Catalysis of transthiolacylation in the active centers of dihydrolipoamide acyltransacetylase components of 2‐oxo acid dehydrogenase complexes

The Escherichia coli 2‐oxoglutarate dehydrogenase complex (OGDHc) comprises multiple copies of three enzymes—E1o, E2o, and E3—and transthioesterification takes place within the catalytic domain of E2o. The succinyl group from the thiol ester of S8‐succinyldihydrolipoyl‐E2o is transferred to the thiol group of coenzyme A (CoA), forming the all‐important succinyl‐CoA. Here, we report mechanistic studies of enzymatic transthioesterification on OGDHc. Evidence is provided for the importance of His375 and Asp374 in E2o for the succinyl transfer reaction. The magnitude of the rate acceleration provided by these residues (54‐fold from each with alanine substitution) suggests a role in stabilization of the symmetrical tetrahedral oxyanionic intermediate by formation of two hydrogen bonds, rather than in acid–base catalysis. Further evidence ruling out a role in acid–base catalysis is provided by site‐saturation mutagenesis studies at His375 (His375Trp substitution with little penalty) and substitutions to other potential hydrogen bond participants at Asp374. Taking into account that the rate constant for reductive succinylation of the E2o lipoyl domain (LDo) by E1o and 2‐oxoglutarate (99 s−1) was approximately twofold larger than the rate constant for k cat of 48 s−1 for the overall reaction (NADH production), it could be concluded that succinyl transfer to CoA and release of succinyl‐CoA, rather than reductive succinylation, is the rate‐limiting step. The results suggest a revised mechanism of catalysis for acyl transfer in the superfamily of 2‐oxo acid dehydrogenase complexes, thus provide fundamental information regarding acyl‐CoA formation, so important for several biological processes including post‐translational succinylation of protein lysines. Enzymes 2‐oxoglutarate dehydrogenase (http://www.chem.qmul.ac.uk/iubmb/enzyme/EC1/2/4/2.html); dihydrolipoamide succinyltransferase (http://www.chem.qmul.ac.uk/iubmb/enzyme/EC2/3/1/61.html); dihydrolipoamide dehydrogenase (http://www.chem.qmul.ac.uk/iubmb/enzyme/EC1/8/1/4.html); pyruvate dehydrogenase (http://www.chem.qmul.ac.uk/iubmb/enzyme/EC1/2/4/1.html); dihydrolipoamide acetyltransferase (http://www.chem.qmul.ac.uk/iubmb/enzyme/EC2/3/1/12.html).

The Escherichia coli 2-oxoglutarate dehydrogenase complex (OGDHc) comprises multiple copies of three enzymes-E1o, E2o, and E3-and transthioesterification takes place within the catalytic domain of E2o. The succinyl group from the thiol ester of S8-succinyldihydrolipoyl-E2o is transferred to the thiol group of coenzyme A (CoA), forming the all-important succinyl-CoA. Here, we report mechanistic studies of enzymatic transthioesterification on OGDHc. Evidence is provided for the importance of His375 and Asp374 in E2o for the succinyl transfer reaction. The magnitude of the rate acceleration provided by these residues (54-fold from each with alanine substitution) suggests a role in stabilization of the symmetrical tetrahedral oxyanionic intermediate by formation of two hydrogen bonds, rather than in acid-base catalysis. Further evidence ruling out a role in acid-base catalysis is provided by site-saturation mutagenesis studies at His375 (His375Trp substitution with little penalty) and substitutions to other potential hydrogen bond participants at Asp374. Taking into account that the rate constant for reductive succinylation of the E2o lipoyl domain (LDo) by E1o and 2-oxoglutarate (99 s À1 ) was approximately twofold larger than the rate constant for k cat of 48 s À1 for the overall reaction (NADH production), it could be concluded that succinyl transfer to CoA and release of succinyl-CoA, rather than reductive succinylation, is the rate-limiting step. The results suggest a revised mechanism of catalysis for acyl transfer in the superfamily of 2-oxo acid dehydrogenase complexes, thus provide fundamental information regarding acyl-CoA formation, so important for several biological processes including post-translational succinylation of protein lysines. The 2-oxoglutarate dehydrogenase complex (OGDHc, also known as a-ketoglutarate dehydrogenase complex) catalyzes one of the rate-limiting steps in the tricarboxylic acid cycle (TCA), which is a common pathway for oxidation of fuel molecules, including carbohydrates, fatty acids, and amino acids. The OGDHc is a multienzyme complex composed of multiple copies of three component enzymes: a thiamin diphosphate (ThDP)-dependent 2-oxoglutarate dehydrogenase (E1o; EC 1.2.4.2), dihydrolipoyl succinyltransferase (E2o; EC 2.3.1.61), and dihydrolipoyl dehydrogenase (E3; EC 1.8.1.4) (Fig. 1) [1][2][3]. The E1o and E2o carry out the principal reactions for succinyl-CoA formation, whereas E3 re-oxidizes dihydrolipoamide-E2o to lipoamide-E2o, the redox cofactor covalently amidated onto a lysine residue on the E2o component.
