Nuclear Magnetic Resonance Evidence for the Role of the Flexible Regions of the E1 Component of the Pyruvate Dehydrogenase Complex from Gram-negative Bacteria

Abstract

Most bacterial pyruvate dehydrogenase complexes from either Gram-positive or Gram-negative bacteria have E1 components with an α₂ homodimeric quaternary structure. In a sequel to our previous publications, we present the first NMR study on the flexible regions of the E1 component from Escherichia coli and its biological relevance. We report sequence-specific NMR assignments for 6 residues in the N-terminal 1–55 region and for a glycine in each of the two mobile active center loops of the E1 component, a 200-kDa homodimer. This was accomplished by using site-specific substitutions and appropriate labeling patterns along with a peptide with the sequence corresponding to the N-terminal 1–35 amino acids of the E1 component. To study the functions of these mobile regions, we also examined the spectra in the presence of (a) a reaction intermediate analog known to affect the mobility of the active center loops, (b) an E2 component construct consisting of a lipoyl domain and peripheral subunit binding domain, and (c) a peptide corresponding to the amino acid sequence of the E2 peripheral subunit binding domain. Deductions from the NMR studies are in excellent agreement with our functional finding, providing a clear indication that the N-terminal region of the E1 interacts with the E2 peripheral subunit binding domain and that this interaction precedes reductive acetylation. The results provide the first structural support to the notion that the N-terminal region of the E1 component of this entire class of bacterial pyruvate dehydrogenase complexes is responsible for binding the E2 component.

Introduction

The pyruvate dehydrogenase complex (PDHc)³ is a molecular machine consisting of multiple copies of three distinct proteins in bacteria (5 MDa) (1, 2). In most bacteria (both Gram-negative and Gram-positive) the complex has octahedral symmetry. The PDHc catalyzes the oxidative decarboxylation of pyruvate according to Equation 1.

The components of the Escherichia coli PDHc have the following roles (Equations 2–6) (Refs. 3–5). The thiamin diphosphate (ThDP)-dependent E1 component (E1ec, pyruvate dehydrogenase) is an α₂ homodimer and carries out consecutively pyruvate decarboxylation and reductive acetylation of the E2 component (Equation 3). E2 (E2ec, dihydrolipoamide acetyltransferase) has a dual function. Its covalently attached lipoamide is reductively acetylated by the E1 component and pyruvate, and subsequently, the acetyl group is transferred to coenzyme A (CoA), and the principal metabolic product acetyl-CoA is released (Equation 4). E3 (E3ec, dihydrolipoamide dehydrogenase) has tightly bound FAD and NAD⁺; it regenerates the lipoamide from the reduced dihydrolipoamide form (Equation 5) (Scheme 1).

Our laboratory has been studying the PDHc from E. coli (PDHc-ec) as a representative of this large class of enzymes (6–10). In the x-ray structure of E1ec with ThDP, there were identified three disordered regions corresponding to the N-terminal 1–55 residues, to amino acids spanning residues 401–413 (inner active center loop) and 541–557 (outer active center loop) (7). We have now shown that all three of these regions have important function. To study the function of the N-terminal region of E1ec, we followed up on earlier work of de Kok and co-workers (11, 12) on the related enzyme from Azotobacter vinelandii to elucidate the loci of binary interactions in the E1ec-E2ec complex. In our earlier study (13) we used the E1ec deletion variants (Δ16–25, Δ26–35, Δ36–45, and Δ46–55), single-site substituted variants in the region 7–15, and mass spectrometric analysis to confirm that the N-terminal region, although not seen in any x-ray structures of E1ec (7, 14, 15), is important for both overall activity of the complex and for interaction with the E2ec. We have extended the study to the entire N-terminal region of E1ec and here report the importance of this region in the interaction with E2ec. In a series of reports (9, 14, 16), we also demonstrated (a) the functional importance of the two active center loops in the organization of the active center during the reaction sequence and (b) that loop movement is correlated with catalysis. In this paper we explored the possibility that regions too mobile to be seen in the x-ray structure may give rise to resolvable resonances in the NMR spectrum notwithstanding the size of E1ec (2 × 886 residues for an M_r of 200,000) (3). Our expectations were met; we have for the first time made sequence-specific NMR assignments for six amino acids in the N-terminal region of E1ec, which was made possible by the presence of some remarkably sharp resonances in the two-dimensional HSQC spectra. In addition, we also assigned one glycine resonance in each of the two active center loops. In fact, the total number of observable resonances corresponds well to the number expected on the basis of sequence, and the NMR observations pertain virtually exclusively to the three flexible regions. These “reporters” from the three regions were then examined in the presence of (a) a reaction intermediate analog known to affect the mobility of the active center loops, (b) an E2 component construct consisting of a lipoyl domain and peripheral subunit binding domain (PSBD), (c) a peptide corresponding to the amino acid sequence of the PSBD, and (d) the independently expressed lipoyl domain. The results show that the PSBD is the most important domain for recognition of E1 rather than the lipoyl domain. The results also correlate the functional and structural data from our laboratories.

FIGURE 6.

Far-UV CD spectrum of N-term peptide corresponding to residues 1–35 of E1ec at pH 7 and 5 at 0.10 mg·ml.

SCHEME 1.

Mechanism of the E1 and E2 components of PDHc-ec.

EXPERIMENTAL PROCEDURES

Materials

The Wizard® Plus Miniprep DNA purification system was used for purification of DNA (Promega, Madison, WI). The QuikChange site-directed mutagenesis kit was used for single-site substitution (Stratagene, La Jolla, CA). DNA sequencing was done at the Molecular Resource Facility of the New Jersey Medical School (Newark, NJ). The E. coli AT2457 strain (glycine auxotroph) was from E. coli Genetic Stock Center (Yale University). E. coli BL21(DE3) pLys cells were from Novagen (Novagen, EMD Chemicals, Gibbstown, NJ). The synthetic peptide corresponding to the N-terminal 1–35 amino acids of the E1ec was from SynPep (Dublin, CA). The synthetic peptide with sequence H₂N-ATPLIRRLAREFGVNLAKVKGTGRKGRILREDVQAYVKEAI-OH corresponding to the E2ec PSBD was from CHI Scientific (Maynard, MA). RbCa TXN SALTS used to make super-competent cells was from BIO101, Inc. (Vista, CA). The 18 unlabeled l-amino acids used in the minimal medium were from Sigma, with the exception of l-Lys from EMD Chemicals, Inc. (La Jolla, CA). [¹⁵N]Glycine, [1-¹³C]glycine, and ¹⁵NH₄Cl were from Cambridge Isotope Laboratories, Inc. (Andover, MA).

