slow (G245R-α), slow (G204S-α, A240P-α, F364C-α, Y368C-α, and Y393N-α), and no (T265R-α) assembly. Prolonged induction in E. coli grown in the YTGK medium or lowering of induction temperature from 37 to 28 °C (in the case of T265R-α), however, resulted in the production of mutant E1 proteins. Separation of purified E1 proteins by sucrose density gradient centrifugation showed that the wild-type E1 existed entirely as α2β2 tetramers. In contrast, a subset of E1α missense mutations caused the occurrence of exclusive αβ dimers (Y393N-α and F364C-α) or of both α2β2 tetramers and lower molecular weight species (Y368C-α and T265R-α). Thermal denaturation at 50 °C indicated that mutant E1 proteins aggregated more rapidly than wild type (rate constant, 0.19 min−1), with the T265R-α mutant E1 most severely affected (rate constant, 4.45 min−1). The results establish that the human E1α mutations in the putative thiamine pyrophosphate-binding pocket that are studied, with the exception of G204S-α, have no effect on E1 subunit assembly. The T265R-α mutation adversely impacts both E1α folding and subunit interactions. The mutations involving the C-terminal aromatic residues impede both the kinetics of subunit assembly and the formation of the native α2β2 structure. The mammalian mitochondrial branched-chain α-ketoacid dehydrogenase (BCKD) 1The abbreviations used are: BCKD, branched-chain α-ketoacid dehydrogenase; E1, branched-chain α-ketoacid decarboxylase; E2, dihydrolipoyl transacylase; E3, dihydrolipoamide dehydrogenase; HPLC, high performance liquid chromatography; Hsp, heat shock protein; IPTG, isopropyl β-d-thiogalactopyranoside; KPi, potassium phosphate; MBP, maltose-binding protein; MSUD, maple syrup urine disease; NTA, nitrilotriacetic acid; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; TEV, tobacco etch virus; TPP, thiamine pyrophosphate.1The abbreviations used are: BCKD, branched-chain α-ketoacid dehydrogenase; E1, branched-chain α-ketoacid decarboxylase; E2, dihydrolipoyl transacylase; E3, dihydrolipoamide dehydrogenase; HPLC, high performance liquid chromatography; Hsp, heat shock protein; IPTG, isopropyl β-d-thiogalactopyranoside; KPi, potassium phosphate; MBP, maltose-binding protein; MSUD, maple syrup urine disease; NTA, nitrilotriacetic acid; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; TEV, tobacco etch virus; TPP, thiamine pyrophosphate. complex catalyzes the oxidative decarboxylation of the branched-chain α-ketoacids derived from the branched-chain amino acids, leucine, isoleucine, and valine (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). This multienzyme complex is organized around a dihydrolipoyl transacylase (E2) core, to which a branched-chained α-ketoacid decarboxylase (E1), a dihydrolipoamide dehydrogenase (E3), a specific kinase, and a specific phosphatase are attached through ionic interactions (2Pettit F.H. Yeaman S.J. Reed L.J. Proc. Natl. Acad. Sci. U. S. A. 1978; 75: 4885-4888Crossref Scopus (199) Google Scholar, 3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). E1 is a thiamine pyrophosphate (TPP)-dependent enzyme comprising two E1α and two E1β subunits that assemble into an α2β2 tetramer. E2 is a 24-meric protein consisting of identical lipoic acid-bearing subunits arranged on octahedral 4,3,2-point group symmetry. Each E2 polypeptide contains three independently folded domains (i.e. lipoyl-bearing, E1/E3-binding, and inner core) that are highly conserved among E2 proteins of the α-ketoacid dehydrogenase complexes (4Reed L.J. Hackert M.L. J. Biol. Chem. 1990; 265: 8971-8974Abstract Full Text PDF PubMed Google Scholar). These complexes include pyruvate dehydrogenase, α-ketoglutarate dehydrogenase, and BCKD complexes (3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). E3 is a homodimeric flavoprotein that is common to members of the α-ketoacid dehydrogenase complexes (3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). The kinase and the phosphatase are specific for the BCKD complex and regulate its activity through a reversible phosphorylation (inactivation) and dephosphorylation (activation) cycle (5Harris R.A. Paxton R. Powell S.M. Goodwin G.W. Kuntz M.J. Han A.C. Adv. Enzyme Regul. 