Roles of His291-α and His146-β′ in the Reductive Acylation Reaction Catalyzed by Human Branched-chain α-Ketoacid Dehydrogenase

We report here that alterations of either His291-α or His146-β′ in the active site of human branched-chain α-ketoacid dehydrogenase (E1b) impede both the decarboxylation and the reductive acylation reactions catalyzed by E1b as well as the binding of cofactor thiamin diphosphate (ThDP). In a refined human E1b active-site structure, His291-α, which aligns with His407 in Escherichia coli pyruvate dehydrogenase and His263 in yeast transketolase, is on a largely ordered phosphorylation loop. The imidazole ring of His291-α in E1b coordinates to the terminal phosphate oxygen atoms of bound ThDP. The N3 atom of wild-type His146-β′, which can be protonated, binds a water molecule and points toward the aminopyrimidine ring of ThDP. Remarkably, the H291A-α mutation results in a complete order-to-disorder transition of the loop region, which precludes the binding of the substrate lipoyl-bearing domain to E1b. The H146A-β′ mutation, on the other hand, does not alter the loop structure, but nullifies the reductive acylation activity of E1b. Our results suggest that: 1) His291-α plays a structural rather than a catalytic role in the binding of cofactor ThDP and the lipoyl-bearing domain to E1b, and 2) His146-β′ is an essential catalytic residue, probably functioning as a proton donor in the reductive acylation of lipoamide on the lipoyl-bearing domain.

acyltransfer reaction, to coenzyme A (CoASH) thus yielding an acyl-CoA, a product of the overall reaction catalyzed by the BCKD complex (Reaction 1).
A recent study with the E1 component (E1p) of Escherichia coli pyruvate dehydrogenase complex showed that alteration of the His 407 residue, which is equivalent to His 291 -␣ in human E1b (Fig. 1), inactivates the reductive acetylation of lipoamide covalently attached to LBD of the E2p subunit (8). His 407 in E. coli E1p aligns with His 263 in yeast transketolase (Fig. 1); the latter has been proposed to function as a nucleophile in the transketolase reaction (9). According to this alignment and site-directed mutagenesis data, Nemeria et al. (8) proposed that His 407 in the bacterial E1p is a catalytic residue in the reductive acetylation reaction of lipoamide on the E2p (also fused ␣ and ␤) subunit. However, His 407 is disordered and not visible in the E. coli E1p crystal structure (8), and therefore no structural information is available to support this hypothesis.
We are interested in the roles of the only two histidine residues in the human E1b active site, i.e. His 291 -␣ and His 146 -␤Ј that are equivalent to His 312 -␣ and His 131 -␤Ј in Pseudomonas E1b, respectively (Fig. 1). The H292A-␣ substitution in rat E1b, which aligns with His 291 -␣ in human E1b, was shown to result in reduced ThDP binding and the loss of the reconstituted overall activity of the rat BCKD complex measured by a spectrophotometric assay (5). Similar results were obtained with the corresponding human E1b variant H291A-␣, except that residual overall activity was detected using a more sensitive radiochemical assay (6). In the present study, we show that in a refined human E1b structure His 146 -␤Ј is pointing towards the aminopyrimidine ring of the cofactor ThDP, whereas His 291 -␣, which aligns with His 407 in E. coli E1p and His 263 in yeast transketolase, is on a mostly ordered phosphorylation loop in the human E1b structure. The imidazole ring of His 291 -␣ in E1b coordinates to the terminal phosphate oxygen atoms of the ThDP. The H291A-␣ mutation results in a complete disorder of the loop region and drastically reduces the binding of lipoylated LBD (lip-LBD) to E1b. The H146A mutation does not alter the loop structure, but completely abrogates the reductive acylation (Reaction 3) activity of E1b. Based on these results, we propose that: 1) His 291 -␣ plays a structural instead of a catalytic role in the binding of cofactor ThDP and the lip-LBD to E1b, and 2) His 146 -␤Ј is an essential catalytic residue, probably functioning as a proton donor in the reductive acylation reaction catalyzed by human E1b.