Construction of Single-substituted N-terminal Variants and Expression and Purification of E1ec and Its Variants

See Ref. 17 and supplemental Experimental Procedures.

Expression and Purification of 1-lip E2ec

See Ref. 18 and supplemental Experimental Procedures.

Activity and Related Measurements

The activity of E1ec and its variants was measured after reconstitution to PDHc complex with E2ec and E3ec components using a E1ec:E2ec:E3ec mass ratio of 1:1:0.5 (19). The E1-specific activity of E1ec and its variants was measured in the model reaction with 2,6-dichlorophenolindopenol (DCPIP) as the external oxidizing agent by monitoring the reduction of DCPIP at 600 nm (20).

Size-exclusion Chromatographic Test of the Dimerization of E1ec and Its Variants

Fast protein liquid chromatography Superose 6 10/300 GL column (Amersham Biosciences) was equilibrated with 50 mm KH₂PO₄ (pH 7.0) containing 0.15 m NaCl. The column was calibrated using the following standards: thyroglobulin (669,000), ferritin (440,000), catalase (232,000), aldolase (158,000), bovine serum albumin (67,000), and ovalbumin (43,000). The proteins (1∼2 mg) were eluted at a solvent flow rate of 0.4 ml/min and monitored by UV at 280 nm.

Interaction of E1ec and Its Variants with E2ec Monitored by Size-exclusion Chromatography

E1ec and its variants were incubated with 1-lip E2ec in 50 mm KH₂PO₄ (pH 7.0) containing 0.15 m NaCl at 25 °C (molar ratio of E1ec:1-lip E2ec = 1:1). After 45 min, samples were centrifuged for 10 min at 17,530 × g, and the subcomplex was applied to the column. Circular dichroism (CD) experiments were carried out on an Applied Photophysics Chirascan spectrometer (Leatherhead, UK) at 30 °C as reported earlier from our laboratories (8, 10, 20).

Reductive Acetylation of the Lipoyl Domain and of 1-lip E2ec Monitored by MALDI-TOF/TOF Mass Spectrometry

See Refs. 18 and 21 and supplemental material.

Construction of Plasmid, Expression, and Purification of C-terminal His₆-tag E1ec

The pGS878 plasmid encoding E1ec was used as a template. The XbaI and XhoI restriction sites were created at the beginning and at the end of the E1ec gene by site-directed mutagenesis (22). The DNA fragment of 2307 bp encoding E1ec was subcloned into the pET-22b(+) vector pretreated with the same restriction enzymes. The resulting plasmid pET-22b-E1ec (8028 bp) with His₆-tag at the C-terminal end was transformed into the BL21(DE3) pLysS-competent cells.

For His₆-tag E1ec expression, the overnight culture from a single colony on the plate was inoculated into 5 flasks with 800 ml of LB medium in each containing 50 μg/ml ampicillin, and cells were grown for 1–2 h to reach an A₆₅₀ = 0.800 at 37 °C. The temperature was lowered to 20 °C, and 0.50 mm isopropyl 1-thio-β-d-galactopyranoside was added to start protein expression. Cells were collected after 4–5 h, washed with 20 mm KH₂PO₄ (pH 7.0) containing 0.1 m NaCl, and stored at −20 °C. About 20 g of cells were collected from 4.0 liters of culture.

The His₆-tag E1ec was purified using a Ni²⁺-Sepharose 6 Fast Flow column (GE Healthcare) according to the protocol supplied. Briefly, the cells were dissolved in 20 mm KH₂PO₄ (pH 7.0) containing NaCl (0.10 m), ThDP (1 mm), MgCl₂ (2 mm), benzamidine HCl (1 mm), and phenylmethylsulfonyl fluoride (1 mm). Lysozyme (final concentration of 0.60 mg/ml) was added, and the cells were incubated 20 min on ice. To remove the nucleic acids, DNase (1000 units) and nuclease (500–1000 units) with 5 mm MgCL₂ were added, and cells were incubated for an additional 1 h on ice. Cells were destroyed by sonication, and the supernatant was clarified by centrifugation at 28,978 × g for 30 min. The clarified supernatant was applied to the Ni²⁺-Sepharose 6 Fast Flow column (5–10 ml volume). The His₆-tag E1ec was eluted with elution buffer containing 20 mm KH₂PO₄ (pH 7.4), NaCl (0.5 m), MgCl₂ (2 mm), ThDP (1 mm), and 150 mm imidazole. The fractions containing His₆-tag E1ec were collected, concentrated, and dialyzed against the 20 mm KH₂PO₄ (pH 7.0) containing ThDP (0.20 mm), MgCl₂ (2.0 mm), DTT (1.0 mm), EDTA (0.50 mm), and benzamidine hydrochloride (1 mm). Alternatively, the His₆-tag E1ec can be additionally purified on a DEAE column (see below under “Uniform ¹⁵N labeling of His₆-tag E1ec”).

To create the Q18H, Q33H, and Q38H variants of His₆-tag E1ec, the following primers were used (substituted bases are underlined): Q18H, 5′-CTCGCGACTGGCTCCATGCGATCGAATCG-3′; Q33H, 5′-GGTGTTGAGCGTGCTCATTATCTGATCGACCAACT-3′; Q38H, 5′-AGTATCTGATCGACCATCTGCTTGCTGAAGCCC-3′. To create the G402A and G542A variants of His₆-tag E1ec, the following primers were used (substituted bases are underlined): G402A, 5′CGACGCGGCTGAAGCTAAAAACATCGCGC3′; G542A, 5′-TTACAGCCCGAACGCTCAGCAGTACACCC-3′.

