Inhibition of the Escherichia coli Pyruvate Dehydrogenase Complex E1 Subunit and Its Tyrosine 177 Variants by Thiamin 2-Thiazolone and Thiamin 2-Thiothiazolone Diphosphates EVIDENCE FOR REVERSIBLE TIGHT-BINDING INHIBITION*

Variants of the pyruvate dehydrogenase subunit (E1; EC 1.2.4.1) of the Escherichia coli pyruvate dehydrogenase multienzyme complex with Y177A and Y177F substitutions were created. Both variants displayed pyruvate dehydrogenase multienzyme complex activity at levels of 11% (Y177A E1) and 7% (Y177F E1) of the parental enzyme. The K m values for thiamin diphosphate (ThDP) were 1.58 (cid:1) M (parental E1) and 6.65 (cid:1) M (Y177A E1), whereas the Y177F E1 variant was not saturated at 200 (cid:1) M . According to fluorescence studies, binding of ThDP was unaffected by the Tyr 177 substitutions. The ThDP analogs thiamin 2-thiazolone diphosphate (ThTDP) and thiamin 2-thiothiazolone tight-binding the Y177A and Y177F variants. This analysis revealed that ThTDP and ThTTDP bound to parental E1 via a two-step mechanism, but that ThTDP bound to the Y177A variant via a one-step mechanism. Binding of ThTDP was affected and that of ThTTDP was unaffected by substitutions at Tyr 177 . Addition of ThDP or ThTDP to parental E1 resulted in similar CD spectral changes in the near-UV region. In contrast, binding of ThTTDP to either parental E1 or the Y177A and Y177F variants was accompanied by the appearance of a positive band at 330 nm, indicating that ThTTDP was bound in a chiral environment. In combination with x-ray structural evidence on the location of Tyr 177 , the kinetic and spectro-scopic data suggest that Tyr 177 has a role in stabilization of some transition state(s) in the reaction pathway, starting with the free enzyme and culminating with the first irreversible step (decarboxylation), as well as in reductive acetylation of the dihydrolipoamide acetyltransferase component.

Equipped with an excellent overexpression system for the E1 subunit, we decided to reexamine this issue, in part, to determine whether any special structural features of E. coli PDHc E1 are responsible for the highly specific inhibition. We now present data indicating that the kinetics of inhibition of E. coli PDHc E1 by ThTDP and ThTTDP reflect properties of tightbinding reversible inhibitors.
The recently solved three-dimensional structure of E. coli PDHc E1 revealed the presence of a tyrosine (Tyr 177 ) near the ThDP site, in a highly conserved region of the bacterial E1 subunits. 2 Tyrosine residues involved in ThDP binding were also identified in the crystal structures of human branchedchain ␣-ketoacid dehydrogenase (11), 2-oxoisovalerate dehydrogenase from Pseudomonas putida (12), and benzoylformate decarboxylase from P. putida (13). Therefore, the Y177A and Y177F variants of E. coli PDHc E1 were created, and the binding of ThTDP and ThTTDP to parental E1 and the Y177A and Y177F variants was studied using progress curve analysis, circular dichroism, and fluorescence spectroscopy. The results indicate that Tyr 177 in E. coli PDHc E1 probably interacts with the covalent adduct(s) formed with ThDP during the reaction cycle, but not with ThDP itself. Finally, we report the observation and characterization of a new CD signature, never before reported, formed upon binding of ThTTDP to either the PDHc or the E1 subunit.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Growth Conditions-E. coli strain JRG3456 (aceE mutant with a chloramphenicol resistance cassette in the aceE gene, whereas the E2 and E3 genes are intact; provided by the University of Sheffield) transformed with pGS878 was used for overexpression of the aceE gene encoding the E1 subunit of E. coli PDHc.