Assay for the Reductive Acylation (Reaction 3) of Lipoylated LBD-The assay was modified from that described previously (14). The reaction mixture in a volume of 0.2 ml contained 50 mM potassium phosphate, pH 7.5, 1 mM ThDP, 2 mM MgCl 2 , 0.2 mM [U-14 C]KIV (specific activity 127,500 cpm/nmol), 132 nM E1b (heterotetramers), and 20 M C-terminally His 6 -tagged lip-LBD. The reaction was initiated by the addition of the radiolabeled ␣-ketoacid. After incubation at 22°C for 45 s, radioactivity incorporated into lip-LBD was extracted with Ni 2ϩnitrilotriacetic acid (Ni-NTA) resin (Qiagen, Chatsworth, CA). The resin was washed 3 times with the above phosphate buffer containing 200 mM NaCl and 2 mM ␤-mercaptoethanol. The washed resin containing radiolabeled lip-LBD was added to 2 ml of scintillation mixture, and radioactivity was counted.
Binding Studies by Isothermal Titration Calorimetry (ITC)-Human E1b and lip-LBD, both C-terminally His 6 -tagged, were dialyzed exhaustively against the same reservoir of 50 mM Tris buffer, pH 7.5, 50 mM KCl, 10 mM ␤-mercaptoethanol, 5% glycerol, and 0.2 mM EDTA to remove bound Mg 2ϩ ions and ThDP. Immediately prior to ITC measurements, MgCl 2 and ThDP stock solutions were added to both E1b and lip-LBD to a final concentration of 0.1 mM. Titrations were carried out at 20°C in a MicroCal (Northampton, MA) VP-ITC microcalorimeter. The solution of 1.5 mM lip-LBD in the syringe was added in 7-l increments to the reaction cell containing 1.8 ml of 25 M human E1b (based on the ␣␤ heterodimer). Binding isotherms derived from heat changes were used to calculate the standard free energy of binding (⌬G 0 ) according to the equation: ⌬G 0 ϭ ϪRTlnK a , where R is the gas constant, T the absolute temperature, and K a the association constant. From these binding isotherms, the number of binding sites (n) was obtained, and changes in enthalpy (⌬H 0 ) and entropy (⌬S 0 ) were calculated according to the equation: ⌬G 0 ϭ ⌬H 0 Ϫ T⌬S 0 . Curve fitting and the derivation of thermodynamic parameters were carried out with the ORIGIN software package provided by MicroCal. Concentrations of human E1b heterodimers and lip-LBD monomers were determined by A 280 nm using calculated extinction coefficients (in mg Ϫ1 ml⅐cm Ϫ1 ) of 1.14 for the former and 1.07 for the latter.
Binding Measurements Based on Tryptophan Fluorescence Quenching-Steady-state fluorescence quenching upon ThDP binding (15) to wild-type E1b as well as H291A-␣ and H146A-␤Ј variants was measured using a PerkinElmer (Boston, MA) LS50 B luminescence spectrometer in the photon counting mode. Fluorescence intensities were recorded at 25°C using a 3-ml quartz cuvette at an excitation wavelength of 290 nm and an emission wavelength of 335 nm. Slit widths were set at 5 nm for both excitation and emission. A 290-nm cut-off emission filter was installed to reduce light scattering effects. Protein concentrations for E1b and ThDP (A 235 nm ϭ 11,300 M Ϫ1 cm Ϫ1 , pH Ͼ 7.0) were determined spectrophotometrically as described above. The concentration for all protein samples was 0.23 M (as heterotetramers) in 50 mM potassium phosphate buffer, pH 7.5, 200 mM KCl, and 1 mM FIG. 1. Sequence alignments of conserved active-site histidine residues in the family of ThDP-dependent enzymes. The conserved sequence flanking the ThDP active-site histidine residues (His 291 -␣ and His 146 -␤Ј) of the human E1b were used as query sequences in a gapped BLAST search. Residues identical to human E1b sequences are indicated by dots. The invariant histidine residues are bold-faced. The abbreviations are as follows: HE1b␣ and HE1b␤, the ␣and ␤-subunits, respectively, of human E1b; EcE1p, the subunit of E. coli E1p; PsE1b␣ and PsE1b␤, the ␣and ␤-subunits, respectively, of Pseudomonas E1b; HE1p␣ and HE1p␤, the ␣and ␤-subunits, respectively, of human E1p; HE1k, the subunit of human ␣-ketoglutarate dehydrogenase; YTK, yeast transketolase. The sequence flanking His 146 -␤Ј in human E1b is not conserved in yeast transketolase. MgCl 2 . Fluorescence readings were corrected for dilution and inner filter effects using Equation 1 (16), where, F corr is the corrected fluorescence intensity value, F obs the experimentally measured fluorescence intensity, V 0 the initial volume of the sample, V the volume after adding ThDP, d the path length of the cuvette, A ex the absorption of the sample at the excitation wavelength, and A em the absorption of the sample at the emission wavelength. Three different readings were taken and averaged with the experiment conducted three times (n ϭ 3). The binding data were fitted by nonlinear regression using the program KaleidaGraph (Synergy Software, Essex Junction, VT) according to Equation 2 describing a bimolecular reaction (17), where ⌬F is the (corrected) fluorescence change, F o the fluorescence intensity prior to the addition of ThDP, ⌬F max the maximal fluorescence change, K d the dissociation constant, and [ThDP] the concentration of ThDP in the cuvette. The parameters determined by the fitting procedure were ⌬F max and K d .