Construction of Plasmid, Expression, and Purification of C-terminal-truncated 1-lip E2ec^1–190

The pET-22b(+)-1-lipE2ec (6946 bp) plasmid encoding 1-lip E2ec was used as the template (18), and the amplification primer 5′-GGTGGACTTCAGCTAATTTGGTGAAATCGAAGAAGTG-3′ (the mutated codon is underlined) and its complement were used for site-directed mutagenesis to introduce the TAA stop codon in place of Lys-191, which is located at the N-terminal region of the catalytic domain of the 1-lip E2ec. The 1-lip E2ec^1–190 was well expressed in BL21(DE3) cells at 37 °C with no parental 1-lip E2ec detected by SDS-PAGE. The culture was grown in LB medium with 50 μg/ml ampicillin supplemented with 0.30 mm lipoic acid, and protein expression was induced by 0.50 mm isopropyl 1-thio-β-d-galactopyranoside for 5–6 h at 37 °C. The expressed protein was purified using consecutively a DEAE anion exchange column and a G3000SW TSK size-exclusion column with a high performance liquid chromatography system. About 7 mg of protein was obtained from 6–7 g of E. coli cells (2.4 liters of LB medium). The identity of 1-lip E2ec^1–190 was confirmed by N-terminal sequencing for nine amino acids: AIEIKVPDI.

Preparation of Labeled E1ec Variants for NMR Studies

Uniform ¹⁵N Labeling of His₆-tag E1ec (23)

20 ml of overnight culture from a single colony on the plate was grown in LB medium containing 50 μg/ml of ampicillin. After 15 h of growth at 37 °C, cells were collected by centrifugation (2236 g, 10 min, 4 °C) and washed with M9 minimal medium (24) supplemented with 1g/liter of [¹⁵N]NH₄Cl and sodium acetate (2 mm) to remove the traces of LB medium. The cells were collected by centrifugation, dissolved in 25–30 ml of the minimal medium, and inoculated into 700 ml of the minimal medium supplemented with [¹⁵N]NH₄Cl. Cells were grown to A₆₀₀ = 0.60–0.80 at 37 °C, then the temperature was lowered to 20 °C, and protein expression was induced by 0.50 mm isopropyl 1-thio-β-d-galactopyranoside for 15 h. Cells were collected by centrifugation at 2 236 × g for 7 min and washed with 20 mm KH₂PO₄ (pH 7.0) containing 0.25 mm EDTA and 0.10 m NaCl and stored at −20 °C. The ¹⁵N-labeled E1ec was purified using Ni²⁺-Sepharose 6 Fast flow column similarly to that presented for His₆-tag E1ec above. In addition, the ¹⁵N-labeled E1ec was purified using an anion exchange DEAE column and a 0–0.50 m gradient of NaCl in 20 mm KH₂PO₄ (pH 7.5). The protein was dialyzed against 20 mm KH₂PO₄ (pH 7.0) containing 0.20 mm ThDP, 2.0 mm MgCl₂, 1.0 mm DTT, and 0.50 mm EDTA (dialysis buffer), concentrated using a Centriprep 30 (Millipore) concentrating unit, and stored at −20 °C. Approximately100 mg of purified ¹⁵N-labeled protein was obtained from 3.5 liters of culture.

Selective ¹⁵N Labeling and ¹³C-¹⁵N Double Labeling of His₆-tag E1ec

Selective ¹⁵N and ¹³C,¹⁵N enrichment of glycine residues in E1ec was achieved by transformation of the pET-22(b)-E1ec plasmid into the E. coli strain AT2457 glycine auxotroph. The E. coli AT2457 super-competent cells were prepared using RbCa TXN SALTS. Cells were grown on the minimal salt medium supplemented with 18 unlabeled l-amino acids and 0.35 g·liter⁻¹ [¹⁵N]glycine for ¹⁵N-selective labeling or 0.35 g·liter ⁻¹ [¹⁵N]glycine and 0.35 g·liter ⁻¹ of [1-¹³C]glycine for ¹³C,¹⁵N double-labeling of E1ec, similarly to that presented above for uniform ¹⁵N labeling of E1ec (25, 26). Approximately 46 mg of protein was purified from 1.0 liters of culture.

NMR Spectroscopy

Sample Preparation for NMR Studies

The concentration of His₆-tag E1ec in the NMR sample was measured using Bio-Rad Protein Assay Dye Reagent Concentrate and adjusted to 20 mg/ml (subunit concentration = 211 μm) in 0.4 ml of dialysis buffer (see above), then 0.15 m NaCl and 7% D₂O were added. The pH of the NMR samples was adjusted to 7.0 by the addition of 0.15 m CH₃COOH or 0.15 m NaHCO₃.

NMR Spectroscopy

All NMR experiments were performed on a Varian INOVA 600 MHz spectrometer at 20 °C. Two-dimensional ¹H,¹⁵N TROSY-HSQC (27) spectra were recorded for ¹⁵N-labeled His₆-tag E1ec and all its singly substituted variants and for the [¹⁵N]glycine containing variant. A ¹³C,¹H,¹⁵N triple resonance three-dimensional HNCO (28) NMR experiment was performed for the ¹³C,¹⁵N double-labeled glycine residues in His₆-tag E1ec, and the NH plane from the three-dimensional data was analyzed to assign Gly-47. Two-dimensional ¹H,¹⁵N HSQC (29) spectra were recorded for the 1–35 ¹⁴N synthetic peptide. NMR Pipe (30) and NMR ViewJ (31) were used for data-processing of all spectra.

Two-dimensional HSQC NMR Method to Identify the Region of E1ec Interacting with E2^1–190, Lipoyl Domain, and PSBD

For the HSQC NMR experiment, the ¹⁵N-labeled His₆-tag E1ec (20 mg/ml, concentration of subunits = 211 μm) in 20 mm KH₂PO₄ (pH 7.0) containing 0.20 mm ThDP, 2.0 mm MgCl₂ 1.0 mm DTT, and 0.50 mm EDTA was mixed with 1-lip E2ec^1–190 (422 μm) in the presence of 0.10 m NaCl and 10 mm DTT in a total volume of 0.50 ml, then incubated for 1 h at 25 °C. D₂O (7%) was added before the NMR spectrum was recorded. Similar conditions and concentrations were used for the experiments on His₆-tag E1ec in the presence of PSBD and lipoyl domain.

