FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2023454927> ?p ?o ?g. }

• W2023454927 endingPage "21246" @default.
• W2023454927 startingPage "21240" @default.
• W2023454927 abstract "The derivative of vitamin B1, thiamin pyrophosphate, is a cofactor of enzymes performing catalysis in pathways of energy production. In α2β2-heterotetrameric human pyruvate dehydrogenase, this cofactor is used to cleave the Cα-C(=O) bond of pyruvate followed by reductive acetyl transfer to lipoyl-dihydrolipoamide acetyltransferase. The dynamic nonequivalence of two, otherwise chemically equivalent, catalytic sites has not yet been understood. To understand the mechanism of action of this enzyme, we determined the crystal structure of the holo-form of human pyruvate dehydrogenase at 1.95-Å resolution. We propose a model for the flip-flop action of this enzyme through a concerted ∼2-Å shuttle-like motion of its heterodimers. Similarity of thiamin pyrophosphate binding in human pyruvate dehydrogenase with functionally related enzymes suggests that this newly defined shuttle-like motion of domains is common to the family of thiamin pyrophosphate-dependent enzymes. The derivative of vitamin B1, thiamin pyrophosphate, is a cofactor of enzymes performing catalysis in pathways of energy production. In α2β2-heterotetrameric human pyruvate dehydrogenase, this cofactor is used to cleave the Cα-C(=O) bond of pyruvate followed by reductive acetyl transfer to lipoyl-dihydrolipoamide acetyltransferase. The dynamic nonequivalence of two, otherwise chemically equivalent, catalytic sites has not yet been understood. To understand the mechanism of action of this enzyme, we determined the crystal structure of the holo-form of human pyruvate dehydrogenase at 1.95-Å resolution. We propose a model for the flip-flop action of this enzyme through a concerted ∼2-Å shuttle-like motion of its heterodimers. Similarity of thiamin pyrophosphate binding in human pyruvate dehydrogenase with functionally related enzymes suggests that this newly defined shuttle-like motion of domains is common to the family of thiamin pyrophosphate-dependent enzymes. The thiamin pyrophosphate (TPP) 1The abbreviations used are: TPP, thiamin pyrophosphate; PDC, pyruvate dehydrogenase complex; E1, pyruvate dehydrogenase; E1p, human pyruvate dehydrogenase; E2, dihydrolipoamide acetyltransferase; LTPP, lactyl-TPP; HETPP, 4′-imino-2-(1-hydroxyethyl)TPP; TATPP, transition state adduct of HETPP; PP domain, domain involved in binding of pyrophosphate fragment of TPP; PYR domain, domain involved in binding of the aminopyrimidine ring of TPP. 1The abbreviations used are: TPP, thiamin pyrophosphate; PDC, pyruvate dehydrogenase complex; E1, pyruvate dehydrogenase; E1p, human pyruvate dehydrogenase; E2, dihydrolipoamide acetyltransferase; LTPP, lactyl-TPP; HETPP, 4′-imino-2-(1-hydroxyethyl)TPP; TATPP, transition state adduct of HETPP; PP domain, domain involved in binding of pyrophosphate fragment of TPP; PYR domain, domain involved in binding of the aminopyrimidine ring of TPP.-dependent enzymes perform a wide range of catalytic functions in the pathways of energy production, including decarboxylation of α-keto acids followed by transketolation. The enzymes that have been structurally characterized so far, 2-oxoisovalerate dehydrogenase from Pseudomonas putida (1AEvarsson A. Seger K. Turley S. Sokatch J.R. Hol W.G.L. Nature Struct. Biol. 1999; 6: 785-792Crossref PubMed Scopus (119) Google Scholar), human branched-chain α-ketoacid dehydrogenase (2AEvarsson A. Chuang J.L. Wynn R.M. Turley S. Chuang D.T. Hol W.G.J. Structure. 2000; 8: 277-291Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar), bacterial pyruvate dehydrogenase (3Arjunan P. Nemeria N. Brunskill A. Chandrasekhar K. Sax M. Yan Y. Jordan F. Guest J.R. Furey W. Biochemistry. 2002; 41: 5213-5221Crossref PubMed Scopus (119) Google Scholar), transketolase (4Fiedler E. Thorell S. Sandalova T. Golbik R. Konig S. Schneider G. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 591-595Crossref PubMed Scopus (104) Google Scholar), pyruvate decarboxylase (5Dyda F. Furey W. Swaminatythan S. Sax M. Farrenkopf B. Jordan F. Biochemistry. 1993; 32: 6165-6170Crossref PubMed Scopus (224) Google Scholar), benzoylformate decarboxylase (6Hasson M.S. Muscate A. McLeish M.J. Polovnikova L.S. Gerlt J.A. Kenyon G.L. Petsko G.A. Ringe D. Biochemistry. 1998; 37: 9918-9930Crossref PubMed Scopus (163) Google Scholar), acetohydroxyacid synthase (7Pang S.S. Duggleby R.G. Guddat L.W. J. Mol. Biol. 2002; 317: 249-262Crossref PubMed Scopus (186) Google Scholar), pyruvate oxidase (8Muller Y.A. Schumacher G. Rudolph R. Schulz G.E. J. Mol. Biol. 1994; 237: 315-335Crossref PubMed Scopus (88) Google Scholar), and pyruvate:ferredoxin oxidoreductase (9Chabriere E. Vernede X. Guigliarelli B. Charon M.-H. Hatchikian E.C. Fontecilla-Camps J.C. Science. 