Crystallization of Wild-type and Mutant E1b
Proteins-Wild-type and mutant E1b proteins (C-terminally His-tagged on the ␤-subunit) were produced as described (3). Crystals were grown at 20°C via the vapor diffusion method by mixing equal volumes of E1b (20 -25 mg/ml) in 50 mM Na-HEPES buffer, pH 7.5, 250 mM KCl, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 20 mM dithiothreitol, and 5% (v/v) glycerol with a well solution (1.4 -1.6 M ammonium sulfate, 0.1 M sodium citrate, pH 5.8, 20 mM ␤-mercaptoethanol). MgCl 2 or MnCl 2 and ThDP at 4 mM each were added to both the well solution and the cryo-buffer (see below). Serially diluted crushed crystals were used for microseeding 1 day after the drops were set up. Crystals appeared 1 day after seeding and grew to a maximum size of 120 ϫ 800 m within 10 days, which were stabilized for 12 h by soaking in fresh well solution. Crystals were cryo-protected by stepwise exchanges with a cryo-buffer containing 1.6 M ammonium sulfate, 50 mM Na-HEPES, pH 7.5, 100 mM sodium citrate, pH 5.8, 100 mM KCl, 50 mM dithiothreitol, and 20% (v/v) glycerol. Mn 2ϩ ions could replace Mg 2ϩ required for the binding of ThDP to E1b. The presence of Mn 2ϩ ions in crystals resulted in improved x-ray diffraction qualities without affecting the catalytic properties (data not shown). Crystals obtained with this procedure exhibited the symmetry of space group P3 1 21 with cell parameters of ϳ145 ϫ 145 ϫ 69 Å and contained one ␣␤ heterodimer per asymmetric unit. They diffracted x-rays significantly better (up to a minimum Bragg spacing, d min , of 1.6 Å) than those described by AEvarsson et al. (d min of 2.7 Å) (3).
X-ray Crystallography-Crystals were flash-cooled in liquid propane and kept at about 100K during data collection at beamlines 19ID and 19BM (Advanced Photon Source (APS), Argonne National Laboratory, Argonne, IL), for the H291A-␣ and H146A-␤Ј mutants. Data from the wild-type crystal were collected at 110K using CuK␣ radiation from an in-house rotating anode source (Rigaku RU-300, Japan) and an R-AXIS-IV detector (MSC Industrial Supply, Houston, TX). Data sets were processed with the HKL2000 package (18). Complete data processing statistics are listed in Table I (part A).
The E1b wild-type structure was determined by molecular replacement using the previously determined lower resolution E1b structure (PDB code 1DTW) as the search model. Mutant structures were subsequently determined by Fourier techniques. Refinement of the models was carried out in the program package CNS 1.1 (19) with a random subset of all data set aside for the calculation of free R factors. The refinement protocol consisted of an initial simulated annealing step to remove model bias, followed by cycles of conjugate gradient minimization and calculation of anisotropic displacement parameters interspersed with manual adjustments to the model using the program O (20). The electron density clearly showed the presence of several glycerol molecules in the crystal structure. After the refinement of the protein part was complete, solvent molecules were added where stereochemically reasonable. The quality of the models was finally checked against composite simulated annealing 2F o Ϫ F c omit maps (calculated with CNS), which were found to be nearly identical to regular Aweighted 2F o Ϫ F c electron density maps. Residues in flexible regions (particularly at the termini and in some surface-exposed loop regions) that did not have corresponding electron density were excluded from the models. Residues with corresponding electron density for the main chain atoms but not for the side chain atoms were added with the side chains in the most favorable rotamer conformations that did not lead to steric clashes. Single residues with little corresponding electron densities were included in the model, when they were flanked by other residues with well-defined electron density. Two residues (Ile 266 -␣ and Arg 255 -␤) had main chain and angles that would place them in the disallowed regions in the Ramachandran plot. The corresponding electron densities were, however, very well defined and thus their conformations were dictated by the structural context. Complete refinement statistics are listed in Table I (part B).