Two-dimensional HSQC-TROSY NMR to Identify the Region of E1ec Interacting with C2α-phosphonolactyl-ThDP

Two-dimensional HSQC-TROSY spectra were recorded for the His₆-tag E1ec and its G402A and G542A variants (20 mg/ml, concentration of subunits = 211 μm) in 20 mm KH₂PO₄ (pH 7.0) containing 0.20 mm ThDP, 2.0 mm MgCl₂ 1.0 mm DTT, and 0.50 mm EDTA. To identify the region of E1ec interacting with C2α-phosphonolactyl-ThDP (PLThDP), apoE1ec (20 mg/ml) was prepared by dialysis against 20 mm KH₂PO₄ (pH 7.0) containing 2.0 mm MgCl₂ 1.0 mm DTT, and 0.50 mm EDTA, and apoenzyme was titrated by CD to saturation with 1 mm PLThDP, then 0.15 m NaCl and 7% D₂O were added for NMR experiment.

RESULTS AND DISCUSSION

Functional Evidence for the Importance of the Entire N-terminal E1ec Region in Its Interaction with E2ec

The sequences of E1 components from Gram-negative bacteria reveal a large percentage of highly conserved acidic amino acid residues in their N-terminal 1–55 region (13). The patch of highly conserved positively charged residues in the PSBD of E2ec led us to hypothesize that the negatively charged residues in the N-terminal region of E1ec interact with the positively charged residues on the PSBD (32). Earlier, we reported that the negatively charged residues in the N-terminal region of E1ec Asp-7, Asp-9, Glu-12, Asp-15, and a positively charged Arg-14 are important for the overall PDHc activity and for the interaction with E2ec (13). To extend our study to the entire N-terminal E1ec region, the E21R, E26R, E27R, E30R, D37R, and E42R variants were created and analyzed. Because the nature of the substitution of the negatively charged residues such as Glu-12 to Arg, Asp, or Gln appeared to make a modest difference in the overall PDHc activity (Table 1), we created E1ec variants with Arg substitutions, assuming that substitution of the negatively charged residue to a positive will be informative.

TABLE 1.

Kinetics and assembly of N-terminal-substituted E1ec

N-terminal substitution in E1eca	Overall PDHc activityb	E1-specific activityc	Kd,MAP	Ability to form E1-E2 complexd
	%	%	μm
none	100	100	0.159	+
D7A	0.14	90		+ −
D9A	Not detected	88		+ −
D9R	3	29	0.199	+ −
P10A	100	80		+
I11A	94	62		+
E12R	3.4	83	0.100	+ −
E12D	2.3	10		+ −
E12Q	3.3	41		+ −
T13A	13.0	74		+
R14A	0.25	100		+ −
D15A	0.19	84		+ −
E21R	69	89		+
E26R	87	89		+
E27R	2.7	67	0.068	+ −
E30R	2.9	61	0.296	+ −
D37R	74	75		+
E42R	100	86		+

^a Italics indicate data from Ref. 13.

^b The k_cat for E1ec in the overall PDHc reaction after reconstitution with E2ec and E3ec was 90 s⁻¹.

^c The k_cat for E1ec in the model reaction with DCPIP was 1.20 s⁻¹.

^d +, only E1ec-E2ec complex; + −, E1ec and E1ec-E2ec subcomplex. Strength of subunit interaction: + > + −.

Negatively Charged Residues in the Entire N-terminal Region of E1ec Have No Role in Either Formation of the First Covalent Intermediate or in Decarboxylation

In addition to the negatively charged residues studied earlier (Asp-7, Asp-9, Glu-12, Asp-15), Glu-27 and Glu-30 displayed low overall activity (2.7 and 2.9%, respectively) on reconstitution with E2ec-E3ec subcomplex (Table 1), whereas the others had activities of 69–100%. However, none of the substitutions in E1ec here reported produced significant activity reduction according to the E1-specific 2,6-dichlorophenolindopenol assay, suggesting that the reactions were little affected through decarboxylation (Scheme 1, reactions k₁, k₂, and k₃). The addition of substrate analogue methyl acetylphosphonate (MAP) to ThDP to the E27R and E30R variants revealed formation of a tetrahedral intermediate PLThDP by CD via its 1′,4′-iminopyrimidine tautomeric form, as demonstrated before (33, 34). The values of K_d,_MAP of 0.068 μm (E27R) and 0.296 μm (E30R) were similar to the K_d,_MAP = 0.159 μm for E1ec, demonstrating that the formation of the first predecarboxylation intermediate was not affected by these substitutions either (Scheme 1, k₁, k₂).

Reductive Acetylation of the Lipoyl Domain and of Intact E2ec Are Affected Differentially by N-terminal E1ec Variants

As pointed out in our earlier publication on this topic, there are two loci of interaction between E1ec and E2ec; (a) between ThDP and the lipoyl domain, the site of reductive acetylation, and (b) between the N-terminal region of E1ec and E2ec (suggested to be the PSBD, perhaps with additional interaction with the N-terminal region of the catalytic domain). We had developed a MALDI-TOF mass spectroscopic method to monitor the reductive acetylation of either the independently expressed lipoyl domain or of the entire E2ec by E1ec and its variants in the presence of pyruvate (18, 21). Recently, we extended this method to MALDI-TOF-TOF and Fourier transform mass spectrometry. On incubation of E27R and E30R E1ec variants with lipoyl domain (0.30 or 0.60 mm) in the presence of pyruvate and ThDP, only the acetylated form of the lipoyl domain was detected within 1 min of incubation as with E1ec (a molecular mass of acetylated dihydrolipoyl domain of 9,019.2 Da was detected resulting from the mass of the lipoyl domain of 8,975 Da + 44).