Construction of the Y177A and Y177F Variants of E1-Plasmid DNA was purified according to the Wizard 373 DNA purification system (Promega). Mutagenesis reactions were carried out using the QuikChange site-directed mutagenesis kit (Stratagene). Polymerase chain reaction was carried out for 16 cycles in a mini-cycler (MJ Research, Inc.) using double-stranded pGS878 DNA, two synthetic mutagenic primers complementary to opposite strands of the DNA, and the reagents supplied with the QuikChange site-directed mutagenesis kit. The following synthetic oligonucleotides (and their complements) were used as mutagenic primers (mismatched bases are underlined, and mutant codons are shown in boldface type): 5Ј-TGGCCTCTCTTCCTT-TCCGCACCCGA-3Ј (for Y177F) and 5Ј-TGGCCTCTCTTCCGCTCCG-CACCCGA-3Ј (for Y177A).
Overexpression of Parental PDHc E1 and the Y177A and Y177F Variants-E. coli cells transformed with the corresponding plasmids were grown for 16 h at 37°C in LB medium containing 50 g/ml ampicillin and were used to inoculate 1000 ml of the same medium. The cells were grown to A 650 ϭ 0.6 -0.7 and then induced with isopropyl-␤-D-thiogalactopyranoside (1 mM final concentration), harvested after 5-6 h, washed with 20 mM KH 2 PO 4 (pH 7.0), and stored at Ϫ20°C.
Purification of E1 and the Y177A and Y177F Variants-Purification of parental E1 and the tyrosine variants was carried out following the protocol described previously (14). E1 was also resolved from a singlelipoyl domain PDHc as reported previously (2).
Synthesis of Thiamin 2-Thiazolone and Thiamin 2-Thiothiazolone Diphosphates-This was carried out following previously described procedures (6,15). Orthophosphoric acid (85%, 5 ml in a 50-ml roundbottom flask) was polymerized by heating until the solution became faintly cloudy. Upon cooling to room temperature, a clear syrupy liquid resulted, to which was added thiamin 2-thiothiazolone (1.0 g). This mixture was then heated at 100°C for 15 min while being stirred with a glass rod and cooled to room temperature. A mixture of 4 ml of water and 1 ml of phosphoric acid was added, and the mixture was stirred for 2 h. Next, the pH was adjusted to 2.0 with 6 N NaOH, and the solution was applied to a column of Amberlite CG-50 H ϩ cation-exchange resin and eluted with water, with the fractions being monitored by TLC on cellulose plates. Activity and Related Measurements-The E1 activity was measured by reconstituting holo-PDHc activity with added E2⅐E3 subcomplex (the molar ratio of E1 to E2⅐E3 subcomplex was 1:5) using a Varian DMS 300 spectrophotometer or a Cobas-Bio centrifugal analyzer (Roche Molecular Biochemicals), monitoring the pyruvate-dependent reduction of NAD ϩ at 340 nm. The reaction medium contained (in a 1-ml (DMS 300) or 0.25-ml (Cobas Bio) test volume) 0.1 M Tris-HCl (pH 8.0), 1 mM MgCl 2 , 2 mM sodium pyruvate, 2.5 mM NAD ϩ , 0.1-0.2 mM CoA, 0.2 mM ThDP, and 2.6 mM dithiothreitol at 30°C. The reaction was initiated by adding CoA. Steady-state velocities were taken from the linear portion of the progress curve. One unit of activity is defined as the amount of NADH produced (mol/min/mg of E1).
Inhibition of Parental E1 and the Y177A and Y177F Variants by Thiamin 2-Thiazolone and Thiamin 2-Thiothiazolone Diphosphates-ThTDP and ThTTDP were dissolved in 20 mM KH 2 PO 4 (pH 7.0), and their concentrations were determined using ⑀ 235 ϭ 10,600 M Ϫ1 cm Ϫ1 for ThTDP and ⑀ 319 ϭ 10,900 M Ϫ1 cm Ϫ1 for ThTTDP (4). Parental E1 (0.035-0.10 M) and the tyrosine variants (0.22-0.44 M) were incubated in 50 mM KH 2 PO 4 (pH 7.0) with 2 mM MgCl 2 and 0.002-3.5 M ThTDP or 0.005-3.5 M ThTTDP in a total volume of 0.2 ml at 25°C. After 30 min, the E2⅐E3 subcomplex and all components required for assaying the PDHc activity were added to a volume of 1.0 ml, and the reaction was started by the addition of CoA.
