Differential Effects of Two Mutations at Arginine-234 in the α Subunit of Human Pyruvate Dehydrogenase

Archives of Biochemistry and Biophysics Vol. 395, No. 1, November 1, pp. 121–128, 2001 doi:10.1006/abbi.2001.2576, available online at http://www.idealibrary.com on

Differential Effects of Two Mutations at Arginine-234 in the ␣ Subunit of Human Pyruvate Dehydrogenase Scott J. Jacobia, Lioubov G. Korotchkina, and Mulchand S. Patel 1 Department of Biochemistry, School of Medicine and Biomedical Sciences, State University of New York at Buffalo, Buffalo, New York 14214

Received July 24, 2001; published online October 8, 2001

The most common mutation in the ␣ subunit of the pyruvate dehydrogenase (E1) component of the human pyruvate dehydrogenase complex (PDC) is arginine-234 to glycine and glutamine in 12 and 3 patients, respectively. Interestingly, these two mutations at the same amino acid position cause E1 (and hence PDC) deficiency by apparently different mechanisms. Recombinant human R234Q E1 had similar V max (25.7 ⴞ 4.4 units/mg E1) and apparent K m (101 ⴞ 4 nM) values for TPP as recombinant wild-type human E1, while R234G E1 had no significant change in V max (33.6 ⴞ 4.7 units/mg E1) but had a 7-fold increase in its apparent K m value for TPP (497 ⴞ 25 nM). Both of the R234 mutant proteins had similar apparent K m values for pyruvate. Both R234Q and R234G mutant proteins displayed similar phosphorylation rates of sites 1 and 2 by pyruvate dehydrogenase kinase 2 (PDK2) and site 3 by PDK1 compared to wild-type E1. Phosphorylated R234Q E1, R234G E1, and wild-type E1 also had similar dephosphorylation rates of sites 1 and 2 by phosphopyruvate dehydrogenase phosphatase 1. The rate of dephosphorylation of site 3 was about 50% for R234Q E1 and without a significant change for R234G E1 compared to the wild type. The data indicate that the patients with the R234G E1 mutation are symptomatic due to a decreased ability of this mutant protein to bind TPP, whereas the patients with the R234Q E1 mutation are symptomatic due to a decreased rate of dephosphorylation of site 3, hence keeping the enzyme in a phosphorylated/inactivated form. © 2001

pyruvate dehydrogenase deficiency; thiamine pyrophosphate; phosphorylation; dephosphorylation; heat stability.

Mammalian pyruvate dehydrogenase complex (PDC)2 is composed of three catalytic subunits, pyruvate dehydrogenase (E1), dihydrolipoamide acetyltransferase (E2), and dihydrolipoamide dehydrogenase (E3), an E3-binding protein (E3BP), which facilitates the binding of E3 to the E2 core, and two regulatory components, pyruvate dehydrogenase kinase (PDK) and phosphopyruvate dehydrogenase phosphatase (PDP) (1–3). PDC catalyzes the oxidative decarboxylation of pyruvate, using CoASH and NAD ⫹, to acetyl-CoA with the concomitant formation of carbon dioxide and NADH (1, 4). E1, an ␣ 2␤ 2 heterotetramer, catalyzes the rate-limiting step in the PDC reaction. E1 activity consists of two partial reactions, the decarboxylation of pyruvate and the reductive acetylation of lipoyl moieties covalently linked to E2. PDC is controlled posttranslationally by phosphorylation (inactivation) and dephosphorylation (reactivation) of the E1 component by several isoforms of PDK and PDP, respectively. E1 can be phosphorylated by PDK on three serine residues, serine 264 (site 1), serine 271 (site 2), and serine 203 (site 3), on each of its ␣ subunits. Phosphorylation of any one site on either ␣

Academic Press

Key Words: pyruvate dehydrogenase complex; pyruvate dehydrogenase; point mutation; genetic defects;

1

To whom correspondence should be addressed at Department of Biochemistry, School of Medicine and Biomedical Sciences, State University of New York at Buffalo, 140 Farber Hall, 3435 Main Street, Buffalo, NY 14214. Fax: (716) 829-2725. E-mail: [email protected]. 0003-9861/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

2

Abbreviations used: PDC, pyruvate dehydrogenase complex; E1, pyruvate dehydrogenase; E2, dihydrolipoamide acetyltransferase; E3, dihydrolipoamide dehydrogenase; E3BP, E3-binding protein; PDK, pyruvate dehydrogenase kinase; PDP, phosphopyruvate dehydrogenase phosphatase; TPP, thiamine pyrophosphate; PCR, polymerase chain reaction; Ni-NTA, Ni-nitrilotriacetate-agarose; DTT, dithiothreitol; DCPIP, dichlorophenolindophenol; TCA, trichloroacetic acid. 121