Incubation of the intact 1-lip E2ec component with E1ec in the presence of pyruvate for 2 min at 25 °C followed by tryptic digestion led to the appearance of the lipoyl domain with a mass of 10,160.9 Da, clearly signaling an acetylated lipoyl domain (the molecular mass of the unacetylated lipoyl domain liberated by trypsin is 10,119 Da). The time course for reductive acetylation as monitored by MALDI-TOF/TOF was significantly different between E1ec and its D9R, E12R, and E30R variants. Within 2 min, only the acetylated lipoyl domain was detected with E1ec, but with the D9R, E12R, and E30R variants, we could monitor the time course of interconversion of the unacetylated and acetylated forms of the lipoyl domain of 1-lip E2ec. With increasing incubation time, the peak for the acetylated form of the lipoyl domain increased, whereas the peak corresponding to the unacetylated form diminished in size. Even after 30 min of incubation, the unacetylated forms were still observed. A typical time-course of the reductive acetylation of 1-lip E2ec by D9R E1ec and pyruvate is shown in Fig. 1. The E12R and E30R E1ec variants displayed the same behavior, indicating that substitutions of the negatively charged residues in the N-terminal region of E1ec affect its assembly with the E2ec and, as a result, the overall PDHc reaction rate. According to these experiments, binding of the lipoyl domain of E2ec to E1ec does not limit the rate of reductive acetylation of intact E2ec or of the overall rate of the complex, as is also supported by the structural studies presented below.

FIGURE 1.

Time course of reductive acetylation of 1-lip E2ec by pyruvate and D9R E1ec monitored by MALDI-TOF/TOF mass spectrometry. A, 2 min. B, 5 min; C, 15 min. D, 30 min. The lower mass peak corresponds to the unacetylated lipoyl domain excised by trypsin, and the higher mass corresponds to the acetylated domain.

The E1ec-E2ec Interaction Is Affected by N-terminal E1ec Substitutions According to Size-exclusion Chromatography

Using an analytical size-exclusion chromatographic column calibrated with standards in the mass range of 67,000- 669,000 Da, we demonstrated that all N-terminal-substituted variants are eluted from the column as a dimer, similar to E1ec (retention time of 38.96 min, which corresponds to a molecular weight of dimer; theory, 198,948 Da) (see supplemental Fig. S1).

On assembly of E1ec with E2ec, only one peak corresponding to their subcomplex was detected by analytical size-exclusion chromatography and SDS-PAGE (not shown). The E21R, E26R, D37R, and E42R E1ec variants were apparently converted quantitatively to the complex similarly to E1ec, perfectly accounting for their high overall PDHc activity (69–87% in Table 1). Incubation of E2ec with the D7A, D9A, D9R, E12R, E12D, E12Q, R14A, D15A, E27R, and E30R E1ec variants gave two peaks; according to SDS-PAGE, the first peak to elute contained the E1ec-E2ec subcomplex, and the second peak contained only E1ec. Hence, whereas these variants displayed impaired interaction with E2ec (explaining their low overall PDHc activity), their normal E1-specific activity could be attributed to the fact that the latter is measured in the absence of E2 and E3 components (Scheme 1, k₁, k₂, k₃).

Structural Evidence for the Interaction of the Flexible Regions of E1ec with E2ec

To undertake the NMR studies reported below (the ultimate goal of this section), we created a His₆-tag E1ec to enable thorough purification needed for the NMR studies and demonstrated that this construct could function analogously to the parental E1ec (see supplemental Experimental Procedures).

Sequence-specific Resonance Assignments in the Flexible Regions of E1ec

The lack of interpretable electron density for residues 1–55, 401–413, and 541–557 in the high resolution x-ray structure of E1ec (7) implied that these regions are highly disordered, at the same time suggesting that perhaps they could be sufficiently mobile to give rise to narrow NMR signals. Guided by this hypothesis, the ¹H,¹⁵N TROSY-HSQC spectrum of a uniformly ¹⁵N-labeled His₆-tag E1ec was recorded and processed (Fig. 2a). In this spectrum a tryptophan NH resonance was observed with an ¹H chemical shift of ∼10.12 ppm and ¹⁵N chemical shift of ∼130.55 ppm, assignable to the sole tryptophan Trp-16 present in the three missing regions (7) (Table 2).

FIGURE 2.

¹H,¹⁵N TROSY-HSQC spectra of His₆-tag E1ec. a, uniformly ¹⁵N-labeled type His₆-tag E1ec; b, glycine residues only ¹⁵N-labeled; c, spectra a and b superimposed showing perfect overlap.

TABLE 2.

¹H^N and ¹⁵N chemical shifts of assigned residues of the flexible regions of His₆-tag E1ec compared with the reference chemical shift of random coil

Residue	Random coli chemical shifts at pH 5a, 15N	Residue	Observed chemical shifts at pH 7
			1HN	15N
	ppm		ppm	ppm
Glycine	109.30	Gly-28	8.454	110.9
		Gly-47	8.294	110.3
		Gly-402	8.389	110.4
		Gly-542	8.208	109.9
		Gly-28	8.454	110.9
Glutamine	111.84	Gln-18	7.454	113.7
		Gln-33	7.596	114.1
		Gln-38	7.692	114.8
Tryptophan	129.38	Trp-16	10.12	130.6

^a Random coil chemical shifts reported for glycine in Ref. 40 for Gln and Trp side chains from the Biological Magnetic Resonance Data Bank at the University of Wisconsin-Madison.

A Glycine Auxotroph Assisted in the Assignment of Resonances to Glycines

There were well resolved resonances observed at ¹H 8–8.6 ppm, ¹⁵N 108–111 ppm, the region most often populated by glycines. We confirmed this hypothesis for E1ec by the use of a glycine auxotroph, enabling us to label with ¹⁵N only the glycine backbone nitrogen atoms but no other amino acids, as seen in Fig. 2b. This spectrum when superimposed on the spectrum of the uniformly ¹⁵N-labeled His₆-tag E1ec (Fig. 2c) enabled us to identify glycine residues in the mobile regions.

Assignment of the NH Pertinent to Gly-47

There are five adjacent Gly-Gly pairs in the E1ec sequence (46–47, 104–105, 133–134, 362–363, and 625–626), of which only Gly-46–Gly-47 is located in the flexible regions according to crystallographic B factors (for the other Gly-Gly pairs there is well defined electron density). To identify the glycine resonance corresponding to position 47, a ¹³C,¹H,¹⁵N triple resonance three-dimensional HNCO NMR experiment was employed. All glycines in His₆-tag E1ec were labeled by having an equimolar concentration of ¹⁵N-labeled and 1-¹³C-labeled glycine in the medium (see “Experimental Procedures”), and the NH plane was processed from the three-dimensional data. The ¹H,¹⁵N TROSY-HSQC spectrum of His₆-tag E1ec was compared with the non-TROSY spectrum of the same protein (Fig. 3a). Because there is only a single pair of adjacent glycines in the mobile regions, there should be only one ¹H,¹⁵N (of Gly-47) directly bonded to a ¹³CO (of Gly-46), giving rise to a unique resonance in the three-dimensional HNCO NMR experiment, as is indeed the case in Fig. 3b. This resonance was superimposed on the ¹H,¹⁵N TROSY-HSQC spectrum of the uniformly ¹⁵N-labeled His₆-tag E1ec, and the NH corresponding to residue Gly-47 was assigned (Fig. 3c expanded in Fig. 3d, Table 2).