1986; 25: 219-237Crossref PubMed Scopus (73) Google Scholar).In patients with maple syrup urine disease (MSUD) or branched-chain ketoaciduria, the activity of the BCKD complex is deficient. This leads to clinical manifestations including often fatal ketoacidosis, neurological derangements, and mental retardation (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). The molecular genetics of MSUD are heterogeneous as mutations in the E1α, E1β, E2, and E3 genes have been described (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar, 6Cox R.P. Chuang J.L. Chuang D.T. Rosenberg R.N. Prusiner S.B. DiMauro S. Barchi R.L. Kunkel L.M. The Molecular and Genetic Basis of Neurological Disease. 2nd Ed. Butterworth-Heinemann, Boston1997: 1175-1193Google Scholar). Based on the locus affected, genetic subtypes of MSUD have been proposed, with type IA referring to mutations in the E1α gene, type IB to the E1β gene, type II to the E2 gene, and type III to the E3 gene (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). It has been suggested that certain type IA MSUD missense mutations, for example Y393N-α (7Fisher C.R. Chuang J.L. Cox R.P. Fisher C.W. Star R.A. Chuang D.T. J. Clin. Invest. 1991; 88: 1034-1037Crossref PubMed Scopus (27) Google Scholar) and Y368C-α (8Chuang J.L. Fisher C.R. Cox R.P. Chuang D.T. Am. J. Hum. Genet. 1994; 55: 297-304PubMed Google Scholar), may impede the assembly of mutant E1α with normal E1β subunit, resulting in the degradation of E1 subunits in patient's cells.We have recently established that chaperonins GroEL and GroES are essential for efficient folding and assembly of the E1 tetramer inEscherichia coli (9Wynn R.M. Davie J.R. Cox R.P. Chuang D.T. J. Biol. Chem. 1992; 267: 12400-12403Abstract Full Text PDF PubMed Google Scholar) and the E2 24-mer in vitro(10Wynn R.M. Davie J.R. Zhi W. Cox R.P. Chuang D.T. Biochemistry. 1994; 33: 8962-8968Crossref PubMed Scopus (14) Google Scholar). To gain insight into the biochemical basis of the apparently impaired assembly of E1 in type IA MSUD, we have co-expressed both mature mutant E1α and normal E1β in E. colico-transformed with a second plasmid overproducing chaperonins GroEL and GroES. Pulse-chase labeling of both E1 subunits was carried out to measure the kinetics of assembly of the mutant E1α with normal E1β in the bacterial cell. The results showed a marked reduction in the rate of E1 assembly in certain E1α mutants compared with normal. It was also found that a subset of E1α mutations affect the assembly state of mutant E1 after the prolonged induction in E. coli. Thermostability and protease digestion studies further indicated these slowly assembled mutant E1 proteins had less stable conformations than the wild type. These results define the residues that are critical for subunit interactions and stability of E1 and have implications for understanding chaperonin-mediated biogenesis of hetero-oligomeric structures.DISCUSSIONThe major focus of this investigation is to characterize the effect of human E1α mutations in type IA MSUD on the assembly and stability of mutant E1 proteins. For these studies, we have developed efficient bacterial expression systems for folding and assembly of E1 α2β2 tetramers. We showed previously that co-expression of mature MBP-E1α and E1β sequences of the human BCKD complex in the same E. coli cells is essential for MBP-E1 assembly; however, the yield was very low (12Davie J.R. Wynn R.M. Cox R.P. Chuang D.T. J. Biol. Chem. 1992; 267: 16601-16606Abstract Full Text PDF PubMed Google Scholar). Co-transformation of a second plasmid that overexpressed GroEL and GroES into the sameE. coli cell resulted in a 500-fold increase in the yield of active MBP-E1 tetramers (9Wynn R.M. Davie J.R. Cox R.P. Chuang D.T. J. Biol. Chem. 1992; 267: 12400-12403Abstract Full Text PDF PubMed Google Scholar). In the present study, a (His)6affinity tag is fused to the mature E1α N terminus through a TEV protease recognition site. Co-transformation with the pGroESL plasmid was also found necessary and sufficient for productive folding and assembly of (His)6-E1. The results argue against the suggestion that the dependence of human E1 on chaperonins for a high yield is due to the presence of the MBP sequence in the chimeric E1α polypeptide (23Lessard I.A.D. Persham R.N. J. Biol. Chem. 1994; 269: 10378-10383Abstract Full Text PDF PubMed Google Scholar). Our recent in vitro refolding results indicate that the reconstitution of untagged E1, MBP-E1, and (His)6-E1 show the same chaperonin-dependent kinetics. 