Recent studies revealed multiple roles for the mitochondrial OGDHc, and their better understanding could be important for bringing new insight into human diseases. The unique property of the mitochondrial OGDHc to produce the reactive oxygen species superoxide and hydrogen peroxide (H 2 O 2 ) from its substrate 2-oxoglutarate (OG) had been attributed earlier to the flavin cofactor tightly bound to the E3 component [4,5]. Our studies revealed that human E1o (E1o-h) can produce the ThDP-enamine radical and superoxide anion from OG and from the next higher homologue 2-oxoadipate (OA) by one-electron oxidation of the ThDP-enamine intermediate with dioxygen [6,7]. The efficiency of superoxide/H 2 O 2 production by OGDHc was 7 times larger from OA than from OG making the OGDHc one of the important reactive oxygen species producers among 2-oxo acid dehydrogenase complexes in mitochondria [5,7].
In a role different from its role in the TCA cycle, the OGDHc could also serve as a producer of succinyl groups in neurons and neuronal cell lines for reversible post-translational modification of the cytosolic and mitochondrial enzymes, including those from the TCA, by succinylation, hence playing a role in neurodegenerative diseases [8]. The post-translational modification of histone proteins by succinylation is regarded as very important because it can directly regulate gene expression [9,10]. Recently, evidence of the interaction between the nuclear OGDHc and lysine acetyltransferase 2A displayed a role of lysine acetyltransferase as a carrier of succinyl groups produced by nuclear OGDHc for direct histone H3 succinylation [9].
Our own recent studies suggested that human 2oxoadipate dehydrogenase (hE1a, also known as DHTKD1-encoded protein), which is involved in the oxidative decarboxylation of 2-oxoadipate to glutaryl- CoA on the final degradative pathway of L-lysine, L-hydroxylysine, and L-tryptophan, has recruited the hE2o and hE3 components of the OGDHc for its function [11]. In other words, the hE2o could serve as a source of both succinyl-CoA and glutaryl-CoA in mitochondria and could be linked to the lysine posttranslational modification by glutarylation earlier reported [12].
The enzymatic mechanism being studied in this research is responsible for synthesis of important acyl-CoA metabolites involved in post-translational modification of proteins. The results provide important baseline residue-specific contribution to catalysis of succinyl-CoA formation in the active center of the Escherichia coli E2o that would be applicable to all E2o components due to high sequence identities reported for all E2o core/catalytic domains [13][14][15][16][17][18][19]. The major goal of this study was to address the role of the highly conserved His375 in the active center of the E. coli E2o. The His375Ala substitution was found to decrease the catalytic efficiency (k cat /K m,OG ) by 54fold compared to unsubstituted E2o. This value was approximately 10 000 times smaller than that determined on Ala substitution of the corresponding His195 in chloramphenicol acetyltransferase (CAT, k cat /K m = 53 9 10 À4 -fold smaller for the His195Ala variant than wild-type) [20], placing in doubt the earlier deduction of an acid-base catalyst role of His375, assigned by comparison with CAT [21,22]. To gain more insight regarding the function of His375, sitesaturation mutagenesis was carried out at this residue, whose results suggested that the His375Trp and His375Gly substitutions at E2o were acceptable, reducing the catalytic efficiency only by two-and fivefold, respectively. Similarly, the Asp374Asn substitution in the E2o active center (10 times less efficient variant) rescued some of the activity displayed by the Asp374Ala substitution (54-times less efficient variant), consistent with participation of Asp374 in hydrogen bond formation. The totality of our studies suggested that there is no proton transfer in the rate-limiting step in which both His375 and Asp374 participate; rather, they may stabilize by hydrogen bonds a transition state, probably resembling a tetrahedral oxyanionic intermediate.

Expression and purification of E1o
Expression and purification of E1o was as reported earlier with some modifications [23]. The E1o component was purified using a Ni-Sepharose 6 Fast Flow column equilibrated with 20 mM KH 2 PO 4 (pH 7.5) containing 0.20 M KCl, 0.20 mM ThDP, 1.0 mM MgCl 2 , 1.0 mM benzamidineÁHCl (buffer A), and 30 mM imidazole. After the protein was applied to the column, the column was washed with 500 mL of 30 mM imidazole and then with 500 mL of 50 mM imidazole, both in buffer A. The E1o was eluted with 150 mM imidazole in buffer A and was dialyzed against 2000 mL of 20 mM KH 2 PO 4 (pH 7.5) containing 0.35 mM KCl, 0.20 mM ThDP, 1.0 mM MgCl 2 , and 1.0 mM benzamidineÁHCl for 15 h at 4°C. For E1o storage, buffer was exchanged to 20 mM KH 2 PO 4 (pH 7.5) containing 0.20 M KCL, 0.20 mM ThDP, and 1.0 mM MgCl 2 by ultrafiltration using a concentrating unit with a 30-kDa cutoff. The E1o was stored at À80°C.

Expression and purification of the E2o and its variants
Expression and purification of E2o was as reported earlier for human E2p [24] with some modifications. AG1 cells were grown in LB medium supplemented with 35 lgÁmL À1 of chloramphenicol, and protein expression was induced by 0.50 mM IPTG for 5 h at 37°C. Harvested cells were resuspended in 20 mM KH 2 PO 4 (pH 7.5) containing 0.30 M NaCl, 1 mM ben-zamidineÁHCl, 1 mM DTT, and two protease inhibitor cocktail tablets. The resulting cell suspension was incubated with lysozyme (0.60 mgÁmL À1 ) for 20 min on ice. Next, MgCl 2 (5 mM) and 1000 U each of DNase I and micrococcal nuclease were added and the cells were incubated for an additional 20 min on ice. Cells were then disrupted by ultrasonication at a setting of 6 using 10-s pulse 'on' and 30-s pulse 'off 'for total time of 5 min. PEG-8000 (50% w/v) was added dropwise to the clarified lysate to 6% (v/v), and the precipitated pellet was dissolved in 20 mM KH 2 PO 4 (pH 7.5) containing 0.30 M NaCl, 1 mM DTT, 1.0 mM benza-midineÁHCl, and 1 mM EDTA. The E2o was purified using a Sephacryl S-300 high-resolution size-exclusion column equilibrated with 20 mM KH 2 PO 4 (pH 7.5) containing 0.30 M NaCl, 1 mM DTT, 1.0 mM benza-midineÁHCl, and 1 mM EDTA. Fractions containing E2o according to PAGE were collected, and E2o was precipitated by ultracentrifugation for 4.0 h at 140 000 g. The pellet was dissolved in 20 mM KH 2 PO 4 (pH 7.5) containing 0.40 M NaCl, 0.50 mM EDTA, 1.0 mM DTT, and 1.0 mM benzamidineÁHCl for 15 h at 4°C. The clarified E2o was stored at À80°C.