FIGURE 3.

Resonance assignment for Gly-47. a, shown is a ¹H,¹⁵N TROSY-HSQC spectrum of uniformly ¹⁵N-labeled His₆-tag E1ec aligned to non-TROSY spectrum. b, shown is a ¹³C,¹H,¹⁵N triple resonance three-dimensional HNCO spectrum of ¹³C,¹⁵N-labeled glycine in His₆-tag E1ec; note that only Gly-46–Gly-47 are adjacent glycines in the mobile regions; c and d, shown are superimposed spectra with a and b.

Assignment of Sole Glycines in the Mobile Active Center Loops

Fortuitously, the inner loop (401–413) and the outer loop (541–557) each has only a single glycine, and these were converted to alanine to help assign the resonance to residues Gly-402 (inner loop) and Gly-542 (outer loop) by the difference of the spectra of E1ec and the G402A (activity in complex = 54.2%) or G542A (activity in complex = 20.1%) variants (Fig. 4, a and b, and Table 2).

FIGURE 4.

Resonance assignment for Gly-402 (inner loop) and Gly-542 (outer loop) of His₆-tag E1ec in ¹H,¹⁵N TROSY-HSQC spectra. a and b, superimposed spectra of His₆-tag E1ec and G402A variant. c and d, superimposed spectra of His₆-tag E1ec and G542A variant.

Sequence-specific Assignment of Glutamine Side-chain Resonances in the N-terminal Region of His₆-tag E1ec

Such side chains are characterized by a pair of resonances with the same ¹⁵N but different ¹H chemical shifts, due to slow rotation around the amide bond on the NMR time scale. The chemical shift regions are 6.6–7.7 ppm for ¹H and 115–118 ppm for ¹⁵N in the spectrum of His₆-tag E1ec. Site-directed mutagenesis was carried out,substituting glutamine with histidine at positions 18, 33, and 38, and uniformly ¹⁵N-labeled enzymes were prepared for each E1ec variant. The activities of the PDHc-ec complex with these substitutions were (in parentheses, % compared with His₆-tag E1ec) Q18H (100%), Q33H (55%), and Q38H (66%), indicating no significant effect on the activity (ThDP enzymes accelerate the rates by 10^12–13 or more (35); such reductions in activity as reported here for all of the variants make only minor or insignificant changes in the structure and will not complicate interpretation of the resonance assignments). The dispersion of the resonances was essentially superimposable on those for E1ec, confirming no change in secondary or tertiary structure with the substitutions. Next, a ¹H,¹⁵N TROSY-HSQC spectrum was acquired for each variant and compared with the E1ec spectrum (Fig. 5, a–f). We could assign each resonance to a specific glutamine in the sequence from such comparisons (Table 2); with each substitution a set of two peaks with the same ¹⁵N chemical shift but different ¹H chemical shift was missing.

FIGURE 5.

Resonance assignment of three sets of Gln side chains in ¹H,¹⁵N TROSY-HSQC spectra. a and b, shown are superimposed spectra of His₆-tag E1ec (black) and Q18H variant (red). Also shown are superimposed spectra of His₆-tag E1ec (black) and Q33H variant (red) (c and d) and superimposed spectra of His₆-tag E1ec (black) and Q38H variant (red) (e and f).

A Synthetic Peptide with Sequence Corresponding to the N-terminal 1–35 Residues Enabled Assignment of Gly-28 and Confirmed That of Trp-16

A peptide corresponding to the N-terminal 1–35 amino acids of E1ec was synthesized (SERFPNDVDPIETRDWLQAIESVIREEGVERAQYL) (N-term peptide; mass, 4161.4 Da; theory, 4161.6 Da). The CD spectrum of N-term peptide exhibits the negative band at 200 nm at pH 7.0, characteristic of unfolded peptides. Two negative bands were observed at 222 and 206 nm at pH 5.0, characteristic of α-helix conformation (Fig. 6). To determine whether the synthesized peptide forms a functional binding domain, the E2ec-E3ec subcomplex was incubated with 0.050–1 mm concentrations of N-term peptide. On reconstitution of E2ec-E3ec subcomplex with E1ec, ∼40% of activity was lost, indicating that the N-term peptide competes with E1ec for binding to E2ec. The limited inhibition suggests that N-term peptide is insufficient in length to fully mimic the N-terminal region of E1ec.

We next carried out a two-dimensional ¹H,¹⁵N HSQC experiment (non-TROSY this time to provide better sensitivity) of the unlabeled N-term peptide (7.6 mm) (supplemental Fig. S2). The peptide gave an excellent two-dimensional HSQC spectrum notwithstanding the low ¹⁵N natural abundance (0.36%). The spectrum was aligned to the TROSY spectrum of His₆-tag E1ec for comparison, leading to several conclusions; 1) the narrow dispersion of the resonances suggests that both the synthetic peptide and the N-terminal region of E1ec (and dynamic regions in general) are in the random conformation; 2, two residues are immediately assignable, Trp-16 and Gly-28 NH (Table 2), also providing ¹H and ¹⁵N chemical shifts for the random coil form of the Trp-16 side chain.