2J. L. Chuang, R. M. Wynn, and D. T. Chuang, manuscript in preparation. The findings further established that productive folding and assembly of mature human E1 have an absolute requirement for enrichment for chaperonins GroEL and GroES and are not affected by the presence of affinity tags. Thus, the pulse-chase labeling protocol developed in this study provides the first approximation of the rate of E1 subunit assembly under optimal conditions through the augmentation of bacterial chaperonins that are homologue of mitochondrial chaperonins Hsp60 and Hsp10, respectively.The wild-type and mutant human E1α and the wild-type E1β subunits are expressed at relatively equal efficiencies and are stable, as indicated by Western blotting of the total crude lysates prepared 12 h after induction (Fig. 3 A). This allows one to follow the fates of E1α and E1β subunits synthesized during the 1-min window of pulse-labeling. The presence of the (His)6-tag in the wild-type and mutant E1α subunits facilitates the isolation of the 35S-labeled subunit. One can measure the kinetics of the E1β assembly with the wild-type and mutant E1α subunits by the co-purification of the untagged E1β with (His)6-E1α with the Ni2+-NTA resin as a function of time. The equally strong signals of E1α and E1β subunits during the 2-h chase (Fig. 2 A) indicate that both subunits are efficiently synthesized. The total numbers of cysteine and methionine residues in the E1α and E1β subunits are similar, which are 16 and 15, respectively. The autoradiogram of the co-purified E1 subunits as separated by SDS-PAGE cannot discern the assembly state of the associated subunits. However, size fractionation of the pulse-chase labeled products by HPLC show that the wild-type E1α and E1β assemble during the 2-h chase occur predominantly as inactive αβ dimers, which are later converted to active α2β2 tetramers (Fig. 4, A andB). It is noteworthy that a significant amount of the wild-type αβ dimeric intermediate was observed only when E. coli cells were grown on the C2 broth minimal medium. When the bacterial cells were cultured on the YTGK medium, wild-type E1 was expressed predominantly as α2β2 tetramers (Fig. 5) with little or no accumulation of αβ dimers during the 16-h induction period (data no shown). The factors responsible for the apparent effects of culture media on the accumulation of the wild-type dimeric intermediate during E1 assembly are currently unknown. The major differences between the two bacterial culture media lie in the fact that the C2 minimal medium is low in the content of SO42− and amino acids when compared with the YTGK medium (18Guzman-Verduzco L.M. Kupersztoch Y.M. J. Bacteriol. 1987; 169: 5201-5208Crossref PubMed Google Scholar). Possible effects of these ingredients on the dimerization of wild-type αβ dimers are under investigation.It is of interest that a weak GroEL signal co-purifies with wild-type (His)6-E1α and E1β at 120 min into the chase (Fig.2 A). This apparent ternary complex is also observed in the early fractions of the Imidazole gradient during purification of wild-type (His)6-E1 (Fig. 1). The GroEL-E1α-E1β ternary complex is a productive intermediate at a later step of the chaperonin-mediated assembly of E1 α2β2tetramers.2 Only a GroEL-E1α binary complex is observed in F364C-α because the assembly of E1β with mutant E1α did not occur within the 2-h chase. The weak and sub-stoichiometric signal of GroEL relative to E1 subunits is a result of isotopic dilution by the overabundance of unlabeled GroEL in E. coli.The assembled mutant αβ dimers were produced in E. coligrown on the YTGK medium after the 16-h induction with "slow assembly" E1α mutants including Y393N-α and F364C-α. The results indicate that this group of mutations not only reduces the rate of the assembly of the mutant E1α with normal E1β but also prevents conversion of the dimeric assembly intermediate into the stable α2β2 structure