The E2o variants with Thr323Ala, Thr323Ser, Asp374Ala, Asp374Asn, His375Ala, His375Cys, His375Asn, Arg376Ala, and Asp379Ala substitutions were created using the pCA24N-E2o plasmid as a template and two synthetic oligonucleotide primers complementary to the opposite strands of the DNA with the QuikChange Site-Directed Mutagenesis Kit and protocol supplied by the manufacturer (Stratagene, La Jolla, CA, USA). The following oligonucleotide primers and their complements were used (mismatched bases are underlined, and mutated codons are shown in boldface type) as listed in Table 1.
The presence of substitutions was verified by DNA sequencing with specific primers at the Molecular Resource Facility of Rutgers New Jersey Medical School.

Expression and purification of LDo and E2o 1-176 didomain
Expression and purification of the E2o lipoyl domain (LDo, residues 1-105 in E2o) and E2o 1-176 didomain was as reported earlier [15,25] for E. coli LDp and E2p  . To ensure full lipoylation of the LDo and the E2o 1-176 didomain, these E2o-derived domains were lipoylated in vitro by an E. coli protein lipoyl ligase as reported earlier [26]. Lipoylation of each E2derived domain was confirmed by FT-MS.

Expression and purification of the E2o catalytic domain and its His375Ala and Asp374Ala variants
For expression of the CDo, DNA encoding residues 93-404 corresponding to the catalytic domain, subunit binding domain and linker connecting them in wildtype E2o was synthesized and was cloned into pD451-SR vector by DNA 2.0 Inc. (Menlo Park, CA, USA). DNA encoding CDo with His 6 tag at the C-terminal end was introduced into BL21 (DE3) cells (Novagen, San Diego, CA, USA), and cells were grown in LB medium supplemented with 35 lgÁmL À1 of kanamycin at 37°C overnight. 15 mL of the overnight culture was inoculated into 800 mL of the LB medium supplemented with kanamycin (35 lgÁmL À1 ), and cells were grown to OD 600 of 0.5-0.6 at 37°C, and then, IPTG (0.5 mM) was added and the cells were grown for an additional 5 h at 37°C. The cells were then washed with 50 mM KH 2 PO 4 (pH 7.5) containing 0.15 M NaCl and were stored at À20°C. The harvested cells were resuspended in 20 mM KH 2 PO 4 (pH 7.5) containing 0.30 M NaCl, 1 mM benzamidineÁHCl, 1 mM DTT, and two protease inhibitor cocktail tablets. Cells were treated with lysozyme (0.6 mgÁmL À1 ) at 4°C for 20 min and then with DNase I (NEB) and micrococcal nuclease (NEB) (1000 units of each) for an additional 20 min at 4°C. Cells were disrupted by a sonic dismembrator with 10-s pulse 'on' and 30-s pulse 'off' for 5 min. The supernatant was centrifuged at 30 000 g for 20 min. The clarified lysate was loaded onto a 10 mL Ni-Sepharose 6 Fast Flow column (GE Healthcare) pre-equilibrated with 20 mM KH 2 PO 4 (pH 7.5) containing 0.30 M NaCl, 1.0 mM benzamidineÁHCl, and 10 mM imidazole. The column was first washed with 35 mM imidazole in the same buffer solution. The CDo was eluted with 150 mM imidazole. The fractions containing CDo were dialyzed against 2,000 mL of 20 mM KH 2 PO 4 , containing 0.45 M NaCl and 1.0 mM benzamidineÁHCl for overnight at 4°C. The CDo was concentrated by ultrafiltration with a 10-kDa MW cutoff concentrating unit and was stored at À80°C.

Substitution
Primer sequence For construction of Asp374Ala and His375Ala CDo variants, the pD451-SR plasmid encoding CDo was used as a template, and the amplification primers such as 5 0 -GCGCTGTC-CTACGATGCGCGTCTGATCG AT-3 0 (for His375Ala) and 5 0 -CTGGCGCTGTCCT A-CGCGCACCGTCTG ATC-3 0 (for Asp374Ala) and their complements were used for site-directed mutagenesis (the mismatched bases are underlined, and mutated codons are shown in boldface type). The QuikChange Site-Directed Mutagenesis Kit was used for single-site substitution (Stratagene). The constructed domains are shown in Fig. 2D.

Expression and purification of E3
Expression and purification of E3 was carried out as reported earlier [27].