Oxidative Decarboxylation (Reaction 1) Catalyzed by Wildtype and Mutant E1b-Both
His 291 -␣ and His 146 -␤Ј residues were mutated, and effects of these mutations on the overall reaction (Reaction 1) catalyzed by the reconstituted BCKD complex were measured radiochemically in the presence of excess lip-E2b and E3. Table II shows that H146A-␤Ј and H146N-␤Ј substitutions completely abolish the overall reaction catalyzed by the reconstituted BCKD complex, with KIV or ThDP as the variable substrate or cofactor, respectively. H291A-␣, H291N-␣, and H291Q-␣ mutations result in significant increases in K m for KIV and ThDP, indicating reduced binding affinity for both the substrate and the cofactor. Concomitantly, k cat is reduced to ϳ5-10% of the wild type, as measured with varying concentrations of KIV or ThDP. The marked reductions in catalytic efficiencies associated with the His 291 -␣ and His 146 -␤Ј substitutions establish the critical role of these two residues in the overall reaction catalyzed by the BCKD complex.
His 291 -␣ Is an Essential Residue for the Binding of LBD to E1b-The interactions of wild-type and mutant E1b with lip-LBD were studied by ITC. Fig. 4 shows that wild-type E1b readily binds to lip-LBD, but not unlipoylated LBD, with K d ϭ 15.6 M, and ⌬H 0 ϭ Ϫ1.32 kcal/mol (Table IV). The His146A-␤Ј variant binds to lip-LBD with affinity similar to the wild type. The ⌬H 0 of Ϫ0.69 kcal/mol with the His146A-␤Ј mutant is smaller than that obtained with wild-type E1b. For the His291A-␣ E1b, no binding to lip-LBD was observed. The fact that this mutant exhibits 6% of the wild-type overall activity (Table II) suggests that low affinity binding of lip-LBD to this  E1b variant is likely to occur, however, the affinity is too low to be detected by the present ITC measurements. The ITC data therefore indicate that the His291A-␣ residue is essential for the interactions of E1b with the lip-LBD of E2b. Refined Active-site Structure of Human E1b-Our crystallization conditions yielded crystals that diffract to a significantly higher resolution (up to 1.65 Å for selected crystals) than those reported by AEvarsson et al. (2.7 Å, PDB code 1DTW). This improvement allowed a more detailed characterization of the conformation at the E1b active site including bound water molecules, which is a pre-requisite to dissect the reaction mechanism of this enzyme (Fig. 5). The crystal structure of holo-E1b with bound Mn-ThDP was determined using data to a minimum Bragg spacing, d min , of 1.81 Å ( Table I). As shown in Fig.  5, there are only two histidine residues, His 291 -␣ and His 146 -␤Ј that are within hydrogen-bonding distance to the bound ThDP in the human E1b active site. The N1 nitrogen atom of His 291 -␣ is hydrogen-bonded to the O1 water molecule that in turn interacts with the two terminal phosphate oxygen atoms of ThDP. The same O1 water molecule is further oriented through coordination with the side chain of Arg 287 -␣. Another water molecule O2 coordinates directly to N3 nitrogen atom of the same histidine residue. His 146 -␤Ј is held in place through interactions of its N3 nitrogen atom with the O4 water molecule, with the latter simultaneously hydrogen-bonded to the hydroxyl group of Tyr 102 -␤Ј. The side chain of Tyr 102 -␤Ј is packed against the aminopyrimidine ring of the cofactor with the side chain of Leu 164 -␣ (not shown), which is wedged in between FIG. 4. ITC measurements for lip-LBD binding to wild-type, H146A-␤ and H291A-␣ human E1b. ITC experiments were carried out in a MicroCal VP-ITC microcalorimeter by consecutively injecting aliquots of 1.5 mM lip-LBD or unlipoylated LBD into the reaction cell containing 25 M wild-type or mutant human E1b. Binding isotherms for wild-type (E), H146A-␤Ј (q), and H291A-␣ (OE) were obtained by plotting heat changes against the molar ratio of lip-LBD, as derived from the integrated raw data. The data were fit using the ORIGIN software supplied by the manufacturer. Wild-type E1b and the His146-␤Ј variant show similar affinity for lip-LBD with dissociation constants (K d ) of 2.52 ϫ 10 Ϫ5 M and 1.56 ϫ 10 Ϫ5 M, respectively. The binding of the H291A-␣ mutant to lip-LBD cannot be detected by ITC as indicated by the absence of heat changes. Binding of unlipoylated LBD (‚) to wild-type E1b also cannot be detected.