122

JACOBIA, KOROTCHKINA, AND PATEL

subunit is sufficient for inactivation of the E1 component (and hence the entire PDC), while dephosphorylation of all three sites is necessary for reactivation of E1 (5). It has been shown previously that the phosphorylation rates of the serine residues are site-specific (5, 6). It was also shown that the order of dephosphorylation is random (6, 7). PDC deficiency results in variable clinical presentation in different patients. Patients can develop symptoms ranging from intermittent ataxia to a progressive disease with severe lactic acidosis, neurological dysfunction, and early death (8). Since glucose is the major source of energy in the brain and PDC deficiency disrupts the linkage between glycolysis and the tricarboxylic acid cycle, deficiency results in the reduction of energy production and thus dysfunction in the brain. The majority of PDC deficiencies results from mutations in the E1 component (approximately 130 patients) (8). Of these patients, approximately 70 have been shown to have point mutations in the E1 component, all of which were in the ␣ subunit. E1␣ deficiency is an X-linked disorder. Analysis of these point mutations may provide the biochemical basis for the structure-function relationship of E1. Fourteen male patients and one female patient (all unrelated except for one pair of brothers and a sister) were reported to have mutations of arginine-234 in the ␣ subunit of E1 to either glycine or glutamine (8 –14). The arginine-234 to glycine mutation reported in 12 patients and 4 of their mothers is the most common mutation found in E1␣. Three patients were found to have arginine-234 to glutamine mutation. PDC activity in cultured skin fibroblasts and tissue specimens of R234G E1 patients varied greatly (0 –100% activity of control) in different tissues from several patients investigated and differed among these patients (9, 10, 12, 13, 15). The amounts of E1␣ and E1␤ detected by immunoblotting in several of these patients were also variable (9, 15). Wexler et al. (9) suggested that the lack of activity and immunoreactivity was due to the loss of E1 ␣-␤ binding, which incurred loss of protein stability. Cultured fibroblast cells from two male patients, both of whom had an arginine to glutamine mutation at residue 234 in E1␣, displayed only 38% of PDC activity compared to controls (11). An earlier report suggested that the loss of activity may be due to the inability of the patient’s E1 to be dephosphorylated (16). The ramifications of these two substitution mutations (R234G and R234Q) in these E1-deficient patients are clear, but the biochemical basis of impaired PDC activity causing severe lactic acidosis and premature death in these male patients is not. It was the goal of this study to kinetically characterize recombinant E1 mutant proteins with the above two mutations.

EXPERIMENTAL PROCEDURES Plasmid constructs. Mutant E1␣ cDNA constructs were made using the polymerase chain reaction (PCR) and a megaprimer protocol with the recombinant expression vector pQE-9:E1␣/␤ as template (17, 18). The mutagenic primers (sense) and the 3⬘ primer (antisense) used to create the mutant megaprimers were as follows: 5⬘-GATGGAATGGATATCCTGTGCGTCGGAGAGGCAACAAGGTTTG-3⬘ (R234G E1), 5⬘-GATGGAATGGATATCCTGTGCGTCCAAGAGGCAACAAGGTTTG-3⬘ (R234Q E1), and 5⬘-GCGCGCGGATCCGTTAACTGACTGACTTAAACTT-3⬘ (3⬘ primer; BamHI site in bold). The product from these reactions was the ⬃1.1-kb E1␣ mutant cDNA, which was cloned into the BamHI sites of a pQE-9 vector. Colonies having an insert were analyzed by restriction digest with BamHI and PvuII, separately, to check for size of the insert and correct orientation. The pQE-9:E1␣ vector was then digested and a 1.85-kb fragment containing E1␣ cDNA was then ligated to a 3.8-kb pQE-3:E1␤ wild-type fragment (XhoI and PvuI digested). The resulting ligations formed the vectors pQE-9:E1 R234G ␣/␤ and pQE-9:E1 R234Q ␣/␤. The ligation mixtures were transformed into an Escherichia coli M15 cell line containing pDM1.1 and selected for with ampicillin and kanamycin. Presence of the ␣ and ␤ inserts and the size of the vectors were checked by digestion of the resulting plasmids with BamHI and XhoI, separately, respectively. The entire E1␣ cDNA was sequenced as well as the 5⬘-histidine tag. Overexpression and purification of human wild-type and mutant E1s. The wild-type and mutant E1 proteins encoded by these plasmids were overexpressed and purified as described previously with the following modification: after application of the clarified cell lysate to the Ni-nitriloacetate-agarose (Ni-NTA) column, it was washed overnight with potassium phosphate buffer, pH 7.0, 300 mM KCl, 50 mM imidazole (18). The protein was eluted with an imidazole gradient from 50 to 200 mM. Further purification of E1 was done using a DEAE-Sephadex A-25 column (18). Protein was measured by the BioRad protein assay with bovine serum albumin as the standard. Purity of the purified enzyme protein was measured by densiotometry following separation on a 12% SDS polyacrylamide gel and staining with Coomassie blue. Overexpression and purification of human E2-E3BP and human E3. In order to measure human E1 activity using PDC reconstitution assay, purified E2-E3 binding protein subcomplex (E2-E3BP) and E3 are required. E3 was overexpressed and purified by Dr. Hong of this lab as described previously (19) with a different expression system, XL1-Blue cells with pPROEX-E3 vector (Y. S. Hong and M. S. Patel, unpublished). Protein concentration and purity were measured as described above for E1. The E3 protein had a purity of ⬎95%. A coexpression vector containing the cDNAs encoding the human E2 component and the E3BP was a gift from Dr. Robert A. Harris (Indiana University School of Medicine). The E2-E3BP subcomplex was overexpressed in E. coli and purified based on methods previously published (20, 21). Protein concentration and purity were measured as described above. The E2-E3BP subcomplex had a purity of ⬎90%. Overexpression and purification of rat pyruvate dehydrogenase kinases: Isoforms 1 and 2. BL21(DE3) E. coli cells harboring the expression vectors containing the cDNAs for rat PDK1 and PDK2 individually were generous gifts from Dr. Robert A. Harris. Overexpression and purification of PDK1 and PDK2 were done as previously published with a few minor modifications as stated below (22). BL21(DE3) E. coli cells were grown in 2 liters of LB containing 35 ␮g/ml kanamycin at 37°C to an optical density of 0.6 and then they were induced with 200 ␮g/ml isopropyl ␤-thiogalactopyranoside overnight at 25°C. The cells were pelleted and resuspended in 50 mM potassium phosphate buffer, 100 mM KCl, 5 mM ␤-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM bensamidine, 0.25