2001; 294: 2559-2563Crossref PubMed Scopus (116) Google Scholar), have shown a common mechanism of TPP activation by (i) forming the ionic N-H···O- hydrogen bonding between the N1′ atom of the aminopyrimidine ring of the coenzyme and an intrinsic γ-carboxylate group of glutamate and (ii) imposing an “active” V-conformation that brings the N4′ atom of the aminopyrimidine to the distance required for the intramolecular C-H···N hydrogen-bonding with the thiazolium C2 atom (Fig. 1). Within these two hydrogen bonds that rapidly exchange protons, protonation of the N1′ atom of the aminopyrimidine system is strictly connected with the deprotonation of the 4′-amino group in that system and eventually abstraction of the proton from C2 and formation of the reactive 4′-amino-C2-carbanion (Fig. 1a) (10Kern D. Kern G. Neef H. Tittmann K. Killenberg-Jabs M. Wikner C. Schneider G. Hubner G. Science. 1997; 275: 67-70Crossref PubMed Scopus (236) Google Scholar). This reactive C2 atom of TPP is the nucleophile that attacks the carbonyl carbon of different substrates used in the family of TPP-dependent enzymes. Within pyruvate dehydrogenase (E1), the first catalytic component enzyme of pyruvate dehydrogenase complex (PDC), this substrate is pyruvate (S1). The cleavage of the central Cα-C(=O) bond of this substrate proceeds from induction of the intermediate, 4′-imino-2-(2-hydroxypropionyl)thiamin pyrophosphate, i.e. lactyl-TPP (LTPP) (Fig. 1b), followed by conversion to 4′-imino-2-(1-hydroxyethyl) thiamin pyrophosphate (HETPP) with release of carbon dioxide (P1) (Fig. 1c). The fate of this active C2-α-carbanion/enamine HETPP differs among various TPP-dependent enzymes depending on the nature of the second substrate (S2). In E1, catalysis proceeds through reductive acetyl transfer between HETPP and lipoic acid covalently bound to the second catalytic component of PDC, i.e. dihydrolipoamide acetyltransferase (E2) with formation of a transition state adduct (TATPP) (Fig. 1d) leading to the release of the second product S-acetyldihydrolipoamide-E2 (P2) (11Pan K. Jordan F. Biochemistry. 1998; 37: 1357-1364Crossref PubMed Scopus (29) Google Scholar). The deprotonation of the hydroxyl group at the C2-α carbon atom concomitant with the acetyl transfer to E2 brings the N4′-proton back to the aminopyrimidine moiety, and TPP returns to its active form to start the next cycle of catalytic reactions as presented in Fig. 1a. The E1 component of human PDC, designated E1p, is a TPP-dependent α2β2 heterotetramer with a molecular mass of 154 kDa, which has two catalytic sites, each providing TPP and magnesium ion as cofactors and each formed on the interface between the α and β subunits, where those interfaces are referred to as domains. The chemical equivalence of these two catalytic sites is distinct from the dynamic equivalence of these two sites. The dynamic nonequivalence of these active sites during catalytic action was first detected by spectral and kinetic studies of pigeon E1 (12Khailova L.S. Korotchkina L.G. Biochem. Int. 1985; 11: 509-516PubMed Google Scholar, 13Khailova L.S. Korotchkina L.G. Severin S.E. Biswanger H. Ullrich J. Biochemistry and Physiology of TDP Enzymes. VCH Weinheim, Blaubeuren, Germany1990: 251-265Google Scholar). The flip-flop mechanism was suggested, according to which two active sites affect each other and in which different steps of the catalytic reaction are performed in each of the sites at any given moment. The interactions between the active sites during catalysis and their dynamic nonequivalence are known from biochemical studies for several other TPP-dependent enzymes as well, including transketolase (14Kovina M.V. Kochetov G.A. FEBS Lett. 1998; 440: 81-87Crossref PubMed Scopus (26) Google Scholar), pyruvate decarboxylase (15Sergienko S.A. Jordan F. Biochemistry. 2002; 41: 3952-3967Crossref PubMed Scopus (33) Google Scholar), benzoylformate decarboxylase (16Sergienko E.A. Wang J. Polovnikova L. Hasson M.S. McLeish M.J. Kenyon G.L. Jordan F. Biochemistry. 2000; 39: 13862-13869Crossref PubMed Scopus (43) Google Scholar), and Escherichia coli PDC-E1 (17Yi J. Nemeria N. McNally A. Jordan F. J. Biol. Chem. 1996; 271: 33192-33200Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar). To understand the flip-flop action of E1p, we have determined the crystal structure in the holoenzyme form at the level of 1.95 Å, and based on that structure, we propose a model for enzyme action common to the family of TPP-dependent enzymes. Protein Expression and Purification—A prokaryotic coexpression vector of pQE-9-E1(E1α/HisE1β) was constructed encoding the 361-amino-acid residue sequence of human E1α with an additional 4 amino acid residues added at the N terminus, and separately, the 329-amino-acid residue sequence of E1β with an additional 12-amino-acid residue sequence at the N terminus (18Korotchkina L.G. Tucker M.M. Thekkumkara T.J. Madhusudhan K.T. Pons G. Kim H. Patel M.S. Protein Expression Purif. 