Circular Dichroism-CD spectra were recorded on an Aviv Model 202 CD spectrometer at 25°C. A 1-cm path length quartz cuvette was used for the near-UV region (250 -350 nm). Data were collected at a wavelength step of 1.0 nm, an integration time of 1 s, and a bandwidth of 1 nm. The instrument was calibrated using (ϩ)-10-camphorsulfonic acid as a reference standard. The CD signal was recorded as [⌰], where [⌰] is the ellipticity measured in millidegrees. Prior to the CD experiment, parental E1 and its tyrosine variants were chromatographed on a Sephadex G-25 column equilibrated with 10 mM KH 2 PO 4 (pH 7.0) to exclude ThDP. For recording CD spectra, parental E1 (1.5 mg/ml) and the tyrosine variants (0.33-1.8 mg/ml) were dissolved in 10 mM KH 2 PO 4 (pH 7.0) containing 1 mM MgCl 2 and variable concentrations of ThTDP or ThTTDP. The base line was normally recorded using 10 mM KH 2 PO 4 . Data fitting and calculations utilized the KaleidaGraph, SigmaPlot, and DeltaGraph computer programs.
Fluorescence Spectroscopy-The fluorescence spectra of parental E1 and its tyrosine variants were recorded at 25°C using an SLM8100 spectrofluorometer. The excitation wavelength was 290 nm, and the emission spectra were recorded in the 300 -450-nm range in 3-ml quartz cuvettes. The integration time was 0.5 s, and the scan rate was 1 nm/min. The concentration of parental E1 and its tyrosine variants was 0.05 mg/ml in 20 mM KH 2 PO 4 (pH 7.0). The excitation and emission monochromator slit widths were 4 nm. The inner filter effect was corrected with the absorption spectrum of ThDP, ThTDP, or ThTTDP with Equation 1 (16,17), where F i(corr) is the corrected value of the fluorescence intensity at a given point in the titration, F i(obs) is the experimentally measured fluorescence intensity, V o is the initial volume of sample, V i is the volume at a given point of titration (V i /V o is the dilution factor), A ex is the absorbance of the sample at the excitation wavelength, and A em is the absorbance at the emission wavelength of protein fluorescence. The K d value for ThDP was calculated using Equation 2 (17), where (⌬F/F o ) ϫ 100 is the percent fluorescence quenching following the addition of ThDP. The concentrations of ThTDP and ThTTDP used for the titration were comparable to the concentration of enzyme. Quadratic Equation 3 (18) was used to calculate K d values, where (⌬F/F o )/(⌬F max /F o ) is the relative fluorescence; E t and L are the total concentrations of E1 and ThTDP (or ThTTDP) used for the titration, respectively; K d is the dissociation constant; and ((⌬F/F o )/(⌬F max / F o )) ϫ 100 is the percent of fluorescence quenched.