FIG. 3. ThDP binding to wild-type and mutant human E1b
proteins measured by tryptophan fluorescence quenching. Incremental amounts of ThDP were added to a solution of 0.23 M human E1b. Samples were excited at 290 nm, and emission intensity at 335 nm was measured. Changes in tryptophan fluorescence due to ThDP binding were normalized for the dilution of the sample and for inner filter effects. The data were plotted for wild-type (Ⅺ), H146A-␤Ј (q), and H291A-␣ (E) E1b as % quenching of fluorescence versus ThDP concentrations. The % quenching represents ⌬F/F o , where F o is fluorescence intensity prior to the addition of ThDP and ⌬F is the decrease in fluorescence at a given ThDP concentration. The data were fitted as described under "Experimental Procedures." Dissociation constants (K d ) for ThDP are shown in Table III. the two rings of ThDP, approaching from the other side of the aminopyrimidine ring. These interactions contribute to the formation of the V-conformation (torsion angles ⌽ T ϭ 100°, ⌽ P ϭ Ϫ71°) for the cofactor (3). Overlapping densities are observed between the N1Ј atom of the ThDP aminopyrimidine ring and the side-chain carbonyl group of the invariant Glu 76 -␤Ј. This interaction results in the increased basicity of the 4Ј-NH 2 group, which is required for efficient deprotonation of the C2 atom of the aminopyrimidine ring for the formation of enamine-ThDP (21).
An important structural feature of our E1b crystals is the visibility of the loop region in wild-type E1b between Tyr 286 -␣ and Gln 312 -␣ including Ser 292 -␣ (phosphorylation site 1), with electron density interrupted between Thr 293 -␣ and Ser 294 -␣, at Ala 299 -␣, and between Ser 302 -␣ and Asn 307 -␣ (Fig. 6A). Remarkably, the H291A-␣ mutation renders the entire phosphorylation loop disordered in the active site of this mutant (Fig.  6B). Residual electron densities observed in Tyr 286 -␣ and Arg 287 -␣ are accompanied by the complete absence of densities between Ile 288 -␣ and Gln 312 -␣ (Fig. 6B). In contrast, the H146A-␤Ј substitution does not result in significant conformational alterations from the wild type in the phosphorylation loop region or the vicinity of this histidine residue (Fig. 6C). However, the O4 water molecule hydrogen-bonded to His 146 -␤Ј in the wild-type human E1b (Fig. 5) is absent from the H146A-␤Јvariant. This change abolishes water-mediated interactions between Tyr 102 -␤Ј and His 146 -␤Ј in the wild-type active-site structure. DISCUSSION In the earlier study with D,L-S-methyllipoic acid methyl ester, a tetrahedral adduct at the C2 position of ThDP was formed between enamine-ThDP and the methyl ester (7). The isolation of this tetrahedral intermediate led Jordan's group to postulate the presence of a proton donor in E1, which facilitates the disulfide bond scission of lipoic acid by electrophilic catalysis at the S6 atom (7,22). The E1b structure of Pseudomonas BCKD complex recently showed the presence of two histidine residues (His 312 -␣ and His 131 -␤Ј) flanking the cofactor ThDP in the active-site channel (2). It was proposed, based on the D,L-S-methyllipoic acid model, that either histidine residue can be a candidate proton donor during reductive acylation of lipoamide attached to the E2b subunit. The present study focused on the equivalent histidine residues in human E1b in order to shed light on their roles during the reductive acylation catalytic cycle.