1995; 6: 79-90Crossref PubMed Scopus (37) Google Scholar). Selenomethionine-E1p derivative was expressed in methionine auxotroph DL41 cells transformed with pDM1.1 and pQE-9-E1 (E1α/HisE1β) vector and grown at 25 °C in medium containing 30 mg/liter seleno-l-methionine (19Doublie S. Methods Enzymol. 1997; 276: 523-530Crossref PubMed Scopus (788) Google Scholar). Cell lysates were eluted through a nickel-nitrilotriacetic acid column and further purified by DEAE Sephadex A-25 and Superdex 200-HP chromatography with inclusion of 10 mm dithiothreitol at all times. In the final step, protein fractions were dialyzed against 50 mm potassium phosphate buffer, pH 7.0, 10 mm dithiothreitol, 0.2 mm TPP, and 0.2 mm MgCl2. Selenomethionine-E1p derivative was determined to be as active as the α2β2 wild-type E1p. The final preparation was ∼95% pure as judged by densitometry of SDS-polyacrylamide gels. Crystallization and Data Collection—Crystals of the holo-form of E1p were grown from modified crystallization conditions described in Ref. 20Ciszak E. Korotchkina L.G. Hong Y.-S. Joachimiak A. Patel M.S. Acta Crystallogr. Sect. D Biol. Crystallogr. 2001; 57: 465-468Crossref PubMed Scopus (3) Google Scholar. Crystallization droplets consisted of 2 μl of protein solution at 17 mg/ml mixed with 2 μl of reservoir solution of 200 mm NaSCN, 12% polyethylene glycol 3350, 10% 1,2-propanediol (v/v), and 50 mm potassium phosphate buffer, pH 8.0. For data collection, crystals were flash-frozen directly in liquid nitrogen and transferred to the x-ray beam for crystallography. A set of data at three wavelengths (λ1 = 0.97978, λ2 = 0.97962, and at high energy, remote λ3 = 0.96439 Å) for multiwavelength anomalous dispersion method was collected on a single crystal at the BM-19 Advanced Photon Source in Argonne, IL. Diffraction data were integrated and reduced with the HKL2000 program (21Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38253) Google Scholar) in the space group P 212121, a = 64.33, b = 126.89, c = 190.64 Å with E1p tetramer in the asymmetric unit. Structure Determination—The search for the structure of E1p included (i) multiwavelength anomalous dispersion method and (ii) molecular replacement based on the search model consisting of the reconstructed tetramer of the branched-chain α-ketoacid dehydrogenase (1DTW) with all amino acids replaced for the human E1p sequence. In the first approach, the search for initial phasing with the use of Se-Met data resulted in 35% of the total number of selenium atoms in the structure. In the second approach, the molecular replacement with MOLREP (22Vagin A. Tepyakov A. J. Appl. Crystallogr. 1997; 30: 1022-1025Crossref Scopus (4098) Google Scholar) gave a solution of both rotation and translation functions with the low correlation value of 0.20 and initial R factor of 0.57. The calculated electron density map using phases from the model was of poor quality with ∼30% of the main chain residues that were untraceable. After combining electron density maps derived from the Se-Met phasing with those from the molecular replacement, the core of this initial model was subjected to rigid body refinement. This first step of refinement was followed by slow cooling from 5000 K using the maximum likelihood function target and the restraints on the presence of the 2-fold non-crystallographic symmetry in the model. After improving the electron density maps, fragments 1–29, 57–60, 127–129, 204–219, 245–248, and 272–345 from both α subunits and 30–35, 109–113, 177–193, 226–231, 280–285, and 304–310 from both β subunits were built de novo into the electron density maps. Subsequent rounds of model building and refinement were performed in XtalView (23McRee D.E. J. Struct. Biol. 1999; 125: 156-165Crossref PubMed Scopus (2016) Google Scholar) and CNS (24Brunger A.T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16919) Google Scholar). The E1p structure included coordinates for all 1380 amino acids of the mature protein and a total of four additional sequence residues located at the N termini of E1α and E1β subunits. The current model also includes 2 cis amino acids (Gly134 within both α subunits and Pro243 within both β subunits), 2 TPP molecules, 2 Mg2+ and two K+ ions, and 742 water molecules. The model was analyzed with PROCHECK (25Laskowski R.A. MacArthur M.W. Moss D.S. Thornton J.M. J. Appl. Crystallogr. 