In the OGDHc assay with CDo and LDo replacing E2o, the E1o (60 lg, 1.4 lM subunits) in 50 mM KH 2 PO 4 (pH 7.0) was first preincubated with independently expressed CDo or CDo variants (60 lg, 4.9 lM subunits) and E3 (60 lg, 2.9 lM subunits) at a mass ratio of 1 : 1 : 1 (lg : lg : lg) at 25°C. An aliquot with 5 lg of E1o (0.11 lM subunits) was then withdrawn and was mixed with varying concentrations of LDo in a cuvette containing all components needed for OGDHc assay. After 1 min of equilibration at 30°C, the OGDHc reaction was initiated by the addition of OG (2 mM) and CoA (0.13 mM) as above.
Site-saturation mutagenesis on the His375 residue in the E2o core domain The site-saturation mutagenesis technique was used to create a library at the His375 position in the E2o core domain using a modified procedure for QuikChange Site-Directed Mutagenesis Kit and using NNS primers 5 0 -GGCGCTGTCCTACGATNNSCGTCTGATCG-3 0 (His375X) where N = A/T/G/C and S = G/C. The pCA24N-E2o plasmid DNA was used as template. The library was created and transformed into E. coli BL21 (DE3)-competent cells.
The library was plated onto an LB agar plate containing chloramphenicol antibiotic (50 lgÁmL À1 ) to obtain single colonies. Approximately 270 single colonies were picked and inoculated into a 96-well plate containing LB medium and chloramphenicol (50 lgÁmL À1 ) to make master plates. The master plates were incubated at 37°C at 250 rpm. Next, duplicate plates were constructed by inoculating (509 dilution) deep-well plates that contained 1 mL LB media, chloramphenicol, D,L-lipoic acid (0.3 mM), and IPTG (1 mM) for E2o. The deep-well plates were incubated at 37°C at 250 rpm. The cells were centrifuged at 1600 g, and the media were discarded. The cell pellets were stored À20°C until use. The cell pellets were resuspended and lysed in a buffer containing KH 2 PO 4 (20 mM, pH 7.5), NaCl (0.15 M), DTT (1 mM), protease inhibitor cocktail tablet, lysozyme (1 mgÁmL À1 ), DNAse (0.1 mgÁmL À1 ), and nuclease (1000U) and incubated for 1 h at 37°C. The lysate was centrifuged twice at 1600 g for 30 min, and the clarified supernatant was used for the subsequent steps.
The E1o and E3 components were overexpressed separately each in 1L of culture medium following published protocols and lysed [27]. The lysates of E1o, E2o, and E3 were reconstituted in a 96-well plate (total volume 300 lL) in a buffer containing KH 2 PO 4 (20 mM, pH 7.5), NaCl (0.15 M), MgCl 2 (1 mM), and ThDP (0.1 mM) in a ratio of 1 : 2 : 1 as per OD 600 nm /lL for 1 h and mixed well. The first column on the plate was the wild-type OGDHc complex.
Those E2 variants were chosen that had similar or higher activity than the wild-type [(OD 340 final variant x À OD 340 initial variant x )/(OD 340 final variant ÀOD 340 initial variant )]. The candidates were rescreened, and the positive variants were sequenced.
The rescreening conditions were identical to the screening assay with a minor modification. First, identified clones were traced back to the master plate and restreaked on LB agar plates containing chloramphenicol (50 lgÁmL À1 ). The first column was reserved for wild-type, and two columns or 16 colonies from a single E2o-variant were inoculated into a 96-well plate containing LB medium and antibiotic. The false-positives were discarded. The positive clones were sequenced, overexpressed, and purified following published procedures and as described above [24]. Next, the purified E2o variants were assembled with purified E1o and E3 and the steady-state kinetic parameters were determined.

Fluorescence spectroscopy
To determine the dissociation constant of CoA and succinyl-CoA with the E2o His375Trp variant, to the protein (0.24 mgÁmL À1 , concentration of active centers of 3.5 lM) in a mixture of 50 mM KH 2 PO 4 and 50 mM Tris (pH 8.0), containing 0.1 M NH 4 Cl, 1 mM DTT, and glycerol (1% v/v), was added CoA (3-570 lM) or succinyl-CoA (5-380 lM), and the fluorescence spectra were recorded after each addition at 25°C using a Cary Eclipse spectrofluorimeter. The excitation wavelength was 295 nm, and the emission spectra were recorded in the range of 300-450 nm in 3-mL quartz cuvettes. The K d values for CoA were calculated using Eqn (1).
where (F o ÀF i )/F o is the relative fluorescence quenching following the addition of CoA or suc-CoA, n is the Hill coefficient, and for n = 1.0, the value of S 0.5 is equal to K d .

Reductive succinylation of the lipoyl domain by E1o and OG
For this experiment, the LDo (100 lM) was incubated with E1o (0.015 lM) in 35 mM (NH 4 ) 2 CO 3 (pH 7.5) containing 0.50 mM MgCl 2 and 0.1 mM ThDP. The reaction was started by addition of 2 mM of OG. Aliquots of 10 lL were withdrawn at times of 5-600 s and were diluted into 1 mL of 50% methanol and 0.1% formic acid to stop the reaction, and then, the samples were analyzed by ESI FT-MS. The fraction of succinylated LDo at different times was determined by taking a ratio of the relative intensity of the mass of the succinylated form to the total relative intensity (sum of unsuccinylated and succinylated LDo). The time dependence of this fraction was plotted, and the data were fitted to a single exponential (Eqn 2) using Sigma Plot 10.0. The rate constant was calculated from the linear fit to the initial rate conditions.