of wild-type E1. The production of these mutant αβ dimers apparently is not affected by growth media, as the expression of mutant E1 carrying these mutations in E. coli grown in the C2 broth minimal medium also resulted in the expression of exclusive dimers. In vitroreconstitution of the 6 m urea-denatured mutant Y393N-α E1 in the presence of chaperonins GroEL and GroES also resulted exclusively in mutant αβ dimers.2 The unstable Y393N-α dimeric intermediate as demonstrated by its propensity for thermal aggregation and proteolytic digestion compared with the wild-type tetramer explains the markedly reduced levels E1α and E1β subunits in cells from Mennonite MSUD patients homozygous for the Y393N-α mutation (24Matsuda I. Nobunkuni Y. Mitsubuchi H. Indo Y. Indo F. Asaka J. Harada H. Biochem. Biophys. Res. Commun. 1990; 172: 451-464Crossref Scopus (29) Google Scholar, 25Fisher C.R. Fisher C.W. Chuang D.T. Cox R.P. Am. J. Hum. Genet. 1991; 49: 429-434PubMed Google Scholar). The present study establishes that C-terminal aromatic residues (F364-α, Y368-α, and Y393-α) in the E1α subunit are crucial for proper E1 assembly. The important roles of the C terminus in subunit assembly and protein interactions have been demonstrated. For example, the C-terminal 25 amino acid residues of the herpes simplex virus type 1 UL26.5 protein are required for the assembly of the icosahedral capsid shell (26Kennard J. Rixon F.J. McDougall I.M. Tatman J.D. Preston V.G. J. Gen. Virol. 1995; 76: 1611-1621Crossref PubMed Scopus (59) Google Scholar). In the case of the E2 core of the related pyruvate dehydrogenase complex fromAzotobacter vinelandii, the C-terminal residues 632–637 comprise a 310-like helix (H6) which acts as a "hydrophobic knob" that fits into a "hole" in the 2-fold related subunit to produce the 24-mer cubic assembly (27Mattevi A. Obmodova G. Schulze E. Kalk K.H. Westphal A.H. deKok A. Hol W.G.J. Science. 1992; 255: 1544-1550Crossref PubMed Scopus (227) Google Scholar). Introduction of a polyhistidine tag into the C terminus of BCKD-E2, which is highly homologous to the bacterial pyruvate dehydrogenase-E2, results in the formation of stable trimers instead of the native 24-mer structure (data not shown).The T265R-α subunit, when expressed at 37 °C, was largely insoluble even in the presence of excess chaperonins GroEL and GroES. This was indicated by the rapid disappearance of the mutant E1α signal in the soluble fraction during the 2-h chase (Fig.2 A); however, the level of the T265R-α subunits in total crude lysates was comparable to that of wild type (Fig. 3 A). These results strongly suggest that the T265R-α residue is important for proper folding of the E1α subunit. It is also of interest that lowering of the induction temperature from 37 to 28 °C resulted in the production of a significant amount of assembled mutant E1 protein carrying the T265R-α mutation. The yield of the mutant E1 was 5 mg/liter culture at 28 °C compared with 20–40 mg/liter culture for the wild-type E1 expressed at 37 °C (data not shown). The finding is consistent with the thesis that lowering the expression temperature slows the folding kinetics of the nascent peptide, thereby reducing the probability of the off-pathway folding reactions, as demonstrated by the expression of rabbit muscle phosphorylase (28Browner M.F. Rosor P. Tugendreich S. Fletterick R.J. Protein Eng. 1991; 4: 351-357Crossref PubMed Scopus (38) Google Scholar) and the human E1β (13Wynn R.M. Chuang J.L. Davie J.R. Fisher C.W. Hale M.A. Cox R.P. Chuang D.T. J. Biol. Chem. 1992; 267: 1881-1889Abstract Full Text PDF PubMed Google Scholar). However, the assembled T265R-α mutant E1 has a grossly altered conformation, which renders it very unstable as indicated by its most rapid thermal aggregation at 50 °C among the nine E1α mutants studied. This unstable conformation is also manifested by the apparent dissociation of the mutant tetramers to lower molecular weight species as detected by sucrose density gradient centrifugation. The current data indicate that the T265R-α residue also plays a key role in subunit interactions and are consistent with the location of this residue at the putative subunit-interaction site conserved between BCKD and pyruvate dehydrogenase E1 proteins (29Wexler I.D. Hemalatha S.G. Patel M.S. FEBS Lett. 1991; 282: 209-213Crossref PubMed Scopus (45) Google Scholar).The crystal structure of the E1 α2β2 has not been solved, but structures are known for the related TPP-dependent proteins transketolase (30Nikkola M. Lindqvist Y. Schneider G. J. Mol. Biol. 1994; 238: 387-404Crossref PubMed Scopus (160) Google Scholar) and pyruvate decarboxylase (31Dyda F. Furey W. Swaminathan S. Sax M. Farrenkopf B. Jordan F. Biochemistry. 