RESULTS
Characterization of the Y177A and Y177F Variants of PDHc E1-Site-directed substitutions were introduced into the E. coli E1 subunit at Tyr 177 in the highly conserved region of bacterial E1 (see BLAST alignment in Fig. 1), and the Y177A and Y177F variants were created (Table I). Furthermore, a BLAST search for short nearly exact matches did not reveal any related sequences in the heterotetrameric ␣ 2 ␤ 2 E1 subunits, so the highly conserved sequence in Fig. 1 is present only in the ␣ 2 ho-modimers. The activities of parental E1 and the Y177A and Y177F variants measured with different assays are presented in Table I. After reconstitution with the E2⅐E3 subcomplex, compared with parental E1, the Y177A variant retained 11% and the Y177F variant retained 7% activity according to the overall PDHc reaction assay in the presence of 0.2 mM ThDP. The catalytic constant (k cat ) was ϳ10-fold lower for the Y177A variant compared with the value for parental E1 (Table I). Similar relative activity (7%) was obtained for the Y177A variant using 4,4Ј-dithiodipyridine as a substitute for the E2 in PDHc, an E1-specific assay developed in this laboratory (data not shown) (19). According to the 2,6-DCPIP assay (a different E1-specific assay that measures E1 activity in the absence of the E2⅐E3 subcomplex as a second substrate and that is used extensively in the study of ThDP-dependent enzymes (9,20,21)), compared with parental E1, the activities were 20% (Y177A E1) and 49% (Y177F E1) with pyruvate as substrate and 49% (Y177A E1) and 68% (Y177F E1) with 2␣-hydroxyethyl-ThDP as substrate. These results suggest that both the interaction of E1 with the E2⅐E3 subcomplex and, to a lesser extent, the decarboxylation of pyruvate (the latter concluded on the basis of the 2,6-DCPIP experiments) are affected by the substitutions. The steady-state kinetic parameters for ThDP and pyruvate with parental E1 and the Y177A and Y177F variants are presented in Table II. The apparent K m for ThDP measured in the overall PDHc reaction in the presence of 0.2-200 M ThDP was 6.65 M for the Y177A variant compared with 1.58 M for parental E1. The Y177F variant was not saturated by ThDP even at a concentration Ͼ200 M, indicating that its activity could be increased with elevated ThDP concentrations (data not shown). A weak positive cooperativity (n H ϭ 1.38) was observed for ThDP binding to parental E1, which, however, changed to a negative cooperativity (n H Ͻ 1.0) for the Y177F variant. We previously reported that for singlelipoyl domain PDHc (a single lipoyl domain rather than the three lipoyl domains found in wild-type E. coli PDHc), variants with the G231A, N258Q, and C259S substitutions in the ThDPbinding fold of the E1 subunit were not saturated by concentrations of ThDP as high as 5-10 mM and exhibited negative cooperativity at all concentrations of ThDP studied (2). Similar K m values for pyruvate were obtained for parental E1 (0.515 mM) and the Y177A (0.280 mM) and Y177F (0.531 mM) variants    (Fig. 2, A and B) were required to inactivate parental E1. With the Y177A variant, 1.61-1.75 mol of ThTDP/mol of Y177A E1 monomer and 1.34 -1.55 mol of ThTTDP/mol of Y177A E1 monomer were required for inactivation. The Y177F variant gave a molar ratio of inactivation by ThTDP similar to that observed with the Y177A variant. 2) According to the Ackermann-Potter plot, using E1 resolved from single-lipoyl domain PDHc, ThTDP was found to behave as a tight-binding inhibitor (Fig. 3). A plot of velocity versus concentration of E1 for different concentrations of ThTDP was curved with an asymptote (Fig. 3). According to the Ackermann-Potter plot (22), the asymptote intersects the x axis at E t ϭ I t . As shown in Fig. 3 (inset), for a ThTDP concentration of 0.03 M, the total concentration of E1 titrated was 0.035 M, indicating that the concentration of Th-TDP required for inhibition is comparable to the concentration of enzyme, a diagnostic for tight-binding inhibition. When ThTTDP was tested at a concentration of 0.10 M, the total amount of E1 titrated was ϳ0.021 M, indicating that at least a 5-fold molar excess of ThTTDP is required to inhibit E1 (data not shown). 3) Inhibition of parental E1 by ThTDP and ThTTDP was reversible. About 10% (ThTDP inactivation, followed by 16 h of dialysis) and 34% (ThTTDP inactivation, followed by 48 h of dialysis) of the activity of parental E1 was recovered upon dialysis against 50 mM potassium phosphate buffer (pH 7.5).