Succinyltransferase reaction of the core domain and its His375Ala and Asp374Ala variants in the reverse direction The following protocol outlines simultaneous detection of succinylated and unsuccinylated LDo forms when CDo was used as catalyst in the reverse reaction. The reaction mixture in 300 lL of the 35 mM NH 4 HCO 3 (pH 7.5) contained the following: 40 lM LDo lipoylated in vitro; 0.10 mM TCEP to keep LDo in the reduced form; 0.15 mM succinyl-CoA; and 0.05 lM CDo. The reaction was started by addition of CDo after 40 s of equilibration of the reaction assay and was conducted at 30°C. Aliquots of 20 lL were withdrawn at times of 5-120 s and were diluted into 1 mL of 50% methanol and 0.1% formic acid to quench the reaction, and then, samples were analyzed by ESI FT-MS. The fraction of succinylated LDo at different times was determined by taking a ratio of the relative intensity of the mass corresponding to the succinylated form to the total relative intensity (sum of intensities for the unsuccinylated and succinylated LDo). The time dependence of this fraction was plotted, and data were fitted to a single exponential (Eqn 2) using Sigma Plot 10.0. The rate constant was calculated from the linear fit to the initial rate conditions.

Results and Discussion
Reductive succinylation of the lipoyl domain by E1o and OG is not the rate-limiting step in the OGDHc reaction Reductive succinylation of the lipoyl moiety covalently attached to the lipoyl domain on E2o is the final step involving ThDP-bound covalent intermediates formed on the first E1o component of the OGDHc. It is a concomitant oxidation of the OG-derived enamine and reductive acylation of E2o (Fig. 2, see Fig. 2D for C-terminally truncated E2o proteins created by DNA manipulations and employed in these experiments). According to Fig. 1, the S8-succinyldihydrolipoyl-E2o is the product of the reductive succinylation of the lipoyl-E2o by E1o and OG [28]. Next, transfer of the succinyl group from the thiol ester of S8-succinyldihydrolipoyl-E2o to CoA takes place exclusively at the E2o core domain active centers (Fig. 2C), so there is no oxidation-reduction involved in succinyl transfer, rather it represents a conversion of one thiol ester to another. Earlier, we developed an appropriate model reaction [26] to study the rate of reductive acetyl transfer from E. coli pyruvate dehydrogenase (E1p) to the lipoyl domain of the dihydrolipoyl transacetylase (E2p), which maintains both the chemistry and the intercomponent communication due to its specific recognition of E1p [29][30][31]. The method developed takes advantage of our ability to detect and quantify LD-E2p and acetyl-LD-E2p simultaneously in the quenched reaction mixtures using Fourier transform mass spectrometry (FT-MS). Analogously, the reductive succinylation of LDo by E1o and OG under steady-state conditions here examined provides an appropriate model reaction to study the communication between the E1o and E2o components. The LDo employed in these studies was fully lipoylated in vitro by E. coli lipoyl protein ligase (see Materials and methods), and lipoylation was confirmed by FT-MS [26]. Two LDo forms were identified by FT-MS: an LDo form with mass of 11421.2 Da that correlates well with the theoretical mass of LDo = 11420.4 Da, and an LDo form with mass of 11, 435.2 Da that differs from the previous form by 14 Da, possibly due to methylation of His 6 Átag as reported in the literature [32]. Both LDo forms could be nearly completely reductively succinylated by E1o and OG (>90% succinylation) (Fig. 3A) Fig. 3A for a typical experiment) and was approximately twice as large as the k cat of 48 s À1 determined from the overall OGDHc activity measurements in Table 2. These data indicate that with a sufficient amount of LDo in the reaction assay, the succinyl transfer from E1o to E2o is not a rate-limiting step (48 s À1 is k cat for overall activity vs. 99 s À1 for reductive succinylation). Importantly, the substrate analogues, 2-oxoadipate (k glutarylation = 22 s À1 ), 2-oxovalerate (k butyrylation = 0.0084 s À1 ), and pyruvate (k acetylation = 0.048 s À1 ), could also be used by E1o as substrates for reductive acylation of LDo; however, their efficiency was lower than that with OG (see Fig. 3A for OG and 3B for OA). Finally, this model reaction is an important control to assure that the acyl group derived from 2-oxoglutarate on E1o could be transferred to LDo before acyl-CoA formation in the active centers of E2o and could lead to determination of the catalytic rate constants for acyl transfer.