1993; 32: 6165-6170Crossref PubMed Scopus (225) Google Scholar) from Saccharomyces cerevisiae and pyruvate oxidase from Lactobacillus plantarum (32Muller Y.A. Schulz G.E. Science. 1993; 259: 965-967Crossref PubMed Scopus (215) Google Scholar). The transketolase is a homodimer, whereas the human E1 is a tetramer made up of two non-identical subunits. Sequence alignment between the two enzymes shows that the highly conserved TPP-binding pocket in E1 is composed of residues from both E1α and E1β subunits (33Robinson B.H. Chun K. FEBS Lett. 1993; 328: 99-102Crossref PubMed Scopus (33) Google Scholar, 34Wynn R.M. Davie J.R. Meng M. Chuang D.T. Patel M.S. Roche T.E. Harris R.A. α-Ketoacid Dehydrogenase Complexes. Birkhäuser Verlag, Basel1996: 101-117Google Scholar). Aromatic residues from E1β form a hydrophobic pocket to accommodate the pyrimidium and thiazolium rings of cofactor TPP. On the other hand, the highly conserved TPP-binding motif GDG(X)22–28NN, which was first described by Hawkins et al. (35Hawkins C.F. Borges A. Perham R.N. FEBS Lett. 1989; 255: 77-82Crossref PubMed Scopus (255) Google Scholar) and is essential for binding the pyrophosphate moiety, is located in the E1α subunit (Fig.8). It should be mentioned that a D440E mutation introduced via mutagenesis into this motif in pyruvate decarboxylase from Zymomonas mobilis yielded a homodimeric enzyme with reduced affinity for TPP, in contrast to the wild-type enzyme which exists as a homotetramer (36Candy J.M. Duggleby R.G. Biochem. J. 1994; 300: 7-13Crossref PubMed Scopus (35) Google Scholar). It was proposed that deficient TPP binding may have caused a failure in the conversion of the mutant dimeric forms into native tetramers. However, the occurrence of αβ dimers in the mutant E1 carrying Y393N-α and F364C-α substitutions is likely through a different mechanism, since these residues are not involved in TPP binding and are located in the C-terminal region. G204S-α, R220W-α, and N222S-α that are affected in type IA MSUD are residues within the TPP-binding motif (Fig. 8). Specifically, N222S-α aligns with Asn-187 in the yeast transketolase and provides a ligand to this pentameric coordination involving the Mg2+ cation. The N222S-α mutation conceivably disrupts the pentameric coordination, resulting in the inability of E1 to bind the pyrophosphate moiety of TPP and the loss of E1 catalytic function. However, the N222S-α mutation is without effect on the assembly of the mutant E1α with normal E1β, as determined by pulse-chase labeling (Fig. 2 A). The R220W-α mutation, which is also located in the pyrophosphate moiety binding site, also has no adverse effect on E1 subunit assembly. In contrast, G204S-α mutation, which is presumably located at the interface between the two non-identical subunits of E1, based on the yeast transketolase structure (37Sunderstro¨m M. Lindqvist Y. Schneider G. FEBS Lett. 1992; 313: 229-231Crossref PubMed Scopus (67) Google Scholar), impedes the assembly of the mutant E1α subunit with the normal E1β unit. The mammalian mitochondrial branched-chain α-ketoacid dehydrogenase (BCKD) 1The abbreviations used are: BCKD, branched-chain α-ketoacid dehydrogenase; E1, branched-chain α-ketoacid decarboxylase; E2, dihydrolipoyl transacylase; E3, dihydrolipoamide dehydrogenase; HPLC, high performance liquid chromatography; Hsp, heat shock protein; IPTG, isopropyl β-d-thiogalactopyranoside; KPi, potassium phosphate; MBP, maltose-binding protein; MSUD, maple syrup urine disease; NTA, nitrilotriacetic acid; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; TEV, tobacco etch virus; TPP, thiamine pyrophosphate.1The abbreviations used are: BCKD, branched-chain α-ketoacid dehydrogenase; E1, branched-chain