Determination of K i Values for ThTDP and ThTTDP Using Reaction Progress Curve Analysis-The progress of the overall PDHc reaction was monitored by recording the formation of NADH at 340 nm as a function of time (23,24). A family of progress curves for the reaction, started by the addition of parental E1 and the E2⅐E3 subcomplex to assay medium containing 30 M ThDP and different concentrations of Th-TDP, is shown in Fig. 4A. In the absence of ThTDP, a linear rate of NADH release was observed during the first 10 min of the reaction (Fig. 4A). In the presence of ThTDP (Fig. 4A) and ThTTDP (data not shown), a time-dependent decrease in NADH production was observed, indicating a slow attainment of equilibrium between parental E1, ThDP, and ThTDP or parental E1, ThDP, and ThTTDP. Since the concentrations of ThTDP and ThTTDP were at least 10-fold higher than the concentration of parental E1, resulting in the conversion of Ͻ5% of pyruvate to product (P) during the reaction, the progress curves were treated according to Equation 4, where v o is the initial velocity, v s is the steady-state velocity, and k is the pseudo first-order rate constant for the approach to the steady-state phase. The progress curve analysis revealed that the initial velocity of NADH production (v o ) decreased with increasing concentration of ThTDP or ThTTDP (data not shown), indicating that the equilibria drawn below are reached rapidly: E1 ϩ ThDP 7 E1⅐ThDP and E1 ϩ ThTDP 7 E1⅐ThTDP or E1 ϩ ThTTDP 7 E1⅐ThTTDP. The dependence of the pseudo first-order rate constant (k app ) on the concentration of ThTDP or ThTTDP was hyperbolic, indicating saturation (Fig. 4, B and C). The data suggest that the inhibition of parental E1 by ThTDP and ThTTDP follows mechanism A (Scheme 1) (24, 25) for competitive tight-binding inhibitors. According to mechanism A, the rate of combination between parental E1 and ThDP or ThTDP (or parental E1 and ThDP or ThTTDP) may be similar, but the binding of ThTDP or ThTTDP to parental E1 is accompanied by a slow conformational change of the E1⅐ThTDP* and E1⅐ThTTDP* complexes to the E1⅐ThTDP and E1⅐ThTTDP complexes. The rate constants (k 2 and k Ϫ2 ) and K i values (k Ϫ1 /k 1 ) that characterize the equilibrium between parental E1 and the E1⅐ThTDP and E1⅐ThTTDP complexes were calculated using Equation 5 and are presented in Table III.
The progress curves were also recorded for the Y177A variant in the presence of ThTDP. A plot of k app against ThTDP concentration is linear, indicating a simpler mechanism of inhibition than that outlined for parental E1 (Fig. 5). Mechanism B describes the inhibition of the Y177A variant by ThTDP (Scheme 2) (24, 25).
According to mechanism B, the Y177A E1 ϩ ThTDP 7 Y177A E1⅐ThTDP equilibrium is reached at a significantly slower rate compared with the E1 ϩ ThDP 7 E1⅐ThDP equilibrium, indicating that the Y177A substitution in the E1 subunit affects ThTDP binding. A plot of k app against ThTDP concentration was fitted to Equation 6, providing the following values: k 1 ϭ 0.221 Ϯ 0.014 M Ϫ1 min Ϫ1 , k Ϫ1 ϭ 0.01 Ϯ 0.005 min Ϫ1 , and K i ϭ 0.045 M.  Fig. 6. Similar spectra were obtained for parental E1 titrated with ThDP (data not shown). The CD spectra displayed positive CD bands in the 259 and 265 nm regions, which increased in magnitude upon addition of ThDP or ThTDP, as well as an extensive negative band in the 268 -285 nm region (Fig. 6), similar to that reported for E1 isolated from pigeon breast muscle (26) and human recombinant E1 (27). The additional positive CD bands in the 287 and 293 nm regions are likely due to tyrosine and/or tryptophan and were also observed with parental E1 (Fig. 6). No CD bands were in evidence in the 300 -350 nm region of the spectra. In contrast, the addition of ThTTDP to parental E1 (Fig. 7) produced significant changes in the CD spectrum of E1, which were different from those produced by the addition of ThDP and ThTDP. The addition of ThTTDP decreased the positive CD bands in the 259 and 265 nm regions while generating a broad positive CD band with a maximum near 330 nm (Fig. 7). The amplitude of the