Residues Asp374 and His375 are the key catalytic residues in the active site of E2o Next, the residue-specific contribution to catalysis of succinyl-CoA formation in the active center of the E. coli E2o was studied. Results of this study would be applicable to all E2o components due to high sequence identities reported for all E2o core/catalytic domains [13][14][15][16][17][18][19]. Putative residues from the active center of the E2o involved in succinyl transfer and substrate binding were identified from the X-ray structure of the truncated cubic core of the E. coli E2o, known as the E2o catalytic or core domain [13,14] (Fig. 2A,  B). The following E2o active center variants were created to determine their individual contributions to the catalytic efficiency of the OGDHc complexes assembled from E1o, E2o (or its active center variants), and E3: Thr323Ala, Thr323Ser, Asp374Ala, Asp374Asn, His375Ala, His375Cys, His375Asn, Arg376Ala, and Asp379Ala (Fig. 2B, Table 2). The activity of the assembled OGDHc variants was measured in the overall assay for NADH production and in an E1-specific assay to study OG oxidation by OGDHc in the presence of the artificial electron acceptor 2,6-dichlorophenolindophenol (DCPIP). The E1-specific activity does not need the presence of the E2o and E3 components; however, substitutions in E2o could potentially affect assembly into OGDHc. As shown in Table 2, the Thr323 E2o substitutions to Ala or Ser do not significantly affect the NADH production as 45% (Thr323Ala E2o) and 43% (Thr323Ser E2o) of E2o activities were detected. From these results, it is unlikely that Thr323 in E2o is a catalytically important residue, contrary to the suggestion made based on the X-ray studies [7]. Similarly, the Arg376Ala E2o substitution led to 56% OGDHc activity remaining. In contrast, the Asp374Ala and Asp374Asn E2o substitutions revealed 2.1% and 9.7% of OGDHc The trace is a nonlinear regression fit to a single exponential rise to maximum (see Eqn 2), and the dashed line represents a linear fit to initial rate conditions. The rate constants were calculated from the initial slope. activity remaining, respectively. Approximately, 9.5% of the OGDHc activity was detected for Asp379Alasubstituted E2o ( Table 2). The His375Ala, His375Cys, and His375Asn E2o substitutions at the highly conserved histidine 375 residue led to 2.1%, 1.0%, and 0.5% activity remaining, respectively, but did not abolish the OGDHc activity. These findings are in agreement with data reported from Rutgers earlier for substitutions at residue His339 in the E2p of the E. coli pyruvate dehydrogenase complex (PDHc), a position analogous to His375 in E2o [33]. The activity of the E. coli PDHc was reduced to approximately 5.6% with the His399Ala E2p and 2.8% with the His399Cys E2p substitution, clearly indicating that cysteine substitution led to a lower activity of both E2os as compared with the Ala substitution [33]. When OGDH complexes assembled with E2o and its variants were tested with pyruvate (25 mM) or OV (45 mM) as potential substrates for chemo-enzymatic synthesis of acyl-CoA analogues, no NADH production was detected, indicating that neither acetyl-CoA nor butyryl-CoA could be produced (data not presented). It is also noteworthy that neither the E1o-specific activity of the reconstituted OGDH complexes, nor indeed the complex assembly itself, was affected by the indicated E2o substitutions ( Table 2).
The following could be concluded: two residues, Asp374 and His375 from the E2o active center are catalytically most important, as substitution of either residue for Ala impaired E2o catalytic efficiency by 54fold, while a 107-fold reduction was observed for the His375Cys E2o variant. Compared to the unsubstituted E2o, the His375Asn substitution lowered the catalytic efficiency by 178-fold, while the Asp374Asn E2o lowered it by only 11-fold. The rate retardation resulting from Ala substitution of all five active center residues here tested is~125 000-fold [this number represents the product of (k cat /K m ) wild-type /(k cat /K m ) variant ] for each Ala-substituted variant. This number could be compared to the 30 000-fold rate acceleration for acyl transfer between two aliphatic thiols reported by Hupe and Jencks in model systems [34].

Functional importance of Asp374 and His375 in independently expressed E2o catalytic domain
An alternative approach to study coupling of the E2o active center with the peripheral E1o and E3 components was developed by us recently [15]. It employs independently expressed E2o catalytic domain (CDo) in combination with a lipoyl domain source originated from E2o, such as lipoyl domain by itself (LDo) or Cterminally truncated E2o proteins, such as E2o 1-176 didomain, which consists of LDo, an E1o/E3 binding domain (peripheral subunit binding domain, PSBD) and flexible linkers connecting them (see Fig. 2D). For this approach, first, the functional competence of the independently expressed CDo in the overall OGDHc reaction needed to be proven. As shown in Figs 4 and 5, NADH production was detected on mixing of the CDo with varying concentrations of the LDo (0-70 lM) (Fig. 4A), or of E2o 1-176 didomain (0 -16 lM)  : lg). The basis of using lg ratio instead of lM is to measure absolute amount of protein and be consistent as we use different constructs of the enzyme in the following experiments. Activity was calculated per mg of E1o. No OGDHc activity was detected for E1o or E3 in the absence of E2o. b E1o-specific activity of the assembled OGDH complexes was measured in a DCPIP assay and was calculated per mg of E1o [23]. All assays were carried out in triplicates. The E2o* variants are the ones identified by site-saturation mutagenesis.
( Fig. 5), with E1o and E3 components added to complete the OGDH-like complex. While the rate of NADH production was proportional to the concentration of the lipoyl domain source, no saturation could be reached with addition of either LDo or E2o 1-176 didomain. Nor could saturation be reached with the Asp374Ala and His375Ala CDo variants (Fig. 4B,C).
In the absence of saturation apparent in Figs 4 and 5, the values of the second-order rate constants k cat /K m could be calculated from the initial slope for NADH production: 0.124 9 10 6 M À1 Ás À1 (CDo + LDo domains), 0.192 9 10 6 M À1 Ás À1 (CDo + E2o 1-176 didomain), 0.028 9 10 6 M À1 Ás À1 (CDo Asp374Ala + LDo), and 0.019 9 10 6 M À1 Ás À1 (CDo His375Ala + LDo), in comparison with 1.07 9 10 6 M À1 Ás À1 for E2o by itself (Table 3). Again, the catalytic efficiency of the Asp374Ala and His375Ala CDo variants was 4.4-and 6.5-fold lower than that for unsubstituted CDo. These results further affirm the contribution of Asp374 and His375 to E2o catalysis. Another important conclusion from this study is that the reaction intermediates could be transferred between the active centers of E1o, CDo, and E3 in the presence of lipoyl domain source in the reaction assay with no covalent bond linking CDo and lipoyl domain source. This is the second example from Rutgers, as similar findings were reported earlier for the E. coli E2p [15]. In support of these striking conclusions, the following controls were performed: (a) No OGDHc activity was detected for LDo on its own in the absence of CDo; (b) negligible or no OGDHc activity could be detected for a mixture of E1o, E3, and CDo, but in the absence of LDo, thereby excluding any contamination from intrinsic wild-type OGDHc. The results indicate that it will be sufficient to use independently expressed E2o domains for future active center engineering of E2o.