α-ketoacid decarboxylase; E2, dihydrolipoyl transacylase; E3, dihydrolipoamide dehydrogenase; HPLC, high performance liquid chromatography; Hsp, heat shock protein; IPTG, isopropyl β-d-thiogalactopyranoside; KPi, potassium phosphate; MBP, maltose-binding protein; MSUD, maple syrup urine disease; NTA, nitrilotriacetic acid; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; TEV, tobacco etch virus; TPP, thiamine pyrophosphate. complex catalyzes the oxidative decarboxylation of the branched-chain α-ketoacids derived from the branched-chain amino acids, leucine, isoleucine, and valine (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). This multienzyme complex is organized around a dihydrolipoyl transacylase (E2) core, to which a branched-chained α-ketoacid decarboxylase (E1), a dihydrolipoamide dehydrogenase (E3), a specific kinase, and a specific phosphatase are attached through ionic interactions (2Pettit F.H. Yeaman S.J. Reed L.J. Proc. Natl. Acad. Sci. U. S. A. 1978; 75: 4885-4888Crossref Scopus (199) Google Scholar, 3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). E1 is a thiamine pyrophosphate (TPP)-dependent enzyme comprising two E1α and two E1β subunits that assemble into an α2β2 tetramer. E2 is a 24-meric protein consisting of identical lipoic acid-bearing subunits arranged on octahedral 4,3,2-point group symmetry. Each E2 polypeptide contains three independently folded domains (i.e. lipoyl-bearing, E1/E3-binding, and inner core) that are highly conserved among E2 proteins of the α-ketoacid dehydrogenase complexes (4Reed L.J. Hackert M.L. J. Biol. Chem. 1990; 265: 8971-8974Abstract Full Text PDF PubMed Google Scholar). These complexes include pyruvate dehydrogenase, α-ketoglutarate dehydrogenase, and BCKD complexes (3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). E3 is a homodimeric flavoprotein that is common to members of the α-ketoacid dehydrogenase complexes (3Yeaman S.J. Biochem. J. 1989; 247: 625-632Crossref Scopus (260) Google Scholar). The kinase and the phosphatase are specific for the BCKD complex and regulate its activity through a reversible phosphorylation (inactivation) and dephosphorylation (activation) cycle (5Harris R.A. Paxton R. Powell S.M. Goodwin G.W. Kuntz M.J. Han A.C. Adv. Enzyme Regul. 1986; 25: 219-237Crossref PubMed Scopus (73) Google Scholar). In patients with maple syrup urine disease (MSUD) or branched-chain ketoaciduria, the activity of the BCKD complex is deficient. This leads to clinical manifestations including often fatal ketoacidosis, neurological derangements, and mental retardation (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). The molecular genetics of MSUD are heterogeneous as mutations in the E1α, E1β, E2, and E3 genes have been described (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar, 6Cox R.P. Chuang J.L. Chuang D.T. Rosenberg R.N. Prusiner S.B. DiMauro S. Barchi R.L. Kunkel L.M. The Molecular and Genetic Basis of Neurological Disease. 2nd Ed. Butterworth-Heinemann, Boston1997: 1175-1193Google Scholar). Based on the locus affected, genetic subtypes of MSUD have been proposed, with type IA referring to mutations in the E1α gene, type IB to the E1β gene, type II to the E2 gene, and type III to the E3 gene (1Chuang D.T. Shih V.E. Scriver C.R. Beuadet A.L. Sly W.S. Valle D. The Metabolic and Molecular Bases of Inherited Disease. 7th Ed. McGraw-Hill Inc., New York1995: 1333-1377Google Scholar). It has been suggested that certain type IA MSUD missense mutations, for example Y393N-α (7Fisher C.R. Chuang J.L. Cox R.P. Fisher C.W. Star R.A. Chuang D.T. J. Clin. Invest. 1991; 88: 1034-1037Crossref PubMed Scopus (27) Google Scholar) and Y368C-α (8Chuang J.L. Fisher C.R. Cox R.P. Chuang D.T. Am. J. Hum. Genet. 1994; 55: 297-304PubMed Google Scholar), may impede the assembly of mutant E1α with normal E1β subunit, resulting in the degradation of E1 subunits in patient's cells. We have recently established that chaperonins GroEL and GroES are essential for efficient folding and assembly of the E1 tetramer inEscherichia coli (9Wynn R.M. Davie J.R. Cox R.P. Chuang D.T. J. Biol. Chem. 1992; 267: 12400-12403Abstract Full Text PDF PubMed Google Scholar) and the E2 24-mer in vitro(10Wynn