positive band at 330 nm increased with increasing ThTTDP concentration, exhibiting saturation with inhibitor ( Fig. 7, inset). Fitting the data to Equation 7 provides an estimate for K d of 0.450 Ϯ 0.258 M for the binding of ThTTDP to parental E1, where E t and I t are the total concentrations of parental E1 and    Fluorescence Studies of the Binding of ThDP, ThTDP, and ThTTDP to Parental E1 and Its Tyrosine 177 Variants-It was reported previously that E. coli PDHc and E1 resolved from PDHc have an intrinsic fluorescence, probably due to tryptophan, that diminishes upon ThDP binding in the presence or absence of pyruvate (29,30). We have now shown that, upon addition of ThDP, ThTDP, or ThTTDP (Fig. 8), the intrinsic fluorescence of parental E1 was diminished. An emission maximum at 334 nm resulted from excitation at 290 nm, which very likely corresponds to tryptophan residues. The fluorescence quenching resulting from addition of ThDP, ThTDP, or ThTTDP (Fig. 8) exhibited saturation behavior, which was fitted to Equations 1-3 (see "Experimental Procedures") to estimate the K d values. The K d obtained for the binding of ThDP to parental E1 (Table IV) is in good agreement with the data presented previously for E. coli PDHc (K d ϭ 1.49 M) (29) and with our value of 3.73 M reported for single-lipoyl domain PDHc (2). The K d values for ThTDP and ThTTDP for parental E1 were at least 10 times lower than the K d for ThDP itself (Table IV). According to the data in Table IV, the Y177A and Y177F substitutions did not significantly affect the binding of ThDP. The K d values for ThDP binding were 2.01 M (Y177A) and 3.87 M (Y177F), nearly the same as the K d value of 1.84 M obtained for parental E1.
Titration of the Y177A and Y177F variants with ThTDP quenched 8 and 7% of the fluorescence, respectively, compared with 16% for parental E1, indicating that Tyr 177 substitution in the E1 subunit affects ThTDP binding. The binding of ThTTDP to the Y177A and Y177F variants of E1 was not affected and exhibited 45 and 40% fluorescence quenching, respectively, similar to that observed with parental E1 (57%) ( Table IV). The K d values were 0.113 Ϯ 0.035 M (parental E1), 0.052 Ϯ 0.007 M (Y177A E1), and 0.029 Ϯ 0.040 M (Y177F E1), lower than the corresponding values obtained from the CD titrations. The pronounced quenching of the fluorescence of parental E1 and the tyrosine 177 variants by ThTTDP could be explained by the different substituents at the C-2 atom in the thiazolium ring of ThDP, ThTDP, and ThTTDP, specifically the CϭS bond in the last one. It was reported that the thiol of N-acetylcysteine makes it a stronger quencher of the fluorescence of 3-methylindole compared with the quenching produced by other amino acids tested (31). We ran a model study in which we monitored changes in fluorescence of tryptophan (2 M) upon addition of ThTDP and ThTTDP. No significant changes were observed with concentrations as high as 100 M, indicating that specific binding of ThTDP and ThTTDP to parental E1 and the tyrosine 177 variants contributed to the observed quenching of protein fluorescence. DISCUSSION The results of this study suggest that ThTDP and ThTTDP exhibit kinetic properties characteristic of tight-binding reversible inhibitors according to the following criteria. 1) Titration of parental E1 by ThTDP and ThTTDP revealed that 0.8 -0.94 mol of ThTDP/mol of E1 monomer and ϳ2 mol of ThTTDP/mol of E1 monomer are required to inactivate E1. 2) The Ackermann-Potter plot is consistent with ThTDP behaving as a tight-binding inhibitor. 3) Inhibition of parental E1 by ThTDP and ThTTDP is partially reversible in dialysis experiments. Additional evidence of reversibility of ThTTDP binding to E1 was obtained in experiments in which the CD band at 330 nm induced by ThTTDP was reduced in magnitude upon addition of ThTDP. 4) According to reaction progress curve analysis, ThTDP and ThTTDP behave as competitive tight-binding inhibitors with a two-step mechanism of inhibition shown in Scheme 1. It is noteworthy that ThTTDP was also reported to be a tight-binding inhibitor of human E1 (9).