ARGININE-234 E1 ␣ MUTATIONS

␮g/ml leupeptin. The cells were lysed by two passes through a French press at 16,000 psi. The clarified supernatant was loaded onto a Ni-NTA column preequilibrated with 50 mM potassium phosphate buffer, 100 mM KCl, 5 mM ␤-mercaptoethanol and washed overnight with 20 mM imidazole. The protein was eluted with a 20 –200 mM imidazole gradient. Fractions containing PDK were pooled, concentrated using an UltraFree-15 centrifugal filter from Amicon, and dialyzed against two changes of 2 liters of 50 mM potassium phosphate buffer, 5 mM DTT, 0.5 mM EDTA. PDK purity was determined to be ⬎95% as described above. Overexpression and purification of rat phospho-E1 phosphatase: Isoform I. A pET-28a vector containing the cDNA for rat PDP1 (pPDP) was a generous gift of Dr. Robert A. Harris. The PDP1 purification was a variation of Huang et al. (23). BL21(DE3) E. coli cells containing pPDP and pGroESL, a vector encoding the chaperonins GroES and GroEL, were selected for by growth overnight at 37°C on LB plates containing 50 ␮g/ml kanamycin and 35 ␮g/ml chloramphenicol. The induced cells were collected by centrifugation and resuspended in 50 mM potassium phosphate buffer, pH 8.0, 10 mM ␤-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM benzamidine, 20 ␮g/ml leupeptin, 1% aprotinin. The suspension was pressed through a French press twice at 1600 psi. The clarified supernatant was then applied to a Ni-NTA column (15 ml) previously equilibrated with the wash buffer, 50 mM potassium phosphate, pH 8.0, 300 mM KCl, and washed overnight with the addition of 10 mM imidazole. The protein was eluted from the column with an imidazole gradient from 10 to 100 mM. The fractions containing PDP1 were pooled, concentrated with an UltraFree-15 centrifugal filter from Amicon, and dialyzed against 50 mM potassium phosphate buffer, pH 8.0, 5 mM DTT. PDP1 protein concentration and purity were measured as described above. The purity of PDP1 was determined to be approximately 80%. Kinetic analysis. E1 activity was measured by 2,6-dichlorophenolindophenol (DCPIP) assay, where the reduction of the artificial electron acceptor, DCPIP, is measured at an absorbance of 600 nm (21, 24). Overall PDC activity was measured in reconstitution assay by measuring the formation of NADH when E1 was reconstituted with the E2-E3BP subcomplex and E3 (25). PDC reconstitution assays entailed incubating ⬃1 ␮g of E1 with excess amounts of both E2-E3BP and E3 (3 ␮g each) in 50 mM potassium phosphate, pH 7.5, 2 mM MgCl 2, 1.4 mM NAD ⫹, 500 ␮M TPP, 4 mM cysteine, and 156 ␮M CoASH in a 1-ml reaction volume. The reaction was started by the addition of pyruvate to a final concentration of 2 mM and the change in absorbance at 340 nm/min was measured to calculate activity. For measurement of the apparent K m and V max values for TPP and pyruvate, varying concentrations of TPP and pyruvate were added. The data were plotted on a Lineweaver-Burke plot using Sigma Plot (Jandel Scientific, CA) and a least-squares regression line was extrapolated from the data. Thermostability of E1 proteins at 37°C. Thermostability of the wild-type E1 and the R234Q E1 was measured in both the presence and absence of the E2-E3BP subcomplex. In the absence of the E2-E3BP subcomplex, E1 protein (2.1 ␮M) was incubated at 37°C in 50 mM potassium phosphate buffer, pH 7.0, with TPP concentrations from 0 to 1 mM (and 1 mM MgCl 2). In the presence of E2-E3BP subcomplex, E1 protein (0.8 ␮M, calculated as tetramer E1) was incubated with 0.05 ␮M E2-E3BP (calculated as 60-mer E1) at 37°C with 0 –1 mM TPP (and 1 mM MgCl 2). At various times, 1 ␮g of E1 protein was removed and reconstituted with 3 ␮g of E2-E3BP and 3 ␮g of E3, and PDC activity was measured. PDC inactivation assay. E1 phosphorylation by rat PDK2 in the absence of the E2-E3BP subcomplex was as published (6). Briefly, 10 ␮g of E1 was incubated in 20 mM potassium phosphate buffer, pH 7.0, containing 1 mM MgCl 2, 0.1 mM EDTA, 2 mM DTT, 50 ␮M ATP with 100 ng of PDK2 at 30°C. At various times, aliquots containing