The role of His 291 -␣ was investigated by measuring the affinity of the H291A-␣ mutant for cofactor ThDP. The K d (39.3 M) of this variant, as determined by tryptophan fluorescence quenching, is an order of magnitude higher than that of the wild-type at 1.52 M. These values are comparable to the K m values of 24 M and 0.66 M for the H291A-␣ variant and wild-type E1b, respectively (6). In addition to electrostatic stabilization of the phosphate group by a putative positively charged His 291 -␣, the N1 atom of this residue establishes a water-mediated hydrogen bond to the terminal phosphate oxygen atoms (Fig. 5). The reduced affinity of the H291A-␣ mutant for ThDP likely results from the loss of the stabilizing interaction between the cofactor and the His 291 -␣ side chain.  5. Refined structure of the human E1b active site at the interface between ␣and ␤-subunits. 2F o Ϫ F c electron densities (in green) are contoured at 1. Only two histidine residues are within 5-Å distance from the C2 atom of the bound ThDP. His 146 -␤ is hydrogenbonded to the O4 water molecular, whereas His 291 -␣ forms hydrogen bonds to the O1 and O2 water molecules (in red spheres); the former in turn coordinates to the terminal phosphate oxygen of ThDP. The channel leading to the activated C2 atom of ThDP lies at the interface between the ␣and ␤Ј-subunits, such that these two histidine residues flank opposite sides of the channel. A Mn 2ϩ ion is bound at the metal ion binding site in place of the common Mg 2ϩ ion. Good electron density is present for Ser 292 -␣ (phosphorylation site 1), which is positioned at the opening of the channel. Carbon atoms are in gold, ThDP in green, oxygen atoms in red, nitrogen atoms in blue, phosphorous atoms in magenta, and sulfur atoms in yellow. Graphics were generated with the programs BobScript (24) and PovRay (Persistence of Vision, v3.02, POV-Team, www.povray.org).
The impaired interactions of the H291A-␣ mutant with ThDP account for the reduced rate of decarboxylation (Reaction 2) with this variant relative to the wild-type. It is of significant interest that the H291A-␣ mutant exhibits severely impaired binding to lip-LBD. From the striking absence of visible electron density for most of the phosphorylation loop in crystals of this variant, we conclude that the above indirect interactions between the side chain of His 291 -␣ and the terminal phosphate oxygen atoms of ThDP are vital for the stabilization of this loop (Fig. 6B). An asparagine residue in place of a histidine at position 291-␣ would be able to establish direct or indirect hydrogen bonding interactions with ThDP, but it would not exhibit electrostatic stabilization. Partial stabilization of the phosphorylation loop in this variant would explain the modestly reduced affinity for ThDP (K m ϭ 1.4 M) of the H291N-␣ mutant, compared with the wild type (Table II). Similar mechanisms can be invoked for the H291Q-␣ variant. Taken together, our results suggest that for binding to lip-LBD, the phosphorylation loop must be in a specific conformation conferred by the interactions between ThDP and His 291 -␣. The markedly decreased rate of reductive acylation with the H291A-␣ variant most likely results from its inability to bind lip-LBD, and appears to be largely responsible for the marginal overall activity reconstituted with this mutant (Table II). As discussed above, based on the Pseudomonas E1b structure (2), for the reductive acylation to occur the lip-LBD must penetrate the 20-Å E1b active-site channel, so as to facilitate an efficient acyltransfer from enamine-ThDP to the S8 atom on the dithiolane ring of lipoamide attached to lip-LBD. The current data therefore strongly support a structural role for His 291 -␣ in the reductive acylation reaction catalyzed by human E1b. The weak binding of the H291A-␣ ⌭1b to lip-LBD as measured by ITC is similar to that observed with the equivalent H407A variant of E. coli E1p (8). The K d value (15.6 M) and thermodynamic parameters (⌬G 0 , ⌬H 0 , and T⌬S 0 ) for the binding of wild-type human E1b to lip-LBD determined by the same method are also analogous to those measured with E. coli E1p and its cognate lip-LBD (8).
His 407 in E. coli E1p was proposed to function as a catalytic residue that assists in the protonation of the dithiolane sulfur atom of lipoamide on the E2p subunit (8). In contrast to this hypothesis, our present data argue against the equivalent His 291 -␣ in human E1b serving as a putative proton donor in the E1b-mediated reductive acylation reaction, since substitution of His 291 -␣ with an asparagine or glutamine, which cannot be a proton donor, does not nullify the activity for reductive acylation in these mutants. The basis for the discrepancy between the human E1b and E. coli E1p studies is presently unknown. However, the proposed catalytic role of His 407 in E. coli in activating E2p-attached lipoamide during reductive acetylation cannot be substantiated from the structural viewpoint, since the region that carries His 407 is not visible in the E. coli E1p structure (23).
On the other hand, our data show that the only other conserved histidine residue within the 4.75-Å distance of enamine-ThDP in the E1b active site, i.e. His 146 -␤Ј is a critical catalytic residue. The decarboxylation reaction is severely reduced in both H146A-␤Ј and H146N-␤Ј mutants, indicating that this histidine is involved in but not essential for this reaction (