1993; 2: 283Crossref Google Scholar) showing 99.5% of the residues in the allowable regions of the Ramachandran plot except for Asp218 and Met200 from the α subunits and Arg239 from the β subunits. In each case, those residues have well defined electron density. The final statistics of data collection and refinement are shown in Table I.Table ISummary of data collection and refinement statistics Values within parentheses refer to the reflections and refinement from the 1.95–45.2-Å resolution range.Se-Met λ1Se-Met λ2Se-Met λ3Resolution range (Å)2.40-45.22.10-45.22.50-45.2Completeness (%)94.695.896.6ReflectionsaFriedel pairs not merged for anomalous Se-Met data111,685175,40898,424Redundancies2.382.492.36RsymbRsym=Σ|I−〈I〉|/ΣI (%)13.616.513.7I/σ(I)6.89.96.7Anomalous phasing powercPhasing power=〈F/lack of closure〉 for all phased reflections to 3.5 Å, where phasing power values where determined by CNS (24)2.412.051.60Refinement statisticsResolution range (Å)2.1-45.2 (1.95-45.2)Completeness (%)95.8 (80.3)RcrystdRcryst=Σ|Fo−Fc|/ΣFo0.202 (0.213)RfreeeRfree=Σ|Fo−Fc|/ΣFo, where Fo values are test set amplitudes (10.0%) not used in refinement0.244 (0.251)No. of protein atoms10698No. of water molecules742Overall B value (Å2)18.7 (18.1)r.m.s.d.fRoot mean square deviationBond lengths (Å)0.006 (0.006)Bond angles (°)1.2 (1.2)Dihedral angles (°)22.3 (22.3)Improper angles (°)0.79 (0.79)a Friedel pairs not merged for anomalous Se-Met datab Rsym=Σ|I−〈I〉|/ΣIc Phasing power=〈F/lack of closure〉 for all phased reflections to 3.5 Å, where phasing power values where determined by CNS (24Brunger A.T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16919) Google Scholar)d Rcryst=Σ|Fo−Fc|/ΣFoe Rfree=Σ|Fo−Fc|/ΣFo, where Fo values are test set amplitudes (10.0%) not used in refinementf Root mean square deviation Open table in a new tab E1p Fold—The structure of human E1p determined at 1.95-Å resolution is the first structure reported for mammalian PDC components. As shown in Fig. 2, E1p has four subunits, namely α and α′ (alpha subunits) and β and β′ (beta subunits), that are arranged tetrahedrally with the α and β subunits having their α′ and β′ counterparts related by non-crystallographic 2-fold. Each alpha subunit is composed of a parallel six-stranded β-sheet that is packed against 5 helices out of the total of 10 helices contributing to the content of this subunit. This β-sheet and the five helices are involved in binding of Mg2+ and the pyrophosphate fragment of TPP and are termed the PP and PP′ domain. Other specific secondary structure elements of the alpha subunits include the 28-residue N-terminal fragments and the 36-residue long C-terminal fragments, both being observed in extended conformations, and the 12-amino-acid helical fragment (274–285) found to involve phosphorylation in the alpha subunits (26Korotchkina L.G. Patel M.S. J. Biol. Chem. 2001; 276: 5731-5738Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar). The beta subunits, on the other hand, are composed of two similar sized domains corresponding to the N- and C-terminal halves of the 329-amino-acid polypeptide chain. The first domain of each beta subunit (residues 1–194) has a parallel six-stranded β-sheet surrounded by seven α-helices. This domain of the beta subunits is involved in binding the aminopyrimidine ring of TPP, and it is termed the PYR and PYR′ domain. The second domain of each beta subunit (residues 195–329), termed here the C and C′ domain, has a parallel four-stranded β-sheet with one edge comprised of an antiparallel β-strand braced with four helices, two of them on each side of this central β-sheet. It has been proposed, based on the studies of other heterotetrameric E1s, that E1p is bound to E2 of PDC through the C and C′ domain (1AEvarsson A. Seger K. Turley S. Sokatch J.R. Hol W.G.L. Nature Struct. Biol. 1999; 6: 785-792Crossref PubMed Scopus (119) Google Scholar, 27Perham R.N. Patel M.S. Roche T.E. Harris R.A. Alpha-Keto Acid Dehydrogenase Complexes. Birkhäuser Verlag, Basel, Switzerland1996: 1-15Crossref Google Scholar). Additional structural elements in E1p are two K+ ions, one in each PYR and PYR′ domain of the β subunits, also shown in Fig. 2. Each K+ site is found at the negative end of the helix ranging from residue 150 to 162 (Fig. 3). Two pairs of symmetry-related K+ binding sites were first reported in the homologous structure of human branched-chain α-ketoacid dehydrogenase (2AEvarsson A. Chuang J.L. Wynn R.M. Turley S. Chuang D.T. Hol W.G.J. Structure. 2000; 8: 277-291Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar). One pair of K+ sites was in the place corresponding to the K+ sites found in E1p, and the other pair was placed in the regions of the loops in the alpha subunits (corresponding to loops 133–139 in E1p) that are involved in the cofactor binding and enzyme action. Both pairs of K+ binding sites were reported to have structural roles in stabilizing those particular loops (2AEvarsson A. Chuang J.L. Wynn R.M. Turley S. Chuang D.T. Hol W.G.J. Structure. 