123

1 ␮g of E1 were removed and activity was measured in a PDC reconstitution assay. Phosphorylation of E1 proteins. Incorporation of [ 32P]PO 4 from 32 [ P]ATP into E1 was measured as described previously (6). E1 (10 ␮g) was incubated in 20 mM potassium phosphate buffer, pH 7.0, containing 10 ␮g E2-E3BP, 1 mM MgCl 2, 0.1 mM EDTA, 2 mM DTT, 200 ␮M ATP (specific radioactivity of 460 –960 cpm/pmol ATP) with 10 –100 ng of recombinant rat PDK1 or recombinant rat PDK2 at 30°C. At various times, aliquots were removed and spotted on 10% trichloroacetic acid (TCA) presoaked filter papers, incubated in 10% TCA, 10 mM pyrophosphate solution for 1 h, washed three times with 10% TCA for 20 min, washed three times with ethanol for 5 min, and air-dried. The radioactivity associated with the dried filters was measured by a Beckman LS 6500 Multi-Purpose scintillation counter. For site-specific phosphorylation of site 3, E1 protein was first phosphorylated at sites 1 and 2 as noted above by PDK2 in the presence of the E2-E3BP subcomplex with 200 ␮M ATP (unlabeled). After sites 1 and 2 were fully phosphorylated, 100 –330 ng PDK1 and [ 32P]ATP (final specific radioactivity of 1460 –1500 cpm/pmol ATP) were added and incorporation of [ 32P]PO 4 into site 3 only was measured as noted above or the reaction was allowed to proceed till complete phosphorylation of site 3 to study dephosphorylation of site 3 only. PDC reactivation assay. Dephosphorylation of phospho-E1 was done as described previously (6). Phosphorylation of E1 was stopped by depletion of ATP by incubating 0.4 mM glucose and 2 ␮g of yeast hexokinase in the phosphorylation reaction mixture for 10 min at room temperature. Then 1.2 ␮g of rat PDP1 was added to the above reaction mixture with the addition of 10 mM MgCl 2 and 100 ␮M CaCl 2 and incubated at 30°C. Aliquots of 1 ␮g of E1 were taken at various times, and recovery of activity was measured by reconstitution assay. For both the phosphorylation and dephosphorylation assays, the data were plotted and the curves were fitted to the data using Sigma Plot (Jandel Scientific, CA). The k app values were determined from the curve by the program. Dephosphorylation of phospho-E1 proteins. Phosphorylation of site 3 was performed as described above. ATP was depleted by incubation with 1 mM glucose and 2 ␮g of yeast hexokinase for15 min. MgCl 2 and CaCl 2 were added to the phosphorylated E1 to final concentrations of 10 mM and 100 ␮M, respectively. The reactions were started by the addition of 0.2 ␮g PDP1. At various times, aliquots were removed, boiled in SDS-PAGE sample buffer, and separated on 10% polyacrylamide gels. The amount of the remaining 32 P-E1␣ was determined by phosphoimaging of the SDS-PAGE gels.

RESULTS

E1 Purification Expression and purification of wild-type E1 were performed as described under Experimental Procedures and resulted in ⬃15 mg of protein from 6 liters of cells with a purity of 75% after the Ni-NTA chromatography. Both the R234G E1 and the R234Q E1 mutant proteins showed the same elution profile from the Ni-NTA column as the wild type, albeit lesser amounts of mutant proteins (⬃4 and 6 mg of E1 from 12 liters of cells for R234G and R234Q E1s, respectively), with purities of ⬃50% for both. In order to obtain highly pure E1 preparations, DEAE-Sephadex A-25 chromatography was employed. The DEAE-Sephadex A-25 matrix worked well to remove contaminants, but it also resulted in a low recovery (⬃20%) of E1 protein. The

124

JACOBIA, KOROTCHKINA, AND PATEL TABLE I

Kinetic Parameters of Wild-Type and Mutant E1s Using PDC Reconstitution and DCPIP Assays PDC reconstitution assay

E1 Type

V max a (units/mg E1)

k cat (s ⫺1)

K mTPP a (nM)

WT R234G R234Q

26.5 ⫾ 5.0 33.6 ⫾ 4.7 25.7 ⫾ 4.4

68.0 86.2 66.0

73 ⫾ 13 497 ⫾ 25 101 ⫾ 4

a

(s

⫺1

k cat/K m M ⫺1 ⫻ 10 8) 9.3 1.7 6.5

K mPyr a (␮M) 50 ⫾ 15 44 ⫾ 3 41 ⫾ 3

(s

⫺1

k cat/K m M ⫺1 ⫻ 10 6) 1.4 2.0 1.6

DCPIP assay Activity a (mU/mg E1) 249 ⫾ 23 272 ⫾ 57 206 ⫾ 63

The results are means ⫾ SD of at least four individual determinations. K m values are apparent K m values.

wild-type E1, R234G E1, and R234Q E1 proteins were purified to ⬎97% homogeneity, as judged by densiotometry after separation on a 12% SDS-PAGE and staining with Coomassie blue (results not shown).

that the ability of the mutant proteins to bind to the E2-E3BP subcomplex varied little from the wild type (results not shown). Thermostability of E1 Proteins at 37°C