Further evidence for the catalytic importance of Asp374 and His375 from the rate of succinyl transfer catalyzed by the E2o catalytic domain To further substantiate the result of the His375Ala and Asp374Ala substitutions in the catalytic domain of E2o, an experiment was designed to study the succinyl transfer per se in the reverse reaction according to Eqn (3), thus enabling us to determine rate constants for the reaction taking place in the E2o active center exclusively. A mixture of dihydro-LDo [LDo reduced by tris(2-carboxyethyl)phosphine, TCEP] and CDo or its variants was reacted with succinyl-CoA according to Eqn (4) where CDo or variants are the catalysts for the reaction.
CDo succinyl-CoA Direct measurement of the masses of dihydro-LDo and succinyldihydro-LDo by FT-MS was used to quantify the rate of succinyl transfer from succinyl-CoA to the dihydro-LDo. At concentrations of dihydro-LDo (40 lM), succinyl-CoA (0.15 mM), and CDo (0.05 lM), the dihydro-LDo is a substrate, CDo is a catalyst, and succinyldihydro-LDo is a product in Eqn (3). The reaction was stopped at different times (5-120 s) by diluting into 50% methanol and 0.1% formic acid, and the samples were analyzed by FT-MS. The ratio of succinyldihydro-LDo / total LDo (the sum of succinyldihydro-LDo and dihydro-LDo) was plotted versus time (Fig. 6A). As seen from Fig. 6A, approximately 60% of dihydro-LDo is converted to succinyldihydro-LDo with a rate constant of 19.9 s À1 as calculated from a linear fit to the initial rate of succinyldihydro-LDo formation ( Table 3). It should be noted that some time-dependent formation of succinyldihydro-LDo was detected by FT-MS even in the absence of CDo in the reaction assay. While this nonenzymatic reaction was insignificant with unsubstituted CDo, the slope of this nonenzymatic reaction was taken into account when rate constants were calculated for the His375Ala and Asp374Ala CDo variants (Fig. 6B,C). As seen in Fig. 6 and Table 3, the rate constant of 1.86 s À1 (Asp374Ala) and of 4.9 s À1 (His375Ala) was 11 times and 4 times smaller, respectively, compared to the rate constant of 19.9 s À1 for unsubstituted CDo. A ratio of succinyldihydro-LDo/total LDo of~0.55-0.60 (reproducible with wild-type and variant CDos in Fig. 6) was observed. The data above, and presented in Table 3, support the conclusion that Asp374 and His375 are important residues for succinyl transfer in both directions, as they need to be, while also suggesting an equilibrium constant near unity for the reaction in Eqn (3).

Site-saturation mutagenesis on the E2o active center His375
To gain further insight into the possible role of His375, we carried out site-saturation mutagenesis on this residue and screened for possible variants that display OGDHc activity.
The following His375 variants were identified: His375Trp E2o with a significant retention of the overall NADH activity (60%); and His375Gly E2o with significantly reduced NADH activity (13%) compared to unsubstituted E2o (Table 2). Steady-state kinetic experiments revealed an approximately twofold reduction in k cat for the His375Trp E2o (29 s À1 ) and about eightfold reduction for the His375Gly E2o (6 s À1 ) compared to E2o (48 s À1 ) ( Table 2). No significant changes in values of K m with respect to OG or CoA were detected for the variants in comparison to the unsubstituted E2o (Table 2). To further test whether His375 participated as a general acid-base catalyst, the pH dependence of the overall NADH activity of the His375Trp E2o and of the unsubstituted E2o, both assembled with the E1o and E3 components, was compared and revealed no difference in their shapes. The pK app for the alkaline limb of the curve was 8.8 for E2o and 8.7 for the His375Trp E2o variant (Fig. 7). This parallel behavior of the pH dependence with His   or Trp effectively rules out an acid-base role for His375.
As there is no tryptophan residue in the E2o active center, we could also utilize the intrinsic fluorescence of the newly installed Trp in position 375 of the His375Trp E2o variant to determine the nearly equal K d of CoA (61.2 lM) and succinyl-CoA (47 lM) from quenching of fluorescence on addition of either compound (Fig. 8). As control, the E2o experienced no fluorescence quenching, thus also confirming substitution at the active center. In view of the modest decrease in k cat /K m with the Trp substitution at position 375, these K d values are appropriate for the unsubstituted E2o, and indicate similar binding affinity of E2o for CoA and succinyl-CoA, where both succinyl and CoA moieties appear to contribute to binding.

Conclusion
This paper is directed to an elucidation of the fundamental mechanism of the transthioesterification reaction carried out by the E2o of the E. coli OGDHc that would be applicable to all E2o components due to high sequence identities reported for the E2 catalytic domains [13][14][15][16][17][18][19]. Important classes of enzymes carrying out similar reactions include inteins in expressed protein ligation [35] and many reactions on the polyketide pathways [36]. Based on our studies, the following important results emerged.