2000; 8: 277-291Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar). However, only the K+ ions from the PYR and PYR′ domains have been confirmed in E1p. Rather, stabilization of loops in the alpha subunits in E1p is achieved by engaging Thr87 in a pair of hydrogen bonds with the oxygen and nitrogen atoms of Gly133 and Gly134 located in the cis-peptide bond between these two residues. Properties of the Subunit Interactions—In E1p, the PP and PYR domains of one α,β-heterodimer and the PP′ and PYR′ domains of the symmetry-related α′β′-heterodimer form predominantly hydrophobic contacts. This association of apolar portions of surfaces of the alpha and beta subunits is facilitated through the GΦXXG motif (where the Φ position is often a β-branched chain amino acid) by permitting a pair of helices, 140–155 of the PP domain and 59–72 of the PYR domain, and symmetry-related counterparts, to pack tightly together (Fig. 4, a and b). In the PP-helix, this GΦXXG motif has Pro143 replacing the glycine, Leu144, Gly145, Ala146, and Gly147, whereas in the PYR-helix includes Gly64, Ile65, Ala66, Val67, and Gly68. Both helices form a triad of close main chain contacts between Leu144···Ile65, Gly147···Gly68, and Leu144···Gly68 that range from 3.8 to 4.2 Å. Furthermore, in E1p, this motif extends beyond GΦXXG to Ile148 in the PP domain and Ala69 in the PYR domain, thereby adding a stretch of additional hydrophobic contacts between main chain atoms of Gly147···Ala69, Ile148···Gly68, and Ile148···Ala69 (4.4–4.9 Å). The GΦXXG motif is well known in membrane proteins and was first noticed in non-membrane β-subunit tetramers of pyruvate dehydrogenase from Pyrobaculum aerophilum (28Kleiger G. Perry J. Eisenberg D. Biochemistry. 2001; 40: 14484-14492Crossref PubMed Scopus (16) Google Scholar). Considering all interactions described above, it is suggested that these rigid connections of the PP with the PYR domain, and also their symmetry-related counterparts, are structural elements that promote formation of independent pairs of αβ-heterodimers and that are required in functional α2β2-heterotetramers. In the dynamic protein environment of E1p conducting catalysis, these rigid hydrophobic connections of the PP with the PYR domains are key structural elements required for a concerted movement of each heterodimer in its entirety. On the contrary, flexible connections between these tightly associated PP-PYR domains and their symmetry-related PP′-PYR′ domains are facilitated by two sites containing the cofactors, in such a manner that the aminopyrimidine ring of one cofactor molecule connects the PP with the PYR′ domain, whereas the aminopyrimidine ring of the other cofactor molecule connects the PP′ with the PYR domain. The pyrophosphate termini of each cofactor are nestled exclusively in the PP and PP′ domains by forming hydrogen bonds with the hydroxyl group of Tyr89, the guanidinium group of Arg90, and the nitrogen atoms of Gly168 and Ala169 (Fig. 4c). The binding of pyrophosphate termini is completed by coordination to Mg2+ and by four hydrogen bonds with water molecules W2–W5 serving as proton donors. Completing coordination of Mg2+ is achieved through ligation to the main chain oxygen atom of Tyr198,to the side-chain Oδ1 atoms of Asp167 and Asn196, all being located in the PP domains, and to a single water molecule W1. The latter two amino acids are terminal residues of the 166GDGX26NN196 motif and are known to be present in all TPP-requiring enzymes providing the binding for this cofactor (29Hawkins C.F. Borges A. Perham R.N. FEBS Lett. 1989; 255: 77-82Crossref PubMed Scopus (255) Google Scholar). The aminopyrimidine ring of each cofactor forms three hydrogen bonds: N1′···Oϵ2 of the carboxylate group of Glu59 (2.6 Å to both PYR and PYR′ domains), N4′···O of Gly136 (2.9 and 2.8 Å to PP and PP′ domains, respectively), and N3′···N of Val138 (3.2 and 3.3 Å to PP and PP′ domains, respectively). Val138 also acts as a structural wedge locked in between the thiazolium and aminopyrimidine moieties with contacts to Cβ and Cγ2 atoms varying from 3.6–3.9 Å in both cofactor sites. Uniquely, Glu59 is the first residue of each 59–72 helix of the PYR domains, and Gly136 and Val138 are the residues immediately preceding the beginning of each 140–155 helix of the PP and PP′ domains, all being involved in GΦXXG tight hydrophobic interdomain contacts (Fig. 4b). The carboxylate side chain of Glu59 is the local proton acceptor that enables the N1′ atom to react in the proton translocation within the aminopyrimidine ring, thus enabling deprotonation of the N4′-amino group of the cofactor as shown in Fig. 1. We postulate that in the dynamic protein environment, conducting this proton transfer from the N1′ atom of the cofactor to the protein and vice versa is accomplished by change of distances between the N1′ and Oγ2 atoms of Glu59. A search within the Cambridge Structural Database of small molecule structures determined by neutron and x-ray diffraction methods (30Allen F.H. Davies J.E. Galloy J.J. Kennard O. Macrae C.F. Mitchell E.M. Mitchell G.F. Smith J.M. Watson D.G. J. Chem. Inf. Comput. Sci. 1991; 31: 187-195Crossref Scopus (1378) Google Scholar) for N···O hydrogen bonds similar to that observed between the N1′ atom and Oϵ2 atom of Glu59 revealed that the proton is typically delocalized between the oxygen and nitrogen atoms in hydrogen bonds of 2.6 Å and shorter. On the contrary, longer N···O hydrogen bonds point to a deprotonation of the N1′ atom of the aminopyrimidine moiety, with that proton being transferred to the γ-carboxylate group of Glu59. These findings support our postulated proton translocation mechanism that is understood