Kinetic Analysis The specific activities of wild-type, R234G, and R234Q E1 proteins were measured by DCPIP and PDC reconstitution assays. E1 activity measured by DCPIP assay showed that the R234G and R234Q E1 mutants displayed ⬃83 and ⬃110% activity, respectively, compared to wild-type E1 (Table I), showing that the R234 mutant E1 activities were not significantly different from the wild-type E1 activity. In PDC reconstitution assay, the V max values for the wild-type, R234G, and R234Q E1 proteins did not differ from one another (Table I). The apparent K m values of both TPP and pyruvate were determined in a PDC reconstitution assay (Table I). The apparent K m value for TPP of R234G E1 was approximately 7-fold higher than wild-type E1, while the apparent K m value for the R234Q E1 was similar to the wild-type. The apparent K m values for pyruvate were similar for all three E1s. These findings show that mutation of arginine-234 to glycine but not glutamine affects the ability of the mutant E1 to bind TPP. It is also of interest to note that the R234G E1 protein, though having a 7-fold increase in apparent K m for TPP, does not have an effect on pyruvate binding. This would suggest that the R234G mutation does not cause a gross conformational change, and that it may be an important residue for binding/positioning of TPP in the active site. The E1 heterotetramer binds to the E1-binding domain of the E2 component via its E1␤ subunit (26). It is unlikely that R234Q or R234G mutation significantly affected the ability of mutant E1s to bind to E2 as activities of mutant E1s did not differ from that of the wild-type E1. Additionally, titration of a constant amount of E2-E3BP subcomplex with a variable amount of E1 protein and subsequent measurement of PDC activity by reconstitution assay demonstrated

It was found that the R234Q E1 mutant in the absence of the E2-E3BP subcomplex was less stable at 37°C than was the wild type (Fig. 1). TPP concentrations greater than 130 ␮M were able to increase the stability of R234Q E1 to that of the wild-type (as measured by PDC activity) (Fig. 1). In the presence of the E2-E3BP subcomplex, a condition that mimics the in vivo environment, R234Q E1 displayed a similar thermostability as the wild-type E1 protein (Fig. 1). R234G E1 completely lost its activity after a 60-min incubation at 37°C in the absence of E2-E3BP, which was rescued by 0.5 mM TPP (101 ⫾ 6%, results not shown).

FIG. 1. Stability of wild-type and R234Q E1s at 37°C in the absence and presence of the E2-E3BP subcomplex. 䊐, Wild-type E1; E, R234Q E1; Œ, R234Q E1 with 13 ␮M TPP; ‚, R234Q E1 with 130 ␮M TPP; }, R234Q with 520 ␮M TPP; F, R234Q E1 in the presence of the E2-E3BP subcomplex. Aliquots of E1 (1 ␮g) incubated at 37°C for varying lengths of time were reconstituted into PDC and activity was measured by PDC reconstitution assay.

ARGININE-234 E1 ␣ MUTATIONS

FIG. 2. Inactivation of PDC by phosphorylation of E1. E, Wild-type E1; 䊐, R234G E1; ‚, R234Q E1. E1 proteins were incubated upto 15 min with PDK2 in the absence of the E2-E3BP and at various times, aliquots were removed (1 or 2 ␮g) and reconstituted into PDC and activity was measured by PDC reconstitution assay.

125

proteins. Figure 3B indicates that the rate of phosphorylation of site 2 of R234G was less than that for the wild-type E1, which may result in the increase of the amount of active E1. These results eliminated the possibility that an increase in the rate of phosphorylation of sites 1 or 2 by PDK2 in the presence of the E2-E3BP subcomplex was causing more inactive E1 protein in patients with either the R234Q or the R234G mutation (Figs. 3A and 3B). PDK1 in the presence of the E2-E3BP subcomplex phosphorylates sites 1, 2, and 3 (L. G. Korotchkina and M. S. Patel, unpublished data). Since PDK2 in the presence of the E2-E3BP subcomplex is able to phosphorylate sites 1 and 2, it was possible to specifically phosphorylate site 3 by PDK1 in the presence of [ 32P]ATP. This site 3-specific phosphorylation (Figs. 3C and 3D) was similar for the R234Q, R234G, and the wild-type E1s. Finally, this suggested that the lack of E1 activity found in the patients was not due to an

In the presence of E2-E3BP R234G was less stable than R234Q (80 ⫾ 6% after a 60-min incubation), but reached the wild-type stability in the presence of both E2-E3BP and 0.5 mM TPP (results not shown). In the absence and presence of the E2-E3BP subcomplex the wild-type E1 did not show any detectable loss of activity during 120 min of incubation at 37°C. PDC Inactivation and Phosphorylation of the E1 Proteins To measure the rate of inactivation by phosphorylation of site 1 only, E1 proteins were incubated with PDK2 in the absence of the E2-E3BP subcomplex. Only site 1 of E1 was shown previously to be phosphorylated when using this protocol. The results in Fig. 2 show that both mutant E1 proteins were phosphorylated/ inactivated at a rate similar to that of wild-type protein. This suggested that an increase in the rate of phosphorylation of site 1 of these mutant proteins was not the cause for reduction in PDC activity in the R234 patients reported. With the knowledge that the rate of phosphorylation of site 1 was not different between the wild-type and the mutant E1 proteins, investigation of the phosphorylation of site 2 was done by measuring the incorporation of [ 32P]PO 4 from [ 32P]ATP into sites 1 and 2 of E1 by PDK2 in the presence of the E2-E3BP subcomplex. Under the experimental conditions used, PDK2 phosphorylated only sites 1 and 2 of E1 (L. G. Korotchkina and M. S. Patel, unpublished data). Figure 3A shows that the rate of incorporation of [ 32P]PO 4 into sites 1 and 2 was similar for the wild-type and R234Q E1