Several methods were used to determine the strength of binding of ThTDP and ThTTDP to E. coli E1, including reaction progress curve analysis, CD, and fluorescence spectroscopy. Importantly, progress curve analysis could yield K i values (nanomolar range) inaccessible by other methods. The lower limit of K i for ThTDP was 0.003 M. The graphical procedure of Henderson (36) led to a K i of 0.013 M for parental E1 and E1 resolved from single-lipoyl domain PDHc (data not shown). Gutowski and Lienhard (4) reported a K i of 0.5 nM for inhibition of E. coli PDHc by ThTDP. According to these results, ThTDP binds more tightly to E. coli E1 than to several other ThDPdependent enzymes such as yeast transketolase (K i ϭ 28 nM) (5), wheat germ pyruvate decarboxylase (K i ϭ 2 M) (6), and E. coli pyruvate oxidase (K d ϭ 0.2 M) (7). A K i of 0.064 M was determined for the binding of ThTTDP to E. coli E1 using progress curve analysis, a value significantly higher than that for ThTDP, but very similar to the K i of 0.0737 Ϯ 0.001 M reported for the binding of ThTTDP to human E1 (9). For the binding of ThTTDP to parental E1, CD provided a K d of 0.450 M, and fluorescence spectroscopy provided a K d of 0.113 M. For comparison, a K d of 7 M was estimated for the binding of ThTTDP to brewers' yeast pyruvate decarboxylase (8).
An important new finding from our studies is the observation and characterization of a positive CD band in the 330 nm region upon addition of ThTTDP to parental E1. The same band was not observed upon addition of either ThDP and ThTDP to E1; but we showed (by competition experiments) that all three ThDP analogs occupy the same locus. The addition of ThTDP to a sample in which the 330 nm CD band was evident from mixing ThTTDP and E1 diminished the amplitude of the band. In addition, the 330 nm CD band was also observed when ThTTDP was added to 1) E1 resolved from PDHc, 2) E. coli single-lipoyl domain PDHc, and 3) PDHc from mammalian sources. We wish to emphasize that this positive band is quite distinct from the broad negative band at 300 -350 nm with a maximum at 320 nm, which was assigned to the formation of a charge transfer complex between ThDP and transketolase (32,33) and also reported for pigeon breast muscle E1 (26) and mammalian E1 (27), but only as a very weak band at best for E. coli E1 (2). The appearance of a broad positive CD band at 330 nm was observed upon addition of ThTTDP only; and more likely, it is related to the formation of chiral ThTTDP at the active center by virtue of the V conformation enforced on all ThDP enzymes studied to date (34). We also note that the absorption spectrum of ThTTDP has a maximum at 319 nm, and this electronic transition could be the source of the observation. Finally, we also note that, so far, we have not observed the same 330 nm CD band on the related yeast pyruvate decarboxylase when ThTTDP was added to its apo-E91D var-iant at the highest concentrations. Therefore, to date, we have observed the new band only for solutions of ThTTDP and PDHc or the E1 component.
The results on the Tyr 177 variants suggest that this residue does affect the potent binding of ThTDP. 1) The molar ratio of ThTDP or ThTTDP inhibitor to Y177A variant monomer is the same, in contrast to that observed with parental E1, which is inactivated with a 1:1 stoichiometry. 2) According to reaction progress curve analysis, ThTDP binds to Y177A in a one-step mechanism, in contrast to the two-step mechanism observed for ThTDP and ThTTDP binding to parental E1. This result suggests that the conformational change characterizing the two-step binding mechanism is missing in the Y177A variant, i.e. Tyr 177 is probably involved in this conformational change.