from the motion of the protein subunits. Considering all interactions of the aminopyrimidine ring and the dynamic protein environment conducting catalysis, it is suggested that the two triads of residues Glu59, Gly136, and Val138, one located in the PP and the PYR′ domains and the other located in the PP′ and PYR domains, can detect catalytic states of the cofactors to govern the protonation of the aminopyrimidine fragment, thus transducing catalytic action of E1p into movement of the pairs of helices connected by the GΦXXG motif. From this description of connectivity between the domains, it is interesting to realize that a channel at each access site to the cofactor can close or open to provide accessibility to either pyruvate (S1) or lipoyl moiety (S2), through a “shuttle-like” movement of tightly connected PP-PYR or PP′-PYR′ heterodomains. Flip-Flop Action of E1p—Although the crystal structure of E1p is static, i.e. derived from a crystal grown under conditions of cofactor-saturation in the absence of substrates, the kinetic model of enzyme action can nonetheless be inferred. The pattern of contacts among the catalytic domains supports structural rigidity and flexibility respectively required for a concerted shuttle-like motion of the heterodimers. We propose that this motion is the basis for the flip-flop action of E1p. Furthermore, the functional nonequivalence of two chemically equivalent active sites reported in spectroscopic studies (12Khailova L.S. Korotchkina L.G. Biochem. Int. 1985; 11: 509-516PubMed Google Scholar, 13Khailova L.S. Korotchkina L.G. Severin S.E. Biswanger H. Ullrich J. Biochemistry and Physiology of TDP Enzymes. VCH Weinheim, Blaubeuren, Germany1990: 251-265Google Scholar) can now be understood as two complete catalytic reaction cycles that are conducted separately in each cofactor-binding site, as shown schematically in Fig. 5. These catalytic reaction cycles are in alternate phases simultaneously involving pyruvate decarboxylation (TPP → LTPP → HETPP) in one site and reductive acetylation of lipoyl-E2 (HETPP → TATPP → TPP) in the second site followed by reductive acetylation of lipoyl-E2 in that one site and pyruvate decarboxylation in the second site. Starting with activated cofactor and pyruvate in the first active site formed, for example, between the PP and PYR′ domains shown in Fig. 5, the formation of LTPP intermediate causes the geometrical changes of the cofactor imposed by the presence of a bulky lactyl group at the C2 atom. Support for this hypothesis comes from the molecular modeling studies of Lobell and Crout (31Lobell M. Crout D.H. J. Am. Chem. Soc. 1996; 118: 1867-1873Crossref Scopus (71) Google Scholar) that detailed that the thiazolium ring rotates to remove steric hindrance and the carboxylate group rotates to maximize the overlap of the orbital containing the pair of electrons of the carbon-carboxylate bond with the π-electrons of the thiazolium ring of LTPP. These rotations induce a movement by ∼2 Å of Val138 wedged in the space between the thiazolium and aminopyrimidine rings, where that movement entrains the connected 140–155 helix with subsequent shift of the PP domain with respect to the PYR′ domain, which remains stationary. That shift of the PP domain causes entrained movement of the PYR domain, which has Glu59 bound to the N1′ atom of the cofactor of the second active site. This interlocked translational motion of the PP and PYR domains of the αβ-heterodimer along the stationary PP′ and PYR′ domains of the α′β′-heterodimer leads then to closing the channel to the first active site, thus isolating the site for the reaction of decarboxylation of pyruvate, which is followed by a cascade of temporally connected events in the second active site. These events are (i) opening the access channel to the second cofactor site so that the second substrate lipoyl-E2 can access HETPP to conduct acetyl transfer to E2 and (ii) extending the distance between the N1′ atom of the second cofactor and Oϵ2 atom of Glu59 of the PYR domain. The lengthening of the latter distance leads to translocation of the proton at the N1′ atom back to the Oϵ2 atom of Glu59 that is correlated with deprotonation of the C2-α-hydroxyalkyl group at TATPP, thereby promoting release of the second product S-acetyldihydrolipoamide-E2 (P2) to recreate activated TPP. It is interesting to note that this proposed shuttle-like motion of the heterodomains is driven by catalytic events occurring in both catalytic sites simultaneously, resulting in “pull-push”-like mechanics at the atomic level. The movement of Val138 that follows the geometric changes of the cofactor (TPP → LTPP) in the first active site acts to pull the domains in a discrete direction, whereas at the same time, the lipoyl-E2 substrate conducting catalysis in the second active site (HETPP → TATPP) acts to push the domains in the direction initiated by the motion of Val138. This concerted shuttle-like movement of the heterodimers provides an