FIG. 3. Phosphorylation of wild-type, R234Q, and R234G E1s. (A) Phosphorylation of sites 1 and 2 of wild-type and R234Q E1s by PDK2. (B) Phosphorylation of sites 1 and 2 of wild-type and R234G E1s by PDK2. (C) Site 3-specific phosphorylation by PDK1 after phosphorylation by PDK2 of wild-type and R234Q E1s. (D) Site 3-specific phosphorylation by PDK1 after phosphorylation by PDK2 of wild-type and R234G E1s. E, Wild-type E1; ‚, R234Q E1 and 䊐, R234G E1. Ten micrograms of E1 was phosphorylated by 10 ng of PDK2 (A), 100 ng of PDK2 (B), 100 ng of PDK1 after phosphorylation with 100 ng of PDK2 (C), and 330 ng of PDK1 after phosphorylation with 100 ng of PDK2 (D) as described under Experimental Procedures. Aliquots of 1 ␮g E1 were removed from the phosphorylation reaction mix, placed on TCA filters, and washed, and incorporated radioactive phosphate was measured in a scintillation counter.

126

JACOBIA, KOROTCHKINA, AND PATEL

and then PDK1 was used to specifically phosphorylate site 3 with [ 32P]ATP. This site 3-specific 32P-labeled phospho-E1 was then used as a substrate for PDP1catalyzed dephosphorylation. R234G E1 was dephosphorylated with the rate not significantly different from that of the wild-type E1 (88%). However, the dephosphorylation rate by PDP1 of site 3 for the R234Q mutant E1 was only 50% compared to the wild-type protein (Fig. 4B). This finding in conjunction with the fact that phosphorylation of any one of the phosphorylation sites causes inactivation of the E1 component demonstrates that the reduction in dephosphorylation of R234Q mutant protein provides the biochemical basis for the reduced E1 activity (and hence PDC activity) in skin fibroblasts of these two patients. DISCUSSION

FIG. 4. Dephosphorylation of sites 1, 2, and 3 of wild-type, R234G, and R234Q E1s by PDP1. (A) PDC activity after dephosphorylation of site 1 and site 2 of phospho-E1 by PDP1. (B) Site 3-specific dephosphorylation by PDP1. E, Wild-type E1; 䊐, R234G E1; ‚, R234Q E1. Aliquots of E1 (1 ␮g) from the dephosphorylation reaction mixture were separated on a 10% SDS acrylamide gel and dried, and the amount of residual radioactive phosphate was determined.

increase in the rate of phosphorylation/inactivation of any of the three phosphorylation sites. PDC Dephosphorylation and Reactivation of the Phospho-E1 Proteins The ability of PDP1 to dephosphorylate sites 1 and 2 of E1 was investigated for the wild-type and mutant E1 proteins. It is known that dephosphorylation of all 3 phosphorylation sites is necessary for reactivation of the E1 component and thus PDC activity (6). Figure 4A represents the reactivation of phosphorylated E1 (sites 1 and 2) measured by PDC reconstitution assay after incubation of phospho-E1 with PDP1 for varying lengths of time. It was found that the R234Q and R234G mutant E1 proteins were dephosphorylated/ reactivated at sites 1 and 2 by PDP1 at rates similar to that of wild-type E1 (k app of reactivation were 0.160 min ⫺1 for the wild-type E1, 0.156 min ⫺1 for R234Q E1, and 0.138 min ⫺1 for R234G E1) (Fig. 4A). To investigate the dephosphorylation rate for site 3, sites 1 and 2 were phosphorylated with unlabeled ATP by PDK2 in the presence of the E2-E3BP subcomplex

There are a total of 24 arginine residues in mature human E1␣. Some of these residues have been identified to be mutated in patients displaying E1 deficiency, namely R59S, R98W, R98Q, R112Q, R234G, R234Q, R259H, R273C, R273H, R273L, R349C, and R349H (8 –15, 27–31). Peripheral mononuclear cells from one patient with a R59S mutation in E1␣ displayed E1 activities of only about 7% compared to control and is thought to also play a role in E1 stability (15). Mutation of R98W caused an instability of the E1 protein with E1 activities of only about 10% (27). Patients having the R259H mutation had normal levels of E1 protein; however, they had low levels of E1 activity (32). Expression of R273H E1␣ mutant protein in human fibroblasts lacking endogenous E1␣ mRNA showed that the R273H E1 protein was devoid of activity (29). Five female patients with an R349C mutation have been identified. This mutation is thought to affect the stability of E1 (30). However, the biochemical basis for these mutations is not elucidated. In order to elucidate the biochemical basis for the two specific substitutions at R234, we used site-directed mutagenesis to recreate these mutations. Upon kinetic characterization of the V max values, apparent K m values for TPP, and the apparent K m values for pyruvate, we found that the wild-type and the mutant R234G E1 proteins only differed in their apparent K m value for TPP. This difference, a 7-fold increase in the apparent K m for TPP for R234G E1, is sufficient to cause symptoms in patients. The TPP concentration in intact cells has been determined to be approximately 200 nM (33) but its concentration in the mitochondria is not known. Assuming that a similar level of TPP exists in mitochondria, R234G mutant E1 having K m of about 500 nM will not even reach its half-maximal activity. Additionally, this mutation could change the local conformation of the E1 mutant causing its less stability than the wild-type and R234Q mutant E1.