Identification of Residues/Regions of E1ec Interacting with E2ec

N-Terminal E1ec Residues That Interact with an N-terminal E2ec Construct According to NMR

The E2ec component is composed of five domains. Starting from the N-terminal end, there are three lipoyl domains (LD) followed by the PSBD and the catalytic or core domain (Scheme 2). Although the lipoyl domain undergoes reductive acetylation by the E1ec and pyruvate, the PSBD has been long believed to be important as an anchor for both the E1ec and E3ec components. We used here a single lipoyl domain construct of E2ec (1-lip E2ec, Ref. 36) that can indeed be reconstituted to active PDHc and is only modestly less active than the wild-type three-lipoyl domain enzyme. One of the continuing challenges to those working with these important and ever-fascinating complexes is the high degree of oligomerization of the E2 components (24-mer for E2ec). This oligomerization renders the full-length E2ec difficult to handle. To circumvent the oligomerization issue, starting with the 1-lip E2ec, we created a construct consisting of the lipoyl domain and the PSBD for a total length of 190 amino acids from the N-terminal amino acid (E2ec^1–190, see the supplemental Experimental Procedures). Such a construct is usually denoted as a “didomain.” To gain insight to the residues of E1ec reacting with the didomain, we inspected the two-dimensional HSQC spectrum of E1ec on the addition of E2ec^1–190. The premise of the experiment is that, should the N-terminal region become organized (less mobile or dynamic) on complexation with the E2ec^1–190, the resonances pertinent to the N-terminal region will lose much (or all) of their intensity. This expectation is indeed met in Fig. 7; 1) the side-chain NH resonance for Trp-16, 2) the NH resonance for Gly-47, and 3) the pairs of resonances for the side chains of glutamines 18, 33, and 38 are all reduced/absent in the complex (red) compared with the E1ec spectrum per se. We believe that these residues of E1ec indeed interact with the E2ec construct and their positions (16, 18, 33, 38, and 47) give us confidence that the entire N-terminal region does participate in the complex in a dramatic fashion. It is also important to point out that there are other mobile E1ec resonances unperturbed by the complexation with E2ec^1–190, presumably located at other flexible regions of the E1ec (see below), but also providing excellent control for the interpretation that the observations do pertain selectively to the N-terminal region.

SCHEME 2.

Domain architecture of E2ec.

FIGURE 7.

NMR identification of residues of E1ec interacting with the E2ec^1–190. a, His₆-tag E1ec in green is superimposed on the spectrum (red) of His₆-tag E1ec in the presence of E2ec^1–190. b, shown is an expanded view of the glycine region. c, shown is an expanded view of the glutamine region.

N-Terminal E1ec Residues That Interact with PSBD

A two-dimensional HSQC spectrum was recorded for His₆-tag E1ec in the presence of a 2-fold molar excess of a synthetic peptide with a sequence corresponding to the E2ec PSBD: H₂N-ATPLIRRLAREFGVNLAKVKGTGRKGRILREDVQAYVKEAI-OH (CHI Scientific, Maynard, MA). The spectrum (Fig. 8) appeared very similar to that obtained with E2ec^1–190, indicating that the PSBD of the didomain is mostly responsible for E1ec-E2ec recognition.

FIGURE 8.

Two-dimensional ¹H,¹⁵N TROSY-HSQC spectrum of His₆-tag E1ec with synthetic peptide corresponding to E2 PSBD. a, His₆-tag E1ec-synthetic PSBD is shown in blue; 65 resonances remain of 109. b, spectrum in a is superimposed with His₆-tag E1ec-E2^1–190 subcomplex in red.

N-Terminal E1ec Residues That Interact with Independently Expressed Lipoyl Domain

When the experiment outlined in the previous paragraph was repeated in the presence of a 2-fold molar excess of independently expressed lipoyl domain over His₆-tag E1ec, 106 resonances were detected of 109 total detected for His₆-tag E1ec (see supplemental Fig. S3, a and b). None of the resonances assigned in the N-terminal region and found to interact with either didomain or PSBD was affected by the lipoyl domain.

Interaction of the Flexible Active Center Loops of E1ec with the Stable Predecarboxylation Intermediate PLThDP

Having assigned resonances that could act as reporters of each of the two active center loops (Gly-402 and Gly-542), we are in an excellent position to determine any changes in mobility on the addition of PLThDP, a stable predecarboxylation intermediate that resembles LThDP in Scheme 1 (k₁, k₂). We have very strong evidence from both x-ray and dynamic measurements (9, 14, 16, 43) that in the presence of PLThDP these two loops become organized and are now observed in the x-ray structure. In fact, the addition of PLThDP to His₆-tag E1ec, selectively broadens the NH resonances corresponding to Gly-402 and Gly-542 but not those resonances assigned to the N-terminal residues (Fig. 9). This experiment is not only important to support all of the previous data on this important dynamic property but also serves as an outstanding control experiment, indicating that the observations are indeed selective to the particular interaction loci; the PLThDP interacts with the active center loops but not with the N-terminal region.

FIGURE 9.

Two-dimensional ¹H,¹⁵N TROSY-HSQC spectrum of His₆-tag E1ec with the reaction intermediate analog PLThDP. a, His₆-tag E1ec in the presence of PLThDP. N-terminal residues are still present on the spectrum. b, His₆-tag E1ec in the presence of E2ec^1–190. Resonances for Gly-402 and Gly-542 from two dynamic loops are still present on the spectrum, confirming that these two loops are not involved in E1-E2 assembly.