explanation for the flip-flop mechanism of catalysis. This also implies that this newly defined shuttle-like motion of E1p heterodimers ensures highly controlled depletion rates of both catalytic substrates. General Implications of E1p Structure for Flip-Flop Action— The structure of E1p is proposed as a model for action of other TPP-dependent enzymes because E1p shares structural components similar to other enzymes in this family. A search for related enzymes in the Protein Data Bank 2Protein Data Bank accession codes: 1DTW, 1QS0, 1L8A, 1GPU, 1ZPD, 1QPB, 1BFD, 1JSC, 1POW, 1KEK (www.rcsb.org). revealed that 2-oxoisovalerate dehydrogenase, transketolase, branched-chain α-ketoacid dehydrogenase E. coli pyruvate dehydrogenase, benzoylformate decarboxylase, acetohydroxyacid synthase, pyruvate oxidase, and pyruvate:ferredoxin oxidoreductase (1AEvarsson A. Seger K. Turley S. Sokatch J.R. Hol W.G.L. Nature Struct. Biol. 1999; 6: 785-792Crossref PubMed Scopus (119) Google Scholar, 2AEvarsson A. Chuang J.L. Wynn R.M. Turley S. Chuang D.T. Hol W.G.J. Structure. 2000; 8: 277-291Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar, 3Arjunan P. Nemeria N. Brunskill A. Chandrasekhar K. Sax M. Yan Y. Jordan F. Guest J.R. Furey W. Biochemistry. 2002; 41: 5213-5221Crossref PubMed Scopus (119) Google Scholar, 4Fiedler E. Thorell S. Sandalova T. Golbik R. Konig S. Schneider G. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 591-595Crossref PubMed Scopus (104) Google Scholar, 5Dyda F. Furey W. Swaminatythan S. Sax M. Farrenkopf B. Jordan F. Biochemistry. 1993; 32: 6165-6170Crossref PubMed Scopus (224) Google Scholar, 6Hasson M.S. Muscate A. McLeish M.J. Polovnikova L.S. Gerlt J.A. Kenyon G.L. Petsko G.A. Ringe D. Biochemistry. 1998; 37: 9918-9930Crossref PubMed Scopus (163) Google Scholar, 7Pang S.S. Duggleby R.G. Guddat L.W. J. Mol. Biol. 2002; 317: 249-262Crossref PubMed Scopus (186) Google Scholar, 8Muller Y.A. Schumacher G. Rudolph R. Schulz G.E. J. Mol. Biol. 1994; 237: 315-335Crossref PubMed Scopus (88) Google Scholar, 9Chabriere E. Vernede X. Guigliarelli B. Charon M.-H. Hatchikian E.C. Fontecilla-Camps J.C. Science. 2001; 294: 2559-2563Crossref PubMed Scopus (116) Google Scholar) also have pairs of the PP and PYR domains that provide for the two cofactor-binding sites, despite the fact that these domains can be distributed over two polypeptide chains as observed in the homologous α2β2 tetrameric branched-chain 2-oxoacid dehydrogenases or over a single chain, or often separated by domains that provide other functions. The association patterns between the pairs of PP and PYR domains in these enzymes resemble those of E1p. These include tight hydrophobic contacts between the domains facilitated by the GΦXXG motif and rather flexible contacts between the domains providing for the cofactor sites. Structure-derived sequence alignment of the fragment of E1p that contains Glu59 in the 59–72 helix of the PYR domains and Gly136 and Val138 in the 140–155 helix of the PP domains shown in Fig. 4d revealed that glutamate at position 59 is unique to all TPP-dependent enzymes and that Gly136 and Val138 are not conserved throughout this family of enzymes. Analysis of structures of these enzymes reveals, however, that the residues occupying positions at Gly136 and Val138 have similar functions in forming hydrogen bonds to the N4′ and N3′ atoms of cofactors via their main chain atoms and in protonation of the aminopyrimidine ring that drives the flip-flop action. As regards the role of the side chain of Val138, this structural comparison reveals that the side chain of the residue at the position of Val138 acts as a structural wedge located between the thiazolium and aminopyrimidine rings, thereby providing hydrophobic interactions to both ring systems. Therefore, we conclude that all TPP-dependent enzymes have common structural elements needed for shuttle-like motion of the domains that can support dynamic nonequivalence of otherwise chemically equivalent cofactor binding sites. Finally, E1p is a dynamic molecular machine scanning the entrance to the binding sites centered on the C2 atom for substrate-accessibility, i.e. pyruvate and lipoyl moiety of E2 (lipoamide), accepting them when required and processing them with the release of products. In conclusion, E1p provides an understanding of how this enzyme can use domain movements to select and process particular substrates during its catalytic cycle within a crowded environment such as the mitochondrial matrix. We thank Norma Duke at the Structural Biology Center for generous help with data collection on the beamline BM-19." @default.
• W2023454927 created "2016-06-24" @default.
• W2023454927 creator A5002784711 @default.
• W2023454927 creator A5022897040 @default.
• W2023454927 creator A5041946088 @default.
• W2023454927 creator A5062225146 @default.
• W2023454927 creator A5073619227 @default.
• W2023454927 date "2003-06-01" @default.
• W2023454927 modified "2023-10-14" @default.