3) The percent fluorescence quenching by ThTDP of the Y177A (8%) and Y177F (7%) variants is smaller than the 16% observed with parental E1, also indicating that Tyr 177 in the E1 subunit may be involved in ThTDP binding. The K d value for ThTDP binding determined by fluorescence quenching for parental E1 (0.117 M) is much greater than the K i of 0.003 M determined by progress curve analysis, indicating that additional step(s)   M) and also similar to the K i (0.064 M) determined by progress curve analysis. The CD signature at 330 nm was similar for the Tyr 177 variants of E1 and parental E1 and does not contradict the major conclusion that Tyr substitution does not affect ThTTDP binding. We had previously reported that cysteine 259 in the E1 subunit is important for ThTTDP binding (14) on the basis of inactivation by ThTTDP of E1 resolved from the C259N and C259S E1 variants of single-lipoyl domain PDHc. However, as mentioned above, the CD spectra of the C259S and C259N variants of E1 also exhibited the CD band at 330 nm upon addition of ThTTDP, clearly indicating that ThTTDP binds to these cysteine variants of E1. Further explanations of the observation will have to await solution of the x-ray structure of the E1⅐inhibitor complexes.
According to the high-resolution structure of the PDHc E1 component from E. coli, Tyr 177 is located in the active site cavity, 2 with the Tyr 177 phenolic oxygen atom located at 6 Å from the thiazolium C-2 atom, too far to assist in the first essential deprotonation step required to trigger the reaction (Fig. 9). There is a water molecule hydrogen-bonded (3 Å) to the C-2H and N-4Ј atoms, but too far from the Tyr 177 side chain oxygen atom (3.7 Å) to form a good hydrogen bond with it. Almost certainly, this water molecule would be replaced by substrate binding. The position of Tyr 177 in the active site of E. coli E1 is different from that observed in the crystal structure of human branched-chain ␣-ketoacid dehydrogenase (11) and 2-oxoisovalerate dehydrogenase from P. putida (12). In the crystal structure of human branched-chain ␣-ketoacid dehydrogenase, there are three tyrosine residues near ThDP: Tyr ␣113 , Tyr ␣224 , and Tyr ␤Ј102 (11). Tyr ␣113 is in direct contact with the oxygen atom of the ␤-phosphate group of ThDP; Tyr ␤Ј102 is packed against one side of the 4Ј-aminopyrimidine ring of the ThDP, with Leu ␣164 approaching the other side of the ring. It was suggested that these hydrophobic residues maintain the cofactor in the V conformation. Tyr ␣224 is coordinated to the Mg 2ϩ ion. The Tyr ␤Ј102 -to-Ala substitution showed that Tyr ␤Ј102 is not critical for ThDP binding (10). The Tyr ␣224 -to-Ala substitution, by contrast, resulted in a markedly decreased ability to bind ThDP and an increased K m for cofactor binding. The Tyr ␣113 -to-Ala substitution resulted in a Ͼ50 -100-fold increase in the K m for ThDP, indicating that Tyr ␣113 is indeed important for ThDP binding (10). In 2-oxoisovalerate dehydrogenase from P. putida, Tyr ␤Ј88 , similar to Tyr ␤Ј102 in the human enzyme, is stacked against the 4Ј-aminopyrimidine ring with other hydrophobic residues (12). Tyr ␣133 in the P. putida enzyme has a role similar to that of Tyr ␣113 in the human enzyme, contributing to the binding of the cofactor through a hydrogen bond with a phosphate oxygen. In the crystal structure of E. coli E1, Phe 602 is stacked against the 4Ј-aminopyrimidine ring, 2 with a role analogous to that of Tyr ␤Ј88 and Tyr ␤Ј102 in the two ␣ 2 ␤ 2 E1 subunits. In summary, Tyr 177 in the E. coli E1 enzyme has a different role than the tyrosine residues reported at the active center of the two published structures of the branched-chain pyruvate dehydrogenases with different quaternary structures.
In conclusion, the combination of the x-ray structural evidence on the location of Tyr 177 with the kinetic and spectroscopic data suggests that Tyr 177 has a role in stabilizing some transition state(s) in the reaction pathway, starting with the free enzyme and culminating with the first irreversible step, decarboxylation (k cat /K m -type effects), as well as in some step in reductive acetylation of lipoyl-E2. The results with the Y177F variant suggest that without the hydrogen-bonding ability of the tyrosine, the bulky side chain at position 177 is detrimental to catalysis.