Quantification of Residues Identified in the Flexible E1ec Regions

Given that we have only made sequence specific resonance assignment to 8 of the 886 amino acids present in E1ec, it is important to address the issue of specificity of the observations. First of all, we have several reporters spanning residues 1–47 and 1 in each of the two active center loops. A summary of the resonances observed under different conditions is given in Table 3. In the region of chemical shifts presented in the two-dimensional ¹H,¹⁵N HSQC spectra, we expect to see resonances pertinent to (a) main-chain NHs, one for each peptide bond with the exception of those to proline, (b) two side-chain NHs for each Asn and Gln and one for each Trp. The number of resonances accounting for these is 64 for the N-terminal 1–55 residues, 19 for the inner (401–413), and 26 for the outer (541–557) active center loops for a total of 109. Using the program NMRviewJ peak counting with the same noise filtration for all spectra, we estimate the presence of 103 resonances for E1ec in the presence of saturating ThDP and 101 resonances in the absence of ThDP (i.e. apoenzyme). Apparently, in terms of number of resonances observed, the numbers anticipated in the three flexible regions account for all of the observations within experimental error; coupled with the sequence specific assignments, we have strong evidence that these indeed are the regions observed in the spectra (Table 3). For the complexes, we conclude the following. (i) The addition of the E2^1–190 didomain results in the elimination of 44 resonances (of the 103 total observed and of the 64 calculated for the N-terminal region); apparently, not all residues in the N-terminal region are highly ordered in the complex. (ii) The addition of the synthetic peptide corresponding to PSBD results in the elimination of 38 residues, the same ones as missing on addition of E2^1–190 in (i). (iii) The addition of independently expressed lipoyl domain results in no assigned resonance being eliminated within experimental error. These numbers strongly support the conclusion that the PSBD is the important moiety of the didomain that binds to the E1ec, specifically to its N-terminal regions, whereas the lipoyl domain binds very weakly at best. (iv) The addition of PLThDP results in the elimination of 33 resonances (of the 101 total observed and of the 45 estimated for the inner and outer dynamic active center loops), and the specificity is confirmed by the elimination of the resonances assigned to Gly-402 and Gly-542, reporting on the inner and outer loops, respectively, but not any of those assigned to specific residues in the N-terminal region. These results confirm our series of studies on these loops and, equally importantly, point to the specificity of the NMR observations.

TABLE 3.

Experimentally estimated and theoretical number of resonances on E1ec in the three flexible regions in the ¹⁵N-¹H HSQC spectra

	N-terminal region	Inner loop	Outer loop
No. of residuesa	55 (1–55)	13 (401–413)	17 (541–557)
No. of Gln	3	1	4
No. of Asn	2	2	1
No. of Pro	2	0	1
No. of Trp	1	0	0
Theoretical no. of resonances	64	19	26
No. of resonances estimated (experimental/theoretical)	103/109b
	101/109c
No. of resonances undetected on complexation	44d; 38e; 33f

^a Each backbone NH contributes one resonance. The side chains of glutamine and asparagine give rise to two resonances on two-dimensional HSQC spectra. For each proline present, the total number of resonances was decreased by one. The Trp gives rise to the indole NH resonance.

^b Spectrum of His₆-tag E1ec.

^c Spectrum of His₆-tag E1ec without ThDP.

^d Spectrum of His₆-tag E1ec in the presence of E2ec^1–190.

^e Spectrum of His₆-tag E1ec in the presence of synthetic peptide corresponding to PSBD.

^f Spectrum of His₆-tag E1ec in the presence of PLThDP.

Summary and Conclusions

These NMR results provide the first structural support to the notion that the entire N-terminal region (residues 1–55) of the E1ec is responsible for binding the E2ec component. The N-terminal region of E1ec, although not seen in the x-ray structures, gives well resolved resonances and permitted sequence specific assignment for six resonances (Trp-16, Gln-18, Gly-28, Gln-33, Gln-38, and Gly-47). These assignments complement the biochemical findings on this region; substitution of several residues in this region dramatically reduced the activity of the reconstituted complex and affected intercomponent assembly (interaction of E1ec with E2ec). It is also clear, however, that not all residues in this N-terminal region have a role in assembly. For example, the chemical shift of the resonance for Gly-28 does not change under various conditions.

The narrow dispersion of the NMR resonances observed both in the N-terminal 1–35 synthetic peptide and in the N-terminal region of E1ec suggest that both are in the random conformation, i.e. they are disordered in the absence of E1-E2 assembly.

Interaction of E1ec with either E2ec^1–190 or PSBD, but not with the lipoyl domain, leads to reduction/absence of the side-chain NH resonance for Trp-16, the NH resonance for Gly-47, and of three pairs of resonances for the side chains of glutamines 18, 33, and 38 and indicate that residues from the entire N-terminal region participate in the complex and that the PSBD, but not the lipoyl domain, is essential for E1-E2 interaction.

Concerning the effect of the N-terminal substitutions on individual reaction steps, we conclude the following. (a) None of the substitutions affected the rate of the E1-specific reaction, i.e. the reaction was unaffected through pyruvate decarboxylation. (b) Similarly, we showed that these substitutions do not affect formation of the first covalent intermediate LThDP. (c) Studies on the reductive acetylation of E2ec suggested that binding of the N-terminal region of E1ec to the PSBD of E2ec is essential before reductive acetylation could take place, complementing the NMR results.

The N-terminal region of E1ec is predicted to have a propensity for forming two α-helices (residues 11–26 and 29–45) and a small loop (residues 27 and 28) joining the two α-helices (Fig. 10, a helix-turn-helix motif), similar to the model suggested for the N-terminal region of the E1 component from the related bacterium A. vinelandii (12). According to the structural and functional analysis here presented, if this is the true secondary structure of the N-terminal region of the E1ec in the complex, both helices participate in the interaction with E2ec. Should assembly indeed require that the N-terminal region assume a secondary structure modeled in Fig. 10 and predicted, one could speculate that the residues “silent” to substitution do not participate in assembly.

FIGURE 10.

A, the predicted secondary structure of the N-terminal 1–55 region (program from Protein Homology/analogY Recognition Engine) is shown. B, the Cα atoms were generated from ExPASy Proteomics server. Main-chain and side-chain atoms were generated with the program COOT (41); the final figure was with RIBBONS (42).

The addition of PLThDP to His₆-tag E1ec selectively broadens the NH resonances corresponding to Gly-402 and Gly-542, consistent with its effect on the active center loops, but does not affect those resonances assigned to the N-terminal residues. This indicates that the NMR observations are selective to the particular interaction locus.

The results here reported have broader significance/consequence to other members of this important class of multienzyme complexes. (i) The N-terminal region of the E1 component of the E. coli 2-oxoglutarate dehydrogenase complex was also suggested to interact with its E2 component (37). (ii) For this class of bacterial 2-oxoacid dehydrogenase complexes, the PSBD is the most important domain for recognition of E1, rather than the lipoyl domain, as was suggested by others (38, 39). To our knowledge our results here presented offer the first experimental evidence for this conclusion. (ii) The results also clearly show that with such large proteins, the NMR and x-ray results are indeed complementary; what is not seen in the x-ray may be visible in the NMR spectrum and vice versa. In particular, the N-terminal 1–55 residues and the two active center loops not seen in the x-ray structure have now been shown to provide resolvable resonances in the NMR spectrum of E1ec.