• W2023454927 title "Structural Basis for Flip-Flop Action of Thiamin Pyrophosphate-dependent Enzymes Revealed by Human Pyruvate Dehydrogenase" @default.
• W2023454927 cites W1539796472 @default.
• W2023454927 cites W1580601386 @default.
• W2023454927 cites W1726597626 @default.
• W2023454927 cites W1964871086 @default.
• W2023454927 cites W1970828693 @default.
• W2023454927 cites W1970883608 @default.
• W2023454927 cites W1979058406 @default.
• W2023454927 cites W1986191025 @default.
• W2023454927 cites W1994373739 @default.
• W2023454927 cites W1995017064 @default.
• W2023454927 cites W1996726147 @default.
• W2023454927 cites W2006766324 @default.
• W2023454927 cites W2010864272 @default.
• W2023454927 cites W2018779619 @default.
• W2023454927 cites W2050836805 @default.
• W2023454927 cites W2064122330 @default.
• W2023454927 cites W2074493971 @default.
• W2023454927 cites W2076164532 @default.
• W2023454927 cites W2080065789 @default.
• W2023454927 cites W2082270688 @default.
• W2023454927 cites W2084711973 @default.
• W2023454927 cites W2094767163 @default.
• W2023454927 cites W2097493124 @default.
• W2023454927 cites W2132886727 @default.
• W2023454927 cites W2135839939 @default.
• W2023454927 cites W2140655362 @default.
• W2023454927 cites W2144285975 @default.
• W2023454927 cites W2157935454 @default.
• W2023454927 cites W2159081899 @default.
• W2023454927 doi "https://doi.org/10.1074/jbc.m300339200" @default.
• W2023454927 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/12651851" @default.
• W2023454927 hasPublicationYear "2003" @default.
• W2023454927 type Work @default.
• W2023454927 sameAs 2023454927 @default.
• W2023454927 citedByCount "141" @default.
• W2023454927 countsByYear W20234549272012 @default.
• W2023454927 countsByYear W20234549272013 @default.
• W2023454927 countsByYear W20234549272014 @default.
• W2023454927 countsByYear W20234549272015 @default.
• W2023454927 countsByYear W20234549272016 @default.
• W2023454927 countsByYear W20234549272017 @default.
• W2023454927 countsByYear W20234549272018 @default.
• W2023454927 countsByYear W20234549272019 @default.
• W2023454927 countsByYear W20234549272020 @default.
• W2023454927 countsByYear W20234549272021 @default.
• W2023454927 countsByYear W20234549272022 @default.
• W2023454927 countsByYear W20234549272023 @default.
• W2023454927 crossrefType "journal-article" @default.
• W2023454927 hasAuthorship W2023454927A5002784711 @default.
• W2023454927 hasAuthorship W2023454927A5022897040 @default.
• W2023454927 hasAuthorship W2023454927A5041946088 @default.
• W2023454927 hasAuthorship W2023454927A5062225146 @default.
• W2023454927 hasAuthorship W2023454927A5073619227 @default.
• W2023454927 hasBestOaLocation W20234549271 @default.
• W2023454927 hasConcept C101972281 @default.
• W2023454927 hasConcept C181199279 @default.
• W2023454927 hasConcept C185592680 @default.
• W2023454927 hasConcept C197957613 @default.
• W2023454927 hasConcept C2776043855 @default.
• W2023454927 hasConcept C2776317432 @default.
• W2023454927 hasConcept C2780047943 @default.
• W2023454927 hasConcept C55493867 @default.
• W2023454927 hasConcept C88099638 @default.
• W2023454927 hasConceptScore W2023454927C101972281 @default.
• W2023454927 hasConceptScore W2023454927C181199279 @default.
• W2023454927 hasConceptScore W2023454927C185592680 @default.
• W2023454927 hasConceptScore W2023454927C197957613 @default.
• W2023454927 hasConceptScore W2023454927C2776043855 @default.
• W2023454927 hasConceptScore W2023454927C2776317432 @default.
• W2023454927 hasConceptScore W2023454927C2780047943 @default.
• W2023454927 hasConceptScore W2023454927C55493867 @default.
• W2023454927 hasConceptScore W2023454927C88099638 @default.
• W2023454927 hasIssue "23" @default.
• W2023454927 hasLocation W20234549271 @default.
• W2023454927 hasLocation W20234549272 @default.
• W2023454927 hasOpenAccess W2023454927 @default.
• W2023454927 hasPrimaryLocation W20234549271 @default.
• W2023454927 hasRelatedWork W1974913328 @default.
• W2023454927 hasRelatedWork W2023454927 @default.
• W2023454927 hasRelatedWork W2025867216 @default.
• W2023454927 hasRelatedWork W2038758776 @default.
• W2023454927 hasRelatedWork W2099376584 @default.
• W2023454927 hasRelatedWork W2404858570 @default.
• W2023454927 hasRelatedWork W2466594997 @default.
• W2023454927 hasRelatedWork W2994021485 @default.
• W2023454927 hasRelatedWork W4247474351 @default.
• W2023454927 hasRelatedWork W777219313 @default.
• W2023454927 hasVolume "278" @default.

• next