ARGININE-234 E1 ␣ MUTATIONS

The previously published studies on the R234G E1 mutant protein from patients have varied considerably between groups and between patients studied within groups (15). This may have resulted from instability of the mutant protein in cultured skin fibroblasts and tissue specimens obtained at autopsy. Interestingly, substitution of the arginine-234 by a glutamine residue did not affect the apparent K m value for TPP significantly or any of the other kinetic parameters tested. It then became necessary to investigate the ability of this mutant protein, R234Q E1, to bind to the E2-E3BP subcomplex, its ability to be phosphorylated/dephosphorylated, and its thermostability. The thermostability of R234Q E1 was decreased in the absence of the E2-E3BP subcomplex but in the presence of the E2-E3BP subcomplex its stability was similar to that of the wild type (Fig. 1). Furthermore, the rates of phosphorylation of sites 1, 2, and 3 of this mutant protein were similar to those of the wild-type E1 (Figs. 2 and 3). The rates of dephosphorylation of sites 1 and 2 were similar for the wild-type and R234Q E1 proteins (Fig. 4A); however, the rate of dephosphorylation of site 3 for the R234Q E1 protein was only about 50% of the wild type (Fig. 4B). The study of deficiency of phospho-E1 phosphatase in patients with congenital lactic acidemia resulted in the finding that a 35–50% decrease (compared to control) in the phosphatase activity was the cause for deficiency in these patients (34). This supports the claim that a 50% decrease in the rate of dephosphorylation of site 3 was enough to cause E1 deficiency in the R234Q E1 patients. It is known that phosphorylation of just one of the three phosphorylation sites is sufficient for inactivation of the E1 component (and thus PDC), whereas dephosphorylation of all three sites is necessary for reactivation of the protein (5). Kitano et al. (16) investigated the phosphorylation state of E1 protein in two patients and controls by comparing the migration of all of the PDC components on 2D-SDS gel electrophoresis. In their study, they found that patient extracts had more E1 protein in the phosphorylated form (16). Our study expands upon the hypothesis of Kitano et al. (16), who first suggested an inability of dephosphorylation of the E1 protein from patients displaying primary lactic acidemia. These patients were later found to have the R234Q E1 mutation (11). Our findings show that the reduction in dephosphorylation of site 3 was the primary cause of reduced PDC activity in the R234Q E1 patients. In the absence of a crystal structure for human E1, it is not possible to comment on the nature of interaction between site 3 and R234 in these two patients. R234 is conserved in mammalian E1␣ (including somatic and testis-specific isoforms) and nematode E1␣ but not in E1 of yeast, plant, or gram-positive bacterial PDC and not in BCKDC (9). It was suggested previously that R234 belongs to the stretch of amino acid residues

127

forming the ␣-helical structure (225–248) with positively charged residues positioned on one side of the helix (9). The helix structure was proposed to be conserved within the E1␣ of different PDC and BCKDC. Interestingly the conservation of R234 correlates with the presence of phosphorylation site 3 that is found in mammalian E1␣-PDC and absent in yeast and plant E1-PDC and BCKDC. The only exception is the nematode E1, which has R234 in E1␣ of type I and type II, but S203 (site 3) is present in type II only but not in type I E1␣ (GenBank: AAA29377.1 and AAA29376.1). The reason for this exception remains elusive. In summary, our findings clearly show that different substitutions at the same amino acid position (i.e., arginine-234 to either glycine or glutamine) can cause different functional defects in the E1 protein. ACKNOWLEDGMENTS This study was supported in part by U.S. Public Health Service Grant DK20478. We are grateful to Dr. Young Soo Hong of this laboratory for supplying highly pure E3. We also thank Drs. Murray Ettinger and Daniel Kosman of this department for their helpful discussions throughout this study.