(a) Identification of catalytically important Asp374 and His375 residues of the E2o catalytic domain and the rate acceleration provided by these residues. The other residues tested displayed much smaller contribution to catalysis. Using the E2o variants substituted by alanine at the putative catalytic center for transthioesterification, NADH production as a measure of the complex activity indicated that the five residues tested accounted for 2.23 (Thr323Ala), 54 (Asp374Ala and His375Ala CDo. The reaction assay was similar to that in B except of 40 lM dihydrolipoyl-LDo was used. The traces are the nonlinear regression fit to a single exponential rise to maximum, and the dashed line represents a linear fit to initial rate conditions. For rate constants calculation in B and C, the slope in the control experiment was subtracted from that in the experiment. each), 1.78 (Arg376Ala), and 11 (Asp379Ala)-fold rate accelerations for a total of~125 000-fold. This is the first identification of partial catalytic rate constants for this important reaction and raises the question: Which likely mechanism is consistent with the results? (b) A closer look at the transthioesterification mechanism was provided by the use of independently expressed E2o domains (LDo, E2o (d) These results raise the question: Which likely mechanism of transthiolacylation is consistent with the experimental findings? Based on numerous precedents, there are two prominent likely mechanisms to account for the transfer of acyl group between two thiols depicted by the step characterized by the rate constants k 6 and k -6 in Fig. 1. A general acid-base mechanism would suggest that the His375 converts the thiol of the attacking nucleophile (CoASH) to a thiolate anion (CoAS -) as depicted on pathway B in Fig. 9. The activated thiolate then attacks the carbonyl atom of the succinyldihydrolipoyl-E2 to form a tetrahedral intermediate, which is stabilized by the The reaction was initiated by CoA (0.13 mM) and OG (2 mM) and NADH production was recorded at 30°C for 1 min. The values of activity were plotted to a curve defined by one ionizing group according to Eqn (4).
where x is the value of pH. hydroxyl side chain of Thr323 [14]. This notion was long accepted based on analogy with a mechanism developed for chloramphenicol acetyltransferase (CAT) [20,21,37]. When the analogous His195 in chloramphenicol acetyltransferase was replaced by Ala, a 9 9 10 5 -fold decrease in k cat and virtually no change in K m was observed, that is, a (k cat /K m ) wild-type CAT /(k cat /K m ) His195Ala CAT = 500 000 [15]. In contrast, the (k cat /K m ) unsubstituted E2o / (k cat /K m ) His375Ala = 54. This difference strongly suggests different functions for the highly conserved active center histidine on the two enzymes.
As a 'gold standard' for acid-base catalysis by a histidine side chain, we also present comparison with the much studied His64 at the catalytic triad of the subtilisin family of serine proteases, where the His64Ala substitution led to a reduction of 10 6 -fold in k cat /K m , a number similar to that observed for CAT [38].
As an alternative to acid-base catalysis, for a protein with a pK a for the cysteine thiol group near pH 7.0, one could envision direct attack by a thiolate anion on the thiol ester carbon, forming an essentially symmetrical tetrahedral oxyanionic intermediate, in which the central carbon atom is flanked by two C-S bonds, where either C-S bond has nearly equal probability for cleavage (Fig. 1, lower left and Fig. 9 pathway A) [34]. We suggest that the nearly equal magnitude of the rate acceleration provided by the His375 and the Asp374 residues, more likely, reflects their roles in stabilizing the oxyanion by two hydrogen bonds (Fig. 9, pathway A), in other words creating a so-called 'oxyanion hole'. It had been demonstrated some years ago that in subtilisin, where a putative transition state stabilizing oxyanion hole is created by one main chain hydrogen bond donor and one side chain hydrogen bond donor (Asn155), substitution of Asn155 to Leu had small effect on the K m but reduced the k cat by a factor quoted as 200-to 300-fold [39]. This is very similar in magnitude to the k cat /K m reductions observed (54 for each) with the Asp374Ala and His375Ala substitutions reported in this study (again no change in K m was observed). Our results also concur with homology modeling studies between the enzymes chloramphenicol acetyl transferase and dihydrolipoamide acetyl transferase about existence of a His-Asp-Gly consensus in the catalytic core which is likely to play an important role [21].
While well accepted, it has been difficult to obtain experimental proof for the existence of an oxyanion in solution; as an example, proton inventory solvent kinetic isotope effect studies, a very subtle method, failed to do so [40].
The proposed hydrogen bond donation by Asp374 and His375 was tested by studying the alternative Asp374Asn substitution in E2o that allows hydrogen bond donation similar to His375 ( Table 2). The His375Asn substitution led to 4.0-fold further reduction in the E2o catalytic efficiency (to 0.5% of the remaining activity) compared to the His375Ala substitution (to 2.1%). In contrast, the Asp374Asn substitution partially rescued the E2o catalytic efficiency (9.7%) compared to Asp374Ala E2o (2.1% activity) and suggested that Asp374 is more likely to be involved in hydrogen bond formation. The result also raises an important implication that the carboxyl group of Asp374 has an unusually elevated pK a as only in the conjugate acid form could it serve as a hydrogen bond donor. As this step appears to be part of rate limitation, we further speculate that the Fig. 9. Likely mechanisms of succinyl transfer from dihydrolipoamide-E2o to CoASH. Pathway A: direct attack by the conjugate base thiolate anion of CoASH assuming a low pK a for the CoASH, or by the thiol form itself. Pathway B: initial proton transfer from CoASH to His375 forming the conjugate base CoAS À , which is the attacking agent. Both pathways then proceed by an oxyanionic tetrahedral intermediate consistent with model studies [34].