REFERENCES 1. Reed, L. J. (1974) Acc. Chem. Res. 7, 40 – 46. 2. De Marcucci, O., and Lindsay, J. G. (1985) Eur. J. Biochem. 149, 641– 648. 3. Jilka, J. M., Rahmatullah, M., Kazemi, M., and Roche, T. E. (1986) J. Biol. Chem. 261, 1858 –1867. 4. Patel, M. S., and Roche, T. E. (1990) FASEB J. 4, 3224 –3233. 5. Yeaman, S. J., Hutcheson, E. T., Roche, T. E., Pettit, F. H., Brown, J. R., Reed, L. J., Watson, D. C., and Dixon, G. H. (1978) Biochemistry 17, 2364 –2370. 6. Korotchkina, L. G., and Patel, M. S. (1995) J. Biol. Chem. 270, 14297–14304. 7. Teague, W. M., Pettit, F. H., Wu, T. L., Silberman, S. R., and Reed, L. J. (1982) Biochemistry 21, 5585–5592. 8. Lissens, W., De Meirleir, L., Seneca, S., Liebaers, I., Brown, G. K., Brown, R. M., Ito, M., Naito, E., Kuroda, Y., Kerr, D. S., Wexler, I. D., Patel, M. S., Robinson, B. H., and Seyda, A. (2000) Hum. Mutat. 15, 209 –219. 9. Wexler, I. D., Hemalatha, S. G., Liu, T. C., Berry, S. A., Kerr, D. S., and Patel, M. S. (1992) Pediatr. Res. 32, 169 –174. 10. Chun, K., MacKay, N., Petrova-Benedict, R., and Robinson, B. H. (1993) Hum. Mol. Genet. 2, 449 – 454. 11. Awata, H., Endo, F., Tanoue, A., Kitano, A., and Matsuda, I. (1994) J. Inherit. Metab. Dis. 17, 189 –195. 12. Chun, K., MacKay, N., Petrova-Benedict, R., Federico, A., Fois, A., Cole, D. E., Robertson, E., and Robinson, B. H. (1995) Am. J. Hum. Genet. 56, 558 –569. 13. Lissens, W., Desguerre, I., Benelli, C., Marsac, C., Fouque, F., Haenggeli, C., Ponsot, G., Seneca, S., Liebaers, I., and De Meirleir, L. (1995) Hum. Mol. Genet. 4, 307–308. 14. Kerr, D. S., Wexler, I. D., Tripatara, A., and Patel, M. S. (1996) in Alpha-Keto Acid Dehydrogenase Complexes (Patel, M. S., Roche, T. E., and Harris, R. A., Eds.), pp. 249 –269, Birkhauser, Basel.

128

JACOBIA, KOROTCHKINA, AND PATEL

15. Marsac, C., Benelli, C., Desguerre, I., Diry, M., Fouque, F., De Meirleir, L., Ponsot, G., Seneca, S., Poggi, F., Saudubray, J. M., Zabot, M. T., Fontan, D., and Lissens, W. (1997) Hum. Genet. 99, 785–792. 16. Kitano, A., Endo, F., and Matsuda, I. (1990) Neurology 40, 1312–1314. 17. Landt, O., Grunert, H. P., and Hahn, U. (1990) Gene 96, 125–128. 18. Korotchkina, L. G., Tucker, M. M., Thekkumkara, T. J., Madhusudhan, K. T., Pons, G., Kim, H., and Patel, M. S. (1995) Protein Expression Purif. 6, 79 –90. 19. Liu, T.-C., Korotchkina, L. G., Hyatt, S., Vettakkorumakankav, N. N., and Patel M. S. (1995) J. Biol. Chem. 270, 15545–15550. 20. Yang, D., Song, J., Wagenknecht, T., and Roche, T. E. (1997) J. Biol. Chem. 272, 6361– 6369. 21. Korotchkina, L. G., and Patel, M. S. (2001) J. Biol. Chem. 276, 5731–5738. 22. Bowker-Kinley, M. M., Davis, W. I., Wu, P., Harris, R. A., and Popov, K. M. (1998) Biochem. J. 329, 191–196. 23. Huang, B., Gudi, R, Wu, P., Harris, R. A., Hamilton, J., and Popov, K. M. (1998) J. Biol. Chem. 273, 17680 –17688. 24. Khailova, L. S., Bernhardt, R., and Huebner, G. (1977) Biokhimiia (Russian) 42, 113–117. 25. Roche, T. E., and Reed, L. J. (1972) Biochem. Biophys. Res. Commun. 48, 840 – 846.

26. Wynn, R. M., Chuang, J. L., Davie, J. R., Fisher, C. W., Hale, M. A., Cox, R. P., and Chuang, D. T. (1992) J. Biol. Chem. 267, 1881–1887. 27. Fujii, T., Van Coster, R. N., Old, S. E., Medori, R., Winter, S., Gubits, R. M., Matthews, P. M., Brown, R. M., Brown, G. K., and Dahl, H. H. (1994) Ann. Neurol. 36, 83– 89. 28. Lissens, W., Mahmoud, K. Z., El-Gindi, E., Abdel-Sattar, A., Seneca, S., Van Steirteghem, A., and Liebaers, I. (1999) Mol. Hum. Reprod. 5, 10 –13. 29. Otero, L. J., Brown, R. M., and Brown, G. K. (1998) Hum. Mutat. 12, 114 –121. 30. Fujii, T., Garcia Alvarez, M. B., Sheu, K. F., Kranz-Eble, P. J., and De Vivo, D. C. (1996) Pediatr. Neurol. 14, 328 –334. 31. Wexler, I. D., Hemalatha, S. G., McConnell, J., Buist, N. R., Dahl, H. H., Berry, S. A., Cederbaum, S. D., Patel, M. S., and Kerr, D. S. (1997) Neurology 49, 1655–1661. 32. Lissens, W., Vreken, P., Barth, P. G., Wijburg, F. A., Ruitenbeek, W., Wanders, R. J., Seneca, S., Liebaers, I., and De Meirleir, L. (1999) Eur. J. Pediatr. 158, 853– 857. 33. Bettendorff, L., Peeters, M., Jouan, C., Wins, P., and Schoffeniels, E. (1991) Anal. Biochem. 198, 52–59. 34. Ito, M., Kobashi, H., Naito, E., Saijo, T., Takeda, E., Huq, A. H., and Kuroda, Y. (1992) Clin. Chim. Acta 209, 1–7.